AuroraMiddleware/v8
1
0
Fork 0
Code Issues 2 Pull Requests Releases Wiki Activity
14,792 Commits 1 Branch 0 Tags 839 MiB
7b7e3658f7
Commit Graph

1125 Commits

Author SHA1 Message Date
dslomov@chromium.org
1a67b7f86a External Array renaming and boilerplate scrapping 
Replaced symbolic names with correct JS name (byte -> int8, unsigned int -> uint32 etc).
Using macros to scrap the boilerplate
BUG=
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/145133013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18835 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-24 16:01:15 +00:00
vegorov@chromium.org
c86da7e5dc Fix HGraphBuilder::AddSimulate to pass removable down correctly 
It was passed as a position argument previously making all simulates added with it conservatively FIXED_SIMULATE.

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/146013002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18822 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-24 12:22:13 +00:00
verwaest@chromium.org
21532ddfdc Reland ArrayPop / ArrayPush. 
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/138443012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18814 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-24 11:47:53 +00:00
jkummerow@chromium.org
ee4e034d70 Revert broken ArrayPop changes 
This reverts:
r18749 "Reland (and fix) "Add hydrogen support for ArrayPop, and remove the handwritten call stubs."",
r18790 "Remove ArrayPush from the custom call generators, and instead call directly to the handler in crankshaft.", and
r18798 "MIPS: Remove ArrayPush from the custom call generators, and instead call directly to the handler in crankshaft."

For causing crashes on Canary.

BUG=chromium:337686
LOG=N
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/146003006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18805 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-24 08:32:50 +00:00
verwaest@chromium.org
6b60546b16 Remove ArrayPush from the custom call generators, and instead call directly to the handler in crankshaft. 
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/137693003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18790 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-23 16:20:25 +00:00
hpayer@chromium.org
83a1df2354 Remove Heap::MaxRegularSpaceAllocationSize and use Page::MaxRegularHeapObjectSize instead. 
BUG=
R=mstarzinger@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/141653016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18776 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-23 13:02:27 +00:00
verwaest@chromium.org
f30330325e Reland (and fix) "Add hydrogen support for ArrayPop, and remove the handwritten call stubs." 
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/144913003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18749 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-22 13:22:58 +00:00
bmeurer@chromium.org
c06d248248 Refactor fast path for empty constant strings in BinaryOp. 
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/143003005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18742 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-22 12:30:00 +00:00
mvstanton@chromium.org
1b3280c491 Revert "Add hydrogen support for ArrayPop, and remove the handwritten call stubs." 
This reverts commit r18709, due to deopt fuzzer issue.
TBR=verwaest@chromium.org

Review URL: https://codereview.chromium.org/143983010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18731 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-22 10:41:23 +00:00
rossberg@chromium.org
0d906a8bdb Zonify types in compiler frontend 
Clean up some zone/isolate handling in AST and its visitors on the way.

(Based on https://codereview.chromium.org/103743004/)

R=jkummerow@chromium.org, titzer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/102563004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18719 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-21 16:22:52 +00:00
verwaest@chromium.org
2d9a4eb355 Add hydrogen support for ArrayPop, and remove the handwritten call stubs. 
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/137783023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18709 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-21 12:42:24 +00:00
bmeurer@chromium.org
4bae658948 Fix invalid type assumption in GenerateNumberToString(). 
We don't have any type feedback in %_NumberToString(), so it's
a bug to assume Type::Number here.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/143893002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18704 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-21 10:35:20 +00:00
mvstanton@chromium.org
04b1baa4c4 We no longer need to recover type cells from the oracle. 
We only need the values within them. Function calls to Array from optimized code needed the cell in the past, but no longer.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/141893002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18682 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-20 09:48:05 +00:00
mvstanton@chromium.org
b2bea46245 Minor bugfix in building inlined Array: bad argument to JSArrayBuilder. 
An HConstant pointing to a Cell rather than an AllocationSite
was passed. The argument wasn't used because of fortuitous
flags. An assert was added to protect the argument.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/141533003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18666 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-17 12:18:57 +00:00
mvstanton@chromium.org
155ef100e9 Fix logic error in assert in IsUndeclaredGlobal() 
Recent changes in IC logic meant that CallStubs no longer use the Contextual bit. IsUndeclaredGlobal() needed to adjust for that.

