SPIRV-Tools/source/val/validate.cpp

518 lines
21 KiB
C++
Raw Normal View History

2016-01-07 18:44:22 +00:00
// Copyright (c) 2015-2016 The Khronos Group 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/val/validate.h"
Support SPIR-V 1.4 (#2550) * SPIR-V 1.4 headers, add SPV_ENV_UNIVERSAL_1_4 * Support --target-env spv1.4 in help for command line tools * Support asm/dis of UniformId decoration * Validate UniformId decoration * Fix version check on instructions and operands Also register decorations used with OpDecorateId * Extension lists can differ between enums that match Example: SubgroupMaskEq vs SubgroupMaskEqKHR * Validate scope value for Uniform decoration, for SPIR-V 1.4 * More unioning of exts * Preserve grammar order within an enum value * 1.4: Validate OpSelect over composites * Tools default to 1.4 * Add asm/dis test for OpCopyLogical * 1.4: asm/dis tests for PtrEqual, PtrNotEqual, PtrDiff * Basic asm/Dis test for OpCopyMemory * Test asm/dis OpCopyMemory with 2-memory access Add asm/dis tests for OpCopyMemorySized Requires grammar update to add second optional memory access operand to OpCopyMemory and OpCopyMemorySized * Validate one or two memory accesses on OpCopyMemory* * Check av/vis on CopyMemory source and target memory access This is a proposed rule. See https://gitlab.khronos.org/spirv/SPIR-V/issues/413 * Validate operation for OpSpecConstantOp * Validate NonWritable decoration Also permit NonWritable on members of UBO and SSBO. * SPIR-V 1.4: NonWrtiable can decorate Function and Private vars * Update optimizer CLI tests for SPIR-V 1.4 * Testing tools: Give expected SPIR-V version in message * SPIR-V 1.4 validation for entry point interfaces * Allow only unique interfaces * Allow all global variables * Check that all statically used global variables are listed * new tests * Add validation fixture CompileFailure * Add 1.4 validation for pointer comparisons * New tests * Validate with image operands SignExtend, ZeroExtend Since we don't actually know the image texel format, we can't fully validate. We need more context. But we can make sure we allow the new image operands in known-good cases. * Validate OpCopyLogical * Recursively checks subtypes * new tests * Add SPIR-V 1.4 tests for NoSignedWrap, NoUnsignedWrap * Allow scalar conditions in 1.4 with OpSelect * Allows scalar conditions with vector operands * new tests * Validate uniform id scope as an execution scope * Validate the values of memory and execution scopes are valid scope values * new test * Remove SPIR-V 1.4 Vulkan 1.0 environment * SPIR-V 1.4 requires Vulkan 1.1 * FIX: include string for spvLog * FIX: validate nonwritable * FIX: test case suite for member decorate string * FIX: test case for hlsl functionality1 * Validation test fixture: ease debugging * Use binary version for SPIR-V 1.4 specific features * Switch checks based on the SPIR-V version from the target environment to instead use the version from the binary * Moved header parsing into the ValidationState_t constructor (where version based features are set) * Added new versions of tests that assemble a 1.3 binary and validate a 1.4 environment * Fix test for update to SPIR-V 1.4 headers * Fix formatting * Ext inst lookup: Add Vulkan 1.1 env with SPIR-V 1.4 * Update spirv-val help * Operand version checks should use module version Use the module version instead of the target environment version. * Fix comment about two-access form of OpCopyMemory
2019-05-07 16:27:18 +00:00
#include <algorithm>
2016-03-16 21:20:02 +00:00
#include <cassert>
#include <cstdio>
#include <functional>
#include <iterator>
#include <memory>
2016-03-16 21:20:02 +00:00
#include <sstream>
#include <string>
#include <vector>
#include "source/binary.h"
#include "source/diagnostic.h"
#include "source/enum_string_mapping.h"
#include "source/extensions.h"
#include "source/instruction.h"
#include "source/opcode.h"
#include "source/operand.h"
#include "source/spirv_constant.h"
#include "source/spirv_endian.h"
#include "source/spirv_target_env.h"
#include "source/spirv_validator_options.h"
#include "source/val/construct.h"
#include "source/val/function.h"
#include "source/val/instruction.h"
#include "source/val/validation_state.h"
#include "spirv-tools/libspirv.h"
namespace {
// TODO(issue 1950): The validator only returns a single message anyway, so no
// point in generating more than 1 warning.
