Ben L. Titzer [Wed, 13 May 2015 13:08:25 +0000 (15:08 +0200)]
[turbofan] Turn off type feedback in turbo mode until StoreNamed issue is resolved.
TBR=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1130073006
Cr-Commit-Position: refs/heads/master@{#28394}
bmeurer [Wed, 13 May 2015 12:53:54 +0000 (05:53 -0700)]
[turbofan] Lower more number operations to integer operations.
If the typer was able to prove that a NumberAdd/Sub/Mul/Div/Mod always
produces a Signed32/Unsigned32 value, and the inputs are
Signed32/Unsigned32, we can lower the node to the corresponding integer
operation instead, no matter what the uses are.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1140933002
Cr-Commit-Position: refs/heads/master@{#28393}
titzer [Wed, 13 May 2015 12:29:38 +0000 (05:29 -0700)]
[turbofan] Add FrameStates before all property accesses.
R=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1133303006
Cr-Commit-Position: refs/heads/master@{#28392}
wingo [Wed, 13 May 2015 11:45:04 +0000 (04:45 -0700)]
Use ExpressionClassifier to identify valid arrow function formals
R=dslomov@chromium.org
LOG=N
BUG=
Review URL: https://codereview.chromium.org/
1138153003
Cr-Commit-Position: refs/heads/master@{#28391}
chunyang.dai [Wed, 13 May 2015 11:18:41 +0000 (04:18 -0700)]
X87: Resolve references to "this" the same way as normal variables
port
bd56d279b612f92d2d1d25f92d32fe7f42a455ca (R28340).
original commit message:
Make the parser handle references to "this" as unresolved variables, so the
same logic as for the rest of function parameters is used for the receiver.
Minor additions to the code generation handle copying the receiver to the
context, along with the rest of the function parameters.
Based on work by Adrian Perez de Castro <aperez@igalia.com>.
BUG=
Review URL: https://codereview.chromium.org/
1136953010
Cr-Commit-Position: refs/heads/master@{#28390}
bmeurer [Wed, 13 May 2015 11:17:31 +0000 (04:17 -0700)]
[turbofan] Use frame state before for shift operations as well.
This was already done for other binary operations, so it's basically
copying the existing functionality to shift left and shift right
logical/arithmetic.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1140883003
Cr-Commit-Position: refs/heads/master@{#28389}
chunyang.dai [Wed, 13 May 2015 11:16:25 +0000 (04:16 -0700)]
X87: New hydrogen instruction to reduce cost of growing an array on keyed stores.
port
3bce9c3afbe792167a54db49b16679289e0aea8f (r28359).
original commit message:
HMaybeGrowElements moves the situation where you actually have to grow
into deferred code. This means crankshaft doesn't have to spill registers
just to make the bounds comparison to see if it'll need to grow or not.
It makes the growing case a bit more expensive, but reduces the cost of
the general case.
BUG=
Review URL: https://codereview.chromium.org/
1124093008
Cr-Commit-Position: refs/heads/master@{#28388}
chunyang.dai [Wed, 13 May 2015 11:15:42 +0000 (04:15 -0700)]
X87: [strong] Check arity of functions
port
3226e980200035f37c86fef4b11fb307e17764a2 (28346).
original commit message:
[strong] Check arity of functions
In strong mode it is an error to call a function with too few
arguments.
This is enforced inside the ArgumentsAdaptorTrampoline.
This does not yet handle rest parameter
BUG=
Review URL: https://codereview.chromium.org/
1139913007
Cr-Commit-Position: refs/heads/master@{#28387}
chunyang.dai [Wed, 13 May 2015 11:14:35 +0000 (04:14 -0700)]
X87: Add a MathFloor stub generated with TurboFan
port
abc35080b33cc31e47b9913e64ac99ec687c6b65 (r28339)
original commit message:
This stub will be used as the basis of a Math.floor-specific CallIC to
detect and track calls to floor that return -0.
Along the way:
- Create a TurboFanCodeStub super class from which the StringLength and
MathRound TF stubs derive.
- Fix the ugly hack that passes the first stub parameter as the "this"
pointer in the the TF-compiled JS function.
- Fix bugs in the ia32/x64 disassembler.
BUG=
Review URL: https://codereview.chromium.org/
1134323002
Cr-Commit-Position: refs/heads/master@{#28386}
yangguo [Wed, 13 May 2015 10:18:35 +0000 (03:18 -0700)]
Prevent stack overflow in the serializer/deserializer.
We keep an eye on the recursion depth. Once it exceeds a limit, we serialize
only the object header and size, but defer serializing the object body for
after we have unwound the stack.
R=mvstanton@chromium.org
Review URL: https://codereview.chromium.org/
1125073004
Cr-Commit-Position: refs/heads/master@{#28385}
machenbach [Wed, 13 May 2015 10:00:43 +0000 (03:00 -0700)]
[test] Make perf runner able to pass extra flags to d8.
TEST=tools/run_perf.py --arch x64 --extra-flags="--turbo --turbo-verify-allocation --turbo-greedy-regalloc" /path/to/SunSpider.json
BUG=484208
LOG=n
TBR=tandrii@chromium.org
NOTRY=true
Review URL: https://codereview.chromium.org/
1128933007
Cr-Commit-Position: refs/heads/master@{#28384}
bmeurer [Wed, 13 May 2015 07:38:42 +0000 (00:38 -0700)]
[turbofan] Add frame state before JavaScript comparisons.
Use these check points to optimize comparisons where we already know
that one side cannot be a String (or turn into a string via
ToPrimitive).
Also remove bunch of useless DoNotCrash tests for the scheduler that are
painful to maintain and add almost no value.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1140583004
Cr-Commit-Position: refs/heads/master@{#28383}
dehrenberg [Wed, 13 May 2015 04:32:12 +0000 (21:32 -0700)]
Implement %TypedArray%.{fill,find,findIndex}
This patch adds three methods to TypedArrays which are already
implemented for arrays. The implementations are made by calling
out to the underlying code used by Arrays.
