bmeurer [Tue, 7 Jul 2015 07:28:02 +0000 (00:28 -0700)]
[deoptimizer] Properly evict TurboFan OSR code objects on eager deopts.
TurboFan OSR installs the CompileOptimized builtin on JSFunctions, which
means that we never evict the OSR code objects for such functions from
eager deopts.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1220813018
Cr-Commit-Position: refs/heads/master@{#29507}
machenbach [Tue, 7 Jul 2015 06:41:18 +0000 (23:41 -0700)]
Revert of Fix bug when transferring SharedArrayBuffer to multiple Workers. (patchset #3 id:40001 of https://codereview.chromium.org/
1215233004/)
Reason for revert:
[Sheriff] Test hangs sometimes and times out flakily. E.g.: http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20nosse3/builds/4551/steps/Check%20%28flakes%29/logs/d8-worker-sharedarray..
Original issue's description:
> Fix bug when transferring SharedArrayBuffer to multiple Workers.
>
> Previously, the serialization code would call Externalize for every transferred
> ArrayBuffer or SharedArrayBuffer, but that function can only be called once. If
> the buffer is already externalized, we should call GetContents instead.
>
> Also fix use-after-free bug when transferring ArrayBuffers. The transferred
> ArrayBuffer must be internalized in the new isolate, or be managed by the
> Shell. The current code gives it to the isolate externalized and frees it
> immediately afterward when the SerializationData object is destroyed.
>
> BUG=chromium:497295
> R=jarin@chromium.org
> LOG=n
>
> Committed: https://crrev.com/
dd7962bf7838f8379ba776ee6b7b0e4d3bec2140
> Cr-Commit-Position: refs/heads/master@{#29499}
TBR=jarin@chromium.org,jochen@chromium.org,binji@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:497295
Review URL: https://codereview.chromium.org/
1224843008
Cr-Commit-Position: refs/heads/master@{#29506}
v8-autoroll [Tue, 7 Jul 2015 03:42:17 +0000 (20:42 -0700)]
Update V8 DEPS.
Rolling v8/third_party/icu to
c81a1a3989c3b66fa323e9a6ee7418d7c08297af
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1213043007
Cr-Commit-Position: refs/heads/master@{#29505}
cdai2 [Tue, 7 Jul 2015 01:05:28 +0000 (09:05 +0800)]
X87: [turbofan] Enable tail calls for %_CallRuntime.
port
1fa4285e1ce16cfadf8c40a0993491ec4e2bbbe0 (r29436).
original commit message:
This involves:
- Enabling the tail call optimization reducer in all cases.
- Adding an addition flag to CallFunctionParameters to mark call sites
that can be tail-called enabled.
- Only set the tail-call flag for %_CallFunction.
BUG=
R=weiliang.lin@intel.com
Review URL: https://codereview.chromium.org/
1228463003
Cr-Commit-Position: refs/heads/master@{#29504}
mbrandy [Mon, 6 Jul 2015 19:26:45 +0000 (12:26 -0700)]
PPC: Debugger: use debug break slots instead of ICs (except for calls).
Port
a8a4c364c2ad4c1e5f4c6e1f1580d75ff1a35a4f
R=yangguo@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1216863005
Cr-Commit-Position: refs/heads/master@{#29503}
mbrandy [Mon, 6 Jul 2015 19:22:23 +0000 (12:22 -0700)]
PPC: Fix "Support for global var shortcuts in script contexts."
R=ishell@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1212343007
Cr-Commit-Position: refs/heads/master@{#29502}
mbrandy [Mon, 6 Jul 2015 19:21:16 +0000 (12:21 -0700)]
PPC: Fix "[turbofan] Add Uint64LessThanOrEqual to 64-bit TurboFan backends."
R=titzer@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1218073008
Cr-Commit-Position: refs/heads/master@{#29501}
mstarzinger [Mon, 6 Jul 2015 18:02:41 +0000 (11:02 -0700)]
[turbofan] Port initialization of new.target variable.
This implements the proper initialization of the new.target internal
variable in the AstGraphBuilder. For now this uses a runtime call that
cannot handle inlined frames correctly.
R=arv@chromium.org
Review URL: https://codereview.chromium.org/
1212813008
Cr-Commit-Position: refs/heads/master@{#29500}
binji [Mon, 6 Jul 2015 17:17:59 +0000 (10:17 -0700)]
Fix bug when transferring SharedArrayBuffer to multiple Workers.
Previously, the serialization code would call Externalize for every transferred
ArrayBuffer or SharedArrayBuffer, but that function can only be called once. If
the buffer is already externalized, we should call GetContents instead.
Also fix use-after-free bug when transferring ArrayBuffers. The transferred
ArrayBuffer must be internalized in the new isolate, or be managed by the
Shell. The current code gives it to the isolate externalized and frees it
immediately afterward when the SerializationData object is destroyed.
BUG=chromium:497295
R=jarin@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1215233004
Cr-Commit-Position: refs/heads/master@{#29499}
ishell [Mon, 6 Jul 2015 16:36:28 +0000 (09:36 -0700)]
Support for global var shortcuts in script contexts.
Review URL: https://codereview.chromium.org/
1218783005
Cr-Commit-Position: refs/heads/master@{#29498}
arv [Mon, 6 Jul 2015 16:20:35 +0000 (09:20 -0700)]
Revert of Revert of [es6] Bound function names (patchset #1 id:1 of https://codereview.chromium.org/
1225793002/)
Reason for revert:
This will prevent rolls. Fixing the root issue instead.
Original issue's description:
> Revert of [es6] Bound function names (patchset #1 id:1 of https://codereview.chromium.org/
1195983002/)
>
> Reason for revert:
> Incorrect behavior
>
> Original issue's description:
> > [es6] Bound function names
> >
> > https://people.mozilla.org/~jorendorff/es6-draft.html#sec-function.prototype.bind
> >
> > Bound functions should have a name based on the function that was
> > bound.