In fact, now the CL has morphed to remove the notion of storing contextual state in the IC at all, it just becomes some extra ic state of the load ic. This took some adjustment in harmony code to use the global receiver for certain stores.

Now it's clearer that only LoadICs actually record any information about contextual or not.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/140943002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18660 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-17 11:08:24 +00:00
dslomov@chromium.org
5da41be7b8 Implement in-heap backing store for typed arrays. 
This adds a fixed array sub-type that will represent a backing store for
typed arrays allocated with TypedArray(length) construtor.

R=mvstanton@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/101413006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18651 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-16 17:08:45 +00:00
dslomov@chromium.org
34eeeb8953 Revert "Implement in-heap backing store for typed arrays." 
This reverts commit r18649 for breaking Linux/nosnap and Win64 tests.

TBR=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/140793003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18650 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-16 16:00:36 +00:00
dslomov@chromium.org
97040ce67b Implement in-heap backing store for typed arrays. 
This adds a fixed array sub-type that will represent a backing store for
typed arrays allocated with TypedArray(length) construtor.

R=mvstanton@chromium.org, verwaest@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=18646

Review URL: https://codereview.chromium.org/101413006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18649 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-16 15:01:27 +00:00
dslomov@chromium.org
95f572389e Revert "Implement in-heap backing store for typed arrays." 
This reverts commit r18646 for breaking Win32 build.

TBR=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/132233012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18647 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-16 14:26:15 +00:00
dslomov@chromium.org
0c960c2e96 Implement in-heap backing store for typed arrays. 
This adds a fixed array sub-type that will represent a backing store for
typed arrays allocated with TypedArray(length) construtor.

R=mvstanton@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/101413006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18646 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-16 14:18:37 +00:00
jarin@chromium.org
19d832719e This is a preview of a first step towards unification of the hydrogen 
call machinery.  The change replaces CallNamed, CallKeyed,
CallConstantFunction and CallKnownGlobal hydrogen instructions with two
new instructions with a more lower level semantics:

1. CallJSFunction for direct calls of JSFunction objects (no
   argument adaptation)

2. CallWithDescriptor for calls of a given Code object according to
   the supplied calling convention.

Details:

CallJSFunction should be straightforward, the main difference from the
existing InvokeFunction instruction is the absence of argument adaptor
handling. (As a next step, we will replace InvokeFunction with an
equivalent hydrogen code.)

For CallWithDescriptor, the calling conventions are represented by a
tweaked version of CallStubInterfaceDescriptor. In addition to the
parameter-register mapping, we also define parameter-representation
mapping there. The CallWithDescriptor instruction has variable number of
parameters now - this required some simple tweaks in Lithium, which
assumed fixed number of arguments in some places.

The calling conventions used in the calls are initialized in the
CallDescriptors class (code-stubs.h, <arch>/code-stubs-<arch>.cc), and
they live in a new table in the Isolate class. I should say I am not
quite sure about Representation::Integer32() representation for some of
the params of ArgumentAdaptorCall - it is not clear to me wether the
params could not end up on the stack and thus confuse the GC.

The change also includes an earlier small change to argument adaptor
(https://codereview.chromium.org/98463007) that avoids passing a naked
pointer to the code entry as a parameter. I am sorry for packaging that
with an already biggish change.