static uint32_t kDefaultMaxNumOfWarnings = 1;
} // namespace
namespace spvtools {
namespace val {
namespace {
// Parses OpExtension instruction and registers extension.
void RegisterExtension(ValidationState_t& _,
const spv_parsed_instruction_t* inst) {
const std::string extension_str = spvtools::GetExtensionString(inst);
Extension extension;
if (!GetExtensionFromString(extension_str.c_str(), &extension)) {
// The error will be logged in the ProcessInstruction pass.
return;
}
_.RegisterExtension(extension);
}
// Parses the beginning of the module searching for OpExtension instructions.
// Registers extensions if recognized. Returns SPV_REQUESTED_TERMINATION
// once an instruction which is not SpvOpCapability and SpvOpExtension is
// encountered. According to the SPIR-V spec extensions are declared after
// capabilities and before everything else.
spv_result_t ProcessExtensions(void* user_data,
const spv_parsed_instruction_t* inst) {
const SpvOp opcode = static_cast<SpvOp>(inst->opcode);
if (opcode == SpvOpCapability) return SPV_SUCCESS;
if (opcode == SpvOpExtension) {
ValidationState_t& _ = *(reinterpret_cast<ValidationState_t*>(user_data));
RegisterExtension(_, inst);
return SPV_SUCCESS;
}
// OpExtension block is finished, requesting termination.
return SPV_REQUESTED_TERMINATION;
}
spv_result_t ProcessInstruction(void* user_data,
const spv_parsed_instruction_t* inst) {
Basic SSA Validation Most uses of an ID must occur after the definition of the ID. Forward references are allowed for things like OpName, OpDecorate, and various cases of control-flow instructions such as OpBranch, OpPhi, and OpFunctionCall. TODO: Use CFG analysis for SSA checks. In particular, an ID defined inside a function body is only usable inside that function body. Also, use dominator info to catch some failing cases. Also: * Validator test cases use (standard) assignment form. * Update style to more closely follow the Google C++ style guide * Remove color-diagnostics flag. This is enabled by default on terminals with color. Prints hidden ASCII for terminals that can't handle color(Emacs) * Pass functors to SSAPass to check if the operand can be forward referenced based on its index value * Return SPV_ERROR_INVALID_ID for ID related errors spvBinaryParse returned SPV_ERROR_INVALID_BINARY for all types of errors. Since spvBinaryParse does some ID validation, this was returning inappropriate error codes for some tests. * Common fixture for validation tests. It only runs certian validation passes. * Add a SPV_VALIDATE_SSA_BIT for testing purposes * Fixtures now return error codes * Add OpName support in diag message and unit tests * Binary parsing can fail with invalid ID or invalid binary error code Tests include: * OpDecorate * OpName * OpMemberName * OpBranchConditional * OpSelectionMerge * OpMemberDecorate * OpGroupDecorate * OpDeviceEnqueue * Enable several tests failing in ID validation.
