machenbach [Tue, 20 Jan 2015 15:00:54 +0000 (07:00 -0800)]
Make clang the default on linux.
TBR=jochen@chromium.org
Review URL: https://codereview.chromium.org/
781983002
Cr-Commit-Position: refs/heads/master@{#26164}
weiliang.lin [Tue, 20 Jan 2015 14:59:47 +0000 (06:59 -0800)]
[x86] Avoid memory form of PUSH/CALL for ATOM.
Review URL: https://codereview.chromium.org/
853703002
Cr-Commit-Position: refs/heads/master@{#26163}
jochen [Tue, 20 Jan 2015 14:33:58 +0000 (06:33 -0800)]
Add initial support for v8_target_arch == "arm" to GN config
BUG=none
R=eseidel@chromium.org,brettw@chromium.org,machenbach@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
863683002
Cr-Commit-Position: refs/heads/master@{#26162}
jkummerow [Tue, 20 Jan 2015 14:16:00 +0000 (06:16 -0800)]
Replace generic KeyedStore stub with megamorphic version
This is a follow-up to https://codereview.chromium.org/
859943003.
Review URL: https://codereview.chromium.org/
847963004
Cr-Commit-Position: refs/heads/master@{#26161}
dcarney [Tue, 20 Jan 2015 14:14:29 +0000 (06:14 -0800)]
ensure trace extension works from optimized code when profiler is not enabled
BUG=
Review URL: https://codereview.chromium.org/
857953002
Cr-Commit-Position: refs/heads/master@{#26160}
Benedikt Meurer [Tue, 20 Jan 2015 13:59:07 +0000 (14:59 +0100)]
[x86] Use AVX in Crankshaft when available.
R=verwaest@chromium.org
Review URL: https://codereview.chromium.org/
860003002
Cr-Commit-Position: refs/heads/master@{#26159}
yangguo [Tue, 20 Jan 2015 13:53:34 +0000 (05:53 -0800)]
Use fast_sqrt instead of std::sqrt in simulators.
This prevents clang from inlining and returning inconsistent results.
R=bmeurer@chromium.org
BUG=v8:3802
LOG=N
Review URL: https://codereview.chromium.org/
831393006
Cr-Commit-Position: refs/heads/master@{#26158}
verwaest [Tue, 20 Jan 2015 13:41:17 +0000 (05:41 -0800)]
Remove ForceDelete
This method circumvented JS semantics, and should not be used.
BUG=
Review URL: https://codereview.chromium.org/
854493004
Cr-Commit-Position: refs/heads/master@{#26157}
dcarney [Tue, 20 Jan 2015 13:27:46 +0000 (05:27 -0800)]
fix debug checks for api callbacks returning symbols
BUG=
Review URL: https://codereview.chromium.org/
854363003
Cr-Commit-Position: refs/heads/master@{#26156}
mstarzinger [Tue, 20 Jan 2015 13:03:41 +0000 (05:03 -0800)]
Remove early bailout from Runtime_CompileOptimized.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/
857853004
Cr-Commit-Position: refs/heads/master@{#26155}
bmeurer [Tue, 20 Jan 2015 12:53:44 +0000 (04:53 -0800)]
[ia32] Require SSE2.
This way we stay in sync with what Chromium does and we reduce the
number of bugs caused by 80-to-{32,64}-bit rounding errors. It's also
a requirement for switching to signaling NaN for FixedDoubleArray holes.
R=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/
864483002
Cr-Commit-Position: refs/heads/master@{#26154}
jkummerow [Tue, 20 Jan 2015 12:49:34 +0000 (04:49 -0800)]
Let KeyedStoreIC go megamorphic (instead of generic) when the same map misses twice
BUG=v8:3826
LOG=n
R=verwaest@chromium.org
Review URL: https://codereview.chromium.org/
859943003
Cr-Commit-Position: refs/heads/master@{#26153}
ulan [Tue, 20 Jan 2015 11:38:42 +0000 (03:38 -0800)]
LCodeGen::CallKnownFunction gets the function in register.
There is no need to embed function pointer in code.
BUG=v8:3823
LOG=N
Review URL: https://codereview.chromium.org/
817483005
Cr-Commit-Position: refs/heads/master@{#26152}
jochen [Tue, 20 Jan 2015 11:31:44 +0000 (03:31 -0800)]
Add d8 target to BUILD.gn
R=machenbach@chromium.org,brettw@chromium.org
LOG=n
BUG=none
Review URL: https://codereview.chromium.org/
846743004
Cr-Commit-Position: refs/heads/master@{#26151}
machenbach [Tue, 20 Jan 2015 10:17:21 +0000 (02:17 -0800)]
Add no-test-harness mode to test driver.
BUG=
Review URL: https://codereview.chromium.org/
855993003
Cr-Commit-Position: refs/heads/master@{#26150}
titzer [Tue, 20 Jan 2015 09:45:02 +0000 (01:45 -0800)]
[turbofan] First version of loop peeling.