Performance implications:

Locally, I see a small regression (.2% or so). It is hard to say where
exactly it comes from, but I do see inefficient call sequences to the
adaptor trampoline. For example:

;;; <@78,#24> constant-t
bf85aa515a     mov edi,0x5a51aa85          ;; debug: position 29
;;; <@72,#53> load-named-field
8b7717         mov esi,[edi+0x17]          ;; debug: position 195
;;; <@80,#51> constant-s
b902000000     mov ecx,0x2                 ;; debug: position 195
;;; <@81,#51> gap
894df0         mov [ebp+0xf0],ecx
;;; <@82,#103> constant-i
bb01000000     mov ebx,0x1
;;; <@84,#102> constant-i
b902000000     mov ecx,0x2
;;; <@85,#102> gap
89d8           mov eax,ebx
89cb           mov ebx,ecx
8b4df0         mov ecx,[ebp+0xf0]
;;; <@86,#58> call-with-descriptor
e8ef57fcff     call ArgumentsAdaptorTrampoline  (0x2d80e6e0)    ;; code: BUILTIN

Note the silly handling of ecx; the hydrogen for this code is:

0 4 s27 Constant 1  range:1_1 <|@
0 3 t30 Constant 0x5bc1aa85 <JS Function xyz (SharedFunctionInfo 0x5bc1a919)> type:object <|@
0 1 t36 LoadNamedField t30.[in-object]@24 <|@
0 1 t38 Constant 0x2300e6a1 <Code> <|@
0 1 i102 Constant 2  range:2_2 <|@
0 1 i103 Constant 1  range:1_1 <|@
0 2 t41 CallWithDescriptor t38 t30 t36 s27 i103 i102 #2 changes[*] <|@

BUG=
R=verwaest@chromium.org, danno@chromium.org

Review URL: https://codereview.chromium.org/104663004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18626 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-15 17:00:35 +00:00
jkummerow@chromium.org
5298a078c6 Revert "Eliminatable CheckMaps replaced with if(true) or if(false)." 
This reverts r18592 for breaking the GC stress bots.

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/137783011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18611 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-15 09:53:54 +00:00
ishell@chromium.org
6b389902e1 Revert "Polymorphic named calls optimized for the case of repetitive call targets." for now because it prevents better results for checkmap elimination. 
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/138453002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18597 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-14 16:37:09 +00:00
verwaest@chromium.org
72125bafcc Remove HCallGlobal and merge uses with HCallNamed. 
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/134333007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18595 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-14 16:15:52 +00:00
ishell@chromium.org
952fbe475f Eliminatable CheckMaps replaced with if(true) or if(false). 
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/130613003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18592 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-14 16:06:40 +00:00
verwaest@chromium.org
1257ba358c Remove CALL_AS_FUNCTION and CALL_AS_METHOD. 
BUG=
R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/136403005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18590 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-14 14:36:24 +00:00
verwaest@chromium.org
f2245a9cf9 Make the strict-mode calling convention for contextual calls the default one. 
BUG=
R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/131663003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18581 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-14 12:04:10 +00:00
hpayer@chromium.org
f01dc250dc Handlify AllocationSite AddDependentCompilationInfo. 
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/136703002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18570 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-13 14:54:33 +00:00
mvstanton@chromium.org
854d085996 More efficient use of space in AllocationSite. 
We can eliminate one word from the object by sharing the pretenuring decision and the found memento count.

R=bmeurer@chromium.org, hpayer@chromium.org

Review URL: https://codereview.chromium.org/132063004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18554 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-13 10:28:01 +00:00
ishell@chromium.org
5c65572dd2 "dead" and "osr" block flags added to trace output for unreachable and osr entry blocks respectively. 
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/135943004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18550 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-13 08:45:26 +00:00
jarin@chromium.org
c0f622a45b Fix of Hydrogen environment building for function "apply" calls. 
BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/133773002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18548 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-11 13:59:04 +00:00
rossberg@chromium.org
cb28b7f837 Retry "Templatise type representation" after making clang happy 
The only thing different now is line types.h:208/236, which had a static_cast<Type*> before.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/133683002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18533 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-10 12:19:01 +00:00
hpayer@chromium.org
4ef951cf43 Allocation site pretenuring. 
Pretenuring decisions are made based on allocation site lifetime statistics.

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

Review URL: https://codereview.chromium.org/96783002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18532 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-10 12:11:54 +00:00
machenbach@chromium.org
8ffcd2a281 [Sheriff] Revert "Templatise type representation" and "Fix Mac warnings". 
This reverts commit r18521 and r18522 for breaking mac and win builders.

