summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/radeon/radeon_span.c
blob: d603f52df7be7156266644c056e6b24f256c70e7 (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
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
/**************************************************************************

Copyright (C) The Weather Channel, Inc.  2002.  All Rights Reserved.
Copyright 2000, 2001 ATI Technologies Inc., Ontario, Canada, and
                     VA Linux Systems Inc., Fremont, California.

The Weather Channel (TM) funded Tungsten Graphics to develop the
initial release of the Radeon 8500 driver under the XFree86 license.
This notice must be preserved.

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 (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 NONINFRINGEMENT.
IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS 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:
 *   Kevin E. Martin <martin@valinux.com>
 *   Gareth Hughes <gareth@valinux.com>
 *   Keith Whitwell <keith@tungstengraphics.com>
 *
 */

#include "main/glheader.h"
#include "swrast/swrast.h"

#include "radeon_common.h"
#include "radeon_lock.h"
#include "radeon_span.h"

#define DBG 0

static void radeonSetSpanFunctions(struct radeon_renderbuffer *rrb);


/* r200 depth buffer is always tiled - this is the formula
   according to the docs unless I typo'ed in it
*/
#if defined(RADEON_COMMON_FOR_R200)
static GLubyte *r200_depth_2byte(const struct radeon_renderbuffer * rrb,
				 GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    GLint offset;
    if (rrb->has_surface) {
	offset = x * rrb->cpp + y * rrb->pitch;
    } else {
	GLuint b;
	offset = 0;
	b = (((y  >> 4) * (rrb->pitch >> 8) + (x >> 6)));
	offset += (b >> 1) << 12;
	offset += (((rrb->pitch >> 8) & 0x1) ? (b & 0x1) : ((b & 0x1) ^ ((y >> 4) & 0x1))) << 11;
	offset += ((y >> 2) & 0x3) << 9;
	offset += ((x >> 3) & 0x1) << 8;
	offset += ((x >> 4) & 0x3) << 6;
	offset += ((x >> 2) & 0x1) << 5;
	offset += ((y >> 1) & 0x1) << 4;
	offset += ((x >> 1) & 0x1) << 3;
	offset += (y & 0x1) << 2;
	offset += (x & 0x1) << 1;
    }
    return &ptr[offset];
}

static GLubyte *r200_depth_4byte(const struct radeon_renderbuffer * rrb,
				 GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    GLint offset;
    if (rrb->has_surface) {
	offset = x * rrb->cpp + y * rrb->pitch;
    } else {
	GLuint b;
	offset = 0;
	b = (((y & 0x7ff) >> 4) * (rrb->pitch >> 7) + (x >> 5));
	offset += (b >> 1) << 12;
	offset += (((rrb->pitch >> 7) & 0x1) ? (b & 0x1) : ((b & 0x1) ^ ((y >> 4) & 0x1))) << 11;
	offset += ((y >> 2) & 0x3) << 9;
	offset += ((x >> 2) & 0x1) << 8;
	offset += ((x >> 3) & 0x3) << 6;
	offset += ((y >> 1) & 0x1) << 5;
	offset += ((x >> 1) & 0x1) << 4;
	offset += (y & 0x1) << 3;
	offset += (x & 0x1) << 2;
    }
    return &ptr[offset];
}
#endif

/* r600 tiling
 * two main types:
 * - 1D (akin to macro-linear/micro-tiled on older asics)
 * - 2D (akin to macro-tiled/micro-tiled on older asics)
 * only 1D tiling is implemented below
 */
#if defined(RADEON_COMMON_FOR_R600)
static inline GLint r600_1d_tile_helper(const struct radeon_renderbuffer * rrb,
					GLint x, GLint y, GLint is_depth, GLint is_stencil)
{
    GLint element_bytes = rrb->cpp;
    GLint num_samples = 1;
    GLint tile_width = 8;
    GLint tile_height = 8;
    GLint tile_thickness = 1;
    GLint pitch_elements = rrb->pitch / element_bytes;
    GLint height = rrb->base.Height;
    GLint z = 0;
    GLint sample_number = 0;
    /* */
    GLint tile_bytes;
    GLint tiles_per_row;
    GLint tiles_per_slice;
    GLint slice_offset;
    GLint tile_row_index;
    GLint tile_column_index;
    GLint tile_offset;
    GLint pixel_number = 0;
    GLint element_offset;
    GLint offset = 0;

