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
|
/*
* Copyright © 2015 Intel Corporation
*
* 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 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#include <assert.h>
#include <stdbool.h>
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
#include "anv_private.h"
#include "vk_format_info.h"
#include "vk_util.h"
/** \file anv_cmd_buffer.c
*
* This file contains all of the stuff for emitting commands into a command
* buffer. This includes implementations of most of the vkCmd*
* entrypoints. This file is concerned entirely with state emission and
* not with the command buffer data structure itself. As far as this file
* is concerned, most of anv_cmd_buffer is magic.
*/
/* TODO: These are taken from GLES. We should check the Vulkan spec */
const struct anv_dynamic_state default_dynamic_state = {
.viewport = {
.count = 0,
},
.scissor = {
.count = 0,
},
.line_width = 1.0f,
.depth_bias = {
.bias = 0.0f,
.clamp = 0.0f,
.slope = 0.0f,
},
.blend_constants = { 0.0f, 0.0f, 0.0f, 0.0f },
.depth_bounds = {
.min = 0.0f,
.max = 1.0f,
},
.stencil_compare_mask = {
.front = ~0u,
.back = ~0u,
},
.stencil_write_mask = {
.front = ~0u,
.back = ~0u,
},
.stencil_reference = {
.front = 0u,
.back = 0u,
},
.line_stipple = {
.factor = 0u,
.pattern = 0u,
},
};
/**
* Copy the dynamic state from src to dest based on the copy_mask.
*
* Avoid copying states that have not changed, except for VIEWPORT, SCISSOR and
* BLEND_CONSTANTS (always copy them if they are in the copy_mask).
*
* Returns a mask of the states which changed.
*/
anv_cmd_dirty_mask_t
anv_dynamic_state_copy(struct anv_dynamic_state *dest,
const struct anv_dynamic_state *src,
anv_cmd_dirty_mask_t copy_mask)
{
anv_cmd_dirty_mask_t changed = 0;
if (copy_mask & ANV_CMD_DIRTY_DYNAMIC_VIEWPORT) {
dest->viewport.count = src->viewport.count;
typed_memcpy(dest->viewport.viewports, src->viewport.viewports,
src->viewport.count);
changed |= ANV_CMD_DIRTY_DYNAMIC_VIEWPORT;
}
if (copy_mask & ANV_CMD_DIRTY_DYNAMIC_SCISSOR) {
dest->scissor.count = src->scissor.count;
typed_memcpy(dest->scissor.scissors, src->scissor.scissors,
src->scissor.count);
changed |= ANV_CMD_DIRTY_DYNAMIC_SCISSOR;
}
if (copy_mask & ANV_CMD_DIRTY_DYNAMIC_BLEND_CONSTANTS) {
typed_memcpy(dest->blend_constants, src->blend_constants, 4);
changed |= ANV_CMD_DIRTY_DYNAMIC_BLEND_CONSTANTS;
}
#define ANV_CMP_COPY(field, flag) \
if (copy_mask & flag) { \
if (dest->field != src->field) { \
dest->field = src->field; \
changed |= flag; \
} \
}
ANV_CMP_COPY(line_width, ANV_CMD_DIRTY_DYNAMIC_LINE_WIDTH);
ANV_CMP_COPY(depth_bias.bias, ANV_CMD_DIRTY_DYNAMIC_DEPTH_BIAS);
ANV_CMP_COPY(depth_bias.clamp, ANV_CMD_DIRTY_DYNAMIC_DEPTH_BIAS);
ANV_CMP_COPY(depth_bias.slope, ANV_CMD_DIRTY_DYNAMIC_DEPTH_BIAS);
ANV_CMP_COPY(depth_bounds.min, ANV_CMD_DIRTY_DYNAMIC_DEPTH_BOUNDS);
ANV_CMP_COPY(depth_bounds.max, ANV_CMD_DIRTY_DYNAMIC_DEPTH_BOUNDS);
ANV_CMP_COPY(stencil_compare_mask.front, ANV_CMD_DIRTY_DYNAMIC_STENCIL_COMPARE_MASK);
ANV_CMP_COPY(stencil_compare_mask.back, ANV_CMD_DIRTY_DYNAMIC_STENCIL_COMPARE_MASK);
ANV_CMP_COPY(stencil_write_mask.front, ANV_CMD_DIRTY_DYNAMIC_STENCIL_WRITE_MASK);
ANV_CMP_COPY(stencil_write_mask.back, ANV_CMD_DIRTY_DYNAMIC_STENCIL_WRITE_MASK);
ANV_CMP_COPY(stencil_reference.front, ANV_CMD_DIRTY_DYNAMIC_STENCIL_REFERENCE);
ANV_CMP_COPY(stencil_reference.back, ANV_CMD_DIRTY_DYNAMIC_STENCIL_REFERENCE);
ANV_CMP_COPY(line_stipple.factor, ANV_CMD_DIRTY_DYNAMIC_LINE_STIPPLE);
ANV_CMP_COPY(line_stipple.pattern, ANV_CMD_DIRTY_DYNAMIC_LINE_STIPPLE);
#undef ANV_CMP_COPY
return changed;
}
static void
anv_cmd_state_init(struct anv_cmd_buffer *cmd_buffer)
{
struct anv_cmd_state *state = &cmd_buffer->state;
memset(state, 0, sizeof(*state));
state->current_pipeline = UINT32_MAX;
state->restart_index = UINT32_MAX;
state->gfx.dynamic = default_dynamic_state;
}
static void
anv_cmd_pipeline_state_finish(struct anv_cmd_buffer *cmd_buffer,
struct anv_cmd_pipeline_state *pipe_state)
{