BUG=
Review URL: https://codereview.chromium.org/
816053002
Cr-Commit-Position: refs/heads/master@{#26149}
cdai2 [Tue, 20 Jan 2015 06:47:52 +0000 (14:47 +0800)]
X87: add stub for api function calls with known number of parameters
port
8c990d1716d33db4834b7e3c115874d482291e8c
original commit message:
add stub for api function calls with known number of parameters
BUG=
R=weiliang.lin@intel.com
Review URL: https://codereview.chromium.org/
852263004
Cr-Commit-Position: refs/heads/master@{#26148}
Akos Palfi [Mon, 19 Jan 2015 23:39:07 +0000 (00:39 +0100)]
MIPS: Add fast-path OOB support to KeyedLoadIC_Generic.
Port
79ed8f173ce1624f6d7821ea6b44ce75efa97912
BUG=
R=akos.palfi@imgtec.com
Review URL: https://codereview.chromium.org/
858893002
Cr-Commit-Position: refs/heads/master@{#26147}
ishell [Mon, 19 Jan 2015 17:49:13 +0000 (09:49 -0800)]
Massive renaming of PropertyType values and other implied stuff.
PropertyKind:
DATA -> kData
ACCESSOR -> kAccessor
PropertyType:
FIELD -> DATA
CONSTANT -> DATA_CONSTANT
ACCESSOR_FIELD -> ACCESSOR
CALLBACKS -> ACCESSOR_CONSTANT
PropertyLocation:
IN_OBJECT -> kField
IN_DESCRIPTOR -> kDescriptor
StoreMode:
FORCE_IN_OBJECT -> FORCE_FIELD
FieldDescriptor -> DataDescriptor
ConstantDescriptor -> DataConstantDescriptor
CallbacksDescriptor -> AccessorConstantDescriptor
Review URL: https://codereview.chromium.org/
856503002
Cr-Commit-Position: refs/heads/master@{#26146}
mstarzinger [Mon, 19 Jan 2015 16:51:39 +0000 (08:51 -0800)]
Remove exorbitant duplication of DebuggerHasBreakpoints.
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/
804713006
Cr-Commit-Position: refs/heads/master@{#26145}
machenbach [Mon, 19 Jan 2015 16:23:23 +0000 (08:23 -0800)]
Partially reland Auto-generate v8 version based on tags.
This relands parts of
https://codereview.chromium.org/
843913009
It prepares for using this script outside of v8, e.g. in a
chromium hook.
The script is intended to run as a hook and will create
version_gen.cc if the content has changed.
Changes to gyp and gn files can land as a follow up, once
calling the hook on the chromium side has landed.
BUG=chromium:446166
LOG=n
Review URL: https://codereview.chromium.org/
830093003
Cr-Commit-Position: refs/heads/master@{#26144}
Michael Starzinger [Mon, 19 Jan 2015 16:16:56 +0000 (17:16 +0100)]
Disable new test failing since
79748e3f7ced.
TBR=titzer@chromium.org
TEST=mjsunit/compiler/opt-next-call-turbo
Review URL: https://codereview.chromium.org/
862533002
Cr-Commit-Position: refs/heads/master@{#26143}
mstarzinger [Mon, 19 Jan 2015 15:51:50 +0000 (07:51 -0800)]
Remove overzealous check from %OptimizeFunctionOnNextCall.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/
863443003
Cr-Commit-Position: refs/heads/master@{#26142}
Ben L. Titzer [Mon, 19 Jan 2015 15:35:03 +0000 (16:35 +0100)]
[turbofan] Rename IrOpcode predicate IsLeafOpcode to IsConstantOpcode.
R=mstarzinger@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
863513002
Cr-Commit-Position: refs/heads/master@{#26141}
titzer [Mon, 19 Jan 2015 15:19:43 +0000 (07:19 -0800)]
[turbofan] Improve loop analysis to handle more than 32 loops.
R=mstarzinger@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
855653002
Cr-Commit-Position: refs/heads/master@{#26140}
svenpanne [Mon, 19 Jan 2015 14:41:39 +0000 (06:41 -0800)]
Unobscurified OFStream.
Use simple HAS-A relationship instead of obscure multiple inheritance.
In theory, UBSan should be totally happy with this, but it still isn't.
This seems to be a bug in UBSan, see e.g.
http://lists.cs.uiuc.edu/pipermail/cfe-dev/2014-December/040225.html
BUG=chromium:448102
LOG=y
Review URL: https://codereview.chromium.org/
859773002
Cr-Commit-Position: refs/heads/master@{#26139}
marja [Mon, 19 Jan 2015 14:29:01 +0000 (06:29 -0800)]
Script streaming: don't produce (empty) cached data when debugger forces eagerness.
See CompileTopLevel for similar logic which suppresses producing parser cache
when the debugger is active.
BUG=441130
LOG=n
Review URL: https://codereview.chromium.org/
845643005
Cr-Commit-Position: refs/heads/master@{#26138}
dcarney [Mon, 19 Jan 2015 14:17:23 +0000 (06:17 -0800)]
add stub for api function calls with known number of parameters
BUG=449930
LOG=N
Review URL: https://codereview.chromium.org/
859783002
Cr-Commit-Position: refs/heads/master@{#26137}
dcarney [Mon, 19 Jan 2015 14:16:13 +0000 (06:16 -0800)]
delete flaky test MultithreadedParallelIsolates
BUG=3215
LOG=N
Review URL: https://codereview.chromium.org/
855333002
Cr-Commit-Position: refs/heads/master@{#26136}
Michael Starzinger [Mon, 19 Jan 2015 14:15:04 +0000 (15:15 +0100)]
Blacklist computed property name failures with TurboFan.