R=adamk@chromium.org
BUG=v8:3578
LOG=Y
Review URL: https://codereview.chromium.org/
1136663005
Cr-Commit-Position: refs/heads/master@{#28382}
dehrenberg [Tue, 12 May 2015 19:20:58 +0000 (12:20 -0700)]
TypedArray.prototype.copyWithin method
This patch adds the copyWithin method to TypedArrays. For the first
pass, the internals of Array.copyWithin are used. Eventually, a more
efficient form based on memcpy could be used instead.
BUG=v8:3578
LOG=Y
R=adamk@chromium.org, arv@chromium.org, caitpotter88@gmail.com
Review URL: https://codereview.chromium.org/
1131113002
Cr-Commit-Position: refs/heads/master@{#28381}
dslomov [Tue, 12 May 2015 18:12:56 +0000 (11:12 -0700)]
Fix test formatting
TBR=arv@chromium.org
Review URL: https://codereview.chromium.org/
1129713004
Cr-Commit-Position: refs/heads/master@{#28380}
arv [Tue, 12 May 2015 17:37:34 +0000 (10:37 -0700)]
[strong] Add more function arity tests
This is a follow up to https://codereview.chromium.org/
1115263004/
BUG=v8:3956
LOG=N
R=rossberg@chromium.org
Review URL: https://codereview.chromium.org/
1141603002
Cr-Commit-Position: refs/heads/master@{#28379}
ulan [Tue, 12 May 2015 17:25:04 +0000 (10:25 -0700)]
Make transition to reduce memory mode more conservative in idle time handler.
BUG=chromium:486005
LOG=NO
Review URL: https://codereview.chromium.org/
1131943004
Cr-Commit-Position: refs/heads/master@{#28378}
dslomov [Tue, 12 May 2015 17:13:12 +0000 (10:13 -0700)]
Fix the behavior of 'super.foo' assignment when receiver is not an object.
R=arv@chromium.org,verwaest@chromium.org
BUG=v8:4097
LOG=N
Review URL: https://codereview.chromium.org/
1132203005
Cr-Commit-Position: refs/heads/master@{#28377}
mvstanton [Tue, 12 May 2015 16:51:14 +0000 (09:51 -0700)]
Make LoadFastElementStub a HandlerStub.
BUG=
Review URL: https://codereview.chromium.org/
1132493006
Cr-Commit-Position: refs/heads/master@{#28376}
mbrandy [Tue, 12 May 2015 16:48:16 +0000 (09:48 -0700)]
PPC: Fix '[strong] Disallow implicit conversions for comparison'
Do not trash type register in kIsNotStringMask test.
R=conradw@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1124333013
Cr-Commit-Position: refs/heads/master@{#28375}
mbrandy [Tue, 12 May 2015 16:43:24 +0000 (09:43 -0700)]
PPC: Fix '[strong] Disallow implicit conversions for comparison'
Fix build break.
R=conradw@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1130603004
Cr-Commit-Position: refs/heads/master@{#28374}
hpayer [Tue, 12 May 2015 15:48:04 +0000 (08:48 -0700)]
Get more debugging data when crashing in Deoptimizer::GetOutputInfo.
BUG=
Review URL: https://codereview.chromium.org/
1138493004
Cr-Commit-Position: refs/heads/master@{#28373}
mbrandy [Tue, 12 May 2015 15:42:39 +0000 (08:42 -0700)]
PPC: New hydrogen instruction to reduce cost of growing an array on keyed stores.
Port
3bce9c3afbe792167a54db49b16679289e0aea8f
Original commit message:
HMaybeGrowElements moves the situation where you actually have to grow
into deferred code. This means crankshaft doesn't have to spill registers
just to make the bounds comparison to see if it'll need to grow or not.
It makes the growing case a bit more expensive, but reduces the cost of
the general case.
R=mvstanton@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1132743004
Cr-Commit-Position: refs/heads/master@{#28372}
yurys [Tue, 12 May 2015 15:40:31 +0000 (08:40 -0700)]
Provide accessor for object internal properties that doesn't require debugger to be active
Some of the DevTools' clients need to inspect JS objects without enabling debugger. This CL allows to inspect object's internal properties without enabling debugger and instantiating debug context.
Note that now debug context can be created lazily if v8::Debug::GetDebugContext is called when there is no debug listener. This is fragile and has already resulted in some subtle error. I'm going to fix that in a separate CL.
BUG=chromium:481845
LOG=Y
Review URL: https://codereview.chromium.org/
1134193002
Cr-Commit-Position: refs/heads/master@{#28371}
conradw [Tue, 12 May 2015 15:23:53 +0000 (08:23 -0700)]
[strong] Disallow implicit conversions for comparison
Implements the strong mode proposal's restrictions on implicit conversions
for the binary >, >=, <, and <= operators.
BUG=v8:3956
LOG=N
Review URL: https://codereview.chromium.org/
1130283002
Cr-Commit-Position: refs/heads/master@{#28370}
verwaest [Tue, 12 May 2015 15:11:18 +0000 (08:11 -0700)]
Mark internal AccessorInfo properties as "special data properties" to ensure correct strict-mode handling.
BUG=
Review URL: https://codereview.chromium.org/
1123163005
Cr-Commit-Position: refs/heads/master@{#28369}
yangguo [Tue, 12 May 2015 14:28:40 +0000 (07:28 -0700)]
Fix test-heap/BootstrapperExports on no-snap builds
NOTRY=true
NOTREECHECKS=true
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1137143003
Cr-Commit-Position: refs/heads/master@{#28368}
yangguo [Tue, 12 May 2015 14:00:47 +0000 (07:00 -0700)]
Reland "Use function wrapper argument to expose internal arrays to native scripts."