for (uint32_t i = 0; i < ARRAY_SIZE(pipe_state->push_descriptors); i++) {
if (pipe_state->push_descriptors[i]) {
anv_descriptor_set_layout_unref(cmd_buffer->device,
pipe_state->push_descriptors[i]->set.layout);
vk_free(&cmd_buffer->pool->alloc, pipe_state->push_descriptors[i]);
}
}
}
static void
anv_cmd_state_finish(struct anv_cmd_buffer *cmd_buffer)
{
struct anv_cmd_state *state = &cmd_buffer->state;
anv_cmd_pipeline_state_finish(cmd_buffer, &state->gfx.base);
anv_cmd_pipeline_state_finish(cmd_buffer, &state->compute.base);
vk_free(&cmd_buffer->pool->alloc, state->attachments);
}
static void
anv_cmd_state_reset(struct anv_cmd_buffer *cmd_buffer)
{
anv_cmd_state_finish(cmd_buffer);
anv_cmd_state_init(cmd_buffer);
}
static VkResult anv_create_cmd_buffer(
struct anv_device * device,
struct anv_cmd_pool * pool,
VkCommandBufferLevel level,
VkCommandBuffer* pCommandBuffer)
{
struct anv_cmd_buffer *cmd_buffer;
VkResult result;
cmd_buffer = vk_alloc(&pool->alloc, sizeof(*cmd_buffer), 8,
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (cmd_buffer == NULL)
return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
cmd_buffer->batch.status = VK_SUCCESS;
cmd_buffer->_loader_data.loaderMagic = ICD_LOADER_MAGIC;
cmd_buffer->device = device;
cmd_buffer->pool = pool;
cmd_buffer->level = level;
result = anv_cmd_buffer_init_batch_bo_chain(cmd_buffer);
if (result != VK_SUCCESS)
goto fail;
anv_state_stream_init(&cmd_buffer->surface_state_stream,
&device->surface_state_pool, 4096);
anv_state_stream_init(&cmd_buffer->dynamic_state_stream,
&device->dynamic_state_pool, 16384);
anv_cmd_state_init(cmd_buffer);
if (pool) {
list_addtail(&cmd_buffer->pool_link, &pool->cmd_buffers);
} else {
/* Init the pool_link so we can safefly call list_del when we destroy
* the command buffer
*/
list_inithead(&cmd_buffer->pool_link);
}
*pCommandBuffer = anv_cmd_buffer_to_handle(cmd_buffer);
return VK_SUCCESS;
fail:
vk_free(&cmd_buffer->pool->alloc, cmd_buffer);
return result;
}
VkResult anv_AllocateCommandBuffers(
VkDevice _device,
const VkCommandBufferAllocateInfo* pAllocateInfo,
VkCommandBuffer* pCommandBuffers)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_cmd_pool, pool, pAllocateInfo->commandPool);
VkResult result = VK_SUCCESS;
uint32_t i;
for (i = 0; i < pAllocateInfo->commandBufferCount; i++) {
result = anv_create_cmd_buffer(device, pool, pAllocateInfo->level,
&pCommandBuffers[i]);
if (result != VK_SUCCESS)
break;
}
if (result != VK_SUCCESS) {
anv_FreeCommandBuffers(_device, pAllocateInfo->commandPool,
i, pCommandBuffers);
for (i = 0; i < pAllocateInfo->commandBufferCount; i++)
pCommandBuffers[i] = VK_NULL_HANDLE;
}
return result;
}
static void
anv_cmd_buffer_destroy(struct anv_cmd_buffer *cmd_buffer)
{
list_del(&cmd_buffer->pool_link);
anv_cmd_buffer_fini_batch_bo_chain(cmd_buffer);
anv_state_stream_finish(&cmd_buffer->surface_state_stream);
anv_state_stream_finish(&cmd_buffer->dynamic_state_stream);
anv_cmd_state_finish(cmd_buffer);
vk_free(&cmd_buffer->pool->alloc, cmd_buffer);
}
void anv_FreeCommandBuffers(
VkDevice device,
VkCommandPool commandPool,
uint32_t commandBufferCount,
const VkCommandBuffer* pCommandBuffers)
{
for (uint32_t i = 0; i < commandBufferCount; i++) {
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, pCommandBuffers[i]);
if (!cmd_buffer)
continue;
anv_cmd_buffer_destroy(cmd_buffer);
}
}
VkResult
anv_cmd_buffer_reset(struct anv_cmd_buffer *cmd_buffer)
{
cmd_buffer->usage_flags = 0;
anv_cmd_buffer_reset_batch_bo_chain(cmd_buffer);
anv_cmd_state_reset(cmd_buffer);
anv_state_stream_finish(&cmd_buffer->surface_state_stream);
anv_state_stream_init(&cmd_buffer->surface_state_stream,
&cmd_buffer->device->surface_state_pool, 4096);
anv_state_stream_finish(&cmd_buffer->dynamic_state_stream);
anv_state_stream_init(&cmd_buffer->dynamic_state_stream,
&cmd_buffer->device->dynamic_state_pool, 16384);
return VK_SUCCESS;
}
VkResult anv_ResetCommandBuffer(
VkCommandBuffer commandBuffer,
VkCommandBufferResetFlags flags)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