2015-11-18 20:43:43 +00:00
ValidationState_t& _ = *(reinterpret_cast<ValidationState_t*>(user_data));
auto* instruction = _.AddOrderedInstruction(inst);
_.RegisterDebugInstruction(instruction);
2016-01-05 19:56:02 +00:00
return SPV_SUCCESS;
Basic SSA Validation Most uses of an ID must occur after the definition of the ID. Forward references are allowed for things like OpName, OpDecorate, and various cases of control-flow instructions such as OpBranch, OpPhi, and OpFunctionCall. TODO: Use CFG analysis for SSA checks. In particular, an ID defined inside a function body is only usable inside that function body. Also, use dominator info to catch some failing cases. Also: * Validator test cases use (standard) assignment form. * Update style to more closely follow the Google C++ style guide * Remove color-diagnostics flag. This is enabled by default on terminals with color. Prints hidden ASCII for terminals that can't handle color(Emacs) * Pass functors to SSAPass to check if the operand can be forward referenced based on its index value * Return SPV_ERROR_INVALID_ID for ID related errors spvBinaryParse returned SPV_ERROR_INVALID_BINARY for all types of errors. Since spvBinaryParse does some ID validation, this was returning inappropriate error codes for some tests. * Common fixture for validation tests. It only runs certian validation passes. * Add a SPV_VALIDATE_SSA_BIT for testing purposes * Fixtures now return error codes * Add OpName support in diag message and unit tests * Binary parsing can fail with invalid ID or invalid binary error code Tests include: * OpDecorate * OpName * OpMemberName * OpBranchConditional * OpSelectionMerge * OpMemberDecorate * OpGroupDecorate * OpDeviceEnqueue * Enable several tests failing in ID validation.
2015-11-18 20:43:43 +00:00
}
spv_result_t ValidateForwardDecls(ValidationState_t& _) {
if (_.unresolved_forward_id_count() == 0) return SPV_SUCCESS;
std::stringstream ss;
std::vector<uint32_t> ids = _.UnresolvedForwardIds();
std::transform(
std::begin(ids), std::end(ids),
std::ostream_iterator<std::string>(ss, " "),
bind(&ValidationState_t::getIdName, std::ref(_), std::placeholders::_1));
auto id_str = ss.str();
return _.diag(SPV_ERROR_INVALID_ID, nullptr)
<< "The following forward referenced IDs have not been defined:\n"
<< id_str.substr(0, id_str.size() - 1);
}
std::vector<std::string> CalculateNamesForEntryPoint(ValidationState_t& _,
const uint32_t id) {
auto id_descriptions = _.entry_point_descriptions(id);
auto id_names = std::vector<std::string>();
id_names.reserve((id_descriptions.size()));
for (auto description : id_descriptions) id_names.push_back(description.name);
return id_names;
}
spv_result_t ValidateEntryPointNameUnique(ValidationState_t& _,
const uint32_t id) {
auto id_names = CalculateNamesForEntryPoint(_, id);
const auto names =
std::unordered_set<std::string>(id_names.begin(), id_names.end());
if (id_names.size() != names.size()) {
std::sort(id_names.begin(), id_names.end());
for (size_t i = 0; i < id_names.size() - 1; i++) {
if (id_names[i] == id_names[i + 1]) {
return _.diag(SPV_ERROR_INVALID_BINARY, _.FindDef(id))
<< "Entry point name \"" << id_names[i]
<< "\" is not unique, which is not allow in WebGPU env.";
}
}
}
for (const auto other_id : _.entry_points()) {
if (other_id == id) continue;
const auto other_id_names = CalculateNamesForEntryPoint(_, other_id);
for (const auto& other_id_name : other_id_names) {
if (names.find(other_id_name) != names.end()) {
return _.diag(SPV_ERROR_INVALID_BINARY, _.FindDef(id))
<< "Entry point name \"" << other_id_name
<< "\" is not unique, which is not allow in WebGPU env.";
}
}
}
return SPV_SUCCESS;
}
spv_result_t ValidateEntryPointNamesUnique(ValidationState_t& _) {
for (const auto id : _.entry_points()) {
auto result = ValidateEntryPointNameUnique(_, id);
if (result != SPV_SUCCESS) return result;
}
return SPV_SUCCESS;
}
// Entry point validation. Based on 2.16.1 (Universal Validation Rules) of the
// SPIRV spec:
// * There is at least one OpEntryPoint instruction, unless the Linkage
// capability is being used.
