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
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
|
/*
* Copyright © 2011 Red Hat All Rights Reserved.
* Copyright © 2017 Advanced Micro Devices, Inc.
* 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, 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 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 THE COPYRIGHT HOLDERS, AUTHORS
* 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.
*
* The above copyright notice and this permission notice (including the
* next paragraph) shall be included in all copies or substantial portions
* of the Software.
*/
#include "ac_surface.h"
#include "amd_family.h"
#include "addrlib/amdgpu_asic_addr.h"
#include "ac_gpu_info.h"
#include "util/macros.h"
#include "util/u_atomic.h"
#include "util/u_math.h"
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <amdgpu.h>
#include <amdgpu_drm.h>
#include "addrlib/addrinterface.h"
#ifndef CIASICIDGFXENGINE_SOUTHERNISLAND
#define CIASICIDGFXENGINE_SOUTHERNISLAND 0x0000000A
#endif
#ifndef CIASICIDGFXENGINE_ARCTICISLAND
#define CIASICIDGFXENGINE_ARCTICISLAND 0x0000000D
#endif
static unsigned get_first(unsigned x, unsigned y)
{
return x;
}
static void addrlib_family_rev_id(enum radeon_family family,
unsigned *addrlib_family,
unsigned *addrlib_revid)
{
switch (family) {
case CHIP_TAHITI:
*addrlib_family = FAMILY_SI;
*addrlib_revid = get_first(AMDGPU_TAHITI_RANGE);
break;
case CHIP_PITCAIRN:
*addrlib_family = FAMILY_SI;
*addrlib_revid = get_first(AMDGPU_PITCAIRN_RANGE);
break;
case CHIP_VERDE:
*addrlib_family = FAMILY_SI;
*addrlib_revid = get_first(AMDGPU_CAPEVERDE_RANGE);
break;
case CHIP_OLAND:
*addrlib_family = FAMILY_SI;
*addrlib_revid = get_first(AMDGPU_OLAND_RANGE);
break;
case CHIP_HAINAN:
*addrlib_family = FAMILY_SI;
*addrlib_revid = get_first(AMDGPU_HAINAN_RANGE);
break;
case CHIP_BONAIRE:
*addrlib_family = FAMILY_CI;
*addrlib_revid = get_first(AMDGPU_BONAIRE_RANGE);
break;
case CHIP_KAVERI:
*addrlib_family = FAMILY_KV;
*addrlib_revid = get_first(AMDGPU_SPECTRE_RANGE);
break;
case CHIP_KABINI:
*addrlib_family = FAMILY_KV;
*addrlib_revid = get_first(AMDGPU_KALINDI_RANGE);
break;
case CHIP_HAWAII:
*addrlib_family = FAMILY_CI;
*addrlib_revid = get_first(AMDGPU_HAWAII_RANGE);
break;
case CHIP_MULLINS:
*addrlib_family = FAMILY_KV;
*addrlib_revid = get_first(AMDGPU_GODAVARI_RANGE);
break;
case CHIP_TONGA:
*addrlib_family = FAMILY_VI;
*addrlib_revid = get_first(AMDGPU_TONGA_RANGE);
break;
case CHIP_ICELAND:
*addrlib_family = FAMILY_VI;
*addrlib_revid = get_first(AMDGPU_ICELAND_RANGE);
break;
case CHIP_CARRIZO:
*addrlib_family = FAMILY_CZ;
*addrlib_revid = get_first(AMDGPU_CARRIZO_RANGE);
break;
case CHIP_STONEY:
*addrlib_family = FAMILY_CZ;
*addrlib_revid = get_first(AMDGPU_STONEY_RANGE);
break;
case CHIP_FIJI:
*addrlib_family = FAMILY_VI;
*addrlib_revid = get_first(AMDGPU_FIJI_RANGE);
break;
case CHIP_POLARIS10:
*addrlib_family = FAMILY_VI;
*addrlib_revid = get_first(AMDGPU_POLARIS10_RANGE);
break;
case CHIP_POLARIS11:
*addrlib_family = FAMILY_VI;
*addrlib_revid = get_first(AMDGPU_POLARIS11_RANGE);
break;
case CHIP_POLARIS12:
*addrlib_family = FAMILY_VI;
*addrlib_revid = get_first(AMDGPU_POLARIS12_RANGE);
break;
case CHIP_VEGA10:
*addrlib_family = FAMILY_AI;
*addrlib_revid = get_first(AMDGPU_VEGA10_RANGE);
break;
case CHIP_RAVEN:
*addrlib_family = FAMILY_RV;
*addrlib_revid = get_first(AMDGPU_RAVEN_RANGE);
break;
default:
fprintf(stderr, "amdgpu: Unknown family.\n");
}
}
static void *ADDR_API allocSysMem(const ADDR_ALLOCSYSMEM_INPUT * pInput)
{
return malloc(pInput->sizeInBytes);
}
static ADDR_E_RETURNCODE ADDR_API freeSysMem(const ADDR_FREESYSMEM_INPUT * pInput)
{
free(pInput->pVirtAddr);
return ADDR_OK;
}
ADDR_HANDLE amdgpu_addr_create(const struct radeon_info *info,
const struct amdgpu_gpu_info *amdinfo,
uint64_t *max_alignment)
{
ADDR_CREATE_INPUT addrCreateInput = {0};
ADDR_CREATE_OUTPUT addrCreateOutput = {0};
ADDR_REGISTER_VALUE regValue = {0};
ADDR_CREATE_FLAGS createFlags = {{0}};
ADDR_GET_MAX_ALIGNMENTS_OUTPUT addrGetMaxAlignmentsOutput = {0};
ADDR_E_RETURNCODE addrRet;
addrCreateInput.size = sizeof(ADDR_CREATE_INPUT);
addrCreateOutput.size = sizeof(ADDR_CREATE_OUTPUT);
regValue.gbAddrConfig = amdinfo->gb_addr_cfg;
createFlags.value = 0;
addrlib_family_rev_id(info->family, &addrCreateInput.chipFamily, &addrCreateInput.chipRevision);
if (addrCreateInput.chipFamily == FAMILY_UNKNOWN)
return NULL;
if (addrCreateInput.chipFamily >= FAMILY_AI) {
addrCreateInput.chipEngine = CIASICIDGFXENGINE_ARCTICISLAND;
