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
|
/* Copyright (c) 2015-2024 The Khronos Group Inc.
* Copyright (C) 2025 Arm Limited.
*
* 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
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*/
#include <vulkan/utility/vk_format_utils.h>
#include <vulkan/vulkan_core.h>
#include "core_validation.h"
#include "generated/dispatch_functions.h"
#include "state_tracker/tensor_state.h"
#include "state_tracker/cmd_buffer_state.h"
bool CoreChecks::ValidateTensorFormatUsage(VkFormat format, VkTensorUsageFlagsARM usage, VkTensorTilingARM tiling, const char *vuid,
const Location &loc) const {
bool skip = false;
VkTensorFormatPropertiesARM tensor_fmt_props = vku::InitStructHelper();
VkFormatProperties2 fmt_props_2 = vku::InitStructHelper(&tensor_fmt_props);
DispatchGetPhysicalDeviceFormatProperties2Helper(api_version, physical_device, format, &fmt_props_2);
VkFormatFeatureFlags2 tensor_feature_flags{};
if (VK_TENSOR_TILING_OPTIMAL_ARM == tiling) {
tensor_feature_flags = tensor_fmt_props.optimalTilingTensorFeatures;
} else if (VK_TENSOR_TILING_LINEAR_ARM == tiling) {
tensor_feature_flags = tensor_fmt_props.linearTilingTensorFeatures;
}
const std::vector<std::pair<VkTensorUsageFlagBitsARM, VkFormatFeatureFlagBits2>> usage_to_feature_map = {
{ VK_TENSOR_USAGE_TRANSFER_SRC_BIT_ARM, VK_FORMAT_FEATURE_2_TRANSFER_SRC_BIT },
{ VK_TENSOR_USAGE_TRANSFER_DST_BIT_ARM, VK_FORMAT_FEATURE_2_TRANSFER_DST_BIT },
{ VK_TENSOR_USAGE_IMAGE_ALIASING_BIT_ARM, VK_FORMAT_FEATURE_2_TENSOR_IMAGE_ALIASING_BIT_ARM },
{ VK_TENSOR_USAGE_SHADER_BIT_ARM, VK_FORMAT_FEATURE_2_TENSOR_SHADER_BIT_ARM },
{ VK_TENSOR_USAGE_DATA_GRAPH_BIT_ARM, VK_FORMAT_FEATURE_2_TENSOR_DATA_GRAPH_BIT_ARM },
};
for (auto element : usage_to_feature_map) {
auto usage_bit = element.first;
auto feature_bit = element.second;
if (usage & usage_bit && !(tensor_feature_flags & feature_bit)) {
skip |= LogError(vuid, device, loc.dot(Field::usage), "(%s) has bit (%s) set but format features (%s) does not include matching required bit (%s)",
string_VkTensorUsageFlagsARM(usage).c_str(), string_VkTensorUsageFlagsARM(usage_bit).c_str(),
string_VkTensorUsageFlagsARM(tensor_feature_flags).c_str(), string_VkTensorUsageFlagsARM(feature_bit).c_str());
}
}
return skip;
}
bool CoreChecks::ValidateTensorCreateInfo(const VkTensorCreateInfoARM &create_info, const Location &create_info_loc) const {
bool skip = false;
const VkTensorDescriptionARM &description = *create_info.pDescription;
if (create_info.sharingMode == VK_SHARING_MODE_CONCURRENT && create_info.pQueueFamilyIndices) {
skip |= ValidatePhysicalDeviceQueueFamilies(create_info.queueFamilyIndexCount, create_info.pQueueFamilyIndices,
create_info_loc, "VUID-VkTensorCreateInfoARM-sharingMode-09725");
if (!skip) {
uint32_t queue_family_property_count = 0;
DispatchGetPhysicalDeviceQueueFamilyProperties2Helper(api_version, physical_device, &queue_family_property_count,
nullptr);
vvl::unordered_set<uint32_t> queue_family_indices_set;
for (uint32_t i = 0; i < create_info.queueFamilyIndexCount; i++) {
const uint32_t queue_index = create_info.pQueueFamilyIndices[i];
if (queue_family_indices_set.find(queue_index) != queue_family_indices_set.end()) {
skip |=
LogError("VUID-VkTensorCreateInfoARM-sharingMode-09725", device, create_info_loc.dot(Field::sharingMode),
