summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/radeon/radeon_common_context.h
blob: 024e31f8ec7dc1c023bc27f11270f50c5810762b (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

#ifndef COMMON_CONTEXT_H
#define COMMON_CONTEXT_H

#include "main/mm.h"
#include "math/m_vector.h"
#include "texmem.h"
#include "tnl/t_context.h"
#include "main/colormac.h"

#include "radeon_debug.h"
#include "radeon_screen.h"
#include "radeon_drm.h"
#include "dri_util.h"
#include "tnl/t_vertex.h"

#include "dri_metaops.h"
struct radeon_context;

#include "radeon_bocs_wrapper.h"

/* This union is used to avoid warnings/miscompilation
   with float to uint32_t casts due to strict-aliasing */
typedef union { GLfloat f; uint32_t ui32; } float_ui32_type;

struct radeon_context;
typedef struct radeon_context radeonContextRec;
typedef struct radeon_context *radeonContextPtr;


#define TEX_0   0x1
#define TEX_1   0x2
#define TEX_2   0x4
#define TEX_3	0x8
#define TEX_4	0x10
#define TEX_5	0x20

/* Rasterizing fallbacks */
/* See correponding strings in r200_swtcl.c */
#define RADEON_FALLBACK_TEXTURE		0x0001
#define RADEON_FALLBACK_DRAW_BUFFER	0x0002
#define RADEON_FALLBACK_STENCIL		0x0004
#define RADEON_FALLBACK_RENDER_MODE	0x0008
#define RADEON_FALLBACK_BLEND_EQ	0x0010
#define RADEON_FALLBACK_BLEND_FUNC	0x0020
#define RADEON_FALLBACK_DISABLE 	0x0040
#define RADEON_FALLBACK_BORDER_MODE	0x0080
#define RADEON_FALLBACK_DEPTH_BUFFER	0x0100
#define RADEON_FALLBACK_STENCIL_BUFFER  0x0200

#define R200_FALLBACK_TEXTURE           0x01
#define R200_FALLBACK_DRAW_BUFFER       0x02
#define R200_FALLBACK_STENCIL           0x04
#define R200_FALLBACK_RENDER_MODE       0x08
#define R200_FALLBACK_DISABLE           0x10
#define R200_FALLBACK_BORDER_MODE       0x20

#define RADEON_TCL_FALLBACK_RASTER            0x1 /* rasterization */
#define RADEON_TCL_FALLBACK_UNFILLED          0x2 /* unfilled tris */
#define RADEON_TCL_FALLBACK_LIGHT_TWOSIDE     0x4 /* twoside tris */
#define RADEON_TCL_FALLBACK_MATERIAL          0x8 /* material in vb */
#define RADEON_TCL_FALLBACK_TEXGEN_0          0x10 /* texgen, unit 0 */
#define RADEON_TCL_FALLBACK_TEXGEN_1          0x20 /* texgen, unit 1 */
#define RADEON_TCL_FALLBACK_TEXGEN_2          0x40 /* texgen, unit 2 */
#define RADEON_TCL_FALLBACK_TCL_DISABLE       0x80 /* user disable */
#define RADEON_TCL_FALLBACK_FOGCOORDSPEC      0x100 /* fogcoord, sep. spec light */

/* The blit width for texture uploads
 */
#define BLIT_WIDTH_BYTES 1024

/* Use the templated vertex format:
 */
#define COLOR_IS_RGBA
#define TAG(x) radeon##x
#include "tnl_dd/t_dd_vertex.h"
#undef TAG

#define RADEON_RB_CLASS 0xdeadbeef

struct radeon_renderbuffer
{
	struct gl_renderbuffer base;
	struct radeon_bo *bo;
	unsigned int cpp;
	/* unsigned int offset; */
	unsigned int pitch;

	uint32_t draw_offset; /* FBO */
	/* boo Xorg 6.8.2 compat */
	int has_surface;

	GLuint pf_pending;  /**< sequence number of pending flip */
	GLuint vbl_pending;   /**< vblank sequence number of pending flip */
	__DRIdrawable *dPriv;