regValue.blockVarSizeLog2 = 0;
} else {
regValue.noOfBanks = amdinfo->mc_arb_ramcfg & 0x3;
regValue.noOfRanks = (amdinfo->mc_arb_ramcfg & 0x4) >> 2;
regValue.backendDisables = amdinfo->enabled_rb_pipes_mask;
regValue.pTileConfig = amdinfo->gb_tile_mode;
regValue.noOfEntries = ARRAY_SIZE(amdinfo->gb_tile_mode);
if (addrCreateInput.chipFamily == FAMILY_SI) {
regValue.pMacroTileConfig = NULL;
regValue.noOfMacroEntries = 0;
} else {
regValue.pMacroTileConfig = amdinfo->gb_macro_tile_mode;
regValue.noOfMacroEntries = ARRAY_SIZE(amdinfo->gb_macro_tile_mode);
}
createFlags.useTileIndex = 1;
createFlags.useHtileSliceAlign = 1;
addrCreateInput.chipEngine = CIASICIDGFXENGINE_SOUTHERNISLAND;
}
addrCreateInput.callbacks.allocSysMem = allocSysMem;
addrCreateInput.callbacks.freeSysMem = freeSysMem;
addrCreateInput.callbacks.debugPrint = 0;
addrCreateInput.createFlags = createFlags;
addrCreateInput.regValue = regValue;
addrRet = AddrCreate(&addrCreateInput, &addrCreateOutput);
if (addrRet != ADDR_OK)
return NULL;
if (max_alignment) {
addrRet = AddrGetMaxAlignments(addrCreateOutput.hLib, &addrGetMaxAlignmentsOutput);
if (addrRet == ADDR_OK){
*max_alignment = addrGetMaxAlignmentsOutput.baseAlign;
}
}
return addrCreateOutput.hLib;
}
static int surf_config_sanity(const struct ac_surf_config *config)
{
/* all dimension must be at least 1 ! */
if (!config->info.width || !config->info.height || !config->info.depth ||
!config->info.array_size || !config->info.levels)
return -EINVAL;
switch (config->info.samples) {
case 0:
case 1:
case 2:
case 4:
case 8:
break;
default:
return -EINVAL;
}
if (config->is_3d && config->info.array_size > 1)
return -EINVAL;
if (config->is_cube && config->info.depth > 1)
return -EINVAL;
return 0;
}
static int gfx6_compute_level(ADDR_HANDLE addrlib,
const struct ac_surf_config *config,
struct radeon_surf *surf, bool is_stencil,
unsigned level, bool compressed,
ADDR_COMPUTE_SURFACE_INFO_INPUT *AddrSurfInfoIn,
ADDR_COMPUTE_SURFACE_INFO_OUTPUT *AddrSurfInfoOut,
ADDR_COMPUTE_DCCINFO_INPUT *AddrDccIn,
ADDR_COMPUTE_DCCINFO_OUTPUT *AddrDccOut,
ADDR_COMPUTE_HTILE_INFO_INPUT *AddrHtileIn,
ADDR_COMPUTE_HTILE_INFO_OUTPUT *AddrHtileOut)
{
struct legacy_surf_level *surf_level;
ADDR_E_RETURNCODE ret;
AddrSurfInfoIn->mipLevel = level;
AddrSurfInfoIn->width = u_minify(config->info.width, level);
AddrSurfInfoIn->height = u_minify(config->info.height, level);
/* Make GFX6 linear surfaces compatible with GFX9 for hybrid graphics,
* because GFX9 needs linear alignment of 256 bytes.
*/
if (config->info.levels == 1 &&
AddrSurfInfoIn->tileMode == ADDR_TM_LINEAR_ALIGNED &&
AddrSurfInfoIn->bpp) {
unsigned alignment = 256 / (AddrSurfInfoIn->bpp / 8);
assert(util_is_power_of_two(AddrSurfInfoIn->bpp));
AddrSurfInfoIn->width = align(AddrSurfInfoIn->width, alignment);
}
if (config->is_3d)
AddrSurfInfoIn->numSlices = u_minify(config->info.depth, level);
else if (config->is_cube)
AddrSurfInfoIn->numSlices = 6;
else
AddrSurfInfoIn->numSlices = config->info.array_size;
if (level > 0) {
/* Set the base level pitch. This is needed for calculation
* of non-zero levels. */
if (is_stencil)
AddrSurfInfoIn->basePitch = surf->u.legacy.stencil_level[0].nblk_x;
else
AddrSurfInfoIn->basePitch = surf->u.legacy.level[0].nblk_x;
/* Convert blocks to pixels for compressed formats. */
if (compressed)
AddrSurfInfoIn->basePitch *= surf->blk_w;
}
ret = AddrComputeSurfaceInfo(addrlib,
AddrSurfInfoIn,
AddrSurfInfoOut);
if (ret != ADDR_OK) {
return ret;
}
surf_level = is_stencil ? &surf->u.legacy.stencil_level[level] : &surf->u.legacy.level[level];
surf_level->offset = align64(surf->surf_size, AddrSurfInfoOut->baseAlign);
surf_level->slice_size_dw = AddrSurfInfoOut->sliceSize / 4;
surf_level->nblk_x = AddrSurfInfoOut->pitch;
surf_level->nblk_y = AddrSurfInfoOut->height;
switch (AddrSurfInfoOut->tileMode) {
case ADDR_TM_LINEAR_ALIGNED:
surf_level->mode = RADEON_SURF_MODE_LINEAR_ALIGNED;
break;
case ADDR_TM_1D_TILED_THIN1:
surf_level->mode = RADEON_SURF_MODE_1D;
break;
case ADDR_TM_2D_TILED_THIN1:
surf_level->mode = RADEON_SURF_MODE_2D;
break;
default:
assert(0);
}
if (is_stencil)
surf->u.legacy.stencil_tiling_index[level] = AddrSurfInfoOut->tileIndex;
else
surf->u.legacy.tiling_index[level] = AddrSurfInfoOut->tileIndex;
surf->surf_size = surf_level->offset + AddrSurfInfoOut->surfSize;
/* Clear DCC fields at the beginning. */
surf_level->dcc_offset = 0;
/* The previous level's flag tells us if we can use DCC for this level. */
if (AddrSurfInfoIn->flags.dccCompatible &&
(level == 0 || AddrDccOut->subLvlCompressible)) {
