Commit Graph

5594 Commits

Author SHA1 Message Date
ricow@chromium.org
442cb7c21a Fix compile failure with GCC 4.5.0 (issue 1244)
GCC apparently assumes that given that we do a static cast in the next
line the value must actually be in StrictModeFlag enum range (even
though this is actually what we are asserting)


Review URL: http://codereview.chromium.org/6670031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7147 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 11:02:27 +00:00
danno@chromium.org
4beeee3eb3 Fix x64 array length operations in Crankshaft
x64-bit array length loads and bounds checks should be 32-bits

Review URL: http://codereview.chromium.org/6676009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7146 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 10:30:33 +00:00
ricow@chromium.org
c00631b86e Fix presubmit by deleting regress-1240 not deleted by last patch.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7145 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 09:08:52 +00:00
ricow@chromium.org
f2730d2ab8 Revert revision 7143, this causes a number of webkit tests to fail.
This includes a security test. Reverting to investigate further.

Review URL: http://codereview.chromium.org/6673019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7144 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 09:02:54 +00:00
ricow@chromium.org
fa9e57e326 Change __defineGetter__ and __defineSetter__ to respect non-configurable.
This makes us compatible with firefox. Earlier on we were somehow
compatible with safari - which will allow defining a getter even when
an existing getter is present and non-configurable. We would, however,
in addition to overwriting the getter also change configurable to
true. The approach used by firefox seems much more sound, i.e., why
should it be possible to use __defineGetter__ or __defineSetter__ to
overwrite a non-configurable getter or setter respectively.

I will file a bug on the webkit bugtracker. 

Review URL: http://codereview.chromium.org/6658037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7143 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 08:05:59 +00:00
vitalyr@chromium.org
c17aeb55bb Use pthread_kill on solaris.
Landing patch by Ryan Dahl.

TBR=vegorov@chromium.org

Original review: http://codereview.chromium.org/6667030/

Review URL: http://codereview.chromium.org/6667031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7142 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 07:41:45 +00:00
vitalyr@chromium.org
d3caca74ca Sample correct registers on solaris.
Landing patch by Ryan Dahl.

TBR=vegorov@chromium.org

Original review: http://codereview.chromium.org/6676019/

Review URL: http://codereview.chromium.org/6665023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7141 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 07:34:55 +00:00
vitalyr@chromium.org
7b5b511d1a Update platform-solaris to support runtime profiling.
Landing patch by Ryan Dahl.

TBR=vegorov@chromium.org

Original review: http://codereview.chromium.org/6674002/

Review URL: http://codereview.chromium.org/6667025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7140 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-11 01:43:44 +00:00
mikhail.naganov@gmail.com
e0829eee10 Fix bit field for GlobalHandles::Node State in MSVC.
TBR=vitalyr@chromium.org
BUG=1243
TEST=none

Review URL: http://codereview.chromium.org/6658040

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7136 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 17:29:37 +00:00
whesse@chromium.org
b7d7aa8ad2 Fix error in sin-cos.js test introduced in r7129.
Review URL: http://codereview.chromium.org/6659034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7133 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 14:04:18 +00:00
sgjesse@chromium.org
bc9ddf20bb ARM: Port r7089 to ARM
Ensure that there is always enough bytes between consequtive calls in optimized code to write a call instruction at the return points without overlapping.

Add a call to deoptimize all functions after running tests with --stress-opt. This will catch some issues with functions which cannot be forcefully deoptimized. Some of the tests failed on ARM with that change without the rest of the changes in this change.
Review URL: http://codereview.chromium.org/6661022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7132 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 13:58:20 +00:00
mikhail.naganov@gmail.com
1d45870320 Fix for win32 after r7125. For some reason introducing a bit field into GlobalHandles::Node causes a trouble.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7131 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 13:45:35 +00:00
whesse@chromium.org
7175dc09ef Crankshaft: Fix error in range analysis of compare operations.
Review URL: http://codereview.chromium.org/6658035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7130 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 13:34:56 +00:00
whesse@chromium.org
fc8f77e398 X64 Crankshaft: Fix error in computation of sine and cosine.
Review URL: http://codereview.chromium.org/6646047

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7129 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 13:34:23 +00:00
fschneider@chromium.org
48832819b7 Normalize statistics about compilation time and allocation size.
Now we print compilation time and zone allocation per KB of compiled
source code to better compare with previous revisions.

Review URL: http://codereview.chromium.org/6646015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7128 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 13:26:51 +00:00
mikhail.naganov@gmail.com
a7bfc32829 Fix x64 build after r7125
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7127 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 12:58:57 +00:00
mikhail.naganov@gmail.com
ada15030e4 A follow-up to r7125: fix a couple of remaining comments.
Review URL: http://codereview.chromium.org/6626043/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7126 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 12:22:59 +00:00
mikhail.naganov@gmail.com
dee9c00267 Add an interface for an embedder to provide information about native
objects retained by object groups and global handles.

