Commit Graph

4405 Commits

Author SHA1 Message Date
lrn@chromium.org
a556a16f0f Fix bad cast in string-search.h.
Review URL: http://codereview.chromium.org/3530004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5572 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-10-01 09:08:52 +00:00
kaznacheev@chromium.org
1156bbc82a Fine-tune inobject slack tracking.
The current parameters (initial slack = 6, kGenerousAllocationsCount = 16)
caused 15-20% performance regression on 3d-raytrace on x64 
(visible both in standalone shell and in the web harness). This is very
likely some kind of an alignment  problem (it goes away completely with --prof).
A similar problem can be observed on ia32 with initial slack=2.

Increased initial slack to 8 and decreased constructions count to 8. 
This made the regression go away (and did not harm any of the other tests).

Review URL: http://codereview.chromium.org/3595002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5571 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-10-01 08:28:28 +00:00
mikhail.naganov@gmail.com
5a30072502 Fix HeapSnapshotsDiff test, diff implementation, and a bug introduced
during snapshot size optimization.

Sorry, now I figured out that the diff implementation itself was also
incorrect.  Reachable nodes must be filtered from the beginning,
otherwise, an object that is already disconnected, but not discarded
yet, will not appear as a deleted (thankfully, this bug for some
reason had appeared on the x64 port.)

BUG=868
TEST=HeapSnapshotRootPreservedAfterSorting

Review URL: http://codereview.chromium.org/3531005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5570 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-10-01 07:19:23 +00:00
vitalyr@chromium.org
7228d867cb Use existing global cell status as a hint when generating loads.
Review URL: http://codereview.chromium.org/3537003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5569 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 17:39:31 +00:00
mikhail.naganov@gmail.com
44050b12b9 Revert last commit due to check fail on Linux.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5568 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 12:35:03 +00:00
mikhail.naganov@gmail.com
99c276bbbc Fix HeapSnapshotsDiff test and a bug introduced during snapshot size optimization.
The bug wasn't revealed because the test was disabled.

BUG=848
TEST=HeapSnapshotsDiff,HeapSnapshotRootPreservedAfterSorting

Review URL: http://codereview.chromium.org/3572003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5567 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 12:24:56 +00:00
antonm@chromium.org
f8b274eea8 Rename some x64 macros to be more precise about their semantics.
Review URL: http://codereview.chromium.org/3574002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5566 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 11:48:03 +00:00
mikhail.naganov@gmail.com
4e961fdd36 Add fflush after writing to log (Issue 868).
BUG=868

Review URL: http://codereview.chromium.org/3605001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5565 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 10:50:39 +00:00
kmillikin@chromium.org
9190e5f941 Prepare push to trunk. Now working on version 2.4.8.
Review URL: http://codereview.chromium.org/3585003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5562 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 09:35:50 +00:00
kmillikin@chromium.org
c019508bad Cleanup of the parser.
The lazy parsing functions took a host of arguments that can all be
derived from the SharedFunctionInfo, and the SharedFunctionInfo is
always available when parsing lazily.  Change the interface to take a
single CompilationInfo or SharedFunctionInfo argument.

Also remove a flag in the parser that was always false when it was read.

Review URL: http://codereview.chromium.org/3538005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5561 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 09:28:58 +00:00
kmillikin@chromium.org
f21b34ef0f Introduce subclasses of class CompilationInfo.
It was a wart that we had three handle fields, exactly one of which
was non-null; and that we had three overloaded constructors.  Instead,
introduce subclasses and virtual methods.

Remove some unused fields from class CompilationInfo.

Review URL: http://codereview.chromium.org/3566003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5560 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 08:48:37 +00:00
erik.corry@gmail.com
d46fb9d454 New attempt to make the allocation routines 64 bit clean.
This one has been approved by the 64 bit compiler in MSVC
2005 so I hope it also passes the 2008 version.

The --max-new-space-size option is now in kBytes.
The --max-old-space-size option is now in MBytes.

Some issues remain with 64 bit heaps and the counters.  See
http://code.google.com/p/v8/issues/detail?id=887
Review URL: http://codereview.chromium.org/3573005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5559 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-30 07:22:53 +00:00
antonm@chromium.org
9e618ff460 Add another check to be sure non-negative smi fits into Uint32 range.
Review URL: http://codereview.chromium.org/3546003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5558 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 18:15:36 +00:00
kmillikin@chromium.org
4d04e9d5d4 Remove some unused fields from class CompilationInfo.
Review URL: http://codereview.chromium.org/3533007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5557 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 17:38:37 +00:00
antonm@chromium.org
3567207741 Do not invoke indexed interceptor getters for negative indices.
BUG=https://bugs.webkit.org/show_bug.cgi?id=46689

Review URL: http://codereview.chromium.org/3520006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5553 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 15:08:38 +00:00
erik.corry@gmail.com
ef135e5351 Add cast to make Win-64 compile without warnings. This is a commit of http://codereview.chromium.org/3521002 for lrn.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5552 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 14:06:30 +00:00
lrn@chromium.org
3614edebea Added missing file. Mea culpa.
Review URL: http://codereview.chromium.org/3586001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5551 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 12:12:45 +00:00
lrn@chromium.org
2c85faf1ae Refactored string search code.
Made string search state explicit for repreated calls (a StringSearch class).

