Commit Graph

395 Commits

Author SHA1 Message Date
mstarzinger
d016387e81 [turbofan] Add test coverage for deopt within try-catch.
This just contains test, no fixes. Note that some of the tests are
still disabled because they either fail or we don't want ClusterFuzz
to pick up the flag yet.

R=jarin@chromium.org
TEST=cctest/test-run-jsexceptions/Deopt,mjsunit/compiler/try-deopt

Review URL: https://codereview.chromium.org/972943004

Cr-Commit-Position: refs/heads/master@{#26968}
2015-03-03 17:23:42 +00:00
Jaroslav Sevcik
f0b11873a3 [turbofan] Fix deferred replacement in simplified lowering.
Deferred replacement must also replace in the pending node vector.

BUG=chromium:463056
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/966423002

Cr-Commit-Position: refs/heads/master@{#26938}
2015-03-02 12:49:49 +00:00
titzer
d050c331eb [turbofan] Simplify context specialization and fix for OSR.
AstGraphBuilder puts a constant context in from the beginning.
Also fix bug in merging contexts in environment.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/934293002

Cr-Commit-Position: refs/heads/master@{#26745}
2015-02-19 11:36:50 +00:00
titzer
71765afff9 [turbofan] Rename context stack as part of the environment for OSR.
R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/921083004

Cr-Commit-Position: refs/heads/master@{#26705}
2015-02-17 18:10:11 +00:00
jarin
0b8063cec8 [turbofan] Avoid ToNumber conversions if they could deoptimize.
BUG=chromium:454158
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/922623002

Cr-Commit-Position: refs/heads/master@{#26661}
2015-02-16 12:59:20 +00:00
titzer
31637fb396 [turbofan] Use heavy-handed graph duplication to do loop peeling for OSR.
BUG=

Review URL: https://codereview.chromium.org/898353002

Cr-Commit-Position: refs/heads/master@{#26576}
2015-02-11 13:26:45 +00:00
titzer
1db760de7d Reduce the number of iterations in some OSR tests by using an explicit %OptimizeOsr().
R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/913463002

Cr-Commit-Position: refs/heads/master@{#26540}
2015-02-10 09:53:22 +00:00
titzer
4c302ca290 Make it easier to test OSR with %OptimizeOsr() runtime call.
This call triggers OSR for the current function. And also allows explicitly testing OSR on the top-level code.

R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/906243002

Cr-Commit-Position: refs/heads/master@{#26523}
2015-02-09 12:47:43 +00:00
titzer
79cad15cb0 Speed up tests for OSR of for-in and for-of loops.
R=marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/889293003

Cr-Commit-Position: refs/heads/master@{#26378}
2015-02-02 11:52:07 +00:00
titzer
e25a0f7e4e [turbofan] Fix usage of ThisFunction parameter in OSR.
R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/892593002

Cr-Commit-Position: refs/heads/master@{#26355}
2015-01-30 13:09:44 +00:00
titzer
11311c083a [turbofan] Fix OSR compilations of for-in.
R=mstarzinger@chromium.org
LOG=Y
BUG=

Review URL: https://codereview.chromium.org/890543002

Cr-Commit-Position: refs/heads/master@{#26333}
2015-01-29 17:40:23 +00:00
jarin
a4b163a940 [turbofan] Make sure there is space for lazy deopt patching before the constant pool.
BUG=chromium:446647
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/874863003

Cr-Commit-Position: refs/heads/master@{#26327}
2015-01-29 15:30:32 +00:00
titzer
f5479ca675 [turbofan] Gracefully bail out if OSR encounters a loop too deeply nested.
R=jarin@chromium.org
BUG=

Review URL: https://codereview.chromium.org/877553007

Cr-Commit-Position: refs/heads/master@{#26318}
2015-01-29 09:46:36 +00:00
jarin
489b6f7c60 [turbofan] Add missing deopt for the assignment in the for-in statement.
BUG=chromium:416359
LOG=n
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/881303002

