WasmModuleBuilder is a class that is used to build Wasm modules in the
asm.js parser, in the fuzzer, as well as some tests. When it comes to
Wasm tables, WasmModuleBuilder currently supports only basic tables
(before the reftypes proposal) using an ad-hoc indirect-function index
vector.
This CL adds proper support for element sections and tables that use
them in the full potential of the reftypes extension. The new
functionality will only be used in the fuzzer and potentially some tests
in the future. Along this, we drop some functionality from
WasmModuleBuilder that was only used in tests and is redundant with the
new architecture.
Additionally, we remove tables other than externref and funcref from the
fuzzer (which were not supported properly or used anyway). We will
reintroduce them at a later time.
Bug: v8:11954
Change-Id: I0a4f6e7b63b6e3d9f7da03b5202fbf14d8678332
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3122162
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76597}
- Move the compilation-dependency.h header contents into
compilation-dependencies.cc;
- add macro lists to define type checks and casts;
- add invalidated dependency tracing to
the --trace-compilation-dependencies flag (renamed from
--trace-code-dependencies).
Bug: v8:7790
Change-Id: I34b950cd0b79b8d2673b1195599aec763f6b60d1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129420
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76596}
This change implements longer-width SIMD instructions in the x64
assembler by adding 256-bit versions to one of the conversion macros.
This emits mostly floating-point arithmetic and some boolean operations;
see `SSE_UNOP_INSTRUCTION_LIST` and `SSE_BINOP_INSTRUCTION_LIST`.
Design doc: https://docs.google.com/document/d/1VWZbkO5c_DdxlJObmSLN_9zQUZELVgXyudbpzv5WQM0
Change-Id: I36d56ee09d6b71f66734342cb37bfc9d4801d654
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3123648
Reviewed-by: Zhi An Ng <zhin@chromium.org>
Commit-Queue: Shiyu Zhang <shiyu.zhang@intel.com>
Cr-Commit-Position: refs/heads/main@{#76593}
Rolling v8/build: 609fda6..36b1f71
Rolling v8/buildtools: 66ed234..ac4ea2d
Rolling v8/buildtools/third_party/libc++abi/trunk: 996cd74..be47d0e
Rolling v8/third_party/aemu-linux-x64: ddSGEd1PA5UG-ur2YFNiaY2SuddaJRHJB8HprZVuD68C..CPOECXfDP5keozFnmr0QOiSuGL3ELWB3zIOKJ0CkoAwC
Rolling v8/third_party/depot_tools: bd0674c..d4534cc
Rolling v8/third_party/instrumented_libraries: 9a8087b..ee10dbdTBR=v8-waterfall-sheriff@grotations.appspotmail.com,mtv-sf-v8-sheriff@grotations.appspotmail.com
Change-Id: I1425b7839fc335143338ae375ebc966c37a39210
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3131926
Reviewed-by: v8-ci-autoroll-builder <v8-ci-autoroll-builder@chops-service-accounts.iam.gserviceaccount.com>
Commit-Queue: v8-ci-autoroll-builder <v8-ci-autoroll-builder@chops-service-accounts.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#76592}
template specialisations must be defined
outside of class body to prevent the following compilation error:
error: explicit specialization in non-namespace scope
Change-Id: Ic4b74a28cd21d96991ad784fbd3c598668ffc476
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129881
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Milad Fa <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#76591}
This guards against concurrent usages of PageBackend::Lookup() from
HeapRegistry which can race with adding/removing pages.
This race only manifests in debug mode.
Change-Id: If34dbc255faeda085e522501ff2995693cd97b2e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129702
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76590}
This was missing in one of the last refactorings.
No-Try: true
Bug: chromium:1044942
Change-Id: I2c6bfc75251fad61f35a75afec3a1b2682175d68
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3127705
Reviewed-by: Liviu Rau <liviurau@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76589}
The wasm-compile fuzzer sometimes needs to generate a boolean flag from
the input bytes. Since the general {DataRange::get} method results in
undefined behaviour if instantiated with the {bool} type, we are getting
an 8-bit value instead and looking at the least significant bit only.
This CL improves this situation by implementing a template
specialization for {bool} which uses the same trick, and uses that
instead of hand-coding the modulo operation at the call sites.
R=manoskouk@chromium.org
Bug: v8:11879
Change-Id: I6f9ce02dd8d9cd0998b83e081e4c6ca773e6cb53
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129429
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76588}
Most Torque-defined extern classes already use CPP class generation. As
Nico pointed out in [1], it would be nice to convert the remaining
classes and remove this option. This change converts most of those
remaining classes. I know that the future of Torque-defined classes is a
subject of some debate right now, but I think that it's worth doing a
few mechanical changes to reduce the existing variety of options. A
couple of minor fixes in the Torque compiler were required so that it
generates correct code for shapes.
[1] https://docs.google.com/document/d/1q_gZLnXd4bGnCx3IUfbln46K3bSs9UHBGasy9McQtHI/edit#
Bug: v8:8952
Change-Id: I7e6087153a18d6ee80e67926793e8ba8e01d501e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3015666
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Anton Bikineev <bikineev@chromium.org>
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#76586}
Bug: chromium:1209444
Change-Id: I4ec16a718061063dc01ec0d7c4a397c220e684c0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3127718
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76585}
... by removing some obsolete code.
