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
|
/*
* 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"
// Shader functions
VkResult anv_CreateShaderModule(
VkDevice _device,
const VkShaderModuleCreateInfo* pCreateInfo,
VkShaderModule* pShaderModule)
{
ANV_FROM_HANDLE(anv_device, device, _device);
struct anv_shader_module *module;
assert(pCreateInfo->sType == VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO);
assert(pCreateInfo->flags == 0);
module = anv_device_alloc(device, sizeof(*module) + pCreateInfo->codeSize, 8,
VK_SYSTEM_ALLOC_TYPE_API_OBJECT);
if (module == NULL)
return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
module->nir = NULL;
module->size = pCreateInfo->codeSize;
memcpy(module->data, pCreateInfo->pCode, module->size);
*pShaderModule = anv_shader_module_to_handle(module);
return VK_SUCCESS;
}
void anv_DestroyShaderModule(
VkDevice _device,
VkShaderModule _module)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_shader_module, module, _module);
anv_device_free(device, module);
}
VkResult anv_CreateShader(
VkDevice _device,
const VkShaderCreateInfo* pCreateInfo,
VkShader* pShader)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_shader_module, module, pCreateInfo->module);
struct anv_shader *shader;
assert(pCreateInfo->sType == VK_STRUCTURE_TYPE_SHADER_CREATE_INFO);
assert(pCreateInfo->flags == 0);
const char *name = pCreateInfo->pName ? pCreateInfo->pName : "main";
size_t name_len = strlen(name);
if (strcmp(name, "main") != 0) {
anv_finishme("Multiple shaders per module not really supported");
}
shader = anv_device_alloc(device, sizeof(*shader) + name_len + 1, 8,
VK_SYSTEM_ALLOC_TYPE_API_OBJECT);
if (shader == NULL)
return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
shader->module = module;
memcpy(shader->entrypoint, name, name_len + 1);
*pShader = anv_shader_to_handle(shader);
return VK_SUCCESS;
}
void anv_DestroyShader(
VkDevice _device,
VkShader _shader)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_shader, shader, _shader);
anv_device_free(device, shader);
}
VkResult anv_CreatePipelineCache(
VkDevice device,
const VkPipelineCacheCreateInfo* pCreateInfo,
VkPipelineCache* pPipelineCache)
{
pPipelineCache->handle = 1;
stub_return(VK_SUCCESS);
}
void anv_DestroyPipelineCache(
VkDevice _device,
VkPipelineCache _cache)
{
}
size_t anv_GetPipelineCacheSize(
VkDevice device,
VkPipelineCache pipelineCache)
{
stub_return(0);
}
VkResult anv_GetPipelineCacheData(
VkDevice device,
VkPipelineCache pipelineCache,
void* pData)
{
stub_return(VK_UNSUPPORTED);
}
VkResult anv_MergePipelineCaches(
VkDevice device,
VkPipelineCache destCache,
uint32_t srcCacheCount,
const VkPipelineCache* pSrcCaches)
{
stub_return(VK_UNSUPPORTED);
}
void anv_DestroyPipeline(
VkDevice _device,
VkPipeline _pipeline)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_pipeline, pipeline, _pipeline);
anv_compiler_free(pipeline);
anv_reloc_list_finish(&pipeline->batch_relocs, pipeline->device);
anv_state_stream_finish(&pipeline->program_stream);
anv_state_pool_free(&device->dynamic_state_pool, pipeline->blend_state);
anv_device_free(pipeline->device, pipeline);
}
static const uint32_t vk_to_gen_primitive_type[] = {
[VK_PRIMITIVE_TOPOLOGY_POINT_LIST] = _3DPRIM_POINTLIST,
[VK_PRIMITIVE_TOPOLOGY_LINE_LIST] = _3DPRIM_LINELIST,
[VK_PRIMITIVE_TOPOLOGY_LINE_STRIP] = _3DPRIM_LINESTRIP,
[VK_PRIMITIVE_TOPOLOGY_TRIANGLE_LIST] = _3DPRIM_TRILIST,
[VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP] = _3DPRIM_TRISTRIP,
[VK_PRIMITIVE_TOPOLOGY_TRIANGLE_FAN] = _3DPRIM_TRIFAN,
[VK_PRIMITIVE_TOPOLOGY_LINE_LIST_ADJ] = _3DPRIM_LINELIST_ADJ,
[VK_PRIMITIVE_TOPOLOGY_LINE_STRIP_ADJ] = _3DPRIM_LINESTRIP_ADJ,
[VK_PRIMITIVE_TOPOLOGY_TRIANGLE_LIST_ADJ] = _3DPRIM_TRILIST_ADJ,
[VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP_ADJ] = _3DPRIM_TRISTRIP_ADJ,
[VK_PRIMITIVE_TOPOLOGY_PATCH] = _3DPRIM_PATCHLIST_1
};
static void
anv_pipeline_init_dynamic_state(struct anv_pipeline *pipeline,
const VkGraphicsPipelineCreateInfo *pCreateInfo)
{
pipeline->dynamic_state_mask = 0;
if (pCreateInfo->pDynamicState == NULL)
return;
uint32_t count = pCreateInfo->pDynamicState->dynamicStateCount;
struct anv_dynamic_state *dynamic = &pipeline->dynamic_state;
for (uint32_t s = 0; s < count; s++) {
VkDynamicState state = pCreateInfo->pDynamicState->pDynamicStates[s];
assert(state < 32);
pipeline->dynamic_state_mask |= (1u << state);
switch (state) {
case VK_DYNAMIC_STATE_VIEWPORT:
assert(pCreateInfo->pViewportState);
dynamic->viewport.count = pCreateInfo->pViewportState->viewportCount;
typed_memcpy(dynamic->viewport.viewports,
pCreateInfo->pViewportState->pViewports,
pCreateInfo->pViewportState->viewportCount);
break;
case VK_DYNAMIC_STATE_SCISSOR:
assert(pCreateInfo->pViewportState);
dynamic->scissor.count = pCreateInfo->pViewportState->scissorCount;
typed_memcpy(dynamic->scissor.scissors,
pCreateInfo->pViewportState->pScissors,
pCreateInfo->pViewportState->scissorCount);
break;
case VK_DYNAMIC_STATE_LINE_WIDTH:
assert(pCreateInfo->pRasterState);
dynamic->line_width = pCreateInfo->pRasterState->lineWidth;
break;
case VK_DYNAMIC_STATE_DEPTH_BIAS:
assert(pCreateInfo->pRasterState);
dynamic->depth_bias.bias = pCreateInfo->pRasterState->depthBias;
dynamic->depth_bias.clamp = pCreateInfo->pRasterState->depthBiasClamp;
dynamic->depth_bias.slope_scaled =
pCreateInfo->pRasterState->slopeScaledDepthBias;
break;
case VK_DYNAMIC_STATE_BLEND_CONSTANTS:
assert(pCreateInfo->pColorBlendState);
typed_memcpy(dynamic->blend_constants,
pCreateInfo->pColorBlendState->blendConst, 4);
break;
case VK_DYNAMIC_STATE_DEPTH_BOUNDS:
assert(pCreateInfo->pDepthStencilState);
dynamic->depth_bounds.min =
pCreateInfo->pDepthStencilState->minDepthBounds;
dynamic->depth_bounds.max =
pCreateInfo->pDepthStencilState->maxDepthBounds;
break;
case VK_DYNAMIC_STATE_STENCIL_COMPARE_MASK:
assert(pCreateInfo->pDepthStencilState);
dynamic->stencil_compare_mask.front =
pCreateInfo->pDepthStencilState->front.stencilCompareMask;
dynamic->stencil_compare_mask.back =
pCreateInfo->pDepthStencilState->back.stencilCompareMask;
break;
case VK_DYNAMIC_STATE_STENCIL_WRITE_MASK:
assert(pCreateInfo->pDepthStencilState);
dynamic->stencil_write_mask.front =
pCreateInfo->pDepthStencilState->front.stencilWriteMask;
dynamic->stencil_write_mask.back =
pCreateInfo->pDepthStencilState->back.stencilWriteMask;
break;
case VK_DYNAMIC_STATE_STENCIL_REFERENCE:
assert(pCreateInfo->pDepthStencilState);
dynamic->stencil_reference.front =
pCreateInfo->pDepthStencilState->front.stencilReference;
dynamic->stencil_reference.back =
pCreateInfo->pDepthStencilState->back.stencilReference;
break;
default:
assert(!"Invalid dynamic state");
}
}
}
VkResult
anv_pipeline_init(struct anv_pipeline *pipeline, struct anv_device *device,
const VkGraphicsPipelineCreateInfo *pCreateInfo,
const struct anv_graphics_pipeline_create_info *extra)
