The main goal is to untangle Liftoff from the TF-based wasm compiler,
but since the AccessBuilder does not simplify anything but rather adds
complexity I also removed it from the wasm compiler.
Instead, we now bottleneck all offset computations through the new
ObjectAccess helper.
R=titzer@chromium.org
Bug: v8:6600
Change-Id: I362b7b889d68e89da8c30d3fad7b5bab07bee5c8
Reviewed-on: https://chromium-review.googlesource.com/1204090
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55646}
Shrink embedded.cc by writing out octa words instead of bytes. This
halves the size of the generated file from 28MB to 14MB in a debug build
and reduces compile times for the file from ~2s to ~0.6s.
Bug: v8:8129
Change-Id: I90893c7732d83f4eeedee964cd81958201e3b05c
Reviewed-on: https://chromium-review.googlesource.com/1204111
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55645}
V8 does not require the embedder to open a HandleScope before executing
a v8::Task. However, d8 does open such a HandleScope. Because of that
HandleScope we repeatedly found missing HandleScopes only in Chrome
tests and not in d8 tests. The same is true for the context, which is
not set when Chrome calls a v8::Task.
With this CL we create a SealHandleScope around the execution of a
v8::Task, and we set the context to nullptr, so that d8 matches Chrome
better.
Ideally d8 would not open a HandleScope in the first place, and would
not set a context. Both make d8 behave different than Chrome and thereby
may hide bugs from our testing infrastructure. However, the
implementation of the InspectorClient requires them. I think the
SealHandleScope and resetting the context is a good workaround at the
moment. Yang, do you know if we can get rid of the context there in the
long run?
R=yangguo@chromium.org
Change-Id: I8df0fabde7dfdcdc630d20af4c68f141ac3a454c
Reviewed-on: https://chromium-review.googlesource.com/1177742
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55644}
Path names in import/export statements are relative to the file they are in.
This fixes the logic and unblocks using the messages test suite on Android,
which has cases importing files from mjsunit, which import more files from
there.
Bug: chromium:866862
Change-Id: I8d2ff645f69b67fbdaf4a622d06308e55298b0ce
Reviewed-on: https://chromium-review.googlesource.com/1206570
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55643}
This CL changes the call-site of SmiLexicographicCompare to a fast
c call instead of a runtime call. The runtime function is not deleted
as it is still used in InnerArraySort.
The test is also moved from mjsunit to cctest, to make removal of the
runtime function easier in the future.
R=cbruni@chromium.org, jgruber@chromium.org
Bug: v8:7382
Change-Id: Ie961eeb094c13018e9ec28b68f7c444d7f889036
Reviewed-on: https://chromium-review.googlesource.com/1201587
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55642}
This CL exposes IteratorBuiltinsAssembler::IterableToList as a builtin
to reduce generated code duplication. This follows up on CL 1201882.
Change-Id: I848e17bd1b6756de9e898e9d2f8c93d99699df07
Reviewed-on: https://chromium-review.googlesource.com/1206470
Commit-Queue: Hai Dang <dhai@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55641}
This is a reland of 1c48d52bb1.
It turned out that IterableToList doesn't always behave according to
the ES operation with the same name. Specifically, it allows holey arrays
to take its fast path, which produces an output array with holes where
actually "undefined" elements should appear.
This CL changes the version of IterableToList that is used for spreads
(IterableToListWithSymbolLookup) such that holey arrays take the slow path.
It also includes tests for such situations.
Original change's description:
> [interpreter] Add bytecode for leading array spreads.
>
> This CL improves the performance of creating [...a, b] or [...a].
> If the array literal has a leading spread, this CL emits the bytecode
> [CreateArrayFromIterable] to create the literal. CreateArrayFromIterable
> is implemented by [IterableToListDefault] builtin to create the initial
> array for the leading spread. IterableToListDefault has a fast path to
> clone efficiently if the spread is an actual array.
>
> The bytecode generated is now shorter. Bytecode generation is refactored
> into to BuildCreateArrayLiteral, which allows VisitCallSuper to benefit
> from this optimization also.
> For now, turbofan also lowers the bytecode to the builtin.
>
> The idiomatic use of [...a] to clone the array a now performs better
> than a simple for-loop, but still does not match the performance of slice.
>
> Bug: v8:7980
>
> Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
> Change-Id: Ibde659c82d3c7aa1b1777a3d2f6426ac8cc15e35
> Reviewed-on: https://chromium-review.googlesource.com/1181024
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Commit-Queue: Georg Neis <neis@chromium.org>
> Commit-Queue: Hai Dang <dhai@google.com>
> Cr-Commit-Position: refs/heads/master@{#55520}
Bug: v8:7980
Change-Id: I0b5603a12d2b588327658bf0a9b214bd0f22e237
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Reviewed-on: https://chromium-review.googlesource.com/1201882
Commit-Queue: Hai Dang <dhai@google.com>
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55639}
GCC may be buggy in how it handles visibility in this case. The
previous nested implementation resulted in a 'lambda1 declared with
greater visibility than the type of its field lambda2' error.
Unnesting the inner lambda works around the error.
Bug: v8:8126
Change-Id: Id822ca80fec9af27c4adc7ff53be3b6d9478f0d7
Reviewed-on: https://chromium-review.googlesource.com/1206310
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55638}
Corrected register calling syntax in assembler-s390.cc and
test-platform.cc.
Generate_CEntry in builtins-s390.cc expects return buffer to be
preserved in r2, but when built with clang r2 isn't preserved, which breaks
300+ tests. It is fixed by writing r2's value into r8 (preserved)
and loading the value back to r2 after the operation.