Review URL: https://codereview.chromium.org/
1138173002
Cr-Commit-Position: refs/heads/master@{#28367}
yangguo [Tue, 12 May 2015 13:52:26 +0000 (06:52 -0700)]
Migrate error messages, part 10.
R=mvstanton@chromium.org
Committed: https://crrev.com/
8608e619afe2b4514b0577bfb73a153b1550d41f
Cr-Commit-Position: refs/heads/master@{#28357}
Review URL: https://codereview.chromium.org/
1126043004
Cr-Commit-Position: refs/heads/master@{#28366}
yangguo [Tue, 12 May 2015 13:03:59 +0000 (06:03 -0700)]
Revert of Provide accessor for object internal properties that doesn't require debugger to be active (patchset #3 id:40001 of https://codereview.chromium.org/
1126103006/)
Reason for revert:
GC mole issues: https://chromegw.corp.google.com/i/client.v8/builders/V8%20Linux%20-%20gcmole/builds/1950/steps/GCMole%20ia32/logs/stdio
Original issue's description:
> Provide accessor for object internal properties that doesn't require debugger to be active
>
> Some of the DevTools' clients need to inspect JS objects without enabling debugger. This CL allows to inspect object's internal properties without enabling debugger and instantiating debug context.
>
> Note that now debug context can be created lazily if v8::Debug::GetDebugContext is called when there is no debug listener. This is fragile and has already resulted in some subtle error. I'm going to fix that in a separate CL.
>
> BUG=chromium:481845
> LOG=Y
>
> Committed: https://crrev.com/
bdeb0de88c8cf5f2c78f261b45314138f525110d
> Cr-Commit-Position: refs/heads/master@{#28362}
TBR=pfeldman@chromium.org,kozyatinskiy@chromium.org,yurys@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=chromium:481845
Review URL: https://codereview.chromium.org/
1133243002
Cr-Commit-Position: refs/heads/master@{#28365}
titzer [Tue, 12 May 2015 12:57:32 +0000 (05:57 -0700)]
[turbofan] Correctify FrameState before operations in JSTypeFeedbackSpecializer.
Handle missing or improper FrameStates more gracefully.
Uses the operator properties to determine if the node contains a FrameState
before, then checks if a valid bailout id exists.
R=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1135243002
Cr-Commit-Position: refs/heads/master@{#28364}
titzer [Tue, 12 May 2015 12:41:41 +0000 (05:41 -0700)]
[turbofan] Add AdvancedReducer::ReplaceWithValue() method and convert JSInlining to an AdvancedReducer.
Note that this is just a duplication for now. We'll want to get rid of the
NodeProperties::ReplaceWithValue() method in the long run.
R=bmeurer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1135483004
Cr-Commit-Position: refs/heads/master@{#28363}
yurys [Tue, 12 May 2015 12:38:17 +0000 (05:38 -0700)]
Provide accessor for object internal properties that doesn't require debugger to be active
Some of the DevTools' clients need to inspect JS objects without enabling debugger. This CL allows to inspect object's internal properties without enabling debugger and instantiating debug context.
Note that now debug context can be created lazily if v8::Debug::GetDebugContext is called when there is no debug listener. This is fragile and has already resulted in some subtle error. I'm going to fix that in a separate CL.
BUG=chromium:481845
LOG=Y
Review URL: https://codereview.chromium.org/
1126103006
Cr-Commit-Position: refs/heads/master@{#28362}
titzer [Tue, 12 May 2015 12:30:27 +0000 (05:30 -0700)]
[turbofan] Use FrameStatesBeforeAndAfter to simplify handling of before/after frame states in AstGraphBuilder.
R=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1128193005
Cr-Commit-Position: refs/heads/master@{#28361}
hpayer [Tue, 12 May 2015 11:44:46 +0000 (04:44 -0700)]
Implement unaligned allocate and allocate heap numbers in runtime double unaligned.
In follow-up CLs the scavenger and the MC collector should also respect the unalignment of heap numbers.
BUG=
Review URL: https://codereview.chromium.org/
1141523002
Cr-Commit-Position: refs/heads/master@{#28360}
mvstanton [Tue, 12 May 2015 08:47:15 +0000 (01:47 -0700)]
New hydrogen instruction to reduce cost of growing an array on keyed stores.
HMaybeGrowElements moves the situation where you actually have to grow
into deferred code. This means crankshaft doesn't have to spill registers
just to make the bounds comparison to see if it'll need to grow or not.
It makes the growing case a bit more expensive, but reduces the cost of
the general case.
BUG=chromium:484025
LOG=N
Review URL: https://codereview.chromium.org/
1124443004
Cr-Commit-Position: refs/heads/master@{#28359}
yangguo [Tue, 12 May 2015 08:33:27 +0000 (01:33 -0700)]
Revert of Migrate error messages, part 10. (patchset #2 id:20001 of https://codereview.chromium.org/
1126043004/)
Reason for revert:
patch didn't apply correctly.
Original issue's description:
> Migrate error messages, part 10.
>
> R=mvstanton@chromium.org
>
> Committed: https://crrev.com/
8608e619afe2b4514b0577bfb73a153b1550d41f
> Cr-Commit-Position: refs/heads/master@{#28357}
TBR=mvstanton@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1130073005
Cr-Commit-Position: refs/heads/master@{#28358}
yangguo [Tue, 12 May 2015 08:16:55 +0000 (01:16 -0700)]
Migrate error messages, part 10.
R=mvstanton@chromium.org
Review URL: https://codereview.chromium.org/
1126043004
Cr-Commit-Position: refs/heads/master@{#28357}
yurys [Tue, 12 May 2015 07:33:21 +0000 (00:33 -0700)]
Don't create debug context if debug listener is not set
If there had been no debug listener v8::Debug::GetDebugContext would have created new context and wouln't have kept reference to it. This way we may well end up with several debug contexts and disabled debugger.