BUG=
TBR=rossberg@chromium.org

Review URL: https://codereview.chromium.org/132493002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18524 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-09 19:52:15 +00:00
rossberg@chromium.org
1c33a2d840 Templatise type representation 
This is to support both heap- and zone-allocated types in the future (the latter not yet implemented).

Also, handlify the type API some more.

R=titzer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/107933005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18521 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-09 17:12:48 +00:00
jkummerow@chromium.org
839297487f Skip back edge creation for "do ... while(false)" loops 
Drive-by fix: Fix disassembling "prefetch" instruction on ia32

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/131733002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18519 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-09 15:20:25 +00:00
mvstanton@chromium.org
380aa6dc8d Array constructor can be simplified by loading context from JSFunction. 
BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/128683002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18518 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-09 15:07:23 +00:00
mvstanton@chromium.org
7baadbfac1 FunctionLiteral has work to do in the typing phase. 
In crankshaft, we searched un-optimized code for a SharedFunctionInfo
that matches the FunctionLiteral we are processing. Ideally, this
work should be done in the typing phase.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/104883006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18508 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-09 09:00:19 +00:00
ishell@chromium.org
b2616f415b Polymorphic named calls optimized for the case of repetitive call targets. 
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/118483003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18476 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-07 16:59:42 +00:00
verwaest@chromium.org
4615e9edac Reland v8:18458 "Load the global proxy from the context of the target function." 
BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/104013008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18462 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-07 10:46:39 +00:00
rossberg@chromium.org
2879f2104c Revert "Load the global proxy from the context of the target function." 
This reverts commit https://code.google.com/p/v8/source/detail?r=18458, since it exhibits a bug that breaks some tests.

TBR=verwaest@chromium.org
BUG=

Review URL: https://codereview.chromium.org/93863006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18461 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-07 09:55:25 +00:00
verwaest@chromium.org
5b40c38679 Load the global proxy from the context of the target function. 
BUG=
R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/111613003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18458 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-07 08:21:17 +00:00
bmeurer@chromium.org
ad4231c495 Reland "Allocation site support for monomorphic StringAdds in BinaryOps". 
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/106313003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18444 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-02 15:31:27 +00:00
bmeurer@chromium.org
4d88b0b4be Revert "Fix compilation with C++11." and "Allocation site support for monomorphic StringAdds in BinaryOps.". 
This reverts commit r18431 and r18432 for breaking
the Linux nosnapshot build.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/122463004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18434 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-02 10:38:45 +00:00
bmeurer@chromium.org
a91499fec7 Allocation site support for monomorphic StringAdds in BinaryOps. 
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/106453003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18431 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-02 09:17:52 +00:00
bmeurer@chromium.org
a53dcc9da1 Make it possible to assert that certain instance types are compatible wrt HAllocate. 
Turns clear_next_map_word_ into a flag and adds a new method
CompatibleInstanceTypes(). This is used in BuildUncheckedStringAdd()
to ASSERT that there's no difference wrt. HAllocate in using
CONS_STRING_TYPE vs CONS_ASCII_STRING_TYPE.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/110443003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18428 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2014-01-02 06:26:44 +00:00
jkummerow@chromium.org
3c76ecd732 Fix switch statements with non-Smi integer labels and no type feedback 
BUG=chromium:329709
LOG=Y
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/98643010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18370 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-19 14:25:58 +00:00
ishell@chromium.org
e3752b63c0 HStoreKeyed for Smis optimized for x64 + related redundant moves of elements removed 
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/108503004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18362 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-18 18:05:10 +00:00
ishell@chromium.org
8c159ed4ec HStoreNamedField for Smis optimized for x64 
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/108413003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18360 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-18 17:53:50 +00:00
danno@chromium.org
8683920e3f Improve ==/=== comparison in Hydrogen for monomorphic maps 
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/108083004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18357 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-18 17:18:23 +00:00
jkummerow@chromium.org
59c334ebf4 Hydrogen: Re-use regular comparisons infrastructure for switch statements 
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/111573003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18348 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-18 11:58:58 +00:00
verwaest@chromium.org
fb7218dc3d Enable optimization of functions with generic switches. 
R=jkummerow@chromium.org, titzer@chromium.org