return anv_cmd_buffer_reset(cmd_buffer);
}
#define anv_genX_call(devinfo, func, ...) \
switch ((devinfo)->gen) { \
case 7: \
if ((devinfo)->is_haswell) { \
gen75_##func(__VA_ARGS__); \
} else { \
gen7_##func(__VA_ARGS__); \
} \
break; \
case 8: \
gen8_##func(__VA_ARGS__); \
break; \
case 9: \
gen9_##func(__VA_ARGS__); \
break; \
case 10: \
gen10_##func(__VA_ARGS__); \
break; \
case 11: \
gen11_##func(__VA_ARGS__); \
break; \
default: \
assert(!"Unknown hardware generation"); \
}
void
anv_cmd_buffer_emit_state_base_address(struct anv_cmd_buffer *cmd_buffer)
{
anv_genX_call(&cmd_buffer->device->info,
cmd_buffer_emit_state_base_address,
cmd_buffer);
}
void
anv_cmd_buffer_mark_image_written(struct anv_cmd_buffer *cmd_buffer,
const struct anv_image *image,
VkImageAspectFlagBits aspect,
enum isl_aux_usage aux_usage,
uint32_t level,
uint32_t base_layer,
uint32_t layer_count)
{
anv_genX_call(&cmd_buffer->device->info,
cmd_buffer_mark_image_written,
cmd_buffer, image, aspect, aux_usage,
level, base_layer, layer_count);
}
void
anv_cmd_emit_conditional_render_predicate(struct anv_cmd_buffer *cmd_buffer)
{
anv_genX_call(&cmd_buffer->device->info,
cmd_emit_conditional_render_predicate,
cmd_buffer);
}
static bool
mem_update(void *dst, const void *src, size_t size)
{
if (memcmp(dst, src, size) == 0)
return false;
memcpy(dst, src, size);
return true;
}
static void
set_dirty_for_bind_map(struct anv_cmd_buffer *cmd_buffer,
gl_shader_stage stage,
const struct anv_pipeline_bind_map *map)
{
if (mem_update(cmd_buffer->state.surface_sha1s[stage],
map->surface_sha1, sizeof(map->surface_sha1)))
cmd_buffer->state.descriptors_dirty |= mesa_to_vk_shader_stage(stage);
if (mem_update(cmd_buffer->state.sampler_sha1s[stage],
map->sampler_sha1, sizeof(map->sampler_sha1)))
cmd_buffer->state.descriptors_dirty |= mesa_to_vk_shader_stage(stage);
if (mem_update(cmd_buffer->state.push_sha1s[stage],
map->push_sha1, sizeof(map->push_sha1)))
cmd_buffer->state.push_constants_dirty |= mesa_to_vk_shader_stage(stage);
}
void anv_CmdBindPipeline(
VkCommandBuffer commandBuffer,
VkPipelineBindPoint pipelineBindPoint,
VkPipeline _pipeline)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
ANV_FROM_HANDLE(anv_pipeline, pipeline, _pipeline);
switch (pipelineBindPoint) {
case VK_PIPELINE_BIND_POINT_COMPUTE: {
if (cmd_buffer->state.compute.base.pipeline == pipeline)
return;
cmd_buffer->state.compute.base.pipeline = pipeline;
cmd_buffer->state.compute.pipeline_dirty = true;
const struct anv_pipeline_bind_map *bind_map =
&pipeline->shaders[MESA_SHADER_COMPUTE]->bind_map;
set_dirty_for_bind_map(cmd_buffer, MESA_SHADER_COMPUTE, bind_map);
break;
}
case VK_PIPELINE_BIND_POINT_GRAPHICS:
if (cmd_buffer->state.gfx.base.pipeline == pipeline)
return;
cmd_buffer->state.gfx.base.pipeline = pipeline;
cmd_buffer->state.gfx.vb_dirty |= pipeline->vb_used;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_PIPELINE;
anv_foreach_stage(stage, pipeline->active_stages) {
set_dirty_for_bind_map(cmd_buffer, stage,
&pipeline->shaders[stage]->bind_map);
}
/* Apply the dynamic state from the pipeline */
cmd_buffer->state.gfx.dirty |=
anv_dynamic_state_copy(&cmd_buffer->state.gfx.dynamic,
&pipeline->dynamic_state,
pipeline->dynamic_state_mask);
break;
default:
assert(!"invalid bind point");
break;
}
}
void anv_CmdSetViewport(
VkCommandBuffer commandBuffer,
uint32_t firstViewport,
uint32_t viewportCount,
const VkViewport* pViewports)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
const uint32_t total_count = firstViewport + viewportCount;
if (cmd_buffer->state.gfx.dynamic.viewport.count < total_count)
cmd_buffer->state.gfx.dynamic.viewport.count = total_count;
memcpy(cmd_buffer->state.gfx.dynamic.viewport.viewports + firstViewport,
pViewports, viewportCount * sizeof(*pViewports));
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_VIEWPORT;
}
void anv_CmdSetScissor(
VkCommandBuffer commandBuffer,
uint32_t firstScissor,
uint32_t scissorCount,
const VkRect2D* pScissors)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
