summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/r300/r300_render.c
blob: f1b2fae1392e9b92a0466e7ad787051d015be165 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
/**************************************************************************

Copyright (C) 2004 Nicolai Haehnle.

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
on the rights to use, copy, modify, merge, publish, distribute, sub
license, 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 (including the next
paragraph) 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 NON-INFRINGEMENT. IN NO EVENT SHALL
ATI, VA LINUX SYSTEMS AND/OR THEIR SUPPLIERS 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.

**************************************************************************/

/*
 * Authors:
 *   Nicolai Haehnle <prefect_@gmx.net>
 */

#include "glheader.h"
#include "state.h"
#include "imports.h"
#include "enums.h"
#include "macros.h"
#include "context.h"
#include "dd.h"
#include "simple_list.h"

#include "api_arrayelt.h"
#include "swrast/swrast.h"
#include "swrast_setup/swrast_setup.h"
#include "array_cache/acache.h"
#include "tnl/tnl.h"

#include "radeon_ioctl.h"
#include "radeon_state.h"
#include "r300_context.h"
#include "r300_ioctl.h"
#include "r300_state.h"
#include "r300_reg.h"
#include "r300_program.h"

#include "r300_lib.h"


/**********************************************************************
*                     Hardware rasterization
*
* When we fell back to software TCL, we still try to use the
* rasterization hardware for rendering.
**********************************************************************/

static int r300_get_primitive_type(r300ContextPtr rmesa, 
	GLcontext *ctx,
	int start,
	int end,
	int prim)
{
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
   int type=-1, min_vertices=0;
   char *name="UNKNOWN";
   
   if(end<=start)return -1; /* do we need to watch for this ? */
   
	switch (prim & PRIM_MODE_MASK) {
	case GL_POINTS:
   		name="P";
	        type=R300_VAP_VF_CNTL__PRIM_POINTS;
		min_vertices=1;
      		break;		
	case GL_LINES:
   		name="L";
	        type=R300_VAP_VF_CNTL__PRIM_LINES;
		min_vertices=2;
      		break;		
	case GL_LINE_STRIP:
   		name="LS";
	        type=R300_VAP_VF_CNTL__PRIM_LINE_STRIP;
		min_vertices=2;
      		break;
	case GL_LINE_LOOP:
   		name="LL";
		min_vertices=2;
		return -1;
      		break;
    	case GL_TRIANGLES:
   		name="T";
	        type=R300_VAP_VF_CNTL__PRIM_TRIANGLES;
		min_vertices=3;
      		break;
   	case GL_TRIANGLE_STRIP:
   		name="TS";
	        type=R300_VAP_VF_CNTL__PRIM_TRIANGLE_STRIP;
		min_vertices=3;
      		break;
   	case GL_TRIANGLE_FAN:
   		name="TF";
	        type=R300_VAP_VF_CNTL__PRIM_TRIANGLE_FAN;
		min_vertices=3;
      		break;
	case GL_QUADS:
   		name="Q";
	        type=R300_VAP_VF_CNTL__PRIM_QUADS;
		min_vertices=4;
      		break;
	case GL_QUAD_STRIP:
   		name="QS";
	        type=R300_VAP_VF_CNTL__PRIM_QUAD_STRIP;
		min_vertices=4;
      		break;
   	default:
   		fprintf(stderr, "Cannot handle primitive %02x ", prim & PRIM_MODE_MASK);
		return -1;
         	break;
   	}
   #if 0
   fprintf(stderr, "[%d-%d]%s ", start, end, name);
   #endif
   if(start+min_vertices>=end){
	fprintf(stderr, "Not enough vertices\n");
	return -1;
   	}
   return type;
}



/* Immediate implementation - vertex data is sent via command stream */

static GLfloat default_vector[4]={0.0, 0.0, 0.0, 1.0};

