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
|
/* Copyright (c) 2024-2025 LunarG, 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
*
* 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 "gpuav/core/gpuav.h"
#include "gpuav/resources/gpuav_shader_resources.h"
#include "gpuav/resources/gpuav_state_trackers.h"
namespace gpuav {
// Returns the number of bytes to hold 32 bit aligned array of bits.
static uint32_t BitBufferSize(uint32_t num_bits) {
static constexpr uint32_t kBitsPerWord = 32;
return (((num_bits + (kBitsPerWord - 1)) & ~(kBitsPerWord - 1)) / kBitsPerWord) * sizeof(uint32_t);
}
DescriptorHeap::DescriptorHeap(Validator& gpuav, uint32_t max_descriptors) : max_descriptors_(max_descriptors), buffer_(gpuav) {
// If max_descriptors_ is 0, GPU-AV aborted during vkCreateDevice(). We still need to
// support calls into this class as no-ops if this happens.
if (max_descriptors_ == 0) {
return;
}
VkBufferCreateInfo buffer_info = vku::InitStruct<VkBufferCreateInfo>();
buffer_info.size = BitBufferSize(max_descriptors_ + 1); // add extra entry since 0 is the invalid id.
buffer_info.usage = VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | VK_BUFFER_USAGE_SHADER_DEVICE_ADDRESS_BIT;
VmaAllocationCreateInfo alloc_info{};
alloc_info.requiredFlags = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT;
const bool success = buffer_.Create(&buffer_info, &alloc_info);
if (!success) {
return;
}
gpu_heap_state_ = (uint32_t*)buffer_.GetMappedPtr();
memset(gpu_heap_state_, 0, static_cast<size_t>(buffer_info.size));
}
DescriptorHeap::~DescriptorHeap() {
if (max_descriptors_ > 0) {
buffer_.Destroy();
gpu_heap_state_ = nullptr;
}
}
DescriptorId DescriptorHeap::NextId(const VulkanTypedHandle& handle) {
if (max_descriptors_ == 0) {
return 0;
}
DescriptorId result;
// NOTE: valid ids are in the range [1, max_descriptors_] (inclusive)
// 0 is the invalid id.
std::lock_guard guard(lock_);
if (alloc_map_.size() >= max_descriptors_) {
return 0;
}
do {
result = next_id_++;
if (next_id_ > max_descriptors_) {
next_id_ = 1;
}
} while (alloc_map_.count(result) > 0);
alloc_map_[result] = handle;
gpu_heap_state_[result / 32] |= 1u << (result & 31);
return result;
}
void DescriptorHeap::DeleteId(DescriptorId id) {
if (max_descriptors_ > 0) {
std::lock_guard guard(lock_);
// Note: We don't mess with next_id_ here because ids should be assigned in LRU order.
gpu_heap_state_[id / 32] &= ~(1u << (id & 31));
alloc_map_.erase(id);
}
}
struct DescriptorChecksCbState {
vko::BufferRange last_bound_desc_sets_state_ssbo;
};
void DescriptorChecksOnFinishDeviceSetup(Validator& gpuav) {
if (!gpuav.gpuav_settings.shader_instrumentation.descriptor_checks) {
gpuav.shared_resources_manager.GetOrCreate<DescriptorHeap>(gpuav, 0);
return;
}
VkPhysicalDeviceDescriptorIndexingProperties desc_indexing_props = vku::InitStructHelper();
VkPhysicalDeviceProperties2 props2 = vku::InitStructHelper(&desc_indexing_props);
DispatchGetPhysicalDeviceProperties2Helper(gpuav.api_version, gpuav.physical_device, &props2);
uint32_t num_descs = desc_indexing_props.maxUpdateAfterBindDescriptorsInAllPools;
if (num_descs == 0 || num_descs > glsl::kDebugInputBindlessMaxDescriptors) {
num_descs = glsl::kDebugInputBindlessMaxDescriptors;
}
gpuav.shared_resources_manager.GetOrCreate<DescriptorHeap>(gpuav, num_descs);
}
void RegisterDescriptorChecksValidation(Validator& gpuav, CommandBufferSubState& cb) {
if (!gpuav.gpuav_settings.shader_instrumentation.descriptor_checks) {