R=verwaest@chromium.org
TEST=mjsunit/harmony/computed-property-names
Review URL: https://codereview.chromium.org/
862513002
Cr-Commit-Position: refs/heads/master@{#26135}
jochen [Mon, 19 Jan 2015 13:21:06 +0000 (05:21 -0800)]
Revert of Temporarily restore make dependencies (patchset #1 id:1 of https://codereview.chromium.org/
733203008/)
Reason for revert:
make dependencies is stale now, and everybody should have switched over
Original issue's description:
> Temporarily restore make dependencies
>
> We will remove dependencies in about a week again.
>
> BUG=none
> R=machenbach@chromium.org
> LOG=y
TBR=machenbach@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=none
Review URL: https://codereview.chromium.org/
791813007
Cr-Commit-Position: refs/heads/master@{#26134}
yangguo [Mon, 19 Jan 2015 12:46:32 +0000 (04:46 -0800)]
Serializer: reset typed array cell type.
R=mvstanton@chromium.org
BUG=chromium:449937
LOG=N
Review URL: https://codereview.chromium.org/
861583002
Cr-Commit-Position: refs/heads/master@{#26133}
caitpotter88 [Mon, 19 Jan 2015 12:38:01 +0000 (04:38 -0800)]
Report SyntaxError on Token::ILLEGAL in ParseTemplateLiteral
BUG=v8:3820, v8:3821
LOG=N
R=arv@chromium.org, dslomov@chromium.org, marja@chromium.org
Review URL: https://codereview.chromium.org/
857433004
Cr-Commit-Position: refs/heads/master@{#26132}
mstarzinger [Mon, 19 Jan 2015 12:35:05 +0000 (04:35 -0800)]
Allow --always-opt to go further into the pipeline (2).
R=rossberg@chromium.org
Review URL: https://codereview.chromium.org/
855873002
Cr-Commit-Position: refs/heads/master@{#26131}
caitpotter88 [Mon, 19 Jan 2015 12:22:22 +0000 (04:22 -0800)]
Implement IsConstructor() abstract operation
LOG=N
R=dpino@igalia.com, arv@chromium.org, dslomov@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
851163007
Cr-Commit-Position: refs/heads/master@{#26130}
dcarney [Mon, 19 Jan 2015 12:17:09 +0000 (04:17 -0800)]
remove SignatureInfo class
BUG=
Review URL: https://codereview.chromium.org/
855903002
Cr-Commit-Position: refs/heads/master@{#26129}
yangguo [Mon, 19 Jan 2015 11:32:58 +0000 (03:32 -0800)]
Allow nested compile-script histogram scopes.
R=bmeurer@chromium.org
BUG=chromium:449959
LOG=N
Review URL: https://codereview.chromium.org/
822823003
Cr-Commit-Position: refs/heads/master@{#26128}
jochen [Mon, 19 Jan 2015 10:12:28 +0000 (02:12 -0800)]
Add missing if (FLAG_verify_heap) checks around heap verification
BUG=chromium:449943
R=svenpanne@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
848403003
Cr-Commit-Position: refs/heads/master@{#26127}
mvstanton [Mon, 19 Jan 2015 09:31:02 +0000 (01:31 -0800)]
ClusterFuzz fix: %NormalizeElements shouldn't process the global proxy.
BUG=449070
R=yangguo@chromium.org
LOG=N
Review URL: https://codereview.chromium.org/
859713002
Cr-Commit-Position: refs/heads/master@{#26126}
yangguo [Mon, 19 Jan 2015 09:17:12 +0000 (01:17 -0800)]
Correctly load RegExp.source from the holder.
R=mvstanton@chromium.org
BUG=chromium:447561
LOG=N
Review URL: https://codereview.chromium.org/
844193004
Cr-Commit-Position: refs/heads/master@{#26125}
bmeurer [Mon, 19 Jan 2015 08:07:22 +0000 (00:07 -0800)]
[turbofan] Make Factory::NewNumber() always return the minus_zero_value.
TEST=unittests
Review URL: https://codereview.chromium.org/
857783002
Cr-Commit-Position: refs/heads/master@{#26124}
chunyang.dai [Mon, 19 Jan 2015 05:44:00 +0000 (21:44 -0800)]
X87: Add fast-path OOB support to KeyedLoadIC_Generic
port
79ed8f173ce1624f6d7821ea6b44ce75efa97912.
original commit message:
Add fast-path OOB support to KeyedLoadIC_Generic
BUG=
Review URL: https://codereview.chromium.org/
862443002
Cr-Commit-Position: refs/heads/master@{#26123}
chunyang.dai [Mon, 19 Jan 2015 05:06:29 +0000 (21:06 -0800)]
X87: split api call stubs into accessor and function call stubs
port
6950ead0b0a8373441dc8f8d3b2c316e82549ce9
original commit message:
split api call stubs into accessor and function call stubs
BUG=
Review URL: https://codereview.chromium.org/
860593002
Cr-Commit-Position: refs/heads/master@{#26122}
machenbach [Sun, 18 Jan 2015 15:43:02 +0000 (07:43 -0800)]
Revert of Reland Auto-generate v8 version based on tags. (patchset #1 id:1 of https://codereview.chromium.org/
843913009/)
Reason for revert:
The gclient hook of v8 is not executed on this builder:
http://build.chromium.org/p/tryserver.chromium.linux/builders/android_aosp/builds/49765
Original issue's description:
> Reland Auto-generate v8 version based on tags.