#define output_vector(v, i) \
	{ \
	int _i; \
	for(_i=0;_i<v->size;_i++){ \
		efloat(VEC_ELT(v, GLfloat, i)[_i]); \
		} \
	for(_i=v->size;_i<4;_i++){ \
			efloat(default_vector[_i]); \
			} \
	}

static void r300_render_flat_primitive(r300ContextPtr rmesa, 
	GLcontext *ctx,
	int start,
	int end,
	int prim)
{
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
   int k, type;
   LOCAL_VARS
       	
   type=r300_get_primitive_type(rmesa, ctx, start, end, prim);
		
   if(type<0)return;


   start_immediate_packet(end-start, type, 8);

	for(i=start;i<end;i++){
		#if 0
		fprintf(stderr, "* (%f %f %f %f) (%f %f %f %f)\n", 
			VEC_ELT(VB->ObjPtr, GLfloat, i)[0],
			VEC_ELT(VB->ObjPtr, GLfloat, i)[1],
			VEC_ELT(VB->ObjPtr, GLfloat, i)[2],
			VEC_ELT(VB->ObjPtr, GLfloat, i)[3],
			
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[0],
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[1],
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[2],
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[3]
			);
		#endif
		
		
		/* coordinates */
		output_vector(VB->ObjPtr, i);
		
		/* color components */
		output_vector(VB->ColorPtr[0], i);
		}

}

static GLboolean r300_run_flat_render(GLcontext *ctx,
				 struct tnl_pipeline_stage *stage)
{
   r300ContextPtr rmesa = R300_CONTEXT(ctx);
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
   AOS_DATA vb_arrays[2];
   LOCAL_VARS
	
   /* Flush state - make sure command buffer is nice and large */
   r300Flush(ctx);
	
	if (RADEON_DEBUG == DEBUG_PRIMS)
		fprintf(stderr, "%s\n", __FUNCTION__);

   /* setup array of structures data */

   /* Note: immediate vertex data includes all coordinates.
     To save bandwidth use either VBUF or state-based vertex generation */
    /* xyz */
   vb_arrays[0].element_size=4;
   vb_arrays[0].stride=4;
   vb_arrays[0].offset=0; /* Not used */
   vb_arrays[0].format=AOS_FORMAT_FLOAT;
   vb_arrays[0].ncomponents=4;
   vb_arrays[0].reg=REG_COORDS;

    /* color */
   vb_arrays[1].element_size=4;
   vb_arrays[1].stride=4;
   vb_arrays[1].offset=0; /* Not used */
   vb_arrays[1].format=AOS_FORMAT_FLOAT_COLOR;
   vb_arrays[1].ncomponents=4;
   vb_arrays[1].reg=REG_COLOR0;

   
   /* needed before starting 3d operation .. */
   reg_start(R300_RB3D_DSTCACHE_CTLSTAT,0);
	e32(0x0000000a);
   
   reg_start(0x4f18,0);
	e32(0x00000003);
	
   rmesa->hw.vte.cmd[1] = R300_VPORT_X_SCALE_ENA
				| R300_VPORT_X_OFFSET_ENA
				| R300_VPORT_Y_SCALE_ENA
				| R300_VPORT_Y_OFFSET_ENA
				| R300_VTX_W0_FMT;
   R300_STATECHANGE(rmesa, vte);
   
   r300EmitState(rmesa);
   
   FLAT_COLOR_PIPELINE.vertex_shader.matrix[0].length=16;
   memcpy(FLAT_COLOR_PIPELINE.vertex_shader.matrix[0].body.f, ctx->_ModelProjectMatrix.m, 16*4);

   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.length=4;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[0]=0.0;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[1]=0.0;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[2]=1.0;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[3]=0.0;
   
   program_pipeline(PASS_PREFIX &FLAT_COLOR_PIPELINE);
   cp_delay(PASS_PREFIX 15);
   
