Instead of using runtime calls to generate the Array Literal passed to
%reflect_call / %reflect_construct, we create an ArrayLiteral from the
list of arguments, and perform spreads using the interpreter mechanism for
spreading in ArrayLiterals (thus, the spreading becomes inline). This
array literal is still passed to %reflect_call / %reflect_construct as
before.
This cuts the runtime for bench-spread-call.js -> testSpread roughly in
half, and will likely improve further once
https://chromium-review.googlesource.com/c/v8/v8/+/915364 has landed.
BUG=v8:7446
R=neis@chromium.org, adamk@chromium.org
Change-Id: I74a6acd3a60aad422e4ac575275c7b567659d8ad
Reviewed-on: https://chromium-review.googlesource.com/939587
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51678}
Avoid generating ADDs when concatenating the empty string with other
template parts. This prevents the creation of useless feedback slots,
and reduces the number of extra dispatches.
The impact on performance is negligible.
BUG=v8:7415
Change-Id: I7ef3806b53f7252f3a86f7007ae7050ac697c1e3
Reviewed-on: https://chromium-review.googlesource.com/938145
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Caitlin Potter <caitp@igalia.com>
Cr-Commit-Position: refs/heads/master@{#51669}
The memory gains were significant, so despite the bluebird-doxbee
regression, we think it's better to have this patch than not.
See the attached Chromium bug for more discussion.
This is a reland of 20e346bd08.
Original change's description:
> [parser] Remove pretenuring of closures assigned to properties
>
> This pretenuring was added in https://codereview.chromium.org/5220007,
> back when it was necessary in order to allow use of the closure
> as a "constant function" property. This should no longer be the case,
> and the pretenuring causes some unfortunate downstream effects.
>
> This patch removes the parser's setting of this bit. If it doesn't
> cause regressions on the perf bots, followup CLs will remove the
> rest of the support for this feature.
>
> Bug: v8:7442
> Change-Id: I27c43dd4293ce5de921be6c78571e712778d138a
> Reviewed-on: https://chromium-review.googlesource.com/914610
> Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
> Commit-Queue: Adam Klein <adamk@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#51254}
Bug: v8:7442, chromium:814182
Change-Id: I228c59dccef3844803f115749e72ae6c5f286eda
Reviewed-on: https://chromium-review.googlesource.com/938241
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51668}
This makes sure that the object can be identified by a unique instance
type and hence is not accidentally confused with other FixedArrays on
the heap.
R=clemensh@chromium.org
BUG=v8:7509
Change-Id: I66e3d779ff9a323b64f5464bdd5fe02aefe468c6
Reviewed-on: https://chromium-review.googlesource.com/943442
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51665}
This is a first step towards using Maps as store transition handlers.
It is expected for this CL to noticeably regress memory consumption
but most of it should be recovered by the next CL.
Bug: v8:5988
Change-Id: Ic2e301f9ccebc36e699383ded8c8cd284a906ce1
Reviewed-on: https://chromium-review.googlesource.com/928646
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51663}
Follow-up to https://chromium-review.googlesource.com/c/v8/v8/+/941442.
"background" refers to a priority and is inappropriate to refer to
worker threads as many tasks posted to worker threads by v8 are in
fact high priority.
Also took advantage of this rename to make NumberOfWorkerThreads()
return an int instead of size_t. While it is never negative, int is
simpler and Google C++ style guide states to avoid unsigned integers in
such cases (ref. "On Unsigned Integers" @
https://google.github.io/styleguide/cppguide.html#Integer_Types).
The Chromium embedder for that call provided an int which was converted
to size_t for this override and most often casted back down to int on the
v8 side, adding churn, and readability overhead.
R=ahaas@chromium.org
Bug: v8:7310
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Ib5280df73d2846b111d985be65a10b049995ea6a
Reviewed-on: https://chromium-review.googlesource.com/941944
Commit-Queue: Gabriel Charette <gab@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51662}
This patch removes the StringConstructor_ConstructStub builtin,
merging its functionality into the refactored StringConstructor
TurboFan builtin.
This brings us closer to our goal of deprecating the `construct_stub`
field in `SharedFunctionInfo`.
Bug: v8:7503
Change-Id: Ie98520c652f49dda91eff2fc51263611f29e0ebe
Reviewed-on: https://chromium-review.googlesource.com/942882
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51660}
Port 6687007028
Original Commit Message:
Replace bitwise arithmetic with conditional move / select instructions
on ia32, x64, Arm and Arm64. In local tests this improves --noopt Ignition
performance by between 2-5%.