Review URL: https://chromiumcodereview.appspot.com/110123002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18347 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-18 11:44:38 +00:00
svenpanne@chromium.org
84aa5263f3 Remove the last remnants of the TranscendentalCache. 
It was only used for Math.log, and even then only in full code and in %_MathLog. For crankshafted code, Intel already used the FP operations directly, while the ARM/MIPS ports were a bit lazy and simply called the stub. The latter directly call the C library now without any cache. It would be possible to directly generate machine code if somebody has the time, from what I've seen out in the wild it should be only about a dozen instructions.

LOG=y
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/113343003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18344 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-18 10:40:26 +00:00
yurys@chromium.org
cd5ea74700 Replace 'operator*' with explicit 'get' method on SmartPointer 
Made operator* return reference to the raw type, not pointer. New method 'get()' should be used when raw pointer is needed.

Also removed useless inline modifier from the SmaprtPointer methods and added const modifier to the methods that don't change smart pointer.

Made ~SmartPointerBase protected to avoid accidental calls of the non-virtual base class's destructor.

drive-by: fixed use after free in src/factory.cc

BUG=None
LOG=N
R=alph@chromium.org, svenpanne@chromium.org

Review URL: https://codereview.chromium.org/101763003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-09 07:41:20 +00:00
mvstanton@chromium.org
a1a871fb5a Remove unnecessary overflow check in HGraphBuilder::BuildCreateAllocationMemento(). 
With this fix codegen looks like:

mov ecx,[eax+0xf]  ;;; <@52,#38> load-named-field
add ecx,0x2  ;;; <@54,#40> add-i
mov [eax+0xf],ecx  ;;; <@56,#41> store-named-field

without it there is an overflow check and jump to deopt.

x64 code looks similar, except there is an (annoying) smi-untag then int32-to-smi around the add operation.

R=bmeurer@chromium.org, hpayer@chromium.org

Review URL: https://codereview.chromium.org/104313003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18255 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-04 10:46:18 +00:00
bmeurer@chromium.org
69b5d5f8f6 Revert "Use constant types to represent the fixed right arg of a MOD." 
This reverts commit r18246 for tanking all benchmarks.

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/104243002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18249 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-04 09:27:48 +00:00
bmeurer@chromium.org
1868ef5de9 Use constant types to represent the fixed right arg of a MOD. 
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/103933002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18246 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-04 08:37:34 +00:00
bmeurer@chromium.org
7e540e6b8a Revert "Use constant types to represent the fixed right arg of a MOD." 
This reverts commit r18228 for crashing on Windows.

TBR=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/101663002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18229 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-03 11:24:56 +00:00
bmeurer@chromium.org
5c6b9500aa Use constant types to represent the fixed right arg of a MOD. 
R=mvstanton@chromium.org, svenpanne@chromium.org

Review URL: https://codereview.chromium.org/99173002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18228 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-03 10:51:57 +00:00
bmeurer@chromium.org
aa83f2900a Fix invalid assertion with OSR in BuildBinaryOperation. 
BUG=v8:3032
LOG=n
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/98623004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18190 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-02 13:12:07 +00:00
bmeurer@chromium.org
a3d1df29f1 Fix HInnerAllocatedObject to use an HValue for the offset. 
R=hpayer@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/98673003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18181 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-12-02 11:24:31 +00:00
jkummerow@chromium.org
833ee3cdb7 A few small refactorings and typo fixes 
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/95333002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18153 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-29 10:02:32 +00:00
mstarzinger@chromium.org
b506e8ce96 Follow-up fix for r18135 in TryCallApply. 
R=yangguo@chromium.org
BUG=chromium:323942
TEST=mjsunit/regress/regress-crbug-323942
LOG=N

Review URL: https://codereview.chromium.org/95583002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18152 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-29 09:59:35 +00:00
titzer@chromium.org
6e5e23d630 Implement global check elimination using the HFlowEngine. 
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/46883009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18137 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-28 15:50:54 +00:00
yangguo@chromium.org
f235194518 Fix bug in inlining Function.apply. 
R=jkummerow@chromium.org
BUG=323942
LOG=Y

