Sync with latest Mesa 3.4 CVS sources

164 files changed, 13216 insertions, 20751 deletions

diff --git a/xc/config/cf/host.def b/xc/config/cf/host.def
index 681b737e4..f10005fc4 100644
--- a/xc/config/cf/host.def
+++ b/xc/config/cf/host.def
@@ -15,10 +15,21 @@
 
 #ifdef i386Architecture
 #define MesaUse3DNow
+#if HasKatmaiSupport
+#  define MesaUseKatmai
 #endif
+#endif
+
+/* To do profiling of the dynamically loaded 'xyz_dri.so' object, turn
+ * this on.
+ * Use 'xc/lib/GL/makeprofile.sh' to make it work.
+ */
+/* #define GlxSoProf YES */
 
-/* To do profiling turn on this plus one of the builtin drivers*/
-/* #define ProfileLibGlx YES */
+#ifdef GlxSoProf
+#  undef DefaultCCOptions
+#  define DefaultCCOptions -ansi GccWarningOptions -pipe -g -p
+#endif
 
 /* Optionally turn these on for debugging */
 /* #define GlxBuiltInTdfx YES */
@@ -33,7 +44,6 @@
 /* Optionally turn this on to force the kernel modules to build */
 /* #define BuildXF86DRM YES */
 
-#define SharedLibFont NO
 #define XnestServer NO
 #define XVirtualFramebufferServer NO
 #define XprtServer NO
diff --git a/xc/config/cf/linux.cf b/xc/config/cf/linux.cf
index 5f23b6c24..d71076c38 100644
--- a/xc/config/cf/linux.cf
+++ b/xc/config/cf/linux.cf
@@ -226,6 +226,21 @@ XCOMM binutils:	(LinuxBinUtilsMajorVersion)
 #  endif
 #endif
 
+/* Support for Intel's Katmai Native Instructions, also known as the
+ * Streaming SIMD Extensions, was introduced in the 2.4.x kernels.
+ */
+#ifndef HasKatmaiSupport
+#  ifdef i386Architecture
+#    if OSMajorVersion > 2 || (OSMajorVersion == 2 && OSMinorVersion >= 4)
+#      define HasKatmaiSupport		YES
+#    else
+#      define HasKatmaiSupport		NO
+#    endif
+#  else
+#    define HasKatmaiSupport		NO
+#  endif
+#endif
+
 /*
  * The Rush extension support.
  */
diff --git a/xc/extras/Mesa/include/GL/gl.h b/xc/extras/Mesa/include/GL/gl.h
index 5a31ddebd..11e67c6f3 100644
--- a/xc/extras/Mesa/include/GL/gl.h
+++ b/xc/extras/Mesa/include/GL/gl.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/include/GL/gl.h,v 1.7 2000/09/24 13:50:05 alanh Exp $ */
 
 
 #ifndef __gl_h_
@@ -44,11 +43,11 @@
 #define OPENSTEP
 #endif
 
-#if defined(_WIN32) && !defined(__WIN32__) && !defined(__CYGWIN__)
+#if defined(_WIN32) && !defined(__WIN32__)
 #define __WIN32__
 #endif
 
-#if !defined(OPENSTEP) && (defined(__WIN32__) && !defined(__CYGWIN__))
+#if !defined(OPENSTEP) && (defined(__WIN32__) || defined(__CYGWIN__))
 #  if defined(_MSC_VER) && defined(BUILD_GL32) /* tag specify we're building mesa as a DLL */
 #    define GLAPI __declspec(dllexport)
 #  elif defined(_MSC_VER) && defined(_DLL) /* tag specifying we're building for DLL runtime support */
@@ -63,7 +62,7 @@
 #  define GLAPIENTRY
 #endif /* WIN32 / CYGWIN bracket */
 
-#if defined(_WIN32) && !defined(_WINGDI_) && !defined(__CYGWIN__) && !defined(_GNU_H_WINDOWS32_DEFINES) && !defined(OPENSTEP)
+#if defined(_WIN32) && !defined(_WINGDI_) && !defined(_GNU_H_WINDOWS32_DEFINES) && !defined(OPENSTEP)
 #include <gl/mesa_wgl.h>
 #endif
 
diff --git a/xc/extras/Mesa/include/GL/glext.h b/xc/extras/Mesa/include/GL/glext.h
index 3df22a373..f721a9fd7 100644
--- a/xc/extras/Mesa/include/GL/glext.h
+++ b/xc/extras/Mesa/include/GL/glext.h
@@ -1,5 +1,3 @@
-/* $XFree86: xc/extras/Mesa/include/GL/glext.h,v 1.4 2000/09/26 15:56:28 tsi Exp $ */
-
 #ifndef __glext_h_
 #define __glext_h_
 
diff --git a/xc/extras/Mesa/include/GL/xmesa.h b/xc/extras/Mesa/include/GL/xmesa.h
index 3baec0eb0..d6ef934f3 100644
--- a/xc/extras/Mesa/include/GL/xmesa.h
+++ b/xc/extras/Mesa/include/GL/xmesa.h
@@ -88,7 +88,7 @@ extern struct Library *XLibBase;
 
 
 #define XMESA_MAJOR_VERSION 3
-#define XMESA_MINOR_VERSION 3
+#define XMESA_MINOR_VERSION 4
 
 
 
diff --git a/xc/extras/Mesa/include/GL/xmesa_xf86.h b/xc/extras/Mesa/include/GL/xmesa_xf86.h
index a142e9524..983f234dd 100644
--- a/xc/extras/Mesa/include/GL/xmesa_xf86.h
+++ b/xc/extras/Mesa/include/GL/xmesa_xf86.h
@@ -25,7 +25,6 @@ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
 SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 **************************************************************************/
-/* $XFree86: xc/extras/Mesa/include/GL/xmesa_xf86.h,v 1.4 2000/03/03 16:02:10 tsi Exp $ */
 
 /*
  * Authors:
diff --git a/xc/extras/Mesa/src/OSmesa/osmesa.c b/xc/extras/Mesa/src/OSmesa/osmesa.c
index e489f2553..4cc3e29d5 100644
--- a/xc/extras/Mesa/src/OSmesa/osmesa.c
+++ b/xc/extras/Mesa/src/OSmesa/osmesa.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/OSmesa/osmesa.c,v 1.3 2000/09/26 15:56:38 tsi Exp $ */
+
 
 /*
  * Off-Screen Mesa rendering / Rendering into client memory space
@@ -1356,7 +1356,7 @@ static line_func choose_line_function( GLcontext *ctx )
    OSMesaContext osmesa = (OSMesaContext) ctx;
 
    if (ctx->Line.SmoothFlag)              return NULL;
-   if (ctx->Texture.Enabled)              return NULL;
+   if (ctx->Texture.ReallyEnabled)        return NULL;
    if (ctx->Light.ShadeModel!=GL_FLAT)    return NULL;
 
    if (ctx->Line.Width==1.0F
@@ -1542,7 +1542,7 @@ static triangle_func choose_triangle_function( GLcontext *ctx )
 
    if (ctx->Polygon.SmoothFlag)     return NULL;
    if (ctx->Polygon.StippleFlag)    return NULL;
-   if (ctx->Texture.Enabled)        return NULL;
+   if (ctx->Texture.ReallyEnabled)  return NULL;
 
    if (ctx->RasterMask==DEPTH_BIT
        && ctx->Depth.Func==GL_LESS
diff --git a/xc/extras/Mesa/src/X/fakeglx.c b/xc/extras/Mesa/src/X/fakeglx.c
index f7dde2533..0b7cf6bf3 100644
--- a/xc/extras/Mesa/src/X/fakeglx.c
+++ b/xc/extras/Mesa/src/X/fakeglx.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X/fakeglx.c,v 1.7 2000/09/26 15:56:38 tsi Exp $ */
+
 
 /*
  * This is an emulation of the GLX API which allows Mesa/GLX-based programs
@@ -857,7 +857,7 @@ Fake_glXChooseVisual( Display *dpy, int screen, int *list )
    int *parselist;
    XVisualInfo *vis;
    int min_ci = 0;
-   int min_red=0, min_green=0, min_blue=0;
+   int min_red=0, min_green=0, min_blue=0, min_alpha=0;
    GLboolean rgb_flag = GL_FALSE;
    GLboolean alpha_flag = GL_FALSE;
    GLboolean double_flag = GL_FALSE;
@@ -921,10 +921,8 @@ Fake_glXChooseVisual( Display *dpy, int screen, int *list )
 	    break;
 	 case GLX_ALPHA_SIZE:
 	    parselist++;
-            {
-               GLint size = *parselist++;
-               alpha_flag = size>0 ? 1 : 0;
-            }
+            min_alpha = *parselist++;
+            alpha_flag = (min_alpha > 0);
 	    break;
 	 case GLX_DEPTH_SIZE:
 	    parselist++;
@@ -1003,6 +1001,16 @@ Fake_glXChooseVisual( Display *dpy, int screen, int *list )
       }
    }
 
+   /* DEBUG
+   printf("glXChooseVisual:\n");
+   printf("  GLX_RED_SIZE = %d\n", min_red);
+   printf("  GLX_GREEN_SIZE = %d\n", min_green);
+   printf("  GLX_BLUE_SIZE = %d\n", min_blue);
+   printf("  GLX_ALPHA_SIZE = %d\n", min_alpha);
+   printf("  GLX_DEPTH_SIZE = %d\n", depth_size);
+   printf("  GLX_STENCIL_SIZE = %d\n", stencil_size);
+   */
+
    /*
     * Since we're only simulating the GLX extension this function will never
     * find any real GL visuals.  Instead, all we can do is try to find an RGB
@@ -1049,15 +1057,32 @@ Fake_glXChooseVisual( Display *dpy, int screen, int *list )
    if (vis) {
       /* Note: we're not exactly obeying the glXChooseVisual rules here.
        * When GLX_DEPTH_SIZE = 1 is specified we're supposed to choose the
-       * largest depth buffer size, which is 32bits/value.  However, we
+       * largest depth buffer size, which is 32bits/value.  Instead, we
        * return 16 to maintain performance with earlier versions of Mesa.
        */
-      if (depth_size == 1)
-         depth_size = DEFAULT_SOFTWARE_DEPTH_BITS;
-      else if (depth_size > 24)
-         depth_size = 31;
+      if (depth_size > 24)
+         depth_size = 31;   /* 32 causes int overflow problems */
       else if (depth_size > 16)
          depth_size = 24;
+      else if (depth_size > 0)
+         depth_size = DEFAULT_SOFTWARE_DEPTH_BITS; /*16*/
+
+      /* If using Glide, make sure we don't try to setup an impossible
+       * visual.  This fixes the Q3 bug in which 24-bit Z was being reported.
+       */
+      {
+         const char *fx = getenv("MESA_GLX_FX");
+         if (fx && fx[0] != 'd')
+            if (depth_size > 16 ||
+                stencil_size > 0 ||
+                (min_red > 1 && min_red > 5) ||
+                (min_green > 1 && min_green > 6) ||
+                (min_blue > 1 && min_blue > 5) ||
+                alpha_flag)
+               return NULL;
+      }
+
+
       /* we only support one size of stencil and accum buffers. */
       if (stencil_size > 0)
          stencil_size = STENCIL_BITS;
@@ -1837,7 +1862,7 @@ struct _glxapi_table *_mesa_GetGLXDispatchTable(void)
    {
       int size = sizeof(struct _glxapi_table) / sizeof(void *);
       (void) size;
-      assert(_glxapi_get_dispatch_table_size() >= size);
+      assert((GLint) _glxapi_get_dispatch_table_size() >= size);
    }
 
    /* initialize the whole table to no-ops */
diff --git a/xc/extras/Mesa/src/X/fakeglx.h b/xc/extras/Mesa/src/X/fakeglx.h
deleted file mode 100644
index 62c4fc435..000000000
--- a/xc/extras/Mesa/src/X/fakeglx.h
+++ /dev/null
@@ -1,141 +0,0 @@
-
-/*
- * Mesa 3-D graphics library
- * Version:  3.1
- * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- * 
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-
-
-
-
-/*
- * GLX API functions which either call fake or real GLX implementations
- */
-
-
-
-#ifndef FAKEGLX_H
-#define FAKEGLX_H
-
-
-#include <X11/Xlib.h>
-#include <X11/Xutil.h>
-#include "GL/glx.h"
-
-
-extern XVisualInfo *Fake_glXChooseVisual( Display *dpy,
-                                          int screen, int *list );
-
-
-extern int Fake_glXGetConfig( Display *dpy, XVisualInfo *visinfo,
-                              int attrib, int *value );
-
-
-extern GLXContext Fake_glXCreateContext( Display *dpy, XVisualInfo *visinfo,
-                                         GLXContext shareList, Bool direct );
-
-
-extern void Fake_glXDestroyContext( Display *dpy, GLXContext ctx );
-
-
-extern void Fake_glXCopyContext( Display *dpy, GLXContext src, GLXContext dst,
-                                 GLuint mask );
-
-
-extern Bool Fake_glXMakeCurrent( Display *dpy, GLXDrawable drawable,
-                                 GLXContext ctx );
-
-
-extern GLXContext Fake_glXGetCurrentContext( void );
-
-
-extern GLXDrawable Fake_glXGetCurrentDrawable( void );
-
-
-extern GLXPixmap Fake_glXCreateGLXPixmap( Display *dpy, XVisualInfo *visinfo,
-                                          Pixmap pixmap );
-
-
-extern void Fake_glXDestroyGLXPixmap( Display *dpy, GLXPixmap pixmap );
-
-
-extern Bool Fake_glXQueryExtension( Display *dpy, int *errorb, int *event );
-
-
-extern Bool Fake_glXIsDirect( Display *dpy, GLXContext ctx );
-
-
-extern void Fake_glXSwapBuffers( Display *dpy, GLXDrawable drawable );
-
-
-extern void Fake_glXCopySubBufferMESA( Display *dpy, GLXDrawable drawable,
-                                       int x, int y, int width, int height );
-
-
-extern Bool Fake_glXQueryVersion( Display *dpy, int *maj, int *min );
-
-
-extern void Fake_glXUseXFont( Font font, int first, int count, int listBase );
-
-
-extern void Fake_glXWaitGL( void );
-
-
-extern void Fake_glXWaitX( void );
-
-
-/* GLX 1.1 and later */
-extern const char *Fake_glXQueryExtensionsString( Display *dpy, int screen );
-
-
-/* GLX 1.1 and later */
-extern const char *Fake_glXQueryServerString( Display *dpy, int screen,
-                                              int name );
-
-
-/* GLX 1.1 and later */
-extern const char *Fake_glXGetClientString( Display *dpy, int name );
-
-
-#ifdef GLX_MESA_release_buffers
-extern Bool Fake_glXReleaseBuffersMESA( Display *dpy, Window w );
-#endif
-
-
-#ifdef GLX_MESA_pixmap_colormap
-extern GLXPixmap Fake_glXCreateGLXPixmapMESA( Display *dpy,
-                                              XVisualInfo *visinfo,
-                                              Pixmap pixmap, Colormap cmap );
-#endif
-
-
-#ifdef GLX_MESA_set_3dfx_mode
-extern GLboolean Fake_glXSet3DfxModeMESA( GLint mode );
-#endif
-
-
-#ifdef GLX_EXT_get_proc_address
-/*extern GLfunction Fake_glXGetProcAddress( const GLubyte *procName );*/
-extern void (*Fake_glXGetProcAddress( const GLubyte *procName ))();
-#endif
-
-#endif
diff --git a/xc/extras/Mesa/src/X/realglx.c b/xc/extras/Mesa/src/X/realglx.c
deleted file mode 100644
index d443087f6..000000000
--- a/xc/extras/Mesa/src/X/realglx.c
+++ /dev/null
@@ -1,238 +0,0 @@
-
-/*
- * Mesa 3-D graphics library
- * Version:  3.1
- * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- * 
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-
-
-
-
-/*
- * Real GLX-encoder functions.  Called from glxapi.c
- *
- * Steven Parker's code for the GLX client API functions should be
- * put in this file.
- *
- * Also, the main API functions in api.c should somehow hook into the
- * GLX-encoding functions...
- */
-
-
-
-#include <X11/Xlib.h>
-#include <X11/Xutil.h>
-#include "realglx.h"
-
-
-
-XVisualInfo *Real_glXChooseVisual( Display *dpy, int screen, int *list )
-{
-   (void) dpy;
-   (void) screen;
-   (void) list;
-   return 0;
-}
-
-
-
-int Real_glXGetConfig( Display *dpy, XVisualInfo *visinfo,
-                       int attrib, int *value )
-{
-   (void) dpy;
-   (void) visinfo;
-   (void) attrib;
-   (void) value;
-   return 0;
-}
-
-
-
-GLXContext Real_glXCreateContext( Display *dpy, XVisualInfo *visinfo,
-                                  GLXContext shareList, Bool direct )
-{
-   (void) dpy;
-   (void) visinfo;
-   (void) shareList;
-   (void) direct;
-   return 0;
-}
-
-
-
-void Real_glXDestroyContext( Display *dpy, GLXContext ctx )
-{
-   (void) dpy;
-   (void) ctx;
-}
-
-
-
-void Real_glXCopyContext( Display *dpy, GLXContext src, GLXContext dst,
-                          GLuint mask )
-{
-   (void) dpy;
-   (void) src;
-   (void) dst;
-   (void) mask;
-}
-
-
-
-Bool Real_glXMakeCurrent( Display *dpy, GLXDrawable drawable, GLXContext ctx )
-{
-   (void) dpy;
-   (void) drawable;
-   (void) ctx;
-   return 0;
-}
-
-
-
-GLXContext Real_glXGetCurrentContext( void )
-{
-   return 0;
-}
-
-
-
-GLXDrawable Real_glXGetCurrentDrawable( void )
-{
-   return 0;
-}
-
-
-
-GLXPixmap Real_glXCreateGLXPixmap( Display *dpy, XVisualInfo *visinfo,
-				   Pixmap pixmap )
-{
-   (void) dpy;
-   (void) visinfo;
-   (void) pixmap;
-   return 0;
-}
-
-
-void Real_glXDestroyGLXPixmap( Display *dpy, GLXPixmap pixmap )
-{
-   (void) dpy;
-   (void) pixmap;
-}
-
-
-
-Bool Real_glXQueryExtension( Display *dpy, int *errorb, int *event )
-{
-   (void) dpy;
-   (void) errorb;
-   (void) event;
-   return 0;
-}
-
-
-
-Bool Real_glXIsDirect( Display *dpy, GLXContext ctx )
-{
-   (void) dpy;
-   (void) ctx;
-   return 0;
-}
-
-
-
-void Real_glXSwapBuffers( Display *dpy, GLXDrawable drawable )
-{
-   (void) dpy;
-   (void) drawable;
-}
-
-
-
-Bool Real_glXQueryVersion( Display *dpy, int *maj, int *min )
-{
-   (void) dpy;
-   (void) maj;
-   (void) min;
-   return 0;
-}
-
-
-
-void Real_glXUseXFont( Font font, int first, int count, int listBase )
-{
-   (void) font;
-   (void) first;
-   (void) count;
-   (void) listBase;
-}
-
-
-typedef struct {
-   struct {
-      int major_opcode;
-   } codes;
-
-
-
-} XExtDisplayInfo;
-
-
-void Real_glXWaitGL( void )
-{
-}
-
-
-
-void Real_glXWaitX( void )
-{
-}
-
-
-
-/* GLX 1.1 and later */
-const char *Real_glXQueryExtensionsString( Display *dpy, int screen )
-{
-   (void) dpy;
-   (void) screen;
-   return 0;
-}
-
-
-
-/* GLX 1.1 and later */
-const char *Real_glXQueryServerString( Display *dpy, int screen, int name )
-{
-   (void) dpy;
-   (void) screen;
-   (void) name;
-   return 0;
-}
-
-
-
-/* GLX 1.1 and later */
-const char *Real_glXGetClientString( Display *dpy, int name )
-{
-   (void) dpy;
-   (void) name;
-   return 0;
-}
diff --git a/xc/extras/Mesa/src/X/realglx.h b/xc/extras/Mesa/src/X/realglx.h
deleted file mode 100644
index d9ecc0f34..000000000
--- a/xc/extras/Mesa/src/X/realglx.h
+++ /dev/null
@@ -1,110 +0,0 @@
-
-/*
- * Mesa 3-D graphics library
- * Version:  3.1
- * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- * 
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-
-
-
-
-#ifndef REALGLX_H
-#define REALGLX_H
-
-
-#include <X11/Xlib.h>
-#include <X11/Xutil.h>
-#include "GL/glx.h"
-
-
-
-extern XVisualInfo *Real_glXChooseVisual( Display *dpy,
-                                          int screen, int *list );
-
-
-extern int Real_glXGetConfig( Display *dpy, XVisualInfo *visinfo,
-                              int attrib, int *value );
-
-
-extern GLXContext Real_glXCreateContext( Display *dpy, XVisualInfo *visinfo,
-                                         GLXContext shareList, Bool direct );
-
-
-extern void Real_glXDestroyContext( Display *dpy, GLXContext ctx );
-
-
-extern void Real_glXCopyContext( Display *dpy, GLXContext src, GLXContext dst,
-                                 GLuint mask );
-
-
-extern Bool Real_glXMakeCurrent( Display *dpy, GLXDrawable drawable,
-                                 GLXContext ctx );
-
-
-extern GLXContext Real_glXGetCurrentContext( void );
-
-
-extern GLXDrawable Real_glXGetCurrentDrawable( void );
-
-
-extern GLXPixmap Real_glXCreateGLXPixmap( Display *dpy, XVisualInfo *visinfo,
-                                          Pixmap pixmap );
-
-
-extern void Real_glXDestroyGLXPixmap( Display *dpy, GLXPixmap pixmap );
-
-
-extern Bool Real_glXQueryExtension( Display *dpy, int *errorb, int *event );
-
-
-extern Bool Real_glXIsDirect( Display *dpy, GLXContext ctx );
-
-
-extern void Real_glXSwapBuffers( Display *dpy, GLXDrawable drawable );
-
-
-extern Bool Real_glXQueryVersion( Display *dpy, int *maj, int *min );
-
-
-extern void Real_glXUseXFont( Font font, int first, int count, int listBase );
-
-
-extern void Real_glXWaitGL( void );
-
-
-extern void Real_glXWaitX( void );
-
-
-/* GLX 1.1 and later */
-extern const char *Real_glXQueryExtensionsString( Display *dpy, int screen );
-
-
-/* GLX 1.1 and later */
-extern const char *Real_glXQueryServerString( Display *dpy, int screen,
-                                              int name );
-
-
-/* GLX 1.1 and later */
-extern const char *Real_glXGetClientString( Display *dpy, int name );
-
-
-#endif
diff --git a/xc/extras/Mesa/src/X/xfonts.c b/xc/extras/Mesa/src/X/xfonts.c
index 2a1aca33c..9b1148db9 100644
--- a/xc/extras/Mesa/src/X/xfonts.c
+++ b/xc/extras/Mesa/src/X/xfonts.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  *
@@ -40,11 +40,9 @@
 #include <X11/Xutil.h>
 #include "GL/gl.h"
 #include "GL/glx.h"
-#include "GL/xmesa.h"
 #include "context.h"
 #include "mem.h"
 #include "xfonts.h"
-#include "xmesaP.h"
 
 
 /* Some debugging info.  */
@@ -222,7 +220,6 @@ static XCharStruct *isvalid(XFontStruct *fs, int which)
 
 void Fake_glXUseXFont( Font font, int first, int count, int listbase )
 {
-  XMesaContext CC;
   Display *dpy;
   Window win;
   Pixmap pixmap;
@@ -230,26 +227,23 @@ void Fake_glXUseXFont( Font font, int first, int count, int listbase )
   XGCValues values;
   unsigned long valuemask;
   XFontStruct *fs;
-
   GLint swapbytes, lsbfirst, rowlength;
   GLint skiprows, skippixels, alignment;
-
   unsigned int max_width, max_height, max_bm_width, max_bm_height;
   GLubyte *bm;
-
   int i;
 
-  CC = XMesaGetCurrentContext();
-  dpy = CC->display;
-  win = CC->xm_buffer->frontbuffer;
+  dpy = glXGetCurrentDisplay();
+  if (!dpy)
+     return;  /* I guess glXMakeCurrent wasn't called */
+  win = RootWindow(dpy, DefaultScreen(dpy));
 
   fs = XQueryFont (dpy, font);
-  if (!fs)
-    {
-      gl_error (CC->gl_ctx, GL_INVALID_VALUE,
-                "Couldn't get font structure information");
+  if (!fs) {
+      gl_error(NULL, GL_INVALID_VALUE,
+               "Couldn't get font structure information");
       return;
-    }
+  }
 
   /* Allocate a bitmap that can fit all characters.  */
   max_width = fs->max_bounds.rbearing - fs->min_bounds.lbearing;
@@ -261,10 +255,10 @@ void Fake_glXUseXFont( Font font, int first, int count, int listbase )
 (GLubyte));
   if (!bm) {
       XFreeFontInfo( NULL, fs, 0 );
-      gl_error (CC->gl_ctx, GL_OUT_OF_MEMORY,
+      gl_error(NULL, GL_OUT_OF_MEMORY,
                 "Couldn't allocate bitmap in glXUseXFont()");
       return;
-    }
+  }
 
 #if 0
   /* get the page info */
@@ -386,16 +380,3 @@ bm_height);
   glPixelStorei(GL_UNPACK_SKIP_PIXELS, skippixels);
   glPixelStorei(GL_UNPACK_ALIGNMENT, alignment);
 }
-
-
-extern void xmesa_xfonts_dummy( void );
-void xmesa_xfonts_dummy( void )
-{
-   /* silence unused var warnings */
-   (void) kernel8;
-   (void) DitherValues;
-   (void) HPCR_DRGB;
-   (void) kernel1;
-}
-
-/* The End. */
diff --git a/xc/extras/Mesa/src/X/xmesa1.c b/xc/extras/Mesa/src/X/xmesa1.c
index 8a84ca9a4..f7400da4b 100644
--- a/xc/extras/Mesa/src/X/xmesa1.c
+++ b/xc/extras/Mesa/src/X/xmesa1.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X/xmesa1.c,v 1.7 2000/09/24 13:50:53 alanh Exp $ */
 
 
 /*
@@ -90,6 +89,7 @@
 _glthread_Mutex _xmesa_lock;
 
 
+
 /*
  * Lookup tables for HPCR pixel format:
  */
@@ -322,7 +322,7 @@ static int bits_per_pixel( XMesaVisual xmv )
 
 /*
  * Determine if a given X window ID is valid (window exists).
- * Do this by calling XGetWindowAttributes() for the window and
+ * Do this by calling XGetIconName() for the window and
  * checking if we catch an X error.
  * Input:  dpy - the display
  *         win - the window to check for existance
@@ -343,12 +343,14 @@ static int window_exists_err_handler( XMesaDisplay* dpy, XErrorEvent* xerr )
 
 static GLboolean window_exists( XMesaDisplay *dpy, Window win )
 {
-   XWindowAttributes wa;
+   char *name;
    int (*old_handler)( XMesaDisplay*, XErrorEvent* );
    WindowExistsFlag = GL_TRUE;
    old_handler = XSetErrorHandler(window_exists_err_handler);
-   XGetWindowAttributes( dpy, win, &wa ); /* dummy request */
+   XGetIconName(dpy, win, &name); /* dummy request */
    XSetErrorHandler(old_handler);
+   if (name)
+      XFree(name);
    return WindowExistsFlag;
 }
 #endif
@@ -2509,7 +2511,7 @@ const char *XMesaGetString( XMesaContext c, int name )
 {
    (void) c;
    if (name==XMESA_VERSION) {
-      return "3.1";
+      return "3.4";
    }
    else if (name==XMESA_EXTENSIONS) {
       return "";
diff --git a/xc/extras/Mesa/src/X/xmesa2.c b/xc/extras/Mesa/src/X/xmesa2.c
index 48cea9497..3b4ca355e 100644
--- a/xc/extras/Mesa/src/X/xmesa2.c
+++ b/xc/extras/Mesa/src/X/xmesa2.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X/xmesa2.c,v 1.7 2000/09/26 15:56:38 tsi Exp $ */
+
 
 /*
  * Mesa/X11 interface, part 2.
@@ -5112,7 +5112,7 @@ static const GLubyte *get_string( GLcontext *ctx, GLenum name )
 #endif
       case GL_VENDOR:
 #ifdef XFree86Server
-         return (const GLubyte *) "Precision Insight, Inc.";
+         return (const GLubyte *) "VA Linux Systems, Inc.";
 #else
          return NULL;
 #endif
diff --git a/xc/extras/Mesa/src/X/xmesa3.c b/xc/extras/Mesa/src/X/xmesa3.c
index 3f6f507c6..ca8912718 100644
--- a/xc/extras/Mesa/src/X/xmesa3.c
+++ b/xc/extras/Mesa/src/X/xmesa3.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X/xmesa3.c,v 1.6 2000/09/26 15:56:39 tsi Exp $ */
+
 
 /*
  * Mesa/X11 interface, part 3.
@@ -61,7 +61,7 @@ static void draw_points_ANY_pixmap( GLcontext *ctx, GLuint first, GLuint last )
 
    if (xmesa->xm_visual->gl_visual->RGBAflag) {
       /* RGB mode */
-      for (i=first;i<=last;i++) {
+      for (i=first;i<last;i++) {
          if (VB->ClipMask[i]==0) {
             register int x, y;
 	    const GLubyte *color = VB->ColorPtr->data[i];
@@ -99,16 +99,16 @@ points_func xmesa_get_points_func( GLcontext *ctx )
    XMesaContext xmesa = (XMesaContext) ctx->DriverCtx;
 
    if (ctx->Point.Size==1.0F && !ctx->Point.SmoothFlag && ctx->RasterMask==0
-       && !ctx->Texture.Enabled) {
+       && !ctx->Texture.ReallyEnabled) {
       if (xmesa->xm_buffer->buffer==XIMAGE) {
-         return NULL; /*draw_points_ximage;*/
+         return (points_func)NULL; /*draw_points_ximage;*/
       }
       else {
          return draw_points_ANY_pixmap;
       }
    }
    else {
-      return NULL;
+      return (points_func)NULL;
    }
 }
 
@@ -670,11 +670,11 @@ line_func xmesa_get_line_func( GLcontext *ctx )
    (void) DitherValues;  /* silence unused var warning */
    (void) kernel1;  /* silence unused var warning */
 
-   if (ctx->Line.SmoothFlag)              return NULL;
-   if (ctx->Texture.Enabled)              return NULL;
-   if (ctx->Light.ShadeModel!=GL_FLAT)    return NULL;
+   if (ctx->Line.SmoothFlag)              return (line_func)NULL;
+   if (ctx->Texture.ReallyEnabled)        return (line_func)NULL;
+   if (ctx->Light.ShadeModel!=GL_FLAT)    return (line_func)NULL;
    /* X line stippling doesn't match OpenGL stippling */
-   if (ctx->Line.StippleFlag==GL_TRUE)    return NULL;
+   if (ctx->Line.StippleFlag==GL_TRUE)    return (line_func)NULL;
 
    if (xmesa->xm_buffer->buffer==XIMAGE
        && ctx->RasterMask==DEPTH_BIT
@@ -696,9 +696,9 @@ line_func xmesa_get_line_func( GLcontext *ctx )
          case PF_DITHER_5R6G5B:
             return flat_DITHER_5R6G5B_z_line;
          case PF_DITHER:
-            return (depth==8) ? flat_DITHER8_z_line : NULL;
+            return (depth==8) ? flat_DITHER8_z_line : (line_func)NULL;
          case PF_LOOKUP:
-            return (depth==8) ? flat_LOOKUP8_z_line : NULL;
+            return (depth==8) ? flat_LOOKUP8_z_line : (line_func)NULL;
          case PF_HPCR:
             return flat_HPCR_z_line;
          default:
@@ -722,9 +722,9 @@ line_func xmesa_get_line_func( GLcontext *ctx )
          case PF_DITHER_5R6G5B:
             return flat_DITHER_5R6G5B_line;
          case PF_DITHER:
-            return (depth==8) ? flat_DITHER8_line : NULL;
+            return (depth==8) ? flat_DITHER8_line : (line_func)NULL;
          case PF_LOOKUP:
-            return (depth==8) ? flat_LOOKUP8_line : NULL;
+            return (depth==8) ? flat_LOOKUP8_line : (line_func)NULL;
          case PF_HPCR:
             return flat_HPCR_line;
 	 default:
@@ -741,5 +741,5 @@ line_func xmesa_get_line_func( GLcontext *ctx )
       return flat_pixmap_line;
    }
 #endif
-   return NULL;
+   return (line_func)NULL;
 }
diff --git a/xc/extras/Mesa/src/X/xmesa4.c b/xc/extras/Mesa/src/X/xmesa4.c
index 2211eaffc..223c653e2 100644
--- a/xc/extras/Mesa/src/X/xmesa4.c
+++ b/xc/extras/Mesa/src/X/xmesa4.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X/xmesa4.c,v 1.6 2000/09/26 15:56:39 tsi Exp $ */
+
 
 /*
  * Mesa/X11 interface, part 4.
@@ -1539,8 +1539,8 @@ triangle_func xmesa_get_triangle_func( GLcontext *ctx )
 
    (void) kernel1;
 
-   if (ctx->Polygon.SmoothFlag)     return NULL;
-   if (ctx->Texture.Enabled)        return NULL;
+   if (ctx->Polygon.SmoothFlag)     return (triangle_func)NULL;
+   if (ctx->Texture.ReallyEnabled)  return (triangle_func)NULL;
 
    if (xmesa->xm_buffer->buffer==XIMAGE) {
       if (   ctx->Light.ShadeModel==GL_SMOOTH
@@ -1570,9 +1570,9 @@ triangle_func xmesa_get_triangle_func( GLcontext *ctx )
                return (depth==8) ? smooth_DITHER8_z_triangle
                                         : smooth_DITHER_z_triangle;
             case PF_LOOKUP:
-               return (depth==8) ? smooth_LOOKUP8_z_triangle : NULL;
+               return (depth==8) ? smooth_LOOKUP8_z_triangle : (triangle_func)NULL;
             default:
-               return NULL;
+               return (triangle_func)NULL;
          }
       }
       if (   ctx->Light.ShadeModel==GL_FLAT
@@ -1602,9 +1602,9 @@ triangle_func xmesa_get_triangle_func( GLcontext *ctx )
                return (depth==8) ? flat_DITHER8_z_triangle
                                         : flat_DITHER_z_triangle;
             case PF_LOOKUP:
-               return (depth==8) ? flat_LOOKUP8_z_triangle : NULL;
+               return (depth==8) ? flat_LOOKUP8_z_triangle : (triangle_func)NULL;
             default:
-               return NULL;
+               return (triangle_func)NULL;
          }
       }
       if (   ctx->RasterMask==0   /* no depth test */
@@ -1631,9 +1631,9 @@ triangle_func xmesa_get_triangle_func( GLcontext *ctx )
                return (depth==8) ? smooth_DITHER8_triangle
                                         : smooth_DITHER_triangle;
             case PF_LOOKUP:
-               return (depth==8) ? smooth_LOOKUP8_triangle : NULL;
+               return (depth==8) ? smooth_LOOKUP8_triangle : (triangle_func)NULL;
             default:
-               return NULL;
+               return (triangle_func)NULL;
          }
       }
 
@@ -1661,13 +1661,13 @@ triangle_func xmesa_get_triangle_func( GLcontext *ctx )
                return (depth==8) ? flat_DITHER8_triangle
                                         : flat_DITHER_triangle;
             case PF_LOOKUP:
-               return (depth==8) ? flat_LOOKUP8_triangle : NULL;
+               return (depth==8) ? flat_LOOKUP8_triangle : (triangle_func)NULL;
             default:
-               return NULL;
+               return (triangle_func)NULL;
          }
       }
 
-      return NULL;
+      return (triangle_func)NULL;
    }
    else {
       /* draw to pixmap */
@@ -1684,6 +1684,6 @@ triangle_func xmesa_get_triangle_func( GLcontext *ctx )
          return flat_pixmap_triangle;
       }
 #endif
-      return NULL;
+      return (triangle_func)NULL;
    }
 }
diff --git a/xc/extras/Mesa/src/X/xmesaP.h b/xc/extras/Mesa/src/X/xmesaP.h
index 8d8000874..0886fa3dc 100644
--- a/xc/extras/Mesa/src/X/xmesaP.h
+++ b/xc/extras/Mesa/src/X/xmesaP.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X/xmesaP.h,v 1.7 2000/09/26 15:56:39 tsi Exp $ */
+
 
 #ifndef XMESAP_H
 #define XMESAP_H
@@ -226,7 +226,7 @@ struct xmesa_buffer {
 
    /* Used to do XAllocColor/XFreeColors accounting: */
    int num_alloced;
-   Pixel alloced_colors[256];
+   unsigned long alloced_colors[256];
 
 #if defined(GLX_DIRECT_RENDERING) && !defined(XFree86Server)
   __DRIdrawablePrivate *driDrawPriv;	/* back pointer to DRI drawable
diff --git a/xc/extras/Mesa/src/X86/3dnow.c b/xc/extras/Mesa/src/X86/3dnow.c
index ef92d27c0..7b642dc04 100644
--- a/xc/extras/Mesa/src/X86/3dnow.c
+++ b/xc/extras/Mesa/src/X86/3dnow.c
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -22,176 +22,163 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X86/3dnow.c,v 1.6 2000/09/26 15:56:39 tsi Exp $ */
 
 /*
  * 3DNow! optimizations contributed by
  * Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
  */
-#if defined(USE_3DNOW_ASM) && defined(USE_X86_ASM)
-#include "3dnow.h"
-
-#include <limits.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <math.h>
 
+#include "glheader.h"
 #include "context.h"
 #include "types.h"
-#include "xform.h"
 #include "vertices.h"
+#include "xform.h"
+#include "3dnow.h"
 
 #ifdef DEBUG
 #include "debug_xform.h"
 #endif
 
 
+#define XFORM_ARGS	GLvector4f *to_vec,				\
+			const GLfloat m[16],				\
+			const GLvector4f *from_vec,			\
+			const GLubyte *mask,				\
+			const GLubyte flag
+
+
+#define DECLARE_XFORM_GROUP( pfx, sz, masked ) \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_general_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_identity_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_3d_no_rot_##masked( XFORM_ARGS );    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_perspective_##masked( XFORM_ARGS );  \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_2d_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_2d_no_rot_##masked( XFORM_ARGS );    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_3d_##masked( XFORM_ARGS );
+
+
+#define ASSIGN_XFORM_GROUP( pfx, cma, sz, masked )			\
+   gl_transform_tab[cma][sz][MATRIX_GENERAL] =				\
+      gl_##pfx##_transform_points##sz##_general_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_IDENTITY] =				\
+      gl_##pfx##_transform_points##sz##_identity_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_3D_NO_ROT] =			\
+      gl_##pfx##_transform_points##sz##_3d_no_rot_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_PERSPECTIVE] =			\
+      gl_##pfx##_transform_points##sz##_perspective_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_2D] =				\
+      gl_##pfx##_transform_points##sz##_2d_##masked;			\
+   gl_transform_tab[cma][sz][MATRIX_2D_NO_ROT] =			\
+      gl_##pfx##_transform_points##sz##_2d_no_rot_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_3D] =				\
+      gl_##pfx##_transform_points##sz##_3d_##masked;
+
+
+
+#define NORM_ARGS	const GLmatrix *mat,				\
+			GLfloat scale,					\
+			const GLvector3f *in,				\
+			const GLfloat *lengths,				\
+			const GLubyte mask[],				\
+			GLvector3f *dest
+
+
+#define DECLARE_NORM_GROUP( pfx, masked ) \
+ extern void _ASMAPI gl_##pfx##_rescale_normals_##masked( NORM_ARGS );			    \
+ extern void _ASMAPI gl_##pfx##_normalize_normals_##masked( NORM_ARGS );		    \
+ extern void _ASMAPI gl_##pfx##_transform_normals_##masked( NORM_ARGS );		    \
+ extern void _ASMAPI gl_##pfx##_transform_normals_no_rot_##masked( NORM_ARGS );		    \
+ extern void _ASMAPI gl_##pfx##_transform_rescale_normals_##masked( NORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_rescale_normals_no_rot_##masked( NORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_normalize_normals_##masked( NORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_normalize_normals_no_rot_##masked( NORM_ARGS );
+
+
+#define ASSIGN_NORM_GROUP( pfx, cma, masked )				\
+   gl_normal_tab[NORM_RESCALE][cma] =					\
+      gl_##pfx##_rescale_normals_##masked;				\
+   gl_normal_tab[NORM_NORMALIZE][cma] =					\
+      gl_##pfx##_normalize_normals_##masked;				\
+   gl_normal_tab[NORM_TRANSFORM][cma] =					\
+      gl_##pfx##_transform_normals_##masked;				\
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT][cma] =				\
+      gl_##pfx##_transform_normals_no_rot_##masked;			\
+   gl_normal_tab[NORM_TRANSFORM | NORM_RESCALE][cma] =			\
+      gl_##pfx##_transform_rescale_normals_##masked;			\
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE][cma] =		\
+      gl_##pfx##_transform_rescale_normals_no_rot_##masked;		\
+   gl_normal_tab[NORM_TRANSFORM | NORM_NORMALIZE][cma] =		\
+      gl_##pfx##_transform_normalize_normals_##masked;			\
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE][cma] =		\
+      gl_##pfx##_transform_normalize_normals_no_rot_##masked;
+
+
+#ifdef USE_3DNOW_ASM
+DECLARE_XFORM_GROUP( 3dnow, 1, raw )
+DECLARE_XFORM_GROUP( 3dnow, 2, raw )
+DECLARE_XFORM_GROUP( 3dnow, 3, raw )
+DECLARE_XFORM_GROUP( 3dnow, 4, raw )
+
+DECLARE_XFORM_GROUP( 3dnow, 1, masked )
+DECLARE_XFORM_GROUP( 3dnow, 2, masked )
+DECLARE_XFORM_GROUP( 3dnow, 3, masked )
+DECLARE_XFORM_GROUP( 3dnow, 4, masked )
+
+DECLARE_NORM_GROUP( 3dnow, raw )
+/*DECLARE_NORM_GROUP( 3dnow, masked )*/
 
 
-#define XFORM_ARGS      GLvector4f *to_vec,             \
-                        const GLfloat m[16],            \
-                        const GLvector4f *from_vec,     \
-                        const GLubyte *mask,            \
-                        const GLubyte flag
-
-
-
-#define DECLARE_XFORM_GROUP(pfx, v, masked) \
- extern void _ASMAPI gl##pfx##_transform_points##v##_general_##masked(XFORM_ARGS);    \
- extern void _ASMAPI gl##pfx##_transform_points##v##_identity_##masked(XFORM_ARGS);   \
- extern void _ASMAPI gl##pfx##_transform_points##v##_3d_no_rot_##masked(XFORM_ARGS);  \
- extern void _ASMAPI gl##pfx##_transform_points##v##_perspective_##masked(XFORM_ARGS);\
- extern void _ASMAPI gl##pfx##_transform_points##v##_2d_##masked(XFORM_ARGS);         \
- extern void _ASMAPI gl##pfx##_transform_points##v##_2d_no_rot_##masked(XFORM_ARGS);  \
- extern void _ASMAPI gl##pfx##_transform_points##v##_3d_##masked(XFORM_ARGS);
-
-
-
-#define ASSIGN_XFORM_GROUP( pfx, cma, vsize, masked )           \
- gl_transform_tab[cma][vsize][MATRIX_GENERAL]                   \
-  = gl##pfx##_transform_points##vsize##_general_##masked;      \
- gl_transform_tab[cma][vsize][MATRIX_IDENTITY]                  \
-  = gl##pfx##_transform_points##vsize##_identity_##masked;     \
- gl_transform_tab[cma][vsize][MATRIX_3D_NO_ROT]                 \
-  = gl##pfx##_transform_points##vsize##_3d_no_rot_##masked;    \
- gl_transform_tab[cma][vsize][MATRIX_PERSPECTIVE]               \
-  = gl##pfx##_transform_points##vsize##_perspective_##masked;  \
- gl_transform_tab[cma][vsize][MATRIX_2D]                        \
-  = gl##pfx##_transform_points##vsize##_2d_##masked;           \
- gl_transform_tab[cma][vsize][MATRIX_2D_NO_ROT]                 \
-  = gl##pfx##_transform_points##vsize##_2d_no_rot_##masked;    \
- gl_transform_tab[cma][vsize][MATRIX_3D]                        \
-  = gl##pfx##_transform_points##vsize##_3d_##masked;
-
-
-
-
-#define NORM_ARGS       const GLmatrix *mat,        \
-                        GLfloat scale,              \
-                        const GLvector3f *in,       \
-                        const GLfloat *lengths,     \
-                        const GLubyte mask[],       \
-                        GLvector3f *dest 
-
-
-
-#define DECLARE_NORM_GROUP(pfx, masked)                                        \
- extern void _ASMAPI gl##pfx##_rescale_normals_##masked## (NORM_ARGS);                \
- extern void _ASMAPI gl##pfx##_normalize_normals_##masked## (NORM_ARGS);              \
- extern void _ASMAPI gl##pfx##_transform_normals_##masked## (NORM_ARGS);              \
- extern void _ASMAPI gl##pfx##_transform_normals_no_rot_##masked## (NORM_ARGS);       \
- extern void _ASMAPI gl##pfx##_transform_rescale_normals_##masked## (NORM_ARGS);      \
- extern void _ASMAPI gl##pfx##_transform_rescale_normals_no_rot_##masked## (NORM_ARGS); \
- extern void _ASMAPI gl##pfx##_transform_normalize_normals_##masked## (NORM_ARGS);    \
- extern void _ASMAPI gl##pfx##_transform_normalize_normals_no_rot_##masked## (NORM_ARGS);
-
-
-
-#define ASSIGN_NORM_GROUP( pfx, cma, masked )                                 \
-   gl_normal_tab[NORM_RESCALE][cma]   =                                       \
-      gl##pfx##_rescale_normals_##masked##;                                  \
-   gl_normal_tab[NORM_NORMALIZE][cma] =                                       \
-      gl##pfx##_normalize_normals_##masked##;                                \
-   gl_normal_tab[NORM_TRANSFORM][cma] =                                       \
-      gl##pfx##_transform_normals_##masked##;                                \
-   gl_normal_tab[NORM_TRANSFORM_NO_ROT][cma] =                                \
-      gl##pfx##_transform_normals_no_rot_##masked##;                         \
-   gl_normal_tab[NORM_TRANSFORM | NORM_RESCALE][cma] =                        \
-      gl##pfx##_transform_rescale_normals_##masked##;                        \
-   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE][cma] =                 \
-      gl##pfx##_transform_rescale_normals_no_rot_##masked##;                 \
-   gl_normal_tab[NORM_TRANSFORM | NORM_NORMALIZE][cma] =                      \
-      gl##pfx##_transform_normalize_normals_##masked##;                      \
-   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE][cma] =               \
-	  gl##pfx##_transform_normalize_normals_no_rot_##masked##;
-
+extern void _ASMAPI gl_v16_3dnow_general_xform( GLfloat *first_vert,
+						const GLfloat *m,
+						const GLfloat *src,
+						GLuint src_stride,
+						GLuint count );
 
 extern void _ASMAPI gl_3dnow_project_vertices( GLfloat *first,
-				       GLfloat *last,
-				       const GLfloat *m,
-				       GLuint stride );
-
-extern void _ASMAPI gl_3dnow_project_clipped_vertices( GLfloat *first,
 					       GLfloat *last,
 					       const GLfloat *m,
-					       GLuint stride,
-					       const GLubyte *clipmask );
+					       GLuint stride );
 
-extern void _ASMAPI gl_v16_3dnow_general_xform( GLfloat *first_vert,
-					const GLfloat *m,
-					const GLfloat *src,
-					GLuint src_stride,
-					GLuint count );
+extern void _ASMAPI gl_3dnow_project_clipped_vertices( GLfloat *first,
+						       GLfloat *last,
+						       const GLfloat *m,
+						       GLuint stride,
+						       const GLubyte *clipmask );
+#endif
 
-void gl_init_3dnow_asm_transforms (void)
-{
-   DECLARE_XFORM_GROUP( _3dnow, 1, raw )
-   DECLARE_XFORM_GROUP( _3dnow, 2, raw )
-   DECLARE_XFORM_GROUP( _3dnow, 3, raw )
-   DECLARE_XFORM_GROUP( _3dnow, 4, raw )
-
-   DECLARE_XFORM_GROUP( _3dnow, 1, masked )
-   DECLARE_XFORM_GROUP( _3dnow, 2, masked )
-   DECLARE_XFORM_GROUP( _3dnow, 3, masked )
-   DECLARE_XFORM_GROUP( _3dnow, 4, masked )
-
-   DECLARE_NORM_GROUP( _3dnow, raw )
-/* DECLARE_NORM_GROUP( _3dnow, masked )
-*/
-
-   ASSIGN_XFORM_GROUP( _3dnow, 0, 1, raw )
-   ASSIGN_XFORM_GROUP( _3dnow, 0, 2, raw )
-   ASSIGN_XFORM_GROUP( _3dnow, 0, 3, raw )
-   ASSIGN_XFORM_GROUP( _3dnow, 0, 4, raw )
-
-   ASSIGN_XFORM_GROUP( _3dnow, CULL_MASK_ACTIVE, 1, masked )
-   ASSIGN_XFORM_GROUP( _3dnow, CULL_MASK_ACTIVE, 2, masked )
-   ASSIGN_XFORM_GROUP( _3dnow, CULL_MASK_ACTIVE, 3, masked )
-   ASSIGN_XFORM_GROUP( _3dnow, CULL_MASK_ACTIVE, 4, masked )
-
-   ASSIGN_NORM_GROUP( _3dnow, 0, raw )
-/* ASSIGN_NORM_GROUP( _3dnow, CULL_MASK_ACTIVE, masked )
-*/
 
-#ifdef DEBUG
-   gl_test_all_transform_functions("3Dnow!");
-   gl_test_all_normal_transform_functions("3Dnow!");
-#endif
+void gl_init_3dnow_transform_asm( void )
+{
+#ifdef USE_3DNOW_ASM
+   ASSIGN_XFORM_GROUP( 3dnow, 0, 1, raw );
+   ASSIGN_XFORM_GROUP( 3dnow, 0, 2, raw );
+   ASSIGN_XFORM_GROUP( 3dnow, 0, 3, raw );
+   ASSIGN_XFORM_GROUP( 3dnow, 0, 4, raw );
 
-   /* Hook in some stuff for vertices.c.
-    */
-   gl_xform_points3_v16_general = gl_v16_3dnow_general_xform;
-   gl_project_v16 = gl_3dnow_project_vertices;
-   gl_project_clipped_v16 = gl_3dnow_project_clipped_vertices;
-} 
+   ASSIGN_XFORM_GROUP( 3dnow, CULL_MASK_ACTIVE, 1, masked );
+   ASSIGN_XFORM_GROUP( 3dnow, CULL_MASK_ACTIVE, 2, masked );
+   ASSIGN_XFORM_GROUP( 3dnow, CULL_MASK_ACTIVE, 3, masked );
+   ASSIGN_XFORM_GROUP( 3dnow, CULL_MASK_ACTIVE, 4, masked );
 
-#else
+   ASSIGN_NORM_GROUP( 3dnow, 0, raw );
+/* ASSIGN_NORM_GROUP( 3dnow, CULL_MASK_ACTIVE, masked ); */
 
+#ifdef DEBUG
+   gl_test_all_transform_functions( "3DNow!" );
+   gl_test_all_normal_transform_functions( "3DNow!" );
+#endif
+#endif
+}
 
-/* silence compiler warning */
-extern void _mesa_3dnow_dummy_function(void);
-void _mesa_3dnow_dummy_function(void)
+void gl_init_3dnow_vertex_asm( void )
 {
-}
+#ifdef USE_3DNOW_ASM
+   gl_xform_points3_v16_general	= gl_v16_3dnow_general_xform;
+   gl_project_v16		= gl_3dnow_project_vertices;
+   gl_project_clipped_v16	= gl_3dnow_project_clipped_vertices;
 
+#if 0
+   gl_test_all_vertex_functions( "3DNow!" );
 #endif
+#endif
+}
diff --git a/xc/extras/Mesa/src/X86/3dnow.h b/xc/extras/Mesa/src/X86/3dnow.h
index c7d3040d5..6e34b6ce4 100644
--- a/xc/extras/Mesa/src/X86/3dnow.h
+++ b/xc/extras/Mesa/src/X86/3dnow.h
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -23,72 +23,17 @@
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-
 /*
  * 3DNow! optimizations contributed by
  * Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
  */
 
-
-#ifndef _3dnow_h
-#define _3dnow_h
-
-
+#ifndef __3DNOW_H__
+#define __3DNOW_H__
 
 #include "xform.h"
 
-
-void gl_init_3dnow_asm_transforms (void);
-
-
-
-
-#if 0
-GLvector4f *gl_project_points( GLvector4f *proj_vec,
-                               const GLvector4f *clip_vec )
-{
-   __asm__ (
-   "   femms                                                              \n"
-   "                                                                      \n"
-   "   movq       (%0),     %%mm0      # x1             | x0              \n"
-   "   movq       8(%0),    %%mm1      # oow            | x2              \n"
-   "                                                                      \n"
-   "1: movq       %%mm1,    %%mm2      # oow            | x2              \n"
-   "   addl       %2,       %0         # next point                       \n"
-   "                                                                      \n"
-   "   punpckhdq  %%mm2,    %%mm2      # oow            | oow             \n"
-   "   addl       $16,      %1         # next point                       \n"
-   "                                                                      \n"
-   "   pfrcp      %%mm2,    %%mm3      # 1/oow          | 1/oow           \n"
-   "   decl       %3                                                      \n"
-   "                                                                      \n"
-   "   pfmul      %%mm3,    %%mm0      # x1/oow         | x0/oow          \n"
-   "   movq       %%mm0,    -16(%1)    # write r0, r1                     \n"
-   "                                                                      \n"
-   "   pfmul      %%mm3,    %%mm1      # 1              | x2/oow          \n"
-   "   movq       (%0),     %%mm0      # x1             | x0              \n"
-   "                                                                      \n"
-   "   movd       %%mm1,    8(%1)      # write r2                         \n"
-   "   movd       %%mm3,    12(%1)     # write r3                         \n"
-   "                                                                      \n"
-   "   movq       8(%0),    %%mm1      # oow            | x2              \n"
-   "   ja         1b                                                      \n"
-   "                                                                      \n"
-   "   femms                                                              \n"
-   "                                                                        "
-   ::"a" (clip_vec->start),
-   "c" (proj_vec->start),
-   "g" (clip_vec->stride),
-   "d" (clip_vec->count)
-   );
-
-   proj_vec->flags |= VEC_SIZE_4;
-   proj_vec->size = 3;
-   proj_vec->count = clip_vec->count;
-   return proj_vec;
-}
-#endif
-
-
+void gl_init_3dnow_transform_asm( void );
+void gl_init_3dnow_vertex_asm( void );
 
 #endif
diff --git a/xc/extras/Mesa/src/X86/3dnow_norm_raw.S b/xc/extras/Mesa/src/X86/3dnow_norm_raw.S
index e074c3337..d844b3c87 100644
--- a/xc/extras/Mesa/src/X86/3dnow_norm_raw.S
+++ b/xc/extras/Mesa/src/X86/3dnow_norm_raw.S
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  *
@@ -21,7 +21,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X86/3dnow_norm_raw.S,v 1.4 2000/09/26 15:56:39 tsi Exp $ */
 
 /*
  * 3Dnow assembly code by Holger Waechtler
@@ -254,7 +253,7 @@ ALIGNTEXT16
 GLOBL GLNAME(gl_3dnow_transform_normalize_normals_no_rot_raw)
 GLNAME(gl_3dnow_transform_normalize_normals_no_rot_raw):
 
- #undef FRAME_OFFSET 
+ #undef FRAME_OFFSET
  #define FRAME_OFFSET 12
 
     PUSH_L     ( EDI )
@@ -396,7 +395,7 @@ ALIGNTEXT16
 GLOBL GLNAME(gl_3dnow_transform_rescale_normals_no_rot_raw)
 GLNAME(gl_3dnow_transform_rescale_normals_no_rot_raw):
 
- #undef FRAME_OFFSET 
+ #undef FRAME_OFFSET
  #define FRAME_OFFSET 12
 
     PUSH_L     ( EDI )
@@ -887,6 +886,3 @@ LLBL (G3R_end):
     POP_L      ( ESI )
     POP_L      ( EDI )
     RET
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_vertex.S b/xc/extras/Mesa/src/X86/3dnow_vertex.S
new file mode 100644
index 000000000..a66432c3a
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/3dnow_vertex.S
@@ -0,0 +1,220 @@
+#include "assyntax.h"
+
+   SEG_TEXT
+
+#define MAT_SY 20
+#define MAT_SZ 40
+#define MAT_TX 48
+#define MAT_TY 52
+#define MAT_TZ 56
+
+
+/*
+ * void gl_v16_3dnow_general_xform( GLfloat *dest,
+ *		        	    const GLfloat *m,
+ *				    const GLfloat *src,
+ *                                  GLuint src_stride,
+ *				    GLuint count )
+ *
+ * These tranformation functions could disappear if the standard ones
+ * took an output stride.
+ */
+GLOBL GLNAME( gl_v16_3dnow_general_xform )
+GLNAME( gl_v16_3dnow_general_xform ):
+
+    PUSH_L    ( EDI )
+    PUSH_L    ( ESI )
+
+    MOV_L     ( REGOFF(12, ESP), EAX )		/* dest                      */
+    MOV_L     ( REGOFF(16, ESP), ESI )		/* mat                       */
+    MOV_L     ( REGOFF(20, ESP), EDX )		/* src                       */
+    MOV_L     ( REGOFF(24, ESP), EDI )		/* src_stride                */
+    MOV_L     ( REGOFF(28, ESP), ECX )		/* count                     */
+
+    FEMMS
+
+    MOVQ      ( REGOFF(MAT_TX, ESI), MM7 )	/* ty          | tx          */
+    MOVQ      ( REGOFF(MAT_TZ, ESI), MM3 )	/* tw          | tz          */
+
+ALIGNTEXT32
+LLBL( v16_3dnow_general_loop ):
+
+    PREFETCHW ( REGOFF(128, EAX) )		/* write alloc 2 verts ahead */
+    PREFETCH  ( REGOFF(32, EDX) )		/* prefetch next cache line  */
+
+    MOVQ      ( REGIND(EDX),  MM0 )		/* x1          | x0          */
+    MOVD      ( REGOFF(8, EDX), MM1 )		/*             | x2          */
+    MOVQ      ( REGIND(ESI),  MM4 )		/* m1          | m0          */
+    PUNPCKHDQ ( MM0, MM2 )			/* x1          |             */
+    MOVQ      ( REGOFF(16, ESI), MM5 )		/* m5          | m4          */
+    PUNPCKLDQ ( MM0, MM0 )			/* x0          | x0          */
+    MOVQ      ( REGOFF(32, ESI), MM6 )		/* m9          | m8          */
+    PFMUL     ( MM0, MM4 )			/* x0*m1       | x0*m0       */
+    PUNPCKHDQ ( MM2, MM2 )			/* x1          | x1          */
+    PFMUL     ( MM2, MM5 )			/* x1*m5       | x1*m4       */
+    PUNPCKLDQ ( MM1, MM1 )			/* x2          | x2          */
+    PFMUL     ( REGOFF(8, ESI), MM0 )		/* x0*m3       | x0*m2       */
+    PFMUL     ( REGOFF(24, ESI), MM2 )		/* x1*m7       | x1*m6       */
+    PFMUL     ( MM1, MM6 )			/* x2*m9       | x2*m8       */
+    PFADD     ( MM4, MM5 )			/* x0*m1+x1*m5 | x0*m0+x1*m4 */
+    PFMUL     ( REGOFF(40, ESI), MM1 )		/* x2*m11      | x2*m10      */
+    PFADD     ( MM0, MM2 )			/* x0*m3+x1*m7 | x0*m2+x1*m6 */
+    PFADD     ( MM5, MM6 )
+    PFADD     ( MM1, MM2 )
+    PFADD     ( MM7, MM6 )			/* r1          | r0          */
+    PFADD     ( MM3, MM2 )			/* r3          | r2          */
+    ADD_L     ( EDI, EDX )			/* next input vertex         */
+    MOVQ      ( MM6, REGIND(EAX) )
+    MOVQ      ( MM2, REGOFF(8, EAX) )
+    ADD_L     ( CONST(64), EAX )		/* next output vertex        */
+    DEC_L     ( ECX )
+    JNE       ( LLBL(v16_3dnow_general_loop) )
+
+    FEMMS
+
+    POP_L     ( ESI )
+    POP_L     ( EDI )
+    RET
+
+
+
+/* Do viewport map and perspective projection.  Args should look like:
+ *
+ * gl_3dnow_project_vertices( float *first_vertex,
+ *                            const float *last_vertex,
+ *                            float *matrix,
+ *                            GLuint stride )
+ *
+ * This routine assumes a sane vertex layout with x,y,z,w as
+ * the first four elements, to be projected in clip-space, to
+ * x/w,y/w,z/w,1/w, and then transformed according to the matrix to
+ * device space.  The device coordinates will overwrite the clip
+ * coordinates as the first four elements of the vertex.
+ *
+ * If projection is required for other elements, such as texcoords,
+ * you will have to code a specialized version of this routine.  See
+ * FX/X86 for examples.
+ *
+ * These routines are simplified versions of the FX code written by
+ * Holger.
+ */
+
+GLOBL GLNAME( gl_3dnow_project_vertices )
+GLNAME( gl_3dnow_project_vertices ):
+
+    PUSH_L    ( EBP )
+    FEMMS
+    PREFETCH  ( REGOFF(8, ESP) )		/* fetch the first vertex    */
+
+    MOV_L     ( REGOFF(8, ESP), ECX )		/* first_vert                */
+    MOV_L     ( REGOFF(12, ESP), EDX )		/* last_vert                 */
+    MOV_L     ( REGOFF(16, ESP), EBP )		/* matrix                    */
+    MOV_L     ( REGOFF(20, ESP), EAX )		/* stride                    */
+
+    MOVD      ( REGOFF(MAT_TX, EBP), MM6 )	/*             | tx          */
+    PUNPCKLDQ ( REGOFF(MAT_TY, EBP), MM6 )	/* ty          | tx          */
+    MOVD      ( REGIND(EBP), MM5 )
+    PUNPCKLDQ ( REGOFF(MAT_SY, EBP), MM5 )	/* vsy         | vsx         */
+    MOVD      ( REGOFF(MAT_SZ, EBP), MM1 )	/*             | vsz         */
+    SUB_L     ( ECX, EDX )			/* last -= first             */
+
+ALIGNTEXT32
+LLBL( v16_3dnow_pv_loop_start ):
+
+    PREFETCH  ( REGOFF(64, ECX) )		/* fetch one/two verts ahead */
+    MOVD      ( REGOFF(12, ECX), MM0 )		/*             | f[3]        */
+    PFRCP     ( MM0, MM0 )			/* oow = 1/f[3]              */
+    MOVD      ( REGOFF(12, ECX), MM7 )		/*             | f[3]        */
+    PFRCPIT1  ( MM0, MM7 )
+    PFRCPIT2  ( MM0, MM7 )			/* oow         | oow         */
+    PUNPCKLDQ ( MM7, MM7 )
+    MOVQ      ( REGIND(ECX), MM2 )		/* f[1]        | f[0]        */
+    PFMUL     ( MM7, MM2 )			/* f[1] * oow  | f[0] * oow  */
+    MOVD      ( REGOFF(8, ECX), MM3 )		/*             | f[2]        */
+    PFMUL     ( MM7, MM3 )			/*             | f[2] * oow  */
+    MOVD      ( REGOFF(MAT_TZ, EBP), MM0 )	/*             | vtz         */
+    PFMUL     ( MM1, MM3 )			/*             | f[2] *= vsz */
+    PFADD     ( MM0, MM3 )			/*             | f[2] += vtz */
+    PFMUL     ( MM5, MM2 )			/* f[1] *= vsy | f[0] *= vsx */
+    PFADD     ( MM6, MM2 )			/* f[1] += vty | f[0] += vtx */
+    PUNPCKLDQ ( MM7, MM3 )			/* f[3] = oow  | f[2]        */
+    MOVQ      ( MM2, REGOFF(0, ECX) )
+    MOVQ      ( MM3, REGOFF(8, ECX) )
+    ADD_L     ( EAX, ECX )			/* f += stride               */
+    SUB_L     ( EAX, EDX )
+    JA        ( LLBL(v16_3dnow_pv_loop_start) )
+
+    FEMMS
+    POP_L     ( EBP )
+    RET
+
+
+
+/*
+ * gl_3dnow_project_clipped_vertices( float *first_vertex,
+ *                                    const float *last_vertex,
+ *                                    float *matrix,
+ *                                    GLuint stride,
+ *                                    const GLubyte *clip_mask )
+ */
+GLOBL GLNAME( gl_3dnow_project_clipped_vertices )
+GLNAME( gl_3dnow_project_clipped_vertices ):
+
+    PUSH_L    ( EBP )
+    PUSH_L    ( ESI )
+
+    FEMMS
+
+    PREFETCH  ( REGOFF(12, ESP) )		/* fetch the first vertex    */
+
+    MOV_L     ( REGOFF(12, ESP), ECX )		/* first_vert                */
+    MOV_L     ( REGOFF(16, ESP), EDX )		/* last_vert                 */
+    MOV_L     ( REGOFF(20, ESP), EBP )		/* matrix                    */
+    MOV_L     ( REGOFF(24, ESP), EAX )		/* stride                    */
+    MOV_L     ( REGOFF(28, ESP), ESI )		/* clip_mask                 */
+
+
+    MOVD      ( REGOFF(MAT_TX, EBP), MM6 )	/*             | tx          */
+    PUNPCKLDQ ( REGOFF(MAT_TY, EBP), MM6 )	/* ty          | tx          */
+    MOVD      ( REGIND(EBP), MM5 )
+    PUNPCKLDQ ( REGOFF(MAT_SY, EBP), MM5 )	/* vsy         | vsx         */
+    MOVD      ( REGOFF(MAT_SZ, EBP), MM1 )	/*             | vsz         */
+
+ALIGNTEXT32
+LLBL( v16_3dnow_pcv_loop_start ):
+
+    CMP_B     ( CONST(0), REGIND(ESI) )
+    JNE       ( LLBL(v16_3dnow_pcv_skip) )
+
+    MOVD      ( REGOFF(12, ECX), MM0)		/*             | f[3]        */
+    PFRCP     ( MM0, MM0 )			/* oow = 1/f[3]              */
+    MOVD      ( REGOFF(12, ECX), MM7)		/*             | f[3]        */
+    PFRCPIT1  ( MM0, MM7 )
+    PFRCPIT2  ( MM0, MM7 )			/* oow         | oow         */
+    PUNPCKLDQ ( MM7, MM7 )
+    MOVQ      ( REGIND(ECX), MM2 )		/* f[1]        | f[0]        */
+    PFMUL     ( MM7, MM2 )			/* f[1] * oow  | f[0] * oow  */
+    MOVD      ( REGOFF(8, ECX), MM3 )		/*             | f[2]        */
+    PFMUL     ( MM7, MM3 )			/*             | f[2] * oow  */
+    MOVD      ( REGOFF(MAT_TZ, EBP), MM0 )	/*             | vtz         */
+    PFMUL     ( MM1, MM3 )			/*             | f[2] *= vsz */
+    PFADD     ( MM0, MM3 )			/*             | f[2] += vtz */
+    PFMUL     ( MM5, MM2 )			/* f[1] *= vsy | f[0] *= vsx */
+    PFADD     ( MM6, MM2 )			/* f[1] += vty | f[0] += vtx */
+    PUNPCKLDQ ( MM7, MM3 )			/* f[3] = oow  | f[2]        */
+    MOVQ      ( MM2, REGOFF(0, ECX) )
+    MOVQ      ( MM3, REGOFF(8, ECX) )
+
+LLBL( v16_3dnow_pcv_skip ):
+
+    ADD_L     ( EAX, ECX )			/* f += stride               */
+    INC_L     ( ESI )				/* next clip_mask            */
+
+    CMP_L     ( ECX, EDX )
+    JNE       ( LLBL(v16_3dnow_pcv_loop_start) )
+
+    FEMMS
+
+    POP_L     ( ESI )
+    POP_L     ( EBP )
+    RET
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_masked1.S b/xc/extras/Mesa/src/X86/3dnow_xform_masked1.S
index 622edd9dd..b9ae99202 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_masked1.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_masked1.S
@@ -1,528 +1,501 @@
-#include "assyntax.h"
 
-   SEG_TEXT
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
 
+#include "assyntax.h"
+#include "xform_args.h"
 
-ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_general_masked)
-GLNAME( gl_3dnow_transform_points1_general_masked ):
+    SEG_TEXT
 
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+#define FRAME_OFFSET	16
 
 
-ALIGNTEXT32
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points1_general_masked )
+GLNAME( gl_3dnow_transform_points1_general_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVQ      ( REGOFF(8, ECX), MM1 )	/* m03             | m02             */
+
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVQ      ( REGOFF(56, ECX), MM3 )	/* m33             | m32             */
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVQ      ( REGOFF(8, ECX), MM1 )   /* m03           | m02                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGM_4 ) )
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVQ      ( REGOFF(56, ECX), MM3 )  /* m33           | m32                */
+ALIGNTEXT16
+LLBL( G3TPGM_2 ):
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGM_4) )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPGM_3 ) )	/* skip vertex                       */
 
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
 
-ALIGNTEXT32
-LLBL(G3TPGM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPGM_3) )      /* skip vertex                        */
+    MOVQ      ( MM4, MM5 )		/* x0              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
+    PFMUL     ( MM1, MM5 )		/* x0*m03          | x0*m02          */
+    PFADD     ( MM2, MM4 )		/* x0*m01+m31      | x0*m00+m30      */
 
-    MOVQ      ( MM4, MM5 )              /* x0            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
+    PFADD     ( MM3, MM5 )		/* x0*m03+m33      | x0*m02+m32      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFMUL     ( MM1, MM5 )              /* x0*m03        | x0*m02             */
-    PFADD     ( MM2, MM4 )              /* x0*m01+m31    | x0*m00+m30         */
+    MOVQ      ( MM5, REGOFF(8, EDX) )	/* write r3, r2                      */
 
-    PFADD     ( MM3, MM5 )              /* x0*m03+m33    | x0*m02+m32         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+LLBL( G3TPGM_3 ):
 
-    MOVQ      ( MM5, REGOFF(8, EDX) )   /* write r3, r2                       */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-LLBL(G3TPGM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    INC_L     ( EBP )			/* next clipmask                     */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPGM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPGM_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPGM_4 ):
 
-LLBL(G3TPGM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_identity_masked)
-GLNAME( gl_3dnow_transform_points1_identity_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(1), REGOFF(16, ECX) )
-    OR_B      ( CONST(1), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points1_perspective_masked )
+GLNAME( gl_3dnow_transform_points1_perspective_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-ALIGNTEXT32
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPM_4 ) )
 
-LLBL(G3TPIM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPIM_3)              /* skip vertex                        */ )
+ALIGNTEXT16
+LLBL( G3TPPM_2 ):
 
-    MOVD      ( REGIND(EAX), MM0 )      /*               | x0                 */
-    MOVD      ( MM0, REGIND(EDX) )      /*               | r0                 */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPPM_3 ) )	/* skip vertex                       */
 
-LLBL(G3TPIM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    MOVD      ( REGIND(EAX), MM4 )	/* 0               | x0              */
+    PFMUL     ( MM0, MM4 )		/* 0               | x0*m00          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( MM3, REGOFF(8, EDX) )	/* write r2  (=m32), r3 (=0)         */
 
-    JA        ( LLBL(G3TPIM_2)              /* cnt > 0 ? -> process next vertex   */ )
+LLBL( G3TPPM_3 ):
 
-LLBL(G3TPIM_4):
-    FEMMS
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TPPM_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TPPM_4 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_2d_masked)
-GLNAME( gl_3dnow_transform_points1_2d_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points1_3d_masked )
+GLNAME( gl_3dnow_transform_points1_3d_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVD      ( REGOFF(8, ECX), MM1 )	/*                 | m02             */
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2M_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3M_4 ) )
 
+ALIGNTEXT16
+LLBL( G3TP3M_2 ):
 
-ALIGNTEXT32
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3M_3 ) )	/* skip vertex                       */
 
-LLBL(G3TP2M_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2M_3) )      /* skip vertex                        */
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
+    MOVQ      ( MM4, MM5 )		/*                 | x0              */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
 
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
-    PFADD     ( MM2, MM4 )              /* x0*m01+m31    | x0*m00+m30         */
+    PFMUL     ( MM1, MM5 )		/*                 | x0*m02          */
+    PFADD     ( MM2, MM4 )		/* x0*m01+m31      | x0*m00+m30      */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    PFADD     ( MM3, MM5 )		/*                 | x0*m02+m32      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-LLBL(G3TP2M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVD      ( MM5, REGOFF(8, EDX) )	/* write r2                          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    INC_L     ( EBP )                   /* next clipmask                      */
+LLBL( G3TP3M_3 ):
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP2M_2) )      /* cnt > 0 ? -> process next vertex   */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TP2M_4):
-    FEMMS
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP3M_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TP3M_4 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_2d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points1_2d_no_rot_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points1_3d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points1_3d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRM_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRM_4 ) )
 
+ALIGNTEXT16
+LLBL( G3TP3NRM_2 ):
 
-ALIGNTEXT32
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3NRM_3))	/* skip vertex                       */
 
-LLBL(G3TP2NRM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2NRM_3)            /* skip vertex                        */ )
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PFMUL     ( MM0, MM4 )		/*                 | x0*m00          */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PFMUL     ( MM0, MM4 )              /*               | x0*m00             */
+    PFADD     ( MM2, MM4 )		/* m31             | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFADD     ( MM2, MM4 )              /* m31           | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    MOVD      ( MM3, REGOFF(8, EDX) )	/* write r2                          */
 
-LLBL(G3TP2NRM_3):
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+LLBL( G3TP3NRM_3 ):
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    JA        ( LLBL(G3TP2NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP3NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-LLBL(G3TP2NRM_4):
-    FEMMS
+LLBL( G3TP3NRM_4 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_3d_masked)
-GLNAME( gl_3dnow_transform_points1_3d_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points1_2d_masked )
+GLNAME( gl_3dnow_transform_points1_2d_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVD      ( REGOFF(8, ECX), MM1 )   /*               | m02                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3M_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2M_4 ) )
 
-ALIGNTEXT32
+ALIGNTEXT16
+LLBL( G3TP2M_2 ):
 
-LLBL(G3TP3M_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3M_3)              /* skip vertex                        */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2M_3 ) )	/* skip vertex                       */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
 
-    MOVQ      ( MM4, MM5 )              /*               | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
+    PFADD     ( MM2, MM4 )		/* x0*m01+m31      | x0*m00+m30      */
 
-    PFMUL     ( MM1, MM5 )              /*               | x0*m02             */
-    PFADD     ( MM2, MM4 )              /* x0*m01+m31    | x0*m00+m30         */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFADD     ( MM3, MM5 )              /*               | x0*m02+m32         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+LLBL( G3TP2M_3 ):
 
-    MOVD      ( MM5, REGOFF(8, EDX) )   /* write r2                           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-LLBL(G3TP3M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    INC_L     ( EBP )			/* next clipmask                     */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    INC_L     ( EBP )                   /* next clipmask                      */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2M_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP3M_2)              /* cnt > 0 ? -> process next vertex   */ )
+LLBL( G3TP2M_4 ):
 
-LLBL(G3TP3M_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_3d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points1_3d_no_rot_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-
-ALIGNTEXT32
+GLOBL GLNAME(gl_3dnow_transform_points1_2d_no_rot_masked)
+GLNAME( gl_3dnow_transform_points1_2d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRM_4 ) )
 
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRM_4) )
+ALIGNTEXT16
+LLBL( G3TP2NRM_2 ):
 
-ALIGNTEXT32
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2NRM_3 ) )	/* skip vertex                       */
 
-LLBL(G3TP3NRM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3NRM_3)            /* skip vertex                        */ )
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PFMUL     ( MM0, MM4 )		/*                 | x0*m00          */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PFMUL     ( MM0, MM4 )              /*               | x0*m00             */
+    PFADD     ( MM2, MM4 )		/* m31             | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFADD     ( MM2, MM4 )              /* m31           | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+LLBL( G3TP2NRM_3 ):
 
-    MOVD      ( MM3, REGOFF(8, EDX) )   /* write r2                           */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-LLBL(G3TP3NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    INC_L     ( EBP )                   /* next clipmask                      */
+    JNZ       ( LLBL( G3TP2NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP3NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+LLBL( G3TP2NRM_4 ):
 
-LLBL(G3TP3NRM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_perspective_masked)
-GLNAME( gl_3dnow_transform_points1_perspective_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points1_identity_masked )
+GLNAME( gl_3dnow_transform_points1_identity_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(1), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_1), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPM_4) )
+ALIGNTEXT16
+LLBL( G3TPIM_2 ):
 
-ALIGNTEXT32
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPIM_3 ) )	/* skip vertex                       */
 
-LLBL(G3TPPM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPPM_3)              /* skip vertex                        */ )
+    MOVD      ( REGIND(EAX), MM0 )	/*                 | x0              */
+    MOVD      ( MM0, REGIND(EDX) )	/*                 | r0              */
 
-    MOVD      ( REGIND(EAX), MM4 )      /* 0             | x0                 */
-    PFMUL     ( MM0, MM4 )              /* 0             | x0*m00             */
+LLBL( G3TPIM_3 ):
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( MM3, REGOFF(8, EDX) )   /* write r2  (=m32), r3 (=0)          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TPPM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TPIM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TPPM_2)              /* cnt > 0 ? -> process next vertex   */ )
+LLBL( G3TPIM_4 ):
 
-LLBL(G3TPPM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
-
-
-
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_masked2.S b/xc/extras/Mesa/src/X86/3dnow_xform_masked2.S
index 8ea66549f..37204c2dc 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_masked2.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_masked2.S
@@ -1,587 +1,568 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
 #include "assyntax.h"
+#include "xform_args.h"
 
-   SEG_TEXT
+    SEG_TEXT
+
+#define FRAME_OFFSET	16
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_general_masked)
+GLOBL GLNAME( gl_3dnow_transform_points2_general_masked )
 GLNAME( gl_3dnow_transform_points2_general_masked ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
-
-    PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    FEMMS
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
-
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
-
-    MOVD      ( REGOFF(8, ECX), MM2 )   /*               | m02                */
-    MOVD      ( REGOFF(24, ECX), MM7 )  /*               | m12                */
-    PSLLQ     ( CONST(32), MM7 )        /* m12           |                    */
-    POR       ( MM7, MM2 )              /* m12           | m02                */
-
-    MOVD      ( REGOFF(12, ECX), MM3 )  /*               | m03                */
-    MOVD      ( REGOFF(28, ECX), MM7 )  /*               | m13                */
-    PSLLQ     ( CONST(32), MM7 )        /* m13           |                    */
-    POR       ( MM7, MM3 )              /* m13           | m03                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
+
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
+
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
+
+    MOVD      ( REGOFF(8, ECX), MM2 )	/*                 | m02             */
+    MOVD      ( REGOFF(24, ECX), MM7 )	/*                 | m12             */
+    PSLLQ     ( CONST(32), MM7 )	/* m12             |                 */
+    POR       ( MM7, MM2 )		/* m12             | m02             */
+
+    MOVD      ( REGOFF(12, ECX), MM3 )	/*                 | m03             */
+    MOVD      ( REGOFF(28, ECX), MM7 )	/*                 | m13             */
+    PSLLQ     ( CONST(32), MM7 )	/* m13             |                 */
+    POR       ( MM7, MM3 )		/* m13             | m03             */
+
+    MOVQ      ( REGOFF(48, ECX), MM4 )	/* m31             | m30             */
+    MOVQ      ( REGOFF(56, ECX), MM5 )	/* m33             | m32             */
+
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGM_4 ) )
 
-    MOVQ      ( REGOFF(48, ECX), MM4 )  /* m31           | m30                */
-    MOVQ      ( REGOFF(56, ECX), MM5 )  /* m33           | m32                */
+ALIGNTEXT16
+LLBL( G3TPGM_2 ):
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGM_4) )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPGM_3 ) )	/* skip vertex                       */
 
-ALIGNTEXT32
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
 
-LLBL(G3TPGM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPGM_3) )      /* skip vertex                        */
+    PFMUL     ( MM0, MM6 )		/* x1*m10          | x0*m00          */
+    PFMUL     ( MM1, MM7 )		/* x1*m11          | x0*m01          */
 
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
+    PFACC     ( MM7, MM6 )		/* x0*m01+x1*m11   | x0*x00+x1*m10   */
+    PFADD     ( MM4, MM6 )		/* x0*...*m11+m31  | x0*...*m10+m30  */
 
-    PFMUL     ( MM0, MM6 )              /* x1*m10        | x0*m00             */
-    PFMUL     ( MM1, MM7 )              /* x1*m11        | x0*m01             */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
 
-    PFACC     ( MM7, MM6 )              /* x0*m01+x1*m11 | x0*x00+x1*m10      */
-    PFADD     ( MM4, MM6 )              /* x0*...*m11+m31| x0*x00+x1*m10+m30  */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
+    PFMUL     ( MM2, MM6 )		/* x1*m12          | x0*m02          */
 
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
+    PFMUL     ( MM3, MM7 )		/* x1*m13          | x0*m03          */
+    PFACC     ( MM7, MM6 )		/* x0*m03+x1*m13   | x0*x02+x1*m12   */
 
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
-    PFMUL     ( MM2, MM6 )              /* x1*m12        | x0*m02             */
+    PFADD     ( MM5, MM6 )		/* x0*...*m13+m33  | x0*...*m12+m32  */
+    MOVQ      ( MM6, REGOFF(8, EDX) )	/* write r3, r2                      */
 
-    PFMUL     ( MM3, MM7 )              /* x1*m13        | x0*m03             */
-    PFACC     ( MM7, MM6 )              /* x0*m03+x1*m13 | x0*x02+x1*m12      */
+ALIGNTEXT16
+LLBL( G3TPGM_3 ):
 
-    PFADD     ( MM5, MM6 )              /* x0*...*m13+m33| x0*x02+x1*m12+m32  */
-    MOVQ      ( MM6, REGOFF(8, EDX) )   /* write r3, r2                       */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TPGM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TPGM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TPGM_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPGM_4 ):
 
-LLBL(G3TPGM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_identity_masked)
-GLNAME( gl_3dnow_transform_points2_identity_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points2_perspective_masked )
+GLNAME( gl_3dnow_transform_points2_perspective_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM1 )	/*                 | m11             */
 
-ALIGNTEXT32
+    PSLLQ     ( CONST(32), MM1 )	/* m11             |                 */
+    POR       ( MM1, MM0 )		/* m11             | m00             */
 
-LLBL(G3TPIM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPIM_3) )      /* skip vertex                        */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                 */
-    MOVQ      ( MM0, REGIND(EDX) )      /* r1            | r0                 */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPM_4 ) )
 
-LLBL(G3TPIM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+ALIGNTEXT16
+LLBL( G3TPPM_2 ):
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPPM_3 ) )	/* skip vertex                       */
 
-    JA        ( LLBL(G3TPIM_2) )      /* cnt > 0 ? -> process next vertex   */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-LLBL(G3TPIM_4):
-    FEMMS
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( MM3, REGOFF(8, EDX) )	/* write r2  (=m32), r3 (=0)         */
+
+LLBL( G3TPPM_3 ):
 
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TPPM_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TPPM_4 ):
+
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_2d_masked)
-GLNAME( gl_3dnow_transform_points2_2d_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points2_3d_masked )
+GLNAME( gl_3dnow_transform_points2_3d_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVQ      ( REGOFF(16, ECX), MM1 )  /* m11           | m10                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2M_4) )
+    MOVD      ( REGOFF(8, ECX), MM2 )	/*                 | m02             */
+    MOVD      ( REGOFF(24, ECX), MM7 )	/*                 | m12             */
+    PSLLQ     ( CONST(32), MM7 )	/* m12             |                 */
+    POR       ( MM7, MM2 )		/* m12             | m02             */
 
-ALIGNTEXT32
+    MOVQ      ( REGOFF(48, ECX), MM4 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM5 )	/*                 | m32             */
 
-LLBL(G3TP2M_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2M_3) )      /* skip vertex                        */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3M_4 ) )
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    MOVD      ( REGOFF(4, EAX), MM5 )   /*               | x1                 */
+ALIGNTEXT16
+LLBL( G3TP3M_2 ):
 
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3M_3 ) )	/* skip vertex                       */
 
-    PUNPCKLDQ ( MM5, MM5 )              /* x1            | x1                 */
-    PFMUL     ( MM1, MM5 )              /* x1*m11        | x1*m10             */
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
 
-    PFADD     ( MM5, MM4 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
-    PFADD     ( MM2, MM4 )              /* x...x1*m11+31 | x0*m00+x1*m10+m30  */
+    PFMUL     ( MM0, MM6 )		/* x1*m10          | x0*m00          */
+    PFMUL     ( MM1, MM7 )		/* x1*m11          | x0*m01          */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    PFACC     ( MM7, MM6 )		/* x0*m01+x1*m11   | x0*x00+x1*m10   */
+    PFADD     ( MM4, MM6 )		/* x0*...*m11+m31  | x0*...*m10+m30  */
 
-LLBL(G3TP2M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
+    PFMUL     ( MM2, MM6 )		/* x1*m12          | x0*m02          */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP2M_2) )      /* cnt > 0 ? -> process next vertex   */
+    PFACC     ( MM7, MM6 )		/* ***trash***     | x0*x02+x1*m12   */
 
-LLBL(G3TP2M_4):
-    FEMMS
+    PFADD     ( MM5, MM6 )		/* ***trash***     | x0*...*m12+m32  */
+    MOVD      ( MM6, REGOFF(8, EDX) )	/* write r2                          */
+
+LLBL( G3TP3M_3 ):
+
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TP3M_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TP3M_4 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_2d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points2_2d_no_rot_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points2_3d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points2_3d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM1 )	/*                 | m11             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM1 )  /*               | m11                */
+    PSLLQ     ( CONST(32), MM1 )	/* m11             |                 */
+    POR       ( MM1, MM0 )		/* m11             | m00             */
 
-    PSLLQ     ( CONST(32), MM1 )        /* m11           |                    */
-    POR       ( MM1, MM0 )              /* m11           | m00                */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRM_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRM_4 ) )
 
-ALIGNTEXT32
+ALIGNTEXT16
+LLBL( G3TP3NRM_2 ):
 
-LLBL(G3TP2NRM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2NRM_3) )    /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3NRM_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    PFADD     ( MM2, MM4 )              /* m31           | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    PFADD     ( MM2, MM4 )		/* x1*m11+m31      | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-LLBL(G3TP2NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    MOVD      ( MM3, REGOFF(8, EDX) )	/* write r2                          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+LLBL( G3TP3NRM_3 ):
 
-    JA        ( LLBL(G3TP2NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TP2NRM_4):
-    FEMMS
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TP3NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TP3NRM_4 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_3d_masked)
-GLNAME( gl_3dnow_transform_points2_3d_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points2_2d_masked )
+GLNAME( gl_3dnow_transform_points2_2d_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    FEMMS
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
-
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVD      ( REGOFF(8, ECX), MM2 )   /*               | m02                */
-    MOVD      ( REGOFF(24, ECX), MM7 )  /*               | m12                */
-    PSLLQ     ( CONST(32), MM7 )        /* m12           |                    */
-    POR       ( MM7, MM2 )              /* m12           | m02                */
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVQ      ( REGOFF(16, ECX), MM1 )	/* m11             | m10             */
 
-    MOVQ      ( REGOFF(48, ECX), MM4 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM5 )  /*               | m32                */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3M_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2M_4 ) )
 
-ALIGNTEXT32
+ALIGNTEXT16
+LLBL( G3TP2M_2 ):
 
-LLBL(G3TP3M_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3M_3) )      /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2M_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    MOVD      ( REGOFF(4, EAX), MM5 )	/*                 | x1              */
 
-    PFMUL     ( MM0, MM6 )              /* x1*m10        | x0*m00             */
-    PFMUL     ( MM1, MM7 )              /* x1*m11        | x0*m01             */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
 
-    PFACC     ( MM7, MM6 )              /* x0*m01+x1*m11 | x0*x00+x1*m10      */
-    PFADD     ( MM4, MM6 )              /* x0*...*m11+m31| x0*x00+x1*m10+m30  */
+    PUNPCKLDQ ( MM5, MM5 )		/* x1              | x1              */
+    PFMUL     ( MM1, MM5 )		/* x1*m11          | x1*m10          */
 
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
+    PFADD     ( MM5, MM4 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
+    PFADD     ( MM2, MM4 )		/* x...x1*m11+31   | x0*...*m10+m30  */
 
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
-    PFMUL     ( MM2, MM6 )              /* x1*m12        | x0*m02             */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFACC     ( MM7, MM6 )              /* ***trash***   | x0*x02+x1*m12      */
+LLBL( G3TP2M_3 ):
 
-    PFADD     ( MM5, MM6 )              /* ***trash***   | x0*x02+x1*m12+m32  */
-    MOVD      ( MM6, REGOFF(8, EDX) )   /* write r2                           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-LLBL(G3TP3M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    INC_L     ( EBP )			/* next clipmask                     */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2M_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP3M_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP2M_4 ):
 
-LLBL(G3TP3M_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_3d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points2_3d_no_rot_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points2_2d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points2_2d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    FEMMS
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM1 )  /*               | m11                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM1 )	/*                 | m11             */
 
-    PSLLQ     ( CONST(32), MM1 )        /* m11           |                    */
-    POR       ( MM1, MM0 )              /* m11           | m00                */
+    PSLLQ     ( CONST(32), MM1 )	/* m11             |                 */
+    POR       ( MM1, MM0 )		/* m11             | m00             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRM_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRM_4 ) )
 
-ALIGNTEXT32
+ALIGNTEXT16
+LLBL( G3TP2NRM_2 ):
 
-LLBL(G3TP3NRM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3NRM_3) )    /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2NRM_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    PFADD     ( MM2, MM4 )              /* x1*m11+m31    | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    PFADD     ( MM2, MM4 )		/* m31             | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    MOVD      ( MM3, REGOFF(8, EDX) )   /* write r2                           */
+LLBL( G3TP2NRM_3 ):
 
-LLBL(G3TP3NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    JA        ( LLBL(G3TP3NRM_2) )    /* cnt > 0 ? -> process next vertex   */
+    JNZ       ( LLBL( G3TP2NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-LLBL(G3TP3NRM_4):
+LLBL( G3TP2NRM_4 ):
 
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_perspective_masked)
-GLNAME( gl_3dnow_transform_points2_perspective_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points2_identity_masked )
+GLNAME( gl_3dnow_transform_points2_identity_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM1 )  /*               | m11                */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    PSLLQ     ( CONST(32), MM1 )        /* m11           |                    */
-    POR       ( MM1, MM0 )              /* m11           | m00                */
-
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPM_4) )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-ALIGNTEXT32
+ALIGNTEXT16
+LLBL( G3TPIM_2 ):
 
-LLBL(G3TPPM_2):
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPIM_3 ) )	/* skip vertex                       */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPPM_3) )      /* skip vertex                        */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVQ      ( MM0, REGIND(EDX) )	/* r1              | r0              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+LLBL( G3TPIM_3 ):
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( MM3, REGOFF(8, EDX) )   /* write r2  (=m32), r3 (=0)          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TPPM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TPIM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TPPM_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPIM_4 ):
 
-LLBL(G3TPPM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_masked3.S b/xc/extras/Mesa/src/X86/3dnow_xform_masked3.S
index fbc66c34b..45686e03f 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_masked3.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_masked3.S
@@ -1,729 +1,709 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
 #include "assyntax.h"
+#include "xform_args.h"
+
+    SEG_TEXT
 
-   SEG_TEXT
+#define FRAME_OFFSET	16
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_general_masked)
+GLOBL GLNAME( gl_3dnow_transform_points3_general_masked )
 GLNAME( gl_3dnow_transform_points3_general_masked ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
-
-    PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    FEMMS
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
 
-    MOVQ      ( REGOFF(32, ECX), MM2 )  /* m21           | m20                */
-    MOVQ      ( REGOFF(48, ECX), MM3 )  /* m31           | m30                */
+    MOVQ      ( REGOFF(32, ECX), MM2 )	/* m21             | m20             */
+    MOVQ      ( REGOFF(48, ECX), MM3 )	/* m31             | m30             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGM_6) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGM_6 ) )
 
     PUSH_L    ( EBP )
     PUSH_L    ( EAX )
     PUSH_L    ( EDX )
     PUSH_L    ( ESI )
 
+ALIGNTEXT16
+LLBL( G3TPGM_2 ):
 
-ALIGNTEXT32
-LLBL(G3TPGM_2):
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPGM_3 ) )	/* skip vertex                       */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPGM_3)          /* skip vertex                        */ )
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM6 )	/*                 | x2              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM6 )   /*               | x2                 */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m10          | x0*m00          */
 
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m10        | x0*m00             */
+    PUNPCKLDQ ( MM6, MM6 )		/* x2              | x2              */
+    PFMUL     ( MM1, MM5 )		/* x1*m11          | x0*m01          */
 
-    PUNPCKLDQ ( MM6, MM6 )              /* x2            | x2                 */
-    PFMUL     ( MM1, MM5 )              /* x1*m11        | x0*m01             */
+    PFMUL     ( MM2, MM6 )		/* x2*m21          | x2*m20          */
+    PFACC     ( MM5, MM4 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
 
-    PFMUL     ( MM2, MM6 )              /* x2*m21        | x2*m20             */
-    PFACC     ( MM5, MM4 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
+    PFADD     ( MM3, MM6 )		/* x2*m21+m31      | x2*m20+m30      */
+    PFADD     ( MM4, MM6 )		/* r1              | r0              */
 
-    PFADD     ( MM3, MM6 )              /* x2*m21+m31    | x2*m20+m30         */
-    PFADD     ( MM4, MM6 )              /* r1            | r0                 */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r0, r1                      */
 
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r0, r1                       */
+ALIGNTEXT16
+LLBL( G3TPGM_3 ):
 
-LLBL(G3TPGM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    JA        ( LLBL(G3TPGM_2)          /* cnt > 0 ? -> process next vertex   */ )
-                                        /* and now the second stripe ...      */
+    JNZ       ( LLBL( G3TPGM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    POP_L     ( ESI )                   /* reset counter & pointers           */
+                                	/* and now the second stripe ...     */
+    POP_L     ( ESI )			/* reset counter & pointers          */
     POP_L     ( EDX )
     POP_L     ( EAX )
     POP_L     ( EBP )
 
-    MOVD      ( REGOFF(8, ECX), MM0 )   /*               | m02                */
-    MOVD      ( REGOFF(24, ECX), MM7 )  /*               | m12                */
+    MOVD      ( REGOFF(8, ECX), MM0 )	/*                 | m02             */
+    MOVD      ( REGOFF(24, ECX), MM7 )	/*                 | m12             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m12           |                    */
-    POR       ( MM7, MM0 )              /* m12           | m02                */
+    PSLLQ     ( CONST(32), MM7 )	/* m12             |                 */
+    POR       ( MM7, MM0 )		/* m12             | m02             */
 
-    MOVD      ( REGOFF(12, ECX), MM1 )  /*               | m03                */
-    MOVD      ( REGOFF(28, ECX), MM7 )  /*               | m13                */
+    MOVD      ( REGOFF(12, ECX), MM1 )	/*                 | m03             */
+    MOVD      ( REGOFF(28, ECX), MM7 )	/*                 | m13             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m13           |                    */
-    POR       ( MM7, MM1 )              /* m13           | m03                */
+    PSLLQ     ( CONST(32), MM7 )	/* m13             |                 */
+    POR       ( MM7, MM1 )		/* m13             | m03             */
 
-    MOVQ      ( REGOFF(40, ECX), MM2 )  /* m23           | m22                */
-    MOVQ      ( REGOFF(56, ECX), MM3 )  /* m33           | m32                */
+    MOVQ      ( REGOFF(40, ECX), MM2 )	/* m23             | m22             */
+    MOVQ      ( REGOFF(56, ECX), MM3 )	/* m33             | m32             */
 
+ALIGNTEXT16
+LLBL( G3TPGM_4 ):
 
-ALIGNTEXT32
-LLBL(G3TPGM_4):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPGM_5)          /* skip vertex                        */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPGM_5 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM6 )   /*               | x2                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM6 )	/*                 | x2              */
 
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m12        | x0*m02             */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m12          | x0*m02          */
 
-    PUNPCKLDQ ( MM6, MM6 )              /* x2            | x2                 */
-    PFMUL     ( MM1, MM5 )              /* x1*m13        | x0*m03             */
+    PUNPCKLDQ ( MM6, MM6 )		/* x2              | x2              */
+    PFMUL     ( MM1, MM5 )		/* x1*m13          | x0*m03          */
 
-    PFMUL     ( MM2, MM6 )              /* x2*m23        | x2*m22             */
-    PFACC     ( MM5, MM4 )              /* x0*m03+x1*m13 | x0*m02+x1*m12      */
+    PFMUL     ( MM2, MM6 )		/* x2*m23          | x2*m22          */
+    PFACC     ( MM5, MM4 )		/* x0*m03+x1*m13   | x0*m02+x1*m12   */
 
-    PFADD     ( MM3, MM6 )              /* x2*m23+m33    | x2*m22+m32         */
-    PFADD     ( MM4, MM6 )              /* r3            | r2                 */
+    PFADD     ( MM3, MM6 )		/* x2*m23+m33      | x2*m22+m32      */
+    PFADD     ( MM4, MM6 )		/* r3              | r2              */
 
-    MOVQ      ( MM6, REGOFF(8, EDX) )   /* write r2, r3                       */
+    MOVQ      ( MM6, REGOFF(8, EDX) )	/* write r2, r3                      */
 
-LLBL(G3TPGM_5):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+ALIGNTEXT16
+LLBL( G3TPGM_5 ):
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    JA        ( LLBL(G3TPGM_4)          /* cnt > 0 ? -> process next vertex   */ )
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-LLBL(G3TPGM_6):
+    JNZ       ( LLBL( G3TPGM_4 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    FEMMS
+LLBL( G3TPGM_6 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_identity_masked)
-GLNAME( gl_3dnow_transform_points3_identity_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points3_perspective_masked )
+GLNAME( gl_3dnow_transform_points3_perspective_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM0 )		/* m11             | m00             */
 
-ALIGNTEXT32
-LLBL(G3TPIM_2):
+    MOVQ      ( REGOFF(32, ECX), MM1 )	/* m21             | m20             */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPIM_3)          /* skip vertex                        */ )
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM1 )   /*               | x2                 */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPM_4 ) )
 
-    MOVQ      ( MM0, REGIND(EDX) )      /* r1            | r0                 */
-    MOVD      ( MM1, REGOFF(8, EDX) )   /*               | r2                 */
+ALIGNTEXT16
+LLBL( G3TPPM_2 ):
 
-LLBL(G3TPIM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPPM_3 ) )	/* skip vertex                       */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-    JA        ( LLBL(G3TPIM_2)          /* cnt > 0 ? -> process next vertex   */ )
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    MOVQ      ( MM5, MM6 )		/*                 | x2              */
 
-LLBL(G3TPIM_4):
-    FEMMS
+    PUNPCKLDQ ( MM5, MM5 )		/* x2              | x2              */
+    PFMUL     ( MM1, MM5 )		/* x2*m21          | x2*m20          */
 
-    POP_L     ( EBP )
-    POP_L     ( EBX )
-    POP_L     ( EDI )
-    POP_L     ( ESI )
+    PFADD     ( MM4, MM5 )		/* x1*m11+x2*m21   | x0*m00+x2*m20   */
+    MOVQ      ( MM5, REGIND(EDX) )	/* write r0, r1                      */
 
-    POP_L     ( ESI )
-    RET
+    MOVQ      ( MM6, MM5 )		/*                 | x2              */
+    PFMUL     ( MM2, MM5 )		/*                 | x2*m22          */
 
+    PFADD     ( MM3, MM5 )		/*                 | x2*m22+m32      */
+    PFSUBR    ( MM7, MM6 )		/* (LO mm7 == 0)   | -x2             */
 
+    MOVD      ( MM5, REGOFF(8, EDX) )	/* write r2                          */
+    MOVD      ( MM6, REGOFF(12, EDX) )	/* write r3                          */
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_2d_masked)
-GLNAME( gl_3dnow_transform_points3_2d_masked ):
+LLBL( G3TPPM_3 ):
 
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-ALIGNTEXT32
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    PUSH_L    ( ESI )
-    PUSH_L    ( EDI )
-    PUSH_L    ( EBX )
-    PUSH_L    ( EBP )
+    JNZ       ( LLBL( G3TPPM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+LLBL( G3TPPM_4 ):
 
     FEMMS
+    POP_L     ( EBP )
+    POP_L     ( EBX )
+    POP_L     ( EDI )
+    POP_L     ( ESI )
+    RET
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_3d_masked )
+GLNAME( gl_3dnow_transform_points3_3d_masked ):
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2M_4) )
+    PUSH_L    ( ESI )
+    PUSH_L    ( EDI )
+    PUSH_L    ( EBX )
+    PUSH_L    ( EBP )
 
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-ALIGNTEXT32
-LLBL(G3TP2M_2):
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2M_3)          /* skip vertex                        */ )
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
 
-    MOVQ      ( REGIND(EAX), MM3 )      /* x1            | x0                 */
-    MOVQ      ( MM3, MM4 )              /* x1            | x0                 */
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
 
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
-    PFMUL     ( MM0, MM3 )              /* x1*m10        | x0*m00             */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    PFMUL     ( MM1, MM4 )              /* x1*m11        | x0*m01             */
-    PFACC     ( MM4, MM3 )              /* x0*m00+x1*m10 | x0*m01+x1*m11      */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
 
-    PFADD     ( MM2, MM3 )              /* x0*...*m10+m30| x0*m01+x1*m11+m31  */
-    MOVQ      ( MM3, REGIND(EDX) )      /* write r0, r1                       */
+    MOVQ      ( REGOFF(32, ECX), MM2 )	/* m21             | m20             */
+    MOVQ      ( REGOFF(48, ECX), MM3 )	/* m31             | m30             */
 
-    MOVD      ( MM5, REGOFF(8, EDX) )   /* write r2 (=x2)                     */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3M_6 ) )
 
-LLBL(G3TP2M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    PUSH_L    ( EBP )
+    PUSH_L    ( EAX )
+    PUSH_L    ( EDX )
+    PUSH_L    ( ESI )
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+ALIGNTEXT16
+LLBL( G3TP3M_2 ):
 
-    JA        ( LLBL(G3TP2M_2)              /* cnt > 0 ? -> process next vertex   */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3M_3 ) )	/* skip vertex                       */
 
-LLBL(G3TP2M_4):
-    FEMMS
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
 
-    POP_L     ( EBP )
-    POP_L     ( EBX )
-    POP_L     ( EDI )
-    POP_L     ( ESI )
+    PFMUL     ( MM0, MM4 )		/* x1*m10          | x0*m00          */
+    MOVD      ( REGOFF(8, EAX), MM6 )	/*                 | x2              */
 
-    POP_L     ( ESI )
-    RET
+    PFMUL     ( MM1, MM5 )		/* x1*m11          | x0*m01          */
+    PUNPCKLDQ ( MM6, MM6 )		/* x2              | x2              */
 
+    PFMUL     ( MM2, MM6 )		/* x2*m21          | x2*m20          */
+    PFACC     ( MM5, MM4 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
 
+    PFADD     ( MM3, MM6 )		/* x2*m21+m31      | x2*m20+m30      */
+    PFADD     ( MM4, MM6 )		/* r1              | r0              */
 
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r0, r1                      */
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_2d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points3_2d_no_rot_masked ):
+LLBL( G3TP3M_3 ):
 
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-ALIGNTEXT32
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    PUSH_L    ( ESI )
-    PUSH_L    ( EDI )
-    PUSH_L    ( EBX )
-    PUSH_L    ( EBP )
+    JNZ       ( LLBL( G3TP3M_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+                                	/* and now the second stripe ...     */
+    MOVD      ( REGOFF(8, ECX), MM0 )	/*                 | m02             */
 
-    FEMMS
+    MOVD      ( REGOFF(24, ECX), MM7 )	/*                 | m12             */
+    PSLLQ     ( CONST(32), MM7 )	/* m12             |                 */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    POR       ( MM7, MM0 )		/* m12             | m02             */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM0 )              /* m11           | m00                */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRM_4) )
+    POP_L     ( ESI )			/* reset counter & pointers          */
+    POP_L     ( EDX )
+    POP_L     ( EAX )
+    POP_L     ( EBP )
 
+ALIGNTEXT16
+LLBL( G3TP3M_4 ):
 
-ALIGNTEXT32
-LLBL(G3TP2NRM_2):
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3M_5 ) )	/* skip vertex                       */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2NRM_3)            /* skip vertex                        */ )
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
+    PFMUL     ( MM0, MM4 )		/* x1*m12          | x0*m02          */
+    MOVD      ( REGOFF(8, EAX), MM6 )	/*                 | x2              */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-    MOVD      ( MM5, REGOFF(8, EDX) )   /* write r2 (=x2)                     */
+    PFMUL     ( MM1, MM5 )		/* x1*m13          | x0*m03          */
+    PFACC     ( MM5, MM4 )		/* x0*m03+x1*m13   | x0*m02+x1*m12   */
 
-    PFADD     ( MM1, MM4 )              /* x1*m11+m31    | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r0, r1                       */
+    PFMUL     ( MM2, MM6 )		/*                 | x2*m22          */
+    PFADD     ( MM3, MM6 )		/*                 | x2*m22+m32      */
 
-LLBL(G3TP2NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    PFADD     ( MM4, MM6 )		/*                 | r2              */
+    MOVD      ( MM6, REGOFF(8, EDX) )	/* write r2                          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+ALIGNTEXT16
+LLBL( G3TP3M_5 ):
 
-    JA        ( LLBL(G3TP2NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TP2NRM_4):
-    FEMMS
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TP3M_4 ) )	/* cnt > 0 ? -> process next vertex  */
 
+LLBL( G3TP3M_6 ):
+
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_3d_masked)
-GLNAME( gl_3dnow_transform_points3_3d_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points3_3d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points3_3d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM0 )		/* m11             | m00             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
+    PUNPCKLDQ ( MM2, MM2 )		/* m22             | m22             */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+    PUNPCKLDQ ( MM3, MM3 )		/* m32             | m32             */
 
-    MOVQ      ( REGOFF(32, ECX), MM2 )  /* m21           | m20                */
-    MOVQ      ( REGOFF(48, ECX), MM3 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3M_6) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRM_4 ) )
 
-    PUSH_L    ( EBP )
-    PUSH_L    ( EAX )
-    PUSH_L    ( EDX )
-    PUSH_L    ( ESI )
+ALIGNTEXT16
+LLBL( G3TP3NRM_2 ):
 
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3NRM_3 ) )	/* skip vertex                       */
 
-ALIGNTEXT32
-LLBL(G3TP3M_2):
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3M_3)              /* skip vertex                        */ )
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    PFMUL     ( MM2, MM5 )		/*                 | x2*m22          */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
+    PFADD     ( MM1, MM4 )		/* x1*m11+m31      | x0*m00+m30      */
+    PFADD     ( MM3, MM5 )		/*                 | x2*m22+m32      */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m10        | x0*m00             */
-    MOVD      ( REGOFF(8, EAX), MM6 )   /*               | x2                 */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r0, r1                      */
+    MOVD      ( MM5, REGOFF(8, EDX) )	/* write r2                          */
 
-    PFMUL     ( MM1, MM5 )              /* x1*m11        | x0*m01             */
-    PUNPCKLDQ ( MM6, MM6 )              /* x2            | x2                 */
+LLBL( G3TP3NRM_3 ):
 
-    PFMUL     ( MM2, MM6 )              /* x2*m21        | x2*m20             */
-    PFACC     ( MM5, MM4 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    PFADD     ( MM3, MM6 )              /* x2*m21+m31    | x2*m20+m30         */
-    PFADD     ( MM4, MM6 )              /* r1            | r0                 */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r0, r1                       */
+    JNZ       ( LLBL( G3TP3NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-LLBL(G3TP3M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+LLBL( G3TP3NRM_4 ):
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    FEMMS
+    POP_L     ( EBP )
+    POP_L     ( EBX )
+    POP_L     ( EDI )
+    POP_L     ( ESI )
+    RET
 
-    JA        ( LLBL(G3TP3M_2)              /* cnt > 0 ? -> process next vertex   */ )
-                                        /* and now the second stripe ...      */
-    MOVD      ( REGOFF(8, ECX), MM0 )   /*               | m02                */
 
-    MOVD      ( REGOFF(24, ECX), MM7 )  /*               | m12                */
-    PSLLQ     ( CONST(32), MM7 )        /* m12           |                    */
 
-    POR       ( MM7, MM0 )              /* m12           | m02                */
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
 
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
-    POP_L     ( ESI )                   /* reset counter & pointers           */
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_2d_masked )
+GLNAME( gl_3dnow_transform_points3_2d_masked ):
 
-    POP_L     ( EDX )
-    POP_L     ( EAX )
+    PUSH_L    ( ESI )
+    PUSH_L    ( EDI )
+    PUSH_L    ( EBX )
+    PUSH_L    ( EBP )
 
-    POP_L     ( EBP )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-ALIGNTEXT32
-LLBL(G3TP3M_4):
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3M_5)              /* skip vertex                        */ )
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m12        | x0*m02             */
-    MOVD      ( REGOFF(8, EAX), MM6 )   /*               | x2                 */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
 
-    PFMUL     ( MM1, MM5 )              /* x1*m13        | x0*m03             */
-    PFACC     ( MM5, MM4 )              /* x0*m03+x1*m13 | x0*m02+x1*m12      */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    PFMUL     ( MM2, MM6 )              /*               | x2*m22             */
-    PFADD     ( MM3, MM6 )              /*               | x2*m22+m32         */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2M_4 ) )
 
-    PFADD     ( MM4, MM6 )              /*               | r2                 */
-    MOVD      ( MM6, REGOFF(8, EDX) )   /* write r2                           */
+ALIGNTEXT16
+LLBL( G3TP2M_2 ):
 
-LLBL(G3TP3M_5):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2M_3 ) )	/* skip vertex                       */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVQ      ( REGIND(EAX), MM3 )	/* x1              | x0              */
+    MOVQ      ( MM3, MM4 )		/* x1              | x0              */
 
-    JA        ( LLBL(G3TP3M_4)              /* cnt > 0 ? -> process next vertex   */ )
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
+    PFMUL     ( MM0, MM3 )		/* x1*m10          | x0*m00          */
 
-LLBL(G3TP3M_6):
-    FEMMS
+    PFMUL     ( MM1, MM4 )		/* x1*m11          | x0*m01          */
+    PFACC     ( MM4, MM3 )		/* x0*m00+x1*m10   | x0*m01+x1*m11   */
+
+    PFADD     ( MM2, MM3 )		/* x0*...*m10+m30  | x0*...*m11+m31  */
+    MOVQ      ( MM3, REGIND(EDX) )	/* write r0, r1                      */
+
+    MOVD      ( MM5, REGOFF(8, EDX) )	/* write r2 (=x2)                    */
 
+ALIGNTEXT16
+LLBL( G3TP2M_3 ):
+
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TP2M_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TP2M_4 ):
+
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_3d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points3_3d_no_rot_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points3_2d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points3_2d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM0 )              /* m11           | m00                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
-    PUNPCKLDQ ( MM2, MM2 )              /* m22           | m22                */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM0 )		/* m11             | m00             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
 
-    PUNPCKLDQ ( MM3, MM3 )              /* m32           | m32                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRM_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRM_4 ) )
 
+ALIGNTEXT16
+LLBL( G3TP2NRM_2 ):
 
-ALIGNTEXT32
-LLBL(G3TP3NRM_2):
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2NRM_3 ) )	/* skip vertex                       */
 
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3NRM_3)            /* skip vertex                        */ )
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    MOVD      ( MM5, REGOFF(8, EDX) )	/* write r2 (=x2)                    */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-    PFMUL     ( MM2, MM5 )              /*               | x2*m22             */
+    PFADD     ( MM1, MM4 )		/* x1*m11+m31      | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r0, r1                      */
 
-    PFADD     ( MM1, MM4 )              /* x1*m11+m31    | x0*m00+m30         */
-    PFADD     ( MM3, MM5 )              /*               | x2*m22+m32         */
+ALIGNTEXT16
+LLBL( G3TP2NRM_3 ):
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r0, r1                       */
-    MOVD      ( MM5, REGOFF(8, EDX) )   /* write r2                           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TP3NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TP2NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP3NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+LLBL( G3TP2NRM_4 ):
 
-LLBL(G3TP3NRM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points3_perspective_masked)
-GLNAME( gl_3dnow_transform_points3_perspective_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points3_identity_masked )
+GLNAME( gl_3dnow_transform_points3_identity_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
-
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM0 )              /* m11           | m00                */
-
-    MOVQ      ( REGOFF(32, ECX), MM1 )  /* m21           | m20                */
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
-
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPM_4) )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-ALIGNTEXT32
-LLBL(G3TPPM_2):
-
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPPM_3)              /* skip vertex                        */ )
-
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
-
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-    MOVQ      ( MM5, MM6 )              /*               | x2                 */
+ALIGNTEXT16
+LLBL( G3TPIM_2 ):
 
-    PUNPCKLDQ ( MM5, MM5 )              /* x2            | x2                 */
-    PFMUL     ( MM1, MM5 )              /* x2*m21        | x2*m20             */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPIM_3 ) )	/* skip vertex                       */
 
-    PFADD     ( MM4, MM5 )              /* x1*m11+x2*m21 | x0*m00+x2*m20      */
-    MOVQ      ( MM5, REGIND(EDX) )      /* write r0, r1                       */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM1 )	/*                 | x2              */
 
-    MOVQ      ( MM6, MM5 )              /*               | x2                 */
-    PFMUL     ( MM2, MM5 )              /*               | x2*m22             */
+    MOVQ      ( MM0, REGIND(EDX) )	/* r1              | r0              */
+    MOVD      ( MM1, REGOFF(8, EDX) )	/*                 | r2              */
 
-    PFADD     ( MM3, MM5 )              /*               | x2*m22+m32         */
-    PFSUBR    ( MM7, MM6 )              /* (LO mm7 == 0) | -x2                */
+LLBL( G3TPIM_3 ):
 
-    MOVD      ( MM5, REGOFF(8, EDX) )   /* write r2                           */
-    MOVD      ( MM6, REGOFF(12, EDX) )  /* write r3                           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-LLBL(G3TPPM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TPIM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TPPM_2)              /* cnt > 0 ? -> process next vertex   */ )
+LLBL( G3TPIM_4 ):
 
-LLBL(G3TPPM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
-
-
-
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_masked4.S b/xc/extras/Mesa/src/X86/3dnow_xform_masked4.S
index 6c09efa21..ab496988f 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_masked4.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_masked4.S
@@ -1,63 +1,87 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
 #include "assyntax.h"
+#include "xform_args.h"
 
-   SEG_TEXT
+    SEG_TEXT
 
-ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_general_masked)
-GLNAME( gl_3dnow_transform_points4_general_masked ):
+#define FRAME_OFFSET	16
 
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_general_masked )
+GLNAME( gl_3dnow_transform_points4_general_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    FEMMS
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
 
-    MOVD      ( REGOFF(32, ECX), MM2 )  /*               | m20                */
-    MOVD      ( REGOFF(48, ECX), MM7 )  /*               | m30                */
+    MOVD      ( REGOFF(32, ECX), MM2 )	/*                 | m20             */
+    MOVD      ( REGOFF(48, ECX), MM7 )	/*                 | m30             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m30           |                    */
-    POR       ( MM7, MM2 )              /* m30           | m20                */
+    PSLLQ     ( CONST(32), MM7 )	/* m30             |                 */
+    POR       ( MM7, MM2 )		/* m30             | m20             */
 
-    MOVD      ( REGOFF(36, ECX), MM3 )  /*               | m21                */
-    MOVD      ( REGOFF(52, ECX), MM7 )  /*               | m31                */
+    MOVD      ( REGOFF(36, ECX), MM3 )	/*                 | m21             */
+    MOVD      ( REGOFF(52, ECX), MM7 )	/*                 | m31             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m31           |                    */
-    POR       ( MM7, MM3 )              /* m31           | m21                */
+    PSLLQ     ( CONST(32), MM7 )	/* m31             |                 */
+    POR       ( MM7, MM3 )		/* m31             | m21             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGM_6) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGM_6 ) )
 
     PUSH_L    ( EBP )
     PUSH_L    ( EAX )
@@ -65,697 +89,664 @@ ALIGNTEXT32
     PUSH_L    ( ESI )
 
 ALIGNTEXT32
+LLBL( G3TPGM_2 ):
 
-LLBL(G3TPGM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPGM_3)              /* skip vertex                        */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPGM_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
 
-    MOVQ      ( REGOFF(8, EAX), MM6 )   /* x3            | x2                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m10        | x0*m00             */
+    MOVQ      ( REGOFF(8, EAX), MM6 )	/* x3              | x2              */
+    PFMUL     ( MM0, MM4 )		/* x1*m10          | x0*m00          */
 
-    MOVQ      ( MM6, MM7 )              /* x3            | x2                 */
-    PFMUL     ( MM1, MM5 )              /* x1*m11        | x0*m01             */
+    MOVQ      ( MM6, MM7 )		/* x3              | x2              */
+    PFMUL     ( MM1, MM5 )		/* x1*m11          | x0*m01          */
 
-    PFMUL     ( MM2, MM6 )              /* x3*m30        | x2*m20             */
-    PFACC     ( MM5, MM4 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
+    PFMUL     ( MM2, MM6 )		/* x3*m30          | x2*m20          */
+    PFACC     ( MM5, MM4 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
 
-    PFMUL     ( MM3, MM7 )              /* x3*m31        | x2*m21             */
-    PFACC     ( MM7, MM6 )              /* x2*m21+x3*m31 | x2*m20+x3*m30      */
+    PFMUL     ( MM3, MM7 )		/* x3*m31          | x2*m21          */
+    PFACC     ( MM7, MM6 )		/* x2*m21+x3*m31   | x2*m20+x3*m30   */
 
-    PFADD     ( MM4, MM6 )              /* r1            | r0                 */
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r0, r1                       */
+    PFADD     ( MM4, MM6 )		/* r1              | r0              */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r0, r1                      */
 
-LLBL(G3TPGM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+LLBL( G3TPGM_3 ):
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    JA        ( LLBL(G3TPGM_2)              /* cnt > 0 ? -> process next vertex   */ )
-                                        /* and now the second stripe ...      */
-    MOVD      ( REGOFF(8, ECX), MM0 )   /*               | m02                */
-    MOVD      ( REGOFF(24, ECX), MM7 )  /*               | m12                */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m12           |                    */
-    POR       ( MM7, MM0 )              /* m12           | m02                */
+    JNZ       ( LLBL( G3TPGM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    MOVD      ( REGOFF(12, ECX), MM1 )  /*               | m03                */
-    MOVD      ( REGOFF(28, ECX), MM7 )  /*               | m13                */
+                                        /* and now the second stripe ...     */
+    MOVD      ( REGOFF(8, ECX), MM0 )	/*                 | m02             */
+    MOVD      ( REGOFF(24, ECX), MM7 )	/*                 | m12             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m13           |                    */
-    POR       ( MM7, MM1 )              /* m13           | m03                */
+    PSLLQ     ( CONST(32), MM7 )	/* m12             |                 */
+    POR       ( MM7, MM0 )		/* m12             | m02             */
 
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
-    MOVD      ( REGOFF(56, ECX), MM7 )  /*               | m32                */
+    MOVD      ( REGOFF(12, ECX), MM1 )	/*                 | m03             */
+    MOVD      ( REGOFF(28, ECX), MM7 )	/*                 | m13             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m32           |                    */
-    POR       ( MM7, MM2 )              /* m32           | m22                */
+    PSLLQ     ( CONST(32), MM7 )	/* m13             |                 */
+    POR       ( MM7, MM1 )		/* m13             | m03             */
 
-    MOVD      ( REGOFF(44, ECX), MM3 )  /*               | m23                */
-    MOVD      ( REGOFF(60, ECX), MM7 )  /*               | m33                */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
+    MOVD      ( REGOFF(56, ECX), MM7 )	/*                 | m32             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m33           |                    */
-    POR       ( MM7, MM3 )              /* m33           | m23                */
+    PSLLQ     ( CONST(32), MM7 )	/* m32             |                 */
+    POR       ( MM7, MM2 )		/* m32             | m22             */
 
-    POP_L     ( ESI )                   /* reset counter & pointers           */
+    MOVD      ( REGOFF(44, ECX), MM3 )	/*                 | m23             */
+    MOVD      ( REGOFF(60, ECX), MM7 )	/*                 | m33             */
+
+    PSLLQ     ( CONST(32), MM7 )	/* m33             |                 */
+    POR       ( MM7, MM3 )		/* m33             | m23             */
+
+    POP_L     ( ESI )			/* reset counter & pointers          */
     POP_L     ( EDX )
     POP_L     ( EAX )
     POP_L     ( EBP )
 
 ALIGNTEXT32
+LLBL( G3TPGM_4 ):
 
-LLBL(G3TPGM_4):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPGM_5)              /* skip vertex                        */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPGM_5 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
 
-    MOVQ      ( REGOFF(8, EAX), MM6 )   /* x3            | x2                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m12        | x0*m02             */
+    MOVQ      ( REGOFF(8, EAX), MM6 )	/* x3              | x2              */
+    PFMUL     ( MM0, MM4 )		/* x1*m12          | x0*m02          */
 
-    MOVQ      ( MM6, MM7 )              /* x3            | x2                 */
-    PFMUL     ( MM1, MM5 )              /* x1*m13        | x0*m03             */
+    MOVQ      ( MM6, MM7 )		/* x3              | x2              */
+    PFMUL     ( MM1, MM5 )		/* x1*m13          | x0*m03          */
 
-    PFMUL     ( MM2, MM6 )              /* x3*m32        | x2*m22             */
-    PFACC     ( MM5, MM4 )              /* x0*m03+x1*m13 | x0*m02+x1*m12      */
+    PFMUL     ( MM2, MM6 )		/* x3*m32          | x2*m22          */
+    PFACC     ( MM5, MM4 )		/* x0*m03+x1*m13   | x0*m02+x1*m12   */
 
-    PFMUL     ( MM3, MM7 )              /* x3*m33        | x2*m23             */
-    PFACC     ( MM7, MM6 )              /* x2*m23+x3*m33 | x2*m22+x3*m32      */
+    PFMUL     ( MM3, MM7 )		/* x3*m33          | x2*m23          */
+    PFACC     ( MM7, MM6 )		/* x2*m23+x3*m33   | x2*m22+x3*m32   */
 
-    PFADD     ( MM4, MM6 )              /* r3            | r2                 */
-    MOVQ      ( MM6, REGOFF(8, EDX) )   /* write r2, r3                       */
+    PFADD     ( MM4, MM6 )		/* r3              | r2              */
+    MOVQ      ( MM6, REGOFF(8, EDX) )	/* write r2, r3                      */
 
-LLBL(G3TPGM_5):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+LLBL( G3TPGM_5 ):
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    JA        ( LLBL(G3TPGM_4)              /* cnt > 0 ? -> process next vertex   */ )
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-LLBL(G3TPGM_6):
-    FEMMS
+    JNZ       ( LLBL( G3TPGM_4 ) )	/* cnt > 0 ? -> process next vertex  */
 
+LLBL( G3TPGM_6 ):
+
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_identity_masked)
-GLNAME( gl_3dnow_transform_points4_identity_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points4_perspective_masked )
+GLNAME( gl_3dnow_transform_points4_perspective_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
+
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM0 )		/* m11             | m00             */
+
+    MOVD      ( REGOFF(40, ECX), MM1 )	/*                 | m22             */
+    MOVD      ( REGOFF(56, ECX), MM7 )	/*                 | m32             */
+
+    PSLLQ     ( CONST(32), MM7 )	/* m32             |                 */
+    POR       ( MM7, MM1 )		/* m32             | m22             */
+
+    MOVQ      ( REGOFF(32, ECX), MM2 )	/* m21             | m20             */
+    PUNPCKLDQ ( MM7, MM7 )		/* 0               | 0               */
+
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPM_4 ) )
 
 ALIGNTEXT32
+LLBL( G3TPPM_2 ):
 
-LLBL(G3TPIM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPIM_3)              /* skip vertex                        */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPPM_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                 */
-    MOVQ      ( MM0, REGIND(EDX) )      /* r1            | r0                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    MOVQ      ( REGOFF(8, EAX), MM1 )   /* x3            | x2                 */
-    MOVQ      ( MM1, REGOFF(8, EDX) )   /* r3            | r2                 */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
 
-LLBL(G3TPIM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    PUNPCKLDQ ( MM5, MM5 )		/* x2              | x2              */
+    PFMUL     ( MM1, MM6 )		/* x3*m32          | x2*m22          */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    PFMUL     ( MM2, MM5 )		/* x2*m21          | x2*m20          */
+    PFADD     ( MM4, MM5 )		/* x1*m11+x2*m21   | x0*m00+x2*m20   */
 
-    JA        ( LLBL(G3TPIM_2)              /* cnt > 0 ? -> process next vertex   */ )
+    MOVQ      ( MM5, REGIND(EDX) )	/* write r0, r1                      */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-LLBL(G3TPIM_4):
-    FEMMS
+    PFSUBR    ( MM7, MM5 )		/*                 | -x2             */
+    PFACC     ( MM5, MM6 )		/* -x2             | x2*m22+x3*m32   */
 
+    MOVQ      ( MM6, REGOFF(8, EDX) )	/* write r2, r3                      */
+
+LLBL( G3TPPM_3 ):
+
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TPPM_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TPPM_4 ):
+
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_2d_masked)
-GLNAME( gl_3dnow_transform_points4_2d_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME(gl_3dnow_transform_points4_3d_masked)
+GLNAME( gl_3dnow_transform_points4_3d_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
-
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
-
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2M_4) )
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
 
-ALIGNTEXT32
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
 
-LLBL(G3TP2M_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2M_3)              /* skip vertex                        */ )
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    MOVQ      ( REGIND(EAX), MM3 )      /* x1            | x0                 */
-    MOVQ      ( MM3, MM4 )              /* x1            | x0                 */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
 
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
-    PFMUL     ( MM0, MM3 )              /* x1*m10        | x0*m00             */
+    MOVD      ( REGOFF(32, ECX), MM2 )	/*                 | m20             */
+    MOVD      ( REGOFF(48, ECX), MM7 )	/*                 | m30             */
 
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
-    PFMUL     ( MM1, MM4 )              /* x1*m11        | x0*m01             */
+    PSLLQ     ( CONST(32), MM7 )	/* m30             |                 */
+    POR       ( MM7, MM2 )		/* m30             | m20             */
 
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x3                 */
-    PFACC     ( MM4, MM3 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
+    MOVD      ( REGOFF(36, ECX), MM3 )	/*                 | m21             */
+    MOVD      ( REGOFF(52, ECX), MM7 )	/*                 | m31             */
 
-    PFMUL     ( MM2, MM6 )              /* x3*m31        | x3*m30             */
-    PFADD     ( MM6, MM3 )              /* r1            | r0                 */
+    PSLLQ     ( CONST(32), MM7 )	/* m31             |                 */
+    POR       ( MM7, MM3 )		/* m31             | m21             */
 
-    MOVQ      ( MM3, REGIND(EDX) )      /* write r0, r1                       */
-    MOVQ      ( MM5, REGOFF(8, EDX) )   /* write r2, r3                       */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3M_6 ) )
 
-LLBL(G3TP2M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    PUSH_L    ( EBP )
+    PUSH_L    ( EAX )
+    PUSH_L    ( EDX )
+    PUSH_L    ( ESI )
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+ALIGNTEXT32
+LLBL( G3TP3M_2 ):
 
-    JA        ( LLBL(G3TP2M_2)              /* cnt > 0 ? -> process next vertex   */ )
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3M_3 ) )	/* skip vertex                       */
 
-LLBL(G3TP2M_4):
-    FEMMS
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( MM4, MM5 )		/* x1              | x0              */
 
-    POP_L     ( EBP )
-    POP_L     ( EBX )
-    POP_L     ( EDI )
-    POP_L     ( ESI )
+    PFMUL     ( MM0, MM4 )		/* x1*m10          | x0*m00          */
+    MOVQ      ( REGOFF(8, EAX), MM6 )	/* x3              | x2              */
 
-    POP_L     ( ESI )
-    RET
+    PFMUL     ( MM1, MM5 )		/* x1*m11          | x0*m01          */
+    MOVQ      ( MM6, MM7 )		/* x3              | x2              */
 
+    PFMUL     ( MM2, MM6 )		/* x3*m30          | x2*m20          */
+    PFACC     ( MM5, MM4 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
 
+    PFMUL     ( MM3, MM7 )		/* x3*m31          | x2*m21          */
+    PFACC     ( MM7, MM6 )		/* x2*m21+x3*m31   | x2*m20+x3*m30   */
 
+    PFADD     ( MM4, MM6 )		/* r1              | r0              */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r0, r1                      */
 
-ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_2d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points4_2d_no_rot_masked ):
+LLBL( G3TP3M_3 ):
 
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-ALIGNTEXT32
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    PUSH_L    ( ESI )
-    PUSH_L    ( EDI )
-    PUSH_L    ( EBX )
-    PUSH_L    ( EBP )
+    JNZ       ( LLBL( G3TP3M_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+                                        /* and now the second stripe ...     */
+    MOVD      ( REGOFF(8, ECX), MM0 )	/*                 | m02             */
+    MOVD      ( REGOFF(24, ECX), MM7 )	/*                 | m12             */
 
-    FEMMS
+    PSLLQ     ( CONST(32), MM7 )	/* m12             |                 */
+    POR       ( MM7, MM0 )		/* m12             | m02             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
+    MOVD      ( REGOFF(56, ECX), MM7 )	/*                 | m32             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM0 )              /* m11           | m00                */
+    PSLLQ     ( CONST(32), MM7 )	/* m32             |                 */
+    POR       ( MM7, MM2 )		/* m32             | m22             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRM_4) )
+    POP_L     ( ESI )			/* reset counter & pointers          */
+    POP_L     ( EDX )
+    POP_L     ( EAX )
+    POP_L     ( EBP )
 
 ALIGNTEXT32
+LLBL( G3TP3M_4 ):
 
-LLBL(G3TP2NRM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP2NRM_3) )    /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3M_5 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM6 )	/* x3              | x2              */
 
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+    PFMUL     ( MM0, MM4 )		/* x1*m12          | x0*m02          */
+    MOVD      ( REGOFF(12, EAX), MM7 )	/*                 | x3              */
 
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x2                 */
-    PFMUL     ( MM1, MM6 )              /* x3*m31        | x3*m30             */
+    PFMUL     ( MM2, MM6 )		/* x3*m32          | x2*m22          */
+    PFACC     ( MM4, MM6 )		/* x0*m03+x1*m13   | x0*m02+x1*m12   */
 
-    PFADD     ( MM6, MM4 )              /* x1*m11+x3*m31 | x0*m00+x3*m30      */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r0, r1                       */
+    PFACC     ( MM7, MM6 )		/* x3              | x2*m22+x3*m32   */
+    MOVQ      ( MM6, REGOFF(8, EDX) )	/* write r2, r3                      */
 
-    MOVQ      ( MM5, REGOFF(8, EDX) )   /* write r2, r3                       */
+LLBL( G3TP3M_5 ):
 
-LLBL(G3TP2NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    JA        ( LLBL(G3TP2NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+    JNZ       ( LLBL( G3TP3M_4 ) )	/* cnt > 0 ? -> process next vertex  */
 
-LLBL(G3TP2NRM_4):
-    FEMMS
+LLBL( G3TP3M_6 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
-ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_3d_masked)
-GLNAME( gl_3dnow_transform_points4_3d_masked ):
 
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_3d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points4_3d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
+
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    FEMMS
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(16, ECX), MM7 )  /*               | m10                */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM0 )		/* m11             | m00             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m10           |                    */
-    POR       ( MM7, MM0 )              /* m10           | m00                */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
+    PSLLQ     ( CONST(32), MM3 )	/* m32             |                 */
+    POR       ( MM3, MM2 )		/* m32             | m22             */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM1 )              /* m11           | m01                */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
 
-    MOVD      ( REGOFF(32, ECX), MM2 )  /*               | m20                */
-    MOVD      ( REGOFF(48, ECX), MM7 )  /*               | m30                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRM_4 ) )
 
-    PSLLQ     ( CONST(32), MM7 )        /* m30           |                    */
-    POR       ( MM7, MM2 )              /* m30           | m20                */
+ALIGNTEXT32
+LLBL( G3TP3NRM_2 ):
 
-    MOVD      ( REGOFF(36, ECX), MM3 )  /*               | m21                */
-    MOVD      ( REGOFF(52, ECX), MM7 )  /*               | m31                */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP3NRM_3 ) )	/* skip vertex                       */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m31           |                    */
-    POR       ( MM7, MM3 )              /* m31           | m21                */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3M_6) )
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    PUSH_L    ( EBP )
-    PUSH_L    ( EAX )
-    PUSH_L    ( EDX )
-    PUSH_L    ( ESI )
+    MOVD      ( REGOFF(12, EAX), MM7 )	/*                 | x3              */
+    PFMUL     ( MM2, MM5 )		/* x3*m32          | x2*m22          */
 
-ALIGNTEXT32
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x3              */
+    PFACC     ( MM7, MM5 )		/* x3              | x2*m22+x3*m32   */
 
-LLBL(G3TP3M_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3M_3) )      /* skip vertex                        */
+    PFMUL     ( MM1, MM6 )		/* x3*m31          | x3*m30          */
+    PFADD     ( MM6, MM4 )		/* x1*m11+x3*m31   | x0*m00+x3*m30   */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( MM4, MM5 )              /* x1            | x0                 */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r0, r1                      */
+    MOVQ      ( MM5, REGOFF(8, EDX) )	/* write r2, r3                      */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m10        | x0*m00             */
-    MOVQ      ( REGOFF(8, EAX), MM6 )   /* x3            | x2                 */
+LLBL( G3TP3NRM_3 ):
 
-    PFMUL     ( MM1, MM5 )              /* x1*m11        | x0*m01             */
-    MOVQ      ( MM6, MM7 )              /* x3            | x2                 */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    PFMUL     ( MM2, MM6 )              /* x3*m30        | x2*m20             */
-    PFACC     ( MM5, MM4 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    PFMUL     ( MM3, MM7 )              /* x3*m31        | x2*m21             */
-    PFACC     ( MM7, MM6 )              /* x2*m21+x3*m31 | x2*m20+x3*m30      */
+    JNZ       ( LLBL( G3TP3NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    PFADD     ( MM4, MM6 )              /* r1            | r0                 */
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r0, r1                       */
+LLBL( G3TP3NRM_4 ):
 
-LLBL(G3TP3M_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    FEMMS
+    POP_L     ( EBP )
+    POP_L     ( EBX )
+    POP_L     ( EDI )
+    POP_L     ( ESI )
+    RET
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
 
-    JA        ( LLBL(G3TP3M_2) )      /* cnt > 0 ? -> process next vertex   */
-                                        /* and now the second stripe ...      */
-    MOVD      ( REGOFF(8, ECX), MM0 )   /*               | m02                */
-    MOVD      ( REGOFF(24, ECX), MM7 )  /*               | m12                */
 
-    PSLLQ     ( CONST(32), MM7 )        /* m12           |                    */
-    POR       ( MM7, MM0 )              /* m12           | m02                */
 
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
-    MOVD      ( REGOFF(56, ECX), MM7 )  /*               | m32                */
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_2d_masked )
+GLNAME( gl_3dnow_transform_points4_2d_masked ):
 
-    PSLLQ     ( CONST(32), MM7 )        /* m32           |                    */
-    POR       ( MM7, MM2 )              /* m32           | m22                */
+    PUSH_L    ( ESI )
+    PUSH_L    ( EDI )
+    PUSH_L    ( EBX )
+    PUSH_L    ( EBP )
 
-    POP_L     ( ESI )                   /* reset counter & pointers           */
-    POP_L     ( EDX )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    POP_L     ( EAX )
-    POP_L     ( EBP )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
+
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(16, ECX), MM7 )	/*                 | m10             */
+
+    PSLLQ     ( CONST(32), MM7 )	/* m10             |                 */
+    POR       ( MM7, MM0 )		/* m10             | m00             */
+
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
+
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM1 )		/* m11             | m01             */
+
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2M_4 ) )
 
 ALIGNTEXT32
+LLBL( G3TP2M_2 ):
 
-LLBL(G3TP3M_4):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3M_5) )      /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2M_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM6 )   /* x3            | x2                 */
+    MOVQ      ( REGIND(EAX), MM3 )	/* x1              | x0              */
+    MOVQ      ( MM3, MM4 )		/* x1              | x0              */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m12        | x0*m02             */
-    MOVD      ( REGOFF(12, EAX), MM7 )  /*               | x3                 */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
+    PFMUL     ( MM0, MM3 )		/* x1*m10          | x0*m00          */
 
-    PFMUL     ( MM2, MM6 )              /* x3*m32        | x2*m22             */
-    PFACC     ( MM4, MM6 )              /* x0*m03+x1*m13 | x0*m02+x1*m12      */
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
+    PFMUL     ( MM1, MM4 )		/* x1*m11          | x0*m01          */
 
-    PFACC     ( MM7, MM6 )              /* x3            | x2*m22+x3*m32      */
-    MOVQ      ( MM6, REGOFF(8, EDX) )   /* write r2, r3                       */
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x3              */
+    PFACC     ( MM4, MM3 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
 
-LLBL(G3TP3M_5):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    PFMUL     ( MM2, MM6 )		/* x3*m31          | x3*m30          */
+    PFADD     ( MM6, MM3 )		/* r1              | r0              */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVQ      ( MM3, REGIND(EDX) )	/* write r0, r1                      */
+    MOVQ      ( MM5, REGOFF(8, EDX) )	/* write r2, r3                      */
 
-    JA        ( LLBL(G3TP3M_4) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP2M_3 ):
 
-LLBL(G3TP3M_6):
-    FEMMS
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TP2M_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
+LLBL( G3TP2M_4 ):
+
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_3d_no_rot_masked)
-GLNAME( gl_3dnow_transform_points4_3d_no_rot_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points4_2d_no_rot_masked )
+GLNAME( gl_3dnow_transform_points4_2d_no_rot_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    FEMMS
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
-
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM0 )              /* m11           | m00                */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(20, ECX), MM7 )	/*                 | m11             */
 
-    PSLLQ     ( CONST(32), MM3 )        /* m32           |                    */
-    POR       ( MM3, MM2 )              /* m32           | m22                */
+    PSLLQ     ( CONST(32), MM7 )	/* m11             |                 */
+    POR       ( MM7, MM0 )		/* m11             | m00             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRM_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRM_4 ) )
 
 ALIGNTEXT32
+LLBL( G3TP2NRM_2 ):
 
-LLBL(G3TP3NRM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TP3NRM_3) )    /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TP2NRM_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    MOVD      ( REGOFF(12, EAX), MM7 )  /*               | x3                 */
-    PFMUL     ( MM2, MM5 )              /* x3*m32        | x2*m22             */
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x2              */
+    PFMUL     ( MM1, MM6 )		/* x3*m31          | x3*m30          */
 
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x3                 */
-    PFACC     ( MM7, MM5 )              /* x3            | x2*m22+x3*m32      */
+    PFADD     ( MM6, MM4 )		/* x1*m11+x3*m31   | x0*m00+x3*m30   */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r0, r1                      */
 
-    PFMUL     ( MM1, MM6 )              /* x3*m31        | x3*m30             */
-    PFADD     ( MM6, MM4 )              /* x1*m11+x3*m31 | x0*m00+x3*m30      */
+    MOVQ      ( MM5, REGOFF(8, EDX) )	/* write r2, r3                      */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r0, r1                       */
-    MOVQ      ( MM5, REGOFF(8, EDX) )   /* write r2, r3                       */
+LLBL( G3TP2NRM_3 ):
 
-LLBL(G3TP3NRM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    JA        ( LLBL(G3TP3NRM_2)            /* cnt > 0 ? -> process next vertex   */ )
+    JNZ       ( LLBL( G3TP2NRM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-LLBL(G3TP3NRM_4):
-    FEMMS
+LLBL( G3TP2NRM_4 ):
 
+    FEMMS
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
 
 
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points4_perspective_masked)
-GLNAME( gl_3dnow_transform_points4_perspective_masked ):
-
-    PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
-
-ALIGNTEXT32
+GLOBL GLNAME( gl_3dnow_transform_points4_identity_masked )
+GLNAME( gl_3dnow_transform_points4_identity_masked ):
 
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
     PUSH_L    ( EBX )
     PUSH_L    ( EBP )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
-    MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )   /* count                            */
-    MOV_L     ( REGOFF(12, EAX), EDI )  /* input stride                     */
-    MOV_L     ( REGOFF(4, EAX), EAX )
-    MOV_L     ( REGOFF(36, ESP), EBP )  /* clipmask                         */
-    MOV_B     ( REGOFF(40, ESP), BL )   /* clip flag                        */
-
-    FEMMS
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(20, ECX), MM7 )  /*               | m11                */
-
-    PSLLQ     ( CONST(32), MM7 )        /* m11           |                    */
-    POR       ( MM7, MM0 )              /* m11           | m00                */
-
-    MOVD      ( REGOFF(40, ECX), MM1 )  /*               | m22                */
-    MOVD      ( REGOFF(56, ECX), MM7 )  /*               | m32                */
-
-    PSLLQ     ( CONST(32), MM7 )        /* m32           |                    */
-    POR       ( MM7, MM1 )              /* m32           | m22                */
-
-    MOVQ      ( REGOFF(32, ECX), MM2 )  /* m21           | m20                */
-    PUNPCKLDQ ( MM7, MM7 )              /* 0             | 0                  */
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
+    MOV_L     ( ESI, ECX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
+    MOV_L     ( ARG_CLIP, EBP )
+    MOV_B     ( ARG_FLAG, BL )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPM_4) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPIM_4 ) )
 
 ALIGNTEXT32
+LLBL( G3TPIM_2 ):
 
-LLBL(G3TPPM_2):
-    TEST_B    ( BL, REGIND(EBP) )       /* mask [i] != clip flag ??           */
-    JNZ       ( LLBL(G3TPPM_3) )      /* skip vertex                        */
+    TEST_B    ( BL, REGIND(EBP) )	/* mask [i] != clip flag ??          */
+    JNZ       ( LLBL( G3TPIM_3 ) )	/* skip vertex                       */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVQ      ( MM0, REGIND(EDX) )	/* r1              | r0              */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
+    MOVQ      ( REGOFF(8, EAX), MM1 )	/* x3              | x2              */
+    MOVQ      ( MM1, REGOFF(8, EDX) )	/* r3              | r2              */
 
-    PUNPCKLDQ ( MM5, MM5 )              /* x2            | x2                 */
-    PFMUL     ( MM1, MM6 )              /* x3*m32        | x2*m22             */
+LLBL( G3TPIM_3 ):
 
-    PFMUL     ( MM2, MM5 )              /* x2*m21        | x2*m20             */
-    PFADD     ( MM4, MM5 )              /* x1*m11+x2*m21 | x0*m00+x2*m20      */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVQ      ( MM5, REGIND(EDX) )      /* write r0, r1                       */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
+    INC_L     ( EBP )			/* next clipmask                     */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    PFSUBR    ( MM7, MM5 )              /*               | -x2                */
-    PFACC     ( MM5, MM6 )              /* -x2           | x2*m22+x3*m32      */
+    JNZ       ( LLBL( G3TPIM_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    MOVQ      ( MM6, REGOFF(8, EDX) )   /* write r2, r3                       */
+LLBL( G3TPIM_4 ):
 
-LLBL(G3TPPM_3):
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-
-    INC_L     ( EBP )                   /* next clipmask                      */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-
-    JA        ( LLBL(G3TPPM_2) )      /* cnt > 0 ? -> process next vertex   */
-
-LLBL(G3TPPM_4):
     FEMMS
-
     POP_L     ( EBP )
     POP_L     ( EBX )
     POP_L     ( EDI )
     POP_L     ( ESI )
-
-    POP_L     ( ESI )
     RET
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_raw1.S b/xc/extras/Mesa/src/X86/3dnow_xform_raw1.S
index 4095fe558..5f097ff87 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_raw1.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_raw1.S
@@ -1,125 +1,141 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
 #include "assyntax.h"
+#include "xform_args.h"
+
+    SEG_TEXT
 
-   SEG_TEXT
+#define FRAME_OFFSET	4
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_general_raw)
-GLNAME(gl_3dnow_transform_points1_general_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_general_raw )
+GLNAME( gl_3dnow_transform_points1_general_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    FEMMS
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGR_3 ) )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGR_3) )
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVQ      ( REGOFF(8, ECX), MM1 )	/* m03             | m02             */
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVQ      ( REGOFF(8, ECX), MM1 )   /* m03           | m02                */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVQ      ( REGOFF(56, ECX), MM3 )	/* m33             | m32             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVQ      ( REGOFF(56, ECX), MM3 )  /* m33           | m32                */
+ALIGNTEXT16
+LLBL( G3TPGR_2 ):
 
-ALIGNTEXT32
-LLBL(G3TPGR_2):
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
+    MOVQ      ( MM4, MM5 )		/* x0              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
 
-    MOVQ      ( MM4, MM5 )              /* x0            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
+    PFMUL     ( MM1, MM5 )		/* x0*m03          | x0*m02          */
+    PFADD     ( MM2, MM4 )		/* x0*m01+m31      | x0*m00+m30      */
 
-    PFMUL     ( MM1, MM5 )              /* x0*m03        | x0*m02             */
-    PFADD     ( MM2, MM4 )              /* x0*m01+m31    | x0*m00+m30         */
+    PFADD     ( MM3, MM5 )		/* x0*m03+m33      | x0*m02+m32      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFADD     ( MM3, MM5 )              /* x0*m03+m33    | x0*m02+m32         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    MOVQ      ( MM5, REGOFF(8, EDX) )	/* write r3, r2                      */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVQ      ( MM5, REGOFF(8, EDX) )   /* write r3, r2                       */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TPGR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TPGR_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPGR_3 ):
 
-LLBL(G3TPGR_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_identity_raw)
-GLNAME(gl_3dnow_transform_points1_identity_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_identity_raw )
+GLNAME( gl_3dnow_transform_points1_identity_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(1), REGOFF(16, ECX) )
-    OR_B      ( CONST(1), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(1), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_1), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    FEMMS
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPIR_4) )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPIR_4) )
+ALIGNTEXT16
+LLBL( G3TPIR_3 ):
 
-LLBL(G3TPIR_3):
+    MOVD      ( REGIND(EAX), MM0 )	/*                 | x0              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVD      ( REGIND(EAX), MM0 )      /*               | x0                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVD      ( MM0, REGIND(EDX) )	/*                 | r0              */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVD      ( MM0, REGIND(EDX) )      /*               | r0                 */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPIR_3 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPIR_3) )        /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPIR_4 ):
 
-LLBL(G3TPIR_4):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
@@ -127,322 +143,280 @@ LLBL(G3TPIR_4):
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_2d_raw)
-GLNAME(gl_3dnow_transform_points1_2d_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_3d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points1_3d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    FEMMS
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRR_3 ) )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2R_3) )
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
+ALIGNTEXT16
+LLBL( G3TP3NRR_2 ):
 
-ALIGNTEXT32
-LLBL(G3TP2R_2):
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PFMUL     ( MM0, MM4 )		/*                 | x0*m00          */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
+    PFADD     ( MM2, MM4 )		/* m31             | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
-    PFADD     ( MM2, MM4 )              /* x0*m01+m31    | x0*m00+m30         */
+    MOVD      ( MM3, REGOFF(8, EDX) )	/* write r2                          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TP3NRR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP2R_2) )        /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP3NRR_3 ):
 
-LLBL(G3TP2R_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_2d_no_rot_raw)
-GLNAME(gl_3dnow_transform_points1_2d_no_rot_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_perspective_raw )
+GLNAME( gl_3dnow_transform_points1_perspective_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRR_3) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPR_3 ) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
+ALIGNTEXT16
+LLBL( G3TPPR_2 ):
 
-ALIGNTEXT32
-LLBL(G3TP2NRR_2):
+    MOVD      ( REGIND(EAX), MM4 )	/* 0               | x0              */
+    PFMUL     ( MM0, MM4 )		/* 0               | x0*m00          */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( MM3, REGOFF(8, EDX) )	/* write r2  (=m32), r3 (=0)         */
 
