Commit Graph

5404 Commits

Author SHA1 Message Date
whesse@chromium.org
b5c3b18f29 X64 Crankshaft: Implement DoubleToI in optimizing compiler.
Review URL: http://codereview.chromium.org/6551009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6889 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 12:02:47 +00:00
sandholm@chromium.org
9d4b0fa932 Improve parseInt.
Review URL: http://codereview.chromium.org/6546055

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6888 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 11:21:15 +00:00
kmillikin@chromium.org
3c02d036a8 Partial refactoring of subgraphs.
Change the interface to the Hydrogen graph builder to appear like it
directly holds a current basic block and a current environment.
Remove some direct accesses to the current subgraph, and remove
subgraph accessors that simply forwarded to the exit block.

Review URL: http://codereview.chromium.org/6532088

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6887 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 10:32:16 +00:00
sgjesse@chromium.org
b0fa94777e Refactored PathTracer in heap.cc.
This is so that it can be reused by LOL code later.

Patch by Mark Lam from Hewlett-Packard Development Company, LP

Review URL: http://codereview.chromium.org/6541044


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6886 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 10:05:30 +00:00
sgjesse@chromium.org
2ab5691ac1 ARM: Add support for just one string argument in the string add stub
Having a string add stub which can handle a string in just one of the arguments is essencial for the type recording binary operation stub when expecting strings. Otherwise string added with e.g. smi will always call the runtime for a type transition which will be back to the same types as the transition code keeps it in string arguments when one argument is a string.

This fixes the regression (especially Delta Blue) caused by replacing the generic binary operation stub with the type recording binary operation stub in the ARM lithium code generator.
Review URL: http://codereview.chromium.org/6551008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6882 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 09:35:37 +00:00
ager@chromium.org
823f9feb05 Revert "Expand fast case of parseInt to include radix == 10 and radix == 0."
Causes Mozilla failures.

TBR=sandholm@chromium.org

Review URL: http://codereview.chromium.org/6542056

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6880 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 09:14:47 +00:00
sandholm@chromium.org
4e2101d928 Expand fast case of parseInt to include radix == 10 and radix == 0.
Review URL: http://codereview.chromium.org/6532087

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6879 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 09:01:06 +00:00
kmillikin@chromium.org
14e185b31b Change the translation of break/continue into Hydrogen.
Resolve break and continue when we see them, rather then deferring
them until later.

Review URL: http://codereview.chromium.org/6541060

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6878 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 08:40:10 +00:00
kmillikin@chromium.org
663da3df40 Fix presubmit failure.
TBR=ager@chromium.org

Review URL: http://codereview.chromium.org/6552001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6877 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 06:41:32 +00:00
kmillikin@chromium.org
c1386c3381 Shorten live ranges for arguments to runtime calls.
Before, the live ranges of the arguments extended to the call itself, and
they were pushed immediately before the call.  Now, they are spilled eagerly
as soon as their value is available and they are spilled to the right place.

The inlined runtime calls in the optimized backend are changed to work as in
all the other backends: they get their arguments untranslated and can choose
their own custom evaluation order.

Review URL: http://codereview.chromium.org/6526047

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6876 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 06:25:01 +00:00
mmaly@chromium.org
4e6635ad66 Fix Xcode project.
TBR= ager@chromium.org

BUG=
TEST=

Review URL: http://codereview.chromium.org/6551006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6875 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 01:10:28 +00:00
mmaly@chromium.org
fb20f7fc75 CallIC and KeyedCallIC not wrapping this for strict mode functions.
Fix CallIC and KeyedCallIC to correctly use Handle<Object>.

Review URL: http://codereview.chromium.org/6523052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6874 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-22 00:39:21 +00:00
antonm@chromium.org
6aec28f447 Minor refactoring: unify lazy function compilation for in loop and no in loop variants.
Review URL: http://codereview.chromium.org/6542017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6873 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 18:28:20 +00:00
mikhail.naganov@gmail.com
3eb151ebea Bug: OS::MemoryMappedFile::open() should not truncate a pre-existing file.
Patch by Mark Lam from Hewlett-Packard Development Company, LP

Review URL: http://codereview.chromium.org/6543039

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6872 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 17:17:26 +00:00
kmillikin@chromium.org
123dbb2f5e Change the baseline compiler to match the Hydrogen graph builder.
The Hydrogen graph translation does not build a branch for unary negation in
an effect context, so the baseline compiler should not do so either.

Review URL: http://codereview.chromium.org/6546050

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6871 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 16:49:39 +00:00
vegorov@chromium.org
58600281e8 Heap::gc_count_, last_gc_count, and kGCsBetweenCleanup should be unsigned
in order to not be vulnerable to overflow issues.

Patch by Mark Lam of Hewlett-Packard Development Company, LP

Review URL: http://codereview.chromium.org/5966001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6870 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 16:11:46 +00:00
ager@chromium.org
c2969bfb1a Implement pixel array elements access in the presence of an
interceptor that does not handle the elements load.

