Turbofan is a lot slower than Crankshaft at constructing TypedArrays,
because we always go to the C++ builtin. Port the builtin to CSA
to improve performance, and to clean up the implementation, which is
split across multiple files and pieces at the moment.
This CL increases the performance with --future to roughly the same
as with crankshaft.
BUG=v8:5977
Change-Id: I5a4c4b544a735a56290b85bf33c2f3718df7e2b8
Reviewed-on: https://chromium-review.googlesource.com/445717
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43518}
This fixes a corner-case in the above predicate that was introduced to
allow fully disabling optimization using %NeverOptimizeFunction. This
property of a function is a transient property (i.e. changes over time),
whereas {UseTurboFan} is designed to be a static property (i.e. gives
same answer over time). Violating this led to cases where functions got
optimization disabled for other reasons would suddenly be baselined.
The correct place to check transient properties is when optimization is
requested.
R=jarin@chromium.org
TEST=mjsunit/never-baseline
Change-Id: I37eb0c70d2b39704be29fd4bda76975bfbede66b
Reviewed-on: https://chromium-review.googlesource.com/447937
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43514}
String::SlowFlatten assumed that ConsStrings with empty first parts have
flattened strings as their second part. TurboFan, however, can create
ConsStrings with empty first parts and arbitrary second parts. With
this CL we call String::Flatten on the second part of a ConsString if
the first part is empty, but only when String::Flatten would not call
String::SlowFlatten.
R=jkummerow@chromium.org
BUG=chromium:696651
Change-Id: I9acb681de1be695e1ec2f6f6d28b9e4dc4344e98
Reviewed-on: https://chromium-review.googlesource.com/448457
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43513}
This CL fixes unused lambda captures for big endian targets when Clang
is used for build V8.
TEST=
BUG=
Review-Url: https://codereview.chromium.org/2725613002
Cr-Commit-Position: refs/heads/master@{#43511}
The maybe-assigned flag of the catch variable was not preserved.
BUG=v8:5636,chromium:696332
Change-Id: I9c55e1b1312bdebc53bc45bc3ca1c982bdbe9846
Reviewed-on: https://chromium-review.googlesource.com/447680
Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43506}
BUG=v8:6022
Change-Id: I54205cb3ecc2dd31ed62e55726f0ec5fcd202c30
Reviewed-on: https://chromium-review.googlesource.com/446349
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43500}
Previously, we over-approximated Scope::scope_calls_eval_ in
arrow functions: if either the outer scope or the arrow function
parameters had a direct eval call, we marked both scopes as calling
eval. This over-approximation kept getting us into trouble, though,
especially when eager or lazy parsing would disagree about the
"calls eval" bit.
This patch instead tracks eval calls accurately, using a boolean on
Scope::Snapshot that is reset as appropriately depending on whether
a particular AssignmentExpression turned out to be an arrow parameter
list or not.
BUG=chromium:691687
Change-Id: I527dc59b4d32a2797805ff26dc9f70b1311377b2
Reviewed-on: https://chromium-review.googlesource.com/446094
Commit-Queue: Adam Klein <adamk@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43499}
This is more useful than always returning undefined.
BUG=v8:1569,v8:5978
Change-Id: Id10cf87f7865db1a85de412460eaead4e4bf3b62
Reviewed-on: https://chromium-review.googlesource.com/446846
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43498}
The order in which things were done wasn't quite correct and lead
to wrong behaviour for certain circular module graphs.
BUG=v8:1569,chromium:694566
Change-Id: I291186e261268c853a30ad891ff362904e0b28ef
Reviewed-on: https://chromium-review.googlesource.com/447399
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43497}
With the params (a, b, ...c) the param / variable declaration order used to be
"temp, temp, c, a, b". Now it is "temp, temp, a, b, c" as you'd expect. This
makes it easier for PreParser to match the parameter order of Parser.
R=verwaest@chromium.org
BUG=v8:5516
Change-Id: I79da04ef3f812bf52c032bed6263c009fecb7988
Reviewed-on: https://chromium-review.googlesource.com/447677
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43490}
Reason for revert:
With the Windows bots fixed in https://chromium-review.googlesource.com/445786 , this should be good to reland. Thanks, Michael!
Original issue's description:
> Revert of [test] Speculatively remove local-tests from archive (patchset #2 id:20001 of https://codereview.chromium.org/2643983002/ )
>
> Reason for revert:
> Breaks all windows bots:
> https://build.chromium.org/p/client.v8/builders/V8%20Win32%20-%20debug/builds/6811
>
> Original issue's description:
> > [test] Remove local-tests from test262 archive and add to .isolate
> >
> > This might help fix the bots, which are broken in e.g.,
> > https://build.chromium.org/p/tryserver.v8/builders/v8_mac_rel_ng_triggered/builds/14011
> >
> > The archive was added in order to transmit test262 tests more rapidly.
> > It doesn't serve much of a purpose for local-tests. I naively added
> > local-tests there out of symmetry. However, the BUILD.gn file does not
> > regenerate an archive when files are only deleted and not added or
> > changed. Since the performance concern is not present for the small
> > volume of local-tests, this patch reverts to the more normal mechanism
> > for sending over dependencies, with test262.isolate.
> >
> > R=adamk
> >
> > Review-Url: https://codereview.chromium.org/2643983002
> > Cr-Commit-Position: refs/heads/master@{#42485}
> > Committed: 9f545ea96f
>
> TBR=adamk@chromium.org,littledan@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Review-Url: https://codereview.chromium.org/2640223003
> Cr-Commit-Position: refs/heads/master@{#42491}
> Committed: 4ffe0850dbTBR=adamk@chromium.org,machenbach@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
Review-Url: https://codereview.chromium.org/2725643002
Cr-Commit-Position: refs/heads/master@{#43488}
Unibrow is currently at Unicode version 7.0.0, which does not
include mongolian vowel separator (\u180E) as white space. In
order to appease test262 at the time however we kept it as a
whitespace.