-    PFMUL     ( MM0, MM4 )              /*               | x0*m00             */
-    PFADD     ( MM2, MM4 )              /* m31           | x0*m00+m30         */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPPR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP2NRR_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPPR_3 ):
 
-LLBL(G3TP2NRR_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_3d_raw)
-GLNAME(gl_3dnow_transform_points1_3d_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_2d_raw )
+GLNAME( gl_3dnow_transform_points1_2d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3R_3) )
-
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVD      ( REGOFF(8, ECX), MM1 )   /*               | m02                */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2R_3 ) )
 
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-ALIGNTEXT32
-LLBL(G3TP3R_2):
-
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
+ALIGNTEXT16
+LLBL( G3TP2R_2 ):
 
-    MOVQ      ( MM4, MM5 )              /*               | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
 
-    PFMUL     ( MM1, MM5 )              /*               | x0*m02             */
-    PFADD     ( MM2, MM4 )              /* x0*m01+m31    | x0*m00+m30         */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
+    PFADD     ( MM2, MM4 )		/* x0*m01+m31      | x0*m00+m30      */
 
-    PFADD     ( MM3, MM5 )              /*               | x0*m02+m32         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVD      ( MM5, REGOFF(8, EDX) )   /* write r2                           */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TP2R_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP3R_2) )        /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP2R_3 ):
 
-LLBL(G3TP3R_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_3d_no_rot_raw)
-GLNAME(gl_3dnow_transform_points1_3d_no_rot_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_2d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points1_2d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRR_3) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRR_3 ) )
 
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-ALIGNTEXT32
-LLBL(G3TP3NRR_2):
+ALIGNTEXT16
+LLBL( G3TP2NRR_2 ):
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    PFMUL     ( MM0, MM4 )              /*               | x0*m00             */
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PFADD     ( MM2, MM4 )              /* m31           | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    PFMUL     ( MM0, MM4 )		/*                 | x0*m00          */
+    PFADD     ( MM2, MM4 )		/* m31             | x0*m00+m30      */
 
-    MOVD      ( MM3, REGOFF(8, EDX) )   /* write r2                           */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2NRR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP3NRR_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP2NRR_3 ):
 
-LLBL(G3TP3NRR_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points1_perspective_raw)
-GLNAME(gl_3dnow_transform_points1_perspective_raw):
+GLOBL GLNAME( gl_3dnow_transform_points1_3d_raw )
+GLNAME( gl_3dnow_transform_points1_3d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
     MOV_L     ( REGOFF(4, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPR_3) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3R_3 ) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVD      ( REGOFF(8, ECX), MM1 )	/*                 | m02             */
 
-ALIGNTEXT32
-LLBL(G3TPPR_2):
-    MOVD      ( REGIND(EAX), MM4 )      /* 0             | x0                 */
-    PFMUL     ( MM0, MM4 )              /* 0             | x0*m00             */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( MM3, REGOFF(8, EDX) )   /* write r2  (=m32), r3 (=0)          */
-
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPPR_2) )        /* cnt > 0 ? -> process next vertex   */
+ALIGNTEXT16
+LLBL( G3TP3R_2 ):
 
-LLBL(G3TPPR_3):
-    FEMMS
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
 
-    POP_L     ( ESI )
-    POP_L     ( EDI )
+    MOVQ      ( MM4, MM5 )		/*                 | x0              */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
 
-    POP_L     ( ESI )
-    RET
+    PFMUL     ( MM1, MM5 )		/*                 | x0*m02          */
+    PFADD     ( MM2, MM4 )		/* x0*m01+m31      | x0*m00+m30      */
 
+    PFADD     ( MM3, MM5 )		/*                 | x0*m02+m32      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
+    MOVD      ( MM5, REGOFF(8, EDX) )	/* write r2                          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
+    JNZ       ( LLBL( G3TP3R_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
+LLBL( G3TP3R_3 ):
 
+    FEMMS
+    POP_L     ( EDI )
+    POP_L     ( ESI )
+    RET
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_raw2.S b/xc/extras/Mesa/src/X86/3dnow_xform_raw2.S
index 5355a16b9..1d785e4fd 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_raw2.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_raw2.S
@@ -1,142 +1,165 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
 #include "assyntax.h"
+#include "xform_args.h"
+
+    SEG_TEXT
 
-   SEG_TEXT
+#define FRAME_OFFSET	4
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_general_raw)
-GLNAME(gl_3dnow_transform_points2_general_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_general_raw )
+GLNAME( gl_3dnow_transform_points2_general_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGR_3) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGR_3 ) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )  /* m10           | m00                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )	/* m10             | m00             */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )  /* m11           | m01                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )	/* m11             | m01             */
 
-    MOVD      ( REGOFF(8, ECX), MM2 )   /*               | m02                */
-    PUNPCKLDQ ( REGOFF(24, ECX), MM2 )  /* m12           | m02                */
+    MOVD      ( REGOFF(8, ECX), MM2 )	/*                 | m02             */
+    PUNPCKLDQ ( REGOFF(24, ECX), MM2 )	/* m12             | m02             */
 
-    MOVD      ( REGOFF(12, ECX), MM3 )  /*               | m03                */
-    PUNPCKLDQ ( REGOFF(28, ECX), MM3 )  /* m13           | m03                */
+    MOVD      ( REGOFF(12, ECX), MM3 )	/*                 | m03             */
+    PUNPCKLDQ ( REGOFF(28, ECX), MM3 )	/* m13             | m03             */
 
-    MOVQ      ( REGOFF(48, ECX), MM4 )  /* m31           | m30                */
-    MOVQ      ( REGOFF(56, ECX), MM5 )  /* m33           | m32                */
+    MOVQ      ( REGOFF(48, ECX), MM4 )	/* m31             | m30             */
+    MOVQ      ( REGOFF(56, ECX), MM5 )	/* m33             | m32             */
 
+ALIGNTEXT16
+LLBL( G3TPGR_2 ):
 
-ALIGNTEXT32
-LLBL(G3TPGR_2):
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
 
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
+    PFMUL     ( MM0, MM6 )		/* x1*m10          | x0*m00          */
+    PFMUL     ( MM1, MM7 )		/* x1*m11          | x0*m01          */
 
-    PFMUL     ( MM0, MM6 )              /* x1*m10        | x0*m00             */
-    PFMUL     ( MM1, MM7 )              /* x1*m11        | x0*m01             */
+    PFACC     ( MM7, MM6 )		/* x0*m01+x1*m11   | x0*x00+x1*m10   */
+    PFADD     ( MM4, MM6 )		/* x0*...*m11+m31  | x0*...*m10+m30  */
 
-    PFACC     ( MM7, MM6 )              /* x0*m01+x1*m11 | x0*x00+x1*m10      */
-    PFADD     ( MM4, MM6 )              /* x0*...*m11+m31| x0*x00+x1*m10+m30  */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
 
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
+    PFMUL     ( MM2, MM6 )		/* x1*m12          | x0*m02          */
 
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
-    PFMUL     ( MM2, MM6 )              /* x1*m12        | x0*m02             */
+    PFMUL     ( MM3, MM7 )		/* x1*m13          | x0*m03          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PFMUL     ( MM3, MM7 )              /* x1*m13        | x0*m03             */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    PFACC     ( MM7, MM6 )		/* x0*m03+x1*m13   | x0*x02+x1*m12   */
+    PFADD     ( MM5, MM6 )		/* x0*...*m13+m33  | x0*...*m12+m32  */
 
-    PFACC     ( MM7, MM6 )              /* x0*m03+x1*m13 | x0*x02+x1*m12      */
-    PFADD     ( MM5, MM6 )              /* x0*...*m13+m33| x0*x02+x1*m12+m32  */
+    MOVQ      ( MM6, REGOFF(8, EDX) )	/* write r3, r2                      */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVQ      ( MM6, REGOFF(8, EDX) )   /* write r3, r2                       */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPGR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPGR_2) )        /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPGR_3 ):
 
-LLBL(G3TPGR_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_identity_raw)
-GLNAME(gl_3dnow_transform_points2_identity_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_perspective_raw )
+GLNAME( gl_3dnow_transform_points2_perspective_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    FEMMS
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPR_3 ) )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPIR_3) )
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-LLBL(G3TPIR_3):
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( MM0, REGIND(EDX) )      /* r1            | r0                 */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+ALIGNTEXT16
+LLBL( G3TPPR_2 ):
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPIR_3) )        /* cnt > 0 ? -> process next vertex   */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-LLBL(G3TPIR_4):
-    FEMMS
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( MM3, REGOFF(8, EDX) )	/* write r2  (=m32), r3 (=0)         */
 
-    POP_L     ( ESI )
-    POP_L     ( EDI )
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPPR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
+LLBL( G3TPPR_3 ):
+
+    FEMMS
+    POP_L     ( EDI )
     POP_L     ( ESI )
     RET
 
@@ -144,68 +167,75 @@ LLBL(G3TPIR_4):
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_2d_raw)
-GLNAME(gl_3dnow_transform_points2_2d_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_3d_raw )
+GLNAME( gl_3dnow_transform_points2_3d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3 ), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    FEMMS
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3R_3 ) )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2R_3) )
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )	/* m10             | m00             */
 
-    MOVQ      ( REGIND(ECX), MM0 )      /* m01           | m00                */
-    MOVQ      ( REGOFF(16, ECX), MM1 )  /* m11           | m10                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )	/* m11             | m01             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    MOVD      ( REGOFF(8, ECX), MM2 )	/*                 | m02             */
+    PUNPCKLDQ ( REGOFF(24, ECX), MM2 )	/* m12             | m02             */
 
-ALIGNTEXT32
-LLBL(G3TP2R_2):
+    MOVQ      ( REGOFF(48, ECX), MM4 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM5 )	/*                 | m32             */
 
-    MOVD      ( REGIND(EAX), MM4 )      /*               | x0                 */
-    MOVD      ( REGOFF(4, EAX), MM5 )   /*               | x1                 */
+ALIGNTEXT16
+LLBL( G3TP3R_2 ):
 
-    PUNPCKLDQ ( MM4, MM4 )              /* x0            | x0                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
 
-    PFMUL     ( MM0, MM4 )              /* x0*m01        | x0*m00             */
-    PUNPCKLDQ ( MM5, MM5 )              /* x1            | x1                 */
+    PFMUL     ( MM0, MM6 )		/* x1*m10          | x0*m00          */
+    PFMUL     ( MM1, MM7 )		/* x1*m11          | x0*m01          */
 
-    PFMUL     ( MM1, MM5 )              /* x1*m11        | x1*m10             */
-    PFADD     ( MM2, MM4 )              /* x...x1*m11+31 | x0*m00+x1*m10+m30  */
+    PFACC     ( MM7, MM6 )		/* x0*m01+x1*m11   | x0*x00+x1*m10   */
+    PFADD     ( MM4, MM6 )		/* x0*...*m11+m31  | x0*...*m10+m30  */
 
-    PFADD     ( MM5, MM4 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    MOVQ      ( MM6, REGIND(EDX) )	/* write r1, r0                      */
+    MOVQ      ( REGIND(EAX), MM6 )	/* x1              | x0              */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVQ      ( MM6, MM7 )		/* x1              | x0              */
+    PFMUL     ( MM2, MM6 )		/* x1*m12          | x0*m02          */
 
-    JA        ( LLBL(G3TP2R_2) )        /* cnt > 0 ? -> process next vertex   */
+    PFACC     ( MM7, MM6 )		/* ***trash***     | x0*x02+x1*m12   */
+    PFADD     ( MM5, MM6 )		/* ***trash***     | x0*...*m12+m32  */
 
-LLBL(G3TP2R_3):
-    FEMMS
+    MOVD      ( MM6, REGOFF(8, EDX) )	/* write r2                          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    POP_L     ( ESI )
-    POP_L     ( EDI )
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
+    JNZ       ( LLBL( G3TP3R_2 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TP3R_3 ):
+
+    FEMMS
+    POP_L     ( EDI )
     POP_L     ( ESI )
     RET
 
@@ -213,144 +243,120 @@ LLBL(G3TP2R_3):
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_2d_no_rot_raw)
-GLNAME(gl_3dnow_transform_points2_2d_no_rot_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_3d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points2_3d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(2), REGOFF(16, ECX) )
-    OR_B      ( CONST(3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3 ), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    FEMMS
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRR_3 ) )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRR_3) )
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+ALIGNTEXT16
+LLBL( G3TP3NRR_2 ):
 
-ALIGNTEXT32
-LLBL(G3TP2NRR_2):
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    PFADD     ( MM2, MM4 )		/* x1*m11+m31      | x0*m00+m30      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-    PFADD     ( MM2, MM4 )              /* m31           | x0*m00+m30         */
+    MOVD      ( MM3, REGOFF(8, EDX) )	/* write r2                          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP2NRR_2) )      /* cnt > 0 ? -> process next vertex   */
+    JNZ       ( LLBL( G3TP3NRR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-LLBL(G3TP2NRR_3):
-    FEMMS
+LLBL( G3TP3NRR_3 ):
 
-    POP_L     ( ESI )
+    FEMMS
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_3d_raw)
-GLNAME(gl_3dnow_transform_points2_3d_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_2d_raw )
+GLNAME( gl_3dnow_transform_points2_2d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3R_3) )
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )  /* m10           | m00                */
-
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )  /* m11           | m01                */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVD      ( REGOFF(8, ECX), MM2 )   /*               | m02                */
-    PUNPCKLDQ ( REGOFF(24, ECX), MM2 )  /* m12           | m02                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2R_3 ) )
 
-    MOVQ      ( REGOFF(48, ECX), MM4 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM5 )  /*               | m32                */
+    MOVQ      ( REGIND(ECX), MM0 )	/* m01             | m00             */
+    MOVQ      ( REGOFF(16, ECX), MM1 )	/* m11             | m10             */
 
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-ALIGNTEXT32
-LLBL(G3TP3R_2):
-
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
+ALIGNTEXT16
+LLBL( G3TP2R_2 ):
 
-    PFMUL     ( MM0, MM6 )              /* x1*m10        | x0*m00             */
-    PFMUL     ( MM1, MM7 )              /* x1*m11        | x0*m01             */
+    MOVD      ( REGIND(EAX), MM4 )	/*                 | x0              */
+    MOVD      ( REGOFF(4, EAX), MM5 )	/*                 | x1              */
 
-    PFACC     ( MM7, MM6 )              /* x0*m01+x1*m11 | x0*x00+x1*m10      */
-    PFADD     ( MM4, MM6 )              /* x0*...*m11+m31| x0*x00+x1*m10+m30  */
+    PUNPCKLDQ ( MM4, MM4 )		/* x0              | x0              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVQ      ( MM6, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( REGIND(EAX), MM6 )      /* x1            | x0                 */
+    PFMUL     ( MM0, MM4 )		/* x0*m01          | x0*m00          */
+    PUNPCKLDQ ( MM5, MM5 )		/* x1              | x1              */
 
-    MOVQ      ( MM6, MM7 )              /* x1            | x0                 */
-    PFMUL     ( MM2, MM6 )              /* x1*m12        | x0*m02             */
+    PFMUL     ( MM1, MM5 )		/* x1*m11          | x1*m10          */
+    PFADD     ( MM2, MM4 )		/* x...x1*m11+31   | x0*..*m10+m30   */
 
-    PFACC     ( MM7, MM6 )              /* ***trash***   | x0*x02+x1*m12      */
-    PFADD     ( MM5, MM6 )              /* ***trash***   | x0*x02+x1*m12+m32  */
+    PFADD     ( MM5, MM4 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
 
-    MOVD      ( MM6, REGOFF(8, EDX) )   /* write r2                           */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    JNZ       ( LLBL( G3TP2R_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP3R_2) )        /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP2R_3 ):
 
-LLBL(G3TP3R_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
@@ -358,129 +364,100 @@ LLBL(G3TP3R_3):
 
 
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_3d_no_rot_raw)
-GLNAME(gl_3dnow_transform_points2_3d_no_rot_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_2d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points2_2d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(7), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRR_3) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRR_3 ) )
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-ALIGNTEXT32
-LLBL(G3TP3NRR_2):
+ALIGNTEXT16
+LLBL( G3TP2NRR_2 ):
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PFADD     ( MM2, MM4 )              /* x1*m11+m31    | x0*m00+m30         */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    PFADD     ( MM2, MM4 )		/* m31             | x0*m00+m30      */
 
-    MOVD      ( MM3, REGOFF(8, EDX) )   /* write r2                           */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r1, r0                      */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2NRR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    JA        ( LLBL(G3TP3NRR_2) )      /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TP2NRR_3 ):
 
-LLBL(G3TP3NRR_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
 
 
 
 
-
 ALIGNTEXT16
-GLOBL GLNAME(gl_3dnow_transform_points2_perspective_raw)
-GLNAME(gl_3dnow_transform_points2_perspective_raw):
+GLOBL GLNAME( gl_3dnow_transform_points2_identity_raw )
+GLNAME( gl_3dnow_transform_points2_identity_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(15), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
 
-ALIGNTEXT32
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(2), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
-    PUSH_L    ( ESI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    FEMMS
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPR_3) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPIR_3 ) )
 
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
-
-ALIGNTEXT32
-LLBL(G3TPPR_2):
+ALIGNTEXT16
+LLBL( G3TPIR_3 ):
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r1, r0                       */
-    MOVQ      ( MM3, REGOFF(8, EDX) )   /* write r2  (=m32), r3 (=0)          */
+    MOVQ      ( MM0, REGIND(EDX) )	/* r1              | r0              */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPIR_3 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPPR_2) )        /* cnt > 0 ? -> process next vertex   */
+LLBL( G3TPIR_4 ):
 
-LLBL(G3TPPR_3):
     FEMMS
-
-    POP_L     ( ESI )
     POP_L     ( EDI )
-
     POP_L     ( ESI )
     RET
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_raw3.S b/xc/extras/Mesa/src/X86/3dnow_xform_raw3.S
index dc4105d9f..5dd2e2294 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_raw3.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_raw3.S
@@ -1,96 +1,115 @@
-#include "assyntax.h"
-
-   SEG_TEXT
 
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
 
-#define  VEC_SIZE_1   1
-#define  VEC_SIZE_2   3
-#define  VEC_SIZE_3   7
-#define  VEC_SIZE_4   15
+#include "assyntax.h"
+#include "xform_args.h"
 
+    SEG_TEXT
 
+#define FRAME_OFFSET	4
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_general_raw)
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_general_raw )
 GLNAME( gl_3dnow_transform_points3_general_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGR_2 ) )
 
     PREFETCHW ( REGIND(EDX) )
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
-    MOVD      ( REGOFF(8, EAX), MM2 )   /*               | x2                */
-
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM2 )	/*                 | x2              */
 
-ALIGNTEXT32
-LLBL(G3TPGR_1):
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+ALIGNTEXT16
+LLBL( G3TPGR_1 ):
 
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
     PREFETCH  ( REGIND(EAX) )
 
-    MOVQ      ( MM0, MM1 )              /* x1            | x0                */
-    PUNPCKLDQ ( MM2, MM2 )              /* x2            | x2                */
+    MOVQ      ( MM0, MM1 )		/* x1              | x0              */
+    PUNPCKLDQ ( MM2, MM2 )		/* x2              | x2              */
 
-    PUNPCKLDQ ( MM0, MM0 )              /* x0            | x0                */
-    MOVQ      ( MM2, MM5 )              /* x2            | x2                */
+    PUNPCKLDQ ( MM0, MM0 )		/* x0              | x0              */
+    MOVQ      ( MM2, MM5 )		/* x2              | x2              */
 
-    PUNPCKHDQ ( MM1, MM1 )              /* x1            | x1                */
-    PFMUL     ( REGOFF(32, ECX), MM2 )  /* x2*m9         | x2*m8             */
+    PUNPCKHDQ ( MM1, MM1 )		/* x1              | x1              */
+    PFMUL     ( REGOFF(32, ECX), MM2 )	/* x2*m9           | x2*m8           */
 
-    MOVQ      ( MM0, MM3 )              /* x0            | x0                */
-    PFMUL     ( REGOFF(40, ECX), MM5 )  /* x2*m11        | x2*m10            */
+    MOVQ      ( MM0, MM3 )		/* x0              | x0              */
+    PFMUL     ( REGOFF(40, ECX), MM5 )	/* x2*m11          | x2*m10          */
 
-    MOVQ      ( MM1, MM4 )              /* x1            | x1                */
-    PFMUL     ( REGIND(ECX), MM0 )      /* x0*m1         | x0*m0             */
+    MOVQ      ( MM1, MM4 )		/* x1              | x1              */
+    PFMUL     ( REGIND(ECX), MM0 )	/* x0*m1           | x0*m0           */
 
-    PFADD     ( REGOFF(48, ECX), MM2 )  /* x2*m9+m13     | x2*m8+m12         */
-    PFMUL     ( REGOFF(16, ECX), MM1 )  /* x1*m5         | x1*m4             */
+    PFADD     ( REGOFF(48, ECX), MM2 )	/* x2*m9+m13       | x2*m8+m12       */
+    PFMUL     ( REGOFF(16, ECX), MM1 )	/* x1*m5           | x1*m4           */
 
-    PFADD     ( REGOFF(56, ECX), MM5 )  /* x2*m11+m15    | x2*m10+m14        */
-    PFADD     ( MM0, MM1 )              /* x0*m1+x1*m5   | x0*m0+x1*m4       */
+    PFADD     ( REGOFF(56, ECX), MM5 )	/* x2*m11+m15      | x2*m10+m14      */
+    PFADD     ( MM0, MM1 )		/* x0*m1+x1*m5     | x0*m0+x1*m4     */
 
-    PFMUL     ( REGOFF(8, ECX), MM3 )   /* x0*m3         | x0*m2             */
-    PFADD     ( MM1, MM2 )              /* result r1     | r0                */
+    PFMUL     ( REGOFF(8, ECX), MM3 )	/* x0*m3           | x0*m2           */
+    PFADD     ( MM1, MM2 )		/* r1              | r0              */
 
-    PFMUL     ( REGOFF(24, ECX), MM4 )  /* x1*m7         | x1*m6             */
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
+    PFMUL     ( REGOFF(24, ECX), MM4 )	/* x1*m7           | x1*m6           */
+    ADD_L     ( CONST(16), EDX )	/* next output vertex                */
 
-    PFADD     ( MM3, MM4 )              /* x0*m3+x1*m7   | x0*m2+x1*m6       */
-    MOVQ      ( MM2, REGOFF(-16, EDX) ) /* write r0, r1                      */
+    PFADD     ( MM3, MM4 )		/* x0*m3+x1*m7     | x0*m2+x1*m6     */
+    MOVQ      ( MM2, REGOFF(-16, EDX) )	/* write r0, r1                      */
 
-    PFADD     ( MM4, MM5 )              /* r3            | r2                */
-    MOVQ      ( MM5, REGOFF(-8, EDX) )  /* write r2, r3                      */
+    PFADD     ( MM4, MM5 )		/* r3              | r2              */
+    MOVQ      ( MM5, REGOFF(-8, EDX) )	/* write r2, r3                      */
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
-    MOVD      ( REGOFF(8, EAX), MM2 )   /*               | x2                */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM2 )	/*                 | x2              */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    JA        ( LLBL(G3TPGR_1) )        /* cnt > 0 ? -> process next vertex  */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPGR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TPGR_2):
+LLBL( G3TPGR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -100,55 +119,81 @@ LLBL(G3TPGR_2):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_identity_raw)
-GLNAME( gl_3dnow_transform_points3_identity_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_perspective_raw )
+GLNAME( gl_3dnow_transform_points3_perspective_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPIR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPR_2 ) )
 
+    PREFETCH  ( REGIND(EAX) )
     PREFETCHW ( REGIND(EDX) )
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
-    MOVD      ( REGOFF(8, EAX), MM1 )   /*               | x2                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-ALIGNTEXT32
-LLBL(G3TPIR_1):
+    MOVQ      ( REGOFF(32, ECX), MM1 )	/* m21             | m20             */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
 
-    PREFETCHW ( REGOFF(32, EDX) )
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    MOVQ      ( MM0, REGOFF(-16, EDX) ) /* r1            | r0                */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVD      ( MM1, REGOFF(-8, EDX) )  /*               | r2                */
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
+ALIGNTEXT16
+LLBL( G3TPPR_1 ):
 
-    MOVD      ( REGOFF(8, EAX), MM1 )   /*               | x2                */
-    JA        ( LLBL(G3TPIR_1) )        /* cnt > 0 ? -> process next vertex  */
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGIND(EAX) )
+
+    PXOR      ( MM7, MM7 )		/* 0               | 0               */
+    MOVQ      ( MM5, MM6 )		/*                 | x2              */
+
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    PFSUB     ( MM5, MM7 )		/*                 | -x2             */
+
+    PFMUL     ( MM2, MM6 )		/*                 | x2*m22          */
+    PUNPCKLDQ ( MM5, MM5 )		/* x2              | x2              */
+
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    PFMUL     ( MM1, MM5 )		/* x2*m21          | x2*m20          */
+
+    PFADD     ( MM3, MM6 )		/*                 | x2*m22+m32      */
+    PFADD     ( MM4, MM5 )		/* x1*m11+x2*m21   | x0*m00+x2*m20   */
 
-ALIGNTEXT32
-LLBL(G3TPIR_2):
+    MOVQ      ( MM5, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    MOVD      ( MM6, REGOFF(-8, EDX) )	/* write r2                          */
+
+    MOVD      ( MM7, REGOFF(-4, EDX) )	/* write r3                          */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
+
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPPR_1 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TPPR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -158,74 +203,87 @@ LLBL(G3TPIR_2):
 
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_2d_raw)
-GLNAME( gl_3dnow_transform_points3_2d_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_3d_raw )
+GLNAME( gl_3dnow_transform_points3_3d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2R_3) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3R_2 ) )
 
     PREFETCH  ( REGIND(EAX) )
+    PREFETCH  ( REGIND(EDX) )
 
-    PREFETCHW ( REGIND(EDX) )
+    MOVD      ( REGOFF(8, ECX), MM7 )	/*                 | m2              */
+    PUNPCKLDQ ( REGOFF(24, ECX), MM7 )	/* m6              | m2              */
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )  /* m10           | m00                */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM1 )	/*                 | x2              */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )  /* m11           | m01                */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
-    MOVQ      ( REGIND(EAX), MM3 )      /* x1            | x0                 */
+ALIGNTEXT16
+LLBL( G3TP3R_1 ):
 
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGIND(EAX) )
 
-ALIGNTEXT32
-LLBL(G3TP2R_2):
+    MOVQ      ( MM0, MM2 )		/* x1              | x0              */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+    PUNPCKLDQ ( MM2, MM2 )		/* x0              | x0              */
+    MOVQ      ( MM0, MM3 )		/* x1              | x0              */
 
-    PREFETCH  ( REGIND(EAX) )
+    PFMUL     ( REGIND(ECX), MM2 )	/* x0*m1           | x0*m0           */
+    PUNPCKHDQ ( MM3, MM3 )		/* x1              | x1              */
 
-    MOVQ      ( MM3, MM4 )              /* x1            | x0                 */
-    PFMUL     ( MM0, MM3 )              /* x1*m10        | x0*m00             */
+    MOVQ      ( MM1, MM4 )		/*                 | x2              */
+    PFMUL     ( REGOFF(16, ECX), MM3 )	/* x1*m5           | x1*m4           */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    PFMUL     ( MM1, MM4 )              /* x1*m11        | x0*m01             */
+    PUNPCKLDQ ( MM4, MM4 )		/* x2              | x2              */
+    PFADD     ( MM2, MM3 )		/* x0*m1+x1*m5     | x0*m0+x1*m4     */
+
+    PFMUL     ( REGOFF(32, ECX), MM4 )	/* x2*m9           | x2*m8           */
+    PFADD     ( REGOFF(48, ECX), MM3 )	/* x0*m1+...+m11   | x0*m0+x1*m4+m12 */
+
+    PFMUL     ( MM7, MM0 )		/* x1*m6           | x0*m2           */
+    PFADD     ( MM4, MM3 )		/* r1              | r0              */
+
+    PFMUL     ( REGOFF(40, ECX), MM1 )	/*                 | x2*m10          */
+    PUNPCKLDQ ( REGOFF(56, ECX), MM1 )	/* m14             | x2*m10          */
+
+    PFACC     ( MM0, MM1 )
 
-    PFACC     ( MM4, MM3 )              /* x0*m00+x1*m10 | x0*m01+x1*m11      */
-    MOVD      ( MM5, REGOFF(-8, EDX) )  /* write r2 (=x2)                     */
+    MOVQ      ( MM3, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    PFACC     ( MM1, MM1 )		/*                 | r2              */
 
-    PFADD     ( MM2, MM3 )              /* x0*...*m10+m30| x0*m01+x1*m11+m31  */
-    MOVQ      ( MM3, REGOFF(-16, EDX) ) /* write r0, r1                       */
+    MOVD      ( MM1, REGOFF(-8, EDX) )	/* write r2                          */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
 
-    MOVQ      ( REGIND(EAX), MM3 )      /* x1            | x0                 */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
+    MOVD      ( REGOFF(8, EAX), MM1 )	/*                 | x2              */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP2R_2) )      /* cnt > 0 ? -> process next vertex   */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP3R_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TP2R_3):
+LLBL( G3TP3R_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -235,70 +293,74 @@ LLBL(G3TP2R_3):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_2d_no_rot_raw)
-GLNAME( gl_3dnow_transform_points3_2d_no_rot_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_3d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points3_3d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRR_2 ) )
 
     PREFETCH  ( REGIND(EAX) )
-
     PREFETCHW ( REGIND(EDX) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
+    PUNPCKLDQ ( MM2, MM2 )		/* m22             | m22             */
 
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
+    MOVD      ( REGOFF(56, ECX), MM3 )	/*                 | m32             */
 
+    PUNPCKLDQ ( MM3, MM3 )		/* m32             | m32             */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
 
-ALIGNTEXT32
-LLBL(G3TP2NRR_1):
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
 
-    PREFETCH  ( REGIND(EAX) )
+ALIGNTEXT16
+LLBL( G3TP3NRR_1 ):
+
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00            */
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
+    PREFETCHW ( REGIND(EAX) )
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    PFADD     ( MM1, MM4 )              /* x1*m11+m31    | x0*m00+m30        */
+    PFADD     ( MM1, MM4 )		/* x1*m11+m31      | x0*m00+m30      */
+    PFMUL     ( MM2, MM5 )		/*                 | x2*m22          */
 
-    MOVQ      ( MM4, REGOFF(-16, EDX) ) /* write r0, r1                      */
-    MOVD      ( MM5, REGOFF(-8, EDX) )  /* write r2 (=x2)                    */
+    PFADD     ( MM3, MM5 )		/*                 | x2*m22+m32      */
+    MOVQ      ( MM4, REGIND(EDX) )	/* write r0, r1                      */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
+    MOVD      ( MM5, REGOFF(-8, EDX) )	/* write r2                          */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
 
-    JA        ( LLBL(G3TP2NRR_1) )      /* cnt > 0 ? -> process next vertex  */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
+    JNZ       ( LLBL( G3TP3NRR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TP2NRR_2):
+LLBL( G3TP3NRR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -308,89 +370,72 @@ LLBL(G3TP2NRR_2):
 
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_3d_raw)
-GLNAME( gl_3dnow_transform_points3_3d_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_2d_raw )
+GLNAME( gl_3dnow_transform_points3_2d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3R_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2R_3) )
 
     PREFETCH  ( REGIND(EAX) )
+    PREFETCHW ( REGIND(EDX) )
 
-    PREFETCH  ( REGIND(EDX) )
-
-    MOVD      ( REGOFF(8, ECX), MM7 )   /*               | m2                */
-    PUNPCKLDQ ( REGOFF(24, ECX), MM7 )  /* m6            | m2                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )	/* m10             | m00             */
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
-    MOVD      ( REGOFF(8, EAX), MM1 )   /*               | x2                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )	/* m11             | m01             */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
+    MOVQ      ( REGIND(EAX), MM3 )	/* x1              | x0              */
 
-ALIGNTEXT32
-LLBL(G3TP3R_1):
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+ALIGNTEXT16
+LLBL( G3TP2R_2 ):
 
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
     PREFETCH  ( REGIND(EAX) )
 
-    MOVQ      ( MM0, MM2 )              /* x1            | x0                */
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
-
-    PUNPCKLDQ ( MM2, MM2 )              /* x0            | x0                */
-    MOVQ      ( MM0, MM3 )              /* x1            | x0                */
-
-    PFMUL     ( REGIND(ECX), MM2 )      /* x0*m1         | x0*m0             */
-    PUNPCKHDQ ( MM3, MM3 )              /* x1            | x1                */
+    MOVQ      ( MM3, MM4 )		/* x1              | x0              */
+    PFMUL     ( MM0, MM3 )		/* x1*m10          | x0*m00          */
 
-    MOVQ      ( MM1, MM4 )              /*               | x2                */
-    PFMUL     ( REGOFF(16, ECX), MM3 )  /* x1*m5         | x1*m4             */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    PFMUL     ( MM1, MM4 )		/* x1*m11          | x0*m01          */
 
-    PUNPCKLDQ ( MM4, MM4 )              /* x2            | x2                */
-    PFADD     ( MM2, MM3 )              /* x0*m1+x1*m5   | x0*m0+x1*m4       */
-
-    PFMUL     ( REGOFF(32, ECX), MM4 )  /* x2*m9         | x2*m8             */
-    PFADD     ( REGOFF(48, ECX), MM3 )  /* x0*m1+...+m11 | x0*m0+x1*m4+m12   */
-
-    PFMUL     ( MM7, MM0 )              /* x1*m6         | x0*m2             */
-    PFADD     ( MM4, MM3 )              /* r1            | r0                */
-
-    PFMUL     ( REGOFF(40, ECX), MM1 )  /*               | x2*m10            */
-    PUNPCKLDQ ( REGOFF(56, ECX), MM1 )  /* m14           | x2*m10            */
-
-    PFACC     ( MM0, MM1 )
+    PFACC     ( MM4, MM3 )		/* x0*m00+x1*m10   | x0*m01+x1*m11   */
+    MOVD      ( MM5, REGOFF(-8, EDX) )	/* write r2 (=x2)                    */
 
-    MOVQ      ( MM3, REGOFF(-16, EDX) ) /* write r0, r1                      */
-    PFACC     ( MM1, MM1 )              /*               | r2                */
+    PFADD     ( MM2, MM3 )		/* x0*...*m10+m30  | x0*...*m11+m31  */
+    MOVQ      ( MM3, REGOFF(-16, EDX) )	/* write r0, r1                      */
 
-    MOVD      ( MM1, REGOFF(-8, EDX) )  /* write r2                          */
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
+    MOVQ      ( REGIND(EAX), MM3 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-    MOVD      ( REGOFF(8, EAX), MM1 )   /*               | x2                */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2R_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    JA        ( LLBL(G3TP3R_1) )        /* cnt > 0 ? -> process next vertex  */
-
-ALIGNTEXT32
-LLBL(G3TP3R_2):
+LLBL( G3TP2R_3 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -400,74 +445,66 @@ LLBL(G3TP3R_2):
 
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_3d_no_rot_raw)
-GLNAME( gl_3dnow_transform_points3_3d_no_rot_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_2d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points3_2d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(3), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_3), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRR_2 ) )
 
     PREFETCH  ( REGIND(EAX) )
-
     PREFETCHW ( REGIND(EDX) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22                */
-    PUNPCKLDQ ( MM2, MM2 )              /* m22           | m22                */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32                */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PUNPCKLDQ ( MM3, MM3 )              /* m32           | m32                */
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
 
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                 */
+ALIGNTEXT16
+LLBL( G3TP2NRR_1 ):
 
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGIND(EAX) )
 
-ALIGNTEXT32
-LLBL(G3TP3NRR_1):
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-
-    PREFETCHW ( REGIND(EAX) )
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00            */
-
-    PFADD     ( MM1, MM4 )              /* x1*m11+m31    | x0*m00+m30        */
-    PFMUL     ( MM2, MM5 )              /*               | x2*m22            */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    PFADD     ( MM3, MM5 )              /*               | x2*m22+m32        */
-    MOVQ      ( MM4, REGIND(EDX) )      /* write r0, r1                      */
+    PFADD     ( MM1, MM4 )		/* x1*m11+m31      | x0*m00+m30      */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
+    MOVQ      ( MM4, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    MOVD      ( MM5, REGOFF(-8, EDX) )	/* write r2 (=x2)                    */
 
-    MOVD      ( MM5, REGOFF(-8, EDX) )  /* write r2                          */
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM5 )	/*                 | x2              */
 
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                */
-    JA        ( LLBL(G3TP3NRR_1) )      /* cnt > 0 ? -> process next vertex  */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2NRR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TP3NRR_2):
+LLBL( G3TP2NRR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -477,90 +514,56 @@ LLBL(G3TP3NRR_2):
 
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points3_perspective_raw)
-GLNAME( gl_3dnow_transform_points3_perspective_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points3_identity_raw )
+GLNAME( gl_3dnow_transform_points3_identity_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(3), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPR_2) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    PREFETCH  ( REGIND(EAX) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPIR_2 ) )
 
     PREFETCHW ( REGIND(EDX) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00               */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00               */
-
-    MOVQ      ( REGOFF(32, ECX), MM1 )  /* m21           | m20               */
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22               */
-
-    MOVD      ( REGOFF(56, ECX), MM3 )  /*               | m32               */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVD      ( REGOFF(8, EAX), MM1 )	/*                 | x2              */
 
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                */
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                */
-
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-
-ALIGNTEXT32
-LLBL(G3TPPR_1):
-
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
-
-    PREFETCH  ( REGIND(EAX) )
+ALIGNTEXT16
+LLBL( G3TPIR_1 ):
 
-    PXOR      ( MM7, MM7 )              /* 0             | 0                 */
-    MOVQ      ( MM5, MM6 )              /*               | x2                */
-
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00            */
-    PFSUB     ( MM5, MM7 )              /*               | -x2               */
-
-    PFMUL     ( MM2, MM6 )              /*               | x2*m22            */
-    PUNPCKLDQ ( MM5, MM5 )              /* x2            | x2                */
-
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
-    PFMUL     ( MM1, MM5 )              /* x2*m21        | x2*m20            */
-
-    PFADD     ( MM3, MM6 )              /*               | x2*m22+m32        */
-    PFADD     ( MM4, MM5 )              /* x1*m11+x2*m21 | x0*m00+x2*m20     */
+    PREFETCHW ( REGOFF(32, EDX) )
 
-    MOVQ      ( MM5, REGOFF(-16, EDX) ) /* write r0, r1                      */
-    MOVD      ( MM6, REGOFF(-8, EDX) )  /* write r2                          */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVD      ( MM7, REGOFF(-4, EDX) )  /* write r3                          */
-    MOVD      ( REGOFF(8, EAX), MM5 )   /*               | x2                */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    MOVQ      ( MM0, REGOFF(-16, EDX) )	/* r1              | r0              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                */
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    MOVD      ( MM1, REGOFF(-8, EDX) )	/*                 | r2              */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    JA        ( LLBL(G3TPPR_1) )        /* cnt > 0 ? -> process next vertex  */
+    MOVD      ( REGOFF(8, EAX), MM1 )	/*                 | x2              */
+    JNZ       ( LLBL( G3TPIR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TPPR_2):
+LLBL( G3TPIR_2 ):
 
     FEMMS
     POP_L     ( EDI )
     POP_L     ( ESI )
     RET
-
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/3dnow_xform_raw4.S b/xc/extras/Mesa/src/X86/3dnow_xform_raw4.S
index 17f584f75..fcabb5949 100644
--- a/xc/extras/Mesa/src/X86/3dnow_xform_raw4.S
+++ b/xc/extras/Mesa/src/X86/3dnow_xform_raw4.S
@@ -1,86 +1,105 @@
-#include "assyntax.h"
-
-   SEG_TEXT
-
 
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
 
-#define  VEC_SIZE_1   1
-#define  VEC_SIZE_2   3
-#define  VEC_SIZE_3   7
-#define  VEC_SIZE_4   15
+#include "assyntax.h"
+#include "xform_args.h"
 
+    SEG_TEXT
 
+#define FRAME_OFFSET	4
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_general_raw)
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_general_raw )
 GLNAME( gl_3dnow_transform_points4_general_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPGR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPGR_2 ) )
 
     PREFETCHW ( REGIND(EDX) )
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
-    MOVQ      ( REGOFF(8, EAX), MM4 )   /* x3            | x2                */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM4 )	/* x3              | x2              */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-ALIGNTEXT32
-LLBL(G3TPGR_1):
-
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+ALIGNTEXT16
+LLBL( G3TPGR_1 ):
 
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
     PREFETCH  ( REGIND(EAX) )
 
-    MOVQ      ( MM0, MM2 )              /* x1            | x0                */
-    MOVQ      ( MM4, MM6 )              /* x3            | x2                */
+    MOVQ      ( MM0, MM2 )		/* x1              | x0              */
+    MOVQ      ( MM4, MM6 )		/* x3              | x2              */
 
-    PUNPCKLDQ ( MM0, MM0 )              /* x0            | x0                */
-    PUNPCKHDQ ( MM2, MM2 )              /* x1            | x1                */
+    PUNPCKLDQ ( MM0, MM0 )		/* x0              | x0              */
+    PUNPCKHDQ ( MM2, MM2 )		/* x1              | x1              */
 
-    MOVQ      ( MM0, MM1 )              /* x0            | x0                */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    MOVQ      ( MM0, MM1 )		/* x0              | x0              */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    PFMUL     ( REGIND(ECX), MM0 )      /* x0*m1         | x0*m0             */
-    MOVQ      ( MM2, MM3 )              /* x1            | x1                */
+    PFMUL     ( REGIND(ECX), MM0 )	/* x0*m1           | x0*m0           */
+    MOVQ      ( MM2, MM3 )		/* x1              | x1              */
 
-    PFMUL     ( REGOFF(8, ECX), MM1 )   /* x0*m3         | x0*m2             */
-    PUNPCKLDQ ( MM4, MM4 )              /* x2            | x2                */
+    PFMUL     ( REGOFF(8, ECX), MM1 )	/* x0*m3           | x0*m2           */
+    PUNPCKLDQ ( MM4, MM4 )		/* x2              | x2              */
 
-    PFMUL     ( REGOFF(16, ECX), MM2 )  /* x1*m5         | x1*m4             */
-    MOVQ      ( MM4, MM5 )              /* x2            | x2                */
+    PFMUL     ( REGOFF(16, ECX), MM2 )	/* x1*m5           | x1*m4           */
+    MOVQ      ( MM4, MM5 )		/* x2              | x2              */
 
-    PFMUL     ( REGOFF(24, ECX), MM3 )  /* x1*m7         | x1*m6             */
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x3                */
+    PFMUL     ( REGOFF(24, ECX), MM3 )	/* x1*m7           | x1*m6           */
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x3              */
 
-    PFMUL     ( REGOFF(32, ECX), MM4 )  /* x2*m9         | x2*m8             */
-    MOVQ      ( MM6, MM7 )              /* x3            | x3                */
+    PFMUL     ( REGOFF(32, ECX), MM4 )	/* x2*m9           | x2*m8           */
+    MOVQ      ( MM6, MM7 )		/* x3              | x3              */
 
-    PFMUL     ( REGOFF(40, ECX), MM5 )  /* x2*m11        | x2*m10            */
+    PFMUL     ( REGOFF(40, ECX), MM5 )	/* x2*m11          | x2*m10          */
     PFADD     ( MM0, MM2 )
 
-    PFMUL     ( REGOFF(48, ECX), MM6 )  /* x3*m13        | x3*m12            */
+    PFMUL     ( REGOFF(48, ECX), MM6 )	/* x3*m13          | x3*m12          */
     PFADD     ( MM1, MM3 )
 
-    PFMUL     ( REGOFF(56, ECX), MM7 )  /* x3*m15        | x3*m14            */
+    PFMUL     ( REGOFF(56, ECX), MM7 )	/* x3*m15          | x3*m14          */
     PFADD     ( MM4, MM6 )
 
     PFADD     ( MM5, MM7 )
@@ -90,17 +109,15 @@ LLBL(G3TPGR_1):
     MOVQ      ( MM6, REGOFF(-16, EDX) )
 
     MOVQ      ( MM7, REGOFF(-8, EDX) )
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                */
-
-    MOVQ      ( REGOFF(8, EAX), MM4 )   /* x3            | x2                */
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1            | x0                */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    JA        ( LLBL(G3TPGR_1) )        /* cnt > 0 ? -> process next vertex  */
+    MOVQ      ( REGOFF(8, EAX), MM4 )	/* x3            | x2                */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPGR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TPGR_2):
+LLBL( G3TPGR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -110,57 +127,82 @@ LLBL(G3TPGR_2):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_identity_raw)
-GLNAME( gl_3dnow_transform_points4_identity_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_perspective_raw )
+GLNAME( gl_3dnow_transform_points4_perspective_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPIR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPPR_2 ) )
 
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM1 )   /* x3            | x2                 */
+    PREFETCH  ( REGIND(EAX) )
+    PREFETCHW ( REGIND(EDX) )
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-ALIGNTEXT32
-LLBL(G3TPIR_1):
+    MOVD      ( REGOFF(40, ECX), MM1 )	/*                 | m22             */
+    PUNPCKLDQ ( REGOFF(56, ECX), MM1 )	/* m32             | m22             */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+    MOVQ      ( REGOFF(32, ECX), MM2 )	/* m21             | m20             */
+    PXOR      ( MM7, MM7 )		/* 0               | 0               */
 
-    PREFETCH  ( REGIND(EAX) )
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
+    MOVD      ( REGOFF(8, EAX), MM3 )	/*                 | x2              */
+
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+
+ALIGNTEXT16
+LLBL( G3TPPR_1 ):
+
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGOFF(32, EAX) )	/* hopefully stride is zero          */
+
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    MOVQ      ( MM0, REGOFF(-16, EDX) ) /* r1            | r0                 */
+    PUNPCKLDQ ( MM5, MM5 )		/* x2              | x2              */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVQ      ( MM1, REGOFF(-8, EDX) )  /* r3            | r2                 */
-    MOVQ      ( REGIND(EAX), MM0 )      /* x1            | x0                 */
+    PFMUL     ( MM2, MM5 )		/* x2*m21          | x2*m20          */
+    PFSUBR    ( MM7, MM3 )		/*                 | -x2             */
 
-    MOVQ      ( REGOFF(8, EAX), MM1 )   /* x3            | x2                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    PFMUL     ( MM1, MM6 )		/* x3*m32          | x2*m22          */
+    PFADD     ( MM4, MM5 )		/* x1*m11+x2*m21   | x0*m00+x2*m20   */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TPIR_1) )        /* cnt > 0 ? -> process next vertex   */
+    PFACC     ( MM3, MM6 )		/* -x2             | x2*m22+x3*m32   */
+    MOVQ      ( MM5, REGOFF(-16, EDX) )	/* write r0, r1                      */
 
-ALIGNTEXT32
-LLBL(G3TPIR_2):
+    MOVQ      ( MM6, REGOFF(-8, EDX) )	/* write r2, r3                      */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
+    MOVD      ( REGOFF(8, EAX), MM3 )	/*                 | x2              */
+
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+
+    JNZ       ( LLBL( G3TPPR_1 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TPPR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -170,77 +212,89 @@ LLBL(G3TPIR_2):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_2d_raw)
-GLNAME( gl_3dnow_transform_points4_2d_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_3d_raw )
+GLNAME( gl_3dnow_transform_points4_3d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2R_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3R_2 ) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )  /* m10           | m00                */
+    MOVD      ( REGOFF(8, ECX), MM6 )	/*                 | m2              */
+    PUNPCKLDQ ( REGOFF(24, ECX), MM6 )	/* m6              | m2              */
 
-    MOVD      ( REGOFF(4, ECX), MM1 )   /*               | m01                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )  /* m11           | m01                */
+    MOVD      ( REGOFF(40, ECX), MM7 )	/*                 | m10             */
+    PUNPCKLDQ ( REGOFF(56, ECX), MM7 )	/* m14             | m10             */
 
-    MOVQ      ( REGOFF(48, ECX), MM2 )  /* m31           | m30                */
+    MOVQ      ( REGIND(EAX), MM2 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM3 )	/* x3              | x2              */
 
-    MOVQ      ( REGIND(EAX), MM3 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
+ALIGNTEXT16
+LLBL( G3TP3R_1 ):
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGOFF(32, EAX) )	/* hopefully array is tightly packed */
 
-ALIGNTEXT32
-LLBL(G3TP2R_1):
+    MOVQ      ( MM2, MM0 )		/* x1              | x0              */
+    MOVQ      ( MM3, MM4 )		/* x3              | x2              */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+    MOVQ      ( MM0, MM1 )		/* x1              | x0              */
+    MOVQ      ( MM4, MM5 )		/* x3              | x2              */
 
-    PREFETCH  ( REGIND(EAX) )
+    PUNPCKLDQ ( MM0, MM0 )		/* x0              | x0              */
+    PUNPCKHDQ ( MM1, MM1 )		/* x1              | x1              */
 
-    MOVQ      ( MM3, MM4 )              /* x1            | x0                 */
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
+    PFMUL     ( REGIND(ECX), MM0 )	/* x0*m1           | x0*m0           */
+    PUNPCKLDQ ( MM3, MM3 )		/* x2              | x2              */
 
-    PFMUL     ( MM1, MM4 )              /* x1*m11        | x0*m01             */
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x3                 */
+    PFMUL     ( REGOFF(16, ECX), MM1 )	/* x1*m5           | x1*m4           */
+    PUNPCKHDQ ( MM4, MM4 )		/* x3              | x3              */
 
-    PFMUL     ( MM0, MM3 )              /* x1*m10        | x0*m00             */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
+    PFMUL     ( MM6, MM2 )		/* x1*m6           | x0*m2           */
+    PFADD     ( MM0, MM1 )		/* x0*m1+x1*m5     | x0*m0+x1*m4     */
 
-    PFACC     ( MM4, MM3 )              /* x0*m01+x1*m11 | x0*m00+x1*m10      */
-    PFMUL     ( MM2, MM6 )              /* x3*m31        | x3*m30             */
+    PFMUL     ( REGOFF(32, ECX), MM3 )	/* x2*m9           | x2*m8           */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    PFADD     ( MM6, MM3 )              /* r1            | r0                 */
-    MOVQ      ( MM5, REGOFF(-8, EDX) )  /* write r2, r3                       */
+    PFMUL     ( REGOFF(48, ECX), MM4 )	/* x3*m13          | x3*m12          */
+    PFADD     ( MM1, MM3 )		/* x0*m1+..+x2*m9  | x0*m0+...+x2*m8 */
 
-    MOVQ      ( MM3, REGOFF(-16, EDX) ) /* write r0, r1                       */
-    MOVQ      ( REGIND(EAX), MM3 )      /* x1            | x0                 */
+    PFMUL     ( MM7, MM5 )		/* x3*m14          | x2*m10          */
+    PFADD     ( MM3, MM4 )		/* r1              | r0              */
 
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    PFACC     ( MM2, MM5 )		/* x0*m2+x1*m6     | x2*m10+x3*m14   */
+    MOVD      ( REGOFF(12, EAX), MM0 )	/*                 | x3              */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
-    JA        ( LLBL(G3TP2R_1) )        /* cnt > 0 ? -> process next vertex   */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    PFACC     ( MM0, MM5 )		/* r3              | r2              */
 
-ALIGNTEXT32
-LLBL(G3TP2R_2):
+    MOVQ      ( MM4, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    MOVQ      ( MM5, REGOFF(-8, EDX) )	/* write r2, r3                      */
+
+    MOVQ      ( REGIND(EAX), MM2 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM3 )	/* x3              | x2              */
+
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP3R_1 ) )	/* cnt > 0 ? -> process next vertex  */
+
+LLBL( G3TP3R_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -250,70 +304,75 @@ LLBL(G3TP2R_2):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_2d_no_rot_raw)
-GLNAME( gl_3dnow_transform_points4_2d_no_rot_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_3d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points4_3d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP2NRR_3) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP3NRR_2 ) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30                */
+    MOVD      ( REGOFF(40, ECX), MM2 )	/*                 | m22             */
+    PUNPCKLDQ ( REGOFF(56, ECX), MM2 )	/* m32             | m22             */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
+    MOVD      ( REGOFF(12, EAX), MM7 )	/*                 | x3              */
 
-ALIGNTEXT32
-LLBL(G3TP2NRR_2):
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+ALIGNTEXT16
+LLBL( G3TP3NRR_1 ):
 
-    PREFETCH  ( REGIND(EAX) )
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGOFF(32, EAX) )	/* hopefully stride is zero          */
+
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
 
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x3              */
+    PFMUL     ( MM2, MM5 )		/* x3*m32          | x2*m22          */
 
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x3                 */
+    PFMUL     ( MM1, MM6 )		/* x3*m31          | x3*m30          */
+    PFACC     ( MM7, MM5 )		/* x3              | x2*m22+x3*m32   */
 
-    PFMUL     ( MM1, MM6 )              /* x3*m31        | x3*m30             */
-    PFADD     ( MM4, MM6 )              /* x1*m11+x3*m31 | x0*m00+x3*m30      */
+    PFADD     ( MM6, MM4 )		/* x1*m11+x3*m31   | x0*m00+x3*m30   */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    MOVQ      ( MM6, REGOFF(-16, EDX) ) /* write r0, r1                       */
-    MOVQ      ( MM5, REGOFF(-8, EDX) )  /* write r2, r3                       */
+    MOVQ      ( MM4, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    MOVQ      ( MM5, REGOFF(-8, EDX) )	/* write r2, r3                      */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVD      ( REGOFF(12, EAX), MM7 )	/*                 | x3              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    JA        ( LLBL(G3TP2NRR_2) )    /* cnt > 0 ? -> process next vertex   */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP3NRR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TP2NRR_3):
+LLBL( G3TP3NRR_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -323,92 +382,75 @@ LLBL(G3TP2NRR_3):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_3d_raw)
-GLNAME( gl_3dnow_transform_points4_3d_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_2d_raw )
+GLNAME( gl_3dnow_transform_points4_2d_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3R_2) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVD      ( REGOFF(8, ECX), MM6 )   /*               | m2                */
-    PUNPCKLDQ ( REGOFF(24, ECX), MM6 )  /* m6            | m2                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2R_2 ) )
 
-    MOVD      ( REGOFF(40, ECX), MM7 )  /*               | m10               */
-    PUNPCKLDQ ( REGOFF(56, ECX), MM7 )  /* m14           | m10               */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(16, ECX), MM0 )	/* m10             | m00             */
 
-    MOVQ      ( REGIND(EAX), MM2 )      /* x1            | x0                */
-    MOVQ      ( REGOFF(8, EAX), MM3 )   /* x3            | x2                */
+    MOVD      ( REGOFF(4, ECX), MM1 )	/*                 | m01             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM1 )	/* m11             | m01             */
 
-ALIGNTEXT32
-LLBL(G3TP3R_1):
+    MOVQ      ( REGOFF(48, ECX), MM2 )	/* m31             | m30             */
 
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
-
-    PREFETCH  ( REGOFF(32, EAX) )       /* hopefully array is tightly packed */
-
-    MOVQ      ( MM2, MM0 )              /* x1            | x0                */
-    MOVQ      ( MM3, MM4 )              /* x3            | x2                */
+    MOVQ      ( REGIND(EAX), MM3 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    MOVQ      ( MM0, MM1 )              /* x1            | x0                */
-    MOVQ      ( MM4, MM5 )              /* x3            | x2                */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PUNPCKLDQ ( MM0, MM0 )              /* x0            | x0                */
-    PUNPCKHDQ ( MM1, MM1 )              /* x1            | x1                */
-    
-    PFMUL     ( REGIND(ECX), MM0 )      /* x0*m1         | x0*m0             */
-    PUNPCKLDQ ( MM3, MM3 )              /* x2            | x2                */
+ALIGNTEXT16
+LLBL( G3TP2R_1 ):
 
-    PFMUL     ( REGOFF(16, ECX), MM1 )  /* x1*m5         | x1*m4             */
-    PUNPCKHDQ ( MM4, MM4 )              /* x3            | x3                */
-
-    PFMUL     ( MM6, MM2 )              /* x1*m6         | x0*m2             */
-    PFADD     ( MM0, MM1 )              /* x0*m1+x1*m5   | x0*m0+x1*m4       */
+    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGIND(EAX) )
 
-    PFMUL     ( REGOFF(32, ECX), MM3 )  /* x2*m9         | x2*m8             */
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
+    MOVQ      ( MM3, MM4 )		/* x1              | x0              */
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
 
-    PFMUL     ( REGOFF(48, ECX), MM4 )  /* x3*m13        | x3*m12            */
-    PFADD     ( MM1, MM3 )              /* x0*m1+..+x2*m9| x0*m0+...+x2*m8   */
-    
-    PFMUL     ( MM7, MM5 )              /* x3*m14        | x2*m10            */
-    PFADD     ( MM3, MM4 )              /* r1            | r0                */
+    PFMUL     ( MM1, MM4 )		/* x1*m11          | x0*m01          */
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x3              */
 
-    PFACC     ( MM2, MM5 )              /* x0*m2+x1*m6   | x2*m10+x3*m14     */
-    MOVD      ( REGOFF(12, EAX), MM0 )  /*               | x3                */
+    PFMUL     ( MM0, MM3 )		/* x1*m10          | x0*m00          */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-    PFACC     ( MM0, MM5 )              /* r3            | r2                */
+    PFACC     ( MM4, MM3 )		/* x0*m01+x1*m11   | x0*m00+x1*m10   */
+    PFMUL     ( MM2, MM6 )		/* x3*m31          | x3*m30          */
 
-    MOVQ      ( MM4, REGOFF(-16, EDX) ) /* write r0, r1                      */
-    MOVQ      ( MM5, REGOFF(-8, EDX) )  /* write r2, r3                      */
+    PFADD     ( MM6, MM3 )		/* r1              | r0              */
+    MOVQ      ( MM5, REGOFF(-8, EDX) )	/* write r2, r3                      */
 
-    MOVQ      ( REGIND(EAX), MM2 )      /* x1            | x0                */
-    MOVQ      ( REGOFF(8, EAX), MM3 )   /* x3            | x2                */
+    MOVQ      ( MM3, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    MOVQ      ( REGIND(EAX), MM3 )	/* x1              | x0              */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    JA        ( LLBL(G3TP3R_1) )        /* cnt > 0 ? -> process next vertex  */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TP2R_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TP3R_2):
+LLBL( G3TP2R_2 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -418,78 +460,68 @@ LLBL(G3TP3R_2):
 
 
 
-
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_3d_no_rot_raw)
-GLNAME( gl_3dnow_transform_points4_3d_no_rot_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_2d_no_rot_raw )
+GLNAME( gl_3dnow_transform_points4_2d_no_rot_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TP3NRR_2) )
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TP2NRR_3 ) )
 
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00               */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00               */
+    MOVD      ( REGIND(ECX), MM0 )	/*                 | m00             */
+    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )	/* m11             | m00             */
 
-    MOVD      ( REGOFF(40, ECX), MM2 )  /*               | m22               */
-    PUNPCKLDQ ( REGOFF(56, ECX), MM2 )  /* m32           | m22               */
+    MOVQ      ( REGOFF(48, ECX), MM1 )	/* m31             | m30             */
 
-    MOVQ      ( REGOFF(48, ECX), MM1 )  /* m31           | m30               */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                */
-    MOVD      ( REGOFF(12, EAX), MM7 )  /*               | x3                */
-
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
-
-ALIGNTEXT32
-LLBL(G3TP3NRR_1):
-
-    PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    PREFETCH  ( REGOFF(32, EAX) )       /* hopefully stride is zero          */
+ALIGNTEXT16
+LLBL( G3TP2NRR_2 ):
 
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00            */
+    PREFETCHW ( REGOFF(32, EDX) )	/* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGIND(EAX) )
 
-    PUNPCKHDQ ( MM6, MM6 )              /* x3            | x3                */
-    PFMUL     ( MM2, MM5 )              /* x3*m32        | x2*m22            */
+    PFMUL     ( MM0, MM4 )		/* x1*m11          | x0*m00          */
+    MOVQ      ( MM5, MM6 )		/* x3              | x2              */
 
-    PFMUL     ( MM1, MM6 )              /* x3*m31        | x3*m30            */
-    PFACC     ( MM7, MM5 )              /* x3            | x2*m22+x3*m32     */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    PUNPCKHDQ ( MM6, MM6 )		/* x3              | x3              */
 
-    PFADD     ( MM6, MM4 )              /* x1*m11+x3*m31 | x0*m00+x3*m30     */
-    ADD_L     ( CONST(16), EDX )        /* next r                            */
+    PFMUL     ( MM1, MM6 )		/* x3*m31          | x3*m30          */
+    PFADD     ( MM4, MM6 )		/* x1*m11+x3*m31   | x0*m00+x3*m30   */
 
-    MOVQ      ( MM4, REGOFF(-16, EDX) ) /* write r0, r1                      */
-    MOVQ      ( MM5, REGOFF(-8, EDX) )  /* write r2, r3                      */
+    MOVQ      ( MM6, REGOFF(-16, EDX) )	/* write r0, r1                      */
+    MOVQ      ( MM5, REGOFF(-8, EDX) )	/* write r2, r3                      */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                */
+    MOVQ      ( REGIND(EAX), MM4 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM5 )	/* x3              | x2              */
 
-    MOVD      ( REGOFF(12, EAX), MM7 )  /*               | x3                */
-    ADD_L     ( EDI, EAX )              /* next vertex                       */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
 
-    DEC_L     ( ESI )                   /* decrement vertex counter          */
-    JA        ( LLBL(G3TP3NRR_1) )      /* cnt > 0 ? -> process next vertex  */
+    JNZ       ( LLBL( G3TP2NRR_2 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TP3NRR_2):
+LLBL( G3TP2NRR_3 ):
 
     FEMMS
     POP_L     ( EDI )
@@ -499,87 +531,57 @@ LLBL(G3TP3NRR_2):
 
 
 
-ALIGNTEXT32
-GLOBL GLNAME(gl_3dnow_transform_points4_perspective_raw)
-GLNAME( gl_3dnow_transform_points4_perspective_raw ):
+ALIGNTEXT16
+GLOBL GLNAME( gl_3dnow_transform_points4_identity_raw )
+GLNAME( gl_3dnow_transform_points4_identity_raw ):
 
     PUSH_L    ( ESI )
-    MOV_L     ( REGOFF(8, ESP), ECX )
-    MOV_L     ( REGOFF(12, ESP), ESI )
-    MOV_L     ( REGOFF(16, ESP), EAX )
-    MOV_L     ( CONST(4), REGOFF(16, ECX) )
-    OR_B      ( CONST(VEC_SIZE_4), REGOFF(20, ECX) )
-    MOV_L     ( REGOFF(8, EAX), EDX )
-    MOV_L     ( EDX, REGOFF(8, ECX) )
+
+    MOV_L     ( ARG_DEST, ECX )
+    MOV_L     ( ARG_MATRIX, ESI )
+    MOV_L     ( ARG_SOURCE, EAX )
+    MOV_L     ( CONST(4), REGOFF(V4F_SIZE, ECX) )
+    OR_B      ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) )
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), EDX )
+    MOV_L     ( EDX, REGOFF(V4F_COUNT, ECX) )
 
     PUSH_L    ( EDI )
 
-    MOV_L     ( REGOFF(4, ECX), EDX )
+    MOV_L     ( REGOFF(V4F_START, ECX), EDX )
     MOV_L     ( ESI, ECX )
-    MOV_L     ( REGOFF(8, EAX), ESI )
-    MOV_L     ( REGOFF(12, EAX), EDI )
-    MOV_L     ( REGOFF(4, EAX), EAX )
-
-    CMP_L     ( CONST(0), ESI )
-    JE        ( LLBL(G3TPPR_2) )
-
-    PREFETCH  ( REGIND(EAX) )
-
-    PREFETCHW ( REGIND(EDX) )
-
-    MOVD      ( REGIND(ECX), MM0 )      /*               | m00                */
-    PUNPCKLDQ ( REGOFF(20, ECX), MM0 )  /* m11           | m00                */
+    MOV_L     ( REGOFF(V4F_COUNT, EAX), ESI )
+    MOV_L     ( REGOFF(V4F_STRIDE, EAX), EDI )
+    MOV_L     ( REGOFF(V4F_START, EAX), EAX )
 
-    MOVD      ( REGOFF(40, ECX), MM1 )  /*               | m22                */
-    PUNPCKLDQ ( REGOFF(56, ECX), MM1 )  /* m32           | m22                */
+    TEST_L    ( ESI, ESI )
+    JZ        ( LLBL( G3TPIR_2 ) )
 
-    MOVQ      ( REGOFF(32, ECX), MM2 )  /* m21           | m20                */
-    PXOR      ( MM7, MM7 )              /* 0             | 0                  */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
+    MOVQ      ( REGOFF(8, EAX), MM1 )	/* x3              | x2              */
 
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
-    MOVD      ( REGOFF(8, EAX), MM3 )   /*               | x2                 */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-
-ALIGNTEXT32
-LLBL(G3TPPR_1):
+ALIGNTEXT16
+LLBL( G3TPIR_1 ):
 
     PREFETCHW ( REGOFF(32, EDX) )       /* prefetch 2 vertices ahead         */
+    PREFETCH  ( REGIND(EAX) )
 
-    PREFETCH  ( REGOFF(32, EAX) )       /* hopefully stride is zero          */
-
-    MOVQ      ( MM5, MM6 )              /* x3            | x2                 */
-    PFMUL     ( MM0, MM4 )              /* x1*m11        | x0*m00             */
-
-    PUNPCKLDQ ( MM5, MM5 )              /* x2            | x2                 */
-    ADD_L     ( CONST(16), EDX )        /* next r                             */
-
-    PFMUL     ( MM2, MM5 )              /* x2*m21        | x2*m20             */
-    PFSUBR    ( MM7, MM3 )              /*               | -x2                */
-
-    PFMUL     ( MM1, MM6 )              /* x3*m32        | x2*m22             */
-    PFADD     ( MM4, MM5 )              /* x1*m11+x2*m21 | x0*m00+x2*m20      */
-
-    PFACC     ( MM3, MM6 )              /* -x2           | x2*m22+x3*m32      */
-    MOVQ      ( MM5, REGOFF(-16, EDX) ) /* write r0, r1                       */
-
-    MOVQ      ( MM6, REGOFF(-8, EDX) )  /* write r2, r3                       */
-    MOVQ      ( REGIND(EAX), MM4 )      /* x1            | x0                 */
+    ADD_L     ( CONST(16), EDX )	/* next r                            */
+    MOVQ      ( MM0, REGOFF(-16, EDX) )	/* r1              | r0              */
 
-    MOVQ      ( REGOFF(8, EAX), MM5 )   /* x3            | x2                 */
-    MOVD      ( REGOFF(8, EAX), MM3 )   /*               | x2                 */
+    MOVQ      ( MM1, REGOFF(-8, EDX) )	/* r3              | r2              */
+    MOVQ      ( REGIND(EAX), MM0 )	/* x1              | x0              */
 
-    ADD_L     ( EDI, EAX )              /* next vertex                        */
-    DEC_L     ( ESI )                   /* decrement vertex counter           */
+    MOVQ      ( REGOFF(8, EAX), MM1 )	/* x3              | x2              */
+    ADD_L     ( EDI, EAX )		/* next vertex                       */
 
-    JA        ( LLBL(G3TPPR_1) )        /* cnt > 0 ? -> process next vertex   */
+    DEC_L     ( ESI )			/* decrement vertex counter          */
+    JNZ       ( LLBL( G3TPIR_1 ) )	/* cnt > 0 ? -> process next vertex  */
 
-ALIGNTEXT32
-LLBL(G3TPPR_2):
+LLBL( G3TPIR_2 ):
 
     FEMMS
     POP_L     ( EDI )
     POP_L     ( ESI )
     RET
-
diff --git a/xc/extras/Mesa/src/X86/assyntax.h b/xc/extras/Mesa/src/X86/assyntax.h
index ae1506701..13c48445c 100644
--- a/xc/extras/Mesa/src/X86/assyntax.h
+++ b/xc/extras/Mesa/src/X86/assyntax.h
@@ -1,5 +1,6 @@
+
 #ifndef __ASSYNTAX_H__
-#define	__ASSYNTAX_H__
+#define __ASSYNTAX_H__
 
 /*
  * Copyright 1992 Vrije Universiteit, The Netherlands
@@ -14,12 +15,12 @@
  * representations about the suitability of this software for any purpose.
  * It is provided "as is" without express or implied warranty.
  *
- * The Vrije Universiteit DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
- * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
- * EVENT SHALL The Vrije Universiteit BE LIABLE FOR ANY SPECIAL, INDIRECT OR
- * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
- * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
- * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+ * The Vrije Universiteit DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS
+ * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS,
+ * IN NO EVENT SHALL The Vrije Universiteit BE LIABLE FOR ANY SPECIAL,
+ * INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM
+ * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE
+ * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
  * PERFORMANCE OF THIS SOFTWARE.
  */
 
@@ -71,11 +72,10 @@
 #define GNU_ASSEMBLER
 #endif
 
-#if (defined(__STDC__) && !defined(UNIXCPP)) || (defined (sun) && defined (i386) \
- && defined (SVR4) && defined (__STDC__) && !defined (__GNUC__)) 
-#define	CONCAT(x, y)	x ## y
+#if (defined(__STDC__) && !defined(UNIXCPP)) || (defined (sun) && defined (i386) && defined (SVR4) && defined (__STDC__) && !defined (__GNUC__))
+#define CONCAT(x, y)	x ## y
 #else
-#define	CONCAT(x, y)	x/**/y
+#define CONCAT(x, y)	x/**/y
 #endif
 
 #ifdef ACK_ASSEMBLER
@@ -83,170 +83,170 @@
 /* Assume we write code for 32-bit protected mode! */
 
 /* Redefine register names for GAS & AT&T assemblers */
-#define	AL	al
-#define	AH	ah
-#define	AX	ax
-#define	EAX	ax
-#define	BL	bl
-#define	BH	bh
-#define	BX	bx
-#define	EBX	bx
-#define	CL	cl
-#define	CH	ch
-#define	CX	cx
-#define	ECX	cx
-#define	DL	dl
-#define	DH	dh
-#define	DX	dx
-#define	EDX	dx
-#define	BP	bp
-#define	EBP	bp
-#define	SI	si
-#define	ESI	si
-#define	DI	di
-#define	EDI	di
-#define	SP	sp
-#define	ESP	sp
-#define	CS	cs
-#define	SS	ss
-#define	DS	ds
-#define	ES	es
-#define	FS	fs
-#define	GS	gs
+#define AL		al
+#define AH		ah
+#define AX		ax
+#define EAX		ax
+#define BL		bl
+#define BH		bh
+#define BX		bx
+#define EBX		bx
+#define CL		cl
+#define CH		ch
+#define CX		cx
+#define ECX		cx
+#define DL		dl
+#define DH		dh
+#define DX		dx
+#define EDX		dx
+#define BP		bp
+#define EBP		bp
+#define SI		si
+#define ESI		si
+#define DI		di
+#define EDI		di
+#define SP		sp
+#define ESP		sp
+#define CS		cs
+#define SS		ss
+#define DS		ds
+#define ES		es
+#define FS		fs
+#define GS		gs
 /* Control Registers */
-#define	CR0	cr0
-#define	CR1	cr1
-#define	CR2	cr2
-#define	CR3	cr3
+#define CR0		cr0
+#define CR1		cr1
+#define CR2		cr2
+#define CR3		cr3
 /* Debug Registers */
-#define	DR0	dr0
-#define	DR1	dr1
-#define	DR2	dr2
-#define	DR3	dr3
-#define	DR4	dr4
-#define	DR5	dr5
-#define	DR6	dr6
-#define	DR7	dr7
+#define DR0		dr0
+#define DR1		dr1
+#define DR2		dr2
+#define DR3		dr3
+#define DR4		dr4
+#define DR5		dr5
+#define DR6		dr6
+#define DR7		dr7
 /* Floating-point Stack */
-#define	ST	st
+#define ST		st
 
-#define	AS_BEGIN	.sect .text; .sect .rom; .sect .data; .sect .bss; .sect .text
+#define AS_BEGIN	.sect .text; .sect .rom; .sect .data; .sect .bss; .sect .text
 
 
-#define	_WTOG		o16	/* word toggle for _W instructions */
-#define	_LTOG			/* long toggle for _L instructions */
-#define	ADDR_TOGGLE	a16
-#define	OPSZ_TOGGLE	o16
-#define	USE16		.use16
-#define	USE32		.use32
+#define _WTOG		o16	/* word toggle for _W instructions */
+#define _LTOG			/* long toggle for _L instructions */
+#define ADDR_TOGGLE	a16
+#define OPSZ_TOGGLE	o16
+#define USE16		.use16
+#define USE32		.use32
 
-#define	CHOICE(a,b,c)	c
+#define CHOICE(a,b,c)	c
 
 #else /* AT&T or GAS */
 
 /* Redefine register names for GAS & AT&T assemblers */
-#define	AL	%al
-#define	AH	%ah
-#define	AX	%ax
-#define	EAX	%eax
-#define	BL	%bl
-#define	BH	%bh
-#define	BX	%bx
-#define	EBX	%ebx
-#define	CL	%cl
-#define	CH	%ch
-#define	CX	%cx
-#define	ECX	%ecx
-#define	DL	%dl
-#define	DH	%dh
-#define	DX	%dx
-#define	EDX	%edx
-#define	BP	%bp
-#define	EBP	%ebp
-#define	SI	%si
-#define	ESI	%esi
-#define	DI	%di
-#define	EDI	%edi
-#define	SP	%sp
-#define	ESP	%esp
-#define	CS	%cs
-#define	SS	%ss
-#define	DS	%ds
-#define	ES	%es
-#define	FS	%fs
-#define	GS	%gs
+#define AL		%al
+#define AH		%ah
+#define AX		%ax
+#define EAX		%eax
+#define BL		%bl
+#define BH		%bh
+#define BX		%bx
+#define EBX		%ebx
+#define CL		%cl
+#define CH		%ch
+#define CX		%cx
+#define ECX		%ecx
+#define DL		%dl
+#define DH		%dh
+#define DX		%dx
+#define EDX		%edx
+#define BP		%bp
+#define EBP		%ebp
+#define SI		%si
+#define ESI		%esi
+#define DI		%di
+#define EDI		%edi
+#define SP		%sp
+#define ESP		%esp
+#define CS		%cs
+#define SS		%ss
+#define DS		%ds
+#define ES		%es
+#define FS		%fs
+#define GS		%gs
 /* Control Registers */
-#define	CR0	%cr0
-#define	CR1	%cr1
-#define	CR2	%cr2
-#define	CR3	%cr3
+#define CR0		%cr0
+#define CR1		%cr1
+#define CR2		%cr2
+#define CR3		%cr3
 /* Debug Registers */
-#define	DR0	%db0
-#define	DR1	%db1
-#define	DR2	%db2
-#define	DR3	%db3
-#define	DR4	%db4
-#define	DR5	%db5
-#define	DR6	%db6
-#define	DR7	%db7
+#define DR0		%db0
+#define DR1		%db1
+#define DR2		%db2
+#define DR3		%db3
+#define DR4		%db4
+#define DR5		%db5
+#define DR6		%db6
+#define DR7		%db7
 /* Floating-point Stack */
-#define _STX0 	%st(0)
-#define _STX1 	%st(1)
-#define _STX2 	%st(2)
-#define _STX3 	%st(3)
-#define _STX4 	%st(4)
-#define _STX5 	%st(5)
-#define _STX6 	%st(6)
-#define _STX7 	%st(7)
-#define ST(x) 	CONCAT(_STX,x)
+#define _STX0		%st(0)
+#define _STX1		%st(1)
+#define _STX2		%st(2)
+#define _STX3		%st(3)
+#define _STX4		%st(4)
+#define _STX5		%st(5)
+#define _STX6		%st(6)
+#define _STX7		%st(7)
+#define ST(x)		CONCAT(_STX,x)
 /* MMX Registers */
-#define MM0 	%mm0
-#define MM1 	%mm1
-#define MM2 	%mm2
-#define MM3 	%mm3
-#define MM4 	%mm4
-#define MM5 	%mm5
-#define MM6 	%mm6
-#define MM7 	%mm7
+#define MM0		%mm0
+#define MM1		%mm1
+#define MM2		%mm2
+#define MM3		%mm3
+#define MM4		%mm4
+#define MM5		%mm5
+#define MM6		%mm6
+#define MM7		%mm7
 /* SSE Registers */
-#define XMM0 	%xmm0
-#define XMM1 	%xmm1
-#define XMM2 	%xmm2
-#define XMM3 	%xmm3
-#define XMM4 	%xmm4
-#define XMM5 	%xmm5
-#define XMM6 	%xmm6
-#define XMM7 	%xmm7
-
-#define	AS_BEGIN
-#define	USE16
-#define	USE32
+#define XMM0		%xmm0
+#define XMM1		%xmm1
+#define XMM2		%xmm2
+#define XMM3		%xmm3
+#define XMM4		%xmm4
+#define XMM5		%xmm5
+#define XMM6		%xmm6
+#define XMM7		%xmm7
+
+#define AS_BEGIN
+#define USE16
+#define USE32
 
 #ifdef GNU_ASSEMBLER
 
-#define	ADDR_TOGGLE	aword
-#define	OPSZ_TOGGLE	word
+#define ADDR_TOGGLE	aword
+#define OPSZ_TOGGLE	word
 
-#define	CHOICE(a,b,c)	b
+#define CHOICE(a,b,c)	b
 
 #else
 /*
  * AT&T ASSEMBLER SYNTAX
  * *********************
  */
-#define	CHOICE(a,b,c)	a
+#define CHOICE(a,b,c)	a
 
-#define	ADDR_TOGGLE	addr16
-#define	OPSZ_TOGGLE	data16
+#define ADDR_TOGGLE	addr16
+#define OPSZ_TOGGLE	data16
 
 #endif /* GNU_ASSEMBLER */
 #endif /* ACK_ASSEMBLER */
 
 
 #if defined(__QNX__) || defined(Lynx) || (defined(SYSV) || defined(SVR4)) && !defined(ACK_ASSEMBLER) || defined(__ELF__) || defined(__GNU__)
-#define GLNAME(a)       a
+#define GLNAME(a)	a
 #else
-#define GLNAME(a)       CONCAT(_,a)
+#define GLNAME(a)	CONCAT(_,a)
 #endif
 
 
@@ -260,59 +260,63 @@
 /* Redefine assembler directives */
 /*********************************/
 #define GLOBL		CHOICE(.globl, .globl, .extern)
+#define GLOBAL		GLOBL
+#define EXTERN		GLOBL
 /*
-#define	ALIGNTEXT32	CHOICE(.align 32, .align ARG2(5,0x90), .align 32)
+#define ALIGNTEXT32	CHOICE(.align 32, .align ARG2(5,0x90), .align 32)
 */
-#define	ALIGNTEXT32	CHOICE(.align 32, .balign 32, .align 32)
-#define	ALIGNTEXT16	CHOICE(.align 16, .balign 16, .align 16)
-#define	ALIGNTEXT8	CHOICE(.align 8, .balign 8, .align 8)
-#define	ALIGNTEXT4	CHOICE(.align 4, .balign 4, .align 4)
-#define	ALIGNTEXT2	CHOICE(.align 2, .balign 2, .align 2)
+#define ALIGNTEXT32	CHOICE(.align 32, .balign 32, .align 32)
+#define ALIGNTEXT16	CHOICE(.align 16, .balign 16, .align 16)
+#define ALIGNTEXT8	CHOICE(.align 8, .balign 8, .align 8)
+#define ALIGNTEXT4	CHOICE(.align 4, .balign 4, .align 4)
+#define ALIGNTEXT2	CHOICE(.align 2, .balign 2, .align 2)
 /* ALIGNTEXT4ifNOP is the same as ALIGNTEXT4, but only if the space is
  * guaranteed to be filled with NOPs.  Otherwise it does nothing.
  */
-#define	ALIGNTEXT32ifNOP	CHOICE(.align 32, .balign ARG2(32,0x90), /*can't do it*/)
-#define	ALIGNTEXT16ifNOP	CHOICE(.align 16, .balign ARG2(16,0x90), /*can't do it*/)
-#define	ALIGNTEXT8ifNOP	CHOICE(.align 8, .balign ARG2(8,0x90), /*can't do it*/)
-#define	ALIGNTEXT4ifNOP	CHOICE(.align 4, .balign ARG2(4,0x90), /*can't do it*/)
-#define	ALIGNDATA32	CHOICE(.align 32, .balign ARG2(32,0x0), .align 32)
-#define	ALIGNDATA16	CHOICE(.align 16, .balign ARG2(16,0x0), .align 16)
-#define	ALIGNDATA8	CHOICE(.align 8, .balign ARG2(8,0x0), .align 8)
-#define	ALIGNDATA4	CHOICE(.align 4, .balign ARG2(4,0x0), .align 4)
-#define	ALIGNDATA2	CHOICE(.align 2, .balign ARG2(2,0x0), .align 2)
-#define	FILE(s)		CHOICE(.file s, .file s, .file s)
-#define	STRING(s)	CHOICE(.string s, .asciz s, .asciz s)
-#define	D_LONG		CHOICE(.long, .long, .data4)
-#define	D_WORD		CHOICE(.value, .short, .data2)
-#define	D_BYTE		CHOICE(.byte, .byte, .data1)
-#define	SPACE		CHOICE(.comm, .space, .space)
-#define	COMM		CHOICE(.comm, .comm, .comm)
-#define	SEG_DATA	CHOICE(.data, .data, .sect .data)
-#define	SEG_TEXT	CHOICE(.text, .text, .sect .text)
-#define	SEG_BSS		CHOICE(.bss, .bss, .sect .bss)
+#define ALIGNTEXT32ifNOP	CHOICE(.align 32, .balign ARG2(32,0x90), /*can't do it*/)
+#define ALIGNTEXT16ifNOP	CHOICE(.align 16, .balign ARG2(16,0x90), /*can't do it*/)
+#define ALIGNTEXT8ifNOP	CHOICE(.align 8, .balign ARG2(8,0x90), /*can't do it*/)
+#define ALIGNTEXT4ifNOP	CHOICE(.align 4, .balign ARG2(4,0x90), /*can't do it*/)
+#define ALIGNDATA32	CHOICE(.align 32, .balign ARG2(32,0x0), .align 32)
+#define ALIGNDATA16	CHOICE(.align 16, .balign ARG2(16,0x0), .align 16)
+#define ALIGNDATA8	CHOICE(.align 8, .balign ARG2(8,0x0), .align 8)
+#define ALIGNDATA4	CHOICE(.align 4, .balign ARG2(4,0x0), .align 4)
+#define ALIGNDATA2	CHOICE(.align 2, .balign ARG2(2,0x0), .align 2)
+#define FILE(s)		CHOICE(.file s, .file s, .file s)
+#define STRING(s)	CHOICE(.string s, .asciz s, .asciz s)
+#define D_LONG		CHOICE(.long, .long, .data4)
+#define D_WORD		CHOICE(.value, .short, .data2)
+#define D_BYTE		CHOICE(.byte, .byte, .data1)
+#define SPACE		CHOICE(.comm, .space, .space)
+#define COMM		CHOICE(.comm, .comm, .comm)
+#define SEG_DATA	CHOICE(.data, .data, .sect .data)
+#define SEG_TEXT	CHOICE(.text, .text, .sect .text)
+#define SEG_BSS		CHOICE(.bss, .bss, .sect .bss)
 
 #ifdef GNU_ASSEMBLER
-#define	D_SPACE(n)	. = . + n
+#define D_SPACE(n)	. = . + n
 #else
-#define	D_SPACE(n)	.space n
+#define D_SPACE(n)	.space n
 #endif
 
 /* Addressing Modes */
 /* Immediate Mode */
-#define	ADDR(a)		CHOICE(CONCAT($,a), CONCAT($,a), a)
-#define	CONST(a)	CHOICE(CONCAT($,a), CONCAT($,a), a)
+#define ADDR(a)		CHOICE(CONCAT($,a), CONCAT($,a), a)
+#define CONST(a)	CHOICE(CONCAT($,a), CONCAT($,a), a)
 
 /* Indirect Mode */
-#define	CONTENT(a)	CHOICE(a, a, (a))	 /* take contents of variable */
-#define	REGIND(a)	CHOICE((a), (a), (a))	 /* Register a indirect */
+#define CONTENT(a)	CHOICE(a, a, (a))	 /* take contents of variable */
+#define REGIND(a)	CHOICE((a), (a), (a))	 /* Register a indirect */
 /* Register b indirect plus displacement a */
-#define	REGOFF(a, b)	CHOICE(a(b), a(b), a(b))
+#define REGOFF(a, b)	CHOICE(a(b), a(b), a(b))
 /* Reg indirect Base + Index + Displacement  - this is mainly for 16-bit mode
  * which has no scaling
  */
-#define	REGBID(b,i,d)	CHOICE(d(b,i), d(b,i), d(b)(i))
+#define REGBID(b,i,d)	CHOICE(d(b,i), d(b,i), d(b)(i))
+/* Reg indirect Base + (Index * Scale) */
+#define REGBIS(b,i,s)	CHOICE((b,i,s), (b,i,s), (b)(i*s))
 /* Reg indirect Base + (Index * Scale) + Displacement */
-#define	REGBISD(b,i,s,d) CHOICE(d(b,i,s), d(b,i,s), d(b)(i*s))
+#define REGBISD(b,i,s,d) CHOICE(d(b,i,s), d(b,i,s), d(b)(i*s))
 /* Displaced Scaled Index: */
 #define REGDIS(d,i,s)	CHOICE(d(,i,s), d(,i,s), d(i * s))
 /* Indexed Base: */
@@ -328,10 +332,10 @@
  * eg. (CRT0_PM | CRT_EM)
  */
 
-#define	EXPR(a)		CHOICE([a], (a), [a])
-#define	ENOT(a)		CHOICE(0!a, ~a, ~a)
-#define	EMUL(a,b)	CHOICE(a\*b, a*b, a*b)
-#define	EDIV(a,b)	CHOICE(a\/b, a/b, a/b)
+#define EXPR(a)		CHOICE([a], (a), [a])
+#define ENOT(a)		CHOICE(0!a, ~a, ~a)
+#define EMUL(a,b)	CHOICE(a\*b, a*b, a*b)
+#define EDIV(a,b)	CHOICE(a\/b, a/b, a/b)
 
 /*
  * We have to beat the problem of commas within arguments to choice.
@@ -339,436 +343,436 @@
  * and other known cpp definitions evaluate arguments before substitution
  * so the following works.
  */
-#define	ARG2(a, b)	a,b
-#define	ARG3(a,b,c)	a,b,c
+#define ARG2(a, b)	a,b
+#define ARG3(a,b,c)	a,b,c
 
 /* Redefine assembler commands */
-#define	AAA		CHOICE(aaa, aaa, aaa)
-#define	AAD		CHOICE(aad, aad, aad)
-#define	AAM		CHOICE(aam, aam, aam)
-#define	AAS		CHOICE(aas, aas, aas)
-#define	ADC_L(a, b)	CHOICE(adcl ARG2(a,b), adcl ARG2(a,b), _LTOG adc ARG2(b,a))
-#define	ADC_W(a, b)	CHOICE(adcw ARG2(a,b), adcw ARG2(a,b), _WTOG adc ARG2(b,a))
-#define	ADC_B(a, b)	CHOICE(adcb ARG2(a,b), adcb ARG2(a,b), adcb ARG2(b,a))
-#define	ADD_L(a, b)	CHOICE(addl ARG2(a,b), addl ARG2(a,b), _LTOG add ARG2(b,a))
-#define	ADD_W(a, b)	CHOICE(addw ARG2(a,b), addw ARG2(a,b), _WTOG add ARG2(b,a))
-#define	ADD_B(a, b)	CHOICE(addb ARG2(a,b), addb ARG2(a,b), addb ARG2(b,a))
-#define	AND_L(a, b)	CHOICE(andl ARG2(a,b), andl ARG2(a,b), _LTOG and ARG2(b,a))
-#define	AND_W(a, b)	CHOICE(andw ARG2(a,b), andw ARG2(a,b), _WTOG and ARG2(b,a))
-#define	AND_B(a, b)	CHOICE(andb ARG2(a,b), andb ARG2(a,b), andb ARG2(b,a))
-#define	ARPL(a,b)	CHOICE(arpl ARG2(a,b), arpl ARG2(a,b), arpl ARG2(b,a))
-#define	BOUND_L(a, b)	CHOICE(boundl ARG2(a,b), boundl ARG2(b,a), _LTOG bound ARG2(b,a))
-#define	BOUND_W(a, b)	CHOICE(boundw ARG2(a,b), boundw ARG2(b,a), _WTOG bound ARG2(b,a))
-#define	BSF_L(a, b)	CHOICE(bsfl ARG2(a,b), bsfl ARG2(a,b), _LTOG bsf ARG2(b,a))
-#define	BSF_W(a, b)	CHOICE(bsfw ARG2(a,b), bsfw ARG2(a,b), _WTOG bsf ARG2(b,a))
-#define	BSR_L(a, b)	CHOICE(bsrl ARG2(a,b), bsrl ARG2(a,b), _LTOG bsr ARG2(b,a))
-#define	BSR_W(a, b)	CHOICE(bsrw ARG2(a,b), bsrw ARG2(a,b), _WTOG bsr ARG2(b,a))
-#define	BT_L(a, b)	CHOICE(btl ARG2(a,b), btl ARG2(a,b), _LTOG bt ARG2(b,a))
-#define	BT_W(a, b)	CHOICE(btw ARG2(a,b), btw ARG2(a,b), _WTOG bt ARG2(b,a))
-#define	BTC_L(a, b)	CHOICE(btcl ARG2(a,b), btcl ARG2(a,b), _LTOG btc ARG2(b,a))
-#define	BTC_W(a, b)	CHOICE(btcw ARG2(a,b), btcw ARG2(a,b), _WTOG btc ARG2(b,a))
-#define	BTR_L(a, b)	CHOICE(btrl ARG2(a,b), btrl ARG2(a,b), _LTOG btr ARG2(b,a))
-#define	BTR_W(a, b)	CHOICE(btrw ARG2(a,b), btrw ARG2(a,b), _WTOG btr ARG2(b,a))
-#define	BTS_L(a, b)	CHOICE(btsl ARG2(a,b), btsl ARG2(a,b), _LTOG bts ARG2(b,a))
-#define	BTS_W(a, b)	CHOICE(btsw ARG2(a,b), btsw ARG2(a,b), _WTOG bts ARG2(b,a))
-#define	CALL(a)		CHOICE(call a, call a, call a)
-#define	CALLF(s,a)	CHOICE(lcall ARG2(s,a), lcall ARG2(s,a), callf s:a)
-#define	CBW		CHOICE(cbtw, cbw, cbw)
-#define	CWDE		CHOICE(cwtd, cwde, cwde)
-#define	CLC		CHOICE(clc, clc, clc)
-#define	CLD		CHOICE(cld, cld, cld)
-#define	CLI		CHOICE(cli, cli, cli)
-#define	CLTS		CHOICE(clts, clts, clts)
-#define	CMC		CHOICE(cmc, cmc, cmc)
-#define	CMP_L(a, b)	CHOICE(cmpl ARG2(a,b), cmpl ARG2(a,b), _LTOG cmp ARG2(b,a))
-#define	CMP_W(a, b)	CHOICE(cmpw ARG2(a,b), cmpw ARG2(a,b), _WTOG cmp ARG2(b,a))
-#define	CMP_B(a, b)	CHOICE(cmpb ARG2(a,b), cmpb ARG2(a,b), cmpb ARG2(b,a))
-#define	CMPS_L		CHOICE(cmpsl, cmpsl, _LTOG cmps)
-#define	CMPS_W		CHOICE(cmpsw, cmpsw, _WTOG cmps)
-#define	CMPS_B		CHOICE(cmpsb, cmpsb, cmpsb)
-#define	CWD		CHOICE(cwtl, cwd, cwd)
-#define	CDQ		CHOICE(cltd, cdq, cdq)
-#define	DAA		CHOICE(daa, daa, daa)
-#define	DAS		CHOICE(das, das, das)
-#define	DEC_L(a)	CHOICE(decl a, decl a, _LTOG dec a)
-#define	DEC_W(a)	CHOICE(decw a, decw a, _WTOG dec a)
-#define	DEC_B(a)	CHOICE(decb a, decb a, decb a)
-#define	DIV_L(a)	CHOICE(divl a, divl a, div a)
-#define	DIV_W(a)	CHOICE(divw a, divw a, div a)
-#define	DIV_B(a)	CHOICE(divb a, divb a, divb a)
-#define	ENTER(a,b)	CHOICE(enter ARG2(a,b), enter ARG2(a,b), enter ARG2(b,a))
-#define	HLT		CHOICE(hlt, hlt, hlt)
-#define	IDIV_L(a)	CHOICE(idivl a, idivl a, _LTOG idiv a)
-#define	IDIV_W(a)	CHOICE(idivw a, idivw a, _WTOG idiv a)
-#define	IDIV_B(a)	CHOICE(idivb a, idivb a, idivb a)
+#define AAA		CHOICE(aaa, aaa, aaa)
+#define AAD		CHOICE(aad, aad, aad)
+#define AAM		CHOICE(aam, aam, aam)
+#define AAS		CHOICE(aas, aas, aas)
+#define ADC_L(a, b)	CHOICE(adcl ARG2(a,b), adcl ARG2(a,b), _LTOG adc ARG2(b,a))
+#define ADC_W(a, b)	CHOICE(adcw ARG2(a,b), adcw ARG2(a,b), _WTOG adc ARG2(b,a))
+#define ADC_B(a, b)	CHOICE(adcb ARG2(a,b), adcb ARG2(a,b), adcb ARG2(b,a))
+#define ADD_L(a, b)	CHOICE(addl ARG2(a,b), addl ARG2(a,b), _LTOG add ARG2(b,a))
+#define ADD_W(a, b)	CHOICE(addw ARG2(a,b), addw ARG2(a,b), _WTOG add ARG2(b,a))
+#define ADD_B(a, b)	CHOICE(addb ARG2(a,b), addb ARG2(a,b), addb ARG2(b,a))
+#define AND_L(a, b)	CHOICE(andl ARG2(a,b), andl ARG2(a,b), _LTOG and ARG2(b,a))
+#define AND_W(a, b)	CHOICE(andw ARG2(a,b), andw ARG2(a,b), _WTOG and ARG2(b,a))
+#define AND_B(a, b)	CHOICE(andb ARG2(a,b), andb ARG2(a,b), andb ARG2(b,a))
+#define ARPL(a,b)	CHOICE(arpl ARG2(a,b), arpl ARG2(a,b), arpl ARG2(b,a))
+#define BOUND_L(a, b)	CHOICE(boundl ARG2(a,b), boundl ARG2(b,a), _LTOG bound ARG2(b,a))
+#define BOUND_W(a, b)	CHOICE(boundw ARG2(a,b), boundw ARG2(b,a), _WTOG bound ARG2(b,a))
+#define BSF_L(a, b)	CHOICE(bsfl ARG2(a,b), bsfl ARG2(a,b), _LTOG bsf ARG2(b,a))
+#define BSF_W(a, b)	CHOICE(bsfw ARG2(a,b), bsfw ARG2(a,b), _WTOG bsf ARG2(b,a))
+#define BSR_L(a, b)	CHOICE(bsrl ARG2(a,b), bsrl ARG2(a,b), _LTOG bsr ARG2(b,a))
+#define BSR_W(a, b)	CHOICE(bsrw ARG2(a,b), bsrw ARG2(a,b), _WTOG bsr ARG2(b,a))
+#define BT_L(a, b)	CHOICE(btl ARG2(a,b), btl ARG2(a,b), _LTOG bt ARG2(b,a))
+#define BT_W(a, b)	CHOICE(btw ARG2(a,b), btw ARG2(a,b), _WTOG bt ARG2(b,a))
+#define BTC_L(a, b)	CHOICE(btcl ARG2(a,b), btcl ARG2(a,b), _LTOG btc ARG2(b,a))
+#define BTC_W(a, b)	CHOICE(btcw ARG2(a,b), btcw ARG2(a,b), _WTOG btc ARG2(b,a))
+#define BTR_L(a, b)	CHOICE(btrl ARG2(a,b), btrl ARG2(a,b), _LTOG btr ARG2(b,a))
+#define BTR_W(a, b)	CHOICE(btrw ARG2(a,b), btrw ARG2(a,b), _WTOG btr ARG2(b,a))
+#define BTS_L(a, b)	CHOICE(btsl ARG2(a,b), btsl ARG2(a,b), _LTOG bts ARG2(b,a))
+#define BTS_W(a, b)	CHOICE(btsw ARG2(a,b), btsw ARG2(a,b), _WTOG bts ARG2(b,a))
+#define CALL(a)		CHOICE(call a, call a, call a)
+#define CALLF(s,a)	CHOICE(lcall ARG2(s,a), lcall ARG2(s,a), callf s:a)
+#define CBW		CHOICE(cbtw, cbw, cbw)
+#define CWDE		CHOICE(cwtd, cwde, cwde)
+#define CLC		CHOICE(clc, clc, clc)
+#define CLD		CHOICE(cld, cld, cld)
+#define CLI		CHOICE(cli, cli, cli)
+#define CLTS		CHOICE(clts, clts, clts)
+#define CMC		CHOICE(cmc, cmc, cmc)
+#define CMP_L(a, b)	CHOICE(cmpl ARG2(a,b), cmpl ARG2(a,b), _LTOG cmp ARG2(b,a))
+#define CMP_W(a, b)	CHOICE(cmpw ARG2(a,b), cmpw ARG2(a,b), _WTOG cmp ARG2(b,a))
+#define CMP_B(a, b)	CHOICE(cmpb ARG2(a,b), cmpb ARG2(a,b), cmpb ARG2(b,a))
+#define CMPS_L		CHOICE(cmpsl, cmpsl, _LTOG cmps)
+#define CMPS_W		CHOICE(cmpsw, cmpsw, _WTOG cmps)
+#define CMPS_B		CHOICE(cmpsb, cmpsb, cmpsb)
+#define CWD		CHOICE(cwtl, cwd, cwd)
+#define CDQ		CHOICE(cltd, cdq, cdq)
+#define DAA		CHOICE(daa, daa, daa)
+#define DAS		CHOICE(das, das, das)
+#define DEC_L(a)	CHOICE(decl a, decl a, _LTOG dec a)
+#define DEC_W(a)	CHOICE(decw a, decw a, _WTOG dec a)
+#define DEC_B(a)	CHOICE(decb a, decb a, decb a)
+#define DIV_L(a)	CHOICE(divl a, divl a, div a)
+#define DIV_W(a)	CHOICE(divw a, divw a, div a)
+#define DIV_B(a)	CHOICE(divb a, divb a, divb a)
+#define ENTER(a,b)	CHOICE(enter ARG2(a,b), enter ARG2(a,b), enter ARG2(b,a))
+#define HLT		CHOICE(hlt, hlt, hlt)
+#define IDIV_L(a)	CHOICE(idivl a, idivl a, _LTOG idiv a)
+#define IDIV_W(a)	CHOICE(idivw a, idivw a, _WTOG idiv a)
+#define IDIV_B(a)	CHOICE(idivb a, idivb a, idivb a)
 /* More forms than this for imul!! */
-#define	IMUL_L(a, b)	CHOICE(imull ARG2(a,b), imull ARG2(a,b), _LTOG imul ARG2(b,a))
-#define	IMUL_W(a, b)	CHOICE(imulw ARG2(a,b), imulw ARG2(a,b), _WTOG imul ARG2(b,a))
-#define	IMUL_B(a)	CHOICE(imulb a, imulb a, imulb a)
-#define	IN_L		CHOICE(inl (DX), inl ARG2(DX,EAX), _LTOG in DX)
-#define	IN_W		CHOICE(inw (DX), inw ARG2(DX,AX), _WTOG in DX)
-#define	IN_B		CHOICE(inb (DX), inb ARG2(DX,AL), inb DX)
+#define IMUL_L(a, b)	CHOICE(imull ARG2(a,b), imull ARG2(a,b), _LTOG imul ARG2(b,a))
+#define IMUL_W(a, b)	CHOICE(imulw ARG2(a,b), imulw ARG2(a,b), _WTOG imul ARG2(b,a))
+#define IMUL_B(a)	CHOICE(imulb a, imulb a, imulb a)
+#define IN_L		CHOICE(inl (DX), inl ARG2(DX,EAX), _LTOG in DX)
+#define IN_W		CHOICE(inw (DX), inw ARG2(DX,AX), _WTOG in DX)
+#define IN_B		CHOICE(inb (DX), inb ARG2(DX,AL), inb DX)
 /* Please AS code writer: use the following ONLY, if you refer to ports<256
  * directly, but not in IN1_W(DX), for instance, even if IN1_ looks nicer
  */
 #if defined (sun)
-#define IN1_L(a)    CHOICE(inl (a), inl ARG2(a,EAX), _LTOG in a)
-#define IN1_W(a)    CHOICE(inw (a), inw ARG2(a,AX), _WTOG in a)
-#define IN1_B(a)    CHOICE(inb (a), inb ARG2(a,AL), inb a)
+#define IN1_L(a)	CHOICE(inl (a), inl ARG2(a,EAX), _LTOG in a)
+#define IN1_W(a)	CHOICE(inw (a), inw ARG2(a,AX), _WTOG in a)
+#define IN1_B(a)	CHOICE(inb (a), inb ARG2(a,AL), inb a)
 #else
-#define	IN1_L(a)	CHOICE(inl a, inl ARG2(a,EAX), _LTOG in a)
-#define	IN1_W(a)	CHOICE(inw a, inw ARG2(a,AX), _WTOG in a)
-#define	IN1_B(a)	CHOICE(inb a, inb ARG2(a,AL), inb a)
+#define IN1_L(a)	CHOICE(inl a, inl ARG2(a,EAX), _LTOG in a)
+#define IN1_W(a)	CHOICE(inw a, inw ARG2(a,AX), _WTOG in a)
+#define IN1_B(a)	CHOICE(inb a, inb ARG2(a,AL), inb a)
 #endif
-#define	INC_L(a)	CHOICE(incl a, incl a, _LTOG inc a)
-#define	INC_W(a)	CHOICE(incw a, incw a, _WTOG inc a)
-#define	INC_B(a)	CHOICE(incb a, incb a, incb a)
-#define	INS_L		CHOICE(insl, insl, _LTOG ins)
-#define	INS_W		CHOICE(insw, insw, _WTOG ins)
-#define	INS_B		CHOICE(insb, insb, insb)
-#define	INT(a)		CHOICE(int a, int a, int a)
-#define	INT3		CHOICE(int CONST(3), int3, int CONST(3))
-#define	INTO		CHOICE(into, into, into)
-#define	IRET		CHOICE(iret, iret, iret)
-#define	IRETD		CHOICE(iret, iret, iretd)
-#define	JA(a)		CHOICE(ja a, ja a, ja a)
-#define	JAE(a)		CHOICE(jae a, jae a, jae a)
-#define	JB(a)		CHOICE(jb a, jb a, jb a)
-#define	JBE(a)		CHOICE(jbe a, jbe a, jbe a)
-#define	JC(a)		CHOICE(jc a, jc a, jc a)
-#define	JE(a)		CHOICE(je a, je a, je a)
-#define	JG(a)		CHOICE(jg a, jg a, jg a)
-#define	JGE(a)		CHOICE(jge a, jge a, jge a)
-#define	JL(a)		CHOICE(jl a, jl a, jl a)
-#define	JLE(a)		CHOICE(jle a, jle a, jle a)
-#define	JNA(a)		CHOICE(jna a, jna a, jna a)
-#define	JNAE(a)		CHOICE(jnae a, jnae a, jnae a)
-#define	JNB(a)		CHOICE(jnb a, jnb a, jnb a)
-#define	JNBE(a)		CHOICE(jnbe a, jnbe a, jnbe a)
-#define	JNC(a)		CHOICE(jnc a, jnc a, jnc a)
-#define	JNE(a)		CHOICE(jne a, jne a, jne a)
-#define	JNG(a)		CHOICE(jng a, jng a, jng a)
-#define	JNGE(a)		CHOICE(jnge a, jnge a, jnge a)
-#define	JNL(a)		CHOICE(jnl a, jnl a, jnl a)
-#define	JNLE(a)		CHOICE(jnle a, jnle a, jnle a)
-#define	JNO(a)		CHOICE(jno a, jno a, jno a)
-#define	JNP(a)		CHOICE(jnp a, jnp a, jnp a)
-#define	JNS(a)		CHOICE(jns a, jns a, jns a)
-#define	JNZ(a)		CHOICE(jnz a, jnz a, jnz a)
-#define	JO(a)		CHOICE(jo a, jo a, jo a)
-#define	JP(a)		CHOICE(jp a, jp a, jp a)
-#define	JPE(a)		CHOICE(jpe a, jpe a, jpe a)
-#define	JPO(a)		CHOICE(jpo a, jpo a, jpo a)
-#define	JS(a)		CHOICE(js a, js a, js a)
-#define	JZ(a)		CHOICE(jz a, jz a, jz a)
-#define	JMP(a)		CHOICE(jmp a, jmp a, jmp a)
-#define	JMPF(s,a)	CHOICE(ljmp ARG2(s,a), ljmp ARG2(s,a), jmpf s:a)
-#define	LAHF		CHOICE(lahf, lahf, lahf)
+#define INC_L(a)	CHOICE(incl a, incl a, _LTOG inc a)
+#define INC_W(a)	CHOICE(incw a, incw a, _WTOG inc a)
+#define INC_B(a)	CHOICE(incb a, incb a, incb a)
+#define INS_L		CHOICE(insl, insl, _LTOG ins)
+#define INS_W		CHOICE(insw, insw, _WTOG ins)
+#define INS_B		CHOICE(insb, insb, insb)
+#define INT(a)		CHOICE(int a, int a, int a)
+#define INT3		CHOICE(int CONST(3), int3, int CONST(3))
+#define INTO		CHOICE(into, into, into)
+#define IRET		CHOICE(iret, iret, iret)
+#define IRETD		CHOICE(iret, iret, iretd)
+#define JA(a)		CHOICE(ja a, ja a, ja a)
+#define JAE(a)		CHOICE(jae a, jae a, jae a)
+#define JB(a)		CHOICE(jb a, jb a, jb a)
+#define JBE(a)		CHOICE(jbe a, jbe a, jbe a)
+#define JC(a)		CHOICE(jc a, jc a, jc a)
+#define JE(a)		CHOICE(je a, je a, je a)
+#define JG(a)		CHOICE(jg a, jg a, jg a)
+#define JGE(a)		CHOICE(jge a, jge a, jge a)
+#define JL(a)		CHOICE(jl a, jl a, jl a)
+#define JLE(a)		CHOICE(jle a, jle a, jle a)
+#define JNA(a)		CHOICE(jna a, jna a, jna a)
+#define JNAE(a)		CHOICE(jnae a, jnae a, jnae a)
+#define JNB(a)		CHOICE(jnb a, jnb a, jnb a)
+#define JNBE(a)		CHOICE(jnbe a, jnbe a, jnbe a)
+#define JNC(a)		CHOICE(jnc a, jnc a, jnc a)
+#define JNE(a)		CHOICE(jne a, jne a, jne a)
+#define JNG(a)		CHOICE(jng a, jng a, jng a)
+#define JNGE(a)		CHOICE(jnge a, jnge a, jnge a)
+#define JNL(a)		CHOICE(jnl a, jnl a, jnl a)
+#define JNLE(a)		CHOICE(jnle a, jnle a, jnle a)
+#define JNO(a)		CHOICE(jno a, jno a, jno a)
+#define JNP(a)		CHOICE(jnp a, jnp a, jnp a)
+#define JNS(a)		CHOICE(jns a, jns a, jns a)
+#define JNZ(a)		CHOICE(jnz a, jnz a, jnz a)
+#define JO(a)		CHOICE(jo a, jo a, jo a)
+#define JP(a)		CHOICE(jp a, jp a, jp a)
+#define JPE(a)		CHOICE(jpe a, jpe a, jpe a)
+#define JPO(a)		CHOICE(jpo a, jpo a, jpo a)
+#define JS(a)		CHOICE(js a, js a, js a)
+#define JZ(a)		CHOICE(jz a, jz a, jz a)
+#define JMP(a)		CHOICE(jmp a, jmp a, jmp a)
+#define JMPF(s,a)	CHOICE(ljmp ARG2(s,a), ljmp ARG2(s,a), jmpf s:a)
+#define LAHF		CHOICE(lahf, lahf, lahf)
 #if !defined(_REAL_MODE) && !defined(_V86_MODE)
-#define	LAR(a, b)	CHOICE(lar ARG2(a, b), lar ARG2(a, b), lar ARG2(b, a))
+#define LAR(a, b)	CHOICE(lar ARG2(a, b), lar ARG2(a, b), lar ARG2(b, a))
 #endif
-#define	LEA_L(a, b)	CHOICE(leal ARG2(a,b), leal ARG2(a,b), _LTOG lea ARG2(b,a))
-#define	LEA_W(a, b)	CHOICE(leaw ARG2(a,b), leaw ARG2(a,b), _WTOG lea ARG2(b,a))
-#define	LEAVE		CHOICE(leave, leave, leave)
-#define	LGDT(a)		CHOICE(lgdt a, lgdt a, lgdt a)
-#define	LIDT(a)		CHOICE(lidt a, lidt a, lidt a)
-#define	LDS(a, b)	CHOICE(ldsl ARG2(a,b), lds ARG2(a,b), lds ARG2(b,a))
-#define	LES(a, b)	CHOICE(lesl ARG2(a,b), les ARG2(a,b), les ARG2(b,a))
-#define	LFS(a, b)	CHOICE(lfsl ARG2(a,b), lfs ARG2(a,b), lfs ARG2(b,a))
-#define	LGS(a, b)	CHOICE(lgsl ARG2(a,b), lgs ARG2(a,b), lgs ARG2(b,a))
-#define	LSS(a, b)	CHOICE(lssl ARG2(a,b), lss ARG2(a,b), lss ARG2(b,a))
-#define	LLDT(a)		CHOICE(lldt a, lldt a, lldt a)
-#define	LMSW(a)		CHOICE(lmsw a, lmsw a, lmsw a)
+#define LEA_L(a, b)	CHOICE(leal ARG2(a,b), leal ARG2(a,b), _LTOG lea ARG2(b,a))
+#define LEA_W(a, b)	CHOICE(leaw ARG2(a,b), leaw ARG2(a,b), _WTOG lea ARG2(b,a))
+#define LEAVE		CHOICE(leave, leave, leave)
+#define LGDT(a)		CHOICE(lgdt a, lgdt a, lgdt a)
+#define LIDT(a)		CHOICE(lidt a, lidt a, lidt a)
+#define LDS(a, b)	CHOICE(ldsl ARG2(a,b), lds ARG2(a,b), lds ARG2(b,a))
+#define LES(a, b)	CHOICE(lesl ARG2(a,b), les ARG2(a,b), les ARG2(b,a))
+#define LFS(a, b)	CHOICE(lfsl ARG2(a,b), lfs ARG2(a,b), lfs ARG2(b,a))
+#define LGS(a, b)	CHOICE(lgsl ARG2(a,b), lgs ARG2(a,b), lgs ARG2(b,a))
+#define LSS(a, b)	CHOICE(lssl ARG2(a,b), lss ARG2(a,b), lss ARG2(b,a))
+#define LLDT(a)		CHOICE(lldt a, lldt a, lldt a)
+#define LMSW(a)		CHOICE(lmsw a, lmsw a, lmsw a)
 #define LOCK		CHOICE(lock, lock, lock)
-#define	LODS_L		CHOICE(lodsl, lodsl, _LTOG lods)
-#define	LODS_W		CHOICE(lodsw, lodsw, _WTOG lods)
-#define	LODS_B		CHOICE(lodsb, lodsb, lodsb)
-#define	LOOP(a)		CHOICE(loop a, loop a, loop a)
-#define	LOOPE(a)	CHOICE(loope a, loope a, loope a)
-#define	LOOPZ(a)	CHOICE(loopz a, loopz a, loopz a)
-#define	LOOPNE(a)	CHOICE(loopne a, loopne a, loopne a)
-#define	LOOPNZ(a)	CHOICE(loopnz a, loopnz a, loopnz a)
+#define LODS_L		CHOICE(lodsl, lodsl, _LTOG lods)
+#define LODS_W		CHOICE(lodsw, lodsw, _WTOG lods)
+#define LODS_B		CHOICE(lodsb, lodsb, lodsb)
+#define LOOP(a)		CHOICE(loop a, loop a, loop a)
+#define LOOPE(a)	CHOICE(loope a, loope a, loope a)
+#define LOOPZ(a)	CHOICE(loopz a, loopz a, loopz a)
+#define LOOPNE(a)	CHOICE(loopne a, loopne a, loopne a)
+#define LOOPNZ(a)	CHOICE(loopnz a, loopnz a, loopnz a)
 #if !defined(_REAL_MODE) && !defined(_V86_MODE)
-#define	LSL(a, b)	CHOICE(lsl ARG2(a,b), lsl ARG2(a,b), lsl ARG2(b,a))
+#define LSL(a, b)	CHOICE(lsl ARG2(a,b), lsl ARG2(a,b), lsl ARG2(b,a))
 #endif
-#define	LTR(a)		CHOICE(ltr a, ltr a, ltr a)
-#define	MOV_SR(a, b)	CHOICE(movw ARG2(a,b), mov ARG2(a,b), mov ARG2(b,a))
-#define	MOV_L(a, b)	CHOICE(movl ARG2(a,b), movl ARG2(a,b), _LTOG mov ARG2(b,a))
-#define	MOV_W(a, b)	CHOICE(movw ARG2(a,b), movw ARG2(a,b), _WTOG mov ARG2(b,a))
-#define	MOV_B(a, b)	CHOICE(movb ARG2(a,b), movb ARG2(a,b), movb ARG2(b,a))
-#define	MOVS_L		CHOICE(movsl, movsl, _LTOG movs)
-#define	MOVS_W		CHOICE(movsw, movsw, _WTOG movs)
-#define	MOVS_B		CHOICE(movsb, movsb, movsb)
-#define	MOVSX_BL(a, b)	CHOICE(movsbl ARG2(a,b), movsbl ARG2(a,b), movsx ARG2(b,a))
-#define	MOVSX_BW(a, b)	CHOICE(movsbw ARG2(a,b), movsbw ARG2(a,b), movsx ARG2(b,a))
-#define	MOVSX_WL(a, b)	CHOICE(movswl ARG2(a,b), movswl ARG2(a,b), movsx ARG2(b,a))
-#define	MOVZX_BL(a, b)	CHOICE(movzbl ARG2(a,b), movzbl ARG2(a,b), movzx ARG2(b,a))
-#define	MOVZX_BW(a, b)	CHOICE(movzbw ARG2(a,b), movzbw ARG2(a,b), movzx ARG2(b,a))
-#define	MOVZX_WL(a, b)	CHOICE(movzwl ARG2(a,b), movzwl ARG2(a,b), movzx ARG2(b,a))
-#define	MUL_L(a)	CHOICE(mull a, mull a, _LTOG mul a)
-#define	MUL_W(a)	CHOICE(mulw a, mulw a, _WTOG mul a)
-#define	MUL_B(a)	CHOICE(mulb a, mulb a, mulb a)
-#define	NEG_L(a)	CHOICE(negl a, negl a, _LTOG neg a)
-#define	NEG_W(a)	CHOICE(negw a, negw a, _WTOG neg a)
-#define	NEG_B(a)	CHOICE(negb a, negb a, negb a)
-#define	NOP		CHOICE(nop, nop, nop)
-#define	NOT_L(a)	CHOICE(notl a, notl a, _LTOG not a)
-#define	NOT_W(a)	CHOICE(notw a, notw a, _WTOG not a)
-#define	NOT_B(a)	CHOICE(notb a, notb a, notb a)
-#define	OR_L(a,b)	CHOICE(orl ARG2(a,b), orl ARG2(a,b), _LTOG or ARG2(b,a))
-#define	OR_W(a,b)	CHOICE(orw ARG2(a,b), orw ARG2(a,b), _WTOG or ARG2(b,a))
-#define	OR_B(a,b)	CHOICE(orb ARG2(a,b), orb ARG2(a,b), orb ARG2(b,a))
-#define	OUT_L		CHOICE(outl (DX), outl ARG2(EAX,DX), _LTOG out DX)
-#define	OUT_W		CHOICE(outw (DX), outw ARG2(AX,DX), _WTOG out DX)
-#define	OUT_B		CHOICE(outb (DX), outb ARG2(AL,DX), outb DX)
+#define LTR(a)		CHOICE(ltr a, ltr a, ltr a)
+#define MOV_SR(a, b)	CHOICE(movw ARG2(a,b), mov ARG2(a,b), mov ARG2(b,a))
+#define MOV_L(a, b)	CHOICE(movl ARG2(a,b), movl ARG2(a,b), _LTOG mov ARG2(b,a))
+#define MOV_W(a, b)	CHOICE(movw ARG2(a,b), movw ARG2(a,b), _WTOG mov ARG2(b,a))
+#define MOV_B(a, b)	CHOICE(movb ARG2(a,b), movb ARG2(a,b), movb ARG2(b,a))
+#define MOVS_L		CHOICE(movsl, movsl, _LTOG movs)
+#define MOVS_W		CHOICE(movsw, movsw, _WTOG movs)
+#define MOVS_B		CHOICE(movsb, movsb, movsb)
+#define MOVSX_BL(a, b)	CHOICE(movsbl ARG2(a,b), movsbl ARG2(a,b), movsx ARG2(b,a))
+#define MOVSX_BW(a, b)	CHOICE(movsbw ARG2(a,b), movsbw ARG2(a,b), movsx ARG2(b,a))
+#define MOVSX_WL(a, b)	CHOICE(movswl ARG2(a,b), movswl ARG2(a,b), movsx ARG2(b,a))
+#define MOVZX_BL(a, b)	CHOICE(movzbl ARG2(a,b), movzbl ARG2(a,b), movzx ARG2(b,a))
+#define MOVZX_BW(a, b)	CHOICE(movzbw ARG2(a,b), movzbw ARG2(a,b), movzx ARG2(b,a))
+#define MOVZX_WL(a, b)	CHOICE(movzwl ARG2(a,b), movzwl ARG2(a,b), movzx ARG2(b,a))
+#define MUL_L(a)	CHOICE(mull a, mull a, _LTOG mul a)
+#define MUL_W(a)	CHOICE(mulw a, mulw a, _WTOG mul a)
+#define MUL_B(a)	CHOICE(mulb a, mulb a, mulb a)
+#define NEG_L(a)	CHOICE(negl a, negl a, _LTOG neg a)
+#define NEG_W(a)	CHOICE(negw a, negw a, _WTOG neg a)
+#define NEG_B(a)	CHOICE(negb a, negb a, negb a)
+#define NOP		CHOICE(nop, nop, nop)
+#define NOT_L(a)	CHOICE(notl a, notl a, _LTOG not a)
+#define NOT_W(a)	CHOICE(notw a, notw a, _WTOG not a)
+#define NOT_B(a)	CHOICE(notb a, notb a, notb a)
+#define OR_L(a,b)	CHOICE(orl ARG2(a,b), orl ARG2(a,b), _LTOG or ARG2(b,a))
+#define OR_W(a,b)	CHOICE(orw ARG2(a,b), orw ARG2(a,b), _WTOG or ARG2(b,a))
+#define OR_B(a,b)	CHOICE(orb ARG2(a,b), orb ARG2(a,b), orb ARG2(b,a))
+#define OUT_L		CHOICE(outl (DX), outl ARG2(EAX,DX), _LTOG out DX)
+#define OUT_W		CHOICE(outw (DX), outw ARG2(AX,DX), _WTOG out DX)
+#define OUT_B		CHOICE(outb (DX), outb ARG2(AL,DX), outb DX)
 /* Please AS code writer: use the following ONLY, if you refer to ports<256
  * directly, but not in OUT1_W(DX), for instance, even if OUT1_ looks nicer
  */
-#define	OUT1_L(a)	CHOICE(outl (a), outl ARG2(EAX,a), _LTOG out a)
-#define	OUT1_W(a)	CHOICE(outw (a), outw ARG2(AX,a), _WTOG out a)
-#define	OUT1_B(a)	CHOICE(outb (a), outb ARG2(AL,a), outb a)
-#define	OUTS_L		CHOICE(outsl, outsl, _LTOG outs)
-#define	OUTS_W		CHOICE(outsw, outsw, _WTOG outs)
-#define	OUTS_B		CHOICE(outsb, outsb, outsb)
-#define	POP_SR(a)	CHOICE(pop a, pop a, pop a)
-#define	POP_L(a)	CHOICE(popl a, popl a, _LTOG pop a)
-#define	POP_W(a)	CHOICE(popw a, popw a, _WTOG pop a)
-#define	POPA_L		CHOICE(popal, popal, _LTOG popa)
-#define	POPA_W		CHOICE(popaw, popaw, _WTOG popa)
-#define	POPF_L		CHOICE(popfl, popfl, _LTOG popf)
-#define	POPF_W		CHOICE(popfw, popfw, _WTOG popf)
-#define	PUSH_SR(a)	CHOICE(push a, push a, push a)
-#define	PUSH_L(a)	CHOICE(pushl a, pushl a, _LTOG push a)
-#define	PUSH_W(a)	CHOICE(pushw a, pushw a, _WTOG push a)
-#define	PUSH_B(a)	CHOICE(push a, pushb a, push a)
-#define	PUSHA_L		CHOICE(pushal, pushal, _LTOG pusha)
-#define	PUSHA_W		CHOICE(pushaw, pushaw, _WTOG pusha)
-#define	PUSHF_L		CHOICE(pushfl, pushfl, _LTOG pushf)
-#define	PUSHF_W		CHOICE(pushfw, pushfw, _WTOG pushf)
-#define	RCL_L(a, b)	CHOICE(rcll ARG2(a,b), rcll ARG2(a,b), _LTOG rcl ARG2(b,a))
-#define	RCL_W(a, b)	CHOICE(rclw ARG2(a,b), rclw ARG2(a,b), _WTOG rcl ARG2(b,a))
-#define	RCL_B(a, b)	CHOICE(rclb ARG2(a,b), rclb ARG2(a,b), rclb ARG2(b,a))
-#define	RCR_L(a, b)	CHOICE(rcrl ARG2(a,b), rcrl ARG2(a,b), _LTOG rcr ARG2(b,a))
-#define	RCR_W(a, b)	CHOICE(rcrw ARG2(a,b), rcrw ARG2(a,b), _WTOG rcr ARG2(b,a))
-#define	RCR_B(a, b)	CHOICE(rcrb ARG2(a,b), rcrb ARG2(a,b), rcrb ARG2(b,a))
-#define	ROL_L(a, b)	CHOICE(roll ARG2(a,b), roll ARG2(a,b), _LTOG rol ARG2(b,a))
-#define	ROL_W(a, b)	CHOICE(rolw ARG2(a,b), rolw ARG2(a,b), _WTOG rol ARG2(b,a))
-#define	ROL_B(a, b)	CHOICE(rolb ARG2(a,b), rolb ARG2(a,b), rolb ARG2(b,a))
-#define	ROR_L(a, b)	CHOICE(rorl ARG2(a,b), rorl ARG2(a,b), _LTOG ror ARG2(b,a))
-#define	ROR_W(a, b)	CHOICE(rorw ARG2(a,b), rorw ARG2(a,b), _WTOG ror ARG2(b,a))
-#define	ROR_B(a, b)	CHOICE(rorb ARG2(a,b), rorb ARG2(a,b), rorb ARG2(b,a))
-#define	REP		CHOICE(rep ;, rep ;, repe)
-#define	REPE		CHOICE(repz ;, repe ;, repe)
-#define	REPNE		CHOICE(repnz ;, repne ;, repne)
-#define	REPNZ		REPNE
-#define	REPZ		REPE
-#define	RET		CHOICE(ret, ret, ret)
-#define	SAHF		CHOICE(sahf, sahf, sahf)
-#define	SAL_L(a, b)	CHOICE(sall ARG2(a,b), sall ARG2(a,b), _LTOG sal ARG2(b,a))
-#define	SAL_W(a, b)	CHOICE(salw ARG2(a,b), salw ARG2(a,b), _WTOG sal ARG2(b,a))
-#define	SAL_B(a, b)	CHOICE(salb ARG2(a,b), salb ARG2(a,b), salb ARG2(b,a))
-#define	SAR_L(a, b)	CHOICE(sarl ARG2(a,b), sarl ARG2(a,b), _LTOG sar ARG2(b,a))
-#define	SAR_W(a, b)	CHOICE(sarw ARG2(a,b), sarw ARG2(a,b), _WTOG sar ARG2(b,a))
-#define	SAR_B(a, b)	CHOICE(sarb ARG2(a,b), sarb ARG2(a,b), sarb ARG2(b,a))
-#define	SBB_L(a, b)	CHOICE(sbbl ARG2(a,b), sbbl ARG2(a,b), _LTOG sbb ARG2(b,a))
-#define	SBB_W(a, b)	CHOICE(sbbw ARG2(a,b), sbbw ARG2(a,b), _WTOG sbb ARG2(b,a))
-#define	SBB_B(a, b)	CHOICE(sbbb ARG2(a,b), sbbb ARG2(a,b), sbbb ARG2(b,a))
-#define	SCAS_L		CHOICE(scasl, scasl, _LTOG scas)
-#define	SCAS_W		CHOICE(scasw, scasw, _WTOG scas)
-#define	SCAS_B		CHOICE(scasb, scasb, scasb)
-#define	SETA(a)		CHOICE(seta a, seta a, seta a)
-#define	SETAE(a)	CHOICE(setae a, setae a, setae a)
-#define	SETB(a)		CHOICE(setb a, setb a, setb a)
-#define	SETBE(a)	CHOICE(setbe a, setbe a, setbe a)
-#define	SETC(a)		CHOICE(setc a, setb a, setb a)
-#define	SETE(a)		CHOICE(sete a, sete a, sete a)
-#define	SETG(a)		CHOICE(setg a, setg a, setg a)
-#define	SETGE(a)	CHOICE(setge a, setge a, setge a)
-#define	SETL(a)		CHOICE(setl a, setl a, setl a)
-#define	SETLE(a)	CHOICE(setle a, setle a, setle a)
-#define	SETNA(a)	CHOICE(setna a, setna a, setna a)
-#define	SETNAE(a)	CHOICE(setnae a, setnae a, setnae a)
-#define	SETNB(a)	CHOICE(setnb a, setnb a, setnb a)
-#define	SETNBE(a)	CHOICE(setnbe a, setnbe a, setnbe a)
-#define	SETNC(a)	CHOICE(setnc a, setnb a, setnb a)
-#define	SETNE(a)	CHOICE(setne a, setne a, setne a)
-#define	SETNG(a)	CHOICE(setng a, setng a, setng a)
-#define	SETNGE(a)	CHOICE(setnge a, setnge a, setnge a)
-#define	SETNL(a)	CHOICE(setnl a, setnl a, setnl a)
-#define	SETNLE(a)	CHOICE(setnle a, setnle a, setnle a)
-#define	SETNO(a)	CHOICE(setno a, setno a, setno a)
-#define	SETNP(a)	CHOICE(setnp a, setnp a, setnp a)
-#define	SETNS(a)	CHOICE(setns a, setns a, setna a)
-#define	SETNZ(a)	CHOICE(setnz a, setnz a, setnz a)
-#define	SETO(a)		CHOICE(seto a, seto a, seto a)
-#define	SETP(a)		CHOICE(setp a, setp a, setp a)
-#define	SETPE(a)	CHOICE(setpe a, setpe a, setpe a)
-#define	SETPO(a)	CHOICE(setpo a, setpo a, setpo a)
-#define	SETS(a)		CHOICE(sets a, sets a, seta a)
-#define	SETZ(a)		CHOICE(setz a, setz a, setz a)
-#define	SGDT(a)		CHOICE(sgdt a, sgdt a, sgdt a)
-#define	SIDT(a)		CHOICE(sidt a, sidt a, sidt a)
-#define	SHL_L(a, b)	CHOICE(shll ARG2(a,b), shll ARG2(a,b), _LTOG shl ARG2(b,a))
-#define	SHL_W(a, b)	CHOICE(shlw ARG2(a,b), shlw ARG2(a,b), _WTOG shl ARG2(b,a))
-#define	SHL_B(a, b)	CHOICE(shlb ARG2(a,b), shlb ARG2(a,b), shlb ARG2(b,a))
-#define	SHLD_L(a,b,c)	CHOICE(shldl ARG3(a,b,c), shldl ARG3(a,b,c), _LTOG shld ARG3(c,b,a))
-#define	SHLD2_L(a,b)	CHOICE(shldl ARG2(a,b), shldl ARG3(CL,a,b), _LTOG shld ARG3(b,a,CL))
-#define	SHLD_W(a,b,c)	CHOICE(shldw ARG3(a,b,c), shldw ARG3(a,b,c), _WTOG shld ARG3(c,b,a))
-#define	SHLD2_W(a,b)	CHOICE(shldw ARG2(a,b), shldw ARG3(CL,a,b), _WTOG shld ARG3(b,a,CL))
-#define	SHR_L(a, b)	CHOICE(shrl ARG2(a,b), shrl ARG2(a,b), _LTOG shr ARG2(b,a))
-#define	SHR_W(a, b)	CHOICE(shrw ARG2(a,b), shrw ARG2(a,b), _WTOG shr ARG2(b,a))
-#define	SHR_B(a, b)	CHOICE(shrb ARG2(a,b), shrb ARG2(a,b), shrb ARG2(b,a))
-#define	SHRD_L(a,b,c)	CHOICE(shrdl ARG3(a,b,c), shrdl ARG3(a,b,c), _LTOG shrd ARG3(c,b,a))
-#define	SHRD2_L(a,b)	CHOICE(shrdl ARG2(a,b), shrdl ARG3(CL,a,b), _LTOG shrd ARG3(b,a,CL))
-#define	SHRD_W(a,b,c)	CHOICE(shrdw ARG3(a,b,c), shrdw ARG3(a,b,c), _WTOG shrd ARG3(c,b,a))
-#define	SHRD2_W(a,b)	CHOICE(shrdw ARG2(a,b), shrdw ARG3(CL,a,b), _WTOG shrd ARG3(b,a,CL))
-#define	SLDT(a)		CHOICE(sldt a, sldt a, sldt a)
-#define	SMSW(a)		CHOICE(smsw a, smsw a, smsw a)
-#define	STC		CHOICE(stc, stc, stc)
-#define	STD		CHOICE(std, std, std)
-#define	STI		CHOICE(sti, sti, sti)
-#define	STOS_L		CHOICE(stosl, stosl, _LTOG stos)
-#define	STOS_W		CHOICE(stosw, stosw, _WTOG stos)
-#define	STOS_B		CHOICE(stosb, stosb, stosb)
-#define	STR(a)		CHOICE(str a, str a, str a)
-#define	SUB_L(a, b)	CHOICE(subl ARG2(a,b), subl ARG2(a,b), _LTOG sub ARG2(b,a))
-#define	SUB_W(a, b)	CHOICE(subw ARG2(a,b), subw ARG2(a,b), _WTOG sub ARG2(b,a))
-#define	SUB_B(a, b)	CHOICE(subb ARG2(a,b), subb ARG2(a,b), subb ARG2(b,a))
-#define	TEST_L(a, b)	CHOICE(testl ARG2(a,b), testl ARG2(a,b), _LTOG test ARG2(b,a))
-#define	TEST_W(a, b)	CHOICE(testw ARG2(a,b), testw ARG2(a,b), _WTOG test ARG2(b,a))
-#define	TEST_B(a, b)	CHOICE(testb ARG2(a,b), testb ARG2(a,b), testb ARG2(b,a))
-#define	VERR(a)		CHOICE(verr a, verr a, verr a)
-#define	VERW(a)		CHOICE(verw a, verw a, verw a)
-#define	WAIT		CHOICE(wait, wait, wait)
-#define	XCHG_L(a, b)	CHOICE(xchgl ARG2(a,b), xchgl ARG2(a,b), _LTOG xchg ARG2(b,a))
-#define	XCHG_W(a, b)	CHOICE(xchgw ARG2(a,b), xchgw ARG2(a,b), _WTOG xchg ARG2(b,a))
-#define	XCHG_B(a, b)	CHOICE(xchgb ARG2(a,b), xchgb ARG2(a,b), xchgb ARG2(b,a))
-#define	XLAT		CHOICE(xlat, xlat, xlat)
-#define	XOR_L(a, b)	CHOICE(xorl ARG2(a,b), xorl ARG2(a,b), _LTOG xor ARG2(b,a))
-#define	XOR_W(a, b)	CHOICE(xorw ARG2(a,b), xorw ARG2(a,b), _WTOG xor ARG2(b,a))
-#define	XOR_B(a, b)	CHOICE(xorb ARG2(a,b), xorb ARG2(a,b), xorb ARG2(b,a))
+#define OUT1_L(a)	CHOICE(outl (a), outl ARG2(EAX,a), _LTOG out a)
+#define OUT1_W(a)	CHOICE(outw (a), outw ARG2(AX,a), _WTOG out a)
+#define OUT1_B(a)	CHOICE(outb (a), outb ARG2(AL,a), outb a)
+#define OUTS_L		CHOICE(outsl, outsl, _LTOG outs)
+#define OUTS_W		CHOICE(outsw, outsw, _WTOG outs)
+#define OUTS_B		CHOICE(outsb, outsb, outsb)
+#define POP_SR(a)	CHOICE(pop a, pop a, pop a)
+#define POP_L(a)	CHOICE(popl a, popl a, _LTOG pop a)
+#define POP_W(a)	CHOICE(popw a, popw a, _WTOG pop a)
+#define POPA_L		CHOICE(popal, popal, _LTOG popa)
+#define POPA_W		CHOICE(popaw, popaw, _WTOG popa)
+#define POPF_L		CHOICE(popfl, popfl, _LTOG popf)
+#define POPF_W		CHOICE(popfw, popfw, _WTOG popf)
+#define PUSH_SR(a)	CHOICE(push a, push a, push a)
+#define PUSH_L(a)	CHOICE(pushl a, pushl a, _LTOG push a)
+#define PUSH_W(a)	CHOICE(pushw a, pushw a, _WTOG push a)
+#define PUSH_B(a)	CHOICE(push a, pushb a, push a)
+#define PUSHA_L		CHOICE(pushal, pushal, _LTOG pusha)
+#define PUSHA_W		CHOICE(pushaw, pushaw, _WTOG pusha)
+#define PUSHF_L		CHOICE(pushfl, pushfl, _LTOG pushf)
+#define PUSHF_W		CHOICE(pushfw, pushfw, _WTOG pushf)
+#define RCL_L(a, b)	CHOICE(rcll ARG2(a,b), rcll ARG2(a,b), _LTOG rcl ARG2(b,a))
+#define RCL_W(a, b)	CHOICE(rclw ARG2(a,b), rclw ARG2(a,b), _WTOG rcl ARG2(b,a))
+#define RCL_B(a, b)	CHOICE(rclb ARG2(a,b), rclb ARG2(a,b), rclb ARG2(b,a))
+#define RCR_L(a, b)	CHOICE(rcrl ARG2(a,b), rcrl ARG2(a,b), _LTOG rcr ARG2(b,a))
+#define RCR_W(a, b)	CHOICE(rcrw ARG2(a,b), rcrw ARG2(a,b), _WTOG rcr ARG2(b,a))
+#define RCR_B(a, b)	CHOICE(rcrb ARG2(a,b), rcrb ARG2(a,b), rcrb ARG2(b,a))
+#define ROL_L(a, b)	CHOICE(roll ARG2(a,b), roll ARG2(a,b), _LTOG rol ARG2(b,a))
+#define ROL_W(a, b)	CHOICE(rolw ARG2(a,b), rolw ARG2(a,b), _WTOG rol ARG2(b,a))
+#define ROL_B(a, b)	CHOICE(rolb ARG2(a,b), rolb ARG2(a,b), rolb ARG2(b,a))
+#define ROR_L(a, b)	CHOICE(rorl ARG2(a,b), rorl ARG2(a,b), _LTOG ror ARG2(b,a))
+#define ROR_W(a, b)	CHOICE(rorw ARG2(a,b), rorw ARG2(a,b), _WTOG ror ARG2(b,a))
+#define ROR_B(a, b)	CHOICE(rorb ARG2(a,b), rorb ARG2(a,b), rorb ARG2(b,a))
+#define REP		CHOICE(rep ;, rep ;, repe)
+#define REPE		CHOICE(repz ;, repe ;, repe)
+#define REPNE		CHOICE(repnz ;, repne ;, repne)
+#define REPNZ		REPNE
+#define REPZ		REPE
+#define RET		CHOICE(ret, ret, ret)
+#define SAHF		CHOICE(sahf, sahf, sahf)
+#define SAL_L(a, b)	CHOICE(sall ARG2(a,b), sall ARG2(a,b), _LTOG sal ARG2(b,a))
+#define SAL_W(a, b)	CHOICE(salw ARG2(a,b), salw ARG2(a,b), _WTOG sal ARG2(b,a))
+#define SAL_B(a, b)	CHOICE(salb ARG2(a,b), salb ARG2(a,b), salb ARG2(b,a))
+#define SAR_L(a, b)	CHOICE(sarl ARG2(a,b), sarl ARG2(a,b), _LTOG sar ARG2(b,a))
+#define SAR_W(a, b)	CHOICE(sarw ARG2(a,b), sarw ARG2(a,b), _WTOG sar ARG2(b,a))
+#define SAR_B(a, b)	CHOICE(sarb ARG2(a,b), sarb ARG2(a,b), sarb ARG2(b,a))
+#define SBB_L(a, b)	CHOICE(sbbl ARG2(a,b), sbbl ARG2(a,b), _LTOG sbb ARG2(b,a))
+#define SBB_W(a, b)	CHOICE(sbbw ARG2(a,b), sbbw ARG2(a,b), _WTOG sbb ARG2(b,a))
+#define SBB_B(a, b)	CHOICE(sbbb ARG2(a,b), sbbb ARG2(a,b), sbbb ARG2(b,a))
+#define SCAS_L		CHOICE(scasl, scasl, _LTOG scas)
+#define SCAS_W		CHOICE(scasw, scasw, _WTOG scas)
+#define SCAS_B		CHOICE(scasb, scasb, scasb)
+#define SETA(a)		CHOICE(seta a, seta a, seta a)
+#define SETAE(a)	CHOICE(setae a, setae a, setae a)
+#define SETB(a)		CHOICE(setb a, setb a, setb a)
+#define SETBE(a)	CHOICE(setbe a, setbe a, setbe a)
+#define SETC(a)		CHOICE(setc a, setb a, setb a)
+#define SETE(a)		CHOICE(sete a, sete a, sete a)
+#define SETG(a)		CHOICE(setg a, setg a, setg a)
+#define SETGE(a)	CHOICE(setge a, setge a, setge a)
+#define SETL(a)		CHOICE(setl a, setl a, setl a)
+#define SETLE(a)	CHOICE(setle a, setle a, setle a)
+#define SETNA(a)	CHOICE(setna a, setna a, setna a)
+#define SETNAE(a)	CHOICE(setnae a, setnae a, setnae a)
+#define SETNB(a)	CHOICE(setnb a, setnb a, setnb a)
+#define SETNBE(a)	CHOICE(setnbe a, setnbe a, setnbe a)
+#define SETNC(a)	CHOICE(setnc a, setnb a, setnb a)
+#define SETNE(a)	CHOICE(setne a, setne a, setne a)
+#define SETNG(a)	CHOICE(setng a, setng a, setng a)
+#define SETNGE(a)	CHOICE(setnge a, setnge a, setnge a)
+#define SETNL(a)	CHOICE(setnl a, setnl a, setnl a)
+#define SETNLE(a)	CHOICE(setnle a, setnle a, setnle a)
+#define SETNO(a)	CHOICE(setno a, setno a, setno a)
+#define SETNP(a)	CHOICE(setnp a, setnp a, setnp a)
+#define SETNS(a)	CHOICE(setns a, setns a, setna a)
+#define SETNZ(a)	CHOICE(setnz a, setnz a, setnz a)
+#define SETO(a)		CHOICE(seto a, seto a, seto a)
+#define SETP(a)		CHOICE(setp a, setp a, setp a)
+#define SETPE(a)	CHOICE(setpe a, setpe a, setpe a)
+#define SETPO(a)	CHOICE(setpo a, setpo a, setpo a)
+#define SETS(a)		CHOICE(sets a, sets a, seta a)
+#define SETZ(a)		CHOICE(setz a, setz a, setz a)
+#define SGDT(a)		CHOICE(sgdt a, sgdt a, sgdt a)
+#define SIDT(a)		CHOICE(sidt a, sidt a, sidt a)
+#define SHL_L(a, b)	CHOICE(shll ARG2(a,b), shll ARG2(a,b), _LTOG shl ARG2(b,a))
+#define SHL_W(a, b)	CHOICE(shlw ARG2(a,b), shlw ARG2(a,b), _WTOG shl ARG2(b,a))
+#define SHL_B(a, b)	CHOICE(shlb ARG2(a,b), shlb ARG2(a,b), shlb ARG2(b,a))
+#define SHLD_L(a,b,c)	CHOICE(shldl ARG3(a,b,c), shldl ARG3(a,b,c), _LTOG shld ARG3(c,b,a))
+#define SHLD2_L(a,b)	CHOICE(shldl ARG2(a,b), shldl ARG3(CL,a,b), _LTOG shld ARG3(b,a,CL))
+#define SHLD_W(a,b,c)	CHOICE(shldw ARG3(a,b,c), shldw ARG3(a,b,c), _WTOG shld ARG3(c,b,a))
+#define SHLD2_W(a,b)	CHOICE(shldw ARG2(a,b), shldw ARG3(CL,a,b), _WTOG shld ARG3(b,a,CL))
+#define SHR_L(a, b)	CHOICE(shrl ARG2(a,b), shrl ARG2(a,b), _LTOG shr ARG2(b,a))
+#define SHR_W(a, b)	CHOICE(shrw ARG2(a,b), shrw ARG2(a,b), _WTOG shr ARG2(b,a))
+#define SHR_B(a, b)	CHOICE(shrb ARG2(a,b), shrb ARG2(a,b), shrb ARG2(b,a))
+#define SHRD_L(a,b,c)	CHOICE(shrdl ARG3(a,b,c), shrdl ARG3(a,b,c), _LTOG shrd ARG3(c,b,a))
+#define SHRD2_L(a,b)	CHOICE(shrdl ARG2(a,b), shrdl ARG3(CL,a,b), _LTOG shrd ARG3(b,a,CL))
+#define SHRD_W(a,b,c)	CHOICE(shrdw ARG3(a,b,c), shrdw ARG3(a,b,c), _WTOG shrd ARG3(c,b,a))
+#define SHRD2_W(a,b)	CHOICE(shrdw ARG2(a,b), shrdw ARG3(CL,a,b), _WTOG shrd ARG3(b,a,CL))
+#define SLDT(a)		CHOICE(sldt a, sldt a, sldt a)
+#define SMSW(a)		CHOICE(smsw a, smsw a, smsw a)
+#define STC		CHOICE(stc, stc, stc)
+#define STD		CHOICE(std, std, std)
+#define STI		CHOICE(sti, sti, sti)
+#define STOS_L		CHOICE(stosl, stosl, _LTOG stos)
+#define STOS_W		CHOICE(stosw, stosw, _WTOG stos)
+#define STOS_B		CHOICE(stosb, stosb, stosb)
+#define STR(a)		CHOICE(str a, str a, str a)
+#define SUB_L(a, b)	CHOICE(subl ARG2(a,b), subl ARG2(a,b), _LTOG sub ARG2(b,a))
+#define SUB_W(a, b)	CHOICE(subw ARG2(a,b), subw ARG2(a,b), _WTOG sub ARG2(b,a))
+#define SUB_B(a, b)	CHOICE(subb ARG2(a,b), subb ARG2(a,b), subb ARG2(b,a))
+#define TEST_L(a, b)	CHOICE(testl ARG2(a,b), testl ARG2(a,b), _LTOG test ARG2(b,a))
+#define TEST_W(a, b)	CHOICE(testw ARG2(a,b), testw ARG2(a,b), _WTOG test ARG2(b,a))
+#define TEST_B(a, b)	CHOICE(testb ARG2(a,b), testb ARG2(a,b), testb ARG2(b,a))
+#define VERR(a)		CHOICE(verr a, verr a, verr a)
+#define VERW(a)		CHOICE(verw a, verw a, verw a)
+#define WAIT		CHOICE(wait, wait, wait)
+#define XCHG_L(a, b)	CHOICE(xchgl ARG2(a,b), xchgl ARG2(a,b), _LTOG xchg ARG2(b,a))
+#define XCHG_W(a, b)	CHOICE(xchgw ARG2(a,b), xchgw ARG2(a,b), _WTOG xchg ARG2(b,a))
+#define XCHG_B(a, b)	CHOICE(xchgb ARG2(a,b), xchgb ARG2(a,b), xchgb ARG2(b,a))
+#define XLAT		CHOICE(xlat, xlat, xlat)
+#define XOR_L(a, b)	CHOICE(xorl ARG2(a,b), xorl ARG2(a,b), _LTOG xor ARG2(b,a))
+#define XOR_W(a, b)	CHOICE(xorw ARG2(a,b), xorw ARG2(a,b), _WTOG xor ARG2(b,a))
+#define XOR_B(a, b)	CHOICE(xorb ARG2(a,b), xorb ARG2(a,b), xorb ARG2(b,a))
 