> >
> > This reverts the revert
f2747ed9b48d0e62c7a30da69825ff926aeedbd2. The original
> > CL was reverted because the Blink layout test broke. I have a CL that disables
> > these tests at: https://codereview.chromium.org/
1196753003/
> >
> > BUG=N
> > LOG=N
> > R=adamk
> > CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_chromium_rel_ng;tryserver.blink:linux_blink_rel
> >
> > Committed: https://crrev.com/
b6d950c979f4348138de0ec54e40dcc48d833926
> > Cr-Commit-Position: refs/heads/master@{#29193}
>
> TBR=adamk@chromium.org,verwaest@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=N
>
> Committed: https://crrev.com/
744e4d4fd9316674682bc6ca30ded5866494cc1c
> Cr-Commit-Position: refs/heads/master@{#29495}
TBR=adamk@chromium.org,verwaest@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=N
Review URL: https://codereview.chromium.org/
1222363002
Cr-Commit-Position: refs/heads/master@{#29497}
jkummerow [Mon, 6 Jul 2015 15:29:23 +0000 (08:29 -0700)]
Fix performance regression introduced in r28961
where bound functions' length was made configurable. The bootstrapper must be kept in sync to avoid polymorphism.
BUG=chromium:500686
LOG=n
R=verwaest@chromium.org
Review URL: https://codereview.chromium.org/
1221383003
Cr-Commit-Position: refs/heads/master@{#29496}
arv [Mon, 6 Jul 2015 15:25:30 +0000 (08:25 -0700)]
Revert of [es6] Bound function names (patchset #1 id:1 of https://codereview.chromium.org/
1195983002/)
Reason for revert:
Incorrect behavior
Original issue's description:
> [es6] Bound function names
>
> https://people.mozilla.org/~jorendorff/es6-draft.html#sec-function.prototype.bind
>
> Bound functions should have a name based on the function that was
> bound.
>
> This reverts the revert
f2747ed9b48d0e62c7a30da69825ff926aeedbd2. The original
> CL was reverted because the Blink layout test broke. I have a CL that disables
> these tests at: https://codereview.chromium.org/
1196753003/
>
> BUG=N
> LOG=N
> R=adamk
> CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_chromium_rel_ng;tryserver.blink:linux_blink_rel
>
> Committed: https://crrev.com/
b6d950c979f4348138de0ec54e40dcc48d833926
> Cr-Commit-Position: refs/heads/master@{#29193}
TBR=adamk@chromium.org,verwaest@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=N
Review URL: https://codereview.chromium.org/
1225793002
Cr-Commit-Position: refs/heads/master@{#29495}
machenbach [Mon, 6 Jul 2015 13:59:30 +0000 (06:59 -0700)]
Revert of Reland: Fix logic for incremental marking steps on tenured allocation (patchset #4 id:60001 of https://codereview.chromium.org/
1077153004/)
Reason for revert:
[Sheriff] Speculative revert, see:
https://code.google.com/p/chromium/issues/detail?id=506875
Original issue's description:
> Reland: Fix logic for incremental marking steps on tenured allocation
>
> BUG=
>
> Committed: https://crrev.com/
5000650bde2ec0bc90d959b529c97aea20385043
> Cr-Commit-Position: refs/heads/master@{#29442}
TBR=hpayer@chromium.org,erikcorry@chromium.org
BUG=chromium:506875
LOG=n
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1212063005
Cr-Commit-Position: refs/heads/master@{#29494}
bmeurer [Mon, 6 Jul 2015 12:56:05 +0000 (05:56 -0700)]
[turbofan] Context specialization is the job of the JSContextSpecialization.
Remove the context specialization hack from the AstGraphBuilder, and
properly specialize to the function context in the context specialization.
And replace the correct context in the JSInliner.
R=mstarzinger@chromium.org
BUG=v8:4273
LOG=n
Review URL: https://codereview.chromium.org/
1218873005
Cr-Commit-Position: refs/heads/master@{#29493}
machenbach [Mon, 6 Jul 2015 12:37:31 +0000 (05:37 -0700)]
[test] Push binaries to separate folders on Android devices.
BUG=chromium:507213
LOG=n
NOTRY=true
TBR=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
1213613005
Cr-Commit-Position: refs/heads/master@{#29492}
machenbach [Mon, 6 Jul 2015 12:17:15 +0000 (05:17 -0700)]
[test] Add android perf trybots.
BUG=chromium:502176
LOG=n
NOTRY=true
Review URL: https://codereview.chromium.org/
1217503008
Cr-Commit-Position: refs/heads/master@{#29491}
bmeurer [Mon, 6 Jul 2015 12:11:50 +0000 (05:11 -0700)]
[turbofan] Fix value output count for the Start node.
The value output count for Start is currently off by 1 for code stubs,
because the CommonOperatorBuilder hardcodes the receiver parameter.
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/
1217553005
Cr-Commit-Position: refs/heads/master@{#29490}
machenbach [Mon, 6 Jul 2015 12:09:36 +0000 (05:09 -0700)]
[test] Port clobber of old test262 archive files.
BUG=v8:4254
LOG=n
NOTRY=true
TBR=jkummerow@chromium.org
NOTREECHECKS=true
Review URL: https://codereview.chromium.org/
1226803002
Cr-Commit-Position: refs/heads/master@{#29489}
mstarzinger [Mon, 6 Jul 2015 11:50:00 +0000 (04:50 -0700)]
[turbofan] Cleanup Parameter creation in AstGraphBuilder.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
1222833002
Cr-Commit-Position: refs/heads/master@{#29488}
yangguo [Mon, 6 Jul 2015 11:15:52 +0000 (04:15 -0700)]
Debugger: use debug break slots instead of ICs (except for calls).