    tile_bytes = tile_width * tile_height * tile_thickness * element_bytes * num_samples;
    tiles_per_row = pitch_elements / tile_width;
    tiles_per_slice = tiles_per_row * (height / tile_height);
    slice_offset = (z / tile_thickness) * tiles_per_slice * tile_bytes;
    tile_row_index = y / tile_height;
    tile_column_index = x / tile_width;
    tile_offset = ((tile_row_index * tiles_per_row) + tile_column_index) * tile_bytes;

    if (is_depth) {
	    GLint pixel_offset = 0;

	    pixel_number |= ((x >> 0) & 1) << 0; // pn[0] = x[0]
	    pixel_number |= ((y >> 0) & 1) << 1; // pn[1] = y[0]
	    pixel_number |= ((x >> 1) & 1) << 2; // pn[2] = x[1]
	    pixel_number |= ((y >> 1) & 1) << 3; // pn[3] = y[1]
	    pixel_number |= ((x >> 2) & 1) << 4; // pn[4] = x[2]
	    pixel_number |= ((y >> 2) & 1) << 5; // pn[5] = y[2]
	    switch (element_bytes) {
	    case 2:
		    pixel_offset = pixel_number * element_bytes * num_samples;
		    break;
	    case 4:
		    /* stencil and depth data are stored separately within a tile.
		     * stencil is stored in a contiguous tile before the depth tile.
		     * stencil element is 1 byte, depth element is 3 bytes.
		     * stencil tile is 64 bytes.
		     */
		    if (is_stencil)
			    pixel_offset = pixel_number * 1 * num_samples;
		    else
			    pixel_offset = (pixel_number * 3 * num_samples) + 64;
		    break;
	    }
	    element_offset = pixel_offset + (sample_number * element_bytes);
    } else {
	    GLint sample_offset;

	    switch (element_bytes) {
	    case 1:
		    pixel_number |= ((x >> 0) & 1) << 0; // pn[0] = x[0]
		    pixel_number |= ((x >> 1) & 1) << 1; // pn[1] = x[1]
		    pixel_number |= ((x >> 2) & 1) << 2; // pn[2] = x[2]
		    pixel_number |= ((y >> 1) & 1) << 3; // pn[3] = y[1]
		    pixel_number |= ((y >> 0) & 1) << 4; // pn[4] = y[0]
		    pixel_number |= ((y >> 2) & 1) << 5; // pn[5] = y[2]
		    break;
	    case 2:
		    pixel_number |= ((x >> 0) & 1) << 0; // pn[0] = x[0]
		    pixel_number |= ((x >> 1) & 1) << 1; // pn[1] = x[1]
		    pixel_number |= ((x >> 2) & 1) << 2; // pn[2] = x[2]
		    pixel_number |= ((y >> 0) & 1) << 3; // pn[3] = y[0]
		    pixel_number |= ((y >> 1) & 1) << 4; // pn[4] = y[1]
		    pixel_number |= ((y >> 2) & 1) << 5; // pn[5] = y[2]
		    break;
	    case 4:
		    pixel_number |= ((x >> 0) & 1) << 0; // pn[0] = x[0]
		    pixel_number |= ((x >> 1) & 1) << 1; // pn[1] = x[1]
		    pixel_number |= ((y >> 0) & 1) << 2; // pn[2] = y[0]
		    pixel_number |= ((x >> 2) & 1) << 3; // pn[3] = x[2]
		    pixel_number |= ((y >> 1) & 1) << 4; // pn[4] = y[1]
		    pixel_number |= ((y >> 2) & 1) << 5; // pn[5] = y[2]
		    break;
	    }
	    sample_offset = sample_number * (tile_bytes / num_samples);
	    element_offset = sample_offset + (pixel_number * element_bytes);
    }
    offset = slice_offset + tile_offset + element_offset;
    return offset;
}

