platform/upstream/v8.git
9 years ago[test] Fix missing heartbeats in test runner.
machenbach [Fri, 5 Jun 2015 09:50:53 +0000 (02:50 -0700)]
[test] Fix missing heartbeats in test runner.

TBR=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/1163523008

Cr-Commit-Position: refs/heads/master@{#28808}

9 years agoUpdate V8 DEPS.
v8-autoroll [Fri, 5 Jun 2015 08:10:08 +0000 (01:10 -0700)]
Update V8 DEPS.

Rolling v8/third_party/icu to 9939a5d5314b6d59d5fb070902d73304c2482f88

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1168573003

Cr-Commit-Position: refs/heads/master@{#28807}

9 years ago[es6] super.prop, eval and lazy functions
arv [Thu, 4 Jun 2015 21:16:18 +0000 (14:16 -0700)]
[es6] super.prop, eval and lazy functions

We used to only store the uses_super_property in the preparse data
logger. Let the logger use NeedsHomeObject instead.

BUG=v8:3768
LOG=N
R=wingo, adamk

Review URL: https://codereview.chromium.org/1164073003

Cr-Commit-Position: refs/heads/master@{#28806}

9 years agoStage ES6 Array and TypedArray methods
adamk [Thu, 4 Jun 2015 20:08:27 +0000 (13:08 -0700)]
Stage ES6 Array and TypedArray methods

Also stages ES6 @@isConcatSpreadable support.

BUG=v8:3578
LOG=y

Review URL: https://codereview.chromium.org/1148983006

Cr-Commit-Position: refs/heads/master@{#28805}

9 years agoUnship harmony tostring
arv [Thu, 4 Jun 2015 17:33:20 +0000 (10:33 -0700)]
Unship harmony tostring

Blink's DOM bindings are not yet ready.

BUG=493137, 239915
LOG=N
R=rossberg, adamk
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_chromium_rel_ng;tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1153263004

Cr-Commit-Position: refs/heads/master@{#28804}

9 years agoImplement %TypedArray%.prototype.{reduce,reduceRight}
dehrenberg [Thu, 4 Jun 2015 16:36:37 +0000 (09:36 -0700)]
Implement %TypedArray%.prototype.{reduce,reduceRight}

This patch re-commits a previous commit after fixing a test to not
run into another known bug.
This reverts commit b104a67ef0a50beab2dc85a6ac5732c696d07698.

This patch implements the last two methods on TypedArrays. These
were previously committed and led to a test failure.

BUG=v8:3578
LOG=Y
R=adamk

Review URL: https://codereview.chromium.org/1159663004

Cr-Commit-Position: refs/heads/master@{#28803}

9 years agoRefactor lexical home object binding
arv [Thu, 4 Jun 2015 16:22:29 +0000 (09:22 -0700)]
Refactor lexical home object binding

Before this we had 3 super related lexical bindings that got injected
into method bodies: .home_object, .this_function,  and new.target.
With this change we get rid of the .home_object one in favor of using
.this_function[home_object_symbol] which allows some simplifications
throughout the code base.

BUG=v8:3768
LOG=N
R=adamk@chromium.org, wingo@igalia.com

Review URL: https://codereview.chromium.org/1154103005

Cr-Commit-Position: refs/heads/master@{#28802}

9 years agoAdd support for Embedded Constant Pools for PPC and Arm
mbrandy [Thu, 4 Jun 2015 14:44:00 +0000 (07:44 -0700)]
Add support for Embedded Constant Pools for PPC and Arm

Embed constant pools within their corresponding Code
objects.

This removes support for out-of-line constant pools in favor
of the new approach -- the main advantage being that it
eliminates the need to allocate and manage separate constant
pool array objects.

Currently supported on PPC and ARM.  Enabled by default on
PPC only.

This yields a 6% improvment in Octane on PPC64.

R=bmeurer@chromium.org, rmcilroy@chromium.org, michael_dawson@ca.ibm.com
BUG=chromium:478811
LOG=Y

Review URL: https://codereview.chromium.org/1162993006

Cr-Commit-Position: refs/heads/master@{#28801}

9 years agoFix more -Wsign-compare bugs with GCC 4.9.2
Djordje.Pesic [Thu, 4 Jun 2015 11:45:08 +0000 (04:45 -0700)]
Fix more -Wsign-compare bugs with GCC 4.9.2

Review URL: https://codereview.chromium.org/1152993005

Cr-Commit-Position: refs/heads/master@{#28800}

9 years agoUpdate V8 DEPS.
v8-autoroll [Thu, 4 Jun 2015 03:26:22 +0000 (20:26 -0700)]
Update V8 DEPS.

Rolling v8/tools/clang to b7e158c9336030c1527dfbce92656a2d7b8cfb60

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1150313007

Cr-Commit-Position: refs/heads/master@{#28799}

9 years agoFix more -Wsign-compare bugs with GCC 4.9.2.
paul.lind [Thu, 4 Jun 2015 01:28:14 +0000 (18:28 -0700)]
Fix more -Wsign-compare bugs with GCC 4.9.2.

BUG=

Review URL: https://codereview.chromium.org/1167613004

Cr-Commit-Position: refs/heads/master@{#28798}

9 years agoAdd new Float32x4 type for SIMD.js.
bbudge [Wed, 3 Jun 2015 23:56:21 +0000 (16:56 -0700)]
Add new Float32x4 type for SIMD.js.

LOG=N
BUG=v8:4124

Review URL: https://codereview.chromium.org/1153373003

Cr-Commit-Position: refs/heads/master@{#28797}

9 years agoImplement Atomics API
binji [Wed, 3 Jun 2015 17:58:29 +0000 (10:58 -0700)]
Implement Atomics API

This is behind the flag "--harmony-atomics", and it only works on
SharedArrayBuffers. This implementation only includes the runtime functions.
The TurboFan implementation will be next.

The draft spec for Atomics can be found here:
https://docs.google.com/document/d/1NDGA_gZJ7M7w1Bh8S0AoDyEqwDdRh4uSoTPSNn77PFk

BUG=
LOG=n

Review URL: https://codereview.chromium.org/1162503002

Cr-Commit-Position: refs/heads/master@{#28796}

9 years agoMIPS64: Fix '[es6] Super call in arrows and eval'.
balazs.kilvady [Wed, 3 Jun 2015 17:57:13 +0000 (10:57 -0700)]
MIPS64: Fix '[es6] Super call in arrows and eval'.

Port 4b8051a02ae9f765b7d803fc7986301abde36007

Original commit message:
This splits the SuperReference AST node into SuperPropertyReference and
SuperCallReference. The super call reference node consists of three
unresolved vars to this, new.target and this_function. These gets
declared when the right function is entered and if it is in use. The
variables gets assigned in FullCodeGenerator::Generate.

