V8 is collecting a growing amount of fuzzers, all of which take substantial
space on the bots and in chromium build archives. This CL improves that
situation by allowing component (shared library) builds for almost all fuzzers.
The parser fuzzer is handled as an exception since it would require exporting a
large number of additional functions.
A component build results in about a 50-100x improvement in file size for each
fuzzer (~50M-100M to around 1.1M).
BUG=chromium:648864
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_chromium_compile_dbg_ng;master.tryserver.chromium.android:android_clang_dbg_recipe
Review-Url: https://codereview.chromium.org/2360983002
Cr-Commit-Position: refs/heads/master@{#39709}
The file formerly known as src/js/harmony-async-await.js does not
expose anything directly to JavaScript code; it just makes a few
functions available on the native context for the parser desugaring
to expand into. Experimental natives have various issues with
larger amounts of code, so this patch moves the JS builtins to
support async functions out of experimental natives and into the
core snapshot. The change would be done eventually anyway, but this
patch does it before removing the flag to support shipping the
feature while avoiding the pitfalls of experimental natives.
Drive by cleanup: remove more unused functions from the whitelist for experimental natives.
BUG=v8:5427,v8:4483
Review-Url: https://codereview.chromium.org/2365123002
Cr-Commit-Position: refs/heads/master@{#39706}
This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:
- Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
- Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
- Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
- Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
- Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
- Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
- Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.
I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.
This reduces the percentage of time spent in making BytecodeArrays
in CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.
BUG=v8:4280
Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
Review-Url: https://codereview.chromium.org/2351763002
Cr-Original-Commit-Position: refs/heads/master@{#39599}
Cr-Commit-Position: refs/heads/master@{#39637}
Reason for revert:
Prime suspect for roll blocker: https://codereview.chromium.org/2362503002/
Original issue's description:
> [Interpreter] Optimize BytecodeArrayBuilder and BytecodeArrayWriter.
>
> This CL optimizes the code in BytecodeArrayBuilder and
> BytecodeArrayWriter by making the following main changes:
>
> - Move operand scale calculation out of BytecodeArrayWriter to the
> BytecodeNode constructor, where the decision on which operands are
> scalable can generally be statically decided by the compiler.
> - Move the maximum register calculation out of BytecodeArrayWriter
> and into BytecodeRegisterOptimizer (which is the only place outside
> BytecodeGenerator which updates which registers are used). This
> avoids the BytecodeArrayWriter needing to know the operand types
> of a node as it writes it.
> - Modify EmitBytecodes to use individual push_backs rather than
> building a buffer and calling insert, since this turns out to be faster.
> - Initialize BytecodeArrayWriter's bytecode vector by reserving 512
> bytes,
> - Make common functions in Bytecodes constexpr so that they
> can be statically calculated by the compiler.
> - Move common functions and constructors in Bytecodes and
> BytecodeNode to the header so that they can be inlined.
> - Change large static switch statements in Bytecodes to const array
> lookups, and move to the header to allow inlining.
>
> I also took the opportunity to remove a number of unused helper
> functions, and rework some others for consistency.
>
> This reduces the percentage of time spent in making BytecodeArrays
> in CodeLoad from ~15% to ~11% according to perf. The
> CoadLoad score increase by around 2%.
>
> BUG=v8:4280
>
> Committed: https://crrev.com/b11a8b4d41bf09d6b3d6cf214fe3fb61faf01a64
> Cr-Commit-Position: refs/heads/master@{#39599}
TBR=mythria@chromium.org,leszeks@chromium.org,rmcilroy@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2360193003
Cr-Commit-Position: refs/heads/master@{#39612}
This CL optimizes the code in BytecodeArrayBuilder and
BytecodeArrayWriter by making the following main changes:
- Move operand scale calculation out of BytecodeArrayWriter to the
BytecodeNode constructor, where the decision on which operands are
scalable can generally be statically decided by the compiler.
- Move the maximum register calculation out of BytecodeArrayWriter
and into BytecodeRegisterOptimizer (which is the only place outside
BytecodeGenerator which updates which registers are used). This
avoids the BytecodeArrayWriter needing to know the operand types
of a node as it writes it.
- Modify EmitBytecodes to use individual push_backs rather than
building a buffer and calling insert, since this turns out to be faster.
- Initialize BytecodeArrayWriter's bytecode vector by reserving 512
bytes,
- Make common functions in Bytecodes constexpr so that they
can be statically calculated by the compiler.
- Move common functions and constructors in Bytecodes and
BytecodeNode to the header so that they can be inlined.
- Change large static switch statements in Bytecodes to const array
lookups, and move to the header to allow inlining.
