platform/upstream/v8.git
9 years agoEagerly escape RegExp.source.
yangguo [Thu, 9 Apr 2015 13:22:17 +0000 (06:22 -0700)]
Eagerly escape RegExp.source.

Escaping used to happen lazily, implemented in an accessor property.
However, native implementation of RegExp methods use .source as well.
This leads to performance regressions. Now we do it eagerly instead.

R=jkummerow@chromium.org
BUG=chromium:436447
LOG=N

Review URL: https://codereview.chromium.org/1070093002

Cr-Commit-Position: refs/heads/master@{#27705}

9 years ago[cq] Add win nosnap shared trybot.
machenbach [Thu, 9 Apr 2015 12:43:28 +0000 (05:43 -0700)]
[cq] Add win nosnap shared trybot.

TBR=tandrii@chromium.org

Review URL: https://codereview.chromium.org/1078703002

Cr-Commit-Position: refs/heads/master@{#27704}

9 years agoRelax heap size assert during compaction
erikcorry [Thu, 9 Apr 2015 12:42:22 +0000 (05:42 -0700)]
Relax heap size assert during compaction

R=ulan@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1078533003

Cr-Commit-Position: refs/heads/master@{#27703}

9 years ago[turbofan] Disable select matching due to bug manifesting on arm.
titzer [Thu, 9 Apr 2015 12:41:15 +0000 (05:41 -0700)]
[turbofan] Disable select matching due to bug manifesting on arm.

R=machenbach@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1077613002

Cr-Commit-Position: refs/heads/master@{#27702}

9 years ago[turbofan] Push layout descriptors to InstructionOperand subclasses.
dcarney [Thu, 9 Apr 2015 11:28:54 +0000 (04:28 -0700)]
[turbofan] Push layout descriptors to InstructionOperand subclasses.

Review URL: https://codereview.chromium.org/1070043002

Cr-Commit-Position: refs/heads/master@{#27701}

9 years agoMIPS: [turbofan] Materialize JSFunction from frame if possible.
balazs.kilvady [Thu, 9 Apr 2015 11:16:02 +0000 (04:16 -0700)]
MIPS: [turbofan] Materialize JSFunction from frame if possible.

Port 725cdc533cb4e31c042d32ce1979012b5bd99ced

Original commit message:
This reduces the overhead of recursive calls when context specialization
is enabled. Based on this it might be possible to further reduce the
overhead by also specializing the call itself.

As a drive-by-fix, port the fast context materialization optimization to
arm and arm64, that was previously only supported on x64 and ia32.

BUG=

Review URL: https://codereview.chromium.org/1074743002

Cr-Commit-Position: refs/heads/master@{#27700}

9 years agoMake TestInternalWeakLists more robust against flags.
mstarzinger [Thu, 9 Apr 2015 11:01:19 +0000 (04:01 -0700)]
Make TestInternalWeakLists more robust against flags.

This makes some tests in test-heap.cc more robust against flags, now
that top-level code can be optimized by the --always-opt flag.

R=hpayer@chromium.org
TEST=cctest/test-heap/TestInternalWeakLists

Review URL: https://codereview.chromium.org/1071063004

Cr-Commit-Position: refs/heads/master@{#27699}

9 years ago[turbofan] Make AllocatedOperand an InstructionOperand::Kind.
dcarney [Thu, 9 Apr 2015 10:40:41 +0000 (03:40 -0700)]
[turbofan] Make AllocatedOperand an InstructionOperand::Kind.

This is preparatory work to have MachineTypes encoded in AllocatedOperands.

Review URL: https://codereview.chromium.org/1075863002

Cr-Commit-Position: refs/heads/master@{#27698}

9 years ago[turbofan] Add poor man's store elimination for storing to fields.
Benedikt Meurer [Thu, 9 Apr 2015 10:29:15 +0000 (12:29 +0200)]
[turbofan] Add poor man's store elimination for storing to fields.

This is a very simple dead store elimination that removes StoreField
nodes which are immediately followed by other StoreField nodes that
store to the same field. Ideally there should be a fully featured store
elimination, which walks over the effect graph starting from the end,
but there are some technical difficulties to solve before we can get to
that, esp. we need to think about "effect producing" operators like
ValueEffect first. Once we have that, it is trivial to remove this temporary
poor man's store elimination.

R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/1070003002

Cr-Commit-Position: refs/heads/master@{#27697}

9 years ago[test-runner] Pass slowest test durations to buildbot.
machenbach [Thu, 9 Apr 2015 09:55:59 +0000 (02:55 -0700)]
[test-runner] Pass slowest test durations to buildbot.

BUG=

Review URL: https://codereview.chromium.org/1070753003

Cr-Commit-Position: refs/heads/master@{#27696}

9 years agoUpdate V8 DEPS.
v8-autoroll [Thu, 9 Apr 2015 09:51:40 +0000 (02:51 -0700)]
Update V8 DEPS.

Rolling v8/tools/clang to 5e71fff34c77c4ee9056fe3f2124c433768e2282

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1069983004

Cr-Commit-Position: refs/heads/master@{#27695}

9 years agoX87: Code cleanup in GenerateRecordCallTarget.
chunyang.dai [Thu, 9 Apr 2015 09:31:53 +0000 (02:31 -0700)]
X87: Code cleanup in GenerateRecordCallTarget.

port 6a222b8ff033eb523149794d37c0d0f29bc8fc40 (r27630)

original commit message:

  Code cleanup in GenerateRecordCallTarget

BUG=

Review URL: https://codereview.chromium.org/1074683003

Cr-Commit-Position: refs/heads/master@{#27694}

9 years agoX87: Make --always-opt also optimize top-level code
chunyang.dai [Thu, 9 Apr 2015 09:30:47 +0000 (02:30 -0700)]
X87: Make --always-opt also optimize top-level code

port 2d281e71ac49058ef5696f4255ff1fa30796bc5e (r27633)

original commit message:

    Make --always-opt also optimize top-level code.

    This enables eager optimization of top-level code with TurboFan and
    extends test coverage by triggering it with the --always-opt flag.
    Script contexts are now also properly allocated in TurboFan.