This is a revert of the revert 88b1c9170a0293cbcc8bdaf57fbe12744b48d7e8

BUG=v8:3768
LOG=N

Review URL: https://codereview.chromium.org/1160973008

Cr-Commit-Position: refs/heads/master@{#28795}

9 years agoPPC: [date] Refactor the %_DateField intrinsic to be optimizable.
mbrandy [Wed, 3 Jun 2015 17:21:57 +0000 (10:21 -0700)]
PPC: [date] Refactor the %_DateField intrinsic to be optimizable.

Port e4782a9b468258344d512a5f7dadbf1584928849

Original commit message:
Previously the %_DateField intrinsic would also check the object and
throw an exception if you happen to pass something that is not a valid
JSDate, which (a) violates our policy for instrinsics and (b) is hard to
optimize in TurboFan (even Crankshaft has a hard time, but there we will
never inline the relevant builtins, so it doesn't show up). The throwing
part is now a separate intrinsics %_ThrowIfNotADate that throws an
exception in full codegen and deoptimizes in Crankshaft, which means the
code for the current use cases is roughly the same (modulo some register
renamings/gap moves).

R=bmeurer@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1170463002

Cr-Commit-Position: refs/heads/master@{#28794}

9 years agoFlatten the Arrays returned and consumed by the v8::Map API
adamk [Wed, 3 Jun 2015 16:32:52 +0000 (09:32 -0700)]
Flatten the Arrays returned and consumed by the v8::Map API

This will significantly simplify the serialization code, as well
as speeding it up (by triggering only a single allocation instead of O(size)
allocations).

BUG=chromium:478263
LOG=y

Review URL: https://codereview.chromium.org/1157843006

Cr-Commit-Position: refs/heads/master@{#28793}

9 years agoFixed noi18n build.
ishell [Wed, 3 Jun 2015 16:23:01 +0000 (09:23 -0700)]
Fixed noi18n build.

Review URL: https://codereview.chromium.org/1159553011

Cr-Commit-Position: refs/heads/master@{#28792}

9 years agoPPC: [es6] Super call in arrows and eval
mbrandy [Wed, 3 Jun 2015 15:50:43 +0000 (08:50 -0700)]
PPC: [es6] Super call in arrows and eval

Port 4b8051a02ae9f765b7d803fc7986301abde36007

Original commit message:
This splits the SuperReference AST node into SuperPropertyReference and
SuperCallReference. The super call reference node consists of three
unresolved vars to this, new.target and this_function. These gets
declared when the right function is entered and if it is in use. The
variables gets assigned in FullCodeGenerator::Generate.

This is a revert of the revert 88b1c9170a0293cbcc8bdaf57fbe12744b48d7e8

R=arv@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1146403010

Cr-Commit-Position: refs/heads/master@{#28791}

9 years agoFixed memory-leak in d8. It did not clean evaluation context used for executing shell...
ishell [Wed, 3 Jun 2015 14:34:50 +0000 (07:34 -0700)]
Fixed memory-leak in d8. It did not clean evaluation context used for executing shell commands.

BUG=chromium:493284
LOG=N

Review URL: https://codereview.chromium.org/1147343004

Cr-Commit-Position: refs/heads/master@{#28790}

9 years agoMicro benchmark for Try-Catch-Finally
hablich [Wed, 3 Jun 2015 12:12:20 +0000 (05:12 -0700)]
Micro benchmark for Try-Catch-Finally

BUG=v8:4131
LOG=n

Review URL: https://codereview.chromium.org/1155493007

Cr-Commit-Position: refs/heads/master@{#28789}

9 years agoFix arrow functions requiring context without slots.
mstarzinger [Wed, 3 Jun 2015 11:32:25 +0000 (04:32 -0700)]
Fix arrow functions requiring context without slots.

This fixes a corner-case where arrow functions that require a context
allocate none, because there are no additional slots allocated. Note
that this didn't happen with true function scopes because they always
had at least the receiver slot.

The outcome was a context chain that no longer was in sync with the
scope chain, hence context slot loads were bogus. This is observable
using the DYNAMIC_LOCAL optimization in all compilers.

R=rossberg@chromium.org,wingo@igalia.com
TEST=mjsunit/harmony/regress/regress-4160
BUG=v8:4160
LOG=N

Review URL: https://codereview.chromium.org/1146063006

Cr-Commit-Position: refs/heads/master@{#28788}

9 years ago[turbofan] Don't lower to NumberModulus unless the inputs are numbers.
bmeurer [Wed, 3 Jun 2015 11:24:14 +0000 (04:24 -0700)]
[turbofan] Don't lower to NumberModulus unless the inputs are numbers.

The IC for modulus is usually way faster than converting the inputs to
numbers and doing a Float64Mod on them.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1170433002

Cr-Commit-Position: refs/heads/master@{#28787}

9 years ago[x64] Smi zero can be used as an immediate.
bmeurer [Wed, 3 Jun 2015 11:23:06 +0000 (04:23 -0700)]
[x64] Smi zero can be used as an immediate.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1150853009

Cr-Commit-Position: refs/heads/master@{#28786}

9 years agoAdd ARM64 suppport to grokdump.py
ulan [Wed, 3 Jun 2015 11:07:24 +0000 (04:07 -0700)]
Add ARM64 suppport to grokdump.py

Constants are taken from chromium/src/google_breakpad/common/minidump_cpu_arm64.h

BUG=

Review URL: https://codereview.chromium.org/1149293005

Cr-Commit-Position: refs/heads/master@{#28785}

9 years ago[turbofan] Make the verifier a little bit more cooperative.
mstarzinger [Wed, 3 Jun 2015 10:46:58 +0000 (03:46 -0700)]
[turbofan] Make the verifier a little bit more cooperative.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1155723005

Cr-Commit-Position: refs/heads/master@{#28784}

9 years agoRemove usage of to-be-deprecated APIs from v8 core
jochen [Wed, 3 Jun 2015 10:27:46 +0000 (03:27 -0700)]
Remove usage of to-be-deprecated APIs from v8 core

Also turn on the macro to disable to-be-deprecated APIs for core

BUG=v8:4134
R=vogelheim@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/1162363005

Cr-Commit-Position: refs/heads/master@{#28783}

9 years ago[date] Refactor the %_DateField intrinsic to be optimizable.
bmeurer [Wed, 3 Jun 2015 08:31:15 +0000 (01:31 -0700)]
[date] Refactor the %_DateField intrinsic to be optimizable.

Previously the %_DateField intrinsic would also check the object and
throw an exception if you happen to pass something that is not a valid
JSDate, which (a) violates our policy for instrinsics and (b) is hard to
optimize in TurboFan (even Crankshaft has a hard time, but there we will
never inline the relevant builtins, so it doesn't show up). The throwing
part is now a separate intrinsics %_ThrowIfNotADate that throws an
exception in full codegen and deoptimizes in Crankshaft, which means the
code for the current use cases is roughly the same (modulo some register
renamings/gap moves).

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/1167813003

Cr-Commit-Position: refs/heads/master@{#28782}

9 years agoVisitObject should use MarkObject.
hpayer [Wed, 3 Jun 2015 08:14:47 +0000 (01:14 -0700)]
VisitObject should use MarkObject.