Review URL: https://codereview.chromium.org/95123003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18135 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-28 15:30:17 +00:00
hpayer@chromium.org
0e95ffddf2 Fix allocation memento bookeeping code for js arrays. 
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/95053002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18124 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-28 13:45:23 +00:00
hpayer@chromium.org
0213fc9035 Added tracing support for pretenuring. 
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/94493002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18117 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-28 10:49:02 +00:00
mvstanton@chromium.org
f3a22f965e The goal is to discover the appropriate heap space for objects created in full 
code. By the time we optimize the code, we'll be able to decide on new or old
space based on the number of surviving objects after one or more gcs.

The mechanism is a "memento" placed behind objects in the heap. It's currently
done for array and object literals, with plans to use mementos for constructed
objects as well (in a later CL).

The feature is behind the flag allocation_site_pretenuring, currently off.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/40063002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18104 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-27 14:03:40 +00:00
dslomov@chromium.org
c3a4d718ce Generate TypedArrayInitialize builtin in hydrogen. 
R=danno@chromium.org

Review URL: https://codereview.chromium.org/59023003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18059 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-25 14:41:46 +00:00
mvstanton@chromium.org
81b22bbf96 A performance regression in array literal creation was caused by refactoring that eliminated a special fast case for shallow arrays. At the same time the general case got a bit slower. This CL restores most of the performance without coding the special fast case. The virtual dispatching is unnecessary because we know what we want to do at compile time. A flag was added to Runtime::CreateArrayLiteral. The flags delivers information about shallowness but also whether or not allocation mementos should be created. This is useful for crankshafted code. 
BUG=v8:3008
LOG=Y
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/77293003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18046 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-25 12:41:27 +00:00
dslomov@chromium.org
21dca31203 Safe HGraphBuilder::Add<> and New<> 
R=danno@chromium.org

Review URL: https://codereview.chromium.org/82943005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-22 19:05:21 +00:00
bmeurer@chromium.org
7158b51125 Fix combined string length computation. 
For 31-bit smis, we don't need to explicitly check whether the
combined string length exceeds the max supported string length,
since the value must be converted to a smi at some point (i.e.
when it is stored into the string length field of the resulting
string), which will emit an overflow check.

For 32-bit smis, we insert an explicit check that the combined
string length does not exceed String::kMaxLength.

This also enables to get rid of the JoinContinuation() usage in
BuildUncheckedStringAdd().

BUG=v8:2990
LOG=n
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/82733003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18005 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-22 12:03:01 +00:00
bmeurer@chromium.org
08df395f4e Get rid of the binop_stub parameter to BuildBinaryOperation(). 
Just ask the graph builder whether we are compiling a stub.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/83073003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18004 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-22 11:49:04 +00:00
svenpanne@chromium.org
b6b84c02b2 Reland "Implement Math.random() purely in JavaScript" plus fixes. 
The main change is that a bit has been added to array buffers to
signal that the backing store has to be freed when the buffer dies.

BUG=316359
LOG=Y
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/82763005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18003 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-22 11:35:39 +00:00
bmeurer@chromium.org
da87c188ad Rework BuildCheckString() to be compatible with the other BuildCheck*() methods. 
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/79343009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17985 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-22 07:27:26 +00:00
danno@chromium.org
8e266c2244 Revert 17963, 17962 and 17955: Random number generator in JS changes 
Revert 17966, 17965 also as collateral damage: Embed trigonometric lookup table.

Due to Heapcheck and valgrind failures that are not yet fixed.

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/80513004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17981 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-21 17:21:00 +00:00
svenpanne@chromium.org
2b1da67263 Implement Math.random() purely in JavaScript. 
This removes tons of architecture-specific code and makes it easy to
experiment with other pseudo-RNG algorithms. The crankshafted code is
extremely good, keeping all things unboxed and doing only minimal
checks, so it is basically equivalent to the handwritten code.

