Commit Graph

10242 Commits

Author SHA1 Message Date
verwaest@chromium.org
400b8b8c44 Slight cleanup of UpdateCache code.
Review URL: https://chromiumcodereview.appspot.com/11737033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13328 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 15:35:00 +00:00
svenpanne@chromium.org
eb9a22cc8b Environment bookkeping has linear time complexity now, not a quadratic one.
This reduces the time take for mjsunit/limit-locals from 56.8s to 15.1s in debug
mode and from 12.0s to 1.6s in release mode.

Note that GrowableBitVector and BitVector should really be merged, and probably
have their allocation strategy parmeterized. The current state of affairs
involving tons of checks and delegation is extremely ugly, and it is far from
clear if all that special casing is a clear win. STL FTW! :-P

Review URL: https://codereview.chromium.org/11775016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13327 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 15:28:20 +00:00
yangguo@chromium.org
61f4012989 Use C++ style type casts.
R=mstarzinger@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11644097

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13326 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 15:02:56 +00:00
mstarzinger@chromium.org
0e46919c32 Fix missing exception check in typed array constructor.
The typed array constructor might fail if the first argument is an
object with a length property. Accessing the property can cause an
exception to be thrown and an explicit check needs to be performed.

R=verwaest@chromium.org
BUG=chromium:168545
TEST=mjsunit/regress/regress-crbug-168545.js

Review URL: https://codereview.chromium.org/11777014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13325 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 14:01:04 +00:00
verwaest@chromium.org
11c2557452 Check interceptor before optimizing load/store
Review URL: https://chromiumcodereview.appspot.com/11753027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13324 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 12:14:36 +00:00
danno@chromium.org
a060bf0e9b MIPS: Elide unnecessary context reload in generated stubs.
Port r13290 (6970ca83)

BUG=
TEST=

Review URL: https://codereview.chromium.org/11773013
Patch from Akos Palfi <palfia@homejinni.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13323 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 10:23:30 +00:00
danno@chromium.org
b510dc58d3 MIPS: Re-land Crankshaft-generated KeyedLoad stubs.
Port r13236 (cd9236c5)

BUG=
TEST=

Review URL: https://codereview.chromium.org/11801002
Patch from Akos Palfi <palfia@homejinni.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13322 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 10:18:25 +00:00
danno@chromium.org
ab301a54b3 MIPS: Remove redundant loads in DoCheckMaps Hoist the loop-invariant load out of the loop and call the other CheckMap function
Port r13253 (5af29105)

BUG=
TEST=

Review URL: https://codereview.chromium.org/11784014
Patch from Akos Palfi <palfia@homejinni.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13321 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 10:10:38 +00:00
danno@chromium.org
4246ac3009 Generalize calling to C++ on stub deopt
Remove code specific to KeyedLoadICs in DoCompiledStubFrame on all platforms, driving stub frame translation by the register parameter information found in a stub's CodeStubInterfaceDescriptor.

Review URL: https://codereview.chromium.org/11635015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13320 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 10:06:11 +00:00
ulan@chromium.org
c4bed03342 Add ARM support to tools/grokdump.py
R=svenpanne@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11565014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13319 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 09:48:04 +00:00
ulan@chromium.org
4403daca1a ARM: generate integer zero in a uniform manner.
ARM generated integer zero as either Operand(0, RelocInfo::NONE32), or
Operand(0), or Operand::Zero(). My change makes it use only
Operand::Zero().

This has no functional impact, it's pure cleanup.

R= ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11745030
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13318 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 09:43:12 +00:00
ulan@chromium.org
b73db972e2 Use RelocInfo::IsNone in a few more places.
I had missed some earlier.

Followup to:
https://chromiumcodereview.appspot.com/11695006/

There are now NONE and NONE64 RelocInfo types, but only ARM uses them
both at the same time. They were added in:
https://chromiumcodereview.appspot.com/11191029/

R= ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11742045
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13317 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 09:15:25 +00:00
yangguo@chromium.org
ee563d7875 MIPS: Fix ARM code for DoModI.
Port r13247 (4dda8131)

BUG=
TEST=

Review URL: https://chromiumcodereview.appspot.com/11782009
Patch from Akos Palfi <palfia@homejinni.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13316 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 08:58:52 +00:00
ulan@chromium.org
74026ea46d MIPS: Enable type feedback for branch statement and function call
Port r13288 (5fa2c889)

Original commit message:
This change associates TypeFeedbackIds with ToBoolean stubs in
full-compiled code on ARM, allowing their information to be used in
Crankshaft. This eliminates unnecessary checks, especially in
DoBranch.