BUG=

Review URL: https://codereview.chromium.org/1151993005

Cr-Commit-Position: refs/heads/master@{#28781}

9 years agoAlso allocate small typed arrays on heap when initialized from an array-like
jochen [Wed, 3 Jun 2015 07:41:28 +0000 (00:41 -0700)]
Also allocate small typed arrays on heap when initialized from an array-like

This means something like new Float32Array([23, 42]) will be allocated on heap.

BUG=v8:3996
R=bmeurer@chromium.org,mstarzinger@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/1144393003

Cr-Commit-Position: refs/heads/master@{#28780}

9 years agoAIX: fix another may be uninitialized failure
michael_dawson [Wed, 3 Jun 2015 07:23:15 +0000 (00:23 -0700)]
AIX: fix another may be uninitialized failure

fix two additional cases in heap.cc where the AIX
compiler is now reporting that a variable may be
uninitialized.

R=svenpanne@chromium.org, mbrandy@us.ibm.com

BUG=

Review URL: https://codereview.chromium.org/1157053007

Cr-Commit-Position: refs/heads/master@{#28779}

9 years agoJust clear semi-space mark bits once before Scavenge
sejunho [Wed, 3 Jun 2015 07:09:21 +0000 (00:09 -0700)]
Just clear semi-space mark bits once before Scavenge

Both PrepareForScavenge and ResetAllocationInfo clear semi-space mark bits before Scavenge. Doing this once is good enough.

BUG=

Review URL: https://codereview.chromium.org/1158623003

Cr-Commit-Position: refs/heads/master@{#28778}

9 years agoRevert of Implement %TypedArray%.prototype.{reduce,reduceRight} (patchset #3 id:40001...
bmeurer [Wed, 3 Jun 2015 05:29:30 +0000 (22:29 -0700)]
Revert of Implement %TypedArray%.prototype.{reduce,reduceRight} (patchset #3 id:40001 of https://codereview.chromium.org/1154423014/)

Reason for revert:
Breaks Win32 mjsunit/harmony/typedarray-reduce, see http://build.chromium.org/p/client.v8/builders/V8%20Win32%20-%20nosnap%20-%20shared/builds/7121/steps/Check/logs/typedarray-reduce for details.

Original issue's description:
> Implement %TypedArray%.prototype.{reduce,reduceRight}
>
> This patch implements the last two methods on TypedArrays. These
> were previously committed and led to a test failure.
>
> BUG=v8:3578
> LOG=Y
> R=adamk
>
> Committed: https://crrev.com/95d779ecc9816ac20c1565bebeeac6a1ef29f1c7
> Cr-Commit-Position: refs/heads/master@{#28773}

TBR=adamk@chromium.org,dehrenberg@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:3578

Review URL: https://codereview.chromium.org/1162043008

Cr-Commit-Position: refs/heads/master@{#28777}

9 years ago[turbofan] Use reference equal to zero instead of a smi check.
bmeurer [Wed, 3 Jun 2015 05:10:44 +0000 (22:10 -0700)]
[turbofan] Use reference equal to zero instead of a smi check.

In typed lowering we can use the ReferenceEqual simplified operator
instead of ObjectIsSmi to check for context extensions. This generates
the desired code.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/1163963003

Cr-Commit-Position: refs/heads/master@{#28776}

9 years ago[turbofan] Ship TF for computed property names.
mstarzinger [Wed, 3 Jun 2015 05:03:57 +0000 (22:03 -0700)]
[turbofan] Ship TF for computed property names.

R=bmeurer@chromium.org
BUG=v8:4158
LOG=N

Review URL: https://codereview.chromium.org/1162353002

Cr-Commit-Position: refs/heads/master@{#28775}

9 years agoUpdate V8 DEPS.
v8-autoroll [Wed, 3 Jun 2015 05:02:49 +0000 (22:02 -0700)]
Update V8 DEPS.

Rolling v8/tools/clang to 12a1473d6de70775c77726b62aed18bfd56818c1

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1166643004

Cr-Commit-Position: refs/heads/master@{#28774}

9 years agoImplement %TypedArray%.prototype.{reduce,reduceRight}
dehrenberg [Wed, 3 Jun 2015 05:01:31 +0000 (22:01 -0700)]
Implement %TypedArray%.prototype.{reduce,reduceRight}

This patch implements the last two methods on TypedArrays. These
were previously committed and led to a test failure.

BUG=v8:3578
LOG=Y
R=adamk

Review URL: https://codereview.chromium.org/1154423014

Cr-Commit-Position: refs/heads/master@{#28773}

9 years agoRevert of Embedded constant pools. (patchset #12 id:220001 of https://codereview...
bmeurer [Wed, 3 Jun 2015 03:02:26 +0000 (20:02 -0700)]
Revert of Embedded constant pools. (patchset #12 id:220001 of https://codereview.chromium.org/1131783003/)

Reason for revert:
Breaks Linux nosnap cctest/test-api/FastReturnValuesWithProfiler, see http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20nosnap%20-%20debug%20-%202/builds/609/steps/Check/logs/FastReturnValuesWithP..

Original issue's description:
> Add support for Embedded Constant Pools for PPC and Arm
>
> Embed constant pools within their corresponding Code
> objects.
>
> This removes support for out-of-line constant pools in favor
> of the new approach -- the main advantage being that it
> eliminates the need to allocate and manage separate constant
> pool array objects.
>
> Currently supported on PPC and ARM.  Enabled by default on
> PPC only.
>
> This yields a 6% improvment in Octane on PPC64.
>
> R=danno@chromium.org, svenpanne@chromium.org, bmeurer@chromium.org, rmcilroy@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
> BUG=chromium:478811
> LOG=Y
>
> Committed: https://crrev.com/a9404029343d65f146e3443f5280c40a97e736af
> Cr-Commit-Position: refs/heads/master@{#28770}

TBR=rmcilroy@chromium.org,ishell@chromium.org,rodolph.perfetta@arm.com,mbrandy@us.ibm.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:478811

Review URL: https://codereview.chromium.org/1155703006

Cr-Commit-Position: refs/heads/master@{#28772}

9 years agoAdd SIMD 128 alignment support to Heap.
bbudge [Tue, 2 Jun 2015 22:56:00 +0000 (15:56 -0700)]
Add SIMD 128 alignment support to Heap.
Adds SIMD 128 alignment sizes and masks.
Adds support in Heap for SIMD alignments and fills.
Reworks cctest so that each test independently aligns its allocation address, rather than depending on the previous tests ending state. Adds test cases for SIMD.

LOG=N
BUG=v8:4124

Committed: https://crrev.com/4347d56a6919ae06a70e4a4a8b2f1179cf47bc7e
Cr-Commit-Position: refs/heads/master@{#28767}

Review URL: https://codereview.chromium.org/1159453004

Cr-Commit-Position: refs/heads/master@{#28771}

9 years agoAdd support for Embedded Constant Pools for PPC and Arm
mbrandy [Tue, 2 Jun 2015 22:50:00 +0000 (15:50 -0700)]
Add support for Embedded Constant Pools for PPC and Arm

Embed constant pools within their corresponding Code
objects.