Review URL: http://codereview.chromium.org/6551001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6869 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 15:59:23 +00:00
fschneider@chromium.org
f13f84ae5d Fix broken build by removing more const qualifiers.
Review URL: http://codereview.chromium.org/6532085

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6867 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 13:27:36 +00:00
fschneider@chromium.org
e9162d53fe Remove const qualifier in a few places in the hydrogen IR files.
This is a preparation step for a larger refactoring of the 
hydrogen instructions.

Review URL: http://codereview.chromium.org/6542042

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6866 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 12:05:17 +00:00
ager@chromium.org
01cd10569f Remove passing tests from ARM test expectations. Yay!
TBR=karlklose@chromium.org

Review URL: http://codereview.chromium.org/6548006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6865 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 11:58:19 +00:00
ager@chromium.org
0bd776a94f Remove passing tests from expectations. Yay!
Review URL: http://codereview.chromium.org/6488028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6864 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 11:55:45 +00:00
whesse@chromium.org
e1448ebd34 X64 Crankshaft: Add ModI and ModT to x64 optimizing compiler.
Review URL: http://codereview.chromium.org/6548005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6863 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 11:32:28 +00:00
ager@chromium.org
ea0dda89bb Unifying the handling of storing and loading from safepoint stack
slots across architectures.

Review URL: http://codereview.chromium.org/6541051

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6862 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 11:29:45 +00:00
karlklose@chromium.org
24b73a04b0 ARM: Implement DoMathRound in the lithium code generator.
Review URL: http://codereview.chromium.org/6519051

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6860 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 10:30:25 +00:00
ager@chromium.org
1e89c4aa67 x64: implement apply with arguments in lithium backend.
Includes the plumbing to make sure that all calls generated by the macroassembler for lithium will generate record a safepoint.

Review URL: http://codereview.chromium.org/6469053

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6859 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-21 07:54:55 +00:00
vegorov@chromium.org
c411adca55 Adjust pointer maps for branches when connecting live ranges.
Review URL: http://codereview.chromium.org/6541024

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6858 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 16:29:41 +00:00
ager@chromium.org
5ec91a4279 x64: Fix safepoint register index calculation.
We leave room for all registers on the stack so we have to add a bit.

This is not nicely factored and I need to do that. This will fix the breakage and I
will fix the rest on Monday.

TBR=lrn@chromium.org

Review URL: http://codereview.chromium.org/6544017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6857 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 16:16:17 +00:00
whesse@chromium.org
4ac9f20601 X64 Crankshaft: Implement MathRound, MathFloor, MathSqrt, and MathPowHalf UnaryMathOperations in optimizing code generator. Remove unneeded bailout on Math.pow(-0, 0.5) from ia32 and x64 platforms - we produce the correct value without bailing out.
Review URL: http://codereview.chromium.org/6532054

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6856 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 14:34:17 +00:00
antonm@chromium.org
d00e3225f7 Minor cleanup.
Do not relookup code object and use optimized_code instead
shadowing existing code local.

Review URL: http://codereview.chromium.org/6541020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6855 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 14:22:29 +00:00
ager@chromium.org
c52e397db6 Fix presubmit.
TBR=lrn@chromium.org

Review URL: http://codereview.chromium.org/6532055

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6854 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 14:10:42 +00:00
ager@chromium.org
b8f0b88b54 x64: Implement regexp literals and string CharCodeAt in lithium backend.
Review URL: http://codereview.chromium.org/6469052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6853 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 14:00:46 +00:00
ager@chromium.org
fc7e79a838 x64: Implement the missing generic load and store operations.
Review URL: http://codereview.chromium.org/6541019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6852 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 12:54:43 +00:00
whesse@chromium.org
4752373976 X64 Crankshaft: Implement LValueOf and (dummy) LArgumentsObject.
Review URL: http://codereview.chromium.org/6542011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6850 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 12:06:52 +00:00
antonm@chromium.org
e79bfcaf3f Use [[DefineOwnProperty]] to put 'constructor' field on the protoype object.
That better follows ECMA-262 (see 13.2 Creating Function Objects) and allows
to ignore nasty JS accessors for 'constructor' property.

BUG=v8:1172
TEST=test/mjsunit/regress/regress-1172.js

Review URL: http://codereview.chromium.org/6531037

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6849 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 10:53:38 +00:00
ricow@chromium.org
95892799e1 Add access checks to Object.preventExtensions + add regression test for 1027.
Object.preventExtensions can currently be used cross-domain. With this
change we follow firefox (IE9 has our current behaviour). In addition
this includes a regression test for 1027 and access tests for
Object.seal and Object.freeze.


Review URL: http://codereview.chromium.org/6534019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6848 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-18 10:39:02 +00:00
mmaly@chromium.org
f0df4a6c9e Revert "This is not wrapped for strict mode and builtin functions."
This reverts commit 6845

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6847 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 21:56:37 +00:00
ager@chromium.org
963472c516 Change behavior of global declarations in the presence of setters.
Call accessors in the global object prototype when initializing global
variables. Function declarations are special cased for compatibility
with Safari and setters are not called for them. If this special
casing was not done webkit layout tests would fail.