   /* We need LOAD_VBPNTR to setup AOS_ATTR fields.. the offsets are irrelevant */
   setup_AOS(PASS_PREFIX vb_arrays, 2);
   
   for(i=0; i < VB->PrimitiveCount; i++){
       GLuint prim = VB->Primitive[i].mode;
       GLuint start = VB->Primitive[i].start;
       GLuint length = VB->Primitive[i].count;
	r300_render_flat_primitive(rmesa, ctx, start, start + length, prim);
   	}
	
   sync_VAP(PASS_PREFIX_VOID);
   
   end_3d(PASS_PREFIX_VOID);
   
   /* Flush state - this reduces the chance that something else will mess with 
      the hardware in between */
   r300Flush(ctx);
   //fprintf(stderr, "\n");
   return GL_FALSE;
}

/* vertex buffer implementation */

/* We use the start part of GART texture buffer for vertices */

/* 8 is somewhat bogus... it is probably something like 24 */
#define R300_MAX_AOS_ARRAYS		8

static void upload_vertex_buffer(r300ContextPtr rmesa, 
	GLcontext *ctx, AOS_DATA *array, int *n_arrays)
{
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   int offset=0, idx=0;
   int i,j;
   radeonScreenPtr rsp=rmesa->radeon.radeonScreen;
   /* Not the most efficient implementation, but, for now, I just want something that
      works */
      /* to do - make single memcpy per column (is it possible ?) */
      /* to do - use dirty flags to avoid redundant copies */
#define UPLOAD_VECTOR(v, r, f)\
	{ \
	 /* Is the data dirty ? */ \
	if (v->flags & ((1<<v->size)-1)) { \
		fprintf(stderr, "size=%d vs stride=%d\n", v->size, v->stride); \
		if(v->size*4==v->stride){\
			/* fast path */  \
			memcpy(rsp->gartTextures.map+offset, v->data, v->stride*VB->Count); \
			} else { \
			for(i=0;i<VB->Count;i++){ \
				/* copy one vertex at a time*/ \
				memcpy(rsp->gartTextures.map+offset+i*v->size*4, VEC_ELT(v, GLfloat, i), v->size*4); \
				} \
			} \
		/* v->flags &= ~((1<<v->size)-1);*/ \
		} \
	array[idx].element_size=v->size; \
	array[idx].stride=v->size; \
	array[idx].format=(f); \
	array[idx].ncomponents=v->size; \
	array[idx].offset=rsp->gartTextures.handle+offset; \
	array[idx].reg=r; \
	offset+=v->size*4*VB->Count; \
	idx++; \
	}
	
UPLOAD_VECTOR(VB->ObjPtr, REG_COORDS, AOS_FORMAT_FLOAT);
UPLOAD_VECTOR(VB->ColorPtr[0], REG_COLOR0, AOS_FORMAT_FLOAT_COLOR);

*n_arrays=idx;
if(idx>=R300_MAX_AOS_ARRAYS){
	fprintf(stderr, "Aieee ! Maximum AOS arrays count exceeded.. \n");
	exit(-1);
	}
}

static void r300_render_vb_flat_primitive(r300ContextPtr rmesa, 
	GLcontext *ctx,
	int start,
	int end,
	int prim)
{
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
   int k, type, n_arrays;
   LOCAL_VARS
       	
   if(end<=start)return; /* do we need to watch for this ? */
   
   type=r300_get_primitive_type(rmesa, ctx, start, end, prim);
   if(type<0)return;

   fire_AOS(PASS_PREFIX end-start, type);
}

static VERTEX_SHADER_FRAGMENT default_vector_vsf={
	length: 4,
	body: { 
		f: {0.0, 0.0, 0.0, 1.0}
		}
	};

static GLboolean r300_run_vb_flat_render(GLcontext *ctx,
				 struct tnl_pipeline_stage *stage)
{
   r300ContextPtr rmesa = R300_CONTEXT(ctx);
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   int i, j, n_arrays;
   AOS_DATA vb_arrays[R300_MAX_AOS_ARRAYS];
   AOS_DATA vb_arrays2[R300_MAX_AOS_ARRAYS];
   LOCAL_VARS
	
	if (RADEON_DEBUG == DEBUG_PRIMS)
		fprintf(stderr, "%s\n", __FUNCTION__);