const uint32_t total_count = firstScissor + scissorCount;
if (cmd_buffer->state.gfx.dynamic.scissor.count < total_count)
cmd_buffer->state.gfx.dynamic.scissor.count = total_count;
memcpy(cmd_buffer->state.gfx.dynamic.scissor.scissors + firstScissor,
pScissors, scissorCount * sizeof(*pScissors));
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_SCISSOR;
}
void anv_CmdSetLineWidth(
VkCommandBuffer commandBuffer,
float lineWidth)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
cmd_buffer->state.gfx.dynamic.line_width = lineWidth;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_LINE_WIDTH;
}
void anv_CmdSetDepthBias(
VkCommandBuffer commandBuffer,
float depthBiasConstantFactor,
float depthBiasClamp,
float depthBiasSlopeFactor)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
cmd_buffer->state.gfx.dynamic.depth_bias.bias = depthBiasConstantFactor;
cmd_buffer->state.gfx.dynamic.depth_bias.clamp = depthBiasClamp;
cmd_buffer->state.gfx.dynamic.depth_bias.slope = depthBiasSlopeFactor;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_DEPTH_BIAS;
}
void anv_CmdSetBlendConstants(
VkCommandBuffer commandBuffer,
const float blendConstants[4])
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
memcpy(cmd_buffer->state.gfx.dynamic.blend_constants,
blendConstants, sizeof(float) * 4);
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_BLEND_CONSTANTS;
}
void anv_CmdSetDepthBounds(
VkCommandBuffer commandBuffer,
float minDepthBounds,
float maxDepthBounds)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
cmd_buffer->state.gfx.dynamic.depth_bounds.min = minDepthBounds;
cmd_buffer->state.gfx.dynamic.depth_bounds.max = maxDepthBounds;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_DEPTH_BOUNDS;
}
void anv_CmdSetStencilCompareMask(
VkCommandBuffer commandBuffer,
VkStencilFaceFlags faceMask,
uint32_t compareMask)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
if (faceMask & VK_STENCIL_FACE_FRONT_BIT)
cmd_buffer->state.gfx.dynamic.stencil_compare_mask.front = compareMask;
if (faceMask & VK_STENCIL_FACE_BACK_BIT)
cmd_buffer->state.gfx.dynamic.stencil_compare_mask.back = compareMask;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_STENCIL_COMPARE_MASK;
}
void anv_CmdSetStencilWriteMask(
VkCommandBuffer commandBuffer,
VkStencilFaceFlags faceMask,
uint32_t writeMask)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
if (faceMask & VK_STENCIL_FACE_FRONT_BIT)
cmd_buffer->state.gfx.dynamic.stencil_write_mask.front = writeMask;
if (faceMask & VK_STENCIL_FACE_BACK_BIT)
cmd_buffer->state.gfx.dynamic.stencil_write_mask.back = writeMask;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_STENCIL_WRITE_MASK;
}
void anv_CmdSetStencilReference(
VkCommandBuffer commandBuffer,
VkStencilFaceFlags faceMask,
uint32_t reference)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
if (faceMask & VK_STENCIL_FACE_FRONT_BIT)
cmd_buffer->state.gfx.dynamic.stencil_reference.front = reference;
if (faceMask & VK_STENCIL_FACE_BACK_BIT)
cmd_buffer->state.gfx.dynamic.stencil_reference.back = reference;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_STENCIL_REFERENCE;
}
void anv_CmdSetLineStippleEXT(
VkCommandBuffer commandBuffer,
uint32_t lineStippleFactor,
uint16_t lineStipplePattern)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
cmd_buffer->state.gfx.dynamic.line_stipple.factor = lineStippleFactor;
cmd_buffer->state.gfx.dynamic.line_stipple.pattern = lineStipplePattern;
cmd_buffer->state.gfx.dirty |= ANV_CMD_DIRTY_DYNAMIC_LINE_STIPPLE;
}
static void
anv_cmd_buffer_bind_descriptor_set(struct anv_cmd_buffer *cmd_buffer,
VkPipelineBindPoint bind_point,
struct anv_pipeline_layout *layout,
uint32_t set_index,
struct anv_descriptor_set *set,
uint32_t *dynamic_offset_count,
const uint32_t **dynamic_offsets)
{
struct anv_descriptor_set_layout *set_layout =
layout->set[set_index].layout;
VkShaderStageFlags stages = set_layout->shader_stages &
(bind_point == VK_PIPELINE_BIND_POINT_COMPUTE ?