/* depth buffers */
static GLubyte *r600_ptr_depth(const struct radeon_renderbuffer * rrb,
			       GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    GLint offset = r600_1d_tile_helper(rrb, x, y, 1, 0);
    return &ptr[offset];
}

static GLubyte *r600_ptr_stencil(const struct radeon_renderbuffer * rrb,
				 GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    GLint offset = r600_1d_tile_helper(rrb, x, y, 1, 1);
    return &ptr[offset];
}

static GLubyte *r600_ptr_color(const struct radeon_renderbuffer * rrb,
			       GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    uint32_t mask = RADEON_BO_FLAGS_MACRO_TILE | RADEON_BO_FLAGS_MICRO_TILE;
    GLint offset;

    if (rrb->has_surface || !(rrb->bo->flags & mask)) {
        offset = x * rrb->cpp + y * rrb->pitch;
    } else {
	    offset = r600_1d_tile_helper(rrb, x, y, 0, 0);
    }
    return &ptr[offset];
}

#else

/* radeon tiling on r300-r500 has 4 states,
   macro-linear/micro-linear
   macro-linear/micro-tiled
   macro-tiled /micro-linear
   macro-tiled /micro-tiled
   1 byte surface 
   2 byte surface - two types - we only provide 8x2 microtiling
   4 byte surface
   8/16 byte (unused)
*/
static GLubyte *radeon_ptr_4byte(const struct radeon_renderbuffer * rrb,
			     GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    uint32_t mask = RADEON_BO_FLAGS_MACRO_TILE | RADEON_BO_FLAGS_MICRO_TILE;
    GLint offset;

    if (rrb->has_surface || !(rrb->bo->flags & mask)) {
        offset = x * rrb->cpp + y * rrb->pitch;
    } else {
        offset = 0;
        if (rrb->bo->flags & RADEON_BO_FLAGS_MACRO_TILE) {
	    if (rrb->bo->flags & RADEON_BO_FLAGS_MICRO_TILE) {
		offset = ((y >> 4) * (rrb->pitch >> 7) + (x >> 5)) << 11;
		offset += (((y >> 3) ^ (x >> 5)) & 0x1) << 10;
		offset += (((y >> 4) ^ (x >> 4)) & 0x1) << 9;
		offset += (((y >> 2) ^ (x >> 4)) & 0x1) << 8;
		offset += (((y >> 3) ^ (x >> 3)) & 0x1) << 7;
		offset += ((y >> 1) & 0x1) << 6;
		offset += ((x >> 2) & 0x1) << 5;
		offset += (y & 1) << 4;
		offset += (x & 3) << 2;
            } else {
		offset = ((y >> 3) * (rrb->pitch >> 8) + (x >> 6)) << 11;
		offset += (((y >> 2) ^ (x >> 6)) & 0x1) << 10;
		offset += (((y >> 3) ^ (x >> 5)) & 0x1) << 9;
		offset += (((y >> 1) ^ (x >> 5)) & 0x1) << 8;
		offset += (((y >> 2) ^ (x >> 4)) & 0x1) << 7;
		offset += (y & 1) << 6;
		offset += (x & 15) << 2;
            }
        } else {
	    offset = ((y >> 1) * (rrb->pitch >> 4) + (x >> 2)) << 5;
	    offset += (y & 1) << 4;
	    offset += (x & 3) << 2;
        }
    }
    return &ptr[offset];
}

static GLubyte *radeon_ptr_2byte_8x2(const struct radeon_renderbuffer * rrb,
				     GLint x, GLint y)
{
    GLubyte *ptr = rrb->bo->ptr;
    uint32_t mask = RADEON_BO_FLAGS_MACRO_TILE | RADEON_BO_FLAGS_MICRO_TILE;
    GLint offset;

    if (rrb->has_surface || !(rrb->bo->flags & mask)) {
        offset = x * rrb->cpp + y * rrb->pitch;
    } else {
        offset = 0;
        if (rrb->bo->flags & RADEON_BO_FLAGS_MACRO_TILE) {
            if (rrb->bo->flags & RADEON_BO_FLAGS_MICRO_TILE) {
		offset = ((y >> 4) * (rrb->pitch >> 7) + (x >> 6)) << 11;
		offset += (((y >> 3) ^ (x >> 6)) & 0x1) << 10;
		offset += (((y >> 4) ^ (x >> 5)) & 0x1) << 9;
		offset += (((y >> 2) ^ (x >> 5)) & 0x1) << 8;
		offset += (((y >> 3) ^ (x >> 4)) & 0x1) << 7;
		offset += ((y >> 1) & 0x1) << 6;
		offset += ((x >> 3) & 0x1) << 5;
		offset += (y & 1) << 4;
		offset += (x & 3) << 2;
            } else {
		offset = ((y >> 3) * (rrb->pitch >> 8) + (x >> 7)) << 11;
		offset += (((y >> 2) ^ (x >> 7)) & 0x1) << 10;
		offset += (((y >> 3) ^ (x >> 6)) & 0x1) << 9;
		offset += (((y >> 1) ^ (x >> 6)) & 0x1) << 8;
		offset += (((y >> 2) ^ (x >> 5)) & 0x1) << 7;
		offset += (y & 1) << 6;
		offset += ((x >> 4) & 0x1) << 5;
                offset += (x & 15) << 2;
            }
        } else {
	    offset = ((y >> 1) * (rrb->pitch >> 4) + (x >> 3)) << 5;
	    offset += (y & 0x1) << 4;
	    offset += (x & 0x7) << 1;
        }
    }
    return &ptr[offset];
}