   /* setup array of structures data */

   upload_vertex_buffer(rmesa, ctx, vb_arrays, &n_arrays);
   fprintf(stderr, "Using %d AOS arrays\n", n_arrays);
   
   reg_start(R300_RB3D_DSTCACHE_CTLSTAT,0);
	e32(0x0000000a);
   
   reg_start(0x4f18,0);
	e32(0x00000003);
   
   r300EmitState(rmesa);
   
   FLAT_COLOR_PIPELINE.vertex_shader.matrix[0].length=16;
   memcpy(FLAT_COLOR_PIPELINE.vertex_shader.matrix[0].body.f, ctx->_ModelProjectMatrix.m, 16*4);

   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.length=4;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[0]=0.0;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[1]=0.0;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[2]=1.0;
   FLAT_COLOR_PIPELINE.vertex_shader.unknown2.body.f[3]=0.0;

   program_pipeline(PASS_PREFIX &FLAT_COLOR_PIPELINE);
      
   set_quad0(PASS_PREFIX 1.0,1.0,1.0,1.0);
   set_init21(PASS_PREFIX 0.0,1.0);
   
   for(i=0; i < VB->PrimitiveCount; i++){
       GLuint prim = VB->Primitive[i].mode;
       GLuint start = VB->Primitive[i].start;
       GLuint length = VB->Primitive[i].count;
       
        /* copy arrays */
        memcpy(vb_arrays2, vb_arrays, sizeof(AOS_DATA)*n_arrays);
	for(j=0;j<n_arrays;j++){
		vb_arrays2[j].offset+=vb_arrays2[j].stride*start*4;
		}
		
        setup_AOS(PASS_PREFIX vb_arrays2, n_arrays);
       
	r300_render_vb_flat_primitive(rmesa, ctx, start, start + length, prim);
   	}
	
   end_3d(PASS_PREFIX_VOID);
   
   /* Flush state - we are done drawing.. */
   r300Flush(ctx);
   fprintf(stderr, "\n");
   return GL_FALSE;
}

/* Textures... */

/* Immediate implementation - vertex data is sent via command stream */

static void r300_render_tex_primitive(r300ContextPtr rmesa, 
	GLcontext *ctx,
	int start,
	int end,
	int prim)
{
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
   int k, type;
   LOCAL_VARS
       	
   type=r300_get_primitive_type(rmesa, ctx, start, end, prim);
		
		fprintf(stderr,"ObjPtr: size=%d stride=%d\n", 
			VB->ObjPtr->size, VB->ObjPtr->stride);
		fprintf(stderr,"ColorPtr[0]: size=%d stride=%d\n", 
			VB->ColorPtr[0]->size, VB->ColorPtr[0]->stride);
   
   if(type<0)return;


   start_immediate_packet(end-start, type, 12);

	for(i=start;i<end;i++){
		#if 0
		fprintf(stderr, "* (%f %f %f %f) (%f %f %f %f)\n", 
			VEC_ELT(VB->ObjPtr, GLfloat, i)[0],
			VEC_ELT(VB->ObjPtr, GLfloat, i)[1],
			VEC_ELT(VB->ObjPtr, GLfloat, i)[2],
			VEC_ELT(VB->ObjPtr, GLfloat, i)[3],
			
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[0],
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[1],
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[2],
			VEC_ELT(VB->ColorPtr[0], GLfloat, i)[3]
			);
		#endif
		
		
		/* coordinates */
		output_vector(VB->ObjPtr, i);
		
		/* color components */
		output_vector(VB->ColorPtr[0], i);
		