AddrDccIn->colorSurfSize = AddrSurfInfoOut->surfSize;
AddrDccIn->tileMode = AddrSurfInfoOut->tileMode;
AddrDccIn->tileInfo = *AddrSurfInfoOut->pTileInfo;
AddrDccIn->tileIndex = AddrSurfInfoOut->tileIndex;
AddrDccIn->macroModeIndex = AddrSurfInfoOut->macroModeIndex;
ret = AddrComputeDccInfo(addrlib,
AddrDccIn,
AddrDccOut);
if (ret == ADDR_OK) {
surf_level->dcc_offset = surf->dcc_size;
surf_level->dcc_fast_clear_size = AddrDccOut->dccFastClearSize;
surf->num_dcc_levels = level + 1;
surf->dcc_size = surf_level->dcc_offset + AddrDccOut->dccRamSize;
surf->dcc_alignment = MAX2(surf->dcc_alignment, AddrDccOut->dccRamBaseAlign);
}
}
/* TC-compatible HTILE. */
if (!is_stencil &&
AddrSurfInfoIn->flags.depth &&
surf_level->mode == RADEON_SURF_MODE_2D &&
level == 0) {
AddrHtileIn->flags.tcCompatible = AddrSurfInfoIn->flags.tcCompatible;
AddrHtileIn->pitch = AddrSurfInfoOut->pitch;
AddrHtileIn->height = AddrSurfInfoOut->height;
AddrHtileIn->numSlices = AddrSurfInfoOut->depth;
AddrHtileIn->blockWidth = ADDR_HTILE_BLOCKSIZE_8;
AddrHtileIn->blockHeight = ADDR_HTILE_BLOCKSIZE_8;
AddrHtileIn->pTileInfo = AddrSurfInfoOut->pTileInfo;
AddrHtileIn->tileIndex = AddrSurfInfoOut->tileIndex;
AddrHtileIn->macroModeIndex = AddrSurfInfoOut->macroModeIndex;
ret = AddrComputeHtileInfo(addrlib,
AddrHtileIn,
AddrHtileOut);
if (ret == ADDR_OK) {
surf->htile_size = AddrHtileOut->htileBytes;
surf->htile_slice_size = AddrHtileOut->sliceSize;
surf->htile_alignment = AddrHtileOut->baseAlign;
}
}
return 0;
}
#define G_009910_MICRO_TILE_MODE(x) (((x) >> 0) & 0x03)
#define G_009910_MICRO_TILE_MODE_NEW(x) (((x) >> 22) & 0x07)
static void gfx6_set_micro_tile_mode(struct radeon_surf *surf,
const struct radeon_info *info)
{
uint32_t tile_mode = info->si_tile_mode_array[surf->u.legacy.tiling_index[0]];
if (info->chip_class >= CIK)
surf->micro_tile_mode = G_009910_MICRO_TILE_MODE_NEW(tile_mode);
else
surf->micro_tile_mode = G_009910_MICRO_TILE_MODE(tile_mode);
}
static unsigned cik_get_macro_tile_index(struct radeon_surf *surf)
{
unsigned index, tileb;
tileb = 8 * 8 * surf->bpe;
tileb = MIN2(surf->u.legacy.tile_split, tileb);
for (index = 0; tileb > 64; index++)
tileb >>= 1;
assert(index < 16);
return index;
}
/**
* This must be called after the first level is computed.
*
* Copy surface-global settings like pipe/bank config from level 0 surface
* computation, and compute tile swizzle.
*/
static int gfx6_surface_settings(ADDR_HANDLE addrlib,
const struct radeon_info *info,
const struct ac_surf_config *config,
ADDR_COMPUTE_SURFACE_INFO_OUTPUT* csio,
struct radeon_surf *surf)
{
surf->surf_alignment = csio->baseAlign;
surf->u.legacy.pipe_config = csio->pTileInfo->pipeConfig - 1;
gfx6_set_micro_tile_mode(surf, info);
/* For 2D modes only. */
if (csio->tileMode >= ADDR_TM_2D_TILED_THIN1) {
surf->u.legacy.bankw = csio->pTileInfo->bankWidth;
surf->u.legacy.bankh = csio->pTileInfo->bankHeight;
surf->u.legacy.mtilea = csio->pTileInfo->macroAspectRatio;
surf->u.legacy.tile_split = csio->pTileInfo->tileSplitBytes;
surf->u.legacy.num_banks = csio->pTileInfo->banks;
surf->u.legacy.macro_tile_index = csio->macroModeIndex;
} else {
surf->u.legacy.macro_tile_index = 0;
}
/* Compute tile swizzle. */
/* TODO: fix tile swizzle with mipmapping for SI */
if ((info->chip_class >= CIK || config->info.levels == 1) &&
config->info.surf_index &&
surf->u.legacy.level[0].mode == RADEON_SURF_MODE_2D &&
!(surf->flags & (RADEON_SURF_Z_OR_SBUFFER | RADEON_SURF_SHAREABLE)) &&
(config->info.samples > 1 || !(surf->flags & RADEON_SURF_SCANOUT))) {
ADDR_COMPUTE_BASE_SWIZZLE_INPUT AddrBaseSwizzleIn = {0};
ADDR_COMPUTE_BASE_SWIZZLE_OUTPUT AddrBaseSwizzleOut = {0};
AddrBaseSwizzleIn.size = sizeof(ADDR_COMPUTE_BASE_SWIZZLE_INPUT);
AddrBaseSwizzleOut.size = sizeof(ADDR_COMPUTE_BASE_SWIZZLE_OUTPUT);
AddrBaseSwizzleIn.surfIndex = p_atomic_inc_return(config->info.surf_index) - 1;
AddrBaseSwizzleIn.tileIndex = csio->tileIndex;
AddrBaseSwizzleIn.macroModeIndex = csio->macroModeIndex;
AddrBaseSwizzleIn.pTileInfo = csio->pTileInfo;
AddrBaseSwizzleIn.tileMode = csio->tileMode;
int r = AddrComputeBaseSwizzle(addrlib, &AddrBaseSwizzleIn,
&AddrBaseSwizzleOut);
if (r != ADDR_OK)
return r;
assert(AddrBaseSwizzleOut.tileSwizzle <=
u_bit_consecutive(0, sizeof(surf->tile_swizzle) * 8));
surf->tile_swizzle = AddrBaseSwizzleOut.tileSwizzle;
}
return 0;
}
/**
* Fill in the tiling information in \p surf based on the given surface config.
*
* The following fields of \p surf must be initialized by the caller:
* blk_w, blk_h, bpe, flags.