As a side effect this change allows to efficiently distinguish debug context from blink contexts by simply comparing handles.
BUG=chromium:482290
LOG=Y
Review URL: https://codereview.chromium.org/
1136733002
Cr-Commit-Position: refs/heads/master@{#28356}
yangguo [Tue, 12 May 2015 07:26:05 +0000 (00:26 -0700)]
Revert of Use function wrapper argument to expose internal arrays to native scripts. (patchset #2 id:20001 of https://codereview.chromium.org/
1127983003/)
Reason for revert:
custom snapshot builder failing.
Original issue's description:
> Use function wrapper argument to expose internal arrays to native scripts.
>
> R=jkummerow@chromium.org
>
> Committed: https://crrev.com/
a9b5a1795449d94387218d25baed2c2b3c4fbadc
> Cr-Commit-Position: refs/heads/master@{#28354}
TBR=jkummerow@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Review URL: https://codereview.chromium.org/
1123353008
Cr-Commit-Position: refs/heads/master@{#28355}
yangguo [Tue, 12 May 2015 06:14:17 +0000 (23:14 -0700)]
Use function wrapper argument to expose internal arrays to native scripts.
R=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/
1127983003
Cr-Commit-Position: refs/heads/master@{#28354}
jochen [Tue, 12 May 2015 04:59:46 +0000 (21:59 -0700)]
Deprecate Isolate::New
Embedders need to always pass CreateParams now
BUG=none
LOG=y
R=vogelheim@chromium.org
Review URL: https://codereview.chromium.org/
1133723004
Cr-Commit-Position: refs/heads/master@{#28353}
v8-autoroll [Tue, 12 May 2015 03:24:43 +0000 (20:24 -0700)]
Update V8 DEPS.
Rolling v8/buildtools to
b73e5f70d7ac6be98fb2555461f631afc90216ce
Rolling v8/third_party/icu to
ce41627e388fb46ab49671bd16a5db81dcd75a71
Rolling v8/tools/clang to
6e259b4afbbd784a775e47f37379e57f002bf0c2
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1124083005
Cr-Commit-Position: refs/heads/master@{#28352}
dehrenberg [Tue, 12 May 2015 01:32:26 +0000 (18:32 -0700)]
Factor out core of Array.forEach and .every, for use in TypedArrays
The idea is to make this the model for future TypedArray methods.
A possible downside could be lower array method performance
if everything gets polymorhpic (but if enough inlining happens, it
should still be fast), but on the upside, this change means that
the TypedArray methods won't create as much code size bloat.
BUG=v8:3578
LOG=Y
R=adamk@chromium.org
CC=arv@chromium.org, caitpotter88@gmail.com
Review URL: https://codereview.chromium.org/
1133503003
Cr-Commit-Position: refs/heads/master@{#28351}
akos.palfi [Mon, 11 May 2015 23:07:43 +0000 (16:07 -0700)]
MIPS64: Improve '[strong] Check arity of functions'
Port
3226e980200035f37c86fef4b11fb307e17764a2
BUG=
Review URL: https://codereview.chromium.org/
1129323007
Cr-Commit-Position: refs/heads/master@{#28350}
caitpotter88 [Mon, 11 May 2015 22:53:41 +0000 (15:53 -0700)]
[es6] stage Object.assign() / --harmony-object
BUG=v8:4007
LOG=N
R=
Review URL: https://codereview.chromium.org/
1128873006
Cr-Commit-Position: refs/heads/master@{#28349}
mbrandy [Mon, 11 May 2015 19:08:49 +0000 (12:08 -0700)]
PPC: [strong] Check arity of functions
Port
3226e980200035f37c86fef4b11fb307e17764a2
Original commit message:
In strong mode it is an error to call a function with too few
arguments.
This is enforced inside the ArgumentsAdaptorTrampoline.
This does not yet handle rest parameters
R=arv@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1132153003
Cr-Commit-Position: refs/heads/master@{#28348}
paul.lind [Mon, 11 May 2015 18:49:54 +0000 (11:49 -0700)]
MIPS: Fix Add a MathFloor stub generated with TurboFan.
Port
abc35080b33cc31e47b9913e64ac99ec687c6b65
BUG=
Review URL: https://codereview.chromium.org/
1133103003
Cr-Commit-Position: refs/heads/master@{#28347}
arv [Mon, 11 May 2015 17:20:43 +0000 (10:20 -0700)]
[strong] Check arity of functions
In strong mode it is an error to call a function with too few
arguments.
This is enforced inside the ArgumentsAdaptorTrampoline.
This does not yet handle rest parameters
BUG=v8:3956
LOG=N
R=rossberg@chromium.org, dslomov@chromium.org
Review URL: https://codereview.chromium.org/
1115263004
Cr-Commit-Position: refs/heads/master@{#28346}
dslomov [Mon, 11 May 2015 16:28:28 +0000 (09:28 -0700)]
[destructuring] Implement basic binding destructuring infrastructure
This patch:
- Refactors Parser::ParseVariableDeclarations
- Introduces Parser::PatternMatcher class
- Implements matching a single variable pattern
- Implements rudimentary matching against object literal pattern
as a proof of concept
R=arv@chromium.org,rossberg@chromium.org
BUG=v8:811
LOG=N
Review URL: https://codereview.chromium.org/
1130623004
Cr-Commit-Position: refs/heads/master@{#28345}
titzer [Mon, 11 May 2015 15:56:07 +0000 (08:56 -0700)]
Add a bailout id just before every variable load in fullcode.