R=rmcilroy@chromium.org, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=chromium:798964
LOG=N
Change-Id: I3cb65bdf2cf48df30e97b8a12966636fa6aca5ec
Reviewed-on: https://chromium-review.googlesource.com/939852
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Joran Siu <joransiu@ca.ibm.com>
Cr-Commit-Position: refs/heads/master@{#51658}
This is a reland of 01db326cc2.
Original change's description:
> [Assembler][x64] Make immediates immutable
>
> On x64, we already pass immediates by value. This CL ensures that this
> is indeed cheap, and it makes immediates immutable.
>
> R=mstarzinger@chromium.org
>
> Bug: v8:7310
> Change-Id: I53a0666d53b9de69d390621298798c03b5190497
> Reviewed-on: https://chromium-review.googlesource.com/934341
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#51613}
Bug: v8:7310
Change-Id: I37f1bd1528c742e8df11b7f524316935dd289f00
Reviewed-on: https://chromium-review.googlesource.com/941621
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51655}
This CL adds a regression benchmark for a fast-path of
String.p.charCodeAt, which is important for node.js.
Bug: v8:7326
Change-Id: I54efaa2988c595dd40e6a55a3464b3ee7de6f07b
Reviewed-on: https://chromium-review.googlesource.com/942885
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51654}
MSVC 2015 and 2017 implement std::is_trivially_copyable, but not
correctly. Hence, reimplement it using more low-level primitives.
For stdlibc++ versions below 5.0, we already have a workaround for the
missing support of std::is_trivially_copyable, but this is an unsound
approximation, because it is ignoring move constructor, move assignment
and copy assignment. Therefore, do not use this approximation for
asserting trivial copyability of a type.
Finally, add unittests for the new is_trivially_copyable
implementations.
R=mstarzinger@chromium.orgCC=loorongjie@gmail.com
Change-Id: I9ee56a65882e8c94b72c9a2d484edd27963a5d89
Reviewed-on: https://chromium-review.googlesource.com/941521
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51651}
and use it for parallel GC as it blocks the main thread.
On Chromium this will result in a TaskPriority::USER_BLOCKING task,
allowing TaskScheduler to prioritize them above all other async work.
R=ahaas@chromium.org
Bug: chromium:817421
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I8c782eb045035ce2899214528deae5a45f0e6600
Reviewed-on: https://chromium-review.googlesource.com/941946
Commit-Queue: Gabriel Charette <gab@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51650}
Some macros in test-assembler-mips have the same name
as newly introduced macros in sid sysroot. We rename
them in order to prevent compilation problems
Change-Id: I84d9562f39c1aca8ac20e979466862228f438425
Reviewed-on: https://chromium-review.googlesource.com/942322
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com>
Cr-Commit-Position: refs/heads/master@{#51649}
Introducing a new status-file variable: mips_arch_variant. With this
variable, in status files will be possible to define selections which
are based on MIPS architecture revisions/variants.
TEST=
BUG=
Change-Id: Ifd682552db2f26be4e56dc94ad50bed063ff14c5
Reviewed-on: https://chromium-review.googlesource.com/941212
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com>
Cr-Commit-Position: refs/heads/master@{#51647}
External references are process-specific and thus need to be accessed
through an indirection (or reloc'd by the linker). This CL moves all
used external references to the builtins constants table and rewrites
accesses to load from there.
In the future, this could be made more efficient by removing levels of
indirection or using the native linker.
Bug: v8:6666
Change-Id: I63491670549654edeb59c60bb833acfdc5a48495
Reviewed-on: https://chromium-review.googlesource.com/939783
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51646}
With a temporary intermediate step to allow adapting embedders before
getting rid of the ExpectedRuntime method altogether.
The method is being renamed to CallOnWorkerThread() as an effort to
go away from "background" nomenclature for worker threads ("background"
usually refers to a priority but worker threads are commonly used for
high priority tasks in v8).
Other CLs will follow to rename other "background" APIs.
Bug: v8:7310
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng;master.tryserver.chromium.linux:linux_chromium_rel_ng
Change-Id: I2fd4eac7458708d4eacb0f4871c982a567a3865e
Reviewed-on: https://chromium-review.googlesource.com/941442
Commit-Queue: Gabriel Charette <gab@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51645}
Also Add vhaddps to x64
Fix haddps for SSE3 scope and disassembler on ia32/x64
Change-Id: If511e6428fa1ce034b4281943dfee1405c9d4ffc
Reviewed-on: https://chromium-review.googlesource.com/939265
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Jing Bao <jing.bao@intel.com>
Cr-Commit-Position: refs/heads/master@{#51642}
For namespace objects, [[GetOwnProperty]] on an uninitialized property
throws a ReferenceError. This was not implemented everywhere. This CL
fixes all such issues I'm aware of.