#endif

#ifndef COMPILE_R300
#ifndef COMPILE_R600
static uint32_t
z24s8_to_s8z24(uint32_t val)
{
   return (val << 24) | (val >> 8);
}

static uint32_t
s8z24_to_z24s8(uint32_t val)
{
   return (val >> 24) | (val << 8);
}
#endif
#endif

/*
 * Note that all information needed to access pixels in a renderbuffer
 * should be obtained through the gl_renderbuffer parameter, not per-context
 * information.
 */
#define LOCAL_VARS						\
   struct radeon_context *radeon = RADEON_CONTEXT(ctx);			\
   struct radeon_renderbuffer *rrb = (void *) rb;		\
   const GLint yScale = ctx->DrawBuffer->Name ? 1 : -1;			\
   const GLint yBias = ctx->DrawBuffer->Name ? 0 : rrb->base.Height - 1;\
   unsigned int num_cliprects;						\
   struct drm_clip_rect *cliprects;					\
   int x_off, y_off;							\
   GLuint p;						\
   (void)p;						\
   radeon_get_cliprects(radeon, &cliprects, &num_cliprects, &x_off, &y_off);

#define LOCAL_DEPTH_VARS				\
   struct radeon_context *radeon = RADEON_CONTEXT(ctx);			\
   struct radeon_renderbuffer *rrb = (void *) rb;	\
   const GLint yScale = ctx->DrawBuffer->Name ? 1 : -1;			\
   const GLint yBias = ctx->DrawBuffer->Name ? 0 : rrb->base.Height - 1;\
   unsigned int num_cliprects;						\
   struct drm_clip_rect *cliprects;					\
   int x_off, y_off;							\
  radeon_get_cliprects(radeon, &cliprects, &num_cliprects, &x_off, &y_off);

#define LOCAL_STENCIL_VARS LOCAL_DEPTH_VARS

#define Y_FLIP(_y) ((_y) * yScale + yBias)

#define HW_LOCK()

#define HW_UNLOCK()

/* XXX FBO: this is identical to the macro in spantmp2.h except we get
 * the cliprect info from the context, not the driDrawable.
 * Move this into spantmp2.h someday.
 */
#define HW_CLIPLOOP()							\
   do {									\
      int _nc = num_cliprects;						\
      while ( _nc-- ) {							\
	 int minx = cliprects[_nc].x1 - x_off;				\
	 int miny = cliprects[_nc].y1 - y_off;				\
	 int maxx = cliprects[_nc].x2 - x_off;				\
	 int maxy = cliprects[_nc].y2 - y_off;

/* ================================================================
 * Color buffer
 */

/* 16 bit, RGB565 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_RGB
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_SHORT_5_6_5

#define TAG(x)    radeon##x##_RGB565
#define TAG2(x,y) radeon##x##_RGB565##y
#if defined(RADEON_COMMON_FOR_R600)
#define GET_PTR(X,Y) r600_ptr_color(rrb, (X) + x_off, (Y) + y_off)
#else
#define GET_PTR(X,Y) radeon_ptr_2byte_8x2(rrb, (X) + x_off, (Y) + y_off)
#endif
#include "spantmp2.h"

/* 16 bit, ARGB1555 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_BGRA
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_SHORT_1_5_5_5_REV

#define TAG(x)    radeon##x##_ARGB1555
#define TAG2(x,y) radeon##x##_ARGB1555##y
#if defined(RADEON_COMMON_FOR_R600)
#define GET_PTR(X,Y) r600_ptr_color(rrb, (X) + x_off, (Y) + y_off)
#else
#define GET_PTR(X,Y) radeon_ptr_2byte_8x2(rrb, (X) + x_off, (Y) + y_off)
#endif
#include "spantmp2.h"