VK_SHADER_STAGE_COMPUTE_BIT : VK_SHADER_STAGE_ALL_GRAPHICS);
VkShaderStageFlags dirty_stages = 0;
if (bind_point == VK_PIPELINE_BIND_POINT_COMPUTE) {
if (cmd_buffer->state.compute.base.descriptors[set_index] != set) {
cmd_buffer->state.compute.base.descriptors[set_index] = set;
dirty_stages |= stages;
}
} else {
assert(bind_point == VK_PIPELINE_BIND_POINT_GRAPHICS);
if (cmd_buffer->state.gfx.base.descriptors[set_index] != set) {
cmd_buffer->state.gfx.base.descriptors[set_index] = set;
dirty_stages |= stages;
}
}
/* If it's a push descriptor set, we have to flag things as dirty
* regardless of whether or not the CPU-side data structure changed as we
* may have edited in-place.
*/
if (set->pool == NULL)
dirty_stages |= stages;
if (dynamic_offsets) {
if (set_layout->dynamic_offset_count > 0) {
uint32_t dynamic_offset_start =
layout->set[set_index].dynamic_offset_start;
anv_foreach_stage(stage, stages) {
struct anv_push_constants *push =
&cmd_buffer->state.push_constants[stage];
uint32_t *push_offsets =
&push->dynamic_offsets[dynamic_offset_start];
/* Assert that everything is in range */
assert(set_layout->dynamic_offset_count <= *dynamic_offset_count);
assert(dynamic_offset_start + set_layout->dynamic_offset_count <=
ARRAY_SIZE(push->dynamic_offsets));
unsigned mask = set_layout->stage_dynamic_offsets[stage];
STATIC_ASSERT(MAX_DYNAMIC_BUFFERS <= sizeof(mask) * 8);
while (mask) {
int i = u_bit_scan(&mask);
if (push_offsets[i] != (*dynamic_offsets)[i]) {
push_offsets[i] = (*dynamic_offsets)[i];
dirty_stages |= mesa_to_vk_shader_stage(stage);
}
}
}
*dynamic_offsets += set_layout->dynamic_offset_count;
*dynamic_offset_count -= set_layout->dynamic_offset_count;
}
}
cmd_buffer->state.descriptors_dirty |= dirty_stages;
cmd_buffer->state.push_constants_dirty |= dirty_stages;
}
void anv_CmdBindDescriptorSets(
VkCommandBuffer commandBuffer,
VkPipelineBindPoint pipelineBindPoint,
VkPipelineLayout _layout,
uint32_t firstSet,
uint32_t descriptorSetCount,
const VkDescriptorSet* pDescriptorSets,
uint32_t dynamicOffsetCount,
const uint32_t* pDynamicOffsets)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
ANV_FROM_HANDLE(anv_pipeline_layout, layout, _layout);
assert(firstSet + descriptorSetCount <= MAX_SETS);
for (uint32_t i = 0; i < descriptorSetCount; i++) {
ANV_FROM_HANDLE(anv_descriptor_set, set, pDescriptorSets[i]);
anv_cmd_buffer_bind_descriptor_set(cmd_buffer, pipelineBindPoint,
layout, firstSet + i, set,
&dynamicOffsetCount,
&pDynamicOffsets);
}
}
void anv_CmdBindVertexBuffers(
VkCommandBuffer commandBuffer,
uint32_t firstBinding,
uint32_t bindingCount,
const VkBuffer* pBuffers,
const VkDeviceSize* pOffsets)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
struct anv_vertex_binding *vb = cmd_buffer->state.vertex_bindings;
/* We have to defer setting up vertex buffer since we need the buffer
* stride from the pipeline. */
assert(firstBinding + bindingCount <= MAX_VBS);
for (uint32_t i = 0; i < bindingCount; i++) {
vb[firstBinding + i].buffer = anv_buffer_from_handle(pBuffers[i]);
vb[firstBinding + i].offset = pOffsets[i];
cmd_buffer->state.gfx.vb_dirty |= 1 << (firstBinding + i);
}
}
void anv_CmdBindTransformFeedbackBuffersEXT(
VkCommandBuffer commandBuffer,
uint32_t firstBinding,
uint32_t bindingCount,
const VkBuffer* pBuffers,
const VkDeviceSize* pOffsets,
const VkDeviceSize* pSizes)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
struct anv_xfb_binding *xfb = cmd_buffer->state.xfb_bindings;
/* We have to defer setting up vertex buffer since we need the buffer
* stride from the pipeline. */
assert(firstBinding + bindingCount <= MAX_XFB_BUFFERS);