 
 /* Floating Point Instructions */
-#define	F2XM1		CHOICE(f2xm1, f2xm1, f2xm1)
-#define	FABS		CHOICE(fabs, fabs, fabs)
-#define	FADD_D(a)	CHOICE(faddl a, faddl a, faddd a)
-#define	FADD_S(a)	CHOICE(fadds a, fadds a, fadds a)
-#define	FADD2(a, b)	CHOICE(fadd ARG2(a,b), fadd ARG2(a,b), fadd ARG2(b,a))
-#define	FADDP(a, b)	CHOICE(faddp ARG2(a,b), faddp ARG2(a,b), faddp ARG2(b,a))
-#define	FIADD_L(a)	CHOICE(fiaddl a, fiaddl a, fiaddl a)
-#define	FIADD_W(a)	CHOICE(fiadd a, fiadds a, fiadds a)
-#define	FBLD(a)		CHOICE(fbld a, fbld a, fbld a)
-#define	FBSTP(a)	CHOICE(fbstp a, fbstp a, fbstp a)
-#define	FCHS		CHOICE(fchs, fchs, fchs)
-#define	FCLEX		CHOICE(fclex, wait; fnclex, wait; fclex)
-#define	FNCLEX		CHOICE(fnclex, fnclex, fclex)
-#define	FCOM(a)		CHOICE(fcom a, fcom a, fcom a)
-#define	FCOM_D(a)	CHOICE(fcoml a, fcoml a, fcomd a)
-#define	FCOM_S(a)	CHOICE(fcoms a, fcoms a, fcoms a)
-#define	FCOMP(a)	CHOICE(fcomp a, fcomp a, fcomp a)
-#define	FCOMP_D(a)	CHOICE(fcompl a, fcompl a, fcompd a)
-#define	FCOMP_S(a)	CHOICE(fcomps a, fcomps a, fcomps a)
-#define	FCOMPP		CHOICE(fcompp, fcompp, fcompp)
-#define	FCOS		CHOICE(fcos, fcos, fcos)
-#define	FDECSTP		CHOICE(fdecstp, fdecstp, fdecstp)
-#define	FDIV_D(a)	CHOICE(fdivl a, fdivl a, fdivd a)
-#define	FDIV_S(a)	CHOICE(fdivs a, fdivs a, fdivs a)
-#define	FDIV2(a, b)	CHOICE(fdiv ARG2(a,b), fdiv ARG2(a,b), fdiv ARG2(b,a))
-#define	FDIVP(a, b)	CHOICE(fdivp ARG2(a,b), fdivp ARG2(a,b), fdivp ARG2(b,a))
-#define	FIDIV_L(a)	CHOICE(fidivl a, fidivl a, fidivl a)
-#define	FIDIV_W(a)	CHOICE(fidiv a, fidivs a, fidivs a)
-#define	FDIVR_D(a)	CHOICE(fdivrl a, fdivrl a, fdivrd a)
-#define	FDIVR_S(a)	CHOICE(fdivrs a, fdivrs a, fdivrs a)
-#define	FDIVR2(a, b)	CHOICE(fdivr ARG2(a,b), fdivr ARG2(a,b), fdivr ARG2(b,a))
-#define	FDIVRP(a, b)	CHOICE(fdivrp ARG2(a,b), fdivrp ARG2(a,b), fdivrp ARG2(b,a))
-#define	FIDIVR_L(a)	CHOICE(fidivrl a, fidivrl a, fidivrl a)
-#define	FIDIVR_W(a)	CHOICE(fidivr a, fidivrs a, fidivrs a)
-#define	FFREE(a)	CHOICE(ffree a, ffree a, ffree a)
-#define	FICOM_L(a)	CHOICE(ficoml a, ficoml a, ficoml a)
-#define	FICOM_W(a)	CHOICE(ficom a, ficoms a, ficoms a)
-#define	FICOMP_L(a)	CHOICE(ficompl a, ficompl a, ficompl a)
-#define	FICOMP_W(a)	CHOICE(ficomp a, ficomps a, ficomps a)
-#define	FILD_Q(a)	CHOICE(fildll a, fildq a, fildq a)
-#define	FILD_L(a)	CHOICE(fildl a, fildl a, fildl a)
-#define	FILD_W(a)	CHOICE(fild a, filds a, filds a)
-#define	FINCSTP		CHOICE(fincstp, fincstp, fincstp)
-#define	FINIT		CHOICE(finit, wait; fninit, wait; finit)
-#define	FNINIT		CHOICE(fninit, fninit, finit)
-#define	FIST_L(a)	CHOICE(fistl a, fistl a, fistl a)
-#define	FIST_W(a)	CHOICE(fist a, fists a, fists a)
-#define	FISTP_Q(a)	CHOICE(fistpll a, fistpq a, fistpq a)
-#define	FISTP_L(a)	CHOICE(fistpl a, fistpl a, fistpl a)
-#define	FISTP_W(a)	CHOICE(fistp a, fistps a, fistps a)
-#define	FLD_X(a)	CHOICE(fldt a, fldt a, fldx a) /* 80 bit data type! */
-#define	FLD_D(a)	CHOICE(fldl a, fldl a, fldd a)
-#define	FLD_S(a)	CHOICE(flds a, flds a, flds a)
-#define	FLD1		CHOICE(fld1, fld1, fld1)
-#define	FLDL2T		CHOICE(fldl2t, fldl2t, fldl2t)
-#define	FLDL2E		CHOICE(fldl2e, fldl2e, fldl2e)
-#define	FLDPI		CHOICE(fldpi, fldpi, fldpi)
-#define	FLDLG2		CHOICE(fldlg2, fldlg2, fldlg2)
-#define	FLDLN2		CHOICE(fldln2, fldln2, fldln2)
-#define	FLDZ		CHOICE(fldz, fldz, fldz)
-#define	FLDCW(a)	CHOICE(fldcw a, fldcw a, fldcw a)
-#define	FLDENV(a)	CHOICE(fldenv a, fldenv a, fldenv a)
-#define	FMUL_S(a)	CHOICE(fmuls a, fmuls a, fmuls a)
-#define	FMUL_D(a)	CHOICE(fmull a, fmull a, fmuld a)
-#define	FMUL2(a, b)	CHOICE(fmul ARG2(a,b), fmul ARG2(a,b), fmul ARG2(b,a))
-#define	FMULP(a, b)	CHOICE(fmulp ARG2(a,b), fmulp ARG2(a,b), fmulp ARG2(b,a))
-#define	FIMUL_L(a)	CHOICE(fimull a, fimull a, fimull a)
-#define	FIMUL_W(a)	CHOICE(fimul a, fimuls a, fimuls a)
-#define	FNOP		CHOICE(fnop, fnop, fnop)
-#define	FPATAN		CHOICE(fpatan, fpatan, fpatan)
-#define	FPREM		CHOICE(fprem, fprem, fprem)
-#define	FPREM1		CHOICE(fprem1, fprem1, fprem1)
-#define	FPTAN		CHOICE(fptan, fptan, fptan)
-#define	FRNDINT		CHOICE(frndint, frndint, frndint)
-#define	FRSTOR(a)	CHOICE(frstor a, frstor a, frstor a)
-#define	FSAVE(a)	CHOICE(fsave a, wait; fnsave a, wait; fsave a)
-#define	FNSAVE(a)	CHOICE(fnsave a, fnsave a, fsave a)
-#define	FSCALE		CHOICE(fscale, fscale, fscale)
-#define	FSIN		CHOICE(fsin, fsin, fsin)
-#define	FSINCOS		CHOICE(fsincos, fsincos, fsincos)
-#define	FSQRT		CHOICE(fsqrt, fsqrt, fsqrt)
-#define	FST_D(a)	CHOICE(fstl a, fstl a, fstd a)
-#define	FST_S(a)	CHOICE(fsts a, fsts a, fsts a)
-#define	FSTP_X(a)	CHOICE(fstpt a, fstpt a, fstpx a)
-#define	FSTP_D(a)	CHOICE(fstpl a, fstpl a, fstpd a)
-#define	FSTP_S(a)	CHOICE(fstps a, fstps a, fstps a)
-#define	FSTP(a)		CHOICE(fstp a, fstp a, fstp a)
-#define	FSTCW(a)	CHOICE(fstcw a, wait; fnstcw a, wait; fstcw a)
-#define	FNSTCW(a)	CHOICE(fnstcw a, fnstcw a, fstcw a)
-#define	FSTENV(a)	CHOICE(fstenv a, wait; fnstenv a, fstenv a)
-#define	FNSTENV(a)	CHOICE(fnstenv a, fnstenv a, fstenv a)
-#define	FSTSW(a)	CHOICE(fstsw a, wait; fnstsw a, wait; fstsw a)
-#define	FNSTSW(a)	CHOICE(fnstsw a, fnstsw a, fstsw a)
-#define	FSUB_S(a)	CHOICE(fsubs a, fsubs a, fsubs a)
-#define	FSUB_D(a)	CHOICE(fsubl a, fsubl a, fsubd a)
-#define	FSUB2(a, b)	CHOICE(fsub ARG2(a,b), fsub ARG2(a,b), fsub ARG2(b,a))
-#define	FSUBP(a, b)	CHOICE(fsubp ARG2(a,b), fsubp ARG2(a,b), fsubp ARG2(b,a))
-#define	FISUB_L(a)	CHOICE(fisubl a, fisubl a, fisubl a)
-#define	FISUB_W(a)	CHOICE(fisub a, fisubs a, fisubs a)
-#define	FSUBR_S(a)	CHOICE(fsubrs a, fsubrs a, fsubrs a)
-#define	FSUBR_D(a)	CHOICE(fsubrl a, fsubrl a, fsubrd a)
-#define	FSUBR2(a, b)	CHOICE(fsubr ARG2(a,b), fsubr ARG2(a,b), fsubr ARG2(b,a))
-#define	FSUBRP(a, b)	CHOICE(fsubrp ARG2(a,b), fsubrp ARG2(a,b), fsubrp ARG2(b,a))
-#define	FISUBR_L(a)	CHOICE(fisubrl a, fisubrl a, fisubrl a)
-#define	FISUBR_W(a)	CHOICE(fisubr a, fisubrs a, fisubrs a)
-#define	FTST		CHOICE(ftst, ftst, ftst)
-#define	FUCOM(a)	CHOICE(fucom a, fucom a, fucom a)
-#define	FUCOMP(a)	CHOICE(fucomp a, fucomp a, fucomp a)
-#define	FUCOMPP		CHOICE(fucompp, fucompp, fucompp)
-#define	FWAIT		CHOICE(wait, wait, wait)
-#define	FXAM		CHOICE(fxam, fxam, fxam)
-#define	FXCH(a)		CHOICE(fxch a, fxch a, fxch a)
-#define	FXTRACT		CHOICE(fxtract, fxtract, fxtract)
-#define	FYL2X		CHOICE(fyl2x, fyl2x, fyl2x)
-#define	FYL2XP1		CHOICE(fyl2xp1, fyl2xp1, fyl2xp1)
+#define F2XM1		CHOICE(f2xm1, f2xm1, f2xm1)
+#define FABS		CHOICE(fabs, fabs, fabs)
+#define FADD_D(a)	CHOICE(faddl a, faddl a, faddd a)
+#define FADD_S(a)	CHOICE(fadds a, fadds a, fadds a)
+#define FADD2(a, b)	CHOICE(fadd ARG2(a,b), fadd ARG2(a,b), fadd ARG2(b,a))
+#define FADDP(a, b)	CHOICE(faddp ARG2(a,b), faddp ARG2(a,b), faddp ARG2(b,a))
+#define FIADD_L(a)	CHOICE(fiaddl a, fiaddl a, fiaddl a)
+#define FIADD_W(a)	CHOICE(fiadd a, fiadds a, fiadds a)
+#define FBLD(a)		CHOICE(fbld a, fbld a, fbld a)
+#define FBSTP(a)	CHOICE(fbstp a, fbstp a, fbstp a)
+#define FCHS		CHOICE(fchs, fchs, fchs)
+#define FCLEX		CHOICE(fclex, wait; fnclex, wait; fclex)
+#define FNCLEX		CHOICE(fnclex, fnclex, fclex)
+#define FCOM(a)		CHOICE(fcom a, fcom a, fcom a)
+#define FCOM_D(a)	CHOICE(fcoml a, fcoml a, fcomd a)
+#define FCOM_S(a)	CHOICE(fcoms a, fcoms a, fcoms a)
+#define FCOMP(a)	CHOICE(fcomp a, fcomp a, fcomp a)
+#define FCOMP_D(a)	CHOICE(fcompl a, fcompl a, fcompd a)
+#define FCOMP_S(a)	CHOICE(fcomps a, fcomps a, fcomps a)
+#define FCOMPP		CHOICE(fcompp, fcompp, fcompp)
+#define FCOS		CHOICE(fcos, fcos, fcos)
+#define FDECSTP		CHOICE(fdecstp, fdecstp, fdecstp)
+#define FDIV_D(a)	CHOICE(fdivl a, fdivl a, fdivd a)
+#define FDIV_S(a)	CHOICE(fdivs a, fdivs a, fdivs a)
+#define FDIV2(a, b)	CHOICE(fdiv ARG2(a,b), fdiv ARG2(a,b), fdiv ARG2(b,a))
+#define FDIVP(a, b)	CHOICE(fdivp ARG2(a,b), fdivp ARG2(a,b), fdivp ARG2(b,a))
+#define FIDIV_L(a)	CHOICE(fidivl a, fidivl a, fidivl a)
+#define FIDIV_W(a)	CHOICE(fidiv a, fidivs a, fidivs a)
+#define FDIVR_D(a)	CHOICE(fdivrl a, fdivrl a, fdivrd a)
+#define FDIVR_S(a)	CHOICE(fdivrs a, fdivrs a, fdivrs a)
+#define FDIVR2(a, b)	CHOICE(fdivr ARG2(a,b), fdivr ARG2(a,b), fdivr ARG2(b,a))
+#define FDIVRP(a, b)	CHOICE(fdivrp ARG2(a,b), fdivrp ARG2(a,b), fdivrp ARG2(b,a))
+#define FIDIVR_L(a)	CHOICE(fidivrl a, fidivrl a, fidivrl a)
+#define FIDIVR_W(a)	CHOICE(fidivr a, fidivrs a, fidivrs a)
+#define FFREE(a)	CHOICE(ffree a, ffree a, ffree a)
+#define FICOM_L(a)	CHOICE(ficoml a, ficoml a, ficoml a)
+#define FICOM_W(a)	CHOICE(ficom a, ficoms a, ficoms a)
+#define FICOMP_L(a)	CHOICE(ficompl a, ficompl a, ficompl a)
+#define FICOMP_W(a)	CHOICE(ficomp a, ficomps a, ficomps a)
+#define FILD_Q(a)	CHOICE(fildll a, fildq a, fildq a)
+#define FILD_L(a)	CHOICE(fildl a, fildl a, fildl a)
+#define FILD_W(a)	CHOICE(fild a, filds a, filds a)
+#define FINCSTP		CHOICE(fincstp, fincstp, fincstp)
+#define FINIT		CHOICE(finit, wait; fninit, wait; finit)
+#define FNINIT		CHOICE(fninit, fninit, finit)
+#define FIST_L(a)	CHOICE(fistl a, fistl a, fistl a)
+#define FIST_W(a)	CHOICE(fist a, fists a, fists a)
+#define FISTP_Q(a)	CHOICE(fistpll a, fistpq a, fistpq a)
+#define FISTP_L(a)	CHOICE(fistpl a, fistpl a, fistpl a)
+#define FISTP_W(a)	CHOICE(fistp a, fistps a, fistps a)
+#define FLD_X(a)	CHOICE(fldt a, fldt a, fldx a) /* 80 bit data type! */
+#define FLD_D(a)	CHOICE(fldl a, fldl a, fldd a)
+#define FLD_S(a)	CHOICE(flds a, flds a, flds a)
+#define FLD1		CHOICE(fld1, fld1, fld1)
+#define FLDL2T		CHOICE(fldl2t, fldl2t, fldl2t)
+#define FLDL2E		CHOICE(fldl2e, fldl2e, fldl2e)
+#define FLDPI		CHOICE(fldpi, fldpi, fldpi)
+#define FLDLG2		CHOICE(fldlg2, fldlg2, fldlg2)
+#define FLDLN2		CHOICE(fldln2, fldln2, fldln2)
+#define FLDZ		CHOICE(fldz, fldz, fldz)
+#define FLDCW(a)	CHOICE(fldcw a, fldcw a, fldcw a)
+#define FLDENV(a)	CHOICE(fldenv a, fldenv a, fldenv a)
+#define FMUL_S(a)	CHOICE(fmuls a, fmuls a, fmuls a)
+#define FMUL_D(a)	CHOICE(fmull a, fmull a, fmuld a)
+#define FMUL2(a, b)	CHOICE(fmul ARG2(a,b), fmul ARG2(a,b), fmul ARG2(b,a))
+#define FMULP(a, b)	CHOICE(fmulp ARG2(a,b), fmulp ARG2(a,b), fmulp ARG2(b,a))
+#define FIMUL_L(a)	CHOICE(fimull a, fimull a, fimull a)
+#define FIMUL_W(a)	CHOICE(fimul a, fimuls a, fimuls a)
+#define FNOP		CHOICE(fnop, fnop, fnop)
+#define FPATAN		CHOICE(fpatan, fpatan, fpatan)
+#define FPREM		CHOICE(fprem, fprem, fprem)
+#define FPREM1		CHOICE(fprem1, fprem1, fprem1)
+#define FPTAN		CHOICE(fptan, fptan, fptan)
+#define FRNDINT		CHOICE(frndint, frndint, frndint)
+#define FRSTOR(a)	CHOICE(frstor a, frstor a, frstor a)
+#define FSAVE(a)	CHOICE(fsave a, wait; fnsave a, wait; fsave a)
+#define FNSAVE(a)	CHOICE(fnsave a, fnsave a, fsave a)
+#define FSCALE		CHOICE(fscale, fscale, fscale)
+#define FSIN		CHOICE(fsin, fsin, fsin)
+#define FSINCOS		CHOICE(fsincos, fsincos, fsincos)
+#define FSQRT		CHOICE(fsqrt, fsqrt, fsqrt)
+#define FST_D(a)	CHOICE(fstl a, fstl a, fstd a)
+#define FST_S(a)	CHOICE(fsts a, fsts a, fsts a)
+#define FSTP_X(a)	CHOICE(fstpt a, fstpt a, fstpx a)
+#define FSTP_D(a)	CHOICE(fstpl a, fstpl a, fstpd a)
+#define FSTP_S(a)	CHOICE(fstps a, fstps a, fstps a)
+#define FSTP(a)		CHOICE(fstp a, fstp a, fstp a)
+#define FSTCW(a)	CHOICE(fstcw a, wait; fnstcw a, wait; fstcw a)
+#define FNSTCW(a)	CHOICE(fnstcw a, fnstcw a, fstcw a)
+#define FSTENV(a)	CHOICE(fstenv a, wait; fnstenv a, fstenv a)
+#define FNSTENV(a)	CHOICE(fnstenv a, fnstenv a, fstenv a)
+#define FSTSW(a)	CHOICE(fstsw a, wait; fnstsw a, wait; fstsw a)
+#define FNSTSW(a)	CHOICE(fnstsw a, fnstsw a, fstsw a)
+#define FSUB_S(a)	CHOICE(fsubs a, fsubs a, fsubs a)
+#define FSUB_D(a)	CHOICE(fsubl a, fsubl a, fsubd a)
+#define FSUB2(a, b)	CHOICE(fsub ARG2(a,b), fsub ARG2(a,b), fsub ARG2(b,a))
+#define FSUBP(a, b)	CHOICE(fsubp ARG2(a,b), fsubp ARG2(a,b), fsubp ARG2(b,a))
+#define FISUB_L(a)	CHOICE(fisubl a, fisubl a, fisubl a)
+#define FISUB_W(a)	CHOICE(fisub a, fisubs a, fisubs a)
+#define FSUBR_S(a)	CHOICE(fsubrs a, fsubrs a, fsubrs a)
+#define FSUBR_D(a)	CHOICE(fsubrl a, fsubrl a, fsubrd a)
+#define FSUBR2(a, b)	CHOICE(fsubr ARG2(a,b), fsubr ARG2(a,b), fsubr ARG2(b,a))
+#define FSUBRP(a, b)	CHOICE(fsubrp ARG2(a,b), fsubrp ARG2(a,b), fsubrp ARG2(b,a))
+#define FISUBR_L(a)	CHOICE(fisubrl a, fisubrl a, fisubrl a)
+#define FISUBR_W(a)	CHOICE(fisubr a, fisubrs a, fisubrs a)
+#define FTST		CHOICE(ftst, ftst, ftst)
+#define FUCOM(a)	CHOICE(fucom a, fucom a, fucom a)
+#define FUCOMP(a)	CHOICE(fucomp a, fucomp a, fucomp a)
+#define FUCOMPP		CHOICE(fucompp, fucompp, fucompp)
+#define FWAIT		CHOICE(wait, wait, wait)
+#define FXAM		CHOICE(fxam, fxam, fxam)
+#define FXCH(a)		CHOICE(fxch a, fxch a, fxch a)
+#define FXTRACT		CHOICE(fxtract, fxtract, fxtract)
+#define FYL2X		CHOICE(fyl2x, fyl2x, fyl2x)
+#define FYL2XP1		CHOICE(fyl2xp1, fyl2xp1, fyl2xp1)
 
 /* New instructions */
-#define	CPUID		CHOICE(D_BYTE ARG2(15, 162), cpuid, D_BYTE ARG2(15, 162))
-#define	RDTSC		CHOICE(D_BYTE ARG2(15, 49), rdtsc, D_BYTE ARG2(15, 49))
+#define CPUID		CHOICE(D_BYTE ARG2(15, 162), cpuid, D_BYTE ARG2(15, 162))
+#define RDTSC		CHOICE(D_BYTE ARG2(15, 49), rdtsc, D_BYTE ARG2(15, 49))
 
 #else /* NASM_ASSEMBLER || MASM_ASSEMBLER is defined */
 
@@ -779,166 +783,167 @@
 	/*					*/
 	/****************************************/
 
-#define P_EAX 	EAX
-#define L_EAX 	EAX
-#define W_AX 	AX
-#define B_AH 	AH
-#define B_AL 	AL
-
-#define P_EBX 	EBX
-#define L_EBX 	EBX
-#define W_BX 	BX
-#define B_BH 	BH
-#define B_BL 	BL
-
-#define P_ECX 	ECX
-#define L_ECX 	ECX
-#define W_CX 	CX
-#define B_CH 	CH
-#define B_CL 	CL
-
-#define P_EDX 	EDX
-#define L_EDX 	EDX
-#define W_DX 	DX
-#define B_DH 	DH
-#define B_DL 	DL
-
-#define P_EBP 	EBP
-#define L_EBP 	EBP
-#define W_BP 	BP
-
-#define P_ESI 	ESI
-#define L_ESI 	ESI
-#define W_SI 	SI
-
-#define P_EDI 	EDI
-#define L_EDI 	EDI
-#define W_DI 	DI
-
-#define P_ESP 	ESP
-#define L_ESP 	ESP
-#define W_SP 	SP
-
-#define W_CS 	CS
-#define W_SS 	SS
-#define W_DS 	DS
-#define W_ES 	ES
-#define W_FS 	FS
-#define W_GS 	GS
-
-#define X_ST 	ST
-#define D_ST 	ST
-#define L_ST 	ST
-
-#define P_MM0 	mm0
-#define P_MM1 	mm1
-#define P_MM2 	mm2
-#define P_MM3 	mm3
-#define P_MM4 	mm4
-#define P_MM5 	mm5
-#define P_MM6 	mm6
-#define P_MM7 	mm7
-
-#define P_XMM0 	xmm0
-#define P_XMM1 	xmm1
-#define P_XMM2 	xmm2
-#define P_XMM3 	xmm3
-#define P_XMM4 	xmm4
-#define P_XMM5 	xmm5
-#define P_XMM6 	xmm6
-#define P_XMM7 	xmm7
-
-#define	CONCAT(x, y)	x ## y
+#define P_EAX		EAX
+#define L_EAX		EAX
+#define W_AX		AX
+#define B_AH		AH
+#define B_AL		AL
+
+#define P_EBX		EBX
+#define L_EBX		EBX
+#define W_BX		BX
+#define B_BH		BH
+#define B_BL		BL
+
+#define P_ECX		ECX
+#define L_ECX		ECX
+#define W_CX		CX
+#define B_CH		CH
+#define B_CL		CL
+
+#define P_EDX		EDX
+#define L_EDX		EDX
+#define W_DX		DX
+#define B_DH		DH
+#define B_DL		DL
+
+#define P_EBP		EBP
+#define L_EBP		EBP
+#define W_BP		BP
+
+#define P_ESI		ESI
+#define L_ESI		ESI
+#define W_SI		SI
+
+#define P_EDI		EDI
+#define L_EDI		EDI
+#define W_DI		DI
+
+#define P_ESP		ESP
+#define L_ESP		ESP
+#define W_SP		SP
+
+#define W_CS		CS
+#define W_SS		SS
+#define W_DS		DS
+#define W_ES		ES
+#define W_FS		FS
+#define W_GS		GS
+
+#define X_ST		ST
+#define D_ST		ST
+#define L_ST		ST
+
+#define P_MM0		mm0
+#define P_MM1		mm1
+#define P_MM2		mm2
+#define P_MM3		mm3
+#define P_MM4		mm4
+#define P_MM5		mm5
+#define P_MM6		mm6
+#define P_MM7		mm7
+
+#define P_XMM0		xmm0
+#define P_XMM1		xmm1
+#define P_XMM2		xmm2
+#define P_XMM3		xmm3
+#define P_XMM4		xmm4
+#define P_XMM5		xmm5
+#define P_XMM6		xmm6
+#define P_XMM7		xmm7
+
+#define CONCAT(x, y)	x ## y
 
 #if defined(NASM_ASSEMBLER)
 
-#define ST(n) st ## n
+#define ST(n)		st ## n
 
-#define TBYTE_PTR tword
-#define QWORD_PTR qword
-#define DWORD_PTR dword
-#define WORD_PTR word
-#define BYTE_PTR byte
+#define TBYTE_PTR	tword
+#define QWORD_PTR	qword
+#define DWORD_PTR	dword
+#define WORD_PTR	word
+#define BYTE_PTR	byte
 
 #define OFFSET
 
-#define GLOBL GLOBAL
-#define ALIGNTEXT32 ALIGN 32
-#define ALIGNTEXT16 ALIGN 16
-#define ALIGNTEXT8 ALIGN 8
-#define ALIGNTEXT4 ALIGN 4
-#define ALIGNTEXT2 ALIGN 2
-#define ALIGNTEXT32ifNOP ALIGN 32
-#define ALIGNTEXT16ifNOP ALIGN 16
-#define ALIGNTEXT8ifNOP ALIGN 8
-#define ALIGNTEXT4ifNOP ALIGN 4
-#define ALIGNDATA32 ALIGN 32
-#define ALIGNDATA16 ALIGN 16
-#define ALIGNDATA8 ALIGN 8
-#define ALIGNDATA4 ALIGN 4
-#define ALIGNDATA2 ALIGN 2
+#define GLOBL			GLOBAL
+#define ALIGNTEXT32		ALIGN 32
+#define ALIGNTEXT16		ALIGN 16
+#define ALIGNTEXT8		ALIGN 8
+#define ALIGNTEXT4		ALIGN 4
+#define ALIGNTEXT2		ALIGN 2
+#define ALIGNTEXT32ifNOP	ALIGN 32
+#define ALIGNTEXT16ifNOP	ALIGN 16
+#define ALIGNTEXT8ifNOP		ALIGN 8
+#define ALIGNTEXT4ifNOP		ALIGN 4
+#define ALIGNDATA32		ALIGN 32
+#define ALIGNDATA16		ALIGN 16
+#define ALIGNDATA8		ALIGN 8
+#define ALIGNDATA4		ALIGN 4
+#define ALIGNDATA2		ALIGN 2
 #define FILE(s)
-#define STRING(s) db s
-#define D_LONG dd
-#define D_WORD dw
-#define D_BYTE db
+#define STRING(s)	db s
+#define D_LONG		dd
+#define D_WORD		dw
+#define D_BYTE		db
 /* #define SPACE */
 /* #define COMM */
 #if defined(__WATCOMC__)
 SECTION _TEXT public align=16 class=CODE use32 flat
 SECTION _DATA public align=16 class=DATA use32 flat
-#define SEG_TEXT SECTION _TEXT
-#define SEG_DATA SECTION _DATA
-#define SEG_BSS SECTION .bss
+#define SEG_TEXT	SECTION _TEXT
+#define SEG_DATA	SECTION _DATA
+#define SEG_BSS		SECTION .bss
 #else
-#define SEG_DATA SECTION .data
-#define SEG_TEXT SECTION .text
-#define SEG_BSS SECTION .bss
+#define SEG_DATA	SECTION .data
+#define SEG_TEXT	SECTION .text
+#define SEG_BSS		SECTION .bss
 #endif
 
-#define D_SPACE(n) db n REP 0
+#define D_SPACE(n)	db n REP 0
 
 #define AS_BEGIN
 
-#define NEAR near 	/* Jcc's should be handled better than this... */
+/* Jcc's should be handled better than this... */
+#define NEAR		near
 
 #else /* MASM */
 
-#define TBYTE_PTR tbyte ptr
-#define QWORD_PTR qword ptr
-#define DWORD_PTR dword ptr
-#define WORD_PTR word ptr
-#define BYTE_PTR byte ptr
-
-#define OFFSET offset
-
-#define GLOBL GLOBAL
-#define ALIGNTEXT32 ALIGN 32
-#define ALIGNTEXT16 ALIGN 16
-#define ALIGNTEXT8 ALIGN 8
-#define ALIGNTEXT4 ALIGN 4
-#define ALIGNTEXT2 ALIGN 2
-#define ALIGNTEXT32ifNOP ALIGN 32
-#define ALIGNTEXT16ifNOP ALIGN 16
-#define ALIGNTEXT8ifNOP ALIGN 8
-#define ALIGNTEXT4ifNOP ALIGN 4
-#define ALIGNDATA32 ALIGN 32
-#define ALIGNDATA16 ALIGN 16
-#define ALIGNDATA8 ALIGN 8
-#define ALIGNDATA4 ALIGN 4
-#define ALIGNDATA2 ALIGN 2
+#define TBYTE_PTR	tbyte ptr
+#define QWORD_PTR	qword ptr
+#define DWORD_PTR	dword ptr
+#define WORD_PTR	word ptr
+#define BYTE_PTR	byte ptr
+
+#define OFFSET		offset
+
+#define GLOBL			GLOBAL
+#define ALIGNTEXT32		ALIGN 32
+#define ALIGNTEXT16		ALIGN 16
+#define ALIGNTEXT8		ALIGN 8
+#define ALIGNTEXT4		ALIGN 4
+#define ALIGNTEXT2		ALIGN 2
+#define ALIGNTEXT32ifNOP	ALIGN 32
+#define ALIGNTEXT16ifNOP	ALIGN 16
+#define ALIGNTEXT8ifNOP		ALIGN 8
+#define ALIGNTEXT4ifNOP		ALIGN 4
+#define ALIGNDATA32		ALIGN 32
+#define ALIGNDATA16		ALIGN 16
+#define ALIGNDATA8		ALIGN 8
+#define ALIGNDATA4		ALIGN 4
+#define ALIGNDATA2		ALIGN 2
 #define FILE(s)
-#define STRING(s) db s
-#define D_LONG dd
-#define D_WORD dw
-#define D_BYTE db
+#define STRING(s)	db s
+#define D_LONG		dd
+#define D_WORD		dw
+#define D_BYTE		db
 /* #define SPACE */
 /* #define COMM */
-#define SEG_DATA .DATA
-#define SEG_TEXT .CODE
-#define SEG_BSS .DATA
+#define SEG_DATA	.DATA
+#define SEG_TEXT	.CODE
+#define SEG_BSS		.DATA
 
-#define D_SPACE(n) db n REP 0
+#define D_SPACE(n)	db n REP 0
 
 #define AS_BEGIN
 
@@ -949,520 +954,529 @@ SECTION _DATA public align=16 class=DATA use32 flat
 #if defined(Lynx) || (defined(SYSV) || defined(SVR4)) \
  || (defined(linux) || defined(__OS2ELF__)) && defined(__ELF__) \
  || defined(__FreeBSD__) && __FreeBSD__ >= 3
-#define GLNAME(a) a
+#define GLNAME(a)	a
 #else
-#define GLNAME(a) _ ## a
+#define GLNAME(a)	_ ## a
 #endif
 
 /*
- * 	Addressing Modes
+ *	Addressing Modes
  */
 
 /* Immediate Mode */
-#define P_ADDR(a) 	OFFSET a
-#define X_ADDR(a) 	OFFSET a
-#define D_ADDR(a) 	OFFSET a
-#define L_ADDR(a) 	OFFSET a
-#define W_ADDR(a) 	OFFSET a
-#define B_ADDR(a) 	OFFSET a
-
-#define P_CONST(a) 	a
-#define X_CONST(a) 	a
-#define D_CONST(a) 	a
-#define L_CONST(a) 	a
-#define W_CONST(a) 	a
-#define B_CONST(a) 	a
+#define P_ADDR(a)		OFFSET a
+#define X_ADDR(a)		OFFSET a
+#define D_ADDR(a)		OFFSET a
+#define L_ADDR(a)		OFFSET a
+#define W_ADDR(a)		OFFSET a
+#define B_ADDR(a)		OFFSET a
+
+#define P_CONST(a)		a
+#define X_CONST(a)		a
+#define D_CONST(a)		a
+#define L_CONST(a)		a
+#define W_CONST(a)		a
+#define B_CONST(a)		a
 
 /* Indirect Mode */
-#define P_CONTENT(a) 	a
-#define X_CONTENT(a) 	TBYTE_PTR a
-#define D_CONTENT(a) 	QWORD_PTR a
-#define L_CONTENT(a) 	DWORD_PTR a
-#define W_CONTENT(a) 	WORD_PTR a
-#define B_CONTENT(a) 	BYTE_PTR a
+#define P_CONTENT(a)		a
+#define X_CONTENT(a)		TBYTE_PTR a
+#define D_CONTENT(a)		QWORD_PTR a
+#define L_CONTENT(a)		DWORD_PTR a
+#define W_CONTENT(a)		WORD_PTR a
+#define B_CONTENT(a)		BYTE_PTR a
 
 /* Register a indirect */
-#define P_REGIND(a) 	[a]
-#define X_REGIND(a) 	TBYTE_PTR [a]
-#define D_REGIND(a) 	QWORD_PTR [a]
-#define L_REGIND(a) 	DWORD_PTR [a]
-#define W_REGIND(a) 	WORD_PTR [a]
-#define B_REGIND(a) 	BYTE_PTR [a]
+#define P_REGIND(a)		[a]
+#define X_REGIND(a)		TBYTE_PTR [a]
+#define D_REGIND(a)		QWORD_PTR [a]
+#define L_REGIND(a)		DWORD_PTR [a]
+#define W_REGIND(a)		WORD_PTR [a]
+#define B_REGIND(a)		BYTE_PTR [a]
 
 /* Register b indirect plus displacement a */
-#define P_REGOFF(a, b) 	[b + a]
-#define X_REGOFF(a, b) 	TBYTE_PTR [b + a]
-#define D_REGOFF(a, b) 	QWORD_PTR [b + a]
-#define L_REGOFF(a, b) 	DWORD_PTR [b + a]
-#define W_REGOFF(a, b) 	WORD_PTR [b + a]
-#define B_REGOFF(a, b) 	BYTE_PTR [b + a]
+#define P_REGOFF(a, b)		[b + a]
+#define X_REGOFF(a, b)		TBYTE_PTR [b + a]
+#define D_REGOFF(a, b)		QWORD_PTR [b + a]
+#define L_REGOFF(a, b)		DWORD_PTR [b + a]
+#define W_REGOFF(a, b)		WORD_PTR [b + a]
+#define B_REGOFF(a, b)		BYTE_PTR [b + a]
 
 /* Reg indirect Base + Index + Displacement  - this is mainly for 16-bit mode
  * which has no scaling
  */
-#define P_REGBID(b, i, d) 	[b + i + d]
-#define X_REGBID(b, i, d) 	TBYTE_PTR [b + i + d]
-#define D_REGBID(b, i, d) 	QWORD_PTR [b + i + d]
-#define L_REGBID(b, i, d) 	DWORD_PTR [b + i + d]
-#define W_REGBID(b, i, d) 	WORD_PTR [b + i + d]
-#define B_REGBID(b, i, d) 	BYTE_PTR [b + i + d]
+#define P_REGBID(b, i, d)	[b + i + d]
+#define X_REGBID(b, i, d)	TBYTE_PTR [b + i + d]
+#define D_REGBID(b, i, d)	QWORD_PTR [b + i + d]
+#define L_REGBID(b, i, d)	DWORD_PTR [b + i + d]
+#define W_REGBID(b, i, d)	WORD_PTR [b + i + d]
+#define B_REGBID(b, i, d)	BYTE_PTR [b + i + d]
+
+/* Reg indirect Base + (Index * Scale) */
+#define P_REGBIS(b, i, s)	[b + i * s]
+#define X_REGBIS(b, i, s)	TBYTE_PTR [b + i * s]
+#define D_REGBIS(b, i, s)	QWORD_PTR [b + i * s]
+#define L_REGBIS(b, i, s)	DWORD_PTR [b + i * s]
+#define W_REGBIS(b, i, s)	WORD_PTR [b + i * s]
+#define B_REGBIS(b, i, s)	BYTE_PTR [b + i * s]
 
 /* Reg indirect Base + (Index * Scale) + Displacement */
-#define P_REGBISD(b, i, s, d) 	[b + i * s + d]
-#define X_REGBISD(b, i, s, d) 	TBYTE_PTR [b + i * s + d]
-#define D_REGBISD(b, i, s, d) 	QWORD_PTR [b + i * s + d]
-#define L_REGBISD(b, i, s, d) 	DWORD_PTR [b + i * s + d]
-#define W_REGBISD(b, i, s, d) 	WORD_PTR [b + i * s + d]
-#define B_REGBISD(b, i, s, d) 	BYTE_PTR [b + i * s + d]
+#define P_REGBISD(b, i, s, d)	[b + i * s + d]
+#define X_REGBISD(b, i, s, d)	TBYTE_PTR [b + i * s + d]
+#define D_REGBISD(b, i, s, d)	QWORD_PTR [b + i * s + d]
+#define L_REGBISD(b, i, s, d)	DWORD_PTR [b + i * s + d]
+#define W_REGBISD(b, i, s, d)	WORD_PTR [b + i * s + d]
+#define B_REGBISD(b, i, s, d)	BYTE_PTR [b + i * s + d]
 
 /* Displaced Scaled Index: */
-#define P_REGDIS(d, i, s) 	[i * s + d]
-#define X_REGDIS(d, i, s) 	TBYTE_PTR [i * s + d]
-#define D_REGDIS(d, i, s) 	QWORD_PTR [i * s + d]
-#define L_REGDIS(d, i, s) 	DWORD_PTR [i * s + d]
-#define W_REGDIS(d, i, s) 	WORD_PTR [i * s + d]
-#define B_REGDIS(d, i, s) 	BYTE_PTR [i * s + d]
+#define P_REGDIS(d, i, s)	[i * s + d]
+#define X_REGDIS(d, i, s)	TBYTE_PTR [i * s + d]
+#define D_REGDIS(d, i, s)	QWORD_PTR [i * s + d]
+#define L_REGDIS(d, i, s)	DWORD_PTR [i * s + d]
+#define W_REGDIS(d, i, s)	WORD_PTR [i * s + d]
+#define B_REGDIS(d, i, s)	BYTE_PTR [i * s + d]
 
 /* Indexed Base: */
-#define P_REGBI(b, i) 	[b + i]
-#define X_REGBI(b, i) 	TBYTE_PTR [b + i]
-#define D_REGBI(b, i) 	QWORD_PTR [b + i]
-#define L_REGBI(b, i) 	DWORD_PTR [b + i]
-#define W_REGBI(b, i) 	WORD_PTR [b + i]
-#define B_REGBI(b, i) 	BYTE_PTR [b + i]
+#define P_REGBI(b, i)		[b + i]
+#define X_REGBI(b, i)		TBYTE_PTR [b + i]
+#define D_REGBI(b, i)		QWORD_PTR [b + i]
+#define L_REGBI(b, i)		DWORD_PTR [b + i]
+#define W_REGBI(b, i)		WORD_PTR [b + i]
+#define B_REGBI(b, i)		BYTE_PTR [b + i]
 
 /* Displaced Base: */
-#define P_REGDB(d, b) 	[b + d]
-#define X_REGDB(d, b) 	TBYTE_PTR [b + d]
-#define D_REGDB(d, b) 	QWORD_PTR [b + d]
-#define L_REGDB(d, b) 	DWORD_PTR [b + d]
-#define W_REGDB(d, b) 	WORD_PTR [b + d]
-#define B_REGDB(d, b) 	BYTE_PTR [b + d]
+#define P_REGDB(d, b)		[b + d]
+#define X_REGDB(d, b)		TBYTE_PTR [b + d]
+#define D_REGDB(d, b)		QWORD_PTR [b + d]
+#define L_REGDB(d, b)		DWORD_PTR [b + d]
+#define W_REGDB(d, b)		WORD_PTR [b + d]
+#define B_REGDB(d, b)		BYTE_PTR [b + d]
 
 /* Variable indirect: */
-#define VARINDIRECT(var) 	var
+#define VARINDIRECT(var)	var
 
 /* Use register contents as jump/call target: */
-#define CODEPTR(reg) 	reg
+#define CODEPTR(reg)		reg
 
 /*
- * 	Redefine assembler commands
+ * Redefine assembler commands
  */
 
-#define P_(a) 	P_ ## a
-#define X_(a) 	X_ ## a
-#define D_(a) 	D_ ## a
-#define S_(a) 	L_ ## a
-#define L_(a) 	L_ ## a
-#define W_(a) 	W_ ## a
-#define B_(a) 	B_ ## a
-
-#define	AAA		aaa		
-#define	AAD		aad		
-#define	AAM		aam		
-#define	AAS		aas		
-#define	ADC_L(a, b)	adc L_(b), L_(a)
-#define	ADC_W(a, b)	adc W_(b), W_(a)
-#define	ADC_B(a, b)	adc B_(b), B_(a)
-#define	ADD_L(a, b)	add L_(b), L_(a)
-#define	ADD_W(a, b)	add W_(b), W_(a)
-#define	ADD_B(a, b)	add B_(b), B_(a)
-#define	AND_L(a, b)	and L_(b), L_(a)
-#define	AND_W(a, b)	and W_(b), W_(a)
-#define	AND_B(a, b)	and B_(b), B_(a)
-#define	ARPL(a,b)	arpl W_(b), a
-#define	BOUND_L(a, b)	bound L_(b), L_(a)
-#define	BOUND_W(a, b)	bound W_(b), W_(a)
-#define	BSF_L(a, b)	bsf L_(b), L_(a)
-#define	BSF_W(a, b)	bsf W_(b), W_(a)
-#define	BSR_L(a, b)	bsr L_(b), L_(a)
-#define	BSR_W(a, b)	bsr W_(b), W_(a)
-#define	BT_L(a, b)	bt L_(b), L_(a)
-#define	BT_W(a, b)	bt W_(b), W_(a)
-#define	BTC_L(a, b)	btc L_(b), L_(a)
-#define	BTC_W(a, b)	btc W_(b), W_(a)
-#define	BTR_L(a, b)	btr L_(b), L_(a)
-#define	BTR_W(a, b)	btr W_(b), W_(a)
-#define	BTS_L(a, b)	bts L_(b), L_(a)
-#define	BTS_W(a, b)	bts W_(b), W_(a)
-#define	CALL(a)		call a
-#define	CALLF(s,a)	call far s:a
-#define	CBW		cbw		
-#define	CWDE		cwde		
-#define	CLC		clc		
-#define	CLD		cld		
-#define	CLI		cli		
-#define	CLTS		clts		
-#define	CMC		cmc		
-#define	CMP_L(a, b)	cmp L_(b), L_(a)
-#define	CMP_W(a, b)	cmp W_(b), W_(a)
-#define	CMP_B(a, b)	cmp B_(b), B_(a)
-#define	CMPS_L		cmpsd
-#define	CMPS_W		cmpsw
-#define	CMPS_B		cmpsb
-#define	CWD		cwd		
-#define	CDQ		cdq		
-#define	DAA		daa		
-#define	DAS		das		
-#define	DEC_L(a)	dec L_(a)
-#define	DEC_W(a)	dec W_(a)
-#define	DEC_B(a)	dec B_(a)
-#define	DIV_L(a)	div L_(a)
-#define	DIV_W(a)	div W_(a)
-#define	DIV_B(a)	div B_(a)
-#define	ENTER(a,b)	enter b, a
-#define	HLT		hlt		
-#define	IDIV_L(a)	idiv L_(a)
-#define	IDIV_W(a)	idiv W_(a)
-#define	IDIV_B(a)	idiv B_(a)
-#define	IMUL_L(a, b)	imul L_(b), L_(a)
-#define	IMUL_W(a, b)	imul W_(b), W_(a)
-#define	IMUL_B(a)	imul B_(a)
-#define	IN_L		in EAX, DX
-#define	IN_W		in AX, DX
-#define	IN_B		in AL, DX
-#define	IN1_L(a)	in1 L_(a)
-#define	IN1_W(a)	in1 W_(a)
-#define	IN1_B(a)	in1 B_(a)
-#define	INC_L(a)	inc L_(a)
-#define	INC_W(a)	inc W_(a)
-#define	INC_B(a)	inc B_(a)
-#define	INS_L		ins
-#define	INS_W		ins
-#define	INS_B		ins
-#define	INT(a)		int B_(a)
-#define	INT3		int3		
-#define	INTO		into		
-#define	IRET		iret		
-#define	IRETD		iretd		
-#define	JA(a)		ja NEAR a
-#define	JAE(a)		jae NEAR a
-#define	JB(a)		jb NEAR a
-#define	JBE(a)		jbe NEAR a
-#define	JC(a)		jc NEAR a
-#define	JE(a)		je NEAR a
-#define	JG(a)		jg NEAR a
-#define	JGE(a)		jge NEAR a
-#define	JL(a)		jl NEAR a
-#define	JLE(a)		jle NEAR a
-#define	JNA(a)		jna NEAR a
-#define	JNAE(a)		jnae NEAR a
-#define	JNB(a)		jnb NEAR a
-#define	JNBE(a)		jnbe NEAR a
-#define	JNC(a)		jnc NEAR a
-#define	JNE(a)		jne NEAR a
-#define	JNG(a)		jng NEAR a
-#define	JNGE(a)		jnge NEAR a
-#define	JNL(a)		jnl NEAR a
-#define	JNLE(a)		jnle NEAR a
-#define	JNO(a)		jno NEAR a
-#define	JNP(a)		jnp NEAR a
-#define	JNS(a)		jns NEAR a
-#define	JNZ(a)		jnz NEAR a
-#define	JO(a)		jo NEAR a
-#define	JP(a)		jp NEAR a
-#define	JPE(a)		jpe NEAR a
-#define	JPO(a)		jpo NEAR a
-#define	JS(a)		js NEAR a
-#define	JZ(a)		jz NEAR a
-#define	JMP(a)		jmp a
-#define	JMPF(s,a)	jmpf
-#define	LAHF		lahf		
-#define	LAR(a, b)	lar b, a
-#define	LEA_L(a, b)	lea P_(b), P_(a)
-#define	LEA_W(a, b)	lea P_(b), P_(a)
-#define	LEAVE		leave		
-#define	LGDT(a)		lgdt a
-#define	LIDT(a)		lidt a
-#define	LDS(a, b)	lds b, a
-#define	LES(a, b)	les b, a
-#define	LFS(a, b)	lfs b, a
-#define	LGS(a, b)	lgs b, a
-#define	LSS(a, b)	lss b, a
-#define	LLDT(a)		lldt a
-#define	LMSW(a)		lmsw a
-#define LOCK		lock		
-#define	LODS_L		lodsd
-#define	LODS_W		lodsw
-#define	LODS_B		lodsb
-#define	LOOP(a)		loop a
-#define	LOOPE(a)	loope a
-#define	LOOPZ(a)	loopz a
-#define	LOOPNE(a)	loopne a
-#define	LOOPNZ(a)	loopnz a
-#define	LSL(a, b)	lsl b, a
-#define	LTR(a)		ltr a
-#define	MOV_SR(a, b)	mov S_(b), S_(a)
-#define	MOV_L(a, b)	mov L_(b), L_(a)
-#define	MOV_W(a, b)	mov W_(b), W_(a)
-#define	MOV_B(a, b)	mov B_(b), B_(a)
-#define	MOVS_L		movsd
-#define	MOVS_W		movsw
-#define	MOVS_B		movsb
-#define	MOVSX_BL(a, b)	movsx B_(b), B_(a)
-#define	MOVSX_BW(a, b)	movsx B_(b), B_(a)
-#define	MOVSX_WL(a, b)	movsx W_(b), W_(a)
-#define	MOVZX_BL(a, b)	movzx B_(b), B_(a)
-#define	MOVZX_BW(a, b)	movzx B_(b), B_(a)
-#define	MOVZX_WL(a, b)	movzx W_(b), W_(a)
-#define	MUL_L(a)	mul L_(a)
-#define	MUL_W(a)	mul W_(a)
-#define	MUL_B(a)	mul B_(a)
-#define	NEG_L(a)	neg L_(a)
-#define	NEG_W(a)	neg W_(a)
-#define	NEG_B(a)	neg B_(a)
-#define	NOP		nop		
-#define	NOT_L(a)	not L_(a)
-#define	NOT_W(a)	not W_(a)
-#define	NOT_B(a)	not B_(a)
-#define	OR_L(a,b)	or L_(b), L_(a)
-#define	OR_W(a,b)	or W_(b), W_(a)
-#define	OR_B(a,b)	or B_(b), B_(a)
-#define	OUT_L		out DX, EAX
-#define	OUT_W		out DX, AX
-#define	OUT_B		out DX, AL
-#define	OUT1_L(a)	out1 L_(a)
-#define	OUT1_W(a)	out1 W_(a)
-#define	OUT1_B(a)	out1 B_(a)
-#define	OUTS_L		outsd
-#define	OUTS_W		outsw
-#define	OUTS_B		outsb
-#define	POP_SR(a)	pop S_(a)
-#define	POP_L(a)	pop L_(a)
-#define	POP_W(a)	pop W_(a)
-#define	POPA_L		popad
-#define	POPA_W		popa
-#define	POPF_L		popfd
-#define	POPF_W		popf
-#define	PUSH_SR(a)	push S_(a)
-#define	PUSH_L(a)	push L_(a)
-#define	PUSH_W(a)	push W_(a)
-#define	PUSH_B(a)	push B_(a)
-#define	PUSHA_L		pushad
-#define	PUSHA_W		pusha
-#define	PUSHF_L		pushfd
-#define	PUSHF_W		pushf
-#define	RCL_L(a, b)	rcl L_(b), L_(a)
-#define	RCL_W(a, b)	rcl W_(b), W_(a)
-#define	RCL_B(a, b)	rcl B_(b), B_(a)
-#define	RCR_L(a, b)	rcr L_(b), L_(a)
-#define	RCR_W(a, b)	rcr W_(b), W_(a)
-#define	RCR_B(a, b)	rcr B_(b), B_(a)
-#define	ROL_L(a, b)	rol L_(b), L_(a)
-#define	ROL_W(a, b)	rol W_(b), W_(a)
-#define	ROL_B(a, b)	rol B_(b), B_(a)
-#define	ROR_L(a, b)	ror L_(b), L_(a)
-#define	ROR_W(a, b)	ror W_(b), W_(a)
-#define	ROR_B(a, b)	ror B_(b), B_(a)
-#define	REP		rep		
-#define	REPE		repe		
-#define	REPNE		repne		
-#define	REPNZ		REPNE
-#define	REPZ		REPE
-#define	RET		ret		
-#define	SAHF		sahf		
-#define	SAL_L(a, b)	sal L_(b), L_(a)
-#define	SAL_W(a, b)	sal W_(b), W_(a)
-#define	SAL_B(a, b)	sal B_(b), B_(a)
-#define	SAR_L(a, b)	sar L_(b), L_(a)
-#define	SAR_W(a, b)	sar W_(b), W_(a)
-#define	SAR_B(a, b)	sar B_(b), B_(a)
-#define	SBB_L(a, b)	sbb L_(b), L_(a)
-#define	SBB_W(a, b)	sbb W_(b), W_(a)
-#define	SBB_B(a, b)	sbb B_(b), B_(a)
-#define	SCAS_L		scas
-#define	SCAS_W		scas
-#define	SCAS_B		scas
-#define	SETA(a)		seta a
-#define	SETAE(a)	setae a
-#define	SETB(a)		setb a
-#define	SETBE(a)	setbe a
-#define	SETC(a)		setc a
-#define	SETE(a)		sete a
-#define	SETG(a)		setg a
-#define	SETGE(a)	setge a
-#define	SETL(a)		setl a
-#define	SETLE(a)	setle a
-#define	SETNA(a)	setna a
-#define	SETNAE(a)	setnae a
-#define	SETNB(a)	setnb a
-#define	SETNBE(a)	setnbe a
-#define	SETNC(a)	setnc a
-#define	SETNE(a)	setne a
-#define	SETNG(a)	setng a
-#define	SETNGE(a)	setnge a
-#define	SETNL(a)	setnl a
-#define	SETNLE(a)	setnle a
-#define	SETNO(a)	setno a
-#define	SETNP(a)	setnp a
-#define	SETNS(a)	setns a
-#define	SETNZ(a)	setnz a
-#define	SETO(a)		seto a
-#define	SETP(a)		setp a
-#define	SETPE(a)	setpe a
-#define	SETPO(a)	setpo a
-#define	SETS(a)		sets a
-#define	SETZ(a)		setz a
-#define	SGDT(a)		sgdt a
-#define	SIDT(a)		sidt a
-#define	SHL_L(a, b)	shl L_(b), L_(a)
-#define	SHL_W(a, b)	shl W_(b), W_(a)
-#define	SHL_B(a, b)	shl B_(b), B_(a)
-#define	SHLD_L(a,b,c)	shld
-#define	SHLD2_L(a,b)	shld L_(b), L_(a)
-#define	SHLD_W(a,b,c)	shld
-#define	SHLD2_W(a,b)	shld W_(b), W_(a)
-#define	SHR_L(a, b)	shr L_(b), L_(a)
-#define	SHR_W(a, b)	shr W_(b), W_(a)
-#define	SHR_B(a, b)	shr B_(b), B_(a)
-#define	SHRD_L(a,b,c)	shrd
-#define	SHRD2_L(a,b)	shrd L_(b), L_(a)
-#define	SHRD_W(a,b,c)	shrd
-#define	SHRD2_W(a,b)	shrd W_(b), W_(a)
-#define	SLDT(a)		sldt a
-#define	SMSW(a)		smsw a
-#define	STC		stc		
-#define	STD		std		
-#define	STI		sti		
-#define	STOS_L		stos
-#define	STOS_W		stos
-#define	STOS_B		stos
-#define	STR(a)		str a
-#define	SUB_L(a, b)	sub L_(b), L_(a)
-#define	SUB_W(a, b)	sub W_(b), W_(a)
-#define	SUB_B(a, b)	sub B_(b), B_(a)
-#define	TEST_L(a, b)	test L_(b), L_(a)
-#define	TEST_W(a, b)	test W_(b), W_(a)
-#define	TEST_B(a, b)	test B_(b), B_(a)
-#define	VERR(a)		verr a
-#define	VERW(a)		verw a
-#define	WAIT		wait		
-#define	XCHG_L(a, b)	xchg L_(b), L_(a)
-#define	XCHG_W(a, b)	xchg W_(b), W_(a)
-#define	XCHG_B(a, b)	xchg B_(b), B_(a)
-#define	XLAT		xlat		
-#define	XOR_L(a, b)	xor L_(b), L_(a)
-#define	XOR_W(a, b)	xor W_(b), W_(a)
-#define	XOR_B(a, b)	xor B_(b), B_(a)
-#define	F2XM1		f2xm1		
-#define	FABS		fabs		
-#define	FADD_D(a)	fadd D_(a)
-#define	FADD_S(a)	fadd S_(a)
-#define	FADD2(a, b)	fadd b, a
-#define	FADDP(a, b)	faddp b, a
-#define	FIADD_L(a)	fiadd L_(a)
-#define	FIADD_W(a)	fiadd W_(a)
-#define	FBLD(a)		fbld a
-#define	FBSTP(a)	fbstp a
-#define	FCHS		fchs		
-#define	FCLEX		fclex		
-#define	FNCLEX		fnclex		
-#define	FCOM(a)		fcom a
-#define	FCOM_D(a)	fcom D_(a)
-#define	FCOM_S(a)	fcom S_(a)
-#define	FCOMP(a)	fcomp a
-#define	FCOMP_D(a)	fcomp D_(a)
-#define	FCOMP_S(a)	fcomp S_(a)
-#define	FCOMPP		fcompp		
-#define	FCOS		fcos		
-#define	FDECSTP		fdecstp		
-#define	FDIV_D(a)	fdiv D_(a)
-#define	FDIV_S(a)	fdiv S_(a)
-#define	FDIV2(a, b)	fdiv b, a
-#define	FDIVP(a, b)	fdivp b, a
-#define	FIDIV_L(a)	fidiv L_(a)
-#define	FIDIV_W(a)	fidiv W_(a)
-#define	FDIVR_D(a)	fdivr D_(a)
-#define	FDIVR_S(a)	fdivr S_(a)
-#define	FDIVR2(a, b)	fdivr b, a
-#define	FDIVRP(a, b)	fdivrp b, a
-#define	FIDIVR_L(a)	fidivr L_(a)
-#define	FIDIVR_W(a)	fidivr W_(a)
-#define	FFREE(a)	ffree a
-#define	FICOM_L(a)	ficom L_(a)
-#define	FICOM_W(a)	ficom W_(a)
-#define	FICOMP_L(a)	ficomp L_(a)
-#define	FICOMP_W(a)	ficomp W_(a)
-#define	FILD_Q(a)	fild D_(a)
-#define	FILD_L(a)	fild L_(a)
-#define	FILD_W(a)	fild W_(a)
-#define	FINCSTP		fincstp		
-#define	FINIT		finit		
-#define	FNINIT		fninit		
-#define	FIST_L(a)	fist L_(a)
-#define	FIST_W(a)	fist W_(a)
-#define	FISTP_Q(a)	fistp D_(a)
-#define	FISTP_L(a)	fistp L_(a)
-#define	FISTP_W(a)	fistp W_(a)
-#define	FLD_X(a)	fld X_(a)
-#define	FLD_D(a)	fld D_(a)
-#define	FLD_S(a)	fld S_(a)
-#define	FLD1		fld1		
-#define	FLDL2T		fldl2t		
-#define	FLDL2E		fldl2e		
-#define	FLDPI		fldpi		
-#define	FLDLG2		fldlg2		
-#define	FLDLN2		fldln2		
-#define	FLDZ		fldz		
-#define	FLDCW(a)	fldcw a
-#define	FLDENV(a)	fldenv a
-#define	FMUL_S(a)	fmul S_(a)
-#define	FMUL_D(a)	fmul D_(a)
-#define	FMUL2(a, b)	fmul b, a
-#define	FMULP(a, b)	fmulp b, a
-#define	FIMUL_L(a)	fimul L_(a)
-#define	FIMUL_W(a)	fimul W_(a)
-#define	FNOP		fnop		
-#define	FPATAN		fpatan		
-#define	FPREM		fprem		
-#define	FPREM1		fprem1		
-#define	FPTAN		fptan		
-#define	FRNDINT		frndint		
-#define	FRSTOR(a)	frstor a
-#define	FSAVE(a)	fsave a
-#define	FNSAVE(a)	fnsave a
-#define	FSCALE		fscale		
-#define	FSIN		fsin		
-#define	FSINCOS		fsincos		
-#define	FSQRT		fsqrt		
-#define	FST_D(a)	fst D_(a)
-#define	FST_S(a)	fst S_(a)
-#define	FSTP_X(a)	fstp X_(a)
-#define	FSTP_D(a)	fstp D_(a)
-#define	FSTP_S(a)	fstp S_(a)
-#define	FSTP(a)		fstp a
-#define	FSTCW(a)	fstcw a
-#define	FNSTCW(a)	fnstcw a
-#define	FSTENV(a)	fstenv a
-#define	FNSTENV(a)	fnstenv a
-#define	FSTSW(a)	fstsw a
-#define	FNSTSW(a)	fnstsw a
-#define	FSUB_S(a)	fsub S_(a)
-#define	FSUB_D(a)	fsub D_(a)
-#define	FSUB2(a, b)	fsub b, a
-#define	FSUBP(a, b)	fsubp b, a
-#define	FISUB_L(a)	fisub L_(a)
-#define	FISUB_W(a)	fisub W_(a)
-#define	FSUBR_S(a)	fsubr S_(a)
-#define	FSUBR_D(a)	fsubr D_(a)
-#define	FSUBR2(a, b)	fsubr b, a
-#define	FSUBRP(a, b)	fsubrp b, a
-#define	FISUBR_L(a)	fisubr L_(a)
-#define	FISUBR_W(a)	fisubr W_(a)
-#define	FTST		ftst		
-#define	FUCOM(a)	fucom a
-#define	FUCOMP(a)	fucomp a
-#define	FUCOMPP		fucompp		
-#define	FWAIT		fwait		
-#define	FXAM		fxam		
-#define	FXCH(a)		fxch a
-#define	FXTRACT		fxtract		
-#define	FYL2X		fyl2x		
-#define	FYL2XP1		fyl2xp1		
+#define P_(a)			P_ ## a
+#define X_(a)			X_ ## a
+#define D_(a)			D_ ## a
+#define S_(a)			L_ ## a
+#define L_(a)			L_ ## a
+#define W_(a)			W_ ## a
+#define B_(a)			B_ ## a
+
+#define AAA			aaa
+#define AAD			aad
+#define AAM			aam
+#define AAS			aas
+#define ADC_L(a, b)		adc L_(b), L_(a)
+#define ADC_W(a, b)		adc W_(b), W_(a)
+#define ADC_B(a, b)		adc B_(b), B_(a)
+#define ADD_L(a, b)		add L_(b), L_(a)
+#define ADD_W(a, b)		add W_(b), W_(a)
+#define ADD_B(a, b)		add B_(b), B_(a)
+#define AND_L(a, b)		and L_(b), L_(a)
+#define AND_W(a, b)		and W_(b), W_(a)
+#define AND_B(a, b)		and B_(b), B_(a)
+#define ARPL(a,b)		arpl W_(b), a
+#define BOUND_L(a, b)		bound L_(b), L_(a)
+#define BOUND_W(a, b)		bound W_(b), W_(a)
+#define BSF_L(a, b)		bsf L_(b), L_(a)
+#define BSF_W(a, b)		bsf W_(b), W_(a)
+#define BSR_L(a, b)		bsr L_(b), L_(a)
+#define BSR_W(a, b)		bsr W_(b), W_(a)
+#define BT_L(a, b)		bt L_(b), L_(a)
+#define BT_W(a, b)		bt W_(b), W_(a)
+#define BTC_L(a, b)		btc L_(b), L_(a)
+#define BTC_W(a, b)		btc W_(b), W_(a)
+#define BTR_L(a, b)		btr L_(b), L_(a)
+#define BTR_W(a, b)		btr W_(b), W_(a)
+#define BTS_L(a, b)		bts L_(b), L_(a)
+#define BTS_W(a, b)		bts W_(b), W_(a)
+#define CALL(a)			call a
+#define CALLF(s,a)		call far s:a
+#define CBW			cbw
+#define CWDE			cwde
+#define CLC			clc
+#define CLD			cld
+#define CLI			cli
+#define CLTS			clts
+#define CMC			cmc
+#define CMP_L(a, b)		cmp L_(b), L_(a)
+#define CMP_W(a, b)		cmp W_(b), W_(a)
+#define CMP_B(a, b)		cmp B_(b), B_(a)
+#define CMPS_L			cmpsd
+#define CMPS_W			cmpsw
+#define CMPS_B			cmpsb
+#define CPUID			cpuid
+#define CWD			cwd
+#define CDQ			cdq
+#define DAA			daa
+#define DAS			das
+#define DEC_L(a)		dec L_(a)
+#define DEC_W(a)		dec W_(a)
+#define DEC_B(a)		dec B_(a)
+#define DIV_L(a)		div L_(a)
+#define DIV_W(a)		div W_(a)
+#define DIV_B(a)		div B_(a)
+#define ENTER(a,b)		enter b, a
+#define HLT			hlt
+#define IDIV_L(a)		idiv L_(a)
+#define IDIV_W(a)		idiv W_(a)
+#define IDIV_B(a)		idiv B_(a)
+#define IMUL_L(a, b)		imul L_(b), L_(a)
+#define IMUL_W(a, b)		imul W_(b), W_(a)
+#define IMUL_B(a)		imul B_(a)
+#define IN_L			in EAX, DX
+#define IN_W			in AX, DX
+#define IN_B			in AL, DX
+#define IN1_L(a)		in1 L_(a)
+#define IN1_W(a)		in1 W_(a)
+#define IN1_B(a)		in1 B_(a)
+#define INC_L(a)		inc L_(a)
+#define INC_W(a)		inc W_(a)
+#define INC_B(a)		inc B_(a)
+#define INS_L			ins
+#define INS_W			ins
+#define INS_B			ins
+#define INT(a)			int B_(a)
+#define INT3			int3
+#define INTO			into
+#define IRET			iret
+#define IRETD			iretd
+#define JA(a)			ja NEAR a
+#define JAE(a)			jae NEAR a
+#define JB(a)			jb NEAR a
+#define JBE(a)			jbe NEAR a
+#define JC(a)			jc NEAR a
+#define JE(a)			je NEAR a
+#define JG(a)			jg NEAR a
+#define JGE(a)			jge NEAR a
+#define JL(a)			jl NEAR a
+#define JLE(a)			jle NEAR a
+#define JNA(a)			jna NEAR a
+#define JNAE(a)			jnae NEAR a
+#define JNB(a)			jnb NEAR a
+#define JNBE(a)			jnbe NEAR a
+#define JNC(a)			jnc NEAR a
+#define JNE(a)			jne NEAR a
+#define JNG(a)			jng NEAR a
+#define JNGE(a)			jnge NEAR a
+#define JNL(a)			jnl NEAR a
+#define JNLE(a)			jnle NEAR a
+#define JNO(a)			jno NEAR a
+#define JNP(a)			jnp NEAR a
+#define JNS(a)			jns NEAR a
+#define JNZ(a)			jnz NEAR a
+#define JO(a)			jo NEAR a
+#define JP(a)			jp NEAR a
+#define JPE(a)			jpe NEAR a
+#define JPO(a)			jpo NEAR a
+#define JS(a)			js NEAR a
+#define JZ(a)			jz NEAR a
+#define JMP(a)			jmp a
+#define JMPF(s,a)		jmpf
+#define LAHF			lahf
+#define LAR(a, b)		lar b, a
+#define LEA_L(a, b)		lea P_(b), P_(a)
+#define LEA_W(a, b)		lea P_(b), P_(a)
+#define LEAVE			leave
+#define LGDT(a)			lgdt a
+#define LIDT(a)			lidt a
+#define LDS(a, b)		lds b, a
+#define LES(a, b)		les b, a
+#define LFS(a, b)		lfs b, a
+#define LGS(a, b)		lgs b, a
+#define LSS(a, b)		lss b, a
+#define LLDT(a)			lldt a
+#define LMSW(a)			lmsw a
+#define LOCK			lock
+#define LODS_L			lodsd
+#define LODS_W			lodsw
+#define LODS_B			lodsb
+#define LOOP(a)			loop a
+#define LOOPE(a)		loope a
+#define LOOPZ(a)		loopz a
+#define LOOPNE(a)		loopne a
+#define LOOPNZ(a)		loopnz a
+#define LSL(a, b)		lsl b, a
+#define LTR(a)			ltr a
+#define MOV_SR(a, b)		mov S_(b), S_(a)
+#define MOV_L(a, b)		mov L_(b), L_(a)
+#define MOV_W(a, b)		mov W_(b), W_(a)
+#define MOV_B(a, b)		mov B_(b), B_(a)
+#define MOVS_L			movsd
+#define MOVS_W			movsw
+#define MOVS_B			movsb
+#define MOVSX_BL(a, b)		movsx B_(b), B_(a)
+#define MOVSX_BW(a, b)		movsx B_(b), B_(a)
+#define MOVSX_WL(a, b)		movsx W_(b), W_(a)
+#define MOVZX_BL(a, b)		movzx B_(b), B_(a)
+#define MOVZX_BW(a, b)		movzx B_(b), B_(a)
+#define MOVZX_WL(a, b)		movzx W_(b), W_(a)
+#define MUL_L(a)		mul L_(a)
+#define MUL_W(a)		mul W_(a)
+#define MUL_B(a)		mul B_(a)
+#define NEG_L(a)		neg L_(a)
+#define NEG_W(a)		neg W_(a)
+#define NEG_B(a)		neg B_(a)
+#define NOP			nop
+#define NOT_L(a)		not L_(a)
+#define NOT_W(a)		not W_(a)
+#define NOT_B(a)		not B_(a)
+#define OR_L(a,b)		or L_(b), L_(a)
+#define OR_W(a,b)		or W_(b), W_(a)
+#define OR_B(a,b)		or B_(b), B_(a)
+#define OUT_L			out DX, EAX
+#define OUT_W			out DX, AX
+#define OUT_B			out DX, AL
+#define OUT1_L(a)		out1 L_(a)
+#define OUT1_W(a)		out1 W_(a)
+#define OUT1_B(a)		out1 B_(a)
+#define OUTS_L			outsd
+#define OUTS_W			outsw
+#define OUTS_B			outsb
+#define POP_SR(a)		pop S_(a)
+#define POP_L(a)		pop L_(a)
+#define POP_W(a)		pop W_(a)
+#define POPA_L			popad
+#define POPA_W			popa
+#define POPF_L			popfd
+#define POPF_W			popf
+#define PUSH_SR(a)		push S_(a)
+#define PUSH_L(a)		push L_(a)
+#define PUSH_W(a)		push W_(a)
+#define PUSH_B(a)		push B_(a)
+#define PUSHA_L			pushad
+#define PUSHA_W			pusha
+#define PUSHF_L			pushfd
+#define PUSHF_W			pushf
+#define RCL_L(a, b)		rcl L_(b), L_(a)
+#define RCL_W(a, b)		rcl W_(b), W_(a)
+#define RCL_B(a, b)		rcl B_(b), B_(a)
+#define RCR_L(a, b)		rcr L_(b), L_(a)
+#define RCR_W(a, b)		rcr W_(b), W_(a)
+#define RCR_B(a, b)		rcr B_(b), B_(a)
+#define RDTSC			rdtsc
+#define ROL_L(a, b)		rol L_(b), L_(a)
+#define ROL_W(a, b)		rol W_(b), W_(a)
+#define ROL_B(a, b)		rol B_(b), B_(a)
+#define ROR_L(a, b)		ror L_(b), L_(a)
+#define ROR_W(a, b)		ror W_(b), W_(a)
+#define ROR_B(a, b)		ror B_(b), B_(a)
+#define REP			rep
+#define REPE			repe
+#define REPNE			repne
+#define REPNZ			REPNE
+#define REPZ			REPE
+#define RET			ret
+#define SAHF			sahf
+#define SAL_L(a, b)		sal L_(b), L_(a)
+#define SAL_W(a, b)		sal W_(b), W_(a)
+#define SAL_B(a, b)		sal B_(b), B_(a)
+#define SAR_L(a, b)		sar L_(b), L_(a)
+#define SAR_W(a, b)		sar W_(b), W_(a)
+#define SAR_B(a, b)		sar B_(b), B_(a)
+#define SBB_L(a, b)		sbb L_(b), L_(a)
+#define SBB_W(a, b)		sbb W_(b), W_(a)
+#define SBB_B(a, b)		sbb B_(b), B_(a)
+#define SCAS_L			scas
+#define SCAS_W			scas
+#define SCAS_B			scas
+#define SETA(a)			seta a
+#define SETAE(a)		setae a
+#define SETB(a)			setb a
+#define SETBE(a)		setbe a
+#define SETC(a)			setc a
+#define SETE(a)			sete a
+#define SETG(a)			setg a
+#define SETGE(a)		setge a
+#define SETL(a)			setl a
+#define SETLE(a)		setle a
+#define SETNA(a)		setna a
+#define SETNAE(a)		setnae a
+#define SETNB(a)		setnb a
+#define SETNBE(a)		setnbe a
+#define SETNC(a)		setnc a
+#define SETNE(a)		setne a
+#define SETNG(a)		setng a
+#define SETNGE(a)		setnge a
+#define SETNL(a)		setnl a
+#define SETNLE(a)		setnle a
+#define SETNO(a)		setno a
+#define SETNP(a)		setnp a
+#define SETNS(a)		setns a
+#define SETNZ(a)		setnz a
+#define SETO(a)			seto a
+#define SETP(a)			setp a
+#define SETPE(a)		setpe a
+#define SETPO(a)		setpo a
+#define SETS(a)			sets a
+#define SETZ(a)			setz a
+#define SGDT(a)			sgdt a
+#define SIDT(a)			sidt a
+#define SHL_L(a, b)		shl L_(b), L_(a)
+#define SHL_W(a, b)		shl W_(b), W_(a)
+#define SHL_B(a, b)		shl B_(b), B_(a)
+#define SHLD_L(a,b,c)		shld
+#define SHLD2_L(a,b)		shld L_(b), L_(a)
+#define SHLD_W(a,b,c)		shld
+#define SHLD2_W(a,b)		shld W_(b), W_(a)
+#define SHR_L(a, b)		shr L_(b), L_(a)
+#define SHR_W(a, b)		shr W_(b), W_(a)
+#define SHR_B(a, b)		shr B_(b), B_(a)
+#define SHRD_L(a,b,c)		shrd
+#define SHRD2_L(a,b)		shrd L_(b), L_(a)
+#define SHRD_W(a,b,c)		shrd
+#define SHRD2_W(a,b)		shrd W_(b), W_(a)
+#define SLDT(a)			sldt a
+#define SMSW(a)			smsw a
+#define STC			stc
+#define STD			std
+#define STI			sti
+#define STOS_L			stos
+#define STOS_W			stos
+#define STOS_B			stos
+#define STR(a)			str a
+#define SUB_L(a, b)		sub L_(b), L_(a)
+#define SUB_W(a, b)		sub W_(b), W_(a)
+#define SUB_B(a, b)		sub B_(b), B_(a)
+#define TEST_L(a, b)		test L_(b), L_(a)
+#define TEST_W(a, b)		test W_(b), W_(a)
+#define TEST_B(a, b)		test B_(b), B_(a)
+#define VERR(a)			verr a
+#define VERW(a)			verw a
+#define WAIT			wait
+#define XCHG_L(a, b)		xchg L_(b), L_(a)
+#define XCHG_W(a, b)		xchg W_(b), W_(a)
+#define XCHG_B(a, b)		xchg B_(b), B_(a)
+#define XLAT			xlat
+#define XOR_L(a, b)		xor L_(b), L_(a)
+#define XOR_W(a, b)		xor W_(b), W_(a)
+#define XOR_B(a, b)		xor B_(b), B_(a)
 
-/* New instructions */
-#define	CPUID		D_BYTE 15, 162
-#define	RDTSC		D_BYTE 15, 49
+
+/* Floating Point Instructions */
+#define F2XM1			f2xm1
+#define FABS			fabs
+#define FADD_D(a)		fadd D_(a)
+#define FADD_S(a)		fadd S_(a)
+#define FADD2(a, b)		fadd b, a
+#define FADDP(a, b)		faddp b, a
+#define FIADD_L(a)		fiadd L_(a)
+#define FIADD_W(a)		fiadd W_(a)
+#define FBLD(a)			fbld a
+#define FBSTP(a)		fbstp a
+#define FCHS			fchs
+#define FCLEX			fclex
+#define FNCLEX			fnclex
+#define FCOM(a)			fcom a
+#define FCOM_D(a)		fcom D_(a)
+#define FCOM_S(a)		fcom S_(a)
+#define FCOMP(a)		fcomp a
+#define FCOMP_D(a)		fcomp D_(a)
+#define FCOMP_S(a)		fcomp S_(a)
+#define FCOMPP			fcompp
+#define FCOS			fcos
+#define FDECSTP			fdecstp
+#define FDIV_D(a)		fdiv D_(a)
+#define FDIV_S(a)		fdiv S_(a)
+#define FDIV2(a, b)		fdiv b, a
+#define FDIVP(a, b)		fdivp b, a
+#define FIDIV_L(a)		fidiv L_(a)
+#define FIDIV_W(a)		fidiv W_(a)
+#define FDIVR_D(a)		fdivr D_(a)
+#define FDIVR_S(a)		fdivr S_(a)
+#define FDIVR2(a, b)		fdivr b, a
+#define FDIVRP(a, b)		fdivrp b, a
+#define FIDIVR_L(a)		fidivr L_(a)
+#define FIDIVR_W(a)		fidivr W_(a)
+#define FFREE(a)		ffree a
+#define FICOM_L(a)		ficom L_(a)
+#define FICOM_W(a)		ficom W_(a)
+#define FICOMP_L(a)		ficomp L_(a)
+#define FICOMP_W(a)		ficomp W_(a)
+#define FILD_Q(a)		fild D_(a)
+#define FILD_L(a)		fild L_(a)
+#define FILD_W(a)		fild W_(a)
+#define FINCSTP			fincstp
+#define FINIT			finit
+#define FNINIT			fninit
+#define FIST_L(a)		fist L_(a)
+#define FIST_W(a)		fist W_(a)
+#define FISTP_Q(a)		fistp D_(a)
+#define FISTP_L(a)		fistp L_(a)
+#define FISTP_W(a)		fistp W_(a)
+#define FLD_X(a)		fld X_(a)
+#define FLD_D(a)		fld D_(a)
+#define FLD_S(a)		fld S_(a)
+#define FLD1			fld1
+#define FLDL2T			fldl2t
+#define FLDL2E			fldl2e
+#define FLDPI			fldpi
+#define FLDLG2			fldlg2
+#define FLDLN2			fldln2
+#define FLDZ			fldz
+#define FLDCW(a)		fldcw a
+#define FLDENV(a)		fldenv a
+#define FMUL_S(a)		fmul S_(a)
+#define FMUL_D(a)		fmul D_(a)
+#define FMUL2(a, b)		fmul b, a
+#define FMULP(a, b)		fmulp b, a
+#define FIMUL_L(a)		fimul L_(a)
+#define FIMUL_W(a)		fimul W_(a)
+#define FNOP			fnop
+#define FPATAN			fpatan
+#define FPREM			fprem
+#define FPREM1			fprem1
+#define FPTAN			fptan
+#define FRNDINT			frndint
+#define FRSTOR(a)		frstor a
+#define FSAVE(a)		fsave a
+#define FNSAVE(a)		fnsave a
+#define FSCALE			fscale
+#define FSIN			fsin
+#define FSINCOS			fsincos
+#define FSQRT			fsqrt
+#define FST_D(a)		fst D_(a)
+#define FST_S(a)		fst S_(a)
+#define FSTP_X(a)		fstp X_(a)
+#define FSTP_D(a)		fstp D_(a)
+#define FSTP_S(a)		fstp S_(a)
+#define FSTP(a)			fstp a
+#define FSTCW(a)		fstcw a
+#define FNSTCW(a)		fnstcw a
+#define FSTENV(a)		fstenv a
+#define FNSTENV(a)		fnstenv a
+#define FSTSW(a)		fstsw a
+#define FNSTSW(a)		fnstsw a
+#define FSUB_S(a)		fsub S_(a)
+#define FSUB_D(a)		fsub D_(a)
+#define FSUB2(a, b)		fsub b, a
+#define FSUBP(a, b)		fsubp b, a
+#define FISUB_L(a)		fisub L_(a)
+#define FISUB_W(a)		fisub W_(a)
+#define FSUBR_S(a)		fsubr S_(a)
+#define FSUBR_D(a)		fsubr D_(a)
+#define FSUBR2(a, b)		fsubr b, a
+#define FSUBRP(a, b)		fsubrp b, a
+#define FISUBR_L(a)		fisubr L_(a)
+#define FISUBR_W(a)		fisubr W_(a)
+#define FTST			ftst
+#define FUCOM(a)		fucom a
+#define FUCOMP(a)		fucomp a
+#define FUCOMPP			fucompp
+#define FWAIT			fwait
+#define FXAM			fxam
+#define FXCH(a)			fxch a
+#define FXTRACT			fxtract
+#define FYL2X			fyl2x
+#define FYL2XP1			fyl2xp1
 
 #endif /* NASM_ASSEMBLER, MASM_ASSEMBLER */
 
@@ -1474,182 +1488,188 @@ SECTION _DATA public align=16 class=DATA use32 flat
 	/****************************************/
 
 #if defined(NASM_ASSEMBLER) || defined(MASM_ASSEMBLER)
-#define P_ARG1(a) 	P_ ## a
-#define P_ARG2(a, b) 	P_ ## b, P_ ## a
-#define P_ARG3(a, b, c) 	P_ ## c, P_ ## b, P_ ## a
+#define P_ARG1(a)		P_ ## a
+#define P_ARG2(a, b)		P_ ## b, P_ ## a
+#define P_ARG3(a, b, c)		P_ ## c, P_ ## b, P_ ## a
 #else
-#define P_ARG1(a) 	a
-#define P_ARG2(a, b) 	a, b
-#define P_ARG3(a, b, c) 	a, b, c
+#define P_ARG1(a)		a
+#define P_ARG2(a, b)		a, b
+#define P_ARG3(a, b, c)		a, b, c
 #endif
 
 /* MMX */
-#define MOVD(a, b) 	movd P_ARG2(a, b)
-#define MOVQ(a, b) 	movq P_ARG2(a, b)
+#define MOVD(a, b)		movd P_ARG2(a, b)
+#define MOVQ(a, b)		movq P_ARG2(a, b)
 
-#define PADDB(a, b) 	paddb P_ARG2(a, b)
-#define PADDW(a, b) 	paddw P_ARG2(a, b)
-#define PADDD(a, b) 	paddd P_ARG2(a, b)
+#define PADDB(a, b)		paddb P_ARG2(a, b)
+#define PADDW(a, b)		paddw P_ARG2(a, b)
+#define PADDD(a, b)		paddd P_ARG2(a, b)
 
-#define PADDSB(a, b) 	paddsb P_ARG2(a, b)
-#define PADDSW(a, b) 	paddsw P_ARG2(a, b)
+#define PADDSB(a, b)		paddsb P_ARG2(a, b)
+#define PADDSW(a, b)		paddsw P_ARG2(a, b)
 
-#define PADDUSB(a, b) 	paddusb P_ARG2(a, b)
-#define PADDUSW(a, b) 	paddusw P_ARG2(a, b)
+#define PADDUSB(a, b)		paddusb P_ARG2(a, b)
+#define PADDUSW(a, b)		paddusw P_ARG2(a, b)
 
-#define PSUBB(a, b) 	psubb P_ARG2(a, b)
-#define PSUBW(a, b) 	psubw P_ARG2(a, b)
-#define PSUBD(a, b) 	psubd P_ARG2(a, b)
+#define PSUBB(a, b)		psubb P_ARG2(a, b)
+#define PSUBW(a, b)		psubw P_ARG2(a, b)
+#define PSUBD(a, b)		psubd P_ARG2(a, b)
 
-#define PSUBSB(a, b) 	psubsb P_ARG2(a, b)
-#define PSUBSW(a, b) 	psubsw P_ARG2(a, b)
+#define PSUBSB(a, b)		psubsb P_ARG2(a, b)
+#define PSUBSW(a, b)		psubsw P_ARG2(a, b)
 
-#define PSUBUSB(a, b) 	psubusb P_ARG2(a, b)
-#define PSUBUSW(a, b) 	psubusw P_ARG2(a, b)
+#define PSUBUSB(a, b)		psubusb P_ARG2(a, b)
+#define PSUBUSW(a, b)		psubusw P_ARG2(a, b)
 
-#define PCMPEQB(a, b) 	pcmpeqb P_ARG2(a, b)
-#define PCMPEQW(a, b) 	pcmpeqw P_ARG2(a, b)
-#define PCMPEQD(a, b) 	pcmpeqd P_ARG2(a, b)
+#define PCMPEQB(a, b)		pcmpeqb P_ARG2(a, b)
+#define PCMPEQW(a, b)		pcmpeqw P_ARG2(a, b)
+#define PCMPEQD(a, b)		pcmpeqd P_ARG2(a, b)
 
-#define PCMPGTB(a, b) 	pcmpgtb P_ARG2(a, b)
-#define PCMPGTW(a, b) 	pcmpgtw P_ARG2(a, b)
-#define PCMPGTD(a, b) 	pcmpgtd P_ARG2(a, b)
+#define PCMPGTB(a, b)		pcmpgtb P_ARG2(a, b)
+#define PCMPGTW(a, b)		pcmpgtw P_ARG2(a, b)
+#define PCMPGTD(a, b)		pcmpgtd P_ARG2(a, b)
 
-#define PMULHW(a, b) 	pmulhw P_ARG2(a, b)
-#define PMULLW(a, b) 	pmullw P_ARG2(a, b)
+#define PMULHW(a, b)		pmulhw P_ARG2(a, b)
+#define PMULLW(a, b)		pmullw P_ARG2(a, b)
 
-#define PMADDWD(a, b) 	pmaddwd P_ARG2(a, b)
+#define PMADDWD(a, b)		pmaddwd P_ARG2(a, b)
 
-#define PAND(a, b) 	pand P_ARG2(a, b)
+#define PAND(a, b)		pand P_ARG2(a, b)
 
-#define PANDN(a, b) 	pandn P_ARG2(a, b)
+#define PANDN(a, b)		pandn P_ARG2(a, b)
 
-#define POR(a, b) 	por P_ARG2(a, b)
+#define POR(a, b)		por P_ARG2(a, b)
 
-#define PXOR(a, b) 	pxor P_ARG2(a, b)
+#define PXOR(a, b)		pxor P_ARG2(a, b)
 
-#define PSRAW(a, b) 	psraw P_ARG2(a, b)
-#define PSRAD(a, b) 	psrad P_ARG2(a, b)
+#define PSRAW(a, b)		psraw P_ARG2(a, b)
+#define PSRAD(a, b)		psrad P_ARG2(a, b)
 
-#define PSRLW(a, b) 	psrlw P_ARG2(a, b)
-#define PSRLD(a, b) 	psrld P_ARG2(a, b)
-#define PSRLQ(a, b) 	psrlq P_ARG2(a, b)
+#define PSRLW(a, b)		psrlw P_ARG2(a, b)
+#define PSRLD(a, b)		psrld P_ARG2(a, b)
+#define PSRLQ(a, b)		psrlq P_ARG2(a, b)
 
-#define PSLLW(a, b) 	psllw P_ARG2(a, b)
-#define PSLLD(a, b) 	pslld P_ARG2(a, b)
-#define PSLLQ(a, b) 	psllq P_ARG2(a, b)
+#define PSLLW(a, b)		psllw P_ARG2(a, b)
+#define PSLLD(a, b)		pslld P_ARG2(a, b)
+#define PSLLQ(a, b)		psllq P_ARG2(a, b)
 
-#define PACKSSWB(a, b) 	packsswb P_ARG2(a, b)
-#define PACKSSDW(a, b) 	packssdw P_ARG2(a, b)
-#define PACKUSWB(a, b) 	packuswb P_ARG2(a, b)
+#define PACKSSWB(a, b)		packsswb P_ARG2(a, b)
+#define PACKSSDW(a, b)		packssdw P_ARG2(a, b)
+#define PACKUSWB(a, b)		packuswb P_ARG2(a, b)
 
-#define PUNPCKHBW(a, b) 	punpckhbw P_ARG2(a, b)
-#define PUNPCKHWD(a, b) 	punpckhwd P_ARG2(a, b)
-#define PUNPCKHDQ(a, b) 	punpckhdq P_ARG2(a, b)
-#define PUNPCKLBW(a, b) 	punpcklbw P_ARG2(a, b)
-#define PUNPCKLWD(a, b) 	punpcklwd P_ARG2(a, b)
-#define PUNPCKLDQ(a, b) 	punpckldq P_ARG2(a, b)
+#define PUNPCKHBW(a, b)		punpckhbw P_ARG2(a, b)
+#define PUNPCKHWD(a, b)		punpckhwd P_ARG2(a, b)
+#define PUNPCKHDQ(a, b)		punpckhdq P_ARG2(a, b)
+#define PUNPCKLBW(a, b)		punpcklbw P_ARG2(a, b)
+#define PUNPCKLWD(a, b)		punpcklwd P_ARG2(a, b)
+#define PUNPCKLDQ(a, b)		punpckldq P_ARG2(a, b)
 
-#define EMMS 	emms
+#define EMMS			emms
 
 /* AMD 3DNow! */
-#define PAVGUSB(a, b) 	pavgusb P_ARG2(a, b)
-#define PFADD(a, b) 	pfadd P_ARG2(a, b)
-#define PFSUB(a, b) 	pfsub P_ARG2(a, b)
-#define PFSUBR(a, b) 	pfsubr P_ARG2(a, b)
-#define PFACC(a, b) 	pfacc P_ARG2(a, b)
-#define PFCMPGE(a, b) 	pfcmpge P_ARG2(a, b)
-#define PFCMPGT(a, b) 	pfcmpgt P_ARG2(a, b)
-#define PFCMPEQ(a, b) 	pfcmpeq P_ARG2(a, b)
-#define PFMIN(a, b) 	pfmin P_ARG2(a, b)
-#define PFMAX(a, b) 	pfmax P_ARG2(a, b)
-#define PI2FD(a, b) 	pi2fd P_ARG2(a, b)
-#define PF2ID(a, b) 	pf2id P_ARG2(a, b)
-#define PFRCP(a, b) 	pfrcp P_ARG2(a, b)
-#define PFRSQRT(a, b) 	pfrsqrt P_ARG2(a, b)
-#define PFMUL(a, b) 	pfmul P_ARG2(a, b)
-#define PFRCPIT1(a, b) 	pfrcpit1 P_ARG2(a, b)
-#define PFRSQIT1(a, b) 	pfrsqit1 P_ARG2(a, b)
-#define PFRCPIT2(a, b) 	pfrcpit2 P_ARG2(a, b)
-#define PMULHRW(a, b) 	pmulhrw P_ARG2(a, b)
-
-#define FEMMS	femms
-#define PREFETCH(a) 	prefetch P_ARG1(a)
-#define PREFETCHW(a)    prefetchw P_ARG1(a)
+#define PAVGUSB(a, b)		pavgusb P_ARG2(a, b)
+#define PFADD(a, b)		pfadd P_ARG2(a, b)
+#define PFSUB(a, b)		pfsub P_ARG2(a, b)
+#define PFSUBR(a, b)		pfsubr P_ARG2(a, b)
+#define PFACC(a, b)		pfacc P_ARG2(a, b)
+#define PFCMPGE(a, b)		pfcmpge P_ARG2(a, b)
+#define PFCMPGT(a, b)		pfcmpgt P_ARG2(a, b)
+#define PFCMPEQ(a, b)		pfcmpeq P_ARG2(a, b)
+#define PFMIN(a, b)		pfmin P_ARG2(a, b)
+#define PFMAX(a, b)		pfmax P_ARG2(a, b)
+#define PI2FD(a, b)		pi2fd P_ARG2(a, b)
+#define PF2ID(a, b)		pf2id P_ARG2(a, b)
+#define PFRCP(a, b)		pfrcp P_ARG2(a, b)
+#define PFRSQRT(a, b)		pfrsqrt P_ARG2(a, b)
+#define PFMUL(a, b)		pfmul P_ARG2(a, b)
+#define PFRCPIT1(a, b)		pfrcpit1 P_ARG2(a, b)
+#define PFRSQIT1(a, b)		pfrsqit1 P_ARG2(a, b)
+#define PFRCPIT2(a, b)		pfrcpit2 P_ARG2(a, b)
+#define PMULHRW(a, b)		pmulhrw P_ARG2(a, b)
+
+#define FEMMS			femms
+#define PREFETCH(a)		prefetch P_ARG1(a)
+#define PREFETCHW(a)		prefetchw P_ARG1(a)
 
 /* Intel SSE */
-#define ADDPS(a, b)	addps P_ARG2(a, b)
-#define ADDSS(a, b)	addss P_ARG2(a, b)
-#define ANDNPS(a, b)	andnps P_ARG2(a, b)
-#define ANDPS(a, b)	andps P_ARG2(a, b)
-/*
- NASM only knows the pseudo ops for these.
-	#define CMPPS(a, b, c)	cmpps P_ARG3(a, b, c)
-	#define CMPSS(a, b, c)	cmpss P_ARG3(a, b, c)
+#define ADDPS(a, b)		addps P_ARG2(a, b)
+#define ADDSS(a, b)		addss P_ARG2(a, b)
+#define ANDNPS(a, b)		andnps P_ARG2(a, b)
+#define ANDPS(a, b)		andps P_ARG2(a, b)
+/* NASM only knows the pseudo ops for these.
+#define CMPPS(a, b, c)		cmpps P_ARG3(a, b, c)
+#define CMPSS(a, b, c)		cmpss P_ARG3(a, b, c)
 */
-#define CMPEQPS(a, b)	cmpeqps P_ARG2(a, b)
-#define CMPLTPS(a, b)	cmpltps P_ARG2(a, b)
-#define CMPLEPS(a, b)	cmpleps P_ARG2(a, b)
+#define CMPEQPS(a, b)		cmpeqps P_ARG2(a, b)
+#define CMPLTPS(a, b)		cmpltps P_ARG2(a, b)
+#define CMPLEPS(a, b)		cmpleps P_ARG2(a, b)
 #define CMPUNORDPS(a, b)	cmpunordps P_ARG2(a, b)
-#define CMPNEQPS(a, b)	cmpneqps P_ARG2(a, b)
-#define CMPNLTPS(a, b)	cmpnltps P_ARG2(a, b)
-#define CMPNLEPS(a, b)	cmpnleps P_ARG2(a, b)
-#define CMPORDPS(a, b)	cmpordps P_ARG2(a, b)
-#define CMPEQSS(a, b)	cmpeqss P_ARG2(a, b)
-#define CMPLTSS(a, b)	cmpltss P_ARG2(a, b)
-#define CMPLESS(a, b)	cmpless P_ARG2(a, b)
+#define CMPNEQPS(a, b)		cmpneqps P_ARG2(a, b)
+#define CMPNLTPS(a, b)		cmpnltps P_ARG2(a, b)
+#define CMPNLEPS(a, b)		cmpnleps P_ARG2(a, b)
+#define CMPORDPS(a, b)		cmpordps P_ARG2(a, b)
+#define CMPEQSS(a, b)		cmpeqss P_ARG2(a, b)
+#define CMPLTSS(a, b)		cmpltss P_ARG2(a, b)
+#define CMPLESS(a, b)		cmpless P_ARG2(a, b)
 #define CMPUNORDSS(a, b)	cmpunordss P_ARG2(a, b)
-#define CMPNEQSS(a, b)	cmpneqss P_ARG2(a, b)
-#define CMPNLTSS(a, b)	cmpnltss P_ARG2(a, b)
-#define CMPNLESS(a, b)	cmpnless P_ARG2(a, b)
-#define CMPORDSS(a, b)	cmpordss P_ARG2(a, b)
-#define COMISS(a, b)	comiss P_ARG2(a, b)
-#define CVTPI2PS(a, b)	cvtpi2ps P_ARG2(a, b)
-#define CVTPS2PI(a, b)	cvtps2pi P_ARG2(a, b)
-#define CVTSI2SS(a, b)	cvtsi2ss P_ARG2(a, b)
-#define CVTSS2SI(a, b)	cvtss2si P_ARG2(a, b)
-#define CVTTPS2PI(a, b)	cvttps2pi P_ARG2(a, b)
-#define CVTTSS2SI(a, b)	cvttss2si P_ARG2(a, b)
-#define DIVPS(a, b)	divps P_ARG2(a, b)
-#define DIVSS(a, b)	divss P_ARG2(a, b)
-#define FXRSTOR(a)	fxrstor P_ARG1(a)
-#define FXSAVE(a)	fxsave P_ARG1(a)
-#define LDMXCSR(a)	ldmxcsr P_ARG1(a)
-#define MAXPS(a, b)	maxps P_ARG2(a, b)
-#define MAXSS(a, b)	maxss P_ARG2(a, b)
-#define MINPS(a, b)	minps P_ARG2(a, b)
-#define MINSS(a, b)	minss P_ARG2(a, b)
-#define MOVAPS(a, b)	movaps P_ARG2(a, b)
-#define MOVHLPS(a, b)	movhlps P_ARG2(a, b)
-#define MOVHPS(a, b)	movhps P_ARG2(a, b)
-#define MOVLHPS(a, b)	movlhps P_ARG2(a, b)
-#define MOVLPS(a, b)	movlps P_ARG2(a, b)
-#define MOVMSKPS(a, b)	movmskps P_ARG2(a, b)
-#define MOVSS(a, b)	movss P_ARG2(a, b)
-#define MOVUPS(a, b)	movups P_ARG2(a, b)
-#define MULPS(a, b)	mulps P_ARG2(a, b)
-#define MULSS(a, b)	mulss P_ARG2(a, b)
-#define ORPS(a, b)	orps P_ARG2(a, b)
-#define RCPPS(a, b)	rcpps P_ARG2(a, b)
-#define RCPSS(a, b)	rcpss P_ARG2(a, b)
-#define RSQRTPS(a, b)	rsqrtps P_ARG2(a, b)
-#define RSQRTSS(a, b)	rsqrtss P_ARG2(a, b)
-#define SHUFPS(a, b, c)	shufps P_ARG3(a, b, c)
-#define SQRTPS(a, b)	sqrtps P_ARG2(a, b)
-#define SQRTSS(a, b)	sqrtss P_ARG2(a, b)
-#define STMXCSR(a)	stmxcsr P_ARG1(a)
-#define SUBPS(a, b)	subps P_ARG2(a, b)
-#define UCOMISS(a, b)	ucomiss P_ARG2(a, b)
-#define UNPCKHPS(a, b)	unpckhps P_ARG2(a, b)
-#define UNPCKLPS(a, b)	unpcklps P_ARG2(a, b)
-#define XORPS(a, b)	xorps P_ARG2(a, b)
-
+#define CMPNEQSS(a, b)		cmpneqss P_ARG2(a, b)
+#define CMPNLTSS(a, b)		cmpnltss P_ARG2(a, b)
+#define CMPNLESS(a, b)		cmpnless P_ARG2(a, b)
+#define CMPORDSS(a, b)		cmpordss P_ARG2(a, b)
+#define COMISS(a, b)		comiss P_ARG2(a, b)
+#define CVTPI2PS(a, b)		cvtpi2ps P_ARG2(a, b)
+#define CVTPS2PI(a, b)		cvtps2pi P_ARG2(a, b)
+#define CVTSI2SS(a, b)		cvtsi2ss P_ARG2(a, b)
+#define CVTSS2SI(a, b)		cvtss2si P_ARG2(a, b)
+#define CVTTPS2PI(a, b)		cvttps2pi P_ARG2(a, b)
+#define CVTTSS2SI(a, b)		cvttss2si P_ARG2(a, b)
+#define DIVPS(a, b)		divps P_ARG2(a, b)
+#define DIVSS(a, b)		divss P_ARG2(a, b)
+#define FXRSTOR(a)		fxrstor P_ARG1(a)
+#define FXSAVE(a)		fxsave P_ARG1(a)
+#define LDMXCSR(a)		ldmxcsr P_ARG1(a)
+#define MAXPS(a, b)		maxps P_ARG2(a, b)
+#define MAXSS(a, b)		maxss P_ARG2(a, b)
+#define MINPS(a, b)		minps P_ARG2(a, b)
+#define MINSS(a, b)		minss P_ARG2(a, b)
+#define MOVAPS(a, b)		movaps P_ARG2(a, b)
+#define MOVHLPS(a, b)		movhlps P_ARG2(a, b)
+#define MOVHPS(a, b)		movhps P_ARG2(a, b)
+#define MOVLHPS(a, b)		movlhps P_ARG2(a, b)
+#define MOVLPS(a, b)		movlps P_ARG2(a, b)
+#define MOVMSKPS(a, b)		movmskps P_ARG2(a, b)
+#define MOVNTPS(a, b)		movntps P_ARG2(a, b)
+#define MOVNTQ(a, b)		movntq P_ARG2(a, b)
+#define MOVSS(a, b)		movss P_ARG2(a, b)
+#define MOVUPS(a, b)		movups P_ARG2(a, b)
+#define MULPS(a, b)		mulps P_ARG2(a, b)
+#define MULSS(a, b)		mulss P_ARG2(a, b)
+#define ORPS(a, b)		orps P_ARG2(a, b)
+#define RCPPS(a, b)		rcpps P_ARG2(a, b)
+#define RCPSS(a, b)		rcpss P_ARG2(a, b)
+#define RSQRTPS(a, b)		rsqrtps P_ARG2(a, b)
+#define RSQRTSS(a, b)		rsqrtss P_ARG2(a, b)
+#define SHUFPS(a, b, c)		shufps P_ARG3(a, b, c)
+#define SQRTPS(a, b)		sqrtps P_ARG2(a, b)
+#define SQRTSS(a, b)		sqrtss P_ARG2(a, b)
+#define STMXCSR(a)		stmxcsr P_ARG1(a)
+#define SUBPS(a, b)		subps P_ARG2(a, b)
+#define UCOMISS(a, b)		ucomiss P_ARG2(a, b)
+#define UNPCKHPS(a, b)		unpckhps P_ARG2(a, b)
+#define UNPCKLPS(a, b)		unpcklps P_ARG2(a, b)
+#define XORPS(a, b)		xorps P_ARG2(a, b)
+
+#define PREFETCHNTA(a)		prefetchnta P_ARG1(a)
+#define PREFETCHT0(a)		prefetcht0 P_ARG1(a)
+#define PREFETCHT1(a)		prefetcht1 P_ARG1(a)
+#define PREFETCHT2(a)		prefetcht2 P_ARG1(a)
+#define SFENCE			sfence
 
 /* Added by BrianP for FreeBSD (per David Dawes) */
 #if !defined(NASM_ASSEMBLER) && !defined(MASM_ASSEMBLER) && !defined(__bsdi__)
-#define LLBL(a) CONCAT(.L,a)
+#define LLBL(a)		CONCAT(.L,a)
 #else
-#define LLBL(a) a
+#define LLBL(a)		a
 #endif
 
 
diff --git a/xc/extras/Mesa/src/X86/clip_args.h b/xc/extras/Mesa/src/X86/clip_args.h
new file mode 100644
index 000000000..1196e1499
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/clip_args.h
@@ -0,0 +1,75 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * Clip test function interface for assembly code.  Simply define
+ * FRAME_OFFSET to the number of bytes pushed onto the stack before
+ * using the ARG_* argument macros.
+ *
+ * Gareth Hughes <gareth@valinux.com>
+ */
+
+#ifndef __CLIP_ARGS_H__
+#define __CLIP_ARGS_H__
+
+/* Offsets into GLvector4f
+ */
+#define V4F_DATA	0
+#define V4F_START	4
+#define V4F_COUNT	8
+#define V4F_STRIDE	12
+#define V4F_SIZE	16
+#define V4F_FLAGS	20
+
+/* GLvector4f flags
+ */
+#define VEC_SIZE_1	1
+#define VEC_SIZE_2	3
+#define VEC_SIZE_3	7
+#define VEC_SIZE_4	15
+
+/*
+ * Offsets for clip_func arguments
+ *
+ * typedef GLvector4f *(*clip_func)( GLvector4f *vClip,
+ *	                             GLvector4f *vProj,
+ *	                             GLubyte clipMask[],
+ *	                             GLubyte *orMask,
+ *	                             GLubyte *andMask );
+ */
+
+#define OFFSET_SOURCE	4
+#define OFFSET_DEST	8
+#define OFFSET_CLIP	12
+#define OFFSET_OR	16
+#define OFFSET_AND	20
+
+#define ARG_SOURCE	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
+#define ARG_DEST	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
+#define ARG_CLIP	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
+#define ARG_OR		REGOFF(FRAME_OFFSET+OFFSET_OR, ESP)
+#define ARG_AND		REGOFF(FRAME_OFFSET+OFFSET_AND, ESP)
+
+#endif
diff --git a/xc/extras/Mesa/src/X86/common_x86.c b/xc/extras/Mesa/src/X86/common_x86.c
index fdf1b276c..68b7403c2 100644
--- a/xc/extras/Mesa/src/X86/common_x86.c
+++ b/xc/extras/Mesa/src/X86/common_x86.c
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -25,81 +25,102 @@
 
 
 /*
- *  Check CPU capabilities & initialize optimized funtions for this particular
- *   processor.
+ * Check CPU capabilities & initialize optimized funtions for this particular
+ * processor.
  *
- *  Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
- *  Changed by Andre Werthmann <wertmann@cs.uni-potsdam.de> for using the
- *  new Katmai functions
+ * Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
+ * Changed by Andre Werthmann <wertmann@cs.uni-potsdam.de> for using the
+ * new Katmai functions.
  */
 
 #include <stdlib.h>
 #include <stdio.h>
-#include "common_x86asm.h"
+
+#include "common_x86_asm.h"
+
 
 int gl_x86_cpu_features = 0;
 
-static void message(const char *msg)
+/* No reason for this to be public.
+ */
+extern int gl_identify_x86_cpu_features( void );
+
+
+static void message( const char *msg )
 {
-   if (getenv("MESA_DEBUG"))
-      fprintf(stderr, "%s\n", msg);
+   if ( getenv( "MESA_DEBUG" ) ) {
+      fprintf( stderr, "%s\n", msg );
+   }
 }
 
 
-void gl_init_all_x86_asm (void)
+void gl_init_all_x86_transform_asm( void )
 {
 #ifdef USE_X86_ASM
-   gl_x86_cpu_features = gl_identify_x86_cpu_features ();
-   gl_x86_cpu_features |= GL_CPU_AnyX86;
+   gl_x86_cpu_features = gl_identify_x86_cpu_features();
 
-   if (getenv("MESA_NO_ASM") != 0)
+   if ( getenv( "MESA_NO_ASM" ) ) {
       gl_x86_cpu_features = 0;
-
-   if (gl_x86_cpu_features & GL_CPU_GenuineIntel) {
-      message("GenuineIntel cpu detected.");
    }
 
-   if (gl_x86_cpu_features) {
-      gl_init_x86_asm_transforms ();
+   if ( gl_x86_cpu_features ) {
+      gl_init_x86_transform_asm();
    }
 
 #ifdef USE_MMX_ASM
-   if (gl_x86_cpu_features & GL_CPU_MMX) {
-      char *s = getenv( "MESA_NO_MMX" );
-      if (s == NULL) { 
-         message("MMX cpu detected.");
+   if ( cpu_has_mmx ) {
+      if ( getenv( "MESA_NO_MMX" ) == 0 ) {
+         message( "MMX cpu detected." );
       } else {
-         gl_x86_cpu_features &= (~GL_CPU_MMX); 
+         gl_x86_cpu_features &= ~(X86_FEATURE_MMX);
       }
    }
 #endif
 
-
 #ifdef USE_3DNOW_ASM
-   if (gl_x86_cpu_features & GL_CPU_3Dnow) {
-      char *s = getenv( "MESA_NO_3DNOW" );
-      if (s == NULL) {
-         message("3Dnow cpu detected.");
-         gl_init_3dnow_asm_transforms ();
+   if ( cpu_has_3dnow ) {
+      if ( getenv( "MESA_NO_3DNOW" ) == 0 ) {
+         message( "3Dnow cpu detected." );
+         gl_init_3dnow_transform_asm();
       } else {
-         gl_x86_cpu_features &= (~GL_CPU_3Dnow); 
+         gl_x86_cpu_features &= ~(X86_FEATURE_3DNOW);
       }
    }
 #endif
 
-
 #ifdef USE_KATMAI_ASM
-   if (gl_x86_cpu_features & GL_CPU_Katmai) {
-      char *s = getenv( "MESA_NO_KATMAI" );
-      if (s == NULL) {
-         message("Katmai cpu detected.");
-         gl_init_katmai_asm_transforms ();
+   if ( cpu_has_xmm ) {
+      if ( getenv( "MESA_NO_KATMAI" ) == 0 ) {
+         message( "Katmai cpu detected." );
+         gl_init_katmai_transform_asm();
       } else {
-         gl_x86_cpu_features &= (~GL_CPU_Katmai); 
+         gl_x86_cpu_features &= ~(X86_FEATURE_XMM);
       }
    }
 #endif
-
 #endif
 }
 
+/* Note: the above function must be called before this one, so that
+ * gl_x86_cpu_features gets correctly initialized.
+ */
+void gl_init_all_x86_vertex_asm( void )
+{
+#ifdef USE_X86_ASM
+   if ( gl_x86_cpu_features ) {
+      gl_init_x86_vertex_asm();
+   }
+
+#ifdef USE_3DNOW_ASM
+   if ( cpu_has_3dnow && getenv( "MESA_NO_3DNOW" ) == 0 ) {
+      gl_init_3dnow_vertex_asm();
+   }
+#endif
+
+#ifdef USE_KATMAI_ASM
+   if ( cpu_has_xmm && getenv( "MESA_NO_KATMAI" ) == 0 ) {
+      gl_init_katmai_vertex_asm();
+   }
+#endif
+#endif
+}
diff --git a/xc/extras/Mesa/src/X86/common_x86_asm.S b/xc/extras/Mesa/src/X86/common_x86_asm.S
new file mode 100644
index 000000000..fbde9e30d
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/common_x86_asm.S
@@ -0,0 +1,151 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * Check extended CPU capabilities.  Now justs returns the raw CPUID
+ * feature information, allowing the higher level code to interpret the
+ * results.
+ *
+ * Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
+ *
+ * Cleaned up and simplified by Gareth Hughes <gareth@valinux.com>
+ */
+
+#include "assyntax.h"
+#include "common_x86_features.h"
+
+
+/* Intel vendor string
+ */
+#define GENU	0x756e6547	/* "Genu" */
+#define INEI	0x49656e69	/* "ineI" */
+#define NTEL	0x6c65746e	/* "ntel" */
+
+/* AMD vendor string
+ */
+#define AUTH	0x68747541	/* "Auth" */
+#define ENTI	0x69746e65	/* "enti" */
+#define CAMD	0x444d4163	/* "cAMD" */
+
+
+	SEG_DATA
+
+/* We might want to print out some useful messages.
+ */
+LLBL( found_intel ):	STRING( "Genuine Intel processor found\n\0" )
+LLBL( found_amd ):	STRING( "Authentic AMD processor found\n\0" )
+
+
+	SEG_TEXT
+
+ALIGNTEXT4
+GLOBL GLNAME( gl_identify_x86_cpu_features )
+GLNAME( gl_identify_x86_cpu_features ):
+
+	PUSH_L	( EBX )
+
+	/* Test for the CPUID command.  If the ID Flag bit in EFLAGS
+	 * (bit 21) is writable, the CPUID command is present.
+	 */
+	PUSHF_L
+	POP_L	( EAX )
+	MOV_L	( EAX, ECX )
+	XOR_L	( CONST(0x00200000), EAX )
+	PUSH_L	( EAX )
+	POPF_L
+	PUSHF_L
+	POP_L	( EAX )
+
+	/* Verify the ID Flag bit has been written.
+	 */
+	CMP_L	( ECX, EAX )
+	JZ	( LLBL ( cpuid_done ) )
+
+	/* Get the CPU vendor info.
+	 */
+	XOR_L	( EAX, EAX )
+	CPUID
+
+	/* Test for Intel processors.  We must look for the
+	 * "GenuineIntel" string in EBX, ECX and EDX.
+	 */
+	CMP_L	( CONST(GENU), EBX )
+	JNE	( LLBL( cpuid_amd ) )
+	CMP_L	( CONST(INEI), EDX )
+	JNE	( LLBL( cpuid_amd ) )
+	CMP_L	( CONST(NTEL), ECX )
+	JNE	( LLBL( cpuid_amd ) )
+
+	/* We have an Intel processor, so we can get the feature
+	 * information with an CPUID input value of 1.
+	 */
+	MOV_L	( CONST(0x1), EAX )
+	CPUID
+	MOV_L	( EDX, EAX )
+	JMP	( LLBL( cpuid_done ) )
+
+LLBL( cpuid_amd ):
+
+	/* Test for AMD processors.  We must look for the
+	 * "AuthenticAMD" string in EBX, ECX and EDX.
+	 */
+	CMP_L	( CONST(AUTH), EBX )
+	JNE	( LLBL( cpuid_other ) )
+	CMP_L	( CONST(ENTI), EDX )
+	JNE	( LLBL( cpuid_other ) )
+	CMP_L	( CONST(CAMD), ECX )
+	JNE	( LLBL( cpuid_other ) )
+
+	/* We have an AMD processor, so we can get the feature
+	 * information after we verify that the extended functions are
+	 * supported.
+	 */
+	MOV_L	( CONST(0x80000000), EAX )
+	CPUID
+	TEST_L	( EAX, EAX )
+	JZ	( LLBL ( cpuid_failed ) )
+
+	MOV_L	( CONST(0x80000001), EAX )
+	CPUID
+	MOV_L	( EDX, EAX )
+	JMP	( LLBL ( cpuid_done ) )
+
+LLBL( cpuid_other ):
+
+	/* Test for other processors here when required.
+	 */
+
+LLBL( cpuid_failed ):
+
+	/* If we can't determine the feature information, we must
+	 * return zero to indicate that no platform-specific
+	 * optimizations can be used.
+	 */
+	MOV_L	( CONST(0), EAX )
+
+LLBL ( cpuid_done ):
+
+	POP_L	( EBX )
+	RET
diff --git a/xc/extras/Mesa/src/X86/common_x86asm.h b/xc/extras/Mesa/src/X86/common_x86_asm.h
index a9d901f62..f781d503e 100644
--- a/xc/extras/Mesa/src/X86/common_x86asm.h
+++ b/xc/extras/Mesa/src/X86/common_x86_asm.h
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -23,23 +23,22 @@
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-
 /*
- *  Check CPU capabilities & initialize optimized funtions for this particular
- *   processor.
+ * Check CPU capabilities & initialize optimized funtions for this particular
+ * processor.
+ *
+ * Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
+ * Changed by Andre Werthmann <wertmann@cs.uni-potsdam.de> for using the
+ * new Katmai functions
  *
- *  Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
- *  Changed by Andre Werthmann <wertmann@cs.uni-potsdam.de> for using the
- *  new Katmai functions
+ * Reimplemented by Gareth Hughes <gareth@valinux.com> in a more
+ * future-proof manner, based on code in the Linux kernel.
  */
-#ifndef _common_3dnow_h
-#define _common_3dnow_h
 
-#define  GL_CPU_GenuineIntel     1
-#define  GL_CPU_MMX              2
-#define  GL_CPU_3Dnow            4
-#define  GL_CPU_Katmai           8 /* set if the katmai-instructions are available */
-#define  GL_CPU_AnyX86           16 /* set if x86 asm allowed */
+#ifndef __COMMON_X86_ASM_H__
+#define __COMMON_X86_ASM_H__
+
+#include "common_x86_features.h"
 
 #ifdef HAVE_CONFIG_H
 #include "conf.h"
@@ -56,7 +55,8 @@
 #endif
 
 extern int gl_x86_cpu_features;
-extern void gl_init_all_x86_asm (void);
-extern int gl_identify_x86_cpu_features(void);
+
+extern void gl_init_all_x86_transform_asm( void );
+extern void gl_init_all_x86_vertex_asm( void );
 
 #endif
diff --git a/xc/extras/Mesa/src/X86/common_x86_features.h b/xc/extras/Mesa/src/X86/common_x86_features.h
new file mode 100644
index 000000000..b8e373d89
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/common_x86_features.h
@@ -0,0 +1,76 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * x86 CPUID feature information.  The raw data is returned by
+ * gl_identify_x86_cpu_features() and interpreted with the cpu_has_*
+ * helper macros.
+ *
+ * Gareth Hughes <gareth@valinux.com>
+ */
+
+#ifndef __COMMON_X86_FEATURES_H__
+#define __COMMON_X86_FEATURES_H__
+
+/* Capabilities of CPUs
+ */
+#define X86_FEATURE_FPU		0x00000001
+#define X86_FEATURE_VME		0x00000002
+#define X86_FEATURE_DE		0x00000004
+#define X86_FEATURE_PSE		0x00000008
+#define X86_FEATURE_TSC		0x00000010
+#define X86_FEATURE_MSR		0x00000020
+#define X86_FEATURE_PAE		0x00000040
+#define X86_FEATURE_MCE		0x00000080
+#define X86_FEATURE_CX8		0x00000100
+#define X86_FEATURE_APIC	0x00000200
+#define X86_FEATURE_10		0x00000400
+#define X86_FEATURE_SEP		0x00000800
+#define X86_FEATURE_MTRR	0x00001000
+#define X86_FEATURE_PGE		0x00002000
+#define X86_FEATURE_MCA		0x00004000
+#define X86_FEATURE_CMOV	0x00008000
+#define X86_FEATURE_PAT		0x00010000
+#define X86_FEATURE_PSE36	0x00020000
+#define X86_FEATURE_18		0x00040000
+#define X86_FEATURE_19		0x00080000
+#define X86_FEATURE_20		0x00100000
+#define X86_FEATURE_21		0x00200000
+#define X86_FEATURE_MMXEXT	0x00400000
+#define X86_FEATURE_MMX		0x00800000
+#define X86_FEATURE_FXSR	0x01000000
+#define X86_FEATURE_XMM		0x02000000
+#define X86_FEATURE_26		0x04000000
+#define X86_FEATURE_27		0x08000000
+#define X86_FEATURE_28		0x10000000
+#define X86_FEATURE_29		0x20000000
+#define X86_FEATURE_3DNOWEXT	0x40000000
+#define X86_FEATURE_3DNOW	0x80000000
+
+#define cpu_has_mmx		(gl_x86_cpu_features & X86_FEATURE_MMX)
+#define cpu_has_xmm		(gl_x86_cpu_features & X86_FEATURE_XMM)
+#define cpu_has_3dnow		(gl_x86_cpu_features & X86_FEATURE_3DNOW)
+
+#endif
diff --git a/xc/extras/Mesa/src/X86/common_x86asm.S b/xc/extras/Mesa/src/X86/common_x86asm.S
deleted file mode 100644
index 26b4f4ebf..000000000
--- a/xc/extras/Mesa/src/X86/common_x86asm.S
+++ /dev/null
@@ -1,130 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  3.3
- *
- * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/*
- *  Check extended CPU capabilities.
- *
- *  Written by Holger Waechtler <holger@akaflieg.extern.tu-berlin.de>
- */
-
-
-
-
-#include "assyntax.h"
-
-        SEG_TEXT
-
-
-#define  GL_CPU_GenuineIntel     CONST(1)
-#define  GL_CPU_MMX              CONST(2)
-#define  GL_CPU_3Dnow            CONST(4)
-#define  GL_CPU_Katmai		 CONST(8)
-
-
-GLOBL GLNAME(gl_identify_x86_cpu_features)
-    ALIGNTEXT4
-GLNAME(gl_identify_x86_cpu_features):
-
-    PUSH_L  ( EBX )
-    PUSH_L  ( ESI )
-    XOR_L   ( ESI, ESI )
-
-    PUSHF_L                             /*  test for the cpuid command        */
-    POP_L   ( EAX )                     /*   (if the ID-Bit (Bit 21) in       */
-    MOV_L   ( EAX, EBX )                /*    EEFLAGS is writable, the        */
-    XOR_L   ( CONST(0x00200000), EAX )  /*    cpuid command is present)       */
-    PUSH_L  ( EAX )
-    POPF_L
-    PUSHF_L
-    POP_L   ( EAX )
-    CMP_L   ( EBX, EAX )
-    JZ      ( LLBL ( Gidcpu_end ) )   /*  cpuid command not supported       */
-
-    XOR_L   ( EAX, EAX )
-    CPUID
-
-
-LLBL (Gidcpu_intel):
-    CMP_L  ( CONST(0x756e6547), EBX )  /*  compare                           */
-    JNE    ( LLBL ( Gidcpu_amd ) )     /*  vendor string == 'GenuineIntel' ? */
-    CMP_L  ( CONST(0x49656e69), EDX )
-    JNE    ( LLBL ( Gidcpu_amd ) )
-    CMP_L  ( CONST(0x6c65746e), ECX )
-    JNE    ( LLBL ( Gidcpu_amd ) )
-    OR_L   ( GL_CPU_GenuineIntel, ESI )
-    JMP    ( LLBL ( Gidcpu_katmai ) )
-
-
-LLBL ( Gidcpu_amd ):
-    CMP_L  ( CONST(0x68747541), EBX )  /*  compare vendor string == 'Auth'?  */
-    JNE    ( LLBL ( Gidcpu_other ) )
-    CMP_L  ( CONST(0x69746e65), EDX )  /*  'enti'                            */
-    JNE    ( LLBL ( Gidcpu_other ) )
-    CMP_L  ( CONST(0x444d4163), ECX )  /*  'cAMD'   (= 'AuthenticAMD')       */
-    JE     ( LLBL ( Gidcpu_3dnow ) )
-
-
-LLBL ( Gidcpu_other ):
-
-/*   insert here other vendorstrings with 3dnow capable cpu vendors          */
-
-JMP    ( LLBL ( Gidcpu_mmx ) )
-
-LLBL ( Gidcpu_katmai ):
-    MOV_L   ( CONST(0x1), EAX )
-    CPUID				/* test for Katmai */
-    TEST_L  ( CONST(0x002000000), EDX ) /* is Bit 25 set ? */
-    JZ	    ( LLBL ( Gidcpu_mmx ) )	/* streaming SIMD-extensions not found... */
-    OR_L    ( GL_CPU_Katmai, ESI )	/* Katmai detected ! :) */
-    JMP     ( LLBL ( Gidcpu_mmx ) )	/* check for mmx */
-
-LLBL ( Gidcpu_3dnow ): 
-    MOV_L   ( CONST(0x80000000), EAX )  /*  extended functions supported ?    */
-
-    CPUID
-    TEST_L  ( EAX, EAX )
-    JZ      ( LLBL ( Gidcpu_end ) )
-
-    MOV_L   ( CONST(0x80000001), EAX )
-    CPUID                              /*  test for 3Dnow!                   */
-    TEST_L  ( CONST(0x80000000), EDX )
-    JZ      ( LLBL ( Gidcpu_mmx ) )
-    OR_L    ( GL_CPU_3Dnow, ESI )      /*  (detected 3Dnow! extension)       */
-
-LLBL ( Gidcpu_mmx ):
-    MOV_L   ( CONST(0x1), EAX )
-    CPUID
-    TEST_L  ( CONST(0x00800000), EDX ) /*  Bit 23 set ?                      */
-    JZ      ( LLBL ( Gidcpu_end ) )
-    OR_L    ( GL_CPU_MMX, ESI )        /*  (detected mmx extension)          */
-
-
-LLBL ( Gidcpu_end ):
-    MOV_L   ( ESI, EAX )
-    STC
-    POP_L   ( ESI )
-    POP_L   ( EBX )
-
-    RET
-
diff --git a/xc/extras/Mesa/src/X86/katmai.c b/xc/extras/Mesa/src/X86/katmai.c
index 507a7208c..d0aa79e27 100644
--- a/xc/extras/Mesa/src/X86/katmai.c
+++ b/xc/extras/Mesa/src/X86/katmai.c
@@ -1,19 +1,20 @@
+
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -21,215 +22,199 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X86/katmai.c,v 1.4 2000/09/26 15:56:39 tsi Exp $ */
 
 /*
  * PentiumIII-SIMD (SSE) optimizations contributed by
  * Andre Werthmann <wertmann@cs.uni-potsdam.de>
  */
 
-#if defined(USE_KATMAI_ASM)
-#include "katmai.h"
-
-#include <limits.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <math.h>
-
+#include "glheader.h"
 #include "context.h"
 #include "types.h"
-#include "xform.h"
 #include "vertices.h"
+#include "xform.h"
+#include "katmai.h"
 
 #ifdef DEBUG
 #include "debug_xform.h"
 #endif
 
 
+#define XFORM_ARGS	GLvector4f *to_vec,				\
+			const GLfloat m[16],				\
+			const GLvector4f *from_vec,			\
+			const GLubyte *mask,				\
+			const GLubyte flag
+
+
+#define DECLARE_XFORM_GROUP( pfx, sz, masked ) \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_general_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_identity_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_3d_no_rot_##masked( XFORM_ARGS );    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_perspective_##masked( XFORM_ARGS );  \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_2d_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_2d_no_rot_##masked( XFORM_ARGS );    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_3d_##masked( XFORM_ARGS );
+
+
+#define ASSIGN_XFORM_GROUP( pfx, cma, sz, masked )			\
+   gl_transform_tab[cma][sz][MATRIX_GENERAL] =				\
+      gl_##pfx##_transform_points##sz##_general_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_IDENTITY] =				\
+      gl_##pfx##_transform_points##sz##_identity_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_3D_NO_ROT] =			\
+      gl_##pfx##_transform_points##sz##_3d_no_rot_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_PERSPECTIVE] =			\
+      gl_##pfx##_transform_points##sz##_perspective_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_2D] =				\
+      gl_##pfx##_transform_points##sz##_2d_##masked;			\
+   gl_transform_tab[cma][sz][MATRIX_2D_NO_ROT] =			\
+      gl_##pfx##_transform_points##sz##_2d_no_rot_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_3D] =				\
+      gl_##pfx##_transform_points##sz##_3d_##masked;
+
+
+
+#define NORM_ARGS	const GLmatrix *mat,				\
+			GLfloat scale,					\
+			const GLvector3f *in,				\
+			const GLfloat *lengths,				\
+			const GLubyte mask[],				\
+			GLvector3f *dest
+
+
+#define DECLARE_NORM_GROUP( pfx, masked ) \
+ extern void _ASMAPI gl_##pfx##_rescale_normals_##masked( NORM_ARGS );			    \
+ extern void _ASMAPI gl_##pfx##_normalize_normals_##masked( NORM_ARGS );		    \
+ extern void _ASMAPI gl_##pfx##_transform_normals_##masked( NORM_ARGS );		    \
+ extern void _ASMAPI gl_##pfx##_transform_normals_no_rot_##masked( NORM_ARGS );		    \
+ extern void _ASMAPI gl_##pfx##_transform_rescale_normals_##masked( NORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_rescale_normals_no_rot_##masked( NORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_normalize_normals_##masked( NORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_normalize_normals_no_rot_##masked( NORM_ARGS );
+
+
+#define ASSIGN_NORM_GROUP( pfx, cma, masked )				\
+   gl_normal_tab[NORM_RESCALE][cma] =					\
+      gl_##pfx##_rescale_normals_##masked;				\
+   gl_normal_tab[NORM_NORMALIZE][cma] =					\
+      gl_##pfx##_normalize_normals_##masked;				\
+   gl_normal_tab[NORM_TRANSFORM][cma] =					\
+      gl_##pfx##_transform_normals_##masked;				\
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT][cma] =				\
+      gl_##pfx##_transform_normals_no_rot_##masked;			\
+   gl_normal_tab[NORM_TRANSFORM | NORM_RESCALE][cma] =			\
+      gl_##pfx##_transform_rescale_normals_##masked;			\
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE][cma] =		\
+      gl_##pfx##_transform_rescale_normals_no_rot_##masked;		\
+   gl_normal_tab[NORM_TRANSFORM | NORM_NORMALIZE][cma] =		\
+      gl_##pfx##_transform_normalize_normals_##masked;			\
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE][cma] =		\
+      gl_##pfx##_transform_normalize_normals_no_rot_##masked;
+
+
+#ifdef USE_KATMAI_ASM
+DECLARE_XFORM_GROUP( katmai, 1, raw )
+DECLARE_XFORM_GROUP( katmai, 2, raw )
+DECLARE_XFORM_GROUP( katmai, 3, raw )
+
+DECLARE_XFORM_GROUP( katmai, 1, masked )
+DECLARE_XFORM_GROUP( katmai, 2, masked )
+DECLARE_XFORM_GROUP( katmai, 3, masked )
+
+#if 1
+/* Some functions are not written in SSE-assembly, because the fpu ones are faster */
+extern void gl_katmai_transform_normals_no_rot_raw(NORM_ARGS);
+extern void gl_katmai_transform_rescale_normals_raw(NORM_ARGS);
+extern void gl_katmai_transform_rescale_normals_no_rot_raw(NORM_ARGS);
+
+extern void gl_katmai_transform_points4_general_raw(XFORM_ARGS);
+extern void gl_katmai_transform_points4_general_masked(XFORM_ARGS);
+extern void gl_katmai_transform_points4_identity_masked(XFORM_ARGS);
+extern void gl_katmai_transform_points4_3d_no_rot_masked(XFORM_ARGS);
+extern void gl_katmai_transform_points4_3d_raw(XFORM_ARGS);
+extern void gl_katmai_transform_points4_3d_masked(XFORM_ARGS);
+#else
+DECLARE_NORM_GROUP( katmai, raw )
+#endif
 
 
-#define XFORM_ARGS      GLvector4f *to_vec,             \
-                        const GLfloat m[16],            \
-                        const GLvector4f *from_vec,     \
-                        const GLubyte *mask,            \
-                        const GLubyte flag
-
-
-
-#define DECLARE_XFORM_GROUP(pfx, v, masked) \
- extern void gl_##pfx##_transform_points##v##_general_##masked(XFORM_ARGS);    \
- extern void gl_##pfx##_transform_points##v##_identity_##masked(XFORM_ARGS);   \
- extern void gl_##pfx##_transform_points##v##_3d_no_rot_##masked(XFORM_ARGS);  \
- extern void gl_##pfx##_transform_points##v##_perspective_##masked(XFORM_ARGS);\
- extern void gl_##pfx##_transform_points##v##_2d_##masked(XFORM_ARGS);         \
- extern void gl_##pfx##_transform_points##v##_2d_no_rot_##masked(XFORM_ARGS);  \
- extern void gl_##pfx##_transform_points##v##_3d_##masked(XFORM_ARGS);
-
-
-
-#define ASSIGN_XFORM_GROUP( pfx, cma, vsize, masked )           \
- gl_transform_tab[cma][vsize][MATRIX_GENERAL]                   \
-  = gl_##pfx##_transform_points##vsize##_general_##masked;      \
- gl_transform_tab[cma][vsize][MATRIX_IDENTITY]                  \
-  = gl_##pfx##_transform_points##vsize##_identity_##masked;     \
- gl_transform_tab[cma][vsize][MATRIX_3D_NO_ROT]                 \
-  = gl_##pfx##_transform_points##vsize##_3d_no_rot_##masked;    \
- gl_transform_tab[cma][vsize][MATRIX_PERSPECTIVE]               \
-  = gl_##pfx##_transform_points##vsize##_perspective_##masked;  \
- gl_transform_tab[cma][vsize][MATRIX_2D]                        \
-  = gl_##pfx##_transform_points##vsize##_2d_##masked;           \
- gl_transform_tab[cma][vsize][MATRIX_2D_NO_ROT]                 \
-  = gl_##pfx##_transform_points##vsize##_2d_no_rot_##masked;    \
- gl_transform_tab[cma][vsize][MATRIX_3D]                        \
-  = gl_##pfx##_transform_points##vsize##_3d_##masked;
-
-
-
-
-#define NORM_ARGS       const GLmatrix *mat,        \
-                        GLfloat scale,              \
-                        const GLvector3f *in,       \
-                        const GLfloat *lengths,     \
-                        const GLubyte mask[],       \
-                        GLvector3f *dest 
-
-
-
-#define DECLARE_NORM_GROUP(pfx, masked)                                        \
- extern void gl_##pfx##_rescale_normals_##masked## (NORM_ARGS);                \
- extern void gl_##pfx##_normalize_normals_##masked## (NORM_ARGS);              \
- extern void gl_##pfx##_transform_normals_##masked## (NORM_ARGS);              \
- extern void gl_##pfx##_transform_normals_no_rot_##masked## (NORM_ARGS);       \
- extern void gl_##pfx##_transform_rescale_normals_##masked## (NORM_ARGS);      \
- extern void gl_##pfx##_transform_rescale_normals_no_rot_##masked## (NORM_ARGS); \
- extern void gl_##pfx##_transform_normalize_normals_##masked## (NORM_ARGS);    \
- extern void gl_##pfx##_transform_normalize_normals_no_rot_##masked## (NORM_ARGS);
-
-
-
-#define ASSIGN_NORM_GROUP( pfx, cma, masked )                                 \
-   gl_normal_tab[NORM_RESCALE][cma]   =                                       \
-      gl_##pfx##_rescale_normals_##masked##;                                  \
-   gl_normal_tab[NORM_NORMALIZE][cma] =                                       \
-      gl_##pfx##_normalize_normals_##masked##;                                \
-   gl_normal_tab[NORM_TRANSFORM][cma] =                                       \
-      gl_##pfx##_transform_normals_##masked##;                                \
-   gl_normal_tab[NORM_TRANSFORM_NO_ROT][cma] =                                \
-      gl_##pfx##_transform_normals_no_rot_##masked##;                         \
-   gl_normal_tab[NORM_TRANSFORM | NORM_RESCALE][cma] =                        \
-      gl_##pfx##_transform_rescale_normals_##masked##;                        \
-   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE][cma] =                 \
-      gl_##pfx##_transform_rescale_normals_no_rot_##masked##;                 \
-   gl_normal_tab[NORM_TRANSFORM | NORM_NORMALIZE][cma] =                      \
-      gl_##pfx##_transform_normalize_normals_##masked##;                      \
-   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE][cma] =               \
-      gl_##pfx##_transform_normalize_normals_no_rot_##masked##;
-
-
-extern void gl_katmai_project_vertices( GLfloat *first,
-					GLfloat *last,
-					const GLfloat *m,
-					GLuint stride );
-
-extern void gl_katmai_project_clipped_vertices( GLfloat *first,
+extern void _ASMAPI gl_v16_katmai_general_xform( GLfloat *first_vert,
+						 const GLfloat *m,
+						 const GLfloat *src,
+						 GLuint src_stride,
+						 GLuint count );
+extern void _ASMAPI gl_katmai_project_vertices( GLfloat *first,
 						GLfloat *last,
 						const GLfloat *m,
-						GLuint stride,
-						const GLubyte *clipmask );
-
-extern void gl_v16_katmai_general_xform( GLfloat *first_vert,
-					const GLfloat *m,
-					const GLfloat *src,
-					GLuint src_stride,
-					GLuint count );
+						GLuint stride );
 
+extern void _ASMAPI gl_katmai_project_clipped_vertices( GLfloat *first,
+							GLfloat *last,
+							const GLfloat *m,
+							GLuint stride,
+							const GLubyte *clipmask );
+#endif
 
 
-void gl_init_katmai_asm_transforms (void)
+void gl_init_katmai_transform_asm( void )
 {
-  extern void gl_katmai_transform_rescale_normals_raw(NORM_ARGS);
-  extern void gl_katmai_transform_rescale_normals_no_rot_raw(NORM_ARGS);
-  extern void gl_katmai_transform_normals_no_rot_raw(NORM_ARGS);
-  
-  /* Some functions are not written in SSE-assembly, because the fpu ones are faster */
-  extern void gl_katmai_transform_points4_general_raw(XFORM_ARGS);
-  extern void gl_katmai_transform_points4_general_masked(XFORM_ARGS);
-  extern void gl_katmai_transform_points4_identity_masked(XFORM_ARGS);
-  extern void gl_katmai_transform_points4_3d_no_rot_masked(XFORM_ARGS);
-  extern void gl_katmai_transform_points4_3d_raw(XFORM_ARGS);
-  extern void gl_katmai_transform_points4_3d_masked(XFORM_ARGS);
-
+#ifdef USE_KATMAI_ASM
+   ASSIGN_XFORM_GROUP( katmai, 0, 1, raw );
+   ASSIGN_XFORM_GROUP( katmai, 0, 2, raw );
+   ASSIGN_XFORM_GROUP( katmai, 0, 3, raw );
 
-  DECLARE_XFORM_GROUP( katmai, 1, raw )
-  DECLARE_XFORM_GROUP( katmai, 1, masked )
+   ASSIGN_XFORM_GROUP( katmai, CULL_MASK_ACTIVE, 1, masked );
+   ASSIGN_XFORM_GROUP( katmai, CULL_MASK_ACTIVE, 2, masked );
+   ASSIGN_XFORM_GROUP( katmai, CULL_MASK_ACTIVE, 3, masked );
 
-  DECLARE_XFORM_GROUP( katmai, 2, raw )
-  DECLARE_XFORM_GROUP( katmai, 2, masked )
-
-  DECLARE_XFORM_GROUP( katmai, 3, raw )
-  DECLARE_XFORM_GROUP( katmai, 3, masked )
-
-
-  gl_normal_tab[NORM_TRANSFORM | NORM_RESCALE][0]=gl_katmai_transform_rescale_normals_raw;
-  gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE][0]=gl_katmai_transform_rescale_normals_no_rot_raw; 
-  gl_normal_tab[NORM_TRANSFORM_NO_ROT][0]=gl_katmai_transform_normals_no_rot_raw;
-
-
-  gl_transform_tab[0][4][MATRIX_GENERAL]=gl_katmai_transform_points4_general_raw;
-  gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_GENERAL]=gl_katmai_transform_points4_general_masked;
-
-  gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_IDENTITY]=gl_katmai_transform_points4_identity_masked;
-
-  gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_3D_NO_ROT]=gl_katmai_transform_points4_3d_no_rot_masked;
-
-  gl_transform_tab[0][4][MATRIX_3D]=gl_katmai_transform_points4_3d_raw;
-  gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_3D]=gl_katmai_transform_points4_3d_masked;
-
-
-  ASSIGN_XFORM_GROUP( katmai, 0, 1, raw )
-  ASSIGN_XFORM_GROUP( katmai, CULL_MASK_ACTIVE, 1, masked )
-
-  ASSIGN_XFORM_GROUP( katmai, 0, 2, raw )
-  ASSIGN_XFORM_GROUP( katmai, CULL_MASK_ACTIVE, 2, masked )
-
-  ASSIGN_XFORM_GROUP( katmai, 0, 3, raw )
-  ASSIGN_XFORM_GROUP( katmai, CULL_MASK_ACTIVE, 3, masked )
-
-
-  /* TODO ! (some parts of it) */
-
-  /*
-   DECLARE_NORM_GROUP( katmai, raw )
-
-   ASSIGN_NORM_GROUP( katmai, 0, raw )
-  */
+#if 1
+   /* TODO: Finish these off.
+    */
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT][0] =
+      gl_katmai_transform_normals_no_rot_raw;
+   gl_normal_tab[NORM_TRANSFORM | NORM_RESCALE][0] =
+      gl_katmai_transform_rescale_normals_raw;
+   gl_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE][0] =
+      gl_katmai_transform_rescale_normals_no_rot_raw;
+
+   gl_transform_tab[0][4][MATRIX_GENERAL] =
+      gl_katmai_transform_points4_general_raw;
+   gl_transform_tab[0][4][MATRIX_3D] =
+      gl_katmai_transform_points4_3d_raw;
+
+   gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_GENERAL] =
+      gl_katmai_transform_points4_general_masked;
+   gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_IDENTITY] =
+      gl_katmai_transform_points4_identity_masked;
+   gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_3D_NO_ROT] =
+      gl_katmai_transform_points4_3d_no_rot_masked;
+   gl_transform_tab[CULL_MASK_ACTIVE][4][MATRIX_3D] =
+      gl_katmai_transform_points4_3d_masked;
+#else
+   ASSIGN_NORM_GROUP( katmai, 0, raw );
+#endif
 
 #ifdef DEBUG
-  gl_test_all_transform_functions("Katmai!");
-  gl_test_all_normal_transform_functions("Katmai!");
+   gl_test_all_transform_functions( "SSE" );
+   gl_test_all_normal_transform_functions( "SSE" );
+#endif
 #endif
-   
-   /* done */
-
-   /* Hook in some stuff for vertices.c.
-    */
-   gl_xform_points3_v16_general = gl_v16_katmai_general_xform;
-   
-   /* test, if it works correctly ! */
-   gl_project_v16 = gl_katmai_project_vertices;
-   
-   /* test, if it works correctly ! */
-   gl_project_clipped_v16 = gl_katmai_project_clipped_vertices;
 }
 
-#else
-
-
-/* silence compiler warning */
-extern void _mesa_katmai_dummy_function(void);
-void _mesa_katmai_dummy_function(void)
+void gl_init_katmai_vertex_asm( void )
 {
-}
-
+#ifdef USE_KATMAI_ASM
+   gl_xform_points3_v16_general	= gl_v16_katmai_general_xform;
+#if 0
+   /* GH: These are broken.  I'm fixing them now.
+    */
+   gl_project_v16		= gl_katmai_project_vertices;
+   gl_project_clipped_v16	= gl_katmai_project_clipped_vertices;
+#endif
 
+#if 0
+   gl_test_all_vertex_functions( "SSE" );
+#endif
 #endif
+}
diff --git a/xc/extras/Mesa/src/X86/katmai.h b/xc/extras/Mesa/src/X86/katmai.h
index 40d769f4d..0deeb629b 100644
--- a/xc/extras/Mesa/src/X86/katmai.h
+++ b/xc/extras/Mesa/src/X86/katmai.h
@@ -1,19 +1,20 @@
+
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -22,22 +23,17 @@
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-
 /*
  * PentiumIII-SIMD (SSE) optimizations contributed by
  * Andre Werthmann <wertmann@cs.uni-potsdam.de>
  */
 
-
 #ifndef __KATMAI_H__
 #define __KATMAI_H__
 
-
-
 #include "xform.h"
 
+void gl_init_katmai_transform_asm( void );
+void gl_init_katmai_vertex_asm( void );
 
-void gl_init_katmai_asm_transforms (void);
-
-
-#endif   /* __KATMAI_H__ */
+#endif
diff --git a/xc/extras/Mesa/src/X86/katmai_norm_raw.S b/xc/extras/Mesa/src/X86/katmai_norm_raw.S
index fdd720de7..02fafd8aa 100644
--- a/xc/extras/Mesa/src/X86/katmai_norm_raw.S
+++ b/xc/extras/Mesa/src/X86/katmai_norm_raw.S
@@ -3,9 +3,9 @@
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -28,15 +28,15 @@
 #define V3F_COUNT       8
 #define V3F_STRIDE      12
 
-#define MAT_INV         64
+#define MAT_INV         4
 
 #define M(i)    REGOFF(i * 4, EDX)
 #define S(i)	REGOFF(i * 4, ESI)
 #define D(i)	REGOFF(i * 4, EDI)
 #define STRIDE  REGOFF(12, ESI)
-	
+
 #undef FRAME_OFFSET
-		
+
 /*
  *	Offsets for norm_func arguments
  *
@@ -107,7 +107,7 @@ LLBL(K_G3TRNNRR_finish):
 	RET
 #undef FRAME_OFFSET
 
-	
+
 
 ALIGNTEXT16
 GLOBL GLNAME(gl_katmai_transform_rescale_normals_raw)
@@ -136,7 +136,7 @@ GLNAME(gl_katmai_transform_rescale_normals_raw):
 
 	MOV_L( REGOFF(V3F_START, EDI), EDI )		/* ptr to first dest vertex */
 	ADD_L( EDI, ECX ) 				/* count += dest ptr */
-	
+
 ALIGNTEXT32
 	MOVSS	( M(0), XMM0 )				/* m0 */
 	MOVSS	( M(4), XMM1 )				/* m4 */
@@ -144,7 +144,7 @@ ALIGNTEXT32
 
 	MOVSS	( ARG_SCALE, XMM4 )			/* scale */
 	SHUFPS	( CONST(0x0), XMM4, XMM4 )		/* scale | scale */
-	
+
 	MULPS	( XMM4, XMM0 )				/* m4*scale | m0*scale */
 	MOVSS	( M(1), XMM1 )				/* m1 */
 	MOVSS	( M(5), XMM2 )				/* m5 */
@@ -159,7 +159,7 @@ ALIGNTEXT32
 	MULSS	( ARG_SCALE, XMM6 )			/* m8*scale */
 	MOVSS	( M(9), XMM7 )				/* m9 */
 	MULSS	( ARG_SCALE, XMM7 )			/* m9*scale */
-	
+
 ALIGNTEXT32
 LLBL(K_G3TRNR_top):
 	MOVSS	( S(0), XMM3 )				/* ux */
@@ -187,7 +187,7 @@ LLBL(K_G3TRNR_top):
 	ADDSS	( XMM4, XMM3 )
 	ADDSS	( XMM5, XMM3 )
 	MOVSS	( XMM3, D(2) )
-	
+
 LLBL(K_G3TRNR_skip):
 	ADD_L	( CONST(12), EDI )
 	ADD_L	( EAX, ESI )
@@ -234,7 +234,7 @@ ALIGNTEXT32
 	MOVSS( M(5), XMM1 )				/* m5 */
 	UNPCKLPS( XMM1, XMM0 )				/* m5 | m0 */
 	MOVSS( M(10), XMM1 )				/* m10 */
-	
+
 ALIGNTEXT32
 LLBL(K_G3TNNRR_top):
 	MOVLPS( S(0), XMM2 )				/* uy | ux */
@@ -244,7 +244,7 @@ LLBL(K_G3TNNRR_top):
 	MOVSS( S(2), XMM2 )				/* uz */
 	MULSS( XMM1, XMM2 )				/* uz*m10 */
 	MOVSS( XMM2, D(2) )
-	
+
 LLBL(K_G3TNNRR_skip):
 	ADD_L	( CONST(12), EDI )
 	ADD_L	( EAX, ESI )
@@ -256,4 +256,3 @@ LLBL(K_G3TNNRR_finish):
 	POP_L	( ESI )
 	RET
 #undef FRAME_OFFSET
-
diff --git a/xc/extras/Mesa/src/X86/vertex_katmai.S b/xc/extras/Mesa/src/X86/katmai_vertex.S
index a8b08902a..90fe52fae 100644
--- a/xc/extras/Mesa/src/X86/vertex_katmai.S
+++ b/xc/extras/Mesa/src/X86/katmai_vertex.S
@@ -1,17 +1,16 @@
 #include "assyntax.h"
 
+	SEG_TEXT
+
 #define MAT_SY 20
 #define MAT_SZ 40
 #define MAT_TX 48
 #define MAT_TY 52
 #define MAT_TZ 56
-	
-	
-	SEG_TEXT
 
 
 /*
- * void gl_v16_katmai_general_xform (GLfloat *dest,
+ * void gl_v16_katmai_general_xform( GLfloat *dest,
  *		        	     const GLfloat *m,
  *				     const GLfloat *src,
  *                                   GLuint src_stride,
@@ -23,51 +22,51 @@ ALIGNTEXT16
 GLOBL GLNAME( gl_v16_katmai_general_xform )
 GLNAME( gl_v16_katmai_general_xform ):
 
-	PUSH_L( EDI )
- 	PUSH_L( ESI )
+    PUSH_L( EDI )
+    PUSH_L( ESI )
 
- 	MOV_L( REGOFF(12, ESP), EAX )	/* f - destination */
- 	MOV_L( REGOFF(16, ESP), ESI )	/* m - matrix */
- 	MOV_L( REGOFF(20, ESP), EDX )	/* obj - source */
- 	MOV_L( REGOFF(24, ESP), EDI )	/* obj_stride */
- 	MOV_L( REGOFF(28, ESP), ECX )	/* count */
+    MOV_L( REGOFF(12, ESP), EAX )		/* destination                       */
+    MOV_L( REGOFF(16, ESP), ESI )		/* matrix                            */
+    MOV_L( REGOFF(20, ESP), EDX )		/* source                            */
+    MOV_L( REGOFF(24, ESP), EDI )		/* src_stride                        */
+    MOV_L( REGOFF(28, ESP), ECX )		/* count                             */
 
- 	MOVUPS( REGOFF(0, ESI), XMM4 )	/* x14, x13, x12, x11 => xmm4 */
- 	MOVUPS( REGOFF(16, ESI), XMM5 )	/* x24, x23, x22, x21 => xmm5 */
- 	MOVUPS( REGOFF(32, ESI), XMM6 )	/* x34, x33, x32, x31 => xmm6 */
- 	MOVUPS( REGOFF(48, ESI), XMM7 )	/* x44, x43, x42, x41 => xmm7 */
+    MOVAPS( REGOFF(0, ESI), XMM4 )		/* m3     | m2     | m1     | m0     */
+    MOVAPS( REGOFF(16, ESI), XMM5 )		/* m7     | m6     | m5     | m4     */
+    MOVAPS( REGOFF(32, ESI), XMM6 )		/* m11    | m10    | m9     | m8     */
+    MOVAPS( REGOFF(48, ESI), XMM7 )		/* m15    | m14    | m13    | m12    */
 
 ALIGNTEXT32
-LLBL(loop1):	
+LLBL(v16_katmai_general_loop):
 
- 	MOVSS( REGOFF(0, EDX), XMM0 )	/* s1 => xmm0 */
- 	SHUFPS( CONST(0x0), XMM0, XMM0 )	/* xmm0 shuffle to other 96 bits */
- 	MULPS( XMM4, XMM0 )		/* x14*s1, x13*s1, x12*s1, x11*s1 */
+    MOVSS( REGOFF(0, EDX), XMM0 )		/*        |        |        | x0     */
+    SHUFPS( CONST(0x0), XMM0, XMM0 )		/* x0     | x0     | x0     | x0     */
+    MULPS( XMM4, XMM0 )				/* x0*m3  | x0*m2  | x0*m1  | x0*m0  */
 
- 	MOVSS( REGOFF(4, EDX), XMM1 )	/* s2 => xmm1 */
- 	SHUFPS( CONST(0x0), XMM1, XMM1 )	/* xmm1 shuffle to other 96 bits */
- 	MULPS( XMM5, XMM1 )		/* x24*s2, x23*s2, x22*s2, x21*s2 */
+    MOVSS( REGOFF(4, EDX), XMM1 )		/*        |        |        | x1     */
+    SHUFPS( CONST(0x0), XMM1, XMM1 )		/* x1     | x1     | x1     | x1     */
+    MULPS( XMM5, XMM1 )				/* x1*m7  | x1*m6  | x1*m5  | x1*m4  */
 
- 	MOVSS( REGOFF(8, EDX), XMM2 )	/* s3 => xmm2 */
- 	SHUFPS( CONST(0x0), XMM2, XMM2 )	/* xmm2 shuffle to other 96 bits */
- 	MULPS( XMM6, XMM2 )		/* x34*s3, x33*s3, x32*s3, x31*s3 */
+    MOVSS( REGOFF(8, EDX), XMM2 )		/*        |        |        | x2     */
+    SHUFPS( CONST(0x0), XMM2, XMM2 )		/* x2     | x2     | x2     | x2     */
+    MULPS( XMM6, XMM2 )				/* x2*m11 | x2*m10 | x2*m9  | x2*m8  */
 
- 	ADDPS( XMM1, XMM0 )		/* xmm0 + xmm1 => xmm0 */
- 	ADDPS( XMM2, XMM0 )		/* xmm0 + xmm2 => xmm0 */
- 	ADDPS( XMM7, XMM0 )		/* xmm0 + xmm7 => xmm0 */
+    ADDPS( XMM1, XMM0 )
+    ADDPS( XMM2, XMM0 )
+    ADDPS( XMM7, XMM0 )				/* r3     | r2     | r1     | r1     */
 
- 	MOVUPS( XMM0, REGOFF(0, EAX) ) 	/* xmm0 => dest( f11, f12, f13, f14 ) */
+    MOVAPS( XMM0, REGOFF(0, EAX) )
 
- 	ADD_L   ( CONST(64), EAX )	/* f + 64 bytes => f */
-        ADD_L   ( EDI, EDX )		/* obj + obj_stride => obj */
+    ADD_L( CONST(64), EAX )			/* next output vertex                */
+    ADD_L( EDI, EDX )				/* next input vertex                 */
 