Review URL: http://codereview.chromium.org/3467010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5550 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 10:57:23 +00:00
kmillikin@chromium.org
d9d49052f9 Add an AST node type enum to AST nodes.
When inspecting an AST node in the debugger, it can be difficult to
find its concrete type.  Add a "node_type" virtual function to all the
concrete AST nodes.

Use a macro to generate the definition of node_type as well as Accept
and the AsXXX type testing/casting functions.

Review URL: http://codereview.chromium.org/3431026

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5549 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-29 07:51:47 +00:00
vitalyr@chromium.org
e0feeec039 Mark ptest instruction as requiring SSE4.1.
I also added a flag like the ones we have for other SSE versions.

Review URL: http://codereview.chromium.org/3505003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5548 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 14:56:36 +00:00
vitalyr@chromium.org
9aa65c29dc Fix external array iteration used in array concat.
We should use Factory and not Heap there.

Review URL: http://codereview.chromium.org/3380031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5547 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 14:49:29 +00:00
erik.corry@gmail.com
2d0c23bc82 Revert attempt to make heap size 32/64 clean. This change needs to
be done from Windows where the compiler is stricter about truncating
changes.
Review URL: http://codereview.chromium.org/3454035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5545 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 12:49:25 +00:00
erik.corry@gmail.com
d6313dbbb7 More changes to 32/64 bit correctness.
TBR=vegorov
Review URL: http://codereview.chromium.org/3412038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5544 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 12:18:19 +00:00
erik.corry@gmail.com
28d7064856 More correctness around PrintF and 32/64 bit values.
Fix test after 64 bit heap size change.
Review URL: http://codereview.chromium.org/3432032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5543 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 11:44:30 +00:00
erik.corry@gmail.com
f1635606df Use intptr_t instead of int for heap sizes. This is a step towards
removing the 1Gbyte limit.
Review URL: http://codereview.chromium.org/3418035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5542 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 10:35:47 +00:00
lrn@chromium.org
8f7bfea235 Avoid logging preparse-data inside lazily compiled functions.
Reduces size of preparser data significantly when there are nested functions.
Also allows us to drop the "skip" fields of function entries,
that tells us how much preparse-data to skip when skipping the function source.

Review URL: http://codereview.chromium.org/3412034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5541 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-28 07:50:32 +00:00
vitalyr@chromium.org
1982f9d257 Custom call IC for Math.abs.
Review URL: http://codereview.chromium.org/3446024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5538 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-27 23:41:41 +00:00
podivilov@chromium.org
0012576e2b Add Debug::CancelDebugBreak method to api
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5535 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-27 15:09:55 +00:00
whesse@chromium.org
9c31eb8229 Prepare push of version 2.4.6 to trunk. Now working on version 2.4.7.
Review URL: http://codereview.chromium.org/3470016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5531 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-27 10:58:18 +00:00
vegorov@chromium.org
9045e2e60f Fix more GC unsafe places
Review URL: http://codereview.chromium.org/3499001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5530 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-27 10:29:25 +00:00
whesse@chromium.org
25f31f13fa Fix error in x64 inlined optimized shift operators.
Review URL: http://codereview.chromium.org/3496007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5529 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-27 07:24:01 +00:00
mikhail.naganov@gmail.com
08967bb4aa fixed
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5528 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-27 05:25:31 +00:00
dimich@chromium.org
589eb4a861 Add CODE_POINTER_ALIGN, use it in Page to align generated code.
The object's space in Page starts after Page header and is aligned to kMapAlignment which is 32 bytes on 32-bit and 8 bytes on 64-bit.

In case of 64-bit target, the current page header size is exactly 32 bytes so we get the code magically aligned at 32 bytes but it is better to have a separate CODE_POINTER_ALIGN macro to make sure the object space in Page is aligned properly for both maps and code.

There could be a small waste of bytes sometimes (since both Page header and Code header sizes are aligned separately) but it seems the optimal one would involve cross-dependencies between .h files and not clear if it's worth it.

This is a back-port from Isolates branch.

Review URL: http://codereview.chromium.org/3461021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5526 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 21:48:44 +00:00
kaznacheev@chromium.org
c39194e06f Implemented missing instructions in ia32 and x64 disassembler.
ia32: fld(i), fldpi
x64: fld(i), fldpi, cvtsd2si, cvttsd2si

Review URL: http://codereview.chromium.org/3471011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5525 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 12:55:17 +00:00
mikhail.naganov@gmail.com
0a539dc181 Fix presubmit error.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5524 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 11:50:50 +00:00
mikhail.naganov@gmail.com
8f22667805 Provide more functions to CPU profiler (fix issue 858).
The cause for missing functions is that some of them are created
from compiled code (see FastNewClosureStub), and thus not get
registered in profiler's code map.

My solution is to hook on GC visitor to provide JS functions
addresses to profiler, only if it is enabled.