When benchmarks are run without parallel recompilation, we get a few
percent regression on SunSpider's string-validate-input and
string-base64, but these benchmarks run so fast that the overall
SunSpider score is hardly affected and within the usual jitter. Note
that these benchmarks actually run even faster when we don't
crankshaft at all on the main thread (the regression is not caused by
bad code, it is caused by Crankshaft needing a few hundred microsecond
for compilation of a trivial function). Luckily, when parallel
recompilation is enabled, i.e. in the browser, we see no regression at
all!

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/68723002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17955 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-21 09:55:15 +00:00
bmeurer@chromium.org
d38abc3bba Remove the first_right_arg hack for HMod. 
Generate a proper subgraph in BuildBinaryOperation instead.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/77053003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17919 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-20 12:43:33 +00:00
bmeurer@chromium.org
4c4d911cde Convert constant numbers at compile time. 
TEST=mjsunit/number-tostring
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/77553005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17917 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-20 12:35:21 +00:00
bmeurer@chromium.org
defee14d76 Make BuildBinaryOperation use ReturnValue instead of ReturnInstruction. 
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/78183002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17914 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-20 12:00:57 +00:00
bmeurer@chromium.org
6826b74708 Replace unsafe uses of Add<> with AddUncasted<>. 
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/78213002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17913 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-20 11:57:49 +00:00
bmeurer@chromium.org
da6d5586f9 Don't generate useless string checks for string adds. 
If we know that one side of a string add is definitely a string
(i.e. if it's a string constant), then we don't need to emit a
string check for the argument.

This adds a new BuildCheckString() method to the graph builder,
which does "the right thing".

TEST=mjsunit/string-add
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/78063002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17912 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-20 11:53:09 +00:00
danno@chromium.org
06c7620302 Fixed crashes exposed though fuzzing. 
The %_OneByteSeqStringSetChar intrinsic expects its arguments to be checked before being called for efficiency reasons, but the fuzzer provided no such checks. Now the intrinsic is robust to bad input if FLAG_debug_code is set.

R=yangguo@chromium.org
TEST=test/mjsunit/regress/regress-320948.js
BUG=chromium:320948
LOG=Y

Review URL: https://codereview.chromium.org/72813004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17886 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-19 16:41:07 +00:00
yangguo@chromium.org
e2563d7a8e Make number of available threads isolate-dependent and expose it to ResourceConstraints. 
R=svenpanne@chromium.org
BUG=v8:2991
LOG=Y

Review URL: https://codereview.chromium.org/68203029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17866 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-19 11:52:47 +00:00
ishell@chromium.org
53bd540ee0 Constant-folding through HForceRepresentation fix. 
Reverts changes in HValue::IsInteger32Constant() made in https://code.google.com/p/v8/source/detail?r=17787

R=bmeurer@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/68493005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17863 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-19 11:41:04 +00:00
vegorov@chromium.org
0ec0d5fc32 Reapply r11765 
Add static_cast<int> in HPositionInfo::position to please Win64 build.

Improvements in positions handling in optimizing compiler.

- When building binary arithmetic op and comparison restore source position of the operation itself before building operation itself after it was changed by building operands. This ensures that position recorded for operation points to the operation token instead of pointing to the rightmost operand;

- Add support for recording operands' positions and use these positions when inserting HChange instructions;

- When generating hydrogen.cfg emit H-instruction position as BCI (previously 0 was emitted), additionally on every lithium instruction emit annotation pointing to corresponding hydrogen-instruction. This allows to easily reach from deopt_id to lithium instruction and from it to hydrogen instruction and source position.