>
> This relands the CL
> https://codereview.chromium.org/
797503007/.
>
> It runs the version generation two times. First during
> runhooks as a fallback for recipes that loose git context
> (e.g. android_aosp). Second during compilation like in the
> original CL. In case of failures, the result from the
> runhooks call will be reused.
>
> BUG=chromium:446166
> LOG=n
>
> Committed: https://crrev.com/
d3d6e38b2cba73b52bd5ed00f4441a7db3284a19
> Cr-Commit-Position: refs/heads/master@{#26120}
TBR=jochen@chromium.org,jkummerow@chromium.org,tandrii@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:446166
Review URL: https://codereview.chromium.org/
855263002
Cr-Commit-Position: refs/heads/master@{#26121}
machenbach [Sun, 18 Jan 2015 11:25:41 +0000 (03:25 -0800)]
Reland Auto-generate v8 version based on tags.
This relands the CL
https://codereview.chromium.org/
797503007/.
It runs the version generation two times. First during
runhooks as a fallback for recipes that loose git context
(e.g. android_aosp). Second during compilation like in the
original CL. In case of failures, the result from the
runhooks call will be reused.
BUG=chromium:446166
LOG=n
Review URL: https://codereview.chromium.org/
843913009
Cr-Commit-Position: refs/heads/master@{#26120}
machenbach [Sat, 17 Jan 2015 13:45:27 +0000 (05:45 -0800)]
Revert of [turbofan] Fix bunch of tests failing with --turbo-deoptimization. (patchset #1 id:1 of https://codereview.chromium.org/
786333004/)
Reason for revert:
This changes layout test expectations. I ran a bisect over the layout tests:
Bad:
http://build.chromium.org/p/client.v8/builders/V8-Blink%20Linux%2032/builds/1600
Good:
http://build.chromium.org/p/client.v8/builders/V8-Blink%20Linux%2032/builds/1599
If this is intentional please first land a CL with manualrebaseline test expectations for the changed tests and then reland.
Original issue's description:
> [turbofan] Fix bunch of tests failing with --turbo-deoptimization.
>
> R=bmeurer@chromium.org
>
> Committed: https://crrev.com/
e9e772121b36697821dbfff61f36e0a68367f21c
> Cr-Commit-Position: refs/heads/master@{#26115}
TBR=bmeurer@chromium.org,jarin@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
804333005
Cr-Commit-Position: refs/heads/master@{#26119}
akos.palfi [Sat, 17 Jan 2015 09:02:07 +0000 (01:02 -0800)]
MIPS: Fix after 'MIPS: split api call stubs into accessor and function call stubs'.
Port
6950ead0b0a8373441dc8f8d3b2c316e82549ce9
This commit fixes the remaining test failures after:
ec52f8df7f0956ecaed0422110847ee5634c1304
BUG=
Review URL: https://codereview.chromium.org/
855183003
Cr-Commit-Position: refs/heads/master@{#26118}
v8-autoroll [Sat, 17 Jan 2015 04:29:07 +0000 (20:29 -0800)]
Update V8 DEPS.
Rolling v8/tools/clang to
97bfed3644c03f48864c479b8714f08ed80c5d13
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
851413002
Cr-Commit-Position: refs/heads/master@{#26117}
akos.palfi [Sat, 17 Jan 2015 03:21:42 +0000 (19:21 -0800)]
MIPS: split api call stubs into accessor and function call stubs
Port
6950ead0b0a8373441dc8f8d3b2c316e82549ce9
Note: This port has known issues, there are about 40 test failures. The purpose of this commit is to unbreak the MIPS build.
BUG=
Review URL: https://codereview.chromium.org/
853323002
Cr-Commit-Position: refs/heads/master@{#26116}
jarin [Fri, 16 Jan 2015 22:23:23 +0000 (14:23 -0800)]
[turbofan] Fix bunch of tests failing with --turbo-deoptimization.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
786333004
Cr-Commit-Position: refs/heads/master@{#26115}
yangguo [Fri, 16 Jan 2015 22:22:13 +0000 (14:22 -0800)]
Use back reference map to find references to global proxy.
R=mvstanton@chromium.org
Review URL: https://codereview.chromium.org/
825933004
Cr-Commit-Position: refs/heads/master@{#26114}
jarin [Fri, 16 Jan 2015 22:21:08 +0000 (14:21 -0800)]
[turbofan] Add frame state to the PlainPrimitiveToNumber unit test.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/
792043007
Cr-Commit-Position: refs/heads/master@{#26113}
mstarzinger [Fri, 16 Jan 2015 22:20:00 +0000 (14:20 -0800)]
Allow --always-opt to go further into the pipeline (1).