-        DEC_L   ( ECX )			/* count - 1 => count */
-        JNE     ( LLBL(loop1) )		/* if not zero, go back to LABL(loop1) */
+    DEC_L( ECX )
+    JNE( LLBL(v16_katmai_general_loop) )
 
- 	POP_L( ESI )
- 	POP_L( EDI )
+    POP_L( ESI )
+    POP_L( EDI )
 
- 	RET
+    RET
 
 
 /* void gl_katmai_project_vertices(GLfloat *first,
@@ -80,7 +79,7 @@ LLBL(loop1):
 ALIGNTEXT16
 GLOBL GLNAME( gl_katmai_project_vertices )
 GLNAME( gl_katmai_project_vertices ):
-	
+
 	PUSH_L( EBP )
 
 	MOV_L( REGOFF(8, ESP), ECX )	/* first_vert */
@@ -88,9 +87,9 @@ GLNAME( gl_katmai_project_vertices ):
 	MOV_L( REGOFF(16, ESP), EBP )	/* matrix */
 	MOV_L( REGOFF(20, ESP), EAX )	/* stride */
 
-	
+
 ALIGNTEXT32
-	MOVUPS( REGOFF(MAT_TX, EBP), XMM0 )	/* (x44), x43, x42, x41 => xmm0 */
+	MOVAPS( REGOFF(MAT_TX, EBP), XMM0 )	/* (x44), x43, x42, x41 => xmm0 */
 	MOVSS( REGOFF(0, EBP), XMM1 )		/* -, -, -, x11 => xmm1 */
 	UNPCKLPS( REGOFF(MAT_SY, EBP), XMM1 )	/* -, -, x22, x11 => xmm1 */
 	SHUFPS( CONST(0x44), REGOFF(MAT_SZ, EBP), XMM1 )
@@ -99,7 +98,7 @@ ALIGNTEXT32
 	SUB_L( ECX, EDX )			/* last -= first */
 
 LLBL(v16_katmai_pv_loop_start):
-	MOVUPS( REGOFF(0, ECX), XMM3 )		/* f[3], f[2], f[1], f[0] */
+	MOVAPS( REGOFF(0, ECX), XMM3 )		/* f[3], f[2], f[1], f[0] */
 
 	MOVSS( REGOFF(12, ECX), XMM2 )		/* -, -, -, f[3] */
 	SHUFPS( CONST(0x0), XMM2, XMM2 )	/* f[3], f[3], f[3], f[3] */
@@ -109,7 +108,7 @@ LLBL(v16_katmai_pv_loop_start):
 	MULPS( XMM3, XMM1 )			/* -, x33*1/f[3]*f[2]... */
 	ADDPS( XMM0, XMM1 )			/* -, x33*1/f[3]*f[2]+x43... */
 
-	MOVUPS( XMM1, REGOFF(0, ECX) )		/* back to f */
+	MOVAPS( XMM1, REGOFF(0, ECX) )		/* back to f */
 	MOVSS( XMM2, REGOFF(12, ECX) )		/* 1/f[3] into f[3] ! */
 
 
@@ -133,7 +132,7 @@ LLBL(v16_katmai_pv_loop_start):
 ALIGNTEXT16
 GLOBL GLNAME( gl_katmai_project_clipped_vertices )
 GLNAME( gl_katmai_project_clipped_vertices ):
-	
+
 	PUSH_L( EBP )
 	PUSH_L( ESI )
 
@@ -142,12 +141,12 @@ GLNAME( gl_katmai_project_clipped_vertices ):
 	MOV_L( REGOFF(20, ESP), EBP )	/* matrix */
 	MOV_L( REGOFF(24, ESP), EAX )	/* stride */
 	MOV_L( REGOFF(28, ESP), ESI )	/* clip_mask */
-	
 
-	
+
+
 ALIGNTEXT32
 
-	MOVUPS( REGOFF(MAT_TX, EBP), XMM0 )	/* (x44), x43, x42, x41 => xmm0 */
+	MOVAPS( REGOFF(MAT_TX, EBP), XMM0 )	/* (x44), x43, x42, x41 => xmm0 */
 	MOVSS( REGOFF(0, EBP), XMM1 )		/* -, -, -, x11 => xmm1 */
 	UNPCKLPS( REGOFF(MAT_SY, EBP), XMM1 )	/* -, -, x22, x11 => xmm1 */
 	SHUFPS( CONST(0x44), REGOFF(MAT_SZ, EBP), XMM1 )
@@ -157,7 +156,7 @@ LLBL(v16_katmai_pcv_loop_start):
 	CMP_B ( CONST(0), REGIND(ESI) )		/* clip_mask == 0 ? */
 	JNE( LLBL(v16_katmai_pcv_skip) )	/* no -> skip ! */
 
-	MOVUPS( REGOFF(0, ECX), XMM3 )		/* f[3], f[2], f[1], f[0] */
+	MOVAPS( REGOFF(0, ECX), XMM3 )		/* f[3], f[2], f[1], f[0] */
 
 	MOVSS( REGOFF(12, ECX), XMM2 )		/* -, -, -, f[3] */
 	SHUFPS( CONST(0x0), XMM2, XMM2 )	/* f[3], f[3], f[3], f[3] */
@@ -167,7 +166,7 @@ LLBL(v16_katmai_pcv_loop_start):
 	MULPS( XMM3, XMM1 )			/* -, x33*1/f[3]*f[2]... */
 	ADDPS( XMM0, XMM1 )			/* -, x33*1/f[3]*f[2]+x43... */
 
-	MOVUPS( XMM1, REGOFF(0, ECX) )		/* back to f */
+	MOVAPS( XMM1, REGOFF(0, ECX) )		/* back to f */
 	MOVSS( XMM2, REGOFF(12, ECX) )		/* 1/f[3] into f[3] ! */
 
 LLBL(v16_katmai_pcv_skip):
@@ -178,9 +177,6 @@ LLBL(v16_katmai_pcv_skip):
 	JNE( LLBL(v16_katmai_pcv_loop_start) )
 						/* no -> go on with next vertex */
 
-	POP_L( ESI )	
+	POP_L( ESI )
 	POP_L( EBP )
 	RET
-
-
-
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_masked1.S b/xc/extras/Mesa/src/X86/katmai_xform_masked1.S
index 0408fcf81..4f5f62b80 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_masked1.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_masked1.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_masked1.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -68,7 +65,7 @@ GLNAME( gl_katmai_transform_points1_general_masked ):
 
     MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
-	
+
     MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
 
@@ -89,19 +86,19 @@ GLNAME( gl_katmai_transform_points1_general_masked ):
 
 
 ALIGNTEXT32
-    MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0  */
-    MOVUPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
+    MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0  */
+    MOVAPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
 
 ALIGNTEXT32
 LLBL(K_GTP1GM_top):
     TEST_B( BL, REGIND(EBP) )
     JNZ( LLBL(K_GTP1GM_skip) )
-	
+
     MOVSS( S(0), XMM2 )				/* ox */
     SHUFPS( CONST(0x0), XMM2, XMM2 )		/* ox | ox | ox | ox */
     MULPS( XMM0, XMM2 )				/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
     ADDPS( XMM1, XMM2 )				/* + | + | + | + */
-    MOVUPS( XMM2, D(0) )
+    MOVAPS( XMM2, D(0) )
 
 LLBL(K_GTP1GM_skip):
     INC_L     ( EBP )
@@ -195,7 +192,7 @@ GLNAME(gl_katmai_transform_points1_3d_no_rot_masked):
 
     MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
-	
+
     MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
 
@@ -220,12 +217,12 @@ ALIGNTEXT32
     MOVSS( M(12), XMM1 )			/* m12 */
     MOVSS( M(13), XMM2 )			/* m13 */
     MOVSS( M(14), XMM3 )			/* m14 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13DNRM_top):
     TEST_B( BL, REGIND(EBP) )
     JNZ( LLBL(K_GTP13DNRM_skip) )
-	
+
     MOVSS( S(0), XMM4 )				/* ox */
     MULSS( XMM0, XMM4 )				/* ox*m0 */
     ADDSS( XMM1, XMM4 )				/* ox*m0+m12 */
@@ -248,7 +245,7 @@ LLBL(K_GTP13DNRM_finish):
     POP_L    ( ESI )
     RET
 #undef FRAME_OFFSET
-    
+
 
 ALIGNTEXT4
 GLOBL GLNAME(gl_katmai_transform_points1_perspective_masked)
@@ -287,12 +284,12 @@ ALIGNTEXT32
     XORPS( XMM0, XMM0 )				/* 0 | 0 | 0 | 0 */
     MOVSS( M(0), XMM1 )				/* m0 */
     MOVSS( M(14), XMM2 )			/* m14 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13PM_top):
     TEST_B   ( BL, REGIND(EBP) )
     JNZ      ( LLBL(K_GTP13PM_skip) )
-	
+
     MOVSS( S(0), XMM3 )				/* ox */
     MULSS( XMM1, XMM3 )				/* ox*m0 */
     MOVSS( XMM3, D(0) )				/* ox*m0->D(0) */
@@ -300,7 +297,7 @@ LLBL(K_GTP13PM_top):
 
     MOVSS( XMM0, D(1) )
     MOVSS( XMM0, D(3) )
-	
+
 LLBL(K_GTP13PM_skip):
     INC_L( EBP )
     ADD_L( CONST(16), EDI )
@@ -326,7 +323,7 @@ GLNAME(gl_katmai_transform_points1_2d_masked):
     PUSH_L( EDI )
     PUSH_L( EBX )
     PUSH_L( EBP )
-	
+
     MOV_L( ARG_SOURCE, ESI )			/* ptr to source GLvector4f */
     MOV_L( ARG_DEST, EDI )			/* ptr to dest GLvector4f */
 
@@ -353,18 +350,18 @@ GLNAME(gl_katmai_transform_points1_2d_masked):
 ALIGNTEXT32
     MOVLPS( M(0), XMM0 )			/* m1  | m0  */
     MOVLPS( M(12), XMM1 )			/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13P2DM_top):
     TEST_B( BL, REGIND(EBP) )
     JNZ( LLBL(K_GTP13P2DM_skip) )
-	
+
     MOVSS( S(0), XMM2 )				/* ox */
     SHUFPS( CONST(0x0), XMM2, XMM2 )		/* ox | ox | ox | ox */
     MULPS( XMM0, XMM2 )				/* - | - | ox*m1 | ox*m0 */
     ADDPS( XMM1, XMM2 )				/* - | - | ox*m1+m13 | ox*m0+m12 */
     MOVLPS( XMM2, D(0) )
-	
+
 LLBL(K_GTP13P2DM_skip):
     INC_L    ( EBP )
     ADD_L    ( CONST(16), EDI )
@@ -396,7 +393,7 @@ GLNAME(gl_katmai_transform_points1_2d_no_rot_masked):
 
 	MOV_L( ARG_MATRIX, EDX )			/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP )				/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )		/* source count */
 	MOV_B( ARG_FLAG, BL )				/* clip mask flags */
 
@@ -419,18 +416,18 @@ ALIGNTEXT32
 	MOVSS( M(0), XMM0 )			/* m0 */
 	MOVSS( M(12), XMM1 )			/* m12 */
 	MOVSS( M(13), XMM2 )			/* m13 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13P2DNRM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP13P2DNRM_skip) )
-	
+
 	MOVSS( S(0), XMM3 )			/* ox */
 	MULSS( XMM0, XMM3 )			/* ox*m0 */
 	ADDSS( XMM1, XMM3 )			/* ox*m0+m12 */
 	MOVSS( XMM3, D(0) )
 	MOVSS( XMM2, D(1) )
-	
+
 LLBL(K_GTP13P2DNRM_skip):
 	INC_L( EBP )
 	ADD_L( CONST(16), EDI )
@@ -462,7 +459,7 @@ GLNAME(gl_katmai_transform_points1_3d_masked):
 
 	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
 	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
 
@@ -483,14 +480,14 @@ GLNAME(gl_katmai_transform_points1_3d_masked):
 
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
-	MOVUPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
-		
-ALIGNTEXT32	
+	MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
+	MOVAPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
+
+ALIGNTEXT32
 LLBL(K_GTP13P3DM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP13P3DM_skip) )
-	
+
 	MOVSS( S(0), XMM2 )			/* ox */
 	SHUFPS( CONST(0x0), XMM2, XMM2 )	/* ox | ox | ox | ox */
 	MULPS( XMM0, XMM2 )			/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
@@ -498,7 +495,7 @@ LLBL(K_GTP13P3DM_top):
 	MOVLPS( XMM2, D(0) )			/*   -   |   -   | ->D(1)| ->D(0)*/
 	UNPCKHPS( XMM2, XMM2 )			/* ox*m3+m15 | ox*m3+m15 | ox*m2+m14 | ox*m2+m14 */
 	MOVSS( XMM2, D(2) )			/* ->D(2) */
-	
+
 LLBL(K_GTP13P3DM_skip):
 	INC_L( EBP )
 	ADD_L( CONST(16), EDI )
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_masked2.S b/xc/extras/Mesa/src/X86/katmai_xform_masked2.S
index bfaaa5400..2ede85f2b 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_masked2.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_masked2.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_masked2.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -68,13 +65,13 @@ GLNAME( gl_katmai_transform_points2_general_masked ):
 
     MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
-	
+
     MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
 
     TEST_L( ECX, ECX )
     JZ( LLBL(K_GTP2GM_finish) )			/* count was zero; go to finish */
-	
+
     MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
     OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
 
@@ -89,15 +86,15 @@ GLNAME( gl_katmai_transform_points2_general_masked ):
 
 
 ALIGNTEXT32
-    MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0 */
-    MOVUPS( M(4), XMM1 )			/* m7  | m6  | m5  | m4 */
-    MOVUPS( M(12), XMM2 )			/* m15 | m14 | m13 | m12 */
+    MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0 */
+    MOVAPS( M(4), XMM1 )			/* m7  | m6  | m5  | m4 */
+    MOVAPS( M(12), XMM2 )			/* m15 | m14 | m13 | m12 */
 
 ALIGNTEXT32
 LLBL(K_GTP2GM_top):
     TEST_B( BL, REGIND(EBP) )
     JNZ( LLBL(K_GTP2GM_skip) )
-	
+
     MOVSS( S(0), XMM3 )				/* ox */
     SHUFPS( CONST(0x0), XMM3, XMM3 )		/* ox | ox | ox | ox */
     MULPS( XMM0, XMM3 )				/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
@@ -107,8 +104,8 @@ LLBL(K_GTP2GM_top):
 
     ADDPS( XMM4, XMM3 )
     ADDPS( XMM2, XMM3 )
-    MOVUPS( XMM3, D(0) )	
-	
+    MOVAPS( XMM3, D(0) )
+
 LLBL(K_GTP2GM_skip):
     INC_L     ( EBP )
     ADD_L     ( CONST(16), EDI )
@@ -145,7 +142,7 @@ GLNAME( gl_katmai_transform_points2_identity_masked ):
 
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
-	
+
     MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
     OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
 
@@ -165,12 +162,12 @@ ALIGNTEXT32
 LLBL(K_GTP2IM_top):
     TEST_B( BL, REGIND(EBP) )
     JNZ( LLBL(K_GTP2IM_skip) )
-	
+
     MOV_L( S(0), EDX )
     MOV_L( EDX, D(0) )
     MOV_L( S(1), EDX )
     MOV_L( EDX, D(1) )
-	
+
 LLBL(K_GTP2IM_skip):
     INC_L     ( EBP )
     ADD_L     ( CONST(16), EDI )
@@ -202,7 +199,7 @@ GLNAME(gl_katmai_transform_points2_3d_no_rot_masked):
 
     MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
-	
+
     MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
 
@@ -256,7 +253,7 @@ LLBL(K_GTP23DNRM_finish):
     RET
 #undef FRAME_OFFSET
 
-    
+
 ALIGNTEXT4
 GLOBL GLNAME(gl_katmai_transform_points2_perspective_masked)
 GLNAME(gl_katmai_transform_points2_perspective_masked):
@@ -272,7 +269,7 @@ GLNAME(gl_katmai_transform_points2_perspective_masked):
 
     MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
-	
+
     MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
 
@@ -297,18 +294,18 @@ ALIGNTEXT32
     UNPCKLPS ( XMM2, XMM1 )			/* -  | -  | m5  | m0  */
     MOVSS    ( M(14), XMM3 )			/* m14 */
     XORPS    ( XMM0, XMM0 )			/* 0 | 0 | 0 | 0 */
-    
+
 ALIGNTEXT32
 LLBL(K_GTP23PM_top):
     TEST_B   ( BL, REGIND(EBP) )
     JNZ      ( LLBL(K_GTP23PM_skip) )
-	
+
     MOVLPS( S(0), XMM4 )			/* oy | ox */
     MULPS( XMM1, XMM4 )				/* oy*m5 | ox*m0 */
     MOVLPS( XMM4, D(0) )			/* ->D(1) | ->D(0) */
     MOVSS( XMM3, D(2) )				/* ->D(2) */
     MOVSS( XMM0, D(3) )				/* ->D(3) */
-	
+
 LLBL(K_GTP23PM_skip):
     INC_L( EBP )
     ADD_L( CONST(16), EDI )
@@ -341,7 +338,7 @@ GLNAME(gl_katmai_transform_points2_2d_masked):
 
     MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
     MOV_L( ARG_CLIP, EBP )			/* ptr to clip mask array */
-	
+
     MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
     MOV_B( ARG_FLAG, BL )			/* clip mask flags */
 
@@ -364,7 +361,7 @@ ALIGNTEXT32
     MOVLPS( M(0), XMM0 )			/* m1  | m0 */
     MOVLPS( M(4), XMM1 )			/* m5  | m4 */
     MOVLPS( M(12), XMM2 )			/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP23P2DM_top):
     TEST_B( BL, REGIND(EBP) )
@@ -381,7 +378,7 @@ LLBL(K_GTP23P2DM_top):
     ADDPS( XMM4, XMM3 )
     ADDPS( XMM2, XMM3 )
     MOVLPS( XMM3, D(0) )			/* ->D(1) | ->D(0) */
-	
+
 LLBL(K_GTP23P2DM_skip):
     INC_L    ( EBP )
     ADD_L    ( CONST(16), EDI )
@@ -414,7 +411,7 @@ GLNAME(gl_katmai_transform_points2_2d_no_rot_masked):
 
 	MOV_L( ARG_MATRIX, EDX )			/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP )				/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )		/* source count */
 	MOV_B( ARG_FLAG, BL )				/* clip mask flags */
 
@@ -438,7 +435,7 @@ ALIGNTEXT32
 	MOVSS    ( M(5), XMM2 )				/* m5 */
 	UNPCKLPS ( XMM2, XMM1 )				/* m5 | m0 */
 	MOVLPS   ( M(12), XMM2 )			/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP23P2DNRM_top):
 	TEST_B( BL, REGIND(EBP) )
@@ -448,7 +445,7 @@ LLBL(K_GTP23P2DNRM_top):
 	MULPS( XMM1, XMM0 )				/* oy*m5 | ox*m0 */
 	ADDPS( XMM2, XMM0 )				/* +m13 | +m12 */
 	MOVLPS( XMM0, D(0) )				/* ->D(1) | ->D(0) */
-	
+
 LLBL(K_GTP23P2DNRM_skip):
 	INC_L( EBP )
 	ADD_L( CONST(16), EDI )
@@ -481,7 +478,7 @@ GLNAME(gl_katmai_transform_points2_3d_masked):
 
 	MOV_L( ARG_MATRIX, EDX )		/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
 	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
 
@@ -501,11 +498,11 @@ GLNAME(gl_katmai_transform_points2_3d_masked):
 	ADD_L( EDI, ECX ) 			/* count += dest ptr */
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m2  | m1  | m0 */
-	MOVUPS( M(4), XMM1 )			/* m6  | m5  | m4 */
-	MOVUPS( M(12), XMM2 )			/* m14 | m13 | m12 */
+	MOVAPS( M(0), XMM0 )			/* m2  | m1  | m0 */
+	MOVAPS( M(4), XMM1 )			/* m6  | m5  | m4 */
+	MOVAPS( M(12), XMM2 )			/* m14 | m13 | m12 */
 
-ALIGNTEXT32	
+ALIGNTEXT32
 LLBL(K_GTP23P3DM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP23P3DM_skip) )
@@ -524,7 +521,7 @@ LLBL(K_GTP23P3DM_top):
 	MOVLPS( XMM3, D(0) )			/* ->D(1) | ->D(0) */
 	UNPCKHPS( XMM3, XMM3 )
 	MOVSS( XMM3, D(2) )			/* ->D(2) */
-	
+
 LLBL(K_GTP23P3DM_skip):
 	INC_L( EBP )
 	ADD_L( CONST(16), EDI )
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_masked3.S b/xc/extras/Mesa/src/X86/katmai_xform_masked3.S
index cace27fc1..6bb5659b3 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_masked3.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_masked3.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_masked3.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -88,10 +85,10 @@ GLNAME( gl_katmai_transform_points3_general_masked ):
 
 
 ALIGNTEXT32
-    MOVUPS    ( REGOFF(0, EDX), XMM0 )	/* m0  | m1  | m2  | m3 */
-    MOVUPS    ( REGOFF(16, EDX), XMM1 )	/* m4  | m5  | m6  | m7 */
-    MOVUPS    ( REGOFF(32, EDX), XMM2 )	/* m8  | m9  | m10 | m11 */
-    MOVUPS    ( REGOFF(48, EDX), XMM3 )	/* m12 | m13 | m14 | m15 */
+    MOVAPS    ( REGOFF(0, EDX), XMM0 )	/* m0  | m1  | m2  | m3 */
+    MOVAPS    ( REGOFF(16, EDX), XMM1 )	/* m4  | m5  | m6  | m7 */
+    MOVAPS    ( REGOFF(32, EDX), XMM2 )	/* m8  | m9  | m10 | m11 */
+    MOVAPS    ( REGOFF(48, EDX), XMM3 )	/* m12 | m13 | m14 | m15 */
 
 
 ALIGNTEXT32
@@ -114,7 +111,7 @@ LLBL(K_GTPGM_top):
     ADDPS     ( XMM6, XMM4 )
     ADDPS     ( XMM3, XMM4 )
 
-    MOVUPS    ( XMM4, REGOFF(0, EDI) )
+    MOVAPS    ( XMM4, REGOFF(0, EDI) )
 
 LLBL(K_GTPGM_skip):
     INC_L     ( EBP )
@@ -238,7 +235,7 @@ ALIGNTEXT32
     MOVLPS   ( M(12), XMM2 )			/* - | - | m13 | m12 */
     MOVSS    ( M(10), XMM3 )			/* - | - |  -  | m10 */
     MOVSS    ( M(14), XMM4 )			/* - | - |  -  | m14 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP3DNRM_top):
     TEST_B( BL, REGIND(EBP) )
@@ -253,7 +250,7 @@ LLBL(K_GTP3DNRM_top):
     MULSS    ( XMM3, XMM0 )			/* sz*m10 */
     ADDSS    ( XMM4, XMM0 )			/* +m14 */
     MOVSS    ( XMM0, D(2) )			/* -> D(2) */
-	
+
 LLBL(K_GTP3DNRM_skip):
     INC_L( EBP )
     ADD_L( CONST(16), EDI )
@@ -313,7 +310,7 @@ ALIGNTEXT32
     MOVSS    ( M(10), XMM3 )			/* m10 */
     MOVSS    ( M(14), XMM4 )			/* m14 */
     XORPS    ( XMM6, XMM6 )			/* 0 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP3PM_top):
     TEST_B   ( BL, REGIND(EBP) )
@@ -408,7 +405,7 @@ LLBL(K_GTP3P2DM_top):
     ADDPS    ( XMM4, XMM3 )
     ADDPS    ( XMM2, XMM3 )
     MOVLPS   ( XMM3, D(0) )
-	
+
     MOVSS    ( S(2), XMM3 )
     MOVSS    ( XMM3, D(2) )
 
@@ -468,7 +465,7 @@ ALIGNTEXT32
 	MOVSS    ( M(5), XMM2 )			/* m5 */
 	UNPCKLPS ( XMM2, XMM1 )			/* m5 | m0 */
 	MOVLPS   ( M(12), XMM2 )		/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP3P2DNRM_top):
 	TEST_B( BL, REGIND(EBP) )
@@ -481,7 +478,7 @@ LLBL(K_GTP3P2DNRM_top):
 
 	MOVSS( S(2), XMM0 )
 	MOVSS( XMM0, D(2) )
-	
+
 LLBL(K_GTP3P2DNRM_skip):
 	INC_L( EBP )
 	ADD_L( CONST(16), EDI )
@@ -535,16 +532,16 @@ GLNAME(gl_katmai_transform_points3_3d_masked):
 
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m2  | m1  | m0 */
-	MOVUPS( M(4), XMM1 )			/* m6  | m5  | m4 */
-	MOVUPS( M(8), XMM2 )			/* m10 | m9  | m8 */
-	MOVUPS( M(12), XMM3 )			/* m14 | m13 | m12 */
+	MOVAPS( M(0), XMM0 )			/* m2  | m1  | m0 */
+	MOVAPS( M(4), XMM1 )			/* m6  | m5  | m4 */
+	MOVAPS( M(8), XMM2 )			/* m10 | m9  | m8 */
+	MOVAPS( M(12), XMM3 )			/* m14 | m13 | m12 */
 
 ALIGNTEXT32
 LLBL(K_GTP3P3DM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP3P3DM_skip) )
-	
+
 	MOVSS( S(0), XMM4 )
 	SHUFPS( CONST(0x0), XMM4, XMM4 )	/* ox | ox | ox */
 	MULPS( XMM0, XMM4 )			/* ox*m2 | ox*m1 | ox*m0 */
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_masked4.S b/xc/extras/Mesa/src/X86/katmai_xform_masked4.S
index d5d591bbb..e65930ed5 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_masked4.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_masked4.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_masked4.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -68,7 +65,7 @@ GLNAME( gl_katmai_transform_points4_general_masked ):
 
 	MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP )				/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )		/* source count */
 	MOV_B( ARG_FLAG, BL )				/* clip mask flags */
 
@@ -88,16 +85,16 @@ GLNAME( gl_katmai_transform_points4_general_masked ):
 	ADD_L( EDI, ECX ) 				/* count += dest ptr */
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )				/* m3  | m2  | m1  | m0 */
-	MOVUPS( M(4), XMM1 )				/* m7  | m6  | m5  | m4 */
-	MOVUPS( M(8), XMM2 )				/* m11 | m10 | m9  | m8 */
-	MOVUPS( M(12), XMM3 )				/* m15 | m14 | m13 | m12 */
-	
+	MOVAPS( M(0), XMM0 )				/* m3  | m2  | m1  | m0 */
+	MOVAPS( M(4), XMM1 )				/* m7  | m6  | m5  | m4 */
+	MOVAPS( M(8), XMM2 )				/* m11 | m10 | m9  | m8 */
+	MOVAPS( M(12), XMM3 )				/* m15 | m14 | m13 | m12 */
+
 ALIGNTEXT32
 LLBL(K_GTP4GM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP4GM_skip) )
-	
+
 	MOVSS( S(0), XMM4 )				/* ox */
 	SHUFPS( CONST(0x0), XMM4, XMM4 )		/* ox | ox | ox | ox */
 	MULPS( XMM0, XMM4 )				/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
@@ -117,7 +114,7 @@ LLBL(K_GTP4GM_top):
 	ADDPS( XMM5, XMM4 )				/* ox*m3+oy*m7 | ... */
 	ADDPS( XMM6, XMM4 )				/* ox*m3+oy*m7+oz*m11 | ... */
 	ADDPS( XMM7, XMM4 )				/* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
-	MOVUPS( XMM4, D(0) )				/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
+	MOVAPS( XMM4, D(0) )				/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
 
 LLBL(K_GTP4GM_skip):
 	INC_L     ( EBP )
@@ -169,7 +166,7 @@ GLNAME( gl_katmai_transform_points4_identity_masked ):
 	ADD_L( EDI, ECX ) 				/* count += dest ptr */
 
 	MOV_L( EAX, ARG_SOURCE )
-	
+
 	CMP_L( ESI, EDI )
 	JE( LLBL(K_GTP4IM_finish) )
 
@@ -219,7 +216,7 @@ GLNAME(gl_katmai_transform_points4_3d_no_rot_masked):
 
 	MOV_L( ARG_MATRIX, EDX )			/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP )				/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )		/* source count */
 	MOV_B( ARG_FLAG, BL )				/* clip mask flags */
 
@@ -246,12 +243,12 @@ ALIGNTEXT32
 	MOVSS( M(10), XMM1 )			/* m10 */
 	MOVLPS( M(12), XMM2 )			/* m13 | m12 */
 	MOVSS( M(14), XMM3 )			/* m14 */
-		
+
 ALIGNTEXT32
 LLBL(K_GTP43DNRM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP43DNRM_skip) )
-	
+
 	MOVLPS( S(0), XMM4 )			/* oy | ox */
 	MULPS( XMM0, XMM4 )			/* oy*m5 | ox*m0 */
 	MOVSS( S(3), XMM5 )			/* ow */
@@ -299,7 +296,7 @@ GLNAME(gl_katmai_transform_points4_3d_masked):
 
 	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
 	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-	
+
 	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
 	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
 
@@ -320,16 +317,16 @@ GLNAME(gl_katmai_transform_points4_3d_masked):
 
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
-	MOVUPS( M(4), XMM1 )			/* m7  | m6  | m5  |  m4 */
-	MOVUPS( M(8), XMM2 )			/* m11 | m10 | m9  |  m8 */
-	MOVUPS( M(12), XMM3 )			/* m15 | m14 | m13 | m12 */
-		
-ALIGNTEXT32	
+	MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
+	MOVAPS( M(4), XMM1 )			/* m7  | m6  | m5  |  m4 */
+	MOVAPS( M(8), XMM2 )			/* m11 | m10 | m9  |  m8 */
+	MOVAPS( M(12), XMM3 )			/* m15 | m14 | m13 | m12 */
+
+ALIGNTEXT32
 LLBL(K_GTP43P3DM_top):
 	TEST_B( BL, REGIND(EBP) )
 	JNZ( LLBL(K_GTP43P3DM_skip) )
-	
+
 	MOVSS( S(0), XMM4 )				/* ox */
 	SHUFPS( CONST(0x0), XMM4, XMM4 )		/* ox | ox | ox | ox */
 	MULPS( XMM0, XMM4 )				/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
@@ -349,11 +346,11 @@ LLBL(K_GTP43P3DM_top):
 	ADDPS( XMM5, XMM4 )				/* ox*m3+oy*m7 | ... */
 	ADDPS( XMM6, XMM4 )				/* ox*m3+oy*m7+oz*m11 | ... */
 	ADDPS( XMM7, XMM4 )				/* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
-	MOVUPS( XMM4, D(0) )				/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
+	MOVAPS( XMM4, D(0) )				/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
 
 	MOVSS( S(3), XMM4 )				/* ow */
 	MOVSS( XMM4, D(3) )				/* ->D(3) */
-	
+
 LLBL(K_GTP43P3DM_skip):
 	INC_L( EBP )
 	ADD_L( CONST(16), EDI )
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_raw1.S b/xc/extras/Mesa/src/X86/katmai_xform_raw1.S
index b8e9dcda8..24a567c89 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_raw1.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_raw1.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_raw1.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -84,8 +81,8 @@ GLNAME( gl_katmai_transform_points1_general_raw ):
 
 
 ALIGNTEXT32
-    MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0  */
-    MOVUPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
+    MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0  */
+    MOVAPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
 
 ALIGNTEXT32
 LLBL(K_GTP1GR_top):
@@ -145,7 +142,7 @@ ALIGNTEXT32
 LLBL(K_GTP1IR_top):
     MOV_L( S(0), EDX )
     MOV_L( EDX, D(0) )
-	
+
 LLBL(K_GTP1IR_skip):
     ADD_L     ( CONST(16), EDI )
     ADD_L     ( EAX, ESI )
@@ -195,7 +192,7 @@ ALIGNTEXT32
     MOVSS( M(12), XMM1 )			/* m12 */
     MOVSS( M(13), XMM2 )			/* m13 */
     MOVSS( M(14), XMM3 )			/* m14 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13DNRR_top):
     MOVSS( S(0), XMM4 )				/* ox */
@@ -218,7 +215,7 @@ LLBL(K_GTP13DNRR_finish):
     RET
 #undef FRAME_OFFSET
 
-    
+
 
 ALIGNTEXT4
 GLOBL GLNAME(gl_katmai_transform_points1_perspective_raw)
@@ -254,7 +251,7 @@ ALIGNTEXT32
     XORPS( XMM0, XMM0 )				/* 0 | 0 | 0 | 0 */
     MOVSS( M(0), XMM1 )				/* m0 */
     MOVSS( M(14), XMM2 )			/* m14 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13PR_top):
     MOVSS( S(0), XMM3 )				/* ox */
@@ -264,7 +261,7 @@ LLBL(K_GTP13PR_top):
 
     MOVSS( XMM0, D(1) )
     MOVSS( XMM0, D(3) )
-	
+
 LLBL(K_GTP13PR_skip):
     ADD_L( CONST(16), EDI )
     ADD_L( EAX, ESI )
@@ -310,7 +307,7 @@ GLNAME(gl_katmai_transform_points1_2d_raw):
 ALIGNTEXT32
     MOVLPS( M(0), XMM0 )			/* m1  | m0  */
     MOVLPS( M(12), XMM1 )			/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13P2DR_top):
     MOVSS( S(0), XMM2 )				/* ox */
@@ -318,7 +315,7 @@ LLBL(K_GTP13P2DR_top):
     MULPS( XMM0, XMM2 )				/* - | - | ox*m1 | ox*m0 */
     ADDPS( XMM1, XMM2 )				/* - | - | ox*m1+m13 | ox*m0+m12 */
     MOVLPS( XMM2, D(0) )
-	
+
 LLBL(K_GTP13P2DR_skip):
     ADD_L    ( CONST(16), EDI )
     ADD_L    ( EAX, ESI )
@@ -365,7 +362,7 @@ ALIGNTEXT32
 	MOVSS( M(0), XMM0 )			/* m0 */
 	MOVSS( M(12), XMM1 )			/* m12 */
 	MOVSS( M(13), XMM2 )			/* m13 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP13P2DNRR_top):
 	MOVSS( S(0), XMM3 )			/* ox */
@@ -373,7 +370,7 @@ LLBL(K_GTP13P2DNRR_top):
 	ADDSS( XMM1, XMM3 )			/* ox*m0+m12 */
 	MOVSS( XMM3, D(0) )
 	MOVSS( XMM2, D(1) )
-	
+
 LLBL(K_GTP13P2DNRR_skip):
 	ADD_L( CONST(16), EDI )
 	ADD_L( EAX, ESI )
@@ -419,10 +416,10 @@ GLNAME(gl_katmai_transform_points1_3d_raw):
 
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
-	MOVUPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
-		
-ALIGNTEXT32	
+	MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
+	MOVAPS( M(12), XMM1 )			/* m15 | m14 | m13 | m12 */
+
+ALIGNTEXT32
 LLBL(K_GTP13P3DR_top):
 	MOVSS( S(0), XMM2 )			/* ox */
 	SHUFPS( CONST(0x0), XMM2, XMM2 )	/* ox | ox | ox | ox */
@@ -431,7 +428,7 @@ LLBL(K_GTP13P3DR_top):
 	MOVLPS( XMM2, D(0) )			/*   -   |   -   | ->D(1)| ->D(0)*/
 	UNPCKHPS( XMM2, XMM2 )			/* ox*m3+m15 | ox*m3+m15 | ox*m2+m14 | ox*m2+m14 */
 	MOVSS( XMM2, D(2) )
-	
+
 LLBL(K_GTP13P3DR_skip):
 	ADD_L( CONST(16), EDI )
 	ADD_L( EAX, ESI )
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_raw2.S b/xc/extras/Mesa/src/X86/katmai_xform_raw2.S
index c330b0783..3d37ff687 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_raw2.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_raw2.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_raw2.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -60,7 +57,7 @@ GLNAME( gl_katmai_transform_points2_general_raw ):
 #define FRAME_OFFSET 8
     PUSH_L    ( ESI )
     PUSH_L    ( EDI )
-	
+
     MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
     MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
 
@@ -69,7 +66,7 @@ GLNAME( gl_katmai_transform_points2_general_raw ):
 
     TEST_L( ECX, ECX )
     JZ( LLBL(K_GTP2GR_finish) )			/* count was zero; go to finish */
-	
+
     MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
     OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
 
@@ -83,9 +80,9 @@ GLNAME( gl_katmai_transform_points2_general_raw ):
     ADD_L( EDI, ECX ) 				/* count += dest ptr */
 
 ALIGNTEXT32
-    MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0 */
-    MOVUPS( M(4), XMM1 )			/* m7  | m6  | m5  | m4 */
-    MOVUPS( M(12), XMM2 )			/* m15 | m14 | m13 | m12 */
+    MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0 */
+    MOVAPS( M(4), XMM1 )			/* m7  | m6  | m5  | m4 */
+    MOVAPS( M(12), XMM2 )			/* m15 | m14 | m13 | m12 */
 
 ALIGNTEXT32
 LLBL(K_GTP2GR_top):
@@ -98,8 +95,8 @@ LLBL(K_GTP2GR_top):
 
     ADDPS( XMM4, XMM3 )
     ADDPS( XMM2, XMM3 )
-    MOVUPS( XMM3, D(0) )	
-	
+    MOVAPS( XMM3, D(0) )
+
 LLBL(K_GTP2GR_skip):
     ADD_L     ( CONST(16), EDI )
     ADD_L     ( EAX, ESI )
@@ -151,7 +148,7 @@ LLBL(K_GTP2IR_top):
     MOV_L     ( EDX, D(0) )
     MOV_L     ( S(1), EDX )
     MOV_L     ( EDX, D(1) )
-	
+
 LLBL(K_GTP2IR_skip):
     ADD_L     ( CONST(16), EDI )
     ADD_L     ( EAX, ESI )
@@ -201,7 +198,7 @@ ALIGNTEXT32
     UNPCKLPS ( XMM2, XMM1 )			/* - | - | m5  | m0  */
     MOVLPS   ( M(12), XMM2 )			/* - | - | m13 | m12 */
     MOVSS    ( M(14), XMM3 )			/* - | - |  -  | m14 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP23DNRR_top):
     MOVLPS   ( S(0), XMM0 )			/* - | - |  oy   | ox */
@@ -223,7 +220,7 @@ LLBL(K_GTP23DNRR_finish):
     RET
 #undef FRAME_OFFSET
 
-    
+
 ALIGNTEXT4
 GLOBL GLNAME(gl_katmai_transform_points2_perspective_raw)
 GLNAME(gl_katmai_transform_points2_perspective_raw):
@@ -259,7 +256,7 @@ ALIGNTEXT32
     UNPCKLPS ( XMM2, XMM1 )			/* -  | -  | m5  | m0  */
     MOVSS    ( M(14), XMM3 )			/* m14 */
     XORPS    ( XMM0, XMM0 )			/* 0 | 0 | 0 | 0 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP23PR_top):
     MOVLPS( S(0), XMM4 )			/* oy | ox */
@@ -267,7 +264,7 @@ LLBL(K_GTP23PR_top):
     MOVLPS( XMM4, D(0) )			/* ->D(1) | ->D(0) */
     MOVSS( XMM3, D(2) )				/* ->D(2) */
     MOVSS( XMM0, D(3) )				/* ->D(3) */
-	
+
 LLBL(K_GTP23PR_skip):
     ADD_L( CONST(16), EDI )
     ADD_L( EAX, ESI )
@@ -315,7 +312,7 @@ ALIGNTEXT32
     MOVLPS( M(0), XMM0 )			/* m1  | m0 */
     MOVLPS( M(4), XMM1 )			/* m5  | m4 */
     MOVLPS( M(12), XMM2 )			/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP23P2DR_top):
     MOVSS( S(0), XMM3 )				/* ox */
@@ -378,14 +375,14 @@ ALIGNTEXT32
 	MOVSS    ( M(5), XMM2 )			/* m5 */
 	UNPCKLPS ( XMM2, XMM1 )			/* m5 | m0 */
 	MOVLPS   ( M(12), XMM2 )		/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP23P2DNRR_top):
 	MOVLPS( S(0), XMM0 )			/* oy | ox */
 	MULPS( XMM1, XMM0 )			/* oy*m5 | ox*m0 */
 	ADDPS( XMM2, XMM0 )			/* +m13 | +m12 */
 	MOVLPS( XMM0, D(0) )			/* ->D(1) | ->D(0) */
-	
+
 LLBL(K_GTP23P2DNRR_skip):
 	ADD_L( CONST(16), EDI )
 	ADD_L( EAX, ESI )
@@ -430,11 +427,11 @@ GLNAME(gl_katmai_transform_points2_3d_raw):
 	ADD_L( EDI, ECX ) 			/* count += dest ptr */
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m2  | m1  | m0 */
-	MOVUPS( M(4), XMM1 )			/* m6  | m5  | m4 */
-	MOVUPS( M(12), XMM2 )			/* m14 | m13 | m12 */
+	MOVAPS( M(0), XMM0 )			/* m2  | m1  | m0 */
+	MOVAPS( M(4), XMM1 )			/* m6  | m5  | m4 */
+	MOVAPS( M(12), XMM2 )			/* m14 | m13 | m12 */
 
-ALIGNTEXT32	
+ALIGNTEXT32
 LLBL(K_GTP23P3DR_top):
 	MOVSS( S(0), XMM3 )			/* ox */
 	SHUFPS( CONST(0x0), XMM3, XMM3 )	/* ox | ox | ox */
@@ -450,7 +447,7 @@ LLBL(K_GTP23P3DR_top):
 	MOVLPS( XMM3, D(0) )			/* ->D(1) | ->D(0) */
 	UNPCKHPS( XMM3, XMM3 )
 	MOVSS( XMM3, D(2) )			/* ->D(2) */
-	
+
 LLBL(K_GTP23P3DR_skip):
 	ADD_L( CONST(16), EDI )
 	ADD_L( EAX, ESI )
@@ -462,5 +459,3 @@ LLBL(K_GTP23P3DR_finish):
 	POP_L( ESI )
 	RET
 #undef FRAME_OFFSET
-
-
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_raw3.S b/xc/extras/Mesa/src/X86/katmai_xform_raw3.S
index d9ed22be3..e5176651d 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_raw3.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_raw3.S
@@ -1,14 +1,11 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_raw3.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
+
    SEG_TEXT
 
 
@@ -34,9 +31,9 @@
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
@@ -84,10 +81,10 @@ GLNAME( gl_katmai_transform_points3_general_raw ):
 
 
 ALIGNTEXT32
-    MOVUPS    ( REGOFF(0, EDX), XMM0 )	/* m0  | m1  | m2  | m3 */
-    MOVUPS    ( REGOFF(16, EDX), XMM1 )	/* m4  | m5  | m6  | m7 */
-    MOVUPS    ( REGOFF(32, EDX), XMM2 )	/* m8  | m9  | m10 | m11 */
-    MOVUPS    ( REGOFF(48, EDX), XMM3 )	/* m12 | m13 | m14 | m15 */
+    MOVAPS    ( REGOFF(0, EDX), XMM0 )	/* m0  | m1  | m2  | m3 */
+    MOVAPS    ( REGOFF(16, EDX), XMM1 )	/* m4  | m5  | m6  | m7 */
+    MOVAPS    ( REGOFF(32, EDX), XMM2 )	/* m8  | m9  | m10 | m11 */
+    MOVAPS    ( REGOFF(48, EDX), XMM3 )	/* m12 | m13 | m14 | m15 */
 
 
 ALIGNTEXT32
@@ -107,7 +104,7 @@ LLBL(K_GTPGR_top):
     ADDPS     ( XMM6, XMM4 )
     ADDPS     ( XMM3, XMM4 )
 
-    MOVUPS    ( XMM4, REGOFF(0, EDI) )
+    MOVAPS    ( XMM4, REGOFF(0, EDI) )
 
 LLBL(K_GTPGR_skip):
     ADD_L     ( CONST(16), EDI )
@@ -240,7 +237,7 @@ LLBL(K_GTP3DNRR_finish):
     RET
 #undef FRAME_OFFSET
 
-    
+
 
 ALIGNTEXT4
 GLOBL GLNAME(gl_katmai_transform_points3_perspective_raw)
@@ -279,7 +276,7 @@ ALIGNTEXT32
     MOVSS    ( M(10), XMM3 )			/* m10 */
     MOVSS    ( M(14), XMM4 )			/* m14 */
     XORPS    ( XMM6, XMM6 )			/* 0 */
-    
+
 ALIGNTEXT32
 LLBL(K_GTP3PR_top):
     MOVLPS   ( S(0), XMM0 )			/* oy | ox */
@@ -347,7 +344,7 @@ ALIGNTEXT32
     MOVLPS( M(0), XMM0 )			/* m1  | m0 */
     MOVLPS( M(4), XMM1 )			/* m5  | m4 */
     MOVLPS( M(12), XMM2 )			/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP3P2DR_top):
     MOVSS    ( S(0), XMM3 )			/* ox */
@@ -360,7 +357,7 @@ LLBL(K_GTP3P2DR_top):
     ADDPS    ( XMM4, XMM3 )
     ADDPS    ( XMM2, XMM3 )
     MOVLPS   ( XMM3, D(0) )
-	
+
     MOVSS    ( S(2), XMM3 )
     MOVSS    ( XMM3, D(2) )
 
@@ -412,14 +409,14 @@ ALIGNTEXT32
 	MOVSS    ( M(5), XMM2 )			/* m5 */
 	UNPCKLPS ( XMM2, XMM1 )			/* m5 | m0 */
 	MOVLPS   ( M(12), XMM2 )		/* m13 | m12 */
-	
+
 ALIGNTEXT32
 LLBL(K_GTP3P2DNRR_top):
 	MOVLPS( S(0), XMM0 )			/* oy | ox */
 	MULPS( XMM1, XMM0 )			/* oy*m5 | ox*m0 */
 	ADDPS( XMM2, XMM0 )			/* +m13 | +m12 */
 	MOVLPS( XMM0, D(0) )			/* ->D(1) | ->D(0) */
-	
+
 	MOVSS( S(2), XMM0 )
 	MOVSS( XMM0, D(2) )
 
@@ -470,12 +467,12 @@ GLNAME(gl_katmai_transform_points3_3d_raw):
 
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m2  | m1  | m0 */
-	MOVUPS( M(4), XMM1 )			/* m6  | m5  | m4 */
-	MOVUPS( M(8), XMM2 )			/* m10 | m9  | m8 */
-	MOVUPS( M(12), XMM3 )			/* m14 | m13 | m12 */
+	MOVAPS( M(0), XMM0 )			/* m2  | m1  | m0 */
+	MOVAPS( M(4), XMM1 )			/* m6  | m5  | m4 */
+	MOVAPS( M(8), XMM2 )			/* m10 | m9  | m8 */
+	MOVAPS( M(12), XMM3 )			/* m14 | m13 | m12 */
 
-ALIGNTEXT32	
+ALIGNTEXT32
 LLBL(K_GTP3P3DR_top):
 	MOVSS( S(0), XMM4 )
 	SHUFPS( CONST(0x0), XMM4, XMM4 )	/* ox | ox | ox */
@@ -508,5 +505,3 @@ LLBL(K_GTP3P3DR_finish):
 	POP_L( ESI )
 	RET
 #undef FRAME_OFFSET
-
-
diff --git a/xc/extras/Mesa/src/X86/katmai_xform_raw4.S b/xc/extras/Mesa/src/X86/katmai_xform_raw4.S
index eabec45bc..6cbe22a16 100644
--- a/xc/extras/Mesa/src/X86/katmai_xform_raw4.S
+++ b/xc/extras/Mesa/src/X86/katmai_xform_raw4.S
@@ -1,200 +1,193 @@
-/* $XFree86: xc/extras/Mesa/src/X86/katmai_xform_raw4.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
 #include "assyntax.h"
-
 /** TODO:
   * - insert PREFETCH instructions to avoid cache-misses !
   * - some more optimizations are possible...
   * - for 40-50% more performance in the SSE-functions, the
-  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !	
+  *   data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned !
   */
-	
-   SEG_TEXT
 
+	SEG_TEXT
 
-#define S(i) 	REGOFF(i * 4, ESI)
-#define D(i) 	REGOFF(i * 4, EDI)
-#define M(i) 	REGOFF(i * 4, EDX)
+
+#define S(i)	REGOFF(i * 4, ESI)
+#define D(i)	REGOFF(i * 4, EDI)
+#define M(i)	REGOFF(i * 4, EDX)
 
 /*
- * 	Offsets into GLvector4f
+ *	Offsets into GLvector4f
  */
-#define V4F_DATA 	0
-#define V4F_START 	4
-#define V4F_COUNT 	8
-#define V4F_STRIDE 	12
-#define V4F_SIZE 	16
-#define V4F_FLAGS 	20
-
-#define VEC_SIZE_1   1
-#define VEC_SIZE_2   3
-#define VEC_SIZE_3   7
-#define VEC_SIZE_4   15
+#define V4F_DATA	0
+#define V4F_START	4
+#define V4F_COUNT	8
+#define V4F_STRIDE	12
+#define V4F_SIZE	16
+#define V4F_FLAGS	20
+
+#define VEC_SIZE_1	1
+#define VEC_SIZE_2	3
+#define VEC_SIZE_3	7
+#define VEC_SIZE_4	15
 
 /*
  *	Offsets for transform_func arguments
  *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
+ *	typedef void (*transform_func)( GLvector4f *to_vec,
+ *	                                const GLfloat m[16],
+ *	                                const GLvector4f *from_vec,
  *	                                const GLubyte *clipmask,
  *	                                const GLubyte flag );
  */
-#define OFFSET_DEST 4
-#define OFFSET_MATRIX 8
-#define OFFSET_SOURCE 12
-#define OFFSET_CLIP 16
-#define OFFSET_FLAG 20
+#define OFFSET_DEST	4
+#define OFFSET_MATRIX	8
+#define OFFSET_SOURCE	12
+#define OFFSET_CLIP	16
+#define OFFSET_FLAG	20
 
-#define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-#define ARG_MATRIX 	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
-#define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-#define ARG_CLIP 	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
-#define ARG_FLAG 	REGOFF(FRAME_OFFSET+OFFSET_FLAG, ESP)
+#define ARG_DEST	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
+#define ARG_MATRIX	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
+#define ARG_SOURCE	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
+#define ARG_CLIP	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
+#define ARG_FLAG	REGOFF(FRAME_OFFSET+OFFSET_FLAG, ESP)
 
 
 ALIGNTEXT4
-GLOBL GLNAME(gl_katmai_transform_points4_general_raw)
+GLOBL GLNAME( gl_katmai_transform_points4_general_raw )
 GLNAME( gl_katmai_transform_points4_general_raw ):
 
 #define FRAME_OFFSET 8
-	PUSH_L    ( ESI )
-	PUSH_L    ( EDI )
+	PUSH_L( ESI )
+	PUSH_L( EDI )
 
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
+	MOV_L( ARG_SOURCE, ESI )		/* ptr to source GLvector4f */
+	MOV_L( ARG_DEST, EDI )			/* ptr to dest GLvector4f */
 
-	MOV_L( ARG_MATRIX, EDX ) 			/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )		/* source count */
+	MOV_L( ARG_MATRIX, EDX )		/* ptr to matrix */
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )	/* source count */
 
-	CMP_L( CONST(0), ECX )				/* count == 0 ? */
-	JE( LLBL(K_GTP4GR_finish) )			/* yes -> nothing to do. */
+	CMP_L( CONST(0), ECX )			/* count == 0 ? */
+	JE( LLBL(K_GTP4GR_finish) )		/* yes -> nothing to do. */
 
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )		/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )	/* stride */
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */
 
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )		/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )	/* set dest count */
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )/* set dest size */
 
-	SHL_L( CONST(4), ECX )				/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI )		/* ptr to first source vertex */
+	SHL_L( CONST(4), ECX )			/* count *= 16 */
+	MOV_L( REGOFF(V4F_START, ESI), ESI )	/* ptr to first source vertex */
 
-	MOV_L( REGOFF(V4F_START, EDI), EDI )		/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 				/* count += dest ptr */
+	MOV_L( REGOFF(V4F_START, EDI), EDI )	/* ptr to first dest vertex */
+	ADD_L( EDI, ECX )			/* count += dest ptr */
+
+	MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  | m0 */
+	MOVAPS( M(4), XMM1 )			/* m7  | m6  | m5  | m4 */
+	MOVAPS( M(8), XMM2 )			/* m11 | m10 | m9  | m8 */
+	MOVAPS( M(12), XMM3 )			/* m15 | m14 | m13 | m12 */
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )				/* m3  | m2  | m1  | m0 */
-	MOVUPS( M(4), XMM1 )				/* m7  | m6  | m5  | m4 */
-	MOVUPS( M(8), XMM2 )				/* m11 | m10 | m9  | m8 */
-	MOVUPS( M(12), XMM3 )				/* m15 | m14 | m13 | m12 */
-	
-ALIGNTEXT32
-LLBL(K_GTP4GR_top):
-	MOVSS( S(0), XMM4 )				/* ox */
-	SHUFPS( CONST(0x0), XMM4, XMM4 )		/* ox | ox | ox | ox */
-	MULPS( XMM0, XMM4 )				/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
-
-	MOVSS( S(1), XMM5 )				/* oy */
-	SHUFPS( CONST(0x0), XMM5, XMM5 )		/* oy | oy | oy | oy */
-	MULPS( XMM1, XMM5 )				/* oy*m7 | oy*m6 | oy*m5 | oy*m4 */
-
-	MOVSS( S(2), XMM6 )				/* oz */
-	SHUFPS( CONST(0x0), XMM6, XMM6 )		/* oz | oz | oz | oz */
-	MULPS( XMM2, XMM6 )				/* oz*m11 | oz*m10 | oz*m9 | oz*m8 */
-
-	MOVSS( S(3), XMM7 )				/* ow */
-	SHUFPS( CONST(0x0), XMM7, XMM7 )		/* ow | ow | ow | ow */
-	MULPS( XMM3, XMM7 )				/* ow*m15 | ow*m14 | ow*m13 | ow*m12 */
-
-	ADDPS( XMM5, XMM4 )				/* ox*m3+oy*m7 | ... */
-	ADDPS( XMM6, XMM4 )				/* ox*m3+oy*m7+oz*m11 | ... */
-	ADDPS( XMM7, XMM4 )				/* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
-	MOVUPS( XMM4, D(0) )				/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
-	
-LLBL(K_GTP4GR_skip):
-	ADD_L     ( CONST(16), EDI )
-	ADD_L     ( EAX, ESI )
-	CMP_L     ( ECX, EDI )
-	JNE       ( LLBL(K_GTP4GR_top) )
-	
-LLBL(K_GTP4GR_finish):
-	POP_L     ( EDI )
-	POP_L     ( ESI )
+LLBL( K_GTP4GR_top ):
+	MOVSS( S(0), XMM4 )			/* ox */
+	SHUFPS( CONST(0x0), XMM4, XMM4 )	/* ox | ox | ox | ox */
+	MULPS( XMM0, XMM4 )			/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
+
+	MOVSS( S(1), XMM5 )			/* oy */
+	SHUFPS( CONST(0x0), XMM5, XMM5 )	/* oy | oy | oy | oy */
+	MULPS( XMM1, XMM5 )			/* oy*m7 | oy*m6 | oy*m5 | oy*m4 */
+
+	MOVSS( S(2), XMM6 )			/* oz */
+	SHUFPS( CONST(0x0), XMM6, XMM6 )	/* oz | oz | oz | oz */
+	MULPS( XMM2, XMM6 )			/* oz*m11 | oz*m10 | oz*m9 | oz*m8 */
+
+	MOVSS( S(3), XMM7 )			/* ow */
+	SHUFPS( CONST(0x0), XMM7, XMM7 )	/* ow | ow | ow | ow */
+	MULPS( XMM3, XMM7 )			/* ow*m15 | ow*m14 | ow*m13 | ow*m12 */
+
+	ADDPS( XMM5, XMM4 )			/* ox*m3+oy*m7 | ... */
+	ADDPS( XMM6, XMM4 )			/* ox*m3+oy*m7+oz*m11 | ... */
+	ADDPS( XMM7, XMM4 )			/* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
+	MOVAPS( XMM4, D(0) )			/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
+
+LLBL( K_GTP4GR_skip ):
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL(K_GTP4GR_top) )
+
+LLBL( K_GTP4GR_finish ):
+	POP_L( EDI )
+	POP_L( ESI )
 	RET
 #undef FRAME_OFFSET
 
 
 ALIGNTEXT4
-GLOBL GLNAME(gl_katmai_transform_points4_3d_raw)
-GLNAME(gl_katmai_transform_points4_3d_raw):
+GLOBL GLNAME( gl_katmai_transform_points4_3d_raw )
+GLNAME( gl_katmai_transform_points4_3d_raw ):
 
 #define FRAME_OFFSET 8
 	PUSH_L( ESI )
 	PUSH_L( EDI )
 
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
+	MOV_L( ARG_SOURCE, ESI )		/* ptr to source GLvector4f */
+	MOV_L( ARG_DEST, EDI )			/* ptr to dest GLvector4f */
 
-	MOV_L( ARG_MATRIX, EDX )			/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )		/* source count */
+	MOV_L( ARG_MATRIX, EDX )		/* ptr to matrix */
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )	/* source count */
 
 	TEST_L( ECX, ECX)
-	JZ( LLBL(K_GTP43P3DR_finish) )			/* count was zero; go to finish */
+	JZ( LLBL(K_GTP43P3DR_finish) )		/* count was zero; go to finish */
 
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )		/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )	/* stride */
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */
 
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )		/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )	/* set dest count */
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )/* set dest size */
 
-	SHL_L( CONST(4), ECX )				/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI )		/* ptr to first source vertex */
+	SHL_L( CONST(4), ECX )			/* count *= 16 */
+	MOV_L( REGOFF(V4F_START, ESI), ESI )	/* ptr to first source vertex */
 
-	MOV_L( REGOFF(V4F_START, EDI), EDI )		/* ptr to first dest vertex */
-	ADD_L( EDI, ECX )				/* count += dest ptr */
+	MOV_L( REGOFF(V4F_START, EDI), EDI )	/* ptr to first dest vertex */
+	ADD_L( EDI, ECX )			/* count += dest ptr */
 
+	MOVAPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
+	MOVAPS( M(4), XMM1 )			/* m7  | m6  | m5  |  m4 */
+	MOVAPS( M(8), XMM2 )			/* m11 | m10 | m9  |  m8 */
+	MOVAPS( M(12), XMM3 )			/* m15 | m14 | m13 | m12 */
 
 ALIGNTEXT32
-	MOVUPS( M(0), XMM0 )			/* m3  | m2  | m1  |  m0 */
-	MOVUPS( M(4), XMM1 )			/* m7  | m6  | m5  |  m4 */
-	MOVUPS( M(8), XMM2 )			/* m11 | m10 | m9  |  m8 */
-	MOVUPS( M(12), XMM3 )			/* m15 | m14 | m13 | m12 */
-		
-ALIGNTEXT32	
-LLBL(K_GTP43P3DR_top):
-	MOVSS( S(0), XMM4 )				/* ox */
-	SHUFPS( CONST(0x0), XMM4, XMM4 )		/* ox | ox | ox | ox */
-	MULPS( XMM0, XMM4 )				/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
-
-	MOVSS( S(1), XMM5 )				/* oy */
-	SHUFPS( CONST(0x0), XMM5, XMM5 )		/* oy | oy | oy | oy */
-	MULPS( XMM1, XMM5 )				/* oy*m7 | oy*m6 | oy*m5 | oy*m4 */
-
-	MOVSS( S(2), XMM6 )				/* oz */
-	SHUFPS( CONST(0x0), XMM6, XMM6 )		/* oz | oz | oz | oz */
-	MULPS( XMM2, XMM6 )				/* oz*m11 | oz*m10 | oz*m9 | oz*m8 */
-
-	MOVSS( S(3), XMM7 )				/* ow */
-	SHUFPS( CONST(0x0), XMM7, XMM7 )		/* ow | ow | ow | ow */
-	MULPS( XMM3, XMM7 )				/* ow*m15 | ow*m14 | ow*m13 | ow*m12 */
-
-	ADDPS( XMM5, XMM4 )				/* ox*m3+oy*m7 | ... */
-	ADDPS( XMM6, XMM4 )				/* ox*m3+oy*m7+oz*m11 | ... */
-	ADDPS( XMM7, XMM4 )				/* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
-	MOVUPS( XMM4, D(0) )				/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
-
-	MOVSS( S(3), XMM4 )				/* ow */
-	MOVSS( XMM4, D(3) )				/* ->D(3) */
-		
-LLBL(K_GTP43P3DR_skip):
+LLBL( K_GTP43P3DR_top ):
+	MOVSS( S(0), XMM4 )			/* ox */
+	SHUFPS( CONST(0x0), XMM4, XMM4 )	/* ox | ox | ox | ox */
+	MULPS( XMM0, XMM4 )			/* ox*m3 | ox*m2 | ox*m1 | ox*m0 */
+
+	MOVSS( S(1), XMM5 )			/* oy */
+	SHUFPS( CONST(0x0), XMM5, XMM5 )	/* oy | oy | oy | oy */
+	MULPS( XMM1, XMM5 )			/* oy*m7 | oy*m6 | oy*m5 | oy*m4 */
+
+	MOVSS( S(2), XMM6 )			/* oz */
+	SHUFPS( CONST(0x0), XMM6, XMM6 )	/* oz | oz | oz | oz */
+	MULPS( XMM2, XMM6 )			/* oz*m11 | oz*m10 | oz*m9 | oz*m8 */
+
+	MOVSS( S(3), XMM7 )			/* ow */
+	SHUFPS( CONST(0x0), XMM7, XMM7 )	/* ow | ow | ow | ow */
+	MULPS( XMM3, XMM7 )			/* ow*m15 | ow*m14 | ow*m13 | ow*m12 */
+
+	ADDPS( XMM5, XMM4 )			/* ox*m3+oy*m7 | ... */
+	ADDPS( XMM6, XMM4 )			/* ox*m3+oy*m7+oz*m11 | ... */
+	ADDPS( XMM7, XMM4 )			/* ox*m3+oy*m7+oz*m11+ow*m15 | ... */
+	MOVAPS( XMM4, D(0) )			/* ->D(3) | ->D(2) | ->D(1) | ->D(0) */
+
+	MOVSS( S(3), XMM4 )			/* ow */
+	MOVSS( XMM4, D(3) )			/* ->D(3) */
+
+LLBL( K_GTP43P3DR_skip ):
 	ADD_L( CONST(16), EDI )
 	ADD_L( EAX, ESI )
 	CMP_L( ECX, EDI )
 	JNE( LLBL(K_GTP43P3DR_top) )
 
-LLBL(K_GTP43P3DR_finish):
+LLBL( K_GTP43P3DR_finish ):
 	POP_L( EDI )
 	POP_L( ESI )
 	RET
 #undef FRAME_OFFSET
-
diff --git a/xc/extras/Mesa/src/X86/mmx.h b/xc/extras/Mesa/src/X86/mmx.h
index f0e05cf06..07d66c774 100644
--- a/xc/extras/Mesa/src/X86/mmx.h
+++ b/xc/extras/Mesa/src/X86/mmx.h
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
diff --git a/xc/extras/Mesa/src/X86/mmx_blend.S b/xc/extras/Mesa/src/X86/mmx_blend.S
index e7d6e118c..21fa36ede 100644
--- a/xc/extras/Mesa/src/X86/mmx_blend.S
+++ b/xc/extras/Mesa/src/X86/mmx_blend.S
@@ -350,7 +350,3 @@ LLBL(GMBT_1):
     MOV_L     ( EBP, ESP )
     POP_L     ( EBP )
     RET
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/vertex.S b/xc/extras/Mesa/src/X86/vertex.S
deleted file mode 100644
index 0326c6392..000000000
--- a/xc/extras/Mesa/src/X86/vertex.S
+++ /dev/null
@@ -1,102 +0,0 @@
-#include "assyntax.h"
-
-	
-	SEG_TEXT
-
-/*#define MAT_SX 0*/        /*  accessed by REGIND !! */
-#define MAT_SY 20
-#define MAT_SZ 40
-#define MAT_TX 48
-#define MAT_TY 52
-#define MAT_TZ 56
-
-
-
-
-/*
- * void gl_v16_x86_general_xform (	GLfloat *dest,
- *				const GLfloat *m,
- *				const GLfloat *src,
- *                              GLuint src_stride,
- *				GLuint count )
- */
-
-
-/* This is nothing more glamorous than an objdump of one of Josh's
- * routines hacked to match the above.    
- */
-ALIGNTEXT16
-GLOBL GLNAME( gl_v16_x86_general_xform )
-GLNAME( gl_v16_x86_general_xform ):
-
-
-	PUSH_L ( EDI )
-	PUSH_L ( ESI )
-
-    	MOV_L     ( REGOFF(12, ESP), EAX ) /* dest */
-    	MOV_L     ( REGOFF(16, ESP), ESI ) /* mat */
-    	MOV_L     ( REGOFF(20, ESP), EDX ) /* src */
-    	MOV_L     ( REGOFF(24, ESP), EDI ) /* src_stride */
-    	MOV_L     ( REGOFF(28, ESP), ECX ) /* count */
-
-
-LLBL(v16x86_loop):	
-        FLD_S   ( REGOFF( 0x0, EDX ) )
-        FMUL_S  ( REGOFF( 0x0, ESI ) )
-        FLD_S   ( REGOFF( 0x0, EDX ) )
-        FMUL_S  ( REGOFF( 0x4, ESI ) )
-        FLD_S   ( REGOFF( 0x0, EDX ) )
-        FMUL_S  ( REGOFF( 0x8, ESI ) )
-        FLD_S   ( REGOFF( 0x0, EDX ) )
-        FMUL_S  ( REGOFF( 0xc, ESI ) )
-        FLD_S   ( REGOFF( 0x4, EDX ) )
-        FMUL_S  ( REGOFF( 0x10, ESI ) )
-        FLD_S   ( REGOFF( 0x4, EDX ) )
-        FMUL_S  ( REGOFF( 0x14, ESI ) )
-        FLD_S   ( REGOFF( 0x4, EDX ) )
-        FMUL_S  ( REGOFF( 0x18, ESI ) )
-        FLD_S   ( REGOFF( 0x4, EDX ) )
-        FMUL_S  ( REGOFF( 0x1c, ESI ) )
-        FXCH    ( ST(3) )
-        FADDP   ( ST(0),ST(7) )
-        FXCH    ( ST(1) )
-        FADDP   ( ST(0),ST(5) )
-        FADDP   ( ST(0),ST(3) )
-        FADDP   ( ST(0),ST(1) )
-        FLD_S   ( REGOFF( 0x8, EDX ) )
-        FMUL_S  ( REGOFF( 0x20, ESI ) )
-        FLD_S   ( REGOFF( 0x8, EDX ) )
-        FMUL_S  ( REGOFF( 0x24, ESI ) )
-        FLD_S   ( REGOFF( 0x8, EDX ) )
-        FMUL_S  ( REGOFF( 0x28, ESI ) )
-        FLD_S   ( REGOFF( 0x8, EDX ) )
-        FMUL_S  ( REGOFF( 0x2c, ESI ) )
-        FXCH    ( ST(3) )
-        FADDP   ( ST(0),ST(7) )
-        FXCH    ( ST(1) )
-        FADDP   ( ST(0),ST(5) )
-        FADDP   ( ST(0),ST(3) )
-        FADDP   ( ST(0),ST(1) )
-        FXCH    ( ST(3) )
-        FADD_S  ( REGOFF( 0x30, ESI ) )
-        FXCH    ( ST(2) )
-        FADD_S  ( REGOFF( 0x34, ESI ) )
-        FXCH    ( ST(1) )
-        FADD_S  ( REGOFF( 0x38, ESI ) )
-        FXCH    ( ST(3) )
-        FADD_S  ( REGOFF( 0x3c, ESI ) )
-        FXCH    ( ST(2) )
-        FSTP_S  ( REGOFF( 0x0, EAX ) )
-        FSTP_S  ( REGOFF( 0x4, EAX ) )
-        FXCH    ( ST(1) )
-        FSTP_S  ( REGOFF( 0x8, EAX ) )
-        FSTP_S  ( REGOFF( 0xc, EAX ) )
-        ADD_L   ( CONST(64), EAX )
-        ADD_L   ( EDI, EDX )
-        DEC_L   ( ECX )
-        JNE     ( LLBL(v16x86_loop) )
-
-	POP_L ( ESI )
-	POP_L ( EDI )
-	RET
-
diff --git a/xc/extras/Mesa/src/X86/vertex_3dnow.S b/xc/extras/Mesa/src/X86/vertex_3dnow.S
deleted file mode 100644
index 693ac0812..000000000
--- a/xc/extras/Mesa/src/X86/vertex_3dnow.S
+++ /dev/null
@@ -1,234 +0,0 @@
-#include "assyntax.h"
-
-   SEG_TEXT
-
-
-
-/*#define MAT_SX 0         accessed by REGIND !! */
-#define MAT_SY 20
-#define MAT_SZ 40
-#define MAT_TX 48
-#define MAT_TY 52
-#define MAT_TZ 56
-
-
-
-
-
-/* Do viewport map and perspective projection.  Args should look like:
- *
- * x86_3dnow_project_vertices( float *first_vertex,
- *                             const float *last_vertex,
- *                             float *matrix, 
- *                             GLuint stride )
- *
- * This routine assumes a sane vertex layout with x,y,z,w as
- * the first four elements, to be projected in clip-space, to
- * x/w,y/w,z/w,1/w, and then transformed according to the matrix to
- * device space.  The device coordinates will overwrite the clip
- * coordinates as the first four elements of the vertex.
- *
- * If projection is required for other elements, such as texcoords,
- * you will have to code a specialized version of this routine.  See
- * FX/X86 for examples.
- *
- * These routines are simplified versions of the FX code written by
- * Holger.  
- */
-
-GLOBL GLNAME( gl_3dnow_project_vertices )
-GLNAME( gl_3dnow_project_vertices ):
-
-    PUSH_L    ( EBP )
-    FEMMS
-    PREFETCH  ( REGOFF(8, ESP) )         /* fetch the first vertex */
-
-    MOV_L     ( REGOFF(8, ESP), ECX )    /* first_vert */
-    MOV_L     ( REGOFF(12, ESP), EDX )   /* last_vert */
-    MOV_L     ( REGOFF(16, ESP), EBP )   /* matrix */
-    MOV_L     ( REGOFF(20, ESP), EAX )   /* stride */
-
-    MOVD      ( REGOFF(MAT_TX, EBP), MM6 )      /*             | tx           */
-    PUNPCKLDQ ( REGOFF(MAT_TY, EBP), MM6 )      /*  ty         | tx           */
-    MOVD      ( REGIND(EBP), MM5 )
-    PUNPCKLDQ ( REGOFF(MAT_SY, EBP), MM5 )      /*  vsy        | vsx          */
-    MOVD      ( REGOFF(MAT_SZ, EBP), MM1 )      /*             | vsz          */
-    SUB_L     ( ECX, EDX )	                /* last -= first */
-
-ALIGNTEXT32
-LLBL(v16_3dnow_pv_loop_start):
-
-    PREFETCH  ( REGOFF(64, ECX) )               /* fetch one/two verts ahead */
-    MOVD      ( REGOFF(12, ECX), MM0 )          /*              | f[3]        */
-    PFRCP     ( MM0, MM0 )                      /*  oow = 1/f[3]              */
-    MOVD      ( REGOFF(12, ECX), MM7 )          /*              | f[3]        */
-    PFRCPIT1  ( MM0, MM7 )
-    PFRCPIT2  ( MM0, MM7 )                      /*  oow         | oow         */
-    PUNPCKLDQ ( MM7, MM7 )
-    MOVQ      ( REGIND(ECX), MM2 )              /*  f[1]        | f[0]        */
-    PFMUL     ( MM7, MM2 )                      /*  f[1] * oow  | f[0] * oow  */
-    MOVD      ( REGOFF(8, ECX), MM3 )           /*              | f[2]        */
-    PFMUL     ( MM7, MM3 )                      /*              | f[2] * oow  */
-    MOVD      ( REGOFF(MAT_TZ, EBP), MM0 )      /*              | vtz         */
-    PFMUL     ( MM1, MM3 )                      /*              | f[2] *= vsz */
-    PFADD     ( MM0, MM3 )                      /*              | f[2] += vtz */
-    PFMUL     ( MM5, MM2 )                      /*  f[1] *= vsy | f[0] *= vsx */
-    PFADD     ( MM6, MM2 )                      /*  f[1] += vty | f[0] += vtx */
-    PUNPCKLDQ ( MM7, MM3 )	                /*  f[3] = oow  | f[2] */
-    MOVQ      ( MM2, REGOFF(0, ECX) )
-    MOVQ      ( MM3, REGOFF(8, ECX) ) 
-    ADD_L     ( EAX, ECX )                      /* f += stride */
-    SUB_L     ( EAX, EDX )
-    JA        ( LLBL(v16_3dnow_pv_loop_start) )
-
-    FEMMS
-    POP_L     ( EBP )
-    RET
-
-
-
-
-
-
-
-
-GLOBL GLNAME( gl_3dnow_project_clipped_vertices )
-GLNAME( gl_3dnow_project_clipped_vertices ):
-
-    PUSH_L    ( EBP )
-    PUSH_L    ( ESI )
-
-    FEMMS
-
-    PREFETCH  ( REGOFF(12, ESP) )            /* fetch the first vertex */
-
-    MOV_L     ( REGOFF(12, ESP), ECX )    /* first_vert */
-    MOV_L     ( REGOFF(16, ESP), EDX )     /* last_vert */
-    MOV_L     ( REGOFF(20, ESP), EBP )     /* matrix */
-    MOV_L     ( REGOFF(24, ESP), EAX )     /* stride */
-    MOV_L     ( REGOFF(28, ESP), ESI )     /* clip_mask  */
-
-
-    MOVD      ( REGOFF(MAT_TX, EBP), MM6 )      /*             | tx           */
-    PUNPCKLDQ ( REGOFF(MAT_TY, EBP), MM6 )      /*  ty         | tx           */
-    MOVD      ( REGIND(EBP), MM5 )
-    PUNPCKLDQ ( REGOFF(MAT_SY, EBP), MM5 )      /*  vsy        | vsx          */
-    MOVD      ( REGOFF(MAT_SZ, EBP), MM1 )      /*             | vsz          */
-
-
-ALIGNTEXT32
-LLBL(v16_3dnow_pcv_loop_start):
-
-    CMP_B     ( CONST(0), REGIND(ESI) )
-    JNE       ( LLBL(v16_3dnow_pcv_skip) )
-
-    MOVD      ( REGOFF(12, ECX), MM0)           /*              | f[3]        */
-    PFRCP     ( MM0, MM0 )                      /*  oow = 1/f[3]              */
-    MOVD      ( REGOFF(12, ECX), MM7)           /*              | f[3]        */
-    PFRCPIT1  ( MM0, MM7 )
-    PFRCPIT2  ( MM0, MM7 )                      /*  oow         | oow         */
-    PUNPCKLDQ ( MM7, MM7 )
-    MOVQ      ( REGIND(ECX), MM2 )              /*  f[1]        | f[0]        */
-    PFMUL     ( MM7, MM2 )                      /*  f[1] * oow  | f[0] * oow  */
-    MOVD      ( REGOFF(8, ECX), MM3 )           /*              | f[2]        */
-    PFMUL     ( MM7, MM3 )                      /*              | f[2] * oow  */
-    MOVD      ( REGOFF(MAT_TZ, EBP), MM0 )      /*              | vtz         */
-    PFMUL     ( MM1, MM3 )                      /*              | f[2] *= vsz */
-    PFADD     ( MM0, MM3 )                      /*              | f[2] += vtz */
-    PFMUL     ( MM5, MM2 )                      /*  f[1] *= vsy | f[0] *= vsx */
-    PFADD     ( MM6, MM2 )                      /*  f[1] += vty | f[0] += vtx */
-    PUNPCKLDQ ( MM7, MM3 )	                /*  f[3] = oow  | f[2] */
-    MOVQ      ( MM2, REGOFF(0, ECX) )
-    MOVQ      ( MM3, REGOFF(8, ECX) ) 
-
-LLBL(v16_3dnow_pcv_skip):
-    ADD_L     ( EAX, ECX )                      /* f += stride     */
-    INC_L     ( ESI )                           /* next ClipMask */
-
-    CMP_L     ( ECX, EDX )                     
-    JNE       ( LLBL(v16_3dnow_pcv_loop_start) )
-
-    FEMMS
-
-    POP_L     ( ESI )
-    POP_L     ( EBP )
-    RET
-
-
-
-
-
-/*
- * void gl_3dnow_transform_v16 (GLfloat *dest,
- *		        	const GLfloat *m,
- *				const GLfloat *src,
- *                              GLuint src_stride,
- *				GLuint count )
- *
- * These tranformation functions could disappear if the standard ones
- * took an output stride.  
- */
-GLOBL GLNAME( gl_v16_3dnow_general_xform )
-GLNAME( gl_v16_3dnow_general_xform ):
-
-
-	PUSH_L ( EDI )
-	PUSH_L ( ESI )
-
-    	MOV_L     ( REGOFF(12, ESP), EAX ) /* dest */
-    	MOV_L     ( REGOFF(16, ESP), ESI ) /* mat */
-    	MOV_L     ( REGOFF(20, ESP), EDX ) /* src */
-    	MOV_L     ( REGOFF(24, ESP), EDI ) /* src_stride */
-    	MOV_L     ( REGOFF(28, ESP), ECX ) /* count */
-
-	FEMMS   
-   
-	MOVQ      ( REGOFF(MAT_TX, ESI), MM7 )  /* m31         | m30     */
-	MOVQ	  ( REGOFF(MAT_TZ, ESI), MM3 )  /* m33         | m32   */
-   
-ALIGNTEXT32   
-LLBL(v16_3dnow_general_loop):   
-        PREFETCH  ( REGOFF(128, EAX) )       /* prefetch 2 output verts ahead*/
-        PREFETCH  ( REGOFF(32, EDX) )        /*  */
-
-	MOVQ	  ( REGIND(EDX),  MM0 )	     /* x1          | x0   */
-	MOVD	  ( REGOFF(8, EDX), MM1 )    /*             | x2   */
-	MOVQ	  ( REGIND(ESI),  MM4 )	     /* m00         | m01   */
-	PUNPCKHDQ ( MM0, MM2 )	             /* x1          |       */
-	MOVQ	  ( REGOFF(16, ESI),MM5 )    /* m10         | m11   */
-	PUNPCKLDQ ( MM0, MM0 )	             /* x0          | x0   */
-	MOVQ	  ( REGOFF(32, ESI),MM6 )    /* m20         | m21   */
-	PFMUL	  ( MM0, MM4 )	             /* x0.m01      | x0.m00   */
-	PUNPCKHDQ ( MM2, MM2 )	             /* x1          | x1   */
-	PFMUL	  ( MM2, MM5 )	             /* x1.m11      | x1.m10   */
-	PUNPCKLDQ ( MM1, MM1 )	             /* x2          | x2   */
-	PFMUL	  ( REGOFF(8,ESI), MM0 )     /* x0.m03      | x0.m02   */
-	PFMUL	  ( REGOFF(24,ESI),MM2 )     /* x1.m13      | x1.m12   */
-	PFMUL	  ( MM1, MM6 )	             /* x2.m21      | x2.m20   */
-	PFADD	  ( MM4, MM5 )	             /* x0m01+x1m11 | x0m00+x1m10   */
-	PFMUL	  ( REGOFF(40,ESI),MM1 )     /* x2.m21      | x2.m20   */
-	PFADD	  ( MM0, MM2 )	             /* x0m03+x1m13 | x0m02+x1m12   */
-	PFADD	  ( MM5, MM6 )	             /* (xm012)1    | (xm012)0   */
-	PFADD	  ( MM1, MM2 )	             /* (xm012)3    | (xm012)2   */
-	PFADD	  ( MM7, MM6 )	             /* r1          | r0   */
-	PFADD	  ( MM3, MM2 )	             /* r3          | r2   */
-	ADD_L	  ( EDI, EDX )	             /* next input     */
-	MOVQ	  ( MM6, REGIND(EAX) )       /* */
-	MOVQ	  ( MM2, REGOFF(8,EAX) )     /* */
-	ADD_L	  ( CONST(64), EAX )         /* next output   */
-        DEC_L	  ( ECX )   
-	JNE	  ( LLBL(v16_3dnow_general_loop) )
-
-	FEMMS   
-
-        POP_L     ( ESI )
-        POP_L     ( EDI )
-	RET
-
-
-
-
-
-
-
-
diff --git a/xc/extras/Mesa/src/X86/x86.c b/xc/extras/Mesa/src/X86/x86.c
index 0e531b8df..37abf644b 100644
--- a/xc/extras/Mesa/src/X86/x86.c
+++ b/xc/extras/Mesa/src/X86/x86.c
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -22,13 +22,11 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/X86/x86.c,v 1.6 2000/09/26 15:56:40 tsi Exp $ */
 
 /*
  * Intel x86 assembly code by Josh Vanderhoof
  */
 
-
 #include "glheader.h"
 #include "context.h"
 #include "types.h"
@@ -36,89 +34,104 @@
 #include "xform.h"
 #include "x86.h"
 
-#ifdef USE_X86_ASM
-extern void _ASMAPI gl_v16_x86_cliptest_points4( GLfloat *first_vert,
-					GLfloat *last_vert,
-					GLubyte *or_mask,
-					GLubyte *and_mask,
-					GLubyte *clip_mask );
-
-  
-extern void _ASMAPI gl_v16_x86_general_xform( GLfloat *dest,
-				     const GLfloat *m,
-				     const GLfloat *src,
-				     GLuint src_stride,
-					      GLuint count );
+#ifdef DEBUG
+#include "debug_xform.h"
 #endif
 
 
-#define XFORM_ARGS 	GLvector4f *to_vec, 		\
-			const GLfloat m[16], 		\
-			const GLvector4f *from_vec, 	\
-			const GLubyte *mask, 		\
+#define XFORM_ARGS 	GLvector4f *to_vec, 				\
+			const GLfloat m[16], 				\
+			const GLvector4f *from_vec, 			\
+			const GLubyte *mask, 				\
 			const GLubyte flag
 
-#define DECLARE_XFORM_GROUP(pfx, vsize, masked) \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_general_##masked(XFORM_ARGS);     \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_identity_##masked(XFORM_ARGS);    \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_3d_no_rot_##masked(XFORM_ARGS);   \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_perspective_##masked(XFORM_ARGS);     \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_2d_##masked(XFORM_ARGS);      \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_2d_no_rot_##masked(XFORM_ARGS);   \
- extern void _ASMAPI gl_##pfx##_transform_points##vsize##_3d_##masked(XFORM_ARGS);
-
-#define ASSIGN_XFORM_GROUP( pfx, cma, vsize, masked ) 		\
- gl_transform_tab[cma][vsize][MATRIX_GENERAL] 			\
-  = gl_##pfx##_transform_points##vsize##_general_##masked; 	\
- gl_transform_tab[cma][vsize][MATRIX_IDENTITY] 			\
-  = gl_##pfx##_transform_points##vsize##_identity_##masked; 	\
- gl_transform_tab[cma][vsize][MATRIX_3D_NO_ROT] 		\
-  = gl_##pfx##_transform_points##vsize##_3d_no_rot_##masked; 	\
- gl_transform_tab[cma][vsize][MATRIX_PERSPECTIVE] 		\
-  = gl_##pfx##_transform_points##vsize##_perspective_##masked; 	\
- gl_transform_tab[cma][vsize][MATRIX_2D] 			\
-  = gl_##pfx##_transform_points##vsize##_2d_##masked; 		\
- gl_transform_tab[cma][vsize][MATRIX_2D_NO_ROT] 		\
-  = gl_##pfx##_transform_points##vsize##_2d_no_rot_##masked; 	\
- gl_transform_tab[cma][vsize][MATRIX_3D] 			\
-  = gl_##pfx##_transform_points##vsize##_3d_##masked;
+
+#define DECLARE_XFORM_GROUP( pfx, sz, masked ) \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_general_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_identity_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_3d_no_rot_##masked( XFORM_ARGS );    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_perspective_##masked( XFORM_ARGS );  \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_2d_##masked( XFORM_ARGS );	    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_2d_no_rot_##masked( XFORM_ARGS );    \
+ extern void _ASMAPI gl_##pfx##_transform_points##sz##_3d_##masked( XFORM_ARGS );
+
+
+#define ASSIGN_XFORM_GROUP( pfx, cma, sz, masked )			\
+   gl_transform_tab[cma][sz][MATRIX_GENERAL] =				\
+      gl_##pfx##_transform_points##sz##_general_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_IDENTITY] =				\
+      gl_##pfx##_transform_points##sz##_identity_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_3D_NO_ROT] =			\
+      gl_##pfx##_transform_points##sz##_3d_no_rot_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_PERSPECTIVE] =			\
+      gl_##pfx##_transform_points##sz##_perspective_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_2D] =				\
+      gl_##pfx##_transform_points##sz##_2d_##masked;			\
+   gl_transform_tab[cma][sz][MATRIX_2D_NO_ROT] =			\
+      gl_##pfx##_transform_points##sz##_2d_no_rot_##masked;		\
+   gl_transform_tab[cma][sz][MATRIX_3D] =				\
+      gl_##pfx##_transform_points##sz##_3d_##masked;
 
 
 #ifdef USE_X86_ASM
-   DECLARE_XFORM_GROUP( x86, 2, raw )
-   DECLARE_XFORM_GROUP( x86, 3, raw )
-   DECLARE_XFORM_GROUP( x86, 4, raw )
-   DECLARE_XFORM_GROUP( x86, 2, masked )
-   DECLARE_XFORM_GROUP( x86, 3, masked )
-   DECLARE_XFORM_GROUP( x86, 4, masked )
-
-   extern GLvector4f  * _ASMAPI gl_x86_cliptest_points4( GLvector4f *clip_vec, 
-                                               GLvector4f *proj_vec, 
-                                               GLubyte clipMask[],
-                                               GLubyte *orMask, 
-                                               GLubyte *andMask );
+DECLARE_XFORM_GROUP( x86, 2, raw )
+DECLARE_XFORM_GROUP( x86, 3, raw )
+DECLARE_XFORM_GROUP( x86, 4, raw )
+DECLARE_XFORM_GROUP( x86, 2, masked )
+DECLARE_XFORM_GROUP( x86, 3, masked )
+DECLARE_XFORM_GROUP( x86, 4, masked )
+
+
+extern GLvector4f * _ASMAPI gl_x86_cliptest_points4( GLvector4f *clip_vec,
+						     GLvector4f *proj_vec,
+						     GLubyte clipMask[],
+						     GLubyte *orMask,
+						     GLubyte *andMask );
+
+
+extern void _ASMAPI gl_v16_x86_cliptest_points4( GLfloat *first_vert,
+						 GLfloat *last_vert,
+						 GLubyte *or_mask,
+						 GLubyte *and_mask,
+						 GLubyte *clip_mask );
+
+
+extern void _ASMAPI gl_v16_x86_general_xform( GLfloat *dest,
+					      const GLfloat *m,
+					      const GLfloat *src,
+					      GLuint src_stride,
+					      GLuint count );
 #endif
 
 
-void gl_init_x86_asm_transforms( void )
+void gl_init_x86_transform_asm( void )
 {
 #ifdef USE_X86_ASM
-   ASSIGN_XFORM_GROUP( x86, 0, 2, raw )
-   ASSIGN_XFORM_GROUP( x86, 0, 3, raw )
-   ASSIGN_XFORM_GROUP( x86, 0, 4, raw )
+   ASSIGN_XFORM_GROUP( x86, 0, 2, raw );
+   ASSIGN_XFORM_GROUP( x86, 0, 3, raw );
+   ASSIGN_XFORM_GROUP( x86, 0, 4, raw );
 
-   ASSIGN_XFORM_GROUP( x86, CULL_MASK_ACTIVE, 2, masked )
-   ASSIGN_XFORM_GROUP( x86, CULL_MASK_ACTIVE, 3, masked )
-   ASSIGN_XFORM_GROUP( x86, CULL_MASK_ACTIVE, 4, masked )
+   ASSIGN_XFORM_GROUP( x86, CULL_MASK_ACTIVE, 2, masked );
+   ASSIGN_XFORM_GROUP( x86, CULL_MASK_ACTIVE, 3, masked );
+   ASSIGN_XFORM_GROUP( x86, CULL_MASK_ACTIVE, 4, masked );
 
    /* XXX this function has been found to cause FP overflow exceptions */
    gl_clip_tab[4] = gl_x86_cliptest_points4;
 
 #ifdef DEBUG
-   gl_test_all_transform_functions("x86");
+   gl_test_all_transform_functions( "x86" );
+#endif
 #endif
+}
 
-   gl_cliptest_points4_v16 = gl_v16_x86_cliptest_points4;
-   gl_xform_points3_v16_general = gl_v16_x86_general_xform;
+void gl_init_x86_vertex_asm( void )
+{
+#ifdef USE_X86_ASM
+   gl_xform_points3_v16_general	= gl_v16_x86_general_xform;
+   gl_cliptest_points4_v16	= gl_v16_x86_cliptest_points4;
+
+#if 0
+   gl_test_all_vertex_functions( "x86" );
+#endif
 #endif
 }
diff --git a/xc/extras/Mesa/src/X86/x86.h b/xc/extras/Mesa/src/X86/x86.h
index aaad25b2c..76b2ff566 100644
--- a/xc/extras/Mesa/src/X86/x86.h
+++ b/xc/extras/Mesa/src/X86/x86.h
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
- * 
+ * Version:  3.4
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -27,10 +27,10 @@
  * Intel x86 assembly code by Josh Vanderhoof
  */
 
+#ifndef __X86_H__
+#define __X86_H__
 
-#ifndef X86_H
-#define X86_H
-
-extern void gl_init_x86_asm_transforms(void);
+extern void gl_init_x86_transform_asm( void );
+extern void gl_init_x86_vertex_asm( void );
 