R=danno@chromium.org

Review URL: https://codereview.chromium.org/59703011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17848 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-18 17:24:00 +00:00
dslomov@chromium.org
99133912bd Generate DataViewInitialize built-in in hydrogen. 
R=bmeurer@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/66843011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17831 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-18 13:57:49 +00:00
danno@chromium.org
4ed2538262 Generate KeyedLoadDictionaryElementStub with Hydrogen 
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/19492007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17804 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-15 17:53:35 +00:00
mvstanton@chromium.org
5b9569b574 Regression fix: HForceRepresentation shouldn't be an idef. 
Instead, code sites that are interested in underlying constant integer values
should use HValue::IsInteger32Constant(). The issue is that the infer representation phase shouldn't "see through" HForceRepresentation nodes to an underlying, and less specific representation.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/65643003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17787 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-15 12:10:59 +00:00
verwaest@chromium.org
341d405301 Reland and fix "Add support for keyed-call on arrays of fast elements" 
BUG=
R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/71783003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17782 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-15 10:52:05 +00:00
danno@chromium.org
d62337e127 Add ability to do "else-if" clauses in IfBuilder 
- In an Else block it's possible to add more If<>'s in the same builder that are
  treated as an else if.
- Simplified and cleaned-up some of the IfBuilder's internals.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/66983002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17780 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-15 10:36:02 +00:00
bmeurer@chromium.org
2ee5aa951c Fix missing type feedback check for Generic*String addition. 
TEST=mjsunit/regress/regress-crbug-318671
BUG=318671
LOG=y
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/67473007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17772 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-15 09:13:36 +00:00
rafaelw@chromium.org
25b450b2b7 Revert "Improvements in positions handling in optimizing compiler." (r17765) 
Original issue: https://codereview.chromium.org/49203002/

TBR=vegorov

Review URL: https://codereview.chromium.org/63343003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17768 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-14 21:45:01 +00:00
vegorov@chromium.org
9d440ac4c0 Improvements in positions handling in optimizing compiler. 
- When building binary arithmetic op and comparison restore source position of the operation itself before building operation itself after it was changed by building operands. This ensures that position recorded for operation points to the operation token instead of pointing to the rightmost operand;

- Add support for recording operands' positions and use these positions when inserting HChange instructions;

- When generating hydrogen.cfg emit H-instruction position as BCI (previously 0 was emitted), additionally on every lithium instruction emit annotation pointing to corresponding hydrogen-instruction. This allows to easily reach from deopt_id to lithium instruction and from it to hydrogen instruction and source position.

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/49203002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17765 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-14 21:00:27 +00:00
machenbach@chromium.org
eef8694a7e [Sheriff] Revert "Add support for keyed-call on arrays of fast elements" 
This reverts commit r17746 for breaking layout tests.

TBR=verwaest@chromium.org
BUG=

Review URL: https://codereview.chromium.org/72753002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17751 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-14 15:00:13 +00:00
verwaest@chromium.org
607a175cbc Add support for keyed-call on arrays of fast elements 
R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/23537067

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17746 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-14 13:46:18 +00:00
mvstanton@chromium.org
3cf157b43b Inline zero argument array constructor. 
patch from issue 54583003 (dependent code).

Zero arguments - very easy

1 argument - three special cases:
  a) If length is a constant in valid array length range,
     no need to check it at runtime.
  b) respect DoNotInline feedback on the AllocationSite for
     cases that the argument is not a smi or is an integer
     with a length that should create a dictionary.
  c) if kind feedback is non-holey, and length is non-constant,
     we'd have to generate a lot of code to be correct.
     Don't inline this case.

N arguments - one special case:
  a) If a deopt ever occurs because an input argument isn't
     compatible with the elements kind, then set the
     DoNotInline flag.

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/55933002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17741 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-14 12:05:09 +00:00
mvstanton@chromium.org
026c5507bd Make HForceRepresentation an idef. 
An optimization in the array constructor wasn't firing because the code
would check for constant values, failing to find them through
intervening HForceRepresentation instructions. Repaired this site. A
follow-on CL will provide a better way of making decisions based on
constant HValues.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/59623009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17708 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-13 17:03:11 +00:00
yangguo@chromium.org
e83fd01ce6 Reland "Implement Math.sin, cos and tan using table lookup and spline interpolation." 
This relands r17594 with necessary fixes.

R=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/70003004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17654 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-12 14:43:18 +00:00
yangguo@chromium.org
df9665032e Introduce %_IsMinusZero. 
R=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/63423004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17639 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
 2013-11-12 11:53:13 +00:00