		/* texture coordinates */
		output_vector(VB->TexCoordPtr[0], i);
		}

}

static GLboolean r300_run_tex_render(GLcontext *ctx,
				 struct tnl_pipeline_stage *stage)
{
   r300ContextPtr rmesa = R300_CONTEXT(ctx);
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
   AOS_DATA vb_arrays[3];
   /* Only do 2d textures */
   struct gl_texture_object *to=ctx->Texture.Unit[0].Current2D;
   radeonScreenPtr rsp=rmesa->radeon.radeonScreen;
   LOCAL_VARS
	
   
   /* Flush state - make sure command buffer is nice and large */
   r300Flush(ctx);
   
   fprintf(stderr, "You can enable texture drawing in %s:%s \n", __FILE__, __FUNCTION__);
   return GL_TRUE;
   
	if (RADEON_DEBUG == DEBUG_PRIMS)
		fprintf(stderr, "%s\n", __FUNCTION__);

   /* setup array of structures data */

   /* Note: immediate vertex data includes all coordinates.
     To save bandwidth use either VBUF or state-based vertex generation */
    /* xyzw */
   vb_arrays[0].element_size=4;
   vb_arrays[0].stride=4;
   vb_arrays[0].offset=0; /* Not used */
   vb_arrays[0].format=AOS_FORMAT_FLOAT;
   vb_arrays[0].ncomponents=4;
   vb_arrays[0].reg=REG_COORDS;

    /* color */
   vb_arrays[1].element_size=4;
   vb_arrays[1].stride=4;
   vb_arrays[1].offset=0; /* Not used */
   vb_arrays[1].format=AOS_FORMAT_FLOAT_COLOR;
   vb_arrays[1].ncomponents=4;
   vb_arrays[1].reg=REG_COLOR0;

    /* texture coordinates */
   vb_arrays[2].element_size=4;
   vb_arrays[2].stride=4;
   vb_arrays[2].offset=0; /* Not used */
   vb_arrays[2].format=AOS_FORMAT_FLOAT;
   vb_arrays[2].ncomponents=4;
   vb_arrays[2].reg=REG_TEX0;

   
   /* needed before starting 3d operation .. */
   reg_start(R300_RB3D_DSTCACHE_CTLSTAT,0);
	e32(0x0000000a);
   
   reg_start(0x4f18,0);
	e32(0x00000003);
	
   
   rmesa->hw.vte.cmd[1] = R300_VPORT_X_SCALE_ENA
				| R300_VPORT_X_OFFSET_ENA
				| R300_VPORT_Y_SCALE_ENA
				| R300_VPORT_Y_OFFSET_ENA
				| R300_VTX_W0_FMT;
   R300_STATECHANGE(rmesa, vte);
   
   r300EmitState(rmesa);
   
   SINGLE_TEXTURE_PIPELINE.vertex_shader.matrix[0].length=16;
   memcpy(SINGLE_TEXTURE_PIPELINE.vertex_shader.matrix[0].body.f, ctx->_ModelProjectMatrix.m, 16*4);

   SINGLE_TEXTURE_PIPELINE.vertex_shader.unknown2.length=4;
   SINGLE_TEXTURE_PIPELINE.vertex_shader.unknown2.body.f[0]=0.0;
   SINGLE_TEXTURE_PIPELINE.vertex_shader.unknown2.body.f[1]=0.0;
   SINGLE_TEXTURE_PIPELINE.vertex_shader.unknown2.body.f[2]=1.0;
   SINGLE_TEXTURE_PIPELINE.vertex_shader.unknown2.body.f[3]=0.0;

   /* Put it in the beginning of texture memory */
   SINGLE_TEXTURE_PIPELINE.texture_unit[0].offset=rsp->gartTextures.handle;
   
   /* Fill texture with some random data */
   for(i=0;i<1000;i++)((int *)(rsp->gartTextures.map))[i]=rand();
   