BUG=v8:3147,v8:4269
LOG=N
Review URL: https://codereview.chromium.org/
1218493005
Cr-Commit-Position: refs/heads/master@{#29487}
bmeurer [Mon, 6 Jul 2015 11:11:15 +0000 (04:11 -0700)]
[turbofan] Reland "Add new JSFrameSpecialization reducer." and "Perform OSR deconstruction early and remove type propagation.".
We have to reland these two commits at once, because the first breaks
some asm.js benchmarks without the second. The change was reverted
because of bogus checks in the verifier, which will not work in the
presence of OSR (and where hidden because of the type back propagation
hack in OSR so far). Original messages are below:
[turbofan] Add new JSFrameSpecialization reducer.
The JSFrameSpecialization specializes an OSR graph to the current
unoptimized frame on which we will perform the on-stack replacement.
This is used for asm.js functions, where we cannot reuse the OSR
code object anyway because of context specialization, and so we could as
well specialize to the max instead.
It works by replacing all OsrValues in the graph with their values
in the JavaScriptFrame.
The idea is that using this trick we get better performance without
doing the unsound backpropagation of types to OsrValues later. This
is the first step towards fixing OSR for TurboFan.
[turbofan] Perform OSR deconstruction early and remove type propagation.
This way we don't have to deal with dead pre-OSR code in the graph
and risk optimizing the wrong code, especially we don't make
optimistic assumptions in the dead code that leaks into the OSR code
(i.e. deopt guards are in dead code, but the types propagate to OSR
code via the OsrValue type back propagation).
BUG=v8:4273
LOG=n
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1226673005
Cr-Commit-Position: refs/heads/master@{#29486}
fedor [Mon, 6 Jul 2015 11:00:05 +0000 (04:00 -0700)]
unicode-decoder: fix out-of-band write in utf16
`WriteUtf16Slow` should not assume that the output buffer has enough
bytes to hold both words of surrogate pair. It should pass the number of
remaining bytes to the `Utf8::ValueOf` instead, just as we already do in
`Utf8DecoderBase::Reset`. Otherwise it will attempt to write the trail
uint16_t past the buffer boundary, leading to memory corruption and
possible crash.
Originally reported by: Kris Reeves <kris.re@bbhmedia.com>
BUG=v8:4274
R=danno
R=svenpanne
LOG=y
Review URL: https://codereview.chromium.org/
1226493003
Cr-Commit-Position: refs/heads/master@{#29485}
verwaest [Mon, 6 Jul 2015 10:40:24 +0000 (03:40 -0700)]
Wrap elements.cc in an anonymous namespace
BUG=
Review URL: https://codereview.chromium.org/
1221363002
Cr-Commit-Position: refs/heads/master@{#29484}
machenbach [Mon, 6 Jul 2015 10:35:28 +0000 (03:35 -0700)]
Revert of [test] Move test262-es6 into test262. (patchset #2 id:20001 of https://codereview.chromium.org/
1215303008/)
Reason for revert:
[Sheriff] Breaks test262 on mac
Original issue's description:
> [test] Move test262-es6 into test262.
>
> BUG=v8:4254
> LOG=n
>
> Committed: https://crrev.com/
aaa457b26f6c0f624cf5887e60dc497f6dccabae
> Cr-Commit-Position: refs/heads/master@{#29479}
TBR=rossberg@chromium.org,arv@chromium.org,littledan@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:4254
Review URL: https://codereview.chromium.org/
1227503002
Cr-Commit-Position: refs/heads/master@{#29483}
verwaest [Mon, 6 Jul 2015 10:33:12 +0000 (03:33 -0700)]
Always mark entire prototype descriptor arrays.
BUG=chromium:506952
LOG=n
Review URL: https://codereview.chromium.org/
1226783002
Cr-Commit-Position: refs/heads/master@{#29482}
yangguo [Mon, 6 Jul 2015 10:25:50 +0000 (03:25 -0700)]
Do not truncate message strings.
R=ishell@chromium.org
BUG=chromium:505539
LOG=N
Review URL: https://codereview.chromium.org/
1214373005
Cr-Commit-Position: refs/heads/master@{#29481}
machenbach [Mon, 6 Jul 2015 10:01:27 +0000 (03:01 -0700)]
Revert "[turbofan] Add new JSFrameSpecialization reducer."
Also revert "[turbofan] Perform OSR deconstruction early and remove type propagation."
This reverts commit
b0a852e8c2ce2add1ba8a26d572aff39af0968a3.
This reverts commit
cdbb6c485b8d07fd4ad1cb000d54a937507e3b3e.
NOTRY=true
NOTREECHECKS=true
BUG=v8:4273
LOG=n
TBR=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
1225743002
Cr-Commit-Position: refs/heads/master@{#29480}
machenbach [Mon, 6 Jul 2015 09:24:32 +0000 (02:24 -0700)]
[test] Move test262-es6 into test262.
BUG=v8:4254
LOG=n
Review URL: https://codereview.chromium.org/
1215303008
Cr-Commit-Position: refs/heads/master@{#29479}
bmeurer [Mon, 6 Jul 2015 09:15:14 +0000 (02:15 -0700)]
[turbofan] Perform OSR deconstruction early and remove type propagation.
This way we don't have to deal with dead pre-OSR code in the graph and
risk optimizing the wrong code, especially we don't make optimistic
assumptions in the dead code that leaks into the OSR code (i.e. deopt
guards are in dead code, but the types propagate to OSR code via the
OsrValue type back propagation).
BUG=v8:4273
LOG=n
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1215333005
Cr-Commit-Position: refs/heads/master@{#29478}
verwaest [Mon, 6 Jul 2015 08:53:41 +0000 (01:53 -0700)]
Cleanup Delete backend implementation.
