Commit Graph

371 Commits

Author SHA1 Message Date
lrn@chromium.org
0d12f8f4e5 Fixed remaining bugs in COW-regexp-results.
Review URL: http://codereview.chromium.org/3197011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5322 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-23 13:58:56 +00:00
lrn@chromium.org
0210df5848 Use Copy-on-write arrays for cached regexp results.
Review URL: http://codereview.chromium.org/3158020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5310 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-20 09:37:22 +00:00
ricow@chromium.org
2982f5e320 Change code pointer in function objects to a pointer to the first
instruction.

By changing the pointer to the code object to a pointer to the first
instruction we can call directly this instruction directly instead of
looking up the address through the code object.

Review URL: http://codereview.chromium.org/3156028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5309 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-20 07:10:18 +00:00
vitalyr@chromium.org
3202df6c69 Copy-on-write arrays.
Object model changes
----------------------------------------
New fixed_cow_array_map is used for the elements array of a JSObject
to mark it as COW. The JSObject's map and other fields are not
affected. The JSObject's map still has the "fast elements" bit set. It
means we can do only the receiver map check in keyed loads and the
receiver and the elements map checks in keyed stores. So introducing
COW arrays doesn't hurt performance of these operations. But note that
the elements map check is necessary in all mutating operations because
the "has fast elements" bit now means "has fast elements for reading".
EnsureWritableFastElements can be used in runtime functions to perform
the necessary lazy copying.

Generated code changes
----------------------------------------
Generic keyed load is updated to only do the receiver map check (this
could have been done earlier). FastCloneShallowArrayStub now has two
modes: clone elements and use COW elements. AssertFastElements macro
is added to check the elements when necessary. The custom call IC
generators for Array.prototype.{push,pop} are updated to avoid going
to the slow case (and patching the IC) when calling the builtin should
work.

COW enablement
----------------------------------------
Currently we only put shallow and simple literal arrays in the COW
mode. This is done by the parser.

Review URL: http://codereview.chromium.org/3144002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-16 16:06:46 +00:00
erik.corry@gmail.com
702780ee9a Make the Integer32 type info only cover the signed 32 bit integers.
Fix some bit op bugs introduced last week on IA32:
http://code.google.com/p/chromium/issues/detail?id=52096
Review URL: http://codereview.chromium.org/3151017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-16 15:08:32 +00:00
ager@chromium.org
3e936b547a Remove experimental fast-codegen. We are no longer working on this
approach.

Review URL: http://codereview.chromium.org/3152016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5259 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-13 09:07:09 +00:00
sgjesse@chromium.org
8e0cd6db8a Handle overwriting valueOf on String objects correctly when adding
This adds a check to the fast case string add to ensure that the String object still have the default valueOf function. The default valueOf is sitting on a hidden prototype of String.prototype.

Before using the fast case valueOf the object is checked for a local valueOf property. For slow case objects this check always reports true (the dictionary is not probed, so valueOf might be there) and for fast case objects the descriptor array is checked for the valueOf symbol (just liniar scan). After that the prototype is checked for beeing the initial value of String.prototype. If this all pass (that is the default valueOf is still in place) this result is cached on the map making the check fast the next time.

This is only implemented in the optimizing compiler, as the two usages of %_IsStringWrapperSafeForDefaultValueOf is never hit by the full compiler.

I will port to x64 and ARM when this has been reviewed for ia32.

I will remove the performance counters prior to final commit.

BUG=http://code.google.com/p/v8/issues/detail?id=760
TEST=test/mjsunit/regress/regress-760-1.js
TEST=test/mjsunit/regress/regress-760-2.js

Review URL: http://codereview.chromium.org/3117006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-12 13:43:08 +00:00
sgjesse@chromium.org
fcfe6d74d9 Minor change to for-in
Return (smi) 0 instead of object null from the FILTER_KEY builtin.

Add a test which tests keys being deleted during for-in.
Review URL: http://codereview.chromium.org/3170004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5243 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-11 13:46:10 +00:00
ricow@chromium.org
36e5743674 Put direct code pointers into JSFunction objects. This is a first step
in allowing more flexible compilation and to simplify builtins lookup. 