R=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1139733002
Cr-Commit-Position: refs/heads/master@{#28344}
wingo [Mon, 11 May 2015 15:42:19 +0000 (08:42 -0700)]
Remove Scope::scope_uses_this_ flag
R=rossberg@chromium.org
LOG=N
BUG=
Review URL: https://codereview.chromium.org/
1128963005
Cr-Commit-Position: refs/heads/master@{#28343}
mbrandy [Mon, 11 May 2015 14:34:31 +0000 (07:34 -0700)]
PPC: Resolve references to "this" the same way as normal variables
Port
bd56d279b612f92d2d1d25f92d32fe7f42a455ca
Original commit message:
Make the parser handle references to "this" as unresolved variables, so the
same logic as for the rest of function parameters is used for the receiver.
Minor additions to the code generation handle copying the receiver to the
context, along with the rest of the function parameters.
Based on work by Adrian Perez de Castro <aperez@igalia.com>.
This is a reapplication of https://codereview.chromium.org/
1130733003.
R=wingo@igalia.com, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1134073003
Cr-Commit-Position: refs/heads/master@{#28342}
mbrandy [Mon, 11 May 2015 14:33:27 +0000 (07:33 -0700)]
PPC: Add a MathFloor stub generated with TurboFan
Port
abc35080b33cc31e47b9913e64ac99ec687c6b65
Original commit message:
This stub will be used as the basis of a Math.floor-specific CallIC to
detect and track calls to floor that return -0.
Along the way:
- Create a TurboFanCodeStub super class from which the StringLength and
MathRound TF stubs derive.
- Fix the ugly hack that passes the first stub parameter as the "this"
pointer in the the TF-compiled JS function.
- Fix bugs in the ia32/x64 disassembler.
R=danno@chromium.org, dstence@us.ibm.com, michael_dawson@ca.ibm.com
BUG=
Review URL: https://codereview.chromium.org/
1137033002
Cr-Commit-Position: refs/heads/master@{#28341}
wingo [Mon, 11 May 2015 11:49:48 +0000 (04:49 -0700)]
Resolve references to "this" the same way as normal variables
Make the parser handle references to "this" as unresolved variables, so the
same logic as for the rest of function parameters is used for the receiver.
Minor additions to the code generation handle copying the receiver to the
context, along with the rest of the function parameters.
Based on work by Adrian Perez de Castro <aperez@igalia.com>.
This is a reapplication of https://codereview.chromium.org/
1130733003.
R=rossberg@chromium.org
BUG=v8:2700
LOG=N
Review URL: https://codereview.chromium.org/
1136073002
Cr-Commit-Position: refs/heads/master@{#28340}
danno [Mon, 11 May 2015 11:45:02 +0000 (04:45 -0700)]
Add a MathFloor stub generated with TurboFan
This stub will be used as the basis of a Math.floor-specific CallIC to
detect and track calls to floor that return -0.
Along the way:
- Create a TurboFanCodeStub super class from which the StringLength and
MathRound TF stubs derive.
- Fix the ugly hack that passes the first stub parameter as the "this"
pointer in the the TF-compiled JS function.
- Fix bugs in the ia32/x64 disassembler.
Review URL: https://codereview.chromium.org/
1137703002
Cr-Commit-Position: refs/heads/master@{#28339}
titzer [Mon, 11 May 2015 11:28:50 +0000 (04:28 -0700)]
Add %GetUndetectable() test intrinsic and add tests for undetectables.
R=verwaest@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1132303003
Cr-Commit-Position: refs/heads/master@{#28338}
wingo [Mon, 11 May 2015 11:07:12 +0000 (04:07 -0700)]
Fix more -Werror=sign-compare bugs with GCC 4.9.2
R=svenpanne@chromium.org
LOG=N
BUG=
Review URL: https://codereview.chromium.org/
1128043002
Cr-Commit-Position: refs/heads/master@{#28337}
machenbach [Mon, 11 May 2015 11:03:50 +0000 (04:03 -0700)]
[test] Skip Slimming Paint layout tests.
Also make the expectation lint check happy.
TBR=jochen@chromium.org
NOTRY=true
Review URL: https://codereview.chromium.org/
1137993003
Cr-Commit-Position: refs/heads/master@{#28336}
erikcorry [Mon, 11 May 2015 10:57:55 +0000 (03:57 -0700)]
Postpone counters triggered during GC, and use a HandleScope when calling back.
R=jkummerow@chromium.org,hpayer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1125383007
Cr-Commit-Position: refs/heads/master@{#28335}
machenbach [Mon, 11 May 2015 09:26:37 +0000 (02:26 -0700)]
[test-runner] Improve test execution without tmp files for output.
BUG=chromium:485932
LOG=n
Review URL: https://codereview.chromium.org/
1134703002
Cr-Commit-Position: refs/heads/master@{#28334}
yangguo [Mon, 11 May 2015 08:50:38 +0000 (01:50 -0700)]
Fix build after revert at 28332.
TBR=machenbach@chromium.org
NOTRY=true
NOTREECHECKS=true
Review URL: https://codereview.chromium.org/
1130253006
Cr-Commit-Position: refs/heads/master@{#28333}
machenbach [Mon, 11 May 2015 08:20:02 +0000 (01:20 -0700)]
Revert of Add the concept of a V8 extras exports object (patchset #5 id:80001 of https://codereview.chromium.org/
1128113006/)
Reason for revert:
[Sheriff] Causes gc stress failures:
http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm64%20-%20sim%20-%20gc%20stress/builds/2167
Original issue's description:
> Add the concept of a V8 extras exports object
>
> Exposed to the extras as extrasExports (on the builtins object), on
> which they can put things that should be accessible from C++. Exposed
> to C++ through the V8 API as v8::Context::GetExtrasExportsObject().
>
> Adding a test (in test-api.cc) required adding a simple extra,
> test-extra.js, which we build into the standalone builds.