	/* r6xx+ tiling */
	GLuint tile_config;
	GLint group_bytes;
	GLint num_channels;
	GLint num_banks;
	GLint r7xx_bank_op;
};

struct radeon_framebuffer
{
	struct gl_framebuffer base;

	struct radeon_renderbuffer *color_rb[2];

	GLuint vbl_waited;

	/* buffer swap */
	int64_t swap_ust;
	int64_t swap_missed_ust;

	GLuint swap_count;
	GLuint swap_missed_count;

	/* Drawable page flipping state */
	GLboolean pf_active;
	GLint pf_current_page;
	GLint pf_num_pages;

};


struct radeon_colorbuffer_state {
	GLuint clear;
	int roundEnable;
	struct gl_renderbuffer *rb;
	uint32_t draw_offset; /* offset into color renderbuffer - FBOs */
};

struct radeon_depthbuffer_state {
	GLuint clear;
	struct gl_renderbuffer *rb;
};

struct radeon_scissor_state {
	drm_clip_rect_t rect;
	GLboolean enabled;

	GLuint numClipRects;	/* Cliprects active */
	GLuint numAllocedClipRects;	/* Cliprects available */
	drm_clip_rect_t *pClipRects;
};

struct radeon_stencilbuffer_state {
	GLuint clear;		/* rb3d_stencilrefmask value */
};

struct radeon_state_atom {
	struct radeon_state_atom *next, *prev;
	const char *name;	/* for debug */
	int cmd_size;		/* size in bytes */
        GLuint idx;
	GLuint is_tcl;
        GLuint *cmd;		/* one or more cmd's */
	GLuint *lastcmd;		/* one or more cmd's */
	GLboolean dirty;	/* dirty-mark in emit_state_list */
        int (*check) (GLcontext *, struct radeon_state_atom *atom); /* is this state active? */
        void (*emit) (GLcontext *, struct radeon_state_atom *atom);
};

struct radeon_hw_state {
  	/* Head of the linked list of state atoms. */
	struct radeon_state_atom atomlist;
	int max_state_size;	/* Number of bytes necessary for a full state emit. */
	int max_post_flush_size; /* Number of bytes necessary for post flushing emits */
	GLboolean is_dirty, all_dirty;
};


/* Texture related */
typedef struct _radeon_texture_image radeon_texture_image;

struct _radeon_texture_image {
	struct gl_texture_image base;

	/**
	 * If mt != 0, the image is stored in hardware format in the
	 * given mipmap tree. In this case, base.Data may point into the
	 * mapping of the buffer object that contains the mipmap tree.
	 *
	 * If mt == 0, the image is stored in normal memory pointed to
	 * by base.Data.
	 */
	struct _radeon_mipmap_tree *mt;
	struct radeon_bo *bo;

	int mtlevel; /** if mt != 0, this is the image's level in the mipmap tree */
	int mtface; /** if mt != 0, this is the image's face in the mipmap tree */
};


static INLINE radeon_texture_image *get_radeon_texture_image(struct gl_texture_image *image)
{
	return (radeon_texture_image*)image;
}


typedef struct radeon_tex_obj radeonTexObj, *radeonTexObjPtr;

#define RADEON_TXO_MICRO_TILE               (1 << 3)

/* Texture object in locally shared texture space.
 */
struct radeon_tex_obj {
	struct gl_texture_object base;
	struct _radeon_mipmap_tree *mt;

	/**
	 * This is true if we've verified that the mipmap tree above is complete
	 * and so on.
	 */
	GLboolean validated;
	/* Minimum LOD to be used during rendering */
	unsigned minLod;
	/* Miximum LOD to be used during rendering */
	unsigned maxLod;

	GLuint override_offset;
	GLboolean image_override; /* Image overridden by GLX_EXT_tfp */
	GLuint tile_bits;	/* hw texture tile bits used on this texture */
        struct radeon_bo *bo;

	GLuint pp_txfilter;	/* hardware register values */
	GLuint pp_txformat;
	GLuint pp_txformat_x;
	GLuint pp_txsize;	/* npot only */
	GLuint pp_txpitch;	/* npot only */
	GLuint pp_border_color;
	GLuint pp_cubic_faces;	/* cube face 1,2,3,4 log2 sizes */

