v8/test/cctest/compiler/code-assembler-tester.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

85 lines
2.9 KiB
C
Raw Normal View History

// Copyright 2016 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#ifndef V8_TEST_CCTEST_COMPILER_CODE_ASSEMBLER_TESTER_H_
#define V8_TEST_CCTEST_COMPILER_CODE_ASSEMBLER_TESTER_H_
#include "src/codegen/interface-descriptors.h"
#include "src/compiler/code-assembler.h"
#include "src/compiler/raw-machine-assembler.h"
#include "src/execution/isolate.h"
#include "src/handles/handles.h"
#include "test/cctest/compiler/function-tester.h"
namespace v8 {
namespace internal {
namespace compiler {
class CodeAssemblerTester {
public:
// Test generating code for a stub. Assumes VoidDescriptor call interface.
explicit CodeAssemblerTester(Isolate* isolate, const char* name = "test")
: zone_(isolate->allocator(), ZONE_NAME, kCompressGraphZone),
scope_(isolate),
state_(isolate, &zone_, VoidDescriptor{}, Code::STUB, name,
PoisoningMitigationLevel::kDontPoison) {}
// Test generating code for a JS function (e.g. builtins).
CodeAssemblerTester(Isolate* isolate, int parameter_count,
Code::Kind kind = Code::BUILTIN,
const char* name = "test")
: zone_(isolate->allocator(), ZONE_NAME, kCompressGraphZone),
scope_(isolate),
state_(isolate, &zone_, parameter_count, kind, name,
PoisoningMitigationLevel::kDontPoison) {}
CodeAssemblerTester(Isolate* isolate, Code::Kind kind,
const char* name = "test")
: zone_(isolate->allocator(), ZONE_NAME, kCompressGraphZone),
scope_(isolate),
state_(isolate, &zone_, 0, kind, name,
PoisoningMitigationLevel::kDontPoison) {}
Reland "[arm64] Preparation for padding of arguments" This is a reland of bcf1172992a2d817978ff19a2d814c7629443956 The test was timing out in no snapshot builds, as each CodeAssemblerTester creates a new Context. Reduced the random iterations significantly. Original change's description: > [arm64] Preparation for padding of arguments > > As part of JSSP removal, we need to align the arguments passed to functions > on the stack, by adding a padding slot when the total number of arguments > is odd. > > This patch introduces the kPadArguments flag (which is currently set to > false for all architectures), which will control padding of arguments in > architecture-independent parts of the code (deoptimizer, instruction > selector). > > It also adds some executable tests for tail calls with various stack > parameter counts on the caller and callee sides. > > This will be turned on for arm64 together with arm64-specific changes to > the code generator, the MacroAsembler and the builtins, in a later patch. > > Bug: v8:6644 > Change-Id: I79a5c149123fe8130cedd1ccffec3d9b50361e08 > Reviewed-on: https://chromium-review.googlesource.com/806554 > Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com> > Reviewed-by: Jaroslav Sevcik <jarin@chromium.org> > Reviewed-by: Ross McIlroy <rmcilroy@chromium.org> > Cr-Commit-Position: refs/heads/master@{#50134} TBR=jarin@chromium.org Bug: v8:6644 Change-Id: I795877ed9791e126ffac6841dbbb65189e95d207 Reviewed-on: https://chromium-review.googlesource.com/833046 Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org> Cr-Commit-Position: refs/heads/master@{#50238}
2017-12-18 18:55:23 +00:00
CodeAssemblerTester(Isolate* isolate, CallDescriptor* call_descriptor,
const char* name = "test")
: zone_(isolate->allocator(), ZONE_NAME, kCompressGraphZone),
scope_(isolate),
state_(isolate, &zone_, call_descriptor, Code::STUB, name,
PoisoningMitigationLevel::kDontPoison, Builtins::kNoBuiltinId) {}
CodeAssemblerState* state() { return &state_; }
// Direct low-level access to the machine assembler, for testing only.
RawMachineAssembler* raw_assembler_for_testing() {
return state_.raw_assembler_.get();
}
Handle<Code> GenerateCode() {
return GenerateCode(AssemblerOptions::Default(scope_.isolate()));
}
Handle<Code> GenerateCode(const AssemblerOptions& options) {
if (state_.InsideBlock()) {
CodeAssembler(&state_).Unreachable();
}
Profile-guided optimization of builtins Design doc: https://docs.google.com/document/d/1szInbXZfaErWW70d30hJsOLL0Es-l5_g8d2rXm1ZBqI/edit?usp=sharing V8 can already collect data about how many times each basic block in the builtins is run. This change enables using that data for profile-guided optimization. New comments in BUILD.gn describe how to use this feature. A few implementation details worth mentioning, which aren't covered in the design doc: - BasicBlockProfilerData currently contains an array of RPO numbers. However, this array is always just [0, 1, 2, 3, ...], so this change removes that array. A new DCHECK in BasicBlockInstrumentor::Instrument ensures that the removal is valid. - RPO numbers, while useful for printing data that matches with the stringified schedule, are not useful for matching profiling data with blocks that haven't been scheduled yet. This change adds a new array of block IDs in BasicBlockProfilerData, so that block counters can be used for PGO. - Basic block counters need to be written to a file so that they can be provided to a subsequent run of mksnapshot, but the design doc doesn't specify the transfer format or what file is used. In this change, I propose using the existing v8.log file for that purpose. Block count records look like this: block,TestLessThanHandler,37,29405 This line indicates that block ID 37 in TestLessThanHandler was run 29405 times. If multiple lines refer to the same block, the reader adds them all together. I like this format because it's easy to use: - V8 already has robust logic for creating the log file, naming it to avoid conflicts in multi-process situations, etc. - Line order doesn't matter, and interleaved writes from various logging sources are fine, given that V8 writes each line atomically. - Combining multiple sources of profiling data is as simple as concatenating their v8.log files together. - It is a good idea to avoid making any changes based on profiling data if the function being compiled doesn't match the one that was profiled, since it is common to use profiling data downloaded from a central lab which is updated only periodically. To check whether a function matches, I propose using a hash of the Graph state right before scheduling. This might be stricter than necessary, as some changes to the function might be small enough that the profile data is still relevant, but I'd rather err on the side of not making incorrect changes. This hash is also written to the v8.log file, in a line that looks like this: builtin_hash,LdaZeroHandler,3387822046 Bug: v8:10470 Change-Id: I429e5ce5efa94e01e7489deb3996012cf860cf13 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2220765 Commit-Queue: Seth Brenith <seth.brenith@microsoft.com> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org> Reviewed-by: Tobias Tebbi <tebbi@chromium.org> Cr-Commit-Position: refs/heads/master@{#69008}
2020-07-16 16:37:08 +00:00
return CodeAssembler::GenerateCode(&state_, options, nullptr);
}
Handle<Code> GenerateCodeCloseAndEscape() {
return scope_.CloseAndEscape(GenerateCode());
}
private:
Zone zone_;
HandleScope scope_;
LocalContext context_;
CodeAssemblerState state_;
};
} // namespace compiler
} // namespace internal
} // namespace v8
#endif // V8_TEST_CCTEST_COMPILER_CODE_ASSEMBLER_TESTER_H_