Make the declaration of global const variables in the presence of
callbacks a redeclaration error.

Handle const context slot declarations conflicting with a CALLBACK as
a redeclaration error. That is, unless it is on a context extension
object which is not a real object and therefore conceptually have no
accessors in prototype chains. Accessors in prototype chains of
context extension objects are explicitly ignored in SetProperty.

Review URL: http://codereview.chromium.org/6534029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6846 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 21:04:53 +00:00
mmaly@chromium.org
bb7b014988 This is not wrapped for strict mode and builtin functions.
CallIC and KeyedCallIC do not wrap this when calling builtin
and strict mode functions.

Review URL: http://codereview.chromium.org/6523052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6845 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 17:52:03 +00:00
antonm@chromium.org
749c5ea72e Make OutOfMemory exception thrown from JS call into FatalProcessOutOfMemory as well.
That unifies the behaviour with CALL_HEAP_FUNCTION macro.

BUG=v8:1165

Review URL: http://codereview.chromium.org/6528050

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6844 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 17:21:59 +00:00
ager@chromium.org
65addc5165 Revert change to const and global variable declarations. It causes
may WebKit layout test failures.

I will look into it tomorrow.
TBR=kmillikin@chromium.org

Review URL: http://codereview.chromium.org/6537021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6843 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 16:54:49 +00:00
vegorov@chromium.org
73c2b84b33 Cygwin support
Patch by Bert Belder.

BUG=v8:64
BUG=v8:964
BUG=v8:1029

Review URL: http://codereview.chromium.org/6525028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6842 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 16:33:10 +00:00
ager@chromium.org
dc38755aba Call accessors in the global object prototype when initializing global
variables.

Make the declaration of global const variables in the presence of
callbacks a redeclaration error.

Handle const context slot declarations conflicting with a CALLBACK as
a redeclaration error. That is, unless it is on a context extension
object which is not a real object and therefore conceptually have no
accessors in prototype chains. Accessors in prototype chains of
context extension objects are explicitly ignored in SetProperty.

Review URL: http://codereview.chromium.org/6519050

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6841 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 16:30:15 +00:00
vegorov@chromium.org
b4f3795ae7 Avoid clobbering register containing name in CompilerStoreGlobal.
Review URL: http://codereview.chromium.org/6519052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6840 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 16:02:12 +00:00
whesse@chromium.org
3245a23de0 ARM: Port new version of ParallelMove's GapResolver to ARM. Fix error in vstr(DoubleRegister, MemOperand), where it was implemented as a vldr, in r6830.
Review URL: http://codereview.chromium.org/6311010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6839 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 15:25:38 +00:00
lrn@chromium.org
246560b902 Revert 6832.
The test contains a syntax error that shouldn't be detected, but it sometimes is.

TBR: kmillikin@chromium.org

Review URL: http://codereview.chromium.org/6519049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6836 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 14:13:25 +00:00
kmillikin@chromium.org
b02107284a Fix incorrect deoptimization for logical not in an effect context.
The baseline compiler does not materialize a value for expressions of
the form !expr in an effect context so the graph translation should
not produce such an environment, otherwise we risk targeting it by
deoptimization.

BUG=v8:1167

Review URL: http://codereview.chromium.org/6537018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6833 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 13:05:49 +00:00
lrn@chromium.org
cf50c5e27c Handle exceptions thrown while parsing lazy functions for inlining.
We currently leave the exception as pending without returning a Failure::Exception() value. This is either caught immediately if running with --debug-code, or caught later by an assert in debug mode.

This change makes the pending exception be cleared before returning from the failed optimization attempt.

BUG=v8::1145
TEST=test/mjsunit/regress/regress-1145.js

Review URL: http://codereview.chromium.org/6524039

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6832 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 12:23:18 +00:00
kmillikin@chromium.org
82cdd48b2f Fix a bug in deoptimization after logical expressions in an effect context.
When deoptimizing to after an expression of the form (expr0 || expr1)
or (expr0 && expr1) in an effect context, the unoptimized code could
incorrectly see the value of the expression.

Handle the short-circuit binary operators specially in effect contexts.
This fixes the issue and will generate better code when the left
subexpression is boolean-valued.

BUG=v8:1166

Review URL: http://codereview.chromium.org/6519046

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6831 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 11:06:50 +00:00
whesse@chromium.org
2642168331 ARM: Enable loads and stores of VFP registers with offsets >= 1024 bytes.
Review URL: http://codereview.chromium.org/6530002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6830 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 10:07:13 +00:00
sgjesse@chromium.org
fd0744844e ARM: Don't try to flush the icache when there is nothing to flush
The simulator implementation of the icache did not like size being zero.

BUG=v8:1090
Review URL: http://codereview.chromium.org/6526052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@6829 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-02-17 07:47:05 +00:00