*/
static int gfx6_compute_surface(ADDR_HANDLE addrlib,
const struct radeon_info *info,
const struct ac_surf_config *config,
enum radeon_surf_mode mode,
struct radeon_surf *surf)
{
unsigned level;
bool compressed;
ADDR_COMPUTE_SURFACE_INFO_INPUT AddrSurfInfoIn = {0};
ADDR_COMPUTE_SURFACE_INFO_OUTPUT AddrSurfInfoOut = {0};
ADDR_COMPUTE_DCCINFO_INPUT AddrDccIn = {0};
ADDR_COMPUTE_DCCINFO_OUTPUT AddrDccOut = {0};
ADDR_COMPUTE_HTILE_INFO_INPUT AddrHtileIn = {0};
ADDR_COMPUTE_HTILE_INFO_OUTPUT AddrHtileOut = {0};
ADDR_TILEINFO AddrTileInfoIn = {0};
ADDR_TILEINFO AddrTileInfoOut = {0};
int r;
AddrSurfInfoIn.size = sizeof(ADDR_COMPUTE_SURFACE_INFO_INPUT);
AddrSurfInfoOut.size = sizeof(ADDR_COMPUTE_SURFACE_INFO_OUTPUT);
AddrDccIn.size = sizeof(ADDR_COMPUTE_DCCINFO_INPUT);
AddrDccOut.size = sizeof(ADDR_COMPUTE_DCCINFO_OUTPUT);
AddrHtileIn.size = sizeof(ADDR_COMPUTE_HTILE_INFO_INPUT);
AddrHtileOut.size = sizeof(ADDR_COMPUTE_HTILE_INFO_OUTPUT);
AddrSurfInfoOut.pTileInfo = &AddrTileInfoOut;
compressed = surf->blk_w == 4 && surf->blk_h == 4;
/* MSAA and FMASK require 2D tiling. */
if (config->info.samples > 1 ||
(surf->flags & RADEON_SURF_FMASK))
mode = RADEON_SURF_MODE_2D;
/* DB doesn't support linear layouts. */
if (surf->flags & (RADEON_SURF_Z_OR_SBUFFER) &&
mode < RADEON_SURF_MODE_1D)
mode = RADEON_SURF_MODE_1D;
/* Set the requested tiling mode. */
switch (mode) {
case RADEON_SURF_MODE_LINEAR_ALIGNED:
AddrSurfInfoIn.tileMode = ADDR_TM_LINEAR_ALIGNED;
break;
case RADEON_SURF_MODE_1D:
AddrSurfInfoIn.tileMode = ADDR_TM_1D_TILED_THIN1;
break;
case RADEON_SURF_MODE_2D:
AddrSurfInfoIn.tileMode = ADDR_TM_2D_TILED_THIN1;
break;
default:
assert(0);
}
/* The format must be set correctly for the allocation of compressed
* textures to work. In other cases, setting the bpp is sufficient.
*/
if (compressed) {
switch (surf->bpe) {
case 8:
AddrSurfInfoIn.format = ADDR_FMT_BC1;
break;
case 16:
AddrSurfInfoIn.format = ADDR_FMT_BC3;
break;
default:
assert(0);
}
}
else {
AddrDccIn.bpp = AddrSurfInfoIn.bpp = surf->bpe * 8;
}
AddrDccIn.numSamples = AddrSurfInfoIn.numSamples =
config->info.samples ? config->info.samples : 1;
AddrSurfInfoIn.tileIndex = -1;
/* Set the micro tile type. */
if (surf->flags & RADEON_SURF_SCANOUT)
AddrSurfInfoIn.tileType = ADDR_DISPLAYABLE;
else if (surf->flags & (RADEON_SURF_Z_OR_SBUFFER | RADEON_SURF_FMASK))
AddrSurfInfoIn.tileType = ADDR_DEPTH_SAMPLE_ORDER;
else
AddrSurfInfoIn.tileType = ADDR_NON_DISPLAYABLE;
AddrSurfInfoIn.flags.color = !(surf->flags & RADEON_SURF_Z_OR_SBUFFER);
AddrSurfInfoIn.flags.depth = (surf->flags & RADEON_SURF_ZBUFFER) != 0;
AddrSurfInfoIn.flags.cube = config->is_cube;
AddrSurfInfoIn.flags.fmask = (surf->flags & RADEON_SURF_FMASK) != 0;
AddrSurfInfoIn.flags.display = (surf->flags & RADEON_SURF_SCANOUT) != 0;
AddrSurfInfoIn.flags.pow2Pad = config->info.levels > 1;
AddrSurfInfoIn.flags.tcCompatible = (surf->flags & RADEON_SURF_TC_COMPATIBLE_HTILE) != 0;
/* Only degrade the tile mode for space if TC-compatible HTILE hasn't been
* requested, because TC-compatible HTILE requires 2D tiling.
*/
AddrSurfInfoIn.flags.opt4Space = !AddrSurfInfoIn.flags.tcCompatible &&
!AddrSurfInfoIn.flags.fmask &&
config->info.samples <= 1 &&
(surf->flags & RADEON_SURF_OPTIMIZE_FOR_SPACE);
/* DCC notes:
* - If we add MSAA support, keep in mind that CB can't decompress 8bpp
* with samples >= 4.
* - Mipmapped array textures have low performance (discovered by a closed
* driver team).
*/
AddrSurfInfoIn.flags.dccCompatible =
info->chip_class >= VI &&
!(surf->flags & RADEON_SURF_Z_OR_SBUFFER) &&
!(surf->flags & RADEON_SURF_DISABLE_DCC) &&
!compressed &&
((config->info.array_size == 1 && config->info.depth == 1) ||
config->info.levels == 1);
AddrSurfInfoIn.flags.noStencil = (surf->flags & RADEON_SURF_SBUFFER) == 0;
AddrSurfInfoIn.flags.compressZ = AddrSurfInfoIn.flags.depth;
/* On CI/VI, the DB uses the same pitch and tile mode (except tilesplit)
* for Z and stencil. This can cause a number of problems which we work
* around here:
*
* - a depth part that is incompatible with mipmapped texturing
* - at least on Stoney, entirely incompatible Z/S aspects (e.g.