        GLuint pp_txfilter_1;	/*  r300 */

	/* r700 texture states */
	GLuint SQ_TEX_RESOURCE0;
	GLuint SQ_TEX_RESOURCE1;
	GLuint SQ_TEX_RESOURCE2;
	GLuint SQ_TEX_RESOURCE3;
	GLuint SQ_TEX_RESOURCE4;
	GLuint SQ_TEX_RESOURCE5;
	GLuint SQ_TEX_RESOURCE6;

    GLuint SQ_TEX_RESOURCE7;

	GLuint SQ_TEX_SAMPLER0;
	GLuint SQ_TEX_SAMPLER1;
	GLuint SQ_TEX_SAMPLER2;

	GLuint TD_PS_SAMPLER0_BORDER_RED;
	GLuint TD_PS_SAMPLER0_BORDER_GREEN;
	GLuint TD_PS_SAMPLER0_BORDER_BLUE;
	GLuint TD_PS_SAMPLER0_BORDER_ALPHA;

	GLboolean border_fallback;


};

static INLINE radeonTexObj* radeon_tex_obj(struct gl_texture_object *texObj)
{
	return (radeonTexObj*)texObj;
}

/* occlusion query */
struct radeon_query_object {
	struct gl_query_object Base;
	struct radeon_bo *bo;
	int curr_offset;
	GLboolean emitted_begin;

	/* Double linked list of not flushed query objects */
	struct radeon_query_object *prev, *next;
};

/* Need refcounting on dma buffers:
 */
struct radeon_dma_buffer {
	int refcount;		/* the number of retained regions in buf */
	drmBufPtr buf;
};

struct radeon_aos {
	struct radeon_bo *bo; /** Buffer object where vertex data is stored */
	int offset; /** Offset into buffer object, in bytes */
	int components; /** Number of components per vertex */
	int stride; /** Stride in dwords (may be 0 for repeating) */
	int count; /** Number of vertices */
};

#define DMA_BO_FREE_TIME 100

struct radeon_dma_bo {
  struct radeon_dma_bo *next, *prev;
  struct radeon_bo *bo;
  int expire_counter;
};

struct radeon_dma {
        /* Active dma region.  Allocations for vertices and retained
         * regions come from here.  Also used for emitting random vertices,
         * these may be flushed by calling flush_current();
         */
	struct radeon_dma_bo free;
	struct radeon_dma_bo wait;
	struct radeon_dma_bo reserved;
        size_t current_used; /** Number of bytes allocated and forgotten about */
        size_t current_vertexptr; /** End of active vertex region */
        size_t minimum_size;

        /**
         * If current_vertexptr != current_used then flush must be non-zero.
         * flush must be called before non-active vertex allocations can be
         * performed.
         */
        void (*flush) (GLcontext *);
};

/* radeon_swtcl.c
 */
struct radeon_swtcl_info {

	GLuint RenderIndex;
	GLuint vertex_size;
	GLubyte *verts;

	/* Fallback rasterization functions
	 */
	GLuint hw_primitive;
	GLenum render_primitive;
	GLuint numverts;

	struct tnl_attr_map vertex_attrs[VERT_ATTRIB_MAX];
	GLuint vertex_attr_count;

	GLuint emit_prediction;
        struct radeon_bo *bo;
};

#define RADEON_MAX_AOS_ARRAYS		16
struct radeon_tcl_info {
	struct radeon_aos aos[RADEON_MAX_AOS_ARRAYS];
	GLuint aos_count;
	struct radeon_bo *elt_dma_bo; /** Buffer object that contains element indices */
	int elt_dma_offset; /** Offset into this buffer object, in bytes */
};

struct radeon_ioctl {
	GLuint vertex_offset;
	GLuint vertex_max;
	struct radeon_bo *bo;
	GLuint vertex_size;
};

#define RADEON_MAX_PRIMS 64

struct radeon_prim {
	GLuint start;
	GLuint end;
	GLuint prim;
};

static INLINE GLuint radeonPackColor(GLuint cpp,
                                     GLubyte r, GLubyte g,
                                     GLubyte b, GLubyte a)
{
	switch (cpp) {
	case 2:
		return PACK_COLOR_565(r, g, b);
	case 4:
		return PACK_COLOR_8888(a, r, g, b);
	default:
		return 0;
	}
}

