Commit Graph

192 Commits

Author SHA1 Message Date
verwaest@chromium.org
341d405301 Reland and fix "Add support for keyed-call on arrays of fast elements"
BUG=
R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/71783003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17782 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-15 10:52:05 +00:00
machenbach@chromium.org
eef8694a7e [Sheriff] Revert "Add support for keyed-call on arrays of fast elements"
This reverts commit r17746 for breaking layout tests.

TBR=verwaest@chromium.org
BUG=

Review URL: https://codereview.chromium.org/72753002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17751 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-14 15:00:13 +00:00
verwaest@chromium.org
607a175cbc Add support for keyed-call on arrays of fast elements
R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/23537067

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17746 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-14 13:46:18 +00:00
mvstanton@chromium.org
3cf157b43b Inline zero argument array constructor.
patch from issue 54583003 (dependent code).

Zero arguments - very easy

1 argument - three special cases:
  a) If length is a constant in valid array length range,
     no need to check it at runtime.
  b) respect DoNotInline feedback on the AllocationSite for
     cases that the argument is not a smi or is an integer
     with a length that should create a dictionary.
  c) if kind feedback is non-holey, and length is non-constant,
     we'd have to generate a lot of code to be correct.
     Don't inline this case.

N arguments - one special case:
  a) If a deopt ever occurs because an input argument isn't
     compatible with the elements kind, then set the
     DoNotInline flag.

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/55933002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17741 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-14 12:05:09 +00:00
mvstanton@chromium.org
9d6dddb9ae Simplify behavior of code stubs that accept a variable number of stack
arguments in addition to their parameters. Before, we'd add a special
variable to the environment with the value of a register with the
number of arguments. Now, that register just appears as a parameter to
the code stub.

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/70203002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17680 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-13 10:07:04 +00:00
bmeurer@chromium.org
6f75e92902 Add initial hydrogenized NewStringAddStub.
The new stub is enabled via the --new-string-add flag, which is
disabled by default. For now, it's only a stripped down version
of the native StringAddStub, it's still work-in-progress.

BUG=v8:2990
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/61893009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-12 10:21:08 +00:00
yangguo@chromium.org
b8a77348b2 Age ICs on context dispose notification.
R=ulan@chromium.org
BUG=309723

Review URL: https://codereview.chromium.org/59383012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17526 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-06 15:47:05 +00:00
mvstanton@chromium.org
93fa1939ce Generalize AllocationSite field access in HObjectAccess.
There are simply becoming too many individual field accessors, and
more are coming.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/43563002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17408 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-28 09:36:49 +00:00
ishell@chromium.org
32cdd4d48a More Hydrogen templatization.
Factories added to a set of Hydrogen classes.
Several old-style HXxx::New() replaced with New<HXxx>() and NewUncasted<HXxx>.
Several AddInstruction() calls following New::XXX() replaced with Add<XXX>().
AddLoadNamedField() method added to GraphBuilder.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/34523004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17318 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-22 11:29:05 +00:00
titzer@chromium.org
d63c8e4dc3 Allow constants to be deleted by reinserting them into the graph as needed.
BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/31943002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17299 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-21 14:09:32 +00:00
danno@chromium.org
4c565138b5 Add tool to visualize machine code/lithium.
In the process:
- Add a command-line flag --opt-code-positions to track source position information throughout optimized code.
- Add a subclass of the hydrogen graph builder to ensure that the source position is properly set on the graph builder for all generated hydrogen code.
- Overhaul handling of source positions in hydrogen to ensure they are passed through to generated code consistently and in most cases transparently.

Originally reviewed in this CL: https://codereview.chromium.org/24957003/

Review URL: https://codereview.chromium.org/29123008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17295 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-21 13:35:48 +00:00
bmeurer@chromium.org
100fb55555 Inline number to string conversion for string addition into BinaryOp(Stub).
This fixes a performance regression that was caused by converting the
BinaryOpStub to a Hydrogen code stub. It also fixes a leftover TODO wrt.
the handling of Number*String or String*Number versions of the stub.