"is VK_SHARING_MODE_CONCURRENT but pQueueFamilyIndices[%" PRIu32 "] (%" PRIu32 ") is not unique",
i, queue_index);
break;
} else if (queue_index >= queue_family_property_count) {
skip |=
LogError("VUID-VkTensorCreateInfoARM-sharingMode-09725", device, create_info_loc.dot(Field::sharingMode),
"is VK_SHARING_MODE_CONCURRENT but pQueueFamilyIndices[%" PRIu32 "] (%" PRIu32
") is >= pQueueFamilyPropertyCount (%" PRIu32
") returned by vkGetPhysicalDeviceQueueFamilyProperties2 for the "
"physicalDevice that was used to create device",
i, queue_index, queue_family_property_count);
break;
}
queue_family_indices_set.emplace(queue_index);
}
}
}
// Check usage vs. features
const auto required_bits = VK_TENSOR_USAGE_SHADER_BIT_ARM | VK_TENSOR_USAGE_DATA_GRAPH_BIT_ARM;
if ((description.usage & required_bits) == 0) {
skip |= ValidateTensorFormatUsage(description.format, description.usage, description.tiling,
"VUID-VkTensorCreateInfoARM-pDescription-09728", create_info_loc);
}
return skip;
}
bool CoreChecks::PreCallValidateCreateTensorARM(VkDevice device, const VkTensorCreateInfoARM *pCreateInfo,
const VkAllocationCallbacks *pAllocator, VkTensorARM *pTensor,
const ErrorObject &error_obj) const {
bool skip = false;
const Location create_info_loc = error_obj.location.dot(Field::pCreateInfo);
skip |= ValidateTensorCreateInfo(*pCreateInfo, create_info_loc);
return skip;
}
bool CoreChecks::PreCallValidateCreateTensorViewARM(VkDevice device, const VkTensorViewCreateInfoARM *pCreateInfo,
const VkAllocationCallbacks *pAllocator, VkTensorViewARM *pView,
const ErrorObject &error_obj) const {
bool skip = false;
auto tensor_state_ptr = Get<vvl::Tensor>(pCreateInfo->tensor);
ASSERT_AND_RETURN_SKIP(tensor_state_ptr);
const auto &tensor_state = *tensor_state_ptr;
const Location create_info_loc = error_obj.location.dot(Field::pCreateInfo);
auto valid_usage_flags = VK_TENSOR_USAGE_SHADER_BIT_ARM | VK_TENSOR_USAGE_DATA_GRAPH_BIT_ARM;
skip |= ValidateTensorUsageFlags(VK_NULL_HANDLE, tensor_state, valid_usage_flags, "VUID-VkTensorViewCreateInfoARM-usage-09747",
create_info_loc.dot(Field::tensor));
skip |= ValidateMemoryIsBoundToTensor(LogObjectList(device, pCreateInfo->tensor), tensor_state,
create_info_loc.dot(Field::tensor), "VUID-VkTensorViewCreateInfoARM-tensor-09749");
const VkFormat view_format = pCreateInfo->format;
const VkFormat tensor_format = tensor_state.description.format;
if ((tensor_state.create_info.flags & VK_TENSOR_CREATE_MUTABLE_FORMAT_BIT_ARM) != 0) {
auto view_format_compatibility = vkuFormatCompatibilityClass(view_format);
if (vkuFormatCompatibilityClass(tensor_format) != view_format_compatibility) {
skip |= LogError("VUID-VkTensorViewCreateInfoARM-tensor-09744", pCreateInfo->tensor, create_info_loc.dot(Field::format),
"(%s) is not in the same format compatibility class as the tensor format (%s).",
string_VkFormat(view_format), string_VkFormat(tensor_format));
}
} else if (tensor_format != view_format) {
skip |= LogError("VUID-VkTensorViewCreateInfoARM-tensor-09743", pCreateInfo->tensor, create_info_loc.dot(Field::format),
"(%s) is different from the tensor format (%s).", string_VkFormat(view_format),
string_VkFormat(tensor_format));
}
if ((pCreateInfo->flags & VK_TENSOR_VIEW_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY_BIT_ARM) != 0) {
if (!enabled_features.descriptorBufferCaptureReplay) {