BUG=
TEST=

Review URL: https://chromiumcodereview.appspot.com/11801003
Patch from Akos Palfi <palfia@homejinni.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13315 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-07 08:49:52 +00:00
verwaest@chromium.org
a8962e89b3 Use enum instead of bool for force_generic (MISS / MISS_FORCE_GENERIC)
Review URL: https://chromiumcodereview.appspot.com/11737032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13314 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-04 15:37:59 +00:00
yangguo@chromium.org
4ee20d857b Check for read-only-ness when preparing for array sort.
R=verwaest@chromium.org
BUG=v8:2419

Review URL: https://chromiumcodereview.appspot.com/11759022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13313 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-04 15:24:47 +00:00
svenpanne@chromium.org
8a2cca5207 Replaced a bailout ID assertion with quadratic time complexity by a linear one.
This reduces the time to run our test suite in debug mode considerably (from
8:43 to 4:05 on my local workstation using 32 threads). Note that the assertion
is so fast now that it doesn't need to be hidden behind --enable-slow-asserts.
Furthermore, the bookkeeping of the set is not measurable in all our benchmarks,
so I intentionally avoided any #ifdef chaos to keep things simple.

Review URL: https://codereview.chromium.org/11745027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13312 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-04 12:48:18 +00:00
ulan@chromium.org
0606abbaab Rename RelocInfo::NONE to RelocInfo::NONE32.
This CL only does renaming, nothing else.

Followup to:
https://chromiumcodereview.appspot.com/11695006/

There are now NONE and NONE64 RelocInfo types, but only ARM uses them
both at the same time. They were added in:
https://chromiumcodereview.appspot.com/11191029/

R= ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11744020
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13311 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-04 10:56:24 +00:00
yangguo@chromium.org
f96267aa7b Prepare push to trunk. Now working on version 3.16.3.
R=svenpanne@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11748029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13308 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-04 08:55:22 +00:00
ulan@chromium.org
bb96439743 Cleanup RelocInfo::NONE usage.
There are now NONE and NONE64 RelocInfo types, but only ARM uses them
both at the same time. They were added in:
  https://chromiumcodereview.appspot.com/11191029/

I'll rename NONE to NONE32 in a later CL.

This CL cleans up the RelocInfo::NONE usage by:
 - Using RelocInfo::IsNone when testing for NONE-ness.
 - Using NONE on 32-bit platforms (MIPS and IA32), and NONE64 on 64-bit
   platforms (x64).

This cleans up the code and prevents it from evolving bugs in the future
because NONE32 and NONE64 are used in misleading ways.

R= ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11695006
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13307 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 14:20:08 +00:00
yangguo@chromium.org
ffa936927b Fix Win64 build.
R=dcarney@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11746015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13304 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 13:16:00 +00:00
svenpanne@chromium.org
b74f1b8b4a Synched postmortem script with not-so-recent changes.
Review URL: https://codereview.chromium.org/11744019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13303 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 13:11:47 +00:00
yangguo@chromium.org
e2444edd8f Refactor out assumption that one byte strings are ascii in utf8 processing.
R=yangguo@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11725006
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13302 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 12:59:54 +00:00
svenpanne@chromium.org
252fbddc67 Improved printing of HForceRepresentation.
Review URL: https://codereview.chromium.org/11745011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13301 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 10:05:40 +00:00
yangguo@chromium.org
04ccb975f4 Remove InputBuffer
R=yangguo@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11727004
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13298 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 09:18:01 +00:00
yangguo@chromium.org
a50259f75c Add Makefile options to build for the Raspberry Pi (armv7=0, arm_fpu=vfp2).
With these Makefile changes, it is possible to build V8 on the Raspberry Pi with the following make invocation:

make native armv7=false vfp3=off hardfp=on disassembler=on console=readline snapshot=on gdbjit=on strictaliasing=off

R=danno@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11722003
Patch from Luis Reis <luis.m.reis@gmail.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13297 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 08:56:05 +00:00
svenpanne@chromium.org
840b1a37da Fixed printing of LInstructions in the presence of optional arguments
TBR=yangguo@chromium.org

Review URL: https://codereview.chromium.org/11743012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13294 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-03 07:23:33 +00:00
yangguo@chromium.org
ddf70504cb Relax test expectations in BootUpMemoryUse.
R=dcarney@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11738004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13293 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-02 15:00:12 +00:00
verwaest@chromium.org
537d1d89b0 Move CopyElements to the accessor of the target.
Review URL: https://chromiumcodereview.appspot.com/11416238

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13292 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-02 10:09:42 +00:00
yangguo@chromium.org
bccef0c712 Reland r13275 and 13276 (Remove most uses of StringInputBuffer).
R=dcarney@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11727003
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13291 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-31 11:13:50 +00:00
danno@chromium.org
45a012ec2c Elide unnecessary context reload in generated stubs.
Review URL: https://codereview.chromium.org/11550005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13290 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 16:25:38 +00:00
yangguo@chromium.org
e536abb777 Handle non-constant divisor in MathFloorOfDiv, on ia32/x64
Zheng Liu
zheng.z.liu@intel.com

Review URL: https://chromiumcodereview.appspot.com/11624022
Patch from Zheng Liu <zheng.z.liu@intel.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13289 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 15:52:17 +00:00
ulan@chromium.org
eb27eb0351 Enable type feedback for branch statement and function call
This change associates TypeFeedbackIds with ToBoolean stubs in
full-compiled code on ARM, allowing their information to be used in
Crankshaft. This eliminates unnecessary checks, especially in
DoBranch.