This removes support for out-of-line constant pools in favor
of the new approach -- the main advantage being that it
eliminates the need to allocate and manage separate constant
pool array objects.

Currently supported on PPC and ARM.  Enabled by default on
PPC only.

This yields a 6% improvment in Octane on PPC64.

R=danno@chromium.org, svenpanne@chromium.org, bmeurer@chromium.org, rmcilroy@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=chromium:478811
LOG=Y

Review URL: https://codereview.chromium.org/1131783003

Cr-Commit-Position: refs/heads/master@{#28770}

9 years ago[es6] Super call in arrows and eval
arv [Tue, 2 Jun 2015 22:04:25 +0000 (15:04 -0700)]
[es6] Super call in arrows and eval

This splits the SuperReference AST node into SuperPropertyReference and
SuperCallReference. The super call reference node consists of three
unresolved vars to this, new.target and this_function. These gets
declared when the right function is entered and if it is in use. The
variables gets assigned in FullCodeGenerator::Generate.

This is a revert of the revert 88b1c9170a0293cbcc8bdaf57fbe12744b48d7e8

BUG=v8:3768
LOG=N
R=wingo@igalia.com, adamk@chromium.org

Review URL: https://codereview.chromium.org/1168513004

Cr-Commit-Position: refs/heads/master@{#28769}

9 years agoRevert of Add SIMD 128 alignment support to Heap. (patchset #3 id:40001 of https...
bbudge [Tue, 2 Jun 2015 21:13:14 +0000 (14:13 -0700)]
Revert of Add SIMD 128 alignment support to Heap. (patchset #3 id:40001 of https://codereview.chromium.org/1159453004/)

Reason for revert:
Breaks Linux - arm64 - sim - MSAN
TBR=jochen

Original issue's description:
> Add SIMD 128 alignment support to Heap.
> Adds SIMD 128 alignment sizes and masks.
> Adds support in Heap for SIMD alignments and fills.
> Reworks cctest so that each test independently aligns its allocation address, rather than depending on the previous tests ending state. Adds test cases for SIMD.
>
> LOG=N
> BUG=v8:4124

TBR=hpayer@chromium.org,jochen@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:4124

Review URL: https://codereview.chromium.org/1169453003

Cr-Commit-Position: refs/heads/master@{#28768}

9 years agoAdd SIMD 128 alignment support to Heap.
bbudge [Tue, 2 Jun 2015 20:40:17 +0000 (13:40 -0700)]
Add SIMD 128 alignment support to Heap.
Adds SIMD 128 alignment sizes and masks.
Adds support in Heap for SIMD alignments and fills.
Reworks cctest so that each test independently aligns its allocation address, rather than depending on the previous tests ending state. Adds test cases for SIMD.

LOG=N
BUG=v8:4124

Review URL: https://codereview.chromium.org/1159453004

Cr-Commit-Position: refs/heads/master@{#28767}

9 years agoImplement %TypedArray%.prototype.{toString,toLocaleString,join}
dehrenberg [Tue, 2 Jun 2015 18:58:12 +0000 (11:58 -0700)]
Implement %TypedArray%.prototype.{toString,toLocaleString,join}

Implementations factored out from Array. Tests are derived from
normal array toString tests.

BUG=v8:3578
LOG=Y
R=adamk

Review URL: https://codereview.chromium.org/1166623004

Cr-Commit-Position: refs/heads/master@{#28766}

9 years agoPPC: Refine '[strong] create strong array literals'
mbrandy [Tue, 2 Jun 2015 17:36:49 +0000 (10:36 -0700)]
PPC: Refine '[strong] create strong array literals'

R=dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1156823003

Cr-Commit-Position: refs/heads/master@{#28765}

9 years agoPPC: Build ObjectLiteral constant properties in the numbering phase.
mbrandy [Tue, 2 Jun 2015 17:35:22 +0000 (10:35 -0700)]
PPC: Build ObjectLiteral constant properties in the numbering phase.

Port 450002f3a105d4bb17fd3b1a9d9d54464c01d200

Original commit message:
It's necessary to do this in order to know how many type feedback vector slots
we should allocate for the object literal.

R=mvstanton@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1166473004

Cr-Commit-Position: refs/heads/master@{#28764}

9 years ago[test] Refactor call-tester to use c-signature.h.
titzer [Tue, 2 Jun 2015 15:00:55 +0000 (08:00 -0700)]
[test] Refactor call-tester to use c-signature.h.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1146173005

Cr-Commit-Position: refs/heads/master@{#28763}

9 years agoHandle Delete of element with LookupIterator
verwaest [Tue, 2 Jun 2015 11:52:30 +0000 (04:52 -0700)]
Handle Delete of element with LookupIterator

BUG=v8:4137
LOG=n

Review URL: https://codereview.chromium.org/1157093003

Cr-Commit-Position: refs/heads/master@{#28762}

9 years ago[turbofan] Ship TF for class literals.
mstarzinger [Tue, 2 Jun 2015 11:47:13 +0000 (04:47 -0700)]
[turbofan] Ship TF for class literals.

R=bmeurer@chromium.org
BUG=v8:4158
LOG=N

Review URL: https://codereview.chromium.org/1158883006

Cr-Commit-Position: refs/heads/master@{#28761}

9 years ago[strong] create strong array literals
conradw [Tue, 2 Jun 2015 11:46:04 +0000 (04:46 -0700)]
[strong] create strong array literals

Copied, with permission, from https://codereview.chromium.org/1151853003/

Initial patch set is an unmodified copy, rebased on top of related fixes from
https://codereview.chromium.org/1158933002/

Subsequent patch sets contain fixes for remaining bugs in the CL.

BUG=v8:3956
LOG=N

Review URL: https://codereview.chromium.org/1152093003

Cr-Commit-Position: refs/heads/master@{#28760}

9 years agoMark proxy map as unstable during proxy fixing (freezing, sealing or preventing exten...
ishell [Tue, 2 Jun 2015 11:29:22 +0000 (04:29 -0700)]
Mark proxy map as unstable during proxy fixing (freezing, sealing or preventing extensions).

BUG=chromium:493568
LOG=N

Review URL: https://codereview.chromium.org/1158023003

Cr-Commit-Position: refs/heads/master@{#28759}

9 years agoRemove Execution::CharAt
verwaest [Tue, 2 Jun 2015 11:24:52 +0000 (04:24 -0700)]
Remove Execution::CharAt

BUG=

Review URL: https://codereview.chromium.org/1143153004

Cr-Commit-Position: refs/heads/master@{#28758}

9 years agoUse LookupIterator for GetElementAttributes and friends
verwaest [Tue, 2 Jun 2015 11:10:50 +0000 (04:10 -0700)]
Use LookupIterator for GetElementAttributes and friends

BUG=v8:4137
LOG=n

Review URL: https://codereview.chromium.org/1153583006

Cr-Commit-Position: refs/heads/master@{#28757}

9 years agoFilter out store/slots buffer entries that point into raw data objects.
ishell [Tue, 2 Jun 2015 11:00:45 +0000 (04:00 -0700)]
Filter out store/slots buffer entries that point into raw data objects.