Bug: v8:7470
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I5f024450005c4f4dcb3f41c844ef055f67a9a869
Reviewed-on: https://chromium-review.googlesource.com/937341
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51638}
The assert-guarded comment claiming that ToNumber could not
possibly neuter the target array unfortunately turns out to
have been wishful thinking.
Bug: chromium:816961
Change-Id: Ib98f96f4cd7f33414c0b5a6037bfb881938cc15e
Reviewed-on: https://chromium-review.googlesource.com/939767
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51637}
gcc and clang (and the standard) don't allow implicit conversion of
function pointers to object pointers. MSVC does allow that, and since
system headers require this to work, clang-cl allows it too -- but
it emits a -Wmicrosoft-cast warning (which we currently suppress in
the Chromium build, but which we want to enable.)
As a side effect, when printing a function pointer to a stream, MSVC
(and clang-cl) will pick the operator<<(void*) overload, while gcc
and clang will pick operator<<(bool) since the best allowed conversion
they find is from function pointer to bool.
To prevent the clang-cl warning, we need to make sure that we never
directly print a function pointer to a stream. In v8, this requires
two changes:
1. Give PrintCheckOperand() an explicit specialization for function
pointers and explicitly cast to void* there. This ports
https://codereview.chromium.org/2515283002/ to V8, and also fixes a
bug on non-Windows where DCHECK() of function pointers would print
"(1 vs 1)" instead of the function's addresses.
(The bug remains with member function pointers,
where it's not clear what to print instead of the 1.)
2. has_output_operator<T> must not use operator<< on its argument
in an evaluated context if T is a function pointer. This patch
modifies has_output_operator<> to use an unevaluated context instead,
which is simpler than the current approach (and matches what Chromium's
base does), but changes behavior in minor (boring) ways
(see template-utils-unittest.cc), since operator<<() is now
called with a temporary and only operator<<() implementations callable
with a temporary are considered.
A more complicated but behavior-preserving alternative would be to
add an explicit specialization for function pointers. You can see
this variant in patch set 1 on gerrit.
Bug: chromium:550065
Change-Id: Idc2854d6c258b7fc0b959604006d8952a79eca3d
Reviewed-on: https://chromium-review.googlesource.com/940004
Commit-Queue: Nico Weber <thakis@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51636}
Due to a recent refactoring the function EnsureEventLoopInitialized on
the default platform became obsolete. It does not contain a single line
of code. With this CL we prepare the removal of this function from the
V8 platform API.
R=rmcilroy@chromium.org
Bug: v8:7310
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng;master.tryserver.chromium.linux:linux_chromium_rel_ng
Change-Id: If4d54cd989f8df2f40b322be3b67bb8a482398d0
Reviewed-on: https://chromium-review.googlesource.com/934221
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51635}
This brings us closer to our goal of deprecating the `construct_stub`
field in `SharedFunctionInfo`.
Bug: v8:7503
Change-Id: I9502246f5e463e487a989ab7f8c96c008f0fa9d9
Reviewed-on: https://chromium-review.googlesource.com/941143
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51634}
This brings us closer to our goal of deprecating the `construct_stub`
field in `SharedFunctionInfo`.
Bug: v8:7503
Change-Id: Id7bee0d61830f7bcf2cad4f6e50ff9f24b58a51b
Reviewed-on: https://chromium-review.googlesource.com/939790
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51633}
This CL also reorganizes the Strings test suite
Bug: v8:7340
Change-Id: I54d4d76a16c362e38ebfc9719ac8cb1a490ef3cc
Reviewed-on: https://chromium-review.googlesource.com/941122
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51631}
Next step on deprecating the construct_stub field in the
SharedFunctionInfo.
Bug: v8:7503
Change-Id: Ibcbf9dc0dbcbcf0e26c103167e8a30074cbdd888
Reviewed-on: https://chromium-review.googlesource.com/940942
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51630}
This brings us closer to our goal of deprecating the `construct_stub`
field in `SharedFunctionInfo`.
Bug: v8:7503
Change-Id: I0e995e6ed9e7c04990f76dd8fe8afaaa8d38294b
Reviewed-on: https://chromium-review.googlesource.com/940124
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#51629}