// * No function can be targeted by both an OpEntryPoint instruction and an
// OpFunctionCall instruction.
//
// Additionally enforces that entry points for Vulkan and WebGPU should not have
// recursion. And that entry names should be unique for WebGPU.
spv_result_t ValidateEntryPoints(ValidationState_t& _) {
_.ComputeFunctionToEntryPointMapping();
_.ComputeRecursiveEntryPoints();
if (_.entry_points().empty() && !_.HasCapability(SpvCapabilityLinkage)) {
return _.diag(SPV_ERROR_INVALID_BINARY, nullptr)
<< "No OpEntryPoint instruction was found. This is only allowed if "
"the Linkage capability is being used.";
}
for (const auto& entry_point : _.entry_points()) {
if (_.IsFunctionCallTarget(entry_point)) {
return _.diag(SPV_ERROR_INVALID_BINARY, _.FindDef(entry_point))
<< "A function (" << entry_point
<< ") may not be targeted by both an OpEntryPoint instruction and "
"an OpFunctionCall instruction.";
}
// For Vulkan and WebGPU, the static function-call graph for an entry point
// must not contain cycles.
if (spvIsVulkanOrWebGPUEnv(_.context()->target_env)) {
if (_.recursive_entry_points().find(entry_point) !=
_.recursive_entry_points().end()) {
return _.diag(SPV_ERROR_INVALID_BINARY, _.FindDef(entry_point))
<< "Entry points may not have a call graph with cycles.";
}
}
// For WebGPU all entry point names must be unique.
if (spvIsWebGPUEnv(_.context()->target_env)) {
const auto result = ValidateEntryPointNamesUnique(_);
if (result != SPV_SUCCESS) return result;
}
}
return SPV_SUCCESS;
}
spv_result_t ValidateBinaryUsingContextAndValidationState(
const spv_context_t& context, const uint32_t* words, const size_t num_words,
spv_diagnostic* pDiagnostic, ValidationState_t* vstate) {
auto binary = std::unique_ptr<spv_const_binary_t>(
new spv_const_binary_t{words, num_words});
spv_endianness_t endian;
spv_position_t position = {};
if (spvBinaryEndianness(binary.get(), &endian)) {
return DiagnosticStream(position, context.consumer, "",
SPV_ERROR_INVALID_BINARY)
<< "Invalid SPIR-V magic number.";
}
if (spvIsWebGPUEnv(context.target_env) && endian != SPV_ENDIANNESS_LITTLE) {
return DiagnosticStream(position, context.consumer, "",
SPV_ERROR_INVALID_BINARY)
<< "WebGPU requires SPIR-V to be little endian.";
}
spv_header_t header;
if (spvBinaryHeaderGet(binary.get(), endian, &header)) {
return DiagnosticStream(position, context.consumer, "",
SPV_ERROR_INVALID_BINARY)
<< "Invalid SPIR-V header.";
}
if (header.version > spvVersionForTargetEnv(context.target_env)) {
return DiagnosticStream(position, context.consumer, "",
SPV_ERROR_WRONG_VERSION)
<< "Invalid SPIR-V binary version "
<< SPV_SPIRV_VERSION_MAJOR_PART(header.version) << "."
<< SPV_SPIRV_VERSION_MINOR_PART(header.version)
<< " for target environment "
<< spvTargetEnvDescription(context.target_env) << ".";
}
if (header.bound > vstate->options()->universal_limits_.max_id_bound) {
return DiagnosticStream(position, context.consumer, "",
SPV_ERROR_INVALID_BINARY)
<< "Invalid SPIR-V. The id bound is larger than the max id bound "
<< vstate->options()->universal_limits_.max_id_bound << ".";
}
// Look for OpExtension instructions and register extensions.
