This refactoring is a big step towards separating Turbofan-related code
from backend independent code. This will allow us to include way less
headers from "src/compiler" at various places.
The {WasmCompilationUnit} contained information for Turbofan
compilation, and for Liftoff compilation. This CL tears this apart, such
that {WasmCompilationUnit} holds backend-independent information, plus
a pointer to either {LiftoffCompilationUnit} or
{TurbofanWasmCompilationUnit}. These pointers are opaque, so that
{function-compiler.h}, defining {WasmCompilationUnit}, does not need to
include any Turbofan specific or Liftoff specific headers.
R=ahaas@chromium.org, titzer@chromium.org, mstarzinger@chromium.org
Bug: v8:7570, v8:6600
Change-Id: I024c9a23508ee1b4b3cbe1d068c8e785d732daca
Reviewed-on: https://chromium-review.googlesource.com/1016640
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52684}
This reverts commit f52b4b3b81.
Reason for revert: Makes benchmarks time out:
https://crbug.com/834655
Original change's description:
> [compiler] Temporarily turn some DCHECKs into CHECKs.
>
> This turns most DCHECKs in the register allocator code into CHECKs,
> which hopefully will tell us more about the crashes in crbug 831822.
>
> This CL will eventually be reverted again.
>
> Bug: chromium:831822
> Change-Id: I123ed507949ecab50bb0aaaf9e91978c9c2d8d65
> Reviewed-on: https://chromium-review.googlesource.com/1016762
> Commit-Queue: Georg Neis <neis@chromium.org>
> Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#52673}
TBR=neis@chromium.org,sigurds@chromium.org
Change-Id: Ic2b37a3ce04516c5871b801015153ce84e622e90
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:831822
Reviewed-on: https://chromium-review.googlesource.com/1018860
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52679}
Since external strings are used for things like source strings, we
should tenure them from creation.
Change-Id: I226ab9036836d76d8c17ed168ad97d7f0f824278
Reviewed-on: https://chromium-review.googlesource.com/1006961
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52677}
This turns most DCHECKs in the register allocator code into CHECKs,
which hopefully will tell us more about the crashes in crbug 831822.
This CL will eventually be reverted again.
Bug: chromium:831822
Change-Id: I123ed507949ecab50bb0aaaf9e91978c9c2d8d65
Reviewed-on: https://chromium-review.googlesource.com/1016762
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52673}
This CL adds sort benchmarks for BigUint64Array and BigInt64Array.
They form their own benchmark group separate from the Int and
Float types/element kinds.
R=jgruber@chromium.org
Bug: v8:7382
Change-Id: Id3ce125be36a42c419a2632edc5f930547a1720e
Reviewed-on: https://chromium-review.googlesource.com/1016643
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Simon Zünd <szuend@google.com>
Cr-Commit-Position: refs/heads/master@{#52668}
In ArrayBuiltinsAssembler::VisitAllFastElementsOneKind(), we enumerate
an arrays elements, carefully checking for the "hole" when required.
This code is only called for arrays whose prototype is the initial array
prototype. And the path is only available when the initial array
prototype is free of elements. Since that's the case, we only need to
verify that the initial array prototype remains free of elements during
an iteration with javascript callbacks. We don't need a body of code
that can walk the prototype chain looking for elements visible through
the "hole" value. In practice, this code was never run.
Change-Id: Iba5e275c559d495aa1cf6a4f29d66e2ce475c981
Reviewed-on: https://chromium-review.googlesource.com/1015023
Commit-Queue: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52660}
There are various situations where we explicitly compare a SMI against
another SMI (e.g., BuildIndexedJump). This is also a common pattern for
generated code (e.g., comparing a loop variable with an integer). Instead
of using the generic equality/strict-equality stub for this, which is
expensive, this CL offers a simple comparison stub, repurposing the
TestEqualStrictNoFeedback bytecode to TestReferenceEqual
Bug: v8:5310
Change-Id: Ib2b47cd24d5386cf0d20d3bd794776dc6e3a02a5
Reviewed-on: https://chromium-review.googlesource.com/1007542
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Christian O. Andersson <cricke@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52655}
Previously, if an unbound script was created in a non-inspected context,
but later bound to an inspected one, it never appeared in the
debugger sources.