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/27674002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17290 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-21 12:42:08 +00:00
mstarzinger@chromium.org
701417f955 Make stack_parameter_count a plain register.
R=rossberg@chromium.org, danno@chromium.org

Review URL: https://codereview.chromium.org/28993003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17275 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-18 14:55:21 +00:00
bmeurer@chromium.org
9724f06211 Drop useless HForceRepresentation in BinaryOpStub.
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/27167002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17182 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-14 09:22:34 +00:00
bmeurer@chromium.org
0a9425e756 HIsStringAndBranch does an implicit SMI check if necessary.
No need to test for SMI first using HIsSmiAndBranch.

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/27160002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17175 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-14 08:03:35 +00:00
mvstanton@chromium.org
71af6c86a1 Allow FastCloneShallowObjectStub to use AllocationMementos.
(currently turned off)

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/26827002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17154 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-11 09:25:14 +00:00
bmeurer@chromium.org
eace510108 Get rid of the HInstanceSize instruction.
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/25666006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17116 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-04 19:04:34 +00:00
olivf@chromium.org
66c610398f Reland "Hydrogenisation of binops"
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/25494007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17104 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-04 08:17:11 +00:00
olivf@chromium.org
f454b45d6d Add a flag to print the time it takes to compile HydrogenCodeStubs.
BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/25693003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17071 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-02 11:27:37 +00:00
olivf@chromium.org
9459ed3ab4 Revert "Hydrogenisation of binops"
This reverts r17052-17054 for various build breaks.

TBR=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/25571002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17055 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-01 18:00:02 +00:00
olivf@chromium.org
7873f35eb2 Hydrogenisation of binops
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/24072013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17052 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-01 15:24:56 +00:00
mvstanton@chromium.org
f26bcd9fef Add field nested_sites to AllocationSite. This field is used to maintain
allocation site information for nested array and object literals.
It's not used productively in this CL, merely maintained in a minimal
way. (that comes next :)).

BUG=
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/23463047

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16912 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-24 10:30:41 +00:00
bmeurer@chromium.org
70b8e2d013 Turn the NumberToStringStub into a hydrogen stub.
This adds a BuildLookupNumberStringCache() in Hydrogen, which will
be used by the hydrogen version of StringAddStub, in addition to the
hydrogen version of NumberToStringStub.

R=mvstanton@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=16874

Review URL: https://codereview.chromium.org/23726041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16910 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-24 09:31:07 +00:00
machenbach@chromium.org
576515ebf3 Revert "Turn the NumberToStringStub into a hydrogen stub."
This reverts commit 16874 for breaking the tests.

TBR=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/23440064

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16876 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-23 11:22:35 +00:00
bmeurer@chromium.org
9c48fbda75 Turn the NumberToStringStub into a hydrogen stub.
This adds a BuildLookupNumberStringCache() in Hydrogen, which will
be used by the hydrogen version of StringAddStub, in addition to the
hydrogen version of NumberToStringStub.

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/23726041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16874 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-23 10:58:45 +00:00
mvstanton@chromium.org
1b576e1e86 Some cleanup fixes
* Consolidated CopyJSObject and CopyJSObjectWithAllocationSite
* Factory.h helper for struct maps.
* BuildFastLiteral shouldn't create allocation sites in pretenuring mode.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/24255005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16855 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-20 09:27:40 +00:00
mvstanton@chromium.org
137b43c9a3 Added a DependentCode field to AllocationSite. It's not currently used,
this initial CL is just to get the object layout correct.

BUG=
R=hpayer@chromium.org, mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/23567003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16833 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-19 14:13:34 +00:00
bmeurer@chromium.org
8588b74491 Remove unused context parameter from AddIncrementCounter().
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/23622033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16728 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-16 11:30:10 +00:00
mvstanton@chromium.org
516339cba5 Every place where AllocationMemento is initialized with an
AllocationSite is now checked to be sure a valid Site goes in.

This is temporary code to diagnose chromium bug 284577.

