SPIRV-Tools/source/opt/module.cpp

287 lines
9.6 KiB
C++
Raw Normal View History

// Copyright (c) 2016 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
//
// 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 "source/opt/module.h"
#include <algorithm>
#include <cstring>
#include <ostream>
#include "source/operand.h"
#include "source/opt/ir_context.h"
#include "source/opt/reflect.h"
namespace spvtools {
namespace opt {
uint32_t Module::TakeNextIdBound() {
if (context()) {
if (id_bound() >= context()->max_id_bound()) {
return 0;
}
} else if (id_bound() >= kDefaultMaxIdBound) {
return 0;
}
return header_.bound++;
}
std::vector<Instruction*> Module::GetTypes() {
std::vector<Instruction*> type_insts;
for (auto& inst : types_values_) {
if (IsTypeInst(inst.opcode())) type_insts.push_back(&inst);
}
return type_insts;
}
std::vector<const Instruction*> Module::GetTypes() const {
std::vector<const Instruction*> type_insts;
for (auto& inst : types_values_) {
if (IsTypeInst(inst.opcode())) type_insts.push_back(&inst);
}
return type_insts;
}
std::vector<Instruction*> Module::GetConstants() {
std::vector<Instruction*> const_insts;
for (auto& inst : types_values_) {
if (IsConstantInst(inst.opcode())) const_insts.push_back(&inst);
}
return const_insts;
}
std::vector<const Instruction*> Module::GetConstants() const {
std::vector<const Instruction*> const_insts;
for (auto& inst : types_values_) {
if (IsConstantInst(inst.opcode())) const_insts.push_back(&inst);
}
return const_insts;
}
uint32_t Module::GetGlobalValue(spv::Op opcode) const {
for (auto& inst : types_values_) {
if (inst.opcode() == opcode) return inst.result_id();
}
return 0;
}
void Module::AddGlobalValue(spv::Op opcode, uint32_t result_id,
uint32_t type_id) {
std::unique_ptr<Instruction> newGlobal(
new Instruction(context(), opcode, type_id, result_id, {}));
AddGlobalValue(std::move(newGlobal));
}
void Module::ForEachInst(const std::function<void(Instruction*)>& f,
bool run_on_debug_line_insts) {
#define DELEGATE(list) list.ForEachInst(f, run_on_debug_line_insts)
DELEGATE(capabilities_);
DELEGATE(extensions_);
DELEGATE(ext_inst_imports_);
if (memory_model_) memory_model_->ForEachInst(f, run_on_debug_line_insts);
if (sampled_image_address_mode_)
sampled_image_address_mode_->ForEachInst(f, run_on_debug_line_insts);
DELEGATE(entry_points_);
DELEGATE(execution_modes_);
DELEGATE(debugs1_);
DELEGATE(debugs2_);
DELEGATE(debugs3_);
DELEGATE(ext_inst_debuginfo_);
DELEGATE(annotations_);
DELEGATE(types_values_);
for (auto& i : functions_) {
i->ForEachInst(f, run_on_debug_line_insts,
/* run_on_non_semantic_insts = */ true);
}
#undef DELEGATE
}
void Module::ForEachInst(const std::function<void(const Instruction*)>& f,
bool run_on_debug_line_insts) const {
#define DELEGATE(i) i.ForEachInst(f, run_on_debug_line_insts)
for (auto& i : capabilities_) DELEGATE(i);
for (auto& i : extensions_) DELEGATE(i);
for (auto& i : ext_inst_imports_) DELEGATE(i);
if (memory_model_)
static_cast<const Instruction*>(memory_model_.get())
->ForEachInst(f, run_on_debug_line_insts);
if (sampled_image_address_mode_)
static_cast<const Instruction*>(sampled_image_address_mode_.get())
->ForEachInst(f, run_on_debug_line_insts);
for (auto& i : entry_points_) DELEGATE(i);
for (auto& i : execution_modes_) DELEGATE(i);
for (auto& i : debugs1_) DELEGATE(i);
for (auto& i : debugs2_) DELEGATE(i);
for (auto& i : debugs3_) DELEGATE(i);
for (auto& i : annotations_) DELEGATE(i);
for (auto& i : types_values_) DELEGATE(i);
for (auto& i : ext_inst_debuginfo_) DELEGATE(i);
for (auto& i : functions_) {
static_cast<const Function*>(i.get())->ForEachInst(
f, run_on_debug_line_insts,
/* run_on_non_semantic_insts = */ true);
}
if (run_on_debug_line_insts) {
for (auto& i : trailing_dbg_line_info_) DELEGATE(i);
}
#undef DELEGATE
}
void Module::ToBinary(std::vector<uint32_t>* binary, bool skip_nop) const {
binary->push_back(header_.magic_number);
binary->push_back(header_.version);
// TODO(antiagainst): should we change the generator number?
binary->push_back(header_.generator);
binary->push_back(header_.bound);
binary->push_back(header_.schema);
size_t bound_idx = binary->size() - 2;
DebugScope last_scope(kNoDebugScope, kNoInlinedAt);
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
const Instruction* last_line_inst = nullptr;
bool between_merge_and_branch = false;
bool between_label_and_phi_var = false;
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
auto write_inst = [binary, skip_nop, &last_scope, &last_line_inst,
&between_merge_and_branch, &between_label_and_phi_var,
this](const Instruction* i) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
// Skip emitting line instructions between merge and branch instructions.
auto opcode = i->opcode();
if (between_merge_and_branch && i->IsLineInst()) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
return;
}
if (last_line_inst != nullptr) {
// If the current instruction is OpLine or DebugLine and it is the same
// as the last line instruction that is still effective (can be applied
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
// to the next instruction), we skip writing the current instruction.