Change-Id: I184f0111944b6ad8c0ccc8b97407d702dd97d9d8
Reviewed-on: https://chromium-review.googlesource.com/1204530
Reviewed-by: Junliang Yan <jyan@ca.ibm.com>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Junliang Yan <jyan@ca.ibm.com>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55631}
Since there is no `Value::IsInteger` method in the API, we in the
Node.js project are going to rely on what looks like an implementation
detail of the Integer class. It is currently possible to to call
`Integer::Value` on any Number and the value is cast to an integer.
This commit adds tests for this behavior.
Change-Id: I4de09e7c6e0beac7909e5477f7bfe2ed4c9415b9
Reviewed-on: https://chromium-review.googlesource.com/1200983
Commit-Queue: Michaël Zasso <mic.besace@gmail.com>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55629}
Port CreateBreakIterator and BreakIterator.prototype.resolvedOptions to
C++, refactoring the entire class into another one called
JSV8BreakIterator that would be a subclass of JSObject.
TBR: benedikt@chromium.org
Bug: v8:8111
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I9bd1d82ec34b210c8ed59ea6576548d45a34b8d5
Reviewed-on: https://chromium-review.googlesource.com/1198946
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55627}
Since the benchmark is very fast, a warmup is useful to reduce the noise
created by optimization.
Change-Id: I4902c5c1695099be766d0fcc563c2f5d0892d3a9
Reviewed-on: https://chromium-review.googlesource.com/1204112
Commit-Queue: Hai Dang <dhai@google.com>
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55618}
This fixes an off-by-one in the ordering restriction check for exception
sections in a module. It also adds proper testing for exception handling
of indirect calls. This in turn adds a table section that triggers the
aforementioned bug.
R=clemensh@chromium.org
TEST=mjsunit/wasm/exceptions,unittests/WasmModuleVerifyTest
BUG=v8:8091
Change-Id: Ie44ad4dee1b0c623f069fca7661c4282492b52d6
Reviewed-on: https://chromium-review.googlesource.com/1203993
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55615}
{go} sounds active, as if the state would be changed. {reached} sounds
like an accessor.
Drive-by: Make the accessor const to further clarify that this is just
an accessor.
R=titzer@chromium.org
Bug: v8:8015
Change-Id: I80d5223b9573345d3de386d68871a9e828621fc0
Reviewed-on: https://chromium-review.googlesource.com/1203992
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55614}
Utf*Characterstream caches the data pointer of ExternalStrings through
ExternalStringStream, so lock the strings in ExternalStringStream.
Bug: chromium:877044
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I241caaf64e109b33e2f9982573e11c514410509c
Reviewed-on: https://chromium-review.googlesource.com/1194003
Commit-Queue: Benoit L <lizeb@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55613}
Before, splice was implemented with a C++ fast path and a
comprehensive JavaScript version.
This impl. is entirely in Torque with a fastpath for SMI,
DOUBLE and OBJECT arrays, and a comprehensive slow path.
The same level of "sparse" array support as given by the
array.js implementation is included.
This reland addresses several issues:
* Removed "sparse" array support from splice.
* Addressed ClusterFuzz issue 876443:
The test and code that uses the fix is in this CL.
The fix in isolation can be seen here:
https://chromium-review.googlesource.com/c/v8/v8/+/1199403
* Removed dead code in elements.cc
BUG=chromium:876443, v8:8131, v8:1956, v8:7221
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I2d4a66c24ba1edabeca34e27e6ff8ee6136ed5f1
Reviewed-on: https://chromium-review.googlesource.com/1201783
Commit-Queue: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55610}
Since https://crrev.com/c/1112003, the memory size is stored as size_t
instead of uint32_t in order to support 4GB memories.
Loading it as uint32_t only works on little-endian systems, and only for
memory sizes <4GB. This CL fixes this to load and process the memory
size as pointer-sized value.
Additional platform-specific methods are added to perform a shift by a
constant value. This can be reused to improve the generated code for
other shifts.
R=titzer@chromium.org
Bug: v8:8130, v8:6600
Change-Id: Ifa688a3ed0e2809190571f24bdf47a7f53880b3d
Reviewed-on: https://chromium-review.googlesource.com/1203950
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55609}
Also extend the API to reflect this new feature.
R=jgruber@chromium.org, szuend@google.com, ulan@chromium.org
Bug: v8:8125
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Ic7a7604a8c663ba04b324eb8902ff325a25654e7
Reviewed-on: https://chromium-review.googlesource.com/1202087
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55604}
This fixes exception creation (by the WebAssembly throw operation) so
that it is not observable by JavaScript. Internal properties are now
stored with symbol names instead of string names, which also prevents
them from being accessed or monkey-patched directly by JavaScript.
R=clemensh@chromium.org
TEST=mjsunit/regress/wasm/regress-8094
BUG=v8:8094
Change-Id: I33cb27f4373114cd4db28d9aef23560093e55242
Reviewed-on: https://chromium-review.googlesource.com/1203951
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55602}
WebAssembly.Instantiate would silently fail when the passed in memory
object did not have guard regions even though the compiled module assumes
so. This lead to an inconsitent state and a DCHECK error. Instead, now
throw a LinkError.
Change-Id: I68bab842bcc40d3325aea4b19979d80054ed407c
Reviewed-on: https://chromium-review.googlesource.com/1180892
Commit-Queue: Stephan Herhut <herhut@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55601}
The wasm compiler used Pipeline::GenerateCodeForTesting to generate code
for various stubs. This change adds a dedicated entry point and moves
some common code there.
Bug: v8:8015
Change-Id: Ied628ba14c36e68826cb71d00506994184cc4763
Reviewed-on: https://chromium-review.googlesource.com/1196885
Commit-Queue: Stephan Herhut <herhut@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55599}