Cr-Commit-Position: refs/heads/master@{#26309}
2015-01-28 16:16:24 +00:00
jarin
b4a4c4c591 [turbofan] Only replace nodes eagerly during simplified lowering if the types stay the same.
BUG=chromium:452427
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/871373010

Cr-Commit-Position: refs/heads/master@{#26286}
2015-01-27 09:27:37 +00:00
jarin
4c79f55c32 [turbofan] Handle cyclic dependencies in context typing.
BUG=chromium:451012
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/874983002

Cr-Commit-Position: refs/heads/master@{#26281}
2015-01-27 06:57:41 +00:00
mstarzinger
209b771431 Enable test coverage for test coverage.
R=titzer@chromium.org
TEST=mjsunit/compiler/opt-next-call-turbo

Review URL: https://codereview.chromium.org/822673003

Cr-Commit-Position: refs/heads/master@{#26192}
2015-01-21 14:07:19 +00:00
Michael Starzinger
cecd89c2c6 Disable new test failing since 79748e3f7c.
TBR=titzer@chromium.org
TEST=mjsunit/compiler/opt-next-call-turbo

Review URL: https://codereview.chromium.org/862533002

Cr-Commit-Position: refs/heads/master@{#26143}
2015-01-19 16:17:07 +00:00
mstarzinger
79748e3f7c Remove overzealous check from %OptimizeFunctionOnNextCall.
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/863443003

Cr-Commit-Position: refs/heads/master@{#26142}
2015-01-19 15:52:00 +00:00
jarin
ac04d777d6 [turbofan] Allow deoptimization for JSToNumber operator.
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/841443004

Cr-Commit-Position: refs/heads/master@{#26053}
2015-01-14 13:09:32 +00:00
bmeurer
70b32e4b8a [turbofan] Fix truncation/representation sloppiness wrt. bool/bit.
TEST=cctest,mjsunit,unittests
BUG=v8:3812
LOG=y

Review URL: https://codereview.chromium.org/850013003

Cr-Commit-Position: refs/heads/master@{#26051}
2015-01-14 12:06:56 +00:00
jarin
527e19afd5 [turbofan] Add missing deopt.
BUG=chromium:447567
LOG=n
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/809463005

Cr-Commit-Position: refs/heads/master@{#26033}
2015-01-13 08:40:54 +00:00
titzer
159b14172f [turbofan] Implement OSR for outer loops.
R=bmeurer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/809333002

Cr-Commit-Position: refs/heads/master@{#26020}
2015-01-12 11:39:58 +00:00
bmeurer
fec1bba852 [turbofan] Correctify representation changes to bit.
TEST=cctest/test-representation-change,unittests,mjsunit/compiler/regress-bit-number-constant
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/840953003

Cr-Commit-Position: refs/heads/master@{#25987}
2015-01-08 09:48:41 +00:00
bmeurer
d1c1a3c48f [turbofan] Fix bit representation of NumberConstant.
TEST=mjsunit/compiler/regress-bit-number-constant

Review URL: https://codereview.chromium.org/839813002

Cr-Commit-Position: refs/heads/master@{#25979}
2015-01-07 15:44:22 +00:00
jarin
80a7be5cd9 Restrict representation inference to avoid truncation of phi inputs.
BUG=chromium:446778
LOG=N

Review URL: https://codereview.chromium.org/837153002

Cr-Commit-Position: refs/heads/master@{#25967}
2015-01-07 11:38:54 +00:00
jarin
a9716d9840 Make control reducer revisit newly introduced merges.
TEST=mjsunit/compiler/regress-445876
BUG=chromium:445876
LOG=N

Review URL: https://codereview.chromium.org/830293003

Cr-Commit-Position: refs/heads/master@{#25959}
2015-01-05 16:35:34 +00:00
Benedikt Meurer
17a180842e [turbofan] Don't crash when typing load from a Uint8ClampedArray.
TEST=mjsunit/compiler/regress-446156
BUG=chromium:446156
LOG=y
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/835883003

Cr-Commit-Position: refs/heads/master@{#25957}
2015-01-05 13:43:47 +00:00
jarin
bdf446f590 Do not reduce effect phis for loops.
This prevents eliminating effectful statements before the loop.