(This is a second attempt, the first attempt ran into the problem of undefined ordering of function calls in Windows and Mac optimized builds, see the fixes in code-stubs-hydrogen.cc).

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/23440035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16719 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-13 13:50:16 +00:00
svenpanne@chromium.org
75763061e7 Cleaned up deopt reason messages a bit.
They are now a bit more consistent, and they don't mention the name of a
flag, which will go out of sync some day. ;-)

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/23845003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16514 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-04 07:27:10 +00:00
dcarney@chromium.org
675ac2660d thread isolate for files starting with 'b' and 'c'
R=svenpanne@chromium.org
BUG=

Review URL: https://codereview.chromium.org/23729006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16489 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-03 06:57:16 +00:00
mvstanton@chromium.org
7fea193609 Convert FastNewClosureStub into hydrogen.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/22562002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16356 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-27 11:55:08 +00:00
mvstanton@chromium.org
c9591f005e Store mode for keyed stores should be passed in from type feedback
regardless of the map used in polymorphic stores.

BUG=
R=jkummerow@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/21058003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16323 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-26 12:28:08 +00:00
verwaest@chromium.org
6337281204 Pass checked values to HLoadNamedField, removing the need for extra type-check field.
R=titzer@chromium.org

Review URL: https://chromiumcodereview.appspot.com/22831003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16260 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-21 14:49:56 +00:00
verwaest@chromium.org
ee53b0a5ed Make all load-named-fields depend on their map-check, unless explicitly ignored.
BUG=
R=titzer@chromium.org

Review URL: https://chromiumcodereview.appspot.com/22555004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16139 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-09 18:40:10 +00:00
mvstanton@chromium.org
201c512dc7 Deoptimization is easier to diagnose when there is a text reason.
BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/22339018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16131 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-09 12:50:42 +00:00
svenpanne@chromium.org
bc6fe88030 Desugar bitwise negation into XOR and kill all UnaryOp stuff.
R=mstarzinger@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/22184004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16073 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-06 13:34:51 +00:00
svenpanne@chromium.org
bf71023ba5 Replaced unary negation by multiplication with -1.
This fixes a deopt loop in the Epic Citadel demo and removes some code. Apart from that, this change is performance-neutral.

When we do something similar for BIT_NOT, the whole UnaryOp stuff can go away.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/21782002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16029 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-02 11:56:35 +00:00
loislo@chromium.org
d2c443b774 Extract hardcoded error strings into a single place and replace them with enum.
I'd like to propagate bailout reason to cpu profiler.
So I need to save it into heap object SharedFunctionInfo.
But:
1) all bailout reason strings spread across all the sources.
2) they are native strings and if I convert them into String then I may have a performance issue.
3) one byte is enough for 184 bailout reasons. Otherwise we need 8 bytes for the pointer.

Also I think it would be nice to have error strings collected in one place.
In that case we will get additional benefits:

It allows us to keep this set of messages under control.
It gives us a chance to internationalize them.
It slightly reduces the binary footprint.

From the other hand the developers have to add new strings into that enum.

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/20843012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16024 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-02 09:53:11 +00:00
hpayer@chromium.org
9e8058146e Cleaning up HAllocate space and double alignment selection.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/21074004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16000 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-01 08:34:36 +00:00
danno@chromium.org
4cdffbf4fe Improve instruction creating/adding shorthand in HGraphBuilder
Add multi-argument templates for New, NewUncasted, Add and AddUncasted that
call boilerplate HInstruction::New methods rather than the constructor directly.
This allows for automatic passing of the zone and context for instructions that
need them.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/21356002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15990 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-31 14:58:39 +00:00
bmeurer@chromium.org
8c0f2eabbf Stores to external references don't need write barriers.
This also applies to stores with an external field representation.

Also cleans up the CreateAllocationSiteStub.