BUG=v8:4137
LOG=n
Review URL: https://codereview.chromium.org/
1218813012
Cr-Commit-Position: refs/heads/master@{#29477}
bmeurer [Mon, 6 Jul 2015 08:27:03 +0000 (01:27 -0700)]
[turbofan] Add new JSFrameSpecialization reducer.
The JSFrameSpecialization specializes an OSR graph to the current
unoptimized frame on which we will perform the on-stack replacement.
This is used for asm.js functions, where we cannot reuse the OSR code
object anyway because of context specialization, and so we could as well
specialize to the max instead.
It works by replacing all OsrValues in the graph with their values in
the JavaScriptFrame.
The idea is that using this trick we get better performance without
doing the unsound backpropagation of types to OsrValues later. This is
the first step towards fixing OSR for TurboFan.
R=jarin@chromium.org
BUG=v8:4273
LOG=n
Review URL: https://codereview.chromium.org/
1225683004
Cr-Commit-Position: refs/heads/master@{#29476}
machenbach [Mon, 6 Jul 2015 08:25:55 +0000 (01:25 -0700)]
Revert of Concurrent sweeping of code space. (patchset #4 id:60001 of https://codereview.chromium.org/
1222013002/)
Reason for revert:
[Sheriff] Increased flaky crashes. See:
https://code.google.com/p/v8/issues/detail?id=4275
Original issue's description:
> Concurrent sweeping of code space.
>
> BUG=
>
> Committed: https://crrev.com/
3050b52f57d652dc45c8baf416e174f22dc2c159
> Cr-Commit-Position: refs/heads/master@{#29456}
TBR=jochen@chromium.org,hpayer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1223763003
Cr-Commit-Position: refs/heads/master@{#29475}
jochen [Mon, 6 Jul 2015 07:11:40 +0000 (00:11 -0700)]
Remove usage of to-be-deprecated APIs from tools
BUG=v8:4134
R=bmeurer@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1217123004
Cr-Commit-Position: refs/heads/master@{#29474}
jochen [Mon, 6 Jul 2015 07:09:07 +0000 (00:09 -0700)]
Make v8::Handle as "deprecated soon"
BUG=v8:4131
R=bmeurer@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1224623004
Cr-Commit-Position: refs/heads/master@{#29473}
mstarzinger [Mon, 6 Jul 2015 03:40:22 +0000 (20:40 -0700)]
[turbofan] Fix bogus materialization from frame with OSR.
The context constant cannot be materialized from the frame when we are
compiling for OSR, because the context spill slot contains the current
instead of the outermost context in full-codegen.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
1220013003
Cr-Commit-Position: refs/heads/master@{#29472}
mstarzinger [Mon, 6 Jul 2015 03:39:15 +0000 (20:39 -0700)]
[turbofan] Use OSR value for innermost context value.
This changes the OsrValue insertion in the AstGraphBuilder to emit a
proper OsrValue instead of a special Parameter for the inner context
value at the OSR entry point.
R=titzer@chromium.org
Review URL: https://codereview.chromium.org/
1213043005
Cr-Commit-Position: refs/heads/master@{#29471}
machenbach [Sun, 5 Jul 2015 18:18:45 +0000 (11:18 -0700)]
Revert of Replace reduce-memory mode in idle notification with delayed clean-up GC. (patchset #17 id:320001 of https://codereview.chromium.org/
1218863002/)
Reason for revert:
[Sheriff] Looks like it blocks the roll (bisected). Speculative revert.
https://codereview.chromium.org/
1210293003/
Original issue's description:
> Replace reduce-memory mode in idle notification with delayed clean-up GC.
>
> BUG=490559
> LOG=NO
>
> Committed: https://crrev.com/
0ecd9e1bd2c6b519d4e7285f46cb7e844bc2235c
> Cr-Commit-Position: refs/heads/master@{#29451}
TBR=hpayer@chromium.org,ulan@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=490559
Review URL: https://codereview.chromium.org/
1226703002
Cr-Commit-Position: refs/heads/master@{#29470}
machenbach [Sun, 5 Jul 2015 18:17:35 +0000 (11:17 -0700)]
Revert of Make ARM compiler happy after
0ecd9e1bd (patchset #1 id:1 of https://codereview.chromium.org/
1219863003/)
Reason for revert:
Revert in order to revert
0ecd9e1bd
Original issue's description:
> Make ARM compiler happy after
0ecd9e1bd
>
> BUG=
> TBR=hpayer@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://crrev.com/
bef1cbfdf9991f05ca142448098e8076cb7c2d26
> Cr-Commit-Position: refs/heads/master@{#29454}
TBR=ulan@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1221303002
Cr-Commit-Position: refs/heads/master@{#29469}
verwaest [Fri, 3 Jul 2015 15:06:39 +0000 (08:06 -0700)]
Increment descriptor array slack for prototypes by a constant rather than 50%
Additionally speed up instantiation of ObjectTemplates by preallocating enough space in the descriptor arrays
BUG=v8:4184
LOG=n
Review URL: https://codereview.chromium.org/
1218403002
Cr-Commit-Position: refs/heads/master@{#29468}
verwaest [Fri, 3 Jul 2015 12:38:51 +0000 (05:38 -0700)]
Remove obsolete GetAccessorPair from ElementsAccessor
BUG=v8:4137
LOG=n
Review URL: https://codereview.chromium.org/
1224533003
Cr-Commit-Position: refs/heads/master@{#29467}
hpayer [Fri, 3 Jul 2015 12:13:43 +0000 (05:13 -0700)]
Record code slots that may point to evacuation candidate objects after deoptimizing them.
BUG=chromium:506811
LOG=n
Review URL: https://codereview.chromium.org/
1225573002
Cr-Commit-Position: refs/heads/master@{#29466}
bmeurer [Fri, 3 Jul 2015 11:41:54 +0000 (04:41 -0700)]
[turbofan] Right hand side of shifts needs ToUint32.