This is just a workaround till we found the root cause of the issue, there must be no slots for data object recorded.

BUG=chromium:454297
LOG=N

Review URL: https://codereview.chromium.org/1106983004

Cr-Commit-Position: refs/heads/master@{#28756}

9 years ago[arm64] Use ubfiz in ARM64 instruction selector
martyn.capewell [Tue, 2 Jun 2015 10:57:40 +0000 (03:57 -0700)]
[arm64] Use ubfiz in ARM64 instruction selector

Select ubfiz for (x & mask) << imm where mask is contiguous and imm is non-zero.

BUG=

Review URL: https://codereview.chromium.org/1161643003

Cr-Commit-Position: refs/heads/master@{#28755}

9 years agoUse GetProperty for getting elements.
verwaest [Tue, 2 Jun 2015 10:42:16 +0000 (03:42 -0700)]
Use GetProperty for getting elements.

This also fixes issues with
- kMaxUint32 being a valid length but not index cornercases
- exotic integer objects masking "exotic indexes" even though its in the prototype chain
- concating of holey sloppy arguments

BUG=v8:4137
LOG=n

Review URL: https://codereview.chromium.org/1159433003

Cr-Commit-Position: refs/heads/master@{#28754}

9 years ago[turbofan] Ship TF for dynamic lookup slots.
mstarzinger [Tue, 2 Jun 2015 10:36:21 +0000 (03:36 -0700)]
[turbofan] Ship TF for dynamic lookup slots.

R=bmeurer@chromium.org
BUG=v8:4131
LOG=N

Review URL: https://codereview.chromium.org/1152173011

Cr-Commit-Position: refs/heads/master@{#28753}

9 years agoGlobalDictionary now stores PropertyDetails in PropertyCells.
ishell [Tue, 2 Jun 2015 10:29:12 +0000 (03:29 -0700)]
GlobalDictionary now stores PropertyDetails in PropertyCells.

BUG=

Review URL: https://codereview.chromium.org/1156993018

Cr-Commit-Position: refs/heads/master@{#28752}

9 years agoFix redundant map allocation
rossberg [Tue, 2 Jun 2015 10:15:06 +0000 (03:15 -0700)]
Fix redundant map allocation

R=bmeurer@chromium.org
BUG=chromium:492022,v8:3956
LOG=N

Review URL: https://codereview.chromium.org/1159623008

Cr-Commit-Position: refs/heads/master@{#28751}

9 years ago[turbofan] Optimized lowering of DYNAMIC_GLOBAL lookup slot loads.
mstarzinger [Tue, 2 Jun 2015 09:37:49 +0000 (02:37 -0700)]
[turbofan] Optimized lowering of DYNAMIC_GLOBAL lookup slot loads.

This adds handling of JSLoadDynamicGlobal nodes to JSTypedLowering to
perform extension checks and an inline fast path. The fast path is a
global variable load from the global object.

R=bmeurer@chromium.org
BUG=v8:4131
LOG=N

Review URL: https://codereview.chromium.org/1150723005

Cr-Commit-Position: refs/heads/master@{#28750}

9 years agoBuild ObjectLiteral constant properties in the numbering phase.
mvstanton [Tue, 2 Jun 2015 09:35:21 +0000 (02:35 -0700)]
Build ObjectLiteral constant properties in the numbering phase.

It's necessary to do this in order to know how many type feedback vector slots
we should allocate for the object literal.

BUG=

Review URL: https://codereview.chromium.org/1165463007

Cr-Commit-Position: refs/heads/master@{#28749}

9 years agoX87: [turbofan] First step towards sanitizing for-in and making it optimizable.
chunyang.dai [Tue, 2 Jun 2015 09:07:13 +0000 (02:07 -0700)]
X87: [turbofan] First step towards sanitizing for-in and making it optimizable.

port e2e47f30be06e6ea6b8a3cc90138381914ecd07d (r28711)

original commit message:

    [turbofan] First step towards sanitizing for-in and making it optimizable.

    In a nutshell: The FILTER_KEY builtin is gone, and was replaced by a
    simple runtime call to ForInFilter, which does everything and is even
    cheaper (because FILTER_KEY used to call into the runtime anyway).
    And ForInFilter returns either the name or undefined, which makes it
    possible to remove the control flow construction from the AstGraphBuilder,
    and thereby make both the initialization and the per-loop code of for-in
    optimizable later (in typed lowering).

BUG=

Review URL: https://codereview.chromium.org/1144143005

Cr-Commit-Position: refs/heads/master@{#28748}

9 years ago[turbofan] Enable typed lowering of string addition.
bmeurer [Tue, 2 Jun 2015 08:50:46 +0000 (01:50 -0700)]
[turbofan] Enable typed lowering of string addition.

Unfortunately StringAdd is not pure in V8 because we might throw an
exception if the resulting string length is outside the valid bounds, so
there's no point in having a simplified StringAdd operator.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1164743002

Cr-Commit-Position: refs/heads/master@{#28747}

9 years ago[turbofan] Specify better type for FixedArray::length field.
bmeurer [Tue, 2 Jun 2015 06:18:23 +0000 (23:18 -0700)]
[turbofan] Specify better type for FixedArray::length field.

This avoids redundant smi checks when using the length of a FixedArray,
which is always a positve smi.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1162163003

Cr-Commit-Position: refs/heads/master@{#28746}

9 years agoUpdate V8 DEPS.
v8-autoroll [Tue, 2 Jun 2015 04:07:28 +0000 (21:07 -0700)]
Update V8 DEPS.

Rolling v8/tools/clang to dc542e2b87607a399e0c107b78990aba11f5af48

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1143213004

Cr-Commit-Position: refs/heads/master@{#28745}

9 years ago[es6] Stage array spread
arv [Mon, 1 Jun 2015 23:40:46 +0000 (16:40 -0700)]
[es6] Stage array spread

BUG=v8:3018
LOG=N

Review URL: https://codereview.chromium.org/1165653003

Cr-Commit-Position: refs/heads/master@{#28744}

9 years agoFix Map/Set creation via the API with nosnap build
adamk [Mon, 1 Jun 2015 23:16:27 +0000 (16:16 -0700)]
Fix Map/Set creation via the API with nosnap build

The Map and Set maps get overwritten when collection.js executes, so in
a nosnap build we have to wait until it runs before we grab the maps.
To facilitate that, store the functions in the native context as well.

Review URL: https://codereview.chromium.org/1161363002

Cr-Commit-Position: refs/heads/master@{#28743}

9 years agoCosmetic changes to tests to make it easier to concatenate them.
erikcorry [Mon, 1 Jun 2015 22:46:54 +0000 (15:46 -0700)]
Cosmetic changes to tests to make it easier to concatenate them.