/* 16 bit, RGBA4 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_BGRA
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_SHORT_4_4_4_4_REV

#define TAG(x)    radeon##x##_ARGB4444
#define TAG2(x,y) radeon##x##_ARGB4444##y
#if defined(RADEON_COMMON_FOR_R600)
#define GET_PTR(X,Y) r600_ptr_color(rrb, (X) + x_off, (Y) + y_off)
#else
#define GET_PTR(X,Y) radeon_ptr_2byte_8x2(rrb, (X) + x_off, (Y) + y_off)
#endif
#include "spantmp2.h"

/* 32 bit, xRGB8888 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_BGRA
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_INT_8_8_8_8_REV

#define TAG(x)    radeon##x##_xRGB8888
#define TAG2(x,y) radeon##x##_xRGB8888##y
#if defined(RADEON_COMMON_FOR_R600)
#define GET_VALUE(_x, _y) ((*(GLuint*)(r600_ptr_color(rrb, _x + x_off, _y + y_off)) | 0xff000000))
#define PUT_VALUE(_x, _y, d) { \
   GLuint *_ptr = (GLuint*)r600_ptr_color( rrb, _x + x_off, _y + y_off );		\
   *_ptr = d;								\
} while (0)
#else
#define GET_VALUE(_x, _y) ((*(GLuint*)(radeon_ptr_4byte(rrb, _x + x_off, _y + y_off)) | 0xff000000))
#define PUT_VALUE(_x, _y, d) { \
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );		\
   *_ptr = d;								\
} while (0)
#endif
#include "spantmp2.h"

/* 32 bit, ARGB8888 color spanline and pixel functions
 */
#define SPANTMP_PIXEL_FMT GL_BGRA
#define SPANTMP_PIXEL_TYPE GL_UNSIGNED_INT_8_8_8_8_REV

#define TAG(x)    radeon##x##_ARGB8888
#define TAG2(x,y) radeon##x##_ARGB8888##y
#if defined(RADEON_COMMON_FOR_R600)
#define GET_VALUE(_x, _y) (*(GLuint*)(r600_ptr_color(rrb, _x + x_off, _y + y_off)))
#define PUT_VALUE(_x, _y, d) { \
   GLuint *_ptr = (GLuint*)r600_ptr_color( rrb, _x + x_off, _y + y_off );		\
   *_ptr = d;								\
} while (0)
#else
#define GET_VALUE(_x, _y) (*(GLuint*)(radeon_ptr_4byte(rrb, _x + x_off, _y + y_off)))
#define PUT_VALUE(_x, _y, d) { \
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );		\
   *_ptr = d;								\
} while (0)
#endif
#include "spantmp2.h"

/* ================================================================
 * Depth buffer
 */

/* The Radeon family has depth tiling on all the time, so we have to convert
 * the x,y coordinates into the memory bus address (mba) in the same
 * manner as the engine.  In each case, the linear block address (ba)
 * is calculated, and then wired with x and y to produce the final
 * memory address.
 * The chip will do address translation on its own if the surface registers
 * are set up correctly. It is not quite enough to get it working with hyperz
 * too...
 */

/* 16-bit depth buffer functions
 */
#define VALUE_TYPE GLushort

#if defined(RADEON_COMMON_FOR_R200)
#define WRITE_DEPTH( _x, _y, d )					\
   *(GLushort *)r200_depth_2byte(rrb, _x + x_off, _y + y_off) = d
#elif defined(RADEON_COMMON_FOR_R600)
#define WRITE_DEPTH( _x, _y, d )					\
   *(GLushort *)r600_ptr_depth(rrb, _x + x_off, _y + y_off) = d
#else
#define WRITE_DEPTH( _x, _y, d )					\
   *(GLushort *)radeon_ptr_2byte_8x2(rrb, _x + x_off, _y + y_off) = d
#endif