Currently we lower shifts directly to machine operators, and add an
appropriate Word32And to implement the & 0x1F operation on the right
hand side required by the specification. However for Word32And we assume
Int32 in simplified lowering, which is basically changes the right hand
side bit interpretation for the shifts from Uint32 to Int32, which is
obviously wrong. So now we represent that explicitly by proper
simplified operators for the shifts, which are lowered to machine in
simplified lowering.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1213803008
Cr-Commit-Position: refs/heads/master@{#29465}
yangguo [Fri, 3 Jul 2015 11:11:22 +0000 (04:11 -0700)]
Make deserialize-script-id test more robust.
R=machenbach@chromium.org
BUG=v8:4127
LOG=N
Review URL: https://codereview.chromium.org/
1215123002
Cr-Commit-Position: refs/heads/master@{#29464}
yangguo [Fri, 3 Jul 2015 10:32:27 +0000 (03:32 -0700)]
Make sure the constant pool size is as promised.
LOG=N
R=bmeurer@chromium.org
BUG=chromium:506443
Review URL: https://codereview.chromium.org/
1217673003
Cr-Commit-Position: refs/heads/master@{#29463}
Ilija.Pavlovic [Fri, 3 Jul 2015 10:31:51 +0000 (03:31 -0700)]
MIPS: Disassembler enhancement. Disassembled branch instruction displays branch target absolute address.
TEST=cctest/test-disasm-mips[64]
BUG=
Review URL: https://codereview.chromium.org/
1213553004
Cr-Commit-Position: refs/heads/master@{#29462}
mstarzinger [Fri, 3 Jul 2015 07:47:26 +0000 (00:47 -0700)]
Serializer forces flushing of optimized code map.
This makes sure that the GC fully flushes the optimized code map when
the serializer is preparing a snapshot. Otherwise closures and contexts
could end up in the startup snapshot.
R=hpayer@chromium.org
TEST=cctest/test-serialize/SerializeInternalReference
Review URL: https://codereview.chromium.org/
1215063007
Cr-Commit-Position: refs/heads/master@{#29461}
titzer [Fri, 3 Jul 2015 05:13:57 +0000 (22:13 -0700)]
[turbofan] Add Uint64LessThanOrEqual to 64-bit TurboFan backends.
Also add control inputs to 64-bit integer divide and modulus operations.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
1223613002
Cr-Commit-Position: refs/heads/master@{#29460}
dstence [Thu, 2 Jul 2015 18:46:23 +0000 (11:46 -0700)]
PPC: perf enhancement: omit redundant compare with 0
R=mbrandy@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1222133004
Cr-Commit-Position: refs/heads/master@{#29459}
ulan [Thu, 2 Jul 2015 18:25:42 +0000 (11:25 -0700)]
Fix data-race in default implementation of delayed task.
BUG=
Review URL: https://codereview.chromium.org/
1222203003
Cr-Commit-Position: refs/heads/master@{#29458}
ulan [Thu, 2 Jul 2015 18:24:34 +0000 (11:24 -0700)]
Grow heap slower if --optimize-for-size flag is on.
BUG=chromium:505776
LOG=NO
Review URL: https://codereview.chromium.org/
1221573003
Cr-Commit-Position: refs/heads/master@{#29457}
hpayer [Thu, 2 Jul 2015 16:45:34 +0000 (09:45 -0700)]
Concurrent sweeping of code space.
BUG=
Review URL: https://codereview.chromium.org/
1222013002
Cr-Commit-Position: refs/heads/master@{#29456}
verwaest [Thu, 2 Jul 2015 16:44:17 +0000 (09:44 -0700)]
Move slow classes above fast to directly call ReconfigureImpl, remove friends, make things public
BUG=v8:4137
LOG=n
Review URL: https://codereview.chromium.org/
1225493002
Cr-Commit-Position: refs/heads/master@{#29455}
ulan [Thu, 2 Jul 2015 16:29:17 +0000 (09:29 -0700)]
Make ARM compiler happy after
0ecd9e1bd
BUG=
TBR=hpayer@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1219863003
Cr-Commit-Position: refs/heads/master@{#29454}
mbrandy [Thu, 2 Jul 2015 15:50:56 +0000 (08:50 -0700)]
PPC: [turbofan] Enable tail calls for %_CallRuntime.
Port
1fa4285e1ce16cfadf8c40a0993491ec4e2bbbe0
Original commit message:
This involves:
- Enabling the tail call optimization reducer in all cases.
- Adding an addition flag to CallFunctionParameters to mark call sites
that can be tail-called enabled.
- Only set the tail-call flag for %_CallFunction.
R=danno@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1224533002
Cr-Commit-Position: refs/heads/master@{#29453}
jochen [Thu, 2 Jul 2015 15:42:35 +0000 (08:42 -0700)]
Remove usage of to-be-deprecated APIs from samples
BUG=v8:4134
R=vogelheim@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1219133004
Cr-Commit-Position: refs/heads/master@{#29452}
ulan [Thu, 2 Jul 2015 15:41:29 +0000 (08:41 -0700)]
Replace reduce-memory mode in idle notification with delayed clean-up GC.
BUG=490559
LOG=NO
Review URL: https://codereview.chromium.org/
1218863002
Cr-Commit-Position: refs/heads/master@{#29451}
verwaest [Thu, 2 Jul 2015 15:27:19 +0000 (08:27 -0700)]
Avoid converting key to string for deleting of elements
Additionally rips out (now) unnecessary duplicate code in DefineArrayProperty.
BUG=
Review URL: https://codereview.chromium.org/
1224523002
Cr-Commit-Position: refs/heads/master@{#29450}
conradw [Thu, 2 Jul 2015 15:24:48 +0000 (08:24 -0700)]
[strong] Add test for strong load involving proxies
Keeping this CL separate in case there are more GC-stress problems.