skip |= LogError("VUID-VkTensorViewCreateInfoARM-flags-09745", device, create_info_loc.dot(Field::flags),
"(%s) contains VK_TENSOR_VIEW_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY BIT but the "
"descriptorBufferCaptureReplay feature is not enabled",
string_VkTensorViewCreateFlagsARM(pCreateInfo->flags).c_str());
}
}
if (const auto opaque_capture_descriptor_buffer =
vku::FindStructInPNextChain<VkOpaqueCaptureDescriptorDataCreateInfoEXT>(pCreateInfo->pNext);
opaque_capture_descriptor_buffer &&
!(pCreateInfo->flags & VK_TENSOR_VIEW_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY_BIT_ARM)) {
skip |= LogError("VUID-VkTensorViewCreateInfoARM-pNext-09746", device, create_info_loc.dot(Field::flags),
"(%s) is missing VK_IMAGE_VIEW_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY_BIT_EXT but "
"VkOpaqueCaptureDescriptorDataCreateInfoEXT is in the pNext chain.",
string_VkTensorViewCreateFlagsARM(pCreateInfo->flags).c_str());
}
skip |= ValidateTensorFormatUsage(view_format, tensor_state.create_info.pDescription->usage,
tensor_state.create_info.pDescription->tiling, "VUID-VkTensorViewCreateInfoARM-usage-09748",
create_info_loc);
return skip;
}
bool CoreChecks::ValidateTensorUsageFlags(VkCommandBuffer commandBuffer, vvl::Tensor const &tensor_state,
VkTensorUsageFlagsARM desired, const char *vuid, const Location &tensor_loc) const {
bool skip = false;
LogObjectList objlist(commandBuffer, tensor_state.Handle());
if ((tensor_state.create_info.pDescription->usage & desired) == 0) {
skip |= LogError(vuid, objlist, tensor_loc.dot(Field::usage), "(%s) for tensor (%s) doesn't match requirements.",
string_VkTensorUsageFlagsARM(tensor_state.create_info.pDescription->usage).c_str(),
FormatHandle(tensor_state.Handle()).c_str());
}
return skip;
}
bool CoreChecks::PreCallValidateCmdCopyTensorARM(VkCommandBuffer commandBuffer, const VkCopyTensorInfoARM *pCopyTensorInfo,
const ErrorObject &error_obj) const {
bool skip = false;
auto cb_state_ptr = GetRead<vvl::CommandBuffer>(commandBuffer);
auto src_tensor_state_ptr = Get<vvl::Tensor>(pCopyTensorInfo->srcTensor);
auto dst_tensor_state_ptr = Get<vvl::Tensor>(pCopyTensorInfo->dstTensor);
ASSERT_AND_RETURN_SKIP(cb_state_ptr && src_tensor_state_ptr && dst_tensor_state_ptr);
const auto &src_tensor_state = *src_tensor_state_ptr;
const auto &dst_tensor_state = *dst_tensor_state_ptr;
const auto &cb_state = *cb_state_ptr;
const Location copy_info_loc = error_obj.location.dot(Field::pCopyTensorInfo);
LogObjectList src_objlist(commandBuffer, src_tensor_state.Handle());
LogObjectList dst_objlist(commandBuffer, dst_tensor_state.Handle());
LogObjectList tensors_objlist(src_tensor_state.Handle(), dst_tensor_state.Handle());
skip |= ValidateCmd(cb_state, error_obj.location);
if (pCopyTensorInfo->regionCount != 1) {
skip |= LogError("VUID-VkCopyTensorInfoARM-regionCount-09686", tensors_objlist, copy_info_loc.dot(Field::regionCount),
"(%" PRIu32 ") is not 1", pCopyTensorInfo->regionCount);
}
// currently there can only be 1 region (VUID 09686 above), but this way the code is future-proof
for (uint32_t j = 0; j < pCopyTensorInfo->regionCount; j++) {
const auto ®ion = pCopyTensorInfo->pRegions[j];
if (src_tensor_state.description.dimensionCount != dst_tensor_state.description.dimensionCount) {
skip |= LogError("VUID-VkCopyTensorInfoARM-dimensionCount-09684", tensors_objlist, copy_info_loc,