>
> R=yangguo@chromium.org, jochen@chromium.org
> BUG=
>
> Committed: https://crrev.com/
ad547cea05f3e02c67243b682e933fc53ac763d9
> Cr-Commit-Position: refs/heads/master@{#28317}
TBR=jochen@chromium.org,yangguo@chromium.org,domenic@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1127313005
Cr-Commit-Position: refs/heads/master@{#28332}
yangguo [Mon, 11 May 2015 08:14:54 +0000 (01:14 -0700)]
Call builtin code wrapped in functions from the bootstrapper.
For the moment, we only pass the global object (the one we are setting up).
A few smaller changes were necessary to avoid failures in
test-object-observe/DontLeakContextOnObserve. Otherwise the global object
would be retained by being context allocated, leading to test failure.
R=jkummerow@chromium.org
Review URL: https://codereview.chromium.org/
1132513003
Cr-Commit-Position: refs/heads/master@{#28331}
titzer [Mon, 11 May 2015 08:03:03 +0000 (01:03 -0700)]
[turbofan] Extract frame-states.h from common-operator.h
R=jarin@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1131853002
Cr-Commit-Position: refs/heads/master@{#28330}
v8-autoroll [Mon, 11 May 2015 03:39:29 +0000 (20:39 -0700)]
Update V8 DEPS.
Rolling v8/tools/clang to
5bab78c6ced45a71a8e095a09697ca80492e57e1
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1136043002
Cr-Commit-Position: refs/heads/master@{#28329}
v8-autoroll [Sun, 10 May 2015 03:38:05 +0000 (20:38 -0700)]
Update V8 DEPS.
Rolling v8/tools/clang to
72b418e60c4529a7b160cead01b099de6ea8445f
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1132473006
Cr-Commit-Position: refs/heads/master@{#28328}
jkummerow [Sat, 9 May 2015 10:30:44 +0000 (03:30 -0700)]
Let Runtime_GrowArrayElements accept non-Smi numbers as |key|.
BUG=chromium:485410
LOG=y
R=mvstanton@chromium.org,danno@chromium.org
Review URL: https://codereview.chromium.org/
1132113004
Cr-Commit-Position: refs/heads/master@{#28327}
hpayer [Sat, 9 May 2015 10:29:33 +0000 (03:29 -0700)]
Only double align in PagedSpace::AllocateRawDoubleAligned when allocation succeeded.
BUG=chromium:486003
LOG=n
Review URL: https://codereview.chromium.org/
1136783002
Cr-Commit-Position: refs/heads/master@{#28326}
dehrenberg [Sat, 9 May 2015 01:20:01 +0000 (18:20 -0700)]
Make one copy for all TypedArray methods
This is the first step of converting TypedArrays to the proper
proto chain. There is one copy for each of the Harmony TypedArray
methods, rather than a version for each TypedArray type. This
form prevents accidentally baking in knowledge about a particular
array type into the method definition.
R=adamk@chromium.org, arv@chromium.org, caitpotter88@gmail.com, dslomov@chromium.org, jochen@chromium.org
BUG=v8:4085
LOG=Y
Review URL: https://codereview.chromium.org/
1126313003
Cr-Commit-Position: refs/heads/master@{#28325}
Michael Achenbach [Fri, 8 May 2015 21:32:05 +0000 (23:32 +0200)]
Whitespace CL to test master split.
Cr-Commit-Position: refs/heads/master@{#28324}
titzer [Fri, 8 May 2015 15:44:27 +0000 (08:44 -0700)]
[turbofan] Fix handling of OsrLoopEntry in ControlReducer::ConnectNTL()
R=jarin@chromium.org
LOG=Y
BUG=chromium:485908
Review URL: https://codereview.chromium.org/
1138463004
Cr-Commit-Position: refs/heads/master@{#28323}
erikcorry [Fri, 8 May 2015 15:12:44 +0000 (08:12 -0700)]
If marking deque allocation fails, try to make do with a smaller one
R=hpayer@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1134743002
Cr-Commit-Position: refs/heads/master@{#28322}
adamk [Fri, 8 May 2015 15:03:35 +0000 (08:03 -0700)]
Give ComputeCapacityForSerialization a minimum capacity
This avoids DCHECK failures when passing 0 as the at_least_space_for
argument to HashTableBase::New (allowing converting code from non-serialized
to serialized without changing callsites).
Review URL: https://codereview.chromium.org/
1134573002
Cr-Commit-Position: refs/heads/master@{#28321}
hpayer [Fri, 8 May 2015 14:11:47 +0000 (07:11 -0700)]
Remove explicit double alignment from allocation helper functions.
BUG=
Review URL: https://codereview.chromium.org/
1128323003
Cr-Commit-Position: refs/heads/master@{#28320}
jochen [Fri, 8 May 2015 13:30:55 +0000 (06:30 -0700)]
Add support for on-heap typed arrays to HLoadKeyed::InferRange
BUG=v8:3996
R=bmeurer@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1126413005
Cr-Commit-Position: refs/heads/master@{#28319}
martyn.capewell [Fri, 8 May 2015 12:51:42 +0000 (05:51 -0700)]
[turbofan] Use sbfx in ARM64 instruction selector
Select sbfx for ((x << k) >> k) in ARM64 instruction selector, and similarly
for ubfx. This is a more generic version of the previous sxtb/h selector.
BUG=
Review URL: https://codereview.chromium.org/
1135543002
Cr-Commit-Position: refs/heads/master@{#28318}
domenic [Fri, 8 May 2015 12:06:35 +0000 (05:06 -0700)]
Add the concept of a V8 extras exports object
Exposed to the extras as extrasExports (on the builtins object), on
which they can put things that should be accessible from C++. Exposed
to C++ through the V8 API as v8::Context::GetExtrasExportsObject().
Adding a test (in test-api.cc) required adding a simple extra,
test-extra.js, which we build into the standalone builds.
R=yangguo@chromium.org, jochen@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1128113006
Cr-Commit-Position: refs/heads/master@{#28317}
bmeurer [Fri, 8 May 2015 12:01:46 +0000 (05:01 -0700)]
[clusterfuzz] Length 0 is perfectly fine for BitVector.