* incorrect tiling applied to the stencil part, stencil buffer
* memory accesses that go out of bounds) even without mipmapping
*
* Some piglit tests that are prone to different types of related
* failures:
* ./bin/ext_framebuffer_multisample-upsample 2 stencil
* ./bin/framebuffer-blit-levels {draw,read} stencil
* ./bin/ext_framebuffer_multisample-unaligned-blit N {depth,stencil} {msaa,upsample,downsample}
* ./bin/fbo-depth-array fs-writes-{depth,stencil} / {depth,stencil}-{clear,layered-clear,draw}
* ./bin/depthstencil-render-miplevels 1024 d=s=z24_s8
*/
int stencil_tile_idx = -1;
if (AddrSurfInfoIn.flags.depth && !AddrSurfInfoIn.flags.noStencil &&
(config->info.levels > 1 || info->family == CHIP_STONEY)) {
/* Compute stencilTileIdx that is compatible with the (depth)
* tileIdx. This degrades the depth surface if necessary to
* ensure that a matching stencilTileIdx exists. */
AddrSurfInfoIn.flags.matchStencilTileCfg = 1;
/* Keep the depth mip-tail compatible with texturing. */
AddrSurfInfoIn.flags.noStencil = 1;
}
/* Set preferred macrotile parameters. This is usually required
* for shared resources. This is for 2D tiling only. */
if (AddrSurfInfoIn.tileMode >= ADDR_TM_2D_TILED_THIN1 &&
surf->u.legacy.bankw && surf->u.legacy.bankh &&
surf->u.legacy.mtilea && surf->u.legacy.tile_split) {
assert(!(surf->flags & RADEON_SURF_FMASK));
/* If any of these parameters are incorrect, the calculation
* will fail. */
AddrTileInfoIn.banks = surf->u.legacy.num_banks;
AddrTileInfoIn.bankWidth = surf->u.legacy.bankw;
AddrTileInfoIn.bankHeight = surf->u.legacy.bankh;
AddrTileInfoIn.macroAspectRatio = surf->u.legacy.mtilea;
AddrTileInfoIn.tileSplitBytes = surf->u.legacy.tile_split;
AddrTileInfoIn.pipeConfig = surf->u.legacy.pipe_config + 1; /* +1 compared to GB_TILE_MODE */
AddrSurfInfoIn.flags.opt4Space = 0;
AddrSurfInfoIn.pTileInfo = &AddrTileInfoIn;
/* If AddrSurfInfoIn.pTileInfo is set, Addrlib doesn't set
* the tile index, because we are expected to know it if
* we know the other parameters.
*
* This is something that can easily be fixed in Addrlib.
* For now, just figure it out here.
* Note that only 2D_TILE_THIN1 is handled here.
*/
assert(!(surf->flags & RADEON_SURF_Z_OR_SBUFFER));
assert(AddrSurfInfoIn.tileMode == ADDR_TM_2D_TILED_THIN1);
if (info->chip_class == SI) {
if (AddrSurfInfoIn.tileType == ADDR_DISPLAYABLE) {
if (surf->bpe == 2)
AddrSurfInfoIn.tileIndex = 11; /* 16bpp */
else
AddrSurfInfoIn.tileIndex = 12; /* 32bpp */
} else {
if (surf->bpe == 1)
AddrSurfInfoIn.tileIndex = 14; /* 8bpp */
else if (surf->bpe == 2)
AddrSurfInfoIn.tileIndex = 15; /* 16bpp */
else if (surf->bpe == 4)
AddrSurfInfoIn.tileIndex = 16; /* 32bpp */
else
AddrSurfInfoIn.tileIndex = 17; /* 64bpp (and 128bpp) */
}
} else {
/* CIK - VI */
if (AddrSurfInfoIn.tileType == ADDR_DISPLAYABLE)
AddrSurfInfoIn.tileIndex = 10; /* 2D displayable */
else
AddrSurfInfoIn.tileIndex = 14; /* 2D non-displayable */
/* Addrlib doesn't set this if tileIndex is forced like above. */
AddrSurfInfoOut.macroModeIndex = cik_get_macro_tile_index(surf);
}
}
surf->has_stencil = !!(surf->flags & RADEON_SURF_SBUFFER);
surf->num_dcc_levels = 0;
surf->surf_size = 0;
surf->dcc_size = 0;
surf->dcc_alignment = 1;
surf->htile_size = 0;
surf->htile_slice_size = 0;
surf->htile_alignment = 1;
const bool only_stencil = (surf->flags & RADEON_SURF_SBUFFER) &&
!(surf->flags & RADEON_SURF_ZBUFFER);
/* Calculate texture layout information. */
if (!only_stencil) {
for (level = 0; level < config->info.levels; level++) {
r = gfx6_compute_level(addrlib, config, surf, false, level, compressed,
&AddrSurfInfoIn, &AddrSurfInfoOut,
&AddrDccIn, &AddrDccOut, &AddrHtileIn, &AddrHtileOut);
if (r)
return r;
if (level > 0)
continue;
/* Check that we actually got a TC-compatible HTILE if
* we requested it (only for level 0, since we're not
* supporting HTILE on higher mip levels anyway). */
assert(AddrSurfInfoOut.tcCompatible ||
!AddrSurfInfoIn.flags.tcCompatible ||
AddrSurfInfoIn.flags.matchStencilTileCfg);
if (AddrSurfInfoIn.flags.matchStencilTileCfg) {
if (!AddrSurfInfoOut.tcCompatible) {
AddrSurfInfoIn.flags.tcCompatible = 0;
surf->flags &= ~RADEON_SURF_TC_COMPATIBLE_HTILE;
}
AddrSurfInfoIn.flags.matchStencilTileCfg = 0;
AddrSurfInfoIn.tileIndex = AddrSurfInfoOut.tileIndex;
stencil_tile_idx = AddrSurfInfoOut.stencilTileIdx;
assert(stencil_tile_idx >= 0);
}
r = gfx6_surface_settings(addrlib, info, config,
&AddrSurfInfoOut, surf);
if (r)
return r;
}
}
/* Calculate texture layout information for stencil. */
if (surf->flags & RADEON_SURF_SBUFFER) {
AddrSurfInfoIn.tileIndex = stencil_tile_idx;
AddrSurfInfoIn.bpp = 8;
AddrSurfInfoIn.flags.depth = 0;
AddrSurfInfoIn.flags.stencil = 1;
AddrSurfInfoIn.flags.tcCompatible = 0;
/* This will be ignored if AddrSurfInfoIn.pTileInfo is NULL. */
AddrTileInfoIn.tileSplitBytes = surf->u.legacy.stencil_tile_split;
for (level = 0; level < config->info.levels; level++) {
r = gfx6_compute_level(addrlib, config, surf, true, level, compressed,
&AddrSurfInfoIn, &AddrSurfInfoOut,
&AddrDccIn, &AddrDccOut,
NULL, NULL);
if (r)
return r;
/* DB uses the depth pitch for both stencil and depth. */
if (!only_stencil) {
if (surf->u.legacy.stencil_level[level].nblk_x !=
surf->u.legacy.level[level].nblk_x)
surf->u.legacy.stencil_adjusted = true;
} else {
surf->u.legacy.level[level].nblk_x =
surf->u.legacy.stencil_level[level].nblk_x;
}
if (level == 0) {
if (only_stencil) {
r = gfx6_surface_settings(addrlib, info, config,
&AddrSurfInfoOut, surf);
if (r)
return r;
}
/* For 2D modes only. */
if (AddrSurfInfoOut.tileMode >= ADDR_TM_2D_TILED_THIN1) {
surf->u.legacy.stencil_tile_split =
AddrSurfInfoOut.pTileInfo->tileSplitBytes;
}
}
}
}
/* Recalculate the whole DCC miptree size including disabled levels.