R=mstarzinger@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/21357003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15989 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-31 13:45:51 +00:00
bmeurer@chromium.org
8823e8e89f Get rid of HLinkObjectInList.
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/21345002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15983 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-31 10:47:44 +00:00
hpayer@chromium.org
b6a6fc76dc Allocation space decisions are precisely made in hydrogen.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/21089006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15970 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-31 07:03:16 +00:00
titzer@chromium.org
5dd517ff3b Move representation into HObjectAccess and remove from HLoadNamedField and HStoreNamedField.
BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/18503006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15875 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-25 09:37:52 +00:00
titzer@chromium.org
54cc492225 Add --trace-hydrogen-stubs flag (default = false).
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/20084003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15859 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-24 12:38:52 +00:00
danno@chromium.org
8cbbdd066c Unify SoftDeoptimize and Deoptimize hydrogen instructions
BUG=chromium:258519
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/19528003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15827 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-23 13:35:10 +00:00
verwaest@chromium.org
babce318d1 Eliminate map checks of constant values.
R=ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/19954005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15819 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-23 09:18:42 +00:00
bmeurer@chromium.org
c4f77b3e40 Turn ToNumberStub into a hydrogen code stub.
R=danno@chromium.org, hpayer@chromium.org

Review URL: https://codereview.chromium.org/19793005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15814 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-23 06:02:54 +00:00
verwaest@chromium.org
46ca825561 Avoid repeated rewrites of global store to constant IC due to store of same value.
R=mvstanton@chromium.org

Review URL: https://chromiumcodereview.appspot.com/19663007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15806 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-22 14:15:58 +00:00
mvstanton@chromium.org
ec8c6f4692 Rename AllocationSiteInfo to AllocationMemento
This is just a rename change with the exception of a bug found along the way in
CodeStubGraphBuilder<FastCloneShallowArrayStub>::BuildCodeStub(). There, the
intent is to get the boilerplate object from an AllocationSite. But the wrong
HObjectAccess was used. It only succeeds because it happened to be the same
offset :).

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/19595004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15778 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-19 13:30:49 +00:00
mvstanton@chromium.org
b9f0c06ab2 The gc should be able to traverse all AllocationSites for decision making. The sites are threaded into a weak list. Special problems include:
* Allocations of AllocationSites occur in generated code, so generated code needs to be able to add to the list. For now I have a special hydrogen instruction, though it would be nice to use general purpose instructions.
* The snapshot contains AllocationSites, and these need to be re-threaded into the list on deserialization.

Something nice is that the AllocationSites are only created in old space, so a special new space visitor isn't required.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/18173013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15715 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-17 11:50:24 +00:00
bmeurer@chromium.org
410b4b2db4 Reland "Turn ElementsTransitionAndStore stub into a HydrogenCodeStub".
Fix the invalid array length check, replacing it with a check of
the elements pointer similar to TransitionElementsKindStub.
Refactor common code from ElementsTransitionAndStoreStub and TransitionElementsKindStub into BuildTransitionElementsKind() helper method.
Add test case for the MD5 computation that used to crash before,
and a small test case for the specific issue.

R=danno@chromium.org

Review URL: https://codereview.chromium.org/19367003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15713 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-17 11:37:20 +00:00
bmeurer@chromium.org
a72b5d6c4d Fix invalid array length check in TransitionElementsKindStub.
The stub used to check the length of the JS array to see if
there's a need to duplicate the elements backing store. This
way it will not duplicate the elements array when going from
double to object even if the elements array is not the empty
fixed array. Later on it will then store pointers into a
FixedDoubleArray.

The native code stub used to check whether elements points to
the empty_fixed_array singleton instead of testing the length.
The Hydrogen stub does that as well now.