R=rossberg@chromium.org
Review URL: https://codereview.chromium.org/
841213003
Cr-Commit-Position: refs/heads/master@{#26112}
jkummerow [Fri, 16 Jan 2015 22:18:48 +0000 (14:18 -0800)]
Add fast-path OOB support to KeyedLoadIC_Generic
R=verwaest@chromium.org
Review URL: https://codereview.chromium.org/
852313002
Cr-Commit-Position: refs/heads/master@{#26111}
yangguo [Fri, 16 Jan 2015 22:17:37 +0000 (14:17 -0800)]
Extend and fix tests for custom heap snapshot.
R=vogelheim@chromium.org
Review URL: https://codereview.chromium.org/
856793002
Cr-Commit-Position: refs/heads/master@{#26110}
Michael Achenbach [Fri, 16 Jan 2015 21:56:03 +0000 (22:56 +0100)]
Whitespace change to trigger bots.
Cr-Commit-Position: refs/heads/master@{#26109}
Michael Achenbach [Fri, 16 Jan 2015 20:56:32 +0000 (21:56 +0100)]
Whitespace change to trigger bots.
Cr-Commit-Position: refs/heads/master@{#26108}
machenbach [Fri, 16 Jan 2015 19:17:55 +0000 (11:17 -0800)]
Revert of Revert of [turbofan] Improve memory layout of Node class. (patchset #1 id:1 of https://codereview.chromium.org/
856813003/)
Reason for revert:
Revert didn't help.
Original issue's description:
> Revert of [turbofan] Improve memory layout of Node class. (patchset #1 id:1 of https://codereview.chromium.org/
828803004/)
>
> Reason for revert:
> Speculative revert to see if win XP recovers. Will reland if not.
>
> Original issue's description:
> > [turbofan] Improve memory layout of Node class.
> >
> > Also gets rid of the reinterpret_cast on this.
> >
> > R=svenpanne@chromium.org
> >
> > Committed: https://crrev.com/
f91e8046fad99365c74ccbb39d42de58398dab57
> > Cr-Commit-Position: refs/heads/master@{#26103}
>
> TBR=svenpanne@chromium.org,dcarney@chromium.org,bmeurer@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://crrev.com/
dc6f34d83650dce8faa3e82e87c050f12732d4b3
> Cr-Commit-Position: refs/heads/master@{#26106}
TBR=svenpanne@chromium.org,dcarney@chromium.org,bmeurer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
855723002
Cr-Commit-Position: refs/heads/master@{#26107}
machenbach [Fri, 16 Jan 2015 16:49:58 +0000 (08:49 -0800)]
Revert of [turbofan] Improve memory layout of Node class. (patchset #1 id:1 of https://codereview.chromium.org/
828803004/)
Reason for revert:
Speculative revert to see if win XP recovers. Will reland if not.
Original issue's description:
> [turbofan] Improve memory layout of Node class.
>
> Also gets rid of the reinterpret_cast on this.
>
> R=svenpanne@chromium.org
>
> Committed: https://crrev.com/
f91e8046fad99365c74ccbb39d42de58398dab57
> Cr-Commit-Position: refs/heads/master@{#26103}
TBR=svenpanne@chromium.org,dcarney@chromium.org,bmeurer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
856813003
Cr-Commit-Position: refs/heads/master@{#26106}
mvstanton [Fri, 16 Jan 2015 13:28:22 +0000 (05:28 -0800)]
Vector-based KeyedLoadIC MISS logic needs improvement.
Two issues:
1) We would trace the MISS two times on vector-based KeyedLoadIC
with a string key in MEGAMORPHIC state.
2) There was a confusing asymmetry in the handling of going
GENERIC. This change makes it the same whether
--vector-ics is on or not.
BUG=
Review URL: https://codereview.chromium.org/
839893005
Cr-Commit-Position: refs/heads/master@{#26105}
mstarzinger [Fri, 16 Jan 2015 13:21:59 +0000 (05:21 -0800)]
Remove distinction between DisableTurbofan and DisableCrankshaft.
R=rossberg@chromium.org
Review URL: https://codereview.chromium.org/
826833004
Cr-Commit-Position: refs/heads/master@{#26104}
bmeurer [Fri, 16 Jan 2015 13:16:32 +0000 (05:16 -0800)]
[turbofan] Improve memory layout of Node class.
Also gets rid of the reinterpret_cast on this.
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/
828803004
Cr-Commit-Position: refs/heads/master@{#26103}
mstarzinger [Fri, 16 Jan 2015 12:53:24 +0000 (04:53 -0800)]
Enable test coverage for --turbo-deoptimization mode.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
845193004
Cr-Commit-Position: refs/heads/master@{#26102}
mstarzinger [Fri, 16 Jan 2015 12:38:13 +0000 (04:38 -0800)]
First simple implementation of class literals in TurboFan.