BUG=

Review URL: https://codereview.chromium.org/830923002

Cr-Commit-Position: refs/heads/master@{#25953}
2015-01-03 12:46:00 +00:00
bmeurer
fb2643c858 [turbofan] Truncation of Bit/Word8/16 to Word32 is a no-op.
TEST=mjsunit/compiler/regress-445859
BUG=chromium:445859
LOG=y
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/828313002

Cr-Commit-Position: refs/heads/master@{#25951}
2015-01-02 10:39:10 +00:00
bmeurer
cf866b7c61 [x64] Rearrange code for OOB integer loads.
We cannot just clear the result register optimistically, because the
register allocator might assign the same register to result and buffer.

TEST=mjsunit/compiler/regress-445858
BUG=chromium:445858
LOG=y
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/828303002

Cr-Commit-Position: refs/heads/master@{#25950}
2015-01-02 10:15:40 +00:00
bmeurer
a64ac4575a Fix %NeverOptimizeFunction() intrinsic.
Set a valid reason for disabling optimization when using
%NeverOptimizeFunction.

TEST=mjsunit/compiler/regress-445732
BUG=chromium:445732
LOG=y
TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/832003002

Cr-Commit-Position: refs/heads/master@{#25949}
2015-01-02 08:18:01 +00:00
bmeurer
ef41f70684 [turbofan] Fix invalid bounds check with overflowing offset.
TEST=mjsunit/compiler/regress-445267
BUG=chromium:445267
LOG=y

Review URL: https://codereview.chromium.org/825403002

Cr-Commit-Position: refs/heads/master@{#25945}
2014-12-29 10:01:15 +00:00
bmeurer
b5e8dd0e12 [turbofan] Raise max virtual registers and call parameter limit.
Change InstructionOperand to use a 64-bit field for encoding the operand
information instead of the 32-bit field that was used before. Ideally we
wouldn't use the Zone-allocated bit field at all, and use an integer
instead of the pointer; but that requires fixing the register allocator
first, which will take some time.

TEST=mjsunit/compiler/regress-3786
BUG=v8:3786
LOG=y

Review URL: https://codereview.chromium.org/826673002

Cr-Commit-Position: refs/heads/master@{#25941}
2014-12-25 18:18:04 +00:00
Benedikt Meurer
3f00ce2d59 [turbofan] Fix missing ChangeUint32ToUint64 in lowering of LoadBuffer.
TEST=mjsunit/compiler/regress-444695
BUG=chromium:444695
LOG=y
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/824843002

Cr-Commit-Position: refs/heads/master@{#25932}
2014-12-23 06:54:00 +00:00
Benedikt Meurer
65e69497ab [turbofan] Correctify lowering of Uint8ClampedArray buffer access.
TEST=mjsunit/compiler/regress-444508.js
BUG=chromium:444508
LOG=y
R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/794013004

Cr-Commit-Position: refs/heads/master@{#25916}
2014-12-22 08:27:59 +00:00
Benedikt Meurer
f7e4689061 [turbofan] Fix unsafe out-of-bounds check for checked loads/stores.
BUG=chromium:443744
LOG=y
TEST=mjsunit/compiler/regress-443744
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/804993004

Cr-Commit-Position: refs/heads/master@{#25901}
2014-12-19 12:53:29 +00:00
Benedikt Meurer
14409abc22 [turbofan] Quickfix for invalid number truncation of typed array loads.
TEST=mjsunit/compiler/regress-int32array-outofbounds-nan
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/803483002

Cr-Commit-Position: refs/heads/master@{#25793}
2014-12-12 10:45:38 +00:00
Benedikt Meurer
19418d76df [x64] Fix optimization for certain checked load/stores.
BUG=chromium:439743
LOG=y
TEST=mjsunit
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/733893008