 #endif
diff --git a/xc/extras/Mesa/src/X86/x86_cliptest.S b/xc/extras/Mesa/src/X86/x86_cliptest.S
new file mode 100644
index 000000000..992fec3d7
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_cliptest.S
@@ -0,0 +1,247 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "clip_args.h"
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+/*
+ * Table for clip test.
+ *
+ * 	bit6 = SRC(3) < 0
+ * 	bit5 = SRC(2) < 0
+ * 	bit4 = abs(S(2)) > abs(S(3))
+ * 	bit3 = SRC(1) < 0
+ * 	bit2 = abs(S(1)) > abs(S(3))
+ * 	bit1 = SRC(0) < 0
+ * 	bit0 = abs(S(0)) > abs(S(3))
+ */
+
+	SEG_DATA
+
+clip_table:
+	D_BYTE 0, 1, 0, 2, 4, 5, 4, 6
+	D_BYTE 0, 1, 0, 2, 8, 9, 8, 10
+	D_BYTE 32, 33, 32, 34, 36, 37, 36, 38
+	D_BYTE 32, 33, 32, 34, 40, 41, 40, 42
+	D_BYTE 0, 1, 0, 2, 4, 5, 4, 6
+	D_BYTE 0, 1, 0, 2, 8, 9, 8, 10
+	D_BYTE 16, 17, 16, 18, 20, 21, 20, 22
+	D_BYTE 16, 17, 16, 18, 24, 25, 24, 26
+	D_BYTE 63, 61, 63, 62, 55, 53, 55, 54
+	D_BYTE 63, 61, 63, 62, 59, 57, 59, 58
+	D_BYTE 47, 45, 47, 46, 39, 37, 39, 38
+	D_BYTE 47, 45, 47, 46, 43, 41, 43, 42
+	D_BYTE 63, 61, 63, 62, 55, 53, 55, 54
+	D_BYTE 63, 61, 63, 62, 59, 57, 59, 58
+	D_BYTE 31, 29, 31, 30, 23, 21, 23, 22
+	D_BYTE 31, 29, 31, 30, 27, 25, 27, 26
+
+
+	SEG_TEXT
+
+/*
+ * gl_x86_cliptest_points4
+ *
+ *   AL:  ormask
+ *   AH:  andmask
+ *   EBX: temp0
+ *   ECX: temp1
+ *   EDX: clipmask[]
+ *   ESI: clip[]
+ *   EDI: proj[]
+ *   EBP: temp2
+ */
+
+#if defined(__ELF__) && defined(__PIC__) && !defined(ELFPIC)
+#define ELFPIC
+#endif
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_cliptest_points4 )
+GLNAME( gl_x86_cliptest_points4 ):
+
+#ifdef ELFPIC
+#define FRAME_OFFSET 20
+#else
+#define FRAME_OFFSET 16
+#endif
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBP )
+	PUSH_L( EBX )
+
+#ifdef ELFPIC
+	/* store pointer to clip_table on stack */
+	CALL( LLBL( ctp4_get_eip ) )
+	ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX )
+	MOV_L( REGOFF(clip_table@GOT, EBX), EBX )
+	PUSH_L( EBX )
+	JMP( LLBL( ctp4_clip_table_ready ) )
+
+LLBL( ctp4_get_eip ):
+	/* store eip in ebx */
+	MOV_L( REGIND(ESP), EBX )
+	RET
+
+LLBL( ctp4_clip_table_ready ):
+#endif
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_CLIP, EDX )
+	MOV_L( ARG_OR, EBX )
+
+	MOV_L( ARG_AND, EBP )
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+	MOV_L( EAX, ARG_SOURCE )	/* put stride in ARG_SOURCE */
+
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDX, ECX )
+
+	MOV_L( ECX, ARG_CLIP )		/* put clipmask + count in ARG_CLIP */
+	CMP_L( ECX, EDX )
+
+	MOV_B( REGIND(EBX), AL )
+	MOV_B( REGIND(EBP), AH )
+
+	JZ( LLBL( ctp4_finish ) )
+
+ALIGNTEXT16
+LLBL( ctp4_top ):
+
+	FLD1			/* F3 */
+	FDIV_S( SRC(3) )
+
+	MOV_L( SRC(3), EBP )
+	MOV_L( SRC(2), EBX )
+
+	XOR_L( ECX, ECX )
+	ADD_L( EBP, EBP )	/* ebp = abs(S(3))*2 ; carry = sign of S(3) */
+
+	ADC_L( ECX, ECX )
+	ADD_L( EBX, EBX )	/* ebx = abs(S(2))*2 ; carry = sign of S(2) */
+
+	ADC_L( ECX, ECX )
+	CMP_L( EBX, EBP )	/* carry = abs(S(2))*2 > abs(S(3))*2 */
+
+	ADC_L( ECX, ECX )
+	MOV_L( SRC(1), EBX )
+
+	ADD_L( EBX, EBX )	/* ebx = abs(S(1))*2 ; carry = sign of S(1) */
+
+	ADC_L( ECX, ECX )
+	CMP_L( EBX, EBP )	/* carry = abs(S(1))*2 > abs(S(3))*2 */
+
+	ADC_L( ECX, ECX )
+	MOV_L( SRC(0), EBX )
+
+	ADD_L( EBX, EBX )	/* ebx = abs(S(0))*2 ; carry = sign of S(0) */
+
+	ADC_L( ECX, ECX )
+	CMP_L( EBX, EBP )	/* carry = abs(S(0))*2 > abs(S(3))*2 */
+
+	ADC_L( ECX, ECX )
+
+#ifdef ELFPIC
+	MOV_L( REGIND(ESP), EBP )	/* clip_table */
+
+	MOV_B( REGBI(EBP, ECX), CL )
+#else
+	MOV_B( REGOFF(clip_table,ECX), CL )
+#endif
+
+	OR_B( CL, AL )
+	AND_B( CL, AH )
+
+	TEST_B( CL, CL )
+	MOV_B( CL, REGIND(EDX) )
+
+	JZ( LLBL( ctp4_proj ) )
+
+	FSTP( ST(0) )			/* */
+	JMP( LLBL( ctp4_next ) )
+
+LLBL( ctp4_proj ):
+
+	FLD_S( SRC(0) )			/* F0 F3 */
+	FMUL2( ST(1), ST(0) )
+
+	FLD_S( SRC(1) )			/* F1 F0 F3 */
+	FMUL2( ST(2), ST(0) )
+
+	FLD_S( SRC(2) )			/* F2 F1 F0 F3 */
+	FMUL2( ST(3), ST(0) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F3 */
+	FSTP_S( DST(0) )		/* F1 F2 F3 */
+	FSTP_S( DST(1) )		/* F2 F3 */
+	FSTP_S( DST(2) )		/* F3 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( ctp4_next ):
+
+	INC_L( EDX )
+	ADD_L( CONST(16), EDI )
+
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( EDX, ARG_CLIP )
+
+	JNZ( LLBL( ctp4_top ) )
+
+	MOV_L( ARG_OR, ECX )
+	MOV_L( ARG_AND, EDX )
+
+	MOV_B( AL, REGIND(ECX) )
+	MOV_B( AH, REGIND(EDX) )
+
+LLBL( ctp4_finish ):
+
+	MOV_L( ARG_DEST, EAX )
+#ifdef ELFPIC
+	POP_L( ESI )			/* discard ptr to clip_table */
+#endif
+	POP_L( EBX )
+	POP_L( EBP )
+	POP_L( EDI )
+	POP_L( ESI )
+
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86_vertex.S b/xc/extras/Mesa/src/X86/x86_vertex.S
new file mode 100644
index 000000000..dbe2cfbe4
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_vertex.S
@@ -0,0 +1,298 @@
+#include "assyntax.h"
+
+
+	SEG_TEXT
+
+/*#define MAT_SX 0*/        /*  accessed by REGIND !! */
+#define MAT_SY 20
+#define MAT_SZ 40
+#define MAT_TX 48
+#define MAT_TY 52
+#define MAT_TZ 56
+
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define S(i)		REGOFF(i * 4, ESI)
+#define D(i)		REGOFF(i * 4, EDI)
+#define M(i)		REGOFF(i * 4, EDX)
+
+
+
+/*
+ * void gl_v16_x86_general_xform (	GLfloat *dest,
+ *				const GLfloat *m,
+ *				const GLfloat *src,
+ *                              GLuint src_stride,
+ *				GLuint count )
+ */
+
+
+/* This is nothing more glamorous than an objdump of one of Josh's
+ * routines hacked to match the above.
+ */
+ALIGNTEXT16
+GLOBL GLNAME( gl_v16_x86_general_xform )
+GLNAME( gl_v16_x86_general_xform ):
+
+
+	PUSH_L ( EDI )
+	PUSH_L ( ESI )
+
+    	MOV_L     ( REGOFF(12, ESP), EAX ) /* dest */
+    	MOV_L     ( REGOFF(16, ESP), ESI ) /* mat */
+    	MOV_L     ( REGOFF(20, ESP), EDX ) /* src */
+    	MOV_L     ( REGOFF(24, ESP), EDI ) /* src_stride */
+    	MOV_L     ( REGOFF(28, ESP), ECX ) /* count */
+
+
+LLBL(v16x86_loop):
+        FLD_S   ( REGOFF( 0x0, EDX ) )
+        FMUL_S  ( REGOFF( 0x0, ESI ) )
+        FLD_S   ( REGOFF( 0x0, EDX ) )
+        FMUL_S  ( REGOFF( 0x4, ESI ) )
+        FLD_S   ( REGOFF( 0x0, EDX ) )
+        FMUL_S  ( REGOFF( 0x8, ESI ) )
+        FLD_S   ( REGOFF( 0x0, EDX ) )
+        FMUL_S  ( REGOFF( 0xc, ESI ) )
+        FLD_S   ( REGOFF( 0x4, EDX ) )
+        FMUL_S  ( REGOFF( 0x10, ESI ) )
+        FLD_S   ( REGOFF( 0x4, EDX ) )
+        FMUL_S  ( REGOFF( 0x14, ESI ) )
+        FLD_S   ( REGOFF( 0x4, EDX ) )
+        FMUL_S  ( REGOFF( 0x18, ESI ) )
+        FLD_S   ( REGOFF( 0x4, EDX ) )
+        FMUL_S  ( REGOFF( 0x1c, ESI ) )
+        FXCH    ( ST(3) )
+        FADDP   ( ST(0),ST(7) )
+        FXCH    ( ST(1) )
+        FADDP   ( ST(0),ST(5) )
+        FADDP   ( ST(0),ST(3) )
+        FADDP   ( ST(0),ST(1) )
+        FLD_S   ( REGOFF( 0x8, EDX ) )
+        FMUL_S  ( REGOFF( 0x20, ESI ) )
+        FLD_S   ( REGOFF( 0x8, EDX ) )
+        FMUL_S  ( REGOFF( 0x24, ESI ) )
+        FLD_S   ( REGOFF( 0x8, EDX ) )
+        FMUL_S  ( REGOFF( 0x28, ESI ) )
+        FLD_S   ( REGOFF( 0x8, EDX ) )
+        FMUL_S  ( REGOFF( 0x2c, ESI ) )
+        FXCH    ( ST(3) )
+        FADDP   ( ST(0),ST(7) )
+        FXCH    ( ST(1) )
+        FADDP   ( ST(0),ST(5) )
+        FADDP   ( ST(0),ST(3) )
+        FADDP   ( ST(0),ST(1) )
+        FXCH    ( ST(3) )
+        FADD_S  ( REGOFF( 0x30, ESI ) )
+        FXCH    ( ST(2) )
+        FADD_S  ( REGOFF( 0x34, ESI ) )
+        FXCH    ( ST(1) )
+        FADD_S  ( REGOFF( 0x38, ESI ) )
+        FXCH    ( ST(3) )
+        FADD_S  ( REGOFF( 0x3c, ESI ) )
+        FXCH    ( ST(2) )
+        FSTP_S  ( REGOFF( 0x0, EAX ) )
+        FSTP_S  ( REGOFF( 0x4, EAX ) )
+        FXCH    ( ST(1) )
+        FSTP_S  ( REGOFF( 0x8, EAX ) )
+        FSTP_S  ( REGOFF( 0xc, EAX ) )
+        ADD_L   ( CONST(64), EAX )
+        ADD_L   ( EDI, EDX )
+        DEC_L   ( ECX )
+        JNE     ( LLBL(v16x86_loop) )
+
+	POP_L ( ESI )
+	POP_L ( EDI )
+	RET
+
+
+/*
+ * Table for clip test.
+ *
+ * 	bit6 = S(3) < 0
+ * 	bit5 = S(2) < 0
+ * 	bit4 = abs(S(2)) > abs(S(3))
+ * 	bit3 = S(1) < 0
+ * 	bit2 = abs(S(1)) > abs(S(3))
+ * 	bit1 = S(0) < 0
+ * 	bit0 = abs(S(0)) > abs(S(3))
+ */
+
+
+	SEG_DATA
+
+clip_table:
+	D_BYTE 0, 1, 0, 2, 4, 5, 4, 6
+	D_BYTE 0, 1, 0, 2, 8, 9, 8, 10
+	D_BYTE 32, 33, 32, 34, 36, 37, 36, 38
+	D_BYTE 32, 33, 32, 34, 40, 41, 40, 42
+	D_BYTE 0, 1, 0, 2, 4, 5, 4, 6
+	D_BYTE 0, 1, 0, 2, 8, 9, 8, 10
+	D_BYTE 16, 17, 16, 18, 20, 21, 20, 22
+	D_BYTE 16, 17, 16, 18, 24, 25, 24, 26
+	D_BYTE 63, 61, 63, 62, 55, 53, 55, 54
+	D_BYTE 63, 61, 63, 62, 59, 57, 59, 58
+	D_BYTE 47, 45, 47, 46, 39, 37, 39, 38
+	D_BYTE 47, 45, 47, 46, 43, 41, 43, 42
+	D_BYTE 63, 61, 63, 62, 55, 53, 55, 54
+	D_BYTE 63, 61, 63, 62, 59, 57, 59, 58
+	D_BYTE 31, 29, 31, 30, 23, 21, 23, 22
+	D_BYTE 31, 29, 31, 30, 27, 25, 27, 26
+
+	SEG_TEXT
+
+
+/*
+########################################
+##
+## gl_v16_x86_cliptest_points4
+##
+## Performs cliptesting equivalent to that done by cliptest_v16()
+## in vertices.c
+##
+## This is a hacked version of the original above.
+##
+########################################
+*/
+
+
+
+#define OFFSET_V16_SOURCE	4
+#define OFFSET_V16_LAST		8
+#define OFFSET_V16_OR		12
+#define OFFSET_V16_AND		16
+#define OFFSET_V16_MASK		20
+
+#define ARG_V16_SOURCE	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_SOURCE, ESP)
+#define ARG_V16_LAST	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_LAST, ESP)
+#define ARG_V16_OR	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_OR, ESP)
+#define ARG_V16_AND	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_AND, ESP)
+#define ARG_V16_MASK	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_MASK, ESP)
+
+
+#if defined(__ELF__) && defined(__PIC__) && !defined(ELFPIC)
+#define ELFPIC
+#endif
+
+	GLOBL GLNAME(gl_v16_x86_cliptest_points4)
+	ALIGNTEXT4
+
+GLNAME(gl_v16_x86_cliptest_points4):
+#ifdef ELFPIC
+#define V16_FRAME_OFFSET 20
+#else
+#define V16_FRAME_OFFSET 16
+#endif
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBP )
+	PUSH_L( EBX )
+
+#ifdef ELFPIC
+	/* store pointer to clip_table on stack */
+	CALL( LLBL(v16_ctp4_get_eip) )
+	ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX )
+	MOV_L( REGOFF(clip_table@GOT, EBX), EBX )
+	PUSH_L( EBX )
+	JMP( LLBL(v16_ctp4_clip_table_ready) )
+
+LLBL(v16_ctp4_get_eip):
+	/* store eip in ebx */
+	MOV_L( REGIND(ESP), EBX )
+	RET
+
+LLBL(v16_ctp4_clip_table_ready):
+#endif
+
+	MOV_L( ARG_V16_SOURCE, ESI )            /* ptr to first source vertex */
+	MOV_L( ARG_V16_LAST, EDX ) 	        /* ptr to last source vertex */
+	MOV_L( ARG_V16_OR, EBX )
+	MOV_L( ARG_V16_AND, EBP )
+	MOV_L( ARG_V16_MASK, EDI )
+
+	CMP_L( EDX, ESI )
+
+	MOV_B( REGIND(EBX), AL )
+	MOV_B( REGIND(EBP), AH )
+
+	JZ( LLBL(v16_ctp4_finish) )
+
+	ALIGNTEXT4ifNOP
+LLBL(v16_ctp4_top):
+#if 0
+	FLD1 	/* F0 */
+	FDIV_S( S(3) )
+#endif
+
+	MOV_L( S(3), EBP )
+	MOV_L( S(2), EBX )
+
+	XOR_L( ECX, ECX )
+	ADD_L( EBP, EBP )	/* ebp = abs(S(3))*2 ; carry = sign of S(3) */
+
+	ADC_L( ECX, ECX )
+	ADD_L( EBX, EBX )	/* ebx = abs(S(2))*2 ; carry = sign of S(2) */
+
+	ADC_L( ECX, ECX )
+	CMP_L( EBX, EBP )	/* carry = abs(S(2))*2 > abs(S(3))*2 */
+
+	ADC_L( ECX, ECX )
+	MOV_L( S(1), EBX )
+
+	ADD_L( EBX, EBX )	/* ebx = abs(S(1))*2 ; carry = sign of S(1) */
+
+	ADC_L( ECX, ECX )
+	CMP_L( EBX, EBP )	/* carry = abs(S(1))*2 > abs(S(3))*2 */
+
+	ADC_L( ECX, ECX )
+	MOV_L( S(0), EBX )
+
+	ADD_L( EBX, EBX )	/* ebx = abs(S(0))*2 ; carry = sign of S(0) */
+
+	ADC_L( ECX, ECX )
+	CMP_L( EBX, EBP )	/* carry = abs(S(0))*2 > abs(S(3))*2 */
+
+	ADC_L( ECX, ECX )
+
+#ifdef ELFPIC
+	MOV_L( REGIND(ESP), EBP ) 	/* clip_table */
+
+	MOV_B( REGBI(EBP, ECX), CL )
+#else
+	MOV_B( REGOFF(clip_table,ECX), CL )
+#endif
+
+	OR_B( CL, AL )
+	AND_B( CL, AH )
+
+	MOV_B( CL, REGIND(EDI) )     /* save clipmask */
+	INC_L( EDI )                 /* next clipmask */
+
+#if 0
+	FSTP_S( S(8)   ) 	/* */ 	/* GR_VERTEX_OOW_OFFSET */
+#endif
+
+	ADD_L( CONST(64), ESI )      /* next fxVertex  */
+
+	CMP_L( EDX, ESI )            /* finished? */
+	JNZ( LLBL(v16_ctp4_top) )
+
+	MOV_L( ARG_V16_OR, ECX )
+	MOV_L( ARG_V16_AND, EDX )
+
+	MOV_B( AL, REGIND(ECX) )
+	MOV_B( AH, REGIND(EDX) )
+
+LLBL(v16_ctp4_finish):
+
+#ifdef ELFPIC
+	POP_L( ESI ) 	/* discard ptr to clip_table */
+#endif
+	POP_L( EBX )
+	POP_L( EBP )
+	POP_L( EDI )
+	POP_L( ESI )
+
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86_xform_masked2.S b/xc/extras/Mesa/src/X86/x86_xform_masked2.S
new file mode 100644
index 000000000..1b8240bfb
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_xform_masked2.S
@@ -0,0 +1,610 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "xform_args.h"
+
+	SEG_TEXT
+
+#define FRAME_OFFSET	16
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_general_masked )
+GLNAME( gl_x86_transform_points2_general_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_gm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_gm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_gm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+	FLD_S( SRC(0) )			/* F7 F6 F5 F4 */
+	FMUL_S( MAT(3) )
+
+	FLD_S( SRC(1) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+	FLD_S( SRC(1) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(7) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FXCH( ST(3) )			/* F4 F6 F5 F7 */
+	FADD_S( MAT(12) )
+	FXCH( ST(2) )			/* F5 F6 F4 F7 */
+	FADD_S( MAT(13) )
+	FXCH( ST(1) )			/* F6 F5 F4 F7 */
+	FADD_S( MAT(14) )
+	FXCH( ST(3) )			/* F7 F5 F4 F6 */
+	FADD_S( MAT(15) )
+
+	FXCH( ST(2) )			/* F4 F5 F7 F6 */
+	FSTP_S( DST(0) )		/* F5 F7 F6 */
+	FSTP_S( DST(1) )		/* F7 F6 */
+	FXCH( ST(1) )			/* F6 F7 */
+	FSTP_S( DST(2) )		/* F7 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( x86_p2_gm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_gm_loop ) )
+
+LLBL( x86_p2_gm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_perspective_masked )
+GLNAME( gl_x86_transform_points2_perspective_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_pm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+	MOV_L( MAT(14), EAX )
+
+ALIGNTEXT16
+LLBL( x86_p2_pm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_pm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FSTP_S( DST(0)   )		/* F1 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EAX, DST(2) )
+	MOV_L( CONST(FP_ZERO), DST(3) )
+
+LLBL( x86_p2_pm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_pm_loop ) )
+
+LLBL( x86_p2_pm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_3d_masked )
+GLNAME( gl_x86_transform_points2_3d_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_3dm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_3dm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_3dm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+
+	FLD_S( SRC(1) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 F6 */
+	FADD_S( MAT(13) )
+	FXCH( ST(2) )			/* F6 F4 F5 */
+	FADD_S( MAT(14) )
+
+	FXCH( ST(1) )			/* F4 F6 F5 */
+	FSTP_S( DST(0) )		/* F6 F5 */
+	FXCH( ST(1) )			/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+
+LLBL( x86_p2_3dm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_3dm_loop ) )
+
+LLBL( x86_p2_3dm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_3d_no_rot_masked )
+GLNAME( gl_x86_transform_points2_3d_no_rot_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_3dnrm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+	MOV_L( MAT(14), EAX )
+
+ALIGNTEXT16
+LLBL( x86_p2_3dnrm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_3dnrm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 */
+	FXCH( ST(2) )			/* F1 F4 F5 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 */
+
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EAX, DST(2) )
+
+LLBL( x86_p2_3dnrm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_3dnrm_loop ) )
+
+LLBL( x86_p2_3dnrm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_2d_masked )
+GLNAME( gl_x86_transform_points2_2d_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_2dm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_2dm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_2dm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+
+	FLD_S( SRC(1) )			/* F0 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 */
+	FADD_S( MAT(13) )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+
+LLBL( x86_p2_2dm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_2dm_loop ) )
+
+LLBL( x86_p2_2dm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_2d_no_rot_masked )
+GLNAME( gl_x86_transform_points2_2d_no_rot_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_2dnrm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_2dnrm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_2dnrm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 */
+	FXCH( ST(2) )			/* F1 F4 F5 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 */
+
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+
+LLBL( x86_p2_2dnrm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_2dnrm_loop ) )
+
+LLBL( x86_p2_2dnrm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_identity_masked )
+GLNAME( gl_x86_transform_points2_identity_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_im_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+	CMP_L( ESI, EDI )
+	JE( LLBL( x86_p2_im_done ) )
+
+ALIGNTEXT16
+LLBL( x86_p2_im_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p2_im_skip ) )
+
+	MOV_L( SRC(0), EAX )
+	MOV_L( SRC(1), EDX )
+
+	MOV_L( EAX, DST(0) )
+	MOV_L( EDX, DST(1) )
+
+LLBL( x86_p2_im_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_im_loop ) )
+
+LLBL( x86_p2_im_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86_xform_masked3.S b/xc/extras/Mesa/src/X86/x86_xform_masked3.S
new file mode 100644
index 000000000..54bed7e36
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_xform_masked3.S
@@ -0,0 +1,670 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "xform_args.h"
+
+	SEG_TEXT
+
+#define FRAME_OFFSET	16
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_general_masked )
+GLNAME( gl_x86_transform_points3_general_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_gm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_gm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_gm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+	FLD_S( SRC(0) )			/* F7 F6 F5 F4 */
+	FMUL_S( MAT(3) )
+
+	FLD_S( SRC(1) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+	FLD_S( SRC(1) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(7) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+	FLD_S( SRC(2) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(11) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FXCH( ST(3) )			/* F4 F6 F5 F7 */
+	FADD_S( MAT(12) )
+	FXCH( ST(2) )			/* F5 F6 F4 F7 */
+	FADD_S( MAT(13) )
+	FXCH( ST(1) )			/* F6 F5 F4 F7 */
+	FADD_S( MAT(14) )
+	FXCH( ST(3) )			/* F7 F5 F4 F6 */
+	FADD_S( MAT(15) )
+
+	FXCH( ST(2) )			/* F4 F5 F7 F6 */
+	FSTP_S( DST(0) )		/* F5 F7 F6 */
+	FSTP_S( DST(1) )		/* F7 F6 */
+	FXCH( ST(1) )			/* F6 F7 */
+	FSTP_S( DST(2) )		/* F7 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( x86_p3_gm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_gm_loop ) )
+
+LLBL( x86_p3_gm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_perspective_masked )
+GLNAME( gl_x86_transform_points3_perspective_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_pm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+ALIGNTEXT16
+LLBL( x86_p3_pm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_pm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F0 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F5 F4 */
+	FADDP( ST(0), ST(4) )		/* F1 F2 F5 F4 */
+	FADDP( ST(0), ST(2) )		/* F2 F5 F4 */
+	FLD_S( MAT(14) )		/* F6 F2 F5 F4 */
+	FXCH( ST(1) )			/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(2), EAX )
+	XOR_L( CONST(-2147483648), EAX )/* change sign */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0)   )		/* F5 F6 */
+	FSTP_S( DST(1)   )		/* F6 */
+	FSTP_S( DST(2)   )		/* */
+	MOV_L( EAX, DST(3) )
+
+LLBL( x86_p3_pm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_pm_loop ) )
+
+LLBL( x86_p3_pm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_3d_masked )
+GLNAME( gl_x86_transform_points3_3d_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_3dm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_3dm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_3dm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+
+	FLD_S( SRC(1) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 F6 */
+	FADD_S( MAT(13) )
+	FXCH( ST(2) )			/* F6 F4 F5 */
+	FADD_S( MAT(14) )
+
+	FXCH( ST(1) )			/* F4 F6 F5 */
+	FSTP_S( DST(0) )		/* F6 F5 */
+	FXCH( ST(1) )			/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+
+LLBL( x86_p3_3dm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_3dm_loop ) )
+
+LLBL( x86_p3_3dm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_3d_no_rot_masked )
+GLNAME( gl_x86_transform_points3_3d_no_rot_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_3dnrm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_3dnrm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_3dnrm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F2 F1 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F4 F1 F2 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 F2 */
+	FXCH( ST(2) )			/* F1 F4 F5 F2 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 F2 */
+	FLD_S( MAT(14) )		/* F6 F4 F5 F2 */
+	FXCH( ST(3) )			/* F2 F4 F5 F6 */
+	FADDP( ST(0), ST(3) )		/* F4 F5 F6 */
+
+	FSTP_S( DST(0)   )		/* F5 F6 */
+	FSTP_S( DST(1)   )		/* F6 */
+	FSTP_S( DST(2)   )		/* */
+
+LLBL( x86_p3_3dnrm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_3dnrm_loop ) )
+
+LLBL( x86_p3_3dnrm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_2d_masked )
+GLNAME( gl_x86_transform_points3_2d_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_2dm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+ALIGNTEXT16
+LLBL( x86_p3_2dm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_2dm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+
+	FLD_S( SRC(1) )			/* F0 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 */
+	FADD_S( MAT(13) )
+
+	MOV_L( SRC(2), EAX )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EAX, DST(2) )
+
+LLBL( x86_p3_2dm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_2dm_loop ) )
+
+LLBL( x86_p3_2dm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT4
+GLOBL GLNAME( gl_x86_transform_points3_2d_no_rot_masked )
+GLNAME( gl_x86_transform_points3_2d_no_rot_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_2dnrm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+ALIGNTEXT16
+LLBL( x86_p3_2dnrm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_2dnrm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 */
+
+	FXCH( ST(2) )			/* F1 F4 F5 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 */
+
+	MOV_L( SRC(2), EAX )
+
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EAX, DST(2) )
+
+LLBL( x86_p3_2dnrm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_2dnrm_loop ) )
+
+LLBL( x86_p3_2dnrm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_identity_masked )
+GLNAME( gl_x86_transform_points3_identity_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_im_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need ecx; put dest+count in ARG_DEST */
+	MOV_L( EAX, ARG_SOURCE )
+	MOV_L( ECX, ARG_DEST )
+
+	CMP_L( ESI, EDI )
+	JE( LLBL( x86_p3_im_done ) )
+
+ALIGNTEXT16
+LLBL( x86_p3_im_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p3_im_skip ) )
+
+	MOV_L( SRC(0), EAX )
+	MOV_L( SRC(1), ECX )
+	MOV_L( SRC(2), EDX )
+
+	MOV_L( EAX, DST(0) )
+	MOV_L( ECX, DST(1) )
+	MOV_L( EDX, DST(2) )
+
+LLBL( x86_p3_im_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ARG_DEST, EDI )
+	JNE( LLBL( x86_p3_im_loop ) )
+
+LLBL( x86_p3_im_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86_xform_masked4.S b/xc/extras/Mesa/src/X86/x86_xform_masked4.S
new file mode 100644
index 000000000..3c8755dd8
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_xform_masked4.S
@@ -0,0 +1,710 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "xform_args.h"
+
+	SEG_TEXT
+
+#define FRAME_OFFSET	16
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_general_masked )
+GLNAME( gl_x86_transform_points4_general_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_gm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_gm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_gm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+	FLD_S( SRC(0) )			/* F7 F6 F5 F4 */
+	FMUL_S( MAT(3) )
+
+	FLD_S( SRC(1) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+	FLD_S( SRC(1) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(7) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+	FLD_S( SRC(2) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(11) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FLD_S( SRC(3) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(13) )
+	FLD_S( SRC(3) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+	FLD_S( SRC(3) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(15) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FXCH( ST(3) )			/* F4 F6 F5 F7 */
+	FSTP_S( DST(0) )		/* F6 F5 F7 */
+	FXCH( ST(1) )			/* F5 F6 F7 */
+	FSTP_S( DST(1) )		/* F6 F7 */
+	FSTP_S( DST(2) )		/* F7 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( x86_p4_gm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_gm_loop ) )
+
+LLBL( x86_p4_gm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_perspective_masked )
+GLNAME( gl_x86_transform_points4_perspective_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_pm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( EAX, ARG_SOURCE )	/* need eax; put stride in ARG_SOURCE */
+
+ALIGNTEXT16
+LLBL( x86_p4_pm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_pm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F0 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F6 F1 F0 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F6 F5 F4 */
+	FADDP( ST(0), ST(4) )		/* F1 F6 F5 F4 */
+	FADDP( ST(0), ST(2) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(3) )			/* F2 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(2), EAX )
+	XOR_L( CONST(-2147483648), EAX )/* change sign */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0) )		/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+	MOV_L( EAX, DST(3) )
+
+LLBL( x86_p4_pm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_pm_loop ) )
+
+LLBL( x86_p4_pm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_3d_masked )
+GLNAME( gl_x86_transform_points4_3d_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_3dm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( EAX, ARG_SOURCE )	/* need eax; put stride in ARG_SOURCE */
+
+ALIGNTEXT16
+LLBL( x86_p4_3dm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_3dm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+
+	FLD_S( SRC(1) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(3) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(13) )
+	FLD_S( SRC(3) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(3), EAX )
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0) )		/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+	MOV_L( EAX, DST(3) )
+
+LLBL( x86_p4_3dm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_3dm_loop ) )
+
+LLBL( x86_p4_3dm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_3d_no_rot_masked )
+GLNAME( gl_x86_transform_points4_3d_no_rot_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_3dnrm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( EAX, ARG_SOURCE )	/* need eax; put stride in ARG_SOURCE */
+
+ALIGNTEXT16
+LLBL( x86_p4_3dnrm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_3dnrm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F6 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FLD_S( SRC(3) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(13) )
+	FLD_S( SRC(3) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(3), EAX )
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0) )		/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+	MOV_L( EAX, DST(3) )
+
+LLBL( x86_p4_3dnrm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_3dnrm_loop ) )
+
+LLBL( x86_p4_3dnrm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_2d_masked )
+GLNAME( gl_x86_transform_points4_2d_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_2dm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( EAX, ARG_SOURCE )	/* need eax; put stride in ARG_SOURCE   */
+	MOV_L( ECX, ARG_DEST )		/* need ecx; put dest+count in ARG_DEST */
+
+ALIGNTEXT16
+LLBL( x86_p4_2dm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_2dm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+
+	FLD_S( SRC(1) )			/* F0 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	FLD_S( SRC(3) )			/* F0 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(13) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	MOV_L( SRC(2), EAX )
+	MOV_L( SRC(3), ECX )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+	MOV_L( EAX, DST(2) )
+	MOV_L( ECX, DST(3) )
+
+LLBL( x86_p4_2dm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ARG_DEST, EDI )
+	JNE( LLBL( x86_p4_2dm_loop ) )
+
+LLBL( x86_p4_2dm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_2d_no_rot_masked )
+GLNAME( gl_x86_transform_points4_2d_no_rot_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_2dnrm_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( EAX, ARG_SOURCE )	/* need eax; put stride in ARG_SOURCE   */
+	MOV_L( ECX, ARG_DEST )		/* need ecx; put dest+count in ARG_DEST */
+
+ALIGNTEXT16
+LLBL( x86_p4_2dnrm_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_2dnrm_skip ) )
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(3) )			/* F0 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(13) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	MOV_L( SRC(2), EAX )
+	MOV_L( SRC(3), ECX )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EAX, DST(2) )
+	MOV_L( ECX, DST(3) )
+
+LLBL( x86_p4_2dnrm_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ARG_DEST, EDI )
+	JNE( LLBL( x86_p4_2dnrm_loop ) )
+
+LLBL( x86_p4_2dnrm_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_identity_masked )
+GLNAME( gl_x86_transform_points4_identity_masked ):
+
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( ARG_CLIP, EBP )
+
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+	MOV_B( ARG_FLAG, BL )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_im_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	/* need eax; put stride in ARG_SOURCE */
+	MOV_L( EAX, ARG_SOURCE )
+
+	CMP_L( ESI, EDI )
+	JE( LLBL( x86_p4_im_done ) )
+
+ALIGNTEXT16
+LLBL( x86_p4_im_loop ):
+
+	TEST_B( BL, REGIND(EBP) )
+	JNZ( LLBL( x86_p4_im_skip ) )
+
+	MOV_L( SRC(0), EAX )
+	MOV_L( SRC(1), EDX )
+
+	MOV_L( EAX, DST(0) )
+	MOV_L( EDX, DST(1) )
+
+	MOV_L( SRC(2), EAX )
+	MOV_L( SRC(3), EDX )
+
+	MOV_L( EAX, DST(2) )
+	MOV_L( EDX, DST(3) )
+
+LLBL( x86_p4_im_skip ):
+
+	INC_L( EBP )
+	ADD_L( CONST(16), EDI )
+	ADD_L( ARG_SOURCE, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_im_loop ) )
+
+LLBL( x86_p4_im_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86_xform_raw2.S b/xc/extras/Mesa/src/X86/x86_xform_raw2.S
new file mode 100644
index 000000000..fd6b59617
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_xform_raw2.S
@@ -0,0 +1,535 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "xform_args.h"
+
+	SEG_TEXT
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_general_raw )
+GLNAME( gl_x86_transform_points2_general_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_gr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_gr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+	FLD_S( SRC(0) )			/* F7 F6 F5 F4 */
+	FMUL_S( MAT(3) )
+
+	FLD_S( SRC(1) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+	FLD_S( SRC(1) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(7) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FXCH( ST(3) )			/* F4 F6 F5 F7 */
+	FADD_S( MAT(12) )
+	FXCH( ST(2) )			/* F5 F6 F4 F7 */
+	FADD_S( MAT(13) )
+	FXCH( ST(1) )			/* F6 F5 F4 F7 */
+	FADD_S( MAT(14) )
+	FXCH( ST(3) )			/* F7 F5 F4 F6 */
+	FADD_S( MAT(15) )
+
+	FXCH( ST(2) )			/* F4 F5 F7 F6 */
+	FSTP_S( DST(0) )		/* F5 F7 F6 */
+	FSTP_S( DST(1) )		/* F7 F6 */
+	FXCH( ST(1) )			/* F6 F7 */
+	FSTP_S( DST(2) )		/* F7 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( x86_p2_gr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_gr_loop ) )
+
+LLBL( x86_p2_gr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_perspective_raw )
+GLNAME( gl_x86_transform_points2_perspective_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_pr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( MAT(14), EBX )
+
+ALIGNTEXT16
+LLBL( x86_p2_pr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FSTP_S( DST(0)   )		/* F1 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EBX, DST(2) )
+	MOV_L( CONST(FP_ZERO), DST(3) )
+
+LLBL( x86_p2_pr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_pr_loop ) )
+
+LLBL( x86_p2_pr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_3d_raw )
+GLNAME( gl_x86_transform_points2_3d_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_3dr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_3dr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+
+	FLD_S( SRC(1) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 F6 */
+	FADD_S( MAT(13) )
+	FXCH( ST(2) )			/* F6 F4 F5 */
+	FADD_S( MAT(14) )
+
+	FXCH( ST(1) )			/* F4 F6 F5 */
+	FSTP_S( DST(0) )		/* F6 F5 */
+	FXCH( ST(1) )			/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+
+LLBL( x86_p2_3dr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_3dr_loop ) )
+
+LLBL( x86_p2_3dr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_3d_no_rot_raw )
+GLNAME( gl_x86_transform_points2_3d_no_rot_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_3dnrr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	MOV_L( MAT(14), EBX )
+
+ALIGNTEXT16
+LLBL( x86_p2_3dnrr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 */
+	FXCH( ST(2) )			/* F1 F4 F5 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 */
+
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+	MOV_L( EBX, DST(2) )
+
+LLBL( x86_p2_3dnrr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_3dnrr_loop ) )
+
+LLBL( x86_p2_3dnrr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_2d_raw )
+GLNAME( gl_x86_transform_points2_2d_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_2dr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_2dr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+
+	FLD_S( SRC(1) )			/* F0 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 */
+	FADD_S( MAT(13) )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+
+LLBL( x86_p2_2dr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_2dr_loop ) )
+
+LLBL( x86_p2_2dr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT4
+GLOBL GLNAME( gl_x86_transform_points2_2d_no_rot_raw )
+GLNAME( gl_x86_transform_points2_2d_no_rot_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_2dnrr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p2_2dnrr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 */
+	FXCH( ST(2) )			/* F1 F4 F5 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 */
+
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+
+LLBL( x86_p2_2dnrr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_2dnrr_loop ) )
+
+LLBL( x86_p2_2dnrr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points2_identity_raw )
+GLNAME( gl_x86_transform_points2_identity_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p2_ir_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	CMP_L( ESI, EDI )
+	JE( LLBL( x86_p2_ir_done ) )
+
+ALIGNTEXT16
+LLBL( x86_p2_ir_loop ):
+
+	MOV_L( SRC(0), EBX )
+	MOV_L( SRC(1), EDX )
+
+	MOV_L( EBX, DST(0) )
+	MOV_L( EDX, DST(1) )
+
+LLBL( x86_p2_ir_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p2_ir_loop ) )
+
+LLBL( x86_p2_ir_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
diff --git a/xc/extras/Mesa/src/X86/x86_xform_raw3.S b/xc/extras/Mesa/src/X86/x86_xform_raw3.S
new file mode 100644
index 000000000..b32afee90
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_xform_raw3.S
@@ -0,0 +1,605 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "xform_args.h"
+
+	SEG_TEXT
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_general_raw )
+GLNAME( gl_x86_transform_points3_general_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_gr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_gr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+	FLD_S( SRC(0) )			/* F7 F6 F5 F4 */
+	FMUL_S( MAT(3) )
+
+	FLD_S( SRC(1) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+	FLD_S( SRC(1) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(7) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+	FLD_S( SRC(2) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(11) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FXCH( ST(3) )			/* F4 F6 F5 F7 */
+	FADD_S( MAT(12) )
+	FXCH( ST(2) )			/* F5 F6 F4 F7 */
+	FADD_S( MAT(13) )
+	FXCH( ST(1) )			/* F6 F5 F4 F7 */
+	FADD_S( MAT(14) )
+	FXCH( ST(3) )			/* F7 F5 F4 F6 */
+	FADD_S( MAT(15) )
+
+	FXCH( ST(2) )			/* F4 F5 F7 F6 */
+	FSTP_S( DST(0) )		/* F5 F7 F6 */
+	FSTP_S( DST(1) )		/* F7 F6 */
+	FXCH( ST(1) )			/* F6 F7 */
+	FSTP_S( DST(2) )		/* F7 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( x86_p3_gr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_gr_loop ) )
+
+LLBL( x86_p3_gr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_perspective_raw )
+GLNAME( gl_x86_transform_points3_perspective_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_pr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_pr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F0 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F5 F4 */
+	FADDP( ST(0), ST(4) )		/* F1 F2 F5 F4 */
+	FADDP( ST(0), ST(2) )		/* F2 F5 F4 */
+	FLD_S( MAT(14) )		/* F6 F2 F5 F4 */
+	FXCH( ST(1) )			/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(2), EBX )
+	XOR_L( CONST(-2147483648), EBX )/* change sign */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0) )		/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+	MOV_L( EBX, DST(3) )
+
+LLBL( x86_p3_pr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_pr_loop ) )
+
+LLBL( x86_p3_pr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_3d_raw )
+GLNAME( gl_x86_transform_points3_3d_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_3dr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_3dr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+
+	FLD_S( SRC(1) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 F6 */
+	FADD_S( MAT(13) )
+	FXCH( ST(2) )			/* F6 F4 F5 */
+	FADD_S( MAT(14) )
+
+	FXCH( ST(1) )			/* F4 F6 F5 */
+	FSTP_S( DST(0)   )		/* F6 F5 */
+	FXCH( ST(1) )			/* F5 F6 */
+	FSTP_S( DST(1)   )		/* F6 */
+	FSTP_S( DST(2)   )		/* */
+
+LLBL( x86_p3_3dr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_3dr_loop ) )
+
+LLBL( x86_p3_3dr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_3d_no_rot_raw )
+GLNAME( gl_x86_transform_points3_3d_no_rot_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_3dnrr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_3dnrr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F2 F1 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F4 F1 F2 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 F2 */
+	FXCH( ST(2) )			/* F1 F4 F5 F2 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 F2 */
+	FLD_S( MAT(14) )		/* F6 F4 F5 F2 */
+	FXCH( ST(3) )			/* F2 F4 F5 F6 */
+	FADDP( ST(0), ST(3) )		/* F4 F5 F6 */
+
+	FSTP_S( DST(0)   )		/* F5 F6 */
+	FSTP_S( DST(1)   )		/* F6 */
+	FSTP_S( DST(2)   )		/* */
+
+LLBL( x86_p3_3dnrr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_3dnrr_loop ) )
+
+LLBL( x86_p3_3dnrr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_2d_raw )
+GLNAME( gl_x86_transform_points3_2d_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_2dr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_2dr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+
+	FLD_S( SRC(1) )			/* F0 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FADD_S( MAT(12) )
+	FXCH( ST(1) )			/* F5 F4 */
+	FADD_S( MAT(13) )
+
+	MOV_L( SRC(2), EBX )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EBX, DST(2) )
+
+LLBL( x86_p3_2dr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_2dr_loop ) )
+
+LLBL( x86_p3_2dr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_2d_no_rot_raw )
+GLNAME( gl_x86_transform_points3_2d_no_rot_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_2dnrr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p3_2dnrr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F1 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F4 F1 */
+	FADD_S( MAT(12) )
+	FLD_S( MAT(13) )		/* F5 F4 F1 */
+
+	FXCH( ST(2) )			/* F1 F4 F5 */
+	FADDP( ST(0), ST(2) )		/* F4 F5 */
+
+	MOV_L( SRC(2), EBX )
+
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+	MOV_L( EBX, DST(2) )
+
+LLBL( x86_p3_2dnrr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_2dnrr_loop ) )
+
+LLBL( x86_p3_2dnrr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points3_identity_raw )
+GLNAME(gl_x86_transform_points3_identity_raw ):
+
+#define FRAME_OFFSET 16
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p3_ir_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	CMP_L( ESI, EDI )
+	JE( LLBL( x86_p3_ir_done ) )
+
+ALIGNTEXT16
+LLBL( x86_p3_ir_loop ):
+
+#if 1
+	MOV_L( SRC(0), EBX )
+	MOV_L( SRC(1), EBP )
+	MOV_L( SRC(2), EDX )
+
+	MOV_L( EBX, DST(0) )
+	MOV_L( EBP, DST(1) )
+	MOV_L( EDX, DST(2) )
+#else
+	FLD_S( SRC(0) )
+	FLD_S( SRC(1) )
+	FLD_S( SRC(2) )
+
+	FSTP_S( DST(2) )
+	FSTP_S( DST(1) )
+	FSTP_S( DST(0) )
+#endif
+
+LLBL( x86_p3_ir_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p3_ir_loop ) )
+
+LLBL( x86_p3_ir_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86_xform_raw4.S b/xc/extras/Mesa/src/X86/x86_xform_raw4.S
new file mode 100644
index 000000000..29258acbd
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/x86_xform_raw4.S
@@ -0,0 +1,638 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "assyntax.h"
+#include "xform_args.h"
+
+	SEG_TEXT
+
+#define FP_ONE		1065353216
+#define FP_ZERO		0
+
+#define SRC(i)		REGOFF(i * 4, ESI)
+#define DST(i)		REGOFF(i * 4, EDI)
+#define MAT(i)		REGOFF(i * 4, EDX)
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_general_raw )
+GLNAME( gl_x86_transform_points4_general_raw ):
+
+#define FRAME_OFFSET 8
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_gr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_gr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+	FLD_S( SRC(0) )			/* F7 F6 F5 F4 */
+	FMUL_S( MAT(3) )
+
+	FLD_S( SRC(1) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+	FLD_S( SRC(1) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(7) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+	FLD_S( SRC(2) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(11) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FLD_S( SRC(3) )			/* F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(13) )
+	FLD_S( SRC(3) )			/* F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+	FLD_S( SRC(3) )			/* F3 F2 F1 F0 F7 F6 F5 F4 */
+	FMUL_S( MAT(15) )
+
+	FXCH( ST(3) )			/* F0 F2 F1 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(7) )		/* F2 F1 F3 F7 F6 F5 F4 */
+	FXCH( ST(1) )			/* F1 F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F2 F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F3 F7 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F7 F6 F5 F4 */
+
+	FXCH( ST(3) )			/* F4 F6 F5 F7 */
+	FSTP_S( DST(0) )		/* F6 F5 F7 */
+	FXCH( ST(1) )			/* F5 F6 F7 */
+	FSTP_S( DST(1) )		/* F6 F7 */
+	FSTP_S( DST(2) )		/* F7 */
+	FSTP_S( DST(3) )		/* */
+
+LLBL( x86_p4_gr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_gr_loop ) )
+
+LLBL( x86_p4_gr_done ):
+
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_perspective_raw )
+GLNAME( gl_x86_transform_points4_perspective_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_pr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_pr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F0 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F6 F1 F0 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F6 F5 F4 */
+	FADDP( ST(0), ST(4) )		/* F1 F6 F5 F4 */
+	FADDP( ST(0), ST(2) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(3) )			/* F2 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(2), EBX )
+	XOR_L( CONST(-2147483648), EBX )/* change sign */
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0) )		/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+	MOV_L( EBX, DST(3) )
+
+LLBL( x86_p4_pr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_pr_loop ) )
+
+LLBL( x86_p4_pr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_3d_raw )
+GLNAME( gl_x86_transform_points4_3d_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_3dr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_3dr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+	FLD_S( SRC(0) )			/* F6 F5 F4 */
+	FMUL_S( MAT(2) )
+
+	FLD_S( SRC(1) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(5) )
+	FLD_S( SRC(1) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(6) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(2) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(8) )
+	FLD_S( SRC(2) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(9) )
+	FLD_S( SRC(2) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	FLD_S( SRC(3) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(13) )
+	FLD_S( SRC(3) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(3), EBX )
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0) )		/* F5 F6 */
+	FSTP_S( DST(1) )		/* F6 */
+	FSTP_S( DST(2) )		/* */
+	MOV_L( EBX, DST(3) )
+
+LLBL( x86_p4_3dr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_3dr_loop ) )
+
+LLBL( x86_p4_3dr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME(gl_x86_transform_points4_3d_no_rot_raw)
+GLNAME(gl_x86_transform_points4_3d_no_rot_raw):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_3dnrr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_3dnrr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(2) )			/* F6 F5 F4 */
+	FMUL_S( MAT(10) )
+
+	FLD_S( SRC(3) )			/* F0 F6 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(13) )
+	FLD_S( SRC(3) )			/* F2 F1 F0 F6 F5 F4 */
+	FMUL_S( MAT(14) )
+
+	FXCH( ST(2) )			/* F0 F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(5) )		/* F1 F2 F6 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F2 F6 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F6 F5 F4 */
+
+	MOV_L( SRC(3), EBX )
+
+	FXCH( ST(2) )			/* F4 F5 F6 */
+	FSTP_S( DST(0)   )		/* F5 F6 */
+	FSTP_S( DST(1)   )		/* F6 */
+	FSTP_S( DST(2)   )		/* */
+	MOV_L( EBX, DST(3) )
+
+LLBL( x86_p4_3dnrr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_3dnrr_loop ) )
+
+LLBL( x86_p4_3dnrr_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_2d_raw )
+GLNAME( gl_x86_transform_points4_2d_raw ):
+
+#define FRAME_OFFSET 16
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_2dr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_2dr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+	FLD_S( SRC(0) )			/* F5 F4 */
+	FMUL_S( MAT(1) )
+
+	FLD_S( SRC(1) )			/* F0 F5 F4 */
+	FMUL_S( MAT(4) )
+	FLD_S( SRC(1) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	FLD_S( SRC(3) )			/* F0 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(13) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	MOV_L( SRC(2), EBX )
+	MOV_L( SRC(3), EBP )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0) )		/* F5 */
+	FSTP_S( DST(1) )		/* */
+	MOV_L( EBX, DST(2) )
+	MOV_L( EBP, DST(3) )
+
+LLBL( x86_p4_2dr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_2dr_loop ) )
+
+LLBL( x86_p4_2dr_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_2d_no_rot_raw )
+GLNAME( gl_x86_transform_points4_2d_no_rot_raw ):
+
+#define FRAME_OFFSET 16
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+	PUSH_L( EBP )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_2dnrr_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+ALIGNTEXT16
+LLBL( x86_p4_2dnrr_loop ):
+
+	FLD_S( SRC(0) )			/* F4 */
+	FMUL_S( MAT(0) )
+
+	FLD_S( SRC(1) )			/* F5 F4 */
+	FMUL_S( MAT(5) )
+
+	FLD_S( SRC(3) )			/* F0 F5 F4 */
+	FMUL_S( MAT(12) )
+	FLD_S( SRC(3) )			/* F1 F0 F5 F4 */
+	FMUL_S( MAT(13) )
+
+	FXCH( ST(1) )			/* F0 F1 F5 F4 */
+	FADDP( ST(0), ST(3) )		/* F1 F5 F4 */
+	FADDP( ST(0), ST(1) )		/* F5 F4 */
+
+	MOV_L( SRC(2), EBX )
+	MOV_L( SRC(3), EBP )
+
+	FXCH( ST(1) )			/* F4 F5 */
+	FSTP_S( DST(0)   )		/* F5 */
+	FSTP_S( DST(1)   )		/* */
+	MOV_L( EBX, DST(2) )
+	MOV_L( EBP, DST(3) )
+
+LLBL( x86_p4_2dnrr_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_2dnrr_loop ) )
+
+LLBL( x86_p4_2dnrr_done ):
+
+	POP_L( EBP )
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
+#undef FRAME_OFFSET
+
+
+
+
+ALIGNTEXT16
+GLOBL GLNAME( gl_x86_transform_points4_identity_raw )
+GLNAME( gl_x86_transform_points4_identity_raw ):
+
+#define FRAME_OFFSET 12
+	PUSH_L( ESI )
+	PUSH_L( EDI )
+	PUSH_L( EBX )
+
+	MOV_L( ARG_SOURCE, ESI )
+	MOV_L( ARG_DEST, EDI )
+
+	MOV_L( ARG_MATRIX, EDX )
+	MOV_L( REGOFF(V4F_COUNT, ESI), ECX )
+
+	TEST_L( ECX, ECX )
+	JZ( LLBL( x86_p4_ir_done ) )
+
+	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX )
+	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) )
+
+	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) )
+	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )
+
+	SHL_L( CONST(4), ECX )
+	MOV_L( REGOFF(V4F_START, ESI), ESI )
+
+	MOV_L( REGOFF(V4F_START, EDI), EDI )
+	ADD_L( EDI, ECX )
+
+	CMP_L( ESI, EDI )
+	JE( LLBL( x86_p4_ir_done ) )
+
+ALIGNTEXT16
+LLBL( x86_p4_ir_loop ):
+
+	MOV_L( SRC(0), EBX )
+	MOV_L( SRC(1), EDX )
+
+	MOV_L( EBX, DST(0) )
+	MOV_L( EDX, DST(1) )
+
+	MOV_L( SRC(2), EBX )
+	MOV_L( SRC(3), EDX )
+
+	MOV_L( EBX, DST(2) )
+	MOV_L( EDX, DST(3) )
+
+LLBL( x86_p4_ir_skip ):
+
+	ADD_L( CONST(16), EDI )
+	ADD_L( EAX, ESI )
+	CMP_L( ECX, EDI )
+	JNE( LLBL( x86_p4_ir_loop ) )
+
+LLBL( x86_p4_ir_done ):
+
+	POP_L( EBX )
+	POP_L( EDI )
+	POP_L( ESI )
+	RET
diff --git a/xc/extras/Mesa/src/X86/x86a.S b/xc/extras/Mesa/src/X86/x86a.S
deleted file mode 100644
index ac12457ee..000000000
--- a/xc/extras/Mesa/src/X86/x86a.S
+++ /dev/null
@@ -1,9081 +0,0 @@
-/* $XFree86: xc/extras/Mesa/src/X86/x86a.S,v 1.4 2000/09/26 15:56:40 tsi Exp $ */
-
-#include "assyntax.h"
-
-	SEG_TEXT
-
-#define FP_ONE 1065353216
-#define FP_ZERO 0
-
-#define S(i) 	REGOFF(i * 4, ESI)
-#define D(i) 	REGOFF(i * 4, EDI)
-#define M(i) 	REGOFF(i * 4, EDX)
-
-/*
- * 	Offsets into GLvector4f
- */
-#define V4F_DATA 	0
-#define V4F_START 	4
-#define V4F_COUNT 	8
-#define V4F_STRIDE 	12
-#define V4F_SIZE 	16
-#define V4F_FLAGS 	20
-
-#define VEC_SIZE_1   1
-#define VEC_SIZE_2   3
-#define VEC_SIZE_3   7
-#define VEC_SIZE_4   15
-
-/*
- *	Offsets for transform_func arguments
- *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLfloat m[16], 
- *	                                const GLvector4f *from_vec, 
- *	                                const GLubyte *clipmask,
- *	                                const GLubyte flag );
- */
-#define OFFSET_DEST 4
-#define OFFSET_MATRIX 8
-#define OFFSET_SOURCE 12
-#define OFFSET_CLIP 16
-#define OFFSET_FLAG 20
-
-#define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-#define ARG_MATRIX 	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
-#define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-#define ARG_CLIP 	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
-#define ARG_FLAG 	REGOFF(FRAME_OFFSET+OFFSET_FLAG, ESP)
-
-/*
-########################################
-##
-## Masked versions
-##
-########################################
-*/
-
-/*
-########################################
-##
-## gl_x86_transform_points2_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_general_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_general_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2mgm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mgm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2mgm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p2mgm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2mgm_top) )
-
-
-LLBL(p2mgm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_identity_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_identity_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2mim_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p2mim_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mim_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2mim_skip) )
-
-	MOV_L( S(0), EAX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EAX, D(0) )
-	MOV_L( EDX, D(1) )
-LLBL(p2mim_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2mim_top) )
-
-
-LLBL(p2mim_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2m2dm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2m2dm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-LLBL(p2m2dm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m2dm_top) )
-
-
-LLBL(p2m2dm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_no_rot_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_no_rot_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2m2dnrm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dnrm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2m2dnrm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-LLBL(p2m2dnrm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m2dnrm_top) )
-
-
-LLBL(p2m2dnrm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2m3dm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2m3dm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-LLBL(p2m3dm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m3dm_top) )
-
-
-LLBL(p2m3dm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_no_rot_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_no_rot_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2m3dnrm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	MOV_L( M(14), EAX )
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dnrm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2m3dnrm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EAX, D(2) )
-
-LLBL(p2m3dnrm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m3dnrm_top) )
-
-
-LLBL(p2m3dnrm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_perspective_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_perspective_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p2mpm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	MOV_L( M(14), EAX )
-	ALIGNTEXT4ifNOP
-LLBL(p2mpm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p2mpm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FSTP_S( D(0)   ) 	/* F1 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EAX, D(2) )
-	MOV_L( CONST(FP_ZERO), D(3) )
-
-LLBL(p2mpm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2mpm_top) )
-
-
-LLBL(p2mpm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_general_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_general_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3mgm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mgm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3mgm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p3mgm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3mgm_top) )
-
-
-LLBL(p3mgm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_identity_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_identity_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3mim_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-	MOV_L( ECX, ARG_DEST ) 			/* need ecx; put dest+count in ARG_DEST */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p3mim_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mim_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3mim_skip) )
-
-	MOV_L( S(0), EAX )
-	MOV_L( S(1), ECX )
-	MOV_L( S(2), EDX )
-
-	MOV_L( EAX, D(0) )
-	MOV_L( ECX, D(1) )
-	MOV_L( EDX, D(2) )
-LLBL(p3mim_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ARG_DEST, EDI )
-	JNE( LLBL(p3mim_top) )
-
-
-LLBL(p3mim_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3m2dm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3m2dm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	MOV_L( S(2), EAX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EAX, D(2) )
-
-LLBL(p3m2dm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m2dm_top) )
-
-
-LLBL(p3m2dm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_no_rot_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_no_rot_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3m2dnrm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dnrm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3m2dnrm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	MOV_L( S(2), EAX )
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EAX, D(2) )
-
-LLBL(p3m2dnrm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m2dnrm_top) )
-
-
-LLBL(p3m2dnrm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3m3dm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3m3dm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-LLBL(p3m3dm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m3dm_top) )
-
-
-LLBL(p3m3dm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_no_rot_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_no_rot_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3m3dnrm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dnrm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3m3dnrm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F2 F1 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F4 F1 F2 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 F2 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 F2 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 F2 */
-	FLD_S( M(14) ) 	/* F6 F4 F5 F2 */
-	FXCH( ST(3) ) 	/* F2 F4 F5 F6 */
-	FADDP( ST(0), ST(3) ) 	/* F4 F5 F6 */
-
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-LLBL(p3m3dnrm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m3dnrm_top) )
-
-
-LLBL(p3m3dnrm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_perspective_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_perspective_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p3mpm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mpm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p3mpm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F2 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F2 F5 F4 */
-	FLD_S( M(14) ) 	/* F6 F2 F5 F4 */
-	FXCH( ST(1) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EAX )
-	XOR_L( CONST(-2147483648), EAX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EAX, D(3) )
-
-LLBL(p3mpm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3mpm_top) )
-
-
-LLBL(p3mpm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_general_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_general_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4mgm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mgm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4mgm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(14) )
-	FLD_S( S(3) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(15) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FSTP_S( D(0)   ) 	/* F6 F5 F7 */
-	FXCH( ST(1) ) 	/* F5 F6 F7 */
-	FSTP_S( D(1)   ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p4mgm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4mgm_top) )
-
-
-LLBL(p4mgm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_identity_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_identity_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4mim_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p4mim_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mim_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4mim_skip) )
-
-	MOV_L( S(0), EAX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EAX, D(0) )
-	MOV_L( EDX, D(1) )
-
-	MOV_L( S(2), EAX )
-	MOV_L( S(3), EDX )
-
-	MOV_L( EAX, D(2) )
-	MOV_L( EDX, D(3) )
-LLBL(p4mim_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4mim_top) )
-
-
-LLBL(p4mim_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4m2dm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-	MOV_L( ECX, ARG_DEST ) 			/* need ecx; put dest+count in ARG_DEST */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4m2dm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EAX )
-	MOV_L( S(3), ECX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EAX, D(2) )
-	MOV_L( ECX, D(3) )
-
-LLBL(p4m2dm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ARG_DEST, EDI )
-	JNE( LLBL(p4m2dm_top) )
-
-
-LLBL(p4m2dm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_no_rot_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_no_rot_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4m2dnrm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-	MOV_L( ECX, ARG_DEST ) 			/* need ecx; put dest+count in ARG_DEST */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dnrm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4m2dnrm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EAX )
-	MOV_L( S(3), ECX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EAX, D(2) )
-	MOV_L( ECX, D(3) )
-
-LLBL(p4m2dnrm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ARG_DEST, EDI )
-	JNE( LLBL(p4m2dnrm_top) )
-
-
-LLBL(p4m2dnrm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4m3dm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4m3dm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EAX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EAX, D(3) )
-
-LLBL(p4m3dm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4m3dm_top) )
-
-
-LLBL(p4m3dm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_no_rot_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_no_rot_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4m3dnrm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dnrm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4m3dnrm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EAX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EAX, D(3) )
-
-LLBL(p4m3dnrm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4m3dnrm_top) )
-
-
-LLBL(p4m3dnrm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_perspective_masked)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_perspective_masked):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(p4mpm_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mpm_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(p4mpm_skip) )
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F6 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F6 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F6 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F2 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EAX )
-	XOR_L( CONST(-2147483648), EAX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EAX, D(3) )
-
-LLBL(p4mpm_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4mpm_top) )
-
-
-LLBL(p4mpm_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-
-
-/*
-########################################
-##
-## Unmasked versions
-##
-########################################
-*/
-
-/*
-########################################
-##
-## gl_x86_transform_points2_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_general_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_general_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mgr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mgr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p2mgr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2mgr_top) )
-
-
-LLBL(p2mgr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_identity_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_identity_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mir_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p2mir_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mir_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-LLBL(p2mir_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2mir_top) )
-
-
-LLBL(p2mir_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-LLBL(p2m2dr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m2dr_top) )
-
-
-LLBL(p2m2dr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_no_rot_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_no_rot_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dnrr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dnrr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-LLBL(p2m2dnrr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m2dnrr_top) )
-
-
-LLBL(p2m2dnrr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-LLBL(p2m3dr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m3dr_top) )
-
-
-LLBL(p2m3dr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_no_rot_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_no_rot_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dnrr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dnrr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-LLBL(p2m3dnrr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2m3dnrr_top) )
-
-
-LLBL(p2m3dnrr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_perspective_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_perspective_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mpr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2mpr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FSTP_S( D(0)   ) 	/* F1 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( CONST(FP_ZERO), D(3) )
-
-LLBL(p2mpr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p2mpr_top) )
-
-
-LLBL(p2mpr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_general_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_general_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mgr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mgr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p3mgr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3mgr_top) )
-
-
-LLBL(p3mgr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_identity_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_identity_raw):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mir_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p3mir_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mir_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EBP )
-	MOV_L( S(2), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EBP, D(1) )
-	MOV_L( EDX, D(2) )
-LLBL(p3mir_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3mir_top) )
-
-
-LLBL(p3mir_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	MOV_L( S(2), EBX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-LLBL(p3m2dr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m2dr_top) )
-
-
-LLBL(p3m2dr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_no_rot_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_no_rot_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dnrr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dnrr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	MOV_L( S(2), EBX )
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-LLBL(p3m2dnrr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m2dnrr_top) )
-
-
-LLBL(p3m2dnrr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-LLBL(p3m3dr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m3dr_top) )
-
-
-LLBL(p3m3dr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_no_rot_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_no_rot_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dnrr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dnrr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F2 F1 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F4 F1 F2 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 F2 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 F2 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 F2 */
-	FLD_S( M(14) ) 	/* F6 F4 F5 F2 */
-	FXCH( ST(3) ) 	/* F2 F4 F5 F6 */
-	FADDP( ST(0), ST(3) ) 	/* F4 F5 F6 */
-
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-LLBL(p3m3dnrr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3m3dnrr_top) )
-
-
-LLBL(p3m3dnrr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_perspective_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_perspective_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mpr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mpr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F2 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F2 F5 F4 */
-	FLD_S( M(14) ) 	/* F6 F2 F5 F4 */
-	FXCH( ST(1) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p3mpr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p3mpr_top) )
-
-
-LLBL(p3mpr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_general_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_general_raw):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mgr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mgr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(14) )
-	FLD_S( S(3) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(15) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FSTP_S( D(0)   ) 	/* F6 F5 F7 */
-	FXCH( ST(1) ) 	/* F5 F6 F7 */
-	FSTP_S( D(1)   ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p4mgr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4mgr_top) )
-
-
-LLBL(p4mgr_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_identity_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_identity_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mir_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p4mir_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mir_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EDX )
-
-	MOV_L( EBX, D(2) )
-	MOV_L( EDX, D(3) )
-LLBL(p4mir_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4mir_top) )
-
-
-LLBL(p4mir_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_raw):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4m2dr_top) )
-
-
-LLBL(p4m2dr_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_no_rot_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_no_rot_raw):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dnrr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dnrr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dnrr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4m2dnrr_top) )
-
-
-LLBL(p4m2dnrr_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4m3dr_top) )
-
-
-LLBL(p4m3dr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_no_rot_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_no_rot_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dnrr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dnrr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dnrr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4m3dnrr_top) )
-
-
-LLBL(p4m3dnrr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_perspective_raw)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_perspective_raw):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mpr_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mpr_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F6 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F6 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F6 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F2 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4mpr_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(p4mpr_top) )
-
-
-LLBL(p4mpr_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-
-
-#undef OFFSET_DEST
-#undef OFFSET_MATRIX
-#undef OFFSET_SOURCE
-#undef OFFSET_CLIP
-#undef OFFSET_FLAG
-
-#undef ARG_DEST
-#undef ARG_MATRIX
-#undef ARG_SOURCE
-#undef ARG_CLIP
-#undef ARG_FLAG
-
-#define OFFSET_DEST 4
-#define OFFSET_MATRIX 8
-#define OFFSET_SOURCE 12
-#define OFFSET_STRIDE 16
-#define OFFSET_COUNT 20
-
-#define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-#define ARG_MATRIX 	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
-#define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-#define ARG_STRIDE 	REGOFF(FRAME_OFFSET+OFFSET_STRIDE, ESP)
-#define ARG_COUNT 	REGOFF(FRAME_OFFSET+OFFSET_COUNT, ESP)
-
-/*
-########################################
-##
-## clean (no vertex struct) versions
-##
-########################################
-*/
-
-
-/*
-########################################
-##
-## gl_x86_transform_points2_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_general_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_general_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mgv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mgv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p2mgv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2mgv8_top) )
-
-
-LLBL(p2mgv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_identity_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_identity_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2miv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p2miv8_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p2miv8_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2miv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2miv8_top) )
-
-
-LLBL(p2miv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m2dv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m2dv8_top) )
-
-
-LLBL(p2m2dv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_no_rot_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_no_rot_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dnrv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dnrv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m2dnrv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m2dnrv8_top) )
-
-
-LLBL(p2m2dnrv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m3dv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m3dv8_top) )
-
-
-LLBL(p2m3dv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_no_rot_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_no_rot_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dnrv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dnrv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m3dnrv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m3dnrv8_top) )
-
-
-LLBL(p2m3dnrv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_perspective_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_perspective_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mpv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2mpv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FSTP_S( D(0)   ) 	/* F1 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( CONST(FP_ZERO), D(3) )
-
-LLBL(p2mpv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2mpv8_top) )
-
-
-LLBL(p2mpv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_general_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_general_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mgv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mgv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p3mgv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3mgv8_top) )
-
-
-LLBL(p3mgv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_identity_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_identity_v8):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3miv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p3miv8_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p3miv8_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EBP )
-	MOV_L( S(2), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EBP, D(1) )
-	MOV_L( EDX, D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3miv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3miv8_top) )
-
-
-LLBL(p3miv8_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	MOV_L( S(2), EBX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m2dv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m2dv8_top) )
-
-
-LLBL(p3m2dv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_no_rot_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_no_rot_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dnrv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dnrv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	MOV_L( S(2), EBX )
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m2dnrv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m2dnrv8_top) )
-
-
-LLBL(p3m2dnrv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m3dv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m3dv8_top) )
-
-
-LLBL(p3m3dv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_no_rot_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_no_rot_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dnrv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dnrv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F2 F1 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F4 F1 F2 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 F2 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 F2 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 F2 */
-	FLD_S( M(14) ) 	/* F6 F4 F5 F2 */
-	FXCH( ST(3) ) 	/* F2 F4 F5 F6 */
-	FADDP( ST(0), ST(3) ) 	/* F4 F5 F6 */
-
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m3dnrv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m3dnrv8_top) )
-
-
-LLBL(p3m3dnrv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_perspective_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_perspective_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mpv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mpv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F2 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F2 F5 F4 */
-	FLD_S( M(14) ) 	/* F6 F2 F5 F4 */
-	FXCH( ST(1) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p3mpv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3mpv8_top) )
-
-
-LLBL(p3mpv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_general_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_general_v8):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mgv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mgv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(14) )
-	FLD_S( S(3) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(15) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FSTP_S( D(0)   ) 	/* F6 F5 F7 */
-	FXCH( ST(1) ) 	/* F5 F6 F7 */
-	FSTP_S( D(1)   ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p4mgv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4mgv8_top) )
-
-
-LLBL(p4mgv8_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_identity_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_identity_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4miv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p4miv8_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p4miv8_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EDX )
-
-	MOV_L( EBX, D(2) )
-	MOV_L( EDX, D(3) )
-LLBL(p4miv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4miv8_top) )
-
-
-LLBL(p4miv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_v8):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m2dv8_top) )
-
-
-LLBL(p4m2dv8_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_no_rot_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_no_rot_v8):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dnrv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dnrv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dnrv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m2dnrv8_top) )
-
-
-LLBL(p4m2dnrv8_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m3dv8_top) )
-
-
-LLBL(p4m3dv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_no_rot_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_no_rot_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dnrv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dnrv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dnrv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m3dnrv8_top) )
-
-
-LLBL(p4m3dnrv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_perspective_v8)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_perspective_v8):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mpv8_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mpv8_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F6 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F6 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F6 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F2 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4mpv8_skip):
-	ADD_L( CONST(32), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4mpv8_top) )
-
-
-LLBL(p4mpv8_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-
-
-/*
-########################################
-##
-## gl_x86_transform_points2_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_general_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_general_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mgv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mgv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p2mgv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2mgv12_top) )
-
-
-LLBL(p2mgv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_identity_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_identity_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2miv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p2miv12_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p2miv12_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2miv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2miv12_top) )
-
-
-LLBL(p2miv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m2dv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m2dv12_top) )
-
-
-LLBL(p2m2dv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_no_rot_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_no_rot_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dnrv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dnrv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m2dnrv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m2dnrv12_top) )
-
-
-LLBL(p2m2dnrv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m3dv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m3dv12_top) )
-
-
-LLBL(p2m3dv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_no_rot_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_no_rot_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dnrv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dnrv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m3dnrv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m3dnrv12_top) )
-
-
-LLBL(p2m3dnrv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_perspective_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_perspective_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mpv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2mpv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FSTP_S( D(0)   ) 	/* F1 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( CONST(FP_ZERO), D(3) )
-
-LLBL(p2mpv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2mpv12_top) )
-
-
-LLBL(p2mpv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_general_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_general_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mgv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mgv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p3mgv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3mgv12_top) )
-
-
-LLBL(p3mgv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_identity_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_identity_v12):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3miv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p3miv12_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p3miv12_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EBP )
-	MOV_L( S(2), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EBP, D(1) )
-	MOV_L( EDX, D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3miv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3miv12_top) )
-
-
-LLBL(p3miv12_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	MOV_L( S(2), EBX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m2dv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m2dv12_top) )
-
-
-LLBL(p3m2dv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_no_rot_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_no_rot_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dnrv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dnrv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	MOV_L( S(2), EBX )
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m2dnrv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m2dnrv12_top) )
-
-
-LLBL(p3m2dnrv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m3dv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m3dv12_top) )
-
-
-LLBL(p3m3dv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_no_rot_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_no_rot_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dnrv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dnrv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F2 F1 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F4 F1 F2 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 F2 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 F2 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 F2 */
-	FLD_S( M(14) ) 	/* F6 F4 F5 F2 */
-	FXCH( ST(3) ) 	/* F2 F4 F5 F6 */
-	FADDP( ST(0), ST(3) ) 	/* F4 F5 F6 */
-
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m3dnrv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m3dnrv12_top) )
-
-
-LLBL(p3m3dnrv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_perspective_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_perspective_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mpv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mpv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F2 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F2 F5 F4 */
-	FLD_S( M(14) ) 	/* F6 F2 F5 F4 */
-	FXCH( ST(1) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p3mpv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3mpv12_top) )
-
-
-LLBL(p3mpv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_general_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_general_v12):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mgv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mgv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(14) )
-	FLD_S( S(3) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(15) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FSTP_S( D(0)   ) 	/* F6 F5 F7 */
-	FXCH( ST(1) ) 	/* F5 F6 F7 */
-	FSTP_S( D(1)   ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p4mgv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4mgv12_top) )
-
-
-LLBL(p4mgv12_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_identity_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_identity_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4miv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p4miv12_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p4miv12_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EDX )
-
-	MOV_L( EBX, D(2) )
-	MOV_L( EDX, D(3) )
-LLBL(p4miv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4miv12_top) )
-
-
-LLBL(p4miv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_v12):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m2dv12_top) )
-
-
-LLBL(p4m2dv12_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_no_rot_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_no_rot_v12):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dnrv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dnrv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dnrv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m2dnrv12_top) )
-
-
-LLBL(p4m2dnrv12_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m3dv12_top) )
-
-
-LLBL(p4m3dv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_no_rot_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_no_rot_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dnrv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dnrv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dnrv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m3dnrv12_top) )
-
-
-LLBL(p4m3dnrv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_perspective_v12)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_perspective_v12):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mpv12_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mpv12_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F6 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F6 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F6 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F2 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4mpv12_skip):
-	ADD_L( CONST(48), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4mpv12_top) )
-
-
-LLBL(p4mpv12_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-
-
-/*
-########################################
-##
-## gl_x86_transform_points2_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_general_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_general_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mgv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2mgv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p2mgv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2mgv16_top) )
-
-
-LLBL(p2mgv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_identity_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_identity_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2miv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p2miv16_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p2miv16_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2miv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2miv16_top) )
-
-
-LLBL(p2miv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m2dv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m2dv16_top) )
-
-
-LLBL(p2m2dv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_2d_no_rot_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_2d_no_rot_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m2dnrv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m2dnrv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-
-	MOV_L( CONST(FP_ZERO), D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m2dnrv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m2dnrv16_top) )
-
-
-LLBL(p2m2dnrv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m3dv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m3dv16_top) )
-
-
-LLBL(p2m3dv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_3d_no_rot_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_3d_no_rot_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2m3dnrv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2m3dnrv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p2m3dnrv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2m3dnrv16_top) )
-
-
-LLBL(p2m3dnrv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points2_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points2_perspective_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points2_perspective_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p2mpv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	MOV_L( M(14), EBX )
-	ALIGNTEXT4ifNOP
-LLBL(p2mpv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FSTP_S( D(0)   ) 	/* F1 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( CONST(FP_ZERO), D(3) )
-
-LLBL(p2mpv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p2mpv16_top) )
-
-
-LLBL(p2mpv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_general_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_general_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mgv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mgv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FADD_S( M(12) )
-	FXCH( ST(2) ) 	/* F5 F6 F4 F7 */
-	FADD_S( M(13) )
-	FXCH( ST(1) ) 	/* F6 F5 F4 F7 */
-	FADD_S( M(14) )
-	FXCH( ST(3) ) 	/* F7 F5 F4 F6 */
-	FADD_S( M(15) )
-
-	FXCH( ST(2) ) 	/* F4 F5 F7 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F7 F6 */
-	FSTP_S( D(1)   ) 	/* F7 F6 */
-	FXCH( ST(1) ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p3mgv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3mgv16_top) )
-
-
-LLBL(p3mgv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_identity_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_identity_v16):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3miv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p3miv16_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p3miv16_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EBP )
-	MOV_L( S(2), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EBP, D(1) )
-	MOV_L( EDX, D(2) )
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3miv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3miv16_top) )
-
-
-LLBL(p3miv16_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 */
-	FADD_S( M(13) )
-
-	MOV_L( S(2), EBX )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m2dv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m2dv16_top) )
-
-
-LLBL(p3m2dv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_2d_no_rot_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_2d_no_rot_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m2dnrv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m2dnrv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F4 F1 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 */
-
-	FXCH( ST(2) ) 	/* F1 F4 F5 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 */
-
-	MOV_L( S(2), EBX )
-
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m2dnrv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m2dnrv16_top) )
-
-
-LLBL(p3m2dnrv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FADD_S( M(12) )
-	FXCH( ST(1) ) 	/* F5 F4 F6 */
-	FADD_S( M(13) )
-	FXCH( ST(2) ) 	/* F6 F4 F5 */
-	FADD_S( M(14) )
-
-	FXCH( ST(1) ) 	/* F4 F6 F5 */
-	FSTP_S( D(0)   ) 	/* F6 F5 */
-	FXCH( ST(1) ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m3dv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m3dv16_top) )
-
-
-LLBL(p3m3dv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_3d_no_rot_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_3d_no_rot_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3m3dnrv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3m3dnrv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F1 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F2 F1 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F4 F1 F2 */
-	FADD_S( M(12) )
-	FLD_S( M(13) ) 	/* F5 F4 F1 F2 */
-	FXCH( ST(2) ) 	/* F1 F4 F5 F2 */
-	FADDP( ST(0), ST(2) ) 	/* F4 F5 F2 */
-	FLD_S( M(14) ) 	/* F6 F4 F5 F2 */
-	FXCH( ST(3) ) 	/* F2 F4 F5 F6 */
-	FADDP( ST(0), ST(3) ) 	/* F4 F5 F6 */
-
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-
-	MOV_L( CONST(FP_ONE), D(3) )
-LLBL(p3m3dnrv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3m3dnrv16_top) )
-
-
-LLBL(p3m3dnrv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points3_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points3_perspective_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points3_perspective_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p3mpv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p3mpv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F2 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F2 F5 F4 */
-	FLD_S( M(14) ) 	/* F6 F2 F5 F4 */
-	FXCH( ST(1) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p3mpv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p3mpv16_top) )
-
-
-LLBL(p3mpv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_general
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_general_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_general_v16):
-
-#define FRAME_OFFSET 8
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mgv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mgv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-	FLD_S( S(0) ) 	/* F7 F6 F5 F4 */
-	FMUL_S( M(3) )
-
-	FLD_S( S(1) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(6) )
-	FLD_S( S(1) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(7) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(10) )
-	FLD_S( S(2) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(11) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F7 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(14) )
-	FLD_S( S(3) ) 	/* F3 F2 F1 F0 F7 F6 F5 F4 */
-	FMUL_S( M(15) )
-
-	FXCH( ST(3) ) 	/* F0 F2 F1 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(7) ) 	/* F2 F1 F3 F7 F6 F5 F4 */
-	FXCH( ST(1) ) 	/* F1 F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F2 F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F3 F7 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F7 F6 F5 F4 */
-
-	FXCH( ST(3) ) 	/* F4 F6 F5 F7 */
-	FSTP_S( D(0)   ) 	/* F6 F5 F7 */
-	FXCH( ST(1) ) 	/* F5 F6 F7 */
-	FSTP_S( D(1)   ) 	/* F6 F7 */
-	FSTP_S( D(2)   ) 	/* F7 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(p4mgv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4mgv16_top) )
-
-
-LLBL(p4mgv16_finish):
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_identity
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_identity_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_identity_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4miv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(p4miv16_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(p4miv16_top):
-
-	MOV_L( S(0), EBX )
-	MOV_L( S(1), EDX )
-
-	MOV_L( EBX, D(0) )
-	MOV_L( EDX, D(1) )
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EDX )
-
-	MOV_L( EBX, D(2) )
-	MOV_L( EDX, D(3) )
-LLBL(p4miv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4miv16_top) )
-
-
-LLBL(p4miv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_v16):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-
-	FLD_S( S(1) ) 	/* F0 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(5) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m2dv16_top) )
-
-
-LLBL(p4m2dv16_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_2d_no_rot_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_2d_no_rot_v16):
-
-#define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m2dnrv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m2dnrv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(3) ) 	/* F0 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(13) )
-
-	FXCH( ST(1) ) 	/* F0 F1 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F1 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F5 F4 */
-
-	MOV_L( S(2), EBX )
-	MOV_L( S(3), EBP )
-
-	FXCH( ST(1) ) 	/* F4 F5 */
-	FSTP_S( D(0)   ) 	/* F5 */
-	FSTP_S( D(1)   ) 	/* */
-	MOV_L( EBX, D(2) )
-	MOV_L( EBP, D(3) )
-
-LLBL(p4m2dnrv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m2dnrv16_top) )
-
-
-LLBL(p4m2dnrv16_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-	FLD_S( S(0) ) 	/* F5 F4 */
-	FMUL_S( M(1) )
-	FLD_S( S(0) ) 	/* F6 F5 F4 */
-	FMUL_S( M(2) )
-
-	FLD_S( S(1) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(4) )
-	FLD_S( S(1) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(5) )
-	FLD_S( S(1) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(6) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(2) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m3dv16_top) )
-
-
-LLBL(p4m3dv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d_no_rot
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_3d_no_rot_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_3d_no_rot_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4m3dnrv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4m3dnrv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F6 F5 F4 */
-	FMUL_S( M(10) )
-
-	FLD_S( S(3) ) 	/* F0 F6 F5 F4 */
-	FMUL_S( M(12) )
-	FLD_S( S(3) ) 	/* F1 F0 F6 F5 F4 */
-	FMUL_S( M(13) )
-	FLD_S( S(3) ) 	/* F2 F1 F0 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(5) ) 	/* F1 F2 F6 F5 F4 */
-	FADDP( ST(0), ST(3) ) 	/* F2 F6 F5 F4 */
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(3), EBX )
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4m3dnrv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4m3dnrv16_top) )
-
-
-LLBL(p4m3dnrv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-/*
-########################################
-##
-## gl_x86_transform_points4_perspective
-##
-##
-*/
-	GLOBL GLNAME(gl_x86_transform_points4_perspective_v16)
-	ALIGNTEXT4
-GLNAME(gl_x86_transform_points4_perspective_v16):
-
-#define FRAME_OFFSET 12
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(p4mpv16_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-
-
-	ALIGNTEXT4ifNOP
-LLBL(p4mpv16_top):
-
-
-	FLD_S( S(0) ) 	/* F4 */
-	FMUL_S( M(0) )
-
-	FLD_S( S(1) ) 	/* F5 F4 */
-	FMUL_S( M(5) )
-
-	FLD_S( S(2) ) 	/* F0 F5 F4 */
-	FMUL_S( M(8) )
-	FLD_S( S(2) ) 	/* F1 F0 F5 F4 */
-	FMUL_S( M(9) )
-	FLD_S( S(2) ) 	/* F6 F1 F0 F5 F4 */
-	FMUL_S( M(10) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F6 F5 F4 */
-	FADDP( ST(0), ST(4) ) 	/* F1 F6 F5 F4 */
-	FADDP( ST(0), ST(2) ) 	/* F6 F5 F4 */
-
-	FLD_S( S(3) ) 	/* F2 F6 F5 F4 */
-	FMUL_S( M(14) )
-
-	FADDP( ST(0), ST(1) ) 	/* F6 F5 F4 */
-
-	MOV_L( S(2), EBX )
-	XOR_L( CONST(-2147483648), EBX ) 	/* change sign */
-
-	FXCH( ST(2) ) 	/* F4 F5 F6 */
-	FSTP_S( D(0)   ) 	/* F5 F6 */
-	FSTP_S( D(1)   ) 	/* F6 */
-	FSTP_S( D(2)   ) 	/* */
-	MOV_L( EBX, D(3) )
-
-LLBL(p4mpv16_skip):
-	ADD_L( CONST(64), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(p4mpv16_top) )
-
-
-LLBL(p4mpv16_finish):
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-#undef FRAME_OFFSET
-
-
-
-#undef OFFSET_DEST
-#undef OFFSET_MATRIX
-#undef OFFSET_SOURCE
-#undef OFFSET_STRIDE
-#undef OFFSET_COUNT
-
-#undef ARG_DEST
-#undef ARG_MATRIX
-#undef ARG_SOURCE
-#undef ARG_STRIDE
-#undef ARG_COUNT
-
-
-/*
- * Table for clip test.
- *
- * 	bit6 = S(3) < 0
- * 	bit5 = S(2) < 0
- * 	bit4 = abs(S(2)) > abs(S(3))
- * 	bit3 = S(1) < 0
- * 	bit2 = abs(S(1)) > abs(S(3))
- * 	bit1 = S(0) < 0
- * 	bit0 = abs(S(0)) > abs(S(3))
- */
-
-
-	SEG_DATA
-
-clip_table:
-	D_BYTE 0, 1, 0, 2, 4, 5, 4, 6
-	D_BYTE 0, 1, 0, 2, 8, 9, 8, 10
-	D_BYTE 32, 33, 32, 34, 36, 37, 36, 38
-	D_BYTE 32, 33, 32, 34, 40, 41, 40, 42
-	D_BYTE 0, 1, 0, 2, 4, 5, 4, 6
-	D_BYTE 0, 1, 0, 2, 8, 9, 8, 10
-	D_BYTE 16, 17, 16, 18, 20, 21, 20, 22
-	D_BYTE 16, 17, 16, 18, 24, 25, 24, 26
-	D_BYTE 63, 61, 63, 62, 55, 53, 55, 54
-	D_BYTE 63, 61, 63, 62, 59, 57, 59, 58
-	D_BYTE 47, 45, 47, 46, 39, 37, 39, 38
-	D_BYTE 47, 45, 47, 46, 43, 41, 43, 42
-	D_BYTE 63, 61, 63, 62, 55, 53, 55, 54
-	D_BYTE 63, 61, 63, 62, 59, 57, 59, 58
-	D_BYTE 31, 29, 31, 30, 23, 21, 23, 22
-	D_BYTE 31, 29, 31, 30, 27, 25, 27, 26
-
-	SEG_TEXT
-
-/*
- *	Offsets for clip_func arguments
- *
- *	typedef GLvector4f *(*clip_func)( GLvector4f *vClip, 
- *	                                  GLvector4f *vProj, 
- *	                                  GLubyte clipMask[],
- *	                                  GLubyte *orMask, 
- *	                                  GLubyte *andMask );
- */
-
-#define OFFSET_SOURCE 4
-#define OFFSET_DEST 8
-#define OFFSET_CLIP 12
-#define OFFSET_OR 16
-#define OFFSET_AND 20
-
-#define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-#define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-#define ARG_CLIP 	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
-#define ARG_OR 	REGOFF(FRAME_OFFSET+OFFSET_OR, ESP)
-#define ARG_AND 	REGOFF(FRAME_OFFSET+OFFSET_AND, ESP)
-
-/*
-########################################
-##
-## gl_x86_cliptest_points4
-##
-##   AL:  ormask
-##   AH:  andmask
-##   EBX: temp0
-##   ECX: temp1
-##   EDX: clipmask[]
-##   ESI: clip[]
-##   EDI: proj[]
-##   EBP: temp2
-##
-########################################
-*/
-
-#if defined(__ELF__) && defined(__PIC__) && !defined(ELFPIC)
-#define ELFPIC
-#endif
-
-	GLOBL GLNAME(gl_x86_cliptest_points4)
-	ALIGNTEXT4
-
-GLNAME(gl_x86_cliptest_points4):
-#ifdef ELFPIC
-#define FRAME_OFFSET 20
-#else
-#define FRAME_OFFSET 16
-#endif
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-	PUSH_L( EBP )
-	PUSH_L( EBX )
-
-#ifdef ELFPIC
-	/* store pointer to clip_table on stack */
-	CALL( LLBL(ctp4_get_eip) )
-	ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX )
-	MOV_L( REGOFF(clip_table@GOT, EBX), EBX )
-	PUSH_L( EBX )
-	JMP( LLBL(ctp4_clip_table_ready) )
-
-LLBL(ctp4_get_eip):
-	/* store eip in ebx */
-	MOV_L( REGIND(ESP), EBX )
-	RET
-
-LLBL(ctp4_clip_table_ready):
-#endif
-
-	MOV_L( ARG_SOURCE, ESI )
-	MOV_L( ARG_DEST, EDI )
-
-	MOV_L( ARG_CLIP, EDX )
-	MOV_L( ARG_OR, EBX )
-
-	MOV_L( ARG_AND, EBP )
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* source stride */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-	MOV_L( EAX, ARG_SOURCE ) 		/* put stride in ARG_SOURCE */
-
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDX, ECX )
-
-	MOV_L( ECX, ARG_CLIP ) 			/* put clipmask + count in ARG_CLIP */
-	CMP_L( ECX, EDX )
-
-	MOV_B( REGIND(EBX), AL )
-	MOV_B( REGIND(EBP), AH )
-
-	JZ( LLBL(ctp4_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(ctp4_top):
-	FLD1 	/* F3 */
-	FDIV_S( S(3) )
-
-	MOV_L( S(3), EBP )
-	MOV_L( S(2), EBX )
-
-	XOR_L( ECX, ECX )
-	ADD_L( EBP, EBP )	/* ebp = abs(S(3))*2 ; carry = sign of S(3) */
-
-	ADC_L( ECX, ECX )
-	ADD_L( EBX, EBX )	/* ebx = abs(S(2))*2 ; carry = sign of S(2) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(2))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(1), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(1))*2 ; carry = sign of S(1) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(1))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(0), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(0))*2 ; carry = sign of S(0) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(0))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-
-#ifdef ELFPIC
-	MOV_L( REGIND(ESP), EBP ) 	/* clip_table */
-
-	MOV_B( REGBI(EBP, ECX), CL )
-#else
-	MOV_B( REGOFF(clip_table,ECX), CL )
-#endif
-
-	OR_B( CL, AL )
-	AND_B( CL, AH )
-
-	TEST_B( CL, CL )
-	MOV_B( CL, REGIND(EDX) )
-
-	JZ( LLBL(ctp4_proj) )
-
-	FSTP( ST(0) ) 	/* */
-	JMP( LLBL(ctp4_next) )
-
-LLBL(ctp4_proj):
-	FLD_S( S(0) ) 	/* F0 F3 */
-	FMUL2( ST(1), ST(0) )
-
-	FLD_S( S(1) ) 	/* F1 F0 F3 */
-	FMUL2( ST(2), ST(0) )
-
-	FLD_S( S(2) ) 	/* F2 F1 F0 F3 */
-	FMUL2( ST(3), ST(0) )
-
-	FXCH( ST(2) ) 	/* F0 F1 F2 F3 */
-	FSTP_S( D(0)   ) 	/* F1 F2 F3 */
-	FSTP_S( D(1)   ) 	/* F2 F3 */
-	FSTP_S( D(2)   ) 	/* F3 */
-	FSTP_S( D(3)   ) 	/* */
-
-LLBL(ctp4_next):
-	INC_L( EDX )
-	ADD_L( CONST(16), EDI )
-
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( EDX, ARG_CLIP )
-
-	JNZ( LLBL(ctp4_top) )
-
-	MOV_L( ARG_OR, ECX )
-	MOV_L( ARG_AND, EDX )
-
-	MOV_B( AL, REGIND(ECX) )
-	MOV_B( AH, REGIND(EDX) )
-
-LLBL(ctp4_finish):
-	MOV_L( ARG_DEST, EAX )
-#ifdef ELFPIC
-	POP_L( ESI ) 	/* discard ptr to clip_table */
-#endif
-	POP_L( EBX )
-	POP_L( EBP )
-	POP_L( EDI )
-	POP_L( ESI )
-
-	RET
-
-
-
-
-/*
-########################################
-##
-## gl_v16_x86_cliptest_points4
-##
-## Performs cliptesting equivalent to that done by cliptest_v16()
-## in vertices.c
-##
-## This is a hacked version of the original above.
-##
-########################################
-*/
-
-
-
-#define OFFSET_V16_SOURCE 4
-#define OFFSET_V16_LAST 8
-#define OFFSET_V16_OR 12
-#define OFFSET_V16_AND 16
-#define OFFSET_V16_MASK 20
-
-#define ARG_V16_SOURCE	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_SOURCE, ESP)
-#define ARG_V16_LAST 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_LAST, ESP)
-#define ARG_V16_OR 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_OR, ESP)
-#define ARG_V16_AND 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_AND, ESP)
-#define ARG_V16_MASK 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_MASK, ESP)
-
-
-#if defined(__ELF__) && defined(__PIC__) && !defined(ELFPIC)
-#define ELFPIC
-#endif
-
-	GLOBL GLNAME(gl_v16_x86_cliptest_points4)
-	ALIGNTEXT4
-
-GLNAME(gl_v16_x86_cliptest_points4):
-#ifdef ELFPIC
-#define V16_FRAME_OFFSET 20
-#else
-#define V16_FRAME_OFFSET 16
-#endif
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-	PUSH_L( EBP )
-	PUSH_L( EBX )
-
-#ifdef ELFPIC
-	/* store pointer to clip_table on stack */
-	CALL( LLBL(v16_ctp4_get_eip) )
-	ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX )
-	MOV_L( REGOFF(clip_table@GOT, EBX), EBX )
-	PUSH_L( EBX )
-	JMP( LLBL(v16_ctp4_clip_table_ready) )
-
-LLBL(v16_ctp4_get_eip):
-	/* store eip in ebx */
-	MOV_L( REGIND(ESP), EBX )
-	RET
-
-LLBL(v16_ctp4_clip_table_ready):
-#endif
-
-	MOV_L( ARG_V16_SOURCE, ESI )            /* ptr to first source vertex */
-	MOV_L( ARG_V16_LAST, EDX ) 	        /* ptr to last source vertex */
-	MOV_L( ARG_V16_OR, EBX )
-	MOV_L( ARG_V16_AND, EBP )
-	MOV_L( ARG_V16_MASK, EDI )
-
-	CMP_L( EDX, ESI )
-
-	MOV_B( REGIND(EBX), AL )
-	MOV_B( REGIND(EBP), AH )
-
-	JZ( LLBL(v16_ctp4_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(v16_ctp4_top):
-#if 0
-	FLD1 	/* F0 */
-	FDIV_S( S(3) )
-#endif
-
-	MOV_L( S(3), EBP )
-	MOV_L( S(2), EBX )
-
-	XOR_L( ECX, ECX )
-	ADD_L( EBP, EBP )	/* ebp = abs(S(3))*2 ; carry = sign of S(3) */
-
-	ADC_L( ECX, ECX )
-	ADD_L( EBX, EBX )	/* ebx = abs(S(2))*2 ; carry = sign of S(2) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(2))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(1), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(1))*2 ; carry = sign of S(1) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(1))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(0), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(0))*2 ; carry = sign of S(0) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(0))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-
-#ifdef ELFPIC
-	MOV_L( REGIND(ESP), EBP ) 	/* clip_table */
-
-	MOV_B( REGBI(EBP, ECX), CL )
-#else
-	MOV_B( REGOFF(clip_table,ECX), CL )
-#endif
-
-	OR_B( CL, AL )
-	AND_B( CL, AH )
-
-	MOV_B( CL, REGIND(EDI) )     /* save clipmask */
-	INC_L( EDI )                 /* next clipmask */
-
-#if 0
-	FSTP_S( S(8)   ) 	/* */ 	/* GR_VERTEX_OOW_OFFSET */
-#endif
-
-	ADD_L( CONST(64), ESI )      /* next fxVertex  */
-
-	CMP_L( EDX, ESI )            /* finished? */
-	JNZ( LLBL(v16_ctp4_top) )
-
-	MOV_L( ARG_V16_OR, ECX )
-	MOV_L( ARG_V16_AND, EDX )
-
-	MOV_B( AL, REGIND(ECX) )
-	MOV_B( AH, REGIND(EDX) )
-
-LLBL(v16_ctp4_finish):
-
-#ifdef ELFPIC
-	POP_L( ESI ) 	/* discard ptr to clip_table */
-#endif
-	POP_L( EBX )
-	POP_L( EBP )
-	POP_L( EDI )
-	POP_L( ESI )
-
-	RET
-
diff --git a/xc/extras/Mesa/src/X86/x86a.S.m4 b/xc/extras/Mesa/src/X86/x86a.S.m4
deleted file mode 100644
index 7b543fce6..000000000
--- a/xc/extras/Mesa/src/X86/x86a.S.m4
+++ /dev/null
@@ -1,1751 +0,0 @@
-include(x86flatregs.m4)dnl
-dnl ############################################################
-dnl ## output_x86_transform_functions
-dnl ##
-dnl ## This macro is called twice.  Once with mask checking
-dnl ## versions of FUNCTION_HEAD, LOOP_HEAD, LOOP_TAIL and
-dnl ## FUNCTION_TAIL, and once with versions that don't check
-dnl ## the clip mask.
-dnl ##
-dnl ## The parameters for FUNCTION_HEAD are:
-dnl ## - The function's name.  The mask checking version will
-dnl ##   append "_masked" to this, the unmasked version appends
-dnl ##   "_raw".
-dnl ## - A unique string (for labels).
-dnl ## - The number of integer registers needed.  Can be 0-2.
-dnl ##   If not zero, IREG_0 and/or IREG_1 will be defined.
-dnl ## - The output vector size.
-dnl ############################################################
-define(`output_x86_transform_functions', `dnl
-/*
-########################################
-##
-## gl_x86_transform_points2_general
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_general, p2mg, 0, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-	F_FLD_S( S(0), F6 )
-	F_FMUL_S( M(2), F6 )
-	F_FLD_S( S(0), F7 )
-	F_FMUL_S( M(3), F7 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-	F_FLD_S( S(1), F2 )
-	F_FMUL_S( M(6), F2 )
-	F_FLD_S( S(1), F3 )
-	F_FMUL_S( M(7), F3 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-	F_FADDP( F3, F7 )
-
-	F_FADD_S( M(12), F4 )
-	F_FADD_S( M(13), F5 )
-	F_FADD_S( M(14), F6 )
-	F_FADD_S( M(15), F7 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	F_FSTP_S( F7, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points2_identity
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_identity, p2mi, 1, 2)
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(SHORT_NAME`'_finish) )
-
-LOOP_HEAD
-	MOV_L( S(0), IREG_0 )
-	MOV_L( S(1), EDX )
-
-	MOV_L( IREG_0, D(0) )
-	MOV_L( EDX, D(1) )
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_2d, p2m2d, 0, 2)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-
-	F_FADD_S( M(12), F4 )
-	F_FADD_S( M(13), F5 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points2_2d_no_rot
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_2d_no_rot, p2m2dnr, 0, 2)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FADD_S( M(12), F4 )
-	F_FLD_S( M(13), F5 )
-	F_FADDP( F1, F5 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_3d, p2m3d, 0, 3)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-	F_FLD_S( S(0), F6 )
-	F_FMUL_S( M(2), F6 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-	F_FLD_S( S(1), F2 )
-	F_FMUL_S( M(6), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	F_FADD_S( M(12), F4 )
-	F_FADD_S( M(13), F5 )
-	F_FADD_S( M(14), F6 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points2_3d_no_rot
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_3d_no_rot, p2m3dnr, 1, 3)
-
-	MOV_L( M(14), IREG_0 )
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FADD_S( M(12), F4 )
-	F_FLD_S( M(13), F5 )
-	F_FADDP( F1, F5 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	MOV_L( IREG_0, D(2) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points2_perspective
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points2_perspective, p2mp, 1, 4)
-
-	MOV_L( M(14), IREG_0 )
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F1, D(1) )
-	MOV_L( IREG_0, D(2) )
-	MOV_L( CONST(FP_ZERO), D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_general
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_general, p3mg, 0, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-	F_FLD_S( S(0), F6 )
-	F_FMUL_S( M(2), F6 )
-	F_FLD_S( S(0), F7 )
-	F_FMUL_S( M(3), F7 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-	F_FLD_S( S(1), F2 )
-	F_FMUL_S( M(6), F2 )
-	F_FLD_S( S(1), F3 )
-	F_FMUL_S( M(7), F3 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-	F_FADDP( F3, F7 )
-
-	F_FLD_S( S(2), F0 )
-	F_FMUL_S( M(8), F0 )
-	F_FLD_S( S(2), F1 )
-	F_FMUL_S( M(9), F1 )
-	F_FLD_S( S(2), F2 )
-	F_FMUL_S( M(10), F2 )
-	F_FLD_S( S(2), F3 )
-	F_FMUL_S( M(11), F3 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-	F_FADDP( F3, F7 )
-
-	F_FADD_S( M(12), F4 )
-	F_FADD_S( M(13), F5 )
-	F_FADD_S( M(14), F6 )
-	F_FADD_S( M(15), F7 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	F_FSTP_S( F7, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_identity
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_identity, p3mi, 2, 3)
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(SHORT_NAME`'_finish) )
-
-LOOP_HEAD
-	MOV_L( S(0), IREG_0 )
-	MOV_L( S(1), IREG_1 )
-	MOV_L( S(2), EDX )
-
-	MOV_L( IREG_0, D(0) )
-	MOV_L( IREG_1, D(1) )
-	MOV_L( EDX, D(2) )
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_2d, p3m2d, 1, 3)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-
-	F_FADD_S( M(12), F4 )
-	F_FADD_S( M(13), F5 )
-
-	MOV_L( S(2), IREG_0 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	MOV_L( IREG_0, D(2) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_2d_no_rot
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_2d_no_rot, p3m2dnr, 1, 3)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FADD_S( M(12), F4 )
-	F_FLD_S( M(13), F5 )
-
-	F_FADDP( F1, F5 )
-
-	MOV_L( S(2), IREG_0 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	MOV_L( IREG_0, D(2) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_3d, p3m3d, 0, 3)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-	F_FLD_S( S(0), F6 )
-	F_FMUL_S( M(2), F6 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-	F_FLD_S( S(1), F2 )
-	F_FMUL_S( M(6), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	F_FLD_S( S(2), F0 )
-	F_FMUL_S( M(8), F0 )
-	F_FLD_S( S(2), F1 )
-	F_FMUL_S( M(9), F1 )
-	F_FLD_S( S(2), F2 )
-	F_FMUL_S( M(10), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	F_FADD_S( M(12), F4 )
-	F_FADD_S( M(13), F5 )
-	F_FADD_S( M(14), F6 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_3d_no_rot
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_3d_no_rot, p3m3dnr, 0, 3)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FLD_S( S(2), F2 )
-	F_FMUL_S( M(10), F2 )
-
-	F_FADD_S( M(12), F4 )
-	F_FLD_S( M(13), F5 )
-	F_FADDP( F1, F5 )
-	F_FLD_S( M(14), F6 )
-	F_FADDP( F2, F6 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points3_perspective
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points3_perspective, p3mp, 1, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F5 )
-	F_FMUL_S( M(5), F5 )
-
-	F_FLD_S( S(2), F0 )
-	F_FMUL_S( M(8), F0 )
-	F_FLD_S( S(2), F1 )
-	F_FMUL_S( M(9), F1 )
-	F_FLD_S( S(2), F2 )
-	F_FMUL_S( M(10), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FLD_S( M(14), F6 )
-	F_FADDP( F2, F6 )
-
-	MOV_L( S(2), IREG_0 )
-	XOR_L( CONST(eval(0x80000000)), IREG_0 ) 	/* change sign */
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	MOV_L( IREG_0, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_general
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_general, p4mg, 0, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-	F_FLD_S( S(0), F6 )
-	F_FMUL_S( M(2), F6 )
-	F_FLD_S( S(0), F7 )
-	F_FMUL_S( M(3), F7 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-	F_FLD_S( S(1), F2 )
-	F_FMUL_S( M(6), F2 )
-	F_FLD_S( S(1), F3 )
-	F_FMUL_S( M(7), F3 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-	F_FADDP( F3, F7 )
-
-	F_FLD_S( S(2), F0 )
-	F_FMUL_S( M(8), F0 )
-	F_FLD_S( S(2), F1 )
-	F_FMUL_S( M(9), F1 )
-	F_FLD_S( S(2), F2 )
-	F_FMUL_S( M(10), F2 )
-	F_FLD_S( S(2), F3 )
-	F_FMUL_S( M(11), F3 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-	F_FADDP( F3, F7 )
-
-	F_FLD_S( S(3), F0 )
-	F_FMUL_S( M(12), F0 )
-	F_FLD_S( S(3), F1 )
-	F_FMUL_S( M(13), F1 )
-	F_FLD_S( S(3), F2 )
-	F_FMUL_S( M(14), F2 )
-	F_FLD_S( S(3), F3 )
-	F_FMUL_S( M(15), F3 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-	F_FADDP( F3, F7 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	F_FSTP_S( F7, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_identity
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_identity, p4mi, 1, 4)
-
-	CMP_L( ESI, EDI )
-	JE( LLBL(SHORT_NAME`'_finish) )
-
-LOOP_HEAD
-	MOV_L( S(0), IREG_0 )
-	MOV_L( S(1), EDX )
-
-	MOV_L( IREG_0, D(0) )
-	MOV_L( EDX, D(1) )
-
-	MOV_L( S(2), IREG_0 )
-	MOV_L( S(3), EDX )
-
-	MOV_L( IREG_0, D(2) )
-	MOV_L( EDX, D(3) )
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_2d, p4m2d, 2, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-
-	F_FLD_S( S(3), F0 )
-	F_FMUL_S( M(12), F0 )
-	F_FLD_S( S(3), F1 )
-	F_FMUL_S( M(13), F1 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-
-	MOV_L( S(2), IREG_0 )
-	MOV_L( S(3), IREG_1 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	MOV_L( IREG_0, D(2) )
-	MOV_L( IREG_1, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_2d_no_rot
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_2d_no_rot, p4m2dnr, 2, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F5 )
-	F_FMUL_S( M(5), F5 )
-
-	F_FLD_S( S(3), F0 )
-	F_FMUL_S( M(12), F0 )
-	F_FLD_S( S(3), F1 )
-	F_FMUL_S( M(13), F1 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-
-	MOV_L( S(2), IREG_0 )
-	MOV_L( S(3), IREG_1 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	MOV_L( IREG_0, D(2) )
-	MOV_L( IREG_1, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_3d, p4m3d, 1, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-	F_FLD_S( S(0), F5 )
-	F_FMUL_S( M(1), F5 )
-	F_FLD_S( S(0), F6 )
-	F_FMUL_S( M(2), F6 )
-
-	F_FLD_S( S(1), F0 )
-	F_FMUL_S( M(4), F0 )
-	F_FLD_S( S(1), F1 )
-	F_FMUL_S( M(5), F1 )
-	F_FLD_S( S(1), F2 )
-	F_FMUL_S( M(6), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	F_FLD_S( S(2), F0 )
-	F_FMUL_S( M(8), F0 )
-	F_FLD_S( S(2), F1 )
-	F_FMUL_S( M(9), F1 )
-	F_FLD_S( S(2), F2 )
-	F_FMUL_S( M(10), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	F_FLD_S( S(3), F0 )
-	F_FMUL_S( M(12), F0 )
-	F_FLD_S( S(3), F1 )
-	F_FMUL_S( M(13), F1 )
-	F_FLD_S( S(3), F2 )
-	F_FMUL_S( M(14), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	MOV_L( S(3), IREG_0 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	MOV_L( IREG_0, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_3d_no_rot
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_3d_no_rot, p4m3dnr, 1, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F5 )
-	F_FMUL_S( M(5), F5 )
-
-	F_FLD_S( S(2), F6 )
-	F_FMUL_S( M(10), F6 )
-
-	F_FLD_S( S(3), F0 )
-	F_FMUL_S( M(12), F0 )
-	F_FLD_S( S(3), F1 )
-	F_FMUL_S( M(13), F1 )
-	F_FLD_S( S(3), F2 )
-	F_FMUL_S( M(14), F2 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-	F_FADDP( F2, F6 )
-
-	MOV_L( S(3), IREG_0 )
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	MOV_L( IREG_0, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-/*
-########################################
-##
-## gl_x86_transform_points4_perspective
-##
-##
-*/
-FUNCTION_HEAD(gl_x86_transform_points4_perspective, p4mp, 1, 4)
-
-LOOP_HEAD
-
-	F_FLD_S( S(0), F4 )
-	F_FMUL_S( M(0), F4 )
-
-	F_FLD_S( S(1), F5 )
-	F_FMUL_S( M(5), F5 )
-
-	F_FLD_S( S(2), F0 )
-	F_FMUL_S( M(8), F0 )
-	F_FLD_S( S(2), F1 )
-	F_FMUL_S( M(9), F1 )
-	F_FLD_S( S(2), F6 )
-	F_FMUL_S( M(10), F6 )
-
-	F_FADDP( F0, F4 )
-	F_FADDP( F1, F5 )
-
-	F_FLD_S( S(3), F2 )
-	F_FMUL_S( M(14), F2 )
-
-	F_FADDP( F2, F6 )
-
-	MOV_L( S(2), IREG_0 )
-	XOR_L( CONST(eval(0x80000000)), IREG_0 ) 	/* change sign */
-
-	F_FSTP_S( F4, D(0) )
-	F_FSTP_S( F5, D(1) )
-	F_FSTP_S( F6, D(2) )
-	MOV_L( IREG_0, D(3) )
-
-LOOP_TAIL
-
-FUNCTION_TAIL
-
-')dnl
-#include "assyntax.h"
-
-	SEG_TEXT
-
-`#'define FP_ONE eval(0x3f800000)
-`#'define FP_ZERO 0
-
-`#'if !defined(NASM_ASSEMBLER) && !defined(MASM_ASSEMBLER)
-`#'define LLBL(a) .L ## a
-`#'else
-`#'define LLBL(a) a
-`#'endif
-
-`#'define S(i) 	REGOFF(i * 4, ESI)
-`#'define D(i) 	REGOFF(i * 4, EDI)
-`#'define M(i) 	REGOFF(i * 4, EDX)
-
-/*
- * 	Offsets into GLvector4f
- */
-`#'define V4F_DATA 	0
-`#'define V4F_START 	4
-`#'define V4F_COUNT 	8
-`#'define V4F_STRIDE 	12
-`#'define V4F_SIZE 	16
-`#'define V4F_FLAGS 	20
-
-define(`VEC_DIRTY_0', 1)dnl
-define(`VEC_DIRTY_1', 2)dnl
-define(`VEC_DIRTY_2', 4)dnl
-define(`VEC_DIRTY_3', 8)dnl
-`#'define VEC_SIZE_1   eval(VEC_DIRTY_0)
-`#'define VEC_SIZE_2   eval(VEC_DIRTY_0|VEC_DIRTY_1)
-`#'define VEC_SIZE_3   eval(VEC_DIRTY_0|VEC_DIRTY_1|VEC_DIRTY_2)
-`#'define VEC_SIZE_4   eval(VEC_DIRTY_0|VEC_DIRTY_1|VEC_DIRTY_2|VEC_DIRTY_3)
-
-/*
- *	Offsets for transform_func arguments
- *
- *	typedef void (*transform_func)( GLvector4f *to_vec, 
- *	                                const GLmatrix *mat, 
- *	                                const GLvector4f *from_vec, 
- *	                                const GLubyte *clipmask,
- *	                                const GLubyte flag );
- */
-`#'define OFFSET_DEST 4
-`#'define OFFSET_MATRIX 8
-`#'define OFFSET_SOURCE 12
-`#'define OFFSET_CLIP 16
-`#'define OFFSET_FLAG 20
-
-`#'define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-`#'define ARG_MATRIX 	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
-`#'define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-`#'define ARG_CLIP 	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
-`#'define ARG_FLAG 	REGOFF(FRAME_OFFSET+OFFSET_FLAG, ESP)
-
-/*
-########################################
-##
-## Masked versions
-##
-########################################
-*/
-divert(-1)
-####################
-## FUNCTION_HEAD(
-##   long_name,
-##   short_name,
-##   regs_needed,
-##   output_vector_size
-## );
-##
-##   EAX: stride or temp0
-##   EBX: clipflag
-##   ECX: dst + count or temp1
-##   EDX: matrix
-##   ESI: src[]
-##   EDI: dst[]
-##   EBP: clipmask[]
-##
-####################
-define(`FUNCTION_HEAD', `dnl
-define(`LONG_NAME', $1`'_masked)dnl
-define(`SHORT_NAME', $2`'m)dnl
-define(`TEMP_REGS_NEEDED', $3)dnl
-define(`VEC_SIZE', $4)dnl
-	GLOBL GLNAME(LONG_NAME)
-	ALIGNTEXT4
-GLNAME(LONG_NAME):
-
-`#'define FRAME_OFFSET 16
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	PUSH_L( EBX )
-	PUSH_L( EBP )
-
-	MOV_L( ARG_SOURCE, ESI ) 		/* ptr to source GLvector4f */
-	MOV_L( ARG_DEST, EDI ) 			/* ptr to dest GLvector4f */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_CLIP, EBP ) 			/* ptr to clip mask array */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_B( ARG_FLAG, BL ) 			/* clip mask flags */
-
-	TEST_L( ECX, ECX )
-	JZ( LLBL(SHORT_NAME`'_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(`VEC_SIZE_'VEC_SIZE), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(VEC_SIZE), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-
-dnl	##########
-ifelse(eval(TEMP_REGS_NEEDED > 0), 1,
-`	MOV_L( EAX, ARG_SOURCE ) 		/* need eax; put stride in ARG_SOURCE */
-define(`IREG_0', `EAX')define(`STRIDE', `ARG_SOURCE')dnl
-',
-`define(`STRIDE', `EAX')')dnl
-dnl	##########
-ifelse(eval(TEMP_REGS_NEEDED > 1), 1,
-`	MOV_L( ECX, ARG_DEST ) 			/* need ecx; put dest+count in ARG_DEST */
-define(`IREG_1', `ECX')define(`DEST_END', `ARG_DEST')dnl
-',
-`define(`DEST_END', `ECX')')dnl
-dnl	##########
-')
-
-####################
-## LOOP_HEAD
-####################
-define(`LOOP_HEAD',
-`	ALIGNTEXT4ifNOP
-LLBL(SHORT_NAME`'_top):
-
-	TEST_B( BL, REGIND(EBP) )
-	JNZ( LLBL(SHORT_NAME`'_skip) )
-')
-
-####################
-## LOOP_TAIL
-####################
-define(`LOOP_TAIL',
-`LLBL(SHORT_NAME`'_skip):
-	INC_L( EBP )
-	ADD_L( CONST(16), EDI )
-	ADD_L( STRIDE, ESI )
-	CMP_L( DEST_END, EDI )
-	JNE( LLBL(SHORT_NAME`'_top) )
-')
-
-####################
-## FUNCTION_TAIL
-####################
-define(`FUNCTION_TAIL',
-`LLBL(SHORT_NAME`'_finish):
-	POP_L( EBP )
-	POP_L( EBX )
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-undefine(`LONG_NAME')dnl
-undefine(`SHORT_NAME')dnl
-undefine(`TEMP_REGS_NEEDED')dnl
-undefine(`VEC_SIZE')dnl
-undefine(`IREG_0')dnl
-undefine(`IREG_1')dnl
-undefine(`STRIDE')dnl
-undefine(`DEST_END')dnl
-`#'undef FRAME_OFFSET')
-
-####################
-divert
-output_x86_transform_functions
-
-/*
-########################################
-##
-## Unmasked versions
-##
-########################################
-*/
-divert(-1)
-####################
-## FUNCTION_HEAD(
-##   long_name,
-##   short_name,
-##   regs_needed,
-##   output_vector_size
-## );
-##
-##   EAX: stride
-##   EBX: temp0
-##   ECX: dst + count
-##   EDX: matrix
-##   ESI: src[]
-##   EDI: dst[]
-##   EBP: temp1
-##
-####################
-define(`FUNCTION_HEAD', `dnl
-define(`LONG_NAME', $1`'_raw)dnl
-define(`SHORT_NAME', $2`'r)dnl
-define(`TEMP_REGS_NEEDED', $3)dnl
-define(`VEC_SIZE', $4)dnl
-	GLOBL GLNAME(LONG_NAME)
-	ALIGNTEXT4
-GLNAME(LONG_NAME):
-
-`#'define FRAME_OFFSET eval(TEMP_REGS_NEEDED * 4 + 8)
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-	MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) 	/* ptr to source GLvector4f */
-	MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) 	/* ptr to dest GLvector4f */
-
-ifelse(eval(TEMP_REGS_NEEDED > 0), 1,
-`	PUSH_L( EBX )define(`IREG_0', EBX)
-')dnl
-ifelse(eval(TEMP_REGS_NEEDED > 1), 1,
-`	PUSH_L( EBP )define(`IREG_1', EBP)
-')dnl
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(SHORT_NAME`'_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* stride */
-	OR_L( CONST(`VEC_SIZE_'VEC_SIZE), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-	MOV_L( CONST(VEC_SIZE), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-
-	SHL_L( CONST(4), ECX ) 			/* count *= 16 */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDI, ECX ) 			/* count += dest ptr */
-')
-
-####################
-## LOOP_HEAD
-####################
-define(`LOOP_HEAD',
-`	ALIGNTEXT4ifNOP
-LLBL(SHORT_NAME`'_top):
-')
-
-####################
-## LOOP_TAIL
-####################
-define(`LOOP_TAIL',
-`LLBL(SHORT_NAME`'_skip):
-	ADD_L( CONST(16), EDI )
-	ADD_L( EAX, ESI )
-	CMP_L( ECX, EDI )
-	JNE( LLBL(SHORT_NAME`'_top) )
-')
-
-####################
-## FUNCTION_TAIL
-####################
-define(`FUNCTION_TAIL',
-`LLBL(SHORT_NAME`'_finish):
-ifelse(eval(TEMP_REGS_NEEDED > 1), 1,
-`	POP_L( EBP )
-')dnl
-ifelse(eval(TEMP_REGS_NEEDED > 0), 1,
-`	POP_L( EBX )
-')dnl
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-undefine(`LONG_NAME')dnl
-undefine(`SHORT_NAME')dnl
-undefine(`TEMP_REGS_NEEDED')dnl
-undefine(`VEC_SIZE')dnl
-undefine(`IREG_0')dnl
-undefine(`IREG_1')dnl
-`#'undef FRAME_OFFSET')
-
-
-####################
-divert
-output_x86_transform_functions
-
-`#'undef OFFSET_DEST
-`#'undef OFFSET_MATRIX
-`#'undef OFFSET_SOURCE
-`#'undef OFFSET_CLIP
-`#'undef OFFSET_FLAG
-
-`#'undef ARG_DEST
-`#'undef ARG_MATRIX
-`#'undef ARG_SOURCE
-`#'undef ARG_CLIP
-`#'undef ARG_FLAG
-
-`#'define OFFSET_DEST 4
-`#'define OFFSET_MATRIX 8
-`#'define OFFSET_SOURCE 12
-`#'define OFFSET_STRIDE 16
-`#'define OFFSET_COUNT 20
-
-`#'define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-`#'define ARG_MATRIX 	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
-`#'define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-`#'define ARG_STRIDE 	REGOFF(FRAME_OFFSET+OFFSET_STRIDE, ESP)
-`#'define ARG_COUNT 	REGOFF(FRAME_OFFSET+OFFSET_COUNT, ESP)
-
-/*
-########################################
-##
-## clean (no vertex struct) versions
-##
-########################################
-*/
-divert(-1)
-####################
-## FUNCTION_HEAD(
-##   long_name,
-##   short_name,
-##   regs_needed,
-##   output_vector_size
-## );
-##
-##   EAX: stride
-##   EBX: temp0
-##   ECX: dst + count
-##   EDX: matrix
-##   ESI: src[]
-##   EDI: dst[]
-##   EBP: temp1
-##
-####################
-define(`FUNCTION_HEAD', `dnl
-define(`LONG_NAME', $1`'_v`'DEST_STRIDE)dnl
-define(`SHORT_NAME', $2`'v`'DEST_STRIDE)dnl
-define(`TEMP_REGS_NEEDED', $3)dnl
-define(`VEC_SIZE', $4)dnl
-	GLOBL GLNAME(LONG_NAME)
-	ALIGNTEXT4
-GLNAME(LONG_NAME):
-
-`#'define FRAME_OFFSET eval(TEMP_REGS_NEEDED * 4 + 8)
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-
-ifelse(eval(TEMP_REGS_NEEDED > 0), 1,
-`	PUSH_L( EBX )define(`IREG_0', EBX)
-')dnl
-ifelse(eval(TEMP_REGS_NEEDED > 1), 1,
-`	PUSH_L( EBP )define(`IREG_1', EBP)
-')dnl
-
-	MOV_L( ARG_SOURCE, ESI ) 	/* ptr to source */
-	MOV_L( ARG_DEST, EDI ) 	/* ptr to dest */
-
-	MOV_L( ARG_MATRIX, EDX ) 		/* ptr to matrix */
-	MOV_L( ARG_COUNT, ECX ) 	/* count */
-
-	TEST_L( ECX, ECX)
-	JZ( LLBL(SHORT_NAME`'_finish) ) 	/* count was zero; go to finish */
-
-	MOV_L( ARG_STRIDE, EAX ) 	/* stride */
-')
-
-####################
-## LOOP_HEAD
-####################
-define(`LOOP_HEAD',
-`	ALIGNTEXT4ifNOP
-LLBL(SHORT_NAME`'_top):
-')
-
-####################
-## LOOP_TAIL
-####################
-define(`LOOP_TAIL',
-`ifelse(eval(VEC_SIZE == 1), 1, `	MOV_L( CONST(FP_ZERO), D(1) )
-')dnl
-ifelse(eval(VEC_SIZE <= 2), 1, `	MOV_L( CONST(FP_ZERO), D(2) )
-')dnl
-ifelse(eval(VEC_SIZE <= 3), 1, `	MOV_L( CONST(FP_ONE), D(3) )
-')dnl
-LLBL(SHORT_NAME`'_skip):
-	ADD_L( CONST(eval(DEST_STRIDE*4)), EDI )
-	ADD_L( EAX, ESI )
-	DEC_L( ECX )
-	JNZ( LLBL(SHORT_NAME`'_top) )
-')
-
-####################
-## FUNCTION_TAIL
-####################
-define(`FUNCTION_TAIL',
-`LLBL(SHORT_NAME`'_finish):
-ifelse(eval(TEMP_REGS_NEEDED > 1), 1,
-`	POP_L( EBP )
-')dnl
-ifelse(eval(TEMP_REGS_NEEDED > 0), 1,
-`	POP_L( EBX )
-')dnl
-	POP_L( EDI )
-	POP_L( ESI )
-	RET
-undefine(`LONG_NAME')dnl
-undefine(`SHORT_NAME')dnl
-undefine(`TEMP_REGS_NEEDED')dnl
-undefine(`VEC_SIZE')dnl
-undefine(`IREG_0')dnl
-undefine(`IREG_1')dnl
-`#'undef FRAME_OFFSET')
-
-
-####################
-divert
-
-define(`DEST_STRIDE', 8)dnl
-output_x86_transform_functions
-undefine(`DEST_STRIDE')dnl
-
-define(`DEST_STRIDE', 12)dnl
-output_x86_transform_functions
-undefine(`DEST_STRIDE')dnl
-
-define(`DEST_STRIDE', 16)dnl
-output_x86_transform_functions
-undefine(`DEST_STRIDE')dnl
-
-`#'undef OFFSET_DEST
-`#'undef OFFSET_MATRIX
-`#'undef OFFSET_SOURCE
-`#'undef OFFSET_STRIDE
-`#'undef OFFSET_COUNT
-
-`#'undef ARG_DEST
-`#'undef ARG_MATRIX
-`#'undef ARG_SOURCE
-`#'undef ARG_STRIDE
-`#'undef ARG_COUNT
-
-
-/*
- * Table for clip test.
- *
- * 	bit6 = S(3) < 0
- * 	bit5 = S(2) < 0
- * 	bit4 = abs(S(2)) > abs(S(3))
- * 	bit3 = S(1) < 0
- * 	bit2 = abs(S(1)) > abs(S(3))
- * 	bit1 = S(0) < 0
- * 	bit0 = abs(S(0)) > abs(S(3))
- */
-divert(-1)
-# Vertex buffer clipping flags (from vb.h)
-define(`CLIP_RIGHT_BIT',  `0x01')
-define(`CLIP_LEFT_BIT',   `0x02')
-define(`CLIP_TOP_BIT',    `0x04')
-define(`CLIP_BOTTOM_BIT', `0x08')
-define(`CLIP_NEAR_BIT',   `0x10')
-define(`CLIP_FAR_BIT',    `0x20')
-
-define(`MAGN_X', `(~((($1) & 1) - 1))')
-define(`SIGN_X', `(~(((($1) >> 1) & 1) - 1))')
-define(`MAGN_Y', `(~(((($1) >> 2) & 1) - 1))')
-define(`SIGN_Y', `(~(((($1) >> 3) & 1) - 1))')
-define(`MAGN_Z', `(~(((($1) >> 4) & 1) - 1))')
-define(`SIGN_Z', `(~(((($1) >> 5) & 1) - 1))')
-define(`SIGN_W', `(~(((($1) >> 6) & 1) - 1))')
-
-define(`CLIP_VALUE', `eval(
-	(CLIP_RIGHT_BIT
-	  & ((~SIGN_X($1) & SIGN_W($1))
-	     | (~SIGN_X($1) & ~SIGN_W($1) & MAGN_X($1))
-	     | (SIGN_X($1) & SIGN_W($1) & ~MAGN_X($1))))
-	 | (CLIP_LEFT_BIT
-	    & ((SIGN_X($1) & SIGN_W($1))
-	       | (~SIGN_X($1) & SIGN_W($1) & ~MAGN_X($1))
-	       | (SIGN_X($1) & ~SIGN_W($1) & MAGN_X($1))))
-	 | (CLIP_TOP_BIT
-	    & ((~SIGN_Y($1) & SIGN_W($1))
-	       | (~SIGN_Y($1) & ~SIGN_W($1) & MAGN_Y($1))
-	       | (SIGN_Y($1) & SIGN_W($1) & ~MAGN_Y($1))))
-	 | (CLIP_BOTTOM_BIT
-	    & ((SIGN_Y($1) & SIGN_W($1))
-	       | (~SIGN_Y($1) & SIGN_W($1) & ~MAGN_Y($1))
-	       | (SIGN_Y($1) & ~SIGN_W($1) & MAGN_Y($1))))
-	 | (CLIP_FAR_BIT
-	    & ((~SIGN_Z($1) & SIGN_W($1))
-	       | (~SIGN_Z($1) & ~SIGN_W($1) & MAGN_Z($1))
-	       | (SIGN_Z($1) & SIGN_W($1) & ~MAGN_Z($1))))
-	 | (CLIP_NEAR_BIT
-	    & ((SIGN_Z($1) & SIGN_W($1))
-	       | (~SIGN_Z($1) & SIGN_W($1) & ~MAGN_Z($1))
-	       | (SIGN_Z($1) & ~SIGN_W($1) & MAGN_Z($1))))
-)')
-define(`CLIP_VALUE8',
-`CLIP_VALUE($1 + 0), CLIP_VALUE($1 + 1), CLIP_VALUE($1 + 2), CLIP_VALUE($1 + 3), dnl
-CLIP_VALUE($1 + 4), CLIP_VALUE($1 + 5), CLIP_VALUE($1 + 6), CLIP_VALUE($1 + 7)')
-divert
-
-	SEG_DATA
-
-clip_table:
-	D_BYTE CLIP_VALUE8(0x00)
-	D_BYTE CLIP_VALUE8(0x08)
-	D_BYTE CLIP_VALUE8(0x10)
-	D_BYTE CLIP_VALUE8(0x18)
-	D_BYTE CLIP_VALUE8(0x20)
-	D_BYTE CLIP_VALUE8(0x28)
-	D_BYTE CLIP_VALUE8(0x30)
-	D_BYTE CLIP_VALUE8(0x38)
-	D_BYTE CLIP_VALUE8(0x40)
-	D_BYTE CLIP_VALUE8(0x48)
-	D_BYTE CLIP_VALUE8(0x50)
-	D_BYTE CLIP_VALUE8(0x58)
-	D_BYTE CLIP_VALUE8(0x60)
-	D_BYTE CLIP_VALUE8(0x68)
-	D_BYTE CLIP_VALUE8(0x70)
-	D_BYTE CLIP_VALUE8(0x78)
-
-	SEG_TEXT
-
-/*
- *	Offsets for clip_func arguments
- *
- *	typedef GLvector4f *(*clip_func)( GLvector4f *vClip, 
- *	                                  GLvector4f *vProj, 
- *	                                  GLubyte clipMask[],
- *	                                  GLubyte *orMask, 
- *	                                  GLubyte *andMask );
- */
-
-`#'define OFFSET_SOURCE 4
-`#'define OFFSET_DEST 8
-`#'define OFFSET_CLIP 12
-`#'define OFFSET_OR 16
-`#'define OFFSET_AND 20
-
-`#'define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
-`#'define ARG_DEST 	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
-`#'define ARG_CLIP 	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
-`#'define ARG_OR 	REGOFF(FRAME_OFFSET+OFFSET_OR, ESP)
-`#'define ARG_AND 	REGOFF(FRAME_OFFSET+OFFSET_AND, ESP)
-
-/*
-########################################
-##
-## gl_x86_cliptest_points4
-##
-##   AL:  ormask
-##   AH:  andmask
-##   EBX: temp0
-##   ECX: temp1
-##   EDX: clipmask[]
-##   ESI: clip[]
-##   EDI: proj[]
-##   EBP: temp2
-##
-########################################
-*/
-
-`#'if defined(__ELF__) && defined(__PIC__) && !defined(ELFPIC)
-`#'define ELFPIC
-`#'endif
-
-	GLOBL GLNAME(gl_x86_cliptest_points4)
-	ALIGNTEXT4
-
-GLNAME(gl_x86_cliptest_points4):
-`#'ifdef ELFPIC
-`#'define FRAME_OFFSET 20
-`#'else
-`#'define FRAME_OFFSET 16
-`#'endif
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-	PUSH_L( EBP )
-	PUSH_L( EBX )
-
-`#'ifdef ELFPIC
-	/* store pointer to clip_table on stack */
-	CALL( LLBL(ctp4_get_eip) )
-	ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX )
-	MOV_L( REGOFF(clip_table@GOT, EBX), EBX )
-	PUSH_L( EBX )
-	JMP( LLBL(ctp4_clip_table_ready) )
-
-LLBL(ctp4_get_eip):
-	/* store eip in ebx */
-	MOV_L( REGIND(ESP), EBX )
-	RET
-
-LLBL(ctp4_clip_table_ready):
-`#'endif
-
-	MOV_L( ARG_SOURCE, ESI )
-	MOV_L( ARG_DEST, EDI )
-
-	MOV_L( ARG_CLIP, EDX )
-	MOV_L( ARG_OR, EBX )
-
-	MOV_L( ARG_AND, EBP )
-	MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) 	/* source stride */
-
-	MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) 	/* source count */
-	MOV_L( REGOFF(V4F_START, ESI), ESI ) 	/* ptr to first source vertex */
-
-	OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) 	/* set dest flags */
-	MOV_L( EAX, ARG_SOURCE ) 		/* put stride in ARG_SOURCE */
-
-	MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) 	/* set dest size */
-	MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) 	/* set dest count */
-
-	MOV_L( REGOFF(V4F_START, EDI), EDI ) 	/* ptr to first dest vertex */
-	ADD_L( EDX, ECX )
-
-	MOV_L( ECX, ARG_CLIP ) 			/* put clipmask + count in ARG_CLIP */
-	CMP_L( ECX, EDX )
-
-	MOV_B( REGIND(EBX), AL )
-	MOV_B( REGIND(EBP), AH )
-
-	JZ( LLBL(ctp4_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(ctp4_top):
-	F_PUSH0( FLD1, F3 )
-	F_FDIV_S( S(3), F3 )
-
-	MOV_L( S(3), EBP )
-	MOV_L( S(2), EBX )
-
-	XOR_L( ECX, ECX )
-	ADD_L( EBP, EBP )	/* ebp = abs(S(3))*2 ; carry = sign of S(3) */
-
-	ADC_L( ECX, ECX )
-	ADD_L( EBX, EBX )	/* ebx = abs(S(2))*2 ; carry = sign of S(2) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(2))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(1), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(1))*2 ; carry = sign of S(1) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(1))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(0), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(0))*2 ; carry = sign of S(0) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(0))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-
-`#'ifdef ELFPIC
-	MOV_L( REGIND(ESP), EBP ) 	/* clip_table */
-
-	MOV_B( REGBI(EBP, ECX), CL )
-`#'else
-	MOV_B( REGOFF(clip_table,ECX), CL )
-`#'endif
-
-	OR_B( CL, AL )
-	AND_B( CL, AH )
-
-	TEST_B( CL, CL )
-	MOV_B( CL, REGIND(EDX) )
-
-	JZ( LLBL(ctp4_proj) )
-
-pushdef(`F_STACK', F_STACK)dnl
-	F_FSTP( F3, F3 )
-popdef(`F_STACK')dnl
-	JMP( LLBL(ctp4_next) )
-
-LLBL(ctp4_proj):
-	F_FLD_S( S(0), F0 )
-	F_FMUL2( F3, F0 )
-
-	F_FLD_S( S(1), F1 )
-	F_FMUL2( F3, F1 )
-
-	F_FLD_S( S(2), F2 )
-	F_FMUL2( F3, F2 )
-
-	F_FSTP_S( F0, D(0) )
-	F_FSTP_S( F1, D(1) )
-	F_FSTP_S( F2, D(2) )
-	F_FSTP_S( F3, D(3) )
-
-LLBL(ctp4_next):
-	INC_L( EDX )
-	ADD_L( CONST(16), EDI )
-
-	ADD_L( ARG_SOURCE, ESI )
-	CMP_L( EDX, ARG_CLIP )
-
-	JNZ( LLBL(ctp4_top) )
-
-	MOV_L( ARG_OR, ECX )
-	MOV_L( ARG_AND, EDX )
-
-	MOV_B( AL, REGIND(ECX) )
-	MOV_B( AH, REGIND(EDX) )
-
-LLBL(ctp4_finish):
-	MOV_L( ARG_DEST, EAX )
-`#'ifdef ELFPIC
-	POP_L( ESI ) 	/* discard ptr to clip_table */
-`#'endif
-	POP_L( EBX )
-	POP_L( EBP )
-	POP_L( EDI )
-	POP_L( ESI )
-
-	RET
-
-
-
-
-/*
-########################################
-##
-## gl_v16_x86_cliptest_points4
-##
-## Performs cliptesting equivalent to that done by cliptest_v16()
-## in vertices.c
-##
-## This is a hacked version of the original above.
-##
-########################################
-*/
-
-
-
-`#'define OFFSET_V16_SOURCE 4
-`#'define OFFSET_V16_LAST 8
-`#'define OFFSET_V16_OR 12
-`#'define OFFSET_V16_AND 16
-`#'define OFFSET_V16_MASK 20
-
-`#'define ARG_V16_SOURCE	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_SOURCE, ESP)
-`#'define ARG_V16_LAST 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_LAST, ESP)
-`#'define ARG_V16_OR 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_OR, ESP)
-`#'define ARG_V16_AND 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_AND, ESP)
-`#'define ARG_V16_MASK 	REGOFF(V16_FRAME_OFFSET+OFFSET_V16_MASK, ESP)
-
-
-`#'if defined(__ELF__) && defined(__PIC__) && !defined(ELFPIC)
-`#'define ELFPIC
-`#'endif
-
-	GLOBL GLNAME(gl_v16_x86_cliptest_points4)
-	ALIGNTEXT4
-
-GLNAME(gl_v16_x86_cliptest_points4):
-`#'ifdef ELFPIC
-`#'define V16_FRAME_OFFSET 20
-`#'else
-`#'define V16_FRAME_OFFSET 16
-`#'endif
-	PUSH_L( ESI )
-	PUSH_L( EDI )
-	PUSH_L( EBP )
-	PUSH_L( EBX )
-
-`#'ifdef ELFPIC
-	/* store pointer to clip_table on stack */
-	CALL( LLBL(v16_ctp4_get_eip) )
-	ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX )
-	MOV_L( REGOFF(clip_table@GOT, EBX), EBX )
-	PUSH_L( EBX )
-	JMP( LLBL(v16_ctp4_clip_table_ready) )
-
-LLBL(v16_ctp4_get_eip):
-	/* store eip in ebx */
-	MOV_L( REGIND(ESP), EBX )
-	RET
-
-LLBL(v16_ctp4_clip_table_ready):
-`#'endif
-
-	MOV_L( ARG_V16_SOURCE, ESI )            /* ptr to first source vertex */
-	MOV_L( ARG_V16_LAST, EDX ) 	        /* ptr to last source vertex */
-	MOV_L( ARG_V16_OR, EBX )
-	MOV_L( ARG_V16_AND, EBP )
-	MOV_L( ARG_V16_MASK, EDI )
-
-	CMP_L( EDX, ESI )
-
-	MOV_B( REGIND(EBX), AL )
-	MOV_B( REGIND(EBP), AH )
-
-	JZ( LLBL(v16_ctp4_finish) )
-
-	ALIGNTEXT4ifNOP
-LLBL(v16_ctp4_top):
-`#'if 0
-	F_PUSH0( FLD1, F0 )
-	F_FDIV_S( S(3), F0 )
-`#'endif
-
-	MOV_L( S(3), EBP )
-	MOV_L( S(2), EBX )
-
-	XOR_L( ECX, ECX )
-	ADD_L( EBP, EBP )	/* ebp = abs(S(3))*2 ; carry = sign of S(3) */
-
-	ADC_L( ECX, ECX )
-	ADD_L( EBX, EBX )	/* ebx = abs(S(2))*2 ; carry = sign of S(2) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(2))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(1), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(1))*2 ; carry = sign of S(1) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(1))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-	MOV_L( S(0), EBX )
-
-	ADD_L( EBX, EBX )	/* ebx = abs(S(0))*2 ; carry = sign of S(0) */
-
-	ADC_L( ECX, ECX )
-	CMP_L( EBX, EBP )	/* carry = abs(S(0))*2 > abs(S(3))*2 */
-
-	ADC_L( ECX, ECX )
-
-`#'ifdef ELFPIC
-	MOV_L( REGIND(ESP), EBP ) 	/* clip_table */
-
-	MOV_B( REGBI(EBP, ECX), CL )
-`#'else
-	MOV_B( REGOFF(clip_table,ECX), CL )
-`#'endif
-
-	OR_B( CL, AL )
-	AND_B( CL, AH )
-
-	MOV_B( CL, REGIND(EDI) )     /* save clipmask */
-	INC_L( EDI )                 /* next clipmask */
-
-`#'if 0
-	F_FSTP_S( F0, S(8) ) 	/* GR_VERTEX_OOW_OFFSET */
-`#'endif
-
-	ADD_L( CONST(64), ESI )      /* next fxVertex  */
-
-	CMP_L( EDX, ESI )            /* finished? */
-	JNZ( LLBL(v16_ctp4_top) )
-
-	MOV_L( ARG_V16_OR, ECX )
-	MOV_L( ARG_V16_AND, EDX )
-
-	MOV_B( AL, REGIND(ECX) )
-	MOV_B( AH, REGIND(EDX) )
-
-LLBL(v16_ctp4_finish):
-
-`#'ifdef ELFPIC
-	POP_L( ESI ) 	/* discard ptr to clip_table */
-`#'endif
-	POP_L( EBX )
-	POP_L( EBP )
-	POP_L( EDI )
-	POP_L( ESI )
-
-	RET
-
diff --git a/xc/extras/Mesa/src/X86/x86flatregs.m4 b/xc/extras/Mesa/src/X86/x86flatregs.m4
deleted file mode 100644
index f3cbf39b7..000000000
--- a/xc/extras/Mesa/src/X86/x86flatregs.m4
+++ /dev/null
@@ -1,200 +0,0 @@
-divert(-1)
-############################################################
-## These m4 macros make the x86 floating point register
-## stack look more like normal registers.
-##
-## They do *NOT* know about jump instructions or any other
-## form of conditional execution!
-############################################################
-
-########################################
-## define the stack
-########################################
-define(`F_STACK', `')
-
-########################################
-## print the stack
-########################################
-define(`F_SHOW_STACK', `patsubst(F_STACK, `[0-7]', `F`'\& ')')
-
-########################################
-## define the flat floating point regs
-########################################
-define(`F0', 0)
-define(`F1', 1)
-define(`F2', 2)
-define(`F3', 3)
-define(`F4', 4)
-define(`F5', 5)
-define(`F6', 6)
-define(`F7', 7)
-
-########################################
-## F_PUSH_STACK(REG)
-## push a register
-########################################
-define(`F_PUSH_STACK',
-`ifelse(len(F_STACK), 8, ` [ERROR - STACK OVERFLOW] ')dnl
-ifelse(eval(F_FIND_REG($1) != -1), 1, ` [ERROR - F`'$1 PUSHED TWICE] ')dnl
-define(`F_STACK', eval($1)`'F_STACK)')
-
-########################################
-## F_POP_STACK()
-## pop top of stack
-########################################
-define(`F_POP_STACK', 
-`ifelse(len(F_STACK), 0, ` [ERROR - STACK UNDERFLOW] ')dnl
-define(`F_STACK', substr(F_STACK, 1))')
-
-########################################
-## F_FIND_REG(REG)
-## find register
-########################################
-define(`F_FIND_REG', `index(F_STACK, eval($1))')
-
-########################################
-## F_XCH_TO_TOP(REG)
-## swap to top of stack
-########################################
-define(`F_XCH_TO_TOP',
-`define(`F_INDEX', F_FIND_REG($1))dnl
-ifelse(
-eval(F_INDEX > 0), 1,
-`FXCH( ST('F_FIND_REG($1)`) )dnl
-define(`F_STACK',
- translit(F_STACK,
- substr(F_STACK, 0, 1)`'eval($1),
- eval($1)`'substr(F_STACK, 0, 1)))F_PCOMMENT(`F$1 to top')
-	',
-eval(F_INDEX == -1), 1,
-` [ERROR - F`'$1 NOT ON STACK] ')')
-
-########################################
-## F_PUSH(INSN, SRC, DST)
-## SRC - memory
-## DST - register
-## execute an insn that pushes a register
-########################################
-define(`F_PUSH',
-`$1( $2 )F_PUSH_STACK($3)F_PCOMMENT(`$1 $2, F$3')')
-
-########################################
-## F_PUSHR(INSN, SRC, DST)
-## SRC - register
-## DST - register
-## execute an insn that pushes a register
-########################################
-define(`F_PUSHR',
-`$1( ST(F_FIND_REG($2) )F_PUSH_STACK($3)F_PCOMMENT(`$1 F$2, F$3')')
-
-########################################
-## F_PUSH0(INSN, DST)
-## DST - register
-## execute an insn that pushes a register
-########################################
-define(`F_PUSH0',
-`$1`'F_PUSH_STACK($2)F_PCOMMENT(`$1 F$3')')
-
-########################################
-## F_POP(INSN, SRC, DST)
-## SRC - register
-## DST - memory
-## execute an insn that pops a register
-########################################
-define(`F_POP',
-`F_XCH_TO_TOP($2)dnl
-$1( $3 )F_POP_STACK()F_PCOMMENT(`$1 F$2, $3')')
-
-########################################
-## F_POPR(INSN, SRC, DST)
-## SRC - register
-## DST - register
-## execute an insn that pops a register
-########################################
-define(`F_POPR',
-`F_XCH_TO_TOP($2)dnl
-$1( ST(0), ST(F_FIND_REG($3)) )F_POP_STACK()F_PCOMMENT(`$1 F$2, F$3')')
-
-########################################
-## F_POP1(INSN, SRC, DST)
-## SRC - register
-## DST - register
-## execute an insn that pops a register
-## with st(0) arg omitted
-########################################
-define(`F_POP1',
-`F_XCH_TO_TOP($2)dnl
-$1( ST(F_FIND_REG($3)) )F_POP_STACK()F_PCOMMENT(`$1 F$2, F$3')')
-
-########################################
-## F_EXEC(INSN, SRC, DST)
-## SRC - memory
-## DST - register
-## execute an insn that doesn't change the stack
-########################################
-define(`F_EXEC',
-`F_XCH_TO_TOP($3)dnl
-$1( $2 )F_COMMENT(`$1 $2, F$3')')
-
-########################################
-## F_EXECR(INSN, SRC, DST)
-## SRC - register
-## DST - register
-## execute an insn that doesn't change the stack
-########################################
-define(`F_EXECR',
-`F_XCH_TO_TOP($3)dnl
-$1( ST(F_FIND_REG($2)), ST(0) )F_COMMENT(`$1 F$2, F$3')')
-
-########################################
-## F_EXEC0(INSN, REG)
-## REG - register
-## execute an insn that doesn't change
-## the stack or have any arguments
-## besides the stack top.
-########################################
-define(`F_EXEC0',
-`F_XCH_TO_TOP($2)dnl
-$1`'F_COMMENT(`$1 F$2')')
-
-########################################
-## Some common fp operations -
-########################################
-define(`F_FADD2', `F_EXECR( FADD2, $1, $2 )')
-define(`F_FADD_S', `F_EXEC( FADD_S, $1, $2 )')
-define(`F_FADDP', `F_POPR( FADDP, $1, $2 )')
-
-define(`F_FDIV2', `F_EXECR( FDIV2, $1, $2 )')
-define(`F_FDIV_S', `F_EXEC( FDIV_S, $1, $2 )')
-define(`F_FDIVP', `F_POPR( FDIVP, $1, $2 )')
-
-define(`F_FDIVR2', `F_EXECR( FDIVR2, $1, $2 )')
-define(`F_FDIVR_S', `F_EXEC( FDIVR_S, $1, $2 )')
-define(`F_FDIVRP', `F_POPR( FDIVRP, $1, $2 )')
-
-define(`F_FLD_S', `F_PUSH( FLD_S, $1, $2 )')
-
-define(`F_FMUL2', `F_EXECR( FMUL2, $1, $2 )')
-define(`F_FMUL_S', `F_EXEC( FMUL_S, $1, $2 )')
-define(`F_FMULP', `F_POPR( FMULP, $1, $2 )')
-
-define(`F_FSTP_S', `F_POP( FSTP_S, $1, $2 )')
-define(`F_FSTP', `F_POP1( FSTP, $1, $2 )')
-
-define(`F_FSUB2', `F_EXECR( FSUB2, $1, $2 )')
-define(`F_FSUB_S', `F_EXEC( FSUB_S, $1, $2 )')
-define(`F_FSUBP', `F_POPR( FSUBP, $1, $2 )')
-
-define(`F_FSUBR2', `F_EXECR( FSUBR2, $1, $2 )')
-define(`F_FSUBR_S', `F_EXEC( FSUBR_S, $1, $2 )')
-define(`F_FSUBRP', `F_POPR( FSUBRP, $1, $2 )')
-
-########################################
-## Comments
-########################################
-## define(`F_COMMENT', ` 	/* `$1' */')
-## define(`F_PCOMMENT', ` 	/* `$1' { F_SHOW_STACK} */')
-define(`F_COMMENT', `')
-define(`F_PCOMMENT', ` 	/* F_SHOW_STACK*/')
-
-divert`'dnl
diff --git a/xc/extras/Mesa/src/X86/xform_args.h b/xc/extras/Mesa/src/X86/xform_args.h
new file mode 100644
index 000000000..4c3899acc
--- /dev/null
+++ b/xc/extras/Mesa/src/X86/xform_args.h
@@ -0,0 +1,73 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/*
+ * Transform function interface for assembly code.  Simply define
+ * FRAME_OFFSET to the number of bytes pushed onto the stack before
+ * using the ARG_* argument macros.
+ *
+ * Gareth Hughes <gareth@valinux.com>
+ */
+
+#ifndef __XFORM_ARGS_H__
+#define __XFORM_ARGS_H__
+
+/* Offsets into GLvector4f
+ */
+#define V4F_DATA	0
+#define V4F_START	4
+#define V4F_COUNT	8
+#define V4F_STRIDE	12
+#define V4F_SIZE	16
+#define V4F_FLAGS	20
+
+/* GLvector4f flags
+ */
+#define VEC_SIZE_1	1
+#define VEC_SIZE_2	3
+#define VEC_SIZE_3	7
+#define VEC_SIZE_4	15
+
+/* Offsets for transform_func arguments
+ *
+ * typedef void (*transform_func)( GLvector4f *to_vec,
+ *				   const GLfloat m[16],
+ *				   const GLvector4f *from_vec,
+ *				   const GLubyte *clipmask,
+ *				   const GLubyte flag );
+ */
+#define OFFSET_DEST	4
+#define OFFSET_MATRIX	8
+#define OFFSET_SOURCE	12
+#define OFFSET_CLIP	16
+#define OFFSET_FLAG	20
+
+#define ARG_DEST	REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP)
+#define ARG_MATRIX 	REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP)
+#define ARG_SOURCE 	REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP)
+#define ARG_CLIP 	REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP)
+#define ARG_FLAG 	REGOFF(FRAME_OFFSET+OFFSET_FLAG, ESP)
+
+#endif
diff --git a/xc/extras/Mesa/src/aatritemp.h b/xc/extras/Mesa/src/aatritemp.h
index 976ee9689..e0d623eb0 100644
--- a/xc/extras/Mesa/src/aatritemp.h
+++ b/xc/extras/Mesa/src/aatritemp.h
@@ -303,24 +303,24 @@
 #ifdef DO_STUV1
 #  ifdef DO_SPEC
          gl_write_multitexture_span(ctx, 2, n, startX, iy, z,
-                                    (const GLfloat (*)[MAX_WIDTH]) s,
-                                    (const GLfloat (*)[MAX_WIDTH]) t,
-                                    (const GLfloat (*)[MAX_WIDTH]) u,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) s,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) t,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) u,
                                     (GLfloat (*)[MAX_WIDTH]) lambda,
-                                    rgba, (const GLubyte (*)[4]) spec,
+                                    rgba, (CONST GLubyte (*)[4]) spec,
                                     GL_POLYGON);
 #  else
          gl_write_multitexture_span(ctx, 2, n, startX, iy, z,
-                                    (const GLfloat (*)[MAX_WIDTH]) s,
-                                    (const GLfloat (*)[MAX_WIDTH]) t,
-                                    (const GLfloat (*)[MAX_WIDTH]) u,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) s,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) t,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) u,
                                     lambda, rgba, NULL, GL_POLYGON);
 #  endif
 #elif defined(DO_STUV0)
 #  ifdef DO_SPEC
          gl_write_texture_span(ctx, n, startX, iy, z,
                                s[0], t[0], u[0], lambda[0], rgba,
-                               (const GLubyte (*)[4]) spec, GL_POLYGON);
+                               (CONST GLubyte (*)[4]) spec, GL_POLYGON);
 #  else
          gl_write_texture_span(ctx, n, startX, iy, z,
                                s[0], t[0], u[0], lambda[0],
@@ -408,7 +408,7 @@
          left = ix + 1;
 #ifdef DO_STUV1
          {
-            int j;
+            GLuint j;
             for (j = 0; j < n; j++) {
                s[0][j] = s[0][j + left];
                t[0][j] = t[0][j + left];
@@ -422,17 +422,17 @@
          }
 #  ifdef DO_SPEC
          gl_write_multitexture_span(ctx, 2, n, left, iy, z + left,
-                                    (const GLfloat (*)[MAX_WIDTH]) s,
-                                    (const GLfloat (*)[MAX_WIDTH]) t,
-                                    (const GLfloat (*)[MAX_WIDTH]) u,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) s,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) t,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) u,
                                     lambda, rgba + left,
-                                    (const GLubyte (*)[4]) (spec + left),
+                                    (CONST GLubyte (*)[4]) (spec + left),
                                     GL_POLYGON);
 #  else
          gl_write_multitexture_span(ctx, 2, n, left, iy, z + left,
-                                    (const GLfloat (*)[MAX_WIDTH]) s,
-                                    (const GLfloat (*)[MAX_WIDTH]) t,
-                                    (const GLfloat (*)[MAX_WIDTH]) u,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) s,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) t,
+                                    (CONST GLfloat (*)[MAX_WIDTH]) u,
                                     lambda,
                                     rgba + left, NULL, GL_POLYGON);
 #  endif
@@ -441,7 +441,7 @@
          gl_write_texture_span(ctx, n, left, iy, z + left,
                                s[0] + left, t[0] + left, u[0] + left,
                                lambda[0] + left, rgba + left,
-                               (const GLubyte (*)[4]) (spec + left),
+                               (CONST GLubyte (*)[4]) (spec + left),
                                GL_POLYGON);
 #  else
          gl_write_texture_span(ctx, n, left, iy, z + left,
diff --git a/xc/extras/Mesa/src/accum.c b/xc/extras/Mesa/src/accum.c
index da658df0c..dfed66c18 100644
--- a/xc/extras/Mesa/src/accum.c
+++ b/xc/extras/Mesa/src/accum.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/accum.c,v 1.7 2000/09/26 15:56:29 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -234,6 +234,8 @@ _mesa_Accum( GLenum op, GLfloat value )
          if (ctx->IntegerAccumMode && value != ctx->IntegerAccumScaler)
             rescale_accum(ctx);
             
+         RENDER_START(ctx);
+
          if (ctx->IntegerAccumMode) {
             /* simply add integer color values into accum buffer */
             GLuint j;
@@ -277,6 +279,7 @@ _mesa_Accum( GLenum op, GLfloat value )
          /* restore read buffer = draw buffer (the default) */
          (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer,
                                        ctx->Color.DriverDrawBuffer );
+         RENDER_FINISH(ctx);
 	 break;
 
       case GL_LOAD:
@@ -297,6 +300,7 @@ _mesa_Accum( GLenum op, GLfloat value )
             ctx->IntegerAccumScaler = 0.0;
          }
 
+         RENDER_START(ctx);
          if (ctx->IntegerAccumMode) {
             /* just copy values into accum buffer */
             GLuint j;
@@ -340,6 +344,7 @@ _mesa_Accum( GLenum op, GLfloat value )
          /* restore read buffer = draw buffer (the default) */
          (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer,
                                        ctx->Color.DriverDrawBuffer );
+         RENDER_FINISH(ctx);
 	 break;
 
       case GL_RETURN:
@@ -347,16 +352,18 @@ _mesa_Accum( GLenum op, GLfloat value )
          if (ctx->IntegerAccumMode && value != 1.0)
             rescale_accum(ctx);
 
+         RENDER_START(ctx);
          if (ctx->IntegerAccumMode && ctx->IntegerAccumScaler > 0) {
             /* build lookup table to avoid many floating point multiplies */
             const GLfloat mult = ctx->IntegerAccumScaler;
             static GLchan multTable[32768];
             static GLfloat prevMult = 0.0;
             GLuint j;
-            const GLint max = MIN2((GLint) (256 / mult), 32767);
+            GLint k;
+            const GLint max = MIN2((GLuint) (256 / mult), 32767);
             if (mult != prevMult) {
-               for (j = 0; j < max; j++)
-                  multTable[j] = (GLint) ((GLfloat) j * mult + 0.5F);
+               for (k = 0; k < max; k++)
+                  multTable[k] = (GLint) ((GLfloat) k * mult + 0.5F);
                prevMult = mult;
             }
 
@@ -379,7 +386,7 @@ _mesa_Accum( GLenum op, GLfloat value )
                   _mesa_mask_rgba_span( ctx, width, xpos, ypos, rgba );
                }
                (*ctx->Driver.WriteRGBASpan)( ctx, width, xpos, ypos, 
-                                             (const GLubyte (*)[4])rgba, NULL );
+                                          (CONST GLubyte (*)[4]) rgba, NULL );
                ypos++;
             }
          }
@@ -406,10 +413,11 @@ _mesa_Accum( GLenum op, GLfloat value )
                   _mesa_mask_rgba_span( ctx, width, xpos, ypos, rgba );
                }
                (*ctx->Driver.WriteRGBASpan)( ctx, width, xpos, ypos, 
-                                             (const GLubyte (*)[4])rgba, NULL );
+                                          (CONST GLubyte (*)[4]) rgba, NULL );
                ypos++;
             }
 	 }
+         RENDER_FINISH(ctx);
 	 break;
 
       default:
diff --git a/xc/extras/Mesa/src/attrib.c b/xc/extras/Mesa/src/attrib.c
index a464f0905..93e03802d 100644
--- a/xc/extras/Mesa/src/attrib.c
+++ b/xc/extras/Mesa/src/attrib.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/attrib.c,v 1.6 2000/09/26 15:56:29 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -212,8 +212,8 @@ _mesa_PushAttrib(GLbitfield mask)
       attr->RescaleNormals = ctx->Transform.RescaleNormals;
       attr->Scissor = ctx->Scissor.Enabled;
       attr->Stencil = ctx->Stencil.Enabled;
-      attr->Texture = ctx->Texture.Enabled;
       for (i=0; i<MAX_TEXTURE_UNITS; i++) {
+         attr->Texture[i] = ctx->Texture.Unit[i].Enabled;
          attr->TexGen[i] = ctx->Texture.Unit[i].TexGenEnabled;
       }
       newnode = new_attrib_node( GL_ENABLE_BIT );
@@ -392,6 +392,148 @@ _mesa_PushAttrib(GLbitfield mask)
 
 
 
+static void
+pop_enable_group(GLcontext *ctx, const struct gl_enable_attrib *enable)
+{
+   GLuint i;
+
+#define TEST_AND_UPDATE(VALUE, NEWVALUE, ENUM)		\
+	if ((VALUE) != (NEWVALUE)) {			\
+	   _mesa_set_enable( ctx, ENUM, (NEWVALUE) );	\
+	}
+
+   TEST_AND_UPDATE(ctx->Color.AlphaEnabled, enable->AlphaTest, GL_ALPHA_TEST);
+   TEST_AND_UPDATE(ctx->Transform.Normalize, enable->AutoNormal, GL_NORMALIZE);
+   TEST_AND_UPDATE(ctx->Color.BlendEnabled, enable->Blend, GL_BLEND);
+
+   for (i=0;i<MAX_CLIP_PLANES;i++) {
+      if (ctx->Transform.ClipEnabled[i] != enable->ClipPlane[i])
+         _mesa_set_enable(ctx, (GLenum) (GL_CLIP_PLANE0 + i),
+                          enable->ClipPlane[i]);
+   }
+
+   TEST_AND_UPDATE(ctx->Light.ColorMaterialEnabled, enable->ColorMaterial,
+                   GL_COLOR_MATERIAL);
+   TEST_AND_UPDATE(ctx->Polygon.CullFlag, enable->CullFace, GL_CULL_FACE);
+   TEST_AND_UPDATE(ctx->Depth.Test, enable->DepthTest, GL_DEPTH_TEST);
+   TEST_AND_UPDATE(ctx->Color.DitherFlag, enable->Dither, GL_DITHER);
+   TEST_AND_UPDATE(ctx->Pixel.Convolution1DEnabled, enable->Convolution1D,
+                   GL_CONVOLUTION_1D);
+   TEST_AND_UPDATE(ctx->Pixel.Convolution2DEnabled, enable->Convolution2D,
+                   GL_CONVOLUTION_2D);
+   TEST_AND_UPDATE(ctx->Pixel.Separable2DEnabled, enable->Separable2D,
+                   GL_SEPARABLE_2D);
+   TEST_AND_UPDATE(ctx->Fog.Enabled, enable->Fog, GL_FOG);
+   TEST_AND_UPDATE(ctx->Light.Enabled, enable->Lighting, GL_LIGHTING);
+   TEST_AND_UPDATE(ctx->Line.SmoothFlag, enable->LineSmooth, GL_LINE_SMOOTH);
+   TEST_AND_UPDATE(ctx->Line.StippleFlag, enable->LineStipple,
+                   GL_LINE_STIPPLE);
+   TEST_AND_UPDATE(ctx->Color.IndexLogicOpEnabled, enable->IndexLogicOp,
+                   GL_INDEX_LOGIC_OP);
+   TEST_AND_UPDATE(ctx->Color.ColorLogicOpEnabled, enable->ColorLogicOp,
+                   GL_COLOR_LOGIC_OP);
+   TEST_AND_UPDATE(ctx->Eval.Map1Color4, enable->Map1Color4, GL_MAP1_COLOR_4);
+   TEST_AND_UPDATE(ctx->Eval.Map1Index, enable->Map1Index, GL_MAP1_INDEX);
+   TEST_AND_UPDATE(ctx->Eval.Map1Normal, enable->Map1Normal, GL_MAP1_NORMAL);
+   TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord1, enable->Map1TextureCoord1,
+                   GL_MAP1_TEXTURE_COORD_1);
+   TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord2, enable->Map1TextureCoord2,
+                   GL_MAP1_TEXTURE_COORD_2);
+   TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord3, enable->Map1TextureCoord3,
+                   GL_MAP1_TEXTURE_COORD_3);
+   TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord4, enable->Map1TextureCoord4,
+                   GL_MAP1_TEXTURE_COORD_4);
+   TEST_AND_UPDATE(ctx->Eval.Map1Vertex3, enable->Map1Vertex3,
+                   GL_MAP1_VERTEX_3);
+   TEST_AND_UPDATE(ctx->Eval.Map1Vertex4, enable->Map1Vertex4,
+                   GL_MAP1_VERTEX_4);
+   TEST_AND_UPDATE(ctx->Eval.Map2Color4, enable->Map2Color4, GL_MAP2_COLOR_4);
+   TEST_AND_UPDATE(ctx->Eval.Map2Index, enable->Map2Index, GL_MAP2_INDEX);
+   TEST_AND_UPDATE(ctx->Eval.Map2Normal, enable->Map2Normal, GL_MAP2_NORMAL);
+   TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord1, enable->Map2TextureCoord1,
+                   GL_MAP2_TEXTURE_COORD_1);
+   TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord2, enable->Map2TextureCoord2,
+                   GL_MAP2_TEXTURE_COORD_2);
+   TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord3, enable->Map2TextureCoord3,
+                   GL_MAP2_TEXTURE_COORD_3);
+   TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord4, enable->Map2TextureCoord4,
+                   GL_MAP2_TEXTURE_COORD_4);
+   TEST_AND_UPDATE(ctx->Eval.Map2Vertex3, enable->Map2Vertex3,
+                   GL_MAP2_VERTEX_3);
+   TEST_AND_UPDATE(ctx->Eval.Map2Vertex4, enable->Map2Vertex4,
+                   GL_MAP2_VERTEX_4);
+   TEST_AND_UPDATE(ctx->Transform.Normalize, enable->Normalize, GL_NORMALIZE);
+   TEST_AND_UPDATE(ctx->Transform.RescaleNormals, enable->RescaleNormals,
+                   GL_RESCALE_NORMAL_EXT);
+   TEST_AND_UPDATE(ctx->Pixel.PixelTextureEnabled, enable->PixelTexture,
+                   GL_POINT_SMOOTH);
+   TEST_AND_UPDATE(ctx->Point.SmoothFlag, enable->PointSmooth,
+                   GL_POINT_SMOOTH);
+   TEST_AND_UPDATE(ctx->Polygon.OffsetPoint, enable->PolygonOffsetPoint,
+                   GL_POLYGON_OFFSET_POINT);
+   TEST_AND_UPDATE(ctx->Polygon.OffsetLine, enable->PolygonOffsetLine,
+                   GL_POLYGON_OFFSET_LINE);
+   TEST_AND_UPDATE(ctx->Polygon.OffsetFill, enable->PolygonOffsetFill,
+                   GL_POLYGON_OFFSET_FILL);
+   TEST_AND_UPDATE(ctx->Polygon.SmoothFlag, enable->PolygonSmooth,
+                   GL_POLYGON_SMOOTH);
+   TEST_AND_UPDATE(ctx->Polygon.StippleFlag, enable->PolygonStipple,
+                   GL_POLYGON_STIPPLE);
+   TEST_AND_UPDATE(ctx->Scissor.Enabled, enable->Scissor, GL_SCISSOR_TEST);
+   TEST_AND_UPDATE(ctx->Stencil.Enabled, enable->Stencil, GL_STENCIL_TEST);
+#undef TEST_AND_UPDATE
+
+   /* texture unit enables */
+   for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
+      if (ctx->Texture.Unit[i].Enabled != enable->Texture[i]) {
+         ctx->Texture.Unit[i].Enabled = enable->Texture[i];
+         if (ctx->Driver.Enable) {
+            if (ctx->Driver.ActiveTexture) {
+               (*ctx->Driver.ActiveTexture)(ctx, i);
+            }
+            (*ctx->Driver.Enable)( ctx, GL_TEXTURE_1D,
+                             (GLboolean) (enable->Texture[i] & TEXTURE0_1D) );
+            (*ctx->Driver.Enable)( ctx, GL_TEXTURE_2D,
+                             (GLboolean) (enable->Texture[i] & TEXTURE0_2D) );
+            (*ctx->Driver.Enable)( ctx, GL_TEXTURE_3D,
+                             (GLboolean) (enable->Texture[i] & TEXTURE0_3D) );
+         }
+      }
+
+      if (ctx->Texture.Unit[i].TexGenEnabled != enable->TexGen[i]) {
+         ctx->Texture.Unit[i].TexGenEnabled = enable->TexGen[i];
+         if (ctx->Driver.Enable) {
+            if (ctx->Driver.ActiveTexture) {
+               (*ctx->Driver.ActiveTexture)(ctx, i);
+            }
+            if (enable->TexGen[i] & S_BIT)
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_S, GL_TRUE);
+            else
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_S, GL_FALSE);
+            if (enable->TexGen[i] & T_BIT)
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_T, GL_TRUE);
+            else
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_T, GL_FALSE);
+            if (enable->TexGen[i] & R_BIT)
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_R, GL_TRUE);
+            else
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_R, GL_FALSE);
+            if (enable->TexGen[i] & Q_BIT)
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_Q, GL_TRUE);
+            else
+               (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_Q, GL_FALSE);
+         }
+      }
+   }
+
+   if (ctx->Driver.ActiveTexture) {
+      (*ctx->Driver.ActiveTexture)(ctx, ctx->Texture.CurrentUnit);
+   }
+}
+
+
+
+
 /*
  * This function is kind of long just because we have to call a lot
  * of device driver functions to update device driver state.
@@ -494,116 +636,7 @@ _mesa_PopAttrib(void)
             {
                const struct gl_enable_attrib *enable;
                enable = (const struct gl_enable_attrib *) attr->data;
-
-#define TEST_AND_UPDATE(VALUE, NEWVALUE, ENUM)		\
-	if ((VALUE) != (NEWVALUE)) {			\
-	   _mesa_set_enable( ctx, ENUM, (NEWVALUE) );	\
-	}
-
-               TEST_AND_UPDATE(ctx->Color.AlphaEnabled, enable->AlphaTest, GL_ALPHA_TEST);
-               TEST_AND_UPDATE(ctx->Transform.Normalize, enable->AutoNormal, GL_NORMALIZE);
-               TEST_AND_UPDATE(ctx->Color.BlendEnabled, enable->Blend, GL_BLEND);
-               {
-                  GLuint i;
-                  for (i=0;i<MAX_CLIP_PLANES;i++) {
-                     if (ctx->Transform.ClipEnabled[i] != enable->ClipPlane[i])
-                        _mesa_set_enable( ctx, (GLenum) (GL_CLIP_PLANE0 + i), enable->ClipPlane[i] );
-                  }
-               }
-               TEST_AND_UPDATE(ctx->Light.ColorMaterialEnabled, enable->ColorMaterial, GL_COLOR_MATERIAL);
-               TEST_AND_UPDATE(ctx->Polygon.CullFlag, enable->CullFace, GL_CULL_FACE);
-               TEST_AND_UPDATE(ctx->Depth.Test, enable->DepthTest, GL_DEPTH_TEST);
-               TEST_AND_UPDATE(ctx->Color.DitherFlag, enable->Dither, GL_DITHER);
-               TEST_AND_UPDATE(ctx->Pixel.Convolution1DEnabled, enable->Convolution1D, GL_CONVOLUTION_1D);
-               TEST_AND_UPDATE(ctx->Pixel.Convolution2DEnabled, enable->Convolution2D, GL_CONVOLUTION_2D);
-               TEST_AND_UPDATE(ctx->Pixel.Separable2DEnabled, enable->Separable2D, GL_SEPARABLE_2D);
-               TEST_AND_UPDATE(ctx->Fog.Enabled, enable->Fog, GL_FOG);
-               TEST_AND_UPDATE(ctx->Light.Enabled, enable->Lighting, GL_LIGHTING);
-               TEST_AND_UPDATE(ctx->Line.SmoothFlag, enable->LineSmooth, GL_LINE_SMOOTH);
-               TEST_AND_UPDATE(ctx->Line.StippleFlag, enable->LineStipple, GL_LINE_STIPPLE);
-               TEST_AND_UPDATE(ctx->Color.IndexLogicOpEnabled, enable->IndexLogicOp, GL_INDEX_LOGIC_OP);
-               TEST_AND_UPDATE(ctx->Color.ColorLogicOpEnabled, enable->ColorLogicOp, GL_COLOR_LOGIC_OP);
-               TEST_AND_UPDATE(ctx->Eval.Map1Color4, enable->Map1Color4, GL_MAP1_COLOR_4);
-               TEST_AND_UPDATE(ctx->Eval.Map1Index, enable->Map1Index, GL_MAP1_INDEX);
-               TEST_AND_UPDATE(ctx->Eval.Map1Normal, enable->Map1Normal, GL_MAP1_NORMAL);
-               TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord1, enable->Map1TextureCoord1, GL_MAP1_TEXTURE_COORD_1);
-               TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord2, enable->Map1TextureCoord2, GL_MAP1_TEXTURE_COORD_2);
-               TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord3, enable->Map1TextureCoord3, GL_MAP1_TEXTURE_COORD_3);
-               TEST_AND_UPDATE(ctx->Eval.Map1TextureCoord4, enable->Map1TextureCoord4, GL_MAP1_TEXTURE_COORD_4);
-               TEST_AND_UPDATE(ctx->Eval.Map1Vertex3, enable->Map1Vertex3, GL_MAP1_VERTEX_3);
-               TEST_AND_UPDATE(ctx->Eval.Map1Vertex4, enable->Map1Vertex4, GL_MAP1_VERTEX_4);
-               TEST_AND_UPDATE(ctx->Eval.Map2Color4, enable->Map2Color4, GL_MAP2_COLOR_4);
-               TEST_AND_UPDATE(ctx->Eval.Map2Index, enable->Map2Index, GL_MAP2_INDEX);
-               TEST_AND_UPDATE(ctx->Eval.Map2Normal, enable->Map2Normal, GL_MAP2_NORMAL);
-               TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord1, enable->Map2TextureCoord1, GL_MAP2_TEXTURE_COORD_1);
-               TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord2, enable->Map2TextureCoord2, GL_MAP2_TEXTURE_COORD_2);
-               TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord3, enable->Map2TextureCoord3, GL_MAP2_TEXTURE_COORD_3);
-               TEST_AND_UPDATE(ctx->Eval.Map2TextureCoord4, enable->Map2TextureCoord4, GL_MAP2_TEXTURE_COORD_4);
-               TEST_AND_UPDATE(ctx->Eval.Map2Vertex3, enable->Map2Vertex3, GL_MAP2_VERTEX_3);
-               TEST_AND_UPDATE(ctx->Eval.Map2Vertex4, enable->Map2Vertex4, GL_MAP2_VERTEX_4);
-               TEST_AND_UPDATE(ctx->Transform.Normalize, enable->Normalize, GL_NORMALIZE);
-               TEST_AND_UPDATE(ctx->Transform.RescaleNormals, enable->RescaleNormals, GL_RESCALE_NORMAL_EXT);
-               TEST_AND_UPDATE(ctx->Pixel.PixelTextureEnabled, enable->PixelTexture, GL_POINT_SMOOTH);
-               TEST_AND_UPDATE(ctx->Point.SmoothFlag, enable->PointSmooth, GL_POINT_SMOOTH);
-               TEST_AND_UPDATE(ctx->Polygon.OffsetPoint, enable->PolygonOffsetPoint, GL_POLYGON_OFFSET_POINT);
-               TEST_AND_UPDATE(ctx->Polygon.OffsetLine, enable->PolygonOffsetLine, GL_POLYGON_OFFSET_LINE);
-               TEST_AND_UPDATE(ctx->Polygon.OffsetFill, enable->PolygonOffsetFill, GL_POLYGON_OFFSET_FILL);
-               TEST_AND_UPDATE(ctx->Polygon.SmoothFlag, enable->PolygonSmooth, GL_POLYGON_SMOOTH);
-               TEST_AND_UPDATE(ctx->Polygon.StippleFlag, enable->PolygonStipple, GL_POLYGON_STIPPLE);
-               TEST_AND_UPDATE(ctx->Scissor.Enabled, enable->Scissor, GL_SCISSOR_TEST);
-               TEST_AND_UPDATE(ctx->Stencil.Enabled, enable->Stencil, GL_STENCIL_TEST);
-               if (ctx->Texture.Enabled != enable->Texture) {
-                  ctx->Texture.Enabled = enable->Texture;
-                  if (ctx->Driver.Enable) {
-                     if (ctx->Driver.ActiveTexture)
-                        (*ctx->Driver.ActiveTexture)( ctx, 0 );
-                     (*ctx->Driver.Enable)( ctx, GL_TEXTURE_1D, (GLboolean) (enable->Texture & TEXTURE0_1D) );
-                     (*ctx->Driver.Enable)( ctx, GL_TEXTURE_2D, (GLboolean) (enable->Texture & TEXTURE0_2D) );
-                     (*ctx->Driver.Enable)( ctx, GL_TEXTURE_3D, (GLboolean) (enable->Texture & TEXTURE0_3D) );
-                     if (ctx->Driver.ActiveTexture)
-                        (*ctx->Driver.ActiveTexture)( ctx, 1 );
-                     (*ctx->Driver.Enable)( ctx, GL_TEXTURE_1D, (GLboolean) (enable->Texture & TEXTURE1_1D) );
-                     (*ctx->Driver.Enable)( ctx, GL_TEXTURE_2D, (GLboolean) (enable->Texture & TEXTURE1_2D) );
-                     (*ctx->Driver.Enable)( ctx, GL_TEXTURE_3D, (GLboolean) (enable->Texture & TEXTURE1_3D) );
-                     if (ctx->Driver.ActiveTexture)
-                        (*ctx->Driver.ActiveTexture)( ctx, ctx->Texture.CurrentUnit );
-                  }
-               }
-#undef TEST_AND_UPDATE
-               {
-                  GLuint i;
-                  for (i=0; i<MAX_TEXTURE_UNITS; i++) {
-                     if (ctx->Texture.Unit[i].TexGenEnabled != enable->TexGen[i]) {
-                        ctx->Texture.Unit[i].TexGenEnabled = enable->TexGen[i];
-
-			/* ctx->Enabled recalculated in state change
-                           processing */
-			
-                        if (ctx->Driver.Enable) {
-                           if (ctx->Driver.ActiveTexture)
-                              (*ctx->Driver.ActiveTexture)( ctx, i );
-                           if (enable->TexGen[i] & S_BIT)
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_S, GL_TRUE);
-                           else
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_S, GL_FALSE);
-                           if (enable->TexGen[i] & T_BIT)
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_T, GL_TRUE);
-                           else
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_T, GL_FALSE);
-                           if (enable->TexGen[i] & R_BIT)
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_R, GL_TRUE);
-                           else
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_R, GL_FALSE);
-                           if (enable->TexGen[i] & Q_BIT)
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_Q, GL_TRUE);
-                           else
-                              (*ctx->Driver.Enable)( ctx, GL_TEXTURE_GEN_Q, GL_FALSE);
-                        }
-                     }
-                  }
-                  if (ctx->Driver.ActiveTexture)
-                     (*ctx->Driver.ActiveTexture)( ctx, ctx->Texture.CurrentUnit );
-               }
+               pop_enable_group(ctx, enable);
             }
             break;
          case GL_EVAL_BIT:
diff --git a/xc/extras/Mesa/src/bbox.c b/xc/extras/Mesa/src/bbox.c
index 24a2d86c1..a559a53ed 100644
--- a/xc/extras/Mesa/src/bbox.c
+++ b/xc/extras/Mesa/src/bbox.c
@@ -113,7 +113,7 @@ static void transform_bounds3( GLubyte *orMask, GLubyte *andMask,
       ACC_4V( data[i], dz );
 
    
-   cliptest_bounds( orMask, andMask, (const GLfloat (*)[4])data, 8 );
+   cliptest_bounds( orMask, andMask, (CONST GLfloat (*)[4])data, 8 );
 }
 
 static void transform_bounds2( GLubyte *orMask, GLubyte *andMask, 
@@ -147,7 +147,7 @@ static void transform_bounds2( GLubyte *orMask, GLubyte *andMask,
    ACC_4V( data[2], dy );
    ACC_4V( data[3], dy );
 
-   cliptest_bounds( orMask, andMask, (const GLfloat (*)[4])data, 4 );
+   cliptest_bounds( orMask, andMask, (CONST GLfloat (*)[4])data, 4 );
 }
 
 /* Dummy
diff --git a/xc/extras/Mesa/src/bitmap.c b/xc/extras/Mesa/src/bitmap.c
index eefee932f..d01337d67 100644
--- a/xc/extras/Mesa/src/bitmap.c
+++ b/xc/extras/Mesa/src/bitmap.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -61,6 +61,8 @@ render_bitmap( GLcontext *ctx, GLint px, GLint py,
       return;  /* NULL bitmap is legal, a no-op */
    }
 
+   RENDER_START(ctx);
+
    /* Set bitmap drawing color */
    if (ctx->Visual->RGBAflag) {
       GLint r, g, b, a;
@@ -127,6 +129,8 @@ render_bitmap( GLcontext *ctx, GLint px, GLint py,
    }
 
    gl_flush_pb(ctx);
+
+   RENDER_FINISH(ctx);
 }
 
 
diff --git a/xc/extras/Mesa/src/blend.c b/xc/extras/Mesa/src/blend.c
index 70c114ef9..84141c8bc 100644
--- a/xc/extras/Mesa/src/blend.c
+++ b/xc/extras/Mesa/src/blend.c
@@ -1,20 +1,20 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
- * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ * Version:  3.4
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/blend.c,v 1.7 2000/09/26 15:56:29 tsi Exp $ */
+
 
 
 #ifdef PC_HEADER
@@ -288,7 +288,7 @@ _mesa_BlendEquation( GLenum mode )
 
    ctx->Color.BlendFunc = NULL;
    ctx->NewState |= NEW_RASTER_OPS;
-   
+
    if (ctx->Driver.BlendEquation)
       ctx->Driver.BlendEquation( ctx, mode );
 }
@@ -338,11 +338,31 @@ blend_transparency( GLcontext *ctx, GLuint n, const GLubyte mask[],
             /* 100% alpha, no-op */
          }
          else {
+#if 0
+            /* This is pretty close, but Glean complains */
+            const GLint s = CHAN_MAX - t;
+            const GLint r = (rgba[i][RCOMP] * t + dest[i][RCOMP] * s + 1) >> 8;
+            const GLint g = (rgba[i][GCOMP] * t + dest[i][GCOMP] * s + 1) >> 8;
+            const GLint b = (rgba[i][BCOMP] * t + dest[i][BCOMP] * s + 1) >> 8;
+            const GLint a = (rgba[i][ACOMP] * t + dest[i][ACOMP] * s + 1) >> 8;
+#elif 0
+            /* This is slower but satisfies Glean */
             const GLint s = CHAN_MAX - t;
-            const GLint r = (rgba[i][RCOMP] * t + dest[i][RCOMP] * s) >> 8;
-            const GLint g = (rgba[i][GCOMP] * t + dest[i][GCOMP] * s) >> 8;
-            const GLint b = (rgba[i][BCOMP] * t + dest[i][BCOMP] * s) >> 8;
-            const GLint a = (rgba[i][ACOMP] * t + dest[i][ACOMP] * s) >> 8;
+            const GLint r = (rgba[i][RCOMP] * t + dest[i][RCOMP] * s) / 255;
+            const GLint g = (rgba[i][GCOMP] * t + dest[i][GCOMP] * s) / 255;
+            const GLint b = (rgba[i][BCOMP] * t + dest[i][BCOMP] * s) / 255;
+            const GLint a = (rgba[i][ACOMP] * t + dest[i][ACOMP] * s) / 255;
+#else
+            /* This satisfies Glean and should be reasonably fast */
+            /* Contributed by Nathan Hand */
+#define DIV255(X)  (((X) << 8) + (X) + 256) >> 16
+            const GLint s = CHAN_MAX - t;
+            const GLint r = DIV255(rgba[i][RCOMP] * t + dest[i][RCOMP] * s);
+            const GLint g = DIV255(rgba[i][GCOMP] * t + dest[i][GCOMP] * s);
+            const GLint b = DIV255(rgba[i][BCOMP] * t + dest[i][BCOMP] * s);
+            const GLint a = DIV255(rgba[i][ACOMP] * t + dest[i][ACOMP] * s);
+#undef DIV255
+#endif
             ASSERT(r <= CHAN_MAX);
             ASSERT(g <= CHAN_MAX);
             ASSERT(b <= CHAN_MAX);
@@ -754,22 +774,22 @@ blend_general( GLcontext *ctx, GLuint n, const GLubyte mask[],
 
          /* compute blended color */
          if (ctx->Color.BlendEquation==GL_FUNC_ADD_EXT) {
-            r = Rs * sR + Rd * dR;
-            g = Gs * sG + Gd * dG;
-            b = Bs * sB + Bd * dB;
-            a = As * sA + Ad * dA;
+            r = Rs * sR + Rd * dR + 0.5F;
+            g = Gs * sG + Gd * dG + 0.5F;
+            b = Bs * sB + Bd * dB + 0.5F;
+            a = As * sA + Ad * dA + 0.5F;
          }
          else if (ctx->Color.BlendEquation==GL_FUNC_SUBTRACT_EXT) {
-            r = Rs * sR - Rd * dR;
-            g = Gs * sG - Gd * dG;
-            b = Bs * sB - Bd * dB;
-            a = As * sA - Ad * dA;
+            r = Rs * sR - Rd * dR + 0.5F;
+            g = Gs * sG - Gd * dG + 0.5F;
+            b = Bs * sB - Bd * dB + 0.5F;
+            a = As * sA - Ad * dA + 0.5F;
          }
          else if (ctx->Color.BlendEquation==GL_FUNC_REVERSE_SUBTRACT_EXT) {
-            r = Rd * dR - Rs * sR;
-            g = Gd * dG - Gs * sG;
-            b = Bd * dB - Bs * sB;
-            a = Ad * dA - As * sA;
+            r = Rd * dR - Rs * sR + 0.5F;
+            g = Gd * dG - Gs * sG + 0.5F;
+            b = Bd * dB - Bs * sB + 0.5F;
+            a = Ad * dA - As * sA + 0.5F;
          }
          else {
             /* should never get here */
@@ -790,7 +810,7 @@ blend_general( GLcontext *ctx, GLuint n, const GLubyte mask[],
 
 #if defined(USE_MMX_ASM)
 #include "X86/mmx.h"
-#include "X86/common_x86asm.h"
+#include "X86/common_x86_asm.h"
 #endif
 
 
@@ -810,8 +830,8 @@ static void set_blend_function( GLcontext *ctx )
    /* Hmm.  A table here would have 12^4 == way too many entries.
     * Provide a hook for MMX instead.
     */
-   if (gl_x86_cpu_features & GL_CPU_MMX) {
-      gl_mmx_set_blend_function (ctx);
+   if ( cpu_has_mmx ) {
+      gl_mmx_set_blend_function( ctx );
    }
    else
 #endif
@@ -869,7 +889,7 @@ _mesa_blend_span( GLcontext *ctx, GLuint n, GLint x, GLint y,
    if (!ctx->Color.BlendFunc)
       set_blend_function(ctx);
 
-   (*ctx->Color.BlendFunc)( ctx, n, mask, rgba, (const GLubyte (*)[4])dest );
+   (*ctx->Color.BlendFunc)( ctx, n, mask, rgba, (CONST GLubyte (*)[4])dest );
 }
 
 
@@ -904,5 +924,5 @@ _mesa_blend_pixels( GLcontext *ctx,
    if (!ctx->Color.BlendFunc)
       set_blend_function(ctx);
 
-   (*ctx->Color.BlendFunc)( ctx, n, mask, rgba, (const GLubyte (*)[4])dest );
+   (*ctx->Color.BlendFunc)( ctx, n, mask, rgba, (CONST GLubyte (*)[4])dest );
 }
diff --git a/xc/extras/Mesa/src/buffers.c b/xc/extras/Mesa/src/buffers.c
index e334265a6..214e644cb 100644
--- a/xc/extras/Mesa/src/buffers.c
+++ b/xc/extras/Mesa/src/buffers.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/buffers.c,v 1.4 2000/09/26 15:56:29 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -167,7 +167,7 @@ clear_color_buffer(GLcontext *ctx)
    }
    else {
       /* Color index mode */
-      ASSERT(ctx->Color.IndexMask == ~0);
+      ASSERT(ctx->Color.IndexMask == (GLuint) ~0);
       if (ctx->Visual->IndexBits == 8) {
          /* 8-bit clear */
          GLubyte span[MAX_WIDTH];
@@ -208,15 +208,19 @@ clear_color_buffers(GLcontext *ctx)
       if (bufferBit & ctx->Color.DrawDestMask) {
          if (bufferBit == FRONT_LEFT_BIT) {
             (void) (*ctx->Driver.SetDrawBuffer)( ctx, GL_FRONT_LEFT);
+            (void) (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer, GL_FRONT_LEFT);
          }
          else if (bufferBit == FRONT_RIGHT_BIT) {
             (void) (*ctx->Driver.SetDrawBuffer)( ctx, GL_FRONT_RIGHT);
+            (void) (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer, GL_FRONT_RIGHT);
          }
          else if (bufferBit == BACK_LEFT_BIT) {
             (void) (*ctx->Driver.SetDrawBuffer)( ctx, GL_BACK_LEFT);
+            (void) (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer, GL_BACK_LEFT);
          }
          else {
             (void) (*ctx->Driver.SetDrawBuffer)( ctx, GL_BACK_RIGHT);
+            (void) (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer, GL_BACK_RIGHT);
          }
          
          if (ctx->Color.SWmasking) {
@@ -228,8 +232,9 @@ clear_color_buffers(GLcontext *ctx)
       }
    }
 
-   /* restore default dest buffer */
+   /* restore default read/draw buffers */
    (void) (*ctx->Driver.SetDrawBuffer)( ctx, ctx->Color.DriverDrawBuffer );
+   (void) (*ctx->Driver.SetReadBuffer)( ctx, ctx->ReadBuffer, ctx->Pixel.DriverReadBuffer );
 }
 
 
@@ -287,6 +292,8 @@ _mesa_Clear( GLbitfield mask )
       }
 #endif
 
+      RENDER_START(ctx);
+
       /* do software clearing here */
       if (newMask) {
          if (newMask & ctx->Color.DrawDestMask)   clear_color_buffers(ctx);
@@ -302,6 +309,8 @@ _mesa_Clear( GLbitfield mask )
          _mesa_clear_alpha_buffers( ctx );
       }
 
+      RENDER_FINISH(ctx);
+
 #ifdef PROFILE
       ctx->ClearTime += gl_time() - t0;
       ctx->ClearCount++;
diff --git a/xc/extras/Mesa/src/clip_funcs.h b/xc/extras/Mesa/src/clip_funcs.h
index d2bb45280..e0003bf3d 100644
--- a/xc/extras/Mesa/src/clip_funcs.h
+++ b/xc/extras/Mesa/src/clip_funcs.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/clip_funcs.h,v 1.6 2000/09/26 15:56:29 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
diff --git a/xc/extras/Mesa/src/colortab.c b/xc/extras/Mesa/src/colortab.c
index dff9feae9..425c03b1d 100644
--- a/xc/extras/Mesa/src/colortab.c
+++ b/xc/extras/Mesa/src/colortab.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/colortab.c,v 1.7 2000/09/26 15:56:29 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -324,17 +324,14 @@ _mesa_ColorTable( GLenum target, GLenum internalFormat,
       return;
    }
 
-   if (width > ctx->Const.MaxColorTableSize) {
+   if (width > (GLsizei) ctx->Const.MaxColorTableSize) {
       if (proxy) {
          table->Size = 0;
          table->IntFormat = (GLenum) 0;
          table->Format = (GLenum) 0;
       }
       else {
-         if (width > ctx->Const.MaxColorTableSize)
-            gl_error(ctx, GL_TABLE_TOO_LARGE, "glColorTable(width)");
-         else
-            gl_error(ctx, GL_INVALID_VALUE, "glColorTable(width)");
+         gl_error(ctx, GL_TABLE_TOO_LARGE, "glColorTable(width)");
       }
       return;
    }
@@ -355,7 +352,7 @@ _mesa_ColorTable( GLenum target, GLenum internalFormat,
       if (floatTable) {
          GLfloat tempTab[MAX_COLOR_TABLE_SIZE * 4];
          GLfloat *tableF;
-         GLuint i;
+         GLint i;
 
          _mesa_unpack_float_color_span(ctx, width, table->Format,
                                        tempTab,  /* dest */
@@ -422,7 +419,7 @@ _mesa_ColorTable( GLenum target, GLenum internalFormat,
             return;
          }
          _mesa_unpack_ubyte_color_span(ctx, width, table->Format,
-                                       table->Table,  /* dest */
+                                       (GLubyte *) table->Table,  /* dest */
                                        format, type, data,
                                        &ctx->Unpack, GL_TRUE);
       } /* floatTable */
@@ -523,7 +520,7 @@ _mesa_ColorSubTable( GLenum target, GLsizei start,
    comps = _mesa_components_in_format(table->Format);
    assert(comps > 0);  /* error should have been caught sooner */
 
-   if (start + count > table->Size) {
+   if (start + count > (GLint) table->Size) {
       gl_error(ctx, GL_INVALID_VALUE, "glColorSubTable(count)");
       return;
    }
@@ -541,7 +538,7 @@ _mesa_ColorSubTable( GLenum target, GLsizei start,
    else {
       GLfloat tempTab[MAX_COLOR_TABLE_SIZE * 4];
       GLfloat *tableF;
-      GLuint i;
+      GLint i;
 
       ASSERT(table->TableType == GL_FLOAT);
 
@@ -676,7 +673,7 @@ _mesa_GetColorTable( GLenum target, GLenum format,
    struct gl_texture_unit *texUnit = &ctx->Texture.Unit[ctx->Texture.CurrentUnit];
    struct gl_color_table *table = NULL;
    GLubyte rgba[MAX_COLOR_TABLE_SIZE][4];
-   GLint i;
+   GLuint i;
 
    ASSERT_OUTSIDE_BEGIN_END(ctx, "glGetColorTable");
 
@@ -835,7 +832,7 @@ _mesa_GetColorTable( GLenum target, GLenum format,
          return;
    }
 
-   _mesa_pack_rgba_span(ctx, table->Size, (const GLubyte (*)[]) rgba,
+   _mesa_pack_rgba_span(ctx, table->Size, (CONST GLubyte (*)[4]) rgba,
                         format, type, data, &ctx->Pack, GL_FALSE);
 }
 
diff --git a/xc/extras/Mesa/src/config.c b/xc/extras/Mesa/src/config.c
index 9e50ff4b0..50df728c8 100644
--- a/xc/extras/Mesa/src/config.c
+++ b/xc/extras/Mesa/src/config.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/config.c,v 1.6 2000/09/24 13:50:08 alanh Exp $ */
 
 
 /* Mesa config file parse and execute code.
diff --git a/xc/extras/Mesa/src/config.h b/xc/extras/Mesa/src/config.h
index 1d93ae611..07ff841f6 100644
--- a/xc/extras/Mesa/src/config.h
+++ b/xc/extras/Mesa/src/config.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/config.h,v 1.7 2000/09/26 15:56:30 tsi Exp $ */
+
 
 /*
  * Tunable configuration parameters.
diff --git a/xc/extras/Mesa/src/context.c b/xc/extras/Mesa/src/context.c
index f8a4ef90b..4d4a5974c 100644
--- a/xc/extras/Mesa/src/context.c
+++ b/xc/extras/Mesa/src/context.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  *
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/context.c,v 1.6 2000/09/26 15:56:30 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -210,10 +210,11 @@ _mesa_initialize_visual( GLvisual *vis,
 
    if (depthBits == 0) {
       /* Special case.  Even if we don't have a depth buffer we need
-       * good values for DepthMax for Z vertex transformation purposes.
+       * good values for DepthMax for Z vertex transformation purposes
+       * and for per-fragment fog computation.
        */
-      vis->DepthMax = 1;
-      vis->DepthMaxF = 1.0F;
+      vis->DepthMax = 1 << 16;
+      vis->DepthMaxF = (GLfloat) vis->DepthMax;
    }
    else if (depthBits < 32) {
       vis->DepthMax = (1 << depthBits) - 1;
@@ -451,8 +452,9 @@ alloc_shared_state( void )
    if (!ss)
       return NULL;
 
-   ss->DisplayList = _mesa_NewHashTable();
+   _glthread_INIT_MUTEX(ss->Mutex);
 
+   ss->DisplayList = _mesa_NewHashTable();
    ss->TexObjects = _mesa_NewHashTable();
 
    /* Default Texture objects */
@@ -592,6 +594,8 @@ init_texture_unit( GLcontext *ctx, GLuint unit )
 {
    struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
 
+   texUnit->Enabled = 0;
+   texUnit->ReallyEnabled = 0;
    texUnit->EnvMode = GL_MODULATE;
    texUnit->CombineModeRGB = GL_MODULATE;
    texUnit->CombineModeA = GL_MODULATE;
@@ -1148,9 +1152,10 @@ init_attrib_groups( GLcontext *ctx )
    /* Texture group */
    ctx->Texture.CurrentUnit = 0;      /* multitexture */
    ctx->Texture.CurrentTransformUnit = 0; /* multitexture */
-   ctx->Texture.Enabled = 0;
+   ctx->Texture.ReallyEnabled = 0;
    for (i=0; i<MAX_TEXTURE_UNITS; i++)
       init_texture_unit( ctx, i );
+   ctx->Texture.SharedPalette = GL_FALSE;
    _mesa_init_colortable(&ctx->Texture.Palette);
 
    /* Transformation group */
diff --git a/xc/extras/Mesa/src/convolve.c b/xc/extras/Mesa/src/convolve.c
new file mode 100644
index 000000000..ba25340cf
--- /dev/null
+++ b/xc/extras/Mesa/src/convolve.c
@@ -0,0 +1,415 @@
+
+/*
+ * Mesa 3-D graphics library
+ * Version:  3.3
+ *
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+/*
+ * Image convolution functions.
+ *
+ * Notes: filter kernel elements are indexed by <n> and <m> as in
+ * the GL spec.
+ */
+
+
+#ifdef PC_HEADER
+#include "all.h"
+#else
+#include "glheader.h"
+#include "types.h"
+#endif
+
+
+void
+_mesa_convolve_1d_reduce(GLint srcWidth, const GLfloat src[][4],
+                         GLint filterWidth, const GLfloat filter[][4],
+                         GLfloat dest[][4])
+{
+   const GLint dstWidth = srcWidth - (filterWidth - 1);
+   GLint i, n;
+
+   if (dstWidth <= 0)
+      return;  /* null result */
+
+   for (i = 0; i < dstWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (n = 0; n < filterWidth; n++) {
+         sumR += src[i + n][RCOMP] * filter[n][RCOMP];
+         sumG += src[i + n][GCOMP] * filter[n][GCOMP];
+         sumB += src[i + n][BCOMP] * filter[n][BCOMP];
+         sumA += src[i + n][ACOMP] * filter[n][ACOMP];
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+void
+_mesa_convolve_1d_constant(GLint srcWidth, const GLfloat src[][4],
+                           GLint filterWidth, const GLfloat filter[][4],
+                           const GLfloat borderColor[4], GLfloat dest[][4])
+{
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (n = 0; n < filterWidth; n++) {
+         if (i + n < halfFilterWidth || i + n - halfFilterWidth >= srcWidth) {
+            sumR += borderColor[RCOMP] * filter[n][RCOMP];
+            sumG += borderColor[GCOMP] * filter[n][GCOMP];
+            sumB += borderColor[BCOMP] * filter[n][BCOMP];
+            sumA += borderColor[ACOMP] * filter[n][ACOMP];
+         }
+         else {
+            sumR += src[i + n - halfFilterWidth][RCOMP] * filter[n][RCOMP];
+            sumG += src[i + n - halfFilterWidth][GCOMP] * filter[n][GCOMP];
+            sumB += src[i + n - halfFilterWidth][BCOMP] * filter[n][BCOMP];
+            sumA += src[i + n - halfFilterWidth][ACOMP] * filter[n][ACOMP];
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+void
+_mesa_convolve_1d_replicate(GLint srcWidth, const GLfloat src[][4],
+                            GLint filterWidth, const GLfloat filter[][4],
+                            GLfloat dest[][4])
+{
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (n = 0; n < filterWidth; n++) {
+         if (i + n < halfFilterWidth) {
+            sumR += src[0][RCOMP] * filter[n][RCOMP];
+            sumG += src[0][GCOMP] * filter[n][GCOMP];
+            sumB += src[0][BCOMP] * filter[n][BCOMP];
+            sumA += src[0][ACOMP] * filter[n][ACOMP];
+         }
+         else if (i + n - halfFilterWidth >= srcWidth) {
+            sumR += src[srcWidth - 1][RCOMP] * filter[n][RCOMP];
+            sumG += src[srcWidth - 1][GCOMP] * filter[n][GCOMP];
+            sumB += src[srcWidth - 1][BCOMP] * filter[n][BCOMP];
+            sumA += src[srcWidth - 1][ACOMP] * filter[n][ACOMP];
+         }
+         else {
+            sumR += src[i + n - halfFilterWidth][RCOMP] * filter[n][RCOMP];
+            sumG += src[i + n - halfFilterWidth][GCOMP] * filter[n][GCOMP];
+            sumB += src[i + n - halfFilterWidth][BCOMP] * filter[n][BCOMP];
+            sumA += src[i + n - halfFilterWidth][ACOMP] * filter[n][ACOMP];
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+/*
+ * <src> is the source image width width = srcWidth, height = filterHeight.
+ * <filter> has width <filterWidth> and height <filterHeight>.
+ * <dst> is a 1-D image span of width <srcWidth> - (<filterWidth> - 1).
+ */
+void
+_mesa_convolve_2d_reduce(GLint srcWidth, GLint srcHeight,
+                         const GLfloat src[][4],
+                         GLint filterWidth, GLint filterHeight,
+                         const GLfloat filter[][4],
+                         GLfloat dest[][4])
+{
+   const GLint dstWidth = srcWidth - (filterWidth - 1);
+   GLint i, n, m;
+
+   if (dstWidth <= 0)
+      return;  /* null result */
+
+   /* XXX todo */
+   for (i = 0; i < dstWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (n = 0; n < filterHeight; n++) {
+         for (m = 0; m < filterWidth; m++) {
+            const GLint k = n * srcWidth + i + m;
+            sumR += src[k][RCOMP] * filter[n][RCOMP];
+            sumG += src[k][GCOMP] * filter[n][GCOMP];
+            sumB += src[k][BCOMP] * filter[n][BCOMP];
+            sumA += src[k][ACOMP] * filter[n][ACOMP];
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+void
+_mesa_convolve_2d_constant(GLint srcWidth, GLint srcHeight,
+                           const GLfloat src[][4],
+                           GLint filterWidth, GLint filterHeight,
+                           const GLfloat filter[][4],
+                           GLfloat dest[][4],
+                           const GLfloat borderColor[4])
+{
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n, m;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (m = 0; m < filterHeight; m++) {
+         const GLfloat (*filterRow)[4] = filter + m * filterWidth;
+         for (n = 0; n < filterWidth; n++) {
+            if (i + n < halfFilterWidth ||
+                i + n - halfFilterWidth >= srcWidth) {
+               sumR += borderColor[RCOMP] * filterRow[n][RCOMP];
+               sumG += borderColor[GCOMP] * filterRow[n][GCOMP];
+               sumB += borderColor[BCOMP] * filterRow[n][BCOMP];
+               sumA += borderColor[ACOMP] * filterRow[n][ACOMP];
+            }
+            else {
+               const GLint k = m * srcWidth + i + n - halfFilterWidth;
+               sumR += src[k][RCOMP] * filterRow[n][RCOMP];
+               sumG += src[k][GCOMP] * filterRow[n][GCOMP];
+               sumB += src[k][BCOMP] * filterRow[n][BCOMP];
+               sumA += src[k][ACOMP] * filterRow[n][ACOMP];
+            }
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+void
+_mesa_convolve_2d_replicate(GLint srcWidth, GLint srcHeight,
+                            const GLfloat src[][4],
+                            GLint filterWidth, GLint filterHeight,
+                            const GLfloat filter[][4],
+                            GLfloat dest[][4])
+{
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n, m;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (m = 0; m < filterHeight; m++) {
+         const GLfloat (*filterRow)[4] = filter + m * filterWidth;
+         for (n = 0; n < filterWidth; n++) {
+            if (i + n < halfFilterWidth) {
+               const GLint k = m * srcWidth + 0;
+               sumR += src[k][RCOMP] * filterRow[n][RCOMP];
+               sumG += src[k][GCOMP] * filterRow[n][GCOMP];
+               sumB += src[k][BCOMP] * filterRow[n][BCOMP];
+               sumA += src[k][ACOMP] * filterRow[n][ACOMP];
+            }
+            else if (i + n - halfFilterWidth >= srcWidth) {
+               const GLint k = m * srcWidth + srcWidth - 1;
+               sumR += src[k][RCOMP] * filterRow[n][RCOMP];
+               sumG += src[k][GCOMP] * filterRow[n][GCOMP];
+               sumB += src[k][BCOMP] * filterRow[n][BCOMP];
+               sumA += src[k][ACOMP] * filterRow[n][ACOMP];
+            }
+            else {
+               const GLint k = m * srcWidth + i + n - halfFilterWidth;
+               sumR += src[k][RCOMP] * filterRow[n][RCOMP];
+               sumG += src[k][GCOMP] * filterRow[n][GCOMP];
+               sumB += src[k][BCOMP] * filterRow[n][BCOMP];
+               sumA += src[k][ACOMP] * filterRow[n][ACOMP];
+            }
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+void
+_mesa_convolve_sep_constant(GLint srcWidth, GLint srcHeight,
+                            const GLfloat src[][4],
+                            GLint filterWidth, GLint filterHeight,
+                            const GLfloat rowFilt[][4],
+                            const GLfloat colFilt[][4],
+                            GLfloat dest[][4],
+                            const GLfloat borderColor[4])
+{
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n, m;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (m = 0; m < filterHeight; m++) {
+         for (n = 0; n < filterWidth; n++) {
+            if (i + n < halfFilterWidth ||
+                i + n - halfFilterWidth >= srcWidth) {
+               sumR += borderColor[RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += borderColor[GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += borderColor[BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += borderColor[ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+            else {
+               const GLint k = m * srcWidth + i + n - halfFilterWidth;
+               sumR += src[k][RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += src[k][GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += src[k][BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += src[k][ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
+
+
+void
+_mesa_convolve_sep_reduce(GLint srcWidth, GLint srcHeight,
+                          const GLfloat src[][4],
+                          GLint filterWidth, GLint filterHeight,
+                          const GLfloat rowFilt[][4],
+                          const GLfloat colFilt[][4],
+                          GLfloat dest[][4])
+{
+#if 00
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n, m;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (m = 0; m < filterHeight; m++) {
+         for (n = 0; n < filterWidth; n++) {
+            if (i + n < halfFilterWidth ||
+                i + n - halfFilterWidth >= srcWidth) {
+               sumR += borderColor[RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += borderColor[GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += borderColor[BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += borderColor[ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+            else {
+               const GLint k = m * srcWidth + i + n - halfFilterWidth;
+               sumR += src[k][RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += src[k][GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += src[k][BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += src[k][ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+#endif
+}
+
+
+void
+_mesa_convolve_sep_replicate(GLint srcWidth, GLint srcHeight,
+                             const GLfloat src[][4],
+                             GLint filterWidth, GLint filterHeight,
+                             const GLfloat rowFilt[][4],
+                             const GLfloat colFilt[][4],
+                             GLfloat dest[][4])
+{
+   const GLint halfFilterWidth = filterWidth / 2;
+   GLint i, n, m;
+
+   for (i = 0; i < srcWidth; i++) {
+      GLfloat sumR = 0.0;
+      GLfloat sumG = 0.0;
+      GLfloat sumB = 0.0;
+      GLfloat sumA = 0.0;
+      for (m = 0; m < filterHeight; m++) {
+         for (n = 0; n < filterWidth; n++) {
+            if (i + n < halfFilterWidth) {
+               const GLint k = m * srcWidth + 0;
+               sumR += src[k][RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += src[k][GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += src[k][BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += src[k][ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+            else if (i + n - halfFilterWidth >= srcWidth) {
+               const GLint k = m * srcWidth + srcWidth - 1;
+               sumR += src[k][RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += src[k][GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += src[k][BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += src[k][ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+            else {
+               const GLint k = m * srcWidth + i + n - halfFilterWidth;
+               sumR += src[k][RCOMP] * rowFilt[n][RCOMP] * colFilt[m][RCOMP];
+               sumG += src[k][GCOMP] * rowFilt[n][GCOMP] * colFilt[m][GCOMP];
+               sumB += src[k][BCOMP] * rowFilt[n][BCOMP] * colFilt[m][BCOMP];
+               sumA += src[k][ACOMP] * rowFilt[n][ACOMP] * colFilt[m][ACOMP];
+            }
+         }
+      }
+      dest[i][RCOMP] = sumR;
+      dest[i][GCOMP] = sumG;
+      dest[i][BCOMP] = sumB;
+      dest[i][ACOMP] = sumA;
+   }
+}
diff --git a/xc/extras/Mesa/src/copypix.c b/xc/extras/Mesa/src/copypix.c
index 8b40a8711..2d3c3594a 100644
--- a/xc/extras/Mesa/src/copypix.c
+++ b/xc/extras/Mesa/src/copypix.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/copypix.c,v 1.6 2000/09/26 15:56:30 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -126,7 +126,10 @@ static void copy_rgba_pixels( GLcontext *ctx,
    /* If read and draw buffer are different we must do buffer switching */
    saveReadAlpha = ctx->ReadBuffer->Alpha;
    changeBuffer = ctx->Pixel.ReadBuffer != ctx->Color.DrawBuffer
-               || ctx->DrawBuffer != ctx->ReadBuffer;
+                  || ctx->DrawBuffer != ctx->ReadBuffer;
+
+   (*ctx->Driver.SetReadBuffer)( ctx, ctx->ReadBuffer,
+                                 ctx->Pixel.DriverReadBuffer );
 
    if (overlapping) {
       GLint ssy = sy;
@@ -205,7 +208,7 @@ static void copy_rgba_pixels( GLcontext *ctx,
       if (applyTransferOps) {
          const GLfloat scale = (1.0F / 255.0F);
          GLfloat rgbaFloat[MAX_WIDTH][4];
-         GLuint k;
+         GLint k;
          /* convert ubyte to float */
          for (k = 0; k < width; k++) {
             rgbaFloat[k][RCOMP] = (GLfloat) rgba[k][RCOMP] * scale;
@@ -269,7 +272,7 @@ static void copy_rgba_pixels( GLcontext *ctx,
          MEMCPY(primary_rgba, rgba, 4 * width * sizeof(GLubyte));
 
          for (unit = 0; unit < MAX_TEXTURE_UNITS; unit++) {
-            _mesa_pixeltexgen(ctx, width, (const GLubyte (*)[4]) rgba,
+            _mesa_pixeltexgen(ctx, width, (CONST GLubyte (*)[4]) rgba,
                               s, t, r, q);
             gl_texture_pixels(ctx, unit, width, s, t, r, NULL,
                               primary_rgba, rgba);
@@ -278,17 +281,21 @@ static void copy_rgba_pixels( GLcontext *ctx,
 
       if (quick_draw && dy >= 0 && dy < ctx->DrawBuffer->Height) {
          (*ctx->Driver.WriteRGBASpan)( ctx, width, destx, dy, 
-				       (const GLubyte (*)[4])rgba, NULL );
+				       (CONST GLubyte (*)[4])rgba, NULL );
       }
       else if (zoom) {
          gl_write_zoomed_rgba_span( ctx, width, destx, dy, zspan, 
-				    (const GLubyte (*)[4])rgba, desty);
+				    (CONST GLubyte (*)[4])rgba, desty);
       }
       else {
          gl_write_rgba_span( ctx, width, destx, dy, zspan, rgba, GL_BITMAP );
       }
    }
 