After this change `OnAfterCompile` will be invoked not on the original
script compilation, but when it's actually bound to a context for
execution, which means `Debugger.scriptParsed` will be now sent to the
inspector even for such precompiled scripts.
R=jgruber@chromium.org, kozyatinskiy@chromium.org, yangguo@chromium.org
Bug: v8:7654
Change-Id: Ice13312e425903fb2baf14edab5c566d649a6438
Reviewed-on: https://chromium-review.googlesource.com/1013581
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52652}
Class fields needs to be initialized after `this` is bound, as per the
new spec change:
https://github.com/tc39/proposal-class-fields/pull/92
This CL moves the initialization of `this` from parser desugaring to
the bytecode generator.
Bug: v8:7647
Change-Id: I20f749403e5a4d2f06a39726cf39012ceb541987
Reviewed-on: https://chromium-review.googlesource.com/1014383
Reviewed-by: Mythri Alle <mythria@chromium.org>
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52646}
Throw a TypeError if the length of a concat-spreadable object makes the
total length too large, as specified.
Bug: v8:7652
Change-Id: Ie3f694d64c949703edd733c0310cfb3f64b78a15
Reviewed-on: https://chromium-review.googlesource.com/1013714
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52644}
Add a new permission kRead to PageAllocator::Permission and
OS::MemoryPermission and implement it in platform-*.
Not used yet, because it needs corresponding changes in chromium.
Bug: v8:7464
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I9f84251eff593536cbcc1cde04641d696c79d65c
Reviewed-on: https://chromium-review.googlesource.com/1006756
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52636}
Casting from a floating-point type to an integer type is undefined behavior
if the integral part of the float cannot be represented in the range of the
int.
Bug: v8:3770, chromium:831145
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng;luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I2e85ea8b0f09bbeeb3e0dcc1135fc747fa312f6d
Reviewed-on: https://chromium-review.googlesource.com/1011651
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52631}
Now that tables and stack frames properly root instances, there is no
longer any need to disallow mutations that could unroot instances
while their code is on the stack.
Bug: v8:7232
Change-Id: I907b9522ac12ad7a67fb4124774713b6b3b40bb7
Reviewed-on: https://chromium-review.googlesource.com/1007004
Commit-Queue: Ben Titzer <titzer@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52629}
If the new length is too large, we must throw a TypeError.
Bug: v8:7652
Change-Id: I47268c04405f7a5f5bbc971cd434f2d786af9ca1
Reviewed-on: https://chromium-review.googlesource.com/1013563
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52624}
First version which can compile a very basic code.
Change-Id: I3b98412a5ca39a28f8fe5b60516b82c6981dd187
Reviewed-on: https://chromium-review.googlesource.com/993232
Commit-Queue: Vincent Belliard <vincent.belliard@arm.com>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52622}
The embedder should not need to keep track of the source string.
R=jgruber@chromium.org
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Ie27df755a22fbcae7b6e87a435419d2d8f545558
Reviewed-on: https://chromium-review.googlesource.com/1013482
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52614}
This adds tests for 'oddly' behaving comparison functions.
I.e. functions that cause an element kind change and/or
modify the array. The tests check that sort does not crash in these
instances.
R=jgruber@chromium.org
Bug: v8:7382
Change-Id: I4ac9aa081fda9088d1848a960dc66aba671872e5
Reviewed-on: https://chromium-review.googlesource.com/1010062
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52612}
Per the spec change at [1], Abstract Relational Comparison between a
BigInt and a String converts the String to BigInt via StringToBigInt
before performing the comparison. Before this change, the String was
converted to a Number, and a BigInt/Number comparison was performed.