I also took the opportunity to remove a number of unused helper
functions, and rework some others for consistency.
This reduces the percentage of time spent in making BytecodeArrays
in CodeLoad from ~15% to ~11% according to perf. The
CoadLoad score increase by around 2%.
BUG=v8:4280
Review-Url: https://codereview.chromium.org/2351763002
Cr-Commit-Position: refs/heads/master@{#39599}
This also adds sources missing for PPC and x87, fixes a few
missing files in gyp due to wrong quotation and a few that
were simply not included.
The gn files are now authoritative, but the gcmole gyp and
gn source lists are enforced to match exactly.
This additional enforcement helped finding the bugs above
and will be removed when we deprecate the gyp files.
BUG=614645
NOTRY=true
Review-Url: https://codereview.chromium.org/2352103002
Cr-Commit-Position: refs/heads/master@{#39592}
This is some initial cleanup to keep /src clean. The
AccountingAllocator is actually exclusively used by zones and this
common subfolder makes that more clear.
BUG=v8:5409
Review-Url: https://codereview.chromium.org/2344143003
Cr-Commit-Position: refs/heads/master@{#39558}
Full code uses patching ICs for this feedback, and the interpreter uses
the type feedback vector. It's a good idea to code the vector slots
appropriately as ICs so that the runtime profiler can better gauge if
the function is ready for tiering up from Ignition to TurboFan.
As is, the feedback is stored in "general" slots which can't be
characterized by the runtime profiler into feedback states.
This CL addresses that problem. Note that it's also important to
carefully exclude these slots from the profiler's consideration when
determining if you want to optimize from Full code.
BUG=
Review-Url: https://codereview.chromium.org/2342853002
Cr-Commit-Position: refs/heads/master@{#39555}
Adds template parameters for the TemplateHashMapImpl for the key and
value type, to allow them to be something other than pointers. To keep
the impact of this patch low, uses of TemplateHashMapImpl set these
types to void* to emulate the previous behaviour.
This is part of a wider set of changes discussed in:
https://groups.google.com/forum/#!topic/v8-dev/QLsC0XPYLeM
Review-Url: https://codereview.chromium.org/2343123002
Cr-Commit-Position: refs/heads/master@{#39530}
Reason for revert:
Unblocking roll
Original issue's description:
> [d8] Fix the shared-library build
>
> This commit ensures that the d8 shared library build uses the same logic as
> the standard static build by exporting relevant functions and classes.
>
> BUG=chromium:646337
>
> Committed: https://crrev.com/2c10ca8086a4d595ecf9aa843d2031b068470d65
> Cr-Commit-Position: refs/heads/master@{#39503}
TBR=jochen@chromium.org,vogelheim@chromium.org,bmeurer@chromium.org,titzer@chromium.org,jgruber@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:646337
Review-Url: https://codereview.chromium.org/2356703003
Cr-Commit-Position: refs/heads/master@{#39526}
This commit ensures that the d8 shared library build uses the same logic as
the standard static build by exporting relevant functions and classes.
BUG=chromium:646337
Review-Url: https://codereview.chromium.org/2342563002
Cr-Commit-Position: refs/heads/master@{#39503}
This flag has been flipped off since 52, so it is due for removal.
R=adamk@chromium.org,caitp@igalia.com
BUG=v8:3785
CQ_INCLUDE_TRYBOTS=master.tryserver.v8:v8_linux_noi18n_rel_ng
Review-Url: https://codereview.chromium.org/2268633002
Cr-Commit-Position: refs/heads/master@{#39435}
This CL adds fuzzers for the wasm module sections 'types', 'names',
'globals', 'imports', 'function signatures', 'memory', and 'data', one
fuzzer per section. No fuzzers are added for the other sections because
either there already exists a fuzzer (e.g. wasm-code), or there exist
inter-section dependencies.
To avoid introducing a bunch executables which would make compilation
with make slow, I introduce a single executable
'v8_simple_wasm_section_fuzzer' which calls the fuzzers mentioned above.
This executable is run by the trybots and ensures that the fuzzers
actually compile. For debugging I introduce commandline parameters which
allow to execute the specific fuzzers from 'v8_simple_wasm_section_fuzzer'.
R=titzer@chromium.org, jochen@chromium.org, mstarzinger@chromium.org
Review-Url: https://codereview.chromium.org/2336603002
Cr-Commit-Position: refs/heads/master@{#39413}
The wasm-module-runner is used both in cctests and in fuzzers. As
discussed offline, it is weird to include cctest header files in
fuzzers, so I introduce a new test/common directory which contains the
common files.