+   /* Restore pixel source to be the draw buffer (for blending, etc) */
+   (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer,
+                                 ctx->Color.DriverDrawBuffer );
+
    if (overlapping)
       FREE(prgba);
 }
@@ -336,6 +343,9 @@ static void copy_ci_pixels( GLcontext *ctx,
    changeBuffer = ctx->Pixel.ReadBuffer != ctx->Color.DrawBuffer
                || ctx->DrawBuffer != ctx->ReadBuffer;
 
+   (*ctx->Driver.SetReadBuffer)( ctx, ctx->ReadBuffer,
+                                 ctx->Pixel.DriverReadBuffer );
+
    if (overlapping) {
       GLint ssy = sy;
       pci = (GLuint *) MALLOC(width * height * sizeof(GLuint));
@@ -394,6 +404,10 @@ static void copy_ci_pixels( GLcontext *ctx,
       }
    }
 
+   /* Restore pixel source to be the draw buffer (for blending, etc) */
+   (*ctx->Driver.SetReadBuffer)( ctx, ctx->DrawBuffer,
+                                 ctx->Color.DriverDrawBuffer );
+
    if (overlapping)
       FREE(pci);
 }
@@ -489,7 +503,7 @@ static void copy_depth_pixels( GLcontext *ctx, GLint srcx, GLint srcy,
       if (ctx->Visual->RGBAflag) {
          if (zoom) {
             gl_write_zoomed_rgba_span( ctx, width, destx, dy, zspan, 
-				       (const GLubyte (*)[4])rgba, desty );
+				       (CONST GLubyte (*)[4])rgba, desty );
          }
          else {
             gl_write_rgba_span( ctx, width, destx, dy, zspan, rgba, GL_BITMAP);
@@ -626,9 +640,12 @@ _mesa_CopyPixels( GLint srcx, GLint srcy, GLsizei width, GLsizei height,
 
       ctx->OcclusionResult = GL_TRUE;
 
+      RENDER_START(ctx);
+
       if (ctx->Driver.CopyPixels &&
           (*ctx->Driver.CopyPixels)( ctx, srcx, srcy, width, height,
                                      destx, desty, type )) {
+         RENDER_FINISH(ctx);
          return;
       }
 
@@ -647,6 +664,8 @@ _mesa_CopyPixels( GLint srcx, GLint srcy, GLsizei width, GLsizei height,
       else {
 	 gl_error( ctx, GL_INVALID_ENUM, "glCopyPixels" );
       }
+
+      RENDER_FINISH(ctx);
    }
    else if (ctx->RenderMode == GL_FEEDBACK) {
       GLfloat color[4];
diff --git a/xc/extras/Mesa/src/cull_tmp.h b/xc/extras/Mesa/src/cull_tmp.h
index 463cb59b4..14c83c475 100644
--- a/xc/extras/Mesa/src/cull_tmp.h
+++ b/xc/extras/Mesa/src/cull_tmp.h
@@ -38,7 +38,7 @@ static GLuint TAG(gl_cull_triangles)( struct vertex_buffer *VB,
    const GLubyte face_bit = ctx->Polygon.FrontBit;
    const GLubyte cull_faces = ctx->Polygon.CullBits;
    GLubyte *cullmask = VB->CullMask;
-   GLint i,cullcount = 0;
+   GLint i, cullcount = 0;
    GLint last = count - 3;
    
    (void) parity;
@@ -46,7 +46,7 @@ static GLuint TAG(gl_cull_triangles)( struct vertex_buffer *VB,
       CULL_TRI(DO_CLIP, DO_AREA, i, i+1, i+2, face_bit, 3);
    }
 
-   if (i != count)
+   if (i != (GLint) count)
       cullcount += count - i;
 
    return cullcount;
@@ -77,7 +77,7 @@ static GLuint TAG(gl_cull_triangle_fan)( struct vertex_buffer *VB,
       CULL_TRI(DO_CLIP, DO_AREA, start, i+1, i+2, face_bit, nr);
    }
 
-   if (i != last + 1)
+   if (i != (GLint) (last + 1))
       cullcount += count - i;
 
    return cullcount;
@@ -135,7 +135,7 @@ static GLuint TAG(gl_cull_quads)( struct vertex_buffer *VB,
       CULL_QUAD(DO_CLIP, DO_AREA, i, i+1, i+2, i+3, 4);
    }
 
-   if (i != count)
+   if (i != (GLint) count)
       cullcount += count - i;
 
    return cullcount;
@@ -162,7 +162,7 @@ static GLuint TAG(gl_cull_quad_strip)( struct vertex_buffer *VB,
       CULL_QUAD(DO_CLIP, DO_AREA, i, i+1, i+3, i+2, nr);
    }
 
-   if (i != last + 2)
+   if (i != ((GLint) last + 2))
       cullcount += count - i;
 
    return cullcount;
diff --git a/xc/extras/Mesa/src/cva.c b/xc/extras/Mesa/src/cva.c
index d06890c3c..f4f1a7ba2 100644
--- a/xc/extras/Mesa/src/cva.c
+++ b/xc/extras/Mesa/src/cva.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -244,15 +244,16 @@ void gl_merge_cva( struct vertex_buffer *VB,
 	 translate_4f( VB->ClipPtr, cvaVB->ClipPtr, elt, count);
 
 	 if (VB->ClipOrMask & CLIP_USER_BIT) {
-	    GLubyte or = 0, and = ~0;
+	    GLubyte orMask = 0, andMask = ~0;
 
 	    copy_clipmask( VB->UserClipMask + VB->Start,
-			   &or, &and,
+			   &orMask, &andMask,
 			   cvaVB->UserClipMask,
 			   elt,
 			   VB->Count - VB->Start);
 	    
-	    if (and) VB->ClipAndMask |= CLIP_USER_BIT;
+	    if (andMask)
+               VB->ClipAndMask |= CLIP_USER_BIT;
 	 }
 
 	 if (VB->ClipOrMask) 
@@ -386,7 +387,7 @@ _mesa_LockArraysEXT(GLint first, GLsizei count)
     */
    if (first == 0 && 
        count > 0 && 
-       count <= ctx->Const.MaxArrayLockSize)
+       count <= (GLint) ctx->Const.MaxArrayLockSize)
    {   
       struct gl_cva *cva = &ctx->CVA;
 
@@ -683,7 +684,7 @@ void gl_prepare_arrays_cva( struct vertex_buffer *VB )
       if (disable & VERT_NORM) cva->v.Normal = *VB->store.Normal;
       if (disable & VERT_OBJ_ANY) cva->v.Obj = *VB->store.Obj;
       if (disable & VERT_TEX0_ANY) cva->v.TexCoord[0]= *(VB->store.TexCoord[0]);
-      if (disable & VERT_TEX0_ANY) cva->v.TexCoord[1]= *(VB->store.TexCoord[1]);
+      if (disable & VERT_TEX1_ANY) cva->v.TexCoord[1]= *(VB->store.TexCoord[1]);
       if (disable & VERT_EDGE) cva->v.EdgeFlag = *VB->store.EdgeFlag;
    }
       
diff --git a/xc/extras/Mesa/src/dd.h b/xc/extras/Mesa/src/dd.h
index 90ff664ed..7b269421f 100644
--- a/xc/extras/Mesa/src/dd.h
+++ b/xc/extras/Mesa/src/dd.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/dd.h,v 1.7 2000/09/26 15:56:30 tsi Exp $ */
+
 
 
 #ifndef DD_INCLUDED
@@ -640,6 +640,16 @@ struct dd_function_table {
     * Core Mesa will perform any image format/type conversions that are needed.
     */
 
+   GLboolean (*TestProxyTexImage)(GLcontext *ctx, GLenum target,
+                                  GLint level, GLint internalFormat,
+                                  GLenum format, GLenum type,
+                                  GLint width, GLint height,
+                                  GLint depth, GLint border);
+   /* Called by glTexImage[123]D when user specifies a proxy texture
+    * target.  Return GL_TRUE if the proxy test passes, return GL_FALSE
+    * if the test fails.
+    */
+
    GLboolean (*CompressedTexImage1D)( GLcontext *ctx, GLenum target,
                                       GLint level, GLsizei imageSize,
                                       const GLvoid *data,
@@ -712,8 +722,15 @@ struct dd_function_table {
     */
 
    GLint (*SpecificCompressedTexFormat)(GLcontext *ctx,
-                                        GLint internalFormat,
-                                        GLint numDimensions);
+                                        GLint      internalFormat,
+                                        GLint      numDimensions,
+                                        GLint     *levelp,
+                                        GLsizei   *widthp,
+                                        GLsizei   *heightp,
+                                        GLsizei   *depthp,
+                                        GLint     *borderp,
+                                        GLenum    *formatp,
+                                        GLenum    *typep);
    /* Called to turn a generic texture format into a specific
     * texture format.  For example, if a driver implements
     * GL_3DFX_texture_compression_FXT1, this would map
diff --git a/xc/extras/Mesa/src/debug_xform.c b/xc/extras/Mesa/src/debug_xform.c
index 5926ba191..e38548ee7 100644
--- a/xc/extras/Mesa/src/debug_xform.c
+++ b/xc/extras/Mesa/src/debug_xform.c
@@ -2,19 +2,19 @@
 /*
  * Mesa 3-D graphics library
  * Version:  3.3
- * 
+ *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -22,7 +22,10 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/debug_xform.c,v 1.5 2000/09/26 15:56:30 tsi Exp $ */
+
+/*
+ * Updated for P6 architecture by Gareth Hughes.
+ */
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -39,121 +42,160 @@
 #ifdef DEBUG  /* This code only used for debugging */
 
 
-/*  comment this out to deactivate the cycle counter.
- *  NOTE: it works only on CPUs which know the 'rdtsc' command (586 or higher) 
- *  (hope, you don't try to debug Mesa on a 386 ;) 
+/* Comment this out to deactivate the cycle counter.
+ * NOTE: it works only on CPUs which know the 'rdtsc' command (586 or higher)
+ * (hope, you don't try to debug Mesa on a 386 ;)
  */
 #if defined(__GNUC__) && defined(__i386__) && defined(USE_X86_ASM)
 #define  RUN_XFORM_BENCHMARK
 #endif
 
-#define  TEST_COUNT  100                 /* size of the tested vector array   */
+#define TEST_COUNT		128	/* size of the tested vector array   */
 
-#define  REQUIRED_PRECISION  10          /* allow 4 bits to miss              */
-#define  MAX_PRECISION       24          /* max. precision possible           */
+#define REQUIRED_PRECISION	10	/* allow 4 bits to miss              */
+#define MAX_PRECISION		24	/* max. precision possible           */
 
 
 #ifdef  RUN_XFORM_BENCHMARK
-/*  Modify the the number of tests if you like.
- *  We take the minimum of all results, because every error should be 
- *  positive (time used by other processes, task switches etc).
- *  It is assumed that all calculations are done in the cache.
+/* Overhead of profiling counter in cycles.  Automatically adjusted to
+ * your machine at run time - counter initialization should give very
+ * consistent results.
  */
-#define  BEGIN_RACE(x)                                                        \
-x = 16000000;                                         /*  hope it's enough-*/ \
-for (cycle_i = 0; cycle_i <10; cycle_i++) {                                   \
-   long cycle_tmp1, cycle_tmp2, dummy;                                        \
-   __asm__ ("mov %%eax, %0":"=a" (cycle_tmp1));       /*  cache cycle_tmp1 */ \
-   __asm__ ("mov %%eax, %0":"=a" (cycle_tmp2));       /*  cache cycle_tmp2 */ \
-   __asm__ ("cdq");                                   /*  stall pipeline   */ \
-   __asm__ ("cdq");                                   /*  to avoid pairing */ \
-   __asm__ ("rdtsc":"=a" (cycle_tmp1), "=d" (dummy)); /*  save timestamp   */
-
-
-#define END_RACE(x)                                                           \
-   __asm__ ("cdq");                                                           \
-   __asm__ ("cdq");                                                           \
-   __asm__ ("rdtsc":"=a" (cycle_tmp2), "=d" (dummy));                         \
-   if (x > (cycle_tmp2-cycle_tmp1)) x = cycle_tmp2 - cycle_tmp1;              \
-}                                                                             \
-x -= 13;                      /*   sub the overhead of the counter procedure */
-
-#else
-#define BEGIN_RACE(x) 
-#define END_RACE(x)
-#endif
+static int need_counter = 1;
+static long counter_overhead = 0;
 
+/* Modify the the number of tests if you like.
+ * We take the minimum of all results, because every error should be
+ * positive (time used by other processes, task switches etc).
+ * It is assumed that all calculations are done in the cache.
+ */
 
-static char *mesa_profile = NULL;
-
-
-static GLfloat rnd(void)
-{
-   GLfloat f = (GLfloat)rand() / (GLfloat)RAND_MAX;
-   GLfloat gran = (GLfloat)(1 << 13);
-
-   f = (GLfloat)(GLint)(f * gran) / gran;
-
-   return f * 2.0 - 1.0;
-}
-
-
-
-static int significand_match( GLfloat a, GLfloat b )
-{
-   GLfloat d = a - b;
-   int a_ex, b_ex, d_ex;
-
-   if (d == 0.0F) {
-      return MAX_PRECISION;   /* Exact match */
-   }
-
-   if (a == 0.0F || b == 0.0F) {
-      /*
-       * It would probably be better to check if the
-       * non-zero number is denormalized and return
-       * the index of the highest set bit here.
-       */
-      return 0;
-   }
-
-   frexp(a, &a_ex);
-   frexp(b, &b_ex);
-   frexp(d, &d_ex);
-
-   if (a_ex < b_ex)
-      return a_ex - d_ex;
-   else
-      return b_ex - d_ex;
-}
+#if 1 /* PPro, PII, PIII version */
 
+/* Profiling on the P6 architecture requires a little more work, due to
+ * the internal out-of-order execution.  We must perform a serializing
+ * 'cpuid' instruction before and after the 'rdtsc' instructions to make
+ * sure no other uops are executed when we sample the timestamp counter.
+ */
+#define  INIT_COUNTER()							\
+   do {									\
+      int cycle_i;							\
+      counter_overhead = LONG_MAX;					\
+      for ( cycle_i = 0 ; cycle_i < 4 ; cycle_i++ ) {			\
+	 long cycle_tmp1 = 0, cycle_tmp2 = 0;				\
+	 __asm__ ( "push %%ebx       \n"				\
+		   "xor %%eax, %%eax \n"				\
+		   "cpuid            \n"				\
+		   "rdtsc            \n"				\
+		   "mov %%eax, %0    \n"				\
+		   "xor %%eax, %%eax \n"				\
+		   "cpuid            \n"				\
+		   "pop %%ebx        \n"				\
+		   "push %%ebx       \n"				\
+		   "xor %%eax, %%eax \n"				\
+		   "cpuid            \n"				\
+		   "rdtsc            \n"				\
+		   "mov %%eax, %1    \n"				\
+		   "xor %%eax, %%eax \n"				\
+		   "cpuid            \n"				\
+		   "pop %%ebx        \n"				\
+		   : "=m" (cycle_tmp1), "=m" (cycle_tmp2)		\
+		   : : "eax", "ecx", "edx" );				\
+	 if ( counter_overhead > (cycle_tmp2 - cycle_tmp1) ) {		\
+	    counter_overhead = cycle_tmp2 - cycle_tmp1;			\
+	 }								\
+      }									\
+   } while (0)
+
+#define  BEGIN_RACE(x)							\
+   x = LONG_MAX;							\
+   for ( cycle_i = 0 ; cycle_i < 10 ; cycle_i++ ) {			\
+      long cycle_tmp1 = 0, cycle_tmp2 = 0;				\
+      __asm__ ( "push %%ebx       \n"					\
+		"xor %%eax, %%eax \n"					\
+		"cpuid            \n"					\
+		"rdtsc            \n"					\
+		"mov %%eax, %0    \n"					\
+		"xor %%eax, %%eax \n"					\
+		"cpuid            \n"					\
+		"pop %%ebx        \n"					\
+		: "=m" (cycle_tmp1)					\
+		: : "eax", "ecx", "edx" );
+
+#define END_RACE(x)							\
+      __asm__ ( "push %%ebx       \n"					\
+		"xor %%eax, %%eax \n"					\
+		"cpuid            \n"					\
+		"rdtsc            \n"					\
+		"mov %%eax, %0    \n"					\
+		"xor %%eax, %%eax \n"					\
+		"cpuid            \n"					\
+		"pop %%ebx        \n"					\
+		: "=m" (cycle_tmp2)					\
+		: : "eax", "ecx", "edx" );				\
+      if ( x > (cycle_tmp2 - cycle_tmp1) ) {				\
+	 x = cycle_tmp2 - cycle_tmp1;					\
+      }									\
+   }									\
+   x -= counter_overhead;
+
+#else /* PPlain, PMMX version */
+
+/* To ensure accurate results, we stall the pipelines with the
+ * non-pairable 'cdq' instruction.  This ensures all the code being
+ * profiled is complete when the 'rdtsc' instruction executes.
+ */
+#define  INIT_COUNTER(x)						\
+   do {									\
+      int cycle_i;							\
+      x = LONG_MAX;							\
+      for ( cycle_i = 0 ; cycle_i < 32 ; cycle_i++ ) {			\
+	 long cycle_tmp1, cycle_tmp2, dummy;				\
+	 __asm__ ( "mov %%eax, %0" : "=a" (cycle_tmp1) );		\
+	 __asm__ ( "mov %%eax, %0" : "=a" (cycle_tmp2) );		\
+	 __asm__ ( "cdq" );						\
+	 __asm__ ( "cdq" );						\
+	 __asm__ ( "rdtsc" : "=a" (cycle_tmp1), "=d" (dummy) );		\
+	 __asm__ ( "cdq" );						\
+	 __asm__ ( "cdq" );						\
+	 __asm__ ( "rdtsc" : "=a" (cycle_tmp2), "=d" (dummy) );		\
+	 if ( x > (cycle_tmp2 - cycle_tmp1) )				\
+	    x = cycle_tmp2 - cycle_tmp1;				\
+      }									\
+   } while (0)
+
+#define  BEGIN_RACE(x)							\
+   x = LONG_MAX;							\
+   for ( cycle_i = 0 ; cycle_i < 16 ; cycle_i++ ) {			\
+      long cycle_tmp1, cycle_tmp2, dummy;				\
+      __asm__ ( "mov %%eax, %0" : "=a" (cycle_tmp1) );			\
+      __asm__ ( "mov %%eax, %0" : "=a" (cycle_tmp2) );			\
+      __asm__ ( "cdq" );						\
+      __asm__ ( "cdq" );						\
+      __asm__ ( "rdtsc" : "=a" (cycle_tmp1), "=d" (dummy) );
+
+
+#define END_RACE(x)							\
+      __asm__ ( "cdq" );						\
+      __asm__ ( "cdq" );						\
+      __asm__ ( "rdtsc" : "=a" (cycle_tmp2), "=d" (dummy) );		\
+      if ( x > (cycle_tmp2 - cycle_tmp1) )				\
+	 x = cycle_tmp2 - cycle_tmp1;					\
+   }									\
+   x -= counter_overhead;
 
+#endif
 
+#else
 
+#define INIT_COUNTER(x)
+#define BEGIN_RACE(x)
+#define END_RACE(x)
 
-static void ref_transform( GLvector4f *dst, 
-                           const GLmatrix *mat, 
-                           const GLvector4f *src,
-                           const GLubyte *clipmask, 
-                           const GLubyte flag )
-{
-   int i;
-   GLfloat *s = (GLfloat *)src->start;
-   GLfloat (*d)[4] = (GLfloat (*)[4])dst->start;
-   const GLfloat *m = mat->m;
+#endif
 
-   (void) clipmask;
-   (void) flag;
 
-   for (i = 0; i < src->count; ++i) {
-      GLfloat x = s[0], y = s[1], z = s[2], w = s[3];
-      d[i][0] = m[0]*x + m[4]*y + m[ 8]*z + m[12]*w;
-      d[i][1] = m[1]*x + m[5]*y + m[ 9]*z + m[13]*w;
-      d[i][2] = m[2]*x + m[6]*y + m[10]*z + m[14]*w;
-      d[i][3] = m[3]*x + m[7]*y + m[11]*z + m[15]*w;
-      s = (GLfloat *)((char *)s + src->stride);
-   }
-}
+static char *mesa_profile = NULL;
 
 
 enum { NIL=0, ONE=1, NEG=-1, VAR=2 };
@@ -170,7 +212,7 @@ static int m_identity[16] = {
    NIL, NIL, ONE, NIL,
    NIL, NIL, NIL, ONE
 };
-static int m_2d[16] = {
+static int  m_2d[16]  = {
    VAR, VAR, NIL, VAR,
    VAR, VAR, NIL, VAR,
    NIL, NIL, ONE, NIL,
@@ -229,23 +271,255 @@ static char *mstrings[7] = {
 };
 
 
-static int test_transform_function( transform_func fn, int psize, int mtype,
-                                    int masked, long* cycles )
+
+static int m_norm_identity[16] = {
+   ONE, NIL, NIL, NIL,
+   NIL, ONE, NIL, NIL,
+   NIL, NIL, ONE, NIL,
+   NIL, NIL, NIL, NIL
+};
+static int m_norm_general[16] = {
+   VAR, VAR, VAR, NIL,
+   VAR, VAR, VAR, NIL,
+   VAR, VAR, VAR, NIL,
+   NIL, NIL, NIL, NIL
+};
+static int m_norm_no_rot[16] = {
+   VAR, NIL, NIL, NIL,
+   NIL, VAR, NIL, NIL,
+   NIL, NIL, VAR, NIL,
+   NIL, NIL, NIL, NIL
+};
+static int *norm_templates[8] = {
+   m_norm_no_rot,
+   m_norm_no_rot,
+   m_norm_no_rot,
+   m_norm_general,
+   m_norm_general,
+   m_norm_general,
+   m_norm_identity,
+   m_norm_identity
+};
+static int norm_types[8] = {
+   NORM_TRANSFORM_NO_ROT,
+   NORM_TRANSFORM_NO_ROT | NORM_RESCALE,
+   NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE,
+   NORM_TRANSFORM,
+   NORM_TRANSFORM | NORM_RESCALE,
+   NORM_TRANSFORM | NORM_NORMALIZE,
+   NORM_RESCALE,
+   NORM_NORMALIZE
+};
+static int norm_scale_types[8] = {               /*  rescale factor          */
+   NIL,                                          /*  NIL disables rescaling  */
+   VAR,
+   NIL,
+   NIL,
+   VAR,
+   NIL,
+   VAR,
+   NIL
+};
+static int norm_normalize_types[8] = {           /*  normalizing ?? (no = 0) */
+   0,
+   0,
+   1,
+   0,
+   0,
+   1,
+   0,
+   1
+};
+static char *norm_strings[8] = {
+   "NORM_TRANSFORM_NO_ROT",
+   "NORM_TRANSFORM_NO_ROT | NORM_RESCALE",
+   "NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE",
+   "NORM_TRANSFORM",
+   "NORM_TRANSFORM | NORM_RESCALE",
+   "NORM_TRANSFORM | NORM_NORMALIZE",
+   "NORM_RESCALE",
+   "NORM_NORMALIZE"
+};
+
+
+
+/* ================================================================
+ * Helper functions
+ */
+
+static GLfloat rnd( void )
+{
+   GLfloat f = (GLfloat)rand() / (GLfloat)RAND_MAX;
+   GLfloat gran = (GLfloat)(1 << 13);
+
+   f = (GLfloat)(GLint)(f * gran) / gran;
+
+   return f * 2.0 - 1.0;
+}
+
+static int significand_match( GLfloat a, GLfloat b )
+{
+   GLfloat d = a - b;
+   int a_ex, b_ex, d_ex;
+
+   if ( d == 0.0F ) {
+      return MAX_PRECISION;   /* Exact match */
+   }
+
+   if ( a == 0.0F || b == 0.0F ) {
+      /* It would probably be better to check if the
+       * non-zero number is denormalized and return
+       * the index of the highest set bit here.
+       */
+      return 0;
+   }
+
+   frexp( a, &a_ex );
+   frexp( b, &b_ex );
+   frexp( d, &d_ex );
+
+   if ( a_ex < b_ex )
+      return a_ex - d_ex;
+   else
+      return b_ex - d_ex;
+}
+
+
+
+/* ================================================================
+ * Reference transformations
+ */
+
+static void ref_transform( GLvector4f *dst,
+                           const GLmatrix *mat,
+                           const GLvector4f *src,
+                           const GLubyte *clipmask,
+                           const GLubyte flag )
+{
+   GLuint i;
+   GLfloat *s = (GLfloat *)src->start;
+   GLfloat (*d)[4] = (GLfloat (*)[4])dst->start;
+   const GLfloat *m = mat->m;
+
+   (void) clipmask;
+   (void) flag;
+
+   for ( i = 0 ; i < src->count ; i++ ) {
+      GLfloat x = s[0], y = s[1], z = s[2], w = s[3];
+      d[i][0] = m[0]*x + m[4]*y + m[ 8]*z + m[12]*w;
+      d[i][1] = m[1]*x + m[5]*y + m[ 9]*z + m[13]*w;
+      d[i][2] = m[2]*x + m[6]*y + m[10]*z + m[14]*w;
+      d[i][3] = m[3]*x + m[7]*y + m[11]*z + m[15]*w;
+      s = (GLfloat *)((char *)s + src->stride);
+   }
+}
+
+static void ref_norm_transform_rescale( const GLmatrix *mat,
+					GLfloat scale,
+					const GLvector3f *in,
+					const GLfloat *lengths,
+					const GLubyte mask[],
+					GLvector3f *dest )
+{
+   GLuint i;
+   const GLfloat *s = in->start;
+   const GLfloat *m = mat->inv;
+   GLfloat (*out)[3] = (GLfloat (*)[3])dest->start;
+
+   (void) mask;
+   (void) lengths;
+
+   for ( i = 0 ; i < in->count ; i++ ) {
+      GLfloat x = s[0], y = s[1], z = s[2] ;
+      GLfloat tx = m[0]*x + m[1]*y + m[ 2]*z ;
+      GLfloat ty = m[4]*x + m[5]*y + m[ 6]*z ;
+      GLfloat tz = m[8]*x + m[9]*y + m[10]*z ;
+
+      out[i][0] = tx * scale;
+      out[i][1] = ty * scale;
+      out[i][2] = tz * scale;
+
+      s = (GLfloat *)((char *)s + in->stride);
+   }
+}
+
+static void ref_norm_transform_normalize( const GLmatrix *mat,
+					  GLfloat scale,
+					  const GLvector3f *in,
+					  const GLfloat *lengths,
+					  const GLubyte mask[],
+					  GLvector3f *dest )
+{
+   GLuint i;
+   const GLfloat *s = in->start;
+   const GLfloat *m = mat->inv;
+   GLfloat (*out)[3] = (GLfloat (*)[3])dest->start;
+
+   (void) mask;
+
+   for ( i = 0 ; i < in->count ; i++ ) {
+      GLfloat x = s[0], y = s[1], z = s[2] ;
+      GLfloat tx = m[0]*x + m[1]*y + m[ 2]*z ;
+      GLfloat ty = m[4]*x + m[5]*y + m[ 6]*z ;
+      GLfloat tz = m[8]*x + m[9]*y + m[10]*z ;
+
+      if ( !lengths ) {
+         GLfloat len = tx*tx + ty*ty + tz*tz;
+         if ( len > 1e-20 ) {
+	    /* Hmmm, don't know how we could test the precalculated
+	     * length case...
+	     */
+            scale = 1.0 / sqrt( len );
+            out[i][0] = tx * scale;
+            out[i][1] = ty * scale;
+            out[i][2] = tz * scale;
+         } else {
+            out[i][0] = out[i][1] = out[i][2] = 0;
+         }
+      } else {
+         scale = lengths[i];;
+         out[i][0] = tx * scale;
+         out[i][1] = ty * scale;
+         out[i][2] = tz * scale;
+      }
+
+      s = (GLfloat *)((char *)s + in->stride);
+   }
+}
+
+
+
+/* ================================================================
+ * Vertex transformation tests
+ */
+
+/* Ensure our arrays are correctly aligned.
+ */
+#if defined(__GNUC__)
+#define ALIGN16(x)		x __attribute__ ((aligned (16)))
+#else
+#define ALIGN16(x)		x
+#endif
+static GLfloat ALIGN16(s[TEST_COUNT][5]);
+static GLfloat ALIGN16(d[TEST_COUNT][4]);
+static GLfloat ALIGN16(r[TEST_COUNT][4]);
+
+static int test_transform_function( transform_func func, int psize, int mtype,
+                                    int masked, long *cycles )
 {
    GLvector4f source[1], dest[1], ref[1];
    GLmatrix mat[1];
-   GLfloat s[TEST_COUNT][5], d[TEST_COUNT][4], r[TEST_COUNT][4];
    GLfloat *m;
    GLubyte mask[TEST_COUNT];
    int i, j;
 #ifdef  RUN_XFORM_BENCHMARK
-   int cycle_i;                /* the counter for the benchmarks we run     */
+   int cycle_i;                /* the counter for the benchmarks we run */
 #endif
 
    (void) cycles;
 
-   if (psize > 4) {
-      gl_problem( NULL, "test_transform_function called with psize > 4\n");
+   if ( psize > 4 ) {
+      gl_problem( NULL, "test_transform_function called with psize > 4\n" );
       return 0;
    }
 
@@ -259,9 +533,9 @@ static int test_transform_function( transform_func fn, int psize, int mtype,
    m[2] = 44.0; m[6] =  9.0; m[10] =  7.0; m[14] =  3.0;
    m[3] = 11.0; m[7] = 23.0; m[11] = 91.0; m[15] =  9.0;
 
-   for (i = 0; i < 4; i++) {
-      for (j = 0; j < 4; j++) {
-         switch (templates[mtype][i * 4 + j]) {
+   for ( i = 0 ; i < 4 ; i++ ) {
+      for ( j = 0 ; j < 4 ; j++ ) {
+         switch ( templates[mtype][i * 4 + j] ) {
          case NIL:
             m[j * 4 + i] = 0.0;
             break;
@@ -279,13 +553,13 @@ static int test_transform_function( transform_func fn, int psize, int mtype,
       }
    }
 
-   for (i = 0; i < TEST_COUNT; ++i) {
-      mask[i] = i % 2;            /*  mask every 2nd element          */
+   for ( i = 0 ; i < TEST_COUNT ; i++) {
+      mask[i] = i % 2;				/* mask every 2nd element */
       d[i][0] = s[i][0] = 0.0;
       d[i][1] = s[i][1] = 0.0;
       d[i][2] = s[i][2] = 0.0;
       d[i][3] = s[i][3] = 1.0;
-      for (j = 0; j < psize; j++)
+      for ( j = 0 ; j < psize ; j++ )
          s[i][j] = rnd();
    }
 
@@ -310,47 +584,47 @@ static int test_transform_function( transform_func fn, int psize, int mtype,
    ref->size = 0;
    ref->flags = 0;
 
-   ref_transform(ref, mat, source, NULL, 0);
+   ref_transform( ref, mat, source, NULL, 0 );
 
-   if (mesa_profile) {
-      if (masked) {
-         BEGIN_RACE (*cycles);
-         fn(dest, mat->m, source, mask, 1);
-         END_RACE (*cycles);
+   if ( mesa_profile ) {
+      if ( masked ) {
+         BEGIN_RACE( *cycles );
+         func( dest, mat->m, source, mask, 1 );
+         END_RACE( *cycles );
       } else {
-         BEGIN_RACE (*cycles);
-         fn(dest, mat->m, source, NULL, 0);
-         END_RACE (*cycles);
-      }
+         BEGIN_RACE( *cycles );
+         func( dest, mat->m, source, NULL, 0 );
+         END_RACE( *cycles );
+     }
    }
    else {
-      if (masked) {
-         fn(dest, mat->m, source, mask, 1);
+      if ( masked ) {
+         func( dest, mat->m, source, mask, 1 );
       } else {
-         fn(dest, mat->m, source, NULL, 0);
+         func( dest, mat->m, source, NULL, 0 );
       }
    }
 
-   for (i = 0; i < TEST_COUNT; ++i) {
-      if (masked && mask[i] & 1)
+   for ( i = 0 ; i < TEST_COUNT ; i++ ) {
+      if ( masked && (mask[i] & 1) )
          continue;
 
-      for (j = 0; j < 4; j++) {
-         if (significand_match (d[i][j], r[i][j]) < REQUIRED_PRECISION) {
-            printf("-----------------------------\n");
-            printf ("(i = %i, j = %i)\n", i, j);
-            printf ("%f \t %f \t [diff = %e - %i bit missed]\n",
-                         d[i][0], r[i][0], r[i][0]-d[i][0],
-                         MAX_PRECISION - significand_match (d[i][0], r[i][0]));
-            printf ("%f \t %f \t [diff = %e - %i bit missed]\n",
-                         d[i][1], r[i][1], r[i][1]-d[i][1], 
-                         MAX_PRECISION - significand_match (d[i][1], r[i][1]));
-            printf ("%f \t %f \t [diff = %e - %i bit missed]\n",
-                         d[i][2], r[i][2], r[i][2]-d[i][2], 
-                         MAX_PRECISION - significand_match (d[i][2], r[i][2]));
-            printf ("%f \t %f \t [diff = %e - %i bit missed]\n",
-                         d[i][3], r[i][3], r[i][3]-d[i][3], 
-                         MAX_PRECISION - significand_match (d[i][3], r[i][3]));
+      for ( j = 0 ; j < 4 ; j++ ) {
+         if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) {
+            printf( "-----------------------------\n" );
+            printf( "(i = %i, j = %i)\n", i, j );
+            printf( "%f \t %f \t [diff = %e - %i bit missed]\n",
+		    d[i][0], r[i][0], r[i][0]-d[i][0],
+		    MAX_PRECISION - significand_match( d[i][0], r[i][0] ) );
+            printf( "%f \t %f \t [diff = %e - %i bit missed]\n",
+		    d[i][1], r[i][1], r[i][1]-d[i][1],
+		    MAX_PRECISION - significand_match( d[i][1], r[i][1] ) );
+            printf( "%f \t %f \t [diff = %e - %i bit missed]\n",
+		    d[i][2], r[i][2], r[i][2]-d[i][2],
+		    MAX_PRECISION - significand_match( d[i][2], r[i][2] ) );
+            printf( "%f \t %f \t [diff = %e - %i bit missed]\n",
+		    d[i][3], r[i][3], r[i][3]-d[i][3],
+		    MAX_PRECISION - significand_match( d[i][3], r[i][3] ) );
             return 0;
          }
       }
@@ -360,240 +634,90 @@ static int test_transform_function( transform_func fn, int psize, int mtype,
    return 1;
 }
 
-
-
-
 void gl_test_all_transform_functions( char *description )
 {
    int masked, psize, mtype;
-   long benchmark_tab [2][4][7];
+   long benchmark_tab[2][4][7];
    static int first_time = 1;
 
-   if (first_time) {
+   if ( first_time ) {
       first_time = 0;
-      mesa_profile = getenv("MESA_PROFILE");
+      mesa_profile = getenv( "MESA_PROFILE" );
    }
 
 #ifdef RUN_XFORM_BENCHMARK
-   if (mesa_profile) {
-      printf("transform results after hooking in %s functions:\n", description);
+   if ( mesa_profile ) {
+      if ( need_counter ) {
+	 need_counter = 0;
+	 INIT_COUNTER();
+	 printf( "counter overhead: %ld cycles\n\n", counter_overhead );
+      }
+      printf( "transform results after hooking in %s functions:\n", description );
    }
 #endif
 
-   for (masked = 0; masked <= 1; masked++) {
+   for ( masked = 0 ; masked <= 1 ; masked++ ) {
       int cma = masked ? CULL_MASK_ACTIVE : 0;
       char *cmastring = masked ? "CULL_MASK_ACTIVE" : "0";
 
 #ifdef RUN_XFORM_BENCHMARK
-      if (mesa_profile) {
-         printf ("\n culling: %s \n", masked ? "CULL_MASK_ACTIVE" : "0");
-         for (psize = 1; psize <= 4; psize++) {
-            printf(" p%d\t", psize );
+      if ( mesa_profile ) {
+         printf( "\n culling: %s \n", masked ? "CULL_MASK_ACTIVE" : "0" );
+         for ( psize = 1 ; psize <= 4 ; psize++ ) {
+            printf( " p%d\t", psize );
          }
-         printf("\n--------------------------------------------------------\n");
+         printf( "\n--------------------------------------------------------\n" );
       }
 #endif
 
-      for (mtype = 0; mtype < 7; mtype++) {
-         for (psize = 1; psize <= 4; psize++) {
-            transform_func fn = gl_transform_tab[cma][psize][mtypes[mtype]];
-            long* cycles = &(benchmark_tab [cma][psize-1][mtype]);
+      for ( mtype = 0 ; mtype < 7 ; mtype++ ) {
+         for ( psize = 1 ; psize <= 4 ; psize++ ) {
+            transform_func func = gl_transform_tab[cma][psize][mtypes[mtype]];
+            long *cycles = &(benchmark_tab[cma][psize-1][mtype]);
 
-            if (test_transform_function (fn,psize,mtype,masked,cycles) == 0 ) {
+            if ( test_transform_function( func, psize, mtype,
+					  masked, cycles ) == 0 ) {
                char buf[100];
                sprintf( buf, "gl_transform_tab[%s][%d][%s] failed test (%s)",
                         cmastring, psize, mstrings[mtype], description );
                gl_problem( NULL, buf );
-            } else {
-/*             printf("gl_transform_tab[%s][%d][%s] passed test in %i cycles\n",
-	       cmastring, psize, mstrings[mtype],
-	       benchmark_tab [cma][psize-1][mtype]); 
-*/          }
+            }
 #ifdef RUN_XFORM_BENCHMARK
-            if (mesa_profile)
-               printf(" %li\t", benchmark_tab [cma][psize-1][mtype] );
+            if ( mesa_profile )
+               printf( " %li\t", benchmark_tab[cma][psize-1][mtype] );
 #endif
          }
 #ifdef RUN_XFORM_BENCHMARK
-         if (mesa_profile)
-            printf (" | [%s]\n", mstrings[mtype] );
+         if ( mesa_profile )
+            printf( " | [%s]\n", mstrings[mtype] );
 #endif
       }
 #ifdef RUN_XFORM_BENCHMARK
-      if (mesa_profile)
-         printf ("\n");
+      if ( mesa_profile )
+         printf( "\n" );
 #endif
    }
 }
 
 
 
+/* ================================================================
+ * Normal transformation tests
+ */
 
-
-
-
-
-static void ref_norm_transform_rescale ( const GLmatrix *mat,
-                                    GLfloat scale,
-                                    const GLvector3f *in,
-                                    const GLfloat *lengths,
-                                    const GLubyte mask[],
-                                    GLvector3f *dest )
-{
-   int i;
-   const GLfloat *s = in->start;
-   const GLfloat *m = mat->inv;
-   GLfloat (*out)[3] = (GLfloat (*)[3])dest->start;
-
-   (void) mask;
-   (void) lengths;
-
-   for (i = 0; i < in->count; ++i) {
-      GLfloat x = s[0], y = s[1], z = s[2] ;
-      GLfloat tx = m[0]*x + m[1]*y + m[ 2]*z ;
-      GLfloat ty = m[4]*x + m[5]*y + m[ 6]*z ;
-      GLfloat tz = m[8]*x + m[9]*y + m[10]*z ;
-
-      out[i][0] = tx * scale;
-      out[i][1] = ty * scale;
-      out[i][2] = tz * scale;
-
-      s = (GLfloat *)((char *)s + in->stride);
-   }
-}
-
-
-static void ref_norm_transform_normalize ( const GLmatrix *mat,
-                                      GLfloat scale,
-                                      const GLvector3f *in,
-                                      const GLfloat *lengths,
-                                      const GLubyte mask[],
-                                      GLvector3f *dest )
-{
-   int i;
-   const GLfloat *s = in->start;
-   const GLfloat *m = mat->inv;
-   GLfloat (*out)[3] = (GLfloat (*)[3])dest->start;
-
-   (void) mask;
-
-   for (i = 0; i < in->count; ++i) {
-      GLfloat x = s[0], y = s[1], z = s[2] ;
-      GLfloat tx = m[0]*x + m[1]*y + m[ 2]*z ;
-      GLfloat ty = m[4]*x + m[5]*y + m[ 6]*z ;
-      GLfloat tz = m[8]*x + m[9]*y + m[10]*z ;
-
-      if (!lengths) {
-         GLfloat len = tx*tx + ty*ty + tz*tz;
-         if (len > 1e-20) {
-            scale = 1.0 / sqrt (len);           /*  hmmm, don't know how we   */
-                                                /*  could test the precalcu-  */
-                                                /*  lated length case ...     */
-            out[i][0] = tx * scale;
-            out[i][1] = ty * scale;
-            out[i][2] = tz * scale;
-         } else {
-            out[i][0] = out[i][1] = out[i][2] = 0;
-         }
-      } else {
-         scale = lengths [i];;
-         out[i][0] = tx * scale;
-         out[i][1] = ty * scale;
-         out[i][2] = tz * scale;
-      }
-
-      s = (GLfloat *)((char *)s + in->stride);
-   }
-}
-
-
-
-static int m_norm_identity[16] = {
-   ONE, NIL, NIL, NIL,
-   NIL, ONE, NIL, NIL,
-   NIL, NIL, ONE, NIL,
-   NIL, NIL, NIL, NIL
-};
-static int m_norm_general[16] = {
-   VAR, VAR, VAR, NIL,
-   VAR, VAR, VAR, NIL,
-   VAR, VAR, VAR, NIL,
-   NIL, NIL, NIL, NIL
-};
-static int m_norm_no_rot[16] = {
-   VAR, NIL, NIL, NIL,
-   NIL, VAR, NIL, NIL,
-   NIL, NIL, VAR, NIL,
-   NIL, NIL, NIL, NIL
-};
-static int *norm_templates[8] = {
-   m_norm_no_rot,
-   m_norm_no_rot,
-   m_norm_no_rot,
-   m_norm_general,
-   m_norm_general,
-   m_norm_general,
-   m_norm_identity,
-   m_norm_identity
-};
-static int norm_types[8] = {
-   NORM_TRANSFORM_NO_ROT,
-   NORM_TRANSFORM_NO_ROT | NORM_RESCALE,
-   NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE,
-   NORM_TRANSFORM,
-   NORM_TRANSFORM | NORM_RESCALE,
-   NORM_TRANSFORM | NORM_NORMALIZE,
-   NORM_RESCALE,
-   NORM_NORMALIZE
-};
-static int norm_scale_types[8] = {               /*  rescale factor           */
-   NIL,                                          /*  NIL disables rescaling   */
-   VAR,
-   NIL,
-   NIL,
-   VAR,
-   NIL,
-   VAR,
-   NIL
-};
-static int norm_normalize_types[8] = {           /*  normalizing ??  (no = 0) */
-   0,
-   0,
-   1,
-   0,
-   0,
-   1,
-   0,
-   1
-};
-static char *norm_strings[8] = {
-   "NORM_TRANSFORM_NO_ROT",
-   "NORM_TRANSFORM_NO_ROT | NORM_RESCALE",
-   "NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE",
-   "NORM_TRANSFORM",
-   "NORM_TRANSFORM | NORM_RESCALE",
-   "NORM_TRANSFORM | NORM_NORMALIZE",
-   "NORM_RESCALE",
-   "NORM_NORMALIZE"
-};
-
-
-
-
-
-static int test_norm_function( normal_func fn, int mtype,
-                                    int masked, long* cycles )
+static int test_norm_function( normal_func func, int mtype,
+			       int masked, long *cycles )
 {
    GLvector3f source[1], dest[1], dest2[1], ref[1], ref2[1];
    GLmatrix mat[1];
-   GLfloat s [TEST_COUNT][5], d [TEST_COUNT][3], r [TEST_COUNT][3];
-   GLfloat d2 [TEST_COUNT][3], r2 [TEST_COUNT][3], length [TEST_COUNT];
+   GLfloat s[TEST_COUNT][5], d[TEST_COUNT][3], r[TEST_COUNT][3];
+   GLfloat d2[TEST_COUNT][3], r2[TEST_COUNT][3], length[TEST_COUNT];
    GLfloat scale;
    GLfloat *m;
    GLubyte mask[TEST_COUNT];
    int i, j;
 #ifdef  RUN_XFORM_BENCHMARK
-   int cycle_i;                /* the counter for the benchmarks we run     */
+   int cycle_i;		/* the counter for the benchmarks we run */
 #endif
 
    (void) cycles;
@@ -608,9 +732,9 @@ static int test_norm_function( normal_func fn, int mtype,
 
    scale = 1.0F + rnd () * norm_scale_types[mtype];
 
-   for (i = 0; i < 4; i++) {
-      for (j = 0; j < 4; j++) {
-         switch (norm_templates[mtype][i * 4 + j]) {
+   for ( i = 0 ; i < 4 ; i++ ) {
+      for ( j = 0 ; j < 4 ; j++ ) {
+         switch ( norm_templates[mtype][i * 4 + j] ) {
          case NIL:
             m[j * 4 + i] = 0.0;
             break;
@@ -628,14 +752,16 @@ static int test_norm_function( normal_func fn, int mtype,
       }
    }
 
-   for (i = 0; i < TEST_COUNT; ++i) {
-      mask[i] = i % 2;                         /*  mask every 2nd element     */
+   for ( i = 0 ; i < TEST_COUNT ; i++ ) {
+      mask[i] = i % 2;				/* mask every 2nd element */
       d[i][0] = s[i][0] = d2[i][0] = 0.0;
       d[i][1] = s[i][1] = d2[i][1] = 0.0;
       d[i][2] = s[i][2] = d2[i][2] = 0.0;
-      for (j = 0; j < 3; j++)
+      for ( j = 0 ; j < 3 ; j++ )
          s[i][j] = rnd();
-      length[i] = 1 / sqrt(s[i][0]*s[i][0] + s[i][1]*s[i][1] + s[i][2]*s[i][2]);
+      length[i] = 1 / sqrt( s[i][0]*s[i][0] +
+			    s[i][1]*s[i][1] +
+			    s[i][2]*s[i][2] );
    }
 
    source->data = (GLfloat(*)[3])s;
@@ -668,68 +794,68 @@ static int test_norm_function( normal_func fn, int mtype,
    ref2->stride = sizeof(float[3]);
    ref2->flags = 0;
 
-   if (norm_normalize_types [mtype] == 0) {
-      ref_norm_transform_rescale (mat, scale, source, NULL, NULL, ref);
+   if ( norm_normalize_types[mtype] == 0 ) {
+      ref_norm_transform_rescale( mat, scale, source, NULL, NULL, ref );
    } else {
-      ref_norm_transform_normalize (mat, scale, source, NULL, NULL, ref);
-      ref_norm_transform_normalize (mat, scale, source, length, NULL, ref2);
+      ref_norm_transform_normalize( mat, scale, source, NULL, NULL, ref );
+      ref_norm_transform_normalize( mat, scale, source, length, NULL, ref2 );
    }
 
-   if (mesa_profile) {
-      if (masked) {
-         BEGIN_RACE (*cycles);
-         fn (mat, scale, source, NULL, mask, dest);
-         END_RACE (*cycles);
-         fn (mat, scale, source, length, mask, dest2);
+   if ( mesa_profile ) {
+      if ( masked ) {
+         BEGIN_RACE( *cycles );
+         func( mat, scale, source, NULL, mask, dest );
+         END_RACE( *cycles );
+         func( mat, scale, source, length, mask, dest2 );
       } else {
-         BEGIN_RACE (*cycles);
-         fn (mat, scale, source, NULL, NULL, dest);
-         END_RACE (*cycles);
-         fn (mat, scale, source, length, NULL, dest2);
+         BEGIN_RACE( *cycles );
+         func( mat, scale, source, NULL, NULL, dest );
+         END_RACE( *cycles );
+         func( mat, scale, source, length, NULL, dest2 );
       }
    } else {
-      if (masked) {
-         fn (mat, scale, source, NULL, mask, dest);
-         fn (mat, scale, source, length, mask, dest2);
+      if ( masked ) {
+         func( mat, scale, source, NULL, mask, dest );
+         func( mat, scale, source, length, mask, dest2 );
       } else {
-         fn (mat, scale, source, NULL, NULL, dest);
-         fn (mat, scale, source, length, NULL, dest2);
+         func( mat, scale, source, NULL, NULL, dest );
+         func( mat, scale, source, length, NULL, dest2 );
       }
    }
 
-   for (i = 0; i < TEST_COUNT; ++i) {
-      if (masked && !(mask[i] & 1))
+   for ( i = 0 ; i < TEST_COUNT ; i++ ) {
+      if ( masked && !(mask[i] & 1) )
          continue;
 
-      for (j = 0; j < 3; j++) {
-         if (significand_match (d[i][j], r[i][j]) < REQUIRED_PRECISION) {
-            printf("-----------------------------\n");
-            printf ("(i = %i, j = %i)\n", i, j);
-            printf ("%f \t %f \t [ratio = %e - %i bit missed]\n",
-                         d[i][0], r[i][0], r[i][0]/d[i][0],
-                         MAX_PRECISION - significand_match (d[i][0], r[i][0]));
-            printf ("%f \t %f \t [ratio = %e - %i bit missed]\n",
-                         d[i][1], r[i][1], r[i][1]/d[i][1],
-                         MAX_PRECISION - significand_match (d[i][1], r[i][1]));
-            printf ("%f \t %f \t [ratio = %e - %i bit missed]\n",
-                         d[i][2], r[i][2], r[i][2]/d[i][2],
-                         MAX_PRECISION - significand_match (d[i][2], r[i][2]));
+      for ( j = 0 ; j < 3 ; j++ ) {
+         if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) {
+            printf( "-----------------------------\n" );
+            printf( "(i = %i, j = %i)\n", i, j );
+            printf( "%f \t %f \t [ratio = %e - %i bit missed]\n",
+		    d[i][0], r[i][0], r[i][0]/d[i][0],
+		    MAX_PRECISION - significand_match( d[i][0], r[i][0] ) );
+            printf( "%f \t %f \t [ratio = %e - %i bit missed]\n",
+		    d[i][1], r[i][1], r[i][1]/d[i][1],
+		    MAX_PRECISION - significand_match( d[i][1], r[i][1] ) );
+            printf( "%f \t %f \t [ratio = %e - %i bit missed]\n",
+		    d[i][2], r[i][2], r[i][2]/d[i][2],
+		    MAX_PRECISION - significand_match( d[i][2], r[i][2] ) );
             return 0;
          }
 
-         if (norm_normalize_types [mtype] != 0) {
-            if (significand_match (d2[i][j], r2[i][j]) < REQUIRED_PRECISION) {
-               printf("------------------- precalculated length case ------\n");
-               printf ("(i = %i, j = %i)\n", i, j);
-               printf ("%f \t %f \t [ratio = %e - %i bit missed]\n",
-                         d2[i][0], r2[i][0], r2[i][0]/d2[i][0],
-                         MAX_PRECISION - significand_match (d2[i][0],r2[i][0]));
-               printf ("%f \t %f \t [ratio = %e - %i bit missed]\n",
-                         d2[i][1], r2[i][1], r2[i][1]/d2[i][1], 
-                         MAX_PRECISION - significand_match (d2[i][1],r2[i][1]));
-               printf ("%f \t %f \t [ratio = %e - %i bit missed]\n",
-                         d2[i][2], r2[i][2], r2[i][2]/d2[i][2], 
-                         MAX_PRECISION - significand_match (d2[i][2],r2[i][2]));
+         if ( norm_normalize_types[mtype] != 0 ) {
+            if ( significand_match( d2[i][j], r2[i][j] ) < REQUIRED_PRECISION ) {
+               printf( "------------------- precalculated length case ------\n" );
+               printf( "(i = %i, j = %i)\n", i, j );
+               printf( "%f \t %f \t [ratio = %e - %i bit missed]\n",
+		       d2[i][0], r2[i][0], r2[i][0]/d2[i][0],
+		       MAX_PRECISION - significand_match( d2[i][0], r2[i][0] ) );
+               printf( "%f \t %f \t [ratio = %e - %i bit missed]\n",
+		       d2[i][1], r2[i][1], r2[i][1]/d2[i][1],
+		       MAX_PRECISION - significand_match( d2[i][1], r2[i][1] ) );
+               printf( "%f \t %f \t [ratio = %e - %i bit missed]\n",
+		       d2[i][2], r2[i][2], r2[i][2]/d2[i][2],
+		       MAX_PRECISION - significand_match( d2[i][2], r2[i][2] ) );
                return 0;
             }
          }
@@ -740,68 +866,68 @@ static int test_norm_function( normal_func fn, int mtype,
    return 1;
 }
 
-
-
 void gl_test_all_normal_transform_functions( char *description )
 {
    int masked;
    int mtype;
-   long benchmark_tab [0xf][0x4];
+   long benchmark_tab[0xf][0x4];
    static int first_time = 1;
 
-
-   if (first_time) {
+   if ( first_time ) {
       first_time = 0;
-      mesa_profile = getenv("MESA_PROFILE");
+      mesa_profile = getenv( "MESA_PROFILE" );
    }
 
 #ifdef RUN_XFORM_BENCHMARK
-   if (mesa_profile) {
-      printf ("normal transform results after hooking in %s functions:\n", 
-                                                                description);
+   if ( mesa_profile ) {
+      if ( need_counter ) {
+	 need_counter = 0;
+	 INIT_COUNTER();
+	 printf( "counter overhead: %ld cycles\n\n", counter_overhead );
+      }
+      printf( "normal transform results after hooking in %s functions:\n",
+	      description );
    }
 #endif
 
-
-   for (masked = 0; masked <= 1; masked++) {
+   for ( masked = 0 ; masked <= 1 ; masked++ ) {
       int cma = masked ? CULL_MASK_ACTIVE : 0;
       char *cmastring = masked ? "CULL_MASK_ACTIVE" : "0";
 
 #ifdef RUN_XFORM_BENCHMARK
-      if (mesa_profile) {
-         printf ("\n culling: %s \n", masked ? "CULL_MASK_ACTIVE" : "0");
-         printf ("\n-------------------------------------------------------\n");
+      if ( mesa_profile ) {
+         printf( "\n culling: %s \n", masked ? "CULL_MASK_ACTIVE" : "0" );
+         printf( "\n-------------------------------------------------------\n" );
       }
 #endif
 
-      for (mtype = 0; mtype < 8; mtype++) {
-         normal_func fn = gl_normal_tab [norm_types[mtype]][cma];
-         long* cycles = &(benchmark_tab [mtype][cma]);
+      for ( mtype = 0 ; mtype < 8 ; mtype++ ) {
+         normal_func func = gl_normal_tab[norm_types[mtype]][cma];
+         long *cycles = &(benchmark_tab[mtype][cma]);
 
-         if (test_norm_function (fn, mtype, masked, cycles) == 0 ) {
+         if ( test_norm_function( func, mtype, masked, cycles ) == 0 ) {
             char buf[100];
             sprintf( buf, "gl_normal_tab[%s][%s] failed test (%s)",
                      cmastring, norm_strings[mtype], description );
             gl_problem( NULL, buf );
-         } else {
-/*          printf("gl_normal_tab[%s][%s] passed test in %li cycles (%s)\n",
-                     cmastring, norm_strings[mtype],
-                     benchmark_tab [mtype][cma], description);
-*/       }
+	 }
 
 #ifdef RUN_XFORM_BENCHMARK
-         if (mesa_profile) {
-            printf (" %li\t", benchmark_tab [mtype][cma] );
-            printf (" | [%s]\n", norm_strings[mtype] );
+         if ( mesa_profile ) {
+            printf( " %li\t", benchmark_tab[mtype][cma] );
+            printf( " | [%s]\n", norm_strings[mtype] );
          }
       }
-      if (mesa_profile) { printf ("\n"); }
+      if ( mesa_profile )
+	 printf( "\n" );
 #else
       }
 #endif
    }
+#ifdef RUN_XFORM_BENCHMARK
+   if ( mesa_profile )
+      fflush( stdout );
+#endif
 }
 
-
-
-#endif /*DEBUG*/
+#endif /* DEBUG */
diff --git a/xc/extras/Mesa/src/depth.c b/xc/extras/Mesa/src/depth.c
index ae119018e..600ae105f 100644
--- a/xc/extras/Mesa/src/depth.c
+++ b/xc/extras/Mesa/src/depth.c
@@ -1457,14 +1457,14 @@ _mesa_read_depth_span( GLcontext *ctx,
       /* read from software depth buffer */
       if (ctx->Visual->DepthBits <= 16) {
          const GLushort *zptr = Z_ADDRESS16( ctx, x, y );
-         GLuint i;
+         GLint i;
          for (i = 0; i < n; i++) {
             depth[i] = zptr[i];
          }
       }
       else {
          const GLuint *zptr = Z_ADDRESS32( ctx, x, y );
-         GLuint i;
+         GLint i;
          for (i = 0; i < n; i++) {
             depth[i] = zptr[i];
          }
@@ -1529,14 +1529,14 @@ _mesa_read_depth_span_float( GLcontext *ctx,
       /* read from software depth buffer */
       if (ctx->Visual->DepthBits <= 16) {
          const GLushort *zptr = Z_ADDRESS16( ctx, x, y );
-         GLuint i;
+         GLint i;
          for (i = 0; i < n; i++) {
             depth[i] = (GLfloat) zptr[i] * scale;
          }
       }
       else {
          const GLuint *zptr = Z_ADDRESS32( ctx, x, y );
-         GLuint i;
+         GLint i;
          for (i = 0; i < n; i++) {
             depth[i] = (GLfloat) zptr[i] * scale;
          }
@@ -1545,7 +1545,7 @@ _mesa_read_depth_span_float( GLcontext *ctx,
    else if (ctx->Driver.ReadDepthSpan) {
       /* read from hardware depth buffer */
       GLdepth d[MAX_WIDTH];
-      GLuint i;
+      GLint i;
       assert(n <= MAX_WIDTH);
       (*ctx->Driver.ReadDepthSpan)( ctx, n, x, y, d );
       for (i = 0; i < n; i++) {
diff --git a/xc/extras/Mesa/src/dlist.c b/xc/extras/Mesa/src/dlist.c
index 5c26ba403..ff23ee951 100644
--- a/xc/extras/Mesa/src/dlist.c
+++ b/xc/extras/Mesa/src/dlist.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/dlist.c,v 1.7 2000/09/26 15:56:30 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/drawpix.c b/xc/extras/Mesa/src/drawpix.c
index ecb1b22af..9955367ba 100644
--- a/xc/extras/Mesa/src/drawpix.c
+++ b/xc/extras/Mesa/src/drawpix.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/drawpix.c,v 1.6 2000/09/26 15:56:31 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -125,7 +125,6 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
        && !ctx->Pixel.MinMaxEnabled
        && !ctx->Pixel.HistogramEnabled
        && ctx->Pixel.IndexShift==0 && ctx->Pixel.IndexOffset==0
-       && ctx->Pixel.MapColorFlag==0
        && ctx->Texture.ReallyEnabled == 0
        && unpack->Alignment==1
        && !unpack->SwapBytes
@@ -218,7 +217,8 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
        * skip "skipRows" rows and skip "skipPixels" pixels/row.
        */
 
-      if (format==GL_RGBA && type==GL_UNSIGNED_BYTE) {
+      if (format==GL_RGBA && type==GL_UNSIGNED_BYTE
+          && ctx->Pixel.MapColorFlag==0) {
          if (ctx->Visual->RGBAflag) {
             GLubyte *src = (GLubyte *) pixels
                + (skipRows * rowLength + skipPixels) * 4;
@@ -227,7 +227,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                GLint row;
                for (row=0; row<drawHeight; row++) {
                   (*ctx->Driver.WriteRGBASpan)(ctx, drawWidth, destX, destY,
-                                               (void *) src, NULL);
+                                             (CONST GLubyte (*)[4]) src, NULL);
                   src += rowLength * 4;
                   destY++;
                }
@@ -238,7 +238,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                for (row=0; row<drawHeight; row++) {
                   destY--;
                   (*ctx->Driver.WriteRGBASpan)(ctx, drawWidth, destX, destY,
-                                              (void *) src, NULL);
+                                             (CONST GLubyte (*)[4]) src, NULL);
                   src += rowLength * 4;
                }
             }
@@ -247,7 +247,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                GLint row;
                for (row=0; row<drawHeight; row++) {
                   gl_write_zoomed_rgba_span(ctx, drawWidth, destX, destY,
-                                            zSpan, (void *) src, zoomY0);
+                                    zSpan, (CONST GLubyte (*)[4]) src, zoomY0);
                   src += rowLength * 4;
                   destY++;
                }
@@ -255,7 +255,8 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
          }
          return GL_TRUE;
       }
-      else if (format==GL_RGB && type==GL_UNSIGNED_BYTE) {
+      else if (format==GL_RGB && type==GL_UNSIGNED_BYTE
+               && ctx->Pixel.MapColorFlag==0) {
          if (ctx->Visual->RGBAflag) {
             GLubyte *src = (GLubyte *) pixels
                + (skipRows * rowLength + skipPixels) * 3;
@@ -263,7 +264,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                GLint row;
                for (row=0; row<drawHeight; row++) {
                   (*ctx->Driver.WriteRGBSpan)(ctx, drawWidth, destX, destY,
-                                              (void *) src, NULL);
+                                             (CONST GLubyte (*)[3]) src, NULL);
                   src += rowLength * 3;
                   destY++;
                }
@@ -274,7 +275,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                for (row=0; row<drawHeight; row++) {
                   destY--;
                   (*ctx->Driver.WriteRGBSpan)(ctx, drawWidth, destX, destY,
-                                              (void *) src, NULL);
+                                             (CONST GLubyte (*)[3]) src, NULL);
                   src += rowLength * 3;
                }
             }
@@ -283,7 +284,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                GLint row;
                for (row=0; row<drawHeight; row++) {
                   gl_write_zoomed_rgb_span(ctx, drawWidth, destX, destY,
-                                           zSpan, (void *) src, zoomY0);
+                                    zSpan, (CONST GLubyte (*)[3]) src, zoomY0);
                   src += rowLength * 3;
                   destY++;
                }
@@ -291,7 +292,8 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
          }
          return GL_TRUE;
       }
-      else if (format==GL_LUMINANCE && type==GL_UNSIGNED_BYTE) {
+      else if (format==GL_LUMINANCE && type==GL_UNSIGNED_BYTE
+               && ctx->Pixel.MapColorFlag==0) {
          if (ctx->Visual->RGBAflag) {
             GLubyte *src = (GLubyte *) pixels
                + (skipRows * rowLength + skipPixels);
@@ -307,7 +309,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                      rgb[i][2] = src[i];
 		  }
                   (*ctx->Driver.WriteRGBSpan)(ctx, drawWidth, destX, destY,
-                                              (void *) rgb, NULL);
+                                             (CONST GLubyte (*)[3]) rgb, NULL);
                   src += rowLength;
                   destY++;
                }
@@ -325,7 +327,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                   }
                   destY--;
                   (*ctx->Driver.WriteRGBSpan)(ctx, drawWidth, destX, destY,
-                                              (void *) rgb, NULL);
+                                             (CONST GLubyte (*)[3]) rgb, NULL);
                   src += rowLength;
                }
             }
@@ -341,7 +343,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                      rgb[i][2] = src[i];
 		  }
                   gl_write_zoomed_rgb_span(ctx, drawWidth, destX, destY,
-                                           zSpan, (void *) rgb, zoomY0);
+                                    zSpan, (CONST GLubyte (*)[3]) rgb, zoomY0);
                   src += rowLength;
                   destY++;
                }
@@ -349,7 +351,8 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
          }
          return GL_TRUE;
       }
-      else if (format==GL_LUMINANCE_ALPHA && type==GL_UNSIGNED_BYTE) {
+      else if (format==GL_LUMINANCE_ALPHA && type==GL_UNSIGNED_BYTE
+               && ctx->Pixel.MapColorFlag==0) {
          if (ctx->Visual->RGBAflag) {
             GLubyte *src = (GLubyte *) pixels
                + (skipRows * rowLength + skipPixels)*2;
@@ -367,7 +370,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                      rgba[i][3] = *ptr++;
 		  }
                   (*ctx->Driver.WriteRGBASpan)(ctx, drawWidth, destX, destY,
-                                               (void *) rgba, NULL);
+                                            (CONST GLubyte (*)[4]) rgba, NULL);
                   src += rowLength*2;
                   destY++;
                }
@@ -387,7 +390,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                   }
                   destY--;
                   (*ctx->Driver.WriteRGBASpan)(ctx, drawWidth, destX, destY,
-                                               (void *) rgba, NULL);
+                                            (CONST GLubyte (*)[4]) rgba, NULL);
                   src += rowLength*2;
                }
             }
@@ -405,7 +408,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                      rgba[i][3] = *ptr++;
 		  }
                   gl_write_zoomed_rgba_span(ctx, drawWidth, destX, destY,
-                                            zSpan, (void *) rgba, zoomY0);
+                                   zSpan, (CONST GLubyte (*)[4]) rgba, zoomY0);
                   src += rowLength*2;
                   destY++;
                }
@@ -424,7 +427,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                   ASSERT(drawWidth < MAX_WIDTH);
                   _mesa_map_ci8_to_rgba(ctx, drawWidth, src, rgba);
                   (*ctx->Driver.WriteRGBASpan)(ctx, drawWidth, destX, destY,
-                                               (const GLubyte (*)[4])rgba, 
+                                               (CONST GLubyte (*)[4])rgba, 
 					       NULL);
                   src += rowLength;
                   destY++;
@@ -439,7 +442,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                   _mesa_map_ci8_to_rgba(ctx, drawWidth, src, rgba);
                   destY--;
                   (*ctx->Driver.WriteRGBASpan)(ctx, drawWidth, destX, destY,
-                                               (const GLubyte (*)[4])rgba, 
+                                               (CONST GLubyte (*)[4])rgba, 
                                                NULL);
                   src += rowLength;
                }
@@ -452,7 +455,7 @@ simple_DrawPixels( GLcontext *ctx, GLint x, GLint y,
                   ASSERT(drawWidth < MAX_WIDTH);
                   _mesa_map_ci8_to_rgba(ctx, drawWidth, src, rgba);
                   gl_write_zoomed_rgba_span(ctx, drawWidth, destX, destY,
-                                            zSpan, (void *) rgba, zoomY0);
+                                   zSpan, (CONST GLubyte (*)[4]) rgba, zoomY0);
                   src += rowLength;
                   destY++;
                }
@@ -642,8 +645,9 @@ draw_depth_pixels( GLcontext *ctx, GLint x, GLint y,
       GLint row;
       for (row = 0; row < height; row++, y++) {
          GLdepth zspan[MAX_WIDTH];
-         const GLushort *zptr = _mesa_image_address(&ctx->Unpack,
-                pixels, width, height, GL_DEPTH_COMPONENT, type, 0, row, 0);
+         const GLushort *zptr = (const GLushort *)
+            _mesa_image_address(&ctx->Unpack, pixels, width, height,
+                                GL_DEPTH_COMPONENT, type, 0, row, 0);
          GLint i;
          for (i = 0; i < width; i++)
             zspan[i] = zptr[i];
@@ -655,8 +659,9 @@ draw_depth_pixels( GLcontext *ctx, GLint x, GLint y,
       /* Special case: directly write 32-bit depth values */
       GLint row;
       for (row = 0; row < height; row++, y++) {
-         const GLuint *zptr = _mesa_image_address(&ctx->Unpack,
-                pixels, width, height, GL_DEPTH_COMPONENT, type, 0, row, 0);
+         const GLuint *zptr = (const GLuint *)
+            _mesa_image_address(&ctx->Unpack, pixels, width, height,
+                                GL_DEPTH_COMPONENT, type, 0, row, 0);
          gl_write_rgba_span( ctx, width, x, y, zptr, rgba, GL_BITMAP );
       }
    }
@@ -672,7 +677,7 @@ draw_depth_pixels( GLcontext *ctx, GLint x, GLint y,
          if (ctx->Visual->RGBAflag) {
             if (zoom) {
                gl_write_zoomed_rgba_span(ctx, width, x, y, zspan,
-                                         (const GLubyte (*)[4])rgba, desty);
+                                         (CONST GLubyte (*)[4])rgba, desty);
             }
             else {
                gl_write_rgba_span(ctx, width, x, y, zspan, rgba, GL_BITMAP);
@@ -707,6 +712,11 @@ draw_rgba_pixels( GLcontext *ctx, GLint x, GLint y,
    GLdepth zspan[MAX_WIDTH];
    GLboolean quickDraw;
 
+   if (!_mesa_is_legal_format_and_type(format, type)) {
+      gl_error(ctx, GL_INVALID_ENUM, "glDrawPixels(format or type)");
+      return;
+   }
+
    /* Try an optimized glDrawPixels first */
    if (simple_DrawPixels(ctx, x, y, width, height, format, type, pixels))
       return;
@@ -743,8 +753,8 @@ draw_rgba_pixels( GLcontext *ctx, GLint x, GLint y,
       for (row = 0; row < height; row++, y++) {
          const GLvoid *source = _mesa_image_address(unpack,
                   pixels, width, height, format, type, 0, row, 0);
-         _mesa_unpack_ubyte_color_span(ctx, width, GL_RGBA, (void*) rgba,
-                   format, type, source, unpack, GL_TRUE);
+         _mesa_unpack_ubyte_color_span(ctx, width, GL_RGBA,
+                   (GLubyte *) rgba, format, type, source, unpack, GL_TRUE);
          if ((ctx->Pixel.MinMaxEnabled && ctx->MinMax.Sink) ||
              (ctx->Pixel.HistogramEnabled && ctx->Histogram.Sink))
             continue;
@@ -758,7 +768,7 @@ draw_rgba_pixels( GLcontext *ctx, GLint x, GLint y,
             MEMCPY(primary_rgba, rgba, 4 * width * sizeof(GLubyte));
 
             for (unit = 0; unit < MAX_TEXTURE_UNITS; unit++) {
-               _mesa_pixeltexgen(ctx, width, (const GLubyte (*)[4]) rgba,
+               _mesa_pixeltexgen(ctx, width, (CONST GLubyte (*)[4]) rgba,
                                  s, t, r, q);
                gl_texture_pixels(ctx, unit, width, s, t, r, NULL,
                                  primary_rgba, rgba);
@@ -767,11 +777,11 @@ draw_rgba_pixels( GLcontext *ctx, GLint x, GLint y,
 
          if (quickDraw) {
             (*ctx->Driver.WriteRGBASpan)( ctx, width, x, y,
-                                          (CONST GLubyte (*)[]) rgba, NULL);
+                                          (CONST GLubyte (*)[4]) rgba, NULL);
          }
          else if (zoom) {
             gl_write_zoomed_rgba_span( ctx, width, x, y, zspan, 
-				       (CONST GLubyte (*)[]) rgba, desty );
+				       (CONST GLubyte (*)[4]) rgba, desty );
          }
          else {
             gl_write_rgba_span( ctx, (GLuint) width, x, y, zspan, rgba, GL_BITMAP);
@@ -814,6 +824,7 @@ _mesa_DrawPixels( GLsizei width, GLsizei height,
          return;
       }
 
+      RENDER_START(ctx);
       switch (format) {
 	 case GL_STENCIL_INDEX:
 	    draw_stencil_pixels( ctx, x, y, width, height, type, pixels );
@@ -842,8 +853,8 @@ _mesa_DrawPixels( GLsizei width, GLsizei height,
 	    break;
 	 default:
 	    gl_error( ctx, GL_INVALID_ENUM, "glDrawPixels(format)" );
-            return;
       }
+      RENDER_FINISH(ctx);
    }
    else if (ctx->RenderMode==GL_FEEDBACK) {
       if (ctx->Current.RasterPosValid) {
diff --git a/xc/extras/Mesa/src/enable.c b/xc/extras/Mesa/src/enable.c
index 97554430e..aaf577b1a 100644
--- a/xc/extras/Mesa/src/enable.c
+++ b/xc/extras/Mesa/src/enable.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -284,12 +284,14 @@ void _mesa_set_enable( GLcontext *ctx, GLenum cap, GLboolean state )
       case GL_POINT_SMOOTH:
 	 if (ctx->Point.SmoothFlag!=state) {
             ctx->Point.SmoothFlag = state;
+	    ctx->TriangleCaps ^= DD_POINT_SMOOTH;
             ctx->NewState |= NEW_RASTER_OPS;
          }
 	 break;
       case GL_POLYGON_SMOOTH:
 	 if (ctx->Polygon.SmoothFlag!=state) {
             ctx->Polygon.SmoothFlag = state;
+	    ctx->TriangleCaps ^= DD_TRI_SMOOTH;
             ctx->NewState |= NEW_RASTER_OPS;
          }
 	 break;
@@ -349,48 +351,39 @@ void _mesa_set_enable( GLcontext *ctx, GLenum cap, GLboolean state )
       case GL_TEXTURE_1D:
          if (ctx->Visual->RGBAflag) {
 	    const GLuint curr = ctx->Texture.CurrentUnit;
-	    const GLuint flag = TEXTURE0_1D << (curr * 4);
             struct gl_texture_unit *texUnit = &ctx->Texture.Unit[curr];
 	    ctx->NewState |= NEW_TEXTURE_ENABLE;
             if (state) {
 	       texUnit->Enabled |= TEXTURE0_1D;
-	       ctx->Enabled |= flag;
 	    }
             else {
                texUnit->Enabled &= ~TEXTURE0_1D;
-               ctx->Enabled &= ~flag;
             }
          }
          break;
       case GL_TEXTURE_2D:
          if (ctx->Visual->RGBAflag) {
 	    const GLuint curr = ctx->Texture.CurrentUnit;
-	    const GLuint flag = TEXTURE0_2D << (curr * 4);
             struct gl_texture_unit *texUnit = &ctx->Texture.Unit[curr];
 	    ctx->NewState |= NEW_TEXTURE_ENABLE;
             if (state) {
 	       texUnit->Enabled |= TEXTURE0_2D;
-	       ctx->Enabled |= flag;
 	    }
             else {
                texUnit->Enabled &= ~TEXTURE0_2D;
-               ctx->Enabled &= ~flag;
             }
          }
 	 break;
       case GL_TEXTURE_3D:
          if (ctx->Visual->RGBAflag) {
 	    const GLuint curr = ctx->Texture.CurrentUnit;
-	    const GLuint flag = TEXTURE0_3D << (curr * 4);
             struct gl_texture_unit *texUnit = &ctx->Texture.Unit[curr];
 	    ctx->NewState |= NEW_TEXTURE_ENABLE;
             if (state) {
 	       texUnit->Enabled |= TEXTURE0_3D;
-	       ctx->Enabled |= flag;
 	    }
             else {
                texUnit->Enabled &= ~TEXTURE0_3D;
-               ctx->Enabled &= ~flag;
             }
          }
          break;
@@ -510,35 +503,24 @@ void _mesa_set_enable( GLcontext *ctx, GLenum cap, GLboolean state )
          if (ctx->Extensions.HaveTextureCubeMap) {
             if (ctx->Visual->RGBAflag) {
                const GLuint curr = ctx->Texture.CurrentUnit;
-               const GLuint flag = TEXTURE0_CUBE << (curr * 4);
                struct gl_texture_unit *texUnit = &ctx->Texture.Unit[curr];
                ctx->NewState |= NEW_TEXTURE_ENABLE;
                if (state) {
                   texUnit->Enabled |= TEXTURE0_CUBE;
-                  ctx->Enabled |= flag;
                }
                else {
                   texUnit->Enabled &= ~TEXTURE0_CUBE;
-                  ctx->Enabled &= ~flag;
                }
             }
          }
          else {
-            if (state)
-               gl_error(ctx, GL_INVALID_ENUM, "glEnable");
-            else
-               gl_error(ctx, GL_INVALID_ENUM, "glDisable");
+            gl_error(ctx, GL_INVALID_ENUM, state ? "glEnable" : "glDisable");
             return;
          }
 	 break;
 
       default:
-	 if (state) {
-	    gl_error( ctx, GL_INVALID_ENUM, "glEnable" );
-	 }
-	 else {
-	    gl_error( ctx, GL_INVALID_ENUM, "glDisable" );
-	 }
+         gl_error(ctx, GL_INVALID_ENUM, state ? "glEnable" : "glDisable");
          return;
    }
 
diff --git a/xc/extras/Mesa/src/enums.c b/xc/extras/Mesa/src/enums.c
index 2de4ef06e..c68cf74b8 100644
--- a/xc/extras/Mesa/src/enums.c
+++ b/xc/extras/Mesa/src/enums.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/enums.c,v 1.6 2000/09/26 15:56:31 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -854,7 +854,7 @@ static int compar_nr( const enum_elt **a, const enum_elt **b )
 
 static void sort_enums( void )
 {
-   int i;
+   GLuint i;
    index1 = (enum_elt **)MALLOC( Elements(all_enums) * sizeof(enum_elt *) );
    sorted = 1;
 
diff --git a/xc/extras/Mesa/src/eval.c b/xc/extras/Mesa/src/eval.c
index 2c29dcc27..1cc0cd083 100644
--- a/xc/extras/Mesa/src/eval.c
+++ b/xc/extras/Mesa/src/eval.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/eval.c,v 1.5 2000/09/26 15:56:31 tsi Exp $ */
+
 
 /*
  * eval.c was written by
@@ -109,8 +109,8 @@ static void
 horner_bezier_curve(const GLfloat *cp, GLfloat *out, GLfloat t,
                     GLuint dim, GLuint order)
 {
-  GLfloat s, powert;
-  GLuint i, k, bincoeff;
+  GLfloat s, powert, bincoeff;
+  GLuint i, k;
 
   if(order >= 2)
   { 
@@ -161,8 +161,8 @@ horner_bezier_surf(GLfloat *cn, GLfloat *out, GLfloat u, GLfloat v,
   {
     if(uorder >= 2)
     { 
-      GLfloat s, poweru;
-      GLuint j, k, bincoeff;
+      GLfloat s, poweru, bincoeff;
+      GLuint j, k;
 
       /* Compute the control polygon for the surface-curve in u-direction */
       for(j=0; j<vorder; j++)
diff --git a/xc/extras/Mesa/src/extensions.c b/xc/extras/Mesa/src/extensions.c
index a9522bbd8..f3e771c98 100644
--- a/xc/extras/Mesa/src/extensions.c
+++ b/xc/extras/Mesa/src/extensions.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/extensions.c,v 1.7 2000/09/26 15:56:31 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/feedback.c b/xc/extras/Mesa/src/feedback.c
index 5d14e892e..3d413a978 100644
--- a/xc/extras/Mesa/src/feedback.c
+++ b/xc/extras/Mesa/src/feedback.c
@@ -238,7 +238,7 @@ void gl_feedback_points( GLcontext *ctx, GLuint first, GLuint last )
    const struct vertex_buffer *VB = ctx->VB;
    GLuint i;
 
-   for (i=first;i<=last;i++) {
+   for (i=first;i<last;i++) {
       if (VB->ClipMask[i]==0) {
          FEEDBACK_TOKEN( ctx, (GLfloat) (GLint) GL_POINT_TOKEN );
 	 feedback_vertex( ctx, i, i );
diff --git a/xc/extras/Mesa/src/fog.c b/xc/extras/Mesa/src/fog.c
index e98e67644..a802c087e 100644
--- a/xc/extras/Mesa/src/fog.c
+++ b/xc/extras/Mesa/src/fog.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/fog.c,v 1.8 2000/09/24 13:50:10 alanh Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/fog_tmp.h b/xc/extras/Mesa/src/fog_tmp.h
index 5096b0a99..b3400d4e6 100644
--- a/xc/extras/Mesa/src/fog_tmp.h
+++ b/xc/extras/Mesa/src/fog_tmp.h
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.1
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -193,9 +193,6 @@ static void TAG(fog_ci_vertices)( struct vertex_buffer *VB,
 
 
 
-
-
-
 static void TAG(fog_rgba_vertices)( struct vertex_buffer *VB,
 				    GLuint side,
 				    GLubyte flag)
@@ -226,7 +223,6 @@ static void TAG(fog_rgba_vertices)( struct vertex_buffer *VB,
    VB->ColorPtr = VB->Color[0];
    out = (GLubyte (*)[4])VB->Color[side]->start;
 
-
    if (VB->EyePtr->size > 2) {
       switch (ctx->Fog.Mode) {
       case GL_LINEAR:
@@ -234,12 +230,17 @@ static void TAG(fog_rgba_vertices)( struct vertex_buffer *VB,
 	 for ( i = 0 ; i < n ; i++, STRIDE_F(v, stride), in += in_stride) {
 	    CULLCHECK {
 	       GLfloat f = (end - ABSF(v[2])) * d;
-	       if (f >= 1.0) continue;
-	       if (f < 0) {
+               if (f >= 1.0) {
+                  out[i][0] = in[0];
+                  out[i][1] = in[1];
+                  out[i][2] = in[2];
+               }
+	       else if (f <= 0.0) {
 		  CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(out[i][0], rFog);
 		  CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(out[i][1], gFog);
 		  CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(out[i][2], bFog);
-	       } else {
+	       }
+               else {
 		  t = f * UBYTE_COLOR_TO_FLOAT_COLOR(in[0]) + (1.0F-f)*rFog;
 		  CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(out[i][0], t);
 	       
@@ -293,27 +294,40 @@ static void TAG(fog_rgba_vertices)( struct vertex_buffer *VB,
 	 return;
       }
    }
-   else if (ctx->Fog.Mode == GL_LINEAR)
-   {
-      /* 2-vector vertices */
-      GLubyte r,g,b;
-      GLfloat f = ctx->Fog.End * (ctx->Fog.End - ctx->Fog.Start);
-      CLAMP_FLOAT_COLOR( f );
-      f = 1.0 - f;
-      rFog *= f;
-      bFog *= f;
-      gFog *= f;
-
-      CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(r, rFog);
-      CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(g, gFog);
-      CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(b, bFog);
-
-      for (i = 0 ; i < n ; i++) {
-	 /* CULLCHECK */ {
-	    out[i][0] = r;
-	    out[i][1] = g;
-	    out[i][2] = b;
-	 }
+   else {
+      /* All vertex Z coordinates are zero */
+      if (ctx->Fog.Mode == GL_LINEAR) {
+         /* 2-vector vertices */
+         GLubyte r,g,b;
+         GLfloat f = ctx->Fog.End * (ctx->Fog.End - ctx->Fog.Start);
+         CLAMP_FLOAT_COLOR( f );
+         f = 1.0 - f;
+         rFog *= f;
+         bFog *= f;
+         gFog *= f;
+
+         CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(r, rFog);
+         CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(g, gFog);
+         CLAMPED_FLOAT_COLOR_TO_UBYTE_COLOR(b, bFog);
+
+         for (i = 0 ; i < n ; i++) {
+            /* CULLCHECK */ {
+               out[i][0] = r;
+               out[i][1] = g;
+               out[i][2] = b;
+            }
+         }
+      }
+      else {
+         /* EXP or EXP2 mode */
+         /* f = exp(d*z*z) or f = exp(d*|z|) is one.  Copy in color to out */
+	 for ( i = 0 ; i < n ; i++, STRIDE_F(v,stride), in += in_stride) {
+            /* CULLCHECK */ {
+               out[i][0] = in[0];
+               out[i][1] = in[1];
+               out[i][2] = in[2];
+            }
+         }
       }
    }
 }
diff --git a/xc/extras/Mesa/src/get.c b/xc/extras/Mesa/src/get.c
index a12e6be24..59e86056a 100644
--- a/xc/extras/Mesa/src/get.c
+++ b/xc/extras/Mesa/src/get.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/get.c,v 1.7 2000/09/26 15:56:31 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/glapi.c b/xc/extras/Mesa/src/glapi.c
index c3ac458cb..8a7611e4e 100644
--- a/xc/extras/Mesa/src/glapi.c
+++ b/xc/extras/Mesa/src/glapi.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  *
@@ -1478,11 +1478,13 @@ get_static_proc_offset(const char *funcName)
 static GLvoid *
 get_static_proc_address(const char *funcName)
 {
-   GLint i = get_static_proc_offset(funcName);
-   if (i >= 0)
-      return static_functions[i].Address;
-   else
-      return NULL;
+   GLint i;
+   for (i = 0; static_functions[i].Name; i++) {
+      if (strcmp(static_functions[i].Name, funcName) == 0) {
+         return static_functions[i].Address;
+      }
+   }
+   return NULL;
 }
 
 
@@ -1562,7 +1564,7 @@ _glapi_add_entrypoint(const char *funcName, GLuint offset)
    {
       GLint index = get_static_proc_offset(funcName);
       if (index >= 0) {
-         return (GLboolean) (index == offset);  /* bad offset! */
+         return (GLboolean) (index == (GLint) offset);  /* bad offset! */
       }
    }
 
@@ -1670,7 +1672,7 @@ GLint
 _glapi_get_proc_offset(const char *funcName)
 {
    /* search extension functions first */
-   GLint i;
+   GLuint i;
    for (i = 0; i < NumExtEntryPoints; i++) {
       if (strcmp(ExtEntryTable[i].Name, funcName) == 0) {
          return ExtEntryTable[i].Offset;
@@ -1690,7 +1692,7 @@ const GLvoid *
 _glapi_get_proc_address(const char *funcName)
 {
    /* search extension functions first */
-   GLint i;
+   GLuint i;
    for (i = 0; i < NumExtEntryPoints; i++) {
       if (strcmp(ExtEntryTable[i].Name, funcName) == 0) {
          return ExtEntryTable[i].Address;
diff --git a/xc/extras/Mesa/src/glapinoop.c b/xc/extras/Mesa/src/glapinoop.c
index 43a68c773..a48509d6b 100644
--- a/xc/extras/Mesa/src/glapinoop.c
+++ b/xc/extras/Mesa/src/glapinoop.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/glapinoop.c,v 1.9 2000/09/26 15:56:31 tsi Exp $ */
+
 