BUG=chromium:485952
LOG=n
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1124223006
Cr-Commit-Position: refs/heads/master@{#28316}
dusan.milosavljevic [Fri, 8 May 2015 09:42:42 +0000 (02:42 -0700)]
MIPS [turbofan]: Improve fpu branch assembling for unordered conditions.
TEST=
BUG=
Review URL: https://codereview.chromium.org/
1124023005
Cr-Commit-Position: refs/heads/master@{#28315}
bmeurer [Fri, 8 May 2015 09:34:54 +0000 (02:34 -0700)]
[turbofan] Don't try to push JSToNumber into phis early.
This should not be done during typed lowering, but should be part of a
dedicated truncation/conversion analysis.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1132773002
Cr-Commit-Position: refs/heads/master@{#28314}
hpayer [Fri, 8 May 2015 09:24:29 +0000 (02:24 -0700)]
Initialize sub-array literals first before pointing to it.
BUG=484544
LOG=n
Review URL: https://codereview.chromium.org/
1132763002
Cr-Commit-Position: refs/heads/master@{#28313}
mvstanton [Fri, 8 May 2015 09:13:09 +0000 (02:13 -0700)]
TurboFan: commit dependencies only on update of the opt. code list.
TurboFan compilation was committing dependencies long before the
optimized function made it's way into the optimized code list for
the native context. The problem is that once the code pointer is out
there in dependency arrays, it is eligible for deopt. But the deopt
logic needs the code to be in the optimized code list to fully do it's
job.
BUG=
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1136693003
Cr-Commit-Position: refs/heads/master@{#28312}
domenic [Fri, 8 May 2015 08:44:47 +0000 (01:44 -0700)]
Re-land: Make V8 extras a separate type of native
Instead of making them an extra option that gets passed in and compiled
at the end of the natives file for a given run of js2c, we now make them a
separate run of js2c with a separate natives file output.
This natives file output is then compiled in the bootstrapper. It is not part
of the snapshot (yet), but instead is treated similar to the experimental
natives, just without any of the complexity that comes from tieing the
behavior to flags. We also don't need counterparts to
InitializeExperimentalGlobal and InstallExperimentalNativeFunctions (yet?).
This fixes the issue with https://codereview.chromium.org/
1129743003 by making
the dummy file that is generated for snapshots with no extras (or no experimental
features) nonempty.
R=yangguo@chromium.org, jochen@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1130993003
Cr-Commit-Position: refs/heads/master@{#28311}
bmeurer [Fri, 8 May 2015 08:24:59 +0000 (01:24 -0700)]
[turbofan] Float32Abs and Float64Abs are supported by all backends.
The Float32Abs and Float64Abs operators are supported by all TurboFan
backends, so we no longer need the flags for them.
R=jarin@chromium.org
Review URL: https://codereview.chromium.org/
1132033002
Cr-Commit-Position: refs/heads/master@{#28310}
verwaest [Fri, 8 May 2015 07:28:55 +0000 (00:28 -0700)]
Revert of Allow loading holes from holey smi arrays (patchset #2 id:20001 of https://codereview.chromium.org/
1134483002/)
Reason for revert:
Shouldn't unconditionally load holes since it tanks performance. I'll change it in a follow-up to only do it if the IC ever saw it happen.
Original issue's description:
> Allow loading holes from holey smi arrays
>
> BUG=
>
> Committed: https://crrev.com/
eab5bb5390fab79d063f29398377c6d181963dde
> Cr-Commit-Position: refs/heads/master@{#28298}
TBR=mvstanton@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1131203002
Cr-Commit-Position: refs/heads/master@{#28309}
bmeurer [Fri, 8 May 2015 05:53:44 +0000 (22:53 -0700)]
[turbofan] Don't embed backing store of neutered array buffer.
R=jochen@chromium.org
Review URL: https://codereview.chromium.org/
1127193003
Cr-Commit-Position: refs/heads/master@{#28308}
chunyang.dai [Fri, 8 May 2015 03:57:27 +0000 (20:57 -0700)]
Revert of X87: Resolve references to "this" the same way as normal variables
Reason for revert:
original commit was reverted.
This reverts commit
c3529ce50158546e2f2c25ee888e7e0170a488f1.
original issues's description:
> X87: Resolve references to "this" the same way as normal variables
>
> port
06a792b7cc2db33ffce7244c044a9c05afbb6116 (r28263).
>
> original commit message:
>
> Make the parser handle references to "this" as unresolved variables, so the
> same logic as for the rest of function parameters is used for the receiver.
> Minor additions to the code generation handle copying the receiver to the
> context, along with the rest of the function parameters.
>
> Based on work by Adrian Perez de Castro <aperez@igalia.com>
BUG=
Review URL: https://codereview.chromium.org/
1130913005
Cr-Commit-Position: refs/heads/master@{#28307}
v8-autoroll [Fri, 8 May 2015 03:28:22 +0000 (20:28 -0700)]
Update V8 DEPS.
Rolling v8/buildtools to
b0ede9c89f9d5fbe5387d961ad4c0ec665b6c821
Rolling v8/tools/clang to
bda3bddc01a91a513ec245ed1079b31cf5d7ded3
TBR=machenbach@chromium.org
Review URL: https://codereview.chromium.org/
1124223005
Cr-Commit-Position: refs/heads/master@{#28306}
mbrandy [Thu, 7 May 2015 22:15:38 +0000 (15:15 -0700)]
Revert of PPC: Resolve references to "this" the same way as normal variables (patchset #1 id:1 of https://codereview.chromium.org/
1129803003/)
Reason for revert:
Original commit was reverted.