// This parse should not produce any error messages. Hijack the context and
// replace the message consumer so that we do not pollute any state in input
// consumer.
spv_context_t hijacked_context = context;
hijacked_context.consumer = [](spv_message_level_t, const char*,
const spv_position_t&, const char*) {};
spvBinaryParse(&hijacked_context, vstate, words, num_words,
/* parsed_header = */ nullptr, ProcessExtensions,
/* diagnostic = */ nullptr);
// Parse the module and perform inline validation checks. These checks do
// not require the the knowledge of the whole module.
Support SPIR-V 1.4 (#2550) * SPIR-V 1.4 headers, add SPV_ENV_UNIVERSAL_1_4 * Support --target-env spv1.4 in help for command line tools * Support asm/dis of UniformId decoration * Validate UniformId decoration * Fix version check on instructions and operands Also register decorations used with OpDecorateId * Extension lists can differ between enums that match Example: SubgroupMaskEq vs SubgroupMaskEqKHR * Validate scope value for Uniform decoration, for SPIR-V 1.4 * More unioning of exts * Preserve grammar order within an enum value * 1.4: Validate OpSelect over composites * Tools default to 1.4 * Add asm/dis test for OpCopyLogical * 1.4: asm/dis tests for PtrEqual, PtrNotEqual, PtrDiff * Basic asm/Dis test for OpCopyMemory * Test asm/dis OpCopyMemory with 2-memory access Add asm/dis tests for OpCopyMemorySized Requires grammar update to add second optional memory access operand to OpCopyMemory and OpCopyMemorySized * Validate one or two memory accesses on OpCopyMemory* * Check av/vis on CopyMemory source and target memory access This is a proposed rule. See https://gitlab.khronos.org/spirv/SPIR-V/issues/413 * Validate operation for OpSpecConstantOp * Validate NonWritable decoration Also permit NonWritable on members of UBO and SSBO. * SPIR-V 1.4: NonWrtiable can decorate Function and Private vars * Update optimizer CLI tests for SPIR-V 1.4 * Testing tools: Give expected SPIR-V version in message * SPIR-V 1.4 validation for entry point interfaces * Allow only unique interfaces * Allow all global variables * Check that all statically used global variables are listed * new tests * Add validation fixture CompileFailure * Add 1.4 validation for pointer comparisons * New tests * Validate with image operands SignExtend, ZeroExtend Since we don't actually know the image texel format, we can't fully validate. We need more context. But we can make sure we allow the new image operands in known-good cases. * Validate OpCopyLogical * Recursively checks subtypes * new tests * Add SPIR-V 1.4 tests for NoSignedWrap, NoUnsignedWrap * Allow scalar conditions in 1.4 with OpSelect * Allows scalar conditions with vector operands * new tests * Validate uniform id scope as an execution scope * Validate the values of memory and execution scopes are valid scope values * new test * Remove SPIR-V 1.4 Vulkan 1.0 environment * SPIR-V 1.4 requires Vulkan 1.1 * FIX: include string for spvLog * FIX: validate nonwritable * FIX: test case suite for member decorate string * FIX: test case for hlsl functionality1 * Validation test fixture: ease debugging * Use binary version for SPIR-V 1.4 specific features * Switch checks based on the SPIR-V version from the target environment to instead use the version from the binary * Moved header parsing into the ValidationState_t constructor (where version based features are set) * Added new versions of tests that assemble a 1.3 binary and validate a 1.4 environment * Fix test for update to SPIR-V 1.4 headers * Fix formatting * Ext inst lookup: Add Vulkan 1.1 env with SPIR-V 1.4 * Update spirv-val help * Operand version checks should use module version Use the module version instead of the target environment version. * Fix comment about two-access form of OpCopyMemory
2019-05-07 16:27:18 +00:00
if (auto error = spvBinaryParse(&context, vstate, words, num_words,
/*parsed_header =*/nullptr,
ProcessInstruction, pDiagnostic)) {
2016-09-02 13:51:44 +00:00
return error;
}