R=rossberg@chromium.org,jarin@chromium.org
TEST=cctest/test-run-jsops/ClassLiteral
Review URL: https://codereview.chromium.org/
798873006
Cr-Commit-Position: refs/heads/master@{#26101}
marja.holtta [Fri, 16 Jan 2015 12:01:12 +0000 (04:01 -0800)]
Restore the sanity of AUTHORS. CLA OK.
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
806023003
Cr-Commit-Position: refs/heads/master@{#26100}
Balazs Kilvady [Fri, 16 Jan 2015 11:22:07 +0000 (12:22 +0100)]
MIPS64: fix Make map check in NamedStoreHandlerCompiler::GenerateFieldTypeChecks weak.
BUG=
R=balazs.kilvady@imgtec.com, ulan@chromium.org
Review URL: https://codereview.chromium.org/
854073004
Cr-Commit-Position: refs/heads/master@{#26099}
bmeurer [Fri, 16 Jan 2015 11:04:01 +0000 (03:04 -0800)]
[turbofan] Initial attempt to cleanup Node and related classes.
- Make Node::Inputs and Node::Uses mostly STL compliant.
- Get rid of some pre-C++11 crappiness.
- Start moving unit tests from cctest to unittests.
- TrimInputCount() now tries to reserve inputs slots for
later appending.
- Fix numerous style guide violations.
TEST=cctest,unittests
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/
851263002
Cr-Commit-Position: refs/heads/master@{#26098}
dcarney [Fri, 16 Jan 2015 10:59:01 +0000 (02:59 -0800)]
split api call stubs into accessor and function call stubs
BUG=
Review URL: https://codereview.chromium.org/
836093007
Cr-Commit-Position: refs/heads/master@{#26097}
bmeurer [Fri, 16 Jan 2015 10:40:23 +0000 (02:40 -0800)]
Migrate TRACED_FOREACH to C++11.
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/
851243002
Cr-Commit-Position: refs/heads/master@{#26096}
jochen [Fri, 16 Jan 2015 10:39:16 +0000 (02:39 -0800)]
Add checks to the marking queue
Hopefully we'll catch heap corruption earlier where identifying the
object that holds a stale pointer.
Speaking of staleness, also remove old debugging code.
BUG=chromium:128415
R=ulan@chromium.org,hpayer@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
843013005
Cr-Commit-Position: refs/heads/master@{#26095}
bmeurer [Fri, 16 Jan 2015 10:36:19 +0000 (02:36 -0800)]
[turbofan] Only mark non-smis as references.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/
812033012
Cr-Commit-Position: refs/heads/master@{#26094}
Sven Panne [Fri, 16 Jan 2015 10:12:02 +0000 (11:12 +0100)]
Another attempt to fix regress-crbug-178790.
This time we simply undo the change introduced by the PPC port for
this test. No idea why it should be necessary, and Windows XP
obviously doesn't give us that much stack, anyway.
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
826833003
Cr-Commit-Position: refs/heads/master@{#26093}
Sven Panne [Fri, 16 Jan 2015 09:45:21 +0000 (10:45 +0100)]
PPC aftermath: Fix regress-crbug-178790.
The test fails on XP only, so let's tentatively raise the stack limit more. We probably need to investigate what a tighter limit might be and (more importantly) what the underlying reason for the failure is.
Hopefully 1800kB is enough, we can't test this via try jobs, because we don't have XP try bots. :-/
R=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
791693005
Cr-Commit-Position: refs/heads/master@{#26092}
Sven Panne [Fri, 16 Jan 2015 07:42:00 +0000 (08:42 +0100)]
Contribution of PowerPC port (continuation of
422063005)
Contribution of PowerPC port (continuation of
422063005). The inital patch
covers the core changes to the common files. Subsequent patches will cover
changes to common files to support AIX and to update the ppc directories so
they are current with the changes in the rest of the project.
This is based off of the GitHub repository
https://github.com/andrewlow/v8ppc
BUG=
R=svenpanne@chromium.org, danno@chromium.org, sevnpanne@chromium.org
Review URL: https://codereview.chromium.org/
817143002
Cr-Commit-Position: refs/heads/master@{#26091}
cdai2 [Fri, 16 Jan 2015 05:24:24 +0000 (13:24 +0800)]
X87: ES6 computed property names
port
74e38e34b3caecfb660c3ed1336a58c551215793.
original commit message:
This adds support for computed property names, under the flag
--harmony-computed-property-names, for both object literals and classes.
This is a revert of the revert,
7d48fd9dc29c8f5d7c8b5215f9bd537cdd005eba.
BUG=
R=weiliang.lin@intel.com
Review URL: https://codereview.chromium.org/
825593004
Cr-Commit-Position: refs/heads/master@{#26090}
deftly [Fri, 16 Jan 2015 05:07:10 +0000 (21:07 -0800)]
fix defines for openbsd - build fails without them:
Original version from robert@openbsd.org with modification from myself
BUG=
Review URL: https://codereview.chromium.org/
856553002
Cr-Commit-Position: refs/heads/master@{#26089}
paul.lind [Fri, 16 Jan 2015 04:42:30 +0000 (20:42 -0800)]
MIPS: ES6 computed property names.