This changes a number of places where code objects are assigned to
SharedFunctionInfo objects to also assign this code object to the
JSFunction. In addition, the code flushing is changed slightly to
accomodate this (we need to flush the code from functions pointing to
SharedFunctionInfo objects that has already been flushed).


Review URL: http://codereview.chromium.org/3120006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5233 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-11 08:12:53 +00:00
erik.corry@gmail.com
bdfdf8bee9 Fix fuzzer-found error where left and right were the same register in bitops.
Review URL: http://codereview.chromium.org/3115004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5231 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-10 12:30:14 +00:00
ager@chromium.org
9677b96b7b Fix bug in binary operations code. When loading a 32-bit integer as a
64-bit integer to the stack, the high and low parts of the 64-bit
value were pushed in the wrong order.

TBR=erik.corry@gmail.com
Review URL: http://codereview.chromium.org/3010060

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5216 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-09 14:54:23 +00:00
erik.corry@gmail.com
751ec5d099 IA32: Avoid going into stubs or runtime code for bitops even if the
inputs are heap numbers or the result is a heap number (only with
SSE2).  Make it possible for a deferred code object to work without
spilling all registers.
Review URL: http://codereview.chromium.org/3054047

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5215 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-09 13:12:02 +00:00
lrn@chromium.org
df8b3eb742 Fix cast that fails to build on Windows.
Review URL: http://codereview.chromium.org/3044055

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5213 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-09 12:44:48 +00:00
lrn@chromium.org
53e22e386c Create a new RegExp object for every evaluation of a RegExp literal.
Changes necessary to following ES5 semantics and matching Safari - in ES3
the same RegExp object was generated by each evaluation of the RegExp literal.
Fixes bug 704.

Review URL: http://codereview.chromium.org/3034060

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5198 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-08-06 13:04:27 +00:00
antonm@chromium.org
9d6fd08276 Port faster callbacks invocation to x64.
It's a port of http://code.google.com/p/v8/source/detail?r=3209 to x64 platform.
That allows invocation of callbacks without going into runtime.

Review URL: http://codereview.chromium.org/2801008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5141 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-28 09:36:53 +00:00
ager@chromium.org
79e332010a Fix aliasing problem in inlined stores on x64 and ia32. The receiver
and the value can share a register. We need to remove this aliasing
before modifying the registers.

I haven't managed to generate a stand-alon test case for this
yet. I'll do that as a separate change. This was found while loading
Wave.

Review URL: http://codereview.chromium.org/3039025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5127 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-23 11:55:03 +00:00
whesse@chromium.org
e8c0a459a9 Fix bug in r5123, Comparison(), by unusing results before unconditional jump to smi comparison JumpTarget.
Review URL: http://codereview.chromium.org/3026019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5126 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-23 11:20:59 +00:00
whesse@chromium.org
8bb47c1939 Use static type information when creating the inlined code for comparison operators.
Review URL: http://codereview.chromium.org/3017025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5123 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-23 09:05:46 +00:00
whesse@chromium.org
9deeec0b08 Change asserts to STATIC_ASSERT if they can be checked at compilation time. Just in the codegen-xxx.cc files on all platforms.
Review URL: http://codereview.chromium.org/3017018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5121 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-23 08:25:48 +00:00
ager@chromium.org
dc45fcb232 Port inline in-object property stores from ia32 to x64.
Review URL: http://codereview.chromium.org/3008017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5115 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-22 06:31:49 +00:00
ager@chromium.org
c9c7f8834e Inline in-object property stores on ia32 when in loop and not in top-level code.
Review URL: http://codereview.chromium.org/3046006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5105 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-21 06:59:34 +00:00
whesse@chromium.org
54f68eacbf Stop pushing arguments onto the stack in CompareStub until just before calling runtime.
This is a fixed version of change 5097, which had the problem that LoadFloatOperands tried to load the arguments from the stack.
Review URL: http://codereview.chromium.org/3040010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5103 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-20 12:41:43 +00:00
whesse@chromium.org
24ca3eddd7 Revert 5097 on ia32 platform until problem is diagnosed.
Review URL: http://codereview.chromium.org/3018009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-19 14:26:51 +00:00
whesse@chromium.org
96a7cc4a38 Delay the pushing of arguments until just before calling runtime in CompareStub.
This will allow jumps to runtime that skip useless tests.
Review URL: http://codereview.chromium.org/3016004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5097 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-19 13:35:41 +00:00
ricow@chromium.org
4957326521 Refactor type checks in v8natives.js and runtime.js.
This includes adding a new inline IsSpecObject method to the code
generator.  The old approach was somehow ineffecient since we would
call both IsObject, IsUndetectable and IsFunction to determine if
something was an object according to the spec. This change introduces
a new macro that determines if something is an object according to the
spec (and this does not include null). 