Original issue's description:
> PPC: Resolve references to "this" the same way as normal variables
>
> Port
18619d355192e2699203d12d9ebb9caea107b693
>
> Original commit message:
> Make the parser handle references to "this" as unresolved variables, so the
> same logic as for the rest of function parameters is used for the receiver.
> Minor additions to the code generation handle copying the receiver to the
> context, along with the rest of the function parameters.
>
> Based on work by Adrian Perez de Castro <aperez@igalia.com>.
>
> R=wingo@igalia.com, dstence@us.ibm.com, michael_dawson@ca.ibm.com
> BUG=
>
> Committed: https://crrev.com/
e0ec097b2f49c4c99a2e9490f4b1a09e02abfd4d
> Cr-Commit-Position: refs/heads/master@{#28272}
TBR=dstence@us.ibm.com,michael_dawson@ca.ibm.com,wingo@igalia.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1132513005
Cr-Commit-Position: refs/heads/master@{#28305}
domenic [Thu, 7 May 2015 18:17:57 +0000 (11:17 -0700)]
Fix the build with snapshot=external
R=jochen@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1137473002
Cr-Commit-Position: refs/heads/master@{#28304}
balazs.kilvady [Thu, 7 May 2015 17:21:23 +0000 (10:21 -0700)]
MIPS: Fix Assembler::dd(Label*) implementation.
BUG=
TEST=mjsunit/asm/embenchen/lua_binarytrees
Review URL: https://codereview.chromium.org/
1127393003
Cr-Commit-Position: refs/heads/master@{#28303}
jochen [Thu, 7 May 2015 15:55:23 +0000 (08:55 -0700)]
Only print elements of a typed array if the buffer wasn't neutered
BUG=chromium:485534
R=yangguo@chromium.org
LOG=n
Review URL: https://codereview.chromium.org/
1130063004
Cr-Commit-Position: refs/heads/master@{#28302}
dehrenberg [Thu, 7 May 2015 15:54:16 +0000 (08:54 -0700)]
TypedArray.prototype.every method
BUG=v8:3578
LOG=Y
R=adamk@chromium.org, dslomov@chromium.org, svenpanne@chromium.org
Review URL: https://codereview.chromium.org/
1128273002
Cr-Commit-Position: refs/heads/master@{#28301}
ulan [Thu, 7 May 2015 14:41:53 +0000 (07:41 -0700)]
Add mode to reduce memory usage in idle notification.
While the mutator is active, the idle time handler optimizes for latency by doing only incremental steps and scavenges.
When the mutator becomes inactive, the idle time handler forces few incremental GCs to reclaim memory and then stops until mutator is active again.
BUG=460090
LOG=N
Review URL: https://codereview.chromium.org/
1105293004
Cr-Commit-Position: refs/heads/master@{#28300}
domenic [Thu, 7 May 2015 14:26:38 +0000 (07:26 -0700)]
Revert of Make V8 extras a separate type of native (patchset #4 id:60001 of https://codereview.chromium.org/
1129743003/)
Reason for revert:
https://build.chromium.org/p/client.v8/builders/V8-Blink%20Linux%2064%20%28dbg%29/builds/2745
Original issue's description:
> Make V8 extras a separate type of native
>
> Instead of making them an extra option that gets passed in and compiled
> at the end of the natives file for a given run of js2c, we now make them a
> separate run of js2c with a separate natives file output.
>
> This natives file output is then compiled in the bootstrapper. It is not part
> of the snapshot (yet), but instead is treated similar to the experimental
> natives, just without any of the complexity that comes from tieing the
> behavior to flags. We also don't add counterparts to
> InitializeExperimentalGlobal and InstallExperimentalNativeFunctions, yet.
>
> R=yangguo@chromium.org, jochen@chromium.org
> BUG=
>
> Committed: https://crrev.com/
c93aff4ac63ad9ffb6318e750335208de32b7902
> Cr-Commit-Position: refs/heads/master@{#28296}
TBR=jochen@chromium.org,yangguo@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=
Review URL: https://codereview.chromium.org/
1131903002
Cr-Commit-Position: refs/heads/master@{#28299}
verwaest [Thu, 7 May 2015 14:01:56 +0000 (07:01 -0700)]
Allow loading holes from holey smi arrays
BUG=
Review URL: https://codereview.chromium.org/
1134483002
Cr-Commit-Position: refs/heads/master@{#28298}
ulan [Thu, 7 May 2015 12:50:59 +0000 (05:50 -0700)]
Do not sample memory histograms in short idle notifications.
BUG=chromium:485472
LOG=NO
TBR=jochen@chromium.org
Review URL: https://codereview.chromium.org/
1128363003
Cr-Commit-Position: refs/heads/master@{#28297}
domenic [Thu, 7 May 2015 12:44:06 +0000 (05:44 -0700)]
Make V8 extras a separate type of native
Instead of making them an extra option that gets passed in and compiled
at the end of the natives file for a given run of js2c, we now make them a
separate run of js2c with a separate natives file output.
This natives file output is then compiled in the bootstrapper. It is not part
of the snapshot (yet), but instead is treated similar to the experimental
natives, just without any of the complexity that comes from tieing the
behavior to flags. We also don't add counterparts to
InitializeExperimentalGlobal and InstallExperimentalNativeFunctions, yet.
R=yangguo@chromium.org, jochen@chromium.org
BUG=
Review URL: https://codereview.chromium.org/
1129743003
Cr-Commit-Position: refs/heads/master@{#28296}
kozyatinskiy [Thu, 7 May 2015 11:49:05 +0000 (04:49 -0700)]
[V8] Reland https://codereview.chromium.org/
1121833003/
Chromium tests were prepared in https://codereview.chromium.org/
1118743003/ and https://codereview.chromium.org/
1135493002/.
Blink tests in https://codereview.chromium.org/
1115193002/
LOG=Y
BUG=chromium:480652
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/
1127363002
Cr-Commit-Position: refs/heads/master@{#28295}