BUG=

Review URL: https://codereview.chromium.org/1077523002

Cr-Commit-Position: refs/heads/master@{#27693}

9 years ago[turbofan] cleanup InstructionOperand a little
dcarney [Thu, 9 Apr 2015 09:15:28 +0000 (02:15 -0700)]
[turbofan] cleanup InstructionOperand a little

- ConstantOperand was using a too-small field too store its virtual register
- drop ConvertTo, replace it with simple copy
- split AllocatedOperand off from Immediate and Constant to make assignment clearer, also paving the way for small Immediates
- put zone first in *Operand::New
- driveby: drop delayed ssa deconstruction experiment

R=titzer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1050803002

Cr-Commit-Position: refs/heads/master@{#27692}

9 years agoVectorICs - turn on vector ICs for LoadIC and KeyedLoadIC
mvstanton [Thu, 9 Apr 2015 09:06:00 +0000 (02:06 -0700)]
VectorICs - turn on vector ICs for LoadIC and KeyedLoadIC

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/1075823002

Cr-Commit-Position: refs/heads/master@{#27691}

9 years agoDon't #define snprintf in VS2015 - it's illegal and unneeded.
brucedawson [Thu, 9 Apr 2015 09:02:52 +0000 (02:02 -0700)]
Don't #define snprintf in VS2015 - it's illegal and unneeded.

VS 2015 supplies a conforming snprintf implementation, so #define
snprintf is no longer needed. Also, VS 2015 checks for #define of
snprintf and treats it as a fatal error.

LOG=Y
R=jarin@chromium.org
BUG=440500

Review URL: https://codereview.chromium.org/1078453002

Cr-Commit-Position: refs/heads/master@{#27690}

9 years agoMarkBit cleanup: They have to be accessed through Marking accessors. Avoid arbitrary...
hpayer [Thu, 9 Apr 2015 08:55:37 +0000 (01:55 -0700)]
MarkBit cleanup: They have to be accessed through Marking accessors. Avoid arbitrary abuse of mark bits and make marking explicit.

Added DCHECKs to mark bit transitions to check source state.

BUG=

Review URL: https://codereview.chromium.org/1040443002

Cr-Commit-Position: refs/heads/master@{#27689}

9 years agoBlacklist tests due to optimizing toplevel with --always-opt.
yangguo [Thu, 9 Apr 2015 08:47:28 +0000 (01:47 -0700)]
Blacklist tests due to optimizing toplevel with --always-opt.

Those two tests fail in gc-stress and custom snapshot (embedding mjsunit.js).
This is likely due to different GC timing with the custom snapshot.

R=mstarzinger@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1078653002

Cr-Commit-Position: refs/heads/master@{#27688}

9 years agoDisable more failing tests after f3338dd3b01c.
mstarzinger [Thu, 9 Apr 2015 08:21:38 +0000 (01:21 -0700)]
Disable more failing tests after f3338dd3b01c.

TBR=machenbach@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1075813002

Cr-Commit-Position: refs/heads/master@{#27687}

9 years ago[turbofan] Materialize JSFunction from frame if possible.
bmeurer [Thu, 9 Apr 2015 07:41:06 +0000 (00:41 -0700)]
[turbofan] Materialize JSFunction from frame if possible.

This reduces the overhead of recursive calls when context specialization
is enabled. Based on this it might be possible to further reduce the
overhead by also specializing the call itself.

As a drive-by-fix, port the fast context materialization optimization to
arm and arm64, that was previously only supported on x64 and ia32.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1072743002

Cr-Commit-Position: refs/heads/master@{#27686}

9 years agoRevert of Revert of X87: Reimplement Maps and Sets in JS (patchset #1 id:1 of https...
hablich [Thu, 9 Apr 2015 07:11:12 +0000 (00:11 -0700)]
Revert of Revert of X87: Reimplement Maps and Sets in JS (patchset #1 id:1 of https://codereview.chromium.org/1073723002/)

Reason for revert:
Revert the revert as this commit cannot be the cause for the closed tree.

Original issue's description:
> Revert of X87: Reimplement Maps and Sets in JS (patchset #1 id:1 of https://codereview.chromium.org/1066373002/)
>
> Reason for revert:
> Reverting as it resulted in a closed waterfall.
>
> Original issue's description:
> > X87: Reimplement Maps and Sets in JS
> >
> > port 909500aa1db9789b68e101045a6359a7fcb30e83 (r27605)
> >
> > original commit message:
> >     Previously, the only optimized code path for Maps and Sets was for String keys.
> >     This was achieved through an implementation of various complex operations
> >     in Hydrogen. This approach was neither scalable nor forward-compatible.
> >
> >     This patch adds the necessary intrinsics to implement Maps and Sets almost entirely
> >     in JS. The added intrinsics are:
> >
> >       %_FixedArrayGet
> >       %_FixedArraySet
> >       %_TheHole
> >       %_JSCollectionGetTable
> >       %_StringGetRawHashField
> >
> >     With these additions, as well as a few changes to what's exposed as runtime functions,
> >     most of the C++ code backing Maps and Sets is gone (including both runtime code in
> >     objects.cc and Crankshaft in hydrogen.cc).
> >
> > BUG=
> >
> > Committed: https://crrev.com/56600a35a49ffa5abcba66b14839089de3589ad9
> > Cr-Commit-Position: refs/heads/master@{#27681}
>
> TBR=weiliang.lin@intel.com,chunyang.dai@intel.com
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=
>
> Committed: https://crrev.com/a0486f128109443ed07802fb463c267e53533d81
> Cr-Commit-Position: refs/heads/master@{#27682}

TBR=weiliang.lin@intel.com,chunyang.dai@intel.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/1077543002

Cr-Commit-Position: refs/heads/master@{#27685}

9 years agoX87: JSEntryTrampoline: check for stack space before pushing arguments
chunyang.dai [Thu, 9 Apr 2015 06:51:59 +0000 (23:51 -0700)]
X87: JSEntryTrampoline: check for stack space before pushing arguments

port 146598f44a1efd0f942745abfd6c08a073972c01 (r27614)

original commit message:

  Optimistically pushing a lot of arguments can run into the stack limit of the
  process, at least on operating systems where this limit is close to the limit
  that V8 sets for itself.

BUG=

Review URL: https://codereview.chromium.org/1069283002

Cr-Commit-Position: refs/heads/master@{#27684}

9 years ago[ia32] Introduce BMI instructions.
jing.bao [Thu, 9 Apr 2015 06:50:52 +0000 (23:50 -0700)]
[ia32] Introduce BMI instructions.