Cr-Commit-Position: refs/heads/master@{#25731}
2014-12-10 07:49:12 +00:00
Benedikt Meurer
48a6766e78 [x86] Disable invalid checked load/store optimization.
TEST=mjsunit/compiler/regress-lena
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/784153006

Cr-Commit-Position: refs/heads/master@{#25722}
2014-12-09 14:16:34 +00:00
jarin
322bb23e82 [turbofan] Insert appropriate conversions for typed array stores.
BUG=

Review URL: https://codereview.chromium.org/758643003

Cr-Commit-Position: refs/heads/master@{#25496}
2014-11-25 08:40:29 +00:00
svenpanne
be0fcaa2bc Added test cases for truncating stores.
Review URL: https://codereview.chromium.org/741643003

Cr-Commit-Position: refs/heads/master@{#25474}
2014-11-24 10:42:26 +00:00
yangguo
61bee5c898 Correctly escape RegExp source.
R=ulan@chromium.org
BUG=v8:3229
LOG=N

Review URL: https://codereview.chromium.org/736003002

Cr-Commit-Position: refs/heads/master@{#25457}
2014-11-21 10:50:24 +00:00
danno@chromium.org
9a58807030 [turbofan] Optimize add operations to use 'leal' instruction on x64
Add MemoryOperandMatcher that recognizes node clusters in the form
[%r1 + %r2*SCALE + OFFSET] and explicit support in the x64 Int32Add
selector to use it to translate complex adds to 'leal' instructions.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/704713003

Cr-Commit-Position: refs/heads/master@{#25223}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25223 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-11-07 16:47:45 +00:00
jarin@chromium.org
364cec0034 [turbofan] Fix deoptimization of uint8, uint16 inputs.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/707443003

Cr-Commit-Position: refs/heads/master@{#25214}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25214 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-11-07 14:07:45 +00:00
dcarney@chromium.org
a350f0d608 [turbofan] phis cannot take registers as inputs
BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/704153002

Cr-Commit-Position: refs/heads/master@{#25191}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25191 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-11-06 12:56:44 +00:00
bmeurer@chromium.org
017c518321 [x86] Fix register constraints for multiply high and modulus.
R=jarin@chromium.org
TEST=mjsunit/compiler/regress-register-allocator2

Review URL: https://codereview.chromium.org/697053002

Cr-Commit-Position: refs/heads/master@{#25054}
git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@25054 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-11-03 06:28:12 +00:00
bmeurer@chromium.org
548fb46331 [x86] Fix register constraints for multiply-high.
TEST=mjsunit/compiler,unittests
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/671393002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24862 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-24 09:36:40 +00:00
verwaest@chromium.org
23868b419c Optimize Function.prototype.call
BUG=
R=verwaest@chromium.org, jarin@chromium.org, jkummerow@chromium.org

Review URL: https://codereview.chromium.org/588573002

Patch from Petka Antonov <p.antonov@partner.samsung.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24631 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-15 12:22:15 +00:00
jkummerow@chromium.org
b6f82c734c Remove FLAG_opt_safe_uint32_operations.
It has been turned on by default for a long time, and hydrogenized BinaryOpStubs actually depend on it being turned on.

BUG=v8:3487
LOG=n
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/630023002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24415 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-10-06 13:15:23 +00:00
titzer@chromium.org
cbf66711af Minor compiler pipeline refactoring. Inline UpdateSharedFunctionInfo and make Parser::Parse responsible for setting the strict mode of the CompilationInfo.
R=mstarzinger@chromium.org
BUG=

Review URL: https://codereview.chromium.org/555553003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24001 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-09-17 12:34:46 +00:00
titzer@chromium.org
4c53bb086e Fix more fallout from making OptimizeFunctionOnNextCall work as advertised.
R=jarin@chromium.org
BUG=411262

Review URL: https://codereview.chromium.org/544213002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23745 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-09-05 15:31:33 +00:00
titzer@chromium.org
67b89f9db6 Fix %OptimizeFunctionOnNextCall to actually work when the function has not yet been compiled.
R=yangguo@chromium.org
BUG=

Review URL: https://codereview.chromium.org/543643002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23687 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-09-04 11:27:20 +00:00
Jacob.Bramley@arm.com
da48f1246c ARM64: Fix SHR logic error.
The `right == 0` checks only worked for `0 <= right < 32`. This patch
replaces the checks with simple tests for negative results.