if (i->IsLine()) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
uint32_t operand_index = 0;
if (last_line_inst->WhileEachInOperand(
[&operand_index, i](const uint32_t* word) {
assert(i->NumInOperandWords() > operand_index);
return *word == i->GetSingleWordInOperand(operand_index++);
})) {
return;
}
} else if (!i->IsNoLine() && i->dbg_line_insts().empty()) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
// If the current instruction does not have the line information,
// the last line information is not effective any more. Emit OpNoLine
// or DebugNoLine to specify it.
uint32_t shader_set_id = context()
->get_feature_mgr()
->GetExtInstImportId_Shader100DebugInfo();
if (shader_set_id != 0) {
binary->push_back((5 << 16) |
static_cast<uint16_t>(spv::Op::OpExtInst));
binary->push_back(context()->get_type_mgr()->GetVoidTypeId());
binary->push_back(context()->TakeNextId());
binary->push_back(shader_set_id);
binary->push_back(NonSemanticShaderDebugInfo100DebugNoLine);
} else {
binary->push_back((1 << 16) |
static_cast<uint16_t>(spv::Op::OpNoLine));
}
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
last_line_inst = nullptr;
}
}
if (opcode == spv::Op::OpLabel) {
between_label_and_phi_var = true;
} else if (opcode != spv::Op::OpVariable && opcode != spv::Op::OpPhi &&
!spvtools::opt::IsOpLineInst(opcode)) {
between_label_and_phi_var = false;
}
if (!(skip_nop && i->IsNop())) {
const auto& scope = i->GetDebugScope();
if (scope != last_scope && !between_merge_and_branch) {
// Can only emit nonsemantic instructions after all phi instructions
// in a block so don't emit scope instructions before phi instructions
// for NonSemantic.Shader.DebugInfo.100.
if (!between_label_and_phi_var ||
context()
->get_feature_mgr()
->GetExtInstImportId_OpenCL100DebugInfo()) {
// Emit DebugScope |scope| to |binary|.
auto dbg_inst = ext_inst_debuginfo_.begin();
scope.ToBinary(dbg_inst->type_id(), context()->TakeNextId(),
dbg_inst->GetSingleWordOperand(2), binary);
}
last_scope = scope;
}
i->ToBinaryWithoutAttachedDebugInsts(binary);
}
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
// Update the last line instruction.
between_merge_and_branch = false;
if (spvOpcodeIsBlockTerminator(opcode) || i->IsNoLine()) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
last_line_inst = nullptr;
} else if (opcode == spv::Op::OpLoopMerge ||
opcode == spv::Op::OpSelectionMerge) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
between_merge_and_branch = true;
last_line_inst = nullptr;
} else if (i->IsLine()) {
Propagate OpLine to all applied instructions in spirv-opt (#3951) Based on the OpLine spec, an OpLine instruction must be applied to the instructions physically following it up to the first occurrence of the next end of block, the next OpLine instruction, or the next OpNoLine instruction. ``` OpLine %file 0 0 OpNoLine OpLine %file 1 1 OpStore %foo %int_1 %value = OpLoad %int %foo OpLine %file 2 2 ``` For the above code, the current spirv-opt keeps three line instructions `OpLine %file 0 0`, `OpNoLine`, and `OpLine %file 1 1` in `std::vector<Instruction> dbg_line_insts_` of Instruction class for `OpStore %foo %int_1`. It does not put any line instruction to `std::vector<Instruction> dbg_line_insts_` of `%value = OpLoad %int %foo` even though `OpLine %file 1 1` must be applied to `%value = OpLoad %int %foo` based on the spec. This results in the missing line information for `%value = OpLoad %int %foo` while each spirv-opt pass optimizes the code. We have to put `OpLine %file 1 1` to `std::vector<Instruction> dbg_line_insts_` of both `%value = OpLoad %int %foo` and `OpStore %foo %int_1`. This commit conducts the line instruction propagation and skips emitting the eliminated line instructions at the end, which are the same with PropagateLineInfoPass and RedundantLineInfoElimPass. This commit removes PropagateLineInfoPass and RedundantLineInfoElimPass. KhronosGroup/glslang#2440 is a related PR that stop using PropagateLineInfoPass and RedundantLineInfoElimPass from glslang. When the code in this PR applied, the glslang tests will pass.
2020-10-29 17:06:30 +00:00
last_line_inst = i;
}
};
ForEachInst(write_inst, true);
// We create new instructions for DebugScope and DebugNoLine. The bound must
// be updated.
binary->data()[bound_idx] = header_.bound;
}
uint32_t Module::ComputeIdBound() const {
uint32_t highest = 0;
2017-05-17 18:12:25 +00:00
ForEachInst(
[&highest](const Instruction* inst) {
for (const auto& operand : *inst) {
if (spvIsIdType(operand.type)) {
highest = std::max(highest, operand.words[0]);
}
}
},
true /* scan debug line insts as well */);
return highest + 1;
}
bool Module::HasExplicitCapability(uint32_t cap) {
for (auto& ci : capabilities_) {
uint32_t tcap = ci.GetSingleWordOperand(0);
if (tcap == cap) {
return true;
}
}
return false;
}
uint32_t Module::GetExtInstImportId(const char* extstr) {
for (auto& ei : ext_inst_imports_)
if (!ei.GetInOperand(0).AsString().compare(extstr)) return ei.result_id();
return 0;
}
std::ostream& operator<<(std::ostream& str, const Module& module) {
module.ForEachInst([&str](const Instruction* inst) {
str << *inst;
if (inst->opcode() != spv::Op::OpFunctionEnd) {
str << std::endl;
}
});
return str;
}
} // namespace opt
} // namespace spvtools