BUG=v8:4015
LOG=n

Review URL: https://codereview.chromium.org/1069683002

Cr-Commit-Position: refs/heads/master@{#27683}

9 years agoRevert of X87: Reimplement Maps and Sets in JS (patchset #1 id:1 of https://coderevie...
hablich [Thu, 9 Apr 2015 06:49:30 +0000 (23:49 -0700)]
Revert of X87: Reimplement Maps and Sets in JS (patchset #1 id:1 of https://codereview.chromium.org/1066373002/)

Reason for revert:
Reverting as it resulted in a closed waterfall.

Original issue's description:
> X87: Reimplement Maps and Sets in JS
>
> port 909500aa1db9789b68e101045a6359a7fcb30e83 (r27605)
>
> original commit message:
>     Previously, the only optimized code path for Maps and Sets was for String keys.
>     This was achieved through an implementation of various complex operations
>     in Hydrogen. This approach was neither scalable nor forward-compatible.
>
>     This patch adds the necessary intrinsics to implement Maps and Sets almost entirely
>     in JS. The added intrinsics are:
>
>       %_FixedArrayGet
>       %_FixedArraySet
>       %_TheHole
>       %_JSCollectionGetTable
>       %_StringGetRawHashField
>
>     With these additions, as well as a few changes to what's exposed as runtime functions,
>     most of the C++ code backing Maps and Sets is gone (including both runtime code in
>     objects.cc and Crankshaft in hydrogen.cc).
>
> BUG=
>
> Committed: https://crrev.com/56600a35a49ffa5abcba66b14839089de3589ad9
> Cr-Commit-Position: refs/heads/master@{#27681}

TBR=weiliang.lin@intel.com,chunyang.dai@intel.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/1073723002

Cr-Commit-Position: refs/heads/master@{#27682}

9 years agoX87: Reimplement Maps and Sets in JS
chunyang.dai [Thu, 9 Apr 2015 02:24:04 +0000 (19:24 -0700)]
X87: Reimplement Maps and Sets in JS

port 909500aa1db9789b68e101045a6359a7fcb30e83 (r27605)

original commit message:
    Previously, the only optimized code path for Maps and Sets was for String keys.
    This was achieved through an implementation of various complex operations
    in Hydrogen. This approach was neither scalable nor forward-compatible.

    This patch adds the necessary intrinsics to implement Maps and Sets almost entirely
    in JS. The added intrinsics are:

      %_FixedArrayGet
      %_FixedArraySet
      %_TheHole
      %_JSCollectionGetTable
      %_StringGetRawHashField

    With these additions, as well as a few changes to what's exposed as runtime functions,
    most of the C++ code backing Maps and Sets is gone (including both runtime code in
    objects.cc and Crankshaft in hydrogen.cc).

BUG=

Review URL: https://codereview.chromium.org/1066373002

Cr-Commit-Position: refs/heads/master@{#27681}

9 years agoES6: Number and Boolean prototype should be ordinary objects
arv [Wed, 8 Apr 2015 21:18:31 +0000 (14:18 -0700)]
ES6: Number and Boolean prototype should be ordinary objects

BUG=v8:4001
LOG=N
R=adamk@chromium.org, rossberg@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_chromium_rel_ng;tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1051373002

Cr-Commit-Position: refs/heads/master@{#27680}

9 years agoMIPS: [turbofan] Add new Float32Abs and Float64Abs operators.
balazs.kilvady [Wed, 8 Apr 2015 19:30:04 +0000 (12:30 -0700)]
MIPS: [turbofan] Add new Float32Abs and Float64Abs operators.

Port 9af9f1d0262e2dab66c62331a03f6a76504f3866

Original commit message:
These operators compute the absolute floating point value of some
arbitrary input, and are implemented without any branches (i.e. using
vabs on arm, and andps/andpd on x86).

BUG=

Review URL: https://codereview.chromium.org/1073463003

Cr-Commit-Position: refs/heads/master@{#27679}

9 years agoMIPS: Fix in-object memory slack tracking bug.
dusan.milosavljevic [Wed, 8 Apr 2015 19:03:39 +0000 (12:03 -0700)]
MIPS: Fix in-object memory slack tracking bug.

TEST=
BUG=

Review URL: https://codereview.chromium.org/1067393003

Cr-Commit-Position: refs/heads/master@{#27678}

9 years ago[parser] report better errors for multiple ForBindings in ForIn/Of loops
caitpotter88 [Wed, 8 Apr 2015 18:47:36 +0000 (11:47 -0700)]
[parser] report better errors for multiple ForBindings in ForIn/Of loops

Instead of "unexpected token" errors, report something meatier and more actionable.

BUG=
R=marja@chromium.org, arv@chromium.org
LOG=N

Review URL: https://codereview.chromium.org/1062263002

Cr-Commit-Position: refs/heads/master@{#27677}

9 years agoShip ES6 computed property names
arv [Wed, 8 Apr 2015 17:49:11 +0000 (10:49 -0700)]
Ship ES6 computed property names

BUG=v8:3754
LOG=N
R=dslomov@chromium.org, rossberg@chromium.org

Review URL: https://codereview.chromium.org/1054883003

Cr-Commit-Position: refs/heads/master@{#27676}

9 years agoDisable more failing tests after f3338dd3b01c.
mstarzinger [Wed, 8 Apr 2015 17:28:52 +0000 (10:28 -0700)]
Disable more failing tests after f3338dd3b01c.

TBR=machenbach@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1070843002

Cr-Commit-Position: refs/heads/master@{#27675}

9 years agoCpuProfiler: public API for deopt info in cpu profiler.
loislo [Wed, 8 Apr 2015 16:13:24 +0000 (09:13 -0700)]
CpuProfiler: public API for deopt info in cpu profiler.

BUG=chromium:452067
LOG=n

Committed: https://crrev.com/baf927ff5115ec62a6dad684b9232ed9d3960e3a
Cr-Commit-Position: refs/heads/master@{#27626}

Review URL: https://codereview.chromium.org/1045753002

Cr-Commit-Position: refs/heads/master@{#27674}

9 years agoCheck mark bit of the found object in MarkCompactCollector::IsSlotInBlackObject.
ulan [Wed, 8 Apr 2015 15:58:14 +0000 (08:58 -0700)]
Check mark bit of the found object in MarkCompactCollector::IsSlotInBlackObject.