* This is what addrlib does, but calling addrlib would be a lot more
* complicated.
*/
if (surf->dcc_size && config->info.levels > 1) {
/* The smallest miplevels that are never compressed by DCC
* still read the DCC buffer via TC if the base level uses DCC,
* and for some reason the DCC buffer needs to be larger if
* the miptree uses non-zero tile_swizzle. Otherwise there are
* VM faults.
*
* "dcc_alignment * 4" was determined by trial and error.
*/
surf->dcc_size = align64(surf->surf_size >> 8,
surf->dcc_alignment * 4);
}
/* Make sure HTILE covers the whole miptree, because the shader reads
* TC-compatible HTILE even for levels where it's disabled by DB.
*/
if (surf->htile_size && config->info.levels > 1)
surf->htile_size *= 2;
surf->is_linear = surf->u.legacy.level[0].mode == RADEON_SURF_MODE_LINEAR_ALIGNED;
surf->is_displayable = surf->is_linear ||
surf->micro_tile_mode == RADEON_MICRO_MODE_DISPLAY ||
surf->micro_tile_mode == RADEON_MICRO_MODE_ROTATED;
return 0;
}
/* This is only called when expecting a tiled layout. */
static int
gfx9_get_preferred_swizzle_mode(ADDR_HANDLE addrlib,
ADDR2_COMPUTE_SURFACE_INFO_INPUT *in,
bool is_fmask, AddrSwizzleMode *swizzle_mode)
{
ADDR_E_RETURNCODE ret;
ADDR2_GET_PREFERRED_SURF_SETTING_INPUT sin = {0};
ADDR2_GET_PREFERRED_SURF_SETTING_OUTPUT sout = {0};
sin.size = sizeof(ADDR2_GET_PREFERRED_SURF_SETTING_INPUT);
sout.size = sizeof(ADDR2_GET_PREFERRED_SURF_SETTING_OUTPUT);
sin.flags = in->flags;
sin.resourceType = in->resourceType;
sin.format = in->format;
sin.resourceLoction = ADDR_RSRC_LOC_INVIS;
/* TODO: We could allow some of these: */
sin.forbiddenBlock.micro = 1; /* don't allow the 256B swizzle modes */
sin.forbiddenBlock.var = 1; /* don't allow the variable-sized swizzle modes */
sin.forbiddenBlock.linear = 1; /* don't allow linear swizzle modes */
sin.bpp = in->bpp;
sin.width = in->width;
sin.height = in->height;
sin.numSlices = in->numSlices;
sin.numMipLevels = in->numMipLevels;
sin.numSamples = in->numSamples;
sin.numFrags = in->numFrags;
if (is_fmask) {
sin.flags.color = 0;
sin.flags.fmask = 1;
}
ret = Addr2GetPreferredSurfaceSetting(addrlib, &sin, &sout);
if (ret != ADDR_OK)
return ret;
*swizzle_mode = sout.swizzleMode;
return 0;
}
static int gfx9_compute_miptree(ADDR_HANDLE addrlib,
struct radeon_surf *surf, bool compressed,
ADDR2_COMPUTE_SURFACE_INFO_INPUT *in)
{
ADDR2_MIP_INFO mip_info[RADEON_SURF_MAX_LEVELS] = {};
ADDR2_COMPUTE_SURFACE_INFO_OUTPUT out = {0};
ADDR_E_RETURNCODE ret;
out.size = sizeof(ADDR2_COMPUTE_SURFACE_INFO_OUTPUT);
out.pMipInfo = mip_info;
ret = Addr2ComputeSurfaceInfo(addrlib, in, &out);
if (ret != ADDR_OK)
return ret;
if (in->flags.stencil) {
surf->u.gfx9.stencil.swizzle_mode = in->swizzleMode;
surf->u.gfx9.stencil.epitch = out.epitchIsHeight ? out.mipChainHeight - 1 :
out.mipChainPitch - 1;
surf->surf_alignment = MAX2(surf->surf_alignment, out.baseAlign);
surf->u.gfx9.stencil_offset = align(surf->surf_size, out.baseAlign);
surf->surf_size = surf->u.gfx9.stencil_offset + out.surfSize;
return 0;
}
surf->u.gfx9.surf.swizzle_mode = in->swizzleMode;
surf->u.gfx9.surf.epitch = out.epitchIsHeight ? out.mipChainHeight - 1 :
out.mipChainPitch - 1;
/* CMASK fast clear uses these even if FMASK isn't allocated.
* FMASK only supports the Z swizzle modes, whose numbers are multiples of 4.