Port:
74e38e34b3caecfb660c3ed1336a58c551215793
Original commit message:
This adds support for computed property names, under the flag
--harmony-computed-property-names, for both object literals and
classes.
This is a revert of the revert,
7d48fd9dc29c8f5d7c8b5215f9bd537cdd005eba.
BUG=v8:3754
LOG=Y
Review URL: https://codereview.chromium.org/
829913005
Cr-Commit-Position: refs/heads/master@{#26088}
v8-autoroll [Fri, 16 Jan 2015 04:41:20 +0000 (20:41 -0800)]
Update V8 DEPS.
Rolling v8/build/gyp to
194ec65a55edf24c1976d4265f57e8c90ef5bb2f
Rolling v8/tools/clang to
790f3b26cdce99c8e5751893a18134afd0202ed0
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
814833012
Cr-Commit-Position: refs/heads/master@{#26087}
arv [Fri, 16 Jan 2015 00:20:00 +0000 (16:20 -0800)]
Computed property names: Enable test on windows again
Enable test on windows again and disable the problematic lines only.
BUG=v8:3815
LOG=N
R=dslomov@chromium.org
Review URL: https://codereview.chromium.org/
850773004
Cr-Commit-Position: refs/heads/master@{#26086}
Erik Arvidsson [Thu, 15 Jan 2015 20:51:22 +0000 (15:51 -0500)]
Skip computed property name test on Windows
BUG=v8:3815
LOG=N
TBR=dslomov@chromium.org
Review URL: https://codereview.chromium.org/
854903002
Cr-Commit-Position: refs/heads/master@{#26085}
arv [Thu, 15 Jan 2015 20:02:20 +0000 (12:02 -0800)]
ES6 computed property names
This adds support for computed property names, under the flag
--harmony-computed-property-names, for both object literals and
classes.
This is a revert of the revert,
7d48fd9dc29c8f5d7c8b5215f9bd537cdd005eba.
BUG=v8:3754
LOG=Y
R=dslomov@chromium.org
Review URL: https://codereview.chromium.org/
798243004
Cr-Commit-Position: refs/heads/master@{#26084}
adamk [Thu, 15 Jan 2015 19:18:05 +0000 (11:18 -0800)]
Remove dead TargetCollector code from the parser
This also simplifies target_stack_ and tightens up the types of nodes
that can appear in it.
Best I can tell, this code has been dead since
https://code.google.com/p/v8/source/detail?r=7542
Review URL: https://codereview.chromium.org/
850093004
Cr-Commit-Position: refs/heads/master@{#26083}
bmeurer [Thu, 15 Jan 2015 17:08:41 +0000 (09:08 -0800)]
[turbofan] Improve AstGraphBuilder::BuildToBoolean() for heap constants.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/
852963004
Cr-Commit-Position: refs/heads/master@{#26082}
bmeurer [Thu, 15 Jan 2015 15:29:31 +0000 (07:29 -0800)]
Revert of [turbofan] Make Factory::NewNumber() always return the minus_zero_value. (patchset #1 id:1 of https://codereview.chromium.org/
838263006/)
Reason for revert:
Breaks the tree
Original issue's description:
> [turbofan] Make Factory::NewNumber() always return the minus_zero_value.
>
> TEST=unittests
> R=mstarzinger@chromium.org, dcarney@chromium.org
>
> Committed: https://crrev.com/
55b3e268b2895e2c09230d39669a75b6ab4e7e4d
> Cr-Commit-Position: refs/heads/master@{#26080}
TBR=dcarney@chromium.org,mstarzinger@chromium.org
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
853683009
Cr-Commit-Position: refs/heads/master@{#26081}
Benedikt Meurer [Thu, 15 Jan 2015 15:09:50 +0000 (16:09 +0100)]
[turbofan] Make Factory::NewNumber() always return the minus_zero_value.
TEST=unittests
R=mstarzinger@chromium.org, dcarney@chromium.org
Review URL: https://codereview.chromium.org/
838263006
Cr-Commit-Position: refs/heads/master@{#26080}
bmeurer [Thu, 15 Jan 2015 15:07:29 +0000 (07:07 -0800)]
[turbofan] Turn IrOpcode::IsXXX() into range checks.
TEST=unittests
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/
807573004
Cr-Commit-Position: refs/heads/master@{#26079}
bmeurer [Thu, 15 Jan 2015 13:26:40 +0000 (05:26 -0800)]
Revert of [turbofan] Use PlainPrimitiveToNumber whenever possible. (patchset #1 id:1 of https://codereview.chromium.org/
852763002/)
Reason for revert:
Tanks performance of SQLite and towers.c.
Original issue's description:
> [turbofan] Use PlainPrimitiveToNumber whenever possible.
>
> TEST=cctest,unittests
> R=jarin@chromium.org
>
> Committed: https://crrev.com/
1a5db24e6bf831e61e3f4aa0c252f77e48c3689c
> Cr-Commit-Position: refs/heads/master@{#26071}
TBR=jarin@chromium.org,dcarney@chromium.org
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
852153003
Cr-Commit-Position: refs/heads/master@{#26078}
marja [Thu, 15 Jan 2015 13:10:01 +0000 (05:10 -0800)]
Check author of a CL against the AUTHORS file.