BUG=

Review URL: https://codereview.chromium.org/1069763002

Cr-Commit-Position: refs/heads/master@{#27673}

9 years ago[TurboFan] Fixed handling of CompareIC return type.
svenpanne [Wed, 8 Apr 2015 15:57:06 +0000 (08:57 -0700)]
[TurboFan] Fixed handling of CompareIC return type.

Although all this code might be moved around later, it's a good idea
to unhackify it now. :-)

Review URL: https://codereview.chromium.org/1067193004

Cr-Commit-Position: refs/heads/master@{#27672}

9 years agoTests that carefully checks opt/deopt status shouldn't --always-opt.
mvstanton [Wed, 8 Apr 2015 15:55:58 +0000 (08:55 -0700)]
Tests that carefully checks opt/deopt status shouldn't --always-opt.

If we optimize a function before gathering feedback it may be
peppered with soft deoptimizations. So it can't help but deoptimize.
A judicious reading of the code isn't enough to determine what the
optimization state should be in the face of such chaotic gyrations.

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/1069363003

Cr-Commit-Position: refs/heads/master@{#27671}

9 years agoDisable more failing tests after f3338dd3b01c.
mstarzinger [Wed, 8 Apr 2015 14:16:50 +0000 (07:16 -0700)]
Disable more failing tests after f3338dd3b01c.

TBR=machenbach@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1067413002

Cr-Commit-Position: refs/heads/master@{#27670}

9 years ago[turbofan] Reduce JSLoadProperty and JSStoreProperty of strings to JSLoadNamed and...
titzer [Wed, 8 Apr 2015 13:28:00 +0000 (06:28 -0700)]
[turbofan] Reduce JSLoadProperty and JSStoreProperty of strings to JSLoadNamed and JSStoreNamed.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1038953002

Cr-Commit-Position: refs/heads/master@{#27669}

9 years agoAdd more systematic tests for comparisons.
titzer [Wed, 8 Apr 2015 13:15:34 +0000 (06:15 -0700)]
Add more systematic tests for comparisons.

Because, well, JavaScript is tricky and compilers like folding.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1069673002

Cr-Commit-Position: refs/heads/master@{#27668}

9 years agoFix maybe_string_add for adds that have no type feedback where --always-opt is on.
titzer [Wed, 8 Apr 2015 13:14:26 +0000 (06:14 -0700)]
Fix maybe_string_add for adds that have no type feedback where --always-opt is on.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1071473003

Cr-Commit-Position: refs/heads/master@{#27667}

9 years ago[turbofan] Add JSStackCheck into loop bodies.
mstarzinger [Wed, 8 Apr 2015 12:54:56 +0000 (05:54 -0700)]
[turbofan] Add JSStackCheck into loop bodies.

This allows loopy TurboFan code to be interrupted by placing a stack
check (i.e. JSStackCheck node) into each loop. Note that we currently
limit this to non-asm.js code. Also note that stack checks are actually
placed after loop headers and not at back-branches, which allows us to
reuse existing BailoutIds from Crankshaft.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/1065923002

Cr-Commit-Position: refs/heads/master@{#27666}

9 years agoMIPS64: Unbreak cross build.
dusan.milosavljevic [Wed, 8 Apr 2015 12:53:49 +0000 (05:53 -0700)]
MIPS64: Unbreak cross build.

TEST=
BUG=

Review URL: https://codereview.chromium.org/1066403002

Cr-Commit-Position: refs/heads/master@{#27665}

9 years agoRevert of VectorICs: Turn on vector ICs for LOAD and KEYED_LOAD cases. (patchset...
mvstanton [Wed, 8 Apr 2015 12:52:42 +0000 (05:52 -0700)]
Revert of VectorICs: Turn on vector ICs for LOAD and KEYED_LOAD cases. (patchset #1 id:1 of https://codereview.chromium.org/1070653002/)

Reason for revert:
ARM test failure, need to investigate.

Original issue's description:
> VectorICs: Turn on vector ICs for LOAD and KEYED_LOAD cases.
>
> R=verwaest@chromium.org
> BUG=
>
> Committed: https://crrev.com/2395eda3bb7aca938751938df76e01ac2b89b0a6
> Cr-Commit-Position: refs/heads/master@{#27657}

TBR=verwaest@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=

Review URL: https://codereview.chromium.org/1072463002

Cr-Commit-Position: refs/heads/master@{#27664}

9 years ago[ia32] Fix MacroAssembler::Move for int64 to float64 moves.
mstarzinger [Wed, 8 Apr 2015 12:06:52 +0000 (05:06 -0700)]
[ia32] Fix MacroAssembler::Move for int64 to float64 moves.

R=bmeurer@chromium.org
TEST=test262-es6/language/expressions/addition/S9.3_A4.1_T2

Review URL: https://codereview.chromium.org/1071543003

Cr-Commit-Position: refs/heads/master@{#27663}

9 years ago[turbofan] Add new Float32Abs and Float64Abs operators.
Benedikt Meurer [Wed, 8 Apr 2015 11:54:53 +0000 (13:54 +0200)]
[turbofan] Add new Float32Abs and Float64Abs operators.

These operators compute the absolute floating point value of some
arbitrary input, and are implemented without any branches (i.e. using
vabs on arm, and andps/andpd on x86).

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1066393002

Cr-Commit-Position: refs/heads/master@{#27662}

9 years agoDisable more failing tests after f3338dd3b01c.
mstarzinger [Wed, 8 Apr 2015 11:42:19 +0000 (04:42 -0700)]
Disable more failing tests after f3338dd3b01c.

TBR=machenbach@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1067303003

Cr-Commit-Position: refs/heads/master@{#27661}

9 years ago[es6] Stage unicode escapes in strings, var names etc.
marja [Wed, 8 Apr 2015 11:16:02 +0000 (04:16 -0700)]
[es6] Stage unicode escapes in strings, var names etc.

R=rossberg@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1071533002

Cr-Commit-Position: refs/heads/master@{#27660}

9 years agoCreate result array of %DebugGetLoadedScripts outside the debug context.
yangguo [Wed, 8 Apr 2015 11:14:56 +0000 (04:14 -0700)]
Create result array of %DebugGetLoadedScripts outside the debug context.