   /* Upload texture, a hack, really  we can do a lot better */
   #if 0
   memcpy(rsp->gartTextures.map, to->Image[0][0]->Data, to->Image[0][0]->RowStride*to->Image[0][0]->Height*4);
   #endif

   /* Program RS unit. This needs to be moved into R300 pipeline */   
reg_start(R300_RS_CNTL_0,1);
	/* R300_RS_CNTL_0(4300) */
	e32(0x00040084);
	/* RS_INST_COUNT(4304) */
	e32(0x000000c0);

reg_start(R300_RS_ROUTE_0,0);
	e32(0x00024008);

reg_start(R300_RS_INTERP_0,7);
	/* X_MEM0_0(4310) */
	e32(0x00d10000);
	/* X_MEM0_1(4314) */
	e32(0x00d10044);
	/* X_MEM0_2(4318) */
	e32(0x00d10084);
	/* X_MEM0_3(431c) */
	e32(0x00d100c4);
	/* X_MEM0_4(4320) */
	e32(0x00d10004);
	/* X_MEM0_5(4324) */
	e32(0x00d10004);
	/* X_MEM0_6(4328) */
	e32(0x00d10004);
	/* X_MEM0_7(432c) */
	e32(0x00d10004);

  reg_start(R300_RS_CNTL_0,0);
	e32(0x00040084);

   /* Magic register - note it is right after 20b0 */
   
   reg_start(0x20b4,0);
	e32(0x0000000c);
   
   program_pipeline(PASS_PREFIX &SINGLE_TEXTURE_PIPELINE);
   
   sync_VAP(PASS_PREFIX_VOID);
   cp_delay(PASS_PREFIX 15);
      
   /* We need LOAD_VBPNTR to setup AOS_ATTR fields.. the offsets are irrelevant */
   setup_AOS(PASS_PREFIX vb_arrays, 3);
   
   for(i=0; i < VB->PrimitiveCount; i++){
       GLuint prim = VB->Primitive[i].mode;
       GLuint start = VB->Primitive[i].start;
       GLuint length = VB->Primitive[i].count;
	r300_render_tex_primitive(rmesa, ctx, start, start + length, prim);
   	}
	
   sync_VAP(PASS_PREFIX_VOID);
   
   end_3d(PASS_PREFIX_VOID);
   
   /* Flush state - this reduces the chance that something else will mess with 
      the hardware in between */
   r300Flush(ctx);
   
   fprintf(stderr, "\n");
   return GL_FALSE;
}


/**
 * Called by the pipeline manager to render a batch of primitives.
 * We can return true to pass on to the next stage (i.e. software
 * rasterization) or false to indicate that the pipeline has finished
 * after we render something.
 */
static GLboolean r300_run_render(GLcontext *ctx,
				 struct tnl_pipeline_stage *stage)
{
   r300ContextPtr rmesa = R300_CONTEXT(ctx);
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;
	
	if (RADEON_DEBUG == DEBUG_PRIMS)
		fprintf(stderr, "%s\n", __FUNCTION__);

   #if 1
   	/* Just switch between pipelines.. We could possibly do better.. (?) */
        if(ctx->Texture.Unit[0].Enabled)
        	return r300_run_tex_render(ctx, stage);
		else
        	return r300_run_flat_render(ctx, stage);
   #else
	return GL_TRUE;
   #endif

#if 0
   mgaContextPtr mmesa = MGA_CONTEXT(ctx);
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;