 /*
  * This is part of the reusable GL dispather, see glapi.c for details.
diff --git a/xc/extras/Mesa/src/glext_proto.h b/xc/extras/Mesa/src/glext_proto.h
deleted file mode 100644
index 6f7e21535..000000000
--- a/xc/extras/Mesa/src/glext_proto.h
+++ /dev/null
@@ -1,1757 +0,0 @@
-#ifndef __glext_proto_h_
-#define __glext_proto_h_
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-/*
-** License Applicability. Except to the extent portions of this file are
-** made subject to an alternative license as permitted in the SGI Free
-** Software License B, Version 1.1 (the "License"), the contents of this
-** file are subject only to the provisions of the License. You may not use
-** this file except in compliance with the License. You may obtain a copy
-** of the License at Silicon Graphics, Inc., attn: Legal Services, 1600
-** Amphitheatre Parkway, Mountain View, CA 94043-1351, or at:
-** 
-** http://oss.sgi.com/projects/FreeB
-** 
-** Note that, as provided in the License, the Software is distributed on an
-** "AS IS" basis, with ALL EXPRESS AND IMPLIED WARRANTIES AND CONDITIONS
-** DISCLAIMED, INCLUDING, WITHOUT LIMITATION, ANY IMPLIED WARRANTIES AND
-** CONDITIONS OF MERCHANTABILITY, SATISFACTORY QUALITY, FITNESS FOR A
-** PARTICULAR PURPOSE, AND NON-INFRINGEMENT.
-** 
-** Original Code. The Original Code is: OpenGL Sample Implementation,
-** Version 1.2.1, released January 26, 2000, developed by Silicon Graphics,
-** Inc. The Original Code is Copyright (c) 1991-2000 Silicon Graphics, Inc.
-** Copyright in any portions created by third parties is as indicated
-** elsewhere herein. All Rights Reserved.
-** 
-** Additional Notice Provisions: This software was created using the
-** OpenGL(R) version 1.2.1 Sample Implementation published by SGI, but has
-** not been independently verified as being compliant with the OpenGL(R)
-** version 1.2.1 Specification.
-*/
-
-#if defined(_WIN32) && !defined(APIENTRY) && !defined(__CYGWIN32)
-#define WIN32_LEAN_AND_MEAN 1
-#include <windows.h>
-#endif
-
-#ifndef APIENTRY
-#define APIENTRY
-#endif
-
-/*************************************************************/
-
-/* OpenGL ABI for Linux version number */
-#define GLEXT_VERSION_EXT 5
-#define GL_CONSTANT_COLOR                 0x8001
-#define GL_ONE_MINUS_CONSTANT_COLOR       0x8002
-#define GL_CONSTANT_ALPHA                 0x8003
-#define GL_ONE_MINUS_CONSTANT_ALPHA       0x8004
-#define GL_BLEND_COLOR                    0x8005
-#define GL_FUNC_ADD                       0x8006
-#define GL_MIN                            0x8007
-#define GL_MAX                            0x8008
-#define GL_BLEND_EQUATION                 0x8009
-#define GL_FUNC_SUBTRACT                  0x800A
-#define GL_FUNC_REVERSE_SUBTRACT          0x800B
-#define GL_CONVOLUTION_1D                 0x8010
-#define GL_CONVOLUTION_2D                 0x8011
-#define GL_SEPARABLE_2D                   0x8012
-#define GL_CONVOLUTION_BORDER_MODE        0x8013
-#define GL_CONVOLUTION_FILTER_SCALE       0x8014
-#define GL_CONVOLUTION_FILTER_BIAS        0x8015
-#define GL_REDUCE                         0x8016
-#define GL_CONVOLUTION_FORMAT             0x8017
-#define GL_CONVOLUTION_WIDTH              0x8018
-#define GL_CONVOLUTION_HEIGHT             0x8019
-#define GL_MAX_CONVOLUTION_WIDTH          0x801A
-#define GL_MAX_CONVOLUTION_HEIGHT         0x801B
-#define GL_POST_CONVOLUTION_RED_SCALE     0x801C
-#define GL_POST_CONVOLUTION_GREEN_SCALE   0x801D
-#define GL_POST_CONVOLUTION_BLUE_SCALE    0x801E
-#define GL_POST_CONVOLUTION_ALPHA_SCALE   0x801F
-#define GL_POST_CONVOLUTION_RED_BIAS      0x8020
-#define GL_POST_CONVOLUTION_GREEN_BIAS    0x8021
-#define GL_POST_CONVOLUTION_BLUE_BIAS     0x8022
-#define GL_POST_CONVOLUTION_ALPHA_BIAS    0x8023
-#define GL_HISTOGRAM                      0x8024
-#define GL_PROXY_HISTOGRAM                0x8025
-#define GL_HISTOGRAM_WIDTH                0x8026
-#define GL_HISTOGRAM_FORMAT               0x8027
-#define GL_HISTOGRAM_RED_SIZE             0x8028
-#define GL_HISTOGRAM_GREEN_SIZE           0x8029
-#define GL_HISTOGRAM_BLUE_SIZE            0x802A
-#define GL_HISTOGRAM_ALPHA_SIZE           0x802B
-#define GL_HISTOGRAM_LUMINANCE_SIZE       0x802C
-#define GL_HISTOGRAM_SINK                 0x802D
-#define GL_MINMAX                         0x802E
-#define GL_MINMAX_FORMAT                  0x802F
-#define GL_MINMAX_SINK                    0x8030
-#define GL_TABLE_TOO_LARGE                0x8031
-#define GL_UNSIGNED_BYTE_3_3_2            0x8032
-#define GL_UNSIGNED_SHORT_4_4_4_4         0x8033
-#define GL_UNSIGNED_SHORT_5_5_5_1         0x8034
-#define GL_UNSIGNED_INT_8_8_8_8           0x8035
-#define GL_UNSIGNED_INT_10_10_10_2        0x8036
-#define GL_RESCALE_NORMAL                 0x803A
-#define GL_UNSIGNED_BYTE_2_3_3_REV        0x8362
-#define GL_UNSIGNED_SHORT_5_6_5           0x8363
-#define GL_UNSIGNED_SHORT_5_6_5_REV       0x8364
-#define GL_UNSIGNED_SHORT_4_4_4_4_REV     0x8365
-#define GL_UNSIGNED_SHORT_1_5_5_5_REV     0x8366
-#define GL_UNSIGNED_INT_8_8_8_8_REV       0x8367
-#define GL_UNSIGNED_INT_2_10_10_10_REV    0x8368
-#define GL_COLOR_MATRIX                   0x80B1
-#define GL_COLOR_MATRIX_STACK_DEPTH       0x80B2
-#define GL_MAX_COLOR_MATRIX_STACK_DEPTH   0x80B3
-#define GL_POST_COLOR_MATRIX_RED_SCALE    0x80B4
-#define GL_POST_COLOR_MATRIX_GREEN_SCALE  0x80B5
-#define GL_POST_COLOR_MATRIX_BLUE_SCALE   0x80B6
-#define GL_POST_COLOR_MATRIX_ALPHA_SCALE  0x80B7
-#define GL_POST_COLOR_MATRIX_RED_BIAS     0x80B8
-#define GL_POST_COLOR_MATRIX_GREEN_BIAS   0x80B9
-#define GL_POST_COLOR_MATRIX_BLUE_BIAS    0x80BA
-#define GL_COLOR_TABLE                    0x80D0
-#define GL_POST_CONVOLUTION_COLOR_TABLE   0x80D1
-#define GL_POST_COLOR_MATRIX_COLOR_TABLE  0x80D2
-#define GL_PROXY_COLOR_TABLE              0x80D3
-#define GL_PROXY_POST_CONVOLUTION_COLOR_TABLE 0x80D4
-#define GL_PROXY_POST_COLOR_MATRIX_COLOR_TABLE 0x80D5
-#define GL_COLOR_TABLE_SCALE              0x80D6
-#define GL_COLOR_TABLE_BIAS               0x80D7
-#define GL_COLOR_TABLE_FORMAT             0x80D8
-#define GL_COLOR_TABLE_WIDTH              0x80D9
-#define GL_COLOR_TABLE_RED_SIZE           0x80DA
-#define GL_COLOR_TABLE_GREEN_SIZE         0x80DB
-#define GL_COLOR_TABLE_BLUE_SIZE          0x80DC
-#define GL_COLOR_TABLE_ALPHA_SIZE         0x80DD
-#define GL_COLOR_TABLE_LUMINANCE_SIZE     0x80DE
-#define GL_COLOR_TABLE_INTENSITY_SIZE     0x80DF
-#define GL_CLAMP_TO_EDGE                  0x812F
-#define GL_TEXTURE_MIN_LOD                0x813A
-#define GL_TEXTURE_MAX_LOD                0x813B
-#define GL_TEXTURE_BASE_LEVEL             0x813C
-#define GL_TEXTURE_MAX_LEVEL              0x813D
-#define GL_TEXTURE0_ARB                   0x84C0
-#define GL_TEXTURE1_ARB                   0x84C1
-#define GL_TEXTURE2_ARB                   0x84C2
-#define GL_TEXTURE3_ARB                   0x84C3
-#define GL_ACTIVE_TEXTURE_ARB             0x84E0
-#define GL_CLIENT_ACTIVE_TEXTURE_ARB      0x84E1
-#define GL_MAX_ACTIVE_TEXTURES_ARB        0x84E2
-#define GL_TRANSPOSE_MODELVIEW_MATRIX_ARB 0x84E3
-#define GL_TRANSPOSE_PROJECTION_MATRIX_ARB 0x84E4
-#define GL_TRANSPOSE_TEXTURE_MATRIX_ARB   0x84E5
-#define GL_TRANSPOSE_COLOR_MATRIX_ARB     0x84E6
-#define GL_MULTISAMPLE_ARB                0x809D
-#define GL_SAMPLE_ALPHA_TO_COVERAGE_ARB   0x809E
-#define GL_SAMPLE_ALPHA_TO_ONE_ARB        0x809F
-#define GL_SAMPLE_COVERAGE_ARB            0x80A0
-#define GL_SAMPLE_BUFFERS_ARB             0x80A8
-#define GL_SAMPLES_ARB                    0x80A9
-#define GL_SAMPLE_COVERAGE_VALUE_ARB      0x80AA
-#define GL_SAMPLE_COVERAGE_INVERT_ARB     0x80AB
-#define GL_MULTISAMPLE_BIT_ARB            0x20000000
-#define GL_NORMAL_MAP_ARB                 0x8511
-#define GL_REFLECTION_MAP_ARB             0x8512
-#define GL_TEXTURE_CUBE_MAP_ARB           0x8513
-#define GL_TEXTURE_BINDING_CUBE_MAP_ARB   0x8514
-#define GL_TEXTURE_CUBE_MAP_POSITIVE_X_ARB 0x8515
-#define GL_TEXTURE_CUBE_MAP_NEGATIVE_X_ARB 0x8516
-#define GL_TEXTURE_CUBE_MAP_POSITIVE_Y_ARB 0x8517
-#define GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_ARB 0x8518
-#define GL_TEXTURE_CUBE_MAP_POSITIVE_Z_ARB 0x8519
-#define GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_ARB 0x851A
-#define GL_PROXY_TEXTURE_CUBE_MAP_ARB     0x851B
-#define GL_MAX_CUBE_MAP_TEXTURE_SIZE_ARB  0x851C
-#define GL_COMPRESSED_ALPHA_ARB           0x84E9
-#define GL_COMPRESSED_LUMINANCE_ARB       0x84EA
-#define GL_COMPRESSED_LUMINANCE_ALPHA_ARB 0x84EB
-#define GL_COMPRESSED_INTENSITY_ARB       0x84EC
-#define GL_COMPRESSED_RGB_ARB             0x84ED
-#define GL_COMPRESSED_RGBA_ARB            0x84EE
-#define GL_TEXTURE_COMPRESSION_HINT_ARB   0x84EF
-#define GL_TEXTURE_IMAGE_SIZE_ARB         0x86A0
-#define GL_TEXTURE_COMPRESSED_ARB         0x86A1
-#define GL_NUM_COMPRESSED_TEXTURE_FORMATS_ARB 0x86A2
-#define GL_COMPRESSED_TEXTURE_FORMATS_ARB 0x86A3
-#define GL_ABGR_EXT                       0x8000
-#define GL_CONSTANT_COLOR_EXT             0x8001
-#define GL_ONE_MINUS_CONSTANT_COLOR_EXT   0x8002
-#define GL_CONSTANT_ALPHA_EXT             0x8003
-#define GL_ONE_MINUS_CONSTANT_ALPHA_EXT   0x8004
-#define GL_BLEND_COLOR_EXT                0x8005
-#define GL_POLYGON_OFFSET_EXT             0x8037
-#define GL_POLYGON_OFFSET_FACTOR_EXT      0x8038
-#define GL_POLYGON_OFFSET_BIAS_EXT        0x8039
-#define GL_ALPHA4_EXT                     0x803B
-#define GL_ALPHA8_EXT                     0x803C
-#define GL_ALPHA12_EXT                    0x803D
-#define GL_ALPHA16_EXT                    0x803E
-#define GL_LUMINANCE4_EXT                 0x803F
-#define GL_LUMINANCE8_EXT                 0x8040
-#define GL_LUMINANCE12_EXT                0x8041
-#define GL_LUMINANCE16_EXT                0x8042
-#define GL_LUMINANCE4_ALPHA4_EXT          0x8043
-#define GL_LUMINANCE6_ALPHA2_EXT          0x8044
-#define GL_LUMINANCE8_ALPHA8_EXT          0x8045
-#define GL_LUMINANCE12_ALPHA4_EXT         0x8046
-#define GL_LUMINANCE12_ALPHA12_EXT        0x8047
-#define GL_LUMINANCE16_ALPHA16_EXT        0x8048
-#define GL_INTENSITY_EXT                  0x8049
-#define GL_INTENSITY4_EXT                 0x804A
-#define GL_INTENSITY8_EXT                 0x804B
-#define GL_INTENSITY12_EXT                0x804C
-#define GL_INTENSITY16_EXT                0x804D
-#define GL_RGB2_EXT                       0x804E
-#define GL_RGB4_EXT                       0x804F
-#define GL_RGB5_EXT                       0x8050
-#define GL_RGB8_EXT                       0x8051
-#define GL_RGB10_EXT                      0x8052
-#define GL_RGB12_EXT                      0x8053
-#define GL_RGB16_EXT                      0x8054
-#define GL_RGBA2_EXT                      0x8055
-#define GL_RGBA4_EXT                      0x8056
-#define GL_RGB5_A1_EXT                    0x8057
-#define GL_RGBA8_EXT                      0x8058
-#define GL_RGB10_A2_EXT                   0x8059
-#define GL_RGBA12_EXT                     0x805A
-#define GL_RGBA16_EXT                     0x805B
-#define GL_TEXTURE_RED_SIZE_EXT           0x805C
-#define GL_TEXTURE_GREEN_SIZE_EXT         0x805D
-#define GL_TEXTURE_BLUE_SIZE_EXT          0x805E
-#define GL_TEXTURE_ALPHA_SIZE_EXT         0x805F
-#define GL_TEXTURE_LUMINANCE_SIZE_EXT     0x8060
-#define GL_TEXTURE_INTENSITY_SIZE_EXT     0x8061
-#define GL_REPLACE_EXT                    0x8062
-#define GL_PROXY_TEXTURE_1D_EXT           0x8063
-#define GL_PROXY_TEXTURE_2D_EXT           0x8064
-#define GL_TEXTURE_TOO_LARGE_EXT          0x8065
-#define GL_PACK_SKIP_IMAGES               0x806B
-#define GL_PACK_SKIP_IMAGES_EXT           0x806B
-#define GL_PACK_IMAGE_HEIGHT              0x806C
-#define GL_PACK_IMAGE_HEIGHT_EXT          0x806C
-#define GL_UNPACK_SKIP_IMAGES             0x806D
-#define GL_UNPACK_SKIP_IMAGES_EXT         0x806D
-#define GL_UNPACK_IMAGE_HEIGHT            0x806E
-#define GL_UNPACK_IMAGE_HEIGHT_EXT        0x806E
-#define GL_TEXTURE_3D                     0x806F
-#define GL_TEXTURE_3D_EXT                 0x806F
-#define GL_PROXY_TEXTURE_3D               0x8070
-#define GL_PROXY_TEXTURE_3D_EXT           0x8070
-#define GL_TEXTURE_DEPTH                  0x8071
-#define GL_TEXTURE_DEPTH_EXT              0x8071
-#define GL_TEXTURE_WRAP_R                 0x8072
-#define GL_TEXTURE_WRAP_R_EXT             0x8072
-#define GL_MAX_3D_TEXTURE_SIZE            0x8073
-#define GL_MAX_3D_TEXTURE_SIZE_EXT        0x8073
-#define GL_FILTER4_SGIS                   0x8146
-#define GL_TEXTURE_FILTER4_SIZE_SGIS      0x8147
-#define GL_HISTOGRAM_EXT                  0x8024
-#define GL_PROXY_HISTOGRAM_EXT            0x8025
-#define GL_HISTOGRAM_WIDTH_EXT            0x8026
-#define GL_HISTOGRAM_FORMAT_EXT           0x8027
-#define GL_HISTOGRAM_RED_SIZE_EXT         0x8028
-#define GL_HISTOGRAM_GREEN_SIZE_EXT       0x8029
-#define GL_HISTOGRAM_BLUE_SIZE_EXT        0x802A
-#define GL_HISTOGRAM_ALPHA_SIZE_EXT       0x802B
-#define GL_HISTOGRAM_LUMINANCE_SIZE_EXT   0x802C
-#define GL_HISTOGRAM_SINK_EXT             0x802D
-#define GL_MINMAX_EXT                     0x802E
-#define GL_MINMAX_FORMAT_EXT              0x802F
-#define GL_MINMAX_SINK_EXT                0x8030
-#define GL_TABLE_TOO_LARGE_EXT            0x8031
-#define GL_CONVOLUTION_1D_EXT             0x8010
-#define GL_CONVOLUTION_2D_EXT             0x8011
-#define GL_SEPARABLE_2D_EXT               0x8012
-#define GL_CONVOLUTION_BORDER_MODE_EXT    0x8013
-#define GL_CONVOLUTION_FILTER_SCALE_EXT   0x8014
-#define GL_CONVOLUTION_FILTER_BIAS_EXT    0x8015
-#define GL_REDUCE_EXT                     0x8016
-#define GL_CONVOLUTION_FORMAT_EXT         0x8017
-#define GL_CONVOLUTION_WIDTH_EXT          0x8018
-#define GL_CONVOLUTION_HEIGHT_EXT         0x8019
-#define GL_MAX_CONVOLUTION_WIDTH_EXT      0x801A
-#define GL_MAX_CONVOLUTION_HEIGHT_EXT     0x801B
-#define GL_POST_CONVOLUTION_RED_SCALE_EXT 0x801C
-#define GL_POST_CONVOLUTION_GREEN_SCALE_EXT 0x801D
-#define GL_POST_CONVOLUTION_BLUE_SCALE_EXT 0x801E
-#define GL_POST_CONVOLUTION_ALPHA_SCALE_EXT 0x801F
-#define GL_POST_CONVOLUTION_RED_BIAS_EXT  0x8020
-#define GL_POST_CONVOLUTION_GREEN_BIAS_EXT 0x8021
-#define GL_POST_CONVOLUTION_BLUE_BIAS_EXT 0x8022
-#define GL_POST_CONVOLUTION_ALPHA_BIAS_EXT 0x8023
-#define GL_COLOR_MATRIX_SGI               0x80B1
-#define GL_COLOR_MATRIX_STACK_DEPTH_SGI   0x80B2
-#define GL_MAX_COLOR_MATRIX_STACK_DEPTH_SGI 0x80B3
-#define GL_POST_COLOR_MATRIX_RED_SCALE_SGI 0x80B4
-#define GL_POST_COLOR_MATRIX_GREEN_SCALE_SGI 0x80B5
-#define GL_POST_COLOR_MATRIX_BLUE_SCALE_SGI 0x80B6
-#define GL_POST_COLOR_MATRIX_ALPHA_SCALE_SGI 0x80B7
-#define GL_POST_COLOR_MATRIX_RED_BIAS_SGI 0x80B8
-#define GL_POST_COLOR_MATRIX_GREEN_BIAS_SGI 0x80B9
-#define GL_POST_COLOR_MATRIX_BLUE_BIAS_SGI 0x80BA
-#define GL_POST_COLOR_MATRIX_ALPHA_BIAS_SGI 0x80BB
-#define GL_COLOR_TABLE_SGI                0x80D0
-#define GL_POST_CONVOLUTION_COLOR_TABLE_SGI 0x80D1
-#define GL_POST_COLOR_MATRIX_COLOR_TABLE_SGI 0x80D2
-#define GL_PROXY_COLOR_TABLE_SGI          0x80D3
-#define GL_PROXY_POST_CONVOLUTION_COLOR_TABLE_SGI 0x80D4
-#define GL_PROXY_POST_COLOR_MATRIX_COLOR_TABLE_SGI 0x80D5
-#define GL_COLOR_TABLE_SCALE_SGI          0x80D6
-#define GL_COLOR_TABLE_BIAS_SGI           0x80D7
-#define GL_COLOR_TABLE_FORMAT_SGI         0x80D8
-#define GL_COLOR_TABLE_WIDTH_SGI          0x80D9
-#define GL_COLOR_TABLE_RED_SIZE_SGI       0x80DA
-#define GL_COLOR_TABLE_GREEN_SIZE_SGI     0x80DB
-#define GL_COLOR_TABLE_BLUE_SIZE_SGI      0x80DC
-#define GL_COLOR_TABLE_ALPHA_SIZE_SGI     0x80DD
-#define GL_COLOR_TABLE_LUMINANCE_SIZE_SGI 0x80DE
-#define GL_COLOR_TABLE_INTENSITY_SIZE_SGI 0x80DF
-#define GL_PIXEL_TEXTURE_SGIS             0x8353
-#define GL_PIXEL_FRAGMENT_RGB_SOURCE_SGIS 0x8354
-#define GL_PIXEL_FRAGMENT_ALPHA_SOURCE_SGIS 0x8355
-#define GL_PIXEL_GROUP_COLOR_SGIS         0x8356
-#define GL_PIXEL_TEX_GEN_SGIX             0x8139
-#define GL_PIXEL_TEX_GEN_MODE_SGIX        0x832B
-#define GL_PACK_SKIP_VOLUMES_SGIS         0x8130
-#define GL_PACK_IMAGE_DEPTH_SGIS          0x8131
-#define GL_UNPACK_SKIP_VOLUMES_SGIS       0x8132
-#define GL_UNPACK_IMAGE_DEPTH_SGIS        0x8133
-#define GL_TEXTURE_4D_SGIS                0x8134
-#define GL_PROXY_TEXTURE_4D_SGIS          0x8135
-#define GL_TEXTURE_4DSIZE_SGIS            0x8136
-#define GL_TEXTURE_WRAP_Q_SGIS            0x8137
-#define GL_MAX_4D_TEXTURE_SIZE_SGIS       0x8138
-#define GL_TEXTURE_4D_BINDING_SGIS        0x814F
-#define GL_TEXTURE_COLOR_TABLE_SGI        0x80BC
-#define GL_PROXY_TEXTURE_COLOR_TABLE_SGI  0x80BD
-#define GL_CMYK_EXT                       0x800C
-#define GL_CMYKA_EXT                      0x800D
-#define GL_PACK_CMYK_HINT_EXT             0x800E
-#define GL_UNPACK_CMYK_HINT_EXT           0x800F
-#define GL_TEXTURE_PRIORITY_EXT           0x8066
-#define GL_TEXTURE_RESIDENT_EXT           0x8067
-#define GL_TEXTURE_1D_BINDING_EXT         0x8068
-#define GL_TEXTURE_2D_BINDING_EXT         0x8069
-#define GL_TEXTURE_3D_BINDING_EXT         0x806A
-#define GL_DETAIL_TEXTURE_2D_SGIS         0x8095
-#define GL_DETAIL_TEXTURE_2D_BINDING_SGIS 0x8096
-#define GL_LINEAR_DETAIL_SGIS             0x8097
-#define GL_LINEAR_DETAIL_ALPHA_SGIS       0x8098
-#define GL_LINEAR_DETAIL_COLOR_SGIS       0x8099
-#define GL_DETAIL_TEXTURE_LEVEL_SGIS      0x809A
-#define GL_DETAIL_TEXTURE_MODE_SGIS       0x809B
-#define GL_DETAIL_TEXTURE_FUNC_POINTS_SGIS 0x809C
-#define GL_LINEAR_SHARPEN_SGIS            0x80AD
-#define GL_LINEAR_SHARPEN_ALPHA_SGIS      0x80AE
-#define GL_LINEAR_SHARPEN_COLOR_SGIS      0x80AF
-#define GL_SHARPEN_TEXTURE_FUNC_POINTS_SGIS 0x80B0
-#define GL_UNSIGNED_BYTE_3_3_2_EXT        0x8032
-#define GL_UNSIGNED_SHORT_4_4_4_4_EXT     0x8033
-#define GL_UNSIGNED_SHORT_5_5_5_1_EXT     0x8034
-#define GL_UNSIGNED_INT_8_8_8_8_EXT       0x8035
-#define GL_UNSIGNED_INT_10_10_10_2_EXT    0x8036
-#define GL_TEXTURE_MIN_LOD_SGIS           0x813A
-#define GL_TEXTURE_MAX_LOD_SGIS           0x813B
-#define GL_TEXTURE_BASE_LEVEL_SGIS        0x813C
-#define GL_TEXTURE_MAX_LEVEL_SGIS         0x813D
-#define GL_MULTISAMPLE_SGIS               0x809D
-#define GL_SAMPLE_ALPHA_TO_MASK_SGIS      0x809E
-#define GL_SAMPLE_ALPHA_TO_ONE_SGIS       0x809F
-#define GL_SAMPLE_MASK_SGIS               0x80A0
-#define GL_1PASS_SGIS                     0x80A1
-#define GL_2PASS_0_SGIS                   0x80A2
-#define GL_2PASS_1_SGIS                   0x80A3
-#define GL_4PASS_0_SGIS                   0x80A4
-#define GL_4PASS_1_SGIS                   0x80A5
-#define GL_4PASS_2_SGIS                   0x80A6
-#define GL_4PASS_3_SGIS                   0x80A7
-#define GL_SAMPLE_BUFFERS_SGIS            0x80A8
-#define GL_SAMPLES_SGIS                   0x80A9
-#define GL_SAMPLE_MASK_VALUE_SGIS         0x80AA
-#define GL_SAMPLE_MASK_INVERT_SGIS        0x80AB
-#define GL_SAMPLE_PATTERN_SGIS            0x80AC
-#define GL_RESCALE_NORMAL_EXT             0x803A
-#define GL_VERTEX_ARRAY_EXT               0x8074
-#define GL_NORMAL_ARRAY_EXT               0x8075
-#define GL_COLOR_ARRAY_EXT                0x8076
-#define GL_INDEX_ARRAY_EXT                0x8077
-#define GL_TEXTURE_COORD_ARRAY_EXT        0x8078
-#define GL_EDGE_FLAG_ARRAY_EXT            0x8079
-#define GL_VERTEX_ARRAY_SIZE_EXT          0x807A
-#define GL_VERTEX_ARRAY_TYPE_EXT          0x807B
-#define GL_VERTEX_ARRAY_STRIDE_EXT        0x807C
-#define GL_VERTEX_ARRAY_COUNT_EXT         0x807D
-#define GL_NORMAL_ARRAY_TYPE_EXT          0x807E
-#define GL_NORMAL_ARRAY_STRIDE_EXT        0x807F
-#define GL_NORMAL_ARRAY_COUNT_EXT         0x8080
-#define GL_COLOR_ARRAY_SIZE_EXT           0x8081
-#define GL_COLOR_ARRAY_TYPE_EXT           0x8082
-#define GL_COLOR_ARRAY_STRIDE_EXT         0x8083
-#define GL_COLOR_ARRAY_COUNT_EXT          0x8084
-#define GL_INDEX_ARRAY_TYPE_EXT           0x8085
-#define GL_INDEX_ARRAY_STRIDE_EXT         0x8086
-#define GL_INDEX_ARRAY_COUNT_EXT          0x8087
-#define GL_TEXTURE_COORD_ARRAY_SIZE_EXT   0x8088
-#define GL_TEXTURE_COORD_ARRAY_TYPE_EXT   0x8089
-#define GL_TEXTURE_COORD_ARRAY_STRIDE_EXT 0x808A
-#define GL_TEXTURE_COORD_ARRAY_COUNT_EXT  0x808B
-#define GL_EDGE_FLAG_ARRAY_STRIDE_EXT     0x808C
-#define GL_EDGE_FLAG_ARRAY_COUNT_EXT      0x808D
-#define GL_VERTEX_ARRAY_POINTER_EXT       0x808E
-#define GL_NORMAL_ARRAY_POINTER_EXT       0x808F
-#define GL_COLOR_ARRAY_POINTER_EXT        0x8090
-#define GL_INDEX_ARRAY_POINTER_EXT        0x8091
-#define GL_TEXTURE_COORD_ARRAY_POINTER_EXT 0x8092
-#define GL_EDGE_FLAG_ARRAY_POINTER_EXT    0x8093
-#define GL_GENERATE_MIPMAP_SGIS           0x8191
-#define GL_GENERATE_MIPMAP_HINT_SGIS      0x8192
-#define GL_LINEAR_CLIPMAP_LINEAR_SGIX     0x8170
-#define GL_TEXTURE_CLIPMAP_CENTER_SGIX    0x8171
-#define GL_TEXTURE_CLIPMAP_FRAME_SGIX     0x8172
-#define GL_TEXTURE_CLIPMAP_OFFSET_SGIX    0x8173
-#define GL_TEXTURE_CLIPMAP_VIRTUAL_DEPTH_SGIX 0x8174
-#define GL_TEXTURE_CLIPMAP_LOD_OFFSET_SGIX 0x8175
-#define GL_TEXTURE_CLIPMAP_DEPTH_SGIX     0x8176
-#define GL_MAX_CLIPMAP_DEPTH_SGIX         0x8177
-#define GL_MAX_CLIPMAP_VIRTUAL_DEPTH_SGIX 0x8178
-#define GL_NEAREST_CLIPMAP_NEAREST_SGIX   0x844D
-#define GL_NEAREST_CLIPMAP_LINEAR_SGIX    0x844E
-#define GL_LINEAR_CLIPMAP_NEAREST_SGIX    0x844F
-#define GL_TEXTURE_COMPARE_SGIX           0x819A
-#define GL_TEXTURE_COMPARE_OPERATOR_SGIX  0x819B
-#define GL_TEXTURE_LEQUAL_R_SGIX          0x819C
-#define GL_TEXTURE_GEQUAL_R_SGIX          0x819D
-#define GL_CLAMP_TO_EDGE_SGIS             0x812F
-#define GL_CLAMP_TO_BORDER_SGIS           0x812D
-#define GL_FUNC_ADD_EXT                   0x8006
-#define GL_MIN_EXT                        0x8007
-#define GL_MAX_EXT                        0x8008
-#define GL_BLEND_EQUATION_EXT             0x8009
-#define GL_FUNC_SUBTRACT_EXT              0x800A
-#define GL_FUNC_REVERSE_SUBTRACT_EXT      0x800B
-#define GL_INTERLACE_SGIX                 0x8094
-#define GL_PIXEL_TILE_BEST_ALIGNMENT_SGIX 0x813E
-#define GL_PIXEL_TILE_CACHE_INCREMENT_SGIX 0x813F
-#define GL_PIXEL_TILE_WIDTH_SGIX          0x8140
-#define GL_PIXEL_TILE_HEIGHT_SGIX         0x8141
-#define GL_PIXEL_TILE_GRID_WIDTH_SGIX     0x8142
-#define GL_PIXEL_TILE_GRID_HEIGHT_SGIX    0x8143
-#define GL_PIXEL_TILE_GRID_DEPTH_SGIX     0x8144
-#define GL_PIXEL_TILE_CACHE_SIZE_SGIX     0x8145
-#define GL_DUAL_ALPHA4_SGIS               0x8110
-#define GL_DUAL_ALPHA8_SGIS               0x8111
-#define GL_DUAL_ALPHA12_SGIS              0x8112
-#define GL_DUAL_ALPHA16_SGIS              0x8113
-#define GL_DUAL_LUMINANCE4_SGIS           0x8114
-#define GL_DUAL_LUMINANCE8_SGIS           0x8115
-#define GL_DUAL_LUMINANCE12_SGIS          0x8116
-#define GL_DUAL_LUMINANCE16_SGIS          0x8117
-#define GL_DUAL_INTENSITY4_SGIS           0x8118
-#define GL_DUAL_INTENSITY8_SGIS           0x8119
-#define GL_DUAL_INTENSITY12_SGIS          0x811A
-#define GL_DUAL_INTENSITY16_SGIS          0x811B
-#define GL_DUAL_LUMINANCE_ALPHA4_SGIS     0x811C
-#define GL_DUAL_LUMINANCE_ALPHA8_SGIS     0x811D
-#define GL_QUAD_ALPHA4_SGIS               0x811E
-#define GL_QUAD_ALPHA8_SGIS               0x811F
-#define GL_QUAD_LUMINANCE4_SGIS           0x8120
-#define GL_QUAD_LUMINANCE8_SGIS           0x8121
-#define GL_QUAD_INTENSITY4_SGIS           0x8122
-#define GL_QUAD_INTENSITY8_SGIS           0x8123
-#define GL_DUAL_TEXTURE_SELECT_SGIS       0x8124
-#define GL_QUAD_TEXTURE_SELECT_SGIS       0x8125
-#define GL_SPRITE_SGIX                    0x8148
-#define GL_SPRITE_MODE_SGIX               0x8149
-#define GL_SPRITE_AXIS_SGIX               0x814A
-#define GL_SPRITE_TRANSLATION_SGIX        0x814B
-#define GL_SPRITE_AXIAL_SGIX              0x814C
-#define GL_SPRITE_OBJECT_ALIGNED_SGIX     0x814D
-#define GL_SPRITE_EYE_ALIGNED_SGIX        0x814E
-#define GL_TEXTURE_MULTI_BUFFER_HINT_SGIX 0x812E
-#define GL_POINT_SIZE_MIN_EXT             0x8126
-#define GL_POINT_SIZE_MIN_SGIS            0x8126
-#define GL_POINT_SIZE_MAX_EXT             0x8127
-#define GL_POINT_SIZE_MAX_SGIS            0x8127
-#define GL_POINT_FADE_THRESHOLD_SIZE_EXT  0x8128
-#define GL_POINT_FADE_THRESHOLD_SIZE_SGIS 0x8128
-#define GL_DISTANCE_ATTENUATION_EXT       0x8129
-#define GL_DISTANCE_ATTENUATION_SGIS      0x8129
-#define GL_INSTRUMENT_BUFFER_POINTER_SGIX 0x8180
-#define GL_INSTRUMENT_MEASUREMENTS_SGIX   0x8181
-#define GL_POST_TEXTURE_FILTER_BIAS_SGIX  0x8179
-#define GL_POST_TEXTURE_FILTER_SCALE_SGIX 0x817A
-#define GL_POST_TEXTURE_FILTER_BIAS_RANGE_SGIX 0x817B
-#define GL_POST_TEXTURE_FILTER_SCALE_RANGE_SGIX 0x817C
-#define GL_FRAMEZOOM_SGIX                 0x818B
-#define GL_FRAMEZOOM_FACTOR_SGIX          0x818C
-#define GL_MAX_FRAMEZOOM_FACTOR_SGIX      0x818D
-#define GL_REFERENCE_PLANE_SGIX           0x817D
-#define GL_REFERENCE_PLANE_EQUATION_SGIX  0x817E
-#define GL_DEPTH_COMPONENT16_SGIX         0x81A5
-#define GL_DEPTH_COMPONENT24_SGIX         0x81A6
-#define GL_DEPTH_COMPONENT32_SGIX         0x81A7
-#define GL_FOG_FUNC_SGIS                  0x812A
-#define GL_FOG_FUNC_POINTS_SGIS           0x812B
-#define GL_MAX_FOG_FUNC_POINTS_SGIS       0x812C
-#define GL_FOG_OFFSET_SGIX                0x8198
-#define GL_FOG_OFFSET_VALUE_SGIX          0x8199
-#define GL_IMAGE_SCALE_X_HP               0x8155
-#define GL_IMAGE_SCALE_Y_HP               0x8156
-#define GL_IMAGE_TRANSLATE_X_HP           0x8157
-#define GL_IMAGE_TRANSLATE_Y_HP           0x8158
-#define GL_IMAGE_ROTATE_ANGLE_HP          0x8159
-#define GL_IMAGE_ROTATE_ORIGIN_X_HP       0x815A
-#define GL_IMAGE_ROTATE_ORIGIN_Y_HP       0x815B
-#define GL_IMAGE_MAG_FILTER_HP            0x815C
-#define GL_IMAGE_MIN_FILTER_HP            0x815D
-#define GL_IMAGE_CUBIC_WEIGHT_HP          0x815E
-#define GL_CUBIC_HP                       0x815F
-#define GL_AVERAGE_HP                     0x8160
-#define GL_IMAGE_TRANSFORM_2D_HP          0x8161
-#define GL_POST_IMAGE_TRANSFORM_COLOR_TABLE_HP 0x8162
-#define GL_PROXY_POST_IMAGE_TRANSFORM_COLOR_TABLE_HP 0x8163
-#define GL_IGNORE_BORDER_HP               0x8150
-#define GL_CONSTANT_BORDER_HP             0x8151
-#define GL_REPLICATE_BORDER_HP            0x8153
-#define GL_CONVOLUTION_BORDER_COLOR_HP    0x8154
-#define GL_TEXTURE_ENV_BIAS_SGIX          0x80BE
-#define GL_VERTEX_DATA_HINT_PGI           0x1A22A
-#define GL_VERTEX_CONSISTENT_HINT_PGI     0x1A22B
-#define GL_MATERIAL_SIDE_HINT_PGI         0x1A22C
-#define GL_MAX_VERTEX_HINT_PGI            0x1A22D
-#define GL_COLOR3_BIT_PGI                 0x00010000
-#define GL_COLOR4_BIT_PGI                 0x00020000
-#define GL_EDGEFLAG_BIT_PGI               0x00040000
-#define GL_INDEX_BIT_PGI                  0x00080000
-#define GL_MAT_AMBIENT_BIT_PGI            0x00100000
-#define GL_MAT_AMBIENT_AND_DIFFUSE_BIT_PGI 0x00200000
-#define GL_MAT_DIFFUSE_BIT_PGI            0x00400000
-#define GL_MAT_EMISSION_BIT_PGI           0x00800000
-#define GL_MAT_COLOR_INDEXES_BIT_PGI      0x01000000
-#define GL_MAT_SHININESS_BIT_PGI          0x02000000
-#define GL_MAT_SPECULAR_BIT_PGI           0x04000000
-#define GL_NORMAL_BIT_PGI                 0x08000000
-#define GL_TEXCOORD1_BIT_PGI              0x10000000
-#define GL_TEXCOORD2_BIT_PGI              0x20000000
-#define GL_TEXCOORD3_BIT_PGI              0x40000000
-#define GL_TEXCOORD4_BIT_PGI              0x80000000
-#define GL_VERTEX23_BIT_PGI               0x00000004
-#define GL_VERTEX4_BIT_PGI                0x00000008
-#define GL_PREFER_DOUBLEBUFFER_HINT_PGI   0x1A1F8
-#define GL_CONSERVE_MEMORY_HINT_PGI       0x1A1FD
-#define GL_RECLAIM_MEMORY_HINT_PGI        0x1A1FE
-#define GL_NATIVE_GRAPHICS_HANDLE_PGI     0x1A202
-#define GL_NATIVE_GRAPHICS_BEGIN_HINT_PGI 0x1A203
-#define GL_NATIVE_GRAPHICS_END_HINT_PGI   0x1A204
-#define GL_ALWAYS_FAST_HINT_PGI           0x1A20C
-#define GL_ALWAYS_SOFT_HINT_PGI           0x1A20D
-#define GL_ALLOW_DRAW_OBJ_HINT_PGI        0x1A20E
-#define GL_ALLOW_DRAW_WIN_HINT_PGI        0x1A20F
-#define GL_ALLOW_DRAW_FRG_HINT_PGI        0x1A210
-#define GL_ALLOW_DRAW_MEM_HINT_PGI        0x1A211
-#define GL_STRICT_DEPTHFUNC_HINT_PGI      0x1A216
-#define GL_STRICT_LIGHTING_HINT_PGI       0x1A217
-#define GL_STRICT_SCISSOR_HINT_PGI        0x1A218
-#define GL_FULL_STIPPLE_HINT_PGI          0x1A219
-#define GL_CLIP_NEAR_HINT_PGI             0x1A220
-#define GL_CLIP_FAR_HINT_PGI              0x1A221
-#define GL_WIDE_LINE_HINT_PGI             0x1A222
-#define GL_BACK_NORMALS_HINT_PGI          0x1A223
-#define GL_COLOR_INDEX1_EXT               0x80E2
-#define GL_COLOR_INDEX2_EXT               0x80E3
-#define GL_COLOR_INDEX4_EXT               0x80E4
-#define GL_COLOR_INDEX8_EXT               0x80E5
-#define GL_COLOR_INDEX12_EXT              0x80E6
-#define GL_COLOR_INDEX16_EXT              0x80E7
-#define GL_TEXTURE_INDEX_SIZE_EXT         0x80ED
-#define GL_CLIP_VOLUME_CLIPPING_HINT_EXT  0x80F0
-#define GL_LIST_PRIORITY_SGIX             0x8182
-#define GL_IR_INSTRUMENT1_SGIX            0x817F
-#define GL_CALLIGRAPHIC_FRAGMENT_SGIX     0x8183
-#define GL_TEXTURE_LOD_BIAS_S_SGIX        0x818E
-#define GL_TEXTURE_LOD_BIAS_T_SGIX        0x818F
-#define GL_TEXTURE_LOD_BIAS_R_SGIX        0x8190
-#define GL_SHADOW_AMBIENT_SGIX            0x80BF
-#define GL_INDEX_MATERIAL_EXT             0x81B8
-#define GL_INDEX_MATERIAL_PARAMETER_EXT   0x81B9
-#define GL_INDEX_MATERIAL_FACE_EXT        0x81BA
-#define GL_INDEX_TEST_EXT                 0x81B5
-#define GL_INDEX_TEST_FUNC_EXT            0x81B6
-#define GL_INDEX_TEST_REF_EXT             0x81B7
-#define GL_IUI_V2F_EXT                    0x81AD
-#define GL_IUI_V3F_EXT                    0x81AE
-#define GL_IUI_N3F_V2F_EXT                0x81AF
-#define GL_IUI_N3F_V3F_EXT                0x81B0
-#define GL_T2F_IUI_V2F_EXT                0x81B1
-#define GL_T2F_IUI_V3F_EXT                0x81B2
-#define GL_T2F_IUI_N3F_V2F_EXT            0x81B3
-#define GL_T2F_IUI_N3F_V3F_EXT            0x81B4
-#define GL_ARRAY_ELEMENT_LOCK_FIRST_EXT   0x81A8
-#define GL_ARRAY_ELEMENT_LOCK_COUNT_EXT   0x81A9
-#define GL_CULL_VERTEX_EXT                0x81AA
-#define GL_CULL_VERTEX_EYE_POSITION_EXT   0x81AB
-#define GL_CULL_VERTEX_OBJECT_POSITION_EXT 0x81AC
-#define GL_YCRCB_422_SGIX                 0x81BB
-#define GL_YCRCB_444_SGIX                 0x81BC
-#define GL_FRAGMENT_LIGHTING_SGIX         0x8400
-#define GL_FRAGMENT_COLOR_MATERIAL_SGIX   0x8401
-#define GL_FRAGMENT_COLOR_MATERIAL_FACE_SGIX 0x8402
-#define GL_FRAGMENT_COLOR_MATERIAL_PARAMETER_SGIX 0x8403
-#define GL_MAX_FRAGMENT_LIGHTS_SGIX       0x8404
-#define GL_MAX_ACTIVE_LIGHTS_SGIX         0x8405
-#define GL_CURRENT_RASTER_NORMAL_SGIX     0x8406
-#define GL_LIGHT_ENV_MODE_SGIX            0x8407
-#define GL_FRAGMENT_LIGHT_MODEL_LOCAL_VIEWER_SGIX 0x8408
-#define GL_FRAGMENT_LIGHT_MODEL_TWO_SIDE_SGIX 0x8409
-#define GL_FRAGMENT_LIGHT_MODEL_AMBIENT_SGIX 0x840A
-#define GL_FRAGMENT_LIGHT_MODEL_NORMAL_INTERPOLATION_SGIX 0x840B
-#define GL_FRAGMENT_LIGHT0_SGIX           0x840C
-#define GL_FRAGMENT_LIGHT1_SGIX           0x840D
-#define GL_FRAGMENT_LIGHT2_SGIX           0x840E
-#define GL_FRAGMENT_LIGHT3_SGIX           0x840F
-#define GL_FRAGMENT_LIGHT4_SGIX           0x8410
-#define GL_FRAGMENT_LIGHT5_SGIX           0x8411
-#define GL_FRAGMENT_LIGHT6_SGIX           0x8412
-#define GL_FRAGMENT_LIGHT7_SGIX           0x8413
-#define GL_RASTER_POSITION_UNCLIPPED_IBM  0x19262
-#define GL_TEXTURE_LIGHTING_MODE_HP       0x8167
-#define GL_TEXTURE_POST_SPECULAR_HP       0x8168
-#define GL_TEXTURE_PRE_SPECULAR_HP        0x8169
-#define GL_MAX_ELEMENTS_VERTICES_EXT      0x80E8
-#define GL_MAX_ELEMENTS_INDICES_EXT       0x80E9
-#define GL_PHONG_WIN                      0x80EA
-#define GL_PHONG_HINT_WIN                 0x80EB
-#define GL_FOG_SPECULAR_TEXTURE_WIN       0x80EC
-#define GL_FRAGMENT_MATERIAL_EXT          0x8349
-#define GL_FRAGMENT_NORMAL_EXT            0x834A
-#define GL_FRAGMENT_COLOR_EXT             0x834C
-#define GL_ATTENUATION_EXT                0x834D
-#define GL_SHADOW_ATTENUATION_EXT         0x834E
-#define GL_TEXTURE_APPLICATION_MODE_EXT   0x834F
-#define GL_TEXTURE_LIGHT_EXT              0x8350
-#define GL_TEXTURE_MATERIAL_FACE_EXT      0x8351
-#define GL_TEXTURE_MATERIAL_PARAMETER_EXT 0x8352
-/* reuse GL_FRAGMENT_DEPTH_EXT */
-#define GL_ALPHA_MIN_SGIX                 0x8320
-#define GL_ALPHA_MAX_SGIX                 0x8321
-#define GL_BGR_EXT                        0x80E0
-#define GL_BGRA_EXT                       0x80E1
-#define GL_PARALLEL_ARRAYS_INTEL          0x83F4
-#define GL_VERTEX_ARRAY_PARALLEL_POINTERS_INTEL 0x83F5
-#define GL_NORMAL_ARRAY_PARALLEL_POINTERS_INTEL 0x83F6
-#define GL_COLOR_ARRAY_PARALLEL_POINTERS_INTEL 0x83F7
-#define GL_TEXTURE_COORD_ARRAY_PARALLEL_POINTERS_INTEL 0x83F8
-#define GL_OCCLUSION_TEST_HP              0x8165
-#define GL_OCCLUSION_TEST_RESULT_HP       0x8166
-#define GL_PIXEL_TRANSFORM_2D_EXT         0x8330
-#define GL_PIXEL_MAG_FILTER_EXT           0x8331
-#define GL_PIXEL_MIN_FILTER_EXT           0x8332
-#define GL_PIXEL_CUBIC_WEIGHT_EXT         0x8333
-#define GL_CUBIC_EXT                      0x8334
-#define GL_AVERAGE_EXT                    0x8335
-#define GL_PIXEL_TRANSFORM_2D_STACK_DEPTH_EXT 0x8336
-#define GL_MAX_PIXEL_TRANSFORM_2D_STACK_DEPTH_EXT 0x8337
-#define GL_PIXEL_TRANSFORM_2D_MATRIX_EXT  0x8338
-#define GL_SHARED_TEXTURE_PALETTE_EXT     0x81FB
-#define GL_LIGHT_MODEL_COLOR_CONTROL_EXT  0x81F8
-#define GL_SINGLE_COLOR_EXT               0x81F9
-#define GL_SEPARATE_SPECULAR_COLOR_EXT    0x81FA
-#define GL_COLOR_SUM_EXT                  0x8458
-#define GL_CURRENT_SECONDARY_COLOR_EXT    0x8459
-#define GL_SECONDARY_COLOR_ARRAY_SIZE_EXT 0x845A
-#define GL_SECONDARY_COLOR_ARRAY_TYPE_EXT 0x845B
-#define GL_SECONDARY_COLOR_ARRAY_STRIDE_EXT 0x845C
-#define GL_SECONDARY_COLOR_ARRAY_POINTER_EXT 0x845D
-#define GL_SECONDARY_COLOR_ARRAY_EXT      0x845E
-#define GL_PERTURB_EXT                    0x85AE
-#define GL_TEXTURE_NORMAL_EXT             0x85AF
-#define GL_FOG_COORDINATE_SOURCE_EXT      0x8450
-#define GL_FOG_COORDINATE_EXT             0x8451
-#define GL_FRAGMENT_DEPTH_EXT             0x8452
-#define GL_CURRENT_FOG_COORDINATE_EXT     0x8453
-#define GL_FOG_COORDINATE_ARRAY_TYPE_EXT  0x8454
-#define GL_FOG_COORDINATE_ARRAY_STRIDE_EXT 0x8455
-#define GL_FOG_COORDINATE_ARRAY_POINTER_EXT 0x8456
-#define GL_FOG_COORDINATE_ARRAY_EXT       0x8457
-#define GL_SCREEN_COORDINATES_REND        0x8490
-#define GL_INVERTED_SCREEN_W_REND         0x8491
-#define GL_TANGENT_ARRAY_EXT              0x8439
-#define GL_BINORMAL_ARRAY_EXT             0x843A
-#define GL_CURRENT_TANGENT_EXT            0x843B
-#define GL_CURRENT_BINORMAL_EXT           0x843C
-#define GL_TANGENT_ARRAY_TYPE_EXT         0x843E
-#define GL_TANGENT_ARRAY_STRIDE_EXT       0x843F
-#define GL_BINORMAL_ARRAY_TYPE_EXT        0x8440
-#define GL_BINORMAL_ARRAY_STRIDE_EXT      0x8441
-#define GL_TANGENT_ARRAY_POINTER_EXT      0x8442
-#define GL_BINORMAL_ARRAY_POINTER_EXT     0x8443
-#define GL_MAP1_TANGENT_EXT               0x8444
-#define GL_MAP2_TANGENT_EXT               0x8445
-#define GL_MAP1_BINORMAL_EXT              0x8446
-#define GL_MAP2_BINORMAL_EXT              0x8447
-#define GL_COMBINE_EXT                    0x8570
-#define GL_COMBINE_RGB_EXT                0x8571
-#define GL_COMBINE_ALPHA_EXT              0x8572
-#define GL_RGB_SCALE_EXT                  0x8573
-#define GL_ADD_SIGNED_EXT                 0x8574
-#define GL_INTERPOLATE_EXT                0x8575
-#define GL_CONSTANT_EXT                   0x8576
-#define GL_PRIMARY_COLOR_EXT              0x8577
-#define GL_PREVIOUS_EXT                   0x8578
-#define GL_SOURCE0_RGB_EXT                0x8580
-#define GL_SOURCE1_RGB_EXT                0x8581
-#define GL_SOURCE2_RGB_EXT                0x8582
-#define GL_SOURCE3_RGB_EXT                0x8583
-#define GL_SOURCE4_RGB_EXT                0x8584
-#define GL_SOURCE5_RGB_EXT                0x8585
-#define GL_SOURCE6_RGB_EXT                0x8586
-#define GL_SOURCE7_RGB_EXT                0x8587
-#define GL_SOURCE0_ALPHA_EXT              0x8588
-#define GL_SOURCE1_ALPHA_EXT              0x8589
-#define GL_SOURCE2_ALPHA_EXT              0x858A
-#define GL_SOURCE3_ALPHA_EXT              0x858B
-#define GL_SOURCE4_ALPHA_EXT              0x858C
-#define GL_SOURCE5_ALPHA_EXT              0x858D
-#define GL_SOURCE6_ALPHA_EXT              0x858E
-#define GL_SOURCE7_ALPHA_EXT              0x858F
-#define GL_OPERAND0_RGB_EXT               0x8590
-#define GL_OPERAND1_RGB_EXT               0x8591
-#define GL_OPERAND2_RGB_EXT               0x8592
-#define GL_OPERAND3_RGB_EXT               0x8593
-#define GL_OPERAND4_RGB_EXT               0x8594
-#define GL_OPERAND5_RGB_EXT               0x8595
-#define GL_OPERAND6_RGB_EXT               0x8596
-#define GL_OPERAND7_RGB_EXT               0x8597
-#define GL_OPERAND0_ALPHA_EXT             0x8598
-#define GL_OPERAND1_ALPHA_EXT             0x8599
-#define GL_OPERAND2_ALPHA_EXT             0x859A
-#define GL_OPERAND3_ALPHA_EXT             0x859B
-#define GL_OPERAND4_ALPHA_EXT             0x859C
-#define GL_OPERAND5_ALPHA_EXT             0x859D
-#define GL_OPERAND6_ALPHA_EXT             0x859E
-#define GL_OPERAND7_ALPHA_EXT             0x859F
-#define GL_LIGHT_MODEL_SPECULAR_VECTOR_APPLE 0x85B0
-#define GL_TRANSFORM_HINT_APPLE           0x85B1
-#define GL_FOG_SCALE_SGIX                 0x81FC
-#define GL_FOG_SCALE_VALUE_SGIX           0x81FD
-#define GL_UNPACK_CONSTANT_DATA_SUNX      0x81D5
-#define GL_TEXTURE_CONSTANT_DATA_SUNX     0x81D6
-#define GL_GLOBAL_ALPHA_SUN               0x81D9
-#define GL_GLOBAL_ALPHA_FACTOR_SUN        0x81DA
-#define GL_RESTART_SUN                    0x01
-#define GL_REPLACE_MIDDLE_SUN             0x02
-#define GL_REPLACE_OLDEST_SUN             0x03
-#define GL_TRIANGLE_LIST_SUN              0x81D7
-#define GL_REPLACEMENT_CODE_SUN           0x81D8
-#define GL_REPLACEMENT_CODE_ARRAY_SUN     0x85C0
-#define GL_REPLACEMENT_CODE_ARRAY_TYPE_SUN 0x85C1
-#define GL_REPLACEMENT_CODE_ARRAY_STRIDE_SUN 0x85C2
-#define GL_REPLACEMENT_CODE_ARRAY_POINTER_SUN 0x85C3
-#define GL_R1UI_V3F_SUN                   0x85C4
-#define GL_R1UI_C4UB_V3F_SUN              0x85C5
-#define GL_R1UI_C3F_V3F_SUN               0x85C6
-#define GL_R1UI_N3F_V3F_SUN               0x85C7
-#define GL_R1UI_C4F_N3F_V3F_SUN           0x85C8
-#define GL_R1UI_T2F_V3F_SUN               0x85C9
-#define GL_R1UI_T2F_N3F_V3F_SUN           0x85CA
-#define GL_R1UI_T2F_C4F_N3F_V3F_SUN       0x85CB
-#define GL_BLEND_DST_RGB_EXT              0x80C8
-#define GL_BLEND_SRC_RGB_EXT              0x80C9
-#define GL_BLEND_DST_ALPHA_EXT            0x80CA
-#define GL_BLEND_SRC_ALPHA_EXT            0x80CB
-#define GL_RED_MIN_CLAMP_INGR             0x8560
-#define GL_GREEN_MIN_CLAMP_INGR           0x8561
-#define GL_BLUE_MIN_CLAMP_INGR            0x8562
-#define GL_ALPHA_MIN_CLAMP_INGR           0x8563
-#define GL_RED_MAX_CLAMP_INGR             0x8564
-#define GL_GREEN_MAX_CLAMP_INGR           0x8565
-#define GL_BLUE_MAX_CLAMP_INGR            0x8566
-#define GL_ALPHA_MAX_CLAMP_INGR           0x8567
-#define GL_INTERLACE_READ_INGR            0x8568
-#define GL_INCR_WRAP_EXT                  0x8507
-#define GL_DECR_WRAP_EXT                  0x8508
-#define GL_422_EXT                        0x80CC
-#define GL_422_REV_EXT                    0x80CD
-#define GL_422_AVERAGE_EXT                0x80CE
-#define GL_422_REV_AVERAGE_EXT            0x80CF
-#define GL_NORMAL_MAP_NV                  0x8511
-#define GL_REFLECTION_MAP_NV              0x8512
-#define GL_NORMAL_MAP_EXT                 0x8511
-#define GL_REFLECTION_MAP_EXT             0x8512
-#define GL_TEXTURE_CUBE_MAP_EXT           0x8513
-#define GL_TEXTURE_BINDING_CUBE_MAP_EXT   0x8514
-#define GL_TEXTURE_CUBE_MAP_POSITIVE_X_EXT 0x8515
-#define GL_TEXTURE_CUBE_MAP_NEGATIVE_X_EXT 0x8516
-#define GL_TEXTURE_CUBE_MAP_POSITIVE_Y_EXT 0x8517
-#define GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_EXT 0x8518
-#define GL_TEXTURE_CUBE_MAP_POSITIVE_Z_EXT 0x8519
-#define GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_EXT 0x851A
-#define GL_PROXY_TEXTURE_CUBE_MAP_EXT     0x851B
-#define GL_MAX_CUBE_MAP_TEXTURE_SIZE_EXT  0x851C
-#define GL_WRAP_BORDER_SUN                0x81D4
-#define GL_MAX_TEXTURE_LOD_BIAS_EXT       0x84FD
-#define GL_TEXTURE_FILTER_CONTROL_EXT     0x8500
-#define GL_TEXTURE_LOD_BIAS_EXT           0x8501
-#define GL_TEXTURE_MAX_ANISOTROPY_EXT     0x84FE
-#define GL_MAX_TEXTURE_MAX_ANISOTROPY_EXT 0x84FF
-#define GL_MODELVIEW0_STACK_DEPTH_EXT     GL_MODELVIEW_STACK_DEPTH
-#define GL_MODELVIEW1_STACK_DEPTH_EXT     0x8502
-#define GL_MODELVIEW0_MATRIX_EXT          GL_MODELVIEW_MATRIX
-#define GL_MODELVIEW_MATRIX1_EXT          0x8506
-#define GL_VERTEX_WEIGHTING_EXT           0x8509
-#define GL_MODELVIEW0_EXT                 GL_MODELVIEW
-#define GL_MODELVIEW1_EXT                 0x850A
-#define GL_CURRENT_VERTEX_WEIGHT_EXT      0x850B
-#define GL_VERTEX_WEIGHT_ARRAY_EXT        0x850C
-#define GL_VERTEX_WEIGHT_ARRAY_SIZE_EXT   0x850D
-#define GL_VERTEX_WEIGHT_ARRAY_TYPE_EXT   0x850E
-#define GL_VERTEX_WEIGHT_ARRAY_STRIDE_EXT 0x850F
-#define GL_VERTEX_WEIGHT_ARRAY_POINTER_EXT 0x8510
-#define GL_MAX_SHININESS_NV               0x8504
-#define GL_MAX_SPOT_EXPONENT_NV           0x8505
-#define GL_VERTEX_ARRAY_RANGE_NV          0x851D
-#define GL_VERTEX_ARRAY_RANGE_LENGTH_NV   0x851E
-#define GL_VERTEX_ARRAY_RANGE_VALID_NV    0x851F
-#define GL_MAX_VERTEX_ARRAY_RANGE_ELEMENT_NV 0x8520
-#define GL_VERTEX_ARRAY_RANGE_POINTER_NV  0x8521
-#define GL_REGISTER_COMBINERS_NV          0x8522
-#define GL_VARIABLE_A_NV                  0x8523
-#define GL_VARIABLE_B_NV                  0x8524
-#define GL_VARIABLE_C_NV                  0x8525
-#define GL_VARIABLE_D_NV                  0x8526
-#define GL_VARIABLE_E_NV                  0x8527
-#define GL_VARIABLE_F_NV                  0x8528
-#define GL_VARIABLE_G_NV                  0x8529
-#define GL_CONSTANT_COLOR0_NV             0x852A
-#define GL_CONSTANT_COLOR1_NV             0x852B
-#define GL_PRIMARY_COLOR_NV               0x852C
-#define GL_SECONDARY_COLOR_NV             0x852D
-#define GL_SPARE0_NV                      0x852E
-#define GL_SPARE1_NV                      0x852F
-#define GL_DISCARD_NV                     0x8530
-#define GL_E_TIMES_F_NV                   0x8531
-#define GL_SPARE0_PLUS_SECONDARY_COLOR_NV 0x8532
-#define GL_UNSIGNED_IDENTITY_NV           0x8536
-#define GL_UNSIGNED_INVERT_NV             0x8537
-#define GL_EXPAND_NORMAL_NV               0x8538
-#define GL_EXPAND_NEGATE_NV               0x8539
-#define GL_HALF_BIAS_NORMAL_NV            0x853A
-#define GL_HALF_BIAS_NEGATE_NV            0x853B
-#define GL_SIGNED_IDENTITY_NV             0x853C
-#define GL_SIGNED_NEGATE_NV               0x853D
-#define GL_SCALE_BY_TWO_NV                0x853E
-#define GL_SCALE_BY_FOUR_NV               0x853F
-#define GL_SCALE_BY_ONE_HALF_NV           0x8540
-#define GL_BIAS_BY_NEGATIVE_ONE_HALF_NV   0x8541
-#define GL_COMBINER_INPUT_NV              0x8542
-#define GL_COMBINER_MAPPING_NV            0x8543
-#define GL_COMBINER_COMPONENT_USAGE_NV    0x8544
-#define GL_COMBINER_AB_DOT_PRODUCT_NV     0x8545
-#define GL_COMBINER_CD_DOT_PRODUCT_NV     0x8546
-#define GL_COMBINER_MUX_SUM_NV            0x8547
-#define GL_COMBINER_SCALE_NV              0x8548
-#define GL_COMBINER_BIAS_NV               0x8549
-#define GL_COMBINER_AB_OUTPUT_NV          0x854A
-#define GL_COMBINER_CD_OUTPUT_NV          0x854B
-#define GL_COMBINER_SUM_OUTPUT_NV         0x854C
-#define GL_MAX_GENERAL_COMBINERS_NV       0x854D
-#define GL_NUM_GENERAL_COMBINERS_NV       0x854E
-#define GL_COLOR_SUM_CLAMP_NV             0x854F
-#define GL_COMBINER0_NV                   0x8550
-#define GL_COMBINER1_NV                   0x8551
-#define GL_COMBINER2_NV                   0x8552
-#define GL_COMBINER3_NV                   0x8553
-#define GL_COMBINER4_NV                   0x8554
-#define GL_COMBINER5_NV                   0x8555
-#define GL_COMBINER6_NV                   0x8556
-#define GL_COMBINER7_NV                   0x8557
-/* reuse GL_TEXTURE0_ARB */
-/* reuse GL_TEXTURE1_ARB */
-/* reuse GL_ZERO */
-/* reuse GL_NONE */
-/* reuse GL_FOG */
-#define GL_FOG_DISTANCE_MODE_NV           0x855A
-#define GL_EYE_RADIAL_NV                  0x855B
-#define GL_EYE_PLANE_ABSOLUTE_NV          0x855C
-/* reuse GL_EYE_PLANE */
-#define GL_EMBOSS_LIGHT_NV                0x855D
-#define GL_EMBOSS_CONSTANT_NV             0x855E
-#define GL_EMBOSS_MAP_NV                  0x855F
-#define GL_COMBINE4_NV                    0x8503
-#define GL_SOURCE3_RGB_NV                 0x8583
-#define GL_SOURCE3_ALPHA_NV               0x858B
-#define GL_OPERAND3_RGB_NV                0x8593
-#define GL_OPERAND3_ALPHA_NV              0x859B
-#define GL_COMPRESSED_RGB_S3TC_DXT1_EXT   0x83F0
-#define GL_COMPRESSED_RGBA_S3TC_DXT1_EXT  0x83F1
-#define GL_COMPRESSED_RGBA_S3TC_DXT3_EXT  0x83F2
-#define GL_COMPRESSED_RGBA_S3TC_DXT5_EXT  0x83F3
-#define GL_CULL_VERTEX_IBM                103050
-#define GL_VERTEX_ARRAY_LIST_IBM          103070
-#define GL_NORMAL_ARRAY_LIST_IBM          103071
-#define GL_COLOR_ARRAY_LIST_IBM           103072
-#define GL_INDEX_ARRAY_LIST_IBM           103073
-#define GL_TEXTURE_COORD_ARRAY_LIST_IBM   103074
-#define GL_EDGE_FLAG_ARRAY_LIST_IBM       103075
-#define GL_FOG_COORDINATE_ARRAY_LIST_IBM  103076
-#define GL_SECONDARY_COLOR_ARRAY_LIST_IBM 103077
-#define GL_VERTEX_ARRAY_LIST_STRIDE_IBM   103080
-#define GL_NORMAL_ARRAY_LIST_STRIDE_IBM   103081
-#define GL_COLOR_ARRAY_LIST_STRIDE_IBM    103082
-#define GL_INDEX_ARRAY_LIST_STRIDE_IBM    103083
-#define GL_TEXTURE_COORD_ARRAY_LIST_STRIDE_IBM 103084
-#define GL_EDGE_FLAG_ARRAY_LIST_STRIDE_IBM 103085
-#define GL_FOG_COORDINATE_ARRAY_LIST_STRIDE_IBM 103086
-#define GL_SECONDARY_COLOR_ARRAY_LIST_STRIDE_IBM 103087
-#define GL_PACK_SUBSAMPLE_RATE_SGIX       0x85A0
-#define GL_UNPACK_SUBSAMPLE_RATE_SGIX     0x85A1
-#define GL_PIXEL_SUBSAMPLE_4444_SGIX      0x85A2
-#define GL_PIXEL_SUBSAMPLE_2424_SGIX      0x85A3
-#define GL_PIXEL_SUBSAMPLE_4242_SGIX      0x85A4
-#define GL_YCRCB_SGIX                     0x8318
-#define GL_YCRCBA_SGIX                    0x8319
-#define GL_DEPTH_PASS_INSTRUMENT_SGIX     0x8310
-#define GL_DEPTH_PASS_INSTRUMENT_COUNTERS_SGIX 0x8311
-#define GL_DEPTH_PASS_INSTRUMENT_MAX_SGIX 0x8312
-#define GL_COMPRESSED_RGB_FXT1_3DFX       0x86B0
-#define GL_COMPRESSED_RGBA_FXT1_3DFX      0x86B1
-#define GL_MULTISAMPLE_3DFX               0x86B2
-#define GL_SAMPLE_BUFFERS_3DFX            0x86B3
-#define GL_SAMPLES_3DFX                   0x86B4
-#define GL_MULTISAMPLE_BIT_3DFX           0x20000000
-#define GL_MULTISAMPLE_EXT                0x809D
-#define GL_SAMPLE_ALPHA_TO_MASK_EXT       0x809E
-#define GL_SAMPLE_ALPHA_TO_ONE_EXT        0x809F
-#define GL_SAMPLE_MASK_EXT                0x80A0
-#define GL_1PASS_EXT                      0x80A1
-#define GL_2PASS_0_EXT                    0x80A2
-#define GL_2PASS_1_EXT                    0x80A3
-#define GL_4PASS_0_EXT                    0x80A4
-#define GL_4PASS_1_EXT                    0x80A5
-#define GL_4PASS_2_EXT                    0x80A6
-#define GL_4PASS_3_EXT                    0x80A7
-#define GL_SAMPLE_BUFFERS_EXT             0x80A8
-#define GL_SAMPLES_EXT                    0x80A9
-#define GL_SAMPLE_MASK_VALUE_EXT          0x80AA
-#define GL_SAMPLE_MASK_INVERT_EXT         0x80AB
-#define GL_SAMPLE_PATTERN_EXT             0x80AC
-#define GL_VERTEX_PRECLIP_SGIX            0x83EE
-#define GL_VERTEX_PRECLIP_HINT_SGIX       0x83EF
-#define GL_CONVOLUTION_HINT_SGIX          0x8316
-#define GL_PACK_RESAMPLE_SGIX             0x842C
-#define GL_UNPACK_RESAMPLE_SGIX           0x842D
-#define GL_RESAMPLE_REPLICATE_SGIX        0x842E
-#define GL_RESAMPLE_ZERO_FILL_SGIX        0x842F
-#define GL_RESAMPLE_DECIMATE_SGIX         0x8430
-#define GL_EYE_DISTANCE_TO_POINT_SGIS     0x81F0
-#define GL_OBJECT_DISTANCE_TO_POINT_SGIS  0x81F1
-#define GL_EYE_DISTANCE_TO_LINE_SGIS      0x81F2
-#define GL_OBJECT_DISTANCE_TO_LINE_SGIS   0x81F3
-#define GL_EYE_POINT_SGIS                 0x81F4
-#define GL_OBJECT_POINT_SGIS              0x81F5
-#define GL_EYE_LINE_SGIS                  0x81F6
-#define GL_OBJECT_LINE_SGIS               0x81F7
-
-
-/*************************************************************/
-
-extern void APIENTRY glBlendColor (GLclampf, GLclampf, GLclampf, GLclampf);
-typedef void (APIENTRY * PFNGLBLENDCOLORPROC) (GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha);
-extern void APIENTRY glBlendEquation (GLenum);
-typedef void (APIENTRY * PFNGLBLENDEQUATIONPROC) (GLenum mode);
-extern void APIENTRY glDrawRangeElements (GLenum, GLuint, GLuint, GLsizei, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLDRAWRANGEELEMENTSPROC) (GLenum mode, GLuint start, GLuint end, GLsizei count, GLenum type, const GLvoid *indices);
-extern void APIENTRY glColorTable (GLenum, GLenum, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOLORTABLEPROC) (GLenum target, GLenum internalformat, GLsizei width, GLenum format, GLenum type, const GLvoid *table);
-extern void APIENTRY glColorTableParameterfv (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOLORTABLEPARAMETERFVPROC) (GLenum target, GLenum pname, const GLfloat *params);
-extern void APIENTRY glColorTableParameteriv (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLCOLORTABLEPARAMETERIVPROC) (GLenum target, GLenum pname, const GLint *params);
-extern void APIENTRY glCopyColorTable (GLenum, GLenum, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCOLORTABLEPROC) (GLenum target, GLenum internalformat, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glGetColorTable (GLenum, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPROC) (GLenum target, GLenum format, GLenum type, GLvoid *table);
-extern void APIENTRY glGetColorTableParameterfv (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPARAMETERFVPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetColorTableParameteriv (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPARAMETERIVPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glColorSubTable (GLenum, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOLORSUBTABLEPROC) (GLenum target, GLsizei start, GLsizei count, GLenum format, GLenum type, const GLvoid *data);
-extern void APIENTRY glCopyColorSubTable (GLenum, GLsizei, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCOLORSUBTABLEPROC) (GLenum target, GLsizei start, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glConvolutionFilter1D (GLenum, GLenum, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONFILTER1DPROC) (GLenum target, GLenum internalformat, GLsizei width, GLenum format, GLenum type, const GLvoid *image);
-extern void APIENTRY glConvolutionFilter2D (GLenum, GLenum, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONFILTER2DPROC) (GLenum target, GLenum internalformat, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *image);
-extern void APIENTRY glConvolutionParameterf (GLenum, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERFPROC) (GLenum target, GLenum pname, GLfloat params);
-extern void APIENTRY glConvolutionParameterfv (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERFVPROC) (GLenum target, GLenum pname, const GLfloat *params);
-extern void APIENTRY glConvolutionParameteri (GLenum, GLenum, GLint);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERIPROC) (GLenum target, GLenum pname, GLint params);
-extern void APIENTRY glConvolutionParameteriv (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERIVPROC) (GLenum target, GLenum pname, const GLint *params);
-extern void APIENTRY glCopyConvolutionFilter1D (GLenum, GLenum, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCONVOLUTIONFILTER1DPROC) (GLenum target, GLenum internalformat, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glCopyConvolutionFilter2D (GLenum, GLenum, GLint, GLint, GLsizei, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCONVOLUTIONFILTER2DPROC) (GLenum target, GLenum internalformat, GLint x, GLint y, GLsizei width, GLsizei height);
-extern void APIENTRY glGetConvolutionFilter (GLenum, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETCONVOLUTIONFILTERPROC) (GLenum target, GLenum format, GLenum type, GLvoid *image);
-extern void APIENTRY glGetConvolutionParameterfv (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCONVOLUTIONPARAMETERFVPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetConvolutionParameteriv (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCONVOLUTIONPARAMETERIVPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glGetSeparableFilter (GLenum, GLenum, GLenum, GLvoid *, GLvoid *, GLvoid *);
-typedef void (APIENTRY * PFNGLGETSEPARABLEFILTERPROC) (GLenum target, GLenum format, GLenum type, GLvoid *row, GLvoid *column, GLvoid *span);
-extern void APIENTRY glSeparableFilter2D (GLenum, GLenum, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *, const GLvoid *);
-typedef void (APIENTRY * PFNGLSEPARABLEFILTER2DPROC) (GLenum target, GLenum internalformat, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *row, const GLvoid *column);
-extern void APIENTRY glGetHistogram (GLenum, GLboolean, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETHISTOGRAMPROC) (GLenum target, GLboolean reset, GLenum format, GLenum type, GLvoid *values);
-extern void APIENTRY glGetHistogramParameterfv (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETHISTOGRAMPARAMETERFVPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetHistogramParameteriv (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETHISTOGRAMPARAMETERIVPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glGetMinmax (GLenum, GLboolean, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETMINMAXPROC) (GLenum target, GLboolean reset, GLenum format, GLenum type, GLvoid *values);
-extern void APIENTRY glGetMinmaxParameterfv (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETMINMAXPARAMETERFVPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetMinmaxParameteriv (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETMINMAXPARAMETERIVPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glHistogram (GLenum, GLsizei, GLenum, GLboolean);
-typedef void (APIENTRY * PFNGLHISTOGRAMPROC) (GLenum target, GLsizei width, GLenum internalformat, GLboolean sink);
-extern void APIENTRY glMinmax (GLenum, GLenum, GLboolean);
-typedef void (APIENTRY * PFNGLMINMAXPROC) (GLenum target, GLenum internalformat, GLboolean sink);
-extern void APIENTRY glResetHistogram (GLenum);
-typedef void (APIENTRY * PFNGLRESETHISTOGRAMPROC) (GLenum target);
-extern void APIENTRY glResetMinmax (GLenum);
-typedef void (APIENTRY * PFNGLRESETMINMAXPROC) (GLenum target);
-extern void APIENTRY glTexImage3D (GLenum, GLint, GLint, GLsizei, GLsizei, GLsizei, GLint, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXIMAGE3DPROC) (GLenum target, GLint level, GLint internalformat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glTexSubImage3D (GLenum, GLint, GLint, GLint, GLint, GLsizei, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXSUBIMAGE3DPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glCopyTexSubImage3D (GLenum, GLint, GLint, GLint, GLint, GLint, GLint, GLsizei, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYTEXSUBIMAGE3DPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLint x, GLint y, GLsizei width, GLsizei height);
-extern void APIENTRY glActiveTextureARB (GLenum);
-typedef void (APIENTRY * PFNGLACTIVETEXTUREARBPROC) (GLenum texture);
-extern void APIENTRY glClientActiveTextureARB (GLenum);
-typedef void (APIENTRY * PFNGLCLIENTACTIVETEXTUREARBPROC) (GLenum texture);
-extern void APIENTRY glMultiTexCoord1dARB (GLenum, GLdouble);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1DARBPROC) (GLenum target, GLdouble s);
-extern void APIENTRY glMultiTexCoord1dvARB (GLenum, const GLdouble *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1DVARBPROC) (GLenum target, const GLdouble *v);
-extern void APIENTRY glMultiTexCoord1fARB (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1FARBPROC) (GLenum target, GLfloat s);
-extern void APIENTRY glMultiTexCoord1fvARB (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1FVARBPROC) (GLenum target, const GLfloat *v);
-extern void APIENTRY glMultiTexCoord1iARB (GLenum, GLint);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1IARBPROC) (GLenum target, GLint s);
-extern void APIENTRY glMultiTexCoord1ivARB (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1IVARBPROC) (GLenum target, const GLint *v);
-extern void APIENTRY glMultiTexCoord1sARB (GLenum, GLshort);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1SARBPROC) (GLenum target, GLshort s);
-extern void APIENTRY glMultiTexCoord1svARB (GLenum, const GLshort *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD1SVARBPROC) (GLenum target, const GLshort *v);
-extern void APIENTRY glMultiTexCoord2dARB (GLenum, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2DARBPROC) (GLenum target, GLdouble s, GLdouble t);
-extern void APIENTRY glMultiTexCoord2dvARB (GLenum, const GLdouble *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2DVARBPROC) (GLenum target, const GLdouble *v);
-extern void APIENTRY glMultiTexCoord2fARB (GLenum, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2FARBPROC) (GLenum target, GLfloat s, GLfloat t);
-extern void APIENTRY glMultiTexCoord2fvARB (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2FVARBPROC) (GLenum target, const GLfloat *v);
-extern void APIENTRY glMultiTexCoord2iARB (GLenum, GLint, GLint);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2IARBPROC) (GLenum target, GLint s, GLint t);
-extern void APIENTRY glMultiTexCoord2ivARB (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2IVARBPROC) (GLenum target, const GLint *v);
-extern void APIENTRY glMultiTexCoord2sARB (GLenum, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2SARBPROC) (GLenum target, GLshort s, GLshort t);
-extern void APIENTRY glMultiTexCoord2svARB (GLenum, const GLshort *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD2SVARBPROC) (GLenum target, const GLshort *v);
-extern void APIENTRY glMultiTexCoord3dARB (GLenum, GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3DARBPROC) (GLenum target, GLdouble s, GLdouble t, GLdouble r);
-extern void APIENTRY glMultiTexCoord3dvARB (GLenum, const GLdouble *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3DVARBPROC) (GLenum target, const GLdouble *v);
-extern void APIENTRY glMultiTexCoord3fARB (GLenum, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3FARBPROC) (GLenum target, GLfloat s, GLfloat t, GLfloat r);
-extern void APIENTRY glMultiTexCoord3fvARB (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3FVARBPROC) (GLenum target, const GLfloat *v);
-extern void APIENTRY glMultiTexCoord3iARB (GLenum, GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3IARBPROC) (GLenum target, GLint s, GLint t, GLint r);
-extern void APIENTRY glMultiTexCoord3ivARB (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3IVARBPROC) (GLenum target, const GLint *v);
-extern void APIENTRY glMultiTexCoord3sARB (GLenum, GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3SARBPROC) (GLenum target, GLshort s, GLshort t, GLshort r);
-extern void APIENTRY glMultiTexCoord3svARB (GLenum, const GLshort *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD3SVARBPROC) (GLenum target, const GLshort *v);
-extern void APIENTRY glMultiTexCoord4dARB (GLenum, GLdouble, GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4DARBPROC) (GLenum target, GLdouble s, GLdouble t, GLdouble r, GLdouble q);
-extern void APIENTRY glMultiTexCoord4dvARB (GLenum, const GLdouble *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4DVARBPROC) (GLenum target, const GLdouble *v);
-extern void APIENTRY glMultiTexCoord4fARB (GLenum, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4FARBPROC) (GLenum target, GLfloat s, GLfloat t, GLfloat r, GLfloat q);
-extern void APIENTRY glMultiTexCoord4fvARB (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4FVARBPROC) (GLenum target, const GLfloat *v);
-extern void APIENTRY glMultiTexCoord4iARB (GLenum, GLint, GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4IARBPROC) (GLenum target, GLint s, GLint t, GLint r, GLint q);
-extern void APIENTRY glMultiTexCoord4ivARB (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4IVARBPROC) (GLenum target, const GLint *v);
-extern void APIENTRY glMultiTexCoord4sARB (GLenum, GLshort, GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4SARBPROC) (GLenum target, GLshort s, GLshort t, GLshort r, GLshort q);
-extern void APIENTRY glMultiTexCoord4svARB (GLenum, const GLshort *);
-typedef void (APIENTRY * PFNGLMULTITEXCOORD4SVARBPROC) (GLenum target, const GLshort *v);
-extern void APIENTRY glLoadTransposeMatrixfARB (const GLfloat *);
-typedef void (APIENTRY * PFNGLLOADTRANSPOSEMATRIXFARBPROC) (const GLfloat *m);
-extern void APIENTRY glLoadTransposeMatrixdARB (const GLdouble *);
-typedef void (APIENTRY * PFNGLLOADTRANSPOSEMATRIXDARBPROC) (const GLdouble *m);
-extern void APIENTRY glMultTransposeMatrixfARB (const GLfloat *);
-typedef void (APIENTRY * PFNGLMULTTRANSPOSEMATRIXFARBPROC) (const GLfloat *m);
-extern void APIENTRY glMultTransposeMatrixdARB (const GLdouble *);
-typedef void (APIENTRY * PFNGLMULTTRANSPOSEMATRIXDARBPROC) (const GLdouble *m);
-extern void APIENTRY glSampleCoverageARB (GLclampf, GLboolean);
-typedef void (APIENTRY * PFNGLSAMPLECOVERAGEARBPROC) (GLclampf value, GLboolean invert);
-extern void APIENTRY glSamplePassARB (GLenum);
-typedef void (APIENTRY * PFNGLSAMPLEPASSARBPROC) (GLenum pass);
-extern void APIENTRY glCompressedTexImage3DARB (GLenum, GLint, GLenum, GLsizei, GLsizei, GLsizei, GLint, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOMPRESSEDTEXIMAGE3DARBPROC) (GLenum target, GLint level, GLenum internalformat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLsizei imageSize, const GLvoid *data);
-extern void APIENTRY glCompressedTexImage2DARB (GLenum, GLint, GLenum, GLsizei, GLsizei, GLint, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOMPRESSEDTEXIMAGE2DARBPROC) (GLenum target, GLint level, GLenum internalformat, GLsizei width, GLsizei height, GLint border, GLsizei imageSize, const GLvoid *data);
-extern void APIENTRY glCompressedTexImage1DARB (GLenum, GLint, GLenum, GLsizei, GLint, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOMPRESSEDTEXIMAGE1DARBPROC) (GLenum target, GLint level, GLenum internalformat, GLsizei width, GLint border, GLsizei imageSize, const GLvoid *data);
-extern void APIENTRY glCompressedTexSubImage3DARB (GLenum, GLint, GLint, GLint, GLint, GLsizei, GLsizei, GLsizei, GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOMPRESSEDTEXSUBIMAGE3DARBPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLsizei imageSize, const GLvoid *data);
-extern void APIENTRY glCompressedTexSubImage2DARB (GLenum, GLint, GLint, GLint, GLsizei, GLsizei, GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOMPRESSEDTEXSUBIMAGE2DARBPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLsizei imageSize, const GLvoid *data);
-extern void APIENTRY glCompressedTexSubImage1DARB (GLenum, GLint, GLint, GLsizei, GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOMPRESSEDTEXSUBIMAGE1DARBPROC) (GLenum target, GLint level, GLint xoffset, GLsizei width, GLenum format, GLsizei imageSize, const GLvoid *data);
-extern void APIENTRY glGetCompressedTexImageARB (GLenum, GLint, void *);
-typedef void (APIENTRY * PFNGLGETCOMPRESSEDTEXIMAGEARBPROC) (GLenum target, GLint level, void *img);
-extern void APIENTRY glBlendColorEXT (GLclampf, GLclampf, GLclampf, GLclampf);
-typedef void (APIENTRY * PFNGLBLENDCOLOREXTPROC) (GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha);
-extern void APIENTRY glPolygonOffsetEXT (GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLPOLYGONOFFSETEXTPROC) (GLfloat factor, GLfloat bias);
-extern void APIENTRY glTexImage3DEXT (GLenum, GLint, GLenum, GLsizei, GLsizei, GLsizei, GLint, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXIMAGE3DEXTPROC) (GLenum target, GLint level, GLenum internalformat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glTexSubImage3DEXT (GLenum, GLint, GLint, GLint, GLint, GLsizei, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXSUBIMAGE3DEXTPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glGetTexFilterFuncSGIS (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETTEXFILTERFUNCSGISPROC) (GLenum target, GLenum filter, GLfloat *weights);
-extern void APIENTRY glTexFilterFuncSGIS (GLenum, GLenum, GLsizei, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXFILTERFUNCSGISPROC) (GLenum target, GLenum filter, GLsizei n, const GLfloat *weights);
-extern void APIENTRY glTexSubImage1DEXT (GLenum, GLint, GLint, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXSUBIMAGE1DEXTPROC) (GLenum target, GLint level, GLint xoffset, GLsizei width, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glTexSubImage2DEXT (GLenum, GLint, GLint, GLint, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXSUBIMAGE2DEXTPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glCopyTexImage1DEXT (GLenum, GLint, GLenum, GLint, GLint, GLsizei, GLint);
-typedef void (APIENTRY * PFNGLCOPYTEXIMAGE1DEXTPROC) (GLenum target, GLint level, GLenum internalformat, GLint x, GLint y, GLsizei width, GLint border);
-extern void APIENTRY glCopyTexImage2DEXT (GLenum, GLint, GLenum, GLint, GLint, GLsizei, GLsizei, GLint);
-typedef void (APIENTRY * PFNGLCOPYTEXIMAGE2DEXTPROC) (GLenum target, GLint level, GLenum internalformat, GLint x, GLint y, GLsizei width, GLsizei height, GLint border);
-extern void APIENTRY glCopyTexSubImage1DEXT (GLenum, GLint, GLint, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYTEXSUBIMAGE1DEXTPROC) (GLenum target, GLint level, GLint xoffset, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glCopyTexSubImage2DEXT (GLenum, GLint, GLint, GLint, GLint, GLint, GLsizei, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYTEXSUBIMAGE2DEXTPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint x, GLint y, GLsizei width, GLsizei height);
-extern void APIENTRY glCopyTexSubImage3DEXT (GLenum, GLint, GLint, GLint, GLint, GLint, GLint, GLsizei, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYTEXSUBIMAGE3DEXTPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLint x, GLint y, GLsizei width, GLsizei height);
-extern void APIENTRY glGetHistogramEXT (GLenum, GLboolean, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETHISTOGRAMEXTPROC) (GLenum target, GLboolean reset, GLenum format, GLenum type, GLvoid *values);
-extern void APIENTRY glGetHistogramParameterfvEXT (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETHISTOGRAMPARAMETERFVEXTPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetHistogramParameterivEXT (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETHISTOGRAMPARAMETERIVEXTPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glGetMinmaxEXT (GLenum, GLboolean, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETMINMAXEXTPROC) (GLenum target, GLboolean reset, GLenum format, GLenum type, GLvoid *values);
-extern void APIENTRY glGetMinmaxParameterfvEXT (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETMINMAXPARAMETERFVEXTPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetMinmaxParameterivEXT (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETMINMAXPARAMETERIVEXTPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glHistogramEXT (GLenum, GLsizei, GLenum, GLboolean);
-typedef void (APIENTRY * PFNGLHISTOGRAMEXTPROC) (GLenum target, GLsizei width, GLenum internalformat, GLboolean sink);
-extern void APIENTRY glMinmaxEXT (GLenum, GLenum, GLboolean);
-typedef void (APIENTRY * PFNGLMINMAXEXTPROC) (GLenum target, GLenum internalformat, GLboolean sink);
-extern void APIENTRY glResetHistogramEXT (GLenum);
-typedef void (APIENTRY * PFNGLRESETHISTOGRAMEXTPROC) (GLenum target);
-extern void APIENTRY glResetMinmaxEXT (GLenum);
-typedef void (APIENTRY * PFNGLRESETMINMAXEXTPROC) (GLenum target);
-extern void APIENTRY glConvolutionFilter1DEXT (GLenum, GLenum, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONFILTER1DEXTPROC) (GLenum target, GLenum internalformat, GLsizei width, GLenum format, GLenum type, const GLvoid *image);
-extern void APIENTRY glConvolutionFilter2DEXT (GLenum, GLenum, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONFILTER2DEXTPROC) (GLenum target, GLenum internalformat, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *image);
-extern void APIENTRY glConvolutionParameterfEXT (GLenum, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERFEXTPROC) (GLenum target, GLenum pname, GLfloat params);
-extern void APIENTRY glConvolutionParameterfvEXT (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERFVEXTPROC) (GLenum target, GLenum pname, const GLfloat *params);
-extern void APIENTRY glConvolutionParameteriEXT (GLenum, GLenum, GLint);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERIEXTPROC) (GLenum target, GLenum pname, GLint params);
-extern void APIENTRY glConvolutionParameterivEXT (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLCONVOLUTIONPARAMETERIVEXTPROC) (GLenum target, GLenum pname, const GLint *params);
-extern void APIENTRY glCopyConvolutionFilter1DEXT (GLenum, GLenum, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCONVOLUTIONFILTER1DEXTPROC) (GLenum target, GLenum internalformat, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glCopyConvolutionFilter2DEXT (GLenum, GLenum, GLint, GLint, GLsizei, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCONVOLUTIONFILTER2DEXTPROC) (GLenum target, GLenum internalformat, GLint x, GLint y, GLsizei width, GLsizei height);
-extern void APIENTRY glGetConvolutionFilterEXT (GLenum, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETCONVOLUTIONFILTEREXTPROC) (GLenum target, GLenum format, GLenum type, GLvoid *image);
-extern void APIENTRY glGetConvolutionParameterfvEXT (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCONVOLUTIONPARAMETERFVEXTPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetConvolutionParameterivEXT (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCONVOLUTIONPARAMETERIVEXTPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glGetSeparableFilterEXT (GLenum, GLenum, GLenum, GLvoid *, GLvoid *, GLvoid *);
-typedef void (APIENTRY * PFNGLGETSEPARABLEFILTEREXTPROC) (GLenum target, GLenum format, GLenum type, GLvoid *row, GLvoid *column, GLvoid *span);
-extern void APIENTRY glSeparableFilter2DEXT (GLenum, GLenum, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *, const GLvoid *);
-typedef void (APIENTRY * PFNGLSEPARABLEFILTER2DEXTPROC) (GLenum target, GLenum internalformat, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *row, const GLvoid *column);
-extern void APIENTRY glColorTableSGI (GLenum, GLenum, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOLORTABLESGIPROC) (GLenum target, GLenum internalformat, GLsizei width, GLenum format, GLenum type, const GLvoid *table);
-extern void APIENTRY glColorTableParameterfvSGI (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOLORTABLEPARAMETERFVSGIPROC) (GLenum target, GLenum pname, const GLfloat *params);
-extern void APIENTRY glColorTableParameterivSGI (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLCOLORTABLEPARAMETERIVSGIPROC) (GLenum target, GLenum pname, const GLint *params);
-extern void APIENTRY glCopyColorTableSGI (GLenum, GLenum, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCOLORTABLESGIPROC) (GLenum target, GLenum internalformat, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glGetColorTableSGI (GLenum, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLESGIPROC) (GLenum target, GLenum format, GLenum type, GLvoid *table);
-extern void APIENTRY glGetColorTableParameterfvSGI (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPARAMETERFVSGIPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetColorTableParameterivSGI (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPARAMETERIVSGIPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glPixelTexGenSGIX (GLenum);
-typedef void (APIENTRY * PFNGLPIXELTEXGENSGIXPROC) (GLenum mode);
-extern void APIENTRY glPixelTexGenParameteriSGIS (GLenum, GLint);
-typedef void (APIENTRY * PFNGLPIXELTEXGENPARAMETERISGISPROC) (GLenum pname, GLint param);
-extern void APIENTRY glPixelTexGenParameterivSGIS (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLPIXELTEXGENPARAMETERIVSGISPROC) (GLenum pname, const GLint *params);
-extern void APIENTRY glPixelTexGenParameterfSGIS (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLPIXELTEXGENPARAMETERFSGISPROC) (GLenum pname, GLfloat param);
-extern void APIENTRY glPixelTexGenParameterfvSGIS (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLPIXELTEXGENPARAMETERFVSGISPROC) (GLenum pname, const GLfloat *params);
-extern void APIENTRY glGetPixelTexGenParameterivSGIS (GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETPIXELTEXGENPARAMETERIVSGISPROC) (GLenum pname, GLint *params);
-extern void APIENTRY glGetPixelTexGenParameterfvSGIS (GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETPIXELTEXGENPARAMETERFVSGISPROC) (GLenum pname, GLfloat *params);
-extern void APIENTRY glTexImage4DSGIS (GLenum, GLint, GLenum, GLsizei, GLsizei, GLsizei, GLsizei, GLint, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXIMAGE4DSGISPROC) (GLenum target, GLint level, GLenum internalformat, GLsizei width, GLsizei height, GLsizei depth, GLsizei size4d, GLint border, GLenum format, GLenum type, const GLvoid *pixels);
-extern void APIENTRY glTexSubImage4DSGIS (GLenum, GLint, GLint, GLint, GLint, GLint, GLsizei, GLsizei, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXSUBIMAGE4DSGISPROC) (GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLint woffset, GLsizei width, GLsizei height, GLsizei depth, GLsizei size4d, GLenum format, GLenum type, const GLvoid *pixels);
-extern GLboolean APIENTRY glAreTexturesResidentEXT (GLsizei, const GLuint *, GLboolean *);
-typedef GLboolean (APIENTRY * PFNGLARETEXTURESRESIDENTEXTPROC) (GLsizei n, const GLuint *textures, GLboolean *residences);
-extern void APIENTRY glBindTextureEXT (GLenum, GLuint);
-typedef void (APIENTRY * PFNGLBINDTEXTUREEXTPROC) (GLenum target, GLuint texture);
-extern void APIENTRY glDeleteTexturesEXT (GLsizei, const GLuint *);
-typedef void (APIENTRY * PFNGLDELETETEXTURESEXTPROC) (GLsizei n, const GLuint *textures);
-extern void APIENTRY glGenTexturesEXT (GLsizei, GLuint *);
-typedef void (APIENTRY * PFNGLGENTEXTURESEXTPROC) (GLsizei n, GLuint *textures);
-extern GLboolean APIENTRY glIsTextureEXT (GLuint);
-typedef GLboolean (APIENTRY * PFNGLISTEXTUREEXTPROC) (GLuint texture);
-extern void APIENTRY glPrioritizeTexturesEXT (GLsizei, const GLuint *, const GLclampf *);
-typedef void (APIENTRY * PFNGLPRIORITIZETEXTURESEXTPROC) (GLsizei n, const GLuint *textures, const GLclampf *priorities);
-extern void APIENTRY glDetailTexFuncSGIS (GLenum, GLsizei, const GLfloat *);
-typedef void (APIENTRY * PFNGLDETAILTEXFUNCSGISPROC) (GLenum target, GLsizei n, const GLfloat *points);
-extern void APIENTRY glGetDetailTexFuncSGIS (GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETDETAILTEXFUNCSGISPROC) (GLenum target, GLfloat *points);
-extern void APIENTRY glSharpenTexFuncSGIS (GLenum, GLsizei, const GLfloat *);
-typedef void (APIENTRY * PFNGLSHARPENTEXFUNCSGISPROC) (GLenum target, GLsizei n, const GLfloat *points);
-extern void APIENTRY glGetSharpenTexFuncSGIS (GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETSHARPENTEXFUNCSGISPROC) (GLenum target, GLfloat *points);
-extern void APIENTRY glSampleMaskSGIS (GLclampf, GLboolean);
-typedef void (APIENTRY * PFNGLSAMPLEMASKSGISPROC) (GLclampf value, GLboolean invert);
-extern void APIENTRY glSamplePatternSGIS (GLenum);
-typedef void (APIENTRY * PFNGLSAMPLEPATTERNSGISPROC) (GLenum pattern);
-extern void APIENTRY glArrayElementEXT (GLint);
-typedef void (APIENTRY * PFNGLARRAYELEMENTEXTPROC) (GLint i);
-extern void APIENTRY glColorPointerEXT (GLint, GLenum, GLsizei, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOLORPOINTEREXTPROC) (GLint size, GLenum type, GLsizei stride, GLsizei count, const GLvoid *pointer);
-extern void APIENTRY glDrawArraysEXT (GLenum, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLDRAWARRAYSEXTPROC) (GLenum mode, GLint first, GLsizei count);
-extern void APIENTRY glEdgeFlagPointerEXT (GLsizei, GLsizei, const GLboolean *);
-typedef void (APIENTRY * PFNGLEDGEFLAGPOINTEREXTPROC) (GLsizei stride, GLsizei count, const GLboolean *pointer);
-extern void APIENTRY glGetPointervEXT (GLenum, GLvoid* *);
-typedef void (APIENTRY * PFNGLGETPOINTERVEXTPROC) (GLenum pname, GLvoid* *params);
-extern void APIENTRY glIndexPointerEXT (GLenum, GLsizei, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLINDEXPOINTEREXTPROC) (GLenum type, GLsizei stride, GLsizei count, const GLvoid *pointer);
-extern void APIENTRY glNormalPointerEXT (GLenum, GLsizei, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLNORMALPOINTEREXTPROC) (GLenum type, GLsizei stride, GLsizei count, const GLvoid *pointer);
-extern void APIENTRY glTexCoordPointerEXT (GLint, GLenum, GLsizei, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLTEXCOORDPOINTEREXTPROC) (GLint size, GLenum type, GLsizei stride, GLsizei count, const GLvoid *pointer);
-extern void APIENTRY glVertexPointerEXT (GLint, GLenum, GLsizei, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLVERTEXPOINTEREXTPROC) (GLint size, GLenum type, GLsizei stride, GLsizei count, const GLvoid *pointer);
-extern void APIENTRY glBlendEquationEXT (GLenum);
-typedef void (APIENTRY * PFNGLBLENDEQUATIONEXTPROC) (GLenum mode);
-extern void APIENTRY glSpriteParameterfSGIX (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLSPRITEPARAMETERFSGIXPROC) (GLenum pname, GLfloat param);
-extern void APIENTRY glSpriteParameterfvSGIX (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLSPRITEPARAMETERFVSGIXPROC) (GLenum pname, const GLfloat *params);
-extern void APIENTRY glSpriteParameteriSGIX (GLenum, GLint);
-typedef void (APIENTRY * PFNGLSPRITEPARAMETERISGIXPROC) (GLenum pname, GLint param);
-extern void APIENTRY glSpriteParameterivSGIX (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLSPRITEPARAMETERIVSGIXPROC) (GLenum pname, const GLint *params);
-extern void APIENTRY glPointParameterfEXT (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLPOINTPARAMETERFEXTPROC) (GLenum pname, GLfloat param);
-extern void APIENTRY glPointParameterfvEXT (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLPOINTPARAMETERFVEXTPROC) (GLenum pname, const GLfloat *params);
-extern void APIENTRY glPointParameterfSGIS (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLPOINTPARAMETERFSGISPROC) (GLenum pname, GLfloat param);
-extern void APIENTRY glPointParameterfvSGIS (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLPOINTPARAMETERFVSGISPROC) (GLenum pname, const GLfloat *params);
-extern GLint APIENTRY glGetInstrumentsSGIX (void);
-typedef GLint (APIENTRY * PFNGLGETINSTRUMENTSSGIXPROC) (void);
-extern void APIENTRY glInstrumentsBufferSGIX (GLsizei, GLint *);
-typedef void (APIENTRY * PFNGLINSTRUMENTSBUFFERSGIXPROC) (GLsizei size, GLint *buffer);
-extern GLint APIENTRY glPollInstrumentsSGIX (GLint *);
-typedef GLint (APIENTRY * PFNGLPOLLINSTRUMENTSSGIXPROC) (GLint *marker_p);
-extern void APIENTRY glReadInstrumentsSGIX (GLint);
-typedef void (APIENTRY * PFNGLREADINSTRUMENTSSGIXPROC) (GLint marker);
-extern void APIENTRY glStartInstrumentsSGIX (void);
-typedef void (APIENTRY * PFNGLSTARTINSTRUMENTSSGIXPROC) (void);
-extern void APIENTRY glStopInstrumentsSGIX (GLint);
-typedef void (APIENTRY * PFNGLSTOPINSTRUMENTSSGIXPROC) (GLint marker);
-extern void APIENTRY glFrameZoomSGIX (GLint);
-typedef void (APIENTRY * PFNGLFRAMEZOOMSGIXPROC) (GLint factor);
-extern void APIENTRY glTagSampleBufferSGIX (void);
-typedef void (APIENTRY * PFNGLTAGSAMPLEBUFFERSGIXPROC) (void);
-extern void APIENTRY glReferencePlaneSGIX (const GLdouble *);
-typedef void (APIENTRY * PFNGLREFERENCEPLANESGIXPROC) (const GLdouble *equation);
-extern void APIENTRY glFlushRasterSGIX (void);
-typedef void (APIENTRY * PFNGLFLUSHRASTERSGIXPROC) (void);
-extern void APIENTRY glFogFuncSGIS (GLsizei, const GLfloat *);
-typedef void (APIENTRY * PFNGLFOGFUNCSGISPROC) (GLsizei n, const GLfloat *points);
-extern void APIENTRY glGetFogFuncSGIS (const GLfloat *);
-typedef void (APIENTRY * PFNGLGETFOGFUNCSGISPROC) (const GLfloat *points);
-extern void APIENTRY glImageTransformParameteriHP (GLenum, GLenum, GLint);
-typedef void (APIENTRY * PFNGLIMAGETRANSFORMPARAMETERIHPPROC) (GLenum target, GLenum pname, GLint param);
-extern void APIENTRY glImageTransformParameterfHP (GLenum, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLIMAGETRANSFORMPARAMETERFHPPROC) (GLenum target, GLenum pname, GLfloat param);
-extern void APIENTRY glImageTransformParameterivHP (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLIMAGETRANSFORMPARAMETERIVHPPROC) (GLenum target, GLenum pname, const GLint *params);
-extern void APIENTRY glImageTransformParameterfvHP (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLIMAGETRANSFORMPARAMETERFVHPPROC) (GLenum target, GLenum pname, const GLfloat *params);
-extern void APIENTRY glGetImageTransformParameterivHP (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETIMAGETRANSFORMPARAMETERIVHPPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glGetImageTransformParameterfvHP (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETIMAGETRANSFORMPARAMETERFVHPPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glColorSubTableEXT (GLenum, GLsizei, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOLORSUBTABLEEXTPROC) (GLenum target, GLsizei start, GLsizei count, GLenum format, GLenum type, const GLvoid *data);
-extern void APIENTRY glCopyColorSubTableEXT (GLenum, GLsizei, GLint, GLint, GLsizei);
-typedef void (APIENTRY * PFNGLCOPYCOLORSUBTABLEEXTPROC) (GLenum target, GLsizei start, GLint x, GLint y, GLsizei width);
-extern void APIENTRY glHintPGI (GLenum, GLint);
-typedef void (APIENTRY * PFNGLHINTPGIPROC) (GLenum target, GLint mode);
-extern void APIENTRY glColorTableEXT (GLenum, GLenum, GLsizei, GLenum, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLCOLORTABLEEXTPROC) (GLenum target, GLenum internalFormat, GLsizei width, GLenum format, GLenum type, const GLvoid *table);
-extern void APIENTRY glGetColorTableEXT (GLenum, GLenum, GLenum, GLvoid *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEEXTPROC) (GLenum target, GLenum format, GLenum type, GLvoid *data);
-extern void APIENTRY glGetColorTableParameterivEXT (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPARAMETERIVEXTPROC) (GLenum target, GLenum pname, GLint *params);
-extern void APIENTRY glGetColorTableParameterfvEXT (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCOLORTABLEPARAMETERFVEXTPROC) (GLenum target, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetListParameterfvSGIX (GLuint, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETLISTPARAMETERFVSGIXPROC) (GLuint list, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetListParameterivSGIX (GLuint, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETLISTPARAMETERIVSGIXPROC) (GLuint list, GLenum pname, GLint *params);
-extern void APIENTRY glListParameterfSGIX (GLuint, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLLISTPARAMETERFSGIXPROC) (GLuint list, GLenum pname, GLfloat param);
-extern void APIENTRY glListParameterfvSGIX (GLuint, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLLISTPARAMETERFVSGIXPROC) (GLuint list, GLenum pname, const GLfloat *params);
-extern void APIENTRY glListParameteriSGIX (GLuint, GLenum, GLint);
-typedef void (APIENTRY * PFNGLLISTPARAMETERISGIXPROC) (GLuint list, GLenum pname, GLint param);
-extern void APIENTRY glListParameterivSGIX (GLuint, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLLISTPARAMETERIVSGIXPROC) (GLuint list, GLenum pname, const GLint *params);
-extern void APIENTRY glIndexMaterialEXT (GLenum, GLenum);
-typedef void (APIENTRY * PFNGLINDEXMATERIALEXTPROC) (GLenum face, GLenum mode);
-extern void APIENTRY glIndexFuncEXT (GLenum, GLclampf);
-typedef void (APIENTRY * PFNGLINDEXFUNCEXTPROC) (GLenum func, GLclampf ref);
-extern void APIENTRY glLockArraysEXT (GLint, GLsizei);
-typedef void (APIENTRY * PFNGLLOCKARRAYSEXTPROC) (GLint first, GLsizei count);
-extern void APIENTRY glUnlockArraysEXT (void);
-typedef void (APIENTRY * PFNGLUNLOCKARRAYSEXTPROC) (void);
-extern void APIENTRY glCullParameterdvEXT (GLenum, const GLdouble *);
-typedef void (APIENTRY * PFNGLCULLPARAMETERDVEXTPROC) (GLenum pname, const GLdouble *params);
-extern void APIENTRY glCullParameterfvEXT (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLCULLPARAMETERFVEXTPROC) (GLenum pname, const GLfloat *params);
-extern void APIENTRY glFragmentColorMaterialSGIX (GLenum, GLenum);
-typedef void (APIENTRY * PFNGLFRAGMENTCOLORMATERIALSGIXPROC) (GLenum face, GLenum mode);
-extern void APIENTRY glFragmentLightfSGIX (GLenum, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTFSGIXPROC) (GLenum light, GLenum pname, GLfloat param);
-extern void APIENTRY glFragmentLightfvSGIX (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTFVSGIXPROC) (GLenum light, GLenum pname, const GLfloat *params);
-extern void APIENTRY glFragmentLightiSGIX (GLenum, GLenum, GLint);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTISGIXPROC) (GLenum light, GLenum pname, GLint param);
-extern void APIENTRY glFragmentLightivSGIX (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTIVSGIXPROC) (GLenum light, GLenum pname, const GLint *params);
-extern void APIENTRY glFragmentLightModelfSGIX (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTMODELFSGIXPROC) (GLenum pname, GLfloat param);
-extern void APIENTRY glFragmentLightModelfvSGIX (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTMODELFVSGIXPROC) (GLenum pname, const GLfloat *params);
-extern void APIENTRY glFragmentLightModeliSGIX (GLenum, GLint);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTMODELISGIXPROC) (GLenum pname, GLint param);
-extern void APIENTRY glFragmentLightModelivSGIX (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLFRAGMENTLIGHTMODELIVSGIXPROC) (GLenum pname, const GLint *params);
-extern void APIENTRY glFragmentMaterialfSGIX (GLenum, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLFRAGMENTMATERIALFSGIXPROC) (GLenum face, GLenum pname, GLfloat param);
-extern void APIENTRY glFragmentMaterialfvSGIX (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLFRAGMENTMATERIALFVSGIXPROC) (GLenum face, GLenum pname, const GLfloat *params);
-extern void APIENTRY glFragmentMaterialiSGIX (GLenum, GLenum, GLint);
-typedef void (APIENTRY * PFNGLFRAGMENTMATERIALISGIXPROC) (GLenum face, GLenum pname, GLint param);
-extern void APIENTRY glFragmentMaterialivSGIX (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLFRAGMENTMATERIALIVSGIXPROC) (GLenum face, GLenum pname, const GLint *params);
-extern void APIENTRY glGetFragmentLightfvSGIX (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETFRAGMENTLIGHTFVSGIXPROC) (GLenum light, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetFragmentLightivSGIX (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETFRAGMENTLIGHTIVSGIXPROC) (GLenum light, GLenum pname, GLint *params);
-extern void APIENTRY glGetFragmentMaterialfvSGIX (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETFRAGMENTMATERIALFVSGIXPROC) (GLenum face, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetFragmentMaterialivSGIX (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETFRAGMENTMATERIALIVSGIXPROC) (GLenum face, GLenum pname, GLint *params);
-extern void APIENTRY glLightEnviSGIX (GLenum, GLint);
-typedef void (APIENTRY * PFNGLLIGHTENVISGIXPROC) (GLenum pname, GLint param);
-extern void APIENTRY glDrawRangeElementsEXT (GLenum, GLuint, GLuint, GLsizei, GLenum, const GLvoid *);
-typedef void (APIENTRY * PFNGLDRAWRANGEELEMENTSEXTPROC) (GLenum mode, GLuint start, GLuint end, GLsizei count, GLenum type, const GLvoid *indices);
-extern void APIENTRY glApplyTextureEXT (GLenum);
-typedef void (APIENTRY * PFNGLAPPLYTEXTUREEXTPROC) (GLenum mode);
-extern void APIENTRY glTextureLightEXT (GLenum);
-typedef void (APIENTRY * PFNGLTEXTURELIGHTEXTPROC) (GLenum pname);
-extern void APIENTRY glTextureMaterialEXT (GLenum, GLenum);
-typedef void (APIENTRY * PFNGLTEXTUREMATERIALEXTPROC) (GLenum face, GLenum mode);
-extern void APIENTRY glVertexPointervINTEL (GLint, GLenum, const GLvoid* *);
-typedef void (APIENTRY * PFNGLVERTEXPOINTERVINTELPROC) (GLint size, GLenum type, const GLvoid* *pointer);
-extern void APIENTRY glNormalPointervINTEL (GLenum, const GLvoid* *);
-typedef void (APIENTRY * PFNGLNORMALPOINTERVINTELPROC) (GLenum type, const GLvoid* *pointer);
-extern void APIENTRY glColorPointervINTEL (GLint, GLenum, const GLvoid* *);
-typedef void (APIENTRY * PFNGLCOLORPOINTERVINTELPROC) (GLint size, GLenum type, const GLvoid* *pointer);
-extern void APIENTRY glTexCoordPointervINTEL (GLint, GLenum, const GLvoid* *);
-typedef void (APIENTRY * PFNGLTEXCOORDPOINTERVINTELPROC) (GLint size, GLenum type, const GLvoid* *pointer);
-extern void APIENTRY glPixelTransformParameteriEXT (GLenum, GLenum, GLint);
-typedef void (APIENTRY * PFNGLPIXELTRANSFORMPARAMETERIEXTPROC) (GLenum target, GLenum pname, GLint param);
-extern void APIENTRY glPixelTransformParameterfEXT (GLenum, GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLPIXELTRANSFORMPARAMETERFEXTPROC) (GLenum target, GLenum pname, GLfloat param);
-extern void APIENTRY glPixelTransformParameterivEXT (GLenum, GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLPIXELTRANSFORMPARAMETERIVEXTPROC) (GLenum target, GLenum pname, const GLint *params);
-extern void APIENTRY glPixelTransformParameterfvEXT (GLenum, GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLPIXELTRANSFORMPARAMETERFVEXTPROC) (GLenum target, GLenum pname, const GLfloat *params);
-extern void APIENTRY glSecondaryColor3bEXT (GLbyte, GLbyte, GLbyte);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3BEXTPROC) (GLbyte red, GLbyte green, GLbyte blue);
-extern void APIENTRY glSecondaryColor3bvEXT (const GLbyte *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3BVEXTPROC) (const GLbyte *v);
-extern void APIENTRY glSecondaryColor3dEXT (GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3DEXTPROC) (GLdouble red, GLdouble green, GLdouble blue);
-extern void APIENTRY glSecondaryColor3dvEXT (const GLdouble *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3DVEXTPROC) (const GLdouble *v);
-extern void APIENTRY glSecondaryColor3fEXT (GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3FEXTPROC) (GLfloat red, GLfloat green, GLfloat blue);
-extern void APIENTRY glSecondaryColor3fvEXT (const GLfloat *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3FVEXTPROC) (const GLfloat *v);
-extern void APIENTRY glSecondaryColor3iEXT (GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3IEXTPROC) (GLint red, GLint green, GLint blue);
-extern void APIENTRY glSecondaryColor3ivEXT (const GLint *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3IVEXTPROC) (const GLint *v);
-extern void APIENTRY glSecondaryColor3sEXT (GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3SEXTPROC) (GLshort red, GLshort green, GLshort blue);
-extern void APIENTRY glSecondaryColor3svEXT (const GLshort *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3SVEXTPROC) (const GLshort *v);
-extern void APIENTRY glSecondaryColor3ubEXT (GLubyte, GLubyte, GLubyte);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3UBEXTPROC) (GLubyte red, GLubyte green, GLubyte blue);
-extern void APIENTRY glSecondaryColor3ubvEXT (const GLubyte *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3UBVEXTPROC) (const GLubyte *v);
-extern void APIENTRY glSecondaryColor3uiEXT (GLuint, GLuint, GLuint);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3UIEXTPROC) (GLuint red, GLuint green, GLuint blue);
-extern void APIENTRY glSecondaryColor3uivEXT (const GLuint *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3UIVEXTPROC) (const GLuint *v);
-extern void APIENTRY glSecondaryColor3usEXT (GLushort, GLushort, GLushort);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3USEXTPROC) (GLushort red, GLushort green, GLushort blue);
-extern void APIENTRY glSecondaryColor3usvEXT (const GLushort *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLOR3USVEXTPROC) (const GLushort *v);
-extern void APIENTRY glSecondaryColorPointerEXT (GLint, GLenum, GLsizei, GLvoid *);
-typedef void (APIENTRY * PFNGLSECONDARYCOLORPOINTEREXTPROC) (GLint size, GLenum type, GLsizei stride, GLvoid *pointer);
-extern void APIENTRY glTextureNormalEXT (GLenum);
-typedef void (APIENTRY * PFNGLTEXTURENORMALEXTPROC) (GLenum mode);
-extern void APIENTRY glMultiDrawArraysEXT (GLenum, GLint *, GLsizei *, GLsizei);
-typedef void (APIENTRY * PFNGLMULTIDRAWARRAYSEXTPROC) (GLenum mode, GLint *first, GLsizei *count, GLsizei primcount);
-extern void APIENTRY glMultiDrawElementsEXT (GLenum, const GLsizei *, GLenum, const GLvoid* *, GLsizei);
-typedef void (APIENTRY * PFNGLMULTIDRAWELEMENTSEXTPROC) (GLenum mode, const GLsizei *count, GLenum type, const GLvoid* *indices, GLsizei primcount);
-extern void APIENTRY glFogCoordfEXT (GLfloat);
-typedef void (APIENTRY * PFNGLFOGCOORDFEXTPROC) (GLfloat coord);
-extern void APIENTRY glFogCoordfvEXT (const GLfloat *);
-typedef void (APIENTRY * PFNGLFOGCOORDFVEXTPROC) (const GLfloat *coord);
-extern void APIENTRY glFogCoorddEXT (GLdouble);
-typedef void (APIENTRY * PFNGLFOGCOORDDEXTPROC) (GLdouble coord);
-extern void APIENTRY glFogCoorddvEXT (const GLdouble *);
-typedef void (APIENTRY * PFNGLFOGCOORDDVEXTPROC) (const GLdouble *coord);
-extern void APIENTRY glFogCoordPointerEXT (GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLFOGCOORDPOINTEREXTPROC) (GLenum type, GLsizei stride, const GLvoid *pointer);
-extern void APIENTRY glTangent3b (GLbyte, GLbyte, GLbyte);
-typedef void (APIENTRY * PFNGLTANGENT3BPROC) (GLbyte tx, GLbyte ty, GLbyte tz);
-extern void APIENTRY glTangent3bv (const GLbyte *);
-typedef void (APIENTRY * PFNGLTANGENT3BVPROC) (const GLbyte *v);
-extern void APIENTRY glTangent3d (GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLTANGENT3DPROC) (GLdouble tx, GLdouble ty, GLdouble tz);
-extern void APIENTRY glTangent3dv (const GLdouble *);
-typedef void (APIENTRY * PFNGLTANGENT3DVPROC) (const GLdouble *v);
-extern void APIENTRY glTangent3f (GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTANGENT3FPROC) (GLfloat tx, GLfloat ty, GLfloat tz);
-extern void APIENTRY glTangent3fv (const GLfloat *);
-typedef void (APIENTRY * PFNGLTANGENT3FVPROC) (const GLfloat *v);
-extern void APIENTRY glTangent3i (GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLTANGENT3IPROC) (GLint tx, GLint ty, GLint tz);
-extern void APIENTRY glTangent3iv (const GLint *);
-typedef void (APIENTRY * PFNGLTANGENT3IVPROC) (const GLint *v);
-extern void APIENTRY glTangent3s (GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLTANGENT3SPROC) (GLshort tx, GLshort ty, GLshort tz);
-extern void APIENTRY glTangent3sv (const GLshort *);
-typedef void (APIENTRY * PFNGLTANGENT3SVPROC) (const GLshort *v);
-extern void APIENTRY glBinormal3b (GLbyte, GLbyte, GLbyte);
-typedef void (APIENTRY * PFNGLBINORMAL3BPROC) (GLbyte bx, GLbyte by, GLbyte bz);
-extern void APIENTRY glBinormal3bv (const GLbyte *);
-typedef void (APIENTRY * PFNGLBINORMAL3BVPROC) (const GLbyte *v);
-extern void APIENTRY glBinormal3d (GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLBINORMAL3DPROC) (GLdouble bx, GLdouble by, GLdouble bz);
-extern void APIENTRY glBinormal3dv (const GLdouble *);
-typedef void (APIENTRY * PFNGLBINORMAL3DVPROC) (const GLdouble *v);
-extern void APIENTRY glBinormal3f (GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLBINORMAL3FPROC) (GLfloat bx, GLfloat by, GLfloat bz);
-extern void APIENTRY glBinormal3fv (const GLfloat *);
-typedef void (APIENTRY * PFNGLBINORMAL3FVPROC) (const GLfloat *v);
-extern void APIENTRY glBinormal3i (GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLBINORMAL3IPROC) (GLint bx, GLint by, GLint bz);
-extern void APIENTRY glBinormal3iv (const GLint *);
-typedef void (APIENTRY * PFNGLBINORMAL3IVPROC) (const GLint *v);
-extern void APIENTRY glBinormal3s (GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLBINORMAL3SPROC) (GLshort bx, GLshort by, GLshort bz);
-extern void APIENTRY glBinormal3sv (const GLshort *);
-typedef void (APIENTRY * PFNGLBINORMAL3SVPROC) (const GLshort *v);
-extern void APIENTRY glTangentPointerEXT (GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLTANGENTPOINTEREXTPROC) (GLenum type, GLsizei stride, const GLvoid *pointer);
-extern void APIENTRY glBinormalPointerEXT (GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLBINORMALPOINTEREXTPROC) (GLenum type, GLsizei stride, const GLvoid *pointer);
-extern void APIENTRY glFinishTextureSUNX (void);
-typedef void (APIENTRY * PFNGLFINISHTEXTURESUNXPROC) (void);
-extern void APIENTRY glGlobalAlphaFactorbSUN (GLbyte);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORBSUNPROC) (GLbyte factor);
-extern void APIENTRY glGlobalAlphaFactorsSUN (GLshort);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORSSUNPROC) (GLshort factor);
-extern void APIENTRY glGlobalAlphaFactoriSUN (GLint);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORISUNPROC) (GLint factor);
-extern void APIENTRY glGlobalAlphaFactorfSUN (GLfloat);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORFSUNPROC) (GLfloat factor);
-extern void APIENTRY glGlobalAlphaFactordSUN (GLdouble);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORDSUNPROC) (GLdouble factor);
-extern void APIENTRY glGlobalAlphaFactorubSUN (GLubyte);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORUBSUNPROC) (GLubyte factor);
-extern void APIENTRY glGlobalAlphaFactorusSUN (GLushort);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORUSSUNPROC) (GLushort factor);
-extern void APIENTRY glGlobalAlphaFactoruiSUN (GLuint);
-typedef void (APIENTRY * PFNGLGLOBALALPHAFACTORUISUNPROC) (GLuint factor);
-extern void APIENTRY glReplacementCodeuiSUN (GLuint);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUISUNPROC) (GLuint code);
-extern void APIENTRY glReplacementCodeusSUN (GLushort);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUSSUNPROC) (GLushort code);
-extern void APIENTRY glReplacementCodeubSUN (GLubyte);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUBSUNPROC) (GLubyte code);
-extern void APIENTRY glReplacementCodeuivSUN (const GLuint *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUIVSUNPROC) (const GLuint *code);
-extern void APIENTRY glReplacementCodeusvSUN (const GLushort *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUSVSUNPROC) (const GLushort *code);
-extern void APIENTRY glReplacementCodeubvSUN (const GLubyte *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUBVSUNPROC) (const GLubyte *code);
-extern void APIENTRY glReplacementCodePointer (GLenum, GLsizei, const GLvoid* *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEPOINTERPROC) (GLenum type, GLsizei stride, const GLvoid* *pointer);
-extern void APIENTRY glColor4ubVertex2fSUN (GLubyte, GLubyte, GLubyte, GLubyte, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLCOLOR4UBVERTEX2FSUNPROC) (GLubyte r, GLubyte g, GLubyte b, GLubyte a, GLfloat x, GLfloat y);
-extern void APIENTRY glColor4ubVertex2fvSUN (const GLubyte *, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOLOR4UBVERTEX2FVSUNPROC) (const GLubyte *c, const GLfloat *v);
-extern void APIENTRY glColor4ubVertex3fSUN (GLubyte, GLubyte, GLubyte, GLubyte, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLCOLOR4UBVERTEX3FSUNPROC) (GLubyte r, GLubyte g, GLubyte b, GLubyte a, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glColor4ubVertex3fvSUN (const GLubyte *, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOLOR4UBVERTEX3FVSUNPROC) (const GLubyte *c, const GLfloat *v);
-extern void APIENTRY glColor3fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLCOLOR3FVERTEX3FSUNPROC) (GLfloat r, GLfloat g, GLfloat b, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glColor3fVertex3fvSUN (const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOLOR3FVERTEX3FVSUNPROC) (const GLfloat *c, const GLfloat *v);
-extern void APIENTRY glNormal3fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLNORMAL3FVERTEX3FSUNPROC) (GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glNormal3fVertex3fvSUN (const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLNORMAL3FVERTEX3FVSUNPROC) (const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glColor4fNormal3fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLCOLOR4FNORMAL3FVERTEX3FSUNPROC) (GLfloat r, GLfloat g, GLfloat b, GLfloat a, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glColor4fNormal3fVertex3fvSUN (const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOLOR4FNORMAL3FVERTEX3FVSUNPROC) (const GLfloat *c, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glTexCoord2fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD2FVERTEX3FSUNPROC) (GLfloat s, GLfloat t, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glTexCoord2fVertex3fvSUN (const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD2FVERTEX3FVSUNPROC) (const GLfloat *tc, const GLfloat *v);
-extern void APIENTRY glTexCoord4fVertex4fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD4FVERTEX4FSUNPROC) (GLfloat s, GLfloat t, GLfloat p, GLfloat q, GLfloat x, GLfloat y, GLfloat z, GLfloat w);
-extern void APIENTRY glTexCoord4fVertex4fvSUN (const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD4FVERTEX4FVSUNPROC) (const GLfloat *tc, const GLfloat *v);
-extern void APIENTRY glTexCoord2fColor4ubVertex3fSUN (GLfloat, GLfloat, GLubyte, GLubyte, GLubyte, GLubyte, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD2FCOLOR4UBVERTEX3FSUNPROC) (GLfloat s, GLfloat t, GLubyte r, GLubyte g, GLubyte b, GLubyte a, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glTexCoord2fColor4ubVertex3fvSUN (const GLfloat *, const GLubyte *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD2FCOLOR4UBVERTEX3FVSUNPROC) (const GLfloat *tc, const GLubyte *c, const GLfloat *v);
-extern void APIENTRY glTexCoord2fColor3fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD2FCOLOR3FVERTEX3FSUNPROC) (GLfloat s, GLfloat t, GLfloat r, GLfloat g, GLfloat b, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glTexCoord2fColor3fVertex3fvSUN (const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD2FCOLOR3FVERTEX3FVSUNPROC) (const GLfloat *tc, const GLfloat *c, const GLfloat *v);
-extern void APIENTRY glTexCoord2fNormal3fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD2FNORMAL3FVERTEX3FSUNPROC) (GLfloat s, GLfloat t, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glTexCoord2fNormal3fVertex3fvSUN (const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD2FNORMAL3FVERTEX3FVSUNPROC) (const GLfloat *tc, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glTexCoord2fColor4fNormal3fVertex3fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD2FCOLOR4FNORMAL3FVERTEX3FSUNPROC) (GLfloat s, GLfloat t, GLfloat r, GLfloat g, GLfloat b, GLfloat a, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glTexCoord2fColor4fNormal3fVertex3fvSUN (const GLfloat *, const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD2FCOLOR4FNORMAL3FVERTEX3FVSUNPROC) (const GLfloat *tc, const GLfloat *c, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glTexCoord4fColor4fNormal3fVertex4fSUN (GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLTEXCOORD4FCOLOR4FNORMAL3FVERTEX4FSUNPROC) (GLfloat s, GLfloat t, GLfloat p, GLfloat q, GLfloat r, GLfloat g, GLfloat b, GLfloat a, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z, GLfloat w);
-extern void APIENTRY glTexCoord4fColor4fNormal3fVertex4fvSUN (const GLfloat *, const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLTEXCOORD4FCOLOR4FNORMAL3FVERTEX4FVSUNPROC) (const GLfloat *tc, const GLfloat *c, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUIVERTEX3FSUNPROC) (GLenum rc, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiVertex3fvSUN (const GLenum *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUIVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiColor4ubVertex3fSUN (GLenum, GLubyte, GLubyte, GLubyte, GLubyte, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUICOLOR4UBVERTEX3FSUNPROC) (GLenum rc, GLubyte r, GLubyte g, GLubyte b, GLubyte a, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiColor4ubVertex3fvSUN (const GLenum *, const GLubyte *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUICOLOR4UBVERTEX3FVSUNPROC) (const GLenum *rc, const GLubyte *c, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiColor3fVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUICOLOR3FVERTEX3FSUNPROC) (GLenum rc, GLfloat r, GLfloat g, GLfloat b, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiColor3fVertex3fvSUN (const GLenum *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUICOLOR3FVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *c, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiNormal3fVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUINORMAL3FVERTEX3FSUNPROC) (GLenum rc, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiNormal3fVertex3fvSUN (const GLenum *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUINORMAL3FVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiColor4fNormal3fVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUICOLOR4FNORMAL3FVERTEX3FSUNPROC) (GLenum rc, GLfloat r, GLfloat g, GLfloat b, GLfloat a, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiColor4fNormal3fVertex3fvSUN (const GLenum *, const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUICOLOR4FNORMAL3FVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *c, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiTexCoord2fVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUITEXCOORD2FVERTEX3FSUNPROC) (GLenum rc, GLfloat s, GLfloat t, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiTexCoord2fVertex3fvSUN (const GLenum *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUITEXCOORD2FVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *tc, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiTexCoord2fNormal3fVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUITEXCOORD2FNORMAL3FVERTEX3FSUNPROC) (GLenum rc, GLfloat s, GLfloat t, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiTexCoord2fNormal3fVertex3fvSUN (const GLenum *, const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUITEXCOORD2FNORMAL3FVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *tc, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glReplacementCodeuiTexCoord2fColor4fNormal3fVertex3fSUN (GLenum, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUITEXCOORD2FCOLOR4FNORMAL3FVERTEX3FSUNPROC) (GLenum rc, GLfloat s, GLfloat t, GLfloat r, GLfloat g, GLfloat b, GLfloat a, GLfloat nx, GLfloat ny, GLfloat nz, GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glReplacementCodeuiTexCoord2fColor4fNormal3fVertex3fvSUN (const GLenum *, const GLfloat *, const GLfloat *, const GLfloat *, const GLfloat *);
-typedef void (APIENTRY * PFNGLREPLACEMENTCODEUITEXCOORD2FCOLOR4FNORMAL3FVERTEX3FVSUNPROC) (const GLenum *rc, const GLfloat *tc, const GLfloat *c, const GLfloat *n, const GLfloat *v);
-extern void APIENTRY glBlendFuncSeparateEXT (GLenum, GLenum, GLenum, GLenum);
-typedef void (APIENTRY * PFNGLBLENDFUNCSEPARATEEXTPROC) (GLenum sfactorRGB, GLenum dfactorRGB, GLenum sfactorAlpha, GLenum dfactorAlpha);
-extern void APIENTRY glVertexWeightfEXT (GLfloat);
-typedef void (APIENTRY * PFNGLVERTEXWEIGHTFEXTPROC) (GLfloat weight);
-extern void APIENTRY glVertexWeightfvEXT (const GLfloat *);
-typedef void (APIENTRY * PFNGLVERTEXWEIGHTFVEXTPROC) (const GLfloat *weight);
-extern void APIENTRY glVertexWeightPointerEXT (GLsizei, GLenum, GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLVERTEXWEIGHTPOINTEREXTPROC) (GLsizei size, GLenum type, GLsizei stride, const GLvoid *pointer);
-extern void APIENTRY glFlushVertexArrayRangeNV (void);
-typedef void (APIENTRY * PFNGLFLUSHVERTEXARRAYRANGENVPROC) (void);
-extern void APIENTRY glVertexArrayRangeNV (GLsizei, const GLvoid *);
-typedef void (APIENTRY * PFNGLVERTEXARRAYRANGENVPROC) (GLsizei size, const GLvoid *pointer);
-extern void APIENTRY glCombinerParameterfvNV (GLenum, const GLfloat *);
-typedef void (APIENTRY * PFNGLCOMBINERPARAMETERFVNVPROC) (GLenum pname, const GLfloat *params);
-extern void APIENTRY glCombinerParameterfNV (GLenum, GLfloat);
-typedef void (APIENTRY * PFNGLCOMBINERPARAMETERFNVPROC) (GLenum pname, GLfloat param);
-extern void APIENTRY glCombinerParameterivNV (GLenum, const GLint *);
-typedef void (APIENTRY * PFNGLCOMBINERPARAMETERIVNVPROC) (GLenum pname, const GLint *params);
-extern void APIENTRY glCombinerParameteriNV (GLenum, GLint);
-typedef void (APIENTRY * PFNGLCOMBINERPARAMETERINVPROC) (GLenum pname, GLint param);
-extern void APIENTRY glCombinerInputNV (GLenum, GLenum, GLenum, GLenum, GLenum, GLenum);
-typedef void (APIENTRY * PFNGLCOMBINERINPUTNVPROC) (GLenum stage, GLenum portion, GLenum variable, GLenum input, GLenum mapping, GLenum componentUsage);
-extern void APIENTRY glCombinerOutputNV (GLenum, GLenum, GLenum, GLenum, GLenum, GLenum, GLenum, GLboolean, GLboolean, GLboolean);
-typedef void (APIENTRY * PFNGLCOMBINEROUTPUTNVPROC) (GLenum stage, GLenum portion, GLenum abOutput, GLenum cdOutput, GLenum sumOutput, GLenum scale, GLenum bias, GLboolean abDotProduct, GLboolean cdDotProduct, GLboolean muxSum);
-extern void APIENTRY glFinalCombinerInputNV (GLenum, GLenum, GLenum, GLenum);
-typedef void (APIENTRY * PFNGLFINALCOMBINERINPUTNVPROC) (GLenum variable, GLenum input, GLenum mapping, GLenum componentUsage);
-extern void APIENTRY glGetCombinerInputParameterfvNV (GLenum, GLenum, GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCOMBINERINPUTPARAMETERFVNVPROC) (GLenum stage, GLenum portion, GLenum variable, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetCombinerInputParameterivNV (GLenum, GLenum, GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCOMBINERINPUTPARAMETERIVNVPROC) (GLenum stage, GLenum portion, GLenum variable, GLenum pname, GLint *params);
-extern void APIENTRY glGetCombinerOutputParameterfvNV (GLenum, GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETCOMBINEROUTPUTPARAMETERFVNVPROC) (GLenum stage, GLenum portion, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetCombinerOutputParameterivNV (GLenum, GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETCOMBINEROUTPUTPARAMETERIVNVPROC) (GLenum stage, GLenum portion, GLenum pname, GLint *params);
-extern void APIENTRY glGetFinalCombinerInputParameterfvNV (GLenum, GLenum, GLfloat *);
-typedef void (APIENTRY * PFNGLGETFINALCOMBINERINPUTPARAMETERFVNVPROC) (GLenum variable, GLenum pname, GLfloat *params);
-extern void APIENTRY glGetFinalCombinerInputParameterivNV (GLenum, GLenum, GLint *);
-typedef void (APIENTRY * PFNGLGETFINALCOMBINERINPUTPARAMETERIVNVPROC) (GLenum variable, GLenum pname, GLint *params);
-extern void APIENTRY glResizeBuffersMESA (void);
-typedef void (APIENTRY * PFNGLRESIZEBUFFERSMESAPROC) (void);
-extern void APIENTRY glWindowPos2dMESA (GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLWINDOWPOS2DMESAPROC) (GLdouble x, GLdouble y);
-extern void APIENTRY glWindowPos2dvMESA (const GLdouble *);
-typedef void (APIENTRY * PFNGLWINDOWPOS2DVMESAPROC) (const GLdouble *v);
-extern void APIENTRY glWindowPos2fMESA (GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLWINDOWPOS2FMESAPROC) (GLfloat x, GLfloat y);
-extern void APIENTRY glWindowPos2fvMESA (const GLfloat *);
-typedef void (APIENTRY * PFNGLWINDOWPOS2FVMESAPROC) (const GLfloat *v);
-extern void APIENTRY glWindowPos2iMESA (GLint, GLint);
-typedef void (APIENTRY * PFNGLWINDOWPOS2IMESAPROC) (GLint x, GLint y);
-extern void APIENTRY glWindowPos2ivMESA (const GLint *);
-typedef void (APIENTRY * PFNGLWINDOWPOS2IVMESAPROC) (const GLint *v);
-extern void APIENTRY glWindowPos2sMESA (GLshort, GLshort);
-typedef void (APIENTRY * PFNGLWINDOWPOS2SMESAPROC) (GLshort x, GLshort y);
-extern void APIENTRY glWindowPos2svMESA (const GLshort *);
-typedef void (APIENTRY * PFNGLWINDOWPOS2SVMESAPROC) (const GLshort *v);
-extern void APIENTRY glWindowPos3dMESA (GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLWINDOWPOS3DMESAPROC) (GLdouble x, GLdouble y, GLdouble z);
-extern void APIENTRY glWindowPos3dvMESA (const GLdouble *);
-typedef void (APIENTRY * PFNGLWINDOWPOS3DVMESAPROC) (const GLdouble *v);
-extern void APIENTRY glWindowPos3fMESA (GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLWINDOWPOS3FMESAPROC) (GLfloat x, GLfloat y, GLfloat z);
-extern void APIENTRY glWindowPos3fvMESA (const GLfloat *);
-typedef void (APIENTRY * PFNGLWINDOWPOS3FVMESAPROC) (const GLfloat *v);
-extern void APIENTRY glWindowPos3iMESA (GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLWINDOWPOS3IMESAPROC) (GLint x, GLint y, GLint z);
-extern void APIENTRY glWindowPos3ivMESA (const GLint *);
-typedef void (APIENTRY * PFNGLWINDOWPOS3IVMESAPROC) (const GLint *v);
-extern void APIENTRY glWindowPos3sMESA (GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLWINDOWPOS3SMESAPROC) (GLshort x, GLshort y, GLshort z);
-extern void APIENTRY glWindowPos3svMESA (const GLshort *);
-typedef void (APIENTRY * PFNGLWINDOWPOS3SVMESAPROC) (const GLshort *v);
-extern void APIENTRY glWindowPos4dMESA (GLdouble, GLdouble, GLdouble, GLdouble);
-typedef void (APIENTRY * PFNGLWINDOWPOS4DMESAPROC) (GLdouble x, GLdouble y, GLdouble z, GLdouble w);
-extern void APIENTRY glWindowPos4dvMESA (const GLdouble *);
-typedef void (APIENTRY * PFNGLWINDOWPOS4DVMESAPROC) (const GLdouble *v);
-extern void APIENTRY glWindowPos4fMESA (GLfloat, GLfloat, GLfloat, GLfloat);
-typedef void (APIENTRY * PFNGLWINDOWPOS4FMESAPROC) (GLfloat x, GLfloat y, GLfloat z, GLfloat w);
-extern void APIENTRY glWindowPos4fvMESA (const GLfloat *);
-typedef void (APIENTRY * PFNGLWINDOWPOS4FVMESAPROC) (const GLfloat *v);
-extern void APIENTRY glWindowPos4iMESA (GLint, GLint, GLint, GLint);
-typedef void (APIENTRY * PFNGLWINDOWPOS4IMESAPROC) (GLint x, GLint y, GLint z, GLint w);
-extern void APIENTRY glWindowPos4ivMESA (const GLint *);
-typedef void (APIENTRY * PFNGLWINDOWPOS4IVMESAPROC) (const GLint *v);
-extern void APIENTRY glWindowPos4sMESA (GLshort, GLshort, GLshort, GLshort);
-typedef void (APIENTRY * PFNGLWINDOWPOS4SMESAPROC) (GLshort x, GLshort y, GLshort z, GLshort w);
-extern void APIENTRY glWindowPos4svMESA (const GLshort *);
-typedef void (APIENTRY * PFNGLWINDOWPOS4SVMESAPROC) (const GLshort *v);
-extern void APIENTRY glMultiModeDrawArraysIBM (GLenum, const GLint *, const GLsizei *, GLsizei, GLint);
-typedef void (APIENTRY * PFNGLMULTIMODEDRAWARRAYSIBMPROC) (GLenum mode, const GLint *first, const GLsizei *count, GLsizei primcount, GLint modestride);
-extern void APIENTRY glMultiModeDrawElementsIBM (const GLenum *, const GLsizei *, GLenum, const GLvoid* *, GLsizei, GLint);
-typedef void (APIENTRY * PFNGLMULTIMODEDRAWELEMENTSIBMPROC) (const GLenum *mode, const GLsizei *count, GLenum type, const GLvoid* *indices, GLsizei primcount, GLint modestride);
-extern void APIENTRY glColorPointerListIBM (GLint, GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLCOLORPOINTERLISTIBMPROC) (GLint size, GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glSecondaryColorPointerListIBM (GLint, GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLSECONDARYCOLORPOINTERLISTIBMPROC) (GLint size, GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glEdgeFlagPointerListIBM (GLint, const GLboolean* *, GLint);
-typedef void (APIENTRY * PFNGLEDGEFLAGPOINTERLISTIBMPROC) (GLint stride, const GLboolean* *pointer, GLint ptrstride);
-extern void APIENTRY glFogCoordPointerListIBM (GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLFOGCOORDPOINTERLISTIBMPROC) (GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glIndexPointerListIBM (GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLINDEXPOINTERLISTIBMPROC) (GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glNormalPointerListIBM (GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLNORMALPOINTERLISTIBMPROC) (GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glTexCoordPointerListIBM (GLint, GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLTEXCOORDPOINTERLISTIBMPROC) (GLint size, GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glVertexPointerListIBM (GLint, GLenum, GLint, const GLvoid* *, GLint);
-typedef void (APIENTRY * PFNGLVERTEXPOINTERLISTIBMPROC) (GLint size, GLenum type, GLint stride, const GLvoid* *pointer, GLint ptrstride);
-extern void APIENTRY glTbufferMask3DFX (GLuint);
-typedef void (APIENTRY * PFNGLTBUFFERMASK3DFXPROC) (GLuint mask);
-extern void APIENTRY glSampleMaskEXT (GLclampf, GLboolean);
-typedef void (APIENTRY * PFNGLSAMPLEMASKEXTPROC) (GLclampf value, GLboolean invert);
-extern void APIENTRY glSamplePatternEXT (GLenum);
-typedef void (APIENTRY * PFNGLSAMPLEPATTERNEXTPROC) (GLenum pattern);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif
diff --git a/xc/extras/Mesa/src/glheader.h b/xc/extras/Mesa/src/glheader.h
index 705c25c74..50316b6f5 100644
--- a/xc/extras/Mesa/src/glheader.h
+++ b/xc/extras/Mesa/src/glheader.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/glheader.h,v 1.11 2000/10/26 17:57:48 dawes Exp $ */
 
 
 #ifndef GLHEADER_H
@@ -83,11 +82,11 @@
  * the new src/glheader.h file.
  */
 
-#if defined(_WIN32) && !defined(__WIN32__) && !defined(__CYGWIN__)
+#if defined(_WIN32) && !defined(__WIN32__)
 #	define __WIN32__
 #endif
 
-#if !defined(OPENSTEP) && (defined(__WIN32__) && !defined(__CYGWIN__))
+#if !defined(OPENSTEP) && (defined(__WIN32__) || defined(__CYGWIN__))
 #  pragma warning( disable : 4068 ) /* unknown pragma */
 #  pragma warning( disable : 4710 ) /* function 'foo' not inlined */
 #  pragma warning( disable : 4711 ) /* function 'foo' selected for automatic inline expansion */
@@ -134,13 +133,13 @@
 
 /* compatability guard so we don't need to change client code */
 
-#if defined(_WIN32) && !defined(_WINDEF_) && !defined(_GNU_H_WINDOWS32_BASE) && !defined(OPENSTEP) && !defined(__CYGWIN__)
+#if defined(_WIN32) && !defined(_WINDEF_) && !defined(_GNU_H_WINDOWS32_BASE) && !defined(OPENSTEP)
 #if 0
 #	define CALLBACK GLCALLBACK
-#endif
-typedef int (GLAPIENTRY *PROC)();
 typedef void *HGLRC;
 typedef void *HDC;
+#endif
+typedef int (GLAPIENTRY *PROC)();
 typedef unsigned long COLORREF;
 #endif
 
@@ -164,10 +163,8 @@ typedef unsigned long COLORREF;
 typedef struct tagLAYERPLANEDESCRIPTOR LAYERPLANEDESCRIPTOR, *PLAYERPLANEDESCRIPTOR, *LPLAYERPLANEDESCRIPTOR;
 typedef struct _GLYPHMETRICSFLOAT GLYPHMETRICSFLOAT, *PGLYPHMETRICSFLOAT, *LPGLYPHMETRICSFLOAT;
 typedef struct tagPIXELFORMATDESCRIPTOR PIXELFORMATDESCRIPTOR, *PPIXELFORMATDESCRIPTOR, *LPPIXELFORMATDESCRIPTOR;
-#if 0
 #include <gl/mesa_wgl.h>
 #endif
-#endif
 
 
 
diff --git a/xc/extras/Mesa/src/image.c b/xc/extras/Mesa/src/image.c
index f1781326d..8053558f9 100644
--- a/xc/extras/Mesa/src/image.c
+++ b/xc/extras/Mesa/src/image.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/image.c,v 1.7 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -629,7 +629,7 @@ _mesa_pack_rgba_span( GLcontext *ctx,
    }
    else if (!applyTransferOps && format == GL_RGB && type == GL_UNSIGNED_BYTE) {
       /* common simple case */
-      GLint i;
+      GLuint i;
       GLubyte *dest = (GLubyte *) destination;
       for (i = 0; i < n; i++) {
          dest[0] = srcRgba[i][RCOMP];
@@ -1276,7 +1276,7 @@ _mesa_pack_rgba_span( GLcontext *ctx,
             }
             break;
          case GL_UNSIGNED_SHORT_4_4_4_4:
-            if (format == GL_RGB) {
+            if (format == GL_RGBA) {
                GLushort *dst = (GLushort *) destination;
                for (i=0;i<n;i++) {
                   dst[i] = (((GLint) (rgba[i][RCOMP] * 15.0F)) << 12)
@@ -1285,9 +1285,27 @@ _mesa_pack_rgba_span( GLcontext *ctx,
                          | (((GLint) (rgba[i][ACOMP] * 15.0F))      );
                }
             }
+            else if (format == GL_BGRA) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][BCOMP] * 15.0F)) << 12)
+                         | (((GLint) (rgba[i][GCOMP] * 15.0F)) <<  8)
+                         | (((GLint) (rgba[i][RCOMP] * 15.0F)) <<  4)
+                         | (((GLint) (rgba[i][ACOMP] * 15.0F))      );
+               }
+            }
+            else if (format == GL_ABGR_EXT) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][ACOMP] * 15.0F)) <<  4)
+                         | (((GLint) (rgba[i][BCOMP] * 15.0F)) <<  8)
+                         | (((GLint) (rgba[i][GCOMP] * 15.0F)) << 12)
+                         | (((GLint) (rgba[i][RCOMP] * 15.0F))      );
+               }
+            }
             break;
          case GL_UNSIGNED_SHORT_4_4_4_4_REV:
-            if (format == GL_RGB) {
+            if (format == GL_RGBA) {
                GLushort *dst = (GLushort *) destination;
                for (i=0;i<n;i++) {
                   dst[i] = (((GLint) (rgba[i][RCOMP] * 15.0F))      )
@@ -1296,9 +1314,27 @@ _mesa_pack_rgba_span( GLcontext *ctx,
                          | (((GLint) (rgba[i][ACOMP] * 15.0F)) << 12);
                }
             }
+            else if (format == GL_BGRA) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][BCOMP] * 15.0F))      )
+                         | (((GLint) (rgba[i][GCOMP] * 15.0F)) <<  4)
+                         | (((GLint) (rgba[i][RCOMP] * 15.0F)) <<  8)
+                         | (((GLint) (rgba[i][ACOMP] * 15.0F)) << 12);
+               }
+            }
+            else if (format == GL_ABGR_EXT) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][ACOMP] * 15.0F))      )
+                         | (((GLint) (rgba[i][BCOMP] * 15.0F)) <<  4)
+                         | (((GLint) (rgba[i][GCOMP] * 15.0F)) <<  8)
+                         | (((GLint) (rgba[i][RCOMP] * 15.0F)) << 12);
+               }
+            }
             break;
          case GL_UNSIGNED_SHORT_5_5_5_1:
-            if (format == GL_RGB) {
+            if (format == GL_RGBA) {
                GLushort *dst = (GLushort *) destination;
                for (i=0;i<n;i++) {
                   dst[i] = (((GLint) (rgba[i][RCOMP] * 31.0F)) << 11)
@@ -1307,9 +1343,27 @@ _mesa_pack_rgba_span( GLcontext *ctx,
                          | (((GLint) (rgba[i][ACOMP] *  1.0F))      );
                }
             }
+            else if (format == GL_BGRA) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][BCOMP] * 31.0F)) << 11)
+                         | (((GLint) (rgba[i][GCOMP] * 31.0F)) <<  6)
+                         | (((GLint) (rgba[i][RCOMP] * 31.0F)) <<  1)
+                         | (((GLint) (rgba[i][ACOMP] *  1.0F))      );
+               }
+            }
+            else if (format == GL_ABGR_EXT) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][ACOMP] * 31.0F)) << 11)
+                         | (((GLint) (rgba[i][BCOMP] * 31.0F)) <<  6)
+                         | (((GLint) (rgba[i][GCOMP] * 31.0F)) <<  1)
+                         | (((GLint) (rgba[i][RCOMP] *  1.0F))      );
+               }
+            }
             break;
          case GL_UNSIGNED_SHORT_1_5_5_5_REV:
-            if (format == GL_RGB) {
+            if (format == GL_RGBA) {
                GLushort *dst = (GLushort *) destination;
                for (i=0;i<n;i++) {
                   dst[i] = (((GLint) (rgba[i][RCOMP] * 31.0F))      )
@@ -1318,6 +1372,24 @@ _mesa_pack_rgba_span( GLcontext *ctx,
                          | (((GLint) (rgba[i][ACOMP] *  1.0F)) << 15);
                }
             }
+            else if (format == GL_BGRA) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][BCOMP] * 31.0F))      )
+                         | (((GLint) (rgba[i][GCOMP] * 31.0F)) <<  5)
+                         | (((GLint) (rgba[i][RCOMP] * 31.0F)) << 10)
+                         | (((GLint) (rgba[i][ACOMP] *  1.0F)) << 15);
+               }
+            }
+            else if (format == GL_ABGR_EXT) {
+               GLushort *dst = (GLushort *) destination;
+               for (i=0;i<n;i++) {
+                  dst[i] = (((GLint) (rgba[i][ACOMP] * 31.0F))      )
+                         | (((GLint) (rgba[i][BCOMP] * 31.0F)) <<  5)
+                         | (((GLint) (rgba[i][GCOMP] * 31.0F)) << 10)
+                         | (((GLint) (rgba[i][RCOMP] *  1.0F)) << 15);
+               }
+            }
             break;
          case GL_UNSIGNED_INT_8_8_8_8:
             if (format == GL_RGBA) {
@@ -2280,7 +2352,7 @@ _mesa_unpack_ubyte_color_span( GLcontext *ctx,
                               unpacking);
 
          if (applyTransferOps) {
-            if (ctx->Pixel.MapColorFlag) {
+            if (dstFormat == GL_COLOR_INDEX && ctx->Pixel.MapColorFlag) {
                _mesa_map_ci(ctx, n, indexes);
             }
             if (ctx->Pixel.IndexShift || ctx->Pixel.IndexOffset) {
@@ -2549,7 +2621,7 @@ _mesa_unpack_float_color_span( GLcontext *ctx,
                               unpacking);
 
          if (applyTransferOps) {
-            if (ctx->Pixel.MapColorFlag) {
+            if (dstFormat == GL_COLOR_INDEX && ctx->Pixel.MapColorFlag) {
                _mesa_map_ci(ctx, n, indexes);
             }
             if (ctx->Pixel.IndexShift || ctx->Pixel.IndexOffset) {
@@ -2943,7 +3015,7 @@ _mesa_unpack_depth_span( const GLcontext *ctx, GLuint n, GLdepth *dest,
                          const struct gl_pixelstore_attrib *unpacking,
                          GLboolean applyTransferOps )
 {
-   GLfloat *depth = MALLOC(n * sizeof(GLfloat));
+   GLfloat *depth = (GLfloat *) MALLOC(n * sizeof(GLfloat));
    if (!depth)
       return;
 
@@ -3074,7 +3146,7 @@ _mesa_unpack_image( GLsizei width, GLsizei height, GLsizei depth,
    }
 
    {
-      GLubyte *destBuffer = MALLOC(bytesPerRow * height * depth);
+      GLubyte *destBuffer = (GLubyte *) MALLOC(bytesPerRow * height * depth);
       GLubyte *dst;
       GLint img, row;
       if (!destBuffer)
@@ -3128,9 +3200,8 @@ _mesa_unpack_bitmap( GLint width, GLint height, const GLubyte *pixels,
    width_in_bytes = CEILING( width, 8 );
    dst = buffer;
    for (row = 0; row < height; row++) {
-      GLubyte *src = _mesa_image_address( packing, pixels, width, height,
-                                          GL_COLOR_INDEX, GL_BITMAP,
-                                          0, row, 0 );
+      GLubyte *src = (GLubyte *) _mesa_image_address(packing, pixels, width,
+                                 height, GL_COLOR_INDEX, GL_BITMAP, 0, row, 0);
       if (!src) {
          FREE(buffer);
          return NULL;
@@ -3223,7 +3294,7 @@ _mesa_pack_bitmap( GLint width, GLint height, const GLubyte *source,
    width_in_bytes = CEILING( width, 8 );
    src = source;
    for (row = 0; row < height; row++) {
-      GLubyte *dst = _mesa_image_address( packing, dest, width, height,
+      GLubyte *dst = (GLubyte *) _mesa_image_address( packing, dest, width, height,
                                           GL_COLOR_INDEX, GL_BITMAP,
                                           0, row, 0 );
       if (!dst)
diff --git a/xc/extras/Mesa/src/imaging.c b/xc/extras/Mesa/src/imaging.c
index 5d2250cbc..27fb58a6a 100644
--- a/xc/extras/Mesa/src/imaging.c
+++ b/xc/extras/Mesa/src/imaging.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/imaging.c,v 1.6 2000/09/26 15:56:32 tsi Exp $ */
+
 
 /*
  * Histogram, Min/max and convolution for GL_ARB_imaging subset
@@ -210,7 +210,7 @@ pack_minmax( GLcontext *ctx, CONST GLfloat minmax[2][4],
              const struct gl_pixelstore_attrib *packing )
 {
    const GLint comps = _mesa_components_in_format(format);
-   GLuint luminance[2];
+   GLfloat luminance[2];
 
    if (format == GL_LUMINANCE || format == GL_LUMINANCE_ALPHA) {
       luminance[0] = minmax[0][RCOMP] + minmax[0][GCOMP] + minmax[0][BCOMP];
diff --git a/xc/extras/Mesa/src/light.c b/xc/extras/Mesa/src/light.c
index 9d73e6814..138aefb08 100644
--- a/xc/extras/Mesa/src/light.c
+++ b/xc/extras/Mesa/src/light.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/light.c,v 1.5 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/light.h b/xc/extras/Mesa/src/light.h
index 1a4fa4174..8edcac801 100644
--- a/xc/extras/Mesa/src/light.h
+++ b/xc/extras/Mesa/src/light.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/light.h,v 1.5 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifndef LIGHT_H
 #define LIGHT_H
diff --git a/xc/extras/Mesa/src/lines.c b/xc/extras/Mesa/src/lines.c
index 24b9d5feb..ffbbad416 100644
--- a/xc/extras/Mesa/src/lines.c
+++ b/xc/extras/Mesa/src/lines.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/lines.c,v 1.6 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -1057,7 +1057,7 @@ _mesa_print_line_function(GLcontext *ctx)
    else if (ctx->Driver.LineFunc == null_line)
       printf("null_line\n");
    else
-      printf("Driver func %p\n", ctx->Driver.PointsFunc);
+      printf("Driver func %p\n", ctx->Driver.LineFunc);
 }
 #endif
 
diff --git a/xc/extras/Mesa/src/linetemp.h b/xc/extras/Mesa/src/linetemp.h
index ab0f6657e..2c4eac4b0 100644
--- a/xc/extras/Mesa/src/linetemp.h
+++ b/xc/extras/Mesa/src/linetemp.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/linetemp.h,v 1.9 2000/09/24 13:50:12 alanh Exp $ */
+
 
 /*
  * Line Rasterizer Template
@@ -227,8 +227,8 @@
       z1 = FloatToFixed(VB->Win.data[vert1][2] + ctx->LineZoffset);
    }
    else {
-      z0 = (int) VB->Win.data[vert0][2] + ctx->LineZoffset;
-      z1 = (int) VB->Win.data[vert1][2] + ctx->LineZoffset;
+      z0 = (GLint) (VB->Win.data[vert0][2] + ctx->LineZoffset);
+      z1 = (GLint) (VB->Win.data[vert1][2] + ctx->LineZoffset);
    }
 #endif
 #ifdef PIXEL_ADDRESS
diff --git a/xc/extras/Mesa/src/lnaatemp.h b/xc/extras/Mesa/src/lnaatemp.h
index 872fe09b9..9cde169e3 100644
--- a/xc/extras/Mesa/src/lnaatemp.h
+++ b/xc/extras/Mesa/src/lnaatemp.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/lnaatemp.h,v 1.11 2000/09/24 13:50:12 alanh Exp $ */
 
 /*
  * Antialiased Line Rasterizer Template
@@ -107,8 +106,8 @@
       z1 = FloatToFixed(VB->Win.data[vert1][2] + ctx->LineZoffset);
    }
    else {
-      z0 = (int) VB->Win.data[vert0][2] + ctx->LineZoffset;
-      z1 = (int) VB->Win.data[vert1][2] + ctx->LineZoffset;
+      z0 = (GLint) (VB->Win.data[vert0][2] + ctx->LineZoffset);
+      z1 = (GLint) (VB->Win.data[vert1][2] + ctx->LineZoffset);
    }
 
 #ifdef INTERP_STUV0
diff --git a/xc/extras/Mesa/src/logic.c b/xc/extras/Mesa/src/logic.c
index 2a4bd3aed..872cb48d6 100644
--- a/xc/extras/Mesa/src/logic.c
+++ b/xc/extras/Mesa/src/logic.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/logic.c,v 1.6 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/matrix.c b/xc/extras/Mesa/src/matrix.c
index 3dcc61ad9..7fc3e9404 100644
--- a/xc/extras/Mesa/src/matrix.c
+++ b/xc/extras/Mesa/src/matrix.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/matrix.c,v 1.7 2000/09/26 15:56:32 tsi Exp $ */
+
 
 /*
  * Matrix operations
@@ -763,18 +763,16 @@ static void analyze_from_scratch( GLmatrix *mat )
 
    /* Do the real work
     */
-   if (mask == MASK_IDENTITY) {
+   if (mask == (GLuint) MASK_IDENTITY) {
       mat->type = MATRIX_IDENTITY;
    }
-   else if ((mask & MASK_2D_NO_ROT) == MASK_2D_NO_ROT)  
-   {
+   else if ((mask & MASK_2D_NO_ROT) == (GLuint) MASK_2D_NO_ROT) {
       mat->type = MATRIX_2D_NO_ROT;
       
       if ((mask & MASK_NO_2D_SCALE) != MASK_NO_2D_SCALE)
 	 mat->flags = MAT_FLAG_GENERAL_SCALE;
    }
-   else if ((mask & MASK_2D) == MASK_2D)  
-   {
+   else if ((mask & MASK_2D) == (GLuint) MASK_2D) {
       GLfloat mm = DOT2(m, m);
       GLfloat m4m4 = DOT2(m+4,m+4);
       GLfloat mm4 = DOT2(m,m+4);
@@ -793,8 +791,7 @@ static void analyze_from_scratch( GLmatrix *mat )
 	 mat->flags |= MAT_FLAG_ROTATION;
 
    }
-   else if ((mask & MASK_3D_NO_ROT) == MASK_3D_NO_ROT)
-   {
+   else if ((mask & MASK_3D_NO_ROT) == (GLuint) MASK_3D_NO_ROT) {
       mat->type = MATRIX_3D_NO_ROT;
 
       /* Check for scale */
@@ -805,8 +802,7 @@ static void analyze_from_scratch( GLmatrix *mat )
       } else
 	 mat->flags |= MAT_FLAG_GENERAL_SCALE;
    }
-   else if ((mask & MASK_3D) == MASK_3D)
-   {
+   else if ((mask & MASK_3D) == (GLuint) MASK_3D) {
       GLfloat c1 = DOT3(m,m);
       GLfloat c2 = DOT3(m+4,m+4);
       GLfloat c3 = DOT3(m+8,m+8);
diff --git a/xc/extras/Mesa/src/matrix.h b/xc/extras/Mesa/src/matrix.h
index 2da7687c9..6d989abd1 100644
--- a/xc/extras/Mesa/src/matrix.h
+++ b/xc/extras/Mesa/src/matrix.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/matrix.h,v 1.5 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifndef MATRIX_H
 #define MATRIX_H
diff --git a/xc/extras/Mesa/src/mem.c b/xc/extras/Mesa/src/mem.c
index 9e6a5c2b1..9334b63c4 100644
--- a/xc/extras/Mesa/src/mem.c
+++ b/xc/extras/Mesa/src/mem.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/mem.c,v 1.4 2000/09/26 15:56:32 tsi Exp $ */
+
 
 /*
  * Memory allocation functions.  Called via the MALLOC, CALLOC and
diff --git a/xc/extras/Mesa/src/mem.h b/xc/extras/Mesa/src/mem.h
index e9b1f3136..8935cb1aa 100644
--- a/xc/extras/Mesa/src/mem.h
+++ b/xc/extras/Mesa/src/mem.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/mem.h,v 1.5 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifndef MEM_H
 #define MEM_H
diff --git a/xc/extras/Mesa/src/mmath.h b/xc/extras/Mesa/src/mmath.h
index dae66eac6..fc34fcb08 100644
--- a/xc/extras/Mesa/src/mmath.h
+++ b/xc/extras/Mesa/src/mmath.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/mmath.h,v 1.7 2000/09/24 13:50:14 alanh Exp $ */
+
 
 /*
  * Faster arithmetic functions.  If the FAST_MATH preprocessor symbol is
@@ -158,7 +158,7 @@ static __inline__ int FloatToInt(float f)
    __asm__ ("fistpl %0" : "=m" (r) : "t" (f) : "st");
    return r;
 }
-#elif  defined(__MSC__) && defined(__WIN32__) && !defined(__CYGWIN__)
+#elif  defined(__MSC__) && defined(__WIN32__)
 static __inline int FloatToInt(float f)
 {
    int r;
diff --git a/xc/extras/Mesa/src/pb.c b/xc/extras/Mesa/src/pb.c
index 454882d5b..8cc67e0ea 100644
--- a/xc/extras/Mesa/src/pb.c
+++ b/xc/extras/Mesa/src/pb.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/pb.c,v 1.6 2000/09/26 15:56:32 tsi Exp $ */
+
 
 
 /*
@@ -181,10 +181,10 @@ static void multi_write_rgba_pixels( GLcontext *ctx, GLuint n,
          }
 
          (*ctx->Driver.WriteRGBAPixels)( ctx, n, x, y, 
-					 (const GLubyte (*)[4])rgbaTmp, mask );
+					 (CONST GLubyte (*)[4])rgbaTmp, mask );
          if (ctx->RasterMask & ALPHABUF_BIT) {
             _mesa_write_alpha_pixels( ctx, n, x, y, 
-                                      (const GLubyte (*)[4])rgbaTmp, mask );
+                                      (CONST GLubyte (*)[4])rgbaTmp, mask );
          }
       }
    }
@@ -271,7 +271,7 @@ void gl_flush_pb( GLcontext *ctx )
          if (ctx->Light.Model.ColorControl == GL_SEPARATE_SPECULAR_COLOR
              && ctx->Light.Enabled && ctx->Texture.ReallyEnabled) {
             /* add specular color to primary color */
-            add_colors( PB->count, PB->rgba, (const GLubyte (*)[3]) PB->spec );
+            add_colors( PB->count, PB->rgba, (CONST GLubyte (*)[3]) PB->spec );
          }
 
 	 if (ctx->Fog.Enabled
@@ -284,7 +284,7 @@ void gl_flush_pb( GLcontext *ctx )
 
 	 if (ctx->Color.AlphaEnabled) {
 	    if (_mesa_alpha_test( ctx, PB->count, 
-                                  (const GLubyte (*)[4]) PB->rgba, mask )==0) {
+                                  (CONST GLubyte (*)[4]) PB->rgba, mask )==0) {
 	       goto CleanUp;
 	    }
 	 }
@@ -304,7 +304,7 @@ void gl_flush_pb( GLcontext *ctx )
          
          if (ctx->RasterMask & MULTI_DRAW_BIT) {
             multi_write_rgba_pixels( ctx, PB->count, PB->x, PB->y,
-                                     (const GLubyte (*)[4])PB->rgba, mask );
+                                     (CONST GLubyte (*)[4])PB->rgba, mask );
          }
          else {
             /* normal case: write to exactly one buffer */
@@ -321,11 +321,11 @@ void gl_flush_pb( GLcontext *ctx )
             }
 
             (*ctx->Driver.WriteRGBAPixels)( ctx, PB->count, PB->x, PB->y,
-                                            (const GLubyte (*)[4]) PB->rgba,
+                                            (CONST GLubyte (*)[4]) PB->rgba,
 					    mask );
             if (ctx->RasterMask & ALPHABUF_BIT) {
                _mesa_write_alpha_pixels( ctx, PB->count, PB->x, PB->y, 
-				      (const GLubyte (*)[4]) PB->rgba, mask );
+				      (CONST GLubyte (*)[4]) PB->rgba, mask );
             }
          }
       }
@@ -336,7 +336,7 @@ void gl_flush_pb( GLcontext *ctx )
 
 	 if (ctx->Color.AlphaEnabled) {
 	    if (_mesa_alpha_test( ctx, PB->count, 
-                                  (const GLubyte (*)[4]) PB->rgba, mask )==0) {
+                                  (CONST GLubyte (*)[4]) PB->rgba, mask )==0) {
 	       goto CleanUp;
 	    }
 	 }
@@ -360,7 +360,7 @@ void gl_flush_pb( GLcontext *ctx )
          if (ctx->RasterMask & MULTI_DRAW_BIT) {
             /* Copy mono color to all pixels */
             multi_write_rgba_pixels( ctx, PB->count, PB->x, PB->y,
-                                     (const GLubyte (*)[4]) PB->rgba, mask );
+                                     (CONST GLubyte (*)[4]) PB->rgba, mask );
          }
          else {
             /* normal case: write to exactly one buffer */
diff --git a/xc/extras/Mesa/src/pipeline.c b/xc/extras/Mesa/src/pipeline.c
index f1b4375f9..0a469894d 100644
--- a/xc/extras/Mesa/src/pipeline.c
+++ b/xc/extras/Mesa/src/pipeline.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -414,7 +414,7 @@ void gl_build_immediate_pipeline( GLcontext *ctx )
       gl_print_pipeline( ctx, elt ); 
 }
    
-#define INTERESTED ~(NEW_DRIVER_STATE|NEW_CLIENT_STATE|NEW_TEXTURE_ENABLE)
+#define INTERESTED ~(NEW_DRIVER_STATE|NEW_CLIENT_STATE)
 
 void gl_update_pipelines( GLcontext *ctx )
 {
@@ -438,7 +438,7 @@ void gl_update_pipelines( GLcontext *ctx )
       else 
 	 flags |= VERT_INDEX;
 
-      if (ctx->Texture.Enabled & 0xf) {
+      if (ctx->Texture.ReallyEnabled & 0xf) {
 /* XXX this should also check that the texture is RGBA.  What about Unit[1]?
 	 if (ctx->Texture.Unit[0].EnvMode == GL_REPLACE)
 	    flags &= ~VERT_RGBA;
@@ -446,7 +446,7 @@ void gl_update_pipelines( GLcontext *ctx )
 	 flags |= VERT_TEX0_ANY;
       }
 
-      if (ctx->Texture.Enabled & 0xf0)
+      if (ctx->Texture.ReallyEnabled & 0xf0)
 	 flags |= VERT_TEX1_ANY;
    
       if (ctx->Polygon.Unfilled) 
@@ -489,7 +489,7 @@ void gl_run_pipeline( struct vertex_buffer *VB )
 
    START_FAST_MATH(x);
    
-   for ( VB->Culled = 0; *stages && !VB->Culled ; stages++ ) 
+   for ( VB->Culled = 0; *stages && !VB->Culled ; stages++ )
       (*stages)->run( VB );
       
    END_FAST_MATH(x);
@@ -524,7 +524,7 @@ void gl_print_vert_flags( const char *name, GLuint flags )
 void gl_print_tri_caps( const char *name, GLuint flags ) 
 {
    fprintf(stderr, 
-	   "%s: (0x%x) %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
+	   "%s: (0x%x) %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n",
 	   name,
 	   flags,
 	   (flags & DD_FEEDBACK)            ? "feedback, " : "",
@@ -548,7 +548,9 @@ void gl_print_tri_caps( const char *name, GLuint flags )
 	   (flags & DD_LINE_SW_RASTERIZE)   ? "sw-lines, " : "", 
 	   (flags & DD_TRI_SW_RASTERIZE)    ? "sw-tris, " : "", 
 	   (flags & DD_QUAD_SW_RASTERIZE)   ? "sw-quads, " : "",
-	   (flags & DD_TRI_CULL_FRONT_BACK) ? "cull-all, " : ""
+	   (flags & DD_TRI_CULL_FRONT_BACK) ? "cull-all, " : "",
+	   (flags & DD_STENCIL)             ? "stencil, " : "",
+	   (flags & DD_CLIP_FOG_COORD)      ? "clip-fog-coord, " : ""
       );
 }
 
diff --git a/xc/extras/Mesa/src/pixel.c b/xc/extras/Mesa/src/pixel.c
index 17c6eb6bb..7725f8ddb 100644
--- a/xc/extras/Mesa/src/pixel.c
+++ b/xc/extras/Mesa/src/pixel.c
@@ -104,6 +104,14 @@ _mesa_PixelStorei( GLenum pname, GLint param )
 	    ctx->Pack.SkipRows = param;
 	 }
 	 break;
+      case GL_PACK_SKIP_IMAGES:
+	 if (param<0) {
+	    gl_error( ctx, GL_INVALID_VALUE, "glPixelStore(param)" );
+	 }
+	 else {
+	    ctx->Pack.SkipImages = param;
+	 }
+	 break;
       case GL_PACK_ALIGNMENT:
          if (param==1 || param==2 || param==4 || param==8) {
 	    ctx->Pack.Alignment = param;
@@ -148,6 +156,14 @@ _mesa_PixelStorei( GLenum pname, GLint param )
 	    ctx->Unpack.SkipRows = param;
 	 }
 	 break;
+      case GL_UNPACK_SKIP_IMAGES:
+	 if (param < 0) {
+	    gl_error( ctx, GL_INVALID_VALUE, "glPixelStore(param)" );
+	 }
+	 else {
+	    ctx->Unpack.SkipImages = param;
+	 }
+	 break;
       case GL_UNPACK_ALIGNMENT:
          if (param==1 || param==2 || param==4 || param==8) {
 	    ctx->Unpack.Alignment = param;
diff --git a/xc/extras/Mesa/src/points.c b/xc/extras/Mesa/src/points.c
index cd73dd1c3..c912d20df 100644
--- a/xc/extras/Mesa/src/points.c
+++ b/xc/extras/Mesa/src/points.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/points.c,v 1.6 2000/09/26 15:56:32 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -160,7 +160,7 @@ size1_ci_points( GLcontext *ctx, GLuint first, GLuint last )
    GLuint i;
 
    win = &VB->Win.data[first][0];
-   for (i = first; i <= last; i++) {
+   for (i = first; i < last; i++) {
       if (VB->ClipMask[i] == 0) {
          pbx[pbcount] = (GLint)  win[0];
          pby[pbcount] = (GLint)  win[1];
@@ -186,7 +186,7 @@ size1_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    struct pixel_buffer *PB = ctx->PB;
    GLuint i;
 
-   for (i = first; i <= last; i++) {
+   for (i = first; i < last; i++) {
       if (VB->ClipMask[i] == 0) {
          GLint x, y, z;
          GLint red, green, blue, alpha;
@@ -220,7 +220,7 @@ general_ci_points( GLcontext *ctx, GLuint first, GLuint last )
    GLint radius = isize >> 1;
    GLuint i;
 
-   for (i = first; i <= last; i++) {
+   for (i = first; i < last; i++) {
       if (VB->ClipMask[i] == 0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy;
@@ -269,7 +269,7 @@ general_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    GLint radius = isize >> 1;
    GLuint i;
 
-   for (i = first; i <= last; i++) {
+   for (i = first; i < last; i++) {
       if (VB->ClipMask[i] == 0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy;
@@ -322,7 +322,7 @@ textured_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    struct pixel_buffer *PB = ctx->PB;
    GLuint i;
 
-   for (i = first; i <= last; i++) {
+   for (i = first; i < last; i++) {
       if (VB->ClipMask[i] == 0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy, radius;
@@ -409,7 +409,7 @@ multitextured_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    struct pixel_buffer *PB = ctx->PB;
    GLuint i;
 
-   for (i = first; i <= last; i++) {
+   for (i = first; i < last; i++) {
       if (VB->ClipMask[i] == 0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy;
@@ -550,7 +550,7 @@ antialiased_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    GLuint i;
 
    if (ctx->Texture.ReallyEnabled) {
-      for (i = first; i <= last; i++) {
+      for (i = first; i < last; i++) {
          if (VB->ClipMask[i] == 0) {
             GLint x, y;
             GLint red, green, blue, alpha;
@@ -666,7 +666,7 @@ antialiased_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    }
    else {
       /* Not texture mapped */
-      for (i=first;i<=last;i++) {
+      for (i=first;i<last;i++) {
          if (VB->ClipMask[i]==0) {
             GLint xmin, ymin, xmax, ymax;
             GLint x, y, z;
@@ -737,7 +737,7 @@ dist3(GLfloat *out, GLuint first, GLuint last,
    const GLfloat *p = VEC_ELT(v, GLfloat, first);
    GLuint i;
 
-   for (i = first ; i <= last ; i++, STRIDE_F(p, stride) ) {
+   for (i = first ; i < last ; i++, STRIDE_F(p, stride) ) {
       GLfloat dist = GL_SQRT(p[0]*p[0]+p[1]*p[1]+p[2]*p[2]);
       out[i] = 1.0F / (ctx->Point.Params[0] +
                        dist * (ctx->Point.Params[1] +
@@ -754,7 +754,7 @@ dist2(GLfloat *out, GLuint first, GLuint last,
    const GLfloat *p = VEC_ELT(v, GLfloat, first);
    GLuint i;
 
-   for (i = first ; i <= last ; i++, STRIDE_F(p, stride) ) {
+   for (i = first ; i < last ; i++, STRIDE_F(p, stride) ) {
       GLfloat dist = GL_SQRT(p[0]*p[0]+p[1]*p[1]);
       out[i] = 1.0F / (ctx->Point.Params[0] +
                        dist * (ctx->Point.Params[1] +
@@ -794,7 +794,7 @@ clip_dist(GLfloat *out, GLuint first, GLuint last,
    const GLfloat *from = (GLfloat *)clip_vec->start;
    const GLuint stride = clip_vec->stride;
 
-   for (i = first ; i <= last ; i++ )
+   for (i = first ; i < last ; i++ )
    {
       GLfloat dist = win[i][2];
       out[i] = 1/(ctx->Point.Params[0]+ 
@@ -823,7 +823,7 @@ dist_atten_general_ci_points( GLcontext *ctx, GLuint first, GLuint last )
    else 
       clip_dist( dist, first, last, ctx, VB->ClipPtr );
 
-   for (i=first;i<=last;i++) {
+   for (i=first;i<last;i++) {
       if (VB->ClipMask[i]==0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy;
@@ -885,7 +885,7 @@ dist_atten_general_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    else 
       clip_dist( dist, first, last, ctx, VB->ClipPtr );
 
-   for (i=first;i<=last;i++) {
+   for (i=first;i<last;i++) {
       if (VB->ClipMask[i]==0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy;
@@ -955,7 +955,7 @@ dist_atten_textured_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    else 
       clip_dist( dist, first, last, ctx, VB->ClipPtr );
 
-   for (i=first;i<=last;i++) {
+   for (i=first;i<last;i++) {
       if (VB->ClipMask[i]==0) {
          GLint x0, x1, y0, y1;
          GLint ix, iy;
@@ -1103,7 +1103,7 @@ dist_atten_antialiased_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
       clip_dist( dist, first, last, ctx, VB->ClipPtr );
 
    if (ctx->Texture.ReallyEnabled) {
-      for (i=first;i<=last;i++) {
+      for (i=first;i<last;i++) {
          if (VB->ClipMask[i]==0) {
             GLfloat radius, rmin, rmax, rmin2, rmax2, cscale, alphaf;
             GLint xmin, ymin, xmax, ymax;
@@ -1231,7 +1231,7 @@ dist_atten_antialiased_rgba_points( GLcontext *ctx, GLuint first, GLuint last )
    }
    else {
       /* Not texture mapped */
-      for (i = first; i <= last; i++) {
+      for (i = first; i < last; i++) {
          if (VB->ClipMask[i] == 0) {
             GLfloat radius, rmin, rmax, rmin2, rmax2, cscale, alphaf;
             GLint xmin, ymin, xmax, ymax;
diff --git a/xc/extras/Mesa/src/readpix.c b/xc/extras/Mesa/src/readpix.c
index 4e5e11985..96a3f11b2 100644
--- a/xc/extras/Mesa/src/readpix.c
+++ b/xc/extras/Mesa/src/readpix.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/readpix.c,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -655,7 +655,7 @@ static void read_rgba_pixels( GLcontext *ctx,
          dest = _mesa_image_address( packing, pixels, width, height,
                                      format, type, 0, j, 0);
 
-         _mesa_pack_rgba_span( ctx, readWidth, (const GLubyte (*)[4]) rgba,
+         _mesa_pack_rgba_span( ctx, readWidth, (CONST GLubyte (*)[4]) rgba,
                                format, type, dest, packing, GL_TRUE );
       }
    }
@@ -678,7 +678,7 @@ static void read_rgba_pixels( GLcontext *ctx,
          dest = _mesa_image_address( packing, pixels, width, height,
                                      format, type, 0, j, 0);
 
-         _mesa_pack_rgba_span( ctx, readWidth, (const GLubyte (*)[4]) rgba,
+         _mesa_pack_rgba_span( ctx, readWidth, (CONST GLubyte (*)[4]) rgba,
                                format, type, dest, packing, GL_TRUE );
       }
    }
@@ -705,6 +705,8 @@ _mesa_ReadPixels( GLint x, GLint y, GLsizei width, GLsizei height,
                                  format, type, &ctx->Pack, pixels))
       return;
 
+   RENDER_START(ctx);
+
    switch (format) {
       case GL_COLOR_INDEX:
          read_index_pixels(ctx, x, y, width, height, type, pixels, &ctx->Pack);
@@ -732,4 +734,6 @@ _mesa_ReadPixels( GLint x, GLint y, GLsizei width, GLsizei height,
       default:
 	 gl_error( ctx, GL_INVALID_ENUM, "glReadPixels(format)" );
    }
+
+   RENDER_FINISH(ctx);
 }
diff --git a/xc/extras/Mesa/src/rect.c b/xc/extras/Mesa/src/rect.c
index 5a54a072b..502498c01 100644
--- a/xc/extras/Mesa/src/rect.c
+++ b/xc/extras/Mesa/src/rect.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/rect.c,v 1.5 2000/09/26 15:56:33 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/render_tmp.h b/xc/extras/Mesa/src/render_tmp.h
index b7a23b056..bcc60d42b 100644
--- a/xc/extras/Mesa/src/render_tmp.h
+++ b/xc/extras/Mesa/src/render_tmp.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/render_tmp.h,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
diff --git a/xc/extras/Mesa/src/shade.c b/xc/extras/Mesa/src/shade.c
index 0d794dcdf..e2516ece7 100644
--- a/xc/extras/Mesa/src/shade.c
+++ b/xc/extras/Mesa/src/shade.c
@@ -22,7 +22,8 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/shade.c,v 1.5 2000/09/26 15:56:33 tsi Exp $ */
+
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -203,7 +204,7 @@ void gl_update_lighting_function( GLcontext *ctx )
 
    if (ctx->Visual->RGBAflag) {
       if (ctx->Light.NeedVertices) {
-	 if (ctx->Texture.Enabled && 
+	 if (ctx->Texture.ReallyEnabled && 
 	     ctx->Light.Model.ColorControl==GL_SEPARATE_SPECULAR_COLOR) 
 	    tab = gl_shade_spec_tab;
 	 else
diff --git a/xc/extras/Mesa/src/shade_tmp.h b/xc/extras/Mesa/src/shade_tmp.h
index 72a7f2d82..f6ec08cb7 100644
--- a/xc/extras/Mesa/src/shade_tmp.h
+++ b/xc/extras/Mesa/src/shade_tmp.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/shade_tmp.h,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
diff --git a/xc/extras/Mesa/src/span.c b/xc/extras/Mesa/src/span.c
index 3a3ba0633..a86489f03 100644
--- a/xc/extras/Mesa/src/span.c
+++ b/xc/extras/Mesa/src/span.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/span.c,v 1.7 2000/09/26 15:56:33 tsi Exp $ */
+
 
 /*
  * pixel span rasterization:
@@ -407,10 +407,10 @@ static void multi_write_rgba_span( GLcontext *ctx, GLuint n,
          }
 
          (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, 
-				       (const GLubyte (*)[4]) rgbaTmp, mask );
+				       (CONST GLubyte (*)[4]) rgbaTmp, mask );
          if (ctx->RasterMask & ALPHABUF_BIT) {
             _mesa_write_alpha_span( ctx, n, x, y, 
-                                    (const GLubyte (*)[4])rgbaTmp, mask );
+                                    (CONST GLubyte (*)[4])rgbaTmp, mask );
          }
       }
    }
@@ -475,7 +475,7 @@ void gl_write_rgba_span( GLcontext *ctx,
 
    /* Do the alpha test */
    if (ctx->Color.AlphaEnabled) {
-      if (_mesa_alpha_test( ctx, n, (const GLubyte (*)[4]) rgba, mask )==0) {
+      if (_mesa_alpha_test( ctx, n, (CONST GLubyte (*)[4]) rgba, mask )==0) {
 	 return;
       }
       write_all = GL_FALSE;
@@ -503,7 +503,7 @@ void gl_write_rgba_span( GLcontext *ctx,
    ctx->OcclusionResult = GL_TRUE;
 
    if (ctx->RasterMask & MULTI_DRAW_BIT) {
-      multi_write_rgba_span( ctx, n, x, y, (const GLubyte (*)[4]) rgba, mask );
+      multi_write_rgba_span( ctx, n, x, y, (CONST GLubyte (*)[4]) rgba, mask );
    }
    else {
       /* normal: write to exactly one buffer */
@@ -524,12 +524,12 @@ void gl_write_rgba_span( GLcontext *ctx,
 
       /* write pixels */
       (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, 
-				    (const GLubyte (*)[4]) rgba, 
+				    (CONST GLubyte (*)[4]) rgba, 
 				    write_all ? Null : mask );
 
       if (ctx->RasterMask & ALPHABUF_BIT) {
          _mesa_write_alpha_span( ctx, n, x, y, 
-                                 (const GLubyte (*)[4]) rgba, 
+                                 (CONST GLubyte (*)[4]) rgba, 
                                  write_all ? Null : mask );
       }
 
@@ -588,7 +588,7 @@ void gl_write_monocolor_span( GLcontext *ctx,
       for (i=0;i<n;i++) {
          rgba[i][ACOMP] = color[ACOMP];
       }
-      if (_mesa_alpha_test( ctx, n, (const GLubyte (*)[4])rgba, mask )==0) {
+      if (_mesa_alpha_test( ctx, n, (CONST GLubyte (*)[4])rgba, mask )==0) {
 	 return;
       }
       write_all = GL_FALSE;
@@ -637,7 +637,7 @@ void gl_write_monocolor_span( GLcontext *ctx,
 
       if (ctx->RasterMask & MULTI_DRAW_BIT) {
          multi_write_rgba_span( ctx, n, x, y,
-                                (const GLubyte (*)[4]) rgba, mask );
+                                (CONST GLubyte (*)[4]) rgba, mask );
       }
       else {
          /* normal: write to exactly one buffer */
@@ -657,11 +657,11 @@ void gl_write_monocolor_span( GLcontext *ctx,
 
          /* write pixels */
          (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, 
-				       (const GLubyte (*)[4]) rgba, 
+				       (CONST GLubyte (*)[4]) rgba, 
 				       write_all ? Null : mask );
          if (ctx->RasterMask & ALPHABUF_BIT) {
             _mesa_write_alpha_span( ctx, n, x, y, 
-                                    (const GLubyte (*)[4]) rgba, 
+                                    (CONST GLubyte (*)[4]) rgba, 
                                     write_all ? Null : mask );
          }
       }
@@ -679,7 +679,7 @@ void gl_write_monocolor_span( GLcontext *ctx,
             }
          }
          multi_write_rgba_span( ctx, n, x, y, 
-				(const GLubyte (*)[4]) rgba, mask );
+				(CONST GLubyte (*)[4]) rgba, mask );
       }
       else {
          (*ctx->Driver.WriteMonoRGBASpan)( ctx, n, x, y, mask );
@@ -787,7 +787,7 @@ void gl_write_texture_span( GLcontext *ctx,
 
    /* Do the alpha test */
    if (ctx->Color.AlphaEnabled) {
-      if (_mesa_alpha_test( ctx, n, (const GLubyte (*)[4]) rgba, mask )==0) {
+      if (_mesa_alpha_test( ctx, n, (CONST GLubyte (*)[4]) rgba, mask )==0) {
 	 return;
       }
       write_all = GL_FALSE;
@@ -815,7 +815,7 @@ void gl_write_texture_span( GLcontext *ctx,
    ctx->OcclusionResult = GL_TRUE;
 
    if (ctx->RasterMask & MULTI_DRAW_BIT) {
-      multi_write_rgba_span( ctx, n, x, y, (const GLubyte (*)[4]) rgba, mask );
+      multi_write_rgba_span( ctx, n, x, y, (CONST GLubyte (*)[4]) rgba, mask );
    }
    else {
       /* normal: write to exactly one buffer */
@@ -831,10 +831,10 @@ void gl_write_texture_span( GLcontext *ctx,
          _mesa_mask_rgba_span( ctx, n, x, y, rgba );
       }
 
-      (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, (const GLubyte (*)[4])rgba,
+      (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, (CONST GLubyte (*)[4])rgba,
 				    write_all ? Null : mask );
       if (ctx->RasterMask & ALPHABUF_BIT) {
-         _mesa_write_alpha_span( ctx, n, x, y, (const GLubyte (*)[4]) rgba, 
+         _mesa_write_alpha_span( ctx, n, x, y, (CONST GLubyte (*)[4]) rgba, 
                                  write_all ? Null : mask );
       }
    }
@@ -918,7 +918,7 @@ gl_write_multitexture_span( GLcontext *ctx, GLuint texUnits,
 
    /* Do the alpha test */
    if (ctx->Color.AlphaEnabled) {
-      if (_mesa_alpha_test( ctx, n, (const GLubyte (*)[4])rgba, mask )==0) {
+      if (_mesa_alpha_test( ctx, n, (CONST GLubyte (*)[4])rgba, mask )==0) {
 	 return;
       }
       write_all = GL_FALSE;
@@ -946,7 +946,7 @@ gl_write_multitexture_span( GLcontext *ctx, GLuint texUnits,
    ctx->OcclusionResult = GL_TRUE;
 
    if (ctx->RasterMask & MULTI_DRAW_BIT) {
-      multi_write_rgba_span( ctx, n, x, y, (const GLubyte (*)[4]) rgba, mask );
+      multi_write_rgba_span( ctx, n, x, y, (CONST GLubyte (*)[4]) rgba, mask );
    }
    else {
       /* normal: write to exactly one buffer */
@@ -963,9 +963,9 @@ gl_write_multitexture_span( GLcontext *ctx, GLuint texUnits,
          _mesa_mask_rgba_span( ctx, n, x, y, rgba );
       }
 
-      (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, (const GLubyte (*)[4])rgba, write_all ? Null : mask );
+      (*ctx->Driver.WriteRGBASpan)( ctx, n, x, y, (CONST GLubyte (*)[4])rgba, write_all ? Null : mask );
       if (ctx->RasterMask & ALPHABUF_BIT) {
-         _mesa_write_alpha_span( ctx, n, x, y, (const GLubyte (*)[4])rgba,
+         _mesa_write_alpha_span( ctx, n, x, y, (CONST GLubyte (*)[4])rgba,
                                  write_all ? Null : mask );
       }
    }
diff --git a/xc/extras/Mesa/src/stages.c b/xc/extras/Mesa/src/stages.c
index 859aadac5..adf614270 100644
--- a/xc/extras/Mesa/src/stages.c
+++ b/xc/extras/Mesa/src/stages.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  *
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/stages.c,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -457,7 +457,7 @@ static void do_lighting( struct vertex_buffer *VB )
    /* Make sure we can talk about elements 0..2 in the vector we are
     * lighting.
     */
-   if (VB->Unprojected->size == 2) {
+   if (VB->Unprojected && VB->Unprojected->size == 2) {
       if (VB->Unprojected->flags & VEC_WRITABLE)
 	 gl_vector4f_clean_elem(VB->Unprojected, VB->Count, 2);
       else 
@@ -774,8 +774,8 @@ CONST struct gl_pipeline_stage gl_default_pipeline[] = {
      PIPE_OP_TEX0,
      PIPE_PRECALC|PIPE_IMMEDIATE,
      0,
-     NEW_TEXTURING|NEW_TEXTURE_MATRIX,
-     NEW_TEXTURING|NEW_TEXTURE_MATRIX,
+     NEW_TEXTURING|NEW_TEXTURE_MATRIX|NEW_TEXTURE_ENABLE,
+     NEW_TEXTURING|NEW_TEXTURE_MATRIX|NEW_TEXTURE_ENABLE,
      0, 0,
      DYN_STATE,
      check_texture_0,
@@ -785,8 +785,8 @@ CONST struct gl_pipeline_stage gl_default_pipeline[] = {
      PIPE_OP_TEX1,
      PIPE_PRECALC|PIPE_IMMEDIATE,
      0,
-     NEW_TEXTURING|NEW_TEXTURE_MATRIX,
-     NEW_TEXTURING|NEW_TEXTURE_MATRIX,
+     NEW_TEXTURING|NEW_TEXTURE_MATRIX|NEW_TEXTURE_ENABLE,
+     NEW_TEXTURING|NEW_TEXTURE_MATRIX|NEW_TEXTURE_ENABLE,
      0, 0,
      DYN_STATE,
      check_texture_1,
diff --git a/xc/extras/Mesa/src/state.c b/xc/extras/Mesa/src/state.c
index e111b18db..4c910aa5b 100644
--- a/xc/extras/Mesa/src/state.c
+++ b/xc/extras/Mesa/src/state.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  *
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  *
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/state.c,v 1.5 2000/09/26 15:56:33 tsi Exp $ */
+
 
 /*
  * This file initializes the immediate-mode dispatch table (which may
@@ -774,10 +774,10 @@ static void update_pixel_masking( GLcontext *ctx )
 
 static void update_fog_mode( GLcontext *ctx )
 {
-   int old_mode = ctx->FogMode;
+   const GLuint old_mode = ctx->FogMode;
 
    if (ctx->Fog.Enabled) {
-      if (ctx->Texture.Enabled)
+      if (ctx->Texture.ReallyEnabled)
          ctx->FogMode = FOG_FRAGMENT;
       else if (ctx->Hint.Fog == GL_NICEST)
          ctx->FogMode = FOG_FRAGMENT;
@@ -921,8 +921,9 @@ void gl_update_state( GLcontext *ctx )
       gl_update_client_state( ctx );
 
    if ((ctx->NewState & NEW_TEXTURE_ENABLE) &&
-       (ctx->Enabled & ENABLE_TEX_ANY) != ctx->Texture.Enabled)
+       (ctx->Enabled & ENABLE_TEX_ANY) != ctx->Texture.ReallyEnabled) {
       ctx->NewState |= NEW_TEXTURING | NEW_RASTER_OPS;
+   }
 
    if (ctx->NewState & NEW_TEXTURE_ENV) {
       if (ctx->Texture.Unit[0].EnvMode == ctx->Texture.Unit[0].LastEnvMode &&
@@ -932,21 +933,23 @@ void gl_update_state( GLcontext *ctx )
       ctx->Texture.Unit[1].LastEnvMode = ctx->Texture.Unit[1].EnvMode;
    }
 
-   if (ctx->NewState & NEW_TEXTURE_MATRIX) {
-      ctx->Enabled &= ~(ENABLE_TEXMAT0|ENABLE_TEXMAT1);
+   /* Update ctx->Enabled's ENABLE_TEXMATn flags */
+   if (ctx->NewState & (NEW_TEXTURE_MATRIX | NEW_TEXTURE_ENABLE)) {
+      ctx->Enabled &= ~(ENABLE_TEXMAT0 | ENABLE_TEXMAT1);
 
       for (i=0; i < MAX_TEXTURE_UNITS; i++) {
 	 if (ctx->TextureMatrix[i].flags & MAT_DIRTY_ALL_OVER) {
 	    gl_matrix_analyze( &ctx->TextureMatrix[i] );
 	    ctx->TextureMatrix[i].flags &= ~MAT_DIRTY_DEPENDENTS;
-
-	    if (ctx->Texture.Unit[i].Enabled &&
-		ctx->TextureMatrix[i].type != MATRIX_IDENTITY)
-	       ctx->Enabled |= ENABLE_TEXMAT0 << i;
 	 }
+         if (ctx->Texture.Unit[i].Enabled &&
+             ctx->TextureMatrix[i].type != MATRIX_IDENTITY) {
+            ctx->Enabled |= ENABLE_TEXMAT0 << i;
+         }
       }
    }
 
+   /* Update ctx->Enabled's ENABLE_TEXGENn and ENABLE_TEXn flags */
    if (ctx->NewState & (NEW_TEXTURING | NEW_TEXTURE_ENABLE)) {
       ctx->Texture.NeedNormals = GL_FALSE;
       gl_update_dirty_texobjs(ctx);
@@ -973,9 +976,11 @@ void gl_update_state( GLcontext *ctx )
 	       }
 	    }
 	 }
+         else {
+            ctx->Texture.Unit[i].ReallyEnabled = 0;
+         }
       }
-
-      ctx->Texture.Enabled = ctx->Enabled & ENABLE_TEX_ANY;
+      ctx->Enabled = (ctx->Enabled & ~ENABLE_TEX_ANY) | ctx->Texture.ReallyEnabled;
       ctx->NeedNormals = (ctx->Light.Enabled || ctx->Texture.NeedNormals);
    }
 
@@ -1186,8 +1191,9 @@ void gl_update_state( GLcontext *ctx )
 	 }
 	 ctx->NeedEyeNormals = ctx->NeedEyeCoords;
       }
-      if (ctx->Texture.Enabled || ctx->RenderMode==GL_FEEDBACK) {
-	 if (ctx->Texture.NeedEyeCoords) ctx->NeedEyeCoords = GL_TRUE;
+      if (ctx->Texture.ReallyEnabled || ctx->RenderMode==GL_FEEDBACK) {
+	 if (ctx->Texture.NeedEyeCoords)
+            ctx->NeedEyeCoords = GL_TRUE;
 	 if (ctx->Texture.NeedNormals)
 	    ctx->NeedNormals = ctx->NeedEyeNormals = GL_TRUE;
       }
diff --git a/xc/extras/Mesa/src/texgen_tmp.h b/xc/extras/Mesa/src/texgen_tmp.h
index 8fe16187f..c1fa50816 100644
--- a/xc/extras/Mesa/src/texgen_tmp.h
+++ b/xc/extras/Mesa/src/texgen_tmp.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texgen_tmp.h,v 1.5 2000/09/26 15:56:33 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
diff --git a/xc/extras/Mesa/src/teximage.c b/xc/extras/Mesa/src/teximage.c
index c92ff7937..ecbc0d7fe 100644
--- a/xc/extras/Mesa/src/teximage.c
+++ b/xc/extras/Mesa/src/teximage.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/teximage.c,v 1.10 2000/09/24 13:50:18 alanh Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -52,50 +52,50 @@
 #ifdef DEBUG
 static void PrintTexture(const struct gl_texture_image *img)
 {
-  int i, j, c;
-  GLubyte *data = img->Data;
-
-  if (!data) {
-     printf("No texture data\n");
-     return;
-  }
-
-  switch (img->Format) {
-     case GL_ALPHA:
-     case GL_LUMINANCE:
-     case GL_INTENSITY:
-     case GL_COLOR_INDEX:
-        c = 1;
-        break;
-     case GL_LUMINANCE_ALPHA:
-        c = 2;
-        break;
-     case GL_RGB:
-        c = 3;
-        break;
-     case GL_RGBA:
-        c = 4;
-        break;
-     default:
-        gl_problem(NULL, "error in PrintTexture\n");
-        return;
-  }
-
-
-  for (i = 0; i < img->Height; i++) {
-    for (j = 0; j < img->Width; j++) {
-      if (c==1)
-        printf("%02x  ", data[0]);
-      else if (c==2)
-        printf("%02x%02x  ", data[0], data[1]);
-      else if (c==3)
-        printf("%02x%02x%02x  ", data[0], data[1], data[2]);
-      else if (c==4)
-        printf("%02x%02x%02x%02x  ", data[0], data[1], data[2], data[3]);
-      data += c;
-    }
-    printf("\n");
-  }
+   GLuint i, j, c;
+   GLubyte *data = img->Data;
+
+   if (!data) {
+      printf("No texture data\n");
+      return;
+   }
+
+   switch (img->Format) {
+      case GL_ALPHA:
+      case GL_LUMINANCE:
+      case GL_INTENSITY:
+      case GL_COLOR_INDEX:
+         c = 1;
+         break;
+      case GL_LUMINANCE_ALPHA:
+         c = 2;
+         break;
+      case GL_RGB:
+         c = 3;
+         break;
+      case GL_RGBA:
+         c = 4;
+         break;
+      default:
+         gl_problem(NULL, "error in PrintTexture\n");
+         return;
+   }
+
+
+   for (i = 0; i < img->Height; i++) {
+      for (j = 0; j < img->Width; j++) {
+         if (c==1)
+            printf("%02x  ", data[0]);
+         else if (c==2)
+            printf("%02x%02x  ", data[0], data[1]);
+         else if (c==3)
+            printf("%02x%02x%02x  ", data[0], data[1], data[2]);
+         else if (c==4)
+            printf("%02x%02x%02x%02x  ", data[0], data[1], data[2], data[3]);
+         data += c;
+      }
+      printf("\n");
+   }
 }
 #endif
 
@@ -698,7 +698,7 @@ make_texture_image( GLcontext *ctx,
        && !ctx->Pixel.IndexOffset && !ctx->Pixel.IndexShift
        && srcType == GL_UNSIGNED_BYTE && depth == 1) {
 
-      if (srcFormat == internalFormat ||
+      if ((GLint) srcFormat == internalFormat ||
           (srcFormat == GL_LUMINANCE && internalFormat == 1) ||
           (srcFormat == GL_LUMINANCE_ALPHA && internalFormat == 2) ||
           (srcFormat == GL_RGB && internalFormat == 3) ||
@@ -824,13 +824,14 @@ make_null_texture( struct gl_texture_image *texImage )
       };
 
       GLubyte *imgPtr = texImage->Data;
-      GLint i, j, k;
+      GLuint i, j;
+      GLint k;
       for (i = 0; i < texImage->Height; i++) {
          GLint srcRow = 7 - i % 8;
          for (j = 0; j < texImage->Width; j++) {
             GLint srcCol = j % 32;
             GLint texel = (message[srcRow][srcCol]=='X') ? 255 : 70;
-            for (k=0;k<components;k++) {
+            for (k = 0; k < components; k++) {
                *imgPtr++ = (GLubyte) texel;
             }
          }
@@ -1348,7 +1349,14 @@ copytexsubimage_error_check( GLcontext *ctx, GLuint dimensions,
  */
 static GLint
 get_specific_compressed_tex_format(GLcontext *ctx,
-                                   GLint ifmt, GLint numDimensions)
+                                   GLint ifmt, GLint numDimensions,
+                                   GLint     *levelp,
+                                   GLsizei   *widthp,
+                                   GLsizei   *heightp,
+                                   GLsizei   *depthp,
+                                   GLint     *borderp,
+                                   GLenum    *formatp,
+                                   GLenum    *typep)
 {
    char message[100];
    GLint internalFormat = ifmt;
@@ -1357,22 +1365,14 @@ get_specific_compressed_tex_format(GLcontext *ctx,
        && ctx->Driver.SpecificCompressedTexFormat) {
       /*
        * First, ask the driver for the specific format.
+       * We do this for all formats, since we may want to
+       * fake one compressed format for another.
        */
-      switch (internalFormat) {
-         case GL_COMPRESSED_ALPHA_ARB:
-         case GL_COMPRESSED_LUMINANCE_ARB:
-         case GL_COMPRESSED_LUMINANCE_ALPHA_ARB:
-         case GL_COMPRESSED_INTENSITY_ARB:
-         case GL_COMPRESSED_RGB_ARB:
-         case GL_COMPRESSED_RGBA_ARB:
-            internalFormat = (*ctx->Driver.SpecificCompressedTexFormat)
-                                         (ctx, internalFormat, numDimensions);
-            /* XXX shouldn't we return now? */
-            break;
-         default:
-            /* silence compiler warnings */
-            ;
-      }
+       internalFormat = (*ctx->Driver.SpecificCompressedTexFormat)
+                               (ctx, internalFormat, numDimensions,
+                                levelp,
+                                widthp, heightp, depthp,
+                                borderp, formatp, typep);
    }
 
    /*
@@ -1438,7 +1438,6 @@ get_specific_compressed_tex_format(GLcontext *ctx,
 }
 
 
-
 /*
  * Called from the API.  Note that width includes the border.
  */
@@ -1456,7 +1455,10 @@ _mesa_TexImage1D( GLenum target, GLint level, GLint internalFormat,
       struct gl_texture_image *texImage;
       GLint ifmt;
 
-      ifmt = get_specific_compressed_tex_format(ctx, internalFormat, 1);
+      ifmt = get_specific_compressed_tex_format(ctx, internalFormat, 1,
+                                                &level,
+                                                &width, 0, 0,
+                                                &border, &format, &type);
       if (ifmt < 0) {
          /*
           * The error here is that we were sent a generic compressed
@@ -1536,10 +1538,16 @@ _mesa_TexImage1D( GLenum target, GLint level, GLint internalFormat,
       gl_put_texobj_on_dirty_list( ctx, texObj );
       ctx->NewState |= NEW_TEXTURING;
    }
-   else if (target==GL_PROXY_TEXTURE_1D) {
+   else if (target == GL_PROXY_TEXTURE_1D) {
       /* Proxy texture: check for errors and update proxy state */
-      if (texture_error_check(ctx, target, level, internalFormat,
-                              format, type, 1, width, 1, 1, border)) {
+      GLenum error = texture_error_check(ctx, target, level, internalFormat,
+                                         format, type, 1, width, 1, 1, border);
+      if (!error && ctx->Driver.TestProxyTexImage) {
+         error = !(*ctx->Driver.TestProxyTexImage)(ctx, target, level,
+                                                  internalFormat, format, type,
+                                                  width, 1, 1, border);
+      }
+      if (error) {
          /* if error, clear all proxy texture image parameters */
          if (level>=0 && level<ctx->Const.MaxTextureLevels) {
             clear_proxy_teximage(ctx->Texture.Proxy1D->Image[level]);
@@ -1576,7 +1584,10 @@ _mesa_TexImage2D( GLenum target, GLint level, GLint internalFormat,
       struct gl_texture_image *texImage;
       GLint ifmt;
 
-      ifmt = get_specific_compressed_tex_format(ctx, internalFormat, 2);
+      ifmt = get_specific_compressed_tex_format(ctx, internalFormat, 2,
+                                                &level,
+                                                &width, &height, 0,
+                                                &border, &format, &type);
       if (ifmt < 0) {
          /*
           * The error here is that we were sent a generic compressed
@@ -1667,10 +1678,16 @@ _mesa_TexImage2D( GLenum target, GLint level, GLint internalFormat,
       gl_put_texobj_on_dirty_list( ctx, texObj );
       ctx->NewState |= NEW_TEXTURING;
    }
-   else if (target==GL_PROXY_TEXTURE_2D) {
+   else if (target == GL_PROXY_TEXTURE_2D) {
       /* Proxy texture: check for errors and update proxy state */
-      if (texture_error_check(ctx, target, level, internalFormat,
-                              format, type, 2, width, height, 1, border)) {
+      GLenum error = texture_error_check(ctx, target, level, internalFormat,
+                                    format, type, 2, width, height, 1, border);
+      if (!error && ctx->Driver.TestProxyTexImage) {
+         error = !(*ctx->Driver.TestProxyTexImage)(ctx, target, level,
+                                                  internalFormat, format, type,
+                                                  width, height, 1, border);
+      }
+      if (error) {
          /* if error, clear all proxy texture image parameters */
          if (level>=0 && level<ctx->Const.MaxTextureLevels) {
             clear_proxy_teximage(ctx->Texture.Proxy2D->Image[level]);
@@ -1690,7 +1707,6 @@ _mesa_TexImage2D( GLenum target, GLint level, GLint internalFormat,
 }
 
 
-
 /*
  * Called by the API or display list executor.
  * Note that width and height include the border.
@@ -1710,7 +1726,10 @@ _mesa_TexImage3D( GLenum target, GLint level, GLint internalFormat,
       struct gl_texture_image *texImage;
       GLint ifmt;
 
-      ifmt = get_specific_compressed_tex_format(ctx, internalFormat, 3);
+      ifmt = get_specific_compressed_tex_format(ctx, internalFormat, 3,
+                                                &level,
+                                                &width, &height, &depth,
+                                                &border, &format, &type);
       if (ifmt < 0) {
          /*
           * The error here is that we were sent a generic compressed
@@ -1791,10 +1810,16 @@ _mesa_TexImage3D( GLenum target, GLint level, GLint internalFormat,
       gl_put_texobj_on_dirty_list( ctx, texObj );
       ctx->NewState |= NEW_TEXTURING;
    }
-   else if (target==GL_PROXY_TEXTURE_3D) {
+   else if (target == GL_PROXY_TEXTURE_3D) {
       /* Proxy texture: check for errors and update proxy state */
-      if (texture_error_check(ctx, target, level, internalFormat,
-                              format, type, 3, width, height, depth, border)) {
+      GLenum error = texture_error_check(ctx, target, level, internalFormat,
+                                format, type, 3, width, height, depth, border);
+      if (!error && ctx->Driver.TestProxyTexImage) {
+         error = !(*ctx->Driver.TestProxyTexImage)(ctx, target, level,
+                                                 internalFormat, format, type,
+                                                 width, height, depth, border);
+      }
+      if (error) {
          /* if error, clear all proxy texture image parameters */
          if (level>=0 && level<ctx->Const.MaxTextureLevels) {
             clear_proxy_teximage(ctx->Texture.Proxy3D->Image[level]);
@@ -1915,6 +1940,56 @@ _mesa_get_teximage_from_driver( GLcontext *ctx, GLenum target, GLint level,
 }
 