When compiling on a laptop I like to concatenate the small test files.
This makes a big difference to compile times. These changes make that
easier.

R=ulan@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1163803002

Cr-Commit-Position: refs/heads/master@{#28742}

9 years ago[strong] fix strong object exception messages
conradw [Mon, 1 Jun 2015 22:43:50 +0000 (15:43 -0700)]
[strong] fix strong object exception messages

A mistake in a recent CL has messed up the error messages for strong object
semantics.

BUG=
LOG=N

Review URL: https://codereview.chromium.org/1166433007

Cr-Commit-Position: refs/heads/master@{#28741}

9 years agoRevert of [es6] implement default parameters via desugaring (patchset #19 id:380001...
caitpotter88 [Mon, 1 Jun 2015 18:34:55 +0000 (11:34 -0700)]
Revert of [es6] implement default parameters via desugaring (patchset #19 id:380001 of https://codereview.chromium.org/1127063003/)

Reason for revert:
Broken on arm64

Original issue's description:
> [es6] implement default parameters via desugaring
>
> Stage 1 implementation:
>
> - Parameters can't be referenced before initialized (from left-to-right)
> - SingleNameBindings only, no support for BindingPatterns
>
> Known issues:
>
> - Incorrect scoping (parameter expressions may reference variables declared in function body)
> - Function arity is untouched
> - Hole-checking needs work
> - Rest parameters are broken when mixed with optional arguments
>
> BUG=v8:2160
> LOG=N
> R=arv@chromium.org, rossberg@chromium.org
>
> Committed: https://crrev.com/892c85485881f8be2f17bd83238980f858126576
> Cr-Commit-Position: refs/heads/master@{#28739}

TBR=rossberg@chromium.org,wingo@igalia.com,arv@chromium.org,dslomov@chromium.org,adamk@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:2160

Review URL: https://codereview.chromium.org/1163853002

Cr-Commit-Position: refs/heads/master@{#28740}

9 years ago[es6] implement default parameters via desugaring
caitpotter88 [Mon, 1 Jun 2015 17:10:39 +0000 (10:10 -0700)]
[es6] implement default parameters via desugaring

Stage 1 implementation:

- Parameters can't be referenced before initialized (from left-to-right)
- SingleNameBindings only, no support for BindingPatterns

Known issues:

- Incorrect scoping (parameter expressions may reference variables declared in function body)
- Function arity is untouched
- Hole-checking needs work
- Rest parameters are broken when mixed with optional arguments

BUG=v8:2160
LOG=N
R=arv@chromium.org, rossberg@chromium.org

Review URL: https://codereview.chromium.org/1127063003

Cr-Commit-Position: refs/heads/master@{#28739}

9 years agoEnum DictionaryEntryType removed.
ishell [Mon, 1 Jun 2015 17:07:25 +0000 (10:07 -0700)]
Enum DictionaryEntryType removed.

Dictionary customization should be implemented in respective shape class.

Review URL: https://codereview.chromium.org/1160813009

Cr-Commit-Position: refs/heads/master@{#28738}

9 years agoStarting using GlobalDictionary for global objects instead of NameDictionary.
ishell [Mon, 1 Jun 2015 16:24:59 +0000 (09:24 -0700)]
Starting using GlobalDictionary for global objects instead of NameDictionary.

Review URL: https://codereview.chromium.org/1165603002

Cr-Commit-Position: refs/heads/master@{#28737}

9 years ago/infra-config directory
nodir [Mon, 1 Jun 2015 16:19:37 +0000 (09:19 -0700)]
/infra-config directory

Added /infra with chrome-infra specific files.
/infra/project-config/cr-buildbucket.cfg defines v8 buckets on cr-buildbucket.appspot.com

R=tandrii@chromium.org, machenbach@chromium.org, sergiyb@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1161143002

Cr-Commit-Position: refs/heads/master@{#28736}

9 years agoRevert of [es6] Super call in arrows and eval (patchset #5 id:100001 of https://coder...
arv [Mon, 1 Jun 2015 16:10:08 +0000 (09:10 -0700)]
Revert of [es6] Super call in arrows and eval (patchset #5 id:100001 of https://codereview.chromium.org/1146863007/)

Reason for revert:
Fails

http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20nosnap%20-%20debug%20-%201/builds/579/steps/Check/logs/super

Original issue's description:
> [es6] Super call in arrows and eval
>
> This splits the SuperReference AST node into SuperPropertyReference and
> SuperCallReference. The super call reference node consists of three
> unresolved vars to this, new.target and this_function. These gets
> declared when the right function is entered and if it is in use. The
> variables gets assigned in FullCodeGenerator::Generate.
>
> BUG=v8:3768
> LOG=N
> R=wingo@igalia.com, adamk@chromium.org
>
> Committed: https://crrev.com/673c0516ab96f24343bbb26e0afc2846b5a679df
> Cr-Commit-Position: refs/heads/master@{#28731}

TBR=wingo@igalia.com,adamk@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:3768

Review URL: https://codereview.chromium.org/1161243005

Cr-Commit-Position: refs/heads/master@{#28735}

9 years agoPPC: [turbofan] First step towards sanitizing for-in and making it optimizable.
mbrandy [Mon, 1 Jun 2015 15:59:47 +0000 (08:59 -0700)]
PPC: [turbofan] First step towards sanitizing for-in and making it optimizable.

Port e2e47f30be06e6ea6b8a3cc90138381914ecd07d

Original commit message:
In a nutshell: The FILTER_KEY builtin is gone, and was replaced by a
simple runtime call to ForInFilter, which does everything and is even
cheaper (because FILTER_KEY used to call into the runtime anyway).
And ForInFilter returns either the name or undefined, which makes it
possible to remove the control flow construction from the AstGraphBuilder,
and thereby make both the initialization and the per-loop code of for-in
optimizable later (in typed lowering).

R=bmeurer@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1160973004

Cr-Commit-Position: refs/heads/master@{#28734}

9 years ago[turbofan] Tester improvements; use CSignature and simplify ReturnValueTraits.
titzer [Mon, 1 Jun 2015 15:54:53 +0000 (08:54 -0700)]
[turbofan] Tester improvements; use CSignature and simplify ReturnValueTraits.

R=bmeurer@chromium.org,mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1150083006

Cr-Commit-Position: refs/heads/master@{#28733}

9 years agoIntroducing GlobalDictionary, a backing store for global objects.
ishell [Mon, 1 Jun 2015 15:43:24 +0000 (08:43 -0700)]
Introducing GlobalDictionary, a backing store for global objects.

This updates Dictionary classes hierarchy and introduces GlobalDictionary class but it is not used yet.

