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
|
/*
* Copyright (c) 2020-2025 The Khronos Group Inc.
* Copyright (c) 2020-2025 Valve Corporation
* Copyright (c) 2020-2025 LunarG, Inc.
* Copyright (c) 2020-2025 Google, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*/
#include "../framework/layer_validation_tests.h"
#include "../framework/pipeline_helper.h"
#include "../framework/descriptor_helper.h"
#include "cooperative_matrix_helper.h"
class NegativeGpuAVDescriptorClassGeneralBufferCoopMat : public GpuAVDescriptorClassGeneralBufferCoopMat {};
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, Store) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) coherent buffer SSBO { uint8_t payload[]; };
coopmat<uint8_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseA> matA;
void main() {
coopMatLoad(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
coopMatStore(matA, payload, 0, 32, gl_CooperativeMatrixLayoutRowMajor);
}
)glsl";
BasicComputeTest(cs_source, SPV_SOURCE_GLSL, 256, "VUID-vkCmdDispatch-storageBuffers-06936");
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, Load) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) coherent buffer SSBO { uint8_t payload[]; };
coopmat<uint32_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseAccumulator> matC;
void main() {
coopMatLoad(matC, payload, 1, 16, gl_CooperativeMatrixLayoutColumnMajor);
coopMatStore(matC, payload, 0, 16, gl_CooperativeMatrixLayoutColumnMajor);
}
)glsl";
BasicComputeTest(cs_source, SPV_SOURCE_GLSL, 1024, "VUID-vkCmdDispatch-storageBuffers-06936");
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, Function) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) coherent buffer SSBO { uint8_t payload[]; };
coopmat<uint8_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseA> matA;
void foo() {
coopMatStore(matA, payload, 0, 32, gl_CooperativeMatrixLayoutRowMajor);
}
void main() {
coopMatLoad(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
coopMatStore(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
foo();
}
)glsl";
BasicComputeTest(cs_source, SPV_SOURCE_GLSL, 256, "VUID-vkCmdDispatch-storageBuffers-06936");
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, MidStruct) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) coherent buffer SSBO {
vec4 a;
uint32_t payload[16];
float c;
};
coopmat<uint8_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseA> matA;
void main() {
coopMatLoad(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
coopMatStore(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
}
)glsl";
BasicComputeTest(cs_source, SPV_SOURCE_GLSL, 128, "VUID-vkCmdDispatch-storageBuffers-06936");
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, ElementSize) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) coherent buffer SSBO { uvec4 payload[]; }; // each element is 16 bytes
coopmat<uint8_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseA> matA; // 256 bytes
void main() {
// access 528 (16 * 33) bytes into payload
coopMatLoad(matA, payload, 33, 16, gl_CooperativeMatrixLayoutRowMajor);
coopMatStore(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
}
)glsl";
BasicComputeTest(cs_source, SPV_SOURCE_GLSL, 768, "VUID-vkCmdDispatch-storageBuffers-06936");
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, Mix) {
TEST_DESCRIPTION("Mix OOB from both a CoopMat and a normal store OOB");
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
RETURN_IF_SKIP(InitCooperativeMatrixKHR(true));
CooperativeMatrixHelper helper(*this);
if (!helper.Has16x16UintProperty()) {
GTEST_SKIP() << "16x16 Uint Property not found";
}
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_KHR_shader_subgroup_basic : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) buffer SSBO0 { uint8_t good_payload[1024]; };
layout(set=0, binding=1) buffer SSBO1 { uint8_t bad_payload[1024]; };
coopmat<uint32_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseAccumulator> matC;
void main() {
coopMatLoad(matC, good_payload, 0, 16, gl_CooperativeMatrixLayoutColumnMajor);
coopMatStore(matC, bad_payload, 0, 16, gl_CooperativeMatrixLayoutColumnMajor);
// only want 1 thread to have error
if (subgroupElect()) {
bad_payload[1000] = uint8_t(0);
}
}
)glsl";
OneOffDescriptorSet descriptor_set(m_device, {{0, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 1, VK_SHADER_STAGE_ALL, nullptr},
{1, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 1, VK_SHADER_STAGE_ALL, nullptr}});
vkt::PipelineLayout pipeline_layout(*m_device, {&descriptor_set.layout_});
CreateComputePipelineHelper pipe(*this);
pipe.cs_ = VkShaderObj(this, cs_source, VK_SHADER_STAGE_COMPUTE_BIT, SPV_ENV_VULKAN_1_3);
pipe.cp_ci_.layout = pipeline_layout;
pipe.CreateComputePipeline();
m_errorMonitor->VerifyFound();
vkt::Buffer ssbo(*m_device, 1024, VK_BUFFER_USAGE_STORAGE_BUFFER_BIT, kHostVisibleMemProps);
descriptor_set.WriteDescriptorBufferInfo(0, ssbo, 0, VK_WHOLE_SIZE, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER);
descriptor_set.WriteDescriptorBufferInfo(1, ssbo, 0, 512, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER); // only half bound
descriptor_set.UpdateDescriptorSets();
m_command_buffer.Begin();