Bug: v8:7790
Change-Id: Iad31b60de5905ba05e1c622e81c3234071752e9f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124806
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76584}
Back references to C++ objects may point to objects that never have
their graph nodes materializes through other C++ edges. We can just
create a graph node in this case, and avoid delaying the merging
completetly.
Bug: chromium:1244522
Change-Id: I0e9cb7a89ee90bfba217bc8475ac40bd7fe92a0b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129426
Reviewed-by: Omer Katz <omerkatz@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76583}
... by removing some obsolete code.
Bug: v8:7790
Change-Id: I3a244ef5fc7fe15321e5bb1c9bb2fe794030ba3b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124801
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76582}
It seems like SP on heap does not produce too much memory fragmentation,
therefore we do not need UndoLastAllocationAt.
Bug: v8:11872
Change-Id: Id2e44405329b52c1dcd6cd81bfc72ffba00035ee
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129428
Auto-Submit: Victor Gomes <victorgomes@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76581}
... by removing some obsolete code.
Bug: v8:7790
Change-Id: I722031158d45335f3e086eb335a447fbc5066cac
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124798
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76580}
... by removing some obsolete code.
Bug: v8:7790
Change-Id: I32880d2a4fbd943ea0e485d8e8aff07ac9903e9d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124795
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76576}
... by removing some obsolete code.
Bug: v8:7790
Change-Id: Ie098055a1849de5d853c126e0c7275164f964ce8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124774
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76575}
The log test checks for log positions, which may change when background
serialization / background compilation are enabled.
Fixed: v8:12117
Change-Id: I193c9c23e016fad1e3f06a9f377bb53db84a6988
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3129421
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76573}
Allow %CompileBaseline on architectures that support Sparkplug
independent of runtime flags.
The deicsion based on --sparkplug runtime flag lead to spurious errors
on correctness fuzzers.
Bug: chromium:1244474
Change-Id: I764bd80cd7dff7e72729145c165dc039c594753c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3127719
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76572}
Instead create the appropriate data on demand. Note that this
changes behavior of the default configuration.
Bug: v8:7790
Change-Id: Ia6bfcaace655c0fd72e2dcc0c2547195dc1cc4a7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3123419
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76570}
As with other crashes, also SIGABRT cuts of execution earlier. While
the reason might be interesting in normal fuzzing, in correctness
fuzzing, it leads to spurious reports due to the output differences.
No-Try: true
Bug: chromium:1242193
Change-Id: I6ee9a8e2a0254255d02b0106575931b523fbf666
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124808
Auto-Submit: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76569}
If a stack overflow occurs inside the regexp parser, propagate that
information to the parser.
Bug: v8:896,chromium:1243989
Change-Id: I5ced27ff968ad97764e156643e1980b3a722af1a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3127717
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76568}
This is for tests only, and in those tests it should crash, while
on fuzzers it should silently fail. For those failing cases, we
should clear the exception so that the runtime call isn't confused
Bug: chromium:1244254
Change-Id: I5bb1c50d1538331dd9298911d742530b9769be8d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3127714
Commit-Queue: Patrick Thier <pthier@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76566}
We can now tighten the return type of FindRootMap and remove some
related code.
Bug: v8:7790
Change-Id: I08325e7e4f4c9261c45770f7674b6644cc5c2b80
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3123411
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76564}
These methods are called only during the inlining phase, so even in the
default configuration we follow the same branch as concurrent inlining
and ignore the serialized data. We can thus tighten their return types
and cut down JSBoundFunctionData.
Bug: v8:7790
Change-Id: Ic48f8f2651d684440dc5f6a9934de2ae3a5b5132
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3123410
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76561}
As per the release plan in crbug.com/v8/12142.
Bug: v8:7790,v8:12142
Change-Id: I80e2a3c571681a968ea245d52adfa539e0e7ab7b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3127711
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76560}
.. to consistently support more than a single argument.
Each argument is now a tagged union that may contain an AST string, a
C string, or a JS string handle.
Change-Id: Iac8e40b717dea95a2bc2903449dab56c181702d6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3122086
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76559}
The runtime-function blocking is implemented in V8 behind the
--fuzzing flag since a while now. The legacy blocklist on the fuzzer
side can be removed since some time now - it already diverted.
No-Try: true
Bug: chromium:1044942
Change-Id: I55f92419beb4d4462cbe03918dbf84d9c979862d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124810
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76556}
The CL https://crrev.com/c/2928505 changed tests without changing the
fuzzer with the result that tests fail now.
It's not helpful to switch to using new API methods in generated fuzz
tests, as they'd then not bisect well backwards to older V8 versions.
No-Try: true
Change-Id: Ia307e88b4532bd792091b23374889f2b0e490fce
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3124809
Auto-Submit: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76552}
Land some of the tests for Temporal.PlainDate
All marked as FAIL at this stage.
Bug: v8:11544
Change-Id: I004b7cb34effe1de1735b61c7ac749ae3c8e9bf7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3085624
Commit-Queue: Frank Tang <ftang@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76551}
Land some of the tests for Temporal.Instant
All marked as FAIL at this stage.
Bug: v8:11544
Change-Id: I79d14df47248c708e5d73a0e00e3f7973c521d16
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3086903
Commit-Queue: Frank Tang <ftang@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#76550}