#if defined(RADEON_COMMON_FOR_R200)
#define READ_DEPTH( d, _x, _y )						\
   d = *(GLushort *)r200_depth_2byte(rrb, _x + x_off, _y + y_off)
#elif defined(RADEON_COMMON_FOR_R600)
#define READ_DEPTH( d, _x, _y )						\
   d = *(GLushort *)r600_ptr_depth(rrb, _x + x_off, _y + y_off)
#else
#define READ_DEPTH( d, _x, _y )						\
   d = *(GLushort *)radeon_ptr_2byte_8x2(rrb, _x + x_off, _y + y_off)
#endif

#define TAG(x) radeon##x##_z16
#include "depthtmp.h"

/* 24 bit depth
 *
 * Careful: It looks like the R300 uses ZZZS byte order while the R200
 * uses SZZZ for 24 bit depth, 8 bit stencil mode.
 */
#define VALUE_TYPE GLuint

#if defined(COMPILE_R300)
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   tmp &= 0x000000ff;							\
   tmp |= ((d << 8) & 0xffffff00);					\
   *_ptr = tmp;					\
} while (0)
#elif defined(RADEON_COMMON_FOR_R600)
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)r600_ptr_depth( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   tmp &= 0xff000000;							\
   tmp |= ((d) & 0x00ffffff);					\
   *_ptr = tmp;					\
} while (0)
#elif defined(RADEON_COMMON_FOR_R200)
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)r200_depth_4byte( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   tmp &= 0xff000000;							\
   tmp |= ((d) & 0x00ffffff);						\
   *_ptr = tmp;					\
} while (0)
#else
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );	\
   GLuint tmp = *_ptr;							\
   tmp &= 0xff000000;							\
   tmp |= ((d) & 0x00ffffff);						\
   *_ptr = tmp;					\
} while (0)
#endif

#if defined(COMPILE_R300)
#define READ_DEPTH( d, _x, _y )						\
  do {									\
    d = (*(GLuint*)(radeon_ptr_4byte(rrb, _x + x_off, _y + y_off)) & 0xffffff00) >> 8; \
  }while(0)
#elif defined(RADEON_COMMON_FOR_R600)
#define READ_DEPTH( d, _x, _y )						\
  do {									\
    d = (*(GLuint*)(r600_ptr_depth(rrb, _x + x_off, _y + y_off)) & 0x00ffffff); \
  }while(0)
#elif defined(RADEON_COMMON_FOR_R200)
#define READ_DEPTH( d, _x, _y )						\
  do {									\
    d = *(GLuint*)(r200_depth_4byte(rrb, _x + x_off, _y + y_off)) & 0x00ffffff; \
  }while(0)
#else
#define READ_DEPTH( d, _x, _y )	\
  d = *(GLuint*)(radeon_ptr_4byte(rrb, _x + x_off,	_y + y_off)) & 0x00ffffff;
#endif

#define TAG(x) radeon##x##_z24
#include "depthtmp.h"

/* 24 bit depth, 8 bit stencil depthbuffer functions
 * EXT_depth_stencil
 *
 * Careful: It looks like the R300 uses ZZZS byte order while the R200
 * uses SZZZ for 24 bit depth, 8 bit stencil mode.
 */
#define VALUE_TYPE GLuint

#if defined(COMPILE_R300)
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );		\
   *_ptr = d;								\
} while (0)
#elif defined(RADEON_COMMON_FOR_R600)
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)r600_ptr_depth( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   tmp &= 0xff000000;							\
   tmp |= (((d) >> 8) & 0x00ffffff);					\
   *_ptr = tmp;					\
   _ptr = (GLuint*)r600_ptr_stencil(rrb, _x + x_off, _y + y_off);		\
   tmp = *_ptr;				\
   tmp &= 0xffffff00;							\
   tmp |= (d) & 0xff;							\
   *_ptr = tmp;					\
} while (0)
#elif defined(RADEON_COMMON_FOR_R200)
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)r200_depth_4byte( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = z24s8_to_s8z24(d);					\
   *_ptr = tmp;								\
} while (0)
#else
#define WRITE_DEPTH( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );	\
   GLuint tmp = z24s8_to_s8z24(d);					\
   *_ptr = tmp;					\
} while (0)
#endif