R=danno@chromium.org

Review URL: https://codereview.chromium.org/19289009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15701 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-17 08:32:28 +00:00
danno@chromium.org
5713ca0d8f Revert 15635: Turn ElementsTransitionAndStore stub into a HydrogenCodeStub
Causes breakage in crypto-(sha1|md5) SunSpider

TBR=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/19214002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15674 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-15 15:23:52 +00:00
olivf@chromium.org
bdf4fc96b0 Encapsulate compare nil ic_state.
BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/18602003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15639 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-11 17:20:57 +00:00
bmeurer@chromium.org
6c13f097d6 Turn ElementsTransitionAndStore stub into a HydrogenCodeStub.
R=danno@chromium.org, mvstanton@chromium.org

Review URL: https://codereview.chromium.org/18881004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-11 14:29:00 +00:00
bmeurer@chromium.org
854e9b99fb Insert HTrapAllocationMemento only when required for TransitionElementsKindStub.
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/18292018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15594 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-10 13:24:51 +00:00
bmeurer@chromium.org
bdff9c70b8 Use BuildGrowElementsCapacity for the TransitionElementsKind stub.
R=danno@chromium.org, dslomov@chromium.org

Review URL: https://codereview.chromium.org/18876004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15589 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-10 12:19:02 +00:00
mvstanton@chromium.org
23695eb86e Rename AllocationSite::payload to AllocationSite::transition_info
BUG=
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/18749004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15556 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-08 15:00:12 +00:00
mvstanton@chromium.org
67d9051bcd Create AllocationSite objects, pointed to by AllocationSiteInfo.
This creates a platform where we can do additional things with allocation sites,
other than just aid in reducing array transitions.

BUG=
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/15094018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15545 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-08 10:02:16 +00:00
danno@chromium.org
bd50e6d38f Refactoring and cleanup of control instructions
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/18331004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15513 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-05 10:40:14 +00:00
danno@chromium.org
345cc98a25 Generate StoreGlobal stubs with Hydrogen
- Constants globals are inlined into Hydrogen code using code dependencies that invalidate the Crankshafted code when global PropertyCells or the global object change.
- The more general case generates code that is just as good as the hand-written assembly stubs on all platforms.

R=rossberg@chromium.org, ulan@chromium.org

Committed: http://code.google.com/p/v8/source/detail?r=15419

Review URL: https://codereview.chromium.org/16925008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15512 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-05 10:34:02 +00:00
olivf@chromium.org
240c7aced9 Convert UnaryOpStub to a HydrogenCodeStub
BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/18712002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15506 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-05 09:26:22 +00:00
olivf@chromium.org
950a372834 Revert "Convert UnaryOpStub to a HydrogenCodeStub"
The problem is the HCallConstantFunction which is not context sensitive, so we leak the builtin. We first need a Hydrogen version of __ IvokeBuiltin.

BUG=
R=danno@chromium.org, machenbach@chromium.org

Review URL: https://codereview.chromium.org/18650003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15486 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-03 19:57:25 +00:00
olivf@chromium.org
0d2fed8f26 Convert UnaryOpStub to a HydrogenCodeStub
BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/17229005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15473 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-03 14:56:53 +00:00
titzer@chromium.org
63abcc515b Remove HFixedArrayBaseLength instruction and replace with regular HLoadNamedField.
BUG=

Review URL: https://codereview.chromium.org/18508002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15454 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-02 15:31:17 +00:00
danno@chromium.org
77c20c30a3 Revert r15419: "Generate StoreGlobal stubs with Hydrogen"
TBR=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/18357004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15427 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-01 15:12:21 +00:00
danno@chromium.org
a3bce19868 Generate StoreGlobal stubs with Hydrogen
- Constants globals are inlined into Hydrogen code using code dependencies that invalidate the Crankshafted code when global PropertyCells or the global object change.
- The more general case generates code that is just as good as the hand-written assembly stubs on all platforms.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/16925008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15419 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-01 13:22:13 +00:00
mvstanton@chromium.org
83519ec87a Hydrogen array constructor cleanup and improvements
* Cleanup of LCallNewArray::PrintDataTo() method
* Created HCallNewArray::PrintDataTo()
* Created many more tests in array-constructor-feedback.js
* Removed redundant instructions in
  GenerateRecordCallTarget
* Bugfix in CreateArrayDispatchOneArgument: on a call to
  new Array(0), we'd like to set the type feedback cell to
  a packed elements kind, but we shouldn't do it if the
  cell contains the megamorphic sentinel.
* When used from crankshaft, ArrayConstructorStubs can
  avoid verifying that the function being called is the
  array function from the current native context, relying
  instead on the fact that crankshaft issues an
  HCheckFunction to protect the constructor call. (this
  new minor key is used in LCodeGen::DoCallNewArray(), and
  influences code generation in
  CodeStubGraphBuilderBase::BuildArrayConstructor()).
* Optimization: the array constructor specialized for
  FAST_SMI_ELEMENTS can save some instructions by looking
  up the correct map on the passed in constructor, rather
  than indexing into the array of cached maps per element
  kind.