vk::CmdBindPipeline(m_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipe);
vk::CmdBindDescriptorSets(m_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipeline_layout, 0, 1, &descriptor_set.set_, 0,
nullptr);
vk::CmdDispatch(m_command_buffer, 1, 1, 1);
m_command_buffer.End();
m_errorMonitor->SetDesiredError("VUID-vkCmdDispatch-storageBuffers-06936");
m_default_queue->SubmitAndWait(m_command_buffer);
m_errorMonitor->VerifyFound();
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, Robustness) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
// Not the correct robustness for CoopMat
AddRequiredFeature(vkt::Feature::robustBufferAccess);
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) coherent buffer SSBO { uint8_t payload[]; };
coopmat<uint8_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseA> matA;
void main() {
coopMatLoad(matA, payload, 0, 16, gl_CooperativeMatrixLayoutRowMajor);
coopMatStore(matA, payload, 0, 32, gl_CooperativeMatrixLayoutRowMajor);
}
)glsl";
BasicComputeTest(cs_source, SPV_SOURCE_GLSL, 256, "VUID-vkCmdDispatch-storageBuffers-06936");
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, DynamicStride) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
RETURN_IF_SKIP(InitCooperativeMatrixKHR(true));
CooperativeMatrixHelper helper(*this);
if (!helper.Has16x16UintProperty()) {
GTEST_SKIP() << "16x16 Uint Property not found";
}
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) buffer SSBO {
uint32_t stride;
uint8_t payload[]; // 1024 bytes needed
};
coopmat<uint32_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseAccumulator> matC;
void main() {
coopMatLoad(matC, payload, 0, stride + 4, gl_CooperativeMatrixLayoutColumnMajor);
coopMatStore(matC, payload, 0, 16, gl_CooperativeMatrixLayoutColumnMajor);
}
)glsl";
CreateComputePipelineHelper pipe(*this);
pipe.dsl_bindings_[0] = {0, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 1, VK_SHADER_STAGE_ALL, nullptr};
pipe.cs_ = VkShaderObj(this, cs_source, VK_SHADER_STAGE_COMPUTE_BIT, SPV_ENV_VULKAN_1_2);
pipe.CreateComputePipeline();
vkt::Buffer in_buffer(*m_device, 2048, VK_BUFFER_USAGE_STORAGE_BUFFER_BIT, kHostVisibleMemProps);
pipe.descriptor_set_.WriteDescriptorBufferInfo(0, in_buffer, 0, VK_WHOLE_SIZE, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER);
pipe.descriptor_set_.UpdateDescriptorSets();
uint32_t* buffer_ptr = (uint32_t*)in_buffer.Memory().Map();
buffer_ptr[0] = 64; // stride
m_command_buffer.Begin();
vk::CmdBindPipeline(m_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipe);
vk::CmdBindDescriptorSets(m_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipe.pipeline_layout_, 0, 1,
&pipe.descriptor_set_.set_, 0, nullptr);
vk::CmdDispatch(m_command_buffer, 1, 1, 1);
m_command_buffer.End();
// VUID-vkCmdDispatch-storageBuffers-06936
m_errorMonitor->SetDesiredError("highest out of bounds access was at [4143] bytes");
m_default_queue->SubmitAndWait(m_command_buffer);
m_errorMonitor->VerifyFound();
}
TEST_F(NegativeGpuAVDescriptorClassGeneralBufferCoopMat, DynamicElement) {
SetTargetApiVersion(VK_API_VERSION_1_3);
AddRequiredFeature(vkt::Feature::shaderInt8);
AddRequiredFeature(vkt::Feature::storageBuffer8BitAccess);
RETURN_IF_SKIP(InitCooperativeMatrixKHR(true));
CooperativeMatrixHelper helper(*this);
if (!helper.Has16x16UintProperty()) {
GTEST_SKIP() << "16x16 Uint Property not found";
}
const char* cs_source = R"glsl(
#version 450 core
#pragma use_vulkan_memory_model
#extension GL_KHR_memory_scope_semantics : enable
#extension GL_KHR_cooperative_matrix : enable
#extension GL_EXT_shader_explicit_arithmetic_types : enable
#extension GL_EXT_shader_explicit_arithmetic_types_int8 : enable
layout(local_size_x = 64) in;
layout(set=0, binding=0) buffer SSBO {
int32_t element;
uint8_t payload[]; // 1024 bytes needed
};
coopmat<uint32_t, gl_ScopeSubgroup, 16, 16, gl_MatrixUseAccumulator> matC;
void main() {
coopMatLoad(matC, payload, element, 16, gl_CooperativeMatrixLayoutColumnMajor);
coopMatStore(matC, payload, 0, 16, gl_CooperativeMatrixLayoutColumnMajor);
}
)glsl";
CreateComputePipelineHelper pipe(*this);
pipe.dsl_bindings_[0] = {0, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER, 1, VK_SHADER_STAGE_ALL, nullptr};
pipe.cs_ = VkShaderObj(this, cs_source, VK_SHADER_STAGE_COMPUTE_BIT, SPV_ENV_VULKAN_1_2);
pipe.CreateComputePipeline();
vkt::Buffer in_buffer(*m_device, 2048, VK_BUFFER_USAGE_STORAGE_BUFFER_BIT, kHostVisibleMemProps);
pipe.descriptor_set_.WriteDescriptorBufferInfo(0, in_buffer, 0, VK_WHOLE_SIZE, VK_DESCRIPTOR_TYPE_STORAGE_BUFFER);
pipe.descriptor_set_.UpdateDescriptorSets();
uint32_t* buffer_ptr = (uint32_t*)in_buffer.Memory().Map();
buffer_ptr[0] = 2048; // element
m_command_buffer.Begin();
vk::CmdBindPipeline(m_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipe);
vk::CmdBindDescriptorSets(m_command_buffer, VK_PIPELINE_BIND_POINT_COMPUTE, pipe.pipeline_layout_, 0, 1,
&pipe.descriptor_set_.set_, 0, nullptr);
vk::CmdDispatch(m_command_buffer, 1, 1, 1);
m_command_buffer.End();
// VUID-vkCmdDispatch-storageBuffers-06936
m_errorMonitor->SetDesiredError("highest out of bounds access was at [3075] bytes");
m_default_queue->SubmitAndWait(m_command_buffer);
m_errorMonitor->VerifyFound();
}
|