Test262 has since then been updated. And while this is not an
update of unibrow, we are removing \u180E as white space here.
R=jshin@chromium.org, littledan@chromium.org
BUG=v8:5155
Review-Url: https://codereview.chromium.org/2720953003
Cr-Commit-Position: refs/heads/master@{#43485}
This relaxes the check if the test262 archive should be unpacked or not.
A bug in the swarming isolate processor for windows makes the old check for the data directory pass and not unpack.
With this work-around we actually start running tests on windows.
This also fixes the regexp for exception extraction and handles the windows drive letter.
It also strips whitespace to fix a case where a carriage return was included on windows.
BUG=v8:5872
Change-Id: I363925665b0bad7530a1f93a2ea8f39b056d4174
Reviewed-on: https://chromium-review.googlesource.com/445786
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Daniel Ehrenberg <littledan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43482}
The x64 side is included in https://chromium-review.googlesource.com/c/444226/
BUG=v8:5294
Change-Id: Ie255604c5e38c72e3c2b76e1ca3557a5fde108ee
Reviewed-on: https://chromium-review.googlesource.com/446394
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43481}
The above intrinsic by now has to perform a check whether the prototype
of a derived constructor is actually a constructor function itself. This
is done as part of the {JSGetConstructorCall} operator. The intrinsic
should just reduce down to the operator to maintain correct semantics.
R=bmeurer@chromium.org
TEST=mjsunit/regress/regress-crbug-696622
BUG=chromium:696622
Change-Id: Ia19c188f17ad16b12248db1f01a01b8d7258499b
Reviewed-on: https://chromium-review.googlesource.com/447716
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43479}
The new NewUnmappedArgumentsElements node now takes two inputs:
- the frame holding the arguments (current frame or arguments adaptor frame)
- the length of the suffix of passed arguments to be copied into the backing store
These inputs are computed with two new node types:
ArgumentsFrame()
ArgumentsLength[formal_parameter_count,is_rest_length](Node* arguments_frame)
The node type NewRestParameterElements can now be expressed with NewUnmappedArgumentsElements and an appropriate length and is thus not needed anymore.
In escape analysis, we lower loads from the length field of NewUnmappedArgumentsElements with its length input and if we find out that no write access to the arguments elements exists, we replace element loads with direct stack access and replace the NewUnmappedArgumentsElements node with a node of the new node type ArgumentsElementsState. This corresponds to an ObjectState node and gets translated into a deoptimizer instruction to allocate the backing store. Together with the already existing deoptimizer support for the actual arguments object/rest parameters, this allows to remove all allocations for arguments objects/rest parameters in this case.
In the deoptimizer, we read the actual parameters from the stack while transforming the static deopt info into TranslatedValue objects.
If escape analysis cannot remove the backing store allocation, NewUnmappedArgumentsElements gets lo
BUG=v8:5726
Review-Url: https://codereview.chromium.org/2692753004
Cr-Commit-Position: refs/heads/master@{#43475}
This makes it no longer necessary to ensure that V8 and Blink have non-colliding
tags, which makes it easier for them to evolve independently, and also makes
the wire format more suitable for other V8 embedders, who would not
necessarily be surveyed before V8 introduced a new tag that might collide
with theirs.
BUG=chromium:686159
Review-Url: https://codereview.chromium.org/2709023003
Cr-Commit-Position: refs/heads/master@{#43466}
perf regression. See crbug.com/695653 for more info.
Revert "[SAB] Move Atomics builtins to C++"
This reverts commit 2b9840d86f.
Revert "[SAB] Remove unreachable Uint8Clamped atomics paths"
This reverts commit d1160fb14f.
Revert "Remove tiny unit test for MinSimple/MaxSimple"
This reverts commit 837760ecb7.
Revert "Remove infrastructure for experimental JS natives"
This reverts commit 8cfe45b6f1.
BUG=695653
TBR=hablich@chromium.org
Review-Url: https://codereview.chromium.org/2715223003
Cr-Commit-Position: refs/heads/master@{#43462}
When an instance of a constructor goes dictionary mode, this changes the initial map
of that constructor to also be in dictionary mode. This avoids spurious hidden class
creation, that also results in IC misses.
BUG=
Change-Id: I0e70f822ac345d0224f2092ec473621a603d4cc5
Reviewed-on: https://chromium-review.googlesource.com/446361
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43452}
This extends the existing typer for JS-level comparisons to also be
applied on simplified-level number comparisons. Now that nodes having
such operators are introduced early during graph construction, precise
type information has become more important.
R=jarin@chromium.org
TEST=unittests/TyperTest
Change-Id: Ibcfda36cadf72381b3c63138dd8ad2f6b1f7fbba
Reviewed-on: https://chromium-review.googlesource.com/445879
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43447}
Nodes having the {Throw} operator are just used as terminators for
control-flow leaving the function body with an exception completion,
they turn into basic-block terminators within the schedule. Actually
raising an exception is done solely via runtime calls. Hence {Throw}
nodes no longer need any value input.
R=bmeurer@chromium.org
Change-Id: Id6d8e46b12c4b84f4e7a8ac96623c8efb1f27b26
Reviewed-on: https://chromium-review.googlesource.com/446501
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#43445}