Review URL: https://chromiumcodereview.appspot.com/11635046
Patch from Jay Conrod <dconrod@codeaurora.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13288 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 15:14:44 +00:00
danno@chromium.org
3526c6be98 Use EXTERNAL_REFERENCE instead of RUNTIME_ENTRY for Runtime::PerformGC and
CallApiFunctionAndReturn in X64

Review URL: https://codereview.chromium.org/11694008
Patch from Haitao Feng <haitao.feng@intel.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13287 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 14:36:55 +00:00
ulan@chromium.org
a7f9c491dc Use VLDR instead of VMOVs from GPR when a 64-bit double can't be encoded as a VMOV immediate.
This requires constant blinding before it can be enabled. There are other interesting optimizations that can be added later, detailed in a TODO.

BUG=optimization
R=ulan@chromium.org,mstarzinger@chromium.org, hwennborg@google.com

Review URL: https://chromiumcodereview.appspot.com/11191029
Patch from JF Bastien <jfb@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13286 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 13:34:15 +00:00
danno@chromium.org
2e6a7c733b postmortem: export kSmiShiftSize
Right now it's impossible to figure out from headers that smis are left shifted
by 31 bit on x64, and only then tagged.

R=erik.corry@gmail.com

Review URL: https://codereview.chromium.org/11676005
Patch from Fedor Indutny <fedor@indutny.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13285 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 13:06:08 +00:00
danno@chromium.org
17326e8746 Refactor and improve inlined double-aligned allocations
Change is performance neutral but generates smaller code and encapsulates double alignment in the macro-assembler rather than at the allocation site.

Review URL: https://codereview.chromium.org/11684005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13284 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-28 11:09:16 +00:00
ulan@chromium.org
0aee912480 Prepare push to trunk. Now working on version 3.16.2.
R=danno@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11675006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13281 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-27 15:56:02 +00:00
danno@chromium.org
9307f8ea30 Fix SSE2 debug asserts in LayoutTests
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/11675005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13280 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-27 15:40:34 +00:00
yangguo@chromium.org
2f821f1ed9 Revert r13188, r13194, r13256 (Deferred formatting of error stack trace during GC).
R=ulan@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11678006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13279 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-27 13:12:27 +00:00
yangguo@chromium.org
074038058a Add kNumberStringCacheRootIndex into writable_roots
Review URL: https://chromiumcodereview.appspot.com/11660028
Patch from Haitao Feng <haitao.feng@intel.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13278 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-27 10:11:08 +00:00
yangguo@chromium.org
121f3f6020 Revert r13275 and 13276 (Remove most uses of StringInputBuffer).
This is due to test failures in test-mark-compact/BootUpMemoryUse.

R=ulan@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11688003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13277 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-27 10:03:17 +00:00
yangguo@chromium.org
6e6140728b Fix build warnings.
TBR=dcarney@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11669020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13276 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-24 08:52:32 +00:00
yangguo@chromium.org
7f074acd8d Remove most uses of StringInputBuffer
R=yangguo@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11638037
Patch from Dan Carney <dcarney@google.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-24 08:29:48 +00:00
vegorov@chromium.org
6b7fe028e9 Fix typo in the CheckSSE2OperandIsInt32: use bitwise not instead of integer negation.
R=ulan@chromium.org
BUG=

Review URL: https://codereview.chromium.org/11662009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13274 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-21 18:16:27 +00:00
ulan@chromium.org
b64f834383 Fix x64 MathMinMax for negative untagged int32 arguments.
An untagged int32 has zeros in the upper half even if it is negative.
Using cmpq to compare such numbers will incorrectly ignore the sign.

BUG=164442
R=mvstanton@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11665007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-21 17:52:00 +00:00
adamk@chromium.org
d14b05a38d Basic test for interaction of Object.observe and hidden prototypes
The test simply shows the current behavior, not necessarily what we
want the behavior to be.

Review URL: https://codereview.chromium.org/11635033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13272 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-21 17:40:09 +00:00
svenpanne@chromium.org
6ad15b88be Fixed typing confusion found by GCMole
Review URL: https://codereview.chromium.org/11668006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13271 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-21 14:28:33 +00:00
svenpanne@chromium.org
9c7bb6da23 Fixed FloatingPointHelper::CheckSSE2OperandIsInt32.
BUG=v8:2458

Review URL: https://codereview.chromium.org/11660010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13270 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-21 14:01:53 +00:00
svenpanne@chromium.org
c024ba9a90 Fixed x64 disassembler for cmpb_al
Review URL: https://codereview.chromium.org/11666007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13268 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-12-21 13:50:17 +00:00