{
VkResult result;
pipeline->device = device;
pipeline->layout = anv_pipeline_layout_from_handle(pCreateInfo->layout);
memset(pipeline->shaders, 0, sizeof(pipeline->shaders));
result = anv_reloc_list_init(&pipeline->batch_relocs, device);
if (result != VK_SUCCESS) {
anv_device_free(device, pipeline);
return result;
}
pipeline->batch.next = pipeline->batch.start = pipeline->batch_data;
pipeline->batch.end = pipeline->batch.start + sizeof(pipeline->batch_data);
pipeline->batch.relocs = &pipeline->batch_relocs;
anv_state_stream_init(&pipeline->program_stream,
&device->instruction_block_pool);
for (uint32_t i = 0; i < pCreateInfo->stageCount; i++) {
pipeline->shaders[pCreateInfo->pStages[i].stage] =
anv_shader_from_handle(pCreateInfo->pStages[i].shader);
}
anv_pipeline_init_dynamic_state(pipeline, pCreateInfo);
if (pCreateInfo->pTessellationState)
anv_finishme("VK_STRUCTURE_TYPE_PIPELINE_TESSELLATION_STATE_CREATE_INFO");
if (pCreateInfo->pViewportState)
anv_finishme("VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO");
if (pCreateInfo->pMultisampleState &&
pCreateInfo->pMultisampleState->rasterSamples > 1)
anv_finishme("VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO");
pipeline->use_repclear = extra && extra->use_repclear;
anv_compiler_run(device->compiler, pipeline);
const struct brw_wm_prog_data *wm_prog_data = &pipeline->wm_prog_data;
pipeline->ps_ksp2 = 0;
pipeline->ps_grf_start2 = 0;
if (pipeline->ps_simd8 != NO_KERNEL) {
pipeline->ps_ksp0 = pipeline->ps_simd8;
pipeline->ps_grf_start0 = wm_prog_data->base.dispatch_grf_start_reg;
if (pipeline->ps_simd16 != NO_KERNEL) {
pipeline->ps_ksp2 = pipeline->ps_simd16;
pipeline->ps_grf_start2 = wm_prog_data->dispatch_grf_start_reg_16;
}
} else if (pipeline->ps_simd16 != NO_KERNEL) {
pipeline->ps_ksp0 = pipeline->ps_simd16;
pipeline->ps_grf_start0 = wm_prog_data->dispatch_grf_start_reg_16;
} else {
unreachable("no ps shader");
}
const VkPipelineVertexInputStateCreateInfo *vi_info =
pCreateInfo->pVertexInputState;
pipeline->vb_used = 0;
for (uint32_t i = 0; i < vi_info->bindingCount; i++) {
const VkVertexInputBindingDescription *desc =
&vi_info->pVertexBindingDescriptions[i];
pipeline->vb_used |= 1 << desc->binding;
pipeline->binding_stride[desc->binding] = desc->strideInBytes;
/* Step rate is programmed per vertex element (attribute), not
* binding. Set up a map of which bindings step per instance, for
* reference by vertex element setup. */
switch (desc->stepRate) {
default:
case VK_VERTEX_INPUT_STEP_RATE_VERTEX:
pipeline->instancing_enable[desc->binding] = false;
break;
case VK_VERTEX_INPUT_STEP_RATE_INSTANCE:
pipeline->instancing_enable[desc->binding] = true;
break;
}
}
const VkPipelineInputAssemblyStateCreateInfo *ia_info =
pCreateInfo->pInputAssemblyState;
pipeline->primitive_restart = ia_info->primitiveRestartEnable;
pipeline->topology = vk_to_gen_primitive_type[ia_info->topology];
if (extra && extra->use_rectlist)
pipeline->topology = _3DPRIM_RECTLIST;
return VK_SUCCESS;
}
VkResult
anv_graphics_pipeline_create(
VkDevice _device,
const VkGraphicsPipelineCreateInfo *pCreateInfo,
const struct anv_graphics_pipeline_create_info *extra,
VkPipeline *pPipeline)
{
ANV_FROM_HANDLE(anv_device, device, _device);
switch (device->info.gen) {
case 7:
return gen7_graphics_pipeline_create(_device, pCreateInfo, extra, pPipeline);
case 8:
return gen8_graphics_pipeline_create(_device, pCreateInfo, extra, pPipeline);