R=jarin@chromium.org
BUG=chromium:474297
LOG=N

Review URL: https://codereview.chromium.org/1062143002

Cr-Commit-Position: refs/heads/master@{#27659}

9 years ago[turbofan] Match selects in control reducer (configurable).
titzer [Wed, 8 Apr 2015 11:13:43 +0000 (04:13 -0700)]
[turbofan] Match selects in control reducer (configurable).

R=bmeurer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1061303002

Cr-Commit-Position: refs/heads/master@{#27658}

9 years agoVectorICs: Turn on vector ICs for LOAD and KEYED_LOAD cases.
mvstanton [Wed, 8 Apr 2015 11:12:37 +0000 (04:12 -0700)]
VectorICs: Turn on vector ICs for LOAD and KEYED_LOAD cases.

R=verwaest@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1070653002

Cr-Commit-Position: refs/heads/master@{#27657}

9 years agoBlacklist failing test on arm64 (issue 4016).
yangguo [Wed, 8 Apr 2015 11:11:29 +0000 (04:11 -0700)]
Blacklist failing test on arm64 (issue 4016).

TBR=jkummerow@chromium.org
BUG=v8:4016
LOG=N

Review URL: https://codereview.chromium.org/1071513002

Cr-Commit-Position: refs/heads/master@{#27656}

9 years agoDisable more failing tests after f3338dd3b01c.
mstarzinger [Wed, 8 Apr 2015 10:35:09 +0000 (03:35 -0700)]
Disable more failing tests after f3338dd3b01c.

TBR=yangguo@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1072433002

Cr-Commit-Position: refs/heads/master@{#27655}

9 years agoRevert of Remove support for thread-based recompilation (patchset #1 id:1 of https...
yangguo [Wed, 8 Apr 2015 10:29:23 +0000 (03:29 -0700)]
Revert of Remove support for thread-based recompilation (patchset #1 id:1 of https://codereview.chromium.org/966653002/)

Reason for revert:
speculative revert due to gc-stress timeouts.

Original issue's description:
> Remove support for thread-based recompilation
>
> BUG=v8:3608
> R=yangguo@chromium.org
> LOG=y
>
> Committed: https://crrev.com/ed5db223a19dfe126af012e894582251aa3635d7
> Cr-Commit-Position: refs/heads/master@{#27619}

TBR=jochen@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
BUG=v8:3608
LOG=N

Review URL: https://codereview.chromium.org/1063383004

Cr-Commit-Position: refs/heads/master@{#27654}

9 years agoMake test runner more chatty to avoid it getting killed by buildbot.
machenbach [Wed, 8 Apr 2015 09:53:35 +0000 (02:53 -0700)]
Make test runner more chatty to avoid it getting killed by buildbot.

NOTRY=true
NOTREECHECKS=true

Review URL: https://codereview.chromium.org/1064043002

Cr-Commit-Position: refs/heads/master@{#27653}

9 years agoDisable more failing tests after f3338dd3b01c.
mstarzinger [Wed, 8 Apr 2015 09:51:14 +0000 (02:51 -0700)]
Disable more failing tests after f3338dd3b01c.

TBR=yangguo@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1067353002

Cr-Commit-Position: refs/heads/master@{#27652}

9 years agoDisable failing test after f3338dd3b01c.
mstarzinger [Wed, 8 Apr 2015 09:24:55 +0000 (02:24 -0700)]
Disable failing test after f3338dd3b01c.

R=yangguo@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1070663002

Cr-Commit-Position: refs/heads/master@{#27651}

9 years agoMake GetDebugContext a bit more robust.
svenpanne [Wed, 8 Apr 2015 09:10:10 +0000 (02:10 -0700)]
Make GetDebugContext a bit more robust.

Perhaps we should throw an exception and/or change our external API to
use a MaybeLocal, but that would be a bigger change. For now, we just
return undefined when something goes wrong with the DebugContext,
which is good enough to avoid crashing.

BUG=chromium:474538
LOG=y

Review URL: https://codereview.chromium.org/1065213002

Cr-Commit-Position: refs/heads/master@{#27650}

9 years agoPrevent overzealous bailout due to script context.
mstarzinger [Wed, 8 Apr 2015 09:00:41 +0000 (02:00 -0700)]
Prevent overzealous bailout due to script context.

This is a follow-up to 2d281e71ac49 and prevents bailouts on empty
script contexts in Crankshaft, which don't need allocation. Only
non-empty script contexts should cause a bailout.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1063373003

Cr-Commit-Position: refs/heads/master@{#27649}

9 years ago[x64] Introduce BMI instructions.
jing.bao [Wed, 8 Apr 2015 07:15:51 +0000 (00:15 -0700)]
[x64] Introduce BMI instructions.

BUG=v8:4015
LOG=n

Review URL: https://codereview.chromium.org/1040603002

Cr-Commit-Position: refs/heads/master@{#27648}

9 years agoUnbreak check-name-clashes.py after recent macro renamings.
svenpanne [Wed, 8 Apr 2015 07:04:07 +0000 (00:04 -0700)]
Unbreak check-name-clashes.py after recent macro renamings.

This should have been part of https://codereview.chromium.org/1056863002/. :-/

TBR=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/1070593002

Cr-Commit-Position: refs/heads/master@{#27647}

9 years ago[arm] Use position independent table switches.
bmeurer [Wed, 8 Apr 2015 06:50:10 +0000 (23:50 -0700)]
[arm] Use position independent table switches.

Performance measurements show that the position independent code is
usually slightly faster than the position dependent code, and there
seems to be no noticable regression.

This also gets rid of a lot of support code that was only required to allow
embedding labels into the code stream. And it implies that neither the GC
nor the deserializer need to do anything for jump tables.

R=svenpanne@chromium.org
BUG=v8:3872
LOG=n

Review URL: https://codereview.chromium.org/1069633002

Cr-Commit-Position: refs/heads/master@{#27646}

9 years agoUpdate V8 DEPS.
v8-autoroll [Wed, 8 Apr 2015 03:39:22 +0000 (20:39 -0700)]
Update V8 DEPS.