"dimensionCount for srcTensor (%" PRIu32 ") and dstTensor (%" PRIu32 ") are different",
src_tensor_state.description.dimensionCount, dst_tensor_state.description.dimensionCount);
} else {
for (uint32_t i = 0; i < src_tensor_state.description.dimensionCount; i++) {
if (src_tensor_state.description.pDimensions[i] != dst_tensor_state.description.pDimensions[i]) {
skip |=
LogError("VUID-VkCopyTensorInfoARM-pDimensions-09685", tensors_objlist, copy_info_loc,
"pDimensions[%" PRIu32 "] for srcTensor (%" PRIi64 ") and dstTensor (%" PRIi64 ") are different.",
i, src_tensor_state.description.pDimensions[i], dst_tensor_state.description.pDimensions[i]);
} else {
if (region.pExtent) {
if (static_cast<int64_t>(region.pExtent[i]) != src_tensor_state.description.pDimensions[i]) {
skip |= LogError("VUID-VkCopyTensorInfoARM-pRegions-09689", src_objlist,
copy_info_loc.dot(Field::pRegions).dot(Field::pExtent, i),
"(%" PRIu64 ") is not equal to srcTensor::pDimensions[%" PRIu32 "] (%" PRIi64 ")",
region.pExtent[i], i, src_tensor_state.description.pDimensions[i]);
}
}
}
}
}
if (region.pSrcOffset) {
for (uint32_t i = 0; i < region.dimensionCount; i++) {
if (region.pSrcOffset[i] != 0) {
skip |= LogError("VUID-VkCopyTensorInfoARM-pRegions-09687", src_objlist,
copy_info_loc.dot(Field::pSrcOffset, i), "(%" PRIu64 ") is not zero", region.pSrcOffset[i]);
break;
}
}
}
if (region.pDstOffset) {
for (uint32_t i = 0; i < region.dimensionCount; i++) {
if (region.pDstOffset[i] != 0) {
skip |= LogError("VUID-VkCopyTensorInfoARM-pRegions-09688", dst_objlist,
copy_info_loc.dot(Field::pDstOffset, i), "(%" PRIu64 ") is not zero", region.pDstOffset[i]);
break;
}
}
}
if (region.dimensionCount != 0) {
const uint32_t maxDimensionCount =
std::max(src_tensor_state.description.dimensionCount, dst_tensor_state.description.dimensionCount);
if (region.dimensionCount != maxDimensionCount) {
skip |= LogError("VUID-VkCopyTensorInfoARM-pRegions-09954", tensors_objlist,
copy_info_loc.dot(Field::pRegions, 0).dot(Field::dimensionCount),
"(%" PRIu32 ") is not equal to the max dimensionCount of srcTensor (%" PRIu32
") and dstTensor (%" PRIu32 ")",
region.dimensionCount, src_tensor_state.description.dimensionCount,
dst_tensor_state.description.dimensionCount);
}
} else {
if (region.pSrcOffset || region.pDstOffset || region.pExtent) {
skip |= LogError("VUID-VkTensorCopyARM-dimensionCount-09955", tensors_objlist,
copy_info_loc.dot(Field::pRegions, 0).dot(Field::dimensionCount),
"(%" PRIu32
") is zero, but at least one of pSrcOffset (%p), pDstOffset (%p), pExtent (%p) is not NULL",
region.dimensionCount, region.pSrcOffset, region.pDstOffset, region.pExtent);
}
}
}
skip |= ValidateTensorFormatUsage(src_tensor_state.description.format, VK_TENSOR_USAGE_TRANSFER_SRC_BIT_ARM,
src_tensor_state.description.tiling, "VUID-VkCopyTensorInfoARM-srcTensor-09690",
copy_info_loc.dot(Field::srcTensor));
skip |= ValidateTensorUsageFlags(commandBuffer, src_tensor_state, VK_TENSOR_USAGE_TRANSFER_SRC_BIT_ARM,
"VUID-VkCopyTensorInfoARM-srcTensor-09691", copy_info_loc.dot(Field::srcTensor));
skip |= ValidateTensorFormatUsage(dst_tensor_state.description.format, VK_TENSOR_USAGE_TRANSFER_DST_BIT_ARM,
dst_tensor_state.description.tiling, "VUID-VkCopyTensorInfoARM-dstTensor-09692",
copy_info_loc.dot(Field::dstTensor));
skip |= ValidateTensorUsageFlags(commandBuffer, dst_tensor_state, VK_TENSOR_USAGE_TRANSFER_DST_BIT_ARM,