return;
}
DescriptorSetBindings& desc_set_bindings = cb.shared_resources_cache.GetOrCreate<DescriptorSetBindings>();
desc_set_bindings.on_update_bound_descriptor_sets.emplace_back(
[](Validator& gpuav, CommandBufferSubState& cb, DescriptorSetBindings::BindingCommand& desc_binding_cmd) {
DescriptorChecksCbState& dc_cb_state = cb.shared_resources_cache.GetOrCreate<DescriptorChecksCbState>();
dc_cb_state.last_bound_desc_sets_state_ssbo =
cb.gpu_resources_manager.GetHostVisibleBufferRange(sizeof(glsl::BoundDescriptorSetsStateSSBO));
dc_cb_state.last_bound_desc_sets_state_ssbo.Clear();
auto desc_state_ssbo =
static_cast<glsl::BoundDescriptorSetsStateSSBO*>(dc_cb_state.last_bound_desc_sets_state_ssbo.offset_mapped_ptr);
desc_state_ssbo->descriptor_init_status = gpuav.shared_resources_manager.Get<DescriptorHeap>().GetDeviceAddress();
for (size_t bound_ds_i = 0; bound_ds_i < desc_binding_cmd.bound_descriptor_sets.size(); ++bound_ds_i) {
auto& bound_ds = desc_binding_cmd.bound_descriptor_sets[bound_ds_i];
// Account for gaps in descriptor sets bindings
if (!bound_ds) {
continue;
}
if (bound_ds->IsUpdateAfterBind()) {
continue;
}
desc_state_ssbo->descriptor_set_types[bound_ds_i] = SubState(*bound_ds).GetTypeAddress(gpuav);
}
desc_binding_cmd.descritpor_state_ssbo = dc_cb_state.last_bound_desc_sets_state_ssbo;
});
cb.on_instrumentation_desc_set_update_functions.emplace_back(
[dummy_buffer_range = vko::BufferRange{}](CommandBufferSubState& cb, VkPipelineBindPoint,
VkDescriptorBufferInfo& out_buffer_info, uint32_t& out_dst_binding) mutable {
DescriptorChecksCbState* dc_cb_state = cb.shared_resources_cache.TryGet<DescriptorChecksCbState>();
if (dc_cb_state) {
out_buffer_info.buffer = dc_cb_state->last_bound_desc_sets_state_ssbo.buffer;
out_buffer_info.offset = dc_cb_state->last_bound_desc_sets_state_ssbo.offset;
out_buffer_info.range = dc_cb_state->last_bound_desc_sets_state_ssbo.size;
} else {
// Eventually, no descriptor set was bound in command buffer.
// Instrumenation descriptor set is already defined at this point and needs a binding,
// so just provide a dummy buffer
if (dummy_buffer_range.buffer == VK_NULL_HANDLE) {
dummy_buffer_range = cb.gpu_resources_manager.GetDeviceLocalBufferRange(64);
}
out_buffer_info.buffer = dummy_buffer_range.buffer;
out_buffer_info.offset = dummy_buffer_range.offset;
out_buffer_info.range = dummy_buffer_range.size;
}
out_dst_binding = glsl::kBindingInstDescriptorIndexingOOB;
});
// For every descriptor binding command, update a GPU buffer holding the type of each bound descriptor set
cb.on_pre_cb_submission_functions.emplace_back([](Validator& gpuav, CommandBufferSubState& cb, VkCommandBuffer) {
DescriptorSetBindings& desc_set_bindings = cb.shared_resources_cache.Get<DescriptorSetBindings>();
for (DescriptorSetBindings::BindingCommand& desc_binding_cmd : desc_set_bindings.descriptor_set_binding_commands) {
auto desc_state_ssbo_ptr =
static_cast<glsl::BoundDescriptorSetsStateSSBO*>(desc_binding_cmd.descritpor_state_ssbo.offset_mapped_ptr);
for (size_t bound_ds_i = 0; bound_ds_i < desc_binding_cmd.bound_descriptor_sets.size(); ++bound_ds_i) {
auto& bound_ds = desc_binding_cmd.bound_descriptor_sets[bound_ds_i];
// Account for gaps in descriptor sets bindings
if (!bound_ds) {
continue;
}
DescriptorSetSubState& desc_set_state = SubState(*bound_ds);
desc_state_ssbo_ptr->descriptor_set_types[bound_ds_i] = desc_set_state.GetTypeAddress(gpuav);
}
}
});
}
} // namespace gpuav
|