#if defined(COMPILE_R300)
#define READ_DEPTH( d, _x, _y )						\
  do { \
    d = (*(GLuint*)(radeon_ptr_4byte(rrb, _x + x_off, _y + y_off)));	\
  }while(0)
#elif defined(RADEON_COMMON_FOR_R600)
#define READ_DEPTH( d, _x, _y )						\
  do { \
    d = ((*(GLuint*)(r600_ptr_depth(rrb, _x + x_off, _y + y_off))) << 8) & 0xffffff00; \
    d |= (*(GLuint*)(r600_ptr_stencil(rrb, _x + x_off, _y + y_off))) & 0x000000ff;	\
  }while(0)
#elif defined(RADEON_COMMON_FOR_R200)
#define READ_DEPTH( d, _x, _y )						\
  do { \
    d = s8z24_to_z24s8(*(GLuint*)(r200_depth_4byte(rrb, _x + x_off, _y + y_off)));	\
  }while(0)
#else
#define READ_DEPTH( d, _x, _y )	do {					\
    d = s8z24_to_z24s8(*(GLuint*)(radeon_ptr_4byte(rrb, _x + x_off,	_y + y_off ))); \
  } while (0)
#endif

#define TAG(x) radeon##x##_z24_s8
#include "depthtmp.h"

/* ================================================================
 * Stencil buffer
 */

/* 24 bit depth, 8 bit stencil depthbuffer functions
 */
#ifdef COMPILE_R300
#define WRITE_STENCIL( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte(rrb, _x + x_off, _y + y_off);		\
   GLuint tmp = *_ptr;				\
   tmp &= 0xffffff00;							\
   tmp |= (d) & 0xff;							\
   *_ptr = tmp;					\
} while (0)
#elif defined(RADEON_COMMON_FOR_R600)
#define WRITE_STENCIL( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)r600_ptr_stencil(rrb, _x + x_off, _y + y_off);		\
   GLuint tmp = *_ptr;				\
   tmp &= 0xffffff00;							\
   tmp |= (d) & 0xff;							\
   *_ptr = tmp;					\
} while (0)
#elif defined(RADEON_COMMON_FOR_R200)
#define WRITE_STENCIL( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)r200_depth_4byte(rrb, _x + x_off, _y + y_off);		\
   GLuint tmp = *_ptr;				\
   tmp &= 0x00ffffff;							\
   tmp |= (((d) & 0xff) << 24);						\
   *_ptr = tmp;					\
} while (0)
#else
#define WRITE_STENCIL( _x, _y, d )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte(rrb, _x + x_off, _y + y_off);		\
   GLuint tmp = *_ptr;				\
   tmp &= 0x00ffffff;							\
   tmp |= (((d) & 0xff) << 24);						\
   *_ptr = tmp;					\
} while (0)
#endif

#ifdef COMPILE_R300
#define READ_STENCIL( d, _x, _y )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   d = tmp & 0x000000ff;						\
} while (0)
#elif defined(RADEON_COMMON_FOR_R600)
#define READ_STENCIL( d, _x, _y )					\
do {									\
   GLuint *_ptr = (GLuint*)r600_ptr_stencil( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   d = tmp & 0x000000ff;						\
} while (0)
#elif defined(RADEON_COMMON_FOR_R200)
#define READ_STENCIL( d, _x, _y )					\
do {									\
   GLuint *_ptr = (GLuint*)r200_depth_4byte( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   d = (tmp & 0xff000000) >> 24;					\
} while (0)
#else
#define READ_STENCIL( d, _x, _y )					\
do {									\
   GLuint *_ptr = (GLuint*)radeon_ptr_4byte( rrb, _x + x_off, _y + y_off );		\
   GLuint tmp = *_ptr;				\
   d = (tmp & 0xff000000) >> 24;					\
} while (0)
#endif

#define TAG(x) radeon##x##_z24_s8
#include "stenciltmp.h"


static void map_unmap_rb(struct gl_renderbuffer *rb, int flag)
{
	struct radeon_renderbuffer *rrb = radeon_renderbuffer(rb);
	int r;

	if (rrb == NULL || !rrb->bo)
		return;

	if (flag) {
		if (rrb->bo->bom->funcs->bo_wait)
			radeon_bo_wait(rrb->bo);
		r = radeon_bo_map(rrb->bo, 1);
		if (r) {
			fprintf(stderr, "(%s) error(%d) mapping buffer.\n",
				__FUNCTION__, r);
		}

		radeonSetSpanFunctions(rrb);
	} else {
		radeon_bo_unmap(rrb->bo);
		rb->GetRow = NULL;
		rb->PutRow = NULL;
	}
}

static void
radeon_map_unmap_buffers(GLcontext *ctx, GLboolean map)
{
	GLuint i, j;