for (uint32_t i = 0; i < bindingCount; i++) {
if (pBuffers[i] == VK_NULL_HANDLE) {
xfb[firstBinding + i].buffer = NULL;
} else {
ANV_FROM_HANDLE(anv_buffer, buffer, pBuffers[i]);
xfb[firstBinding + i].buffer = buffer;
xfb[firstBinding + i].offset = pOffsets[i];
xfb[firstBinding + i].size =
anv_buffer_get_range(buffer, pOffsets[i],
pSizes ? pSizes[i] : VK_WHOLE_SIZE);
}
}
}
enum isl_format
anv_isl_format_for_descriptor_type(VkDescriptorType type)
{
switch (type) {
case VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER:
case VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER_DYNAMIC:
return ISL_FORMAT_R32G32B32A32_FLOAT;
case VK_DESCRIPTOR_TYPE_STORAGE_BUFFER:
case VK_DESCRIPTOR_TYPE_STORAGE_BUFFER_DYNAMIC:
return ISL_FORMAT_RAW;
default:
unreachable("Invalid descriptor type");
}
}
struct anv_state
anv_cmd_buffer_emit_dynamic(struct anv_cmd_buffer *cmd_buffer,
const void *data, uint32_t size, uint32_t alignment)
{
struct anv_state state;
state = anv_cmd_buffer_alloc_dynamic_state(cmd_buffer, size, alignment);
memcpy(state.map, data, size);
VG(VALGRIND_CHECK_MEM_IS_DEFINED(state.map, size));
return state;
}
struct anv_state
anv_cmd_buffer_merge_dynamic(struct anv_cmd_buffer *cmd_buffer,
uint32_t *a, uint32_t *b,
uint32_t dwords, uint32_t alignment)
{
struct anv_state state;
uint32_t *p;
state = anv_cmd_buffer_alloc_dynamic_state(cmd_buffer,
dwords * 4, alignment);
p = state.map;
for (uint32_t i = 0; i < dwords; i++)
p[i] = a[i] | b[i];
VG(VALGRIND_CHECK_MEM_IS_DEFINED(p, dwords * 4));
return state;
}
struct anv_state
anv_cmd_buffer_push_constants(struct anv_cmd_buffer *cmd_buffer,
gl_shader_stage stage)
{
struct anv_push_constants *data =
&cmd_buffer->state.push_constants[stage];
struct anv_state state =
anv_cmd_buffer_alloc_dynamic_state(cmd_buffer,
sizeof(struct anv_push_constants),
32 /* bottom 5 bits MBZ */);
memcpy(state.map, data, sizeof(struct anv_push_constants));
return state;
}
struct anv_state
anv_cmd_buffer_cs_push_constants(struct anv_cmd_buffer *cmd_buffer)
{
struct anv_push_constants *data =
&cmd_buffer->state.push_constants[MESA_SHADER_COMPUTE];
struct anv_pipeline *pipeline = cmd_buffer->state.compute.base.pipeline;
const struct brw_cs_prog_data *cs_prog_data = get_cs_prog_data(pipeline);
const struct anv_push_range *range =
&pipeline->shaders[MESA_SHADER_COMPUTE]->bind_map.push_ranges[0];
if (cs_prog_data->push.total.size == 0)
return (struct anv_state) { .offset = 0 };
const unsigned push_constant_alignment =
cmd_buffer->device->info.gen < 8 ? 32 : 64;
const unsigned aligned_total_push_constants_size =
ALIGN(cs_prog_data->push.total.size, push_constant_alignment);
struct anv_state state =
anv_cmd_buffer_alloc_dynamic_state(cmd_buffer,
aligned_total_push_constants_size,
push_constant_alignment);
void *dst = state.map;
const void *src = (char *)data + (range->start * 32);
if (cs_prog_data->push.cross_thread.size > 0) {
memcpy(dst, src, cs_prog_data->push.cross_thread.size);
dst += cs_prog_data->push.cross_thread.size;
src += cs_prog_data->push.cross_thread.size;
}
if (cs_prog_data->push.per_thread.size > 0) {
for (unsigned t = 0; t < cs_prog_data->threads; t++) {
memcpy(dst, src, cs_prog_data->push.per_thread.size);
uint32_t *subgroup_id = dst +
offsetof(struct anv_push_constants, cs.subgroup_id) -
(range->start * 32 + cs_prog_data->push.cross_thread.size);
*subgroup_id = t;
dst += cs_prog_data->push.per_thread.size;
}
}
return state;
}
void anv_CmdPushConstants(
VkCommandBuffer commandBuffer,
VkPipelineLayout layout,
VkShaderStageFlags stageFlags,
uint32_t offset,
uint32_t size,
const void* pValues)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