The code is copied from Chromium presubmit.
BUG=
Review URL: https://codereview.chromium.org/
837503003
Cr-Commit-Position: refs/heads/master@{#26077}
mvstanton [Thu, 15 Jan 2015 12:52:33 +0000 (04:52 -0800)]
Vector-based LoadICs need to handle Smi receivers.
The MISS handler was being called when the receiver was a Smi, instead,
we should recognize the case and use the heap number map.
BUG=
Review URL: https://codereview.chromium.org/
854623002
Cr-Commit-Position: refs/heads/master@{#26076}
mstarzinger [Thu, 15 Jan 2015 12:47:53 +0000 (04:47 -0800)]
Remove redundant test expectation from mjsunit.
R=jkummerow@chromium.org
TEST=mjsunit/nans
Review URL: https://codereview.chromium.org/
850303002
Cr-Commit-Position: refs/heads/master@{#26075}
dcarney [Thu, 15 Jan 2015 12:17:23 +0000 (04:17 -0800)]
[turbofan] add flag for register allocation verification
BUG=
Review URL: https://codereview.chromium.org/
798873008
Cr-Commit-Position: refs/heads/master@{#26074}
Yang Guo [Thu, 15 Jan 2015 11:48:57 +0000 (12:48 +0100)]
Fix number-string cache test case.
TBR=jochen@chromium.org
Review URL: https://codereview.chromium.org/
843913008
Cr-Commit-Position: refs/heads/master@{#26073}
yangguo [Thu, 15 Jan 2015 11:22:04 +0000 (03:22 -0800)]
Fix remaining issues in the custom snapshot.
Math functions:
Some Math functions require typed arrays for their implementation. The embedded
script may call those Math functions. The serializer needs to deal with this.
Added assertion to make sure no other typed array is created when snapshotting.
Number-string cache:
We assume that the initial snapshot does not expand the number-string cache.
This is no longer true for custom heap snapshots.
Bound functions:
Bound functions store the bound arguments in a COW fixed array, including the
bindee function. COW arrays are serialized into the startup snapshot and
referenced in the partial snapshot via partial snapshot cache. However, the
bindee function is context-dependent and must not be part of the startup
snapshot. There is no need for bound functions to use a COW array though.
R=jochen@chromium.org
Review URL: https://codereview.chromium.org/
851073002
Cr-Commit-Position: refs/heads/master@{#26072}
bmeurer [Thu, 15 Jan 2015 09:49:14 +0000 (01:49 -0800)]
[turbofan] Use PlainPrimitiveToNumber whenever possible.
TEST=cctest,unittests
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
852763002
Cr-Commit-Position: refs/heads/master@{#26071}
ulan [Thu, 15 Jan 2015 09:43:25 +0000 (01:43 -0800)]
Make map check in NamedStoreHandlerCompiler::GenerateFieldTypeChecks weak.
BUG=v8:3629
LOG=N
Review URL: https://codereview.chromium.org/
853713002
Cr-Commit-Position: refs/heads/master@{#26070}
dcarney [Thu, 15 Jan 2015 09:05:52 +0000 (01:05 -0800)]
Reland "[turbofan] simplify gap ordering"
BUG=
Review URL: https://codereview.chromium.org/
854703002
Cr-Commit-Position: refs/heads/master@{#26069}
dcarney [Thu, 15 Jan 2015 08:55:51 +0000 (00:55 -0800)]
add register allocator regression test
BUG=
Review URL: https://codereview.chromium.org/
852633002
Cr-Commit-Position: refs/heads/master@{#26068}
machenbach [Thu, 15 Jan 2015 08:01:03 +0000 (00:01 -0800)]
Revert of Auto-generate v8 version based on tags. (patchset #5 id:80001 of https://codereview.chromium.org/
797503007/)
Reason for revert:
Blocks roll on android_aosp:
https://codereview.chromium.org/
851953005/
Original issue's description:
> Auto-generate v8 version based on tags.
>
> BUG=chromium:446166
> LOG=y
>
> Committed: https://crrev.com/
b301b85be895c6fcd1edfe2fd1e60b5abd0ac64d
> Cr-Commit-Position: refs/heads/master@{#26062}
TBR=jochen@chromium.org,jkummerow@chromium.org,tandrii@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=chromium:446166
Review URL: https://codereview.chromium.org/
850263002
Cr-Commit-Position: refs/heads/master@{#26067}
baptiste.afsa [Thu, 15 Jan 2015 06:30:51 +0000 (22:30 -0800)]
[turbofan] Allow 0.0 as immediate for floating-point comparison on arm/arm64.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
850073002
Cr-Commit-Position: refs/heads/master@{#26066}
v8-autoroll [Thu, 15 Jan 2015 06:29:40 +0000 (22:29 -0800)]
Update V8 DEPS.
Rolling v8/build/gyp to
789a019a832029f9a6f8d906a4bcee4736eabc90
Rolling v8/testing/gtest to
be1868139ffe0ccd0e8e3b37292b84c821d9c8ad
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
850223003
Cr-Commit-Position: refs/heads/master@{#26065}