Basic SSA Validation Most uses of an ID must occur after the definition of the ID. Forward references are allowed for things like OpName, OpDecorate, and various cases of control-flow instructions such as OpBranch, OpPhi, and OpFunctionCall. TODO: Use CFG analysis for SSA checks. In particular, an ID defined inside a function body is only usable inside that function body. Also, use dominator info to catch some failing cases. Also: * Validator test cases use (standard) assignment form. * Update style to more closely follow the Google C++ style guide * Remove color-diagnostics flag. This is enabled by default on terminals with color. Prints hidden ASCII for terminals that can't handle color(Emacs) * Pass functors to SSAPass to check if the operand can be forward referenced based on its index value * Return SPV_ERROR_INVALID_ID for ID related errors spvBinaryParse returned SPV_ERROR_INVALID_BINARY for all types of errors. Since spvBinaryParse does some ID validation, this was returning inappropriate error codes for some tests. * Common fixture for validation tests. It only runs certian validation passes. * Add a SPV_VALIDATE_SSA_BIT for testing purposes * Fixtures now return error codes * Add OpName support in diag message and unit tests * Binary parsing can fail with invalid ID or invalid binary error code Tests include: * OpDecorate * OpName * OpMemberName * OpBranchConditional * OpSelectionMerge * OpMemberDecorate * OpGroupDecorate * OpDeviceEnqueue * Enable several tests failing in ID validation.
2015-11-18 20:43:43 +00:00
std::vector<Instruction*> visited_entry_points;
for (auto& instruction : vstate->ordered_instructions()) {
{
// In order to do this work outside of Process Instruction we need to be
// able to, briefly, de-const the instruction.
Instruction* inst = const_cast<Instruction*>(&instruction);
if (inst->opcode() == SpvOpEntryPoint) {
const auto entry_point = inst->GetOperandAs<uint32_t>(1);
const auto execution_model = inst->GetOperandAs<SpvExecutionModel>(0);
const char* str = reinterpret_cast<const char*>(
inst->words().data() + inst->operand(2).offset);
ValidationState_t::EntryPointDescription desc;
desc.name = str;
std::vector<uint32_t> interfaces;
for (size_t j = 3; j < inst->operands().size(); ++j)
desc.interfaces.push_back(inst->word(inst->operand(j).offset));
vstate->RegisterEntryPoint(entry_point, execution_model,
std::move(desc));
if (visited_entry_points.size() > 0) {
for (const Instruction* check_inst : visited_entry_points) {
const auto check_execution_model =
check_inst->GetOperandAs<SpvExecutionModel>(0);
const char* check_str = reinterpret_cast<const char*>(
check_inst->words().data() + inst->operand(2).offset);
const std::string check_name(check_str);
if (desc.name == check_name &&
execution_model == check_execution_model) {
return vstate->diag(SPV_ERROR_INVALID_DATA, inst)
<< "2 Entry points cannot share the same name and "
"ExecutionMode.";
}
}
}
visited_entry_points.push_back(inst);
}
if (inst->opcode() == SpvOpFunctionCall) {
if (!vstate->in_function_body()) {
return vstate->diag(SPV_ERROR_INVALID_LAYOUT, &instruction)
<< "A FunctionCall must happen within a function body.";
}
const auto called_id = inst->GetOperandAs<uint32_t>(2);
if (spvIsWebGPUEnv(context.target_env) &&
!vstate->IsFunctionCallDefined(called_id)) {
return vstate->diag(SPV_ERROR_INVALID_LAYOUT, &instruction)
<< "For WebGPU, functions need to be defined before being "
"called.";
}
vstate->AddFunctionCallTarget(called_id);
}
if (vstate->in_function_body()) {
inst->set_function(&(vstate->current_function()));
inst->set_block(vstate->current_function().current_block());
if (vstate->in_block() && spvOpcodeIsBlockTerminator(inst->opcode())) {
vstate->current_function().current_block()->set_terminator(inst);
}
}
if (auto error = IdPass(*vstate, inst)) return error;
}
if (auto error = CapabilityPass(*vstate, &instruction)) return error;
if (auto error = ModuleLayoutPass(*vstate, &instruction)) return error;
if (auto error = CfgPass(*vstate, &instruction)) return error;
if (auto error = InstructionPass(*vstate, &instruction)) return error;
// Now that all of the checks are done, update the state.