[1] https://github.com/tc39/proposal-bigint/pull/139
Bug: v8:6791
Change-Id: I40b4f4ddc78977adb0d44180eb58e0f9a8a70cb6
Reviewed-on: https://chromium-review.googlesource.com/1004117
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52609}
The "Address" type is V8's general-purpose type for manipulating memory
addresses. Per the C++ spec, pointer arithmetic and pointer comparisons
are undefined behavior except within the same array; since we generally
don't operate within a C++ array, our general-purpose type shouldn't be
a pointer type.
Bug: v8:3770
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng;master.tryserver.blink:linux_trusty_blink_rel
Change-Id: Ib96016c24a0f18bcdba916dabd83e3f24a1b5779
Reviewed-on: https://chromium-review.googlesource.com/988657
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52601}
This allows an embedder to check if a Value is a module namespace object.
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Idffceff451dd5f5c6a53d4cb3ce02c1c2c5b653c
Reviewed-on: https://chromium-review.googlesource.com/1011762
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52597}
This patch moves the desugaring from the parser to the bytecode
generator for super calls that have a spread at a non last position.
This allows us to have the post super() call behavior, such as
initializing instance fields in one place in VisitCallSuper.
Bug: v8:7642
Change-Id: I00a693beb7078a63282359c1121b66bb62c157c8
Reviewed-on: https://chromium-review.googlesource.com/1009907
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Mythri Alle <mythria@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52596}
This is mostly a simple copy & paste of the stub implementation from
code-stubs-arch.cc to builtins-arch.cc.
The conversion allows removal of a special case for the DoubleToIStub
within the compiler & wasm pipelines, and also makes the following
builtins isolate-independent (in conjunction with
https://crrev.com/c/1006581):
TFC BitwiseAnd
TFC BitwiseOr
TFC BitwiseXor
TFC Exponentiate
TFC ShiftLeft
TFC ShiftRight
TFC ShiftRightLogical
TFJ AtomicsAdd
TFJ AtomicsAnd
TFJ AtomicsCompareExchange
TFJ AtomicsExchange
TFJ AtomicsLoad
TFJ AtomicsOr
TFJ AtomicsStore
TFJ AtomicsSub
TFJ AtomicsXor
TFJ MathClz32
TFJ MathImul
TFJ MathPow
TFJ NumberParseInt
TFJ StringFromCharCode
TFJ TypedArrayFrom
TFJ TypedArrayOf
TFJ TypedArrayPrototypeMap
Drive-by: dead code removal & TODOs in code-stubs.h.
Bug: v8:6666
Change-Id: I763cba2242bcadc2d130b0aaa16a9787212b466a
Reviewed-on: https://chromium-review.googlesource.com/1012024
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52591}
This patch adds a micro-benchmark comparing
`string.startsWith(singleCodeUnit)`, `string[0]`,
`string.endsWith(singleCodeUnit)`, and `string[string.length - 1]`.
The benchmark can be used to measure any String#{starts,ends}With
optimizations we implement in the future.
Test:
tools/run_perf.py --binary-override-path=out/x64.release/d8 \
--filter=JSTests/Strings/StringStartsEndsWithComparison \
--extra-flags=--trace-turbo test/js-perf-test/JSTests.json
Bug: v8:7453
Change-Id: I68cad197fbcbfc6b1938fc437776c319ee9f81df
Reviewed-on: https://chromium-review.googlesource.com/1011619
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52589}
This changes DoubleToIStub to return its result on the stack instead
of a specific return register.
In a follow-up, the DoubleToIStub could be converted into a builtin.
Bug: v8:6666
Change-Id: I7852e1586c8f7b56bc5d2545a7bf6238dd2ad650
Reviewed-on: https://chromium-review.googlesource.com/1009702
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52587}
It is not safe to assume the first match is a string just
because the RegExp result is fast.