BUG=v8:3956
LOG=N
Review URL: https://codereview.chromium.org/
1217543006
Cr-Commit-Position: refs/heads/master@{#29449}
conradw [Thu, 2 Jul 2015 15:23:40 +0000 (08:23 -0700)]
[strong] Add tests for loading from super, loading with access checks
BUG=v8:3956
LOG=N
Review URL: https://codereview.chromium.org/
1219663009
Cr-Commit-Position: refs/heads/master@{#29448}
verwaest [Thu, 2 Jul 2015 14:38:37 +0000 (07:38 -0700)]
Distinguish slow from fast sloppy arguments
BUG=v8:4137
LOG=n
Review URL: https://codereview.chromium.org/
1221713003
Cr-Commit-Position: refs/heads/master@{#29447}
ulan [Thu, 2 Jul 2015 14:37:23 +0000 (07:37 -0700)]
Fix memory-leak in default platform implementation of delayed tasks.
BUG=
Review URL: https://codereview.chromium.org/
1221083004
Cr-Commit-Position: refs/heads/master@{#29446}
jochen [Thu, 2 Jul 2015 14:22:57 +0000 (07:22 -0700)]
Only define V8_USE_EXTERNAL_SNAPSHOT_DATA if we actually use a snapshot
BUG=none
R=vogelheim@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1218153006
Cr-Commit-Position: refs/heads/master@{#29445}
verwaest [Thu, 2 Jul 2015 13:29:48 +0000 (06:29 -0700)]
Cleanup API property handling
BUG=v8:4137
LOG=n
Review URL: https://codereview.chromium.org/
1226473003
Cr-Commit-Position: refs/heads/master@{#29444}
ulan [Thu, 2 Jul 2015 13:28:36 +0000 (06:28 -0700)]
Remove redundant check in GCTracer::SampleAllocation.
BUG=
Review URL: https://codereview.chromium.org/
1213593004
Cr-Commit-Position: refs/heads/master@{#29443}
erikcorry [Thu, 2 Jul 2015 13:27:27 +0000 (06:27 -0700)]
Reland: Fix logic for incremental marking steps on tenured allocation
BUG=
Review URL: https://codereview.chromium.org/
1077153004
Cr-Commit-Position: refs/heads/master@{#29442}
Michael Achenbach [Thu, 2 Jul 2015 13:24:07 +0000 (15:24 +0200)]
Whitespace change to test build priorities.
Cr-Commit-Position: refs/heads/master@{#29441}
mstarzinger [Thu, 2 Jul 2015 09:22:46 +0000 (02:22 -0700)]
Prepare profiler test for try-catch being optimizable.
This is an adaptation of test-profile-generator/BailoutReason to when
try-catch statements are optimizable. They will no longer cause any
bailout reason to be reported.
R=yangguo@chromium.org
TEST=cctest/test-profile-generator/BailoutReason
Review URL: https://codereview.chromium.org/
1213113007
Cr-Commit-Position: refs/heads/master@{#29440}
ishell [Thu, 2 Jul 2015 09:05:39 +0000 (02:05 -0700)]
Remove deprecated v8::Object::TurnOnAccessCheck() from the V8 API.
The only right way to enable access checks is to install access check callbacks on an object template via v8::ObjectTemplate::SetAccessCheckCallbacks(). It does not make sense to enable access checks on an arbitrary object.
Review URL: https://codereview.chromium.org/
1217893012
Cr-Commit-Position: refs/heads/master@{#29439}
danno [Thu, 2 Jul 2015 09:04:31 +0000 (02:04 -0700)]
[turbofan] Inline %_FixedArrayGet intrinsic.
Review URL: https://codereview.chromium.org/
1223583003
Cr-Commit-Position: refs/heads/master@{#29438}
dcheng [Thu, 2 Jul 2015 07:25:48 +0000 (00:25 -0700)]
Fix some clang warnings with -Wmissing-braces in v8.
Clang warns if there are missing braces around a subobject
initializer. The most common idiom that triggers this is:
STRUCT s = {0};
if the first field of STRUCT is itself a struct. This can
be more simply written as:
STRUCT s = {};
which also prevents the warning from firing.
BUG=chromium:505297
LOG=n
Review URL: https://codereview.chromium.org/
1219233003
Cr-Commit-Position: refs/heads/master@{#29437}
danno [Thu, 2 Jul 2015 06:20:03 +0000 (23:20 -0700)]
[turbofan] Enable tail calls for %_CallRuntime.
This involves:
- Enabling the tail call optimization reducer in all cases.
- Adding an addition flag to CallFunctionParameters to mark call sites
that can be tail-called enabled.
- Only set the tail-call flag for %_CallFunction.
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/
1216933011
Cr-Commit-Position: refs/heads/master@{#29436}
martyn.capewell [Thu, 2 Jul 2015 04:32:03 +0000 (21:32 -0700)]
[turbofan] Support cmp with shift/extend on ARM64.
Support 32-bit cmp with shift/extend by reusing the existing add/sub shift and
extend code.
Review URL: https://codereview.chromium.org/
1218103005
Cr-Commit-Position: refs/heads/master@{#29435}
chunyang.dai [Thu, 2 Jul 2015 04:27:17 +0000 (21:27 -0700)]
X87: [strong] Implement strong property access semantics.
port
7281f8015181736703d3663c9c570be537f008c1 (r29384)
Fix a bug in R29384.
original commit message:
Revert "Revert relanded strong property access CL"
Regression issues should be solved. Initial patchset is the original, subsequent patchsets are the fixing modifications.
This reverts commit
4ac7be56568f0ac71e7bf10e8bec33c33893e5b5.