   /* Don't handle clipping or indexed vertices or vertex manipulations.
    */
   if (mmesa->RenderIndex != 0 ||
       !mga_validate_render( ctx, VB )) {
      return GL_TRUE;
   }

   tnl->Driver.Render.Start( ctx );
   mmesa->SetupNewInputs = ~0;

   for (i = 0 ; i < VB->PrimitiveCount ; i++)
   {
      GLuint prim = VB->Primitive[i].mode;
      GLuint start = VB->Primitive[i].start;
      GLuint length = VB->Primitive[i].count;

      if (!length)
	 continue;

      mga_render_tab_verts[prim & PRIM_MODE_MASK]( ctx, start, start + length,
						   prim);
   }

   tnl->Driver.Render.Finish( ctx );

   return GL_FALSE;		/* finished the pipe */
#endif
}


/**
 * Called by the pipeline manager once before rendering.
 * We check the GL state here to
 *  a) decide whether we can do the current state in hardware and
 *  b) update hardware registers
 */
#define FALLBACK_IF(expr) \
do {										\
	if (expr) {								\
		if (RADEON_DEBUG & DEBUG_FALLBACKS)				\
			fprintf(stderr, "%s: fallback:%s\n",			\
				__FUNCTION__, #expr);				\
		stage->active = GL_FALSE;					\
		return;								\
	}									\
} while(0)

static void r300_check_render(GLcontext *ctx, struct tnl_pipeline_stage *stage)
{
	r300ContextPtr r300 = R300_CONTEXT(ctx);
	int i;

	if (RADEON_DEBUG & DEBUG_STATE)
		fprintf(stderr, "%s\n", __FUNCTION__);

	/* We only support rendering in hardware for now */
	if (ctx->RenderMode != GL_RENDER) {
		stage->active = GL_FALSE;
		return;
	}

	// I failed to figure out how dither works in hardware,
	// let's just ignore it for now
	//FALLBACK_IF(ctx->Color.DitherFlag);

	/* I'm almost certain I forgot something here */
	FALLBACK_IF(ctx->Color.AlphaEnabled); // GL_ALPHA_TEST
	FALLBACK_IF(ctx->Color.BlendEnabled); // GL_BLEND
	FALLBACK_IF(ctx->Fog.Enabled); // GL_FOG
	FALLBACK_IF(ctx->Line.SmoothFlag); // GL_LINE_SMOOTH
	FALLBACK_IF(ctx->Line.StippleFlag); // GL_LINE_STIPPLE
	FALLBACK_IF(ctx->Point.SmoothFlag); // GL_POINT_SMOOTH
	if (ctx->Extensions.NV_point_sprite || ctx->Extensions.ARB_point_sprite)
		FALLBACK_IF(ctx->Point.PointSprite); // GL_POINT_SPRITE_NV
	FALLBACK_IF(ctx->Polygon.OffsetPoint); // GL_POLYGON_OFFSET_POINT
	FALLBACK_IF(ctx->Polygon.OffsetLine); // GL_POLYGON_OFFSET_LINE
	FALLBACK_IF(ctx->Polygon.OffsetFill); // GL_POLYGON_OFFSET_FILL
	FALLBACK_IF(ctx->Polygon.SmoothFlag); // GL_POLYGON_SMOOTH
	FALLBACK_IF(ctx->Polygon.StippleFlag); // GL_POLYGON_STIPPLE
	FALLBACK_IF(ctx->Stencil.Enabled); // GL_STENCIL_TEST
	FALLBACK_IF(ctx->Multisample.Enabled); // GL_MULTISAMPLE_ARB

	/* One step at a time - let one texture pass.. */
	for (i = 1; i < ctx->Const.MaxTextureUnits; i++)
		FALLBACK_IF(ctx->Texture.Unit[i].Enabled);


	/* let r300_run_render do its job */
	#if 0  
	stage->active = GL_FALSE;
	#endif
}


static void dtr(struct tnl_pipeline_stage *stage)
{
	(void)stage;
}

const struct tnl_pipeline_stage _r300_render_stage = {
	"r300 hw rasterize",
	_NEW_ALL,		/* re-check (always re-check for now) */
	0,			/* re-run (always runs) */
	GL_TRUE,		/* active */
	0, 0,			/* inputs (set in check_render), outputs */
	0, 0,			/* changed_inputs, private */
	dtr,			/* destructor */
	r300_check_render,	/* check */
	r300_run_render		/* run */
};