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/17091002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15383 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-28 13:16:14 +00:00
danno@chromium.org
f95ddc8891 Add templatized convienience functions for adding hydrogen instructions
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/18050004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15360 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-27 14:36:14 +00:00
verwaest@chromium.org
1b5835f69f Remove manual representation annotations for constants.
R=mstarzinger@chromium.org

Review URL: https://chromiumcodereview.appspot.com/16818016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15130 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-13 17:38:10 +00:00
rossberg@chromium.org
94f651bc1b Migrate Compare ICs to new type rep
(Does not yet use common AST expression type field.)

R=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/16361015

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15093 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-12 17:20:37 +00:00
olivf@chromium.org
946ea53264 use compare nil ic only for non-strict equality. strict-equality only has one check and cannot deopt. should therefore not be part of the stub.
BUG=
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/16732002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15045 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-10 15:47:23 +00:00
olivf@chromium.org
3a26dda780 revert accidental dcommit
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15044 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-10 15:43:40 +00:00
olivf@chromium.org
73b7ab9ae6 remove equality kind from compare nil ic
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15042 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-10 15:30:02 +00:00
mvstanton@chromium.org
56f3337ebc Adapt hydrogen-based Array constructor to also support InternalArray and function call
versions.

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/14576005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14948 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-05 10:43:18 +00:00
mvstanton@chromium.org
8a02fd3be3 Turn off allocation site info for crankshafted array constructor calls.
Once we crankshaft a method, we should turn off allocation site info for
constructed arrays. Additionally, the semantics for doing this were
awkward because the constructed array code stubs get an
AllocationSiteMode as a minor key, but it's used as a permission to
determine the final mode locally based on ElementsKind. I refactored
this to a simpler boolean for override or local control.

BUG=
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/16206007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14934 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-04 12:48:51 +00:00
yangguo@chromium.org
7f8a3d803c Make assertion scopes thread safe.
R=svenpanne@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/15691017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14919 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-03 15:32:22 +00:00
olivf@chromium.org
88755fb22a Convert ToBooleanStub to a HydrogenStub. Currently just using the existing HBranch instruction, which is still fully implemented in Lithium. Will refactor HBranch in a next CL.
[
CL breaks mips, since the deoptimizer in Deoptimizer::DoTranslateCommand(...) does not convert smis passed to the stub via a0.
]

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/15302004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14886 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-29 14:49:28 +00:00
mstarzinger@chromium.org
39e9a9030d Make sure IfBuilder::Return clears the current block.
This is a follow-up to an earlier bugfix. It doesn't yet trigger any
known bugs but is a safety precaution since IfBuilder usage increases.

R=jkummerow@chromium.org
BUG=chromium:243868

Review URL: https://codereview.chromium.org/15832005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14882 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-29 12:36:41 +00:00
verwaest@chromium.org
7d32e7451f Don't explicitly pass requested representations to constants; implement ConstantS
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15932011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14874 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-29 10:47:55 +00:00
hpayer@chromium.org
44fd249cd2 Refacturing BuildEmitDeepCopy.
BUG=

Review URL: https://codereview.chromium.org/15700007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14845 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-28 10:44:21 +00:00
verwaest@chromium.org
a64b7f926f Remove HCheckSmi, LCheckSmi and rename LCheckSmiAndReturn to LCheckSmi.
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15841007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14816 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-27 09:58:46 +00:00
verwaest@chromium.org
8fc2101ce0 Fix build
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15881004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14815 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-27 08:50:52 +00:00
verwaest@chromium.org
3cb8f95c91 Tag smi-constants as smi. This also fixes code that copies holes into arrays.
BUG=
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15861009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14814 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-27 08:43:58 +00:00
olivf@chromium.org
7aa74c3dc5 fix some names in compare nil ic
The compare nil ic introduces a mechanism called lightweight miss. The
idea is to insert a direct call to the miss handler instead of going
through the deopt routine. This can be use for uninitialized stubs, to
directly jump to the runtime.