"VUID-VkCopyTensorInfoARM-dstTensor-09693", copy_info_loc.dot(Field::dstTensor));
skip |= ValidateMemoryIsBoundToTensor(src_objlist, src_tensor_state, copy_info_loc.dot(Field::srcTensor),
"VUID-VkCopyTensorInfoARM-srcTensor-09694");
skip |= ValidateMemoryIsBoundToTensor(dst_objlist, dst_tensor_state, copy_info_loc.dot(Field::dstTensor),
"VUID-VkCopyTensorInfoARM-dstTensor-09695");
return skip;
}
bool CoreChecks::PreCallValidateDestroyTensorARM(VkDevice device, VkTensorARM tensor, const VkAllocationCallbacks *pAllocator,
const ErrorObject &error_obj) const {
bool skip = false;
auto tensor_state = Get<vvl::Tensor>(tensor);
ASSERT_AND_RETURN_SKIP(tensor_state);
skip |= ValidateObjectNotInUse(tensor_state.get(), error_obj.location, "VUID-vkDestroyTensorARM-tensor-09730");
return skip;
}
bool CoreChecks::PreCallValidateDestroyTensorViewARM(VkDevice device, VkTensorViewARM tensorView,
const VkAllocationCallbacks *pAllocator, const ErrorObject &error_obj) const {
bool skip = false;
auto tensor_view_state = Get<vvl::TensorView>(tensorView);
ASSERT_AND_RETURN_SKIP(tensor_view_state);
skip |= ValidateObjectNotInUse(tensor_view_state.get(), error_obj.location, "VUID-vkDestroyTensorViewARM-tensorView-09750");
return skip;
}
bool CoreChecks::PreCallValidateGetTensorOpaqueCaptureDescriptorDataARM(VkDevice device,
const VkTensorCaptureDescriptorDataInfoARM *pInfo,
void *pData, const ErrorObject &error_obj) const {
bool skip = false;
if (!enabled_features.descriptorBufferCaptureReplay) {
skip |= LogError("VUID-vkGetTensorOpaqueCaptureDescriptorDataARM-descriptorBufferCaptureReplay-09702", pInfo->tensor,
error_obj.location, "descriptorBufferCaptureReplay feature was not enabled.");
} else if (!enabled_features.descriptorBufferTensorDescriptors) {
skip |= LogError("VUID-vkGetTensorOpaqueCaptureDescriptorDataARM-descriptorBufferCaptureReplay-09702", pInfo->tensor,
error_obj.location, "descriptorBufferTensorDescriptors feature was not enabled.");
} else {
auto tensor_state = Get<vvl::Tensor>(pInfo->tensor);
ASSERT_AND_RETURN_SKIP(tensor_state);
if (!(tensor_state->create_info.flags & VK_TENSOR_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY_BIT_ARM)) {
skip |= LogError("VUID-VkTensorCaptureDescriptorDataInfoARM-tensor-09705", pInfo->tensor,
error_obj.location.dot(Field::pInfo).dot(Field::tensor), "was created with %s.",
string_VkTensorCreateFlagsARM(tensor_state->create_info.flags).c_str());
}
}
if (device_state->physical_device_count > 1 && !enabled_features.bufferDeviceAddressMultiDevice &&
!enabled_features.bufferDeviceAddressMultiDeviceEXT) {
skip |= LogError("VUID-vkGetTensorOpaqueCaptureDescriptorDataARM-device-09704", pInfo->tensor, error_obj.location,
"device was created with multiple physical devices (%" PRIu32
"), but the "
"bufferDeviceAddressMultiDevice feature was not enabled.",
device_state->physical_device_count);
}
return skip;
}
bool CoreChecks::PreCallValidateGetTensorViewOpaqueCaptureDescriptorDataARM(VkDevice device,
const VkTensorViewCaptureDescriptorDataInfoARM *pInfo,
void *pData, const ErrorObject &error_obj) const {
bool skip = false;
if (!enabled_features.descriptorBufferCaptureReplay) {
skip |= LogError("VUID-vkGetTensorViewOpaqueCaptureDescriptorDataARM-descriptorBufferCaptureReplay-09706",
pInfo->tensorView, error_obj.location, "descriptorBufferCaptureReplay feature was not enabled.");