BUG=
Review URL: https://codereview.chromium.org/
1210873007
Cr-Commit-Position: refs/heads/master@{#29434}
chunyang.dai [Thu, 2 Jul 2015 04:08:20 +0000 (21:08 -0700)]
X87: Make context register implicit for CallInterfaceDescriptors
port
7015fd20531ee8498352f82d626c78023a1f6857 (r29402)
original commit message:
Up until now the context register was listed explicitly in each stub's
CallInterfaceDescriptor. This was problematic, because it was listed
first in the list of register parameters--which is fine for Crankshaft,
which is more or less built to handle the context as the first
parameter-- but not ideal for TurboFan, which adds the context at
the end of all function parameters. Now the context register is no
longer in the register list and can be handled appropriately by both
compilers. Specifically, this allows the FunctionType specified for
each CallInterfaceDescriptor to exactly match the parameter register
list.
BUG=
Review URL: https://codereview.chromium.org/
1216543004
Cr-Commit-Position: refs/heads/master@{#29433}
chunyang.dai [Thu, 2 Jul 2015 04:02:33 +0000 (21:02 -0700)]
X87: [turbofan] Add CalleeSavedFPRegisters to CallDescriptor.
port
b4275a70eb1e6b6dc0f3c77b9be6bcfca3f80c20 (r29378)
original commit message:
This change makes possible to save and restore the FP registers
in the Prologue and Return parts for the CallAddress kind functions.
BUG=
Review URL: https://codereview.chromium.org/
1214573006
Cr-Commit-Position: refs/heads/master@{#29432}
v8-autoroll [Thu, 2 Jul 2015 03:26:02 +0000 (20:26 -0700)]
Update V8 DEPS.
Rolling v8/third_party/icu to
7fe225d77f307fdbe24695179a84336ef95c1253
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1225463002
Cr-Commit-Position: refs/heads/master@{#29431}
littledan [Wed, 1 Jul 2015 23:10:13 +0000 (16:10 -0700)]
Include Harmony Array/TypedArray methods unconditionally
Conditionally including Array and TypedArray methods seems to cause
a slowdown in V8 context creation, possibly due to the new code added.
BUG=chromium:504629
R=adamk@chromium.org
LOG=Y
Review URL: https://codereview.chromium.org/
1215863003
Cr-Commit-Position: refs/heads/master@{#29430}
erikcorry [Wed, 1 Jul 2015 17:42:15 +0000 (10:42 -0700)]
GC. Delay/avoid entering high promotion mode
R=hpayer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1217813012
Cr-Commit-Position: refs/heads/master@{#29429}
adamk [Wed, 1 Jul 2015 17:20:48 +0000 (10:20 -0700)]
Remove unused is_class_scope bit from Scope and ScopeInfo
This was added in https://chromium.googlesource.com/v8/v8/+/
4a709dd65,
but the only check for it that remained in the final patch is inside
a DCHECK. It appears that the approach for checking use of class names
in methods evolved quite a bit over the review of the original patch.
Review URL: https://codereview.chromium.org/
1219993002
Cr-Commit-Position: refs/heads/master@{#29428}
verwaest [Wed, 1 Jul 2015 16:49:19 +0000 (09:49 -0700)]
Remove outdated comment
BUG=
Review URL: https://codereview.chromium.org/
1220883006
Cr-Commit-Position: refs/heads/master@{#29427}
binji [Wed, 1 Jul 2015 16:41:48 +0000 (09:41 -0700)]
Change d8 Worker API so it takes a string instead of a function.
This is more consistent with the DOM API, and is clearer w.r.t. which values
are available in the lexical environment of the Worker.
BUG=chromium:497295
R=jarin@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1218553004
Cr-Commit-Position: refs/heads/master@{#29426}
Djordje.Pesic [Wed, 1 Jul 2015 16:15:34 +0000 (09:15 -0700)]
MIPS: disabling rsqrt and recip for mips32r1
Disabling rsqrt and recip for mips32r1 in assembler, disassembler and simulator
Review URL: https://codereview.chromium.org/
1221663006
Cr-Commit-Position: refs/heads/master@{#29425}
Toon Verwaest [Wed, 1 Jul 2015 15:39:09 +0000 (17:39 +0200)]
Ensure we enable/disable the bootstrapping flag
TBR=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/
1213683005.
Cr-Commit-Position: refs/heads/master@{#29424}
verwaest [Wed, 1 Jul 2015 15:34:18 +0000 (08:34 -0700)]
Return this in Context::native_context if receiver is a native context.
Checking for native context is faster than checking for global object. Additionally it speeds up the case were it actually is the native context, while not slowing down the alternative case. The bootstrapper only needs to access the native context from the native context, so this avoids the expensive fallback.
BUG=
Review URL: https://codereview.chromium.org/
1214903017
Cr-Commit-Position: refs/heads/master@{#29423}
Toon Verwaest [Wed, 1 Jul 2015 15:33:22 +0000 (17:33 +0200)]
Use is_bootstrapping everywhere
TBR=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/
1220053002.
Cr-Commit-Position: refs/heads/master@{#29422}
mbrandy [Wed, 1 Jul 2015 14:21:04 +0000 (07:21 -0700)]
PPC: Make context register implicit for CallInterfaceDescriptors
Port
7015fd20531ee8498352f82d626c78023a1f6857
Original commit message:
Up until now the context register was listed explicitly in each stub's
CallInterfaceDescriptor. This was problematic, because it was listed
first in the list of register parameters--which is fine for Crankshaft,
which is more or less built to handle the context as the first
parameter-- but not ideal for TurboFan, which adds the context at
the end of all function parameters. Now the context register is no
longer in the register list and can be handled appropriately by both
compilers. Specifically, this allows the FunctionType specified for
each CallInterfaceDescriptor to exactly match the parameter register
list.
R=danno@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1218463004
Cr-Commit-Position: refs/heads/master@{#29421}
akos.palfi [Wed, 1 Jul 2015 14:19:57 +0000 (07:19 -0700)]
MIPS: Fix '[turbofan] Add CalleeSavedFPRegisters to CallDescriptor.'