This cl
1. Cleans up naming
2. localizes the lightweight miss functionality on the only user (compare nil ic)
3. fixes DoCodegen, to actually call the correct method (uninitialized vs initialized)

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/15806005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14798 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-24 11:44:55 +00:00
verwaest@chromium.org
c9a22a3c8d Replace tagged keys for fast access by smi, and use smi in boundscheck.
BUG=
R=jkummerow@chromium.org

Review URL: https://chromiumcodereview.appspot.com/15763004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14794 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-24 08:52:35 +00:00
titzer@chromium.org
94b4240f5c Introduce ObjectAccess, which is used by LoadNamedField and StoreNamedField to denote what parts of an object are referred to by a given load or store. Refactor HGraphBuilder to use ObjectAccess, which removes the need to manually set GVN flags and simplifies the code as well.
Review URL: https://codereview.chromium.org/14284010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14791 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-24 08:38:21 +00:00
jkummerow@chromium.org
f691f48c14 Avoid creating duplicate entries for a value when merging HSimulates
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/15660003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14766 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-23 09:17:01 +00:00
hpayer@chromium.org
2cbc81a5ce Move global pretenuring flag check to ShouldGloballyPretenure().
BUG=

Review URL: https://codereview.chromium.org/15734007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14763 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-23 08:17:03 +00:00
danno@chromium.org
05e8e0e7b4 Elide hole checks on KeyedLoads of holey double arrays
Improves NavierStokes by about 5%

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/15014020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14630 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-13 07:35:26 +00:00
mvstanton@chromium.org
d7b013de57 Becuase of cross-context calls, hydrogen-based Array constructor needs to ensure
the array constructor pointer passed in matches that of the current context.

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/14846017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14581 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-07 21:01:53 +00:00
verwaest@chromium.org
46d39cabd6 Fix polymorphic to monomorphic load to take representation into account.
Review URL: https://chromiumcodereview.appspot.com/14966005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14565 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-07 10:32:23 +00:00
titzer@chromium.org
dbcd0e925a Remove HLoadElements instruction and replace with use of more general HLoadNamedField. This removes the need for both the hydrogen and lithium instructions and allows both kinds of loads to be GVN'd.
Review URL: https://codereview.chromium.org/14556020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14542 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-03 12:21:16 +00:00
verwaest@chromium.org
51dded6e9c Turn the load field code stub into a hydrogen code stub.
Review URL: https://chromiumcodereview.appspot.com/14847004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14526 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-02 16:32:47 +00:00
verwaest@chromium.org
99e17bb12a Track storage types of instance variables.
Review URL: https://chromiumcodereview.appspot.com/14146005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14464 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-04-26 15:30:41 +00:00
mvstanton@chromium.org
e751ad06d9 Constructed arrays can be created with Hydrogen code stubs. The feature is still off by default (--optimize-constructed-arrays).
BUG=

Review URL: https://codereview.chromium.org/12385014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14441 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-04-25 16:00:32 +00:00
danno@chromium.org
f8ddf3a262 Add monomorphic CompareNilICs and Crankshaft support
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/14367018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14407 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-04-24 11:32:17 +00:00
mstarzinger@chromium.org
6f595c231a Replace CheckBuilder with IfBuilder everywhere.
This deprecates and removes the CheckBuilder which is less powerful
than the generic IfBuilder which can deopt as well by now.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/14075013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14372 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-04-22 11:15:43 +00:00