R=titzer@chromium.org, jochen@chromium.org
Review-Url: https://codereview.chromium.org/2335193002
Cr-Commit-Position: refs/heads/master@{#39411}
With this CL the wasm-code-fuzzer first decodes and interprets the test
case generated by the fuzzer. It then compiles the test case, but only
executes the compiled instance if the interpretation of the test case
was successful. If the compiled instance is executed, then the result of
the execution is compared with the result of the interpretation.
Additionally this CL refactors the CompileAndRunWasmModule function in
wasm-module.cc to resuse code in the call to the interpreter.
R=titzer@chromium.org
Review-Url: https://codereview.chromium.org/2321443002
Cr-Commit-Position: refs/heads/master@{#39351}
- inspector becomes a dependency of v8_base;
- generated public protocol files are placed to gen/v8/include/inspector/<Domain.h>;
- added v8_enable_inspector_override to be used in embedders (gn only);
- combined public headers into v8-inspector.h and v8-inspector-protocol.h.
BUG=chromium:635948
Review-Url: https://codereview.chromium.org/2292053003
Cr-Commit-Position: refs/heads/master@{#39226}
Spec discussion: https://github.com/tc39/ecma402/issues/30
It's in stage 4 and Firefox has already implemented it.
For now, it's added to HARMONY_IN_PROGRESS bucket behind
'--datetime-format-to-parts' flag.
BUG=v8:5244
TEST=intl/date-format/date-format-to-parts.js
TEST=test262/intl402/DateTimeFormat/prototype/formatToParts/*
Review-Url: https://codereview.chromium.org/2273953003
Cr-Commit-Position: refs/heads/master@{#39225}
Reason for revert:
Performance regressions: crbug.com/644087
Clusterfuzz: crbug.com/644074
We'll reland all regexp changes at once when the port is complete and at least performance-neutral, since the partial port requires slow workarounds.
Original issue's description:
> [regexp] Port RegExpConstructor
>
> BUG=v8:5339
>
> Committed: https://crrev.com/0aa80be16b8d876999066545c2922e1291e09610
> Cr-Commit-Position: refs/heads/master@{#39074}
TBR=bmeurer@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=v8:5339
Review-Url: https://codereview.chromium.org/2312743002
Cr-Commit-Position: refs/heads/master@{#39177}
This makes for slightly faster rebuilds when touching parser-base.h
(which changes frequently!). Also takes care of an old TODO,
moving CompileTimeValue into its own file under ast/, where it
properly belongs.
BUG=v8:5294
Review-Url: https://codereview.chromium.org/2305883002
Cr-Commit-Position: refs/heads/master@{#39141}
Our Type class has a semantic and representational dimension.
Much code in src/ast, Crankshaft and Turbofan is based on it.
Going forward in Turbofan we'd like to remove representational information
entirely. To that end, new type AstType has been created to preserve
existing behavior for the benefit of Crankshaft and the AST.
BUG=
Review-Url: https://codereview.chromium.org/2302283002
Cr-Commit-Position: refs/heads/master@{#39135}
Correct a small bug which throws a compile-time assertion for ARMv8
builds since this change:
https://codereview.chromium.org/2223433002
This does not add full support for ARMv8 instructions; it simply sets
the CAN_USE_ARMV7_INSTRUCTIONS define on ARMv8 builds.
BUG=
Review-Url: https://codereview.chromium.org/2293253004
Cr-Commit-Position: refs/heads/master@{#39087}
This way, many files which only need CompilationInfo but not compiler.h
and its dependencies can include just compilation-info.h.
BUG=
Review-Url: https://codereview.chromium.org/2284313003
Cr-Commit-Position: refs/heads/master@{#39038}
For asm.js we now have a dedicated AsmTyper, that uses it's own type
system (which is tailored towards asm.js), and so we don't need the
special asm.js types anymore in the TypeCache. This also moves the
TypeCache into the src/compiler directory, because it doesn't make
sense to use outside anyways.
TBR=ahaas@chromium.orgR=jarin@chromium.org
BUG=v8:5267
Review-Url: https://codereview.chromium.org/2289573002
Cr-Commit-Position: refs/heads/master@{#38964}
Reason for revert:
Significantly tanks parsing. We probably should just keep on doing what we're doing: partially deserialize while resolving variables. If we do scope-info backed resolution after regular resolution based on remaining free variables, we can probably reduce the time-frame of that part. We soon after anyway need to sync with the main thread.
Original issue's description:
> Always deserialize scope infos for parsing
>
> When looking up variables in the ScopeInfo, we did a linear scan of the
> ScopeInfo. Since that's unacceptably slow, a context slot cache was added
> that would speed up repeated lookups of the same variable.