anv_foreach_stage(stage, stageFlags) {
memcpy(cmd_buffer->state.push_constants[stage].client_data + offset,
pValues, size);
}
cmd_buffer->state.push_constants_dirty |= stageFlags;
}
VkResult anv_CreateCommandPool(
VkDevice _device,
const VkCommandPoolCreateInfo* pCreateInfo,
const VkAllocationCallbacks* pAllocator,
VkCommandPool* pCmdPool)
{
ANV_FROM_HANDLE(anv_device, device, _device);
struct anv_cmd_pool *pool;
pool = vk_alloc2(&device->alloc, pAllocator, sizeof(*pool), 8,
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (pool == NULL)
return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
if (pAllocator)
pool->alloc = *pAllocator;
else
pool->alloc = device->alloc;
list_inithead(&pool->cmd_buffers);
*pCmdPool = anv_cmd_pool_to_handle(pool);
return VK_SUCCESS;
}
void anv_DestroyCommandPool(
VkDevice _device,
VkCommandPool commandPool,
const VkAllocationCallbacks* pAllocator)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_cmd_pool, pool, commandPool);
if (!pool)
return;
list_for_each_entry_safe(struct anv_cmd_buffer, cmd_buffer,
&pool->cmd_buffers, pool_link) {
anv_cmd_buffer_destroy(cmd_buffer);
}
vk_free2(&device->alloc, pAllocator, pool);
}
VkResult anv_ResetCommandPool(
VkDevice device,
VkCommandPool commandPool,
VkCommandPoolResetFlags flags)
{
ANV_FROM_HANDLE(anv_cmd_pool, pool, commandPool);
list_for_each_entry(struct anv_cmd_buffer, cmd_buffer,
&pool->cmd_buffers, pool_link) {
anv_cmd_buffer_reset(cmd_buffer);
}
return VK_SUCCESS;
}
void anv_TrimCommandPool(
VkDevice device,
VkCommandPool commandPool,
VkCommandPoolTrimFlags flags)
{
/* Nothing for us to do here. Our pools stay pretty tidy. */
}
/**
* Return NULL if the current subpass has no depthstencil attachment.
*/
const struct anv_image_view *
anv_cmd_buffer_get_depth_stencil_view(const struct anv_cmd_buffer *cmd_buffer)
{
const struct anv_subpass *subpass = cmd_buffer->state.subpass;
if (subpass->depth_stencil_attachment == NULL)
return NULL;
const struct anv_image_view *iview =
cmd_buffer->state.attachments[subpass->depth_stencil_attachment->attachment].image_view;
assert(iview->aspect_mask & (VK_IMAGE_ASPECT_DEPTH_BIT |
VK_IMAGE_ASPECT_STENCIL_BIT));
return iview;
}
static struct anv_descriptor_set *
anv_cmd_buffer_push_descriptor_set(struct anv_cmd_buffer *cmd_buffer,
VkPipelineBindPoint bind_point,
struct anv_descriptor_set_layout *layout,
uint32_t _set)
{
struct anv_cmd_pipeline_state *pipe_state;
if (bind_point == VK_PIPELINE_BIND_POINT_COMPUTE) {
pipe_state = &cmd_buffer->state.compute.base;
} else {
assert(bind_point == VK_PIPELINE_BIND_POINT_GRAPHICS);
pipe_state = &cmd_buffer->state.gfx.base;
}
struct anv_push_descriptor_set **push_set =
&pipe_state->push_descriptors[_set];
if (*push_set == NULL) {
*push_set = vk_zalloc(&cmd_buffer->pool->alloc,
sizeof(struct anv_push_descriptor_set), 8,
VK_SYSTEM_ALLOCATION_SCOPE_OBJECT);
if (*push_set == NULL) {
anv_batch_set_error(&cmd_buffer->batch, VK_ERROR_OUT_OF_HOST_MEMORY);
return NULL;
}
}
struct anv_descriptor_set *set = &(*push_set)->set;
if (set->layout != layout) {
if (set->layout)
anv_descriptor_set_layout_unref(cmd_buffer->device, set->layout);
anv_descriptor_set_layout_ref(layout);
set->layout = layout;
}
set->size = anv_descriptor_set_layout_size(layout);
set->buffer_view_count = layout->buffer_view_count;
set->buffer_views = (*push_set)->buffer_views;
if (layout->descriptor_buffer_size &&
((*push_set)->set_used_on_gpu ||
set->desc_mem.alloc_size < layout->descriptor_buffer_size)) {
/* The previous buffer is either actively used by some GPU command (so
* we can't modify it) or is too small. Allocate a new one.