This information is then used during heap snapshot generation
to provide a more complete memory picture.

This patch will be needed to fix https://bugs.webkit.org/show_bug.cgi?id=53659.

Review URL: http://codereview.chromium.org/6626043

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7125 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 12:05:31 +00:00
mikhail.naganov@gmail.com
943e4f8340 Remove `sfi' abbreviations (a follow-up to r6902).
BUG=none
TEST=none

Review URL: http://codereview.chromium.org/6646022/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7124 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 12:00:27 +00:00
mikhail.naganov@gmail.com
6eb8c9e5a0 Revert r7119..r7121
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7123 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 11:55:03 +00:00
fschneider@chromium.org
241e4d364e Initialize zone lists in the register allocator with more reasonable initial capacities.
Also replace fixed length zone lists with embedded vectors for the fixed live ranges.

Review URL: http://codereview.chromium.org/6660023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7122 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 11:54:55 +00:00
mikhail.naganov@gmail.com
ae38aedae4 lint
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7121 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 11:49:17 +00:00
mikhail.naganov@gmail.com
47a5723704 comment
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7120 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 11:49:07 +00:00
mikhail.naganov@gmail.com
c6d2bb18aa renamed
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7119 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 11:48:56 +00:00
lrn@chromium.org
d57a8b0a07 X64: Change kSmiConstantRegister to r12.
We never use the smi register in places where it matters that r12 as
a base register needs an extra SIB byte. Might as well free up the
unencumbered r15 register instead.

Review URL: http://codereview.chromium.org/6647015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7113 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 10:14:24 +00:00
danno@chromium.org
34c399d205 Prepare push to trunk. Now working on version 3.2.2.
Review URL: http://codereview.chromium.org/6659030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7112 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-10 09:43:58 +00:00
danno@chromium.org
7b29810d97 Add bailout to x64 pixel array lithium instructions.
Stops segfaults in tests while tracking down the problem.

TBR=kmillikin@chromium.org

Review URL: http://codereview.chromium.org/6647024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7111 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 17:14:12 +00:00
vegorov@chromium.org
83616da7f5 Rebuild scope chain from serialized scope info before parsing lazily.
We used to rebuild it in Scope::Analyze but this might lead to mismatch between scopes seen by parser during the first and the second parse of the function.

BUG=v8:1230

Review URL: http://codereview.chromium.org/6646017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7110 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 16:57:03 +00:00
whesse@chromium.org
5dcd32beda X64 Crankshaft: Add inline one-element cache for Instanceof.
Review URL: http://codereview.chromium.org/6621071

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7109 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 15:57:47 +00:00
fschneider@chromium.org
012db51660 Reuse backing store of two zone lists by using Rewind(0) instead of Clear().
Review URL: http://codereview.chromium.org/6647019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7108 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 15:50:24 +00:00
whesse@chromium.org
ca20d718b3 X64 Crankshaft: Fix unrestored context register in Math.power.
Review URL: http://codereview.chromium.org/6649020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7107 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 15:27:07 +00:00
danno@chromium.org
9cca7d2733 Combine typed and pixel arrays.
Worth mentioning:
- Specialized versions of pixel array and store/loads inside the generic stubs have been removed, since to have parity for all external arrays, 8 different versions would have to be inlined/checked.
- There's a new constant in v8.h for external arrays with pixel array elements.

Review URL: http://codereview.chromium.org/6546036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7106 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 15:01:16 +00:00
kmillikin@chromium.org
a05a492254 Remove class HSubgraph.
Review URL: http://codereview.chromium.org/6647018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7100 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 12:37:56 +00:00
kmillikin@chromium.org
d31594feae Refactor construction of switch statements to avoid subgraphs.
Refactor construction of switch statements so it doesn't use class
HSubgraph.

There are also a few improvements.  We do not use an auxiliary list of
comparisons because they're embedded as a linked list in the graph
under construction.  We share a common break block for all breaks from
the same switch.  We do not insert empty blocks unless necessary to
maintain edge-split form.

There is also a bug fix.  The entry to a clause body is a potential
join and must have a join ID set, otherwise deoptimization within the
body can go to an unpredictable place in the unoptimized code.

Review URL: http://codereview.chromium.org/6650021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7099 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 12:06:54 +00:00
sgjesse@chromium.org
23777e1b41 Remove deserializer assert checking object size allocated in large object space
The change r7079 could allocate a few small code objects in large object space. Therefore the assert is no longer valid.
Review URL: http://codereview.chromium.org/6647017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 11:50:03 +00:00
sgjesse@chromium.org
061fa8be11 Add the possibility for a code stub to be non-movable
Non-moveable code-stube are allocated in large object space. They are only required on ARM where the different C-entry stubs are required to never move.