#define MAX_CMD_BUF_SZ (16*1024)

#define MAX_DMA_BUF_SZ (64*1024)

struct radeon_store {
	GLuint statenr;
	GLuint primnr;
	char cmd_buf[MAX_CMD_BUF_SZ];
	int cmd_used;
	int elts_start;
};

struct radeon_dri_mirror {
	__DRIcontext *context;	/* DRI context */
	__DRIscreen *screen;	/* DRI screen */

	drm_context_t hwContext;
	drm_hw_lock_t *hwLock;
	int hwLockCount;
	int fd;
	int drmMinor;
};

typedef void (*radeon_tri_func) (radeonContextPtr,
				 radeonVertex *,
				 radeonVertex *, radeonVertex *);

typedef void (*radeon_line_func) (radeonContextPtr,
				  radeonVertex *, radeonVertex *);

typedef void (*radeon_point_func) (radeonContextPtr, radeonVertex *);

#define RADEON_MAX_BOS 32
struct radeon_state {
	struct radeon_colorbuffer_state color;
	struct radeon_depthbuffer_state depth;
	struct radeon_scissor_state scissor;
	struct radeon_stencilbuffer_state stencil;
};

/**
 * This structure holds the command buffer while it is being constructed.
 *
 * The first batch of commands in the buffer is always the state that needs
 * to be re-emitted when the context is lost. This batch can be skipped
 * otherwise.
 */
struct radeon_cmdbuf {
	struct radeon_cs_manager    *csm;
	struct radeon_cs            *cs;
	int size; /** # of dwords total */
	unsigned int flushing:1; /** whether we're currently in FlushCmdBufLocked */
};

struct radeon_context {
   GLcontext *glCtx;
   radeonScreenPtr radeonScreen;	/* Screen private DRI data */

   /* Texture object bookkeeping
    */
   int                   texture_depth;
   float                 initialMaxAnisotropy;
   uint32_t              texture_row_align;
   uint32_t              texture_rect_row_align;
   uint32_t              texture_compressed_row_align;

  struct radeon_dma dma;
  struct radeon_hw_state hw;
   /* Rasterization and vertex state:
    */
   GLuint TclFallback;
   GLuint Fallback;
   GLuint NewGLState;
   DECLARE_RENDERINPUTS(tnl_index_bitset);	/* index of bits for last tnl_install_attrs */

   /* Drawable, cliprect and scissor information */
   GLuint numClipRects;	/* Cliprects for the draw buffer */
   drm_clip_rect_t *pClipRects;
   unsigned int lastStamp;
   drm_radeon_sarea_t *sarea;	/* Private SAREA data */

   /* Mirrors of some DRI state */
   struct radeon_dri_mirror dri;

   /* Busy waiting */
   GLuint do_usleeps;
   GLuint do_irqs;
   GLuint irqsEmitted;
   drm_radeon_irq_wait_t iw;

   /* Derived state - for r300 only */
   struct radeon_state state;

   struct radeon_swtcl_info swtcl;
   struct radeon_tcl_info tcl;
   /* Configuration cache
    */
   driOptionCache optionCache;

   struct radeon_cmdbuf cmdbuf;

   struct radeon_debug debug;

  drm_clip_rect_t fboRect;
  GLboolean constant_cliprect; /* use for FBO or DRI2 rendering */
  GLboolean front_cliprects;

   /**
    * Set if rendering has occured to the drawable's front buffer.
    *
    * This is used in the DRI2 case to detect that glFlush should also copy
    * the contents of the fake front buffer to the real front buffer.
    */
   GLboolean front_buffer_dirty;

   /**
    * Track whether front-buffer rendering is currently enabled
    *
    * A separate flag is used to track this in order to support MRT more
    * easily.
    */
   GLboolean is_front_buffer_rendering;