	/* color draw buffers */
	for (j = 0; j < ctx->DrawBuffer->_NumColorDrawBuffers; j++)
		map_unmap_rb(ctx->DrawBuffer->_ColorDrawBuffers[j], map);

	/* check for render to textures */
	for (i = 0; i < BUFFER_COUNT; i++) {
		struct gl_renderbuffer_attachment *att =
			ctx->DrawBuffer->Attachment + i;
		struct gl_texture_object *tex = att->Texture;
		if (tex) {
			/* Render to texture. Note that a mipmapped texture need not
			 * be complete for render to texture, so we must restrict to
			 * mapping only the attached image.
			 */
			radeon_texture_image *image = get_radeon_texture_image(tex->Image[att->CubeMapFace][att->TextureLevel]);
			ASSERT(att->Renderbuffer);

			if (map)
				radeon_teximage_map(image, GL_TRUE);
			else
				radeon_teximage_unmap(image);
		}
	}

	map_unmap_rb(ctx->ReadBuffer->_ColorReadBuffer, map);

	/* depth buffer (Note wrapper!) */
	if (ctx->DrawBuffer->_DepthBuffer)
		map_unmap_rb(ctx->DrawBuffer->_DepthBuffer->Wrapped, map);

	if (ctx->DrawBuffer->_StencilBuffer)
		map_unmap_rb(ctx->DrawBuffer->_StencilBuffer->Wrapped, map);
}

static void radeonSpanRenderStart(GLcontext * ctx)
{
	radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
	int i;

	radeon_firevertices(rmesa);

	/* The locking and wait for idle should really only be needed in classic mode.
	 * In a future memory manager based implementation, this should become
	 * unnecessary due to the fact that mapping our buffers, textures, etc.
	 * should implicitly wait for any previous rendering commands that must
	 * be waited on. */
	if (!rmesa->radeonScreen->driScreen->dri2.enabled) {
		LOCK_HARDWARE(rmesa);
		radeonWaitForIdleLocked(rmesa);
	}

	for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
		if (ctx->Texture.Unit[i]._ReallyEnabled)
			ctx->Driver.MapTexture(ctx, ctx->Texture.Unit[i]._Current);
	}

	radeon_map_unmap_buffers(ctx, 1);
}

static void radeonSpanRenderFinish(GLcontext * ctx)
{
	radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
	int i;
	_swrast_flush(ctx);
	if (!rmesa->radeonScreen->driScreen->dri2.enabled) {
		UNLOCK_HARDWARE(rmesa);
	}
	for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
		if (ctx->Texture.Unit[i]._ReallyEnabled)
			ctx->Driver.UnmapTexture(ctx, ctx->Texture.Unit[i]._Current);
	}

	radeon_map_unmap_buffers(ctx, 0);
}

void radeonInitSpanFuncs(GLcontext * ctx)
{
	struct swrast_device_driver *swdd =
	    _swrast_GetDeviceDriverReference(ctx);
	swdd->SpanRenderStart = radeonSpanRenderStart;
	swdd->SpanRenderFinish = radeonSpanRenderFinish;
}

/**
 * Plug in the Get/Put routines for the given driRenderbuffer.
 */
static void radeonSetSpanFunctions(struct radeon_renderbuffer *rrb)
{
	if (rrb->base._ActualFormat == GL_RGB5) {
		radeonInitPointers_RGB565(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_RGB8) {
		radeonInitPointers_xRGB8888(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_RGBA8) {
		radeonInitPointers_ARGB8888(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_RGBA4) {
		radeonInitPointers_ARGB4444(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_RGB5_A1) {
		radeonInitPointers_ARGB1555(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_DEPTH_COMPONENT16) {
		radeonInitDepthPointers_z16(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_DEPTH_COMPONENT24) {
		radeonInitDepthPointers_z24(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_DEPTH24_STENCIL8_EXT) {
		radeonInitDepthPointers_z24_s8(&rrb->base);
	} else if (rrb->base._ActualFormat == GL_STENCIL_INDEX8_EXT) {
		radeonInitStencilPointers_z24_s8(&rrb->base);
	} else {
		fprintf(stderr, "radeonSetSpanFunctions: bad actual format: 0x%04X\n", rrb->base._ActualFormat);
	}
}