Rolling v8/tools/clang to 7ec0c52c7a66cd94c96b8eb37bacd26d152e40e6

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1070533002

Cr-Commit-Position: refs/heads/master@{#27645}

9 years agoMIPS64: Fix typo of 'Always update raw pointers when handling interrupts inside RegEx...
balazs.kilvady [Tue, 7 Apr 2015 21:12:13 +0000 (14:12 -0700)]
MIPS64: Fix typo of 'Always update raw pointers when handling interrupts inside RegExp code.'

Port c67cb287a901ddf03d4ae4dafcf431d09fd3e22c

BUG=

Review URL: https://codereview.chromium.org/1061363003

Cr-Commit-Position: refs/heads/master@{#27644}

9 years agoMIPS64: Make --always-opt also optimize top-level code.
balazs.kilvady [Tue, 7 Apr 2015 21:11:05 +0000 (14:11 -0700)]
MIPS64: Make --always-opt also optimize top-level code.

Port 2d281e71ac49058ef5696f4255ff1fa30796bc5e

Original commit message:
This enables eager optimization of top-level code with TurboFan and
extends test coverage by triggering it with the --always-opt flag.
Script contexts are now also properly allocated in TurboFan.

BUG=

Review URL: https://codereview.chromium.org/1066843005

Cr-Commit-Position: refs/heads/master@{#27643}

9 years agoDisable another debug test under turbo always-opt
adamk [Tue, 7 Apr 2015 20:09:57 +0000 (13:09 -0700)]
Disable another debug test under turbo always-opt

TBR=mstarzinger@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1064133002

Cr-Commit-Position: refs/heads/master@{#27642}

9 years agoUse NumberIsNaN in collections.js and make it inlined
adamk [Tue, 7 Apr 2015 19:52:17 +0000 (12:52 -0700)]
Use NumberIsNaN in collections.js and make it inlined

Review URL: https://codereview.chromium.org/1067933002

Cr-Commit-Position: refs/heads/master@{#27641}

9 years ago[release-tools] Only read from the chromium checkout in v8rel.
machenbach [Tue, 7 Apr 2015 19:42:32 +0000 (12:42 -0700)]
[release-tools] Only read from the chromium checkout in v8rel.

Don't create local branches or otherwise manipulate the
checkout. This reads refs from remote branches and
reads file contents using show. It is faster and requires
less bootstrapping and cleanup.

TBR=tandrii@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1060013003

Cr-Commit-Position: refs/heads/master@{#27640}

9 years ago[es6] emit error when for-in loop declarations are initialized in strict mode
caitpotter88 [Tue, 7 Apr 2015 19:28:33 +0000 (12:28 -0700)]
[es6] emit error when for-in loop declarations are initialized in strict mode

The ES6 grammar forbids the initialization of variable declarations in IterationStatements.

This CL will report `for (var x = y in z)` as a SyntaxError in strict mode (as done in JSC). It is possible that this could break sites in sloppy mode, and so that change can wait.

BUG=
R=
LOG=N

Review URL: https://codereview.chromium.org/1033823002

Cr-Commit-Position: refs/heads/master@{#27639}

9 years agoSimplify collections.js now that it's wrapped in an IIFE
adamk [Tue, 7 Apr 2015 19:00:34 +0000 (12:00 -0700)]
Simplify collections.js now that it's wrapped in an IIFE

Also wrap templates.js in an IIFE to avoid unnecessary pollution
of the builtins object.

Review URL: https://codereview.chromium.org/1067903004

Cr-Commit-Position: refs/heads/master@{#27638}

9 years agoMore robust when allocation fails during compaction
erikcorry [Tue, 7 Apr 2015 17:22:10 +0000 (10:22 -0700)]
More robust when allocation fails during compaction

R=ulan@chromium.org
BUG=chromium:473307
LOG=yes

Review URL: https://codereview.chromium.org/1068833002

Cr-Commit-Position: refs/heads/master@{#27637}

9 years agoDisabled failing tests after 2d281e71ac49.
mstarzinger [Tue, 7 Apr 2015 17:06:12 +0000 (10:06 -0700)]
Disabled failing tests after 2d281e71ac49.

R=titzer@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1065983002

Cr-Commit-Position: refs/heads/master@{#27636}

9 years agoSimplify instanceof optimization; don't duplicate lookup
verwaest [Tue, 7 Apr 2015 16:11:20 +0000 (09:11 -0700)]
Simplify instanceof optimization; don't duplicate lookup

BUG=

Review URL: https://codereview.chromium.org/1066873002

Cr-Commit-Position: refs/heads/master@{#27635}

9 years agoMIPS: JSEntryTrampoline: check for stack space before pushing arguments
balazs.kilvady [Tue, 7 Apr 2015 15:59:00 +0000 (08:59 -0700)]
MIPS: JSEntryTrampoline: check for stack space before pushing arguments

Port 146598f44a1efd0f942745abfd6c08a073972c01

Original commit message:
Optimistically pushing a lot of arguments can run into the stack limit of the process, at least on operating systems where this limit is close to the limit that V8 sets for itself.

BUG=chromium:469768
LOG=y

Review URL: https://codereview.chromium.org/1066843003

Cr-Commit-Position: refs/heads/master@{#27634}

9 years agoMake --always-opt also optimize top-level code.
mstarzinger [Tue, 7 Apr 2015 15:44:16 +0000 (08:44 -0700)]
Make --always-opt also optimize top-level code.

This enables eager optimization of top-level code with TurboFan and
extends test coverage by triggering it with the --always-opt flag.
Script contexts are now also properly allocated in TurboFan.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/1053063003

Cr-Commit-Position: refs/heads/master@{#27633}

9 years agoMIPS: Fix compilation error.
mvstanton [Tue, 7 Apr 2015 15:16:48 +0000 (08:16 -0700)]
MIPS: Fix compilation error.

BUG=
TBR=verwaest@chromium.org

Review URL: https://codereview.chromium.org/1068723006

Cr-Commit-Position: refs/heads/master@{#27632}

9 years agoVector-ICs - speed towards the monomorphic exit as quickly as possible.
mvstanton [Tue, 7 Apr 2015 14:49:59 +0000 (07:49 -0700)]
Vector-ICs - speed towards the monomorphic exit as quickly as possible.

Thanks to some careful assumptions, we can examine the object found at
vector[slot] and trust it's a heap object where the second field is
either a map if it's a WeakCell, or definitely not a map if it's a
Symbol, String or FixedArray. Use this to save a memory read.