This change also corrects a few places where undetectable objects was
not allowed even when they should be (priorly they would use only
IS_SPEC_OBJECT_OR_NULL, which would return false on an undetectable
object, the new IS_SPEC_OBJECT returns true on an undetectable object.

Review URL: http://codereview.chromium.org/2877018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5087 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-16 11:21:08 +00:00
whesse@chromium.org
5794e68f76 Remove unnecessary formatting differences between ia32 and x64 code generators. Mainly just typographical changes.
Review URL: http://codereview.chromium.org/3023001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5080 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-15 14:31:49 +00:00
whesse@chromium.org
dd06f4f102 Avoid a smi check when comparing an unknown to a constant smi for equality on ia32 and x64 platforms.
Review URL: http://codereview.chromium.org/2897013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5077 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-15 10:34:08 +00:00
ager@chromium.org
6744793c90 Landing for Rodolph Perfetta.
Allow compareStub on ARM to take register in reverse order optionally.

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/2850046/show

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5059 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-13 20:54:06 +00:00
erik.corry@gmail.com
5bc0522ad4 Second half of number alloc change accidentally omitted from
last change.
Review URL: http://codereview.chromium.org/2919001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5039 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-08 13:50:13 +00:00
erik.corry@gmail.com
118a421067 Simplify the transitions in the Binary Op ICs. Now a single call
to the runtime will both patch in the more specialized binary op
stub and calculate the answer.  This eliminates the need to call
both the rest of the binary op and the patching runtime call.  The
runtime routines are altered to be more agressive in returning
Smis so we don't get spurious heap numbers as inputs to binary ops
while we are patching the binary op ICs.
Review URL: http://codereview.chromium.org/2843049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5026 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-06 12:56:55 +00:00
erik.corry@gmail.com
8da222b010 Specialize GenericUnaryStub so that it knows whether it needs to
take negative zero into account.
Review URL: http://codereview.chromium.org/2850043

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5018 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-05 11:03:16 +00:00
whesse@chromium.org
94ae008739 Let relational comparison operators quickly compare a string with itself.
Review URL: http://codereview.chromium.org/2802028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@5000 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-07-01 15:18:07 +00:00
whesse@chromium.org
b122fbd363 Decide that different detectable objects are not equal without calling runtime, on all platforms.
Review URL: http://codereview.chromium.org/2805046

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4992 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-30 13:24:47 +00:00
whesse@chromium.org
363060ca23 Ensure that ToPrimitive is called on all objects involved in comparisons <, <=, >, >=. Ensures that ToPrimitive is called when comparing an object to undefined. Fixes bugs on all platforms.
Review URL: http://codereview.chromium.org/2834022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4971 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-29 06:47:19 +00:00
lrn@chromium.org
49c4718443 Fix bug in CheckBothPositiveSmi macro.
Fix lint.

Review URL: http://codereview.chromium.org/2826021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4952 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-25 12:31:49 +00:00
lrn@chromium.org
f0529878d1 Add special case of taking a substring that is the entire string to stub.
Review URL: http://codereview.chromium.org/2800027

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4951 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-25 11:24:41 +00:00
vitalyr@chromium.org
8ab6832203 Add "has fast elements" bit to maps and use it in inlined keyed loads.
A potential issue with this change is creating lots of maps when
objects flip between fast/slow elements modes.  We could add special
transitions to avoid this.  Yet testing this on our benchmarks, gmail,
and wave seems to indicate that this is not a real problem.