This gets rid of pre-computing these stubs and hope that they never move. Also for crankshaft the C-entry stub which saved doubles is not generated in the snapshot so it ends up being generated at runtime and potentially move.
Review URL: http://codereview.chromium.org/6626072

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7097 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 10:38:19 +00:00
ricow@chromium.org
df7bbc2386 Remove unneccesary JumpTableEntry pointer from lithium-codegen-x64.cc
Review URL: http://codereview.chromium.org/6648028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7096 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 10:29:14 +00:00
ricow@chromium.org
7783cc48fd Fix memory leaks on x64
This change uses ZoneObject as base class for our jumptable entry. In
addition this change refactors the JumpTableEntry a bit.

Review URL: http://codereview.chromium.org/6647012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7095 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-09 10:02:47 +00:00
fschneider@chromium.org
4d2c3ee82d Avoid allocation of temporary zone lists when inserting representation changes.
Instead of allocating fresh temporary lists for every instruction, reuse
the same instance and reset it between instructions.

This reduces the amount of zone memory used for inserting the HChange
instructions roughly by half.

Review URL: http://codereview.chromium.org/6628079

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7094 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 15:08:36 +00:00
sgjesse@chromium.org
5a768f0769 Fix release build
Disassemble is not available in release build
Review URL: http://codereview.chromium.org/6635043

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7093 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 15:00:08 +00:00
ricow@chromium.org
da9fe0f961 Remove GetProvidedParametersCount from JavaScriptFrame, it simply calls ComputeParametersCount.
Review URL: http://codereview.chromium.org/6635042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7092 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 14:18:28 +00:00
lrn@chromium.org
a8b41a0edd Fix bug in X64 RegExpExec stub.
Used incorrect register for referencing RegExp data, so it always failed
to match the fast case.
When modifiying the object layout, it was possible to make it crash instead.

BUG=v8:1236
TEST=test/mjsunit/regress/regress-1236.js

Review URL: http://codereview.chromium.org/6635041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7091 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 14:15:25 +00:00
lrn@chromium.org
0c74af3d0f Fix Issue 1234.
Ensure that there is always enough bytes between consequtive calls in
unoptimized code to write a call instruction at the return points
without overlapping.

This handles the case where two return points were only four bytes
apart (because the latter call was to a register).

BUG=v8:1234

Review URL: http://codereview.chromium.org/6624091

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7089 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 11:21:38 +00:00
sgjesse@chromium.org
c3f9e1dee2 Ensure the result is used for the remaining calls to SetElement
Now mark SetElement as must use result

The debugger runs inside its own context so there should be no setters hit. Which is the reason for the live-edit code asserting non-empty handles.
Review URL: http://codereview.chromium.org/6621042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7088 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 11:14:25 +00:00
fschneider@chromium.org
4b912c6af3 Fix assertion failure introduced by my previous change.
Review URL: http://codereview.chromium.org/6621068

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7087 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 10:49:01 +00:00
karlklose@chromium.org
822ddaa7d1 ARM: Implement MathPowStub and DoMathPowHalf.
Review URL: http://codereview.chromium.org/6613015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7086 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 10:29:40 +00:00
fschneider@chromium.org
12ba2dd435 Improve dead phi elimination.
This change splits the existing phi elimination into two phases:
1. Remove redundant phis
2. Remove dead phis with a fixed point iteration.

The new approach allows us to remove dead phis that are connected
in a cycle.

Review URL: http://codereview.chromium.org/6624061

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7085 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-08 10:04:23 +00:00
mmaly@chromium.org
927f341d3c Strict mode arguments do not share binding with formal parameters.
Move strict mode flag from TemporaryScope to Scope so that it can be accessed from variable binding code.
Arguments do not alias in strict mode (ia32, x64 and arm, codegen and full codegen).
Hydrogen tolerates null arguments_shadow().
In codegen-<arch> arguments object is allocated eagerly to capture values before they get modified.

Review URL: http://codereview.chromium.org/6625048/

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7083 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-07 19:23:46 +00:00
kmillikin@chromium.org
4a9056cbce Fix a stack-height mismatch during deoptimization.
When deoptimizing after a conditional expression in an effect context, we
should not see the value of the conditional expression.

BUG=v8:1237

Review URL: http://codereview.chromium.org/6625057

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7082 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-07 17:01:12 +00:00
fschneider@chromium.org
ade4b7506b Make the typefeedback oracle use a NumberDictionary instead of JSObject as its backing store.
This avoids problems when getters/setters are defined on Object.

BUG=v8:1232


Review URL: http://codereview.chromium.org/6625054

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7081 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-07 16:09:56 +00:00