} else if (!enabled_features.descriptorBufferTensorDescriptors) {
skip |= LogError("VUID-vkGetTensorViewOpaqueCaptureDescriptorDataARM-descriptorBufferCaptureReplay-09706",
pInfo->tensorView, error_obj.location, "descriptorBufferTensorDescriptors feature was not enabled.");
} else {
auto tensor_view_state = Get<vvl::TensorView>(pInfo->tensorView);
ASSERT_AND_RETURN_SKIP(tensor_view_state);
if (!(tensor_view_state->create_info.flags & VK_TENSOR_VIEW_CREATE_DESCRIPTOR_BUFFER_CAPTURE_REPLAY_BIT_ARM)) {
skip |= LogError("VUID-VkTensorViewCaptureDescriptorDataInfoARM-tensorView-09709", pInfo->tensorView,
error_obj.location.dot(Field::pInfo).dot(Field::tensor), "was created with %s.",
string_VkTensorViewCreateFlagsARM(tensor_view_state->create_info.flags).c_str());
}
}
if (device_state->physical_device_count > 1 && !enabled_features.bufferDeviceAddressMultiDevice &&
!enabled_features.bufferDeviceAddressMultiDeviceEXT) {
skip |= LogError("VUID-vkGetTensorViewOpaqueCaptureDescriptorDataARM-device-09708", pInfo->tensorView, error_obj.location,
"device was created with multiple physical devices (%" PRIu32
"), but the "
"bufferDeviceAddressMultiDevice feature was not enabled.",
device_state->physical_device_count);
}
return skip;
}
// Validates the buffer is allowed to be protected
bool CoreChecks::ValidateProtectedTensor(const vvl::CommandBuffer &cb_state, const vvl::Tensor &tensor_state,
const Location &tensor_loc, const char *vuid, const char *more_message) const {
// don't use on an unprotected tensor
assert(tensor_state.unprotected == false);
bool skip = false;
// if driver supports protectedNoFault the operation is valid, just has undefined values
if ((!phys_dev_props_core11.protectedNoFault) && (cb_state.unprotected == true)) {
const LogObjectList objlist(cb_state.Handle(), tensor_state.Handle());
skip |= LogError(vuid, objlist, tensor_loc, "(%s) is a protected tensor, but command buffer (%s) is unprotected.%s",
FormatHandle(tensor_state).c_str(), FormatHandle(cb_state).c_str(), more_message);
}
return skip;
}
// Validates the buffer is allowed to be unprotected
bool CoreChecks::ValidateUnprotectedTensor(const vvl::CommandBuffer &cb_state, const vvl::Tensor &tensor_state,
const Location &tensor_loc, const char *vuid, const char *more_message) const {
// don't use on a protected tensor
assert(tensor_state.unprotected == true);
bool skip = false;
// if driver supports protectedNoFault the operation is valid, just has undefined values
if ((!phys_dev_props_core11.protectedNoFault) && (cb_state.unprotected == false)) {
const LogObjectList objlist(cb_state.Handle(), tensor_state.Handle());
skip |= LogError(vuid, objlist, tensor_loc, "(%s) is an unprotected tensor, but command buffer (%s) is protected.%s",
FormatHandle(tensor_state).c_str(), FormatHandle(cb_state).c_str(), more_message);
}
return skip;
}
bool CoreChecks::PreCallValidateGetDeviceTensorMemoryRequirementsARM(VkDevice device,
const VkDeviceTensorMemoryRequirementsARM *pInfo,
VkMemoryRequirements2 *pMemoryRequirements,
const ErrorObject &error_obj) const {
bool skip = false;
const Location loc = error_obj.location.dot(Field::tensors);
if (!enabled_features.tensors) {
skip |=
LogError("VUID-vkGetDeviceTensorMemoryRequirementsARM-tensors-09831", device, loc, "tensors feature is not enabled");
}
return skip;
}
|