Fixes debug mode assertions.
TEST=cctest/test-run-machops/RunChangeUint32ToFloat64_B
BUG=
Review URL: https://codereview.chromium.org/
1211143003
Cr-Commit-Position: refs/heads/master@{#29420}
verwaest [Wed, 1 Jul 2015 14:18:50 +0000 (07:18 -0700)]
Micro-optimize lookupiterator: faster path for fast-mode objects
BUG=chromium:505998
LOG=n
Review URL: https://codereview.chromium.org/
1222543003
Cr-Commit-Position: refs/heads/master@{#29419}
Toon Verwaest [Wed, 1 Jul 2015 14:17:47 +0000 (16:17 +0200)]
Update tests to disable interceptors during bootstrapping
BUG=
R=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/
1215593005.
Cr-Commit-Position: refs/heads/master@{#29418}
mstarzinger [Wed, 1 Jul 2015 13:43:36 +0000 (06:43 -0700)]
Fix CodeFlusher::ProcessOptimizedCodeMaps stale fields.
This makes sure that all pointer fields in an entry in the optimized
code map are considered weak, not just the code object. If just one
field of an entry dies, then the whole entry dies.
R=hpayer@chromium.org
Review URL: https://codereview.chromium.org/
1217813013
Cr-Commit-Position: refs/heads/master@{#29417}
mbrandy [Wed, 1 Jul 2015 13:34:57 +0000 (06:34 -0700)]
PPC: Fix InstanceOfStub's inlined call site logic.
This change makes the patching logic less prone to errors in the
face of variable instruction mov sequences.
R=dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1213383003
Cr-Commit-Position: refs/heads/master@{#29416}
mstarzinger [Wed, 1 Jul 2015 12:29:27 +0000 (05:29 -0700)]
[turbofan] Support preserving of optimized code maps over GCs.
This extends the existing support for --noflush-optimized-code-cache
mechanism to the newly added context-independent code entry in the
optimized code map. Note that this machanism is still off by default.
R=hpayer@chromium.org
Review URL: https://codereview.chromium.org/
1220883002
Cr-Commit-Position: refs/heads/master@{#29415}
verwaest [Wed, 1 Jul 2015 12:19:55 +0000 (05:19 -0700)]
Minor performance improvements to the LookupIterator
This change changes bootstrapping semantics for intercepted global objects. Unlike before, we'll now also call into the interceptor during bootstrapping. This affects properties loaded from within the runtime, such as global.Array and global.Symbol. The embedder will need to make sure that those values are the expected values during bootstrapping.
BUG=chromium:505998
LOG=n
Review URL: https://codereview.chromium.org/
1220813005
Cr-Commit-Position: refs/heads/master@{#29414}
chunyang.dai [Wed, 1 Jul 2015 12:05:17 +0000 (05:05 -0700)]
X87: [es6] Make new.target work in functions.
port
7a63bf77eb7610afdc1a968f7660781e5160ba8d (r29358).
original commit message:
This makes new.target work in [[Call]] and [[Construct]] of ordinary
functions.
We achieve this by introducing a new construct stub for functions that
uses the new.target variable. The construct stub pushes the original
constructor just above the receiver in the construct frame.
BUG=
Review URL: https://codereview.chromium.org/
1217083004
Cr-Commit-Position: refs/heads/master@{#29413}
chunyang.dai [Wed, 1 Jul 2015 12:03:26 +0000 (05:03 -0700)]
X87: Unify the stack layout for construct frames.
port
876ae425980f67f489323671f13d2314b0ce0a91 (r29292).
original commit message:
The stack layout was different for different ports.
BUG=
Review URL: https://codereview.chromium.org/
1219823002
Cr-Commit-Position: refs/heads/master@{#29412}
chunyang.dai [Wed, 1 Jul 2015 11:59:39 +0000 (04:59 -0700)]
X87: VectorICs: Lithium support for vector-based stores.
port
8a3cf4ecef1fce76137270845604292bdbdadcd4 (r29310).
BUG=
Review URL: https://codereview.chromium.org/
1213373002
Cr-Commit-Position: refs/heads/master@{#29411}
chunyang.dai [Wed, 1 Jul 2015 11:47:01 +0000 (04:47 -0700)]
X87: Vector ICs: Like megamorphic keyed koads, use a dummy vector for stores.
port
9e7af9efc5857b3c7e23a77d257f3dfbea597753 (r29280).
original commit message:
It's useful for the megamorphic keyed store case to not require a
vector and slot as input. Analogous to the load case, we have a dummy
one-ic-slot vector to aid. Since the only kind of MISS is for
megamorphic cache stub failures, we don't need the real vector.
The reason is that megamorphic cache stub failures don't result in any
change to the type feedback vector state.
BUG=
Review URL: https://codereview.chromium.org/
1212653011
Cr-Commit-Position: refs/heads/master@{#29410}
chunyang.dai [Wed, 1 Jul 2015 11:39:06 +0000 (04:39 -0700)]
X87: [ic] Record call counts for monomorphic calls made with an IC.
port
c1a4f7477f03ebb0c6889bbf8ea6a4c928e0d413 (r29281).
original commit message:
The idea is that TurboFan can use this information for more intelligent
inlining.
BUG=
Review URL: https://codereview.chromium.org/
1213723005
Cr-Commit-Position: refs/heads/master@{#29409}
chunyang.dai [Wed, 1 Jul 2015 11:31:47 +0000 (04:31 -0700)]
X87: [turbofan] Add basic support for calling to (a subset of) C functions.
port
a58ba8d80179bf5b6b7245590c82e47fda8c8a5e (r29279).
original commit message:
This introduces some initial building blocks for calling out to
C/C++ functions directly from TurboFan generated code objects.
BUG=
Review URL: https://codereview.chromium.org/
1214903007
Cr-Commit-Position: refs/heads/master@{#29408}