BUG=

Review URL: https://codereview.chromium.org/1032163002

Cr-Commit-Position: refs/heads/master@{#27631}

9 years agoCode cleanup in GenerateRecordCallTarget.
mvstanton [Tue, 7 Apr 2015 13:39:32 +0000 (06:39 -0700)]
Code cleanup in GenerateRecordCallTarget.

BUG=

Review URL: https://codereview.chromium.org/1057613003

Cr-Commit-Position: refs/heads/master@{#27630}

9 years agoAdd yangguo@chromium.org to src/snapshot/OWNERS and WATCHLISTS
yangguo [Tue, 7 Apr 2015 13:38:22 +0000 (06:38 -0700)]
Add yangguo@chromium.org to src/snapshot/OWNERS and WATCHLISTS

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/1060683003

Cr-Commit-Position: refs/heads/master@{#27629}

9 years agoRevert of CpuProfiler: public API for deopt info in cpu profiler. (patchset #6 id...
machenbach [Tue, 7 Apr 2015 13:30:49 +0000 (06:30 -0700)]
Revert of CpuProfiler: public API for deopt info in cpu profiler. (patchset #6 id:150001 of https://codereview.chromium.org/1045753002/)

Reason for revert:
[Sheriff] Breaks compile here:
http://build.chromium.org/p/client.v8/builders/V8%20Win32%20-%20nosnap%20-%20shared/builds/6115

Original issue's description:
> CpuProfiler: public API for deopt info in cpu profiler.
>
> BUG=chromium:452067
> LOG=n
>
> Committed: https://crrev.com/baf927ff5115ec62a6dad684b9232ed9d3960e3a
> Cr-Commit-Position: refs/heads/master@{#27626}

TBR=alph@chromium.org,jkummerow@chromium.org,svenpanne@chromium.org,yurys@chromium.org,loislo@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:452067

Review URL: https://codereview.chromium.org/1062053004

Cr-Commit-Position: refs/heads/master@{#27628}

9 years agoFix missing SmiTag in failure path of r27614
jkummerow [Tue, 7 Apr 2015 12:47:39 +0000 (05:47 -0700)]
Fix missing SmiTag in failure path of r27614

BUG=chromium:469768
LOG=n

Review URL: https://codereview.chromium.org/1061263003

Cr-Commit-Position: refs/heads/master@{#27627}

9 years agoCpuProfiler: public API for deopt info in cpu profiler.
loislo [Tue, 7 Apr 2015 12:40:41 +0000 (05:40 -0700)]
CpuProfiler: public API for deopt info in cpu profiler.

BUG=chromium:452067
LOG=n

Review URL: https://codereview.chromium.org/1045753002

Cr-Commit-Position: refs/heads/master@{#27626}

9 years agoDebugger: remove debug command API.
yangguo [Tue, 7 Apr 2015 12:21:34 +0000 (05:21 -0700)]
Debugger: remove debug command API.

Instead, please use v8::Isolate::RequestInterrupt to synchronize
to the main thread.

R=yurys@chromium.org
API=Removed v8::Debug::DebugBreakForCommand
LOG=Y

Review URL: https://codereview.chromium.org/1036863002

Cr-Commit-Position: refs/heads/master@{#27625}

9 years agoARM64: Support sign extend for add and subtract
martyn.capewell [Tue, 7 Apr 2015 11:34:33 +0000 (04:34 -0700)]
ARM64: Support sign extend for add and subtract

Support sxtb and sxth extend operators on add and subtract, as we've
done for ubtx/h. This is similar to ARM support for sxtab/h.

BUG=

Review URL: https://codereview.chromium.org/1064813003

Cr-Commit-Position: refs/heads/master@{#27624}

9 years agoReland "Merge old data and pointer space."
hpayer [Tue, 7 Apr 2015 11:31:57 +0000 (04:31 -0700)]
Reland "Merge old data and pointer space."

This reverts commit cbfcee557503514bef20abdfc78ceb262c9c4d64.

BUG=

Review URL: https://codereview.chromium.org/1051233002

Cr-Commit-Position: refs/heads/master@{#27623}

9 years ago[release-tools] Fix v8rel in branch period.
machenbach [Tue, 7 Apr 2015 11:28:18 +0000 (04:28 -0700)]
[release-tools] Fix v8rel in branch period.

In case the v8 revision in chromium was set to an older
revision that's not in the list of recent v8 releases, the
script will bail out too early. This changes iterates over
10 more revisions, which should cover all exceeding revisions
in branch period.

NOTRY=true
TBR=hablich@chromium.org

Review URL: https://codereview.chromium.org/1067793002

Cr-Commit-Position: refs/heads/master@{#27622}

9 years ago[heap] Assert that code objects are always properly aligned.
bmeurer [Tue, 7 Apr 2015 11:24:15 +0000 (04:24 -0700)]
[heap] Assert that code objects are always properly aligned.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/1059903004

Cr-Commit-Position: refs/heads/master@{#27621}

9 years ago[turbofan] Rework handling of loop exits in loop peeling.
titzer [Tue, 7 Apr 2015 11:00:40 +0000 (04:00 -0700)]
[turbofan] Rework handling of loop exits in loop peeling.

This CL primarily makes the loop peeling algorithm more robust; it no longer damages the graph if the loops are improperly closed.

R=bmeurer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1052753004

Cr-Commit-Position: refs/heads/master@{#27620}

9 years agoRemove support for thread-based recompilation
Jochen Eisinger [Tue, 7 Apr 2015 10:59:43 +0000 (12:59 +0200)]
Remove support for thread-based recompilation

BUG=v8:3608
R=yangguo@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/966653002

Cr-Commit-Position: refs/heads/master@{#27619}

9 years ago[ia32] Match -0 - x with sign bit flip.
bmeurer [Tue, 7 Apr 2015 10:43:56 +0000 (03:43 -0700)]
[ia32] Match -0 - x with sign bit flip.

We can use xorps/xorpd on Intel CPUs to flip the sign bit. Ideally we'd
use an absolute 128-bit constant in the code object, as OCaml/GCC
does, however that requires 128-bit alignment for code objects,
which is not yet implemented. So for now we materialize the mask
inline.