{
Instruction* inst = const_cast<Instruction*>(&instruction);
vstate->RegisterInstruction(inst);
if (inst->opcode() == SpvOpTypeForwardPointer) {
vstate->RegisterForwardPointer(inst->GetOperandAs<uint32_t>(0));
}
}
}
if (!vstate->has_memory_model_specified())
return vstate->diag(SPV_ERROR_INVALID_LAYOUT, nullptr)
<< "Missing required OpMemoryModel instruction.";
if (vstate->in_function_body())
return vstate->diag(SPV_ERROR_INVALID_LAYOUT, nullptr)
<< "Missing OpFunctionEnd at end of module.";
// Catch undefined forward references before performing further checks.
if (auto error = ValidateForwardDecls(*vstate)) return error;
2016-03-16 21:20:02 +00:00
// Calculate reachability after all the blocks are parsed, but early that it
// can be relied on in subsequent pases.
ReachabilityPass(*vstate);
// ID usage needs be handled in its own iteration of the instructions,
// between the two others. It depends on the first loop to have been
// finished, so that all instructions have been registered. And the following
// loop depends on all of the usage data being populated. Thus it cannot live
// in either of those iterations.
// It should also live after the forward declaration check, since it will
// have problems with missing forward declarations, but give less useful error
// messages.
for (size_t i = 0; i < vstate->ordered_instructions().size(); ++i) {
auto& instruction = vstate->ordered_instructions()[i];
if (auto error = UpdateIdUse(*vstate, &instruction)) return error;
}
// Validate individual opcodes.
for (size_t i = 0; i < vstate->ordered_instructions().size(); ++i) {
auto& instruction = vstate->ordered_instructions()[i];
// Keep these passes in the order they appear in the SPIR-V specification
// sections to maintain test consistency.
if (auto error = MiscPass(*vstate, &instruction)) return error;
if (auto error = DebugPass(*vstate, &instruction)) return error;
if (auto error = AnnotationPass(*vstate, &instruction)) return error;
if (auto error = ExtensionPass(*vstate, &instruction)) return error;
if (auto error = ModeSettingPass(*vstate, &instruction)) return error;
if (auto error = TypePass(*vstate, &instruction)) return error;
if (auto error = ConstantPass(*vstate, &instruction)) return error;
if (auto error = MemoryPass(*vstate, &instruction)) return error;
if (auto error = FunctionPass(*vstate, &instruction)) return error;
if (auto error = ImagePass(*vstate, &instruction)) return error;
if (auto error = ConversionPass(*vstate, &instruction)) return error;
if (auto error = CompositesPass(*vstate, &instruction)) return error;
if (auto error = ArithmeticsPass(*vstate, &instruction)) return error;
if (auto error = BitwisePass(*vstate, &instruction)) return error;
if (auto error = LogicalsPass(*vstate, &instruction)) return error;
if (auto error = ControlFlowPass(*vstate, &instruction)) return error;
if (auto error = DerivativesPass(*vstate, &instruction)) return error;
if (auto error = AtomicsPass(*vstate, &instruction)) return error;
if (auto error = PrimitivesPass(*vstate, &instruction)) return error;
if (auto error = BarriersPass(*vstate, &instruction)) return error;
// Group
// Device-Side Enqueue
// Pipe
if (auto error = NonUniformPass(*vstate, &instruction)) return error;
if (auto error = LiteralsPass(*vstate, &instruction)) return error;
}
// Validate the preconditions involving adjacent instructions. e.g. SpvOpPhi
// must only be preceeded by SpvOpLabel, SpvOpPhi, or SpvOpLine.