 
+/*
+ * Get all the mipmap images for a texture object from the device driver.
+ * Actually, only get mipmap images if we're using a mipmap filter.
+ */
+GLboolean
+_mesa_get_teximages_from_driver(GLcontext *ctx,
+                                struct gl_texture_object *texObj)
+{
+   if (ctx->Driver.GetTexImage) {
+      static const GLenum targets[] = {
+         GL_TEXTURE_1D,
+         GL_TEXTURE_2D,
+         GL_TEXTURE_3D,
+         GL_TEXTURE_CUBE_MAP_ARB,
+         GL_TEXTURE_CUBE_MAP_ARB,
+         GL_TEXTURE_CUBE_MAP_ARB
+      };
+      GLboolean needLambda = (texObj->MinFilter != texObj->MagFilter);
+      GLenum target = targets[texObj->Dimensions - 1];
+      if (needLambda) {
+         GLint level;
+         /* Get images for all mipmap levels.  We might not need them
+          * all but this is easier.  We're on a (slow) software path
+          * anyway.
+          */
+         for (level = 0; level <= texObj->P; level++) {
+            struct gl_texture_image *texImg = texObj->Image[level];
+            if (texImg && !texImg->Data) {
+               _mesa_get_teximage_from_driver(ctx, target, level, texObj);
+               if (!texImg->Data)
+                  return GL_FALSE;  /* out of memory */
+            }
+         }
+      }
+      else {
+         GLint level = texObj->BaseLevel;
+         struct gl_texture_image *texImg = texObj->Image[level];
+         if (texImg && !texImg->Data) {
+            _mesa_get_teximage_from_driver(ctx, target, level, texObj);
+            if (!texImg->Data)
+               return GL_FALSE;  /* out of memory */
+         }
+      }
+      return GL_TRUE;
+   }
+   return GL_FALSE;
+}
+
+
+
 void
 _mesa_GetTexImage( GLenum target, GLint level, GLenum format,
                    GLenum type, GLvoid *pixels )
@@ -2078,7 +2153,7 @@ _mesa_GetTexImage( GLenum target, GLint level, GLenum format,
                default:
                   gl_problem( ctx, "bad format in gl_GetTexImage" );
             }
-            _mesa_pack_rgba_span( ctx, width, (const GLubyte (*)[4])rgba,
+            _mesa_pack_rgba_span( ctx, width, (CONST GLubyte (*)[4])rgba,
                                   format, type, dest, &ctx->Pack, GL_TRUE );
          }
       }
@@ -2232,8 +2307,9 @@ _mesa_TexSubImage2D( GLenum target, GLint level,
          /* color index texture */
          GLubyte *dst = texImage->Data
                     + (yoffsetb * texImage->Width + xoffsetb) * texComponents;
-         const GLubyte *src = _mesa_image_address(&ctx->Unpack, pixels,
-                                     width, height, format, type, 0, 0, 0);
+         const GLubyte *src = (const GLubyte *)
+            _mesa_image_address(&ctx->Unpack, pixels, width, height, format,
+                                type, 0, 0, 0);
          GLint row;
          for (row = 0; row < height; row++) {
             _mesa_unpack_index_span(ctx, width, GL_UNSIGNED_BYTE, dst, type,
@@ -2246,8 +2322,9 @@ _mesa_TexSubImage2D( GLenum target, GLint level,
          /* color texture */
          GLubyte *dst = texImage->Data
                     + (yoffsetb * texImage->Width + xoffsetb) * texComponents;
-         const GLubyte *src = _mesa_image_address(&ctx->Unpack, pixels,
-                                     width, height, format, type, 0, 0, 0);
+         const GLubyte *src = (const GLubyte *)
+            _mesa_image_address(&ctx->Unpack, pixels, width, height, format,
+                                type, 0, 0, 0);
          GLint row;
          for (row = 0; row < height; row++) {
             _mesa_unpack_ubyte_color_span(ctx, width, texFormat, dst, format,
@@ -2337,8 +2414,9 @@ _mesa_TexSubImage3D( GLenum target, GLint level,
          /* color index texture */
          GLint img, row;
          for (img = 0; img < depth; img++) {
-            const GLubyte *src = _mesa_image_address(&ctx->Unpack, pixels,
-                                    width, height, format, type, img, 0, 0);
+            const GLubyte *src = (const GLubyte *)
+               _mesa_image_address(&ctx->Unpack, pixels, width, height,
+                                   format, type, img, 0, 0);
             GLubyte *dst = texImage->Data + ((zoffsetb + img) * dstRectArea
                      + yoffsetb * texWidth + xoffsetb) * texComponents;
             for (row = 0; row < height; row++) {
@@ -2353,8 +2431,9 @@ _mesa_TexSubImage3D( GLenum target, GLint level,
          /* color texture */
          GLint img, row;
          for (img = 0; img < depth; img++) {
-            const GLubyte *src = _mesa_image_address(&ctx->Unpack, pixels,
-                                     width, height, format, type, img, 0, 0);
+            const GLubyte *src = (const GLubyte *)
+               _mesa_image_address(&ctx->Unpack, pixels, width, height,
+                                   format, type, img, 0, 0);
             GLubyte *dst = texImage->Data + ((zoffsetb + img) * dstRectArea
                      + yoffsetb * texWidth + xoffsetb) * texComponents;
             for (row = 0; row < height; row++) {
@@ -2712,7 +2791,7 @@ _mesa_CompressedTexImage1DARB(GLenum target, GLint level,
                 gl_error(ctx, GL_INVALID_VALUE, "glCompressedTexImage1DARB(imageSize)");
                 return;
             }
-            texImage->Data = MALLOC(computedImageSize);
+            texImage->Data = (GLubyte *) MALLOC(computedImageSize);
             if (texImage->Data) {
                MEMCPY(texImage->Data, data, computedImageSize);
             }
@@ -2738,8 +2817,14 @@ _mesa_CompressedTexImage1DARB(GLenum target, GLint level,
    }
    else if (target == GL_PROXY_TEXTURE_1D) {
       /* Proxy texture: check for errors and update proxy state */
-      if (texture_error_check(ctx, target, level, internalFormat,
-                              GL_NONE, GL_NONE, 1, width, 1, 1, border)) {
+      GLenum error = texture_error_check(ctx, target, level, internalFormat,
+                                    GL_NONE, GL_NONE, 1, width, 1, 1, border);
+      if (!error && ctx->Driver.TestProxyTexImage) {
+         error = !(*ctx->Driver.TestProxyTexImage)(ctx, target, level,
+                                             internalFormat, GL_NONE, GL_NONE,
+                                             width, 1, 1, border);
+      }
+      if (error) {
          /* if error, clear all proxy texture image parameters */
          if (level>=0 && level<ctx->Const.MaxTextureLevels) {
             clear_proxy_teximage(ctx->Texture.Proxy1D->Image[level]);
@@ -2841,7 +2926,7 @@ _mesa_CompressedTexImage2DARB(GLenum target, GLint level,
                 gl_error(ctx, GL_INVALID_VALUE, "glCompressedTexImage2DARB(imageSize)");
                 return;
             }
-            texImage->Data = MALLOC(computedImageSize);
+            texImage->Data = (GLubyte *) MALLOC(computedImageSize);
             if (texImage->Data) {
                MEMCPY(texImage->Data, data, computedImageSize);
             }
@@ -2867,8 +2952,14 @@ _mesa_CompressedTexImage2DARB(GLenum target, GLint level,
    }
    else if (target == GL_PROXY_TEXTURE_2D) {
       /* Proxy texture: check for errors and update proxy state */
-      if (texture_error_check(ctx, target, level, internalFormat,
-                              GL_NONE, GL_NONE, 1, width, 1, 1, border)) {
+      GLenum error = texture_error_check(ctx, target, level, internalFormat,
+                                GL_NONE, GL_NONE, 2, width, height, 1, border);
+      if (!error && ctx->Driver.TestProxyTexImage) {
+         error = !(*ctx->Driver.TestProxyTexImage)(ctx, target, level,
+                                              internalFormat, GL_NONE, GL_NONE,
+                                              width, height, 1, border);
+      }
+      if (error) {
          /* if error, clear all proxy texture image parameters */
          if (level>=0 && level<ctx->Const.MaxTextureLevels) {
             clear_proxy_teximage(ctx->Texture.Proxy2D->Image[level]);
@@ -2964,7 +3055,7 @@ _mesa_CompressedTexImage3DARB(GLenum target, GLint level,
                 gl_error(ctx, GL_INVALID_VALUE, "glCompressedTexImage3DARB(imageSize)");
                 return;
             }
-            texImage->Data = MALLOC(computedImageSize);
+            texImage->Data = (GLubyte *) MALLOC(computedImageSize);
             if (texImage->Data) {
                MEMCPY(texImage->Data, data, computedImageSize);
             }
@@ -2990,8 +3081,14 @@ _mesa_CompressedTexImage3DARB(GLenum target, GLint level,
    }
    else if (target == GL_PROXY_TEXTURE_3D) {
       /* Proxy texture: check for errors and update proxy state */
-      if (texture_error_check(ctx, target, level, internalFormat,
-                          GL_NONE, GL_NONE, 1, width, height, depth, border)) {
+      GLenum error = texture_error_check(ctx, target, level, internalFormat,
+                            GL_NONE, GL_NONE, 1, width, height, depth, border);
+      if (!error && ctx->Driver.TestProxyTexImage) {
+         error = !(*ctx->Driver.TestProxyTexImage)(ctx, target, level,
+                                             internalFormat, GL_NONE, GL_NONE,
+                                             width, height, depth, border);
+      }
+      if (error) {
          /* if error, clear all proxy texture image parameters */
          if (level>=0 && level<ctx->Const.MaxTextureLevels) {
             clear_proxy_teximage(ctx->Texture.Proxy3D->Image[level]);
diff --git a/xc/extras/Mesa/src/teximage.h b/xc/extras/Mesa/src/teximage.h
index 1ff5e9038..ebc45e690 100644
--- a/xc/extras/Mesa/src/teximage.h
+++ b/xc/extras/Mesa/src/teximage.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/teximage.h,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
+
 
 #ifndef TEXIMAGE_H
 #define TEXIMAGE_H
@@ -64,8 +64,14 @@ _mesa_select_tex_image(GLcontext *ctx, const struct gl_texture_unit *texUnit,
 
 
 extern void
-_mesa_get_teximage_from_driver( GLcontext *ctx, GLenum target, GLint level,
-                                const struct gl_texture_object *texObj );
+_mesa_get_teximage_from_driver(GLcontext *ctx, GLenum target, GLint level,
+                               const struct gl_texture_object *texObj);
+
+
+extern GLboolean
+_mesa_get_teximages_from_driver(GLcontext *ctx,
+                                struct gl_texture_object *texObj);
+
 
 
 /*** API entry point functions ***/
diff --git a/xc/extras/Mesa/src/texobj.c b/xc/extras/Mesa/src/texobj.c
index 75d63fff1..86d438c47 100644
--- a/xc/extras/Mesa/src/texobj.c
+++ b/xc/extras/Mesa/src/texobj.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texobj.c,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -62,6 +62,7 @@ gl_alloc_texture_object( struct gl_shared_state *shared, GLuint name,
 
    if (obj) {
       /* init the non-zero fields */
+      _glthread_INIT_MUTEX(obj->Mutex);
       obj->RefCount = 1;
       obj->Name = name;
       obj->Dimensions = dimensions;
@@ -157,6 +158,20 @@ void gl_free_texture_object( struct gl_shared_state *shared,
 
 
 /*
+ * This is helpful to diagnose incomplete texture problems.
+ */
+#if 0
+static void
+incomplete(const struct gl_texture_object *t, const char *why)
+{
+   printf("Texture Obj %d incomplete because: %s\n", t->Name, why);
+}
+#else
+#define incomplete(a, b)
+#endif
+
+
+/*
  * Examine a texture object to determine if it is complete or not.
  * The t->Complete flag will be set to GL_TRUE or GL_FALSE accordingly.
  */
@@ -170,6 +185,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
 
    /* Always need level zero image */
    if (!t->Image[baseLevel]) {
+      incomplete(t, "Image[baseLevel] == NULL");
       t->Complete = GL_FALSE;
       return;
    }
@@ -213,6 +229,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
           t->NegZ[baseLevel]->Width2 != w ||
           t->NegZ[baseLevel]->Height2 != h) {
          t->Complete = GL_FALSE;
+         incomplete(t, "Non-quare cubemap image");
          return;
       }
    }
@@ -228,6 +245,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
 
       if (minLevel > maxLevel) {
          t->Complete = GL_FALSE;
+         incomplete(t, "minLevel > maxLevel");
          return;
       }
 
@@ -236,10 +254,12 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
          if (t->Image[i]) {
             if (t->Image[i]->Format != t->Image[baseLevel]->Format) {
                t->Complete = GL_FALSE;
+               incomplete(t, "Format[i] != Format[baseLevel]");
                return;
             }
             if (t->Image[i]->Border != t->Image[baseLevel]->Border) {
                t->Complete = GL_FALSE;
+               incomplete(t, "Border[i] != Border[baseLevel]");
                return;
             }
          }
@@ -256,10 +276,12 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
             if (i >= minLevel && i <= maxLevel) {
                if (!t->Image[i]) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "1D Image[i] == NULL");
                   return;
                }
                if (t->Image[i]->Width2 != width ) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "1D Image[i] bad width");
                   return;
                }
             }
@@ -282,14 +304,17 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
             if (i >= minLevel && i <= maxLevel) {
                if (!t->Image[i]) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "2D Image[i] == NULL");
                   return;
                }
                if (t->Image[i]->Width2 != width) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "2D Image[i] bad width");
                   return;
                }
                if (t->Image[i]->Height2 != height) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "2D Image[i] bad height");
                   return;
                }
                if (width==1 && height==1) {
@@ -315,19 +340,23 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
             }
             if (i >= minLevel && i <= maxLevel) {
                if (!t->Image[i]) {
+                  incomplete(t, "3D Image[i] == NULL");
                   t->Complete = GL_FALSE;
                   return;
                }
                if (t->Image[i]->Width2 != width) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "3D Image[i] bad width");
                   return;
                }
                if (t->Image[i]->Height2 != height) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "3D Image[i] bad height");
                   return;
                }
                if (t->Image[i]->Depth2 != depth) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "3D Image[i] bad depth");
                   return;
                }
             }
@@ -353,6 +382,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
                    !t->PosY[i]  || !t->NegY[i] ||
                    !t->PosZ[i]  || !t->NegZ[i]) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "CubeMap Image[i] == NULL");
                   return;
                }
                /* check that all six images have same size */
@@ -362,6 +392,7 @@ _mesa_test_texobj_completeness( const GLcontext *ctx,
                    t->PosZ[i]->Width2!=width || t->PosZ[i]->Height2!=height ||
                    t->NegZ[i]->Width2!=width || t->NegZ[i]->Height2!=height) {
                   t->Complete = GL_FALSE;
+                  incomplete(t, "CubeMap Image[i] bad size");
                   return;
                }
             }
@@ -589,6 +620,8 @@ _mesa_BindTexture( GLenum target, GLuint texName )
    /* Pass BindTexture call to device driver */
    if (ctx->Driver.BindTexture) {
       (*ctx->Driver.BindTexture)( ctx, target, newTexObj );
+      /* Make sure the Driver.UpdateState() function gets called! */
+      ctx->NewState |= NEW_TEXTURING;
    }
 
    if (oldTexObj->Name > 0) {
diff --git a/xc/extras/Mesa/src/texstate.c b/xc/extras/Mesa/src/texstate.c
index 7324ca657..523dbf82d 100644
--- a/xc/extras/Mesa/src/texstate.c
+++ b/xc/extras/Mesa/src/texstate.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texstate.c,v 1.6 2000/09/26 15:56:33 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -1591,7 +1591,7 @@ void
 _mesa_ActiveTextureARB( GLenum target )
 {
    GET_CURRENT_CONTEXT(ctx);
-   GLint maxUnits = ctx->Const.MaxTextureUnits;
+   const GLuint maxUnits = ctx->Const.MaxTextureUnits;
 
    ASSERT_OUTSIDE_BEGIN_END( ctx, "glActiveTextureARB" );
 
@@ -1618,7 +1618,7 @@ void
 _mesa_ClientActiveTextureARB( GLenum target )
 {
    GET_CURRENT_CONTEXT(ctx);
-   GLint maxUnits = ctx->Const.MaxTextureUnits;
+   const GLuint maxUnits = ctx->Const.MaxTextureUnits;
 
    ASSERT_OUTSIDE_BEGIN_END( ctx, "glClientActiveTextureARB" );
 
diff --git a/xc/extras/Mesa/src/texture.c b/xc/extras/Mesa/src/texture.c
index fe4863ff4..580121f47 100644
--- a/xc/extras/Mesa/src/texture.c
+++ b/xc/extras/Mesa/src/texture.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texture.c,v 1.7 2000/09/26 15:56:34 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -2005,8 +2005,8 @@ _mesa_set_texture_sampler( struct gl_texture_object *t )
 #define PROD(A,B)   ( (GLuint)(A) * ((GLuint)(B)+1) )
 
 static INLINE void
-_mesa_texture_combine(CONST GLcontext *ctx,
-                      CONST struct gl_texture_unit *textureUnit,
+_mesa_texture_combine(const GLcontext *ctx,
+                      const struct gl_texture_unit *textureUnit,
                       GLuint n,
                       GLubyte (*primary_rgba)[4],
                       GLubyte (*texel)[4],
@@ -2035,9 +2035,9 @@ _mesa_texture_combine(CONST GLcontext *ctx,
          case GL_CONSTANT_EXT:
             {
                GLubyte (*c)[4] = ccolor[j];
-               GLubyte col = FLOAT_TO_UBYTE(textureUnit->EnvColor[3]);
+               GLubyte alpha = FLOAT_TO_UBYTE(textureUnit->EnvColor[3]);
                for (i = 0; i < n; i++)
-                  c[i][ACOMP] = col;
+                  c[i][ACOMP] = alpha;
                argA[j] = ccolor[j];
             }
             break;
@@ -2045,17 +2045,6 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             gl_problem(NULL, "invalid combine source");
       }
 
-      if (textureUnit->CombineOperandA[j] == GL_ONE_MINUS_SRC_ALPHA) {
-         GLubyte (*src)[4] = argA[j];
-         GLubyte (*dst)[4] = ccolor[j];
-         argA[j] = ccolor[j];
-         for (i = 0; i < n; i++)
-            dst[i][ACOMP] = 255 - src[i][ACOMP];
-      }
-
-      if (j == 2)        /* arg2 has no color component in specification. */
-         break;
-
       switch (textureUnit->CombineSourceRGB[j]) {
          case GL_TEXTURE:
             argRGB[j] = texel;
@@ -2069,16 +2058,13 @@ _mesa_texture_combine(CONST GLcontext *ctx,
          case GL_CONSTANT_EXT:
             {
                GLubyte (*c)[4] = ccolor[j];
-               GLubyte col[4];
-
-               col[RCOMP] = FLOAT_TO_UBYTE(textureUnit->EnvColor[0]);
-               col[GCOMP] = FLOAT_TO_UBYTE(textureUnit->EnvColor[1]);
-               col[BCOMP] = FLOAT_TO_UBYTE(textureUnit->EnvColor[2]);
-
+               const GLubyte red   = FLOAT_TO_UBYTE(textureUnit->EnvColor[0]);
+               const GLubyte green = FLOAT_TO_UBYTE(textureUnit->EnvColor[1]);
+               const GLubyte blue  = FLOAT_TO_UBYTE(textureUnit->EnvColor[2]);
                for (i = 0; i < n; i++) {
-                  c[i][RCOMP] = col[RCOMP];
-                  c[i][GCOMP] = col[GCOMP];
-                  c[i][BCOMP] = col[BCOMP];
+                  c[i][RCOMP] = red;
+                  c[i][GCOMP] = green;
+                  c[i][BCOMP] = blue;
                }
                argRGB[j] = ccolor[j];
             }
@@ -2099,13 +2085,17 @@ _mesa_texture_combine(CONST GLcontext *ctx,
                dst[i][GCOMP] = 255 - src[i][GCOMP];
                dst[i][BCOMP] = 255 - src[i][BCOMP];
             }
-         } else if (textureUnit->CombineOperandRGB[j] == GL_SRC_ALPHA) {
+         }
+         else if (textureUnit->CombineOperandRGB[j] == GL_SRC_ALPHA) {
+            src = argA[j];
             for (i = 0; i < n; i++) {
                dst[i][RCOMP] = src[i][ACOMP];
                dst[i][GCOMP] = src[i][ACOMP];
                dst[i][BCOMP] = src[i][ACOMP];
             }
-         } else {                      /*  GL_ONE_MINUS_SRC_ALPHA  */
+         }
+         else {                      /*  GL_ONE_MINUS_SRC_ALPHA  */
+            src = argA[j];
             for (i = 0; i < n; i++) {
                dst[i][RCOMP] = 255 - src[i][ACOMP];
                dst[i][GCOMP] = 255 - src[i][ACOMP];
@@ -2114,6 +2104,15 @@ _mesa_texture_combine(CONST GLcontext *ctx,
          }
       }
 
+      if (textureUnit->CombineOperandA[j] == GL_ONE_MINUS_SRC_ALPHA) {
+         GLubyte (*src)[4] = argA[j];
+         GLubyte (*dst)[4] = ccolor[j];
+         argA[j] = ccolor[j];
+         for (i = 0; i < n; i++) {
+            dst[i][ACOMP] = 255 - src[i][ACOMP];
+         }
+      }
+
       if (textureUnit->CombineModeRGB == GL_REPLACE &&
           textureUnit->CombineModeA == GL_REPLACE) {
          break;      /*  done, we need only arg0  */
@@ -2130,10 +2129,22 @@ _mesa_texture_combine(CONST GLcontext *ctx,
       case GL_REPLACE:
          {
             const GLubyte (*arg0)[4] = (const GLubyte (*)[4]) argRGB[0];
-            for (i = 0; i < n; i++) {
-               rgba[i][RCOMP] = arg0[i][RCOMP] << RGBshift;
-               rgba[i][GCOMP] = arg0[i][GCOMP] << RGBshift;
-               rgba[i][BCOMP] = arg0[i][BCOMP] << RGBshift;
+            if (RGBshift) {
+               for (i = 0; i < n; i++) {
+                  GLuint r = (GLuint) arg0[i][RCOMP] << RGBshift;
+                  GLuint g = (GLuint) arg0[i][GCOMP] << RGBshift;
+                  GLuint b = (GLuint) arg0[i][BCOMP] << RGBshift;
+                  rgba[i][RCOMP] = MIN2(r, 255);
+                  rgba[i][GCOMP] = MIN2(g, 255);
+                  rgba[i][BCOMP] = MIN2(b, 255);
+               }
+            }
+            else {
+               for (i = 0; i < n; i++) {
+                  rgba[i][RCOMP] = arg0[i][RCOMP];
+                  rgba[i][GCOMP] = arg0[i][GCOMP];
+                  rgba[i][BCOMP] = arg0[i][BCOMP];
+               }
             }
          }
          break;
@@ -2143,9 +2154,12 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             const GLubyte (*arg1)[4] = (const GLubyte (*)[4]) argRGB[1];
             RGBshift = 8 - RGBshift;
             for (i = 0; i < n; i++) {
-               rgba[i][RCOMP] = (GLubyte) (PROD(arg0[i][0], arg1[i][RCOMP]) >> RGBshift);
-               rgba[i][GCOMP] = (GLubyte) (PROD(arg0[i][1], arg1[i][GCOMP]) >> RGBshift);
-               rgba[i][BCOMP] = (GLubyte) (PROD(arg0[i][2], arg1[i][BCOMP]) >> RGBshift);
+               GLuint r = PROD(arg0[i][0], arg1[i][RCOMP]) >> RGBshift;
+               GLuint g = PROD(arg0[i][1], arg1[i][GCOMP]) >> RGBshift;
+               GLuint b = PROD(arg0[i][2], arg1[i][BCOMP]) >> RGBshift;
+               rgba[i][RCOMP] = (GLubyte) MIN2(r, 255);
+               rgba[i][GCOMP] = (GLubyte) MIN2(g, 255);
+               rgba[i][BCOMP] = (GLubyte) MIN2(b, 255);
             }
          }
          break;
@@ -2157,9 +2171,9 @@ _mesa_texture_combine(CONST GLcontext *ctx,
                GLint r = ((GLuint) arg0[i][RCOMP] + arg1[i][RCOMP]) << RGBshift;
                GLint g = ((GLuint) arg0[i][GCOMP] + arg1[i][GCOMP]) << RGBshift;
                GLint b = ((GLuint) arg0[i][BCOMP] + arg1[i][BCOMP]) << RGBshift;
-               rgba[i][RCOMP] = (r > 255) ? 255 : (GLubyte) r;
-               rgba[i][GCOMP] = (g > 255) ? 255 : (GLubyte) g;
-               rgba[i][BCOMP] = (b > 255) ? 255 : (GLubyte) b;
+               rgba[i][RCOMP] = (GLubyte) MIN2(r, 255);
+               rgba[i][GCOMP] = (GLubyte) MIN2(g, 255);
+               rgba[i][BCOMP] = (GLubyte) MIN2(b, 255);
             }
          }
          break;
@@ -2174,9 +2188,9 @@ _mesa_texture_combine(CONST GLcontext *ctx,
                r = (r < 0) ? 0 : r << RGBshift;
                b = (b < 0) ? 0 : b << RGBshift;
                g = (g < 0) ? 0 : g << RGBshift;
-               rgba[i][RCOMP] = (r > 255) ? 255 : (GLubyte) r;
-               rgba[i][GCOMP] = (g > 255) ? 255 : (GLubyte) g;
-               rgba[i][BCOMP] = (b > 255) ? 255 : (GLubyte) b;
+               rgba[i][RCOMP] = (GLubyte) MIN2(r, 255);
+               rgba[i][GCOMP] = (GLubyte) MIN2(g, 255);
+               rgba[i][BCOMP] = (GLubyte) MIN2(b, 255);
             }
          }
          break;
@@ -2187,9 +2201,18 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             const GLubyte (*arg2)[4] = (const GLubyte (*)[4]) argRGB[2];
             RGBshift = 8 - RGBshift;
             for (i = 0; i < n; i++) {
-               rgba[i][RCOMP] = (GLubyte) ((PROD(arg0[i][RCOMP], arg2[i][ACOMP]) + PROD(arg1[i][RCOMP], 255 - arg2[i][ACOMP])) >> RGBshift);
-               rgba[i][GCOMP] = (GLubyte) ((PROD(arg0[i][GCOMP], arg2[i][ACOMP]) + PROD(arg1[i][GCOMP], 255 - arg2[i][ACOMP])) >> RGBshift);
-               rgba[i][BCOMP] = (GLubyte) ((PROD(arg0[i][BCOMP], arg2[i][ACOMP]) + PROD(arg1[i][BCOMP], 255 - arg2[i][ACOMP])) >> RGBshift);
+               GLuint r = (PROD(arg0[i][RCOMP], arg2[i][RCOMP])
+                           + PROD(arg1[i][RCOMP], 255 - arg2[i][RCOMP]))
+                              >> RGBshift;
+               GLuint g = (PROD(arg0[i][GCOMP], arg2[i][GCOMP])
+                           + PROD(arg1[i][GCOMP], 255 - arg2[i][GCOMP]))
+                              >> RGBshift;
+               GLuint b = (PROD(arg0[i][BCOMP], arg2[i][BCOMP])
+                           + PROD(arg1[i][BCOMP], 255 - arg2[i][BCOMP]))
+                              >> RGBshift;
+               rgba[i][RCOMP] = (GLubyte) MIN2(r, 255);
+               rgba[i][GCOMP] = (GLubyte) MIN2(g, 255);
+               rgba[i][BCOMP] = (GLubyte) MIN2(b, 255);
             }
          }
          break;
@@ -2201,8 +2224,17 @@ _mesa_texture_combine(CONST GLcontext *ctx,
       case GL_REPLACE:
          {
             const GLubyte (*arg0)[4] = (const GLubyte (*)[4]) argA[0];
-            for (i = 0; i < n; i++)
-               rgba[i][ACOMP] = arg0[i][ACOMP] << Ashift;
+            if (Ashift) {
+               for (i = 0; i < n; i++) {
+                  GLuint a = (GLuint) arg0[i][ACOMP] << Ashift;
+                  rgba[i][ACOMP] = (GLubyte) MIN2(a, 255);
+               }
+            }
+            else {
+               for (i = 0; i < n; i++) {
+                  rgba[i][ACOMP] = arg0[i][ACOMP];
+               }
+            }
          }
          break;
       case GL_MODULATE:
@@ -2210,8 +2242,10 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             const GLubyte (*arg0)[4] = (const GLubyte (*)[4]) argA[0];
             const GLubyte (*arg1)[4] = (const GLubyte (*)[4]) argA[1];
             Ashift = 8 - Ashift;
-            for (i = 0; i < n; i++)
-               rgba[i][ACOMP] = (GLubyte) (PROD(arg0[i][ACOMP], arg1[i][ACOMP]) >> Ashift);
+            for (i = 0; i < n; i++) {
+               GLuint a = (PROD(arg0[i][ACOMP], arg1[i][ACOMP]) >> Ashift);
+               rgba[i][ACOMP] = (GLubyte) MIN2(a, 255);
+            }
          }
          break;
       case GL_ADD:
@@ -2220,7 +2254,7 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             const GLubyte  (*arg1)[4] = (const GLubyte (*)[4]) argA[1];
             for (i = 0; i < n; i++) {
                GLint a = ((GLint) arg0[i][ACOMP] + arg1[i][ACOMP]) << Ashift;
-               rgba[i][ACOMP] = (a > 255) ? 255 : (GLubyte) a;
+               rgba[i][ACOMP] = (GLubyte) MIN2(a, 255);
             }
          }
          break;
@@ -2231,7 +2265,7 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             for (i = 0; i < n; i++) {
                GLint a = (GLint) arg0[i][ACOMP] + (GLint) arg1[i][ACOMP] - 128;
                a = (a < 0) ? 0 : a << Ashift;
-               rgba[i][ACOMP] = (a > 255) ? 255 : (GLubyte) a;
+               rgba[i][ACOMP] = (GLubyte) MIN2(a, 255);
             }
          }
          break;
@@ -2241,8 +2275,12 @@ _mesa_texture_combine(CONST GLcontext *ctx,
             const GLubyte (*arg1)[4] = (const GLubyte (*)[4]) argA[1];
             const GLubyte (*arg2)[4] = (const GLubyte (*)[4]) argA[2];
             Ashift = 8 - Ashift;
-            for (i=0; i<n; i++)
-               rgba[i][ACOMP] = (GLubyte) ((PROD(arg0[i][ACOMP], arg2[i][ACOMP]) + PROD(arg1[i][ACOMP], 255 - arg2[i][ACOMP])) >> Ashift);
+            for (i=0; i<n; i++) {
+               GLuint a = (PROD(arg0[i][ACOMP], arg2[i][ACOMP])
+                           + PROD(arg1[i][ACOMP], 255 - arg2[i][ACOMP]))
+                              >> Ashift;
+               rgba[i][ACOMP] = (GLubyte) MIN2(a, 255);
+            }
          }
          break;
       default:
@@ -2269,7 +2307,7 @@ _mesa_texture_combine(CONST GLcontext *ctx,
  *                according to the texture environment mode.
  */
 static void apply_texture( CONST GLcontext *ctx,
-                           CONST struct gl_texture_unit *texUnit,
+                           const struct gl_texture_unit *texUnit,
                            GLuint n,
                            GLubyte primary_rgba[][4], GLubyte texel[][4],
                            GLubyte rgba[][4] )
@@ -2645,7 +2683,7 @@ static void apply_texture( CONST GLcontext *ctx,
                gl_problem(ctx, "Bad format in apply_texture (GL_COMBINE_EXT)");
                return;
          }
-         _mesa_texture_combine (ctx, texUnit, n, primary_rgba, texel, rgba);
+         _mesa_texture_combine(ctx, texUnit, n, primary_rgba, texel, rgba);
          break;
 
       default:
@@ -2666,7 +2704,7 @@ void gl_texture_pixels( GLcontext *ctx, GLuint texUnit, GLuint n,
                         GLubyte primary_rgba[][4], GLubyte rgba[][4] )
 {
    GLuint mask = (TEXTURE0_1D | TEXTURE0_2D | TEXTURE0_3D | TEXTURE0_CUBE) << (texUnit * 4);
-   if (ctx->Texture.Enabled & mask) {
+   if (ctx->Texture.ReallyEnabled & mask) {
       const struct gl_texture_unit *textureUnit = &ctx->Texture.Unit[texUnit];
       if (textureUnit->Current && textureUnit->Current->SampleFunc) {
          GLubyte texel[PB_SIZE][4];
@@ -2691,41 +2729,9 @@ void gl_texture_pixels( GLcontext *ctx, GLuint texUnit, GLuint n,
          }
 
          /* fetch texture images from device driver, if needed */
-         {
-            static const GLenum targets[] = {
-               GL_TEXTURE_1D,
-               GL_TEXTURE_2D,
-               GL_TEXTURE_3D,
-               GL_TEXTURE_CUBE_MAP_ARB,
-               GL_TEXTURE_CUBE_MAP_ARB,
-               GL_TEXTURE_CUBE_MAP_ARB
-            };
-            struct gl_texture_object *texObj = textureUnit->Current;
-            GLboolean needLambda = (texObj->MinFilter != texObj->MagFilter);
-            GLenum target = targets[texObj->Dimensions - 1];
-            if (needLambda) {
-               GLint level;
-               /* Get images for all mipmap levels.  We might not need them
-                * all but this is easier.  We're on a (slow) software path
-                * anyway.
-                */
-               for (level = 0; level <= texObj->P; level++) {
-                  struct gl_texture_image *texImg = texObj->Image[level];
-                  if (texImg && !texImg->Data) {
-                     _mesa_get_teximage_from_driver(ctx, target, level, texObj);
-                     if (!texImg->Data)
-                        return;  /* out of memory */
-                  }
-               }
-            }
-            else {
-               GLint level = texObj->BaseLevel;
-               struct gl_texture_image *texImg = texObj->Image[level];
-               if (texImg && !texImg->Data) {
-                  _mesa_get_teximage_from_driver(ctx, target, level, texObj);
-                  if (!texImg->Data)
-                     return;  /* out of memory */
-               }
+         if (ctx->Driver.GetTexImage) {
+            if (!_mesa_get_teximages_from_driver(ctx, textureUnit->Current)) {
+               return;
             }
          }
 
diff --git a/xc/extras/Mesa/src/texture.h b/xc/extras/Mesa/src/texture.h
index 379e7a1d4..113602bf7 100644
--- a/xc/extras/Mesa/src/texture.h
+++ b/xc/extras/Mesa/src/texture.h
@@ -22,7 +22,10 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texture.h,v 1.6 2000/09/26 15:56:34 tsi Exp $ */
+
+
+
+
 
 #ifndef TEXTURE_H
 #define TEXTURE_H
diff --git a/xc/extras/Mesa/src/texutil.c b/xc/extras/Mesa/src/texutil.c
index a12ddd745..6d45a3b2f 100644
--- a/xc/extras/Mesa/src/texutil.c
+++ b/xc/extras/Mesa/src/texutil.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texutil.c,v 1.3 2000/09/26 15:56:34 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -80,7 +80,10 @@
  *   GL_RGBA             GL_UNSIGNED_BYTE                 MESA_A4_R4_G4_B4
  *   GL_BGRA             GL_UNSIGNED_SHORT_4_4_4_4_REV    MESA_A4_R4_G4_B4
  *   GL_BGRA             GL_UNSIGHED_SHORT_1_5_5_5_REV    MESA_A1_R5_G5_B5
+ *   GL_RGBA             GL_UNSIGNED_BYTE                 MESA_A1_R5_G5_B5
  *   GL_BGRA             GL_UNSIGNED_INT_8_8_8_8_REV      MESA_A8_R8_G8_B8
+ *   GL_RGBA             GL_UNSIGNED_BYTE                 MESA_A8_R8_G8_B8
+ *   GL_RGB              GL_UNSIGNED_BYTE                 MESA_A8_R8_G8_B8
  *   more to be added for new drivers...
  *
  * Notes:
@@ -123,8 +126,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
             /* store as 8-bit texels */
             if (wScale == 1 && hScale == 1) {
                /* no scaling needed - fast case */
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                GLubyte *dst = (GLubyte *) dstImage;
@@ -141,8 +145,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                             srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      dst[col] = src[col / wScale];
@@ -160,8 +165,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else {
             /* store as 16-bit texels */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                GLushort *dst = (GLushort *) dstImage;
@@ -182,8 +188,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row, col;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                   for (col = 0; col < dstWidth; col++) {
@@ -203,8 +210,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          if (srcFormat == GL_RGB && srcType == GL_UNSIGNED_SHORT_5_6_5) {
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                GLushort *dst = (GLushort *) dstImage;
@@ -221,8 +229,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLushort *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLushort *src = (const GLushort *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      dst[col] = src[col / wScale];
@@ -234,8 +243,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGB && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
 #ifdef DO_32BIT_STORES
@@ -286,8 +296,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      GLint col3 = (col / wScale) * 3;
@@ -305,8 +316,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case (used by Quake3) */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
 #ifdef DO_32BIT_STORES
@@ -357,8 +369,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -384,8 +397,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          if (srcFormat == GL_BGRA && srcType == GL_UNSIGNED_SHORT_4_4_4_4_REV){
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                GLushort *dst = (GLushort *) dstImage;
@@ -402,8 +416,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLushort *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLushort *src = (const GLushort *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      dst[col] = src[col / wScale];
@@ -415,8 +430,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
 #ifdef DO_32BIT_STORES
@@ -473,8 +489,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -499,11 +516,12 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
 
       case MESA_A1_R5_G5_B5:
          /* store as 16-bit texels (GR_TEXFMT_ARGB_1555) */
-         if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_SHORT_1_5_5_5_REV){
+         if (srcFormat == GL_BGRA && srcType == GL_UNSIGNED_SHORT_1_5_5_5_REV){
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                GLushort *dst = (GLushort *) dstImage;
@@ -520,8 +538,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLushort *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLushort *src = (const GLushort *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      dst[col] = src[col / wScale];
@@ -533,8 +552,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
                GLushort *dst = (GLushort *) dstImage;
@@ -561,8 +581,9 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -591,11 +612,12 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          if (srcFormat == GL_BGRA && srcType == GL_UNSIGNED_INT_8_8_8_8_REV){
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
-               GLuint *dst = dstImage;
+               GLuint *dst = (GLuint *) dstImage;
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   MEMCPY(dst, src, dstWidth * sizeof(GLuint));
@@ -605,12 +627,13 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
             }
             else {
                /* must rescale image */
-               GLuint *dst = dstImage;
+               GLuint *dst = (GLuint *) dstImage;
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLuint *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLuint *src = (const GLuint *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      dst[col] = src[col / wScale];
@@ -622,11 +645,12 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
-               GLuint *dst = dstImage;
+               GLuint *dst = (GLuint *) dstImage;
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint col, col4;
@@ -643,12 +667,13 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
             }
             else {
                /* must rescale image */
-               GLuint *dst = dstImage;
+               GLuint *dst = (GLuint *) dstImage;
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -665,11 +690,12 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGB && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  srcWidth, srcFormat, srcType);
-               GLuint *dst = dstImage;
+               GLuint *dst = (GLuint *) dstImage;
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint col, col3;
@@ -686,12 +712,13 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
             }
             else {
                /* must rescale image */
-               GLuint *dst = dstImage;
+               GLuint *dst = (GLuint *) dstImage;
                GLint row;
                for (row = 0; row < dstHeight; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < dstWidth; col++) {
                      GLint col3 = (col / wScale) * 3;
@@ -711,7 +738,7 @@ _mesa_convert_teximage(MesaIntTexFormat dstFormat,
          }
          if (dstFormat == MESA_FF_R8_G8_B8) {
             /* set alpha bytes to 0xff */
-            GLuint i;
+            GLint i;
             GLubyte *dst = (GLubyte *) dstImage;
             for (i = 0; i < dstWidth * dstHeight; i++) {
                dst[i * 4 + 3] = 0xff;
@@ -792,8 +819,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
             /* store as 8-bit texels */
             if (wScale == 1 && hScale == 1) {
                /* no scaling needed - fast case */
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                   width, srcFormat, srcType);
                GLubyte *dst = (GLubyte *) dstImage
@@ -812,8 +840,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      dst[col] = src[col / wScale];
@@ -831,8 +860,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          else {
             /* store as 16-bit texels */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -855,8 +885,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row, col;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                   for (col = 0; col < width; col++) {
@@ -876,8 +907,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          if (srcFormat == GL_RGB && srcType == GL_UNSIGNED_SHORT_5_6_5) {
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -896,8 +928,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLushort *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLushort *src = (const GLushort *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      dst[col] = src[col / wScale];
@@ -909,8 +942,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGB && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -937,8 +971,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      GLint col3 = (col / wScale) * 3;
@@ -956,8 +991,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case (used by Quake3) */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -984,8 +1020,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -1011,8 +1048,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          if (srcFormat == GL_BGRA && srcType == GL_UNSIGNED_SHORT_4_4_4_4_REV){
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -1031,8 +1069,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLushort *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLushort *src = (const GLushort *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      dst[col] = src[col / wScale];
@@ -1044,8 +1083,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -1074,8 +1114,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -1100,11 +1141,12 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
 
       case MESA_A1_R5_G5_B5:
          /* store as 16-bit texels (GR_TEXFMT_ARGB_1555) */
-         if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_SHORT_1_5_5_5_REV){
+         if (srcFormat == GL_BGRA && srcType == GL_UNSIGNED_SHORT_1_5_5_5_REV){
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -1123,8 +1165,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLushort *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLushort *src = (const GLushort *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      dst[col] = src[col / wScale];
@@ -1136,8 +1179,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLushort *dst = (GLushort *) ((GLubyte *) dstImage
@@ -1166,8 +1210,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      GLint col4 = (col / wScale) * 4;
@@ -1196,8 +1241,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          if (srcFormat == GL_BGRA && srcType == GL_UNSIGNED_INT_8_8_8_8_REV){
             /* special, optimized case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLuint *dst = (GLuint *) ((GLubyte *) dstImage
@@ -1216,8 +1262,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLuint *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLuint *src = (const GLuint *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      dst[col] = src[col / wScale];
@@ -1229,8 +1276,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
          else if (srcFormat == GL_RGBA && srcType == GL_UNSIGNED_BYTE) {
             /* general case */
             if (wScale == 1 && hScale == 1) {
-               const GLubyte *src = _mesa_image_address(packing, srcImage,
-                             srcWidth, srcHeight, srcFormat, srcType, 0, 0, 0);
+               const GLubyte *src = (const GLubyte *)
+                  _mesa_image_address(packing, srcImage, srcWidth, srcHeight,
+                                      srcFormat, srcType, 0, 0, 0);
                const GLint srcStride = _mesa_image_row_stride(packing,
                                                  width, srcFormat, srcType);
                GLuint *dst = (GLuint *) ((GLubyte *) dstImage
@@ -1256,8 +1304,9 @@ _mesa_convert_texsubimage(MesaIntTexFormat dstFormat,
                GLint row;
                for (row = 0; row < height; row++) {
                   GLint srcRow = row / hScale;
-                  const GLubyte *src = _mesa_image_address(packing, srcImage,
-                        srcWidth, srcHeight, srcFormat, srcType, 0, srcRow, 0);
+                  const GLubyte *src = (const GLubyte *)
+                     _mesa_image_address(packing, srcImage, srcWidth,
+                                  srcHeight, srcFormat, srcType, 0, srcRow, 0);
                   GLint col;
                   for (col = 0; col < width; col++) {
                      GLint col4 = (col / wScale) * 4;
diff --git a/xc/extras/Mesa/src/texutil.h b/xc/extras/Mesa/src/texutil.h
index f1fbcfab2..22e916d50 100644
--- a/xc/extras/Mesa/src/texutil.h
+++ b/xc/extras/Mesa/src/texutil.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/texutil.h,v 1.3 2000/09/26 15:56:34 tsi Exp $ */
+
 
 #ifndef TEXUTIL_H
 #define TEXUTIL_H
diff --git a/xc/extras/Mesa/src/triangle.c b/xc/extras/Mesa/src/triangle.c
index d8f402d67..eb59d56ab 100644
--- a/xc/extras/Mesa/src/triangle.c
+++ b/xc/extras/Mesa/src/triangle.c
@@ -1,7 +1,7 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
  * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/triangle.c,v 1.7 2000/09/26 15:56:34 tsi Exp $ */
+
 
 /*
  * Triangle rasterizers
@@ -42,6 +42,7 @@
 #include "mem.h"
 #include "mmath.h"
 #include "span.h"
+#include "teximage.h"
 #include "texstate.h"
 #include "triangle.h"
 #include "types.h"
@@ -221,6 +222,7 @@ static void smooth_rgba_triangle( GLcontext *ctx,
 }
 
 
+
 /*
  * Render an RGB, GL_DECAL, textured triangle.
  * Interpolate S,T only w/out mipmapping or perspective correction.
@@ -239,8 +241,14 @@ static void simple_textured_triangle( GLcontext *ctx, GLuint v0, GLuint v1,
    GLint twidth_log2 = obj->Image[b]->WidthLog2;			\
    GLubyte *texture = obj->Image[b]->Data;				\
    GLint smask = obj->Image[b]->Width - 1;				\
-   GLint tmask = obj->Image[b]->Height - 1;
-   (void) pv;
+   GLint tmask = obj->Image[b]->Height - 1;				\
+   (void) pv;								\
+   if (!texture) {							\
+      if (!_mesa_get_teximages_from_driver(ctx, obj))			\
+         return;							\
+      texture = obj->Image[b]->Data;					\
+      ASSERT(texture);							\
+   }
 
 #define INNER_LOOP( LEFT, RIGHT, Y )				\
 	{							\
@@ -261,7 +269,7 @@ static void simple_textured_triangle( GLcontext *ctx, GLuint v0, GLuint v1,
 		 fft += fdtdx;					\
 	      }							\
               (*ctx->Driver.WriteRGBSpan)( ctx, n, LEFT, Y,	\
-                           (const GLubyte (*)[3]) rgb, NULL );	\
+                           (CONST GLubyte (*)[3]) rgb, NULL );	\
 	   }							\
 	}
 
@@ -290,8 +298,14 @@ static void simple_z_textured_triangle( GLcontext *ctx, GLuint v0, GLuint v1,
    GLint twidth_log2 = obj->Image[b]->WidthLog2;			\
    GLubyte *texture = obj->Image[b]->Data;				\
    GLint smask = obj->Image[b]->Width - 1;				\
-   GLint tmask = obj->Image[b]->Height - 1;
-   (void) pv;
+   GLint tmask = obj->Image[b]->Height - 1;				\
+   (void) pv;								\
+   if (!texture) {							\
+      if (!_mesa_get_teximages_from_driver(ctx, obj))			\
+         return;							\
+      texture = obj->Image[b]->Data;					\
+      ASSERT(texture);							\
+   }
 
 #define INNER_LOOP( LEFT, RIGHT, Y )				\
 	{							\
@@ -322,7 +336,7 @@ static void simple_z_textured_triangle( GLcontext *ctx, GLuint v0, GLuint v1,
 		 fft += fdtdx;					\
 	      }							\
               (*ctx->Driver.WriteRGBSpan)( ctx, n, LEFT, Y,	\
-                           (const GLubyte (*)[3]) rgb, mask );	\
+                           (CONST GLubyte (*)[3]) rgb, mask );	\
 	   }							\
 	}
 
@@ -360,6 +374,12 @@ static void affine_textured_triangle( GLcontext *ctx, GLuint v0, GLuint v1,
    GLint comp, tbytesline, tsize;                                       \
    GLfixed er, eg, eb, ea;                                              \
    GLint tr, tg, tb, ta;                                                \
+   if (!texture) {							\
+      if (!_mesa_get_teximages_from_driver(ctx, obj))			\
+         return;							\
+      texture = obj->Image[b]->Data;					\
+      ASSERT(texture);							\
+   }									\
    if (envmode == GL_BLEND) {                                           \
       /* potential off-by-one error here? (1.0f -> 2048 -> 0) */        \
       er = FloatToFixed(unit->EnvColor[0]);                             \
@@ -648,6 +668,12 @@ static void persp_textured_triangle( GLcontext *ctx, GLuint v0, GLuint v1,
    GLint comp, tbytesline, tsize;                                       \
    GLfixed er, eg, eb, ea;                                              \
    GLint tr, tg, tb, ta;                                                \
+   if (!texture) {							\
+      if (!_mesa_get_teximages_from_driver(ctx, obj))			\
+         return;							\
+      texture = obj->Image[b]->Data;					\
+      ASSERT(texture);							\
+   }									\
    if (envmode == GL_BLEND) {                                           \
       er = FloatToFixed(unit->EnvColor[0]);                             \
       eg = FloatToFixed(unit->EnvColor[1]);                             \
@@ -1017,7 +1043,7 @@ static void general_textured_spec_triangle1( GLcontext *ctx, GLuint v0,
               }							\
 	      gl_write_texture_span( ctx, n, LEFT, Y, zspan,	\
                                    s, t, u, NULL, rgba,		\
-                                   (const GLubyte (*)[4]) spec,	\
+                                   (CONST GLubyte (*)[4]) spec,	\
 	                           GL_POLYGON );		\
 	   }							\
 	}
@@ -1243,7 +1269,7 @@ static void lambda_textured_spec_triangle1( GLcontext *ctx, GLuint v0,
               }								\
 	      gl_write_texture_span( ctx, n, LEFT, Y, zspan,		\
                                      s, t, u, lambda,	 		\
-	                             rgba, (const GLubyte (*)[4]) spec, \
+	                             rgba, (CONST GLubyte (*)[4]) spec, \
                                      GL_POLYGON );			\
 	   }								\
 	}
@@ -1371,9 +1397,9 @@ static void lambda_multitextured_triangle1( GLcontext *ctx, GLuint v0,
 		 }							\
               }								\
 	      gl_write_multitexture_span( ctx, 2, n, LEFT, Y, zspan,	\
-                                      (const GLfloat (*)[MAX_WIDTH]) s,	\
-                                      (const GLfloat (*)[MAX_WIDTH]) t,	\
-                                      (const GLfloat (*)[MAX_WIDTH]) u,	\
+                                      (CONST GLfloat (*)[MAX_WIDTH]) s,	\
+                                      (CONST GLfloat (*)[MAX_WIDTH]) t,	\
+                                      (CONST GLfloat (*)[MAX_WIDTH]) u,	\
                                       (GLfloat (*)[MAX_WIDTH]) lambda,	\
 	                              rgba, NULL, GL_POLYGON );		\
 	   }								\
@@ -1505,6 +1531,8 @@ void gl_set_triangle_function( GLcontext *ctx )
       }
 
       if (ctx->Depth.OcclusionTest &&
+          ctx->DrawBuffer->DepthBuffer &&
+          ctx->Depth.Test &&
           ctx->Depth.Mask == GL_FALSE &&
           ctx->Depth.Func == GL_LESS &&
           !ctx->Stencil.Enabled) {
diff --git a/xc/extras/Mesa/src/tritemp.h b/xc/extras/Mesa/src/tritemp.h
index 54f8cea2e..d0135dc1d 100644
--- a/xc/extras/Mesa/src/tritemp.h
+++ b/xc/extras/Mesa/src/tritemp.h
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/tritemp.h,v 1.9 2000/09/24 13:50:20 alanh Exp $ */
+
 
 /*
  * Triangle Rasterizer Template
diff --git a/xc/extras/Mesa/src/types.h b/xc/extras/Mesa/src/types.h
index 550f4f1e6..9b1e832d7 100644
--- a/xc/extras/Mesa/src/types.h
+++ b/xc/extras/Mesa/src/types.h
@@ -23,7 +23,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/types.h,v 1.7 2000/09/26 15:56:34 tsi Exp $ */
+
 
 #ifndef TYPES_H
 #define TYPES_H
@@ -449,7 +449,7 @@ struct gl_enable_attrib {
    GLboolean RescaleNormals;
    GLboolean Scissor;
    GLboolean Stencil;
-   GLuint Texture;
+   GLuint Texture[MAX_TEXTURE_UNITS];
    GLuint TexGen[MAX_TEXTURE_UNITS];
 };
 
@@ -828,8 +828,8 @@ struct gl_texture_object {
  * Texture units are new with the multitexture extension.
  */
 struct gl_texture_unit {
-   GLuint Enabled;
-   GLuint ReallyEnabled;
+   GLuint Enabled;              /* bitmask of TEXTURE0_1D, _2D, _3D, _CUBE */
+   GLuint ReallyEnabled;        /* 0 or one of TEXTURE0_1D, _2D, _3D, _CUBE */
 
    GLenum EnvMode;	   /* GL_MODULATE, GL_DECAL, GL_BLEND, GL_COMBINE_EXT */
    GLenum LastEnvMode;
@@ -886,9 +886,7 @@ struct gl_texture_attrib {
    GLuint CurrentUnit;			/* Current texture unit */
    GLuint CurrentTransformUnit;		/* Current texture xform unit */
 
-   /* Bitwise-OR of TEXTURE_XD values */
-   GLuint Enabled;		/* Enabled by the user */
-   GLuint ReallyEnabled;	/* Really enabled (w.r.t. completeness, etc) */
+   GLuint ReallyEnabled;     /* Really enabled (w.r.t. completeness, etc) */
 	
    GLuint LastEnabled;	/* Decide whether enabled has really changed */
 
@@ -1963,6 +1961,10 @@ struct gl_context {
 
    /* Dither disable via MESA_NO_DITHER env var */
    GLboolean NoDither;
+
+#ifdef DEBUG
+   GLboolean Rendering;
+#endif
 };
 
 
@@ -2071,7 +2073,44 @@ do {								\
 
 #define Elements(x) sizeof(x)/sizeof(*(x))
 
-#endif
 
+#ifdef DEBUG
+
+#define RENDER_START(CTX)			\
+   do {						\
+      assert(!(CTX)->Rendering);		\
+      (CTX)->Rendering = GL_TRUE;		\
+      if ((CTX)->Driver.RenderStart) {		\
+         (*(CTX)->Driver.RenderStart)(CTX);	\
+      }						\
+   } while (0)
+
+#define RENDER_FINISH(CTX)			\
+   do {						\
+      assert((CTX)->Rendering);			\
+      (CTX)->Rendering = GL_FALSE;		\
+      if ((CTX)->Driver.RenderFinish) {		\
+         (*(CTX)->Driver.RenderFinish)(CTX);	\
+      }						\
+   } while (0)
+
+#else
+
+#define RENDER_START(CTX)			\
+   do {						\
+      if ((CTX)->Driver.RenderStart) {		\
+         (*(CTX)->Driver.RenderStart)(CTX);	\
+      }						\
+   } while (0)
+
+#define RENDER_FINISH(CTX)			\
+   do {						\
+      if ((CTX)->Driver.RenderFinish) {		\
+         (*(CTX)->Driver.RenderFinish)(CTX);	\
+      }						\
+   } while (0)
+
+#endif
 
 
+#endif /* TYPES_H */
diff --git a/xc/extras/Mesa/src/varray.c b/xc/extras/Mesa/src/varray.c
index ce8a4a186..7547ea3bf 100644
--- a/xc/extras/Mesa/src/varray.c
+++ b/xc/extras/Mesa/src/varray.c
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/varray.c,v 1.6 2000/09/26 15:56:34 tsi Exp $ */
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -1124,14 +1123,14 @@ _mesa_InterleavedArrays(GLenum format, GLsizei stride, const GLvoid *pointer)
          glTexCoordPointer( tcomps, GL_FLOAT, stride,
                              (GLubyte *) pointer + i * coffset );
       }
-      for (i = factor; i < ctx->Const.MaxTextureUnits; i++) {
+      for (i = factor; i < (GLint) ctx->Const.MaxTextureUnits; i++) {
          _mesa_ClientActiveTextureARB( (GLenum) (GL_TEXTURE0_ARB + i) );
          _mesa_DisableClientState( GL_TEXTURE_COORD_ARRAY );
       }
    }
    else {
       GLint i;
-      for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
+      for (i = 0; i < (GLint) ctx->Const.MaxTextureUnits; i++) {
          _mesa_ClientActiveTextureARB( (GLenum) (GL_TEXTURE0_ARB + i) );
          _mesa_DisableClientState( GL_TEXTURE_COORD_ARRAY );
       }
diff --git a/xc/extras/Mesa/src/vb.c b/xc/extras/Mesa/src/vb.c
index 33139e8f3..b5555317e 100644
--- a/xc/extras/Mesa/src/vb.c
+++ b/xc/extras/Mesa/src/vb.c
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/vb.c,v 1.6 2000/09/26 15:56:35 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
@@ -207,7 +207,6 @@ void gl_vb_free( struct vertex_buffer *VB )
       ALIGN_FREE( VB->CullMask );
       ALIGN_FREE( VB->NormCullMask );
    } else {
-      if (VB->store.Elt) 
       gl_vector4f_free( VB->store.Obj );          FREE( VB->store.Obj );
       gl_vector3f_free( VB->store.Normal );       FREE( VB->store.Normal );
       gl_vector1ub_free( VB->store.EdgeFlag );    FREE( VB->store.EdgeFlag );
diff --git a/xc/extras/Mesa/src/vb.h b/xc/extras/Mesa/src/vb.h
index 1d32b13f1..62ce7c613 100644
--- a/xc/extras/Mesa/src/vb.h
+++ b/xc/extras/Mesa/src/vb.h
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  3.3
+ * Version:  3.4
  * 
- * Copyright (C) 1999  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2000  Brian Paul   All Rights Reserved.
  * 
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -139,10 +139,10 @@ struct immediate
    GLuint ArrayEltFlush;
    GLuint FlushElt;
 
-   GLuint TF1[2];		/* precalc'ed for glTexCoord */
-   GLuint TF2[2];
-   GLuint TF3[2];
-   GLuint TF4[2];
+   GLuint TF1[MAX_TEXTURE_UNITS];	/* precalc'ed for glTexCoord */
+   GLuint TF2[MAX_TEXTURE_UNITS];
+   GLuint TF3[MAX_TEXTURE_UNITS];
+   GLuint TF4[MAX_TEXTURE_UNITS];
 
    GLuint  Primitive[VB_SIZE];	/* GLubyte would do... */
    GLuint  NextPrimitive[VB_SIZE]; 
diff --git a/xc/extras/Mesa/src/vbcull.c b/xc/extras/Mesa/src/vbcull.c
index 4af76165b..ba9524611 100644
--- a/xc/extras/Mesa/src/vbcull.c
+++ b/xc/extras/Mesa/src/vbcull.c
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/vbcull.c,v 1.6 2000/09/26 15:56:35 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
@@ -834,7 +833,7 @@ GLuint gl_cull_vb( struct vertex_buffer *VB )
       first = i;
       next = VB->NextPrimitive[i];
       prim = in_prim[i];
-      n = cull_funcs[prim]( VB, i, next, parity, (const GLfloat (*)[4])proj );
+      n = cull_funcs[prim]( VB, i, next, parity, (CONST GLfloat (*)[4])proj );
 
       if (n == next - i)
 	 out_prim[i] = GL_POLYGON+1; 
@@ -850,7 +849,7 @@ GLuint gl_cull_vb( struct vertex_buffer *VB )
 					   first, 
 					   VB->Count, 
 					   VB->Ovf, 
-					   (const GLfloat (*)[4])proj );     
+					   (CONST GLfloat (*)[4])proj );     
    }
 
    VB->Primitive = out_prim;
@@ -904,7 +903,7 @@ void gl_dont_cull_vb( struct vertex_buffer *VB )
 				 first, 
 				 VB->Count, 
 				 VB->Ovf,
-				 (const GLfloat (*)[4])proj );
+				 (CONST GLfloat (*)[4])proj );
    }
 
    VB->CullDone = 1;
diff --git a/xc/extras/Mesa/src/vbfill.c b/xc/extras/Mesa/src/vbfill.c
index 1bcea531a..50d5806d5 100644
--- a/xc/extras/Mesa/src/vbfill.c
+++ b/xc/extras/Mesa/src/vbfill.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/vbfill.c,v 1.6 2000/09/26 15:56:35 tsi Exp $ */
+
 
 #ifdef PC_HEADER
 #include "all.h"
diff --git a/xc/extras/Mesa/src/vbindirect.c b/xc/extras/Mesa/src/vbindirect.c
index 04639ff9d..79fb29975 100644
--- a/xc/extras/Mesa/src/vbindirect.c
+++ b/xc/extras/Mesa/src/vbindirect.c
@@ -133,10 +133,10 @@ indexed_render_points( struct vertex_buffer *VB,
       const GLubyte *clip = VB->ClipMask;
       for (i = start ; i < count ; i++) 
 	 if (!clip[elt[i]]) 
-	    ctx->Driver.PointsFunc( ctx, elt[i], elt[i] );
+	    ctx->Driver.PointsFunc( ctx, elt[i], elt[i]+1 );
    } else {
       for (i = start ; i < count ; i++) 
-	 ctx->Driver.PointsFunc( ctx, elt[i], elt[i] );
+	 ctx->Driver.PointsFunc( ctx, elt[i], elt[i]+1 );
    }
 }
 
diff --git a/xc/extras/Mesa/src/vbrender.c b/xc/extras/Mesa/src/vbrender.c
index 263dd34fd..ae3a7141a 100644
--- a/xc/extras/Mesa/src/vbrender.c
+++ b/xc/extras/Mesa/src/vbrender.c
@@ -22,7 +22,7 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/vbrender.c,v 1.7 2000/09/24 13:50:22 alanh Exp $ */
+
 
 /*
  * Render points, lines, and polygons.  The only entry point to this
@@ -151,7 +151,7 @@ static void unfilled_polygon( GLcontext *ctx,
          j = vlist[i];
          if (edge_ptr[j] & 0x3) {
 	    edge_ptr[j] &= ~0x3;
-            (*ctx->Driver.PointsFunc)( ctx, j, j );
+            (*ctx->Driver.PointsFunc)( ctx, j, j+1 );
          }
       }
    }
diff --git a/xc/extras/Mesa/src/vbxform.c b/xc/extras/Mesa/src/vbxform.c
index eea0be897..13d47a398 100644
--- a/xc/extras/Mesa/src/vbxform.c
+++ b/xc/extras/Mesa/src/vbxform.c
@@ -680,7 +680,7 @@ void gl_fixup_cassette( GLcontext *ctx, struct immediate *IM )
 	 IM->NormalLengths = (GLfloat *)MALLOC(sizeof(GLfloat) * VB_SIZE);
 
       calc_normal_lengths( IM->NormalLengths + start, 
-			   (const GLfloat (*)[3])(IM->Normal + start), 
+			   (CONST GLfloat (*)[3])(IM->Normal + start), 
 			   IM->Flag + start, 
 			   IM->Count - start);
       
diff --git a/xc/extras/Mesa/src/vector.c b/xc/extras/Mesa/src/vector.c
index a7daed715..954a4235a 100644
--- a/xc/extras/Mesa/src/vector.c
+++ b/xc/extras/Mesa/src/vector.c
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/vector.c,v 1.5 2000/09/26 15:56:35 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
diff --git a/xc/extras/Mesa/src/vertices.c b/xc/extras/Mesa/src/vertices.c
index da139173f..f57958a1c 100644
--- a/xc/extras/Mesa/src/vertices.c
+++ b/xc/extras/Mesa/src/vertices.c
@@ -2,19 +2,19 @@
 /*
  * Mesa 3-D graphics library
  * Version:  3.3
- * 
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -34,7 +34,7 @@
 
 
 #if defined(USE_X86_ASM)
-#include "X86/common_x86asm.h"
+#include "X86/common_x86_asm.h"
 #endif
 
 
@@ -58,7 +58,7 @@
  */
 
 /* The inline 3dnow code seems to give problems with some peoples
- * compiler/binutils. 
+ * compiler/binutils.
  */
 /*  #undef USE_3DNOW_ASM */
 
@@ -69,15 +69,15 @@
 #endif
 
 
-static void _PROJAPI transform_v16(GLfloat *f,
-			  const GLfloat *m,
-			  const GLfloat *obj,
-			  GLuint obj_stride,
-			  GLuint count )
+static void _PROJAPI transform_v16( GLfloat *f,
+				    const GLfloat *m,
+				    const GLfloat *obj,
+				    GLuint obj_stride,
+				    GLuint count )
 {
    GLuint i;
 
-   for (i = 0 ; i < count ; i++, STRIDE_F(obj, obj_stride), f+=16) 
+   for (i = 0 ; i < count ; i++, STRIDE_F(obj, obj_stride), f+=16)
    {
       const GLfloat ox = obj[0], oy = obj[1], oz = obj[2];
       f[0] = m[0] * ox + m[4] * oy + m[8]  * oz + m[12];
@@ -87,71 +87,27 @@ static void _PROJAPI transform_v16(GLfloat *f,
    }
 }
 
-/* Project all vertices upto but not including last.  Guarenteed to be 
- * at least one such vertex.
- */
-static void _PROJAPI project_verts(GLfloat *first,
-			  GLfloat *last,
-			  const GLfloat *m,
-			  GLuint stride )
-{
-   const GLfloat sx = m[0], sy = m[5], sz = m[10];
-   const GLfloat tx = m[12], ty = m[13], tz = m[14];
-   GLfloat *f;
-   
-   for ( f = first ; f != last ; STRIDE_F(f,stride))
-   {
-      const GLfloat oow = 1.0F / f[3];
-      f[0] = sx * f[0] * oow + tx;
-      f[1] = sy * f[1] * oow + ty;
-      f[2] = sz * f[2] * oow + tz;
-      f[3] = oow;
-   }
-}
-
-static void _PROJAPI project_clipped_verts(GLfloat *first,
-				  GLfloat *last,
-				  const GLfloat *m,
-				  GLuint stride,
-				  const GLubyte *clipmask )
-{
-   const GLfloat sx = m[0], sy = m[5], sz = m[10];
-   const GLfloat tx = m[12], ty = m[13], tz = m[14];
-   GLfloat *f;
-   
-   for ( f = first ; f != last ; STRIDE_F(f,stride), clipmask++)
-   {
-      if (!*clipmask) {
-	 const GLfloat oow = 1.0F / f[3];
-	 f[0] = sx * f[0] * oow + tx;
-	 f[1] = sy * f[1] * oow + ty;
-	 f[2] = sz * f[2] * oow + tz;
-	 f[3] = oow;
-      }
-   }
-}
-
 
 
 
 static void _PROJAPI cliptest_v16( GLfloat *first,
-			  GLfloat *last,
-			  GLubyte *p_clipOr,
-			  GLubyte *p_clipAnd,
-			  GLubyte *clipmask )
-{	
+				   GLfloat *last,
+				   GLubyte *p_clipOr,
+				   GLubyte *p_clipAnd,
+				   GLubyte *clipmask )
+{
    GLubyte clipAnd = (GLubyte) ~0;
    GLubyte clipOr = 0;
    GLfloat *f = first;
    static int i;
    i = 0;
 
-   for ( ; f != last ; f+=16, clipmask++, i++) 
+   for ( ; f != last ; f+=16, clipmask++, i++)
    {
       const GLfloat cx = f[0];
       const GLfloat cy = f[1];
       const GLfloat cz = f[2];
-      const GLfloat cw = f[3];      
+      const GLfloat cw = f[3];
       GLubyte mask = 0;
 
       if (cx >  cw) mask |= CLIP_RIGHT_BIT;
@@ -172,6 +128,51 @@ static void _PROJAPI cliptest_v16( GLfloat *first,
 
 
 
+/* Project all vertices upto but not including last.  Guarenteed to be
+ * at least one such vertex.
+ */
+static void _PROJAPI project_verts( GLfloat *first,
+				    GLfloat *last,
+				    const GLfloat *m,
+				    GLuint stride )
+{
+   const GLfloat sx = m[0], sy = m[5], sz = m[10];
+   const GLfloat tx = m[12], ty = m[13], tz = m[14];
+   GLfloat *f;
+
+   for ( f = first ; f != last ; STRIDE_F(f,stride))
+   {
+      const GLfloat oow = 1.0F / f[3];
+      f[0] = sx * f[0] * oow + tx;
+      f[1] = sy * f[1] * oow + ty;
+      f[2] = sz * f[2] * oow + tz;
+      f[3] = oow;
+   }
+}
+
+static void _PROJAPI project_clipped_verts( GLfloat *first,
+					    GLfloat *last,
+					    const GLfloat *m,
+					    GLuint stride,
+					    const GLubyte *clipmask )
+{
+   const GLfloat sx = m[0], sy = m[5], sz = m[10];
+   const GLfloat tx = m[12], ty = m[13], tz = m[14];
+   GLfloat *f;
+
+   for ( f = first ; f != last ; STRIDE_F(f,stride), clipmask++)
+   {
+      if (!*clipmask) {
+	 const GLfloat oow = 1.0F / f[3];
+	 f[0] = sx * f[0] * oow + tx;
+	 f[1] = sy * f[1] * oow + ty;
+	 f[2] = sz * f[2] * oow + tz;
+	 f[3] = oow;
+      }
+   }
+}
+
+
 GLenum gl_reduce_prim[GL_POLYGON+1] = {
    GL_POINTS,
    GL_LINES,
@@ -183,13 +184,28 @@ GLenum gl_reduce_prim[GL_POLYGON+1] = {
    GL_TRIANGLES,
    GL_TRIANGLES,
    GL_TRIANGLES,
-};  
+};
+
 
-gl_transform_func gl_xform_points3_v16_general = transform_v16;
-gl_cliptest_func gl_cliptest_points4_v16 = cliptest_v16;
-gl_project_clipped_func gl_project_clipped_v16 = project_clipped_verts;
-gl_project_func gl_project_v16 = project_verts;
+gl_transform_func	gl_xform_points3_v16_general;
+gl_cliptest_func	gl_cliptest_points4_v16;
+gl_project_func		gl_project_v16;
+gl_project_clipped_func gl_project_clipped_v16;
 
-void gl_init_vertices()
+
+void gl_init_vertices( void )
 {
+   gl_xform_points3_v16_general	= transform_v16;
+   gl_cliptest_points4_v16	= cliptest_v16;
+   gl_project_v16		= project_verts;
+   gl_project_clipped_v16	= project_clipped_verts;
+
+#if 0
+   /* GH: Add tests/benchmarks for the vertex asm */
+   gl_test_all_vertex_functions( "default" );
+#endif
+
+#ifdef USE_X86_ASM
+   gl_init_all_x86_vertex_asm();
+#endif
 }
diff --git a/xc/extras/Mesa/src/vertices.h b/xc/extras/Mesa/src/vertices.h
index 8bdcf470c..b68646227 100644
--- a/xc/extras/Mesa/src/vertices.h
+++ b/xc/extras/Mesa/src/vertices.h
@@ -2,19 +2,19 @@
 /*
  * Mesa 3-D graphics library
  * Version:  3.3
- * 
+ *
  * Copyright (C) 1999  Brian Paul   All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
  * to deal in the Software without restriction, including without limitation
  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  * and/or sell copies of the Software, and to permit persons to whom the
  * Software is furnished to do so, subject to the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice shall be included
  * in all copies or substantial portions of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
@@ -24,8 +24,8 @@
  */
 
 
-#ifndef VERTICES_H_
-#define VERTICES_H_
+#ifndef __VERTICES_H__
+#define __VERTICES_H__
 
 #ifdef USE_X86_ASM
 #define _PROJAPI _ASMAPI
@@ -36,45 +36,44 @@
 #endif
 
 typedef void (_PROJAPIP gl_transform_func)( GLfloat *first_vert,
-				   const GLfloat *m,
-				   const GLfloat *src,
-				   GLuint src_stride,
-				   GLuint count );
+					    const GLfloat *m,
+					    const GLfloat *src,
+					    GLuint src_stride,
+					    GLuint count );
 
 typedef void (_PROJAPIP gl_cliptest_func)( GLfloat *first_vert,
-				  GLfloat *last_vert, /* use count instead? */
-				  GLubyte *or_mask,
-				  GLubyte *and_mask,
-				  GLubyte *clip_mask );
+					   GLfloat *last_vert, /* use count instead? */
+					   GLubyte *or_mask,
+					   GLubyte *and_mask,
+					   GLubyte *clip_mask );
+
+typedef void (_PROJAPIP gl_project_func)( GLfloat *first,
+					  GLfloat *last,
+					  const GLfloat *m,
+					  GLuint stride );
 
 typedef void (_PROJAPIP gl_project_clipped_func)( GLfloat *first,
-					 GLfloat *last,
-					 const GLfloat *m,
-					 GLuint stride,
-					 const GLubyte *clipmask ); 
+						  GLfloat *last,
+						  const GLfloat *m,
+						  GLuint stride,
+						  const GLubyte *clipmask );
 
-typedef void (_PROJAPIP gl_project_func)( GLfloat *first,
-				 GLfloat *last,
-				 const GLfloat *m,
-				 GLuint stride );
+typedef void (_PROJAPIP gl_vertex_interp_func)( GLfloat t,
+						GLfloat *result,
+						const GLfloat *in,
+						const GLfloat *out );
 
 
 /* At the moment these are used by fastpaths in the FX and MGA drivers.
  */
 extern gl_transform_func       gl_xform_points3_v16_general;
-extern gl_cliptest_func        gl_cliptest_points4_v16;			      
-extern gl_project_clipped_func gl_project_clipped_v16;			      
-extern gl_project_func         gl_project_v16;			      
-
+extern gl_cliptest_func        gl_cliptest_points4_v16;
+extern gl_project_func         gl_project_v16;
+extern gl_project_clipped_func gl_project_clipped_v16;
 
 
 extern GLenum gl_reduce_prim[];
-extern void gl_init_vertices(void);
 
-typedef void (*gl_vertex_interp_func)( GLfloat t, 
-				       GLfloat *result,
-				       const GLfloat *in,
-				       const GLfloat *out );
-				 
+extern void gl_init_vertices( void );
 
 #endif
diff --git a/xc/extras/Mesa/src/xform.c b/xc/extras/Mesa/src/xform.c
index 78309e93c..f057c5058 100644
--- a/xc/extras/Mesa/src/xform.c
+++ b/xc/extras/Mesa/src/xform.c
@@ -52,7 +52,7 @@
 #endif
 
 #ifdef USE_X86_ASM
-#include "X86/common_x86asm.h"
+#include "X86/common_x86_asm.h"
 #endif
 
 clip_func gl_clip_tab[5];
@@ -201,12 +201,12 @@ void gl_init_transformation( void )
    init_dotprod_masked();
 
 #ifdef DEBUG
-   gl_test_all_transform_functions ("default");
-   gl_test_all_normal_transform_functions ("default");
+   gl_test_all_transform_functions( "default" );
+   gl_test_all_normal_transform_functions( "default" );
 #endif
 
 #ifdef USE_X86_ASM
-   gl_init_all_x86_asm ();
+   gl_init_all_x86_transform_asm();
 #endif
 }
 
diff --git a/xc/extras/Mesa/src/xform.h b/xc/extras/Mesa/src/xform.h
index ce2166901..2f8843815 100644
--- a/xc/extras/Mesa/src/xform.h
+++ b/xc/extras/Mesa/src/xform.h
@@ -22,7 +22,10 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/xform.h,v 1.5 2000/09/26 15:56:35 tsi Exp $ */
+
+
+
+
 
 #ifndef XFORM_H
 #define XFORM_H
diff --git a/xc/extras/Mesa/src/xform_tmp.h b/xc/extras/Mesa/src/xform_tmp.h
index 8f521fa11..aacbee841 100644
--- a/xc/extras/Mesa/src/xform_tmp.h
+++ b/xc/extras/Mesa/src/xform_tmp.h
@@ -22,7 +22,6 @@
  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
-/* $XFree86: xc/extras/Mesa/src/xform_tmp.h,v 1.5 2000/09/26 15:56:35 tsi Exp $ */
 
 /*
  * New (3.1) transformation code written by Keith Whitwell.
diff --git a/xc/lib/GL/Imakefile b/xc/lib/GL/Imakefile
index 59af03fed..5814fb1d7 100644
--- a/xc/lib/GL/Imakefile
+++ b/xc/lib/GL/Imakefile
@@ -31,37 +31,43 @@ NormalLintTarget($(PROFSRC))
 #endif
 
 #if BuildXF86DRI && defined(i386Architecture)
-   GLAPI_OBJ = glx/glapi_x86.o
+    GLAPI_OBJ = glx/glapi_x86.o
 #endif
 
-     GLXOBJS = glx/?*.o
-    GLXUOBJS = glx/unshared/?*.o $(GLAPI_OBJ)
-    GLXDOBJS = glx/debugger/?*.o $(GLAPI_OBJ)
-    GLXPOBJS = glx/profiled/?*.o $(GLAPI_OBJ)
-    GLXDONES = glx/DONE
+      GLXOBJS = glx/?*.o
+     GLXUOBJS = glx/unshared/?*.o $(GLAPI_OBJ)
+     GLXDOBJS = glx/debugger/?*.o $(GLAPI_OBJ)
+     GLXPOBJS = glx/profiled/?*.o $(GLAPI_OBJ)
+     GLXDONES = glx/DONE
 
 #if BuildXF86DRI
-     DRIOBJS = dri/XF86dri.o dri/dri_glx.o
-    DRIUOBJS = dri/unshared/XF86dri.o dri/unshared/dri_glx.o
-    DRIDOBJS = dri/debugger/XF86dri.o dri/debugger/dri_glx.o
-    DRIPOBJS = dri/profiled/XF86dri.o dri/profiled/dri_glx.o
-    DRIDONES = dri/DONE
-
-     DRMOBJS = dri/drm/?*.o
-    DRMUOBJS = dri/drm/unshared/?*.o
-    DRMDOBJS = dri/drm/debugger/?*.o
-    DRMPOBJS = dri/drm/profiled/?*.o
-    DRMDONES = dri/drm/DONE
-
-    MESAOBJS = mesa/src/?*.o
+      DRIOBJS = dri/XF86dri.o dri/dri_glx.o
+     DRIUOBJS = dri/unshared/XF86dri.o dri/unshared/dri_glx.o
+     DRIDOBJS = dri/debugger/XF86dri.o dri/debugger/dri_glx.o
+     DRIPOBJS = dri/profiled/XF86dri.o dri/profiled/dri_glx.o
+     DRIDONES = dri/DONE
+
+      DRMOBJS = dri/drm/?*.o
+     DRMUOBJS = dri/drm/unshared/?*.o
+     DRMDOBJS = dri/drm/debugger/?*.o
+     DRMPOBJS = dri/drm/profiled/?*.o
+     DRMDONES = dri/drm/DONE
+
+     MESAOBJS = mesa/src/?*.o
 #ifdef i386Architecture
-    ASM_OBJS =  mesa/src/X86/common_x86asm.o \
-		mesa/src/X86/mmx_blend.o \
-		mesa/src/X86/vertex.o \
-		mesa/src/X86/x86a.o
+     ASM_OBJS = mesa/src/X86/common_x86_asm.o \
+		mesa/src/X86/x86_cliptest.o \
+		mesa/src/X86/x86_vertex.o \
+		mesa/src/X86/x86_xform_masked2.o \
+		mesa/src/X86/x86_xform_masked3.o \
+		mesa/src/X86/x86_xform_masked4.o \
+		mesa/src/X86/x86_xform_raw2.o \
+		mesa/src/X86/x86_xform_raw3.o \
+		mesa/src/X86/x86_xform_raw4.o \
+		mesa/src/X86/mmx_blend.o
 #ifdef MesaUse3DNow
-    ASM_OBJS += mesa/src/X86/3dnow_norm.o \
-		mesa/src/X86/3dnow_norm_raw.o \
+    ASM_OBJS += mesa/src/X86/3dnow_norm_raw.o \
+		mesa/src/X86/3dnow_vertex.o \
 		mesa/src/X86/3dnow_xform_masked1.o \
 		mesa/src/X86/3dnow_xform_masked2.o \
 		mesa/src/X86/3dnow_xform_masked3.o \
@@ -69,12 +75,11 @@ NormalLintTarget($(PROFSRC))
 		mesa/src/X86/3dnow_xform_raw1.o \
 		mesa/src/X86/3dnow_xform_raw2.o \
 		mesa/src/X86/3dnow_xform_raw3.o \
-		mesa/src/X86/3dnow_xform_raw4.o \
-		mesa/src/X86/vertex_3dnow.o
+		mesa/src/X86/3dnow_xform_raw4.o
 #endif
 #ifdef MesaUseKatmai
-    ASM_OBJS += mesa/src/X86/katmai_norm.o \
-		mesa/src/X86/katmai_norm_raw.o \
+    ASM_OBJS += mesa/src/X86/katmai_norm_raw.o \
+		mesa/src/X86/katmai_vertex.o \
 		mesa/src/X86/katmai_xform_masked1.o \
 		mesa/src/X86/katmai_xform_masked2.o \
 		mesa/src/X86/katmai_xform_masked3.o \
@@ -82,8 +87,7 @@ NormalLintTarget($(PROFSRC))
 		mesa/src/X86/katmai_xform_raw1.o \
 		mesa/src/X86/katmai_xform_raw2.o \
 		mesa/src/X86/katmai_xform_raw3.o \
-		mesa/src/X86/katmai_xform_raw4.o \
-		mesa/src/X86/vertex_katmai.o
+		mesa/src/X86/katmai_xform_raw4.o
 #endif
 
     MESAOBJS = mesa/src/?*.o mesa/src/X86/?*.o
diff --git a/xc/lib/GL/mesa/src/Imakefile b/xc/lib/GL/mesa/src/Imakefile
index f8232128f..07d28a17c 100644
--- a/xc/lib/GL/mesa/src/Imakefile
+++ b/xc/lib/GL/mesa/src/Imakefile
@@ -349,12 +349,12 @@ LinkSourceFile(zoom.h, $(MESASRCDIR)/src)
      ASM_SRCS = 
      ASM_OBJS = 
 #ifdef MesaUse3DNow
-     3DNOW_DEFS = -DUSE_3DNOW_ASM
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
 #endif
 #ifdef MesaUseKatmai
-     KATMAI_DEFS = -DUSE_KATMAI_ASM
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
 #endif
-     ASM_DEFS = -DUSE_MMX_ASM -DUSE_X86_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
+     ASM_DEFS = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
 #endif
 
 #ifdef UseCompaqMathLibrary
diff --git a/xc/lib/GL/mesa/src/OSmesa/Imakefile b/xc/lib/GL/mesa/src/OSmesa/Imakefile
index a38ddd585..e09928541 100644
--- a/xc/lib/GL/mesa/src/OSmesa/Imakefile
+++ b/xc/lib/GL/mesa/src/OSmesa/Imakefile
@@ -171,26 +171,38 @@ MESA_INCLUDES = -I. -I.. -I../../include -I../../../../../include
 		osmesa.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../X86/x86.c \
-		../X86/x86a.S \
-		../X86/common_x86.c \
-		../X86/common_x86asm.S \
-		../X86/vertex.S
-
-     X86_OBJS = ../X86/x86.o \
-		../X86/x86a.o \
-		../X86/common_x86.o \
-		../X86/common_x86asm.o \
-		../X86/vertex.o
+     X86_SRCS = ../X86/common_x86.c \
+		../X86/common_x86_asm.S \
+		../X86/x86.c \
+		../X86/x86_cliptest.S \
+		../X86/x86_vertex.S \
+		../X86/x86_xform_masked2.S \
+		../X86/x86_xform_masked3.S \
+		../X86/x86_xform_masked4.S \
+		../X86/x86_xform_raw2.S \
+		../X86/x86_xform_raw3.S \
+		../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../X86/common_x86.o \
+		../X86/common_x86_asm.o \
+		../X86/x86.o \
+		../X86/x86_cliptest.o \
+		../X86/x86_vertex.o \
+		../X86/x86_xform_masked2.o \
+		../X86/x86_xform_masked3.o \
+		../X86/x86_xform_masked4.o \
+		../X86/x86_xform_raw2.o \
+		../X86/x86_xform_raw3.o \
+		../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../X86/mmx_blend.S
 
      MMX_OBJS = ../X86/mmx_blend.o
 
-XCOMM Disabling 3Dnow code for the time being.
-#if 0
+#ifdef MesaUse3DNow
    3DNOW_SRCS = ../X86/3dnow.c \
 		../X86/3dnow_norm_raw.S \
+		../X86/3dnow_vertex.S \
 		../X86/3dnow_xform_masked1.S \
 		../X86/3dnow_xform_masked2.S \
 		../X86/3dnow_xform_masked3.S \
@@ -198,11 +210,11 @@ XCOMM Disabling 3Dnow code for the time being.
 		../X86/3dnow_xform_raw1.S \
 		../X86/3dnow_xform_raw2.S \
 		../X86/3dnow_xform_raw3.S \
-		../X86/3dnow_xform_raw4.S \
-		../X86/vertex_3dnow.S
+		../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../X86/3dnow.o \
 		../X86/3dnow_norm_raw.o \
+		../X86/3dnow_vertex.o \
 		../X86/3dnow_xform_masked1.o \
 		../X86/3dnow_xform_masked2.o \
 		../X86/3dnow_xform_masked3.o \
@@ -210,14 +222,38 @@ XCOMM Disabling 3Dnow code for the time being.
 		../X86/3dnow_xform_raw1.o \
 		../X86/3dnow_xform_raw2.o \
 		../X86/3dnow_xform_raw3.o \
-		../X86/3dnow_xform_raw4.o \
-		../X86/vertex_3dnow.o
+		../X86/3dnow_xform_raw4.o
 #endif
 
+#ifdef MesaUseKatmai
+  KATMAI_SRCS = ../X86/katmai.c \
+		../X86/katmai_norm_raw.S \
+		../X86/katmai_vertex.S \
+		../X86/katmai_xform_masked1.S \
+		../X86/katmai_xform_masked2.S \
+		../X86/katmai_xform_masked3.S \
+		../X86/katmai_xform_masked4.S \
+		../X86/katmai_xform_raw1.S \
+		../X86/katmai_xform_raw2.S \
+		../X86/katmai_xform_raw3.S \
+		../X86/katmai_xform_raw4.S
+
+  KATMAI_OBJS = ../X86/katmai.o \
+		../X86/katmai_norm_raw.o \
+		../X86/katmai_vertex.o \
+		../X86/katmai_xform_masked1.o \
+		../X86/katmai_xform_masked2.o \
+		../X86/katmai_xform_masked3.o \
+		../X86/katmai_xform_masked4.o \
+		../X86/katmai_xform_raw1.o \
+		../X86/katmai_xform_raw2.o \
+		../X86/katmai_xform_raw3.o \
+		../X86/katmai_xform_raw4.o
+#endif
 #endif
 
-      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS)
-      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS)
+      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
+      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
 
          SRCS = $(MESASRCS) $(ASMSRCS)
          OBJS = $(MESAOBJS) $(ASMOBJS)
diff --git a/xc/lib/GL/mesa/src/X86/Imakefile b/xc/lib/GL/mesa/src/X86/Imakefile
index e3239ecd8..3f354c901 100644
--- a/xc/lib/GL/mesa/src/X86/Imakefile
+++ b/xc/lib/GL/mesa/src/X86/Imakefile
@@ -8,14 +8,24 @@ XCOMM $XFree86: xc/lib/GL/mesa/src/X86/Imakefile,v 1.11 2000/09/24 13:51:02 alan
 
 LinkSourceFile(assyntax.h, $(MESASRCDIR)/src/X86)
 LinkSourceFile(common_x86.c, $(MESASRCDIR)/src/X86)
-LinkSourceFile(common_x86asm.S, $(MESASRCDIR)/src/X86)
-LinkSourceFile(common_x86asm.h, $(MESASRCDIR)/src/X86)
+LinkSourceFile(common_x86_asm.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(common_x86_asm.h, $(MESASRCDIR)/src/X86)
+LinkSourceFile(common_x86_features.h, $(MESASRCDIR)/src/X86)
 LinkSourceFile(glapi_x86.S, $(MESASRCDIR)/src/X86)
+
+LinkSourceFile(xform_args.h, $(MESASRCDIR)/src/X86)
+LinkSourceFile(clip_args.h, $(MESASRCDIR)/src/X86)
+
 LinkSourceFile(x86.c, $(MESASRCDIR)/src/X86)
 LinkSourceFile(x86.h, $(MESASRCDIR)/src/X86)
-LinkSourceFile(x86a.S, $(MESASRCDIR)/src/X86)
-LinkSourceFile(vertex.S, $(MESASRCDIR)/src/X86)
-LinkSourceFile(x86flatregs.m4, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_cliptest.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_vertex.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_xform_masked2.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_xform_masked3.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_xform_masked4.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_xform_raw2.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_xform_raw3.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(x86_xform_raw4.S, $(MESASRCDIR)/src/X86)
 
 LinkSourceFile(mmx.h, $(MESASRCDIR)/src/X86)
 LinkSourceFile(mmx_blend.S, $(MESASRCDIR)/src/X86)
@@ -23,6 +33,7 @@ LinkSourceFile(mmx_blend.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow.c, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow.h, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_norm_raw.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(3dnow_vertex.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_xform_masked1.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_xform_masked2.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_xform_masked3.S, $(MESASRCDIR)/src/X86)
@@ -31,11 +42,11 @@ LinkSourceFile(3dnow_xform_raw1.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_xform_raw2.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_xform_raw3.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(3dnow_xform_raw4.S, $(MESASRCDIR)/src/X86)
-LinkSourceFile(vertex_3dnow.S, $(MESASRCDIR)/src/X86)
 
 LinkSourceFile(katmai.c, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai.h, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_norm_raw.S, $(MESASRCDIR)/src/X86)
+LinkSourceFile(katmai_vertex.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_xform_masked1.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_xform_masked2.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_xform_masked3.S, $(MESASRCDIR)/src/X86)
@@ -44,7 +55,6 @@ LinkSourceFile(katmai_xform_raw1.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_xform_raw2.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_xform_raw3.S, $(MESASRCDIR)/src/X86)
 LinkSourceFile(katmai_xform_raw4.S, $(MESASRCDIR)/src/X86)
-LinkSourceFile(vertex_katmai.S, $(MESASRCDIR)/src/X86)
 
 LinkSourceFile(glapioffsets.h, $(MESASRCDIR)/src)
 
@@ -53,50 +63,95 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 #endif
 
 #ifdef i386Architecture
-XCOMM Determine at runtime whether 3dNow, Katmai, MMX, etc are really present.
-      X86_SRCS = x86a.S common_x86.c common_x86asm.S glapi_x86.S x86.c vertex.S
-
-      X86_OBJS = x86a.o common_x86.o common_x86asm.o x86.o vertex.o
-
-      X86_DEFS = -DUSE_X86_ASM
-
-
-      MMX_SRCS = mmx_blend.S
-
-      MMX_OBJS = mmx_blend.o
-
-      MMX_DEFS = -DUSE_MMX_ASM
+XCOMM Determine at runtime whether 3DNow!, Katmai, MMX, etc are really present.
+     X86_SRCS = common_x86.c \
+		common_x86_asm.S \
+		glapi_x86.S \
+		x86.c \
+		x86_cliptest.S \
+		x86_vertex.S \
+		x86_xform_masked2.S \
+		x86_xform_masked3.S \
+		x86_xform_masked4.S \
+		x86_xform_raw2.S \
+		x86_xform_raw3.S \
+		x86_xform_raw4.S
+
+     X86_OBJS = common_x86.o \
+		common_x86_asm.o \
+		x86.o \
+		x86_cliptest.o \
+		x86_vertex.o \
+		x86_xform_masked2.o \
+		x86_xform_masked3.o \
+		x86_xform_masked4.o \
+		x86_xform_raw2.o \
+		x86_xform_raw3.o \
+		x86_xform_raw4.o
+
+     X86_DEFS = -DUSE_X86_ASM
+
+
+     MMX_SRCS = mmx_blend.S
+
+     MMX_OBJS = mmx_blend.o
+
+     MMX_DEFS = -DUSE_MMX_ASM
 
 #ifdef MesaUse3DNow
-    3DNOW_SRCS = 3dnow.c 3dnow_norm_raw.S 3dnow_xform_masked1.S \
-		 3dnow_xform_masked2.S 3dnow_xform_masked3.S \
-		 3dnow_xform_masked4.S 3dnow_xform_raw1.S \
-		 3dnow_xform_raw2.S 3dnow_xform_raw3.S 3dnow_xform_raw4.S \
-		 vertex_3dnow.S
-
-    3DNOW_OBJS = 3dnow.o 3dnow_norm_raw.o 3dnow_xform_masked1.o \
-		 3dnow_xform_masked2.o 3dnow_xform_masked3.o \
-		 3dnow_xform_masked4.o 3dnow_xform_raw1.o \
-		 3dnow_xform_raw2.o 3dnow_xform_raw3.o 3dnow_xform_raw4.o \
-		 vertex_3dnow.o
-
-    3DNOW_DEFS = -DUSE_3DNOW_ASM
+   3DNOW_SRCS = 3dnow.c \
+		3dnow_norm_raw.S \
+		3dnow_vertex.S \
+		3dnow_xform_masked1.S \
+		3dnow_xform_masked2.S \
+		3dnow_xform_masked3.S \
+		3dnow_xform_masked4.S \
+		3dnow_xform_raw1.S \
+		3dnow_xform_raw2.S \
+		3dnow_xform_raw3.S \
+		3dnow_xform_raw4.S \
+
+   3DNOW_OBJS = 3dnow.o \
+		3dnow_norm_raw.o \
+		3dnow_vertex.o \
+		3dnow_xform_masked1.o \
+		3dnow_xform_masked2.o \
+		3dnow_xform_masked3.o \
+		3dnow_xform_masked4.o \
+		3dnow_xform_raw1.o \
+		3dnow_xform_raw2.o \
+		3dnow_xform_raw3.o \
+		3dnow_xform_raw4.o
+
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
 #endif
 
 #ifdef MesaUseKatmai
-    KATMAI_SRCS = katmai.c katmai_norm_raw.S katmai_xform_masked1.S \
-		 katmai_xform_masked2.S katmai_xform_masked3.S \
-		 katmai_xform_masked4.S katmai_xform_raw1.S \
-		 katmai_xform_raw2.S katmai_xform_raw3.S katmai_xform_raw4.S \
-		 vertex_katmai.S
-
-    KATMAI_OBJS = katmai.o katmai_norm_raw.o katmai_xform_masked1.o \
-		 katmai_xform_masked2.o katmai_xform_masked3.o \
-		 katmai_xform_masked4.o katmai_xform_raw1.o \
-		 katmai_xform_raw2.o katmai_xform_raw3.o katmai_xform_raw4.o \
-		 vertex_katmai.o
-
-    KATMAI_DEFS = -DUSE_KATMAI_ASM
+  KATMAI_SRCS = katmai.c \
+		katmai_norm_raw.S \
+		katmai_vertex.S \
+		katmai_xform_masked1.S \
+		katmai_xform_masked2.S \
+		katmai_xform_masked3.S \
+		katmai_xform_masked4.S \
+		katmai_xform_raw1.S \
+		katmai_xform_raw2.S \
+		katmai_xform_raw3.S \
+		katmai_xform_raw4.S \
+
+  KATMAI_OBJS = katmai.o \
+		katmai_norm_raw.o \
+		katmai_vertex.o \
+		katmai_xform_masked1.o \
+		katmai_xform_masked2.o \
+		katmai_xform_masked3.o \
+		katmai_xform_masked4.o \
+		katmai_xform_raw1.o \
+		katmai_xform_raw2.o \
+		katmai_xform_raw3.o \
+		katmai_xform_raw4.o
+
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
 #endif
 
 #endif
@@ -122,8 +177,22 @@ STD_CPP_DEFINES = StandardDefines $(PROJECT_DEFINES)
 SubdirLibraryRule($(OBJS))
 NormalLintTarget($(SRCS))
 
+ObjectFromAsmSource(common_x86_asm, NullParameter)
+
+ObjectFromAsmSource(x86_cliptest, NullParameter)
+ObjectFromAsmSource(x86_vertex, NullParameter)
+ObjectFromAsmSource(x86_xform_masked2, NullParameter)
+ObjectFromAsmSource(x86_xform_masked3, NullParameter)
+ObjectFromAsmSource(x86_xform_masked4, NullParameter)
+ObjectFromAsmSource(x86_xform_raw2, NullParameter)
+ObjectFromAsmSource(x86_xform_raw3, NullParameter)
+ObjectFromAsmSource(x86_xform_raw4, NullParameter)
+
+ObjectFromAsmSource(mmx_blend, NullParameter)
+
 #ifdef MesaUse3DNow
 ObjectFromAsmSource(3dnow_norm_raw, NullParameter)
+ObjectFromAsmSource(3dnow_vertex, NullParameter)
 ObjectFromAsmSource(3dnow_xform_masked1, NullParameter)
 ObjectFromAsmSource(3dnow_xform_masked2, NullParameter)
 ObjectFromAsmSource(3dnow_xform_masked3, NullParameter)
@@ -132,11 +201,11 @@ ObjectFromAsmSource(3dnow_xform_raw1, NullParameter)
 ObjectFromAsmSource(3dnow_xform_raw2, NullParameter)
 ObjectFromAsmSource(3dnow_xform_raw3, NullParameter)
 ObjectFromAsmSource(3dnow_xform_raw4, NullParameter)
-ObjectFromAsmSource(vertex_3dnow, NullParameter)
 #endif
 
 #ifdef MesaUseKatmai
 ObjectFromAsmSource(katmai_norm_raw, NullParameter)
+ObjectFromAsmSource(katmai_vertex, NullParameter)
 ObjectFromAsmSource(katmai_xform_masked1, NullParameter)
 ObjectFromAsmSource(katmai_xform_masked2, NullParameter)
 ObjectFromAsmSource(katmai_xform_masked3, NullParameter)
@@ -145,14 +214,6 @@ ObjectFromAsmSource(katmai_xform_raw1, NullParameter)
 ObjectFromAsmSource(katmai_xform_raw2, NullParameter)
 ObjectFromAsmSource(katmai_xform_raw3, NullParameter)
 ObjectFromAsmSource(katmai_xform_raw4, NullParameter)
-ObjectFromAsmSource(vertex_katmai, NullParameter)
 #endif
 
-ObjectFromAsmSource(mmx_blend, NullParameter)
-
-ObjectFromAsmSource(common_x86asm, NullParameter)
-ObjectFromAsmSource(vertex, NullParameter)
-ObjectFromAsmSource(x86a, NullParameter)
-
 DependTarget()
-
diff --git a/xc/lib/GL/mesa/src/drv/gamma/Imakefile b/xc/lib/GL/mesa/src/drv/gamma/Imakefile
index ff31fdf2b..7282cf2fb 100644
--- a/xc/lib/GL/mesa/src/drv/gamma/Imakefile
+++ b/xc/lib/GL/mesa/src/drv/gamma/Imakefile
@@ -12,6 +12,16 @@ XCOMM $XFree86: xc/lib/GL/mesa/src/drv/gamma/Imakefile,v 1.14 2000/10/20 12:57:2
 ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 #endif
 
+#ifdef i386Architecture
+#ifdef MesaUse3DNow
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
+#endif
+#ifdef MesaUseKatmai
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
+#endif
+  ASM_DEFINES = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
+#endif
+
 #if BuildXF86DRI
   DRI_DEFINES = GlxDefines -DDRIVERTS
  DRI_INCLUDES = -I../../../../dri -I../../../../glx \
@@ -26,17 +36,9 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		-I../../../../dri/drm
 
 
-      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES)
+      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(ASM_DEFINES)
      INCLUDES = -I$(XLIBSRC) -I$(EXTINCSRC) $(MESA_INCLUDES) $(DRI_INCLUDES) 
 
-#if 0
-	LOSRC = ../../../../lowpc.c
-	LOOBJ = ../../../../lowpc.o
-
-	HISRC = ../../../../highpc.c
-	HIOBJ = ../../../../highpc.o
-#endif
-
       DRISRCS = ../../../dri/dri_mesa.c \
 		../../../../dri/dri_tmm.c
 
@@ -53,11 +55,21 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../../../dri/drm/xf86drmRandom.o \
 		../../../../dri/drm/xf86drmSL.o
 
-    GAMMASRCS = gamma_gl.c gamma_xmesa.c gamma_init.c gamma_matrix.c \
-                gamma_inithw.c gamma_texture.c /* gamma_dlist.c */
-
-    GAMMAOBJS = gamma_gl.o gamma_xmesa.o gamma_init.o gamma_matrix.o \
-                gamma_inithw.o gamma_texture.o /* gamma_dlist.o */
+    GAMMASRCS = gamma_gl.c \
+		gamma_xmesa.c \
+		gamma_init.c \
+		gamma_matrix.c \
+		gamma_inithw.c \
+		gamma_texture.c \
+		/* gamma_dlist.c */
+
+    GAMMAOBJS = gamma_gl.o \
+		gamma_xmesa.o \
+		gamma_init.o \
+		gamma_matrix.o \
+		gamma_inithw.o \
+		gamma_texture.o \
+		/* gamma_dlist.o */
 
      MESASRCS = ../../aatriangle.c \
 		../../accum.c \
@@ -208,26 +220,38 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../zoom.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../../X86/x86.c \
-		../../X86/x86a.S \
-		../../X86/common_x86.c \
-		../../X86/common_x86asm.S \
-		../../X86/vertex.S
-
-     X86_OBJS = ../../X86/x86.o \
-		../../X86/x86a.o \
-		../../X86/common_x86.o \
-		../../X86/common_x86asm.o \
-		../../X86/vertex.o
+     X86_SRCS = ../../X86/common_x86.c \
+		../../X86/common_x86_asm.S \
+		../../X86/x86.c \
+		../../X86/x86_cliptest.S \
+		../../X86/x86_vertex.S \
+		../../X86/x86_xform_masked2.S \
+		../../X86/x86_xform_masked3.S \
+		../../X86/x86_xform_masked4.S \
+		../../X86/x86_xform_raw2.S \
+		../../X86/x86_xform_raw3.S \
+		../../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../../X86/common_x86.o \
+		../../X86/common_x86_asm.o \
+		../../X86/x86.o \
+		../../X86/x86_cliptest.o \
+		../../X86/x86_vertex.o \
+		../../X86/x86_xform_masked2.o \
+		../../X86/x86_xform_masked3.o \
+		../../X86/x86_xform_masked4.o \
+		../../X86/x86_xform_raw2.o \
+		../../X86/x86_xform_raw3.o \
+		../../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../../X86/mmx_blend.S
 
      MMX_OBJS = ../../X86/mmx_blend.o
 
-XCOMM Disabling 3Dnow code for the time being.
-#if 0
+#ifdef MesaUse3DNow
    3DNOW_SRCS = ../../X86/3dnow.c \
 		../../X86/3dnow_norm_raw.S \
+		../../X86/3dnow_vertex.S \
 		../../X86/3dnow_xform_masked1.S \
 		../../X86/3dnow_xform_masked2.S \
 		../../X86/3dnow_xform_masked3.S \
@@ -235,11 +259,11 @@ XCOMM Disabling 3Dnow code for the time being.
 		../../X86/3dnow_xform_raw1.S \
 		../../X86/3dnow_xform_raw2.S \
 		../../X86/3dnow_xform_raw3.S \
-		../../X86/3dnow_xform_raw4.S \
-		../../X86/vertex_3dnow.S
+		../../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../../X86/3dnow.o \
 		../../X86/3dnow_norm_raw.o \
+		../../X86/3dnow_vertex.o \
 		../../X86/3dnow_xform_masked1.o \
 		../../X86/3dnow_xform_masked2.o \
 		../../X86/3dnow_xform_masked3.o \
@@ -247,17 +271,51 @@ XCOMM Disabling 3Dnow code for the time being.
 		../../X86/3dnow_xform_raw1.o \
 		../../X86/3dnow_xform_raw2.o \
 		../../X86/3dnow_xform_raw3.o \
-		../../X86/3dnow_xform_raw4.o \
-		../../X86/vertex_3dnow.o
+		../../X86/3dnow_xform_raw4.o
 #endif
 
+#ifdef MesaUseKatmai
+  KATMAI_SRCS = ../../X86/katmai.c \
+		../../X86/katmai_norm_raw.S \
+		../../X86/katmai_vertex.S \
+		../../X86/katmai_xform_masked1.S \
+		../../X86/katmai_xform_masked2.S \
+		../../X86/katmai_xform_masked3.S \
+		../../X86/katmai_xform_masked4.S \
+		../../X86/katmai_xform_raw1.S \
+		../../X86/katmai_xform_raw2.S \
+		../../X86/katmai_xform_raw3.S \
+		../../X86/katmai_xform_raw4.S
+
+  KATMAI_OBJS = ../../X86/katmai.o \
+		../../X86/katmai_norm_raw.o \
+		../../X86/katmai_vertex.o \
+		../../X86/katmai_xform_masked1.o \
+		../../X86/katmai_xform_masked2.o \
+		../../X86/katmai_xform_masked3.o \
+		../../X86/katmai_xform_masked4.o \
+		../../X86/katmai_xform_raw1.o \
+		../../X86/katmai_xform_raw2.o \
+		../../X86/katmai_xform_raw3.o \
+		../../X86/katmai_xform_raw4.o
+#endif
 #endif
 
-      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS)
-      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS)
+#ifdef GlxSoProf
+       LOSRCS = ../../../../lowpc.c
+       HISRCS = ../../../../highpc.c
+
+       LOOBJS = ../../../../lowpc.o
+       HIOBJS = ../../../../highpc.o
+#endif
 
-         SRCS = $(LOWSRC) $(DRISRCS) $(DRMSRCS) $(MESASRCS) $(ASMSRCS) $(GAMMASRCS) $(HISRC)
-         OBJS = $(LOWOBJ) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) $(ASMOBJS) $(GAMMAOBJS) $(HIOBJ)
+      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
+      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
+
+         SRCS = $(LOWSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) \
+		$(ASMSRCS) $(GAMMASRCS) $(HISRCS)
+         OBJS = $(LOWOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) \
+		$(ASMOBJS) $(GAMMAOBJS) $(HIOBJS)
 
 REQUIREDLIBS += MathLibrary
 #if !GlxBuiltInGamma
@@ -285,6 +343,12 @@ ALL_OBJS = $(OBJS)
 ALL_DEPS = DONE
 SharedDepModuleTarget($(LIBNAME),$(ALL_DEPS),$(ALL_OBJS))
 InstallDynamicModule($(LIBNAME),$(MODULEDIR),dri)
+
+#ifdef GlxSoProf
+SOPROF_LIBNAME = _gamma_dri_p
+NormalDepLibraryTarget($(SOPROF_LIBNAME),$(ALL_DEPS),$(ALL_OBJS))
+InstallLibrary($(SOPROF_LIBNAME),$(MODULEDIR)/dri)
+#endif
 #endif
 
 DependTarget()
diff --git a/xc/lib/GL/mesa/src/drv/i810/Imakefile b/xc/lib/GL/mesa/src/drv/i810/Imakefile
index b66d715c1..bfe0bc81d 100644
--- a/xc/lib/GL/mesa/src/drv/i810/Imakefile
+++ b/xc/lib/GL/mesa/src/drv/i810/Imakefile
@@ -12,6 +12,16 @@ XCOMM $XFree86: xc/lib/GL/mesa/src/drv/i810/Imakefile,v 1.11 2000/10/20 12:57:22
 ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 #endif
 
+#ifdef i386Architecture
+#ifdef MesaUse3DNow
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
+#endif
+#ifdef MesaUseKatmai
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
+#endif
+  ASM_DEFINES = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
+#endif
+
 #if BuildXF86DRI
   DRI_DEFINES = GlxDefines -DDRIVERTS
  DRI_INCLUDES = -I../../../../dri -I../../../../glx \
@@ -25,7 +35,7 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 MESA_INCLUDES = -I. -I.. -I../../include \
 		-I../../../../dri/drm
 
-      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES)
+      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(ASM_DEFINES)
      INCLUDES = -I$(XLIBSRC) -I$(EXTINCSRC) $(MESA_INCLUDES) $(DRI_INCLUDES) 
 
       DRISRCS = ../../../dri/dri_mesa.c \
@@ -44,23 +54,27 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../../../dri/drm/xf86drmRandom.o \
 		../../../../dri/drm/xf86drmSL.o
 
-#ifdef GlxSoProf
-       HIOBJS = ../../../../highpc.o
-       LOOBJS = ../../../../lowpc.o
-#else
-       HIOBJS = 
-       LOOBJS =
-#endif
-
-     I810SRCS = i810_xmesa.c	\
-		i810dd.c  	\
-		i810pipeline.c i810span.c i810state.c	\
-		i810tex.c i810tris.c i810vb.c i810fastpath.c i810ioctl.c
+     I810SRCS = i810_xmesa.c \
+		i810dd.c \
+		i810pipeline.c \
+		i810span.c \
+		i810state.c \
+		i810tex.c \
+		i810tris.c \
+		i810vb.c \
+		i810fastpath.c \
+		i810ioctl.c
 
      I810OBJS = i810_xmesa.o \
-		i810dd.o  		\
-		i810pipeline.o i810span.o i810state.o	\
-		i810tex.o i810tris.o i810vb.o i810fastpath.o i810ioctl.o
+		i810dd.o \
+		i810pipeline.o \
+		i810span.o \
+		i810state.o \
+		i810tex.o \
+		i810tris.o \
+		i810vb.o \
+		i810fastpath.o \
+		i810ioctl.o
 
      MESASRCS = ../../aatriangle.c \
 		../../accum.c \
@@ -211,26 +225,38 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../zoom.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../../X86/x86.c \
-		../../X86/x86a.S \
-		../../X86/common_x86.c \
-		../../X86/common_x86asm.S \
-		../../X86/vertex.S
-
-     X86_OBJS = ../../X86/x86.o \
-		../../X86/x86a.o \
-		../../X86/common_x86.o \
-		../../X86/common_x86asm.o \
-		../../X86/vertex.o
+     X86_SRCS = ../../X86/common_x86.c \
+		../../X86/common_x86_asm.S \
+		../../X86/x86.c \
+		../../X86/x86_cliptest.S \
+		../../X86/x86_vertex.S \
+		../../X86/x86_xform_masked2.S \
+		../../X86/x86_xform_masked3.S \
+		../../X86/x86_xform_masked4.S \
+		../../X86/x86_xform_raw2.S \
+		../../X86/x86_xform_raw3.S \
+		../../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../../X86/common_x86.o \
+		../../X86/common_x86_asm.o \
+		../../X86/x86.o \
+		../../X86/x86_cliptest.o \
+		../../X86/x86_vertex.o \
+		../../X86/x86_xform_masked2.o \
+		../../X86/x86_xform_masked3.o \
+		../../X86/x86_xform_masked4.o \
+		../../X86/x86_xform_raw2.o \
+		../../X86/x86_xform_raw3.o \
+		../../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../../X86/mmx_blend.S
 
      MMX_OBJS = ../../X86/mmx_blend.o
 
-XCOMM Disabling 3Dnow code for the time being.
 #ifdef MesaUse3DNow
    3DNOW_SRCS = ../../X86/3dnow.c \
 		../../X86/3dnow_norm_raw.S \
+		../../X86/3dnow_vertex.S \
 		../../X86/3dnow_xform_masked1.S \
 		../../X86/3dnow_xform_masked2.S \
 		../../X86/3dnow_xform_masked3.S \
@@ -238,11 +264,11 @@ XCOMM Disabling 3Dnow code for the time being.
 		../../X86/3dnow_xform_raw1.S \
 		../../X86/3dnow_xform_raw2.S \
 		../../X86/3dnow_xform_raw3.S \
-		../../X86/3dnow_xform_raw4.S \
-		../../X86/vertex_3dnow.S
+		../../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../../X86/3dnow.o \
 		../../X86/3dnow_norm_raw.o \
+		../../X86/3dnow_vertex.o \
 		../../X86/3dnow_xform_masked1.o \
 		../../X86/3dnow_xform_masked2.o \
 		../../X86/3dnow_xform_masked3.o \
@@ -250,22 +276,54 @@ XCOMM Disabling 3Dnow code for the time being.
 		../../X86/3dnow_xform_raw1.o \
 		../../X86/3dnow_xform_raw2.o \
 		../../X86/3dnow_xform_raw3.o \
-		../../X86/3dnow_xform_raw4.o \
-		../../X86/vertex_3dnow.o
+		../../X86/3dnow_xform_raw4.o
+#endif
+
+#ifdef MesaUseKatmai
+  KATMAI_SRCS = ../../X86/katmai.c \
+		../../X86/katmai_norm_raw.S \
+		../../X86/katmai_vertex.S \
+		../../X86/katmai_xform_masked1.S \
+		../../X86/katmai_xform_masked2.S \
+		../../X86/katmai_xform_masked3.S \
+		../../X86/katmai_xform_masked4.S \
+		../../X86/katmai_xform_raw1.S \
+		../../X86/katmai_xform_raw2.S \
+		../../X86/katmai_xform_raw3.S \
+		../../X86/katmai_xform_raw4.S
+
+  KATMAI_OBJS = ../../X86/katmai.o \
+		../../X86/katmai_norm_raw.o \
+		../../X86/katmai_vertex.o \
+		../../X86/katmai_xform_masked1.o \
+		../../X86/katmai_xform_masked2.o \
+		../../X86/katmai_xform_masked3.o \
+		../../X86/katmai_xform_masked4.o \
+		../../X86/katmai_xform_raw1.o \
+		../../X86/katmai_xform_raw2.o \
+		../../X86/katmai_xform_raw3.o \
+		../../X86/katmai_xform_raw4.o
 #endif
+#endif
+
+#ifdef GlxSoProf
+       LOSRCS = ../../../../lowpc.c
+       HISRCS = ../../../../highpc.c
 
+       LOOBJS = ../../../../lowpc.o
+       HIOBJS = ../../../../highpc.o
 #endif
 
-      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS)
-      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS)
+      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
+      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
 
-      COMMONSRCS = ../common/mm.c ../common/hwlog.c
-      COMMONOBJS = ../common/mm.o ../common/hwlog.o
+   COMMONSRCS = ../common/mm.c ../common/hwlog.c
+   COMMONOBJS = ../common/mm.o ../common/hwlog.o
 
-         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) $(ASMSRCS)	\
-		$(COMMONSRCS) $(I810SRCS) $(HISRCS)
-         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) $(ASMOBJS)	\
-		$(COMMONOBJS) $(I810OBJS) $(HIOBJS)
+         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) \
+		$(ASMSRCS) $(COMMONSRCS) $(I810SRCS) $(HISRCS)
+         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) \
+		$(ASMOBJS) $(COMMONOBJS) $(I810OBJS) $(HIOBJS)
 
 REQUIREDLIBS += MathLibrary
 #if !GlxBuiltInI810
@@ -301,6 +359,4 @@ InstallLibrary($(SOPROF_LIBNAME),$(MODULEDIR)/dri)
 #endif
 #endif
 
-
-
 DependTarget()
diff --git a/xc/lib/GL/mesa/src/drv/i810/i810tris.h b/xc/lib/GL/mesa/src/drv/i810/i810tris.h
index 20e7cbd80..547a3f43a 100644
--- a/xc/lib/GL/mesa/src/drv/i810/i810tris.h
+++ b/xc/lib/GL/mesa/src/drv/i810/i810tris.h
@@ -133,6 +133,7 @@ static __inline__ void i810_draw_line( i810ContextPtr imesa,
 {
    GLuint vertsize = imesa->vertsize;
    GLuint *vb = i810AllocDwordsInline( imesa, 2 * vertsize );
+   int j;
 
 #if defined(USE_X86_ASM)
     __asm__ __volatile__( "rep ; movsl"
@@ -144,8 +145,6 @@ static __inline__ void i810_draw_line( i810ContextPtr imesa,
 			  : "0" (vertsize), "S" ((long)v1)
 			  : "memory" );
 #else
-    int j;
-
     for (j = 0 ; j < vertsize ; j++)
         vb[j] = v0->ui[j];
 
diff --git a/xc/lib/GL/mesa/src/drv/mga/Imakefile b/xc/lib/GL/mesa/src/drv/mga/Imakefile
index ba384a1ee..c2fe27a47 100644
--- a/xc/lib/GL/mesa/src/drv/mga/Imakefile
+++ b/xc/lib/GL/mesa/src/drv/mga/Imakefile
@@ -14,12 +14,13 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 
 #ifdef i386Architecture
 #ifdef MesaUse3DNow
-     ASM_DEFS = -DUSE_MMX_ASM -DUSE_X86_ASM -DUSE_3DNOW_ASM
-#else
-     ASM_DEFS = -DUSE_MMX_ASM -DUSE_X86_ASM
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
 #endif
+#ifdef MesaUseKatmai
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
+#endif
+  ASM_DEFINES = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
 #endif
-
 
 #if BuildXF86DRI
   DRI_DEFINES = GlxDefines -DDRIVERTS
@@ -34,9 +35,8 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 MESA_INCLUDES = -I. -I.. -I../../include \
 		-I../../../../dri/drm
 
-      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(ASM_DEFS)
-     INCLUDES = -I$(XLIBSRC) -I$(EXTINCSRC) $(MESA_INCLUDES) $(DRI_INCLUDES) 
-
+      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(ASM_DEFINES)
+     INCLUDES = -I$(XLIBSRC) -I$(EXTINCSRC) $(MESA_INCLUDES) $(DRI_INCLUDES)
 
       DRISRCS = ../../../dri/dri_mesa.c \
 		../../../../dri/dri_tmm.c
@@ -54,16 +54,35 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../../../dri/drm/xf86drmRandom.o \
 		../../../../dri/drm/xf86drmSL.o
 
-
-      MGASRCS = mgatexcnv.c mgadd.c mgafastpath.c		\
-		mgaeltpath.c mgapipeline.c mgaspan.c mgastate.c		\
-		mgatex.c mgatexmem.c mgatris.c mgavb.c mgaioctl.c	\
-		mga_xmesa.c mgabuffers.c
-
-      MGAOBJS = mgatexcnv.o mgadd.o mgafastpath.o		\
-		mgaeltpath.o mgapipeline.o mgaspan.o mgastate.o		\
-		mgatex.o mgatexmem.o mgatris.o mgavb.o mgaioctl.o	\
-		mga_xmesa.o mgabuffers.o
+      MGASRCS = mgabuffers.c \
+		mgadd.c \
+		mgaeltpath.c \
+		mgafastpath.c \
+		mgaioctl.c \
+		mgapipeline.c \
+		mgaspan.c \
+		mgastate.c \
+		mgatex.c \
+		mgatexcnv.c \
+		mgatexmem.c \
+		mgatris.c \
+		mgavb.c \
+		mga_xmesa.c \
+
+      MGAOBJS = mgabuffers.o \
+		mgadd.o \
+		mgaeltpath.o \
+		mgafastpath.o \
+		mgaioctl.o \
+		mgapipeline.o \
+		mgaspan.o \
+		mgastate.o \
+		mgatex.o \
+		mgatexcnv.o \
+		mgatexmem.o \
+		mgatris.o \
+		mgavb.o \
+		mga_xmesa.o
 
      MESASRCS = ../../aatriangle.c \
 		../../accum.c \
@@ -214,17 +233,29 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../zoom.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../../X86/x86.c \
-		../../X86/x86a.S \
-		../../X86/common_x86.c \
-		../../X86/common_x86asm.S \
-		../../X86/vertex.S
-
-     X86_OBJS = ../../X86/x86.o \
-		../../X86/x86a.o \
-		../../X86/common_x86.o \
-		../../X86/common_x86asm.o \
-		../../X86/vertex.o
+     X86_SRCS = ../../X86/common_x86.c \
+		../../X86/common_x86_asm.S \
+		../../X86/x86.c \
+		../../X86/x86_cliptest.S \
+		../../X86/x86_vertex.S \
+		../../X86/x86_xform_masked2.S \
+		../../X86/x86_xform_masked3.S \
+		../../X86/x86_xform_masked4.S \
+		../../X86/x86_xform_raw2.S \
+		../../X86/x86_xform_raw3.S \
+		../../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../../X86/common_x86.o \
+		../../X86/common_x86_asm.o \
+		../../X86/x86.o \
+		../../X86/x86_cliptest.o \
+		../../X86/x86_vertex.o \
+		../../X86/x86_xform_masked2.o \
+		../../X86/x86_xform_masked3.o \
+		../../X86/x86_xform_masked4.o \
+		../../X86/x86_xform_raw2.o \
+		../../X86/x86_xform_raw3.o \
+		../../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../../X86/mmx_blend.S
 
@@ -233,6 +264,7 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 #ifdef MesaUse3DNow
    3DNOW_SRCS = ../../X86/3dnow.c \
 		../../X86/3dnow_norm_raw.S \
+		../../X86/3dnow_vertex.S \
 		../../X86/3dnow_xform_masked1.S \
 		../../X86/3dnow_xform_masked2.S \
 		../../X86/3dnow_xform_masked3.S \
@@ -240,11 +272,11 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/3dnow_xform_raw1.S \
 		../../X86/3dnow_xform_raw2.S \
 		../../X86/3dnow_xform_raw3.S \
-		../../X86/3dnow_xform_raw4.S \
-		../../X86/vertex_3dnow.S
+		../../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../../X86/3dnow.o \
 		../../X86/3dnow_norm_raw.o \
+		../../X86/3dnow_vertex.o \
 		../../X86/3dnow_xform_masked1.o \
 		../../X86/3dnow_xform_masked2.o \
 		../../X86/3dnow_xform_masked3.o \
@@ -252,29 +284,54 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/3dnow_xform_raw1.o \
 		../../X86/3dnow_xform_raw2.o \
 		../../X86/3dnow_xform_raw3.o \
-		../../X86/3dnow_xform_raw4.o \
-		../../X86/vertex_3dnow.o
+		../../X86/3dnow_xform_raw4.o
+#endif
 
+#ifdef MesaUseKatmai
+  KATMAI_SRCS = ../../X86/katmai.c \
+		../../X86/katmai_norm_raw.S \
+		../../X86/katmai_vertex.S \
+		../../X86/katmai_xform_masked1.S \
+		../../X86/katmai_xform_masked2.S \
+		../../X86/katmai_xform_masked3.S \
+		../../X86/katmai_xform_masked4.S \
+		../../X86/katmai_xform_raw1.S \
+		../../X86/katmai_xform_raw2.S \
+		../../X86/katmai_xform_raw3.S \
+		../../X86/katmai_xform_raw4.S
+
+  KATMAI_OBJS = ../../X86/katmai.o \
+		../../X86/katmai_norm_raw.o \
+		../../X86/katmai_vertex.o \
+		../../X86/katmai_xform_masked1.o \
+		../../X86/katmai_xform_masked2.o \
+		../../X86/katmai_xform_masked3.o \
+		../../X86/katmai_xform_masked4.o \
+		../../X86/katmai_xform_raw1.o \
+		../../X86/katmai_xform_raw2.o \
+		../../X86/katmai_xform_raw3.o \
+		../../X86/katmai_xform_raw4.o
 #endif
 #endif
 
 #ifdef GlxSoProf
-       HIOBJS = ../../../../highpc.o
-       LOOBJS = ../../../../lowpc.o
-       HISRCS = ../../../../highpc.c
        LOSRCS = ../../../../lowpc.c
+       HISRCS = ../../../../highpc.c
+
+       LOOBJS = ../../../../lowpc.o
+       HIOBJS = ../../../../highpc.o
 #endif
 
-      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS)
-      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS)
+      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
+      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
 
-      COMMONSRCS = ../common/mm.c ../common/hwlog.c
-      COMMONOBJS = ../common/mm.o ../common/hwlog.o
+   COMMONSRCS = ../common/mm.c ../common/hwlog.c
+   COMMONOBJS = ../common/mm.o ../common/hwlog.o
 
-         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) $(ASMSRCS)	\
-		$(COMMONSRCS) $(MGASRCS) $(HISRCS)
-         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) $(ASMOBJS)	\
-		$(COMMONOBJS) $(MGAOBJS) $(HIOBJS)
+         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) \
+		$(ASMSRCS) $(COMMONSRCS) $(MGASRCS) $(HISRCS)
+         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) \
+		$(ASMOBJS) $(COMMONOBJS) $(MGAOBJS) $(HIOBJS)
 
 REQUIREDLIBS += MathLibrary 
 #if !GlxBuiltInMga
diff --git a/xc/lib/GL/mesa/src/drv/r128/Imakefile b/xc/lib/GL/mesa/src/drv/r128/Imakefile
index 3b04b1580..056297ee3 100644
--- a/xc/lib/GL/mesa/src/drv/r128/Imakefile
+++ b/xc/lib/GL/mesa/src/drv/r128/Imakefile
@@ -14,10 +14,12 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 
 #ifdef i386Architecture
 #ifdef MesaUse3DNow
-  ASM_DEFINES = -DUSE_MMX_ASM -DUSE_X86_ASM -DUSE_3DNOW_ASM
-#else
-  ASM_DEFINES = -DUSE_MMX_ASM -DUSE_X86_ASM
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
 #endif
+#ifdef MesaUseKatmai
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
+#endif
+  ASM_DEFINES = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
 #endif
 
 #if BuildXF86DRI
@@ -37,11 +39,6 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 		-I../common
 #endif
 
-LinkSourceFile(mm.c, ../common)
-LinkSourceFile(mm.h, ../common)
-LinkSourceFile(hwlog.c, ../common)
-LinkSourceFile(hwlog.h, ../common)
-
 MESA_INCLUDES = -I. -I.. -I../../include \
 		-I../../../../dri/drm
 
@@ -246,17 +243,29 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../zoom.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../../X86/x86.c \
-		../../X86/x86a.S \
-		../../X86/common_x86.c \
-		../../X86/common_x86asm.S \
-		../../X86/vertex.S
-
-     X86_OBJS = ../../X86/x86.o \
-		../../X86/x86a.o \
-		../../X86/common_x86.o \
-		../../X86/common_x86asm.o \
-		../../X86/vertex.o
+     X86_SRCS = ../../X86/common_x86.c \
+		../../X86/common_x86_asm.S \
+		../../X86/x86.c \
+		../../X86/x86_cliptest.S \
+		../../X86/x86_vertex.S \
+		../../X86/x86_xform_masked2.S \
+		../../X86/x86_xform_masked3.S \
+		../../X86/x86_xform_masked4.S \
+		../../X86/x86_xform_raw2.S \
+		../../X86/x86_xform_raw3.S \
+		../../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../../X86/common_x86.o \
+		../../X86/common_x86_asm.o \
+		../../X86/x86.o \
+		../../X86/x86_cliptest.o \
+		../../X86/x86_vertex.o \
+		../../X86/x86_xform_masked2.o \
+		../../X86/x86_xform_masked3.o \
+		../../X86/x86_xform_masked4.o \
+		../../X86/x86_xform_raw2.o \
+		../../X86/x86_xform_raw3.o \
+		../../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../../X86/mmx_blend.S
 
@@ -265,6 +274,7 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 #ifdef MesaUse3DNow
    3DNOW_SRCS = ../../X86/3dnow.c \
 		../../X86/3dnow_norm_raw.S \
+		../../X86/3dnow_vertex.S \
 		../../X86/3dnow_xform_masked1.S \
 		../../X86/3dnow_xform_masked2.S \
 		../../X86/3dnow_xform_masked3.S \
@@ -272,11 +282,11 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/3dnow_xform_raw1.S \
 		../../X86/3dnow_xform_raw2.S \
 		../../X86/3dnow_xform_raw3.S \
-		../../X86/3dnow_xform_raw4.S \
-		../../X86/vertex_3dnow.S
+		../../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../../X86/3dnow.o \
 		../../X86/3dnow_norm_raw.o \
+		../../X86/3dnow_vertex.o \
 		../../X86/3dnow_xform_masked1.o \
 		../../X86/3dnow_xform_masked2.o \
 		../../X86/3dnow_xform_masked3.o \
@@ -284,29 +294,62 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/3dnow_xform_raw1.o \
 		../../X86/3dnow_xform_raw2.o \
 		../../X86/3dnow_xform_raw3.o \
-		../../X86/3dnow_xform_raw4.o \
-		../../X86/vertex_3dnow.o
+		../../X86/3dnow_xform_raw4.o
 #endif
 
+#ifdef MesaUseKatmai
+  KATMAI_SRCS = ../../X86/katmai.c \
+		../../X86/katmai_norm_raw.S \
+		../../X86/katmai_vertex.S \
+		../../X86/katmai_xform_masked1.S \
+		../../X86/katmai_xform_masked2.S \
+		../../X86/katmai_xform_masked3.S \
+		../../X86/katmai_xform_masked4.S \
+		../../X86/katmai_xform_raw1.S \
+		../../X86/katmai_xform_raw2.S \
+		../../X86/katmai_xform_raw3.S \
+		../../X86/katmai_xform_raw4.S
+
+  KATMAI_OBJS = ../../X86/katmai.o \
+		../../X86/katmai_norm_raw.o \
+		../../X86/katmai_vertex.o \
+		../../X86/katmai_xform_masked1.o \
+		../../X86/katmai_xform_masked2.o \
+		../../X86/katmai_xform_masked3.o \
+		../../X86/katmai_xform_masked4.o \
+		../../X86/katmai_xform_raw1.o \
+		../../X86/katmai_xform_raw2.o \
+		../../X86/katmai_xform_raw3.o \
+		../../X86/katmai_xform_raw4.o
+#endif
 #endif
 #endif
 
-      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS)
-      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS)
+#ifdef GlxSoProf
+       LOSRCS = ../../../../lowpc.c
+       HISRCS = ../../../../highpc.c
+
+       LOOBJS = ../../../../lowpc.o
+       HIOBJS = ../../../../highpc.o
+#endif
 
-   COMMONSRCS = mm.c hwlog.c
-   COMMONOBJS = mm.o hwlog.o
+      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
+      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
 
-         SRCS = $(DRISRCS) $(DRMSRCS) $(MESASRCS) $(ASMSRCS) \
-		$(COMMONSRCS) $(R128SRCS)
-         OBJS = $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) $(ASMOBJS) \
-		$(COMMONOBJS) $(R128OBJS)
+   COMMONSRCS = ../common/mm.c ../common/hwlog.c
+   COMMONOBJS = ../common/mm.o ../common/hwlog.o
+
+         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) \
+		$(ASMSRCS) $(COMMONSRCS) $(R128SRCS) $(HISRCS)
+         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) \
+		$(ASMOBJS) $(COMMONOBJS) $(R128OBJS) $(HIOBJS)
 
 REQUIREDLIBS += MathLibrary
-#if !GlxBuiltInMga
+#if !GlxBuiltInR128
 REQUIREDLIBS += -L../../../.. -lGL
 #endif
 
+
 #if !GlxUseBuiltInDRIDriver
 #undef DoNormalLib NormalLibGlx
 #undef DoExtraLib SharedLibGlx
@@ -327,6 +370,12 @@ ALL_OBJS = $(OBJS)
 ALL_DEPS = DONE
 SharedDepModuleTarget($(LIBNAME),$(ALL_DEPS),$(ALL_OBJS))
 InstallDynamicModule($(LIBNAME),$(MODULEDIR),dri)
+
+#ifdef GlxSoProf
+SOPROF_LIBNAME = _r128_dri_p
+NormalDepLibraryTarget($(SOPROF_LIBNAME),$(ALL_DEPS),$(ALL_OBJS))
+InstallLibrary($(SOPROF_LIBNAME),$(MODULEDIR)/dri)
+#endif
 #endif
 
 DependTarget()
diff --git a/xc/lib/GL/mesa/src/drv/sis/Imakefile b/xc/lib/GL/mesa/src/drv/sis/Imakefile
index 59f2429ac..9993d6568 100644
--- a/xc/lib/GL/mesa/src/drv/sis/Imakefile
+++ b/xc/lib/GL/mesa/src/drv/sis/Imakefile
@@ -13,26 +13,27 @@ SIS_DEFINES = -DSIS_USE_HW_CULL
 
 XCOMM DEBUG_DEFINES = -DDEBUG_LOCKING
 XCOMM DEBUG_DEFINES = -DSIS_DEBUG
- 
+
 #if Malloc0ReturnsNull
 ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 #endif
 
 #ifdef i386Architecture
 #ifdef MesaUse3DNow
-     ASM_DEFS = -DUSE_MMX_ASM -DUSE_X86_ASM -DUSE_3DNOW_ASM
-#else
-     ASM_DEFS = -DUSE_MMX_ASM -DUSE_X86_ASM
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
 #endif
+#ifdef MesaUseKatmai
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
+#endif
+  ASM_DEFINES = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
 #endif
-
 
 #if BuildXF86DRI
   DRI_DEFINES = GlxDefines $(SIS_DEFINES)
-  DRI_INCLUDES = -I../../../../dri -I../../../../glx \
+ DRI_INCLUDES = -I../../../../dri -I../../../../glx \
 		-I../../../dri \
-      		-I$(TOP)/include -I$(TOP)/include/GL \
-    		-I$(XF86OSSRC) -I$(SERVERSRC)/GL/dri \
+		-I$(TOP)/include -I$(TOP)/include/GL \
+		-I$(XF86OSSRC) -I$(SERVERSRC)/GL/dri \
 		-I$(XF86DRIVERSRC)/sis \
 		-I../../../include -I../.. -I../../X \
                 -I../common/
@@ -57,23 +58,41 @@ LinkSourceFile(xdriP.h, ../../X)
  STEREO_DEFINES = -DSIS_STEREO=0
 #endif
 
-     DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(DEBUG_DEFINES) \
-               $(STEREO_DEFINES)
+      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(ASM_DEFINES) \
+		$(DEBUG_DEFINES) $(STEREO_DEFINES)
      INCLUDES = -I$(XLIBSRC) -I$(EXTINCSRC) $(MESA_INCLUDES) $(DRI_INCLUDES) \
 		-I./misc \
                 -I$(SERVERSRC)/hw/xfree86/drivers/sis
 
-     SISSRCS = sis_render.c sis_mesa.c sis_ctx.c sis_alloc.c \
-            sis_fog.c sis_clear.c sis_stencil.c \
-            sis_texture.c sis_xwin.c sis_xmesa.c sis_span.c sis_fastpath.c \
-            $(DEBUG_SRC) \
-            $(STEREO_SRC)
-
-     SISOBJS = sis_render.o sis_mesa.o sis_ctx.o sis_alloc.o \
-            sis_fog.o sis_clear.o sis_stencil.o \
-            sis_texture.o sis_xwin.o sis_xmesa.o sis_span.o sis_fastpath.o \
-            $(DEBUG_OBJ) \
-            $(STEREO_OBJ)
+      SISSRCS = sis_render.c \
+		sis_mesa.c \
+		sis_ctx.c \
+		sis_alloc.c \
+		sis_fog.c \
+		sis_clear.c \
+		sis_stencil.c \
+		sis_texture.c \
+		sis_xwin.c \
+		sis_xmesa.c \
+		sis_span.c \
+		sis_fastpath.c \
+		$(DEBUG_SRC) \
+		$(STEREO_SRC)
+
+      SISOBJS = sis_render.o \
+		sis_mesa.o \
+		sis_ctx.o \
+		sis_alloc.o \
+		sis_fog.o \
+		sis_clear.o \
+		sis_stencil.o \
+		sis_texture.o \
+		sis_xwin.o \
+		sis_xmesa.o \
+		sis_span.o \
+		sis_fastpath.o \
+		$(DEBUG_OBJ) \
+		$(STEREO_OBJ)
 
       DRISRCS = ../../../dri/dri_mesa.c \
 		../../../../dri/dri_tmm.c
@@ -240,17 +259,29 @@ LinkSourceFile(xdriP.h, ../../X)
 		../../zoom.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../../X86/x86.c \
-		../../X86/x86a.S \
-		../../X86/common_x86.c \
-		../../X86/common_x86asm.S \
-		../../X86/vertex.S
-
-     X86_OBJS = ../../X86/x86.o \
-		../../X86/x86a.o \
-		../../X86/common_x86.o \
-		../../X86/common_x86asm.o \
-		../../X86/vertex.o
+     X86_SRCS = ../../X86/common_x86.c \
+		../../X86/common_x86_asm.S \
+		../../X86/x86.c \
+		../../X86/x86_cliptest.S \
+		../../X86/x86_vertex.S \
+		../../X86/x86_xform_masked2.S \
+		../../X86/x86_xform_masked3.S \
+		../../X86/x86_xform_masked4.S \
+		../../X86/x86_xform_raw2.S \
+		../../X86/x86_xform_raw3.S \
+		../../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../../X86/common_x86.o \
+		../../X86/common_x86_asm.o \
+		../../X86/x86.o \
+		../../X86/x86_cliptest.o \
+		../../X86/x86_vertex.o \
+		../../X86/x86_xform_masked2.o \
+		../../X86/x86_xform_masked3.o \
+		../../X86/x86_xform_masked4.o \
+		../../X86/x86_xform_raw2.o \
+		../../X86/x86_xform_raw3.o \
+		../../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../../X86/mmx_blend.S
 
@@ -259,6 +290,7 @@ LinkSourceFile(xdriP.h, ../../X)
 #ifdef MesaUse3DNow
    3DNOW_SRCS = ../../X86/3dnow.c \
 		../../X86/3dnow_norm_raw.S \
+		../../X86/3dnow_vertex.S \
 		../../X86/3dnow_xform_masked1.S \
 		../../X86/3dnow_xform_masked2.S \
 		../../X86/3dnow_xform_masked3.S \
@@ -266,11 +298,11 @@ LinkSourceFile(xdriP.h, ../../X)
 		../../X86/3dnow_xform_raw1.S \
 		../../X86/3dnow_xform_raw2.S \
 		../../X86/3dnow_xform_raw3.S \
-		../../X86/3dnow_xform_raw4.S \
-		../../X86/vertex_3dnow.S
+		../../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../../X86/3dnow.o \
 		../../X86/3dnow_norm_raw.o \
+		../../X86/3dnow_vertex.o \
 		../../X86/3dnow_xform_masked1.o \
 		../../X86/3dnow_xform_masked2.o \
 		../../X86/3dnow_xform_masked3.o \
@@ -278,17 +310,51 @@ LinkSourceFile(xdriP.h, ../../X)
 		../../X86/3dnow_xform_raw1.o \
 		../../X86/3dnow_xform_raw2.o \
 		../../X86/3dnow_xform_raw3.o \
-		../../X86/3dnow_xform_raw4.o \
-		../../X86/vertex_3dnow.o
+		../../X86/3dnow_xform_raw4.o
 #endif
 
+#ifdef MesaUseKatmai
+  KATMAI_SRCS = ../../X86/katmai.c \
+		../../X86/katmai_norm_raw.S \
+		../../X86/katmai_vertex.S \
+		../../X86/katmai_xform_masked1.S \
+		../../X86/katmai_xform_masked2.S \
+		../../X86/katmai_xform_masked3.S \
+		../../X86/katmai_xform_masked4.S \
+		../../X86/katmai_xform_raw1.S \
+		../../X86/katmai_xform_raw2.S \
+		../../X86/katmai_xform_raw3.S \
+		../../X86/katmai_xform_raw4.S
+
+  KATMAI_OBJS = ../../X86/katmai.o \
+		../../X86/katmai_norm_raw.o \
+		../../X86/katmai_vertex.o \
+		../../X86/katmai_xform_masked1.o \
+		../../X86/katmai_xform_masked2.o \
+		../../X86/katmai_xform_masked3.o \
+		../../X86/katmai_xform_masked4.o \
+		../../X86/katmai_xform_raw1.o \
+		../../X86/katmai_xform_raw2.o \
+		../../X86/katmai_xform_raw3.o \
+		../../X86/katmai_xform_raw4.o
+#endif
 #endif
 
-      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS)
-      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS)
+#ifdef GlxSoProf
+       LOSRCS = ../../../../lowpc.c
+       HISRCS = ../../../../highpc.c
 
-         SRCS = $(DRISRCS) $(DRMSRCS) $(SISSRCS) $(MESASRCS) $(ASMSRCS)
-         OBJS = $(DRIOBJS) $(DRMOBJS) $(SISOBJS) $(MESAOBJS) $(ASMOBJS)
+       LOOBJS = ../../../../lowpc.o
+       HIOBJS = ../../../../highpc.o
+#endif
+
+      ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
+      ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
+
+         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) \
+		$(ASMSRCS) $(SISSRCS) $(HISRCS)
+         OBJS = $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) \
+		$(ASMOBJS) $(SISOBJS) $(HIOBJS)
 
 REQUIREDLIBS += MathLibrary
 #if !GlxBuiltInSIS
@@ -316,6 +382,12 @@ ALL_OBJS = $(OBJS)
 ALL_DEPS = DONE
 SharedDepModuleTarget($(LIBNAME),$(ALL_DEPS),$(ALL_OBJS))
 InstallDynamicModule($(LIBNAME),$(MODULEDIR),dri)
+
+#ifdef GlxSoProf
+SOPROF_LIBNAME = _sis_dri_p
+NormalDepLibraryTarget($(SOPROF_LIBNAME),$(ALL_DEPS),$(ALL_OBJS))
+InstallLibrary($(SOPROF_LIBNAME),$(MODULEDIR)/dri)
+#endif
 #endif
 
 DependTarget()
diff --git a/xc/lib/GL/mesa/src/drv/tdfx/Imakefile b/xc/lib/GL/mesa/src/drv/tdfx/Imakefile
index ae70f81d9..e5866c261 100644
--- a/xc/lib/GL/mesa/src/drv/tdfx/Imakefile
+++ b/xc/lib/GL/mesa/src/drv/tdfx/Imakefile
@@ -12,6 +12,16 @@ XCOMM $XFree86: xc/lib/GL/mesa/src/drv/tdfx/Imakefile,v 1.15 2000/10/28 01:05:22
 ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 #endif
 
+#ifdef i386Architecture
+#ifdef MesaUse3DNow
+   3DNOW_DEFS = -DUSE_3DNOW_ASM
+#endif
+#ifdef MesaUseKatmai
+  KATMAI_DEFS = -DUSE_KATMAI_ASM
+#endif
+  ASM_DEFINES = -DUSE_X86_ASM -DUSE_MMX_ASM $(3DNOW_DEFS) $(KATMAI_DEFS)
+#endif
+
 #if BuildXF86DRI
   DRI_DEFINES = GlxDefines -DFX_GLIDE3 -DDRIVERTS
  DRI_INCLUDES = -I../../../../dri -I../../../../glx \
@@ -19,13 +29,12 @@ ALLOC_DEFINES = -DMALLOC_0_RETURNS_NULL
 		-I$(TOP)/include -I$(TOP)/include/GL \
 		-I$(XF86OSSRC) -I$(SERVERSRC)/GL/dri \
 		-I$(XF86DRIVERSRC)/tdfx \
-		-I../../../include -I../.. -I../../X
+		-I../../../include -I../.. -I../../X -I../common
 #endif
 
-MESA_INCLUDES = -I. -I.. -I../../include \
-		-I../../../../dri/drm
+MESA_INCLUDES = -I. -I.. -I../../include
 
-      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES)
+      DEFINES = $(ALLOC_DEFINES) $(DRI_DEFINES) $(ASM_DEFINES)
      INCLUDES = -I$(XLIBSRC) -I$(EXTINCSRC) $(MESA_INCLUDES) $(DRI_INCLUDES) \
 		-I$(GLIDE3INCDIR)
 
@@ -45,23 +54,41 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../../../dri/drm/xf86drmRandom.o \
 		../../../../dri/drm/xf86drmSL.o
 
-#ifdef GlxSoProf
-       HIOBJS = ../../../../highpc.o
-       LOOBJS = ../../../../lowpc.o
-#else
-       HIOBJS = 
-       LOOBJS =
-#endif
-
-     TDFXSRCS = tdfx_xmesa.c tdfx_init.c tdfx_inithw.c \
-		fxclip.c fxcva.c fxdd.c fxddspan.c fxddtex.c fxfastpath.c \
-		fxglidew.c fxpipeline.c fxrender.c fxsanity.c fxsetup.c \
-		fxtexman.c fxtrifuncs.c fxvsetup.c
-
-     TDFXOBJS = tdfx_xmesa.o tdfx_init.o tdfx_inithw.o \
-		fxclip.o fxcva.o fxdd.o fxddspan.o fxddtex.o fxfastpath.o \
-		fxglidew.o fxpipeline.o fxrender.o fxsanity.o fxsetup.o \
-		fxtexman.o fxtrifuncs.o fxvsetup.o
+     TDFXSRCS = tdfx_xmesa.c \
+		tdfx_init.c \
+		tdfx_inithw.c \
+		fxclip.c \
+		fxcva.c \
+		fxdd.c \
+		fxddspan.c \
+		fxddtex.c \
+		fxfastpath.c \
+		fxglidew.c \
+		fxpipeline.c \
+		fxrender.c \
+		fxsanity.c \
+		fxsetup.c \
+		fxtexman.c \
+		fxtrifuncs.c \
+		fxvsetup.c
+
+     TDFXOBJS = tdfx_xmesa.o \
+		tdfx_init.o \
+		tdfx_inithw.o \
+		fxclip.o \
+		fxcva.o \
+		fxdd.o \
+		fxddspan.o \
+		fxddtex.o \
+		fxfastpath.o \
+		fxglidew.o \
+		fxpipeline.o \
+		fxrender.o \
+		fxsanity.o \
+		fxsetup.o \
+		fxtexman.o \
+		fxtrifuncs.o \
+		fxvsetup.o
 
      MESASRCS = ../../aatriangle.c \
 		../../accum.c \
@@ -212,17 +239,29 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../zoom.o
 
 #ifdef i386Architecture
-     X86_SRCS = ../../X86/x86.c \
-		../../X86/x86a.S \
-		../../X86/common_x86.c \
-		../../X86/common_x86asm.S \
-		../../X86/vertex.S
-
-     X86_OBJS = ../../X86/x86.o \
-		../../X86/x86a.o \
-		../../X86/common_x86.o \
-		../../X86/common_x86asm.o \
-		../../X86/vertex.o
+     X86_SRCS = ../../X86/common_x86.c \
+		../../X86/common_x86_asm.S \
+		../../X86/x86.c \
+		../../X86/x86_cliptest.S \
+		../../X86/x86_vertex.S \
+		../../X86/x86_xform_masked2.S \
+		../../X86/x86_xform_masked3.S \
+		../../X86/x86_xform_masked4.S \
+		../../X86/x86_xform_raw2.S \
+		../../X86/x86_xform_raw3.S \
+		../../X86/x86_xform_raw4.S
+
+     X86_OBJS = ../../X86/common_x86.o \
+		../../X86/common_x86_asm.o \
+		../../X86/x86.o \
+		../../X86/x86_cliptest.o \
+		../../X86/x86_vertex.o \
+		../../X86/x86_xform_masked2.o \
+		../../X86/x86_xform_masked3.o \
+		../../X86/x86_xform_masked4.o \
+		../../X86/x86_xform_raw2.o \
+		../../X86/x86_xform_raw3.o \
+		../../X86/x86_xform_raw4.o
 
      MMX_SRCS = ../../X86/mmx_blend.S
 
@@ -231,6 +270,7 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 #ifdef MesaUse3DNow
    3DNOW_SRCS = ../../X86/3dnow.c \
 		../../X86/3dnow_norm_raw.S \
+		../../X86/3dnow_vertex.S \
 		../../X86/3dnow_xform_masked1.S \
 		../../X86/3dnow_xform_masked2.S \
 		../../X86/3dnow_xform_masked3.S \
@@ -238,11 +278,11 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/3dnow_xform_raw1.S \
 		../../X86/3dnow_xform_raw2.S \
 		../../X86/3dnow_xform_raw3.S \
-		../../X86/3dnow_xform_raw4.S \
-		../../X86/vertex_3dnow.S
+		../../X86/3dnow_xform_raw4.S
 
    3DNOW_OBJS = ../../X86/3dnow.o \
 		../../X86/3dnow_norm_raw.o \
+		../../X86/3dnow_vertex.o \
 		../../X86/3dnow_xform_masked1.o \
 		../../X86/3dnow_xform_masked2.o \
 		../../X86/3dnow_xform_masked3.o \
@@ -250,13 +290,13 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/3dnow_xform_raw1.o \
 		../../X86/3dnow_xform_raw2.o \
 		../../X86/3dnow_xform_raw3.o \
-		../../X86/3dnow_xform_raw4.o \
-		../../X86/vertex_3dnow.o
+		../../X86/3dnow_xform_raw4.o
 #endif
 
 #ifdef MesaUseKatmai
-   KATMAI_SRCS = ../../X86/katmai.c \
+  KATMAI_SRCS = ../../X86/katmai.c \
 		../../X86/katmai_norm_raw.S \
+		../../X86/katmai_vertex.S \
 		../../X86/katmai_xform_masked1.S \
 		../../X86/katmai_xform_masked2.S \
 		../../X86/katmai_xform_masked3.S \
@@ -264,11 +304,11 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/katmai_xform_raw1.S \
 		../../X86/katmai_xform_raw2.S \
 		../../X86/katmai_xform_raw3.S \
-		../../X86/katmai_xform_raw4.S \
-		../../X86/vertex_katmai.S
+		../../X86/katmai_xform_raw4.S
 
-   KATMAI_OBJS = ../../X86/katmai.o \
+  KATMAI_OBJS = ../../X86/katmai.o \
 		../../X86/katmai_norm_raw.o \
+		../../X86/katmai_vertex.o \
 		../../X86/katmai_xform_masked1.o \
 		../../X86/katmai_xform_masked2.o \
 		../../X86/katmai_xform_masked3.o \
@@ -276,17 +316,25 @@ MESA_INCLUDES = -I. -I.. -I../../include \
 		../../X86/katmai_xform_raw1.o \
 		../../X86/katmai_xform_raw2.o \
 		../../X86/katmai_xform_raw3.o \
-		../../X86/katmai_xform_raw4.o \
-		../../X86/vertex_katmai.o
+		../../X86/katmai_xform_raw4.o
+#endif
 #endif
 
+#ifdef GlxSoProf
+       LOSRCS = ../../../../lowpc.c
+       HISRCS = ../../../../highpc.c
+
+       LOOBJS = ../../../../lowpc.o
+       HIOBJS = ../../../../highpc.o
 #endif
 
       ASMSRCS = $(X86_SRCS) $(MMX_SRCS) $(3DNOW_SRCS) $(KATMAI_SRCS)
       ASMOBJS = $(X86_OBJS) $(MMX_OBJS) $(3DNOW_OBJS) $(KATMAI_OBJS)
 
-         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(TDFXSRCS) $(MESASRCS) $(ASMSRCS) $(HISRCS)
-         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(TDFXOBJS) $(MESAOBJS) $(ASMOBJS) $(HIOBJS)
+         SRCS = $(LOSRCS) $(DRISRCS) $(DRMSRCS) $(MESASRCS) \
+		$(ASMSRCS) $(TDFXSRCS) $(HISRCS)
+         OBJS = $(LOOBJS) $(DRIOBJS) $(DRMOBJS) $(MESAOBJS) \
+		$(ASMOBJS) $(TDFXOBJS) $(HIOBJS)
 
 REQUIREDLIBS = -l$(GLIDE3LIBNAME) MathLibrary
 #if !GlxBuiltInTdfx
diff --git a/xc/lib/GL/mesa/src/drv/tdfx/fxfastpath.c b/xc/lib/GL/mesa/src/drv/tdfx/fxfastpath.c
index f1c669e9a..9c41e7ade 100644
--- a/xc/lib/GL/mesa/src/drv/tdfx/fxfastpath.c
+++ b/xc/lib/GL/mesa/src/drv/tdfx/fxfastpath.c
@@ -49,7 +49,7 @@
 #include "fxtexman.h"
 #include "vertices.h"
 #ifdef __i386__
-#include "X86/common_x86asm.h"
+#include "X86/common_x86_asm.h"
 #endif
 
 
diff --git a/xc/lib/GL/mesa/src/drv/tdfx/fxfasttmp.h b/xc/lib/GL/mesa/src/drv/tdfx/fxfasttmp.h
index 6f15462c4..1552d8247 100644
--- a/xc/lib/GL/mesa/src/drv/tdfx/fxfasttmp.h
+++ b/xc/lib/GL/mesa/src/drv/tdfx/fxfasttmp.h
@@ -336,8 +336,8 @@ static void TAG(fx_init_fastpath) (struct fx_fast_tab * tab)
     tab->project_vertices = TAG(fx_project_vertices);
     tab->project_clipped_vertices = TAG(fx_project_clipped_vertices);
 
-#if defined(USE_3DNOW_ASM)
-    if (gl_x86_cpu_features & GL_CPU_3Dnow) {
+#if 0 && defined(USE_3DNOW_ASM)
+    if (gl_x86_cpu_features & X86_FEATURE_3DNOW) {
         extern void TAG(fx_3dnow_project_vertices) (GLfloat * first,
                                                     GLfloat * last,
                                                     const GLfloat * mat,
diff --git a/xc/lib/GL/mesa/src/drv/tdfx/fxglidew.h b/xc/lib/GL/mesa/src/drv/tdfx/fxglidew.h
index 388ca2d7c..0f1f2fd0b 100644
--- a/xc/lib/GL/mesa/src/drv/tdfx/fxglidew.h
+++ b/xc/lib/GL/mesa/src/drv/tdfx/fxglidew.h
@@ -115,7 +115,6 @@ extern txErrorSetCallback_t txErrorSetCallbackPtr;
 
 FX_ENTRY void FX_CALL grEnable(GrEnableMode_t mode);
 
-FX_ENTRY void FX_CALL grEnable(GrEnableMode_t mode);
 /* 
  * General context: 
  */