*/
struct anv_state desc_mem =
anv_state_stream_alloc(&cmd_buffer->dynamic_state_stream,
layout->descriptor_buffer_size, 32);
if (set->desc_mem.alloc_size) {
/* TODO: Do we really need to copy all the time? */
memcpy(desc_mem.map, set->desc_mem.map,
MIN2(desc_mem.alloc_size, set->desc_mem.alloc_size));
}
set->desc_mem = desc_mem;
struct anv_address addr = {
.bo = cmd_buffer->dynamic_state_stream.state_pool->block_pool.bo,
.offset = set->desc_mem.offset,
};
const struct isl_device *isl_dev = &cmd_buffer->device->isl_dev;
set->desc_surface_state =
anv_state_stream_alloc(&cmd_buffer->surface_state_stream,
isl_dev->ss.size, isl_dev->ss.align);
anv_fill_buffer_surface_state(cmd_buffer->device,
set->desc_surface_state,
ISL_FORMAT_R32G32B32A32_FLOAT,
addr, layout->descriptor_buffer_size, 1);
}
return set;
}
void anv_CmdPushDescriptorSetKHR(
VkCommandBuffer commandBuffer,
VkPipelineBindPoint pipelineBindPoint,
VkPipelineLayout _layout,
uint32_t _set,
uint32_t descriptorWriteCount,
const VkWriteDescriptorSet* pDescriptorWrites)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
ANV_FROM_HANDLE(anv_pipeline_layout, layout, _layout);
assert(_set < MAX_SETS);
struct anv_descriptor_set_layout *set_layout = layout->set[_set].layout;
struct anv_descriptor_set *set =
anv_cmd_buffer_push_descriptor_set(cmd_buffer, pipelineBindPoint,
set_layout, _set);
if (!set)
return;
/* Go through the user supplied descriptors. */
for (uint32_t i = 0; i < descriptorWriteCount; i++) {
const VkWriteDescriptorSet *write = &pDescriptorWrites[i];
switch (write->descriptorType) {
case VK_DESCRIPTOR_TYPE_SAMPLER:
case VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER:
case VK_DESCRIPTOR_TYPE_SAMPLED_IMAGE:
case VK_DESCRIPTOR_TYPE_STORAGE_IMAGE:
case VK_DESCRIPTOR_TYPE_INPUT_ATTACHMENT:
for (uint32_t j = 0; j < write->descriptorCount; j++) {
anv_descriptor_set_write_image_view(cmd_buffer->device, set,
write->pImageInfo + j,
write->descriptorType,
write->dstBinding,
write->dstArrayElement + j);
}
break;
case VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER:
case VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER:
for (uint32_t j = 0; j < write->descriptorCount; j++) {
ANV_FROM_HANDLE(anv_buffer_view, bview,
write->pTexelBufferView[j]);
anv_descriptor_set_write_buffer_view(cmd_buffer->device, set,
write->descriptorType,
bview,
write->dstBinding,
write->dstArrayElement + j);
}
break;
case VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER:
case VK_DESCRIPTOR_TYPE_STORAGE_BUFFER:
case VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER_DYNAMIC:
case VK_DESCRIPTOR_TYPE_STORAGE_BUFFER_DYNAMIC:
for (uint32_t j = 0; j < write->descriptorCount; j++) {
assert(write->pBufferInfo[j].buffer);
ANV_FROM_HANDLE(anv_buffer, buffer, write->pBufferInfo[j].buffer);
assert(buffer);
anv_descriptor_set_write_buffer(cmd_buffer->device, set,
&cmd_buffer->surface_state_stream,
write->descriptorType,
buffer,
write->dstBinding,
write->dstArrayElement + j,
write->pBufferInfo[j].offset,
write->pBufferInfo[j].range);
}
break;
default:
break;
}
}
anv_cmd_buffer_bind_descriptor_set(cmd_buffer, pipelineBindPoint,
layout, _set, set, NULL, NULL);
}
void anv_CmdPushDescriptorSetWithTemplateKHR(
VkCommandBuffer commandBuffer,
VkDescriptorUpdateTemplate descriptorUpdateTemplate,
VkPipelineLayout _layout,
uint32_t _set,
const void* pData)
{
ANV_FROM_HANDLE(anv_cmd_buffer, cmd_buffer, commandBuffer);
ANV_FROM_HANDLE(anv_descriptor_update_template, template,
descriptorUpdateTemplate);
ANV_FROM_HANDLE(anv_pipeline_layout, layout, _layout);
assert(_set < MAX_PUSH_DESCRIPTORS);
struct anv_descriptor_set_layout *set_layout = layout->set[_set].layout;
struct anv_descriptor_set *set =
anv_cmd_buffer_push_descriptor_set(cmd_buffer, template->bind_point,
set_layout, _set);
if (!set)
return;
anv_descriptor_set_write_template(cmd_buffer->device, set,
&cmd_buffer->surface_state_stream,
template,
pData);
anv_cmd_buffer_bind_descriptor_set(cmd_buffer, template->bind_point,
layout, _set, set, NULL, NULL);
}
void anv_CmdSetDeviceMask(
VkCommandBuffer commandBuffer,
uint32_t deviceMask)
{
/* No-op */
}
|