BUG=858
TEST=

Review URL: http://codereview.chromium.org/3417019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5523 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 11:45:12 +00:00
erik.corry@gmail.com
a26a9b74ab Revert 5520 due to compilation failure.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5521 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 09:35:19 +00:00
erik.corry@gmail.com
68ef41af29 Small tweaks to make the x64 full code generator generate
more compact code.
Review URL: http://codereview.chromium.org/3443022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5520 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 08:34:52 +00:00
kmillikin@chromium.org
9c503a8330 Fix some inconsistent formatting.
I noticed we sometimes had extra spaces before and after the "const"
keyword.  Probably the result of a search and replace gone wrong.

This is a whitespace only change.

Review URL: http://codereview.chromium.org/3427021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5519 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 08:25:31 +00:00
kaznacheev@chromium.org
cfbc1eed9c Prevent modification of cached normalized maps.
Finally sovles the problem that r5342 attempted to solve.
When adding a stub to a map's code cache we need to make
sure that this map is not used by object that do not need
this stub.

Existing solution had 2 flaws:
1. It checked that the map is cached by asking the current context.
If the object escaped into another context then NormalizedMapCache::Contains
returns false negative.

2. If a map gets evicted from the cache we should not try to modify it
even though Contains returns false.

This patch implements much less fragile solution of the same problem:
A map now has a flag (is_shared) that is set once the map is added
to a cache, stays set even after the cache eviction, and is cleared
if the object goes back to fast mode.

Added a regression test.

Review URL: http://codereview.chromium.org/3472006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5518 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 08:18:33 +00:00
kmillikin@chromium.org
830185b175 Clean up some messiness in Scopes.
For some reason, the scope's arguments and arguments shadow were
variable proxies, which resulted in all references to the arguments
shadow being shared in the AST.  This makes it hard to put per-node
state on the AST nodes.

I took the opportunity to remove Variable::AsVariable which has
confused people in the past, and to rename Variable::slot to the more
accurate Variable::AsSlot.

Review URL: http://codereview.chromium.org/3432022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5517 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-24 07:53:59 +00:00
kaznacheev@chromium.org
5e572e4622 Support dec_b instruction in ia32 and x64 disassembler.
Review URL: http://codereview.chromium.org/3412028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5515 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 19:01:56 +00:00
vitalyr@chromium.org
740aba98fb Add an assert to ensure we never allocate second empty fixed array.
Review URL: http://codereview.chromium.org/3442012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5514 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 12:49:59 +00:00
vitalyr@chromium.org
48fd3dcaf9 Fix copy-on-write assert by setting the new array map early.
BUG=876

Review URL: http://codereview.chromium.org/3466013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5513 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 12:23:35 +00:00
vegorov@chromium.org
622351fedd Fix getOwnPropertyDescriptor() support for index properties.
Add support for index properties with getters, setters or indexed interceptors.

For indexed interceptor case only fix crashes, do not guarantee any semantic soundness. Separate issue opened for this http://code.google.com/p/v8/issues/detail?id=877

BUG=http://code.google.com/p/v8/issues/detail?id=874

Review URL: http://codereview.chromium.org/3462005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5512 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 11:25:01 +00:00
erik.corry@gmail.com
94cf27373c Cleanup of contexts in the full code generator.
Review URL: http://codereview.chromium.org/3449004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5511 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 09:22:45 +00:00
kaznacheev@chromium.org
06c867404b Dynamically determine optimal instance size.
The number of inobject properties used to be derived from the number
of this property assignments in the constructor (and increased by 2 to
allow for properties added later). This very often leads to wasted inobject 
slots.

This patch reclaims some of the unused inobject space by the following method:
 - for each constructor function the first several objects are allocated using the initial 
   ("generous) instance size estimation (this is called 'tracking phase').
 - during the tracking phase map transitions are tracked and actual property counts are collected.
 - at the end of the tracking phase instance sizes in the maps are decreased if necessary
   (starting with the function's initial map and traversing the transition tree).
 - all further allocation use more realistic instance size estimation.

Shrinking generously allocated objects without costly heap traversal is made possible 
by initializing their inobject properties with one_pointer_filler_map (instead of undefined). 

The initial slack for the generous allocation is increased from 2 to 6 which really helps some tests.

Review URL: http://codereview.chromium.org/3329019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5510 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 09:15:26 +00:00
vegorov@chromium.org
7c8bb52737 Fix possible evaluation order problems.
We should not allow handle dereference and GC inside the same expression because order of subexpression evalution are not defined.

Review URL: http://codereview.chromium.org/3398014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5509 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 08:27:51 +00:00
kaznacheev@chromium.org
e69278a70c Unuse labels when bailing out of StubCompiler methods.
There are 3 methods where early return happen before the 
miss label is bound. This is harmless in Release mode, 
in Debug an assertion fails.

Review URL: http://codereview.chromium.org/3405022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5508 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-23 08:06:19 +00:00
antonm@chromium.org
31284c01f8 Fix a build for Win64 with VS2008.
It complaints of type conversions.

Review URL: http://codereview.chromium.org/3396015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5507 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-09-22 13:57:00 +00:00