As drive-by-fix, don't hardcode xmm0 as scratch double register.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1064833002

Cr-Commit-Position: refs/heads/master@{#27618}

9 years agoMove prototype metadata from internal properties to prototype maps
jkummerow [Tue, 7 Apr 2015 10:42:45 +0000 (03:42 -0700)]
Move prototype metadata from internal properties to prototype maps

The motivation is that we prefer to avoid creating internal properties, and we have a usable field on maps ("transitions", which is not used for prototype maps).
This CL also ensures the invariant that prototype maps are never shared, even if they are in dictionary mode.

Review URL: https://codereview.chromium.org/1033653002

Cr-Commit-Position: refs/heads/master@{#27617}

9 years agoRevert of Turn off overapproximation of the weak closure again (patchset #1 id:1...
jochen [Tue, 7 Apr 2015 09:45:58 +0000 (02:45 -0700)]
Revert of Turn off overapproximation of the weak closure again (patchset #1 id:1 of https://codereview.chromium.org/1050443002/)

Reason for revert:
this indeed drops the max major gc time considerable, so turn it back on

Original issue's description:
> Turn off overapproximation of the weak closure again
>
> As long as we still have to process global handles, the impact is not
> yet worthwhile
>
> BUG=v8:3862
> R=hpayer@chromium.org
> LOG=y
>
> Committed: https://crrev.com/294cdc6aecbd7f76be68217da4b3d35901ebce4b
> Cr-Commit-Position: refs/heads/master@{#27570}

TBR=hpayer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:3862

Review URL: https://codereview.chromium.org/1068723003

Cr-Commit-Position: refs/heads/master@{#27616}

9 years agoAlways update raw pointers when handling interrupts inside RegExp code.
yangguo [Tue, 7 Apr 2015 09:44:47 +0000 (02:44 -0700)]
Always update raw pointers when handling interrupts inside RegExp code.

R=mstarzinger@chromium.org
BUG=chromium:469480
LOG=N

Review URL: https://codereview.chromium.org/1034173002

Cr-Commit-Position: refs/heads/master@{#27615}

9 years agoJSEntryTrampoline: check for stack space before pushing arguments
jkummerow [Tue, 7 Apr 2015 09:13:36 +0000 (02:13 -0700)]
JSEntryTrampoline: check for stack space before pushing arguments

Optimistically pushing a lot of arguments can run into the stack limit of the process, at least on operating systems where this limit is close to the limit that V8 sets for itself.

BUG=chromium:469768
LOG=y

Review URL: https://codereview.chromium.org/1056913003

Cr-Commit-Position: refs/heads/master@{#27614}

9 years ago[turbofan] Introduce BranchMatcher and DiamondMatcher helpers.
titzer [Tue, 7 Apr 2015 09:03:28 +0000 (02:03 -0700)]
[turbofan] Introduce BranchMatcher and DiamondMatcher helpers.

R=bmeurer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/1054963002

Cr-Commit-Position: refs/heads/master@{#27613}

9 years ago[turbofan][arm64] Match add with shifted operand for mult by a power of 2 plus 1.
baptiste.afsa [Tue, 7 Apr 2015 08:48:14 +0000 (01:48 -0700)]
[turbofan][arm64] Match add with shifted operand for mult by a power of 2 plus 1.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1056543002

Cr-Commit-Position: refs/heads/master@{#27612}

9 years ago[x64] Match -0 - x with sign bit flip.
bmeurer [Tue, 7 Apr 2015 07:34:55 +0000 (00:34 -0700)]
[x64] Match -0 - x with sign bit flip.

We can use xorps/xorpd on Intel CPUs to flip the sign bit. Ideally we'd
use a RIP-relative 128-bit constant in the code object, as OCaml/GCC
does, however that requires 128-bit alignment for code objects, which is
not yet implemented. So for now we materialize the mask inline.

R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/1046893002

Cr-Commit-Position: refs/heads/master@{#27611}

9 years ago[win32] Improve random mmap address generation on 64-bit.
bmeurer [Tue, 7 Apr 2015 07:13:50 +0000 (00:13 -0700)]
[win32] Improve random mmap address generation on 64-bit.

Currently we only generate addresses in the range 2GiB to 4GiB, because
of a bug, when running in 64-bit mode. With this fix we now use the full
range 2GiB to 4096GiB.

BUG=v8:3997
LOG=n
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/1051303003

Cr-Commit-Position: refs/heads/master@{#27610}

9 years agoRearranged intrinsic declarations according to their implementations.
svenpanne [Tue, 7 Apr 2015 07:12:42 +0000 (00:12 -0700)]
Rearranged intrinsic declarations according to their implementations.

Now every FOR_EACH_INTRINSIC_FOO sub-macro corresponds to a
src/runtime/runtime-FOO.cc file, even the order of runtime functions
within that file has been preserved in the corresponding macro, thanks
to some eye-hurting for/clang/sed/grep madness. ;-)

BUG=v8:3947
LOG=n

Review URL: https://codereview.chromium.org/1056863002

Cr-Commit-Position: refs/heads/master@{#27609}

9 years agoPPC: exclude test for AIX due to issue 2857
michael_dawson [Tue, 7 Apr 2015 07:06:39 +0000 (00:06 -0700)]
PPC: exclude test for AIX due to issue 2857

cctest/test-log/EquivalenceOfLoggingAndTraversal fails on
AIX because it does not correctly handle the full 64 bit
address space supported by AIX.  SKIP until the issue
is resolved.

modified:   test/cctest/cctest.status

R=danno@chromium.org, svenpanne@chromium.org

BUG=

Review URL: https://codereview.chromium.org/1055183002

Cr-Commit-Position: refs/heads/master@{#27608}

9 years agoUpdate V8 DEPS.
v8-autoroll [Tue, 7 Apr 2015 03:52:31 +0000 (20:52 -0700)]
Update V8 DEPS.

Rolling v8/third_party/icu to 10834e84d250eb31a73496bfdb7923eaa43ead51

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/1055173004

Cr-Commit-Position: refs/heads/master@{#27607}

9 years agoFix formatting in full-codegen-x64.cc
adamk [Tue, 7 Apr 2015 00:14:28 +0000 (17:14 -0700)]
Fix formatting in full-codegen-x64.cc

TBR=arv@chromium.org

Review URL: https://codereview.chromium.org/1062853003

Cr-Commit-Position: refs/heads/master@{#27606}