Bug: chromium:831943
Change-Id: Idd40f8b75312f0be54f45f626dc017339033abc6
Reviewed-on: https://chromium-review.googlesource.com/1009325
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Peter Wong <peter.wm.wong@gmail.com>
Cr-Commit-Position: refs/heads/master@{#52578}
This adds benchmarks that sort all element kinds with multiple comparison
functions. This also adds benchmarks that cause the element kind of
the array to change after x amount of comparisons.
The last set of benchmarks that are added, measure performance on
pre-sorted arrays.
R=jgruber@chromium.org, petermarshall@chromium.org
Bug: v8:7382
Change-Id: I620de37f4a28f8096682bb708ea4f7a9a21d5316
Reviewed-on: https://chromium-review.googlesource.com/1009602
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52574}
Looking up line numbers with the JITLineInfoTable would sometimes give
wrong answers. Fix these bugs and add a cctest for this data structure.
Also do some cleanup while we're here like inlining the (empty)
constructor and destructor and removing the empty() method which is
only used unnecessarily anyway, to make the contract of
GetSourceLineNumber a bit clearer.
Also rename the data structure to SourcePositionTable, because it
doesn't just provide info for JIT code, but also bytecode, and 'Info'
is pretty ambiguous.
Bug: v8:7018
Change-Id: I126581c844d85df6b2b3f80f2f5acbce01c16ba1
Reviewed-on: https://chromium-review.googlesource.com/1006795
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52571}
Fixes some failing test262 tests for a corner-case in the spec
where we need to use the buffer constructor from a different realm.
Bug: v8:7512
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I3f1334f6181eaaddf0326156139ac20a970c235b
Reviewed-on: https://chromium-review.googlesource.com/966223
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52560}
It's not possible to use GeneratedCode for this test because the called
stubs use the stdcall calling convention but GeneratedCode does not allow
specifying a custom calling convention.
BUG=v8:7164
Change-Id: Ic28c4313bb2b68103b94e4c1ae7c037a789edce2
Reviewed-on: https://chromium-review.googlesource.com/1008994
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52558}
This is to fix an assertion failure in formatToParts when
Chinese calendar is specified with 'u-ca-chinese'.
See https://github.com/tc39/ecma402/issues/225 . This CL
is a temporary work-around to get v8 match the spec in terms
of the external behavior, but it's not taking the steps in
the spec, yet.
Moreover, the spec may have to be revised as to how to pick the best
match pattern when the default calendar for a locale is different from
the calendar specified via 'u-ca'. How to handle 'related year' part
also needs to be specified.
Bug: chromium:826549
Test: intl/date-format/format-with-extensions
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I1f9a2467e86e71e024bc9babe18f16e49746008e
Reviewed-on: https://chromium-review.googlesource.com/1006915
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Jungshik Shin <jshin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52556}
When using registers during the Liftoff-prologue, we need to make sure
that all reserved registers are correctly pushed to and restored
from stack.
Change-Id: Iac444448cfd99fca70a811cb941d0cf5979d638b
Reviewed-on: https://chromium-review.googlesource.com/1005754
Commit-Queue: Kim-Anh Tran <kimanh@google.com>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52555}
Previously embedder had to create an instance of TracingCpuProfiler explicitly.
The patch makes the profiler created automatically for every isolate.
The profiler has no overhead unless tracing with v8.cpu_profiler category is enabled.
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I9369c2c56bcddc72093eda33dc2bc185c9253b4a
Reviewed-on: https://chromium-review.googlesource.com/1006049
Commit-Queue: Alexei Filippov <alph@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52552}
This CL allows SetPrototypeAdd and ArrayIteratorPrototypeNext
to be called on temporary objects during side effect free evaluation.
Bug: v8:7588
Change-Id: Id77848e48d98c243de91bc6c0fae5a0877e693d4
Reviewed-on: https://chromium-review.googlesource.com/998439
Commit-Queue: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52548}