Review URL: https://codereview.chromium.org/1163673003

Cr-Commit-Position: refs/heads/master@{#28732}

9 years ago[es6] Super call in arrows and eval
arv [Mon, 1 Jun 2015 15:02:31 +0000 (08:02 -0700)]
[es6] Super call in arrows and eval

This splits the SuperReference AST node into SuperPropertyReference and
SuperCallReference. The super call reference node consists of three
unresolved vars to this, new.target and this_function. These gets
declared when the right function is entered and if it is in use. The
variables gets assigned in FullCodeGenerator::Generate.

BUG=v8:3768
LOG=N
R=wingo@igalia.com, adamk@chromium.org

Review URL: https://codereview.chromium.org/1146863007

Cr-Commit-Position: refs/heads/master@{#28731}

9 years agoTiny fix to grokdump heap stats printer
erikcorry [Mon, 1 Jun 2015 14:54:04 +0000 (07:54 -0700)]
Tiny fix to grokdump heap stats printer

R=hpayer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1164693002

Cr-Commit-Position: refs/heads/master@{#28730}

9 years agoReland "Re-enable on-heap typed array allocation"
jochen [Mon, 1 Jun 2015 14:22:58 +0000 (07:22 -0700)]
Reland "Re-enable on-heap typed array allocation"

R=mstarzinger@chromium.org
BUG=v8:3996
LOG=y

Review URL: https://codereview.chromium.org/1144143006

Cr-Commit-Position: refs/heads/master@{#28729}

9 years agoFix issue with --print-ast and class expressions
arv [Mon, 1 Jun 2015 14:06:14 +0000 (07:06 -0700)]
Fix issue with --print-ast and class expressions

Make sure that --print-ast works even when the class expression has no
name.

This also updates the printer to print the properties of the class
literal.

BUG=v8:4138
LOG=N
R=rossberg

Review URL: https://codereview.chromium.org/1165613003

Cr-Commit-Position: refs/heads/master@{#28728}

9 years agoFix bogus insertion of filler in LO-space by String#replace.
mstarzinger [Mon, 1 Jun 2015 13:36:02 +0000 (06:36 -0700)]
Fix bogus insertion of filler in LO-space by String#replace.

R=hpayer@chromium.org
TEST=mjsunit/regress/regress-crbug-493779
BUG=chromium:493779
LOG=N

Review URL: https://codereview.chromium.org/1163793002

Cr-Commit-Position: refs/heads/master@{#28727}

9 years ago[turbofan] First steps towards optimizing for-in loops.
bmeurer [Mon, 1 Jun 2015 11:41:21 +0000 (04:41 -0700)]
[turbofan] First steps towards optimizing for-in loops.

This is basically a port of the majority of optimizations that are
applied to for-in in full codegen. But it is not done during graph
building, but instead during typed lowering, which way less adhoc than
what the other compilers do.

Review URL: https://codereview.chromium.org/1155313008

Cr-Commit-Position: refs/heads/master@{#28726}

9 years ago[turbofan] Disable optimization of dynamic lookup slots.
mstarzinger [Mon, 1 Jun 2015 11:40:13 +0000 (04:40 -0700)]
[turbofan] Disable optimization of dynamic lookup slots.

Note that this essentially disables optimization of "with" as well.

R=hablich@chromium.org
BUG=chromium:491897,chromium:491018
LOG=N

Review URL: https://codereview.chromium.org/1158343003

Cr-Commit-Position: refs/heads/master@{#28725}

9 years ago[strong] Implement per-object restrictions behaviour of delete operator
conradw [Mon, 1 Jun 2015 11:39:01 +0000 (04:39 -0700)]
[strong] Implement per-object restrictions behaviour of delete operator

Implements the strong mode proposal's restrictions on the behaviour of the
delete operator for strong objects.

Setting the strong bit is still wip, so this change will only affect those
objects that have the bit correctly set. The tests reflect this, and will be
expanded as more objects can be marked as strong.

Attempt 2, last version did not work with API.

BUG=v8:3956
LOG=N

Review URL: https://codereview.chromium.org/1156573002

Cr-Commit-Position: refs/heads/master@{#28724}

9 years agoRevert of Re-enable on-heap typed array allocation (patchset #1 id:1 of https://coder...
bmeurer [Mon, 1 Jun 2015 10:58:51 +0000 (03:58 -0700)]
Revert of Re-enable on-heap typed array allocation (patchset #1 id:1 of https://codereview.chromium.org/1166433004/)

Reason for revert:
Win32 breakage

Original issue's description:
> Re-enable on-heap typed array allocation
>
> BUG=v8:3996
> R=mstarzinger@chromium.org
> LOG=y
>
> Committed: https://crrev.com/f91df1f25dec4f1982c40af6118da8b699777475
> Cr-Commit-Position: refs/heads/master@{#28722}

TBR=mstarzinger@chromium.org,jochen@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:3996

Review URL: https://codereview.chromium.org/1158193004

Cr-Commit-Position: refs/heads/master@{#28723}

9 years agoRe-enable on-heap typed array allocation
jochen [Mon, 1 Jun 2015 10:07:09 +0000 (03:07 -0700)]
Re-enable on-heap typed array allocation

BUG=v8:3996
R=mstarzinger@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/1166433004

Cr-Commit-Position: refs/heads/master@{#28722}

9 years agoMIPS64: Fix lithium arithmetic operations for integers to sign-extend result.
dusan.milosavljevic [Mon, 1 Jun 2015 09:47:42 +0000 (02:47 -0700)]
MIPS64: Fix lithium arithmetic operations for integers to sign-extend result.

TEST=mjsunit/asm/double-lo
BUG=

Review URL: https://codereview.chromium.org/1161713003

Cr-Commit-Position: refs/heads/master@{#28721}

9 years agoX87: Make KeyedStores from a sloppy arguments array use a handler.
chunyang.dai [Mon, 1 Jun 2015 09:21:50 +0000 (02:21 -0700)]
X87: Make KeyedStores from a sloppy arguments array use a handler.

port 3a1d7335cf6a623f39c306e0bc0e82ccccc54818 (r28683)

original commit message:

  This finishes up work begun by CL (https://codereview.chromium.org/546683003/) some months ago.

BUG=

Review URL: https://codereview.chromium.org/1163603003

Cr-Commit-Position: refs/heads/master@{#28720}

9 years agoX87: [crankshaft] Record inlined shared function infos instead of closures.
chunyang.dai [Mon, 1 Jun 2015 09:11:28 +0000 (02:11 -0700)]
X87: [crankshaft] Record inlined shared function infos instead of closures.

port 388e791df91c299507d62f91bc4f618a4dc80276 (r28672).

original commit message:

   The list of inlined functions is used in exactly two places - for live
    edit and to prevent code flushing for inlined functions - and those are
    fine with SharedFunctionInfo and don't require a closure.

    This is one additional step towards inlining based on SharedFunctionInfo
    instead of JSFunction.