   /**
    * Track whether front-buffer is the current read target.
    *
    * This is closely associated with is_front_buffer_rendering, but may
    * be set separately.  The DRI2 fake front buffer must be referenced
    * either way.
    */
   GLboolean is_front_buffer_reading;

   struct dri_metaops meta;

   struct {
	struct radeon_query_object *current;
	struct radeon_state_atom queryobj;
   } query;

   struct {
	   void (*get_lock)(radeonContextPtr radeon);
	   void (*update_viewport_offset)(GLcontext *ctx);
	   void (*emit_cs_header)(struct radeon_cs *cs, radeonContextPtr rmesa);
	   void (*swtcl_flush)(GLcontext *ctx, uint32_t offset);
	   void (*pre_emit_atoms)(radeonContextPtr rmesa);
	   void (*pre_emit_state)(radeonContextPtr rmesa);
	   void (*fallback)(GLcontext *ctx, GLuint bit, GLboolean mode);
	   void (*free_context)(GLcontext *ctx);
	   void (*emit_query_finish)(radeonContextPtr radeon);
	   void (*update_scissor)(GLcontext *ctx);
	   unsigned (*check_blit)(gl_format mesa_format);
	   unsigned (*blit)(GLcontext *ctx,
                        struct radeon_bo *src_bo,
                        intptr_t src_offset,
                        gl_format src_mesaformat,
                        unsigned src_pitch,
                        unsigned src_width,
                        unsigned src_height,
                        unsigned src_x_offset,
                        unsigned src_y_offset,
                        struct radeon_bo *dst_bo,
                        intptr_t dst_offset,
                        gl_format dst_mesaformat,
                        unsigned dst_pitch,
                        unsigned dst_width,
                        unsigned dst_height,
                        unsigned dst_x_offset,
                        unsigned dst_y_offset,
                        unsigned reg_width,
                        unsigned reg_height,
                        unsigned flip_y);
	   unsigned (*is_format_renderable)(gl_format mesa_format);
   } vtbl;
};

#define RADEON_CONTEXT(glctx) ((radeonContextPtr)(ctx->DriverCtx))

static inline __DRIdrawable* radeon_get_drawable(radeonContextPtr radeon)
{
	return radeon->dri.context->driDrawablePriv;
}

static inline __DRIdrawable* radeon_get_readable(radeonContextPtr radeon)
{
	return radeon->dri.context->driReadablePriv;
}

/**
 * This function takes a float and packs it into a uint32_t
 */
static INLINE uint32_t radeonPackFloat32(float fl)
{
	union {
		float fl;
		uint32_t u;
	} u;

	u.fl = fl;
	return u.u;
}

/* This is probably wrong for some values, I need to test this
 * some more.  Range checking would be a good idea also..
 *
 * But it works for most things.  I'll fix it later if someone
 * else with a better clue doesn't
 */
static INLINE uint32_t radeonPackFloat24(float f)
{
	float mantissa;
	int exponent;
	uint32_t float24 = 0;

	if (f == 0.0)
		return 0;

	mantissa = frexpf(f, &exponent);

	/* Handle -ve */
	if (mantissa < 0) {
		float24 |= (1 << 23);
		mantissa = mantissa * -1.0;
	}
	/* Handle exponent, bias of 63 */
	exponent += 62;
	float24 |= (exponent << 16);
	/* Kill 7 LSB of mantissa */
	float24 |= (radeonPackFloat32(mantissa) & 0x7FFFFF) >> 7;

	return float24;
}

GLboolean radeonInitContext(radeonContextPtr radeon,
			    struct dd_function_table* functions,
			    const __GLcontextModes * glVisual,
			    __DRIcontext * driContextPriv,
			    void *sharedContextPrivate);

void radeonCleanupContext(radeonContextPtr radeon);
GLboolean radeonUnbindContext(__DRIcontext * driContextPriv);
void radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable,
				 GLboolean front_only);
GLboolean radeonMakeCurrent(__DRIcontext * driContextPriv,
			    __DRIdrawable * driDrawPriv,
			    __DRIdrawable * driReadPriv);
extern void radeonDestroyContext(__DRIcontext * driContextPriv);
void radeon_prepare_render(radeonContextPtr radeon);

#endif