Review URL: http://codereview.chromium.org/2870018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4941 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-24 13:56:35 +00:00
vitalyr@chromium.org
3cafa65462 Track ascii-ness of data in externalized strings.
If a two-byte string only contains ascii characters, then we can save
memory when flattening a cons string containing it. Similarly we can
use this in Array.prototype.join implementation. To track this a new
bit is added to instance type. This bit is used as a hint in generated
code and in runtime functions.

To enable testing a new V8 extension is added controlled by
--expose-externalize-string flag.

Review URL: http://codereview.chromium.org/2762008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4894 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-17 16:19:28 +00:00
ager@chromium.org
2043956c54 Remove the comisd instruction from the ia32 and x64 assemblers. We
should always use ucomisd.

Add missing pop from floating-point stack in case of allocation failure.
Review URL: http://codereview.chromium.org/2831009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4878 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-16 12:32:34 +00:00
ager@chromium.org
ed0fc41723 Change hash computation for transcendental cache to use arithmetic
shifts. Leads to fewer collisions.

Review URL: http://codereview.chromium.org/2809012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4872 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-16 08:29:25 +00:00
ager@chromium.org
9a34337066 Remove direct reference from builtins code to global object. Load the
global object through the context instead.

Review URL: http://codereview.chromium.org/2822003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4857 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-14 12:08:39 +00:00
sgjesse@chromium.org
1c5a9435ab ARM: Change code generation for function return
Generating code for function return on ARM is now aligned with the other platforms. The first non-shadowed return statement encountered will emit code for function return and all other returns including the fall through at the bottom of a function will jump to that.
Review URL: http://codereview.chromium.org/2815003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4856 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-14 11:37:05 +00:00
whesse@chromium.org
317dcabe84 Change tests for string type in RegExpExecStub on all platforms.
Review URL: http://codereview.chromium.org/2772004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4854 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-14 10:21:24 +00:00
ager@chromium.org
89ffd6c2f2 Implement transcendental cache probing in generated code on ARM.
Fixed hash code computation on ia32 and x64 to match the runtime
system (change arithmetic shift right to logical shift right).

Review URL: http://codereview.chromium.org/2731007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4843 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-11 10:18:44 +00:00
whesse@chromium.org
e1e4d985f9 Use static type information on x64 in LikelySmiBinaryOperation.
Refactor check for smi operands, using type information, on ia32 and x64.
Review URL: http://codereview.chromium.org/2771001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4831 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-09 11:15:54 +00:00
whesse@chromium.org
3c42d18a44 Revert change 4812 due to performance problems. This was a change in string addition on ia32.
Review URL: http://codereview.chromium.org/2706001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4813 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-07 13:16:05 +00:00
whesse@chromium.org
dd278e634e Improve generated code for string encoding tests on ia32.
Review URL: http://codereview.chromium.org/2673001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4812 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-07 12:15:10 +00:00
kaznacheev@chromium.org
ff61618c4b Extend CallIC to support non-constant names.
This speeds up constructs like this: 
var zz='replace'; '123'[zz]('3','4');

Review URL: http://codereview.chromium.org/2280007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4804 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-07 08:27:32 +00:00
lrn@chromium.org
d3d295efa7 Add optimized version of memcpy on ia32.
Only used in one place right now.
Still room for tweaking.

Review URL: http://codereview.chromium.org/2582001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4796 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-04 11:30:55 +00:00
whesse@chromium.org
0fc3dca698 Add test_b(Operand, immediate) to ia32 assembler, and use it where possible. Improve comparison to a constant one-character string. Use CmpInstanceType in more places on ia32. Add IsObjectJSObjectType and IsInstanceJSObjectType to ia32 macro assembler, using a single branch for a range test.
Review URL: http://codereview.chromium.org/2586001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4795 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2010-06-04 10:46:04 +00:00