*/
surf->u.gfx9.fmask.swizzle_mode = surf->u.gfx9.surf.swizzle_mode & ~0x3;
surf->u.gfx9.fmask.epitch = surf->u.gfx9.surf.epitch;
surf->u.gfx9.surf_slice_size = out.sliceSize;
surf->u.gfx9.surf_pitch = out.pitch;
surf->u.gfx9.surf_height = out.height;
surf->surf_size = out.surfSize;
surf->surf_alignment = out.baseAlign;
if (in->swizzleMode == ADDR_SW_LINEAR) {
for (unsigned i = 0; i < in->numMipLevels; i++)
surf->u.gfx9.offset[i] = mip_info[i].offset;
}
if (in->flags.depth) {
assert(in->swizzleMode != ADDR_SW_LINEAR);
/* HTILE */
ADDR2_COMPUTE_HTILE_INFO_INPUT hin = {0};
ADDR2_COMPUTE_HTILE_INFO_OUTPUT hout = {0};
hin.size = sizeof(ADDR2_COMPUTE_HTILE_INFO_INPUT);
hout.size = sizeof(ADDR2_COMPUTE_HTILE_INFO_OUTPUT);
hin.hTileFlags.pipeAligned = 1;
hin.hTileFlags.rbAligned = 1;
hin.depthFlags = in->flags;
hin.swizzleMode = in->swizzleMode;
hin.unalignedWidth = in->width;
hin.unalignedHeight = in->height;
hin.numSlices = in->numSlices;
hin.numMipLevels = in->numMipLevels;
ret = Addr2ComputeHtileInfo(addrlib, &hin, &hout);
if (ret != ADDR_OK)
return ret;
surf->u.gfx9.htile.rb_aligned = hin.hTileFlags.rbAligned;
surf->u.gfx9.htile.pipe_aligned = hin.hTileFlags.pipeAligned;
surf->htile_size = hout.htileBytes;
surf->htile_slice_size = hout.sliceSize;
surf->htile_alignment = hout.baseAlign;
} else {
/* DCC */
if (!(surf->flags & RADEON_SURF_DISABLE_DCC) &&
!compressed &&
in->swizzleMode != ADDR_SW_LINEAR) {
ADDR2_COMPUTE_DCCINFO_INPUT din = {0};
ADDR2_COMPUTE_DCCINFO_OUTPUT dout = {0};
ADDR2_META_MIP_INFO meta_mip_info[RADEON_SURF_MAX_LEVELS] = {};
din.size = sizeof(ADDR2_COMPUTE_DCCINFO_INPUT);
dout.size = sizeof(ADDR2_COMPUTE_DCCINFO_OUTPUT);
dout.pMipInfo = meta_mip_info;
din.dccKeyFlags.pipeAligned = 1;
din.dccKeyFlags.rbAligned = 1;
din.colorFlags = in->flags;
din.resourceType = in->resourceType;
din.swizzleMode = in->swizzleMode;
din.bpp = in->bpp;
din.unalignedWidth = in->width;
din.unalignedHeight = in->height;
din.numSlices = in->numSlices;
din.numFrags = in->numFrags;
din.numMipLevels = in->numMipLevels;
din.dataSurfaceSize = out.surfSize;
ret = Addr2ComputeDccInfo(addrlib, &din, &dout);
if (ret != ADDR_OK)
return ret;
surf->u.gfx9.dcc.rb_aligned = din.dccKeyFlags.rbAligned;
surf->u.gfx9.dcc.pipe_aligned = din.dccKeyFlags.pipeAligned;
surf->u.gfx9.dcc_pitch_max = dout.pitch - 1;
surf->dcc_size = dout.dccRamSize;
surf->dcc_alignment = dout.dccRamBaseAlign;
surf->num_dcc_levels = in->numMipLevels;
/* Disable DCC for levels that are in the mip tail.
*
* There are two issues that this is intended to
* address:
*
* 1. Multiple mip levels may share a cache line. This
* can lead to corruption when switching between
* rendering to different mip levels because the
* RBs don't maintain coherency.
*
* 2. Texturing with metadata after rendering sometimes
* fails with corruption, probably for a similar
* reason.
*
* Working around these issues for all levels in the
* mip tail may be overly conservative, but it's what
* Vulkan does.
*
* Alternative solutions that also work but are worse:
* - Disable DCC entirely.
* - Flush TC L2 after rendering.
*/
for (unsigned i = 0; i < in->numMipLevels; i++) {
if (meta_mip_info[i].inMiptail) {
surf->num_dcc_levels = i;
break;
}
}
if (!surf->num_dcc_levels)
surf->dcc_size = 0;
}
/* FMASK */
if (in->numSamples > 1) {
ADDR2_COMPUTE_FMASK_INFO_INPUT fin = {0};
ADDR2_COMPUTE_FMASK_INFO_OUTPUT fout = {0};
fin.size = sizeof(ADDR2_COMPUTE_FMASK_INFO_INPUT);
fout.size = sizeof(ADDR2_COMPUTE_FMASK_INFO_OUTPUT);
ret = gfx9_get_preferred_swizzle_mode(addrlib, in, true, &fin.swizzleMode);
if (ret != ADDR_OK)
return ret;
fin.unalignedWidth = in->width;
fin.unalignedHeight = in->height;
fin.numSlices = in->numSlices;
fin.numSamples = in->numSamples;
fin.numFrags = in->numFrags;
ret = Addr2ComputeFmaskInfo(addrlib, &fin, &fout);
if (ret != ADDR_OK)
return ret;
surf->u.gfx9.fmask.swizzle_mode = fin.swizzleMode;
surf->u.gfx9.fmask.epitch = fout.pitch - 1;
surf->u.gfx9.fmask_size = fout.fmaskBytes;
surf->u.gfx9.fmask_alignment = fout.baseAlign;
}
/* CMASK */
if (in->swizzleMode != ADDR_SW_LINEAR) {
ADDR2_COMPUTE_CMASK_INFO_INPUT cin = {0};
ADDR2_COMPUTE_CMASK_INFO_OUTPUT cout = {0};
cin.size = sizeof(ADDR2_COMPUTE_CMASK_INFO_INPUT);
cout.size = sizeof(ADDR2_COMPUTE_CMASK_INFO_OUTPUT);
cin.cMaskFlags.pipeAligned = 1;
cin.cMaskFlags.rbAligned = 1;
cin.colorFlags = in->flags;
cin.resourceType = in->resourceType;
cin.unalignedWidth = in->width;
cin.unalignedHeight = in->height;
cin.numSlices = in->numSlices;
if (in->numSamples > 1)
cin.swizzleMode = surf->u.gfx9.fmask.swizzle_mode;
else
cin.swizzleMode = in->swizzleMode;
ret = Addr2ComputeCmaskInfo(addrlib, &cin, &cout);
if (ret != ADDR_OK)
return ret;
surf->u.gfx9.cmask.rb_aligned = cin.cMaskFlags.rbAligned;
surf->u.gfx9.cmask.pipe_aligned = cin.cMaskFlags.pipeAligned;
surf->u.gfx9.cmask_size = cout.cmaskBytes;
surf->u.gfx9.cmask_alignment = cout.baseAlign;
}
}
return 0;
}
static int gfx9_compute_surface(ADDR_HANDLE addrlib,
const struct ac_surf_config *config,