default:
unreachable("unsupported gen\n");
}
}
VkResult anv_CreateGraphicsPipelines(
VkDevice _device,
VkPipelineCache pipelineCache,
uint32_t count,
const VkGraphicsPipelineCreateInfo* pCreateInfos,
VkPipeline* pPipelines)
{
VkResult result = VK_SUCCESS;
unsigned i = 0;
for (; i < count; i++) {
result = anv_graphics_pipeline_create(_device, &pCreateInfos[i],
NULL, &pPipelines[i]);
if (result != VK_SUCCESS) {
for (unsigned j = 0; j < i; j++) {
anv_DestroyPipeline(_device, pPipelines[j]);
}
return result;
}
}
return VK_SUCCESS;
}
static VkResult anv_compute_pipeline_create(
VkDevice _device,
const VkComputePipelineCreateInfo* pCreateInfo,
VkPipeline* pPipeline)
{
ANV_FROM_HANDLE(anv_device, device, _device);
switch (device->info.gen) {
case 7:
return gen7_compute_pipeline_create(_device, pCreateInfo, pPipeline);
case 8:
return gen8_compute_pipeline_create(_device, pCreateInfo, pPipeline);
default:
unreachable("unsupported gen\n");
}
}
VkResult anv_CreateComputePipelines(
VkDevice _device,
VkPipelineCache pipelineCache,
uint32_t count,
const VkComputePipelineCreateInfo* pCreateInfos,
VkPipeline* pPipelines)
{
VkResult result = VK_SUCCESS;
unsigned i = 0;
for (; i < count; i++) {
result = anv_compute_pipeline_create(_device, &pCreateInfos[i],
&pPipelines[i]);
if (result != VK_SUCCESS) {
for (unsigned j = 0; j < i; j++) {
anv_DestroyPipeline(_device, pPipelines[j]);
}
return result;
}
}
return VK_SUCCESS;
}
// Pipeline layout functions
VkResult anv_CreatePipelineLayout(
VkDevice _device,
const VkPipelineLayoutCreateInfo* pCreateInfo,
VkPipelineLayout* pPipelineLayout)
{
ANV_FROM_HANDLE(anv_device, device, _device);
struct anv_pipeline_layout *layout;
assert(pCreateInfo->sType == VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO);
layout = anv_device_alloc(device, sizeof(*layout), 8,
VK_SYSTEM_ALLOC_TYPE_API_OBJECT);
if (layout == NULL)
return vk_error(VK_ERROR_OUT_OF_HOST_MEMORY);
layout->num_sets = pCreateInfo->descriptorSetCount;
uint32_t surface_start[VK_SHADER_STAGE_NUM] = { 0, };
uint32_t sampler_start[VK_SHADER_STAGE_NUM] = { 0, };
for (uint32_t s = 0; s < VK_SHADER_STAGE_NUM; s++) {
layout->stage[s].has_dynamic_offsets = false;
layout->stage[s].surface_count = 0;
layout->stage[s].sampler_count = 0;
}
uint32_t num_dynamic_offsets = 0;
for (uint32_t i = 0; i < pCreateInfo->descriptorSetCount; i++) {
ANV_FROM_HANDLE(anv_descriptor_set_layout, set_layout,
pCreateInfo->pSetLayouts[i]);
layout->set[i].layout = set_layout;
layout->set[i].dynamic_offset_start = num_dynamic_offsets;
num_dynamic_offsets += set_layout->num_dynamic_buffers;
for (uint32_t s = 0; s < VK_SHADER_STAGE_NUM; s++) {
if (set_layout->num_dynamic_buffers > 0)
layout->stage[s].has_dynamic_offsets = true;
layout->set[i].stage[s].surface_start = surface_start[s];
surface_start[s] += set_layout->stage[s].surface_count;
layout->set[i].stage[s].sampler_start = sampler_start[s];
sampler_start[s] += set_layout->stage[s].sampler_count;
layout->stage[s].surface_count += set_layout->stage[s].surface_count;
layout->stage[s].sampler_count += set_layout->stage[s].sampler_count;
}
}
*pPipelineLayout = anv_pipeline_layout_to_handle(layout);
return VK_SUCCESS;
}
void anv_DestroyPipelineLayout(
VkDevice _device,
VkPipelineLayout _pipelineLayout)
{
ANV_FROM_HANDLE(anv_device, device, _device);
ANV_FROM_HANDLE(anv_pipeline_layout, pipeline_layout, _pipelineLayout);
anv_device_free(device, pipeline_layout);
}
|