>
> Instead, just always fully convert the ScopeInfo into scopes, so they can
> lookup variables without scanning the ScopeInfo.
>
> This also allows for removing the now unused ContextSlotCache.
>
> R=adamk@chromium.org,verwaest@chromium.org,marja@chromium.org
> BUG=v8:5315
>
> Committed: https://crrev.com/81f824cad18e4dc873a8838943217eb9c9f0c1f0
> Cr-Commit-Position: refs/heads/master@{#38953}
TBR=adamk@chromium.org,marja@chromium.org,jochen@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:5315
Review-Url: https://codereview.chromium.org/2287783003
Cr-Commit-Position: refs/heads/master@{#38958}
When looking up variables in the ScopeInfo, we did a linear scan of the
ScopeInfo. Since that's unacceptably slow, a context slot cache was added
that would speed up repeated lookups of the same variable.
Instead, just always fully convert the ScopeInfo into scopes, so they can
lookup variables without scanning the ScopeInfo.
This also allows for removing the now unused ContextSlotCache.
R=adamk@chromium.org,verwaest@chromium.org,marja@chromium.org
BUG=v8:5315
Review-Url: https://codereview.chromium.org/2280933002
Cr-Commit-Position: refs/heads/master@{#38953}
Introduce a new TypedOptimization reducer that contains the type
based optimization reduction steps, which are not (directly)
related to lowering JavaScript operators based on types (which is
what JSTypedLowering is supposed to do).
This also addresses a chicken-and-egg problem that we see in the
Octane/Mandreel benchmark where type based constant folding isn't
applied to the numeric comparison operators introduced by the
JSTypedLowering itself, and thus gives up to 10% speedup for the
benchmark.
BUG=v8:5267
Review-Url: https://codereview.chromium.org/2280673003
Cr-Commit-Position: refs/heads/master@{#38928}
DuplicateFinder isn't actually used by the Scanner, except for one
convenience function which we should probably remove, also.
BUG=
Review-Url: https://codereview.chromium.org/2281443002
Cr-Commit-Position: refs/heads/master@{#38904}
This flag was shipped on in 52, so it's due for removal. The patch includes
removing the deprecated and unused-in-Blink API Promise::Chain, and many
test updates.
R=adamk@chromium.org
BUG=v8:4633
Review-Url: https://codereview.chromium.org/2267033002
Cr-Commit-Position: refs/heads/master@{#38804}
When doing a component build, some test binaries link against the object
files directly, bypassing the components. This results, however, and
rightly so, in linker warnings. In gyp, we just suppressed them. During
the transition to gn, this was dropped for two binaries.
Here I add the suppressions back in.
Long term, we should either change the tests to go through the public
API, or export the required symbols.
BUG=chromium:633688
R=jkummerow@chromium.org
Review-Url: https://codereview.chromium.org/2261123003
Cr-Commit-Position: refs/heads/master@{#38793}
This makes us able to get rid of dependencies to parser.h from places
which only need the ParseInfo, and also gets rid of the curious Parser
<-> Compiler circular dependency.
Also IWYUd where necessary.
BUG=
Review-Url: https://codereview.chromium.org/2268513002
Cr-Commit-Position: refs/heads/master@{#38777}
Adds support for dumping the stack on Windows. Also enables in-process
stack dumping in d8 to dump the stack on exceptions and signals.
This CL changes the format of stack dumps from:
1: V8_Fatal
2: 0x1ac6ba5
3: v8::internal::interpreter::BytecodeGenerator::Visit(v8::internal::AstNode*)
4: v8::internal::interpreter::BytecodeGenerator::VisitForAccumulatorValue(v8::internal::Expression*)
...
To:
./out/x64.debug/d8(v8::base::debug::StackTrace::StackTrace()+0x1e) [0x1c6ee5e]
./out/x64.debug/d8() [0x1c6ede5]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x10330) [0x7fa01193e330]
./out/x64.debug/d8(v8::base::OS::Abort()+0x12) [0x1c6cea2]
./out/x64.debug/d8() [0x1c67538]
./out/x64.debug/d8() [0x1ac80b5]
./out/x64.debug/d8(v8::internal::interpreter::BytecodeGenerator
::Visit(v8::internal::AstNode*)+0x3cb) [0x1ac323b]
./out/x64.debug/d8(v8::internal::interpreter::BytecodeGenerator
::VisitForAccumulatorValue(v8::internal::Expression*)+0x40) [0x1ac2570]
Review-Url: https://codereview.chromium.org/2248393002
Cr-Commit-Position: refs/heads/master@{#38717}