BUG=

Review URL: https://codereview.chromium.org/1143003003

Cr-Commit-Position: refs/heads/master@{#28719}

9 years agoX87: VectorICs: allocating slots for store ics in ast nodes.
chunyang.dai [Mon, 1 Jun 2015 09:04:41 +0000 (02:04 -0700)]
X87: VectorICs: allocating slots for store ics in ast nodes.

port 5450fc07ba07615a70f5ed8379dc23c3275d6fe3 (r18659)

original commit message:

    Also adapt code generation to pass the slot to the
    store/keyed-store ic. AST nodes ObjectLiteral, Assignment,
    ForEach, Call and CountOperation now include one or more
    feedback vector ic slot ids.

BUG=

Review URL: https://codereview.chromium.org/1155383003

Cr-Commit-Position: refs/heads/master@{#28718}

9 years agoX87: [es6] Support super.property in eval and arrow functions
chunyang.dai [Mon, 1 Jun 2015 08:55:20 +0000 (01:55 -0700)]
X87: [es6] Support super.property in eval and arrow functions

port 44e9810345cea9bfd6861905bc6856db7db5a25c (r28644)

original commit message:

    When we enter a method that needs access to the [[HomeObject]]
    we allocate a local variable `.home_object` and assign it the
    value from the [[HomeObject]] private symbol. Something along
    the lines of:

      method() {
        var .home_object = %ThisFunction()[home_object_symbol];
        ...
      }

BUG=

Review URL: https://codereview.chromium.org/1158543004

Cr-Commit-Position: refs/heads/master@{#28717}

9 years agoX87: Move hash code from hidden string to a private symbol
chunyang.dai [Mon, 1 Jun 2015 08:53:06 +0000 (01:53 -0700)]
X87: Move hash code from hidden string to a private symbol

port eca5b5d7abc0a9028cb9832087fbf2ed59dadf92 (r28622).

original commit message:

   * Hash code is now just done with a private own symbol instead of the hidden string, which predates symbols.
    * In the long run we should do all hidden properties this way and get rid of the
    hidden magic 0-length string with the zero hash code.  The advantages include
    less complexity and being able to do things from JS in a natural way.
    * Initially, the performance of weak set regressed, because it's a little harder
    to do the lookup in C++.  Instead of heroics in C++ to make things faster I
    moved some functionality into JS and got the performance back. JS is supposed to be good at looking up named properties on objects.
    * This also changes hash codes of Smis so that they are always Smis.

    Performance figures are in the comments to the code review.  Summary: Most of js-perf-test/Collections is neutral.  Set and Map with object keys are 40-50% better.  WeakMap is -5% and WeakSet is +9%.

    In the code review comments is a patch with an example of the heroics we could do in C++ to make lookup faster (I hope we don't have to do this.  Instead of checking for the property, then doing a new

    In a similar vein we could give the magic zero hash code to the hash code
    symbol.  Then when we look up the hash code we would sometimes see the table
    with all the hidden properties.  This dual use of the field for either the hash
    code or the table with all hidden properties and the hash code is rather ugly,
    and this CL gets rid of it.  I'd be loath to bring it back.  On the benchmarks quoted above it's slightly slower than moving the hash code lookup to JS like in this CL.

    One worry is that the benchmark results above are more monomorphic than real
    world code, so may be overstating the performance benefits of moving to JS.  I
    think this is part of a general issue we have with handling polymorphic code in
    JS and any solutions there will benefit this solution, which boils down to
    regular property access. Any improvement there will lift all boats.

BUG=

Review URL: https://codereview.chromium.org/1153963010

Cr-Commit-Position: refs/heads/master@{#28716}

9 years agoX87: Move work to omit unnecessary ObjectLiteral stores to the numbering pass.
chunyang.dai [Mon, 1 Jun 2015 08:45:20 +0000 (01:45 -0700)]
X87: Move work to omit unnecessary ObjectLiteral stores to the numbering pass.

port 32de677805877bbf8d697926e09c23963a27e191 (r29615)

original commit message:

    The reason is that this information will be needed to compute the number of
    vector ic slots done at numbering time.

BUG=

Review URL: https://codereview.chromium.org/1165693002

Cr-Commit-Position: refs/heads/master@{#28715}

9 years agoFor Micro-benchmarks for 'with'
hablich [Mon, 1 Jun 2015 07:40:02 +0000 (00:40 -0700)]
For Micro-benchmarks for 'with'

R=arv@chromium.org, mstarzinger@chromium.org
BUG=v8:4131
LOG=n

Review URL: https://codereview.chromium.org/1153363002

Cr-Commit-Position: refs/heads/master@{#28714}

9 years ago[turbofan] New operator for loads of DYNAMIC_[GLOBAL,LOCAL].
mstarzinger [Mon, 1 Jun 2015 07:35:06 +0000 (00:35 -0700)]
[turbofan] New operator for loads of DYNAMIC_[GLOBAL,LOCAL].

This introduces two new operators for loads of variables bound to
Variable::LOOKUP locations. Currently they all still lower to runtime
calls, but will allow optimization during typed lowering.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1155103004

Cr-Commit-Position: refs/heads/master@{#28713}

9 years agoAlso expose DefineOwnProperty
jochen [Mon, 1 Jun 2015 07:26:38 +0000 (00:26 -0700)]
Also expose DefineOwnProperty

In contrast to CreateDataProperty, this will always call out to JS

BUG=475206
R=adamk@chromium.org,verwaest@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/1167473002

Cr-Commit-Position: refs/heads/master@{#28712}

9 years ago[turbofan] First step towards sanitizing for-in and making it optimizable.
bmeurer [Mon, 1 Jun 2015 07:20:50 +0000 (00:20 -0700)]
[turbofan] First step towards sanitizing for-in and making it optimizable.

In a nutshell: The FILTER_KEY builtin is gone, and was replaced by a
simple runtime call to ForInFilter, which does everything and is even
cheaper (because FILTER_KEY used to call into the runtime anyway).
And ForInFilter returns either the name or undefined, which makes it
possible to remove the control flow construction from the AstGraphBuilder,
and thereby make both the initialization and the per-loop code of for-in
optimizable later (in typed lowering).

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1160983004

Cr-Commit-Position: refs/heads/master@{#28711}

9 years agoFix compile failure for AIX
michael_dawson [Mon, 1 Jun 2015 06:51:39 +0000 (23:51 -0700)]
Fix compile failure for AIX

Fix another may be uninitialized compile failure on AIX
in src/heap/heap.cc

R=svenpanne@chromium.org, mbrandy@us.ibm.com

BUG=

Review URL: https://codereview.chromium.org/1153233004

Cr-Commit-Position: refs/heads/master@{#28710}

9 years ago[arm] Fix detection of architecture versions.
jacob.bramley [Mon, 1 Jun 2015 04:26:27 +0000 (21:26 -0700)]
[arm] Fix detection of architecture versions.

Use __ARM_ARCH_6ZK__ instead of __ARM_ARCH_6KZ__. The architecture is
technically called ARMv6KZ, but GCC has always called it 6ZK.

Clang defines __ARM_ARCH_6K__ for this target, and is unaffected.

BUG=v8:3978
LOG=N

Review URL: https://codereview.chromium.org/1145973004

Cr-Commit-Position: refs/heads/master@{#28709}