enum radeon_surf_mode mode,
struct radeon_surf *surf)
{
bool compressed;
ADDR2_COMPUTE_SURFACE_INFO_INPUT AddrSurfInfoIn = {0};
int r;
assert(!(surf->flags & RADEON_SURF_FMASK));
AddrSurfInfoIn.size = sizeof(ADDR2_COMPUTE_SURFACE_INFO_INPUT);
compressed = surf->blk_w == 4 && surf->blk_h == 4;
/* The format must be set correctly for the allocation of compressed
* textures to work. In other cases, setting the bpp is sufficient. */
if (compressed) {
switch (surf->bpe) {
case 8:
AddrSurfInfoIn.format = ADDR_FMT_BC1;
break;
case 16:
AddrSurfInfoIn.format = ADDR_FMT_BC3;
break;
default:
assert(0);
}
} else {
AddrSurfInfoIn.bpp = surf->bpe * 8;
}
AddrSurfInfoIn.flags.color = !(surf->flags & RADEON_SURF_Z_OR_SBUFFER);
AddrSurfInfoIn.flags.depth = (surf->flags & RADEON_SURF_ZBUFFER) != 0;
AddrSurfInfoIn.flags.display = (surf->flags & RADEON_SURF_SCANOUT) != 0;
/* flags.texture currently refers to TC-compatible HTILE */
AddrSurfInfoIn.flags.texture = AddrSurfInfoIn.flags.color ||
surf->flags & RADEON_SURF_TC_COMPATIBLE_HTILE;
AddrSurfInfoIn.flags.opt4space = 1;
AddrSurfInfoIn.numMipLevels = config->info.levels;
AddrSurfInfoIn.numSamples = config->info.samples ? config->info.samples : 1;
AddrSurfInfoIn.numFrags = AddrSurfInfoIn.numSamples;
/* GFX9 doesn't support 1D depth textures, so allocate all 1D textures
* as 2D to avoid having shader variants for 1D vs 2D, so all shaders
* must sample 1D textures as 2D. */
if (config->is_3d)
AddrSurfInfoIn.resourceType = ADDR_RSRC_TEX_3D;
else
AddrSurfInfoIn.resourceType = ADDR_RSRC_TEX_2D;
AddrSurfInfoIn.width = config->info.width;
AddrSurfInfoIn.height = config->info.height;
if (config->is_3d)
AddrSurfInfoIn.numSlices = config->info.depth;
else if (config->is_cube)
AddrSurfInfoIn.numSlices = 6;
else
AddrSurfInfoIn.numSlices = config->info.array_size;
switch (mode) {
case RADEON_SURF_MODE_LINEAR_ALIGNED:
assert(config->info.samples <= 1);
assert(!(surf->flags & RADEON_SURF_Z_OR_SBUFFER));
AddrSurfInfoIn.swizzleMode = ADDR_SW_LINEAR;
break;
case RADEON_SURF_MODE_1D:
case RADEON_SURF_MODE_2D:
if (surf->flags & RADEON_SURF_IMPORTED) {
AddrSurfInfoIn.swizzleMode = surf->u.gfx9.surf.swizzle_mode;
break;
}
r = gfx9_get_preferred_swizzle_mode(addrlib, &AddrSurfInfoIn, false,
&AddrSurfInfoIn.swizzleMode);
if (r)
return r;
break;
default:
assert(0);
}
surf->u.gfx9.resource_type = AddrSurfInfoIn.resourceType;
surf->has_stencil = !!(surf->flags & RADEON_SURF_SBUFFER);
surf->num_dcc_levels = 0;
surf->surf_size = 0;
surf->dcc_size = 0;
surf->htile_size = 0;
surf->htile_slice_size = 0;
surf->u.gfx9.surf_offset = 0;
surf->u.gfx9.stencil_offset = 0;
surf->u.gfx9.fmask_size = 0;
surf->u.gfx9.cmask_size = 0;
/* Calculate texture layout information. */
r = gfx9_compute_miptree(addrlib, surf, compressed, &AddrSurfInfoIn);
if (r)
return r;
/* Calculate texture layout information for stencil. */
if (surf->flags & RADEON_SURF_SBUFFER) {
AddrSurfInfoIn.flags.stencil = 1;
AddrSurfInfoIn.bpp = 8;
if (!AddrSurfInfoIn.flags.depth) {
r = gfx9_get_preferred_swizzle_mode(addrlib, &AddrSurfInfoIn, false,
&AddrSurfInfoIn.swizzleMode);
if (r)
return r;
} else
AddrSurfInfoIn.flags.depth = 0;
r = gfx9_compute_miptree(addrlib, surf, compressed, &AddrSurfInfoIn);
if (r)
return r;
}
surf->is_linear = surf->u.gfx9.surf.swizzle_mode == ADDR_SW_LINEAR;
/* Query whether the surface is displayable. */
bool displayable = false;
r = Addr2IsValidDisplaySwizzleMode(addrlib, surf->u.gfx9.surf.swizzle_mode,
surf->bpe * 8, &displayable);
if (r)
return r;
surf->is_displayable = displayable;
switch (surf->u.gfx9.surf.swizzle_mode) {
/* S = standard. */
case ADDR_SW_256B_S:
case ADDR_SW_4KB_S:
case ADDR_SW_64KB_S:
case ADDR_SW_VAR_S:
case ADDR_SW_64KB_S_T:
case ADDR_SW_4KB_S_X:
case ADDR_SW_64KB_S_X:
case ADDR_SW_VAR_S_X:
surf->micro_tile_mode = RADEON_MICRO_MODE_THIN;
break;
/* D = display. */
case ADDR_SW_LINEAR:
case ADDR_SW_256B_D:
case ADDR_SW_4KB_D:
case ADDR_SW_64KB_D:
case ADDR_SW_VAR_D:
case ADDR_SW_64KB_D_T:
case ADDR_SW_4KB_D_X:
case ADDR_SW_64KB_D_X:
case ADDR_SW_VAR_D_X:
surf->micro_tile_mode = RADEON_MICRO_MODE_DISPLAY;
break;
/* R = rotated. */
case ADDR_SW_256B_R:
case ADDR_SW_4KB_R:
case ADDR_SW_64KB_R:
case ADDR_SW_VAR_R:
case ADDR_SW_64KB_R_T:
case ADDR_SW_4KB_R_X:
case ADDR_SW_64KB_R_X:
case ADDR_SW_VAR_R_X:
surf->micro_tile_mode = RADEON_MICRO_MODE_ROTATED;
break;
/* Z = depth. */
case ADDR_SW_4KB_Z:
case ADDR_SW_64KB_Z:
case ADDR_SW_VAR_Z:
case ADDR_SW_64KB_Z_T:
case ADDR_SW_4KB_Z_X:
case ADDR_SW_64KB_Z_X:
case ADDR_SW_VAR_Z_X:
surf->micro_tile_mode = RADEON_MICRO_MODE_DEPTH;
break;
default:
assert(0);
}
return 0;
}
int ac_compute_surface(ADDR_HANDLE addrlib, const struct radeon_info *info,
const struct ac_surf_config *config,
enum radeon_surf_mode mode,
struct radeon_surf *surf)
{
int r;
r = surf_config_sanity(config);
if (r)
return r;
if (info->chip_class >= GFX9)
return gfx9_compute_surface(addrlib, config, mode, surf);
else
return gfx6_compute_surface(addrlib, info, config, mode, surf);
}
|