if (auto error = ValidateAdjacency(*vstate)) return error;
if (auto error = ValidateEntryPoints(*vstate)) return error;
// CFG checks are performed after the binary has been parsed
// and the CFGPass has collected information about the control flow
if (auto error = PerformCfgChecks(*vstate)) return error;
if (auto error = CheckIdDefinitionDominateUse(*vstate)) return error;
if (auto error = ValidateDecorations(*vstate)) return error;
if (auto error = ValidateInterfaces(*vstate)) return error;
// TODO(dsinclair): Restructure ValidateBuiltins so we can move into the
// for() above as it loops over all ordered_instructions internally.
if (auto error = ValidateBuiltIns(*vstate)) return error;
// These checks must be performed after individual opcode checks because
// those checks register the limitation checked here.
for (const auto& inst : vstate->ordered_instructions()) {
if (auto error = ValidateExecutionLimitations(*vstate, &inst)) return error;
if (auto error = ValidateSmallTypeUses(*vstate, &inst)) return error;
}
return SPV_SUCCESS;
}
} // namespace
spv_result_t ValidateBinaryAndKeepValidationState(
const spv_const_context context, spv_const_validator_options options,
const uint32_t* words, const size_t num_words, spv_diagnostic* pDiagnostic,
std::unique_ptr<ValidationState_t>* vstate) {
spv_context_t hijack_context = *context;
if (pDiagnostic) {
*pDiagnostic = nullptr;
UseDiagnosticAsMessageConsumer(&hijack_context, pDiagnostic);
}
vstate->reset(new ValidationState_t(&hijack_context, options, words,
num_words, kDefaultMaxNumOfWarnings));
return ValidateBinaryUsingContextAndValidationState(
hijack_context, words, num_words, pDiagnostic, vstate->get());
}
} // namespace val
} // namespace spvtools
spv_result_t spvValidate(const spv_const_context context,
const spv_const_binary binary,
spv_diagnostic* pDiagnostic) {
return spvValidateBinary(context, binary->code, binary->wordCount,
pDiagnostic);
}
spv_result_t spvValidateBinary(const spv_const_context context,
const uint32_t* words, const size_t num_words,
spv_diagnostic* pDiagnostic) {
spv_context_t hijack_context = *context;
if (pDiagnostic) {
*pDiagnostic = nullptr;
spvtools::UseDiagnosticAsMessageConsumer(&hijack_context, pDiagnostic);
}
// This interface is used for default command line options.
spv_validator_options default_options = spvValidatorOptionsCreate();
// Create the ValidationState using the context and default options.
spvtools::val::ValidationState_t vstate(&hijack_context, default_options,
words, num_words,
kDefaultMaxNumOfWarnings);
spv_result_t result =
spvtools::val::ValidateBinaryUsingContextAndValidationState(
hijack_context, words, num_words, pDiagnostic, &vstate);
spvValidatorOptionsDestroy(default_options);
return result;
}
spv_result_t spvValidateWithOptions(const spv_const_context context,
spv_const_validator_options options,
const spv_const_binary binary,
spv_diagnostic* pDiagnostic) {
spv_context_t hijack_context = *context;
if (pDiagnostic) {
*pDiagnostic = nullptr;
spvtools::UseDiagnosticAsMessageConsumer(&hijack_context, pDiagnostic);
}
// Create the ValidationState using the context.
spvtools::val::ValidationState_t vstate(&hijack_context, options,
binary->code, binary->wordCount,
kDefaultMaxNumOfWarnings);
return spvtools::val::ValidateBinaryUsingContextAndValidationState(
hijack_context, binary->code, binary->wordCount, pDiagnostic, &vstate);
}