The attached test can detect this error, but the test relies on a broken
flag (--noopt-safe-uint32-operations), so it is skipped for now. See
issue 3487 for details.

BUG=
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/487913005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23243 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-08-20 14:58:18 +00:00
mstarzinger@chromium.org
7a4054b7d7 Allow inlining of functions containing %_Arguments.
R=svenpanne@chromium.org
TEST=mjsunit/compiler/inline-arguments

Review URL: https://codereview.chromium.org/356773002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22060 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-06-27 11:04:35 +00:00
dcarney@chromium.org
bd3f8a524e Revert "Optimize Function.prototype.call"
This reverts commit r21840.

R=danno@chromium.org
LOG=y
BUG=chromium:385565

Review URL: https://codereview.chromium.org/347573002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21887 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-06-18 14:04:41 +00:00
verwaest@chromium.org
7005abf03b Optimize Function.prototype.call
- May inline the function, or call it directly, instead of going through call
- Supports arguments object escaping when it escapes to builtins (preparation for slice.call(arguments, ...) optimization)
- Both .call and .apply now support inlining when calling builtins indirectly

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

Review URL: https://codereview.chromium.org/335683002

Patch from Petka Antonov <p.antonov@partner.samsung.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21840 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-06-13 12:52:23 +00:00
hpayer@chromium.org
de21c8a245 Simplify ConfigureHeap and change --max_new_space_size to --max_semi_space_size.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/271843005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21204 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-05-09 08:38:27 +00:00
hpayer@chromium.org
dde49c9dc3 Set max new space size in tests to proper MB value.
Revert "Limit old space size in test which require a large new space."

This reverts commit r21103.

Revert "Remove max space limits in tests."

This reverts commit r21104.

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/263103006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@21149 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-05-05 16:48:33 +00:00
hpayer@chromium.org
56d0b9757e Remove max space limits in tests.
BUG=

Review URL: https://codereview.chromium.org/263703003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@21104 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-04-30 19:32:47 +00:00
hpayer@chromium.org
3dd05f8fc7 Limit old space size in test which require a large new space.
BUG=

Review URL: https://codereview.chromium.org/265673003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@21103 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-04-30 18:57:25 +00:00
dslomov@chromium.org
f66af4feb4 Refactor optimized in hydrogen only runtime functions.
This splits all runtime function into 3 categories:
1) RUNTIME: implemented in runtime and called from both full and optimized code.
2) RUNTIME_HIDDEN: implemented in runtime, never called directly from JS builtins.
3) INLINE: inlined in both full and optimized code
4) INLINE_OPTIMIZED: inlined in optimized code, implemented in runtime for full code.

R=yangguo@chromium.org, yannguo@chromium.org

Review URL: https://codereview.chromium.org/209353006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20252 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-03-25 14:26:55 +00:00
titzer@chromium.org
3c31102025 First implementation of store elimination.
BUG=
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/100253004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20224 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-03-25 09:06:16 +00:00
ulan@chromium.org
2cb4e78e74 Fix mjsunit/compiler/concurrent-invalidate-transition-map.js test.
BUG=v8:3156
LOG=N
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/180053003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20009 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-03-17 17:00:48 +00:00
svenpanne@chromium.org
be328fd4ce Disable special handling of flooring division by constant until it is fixed for real.
Added a test to check the various division-like operations more exhaustively.

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

Review URL: https://codereview.chromium.org/194863002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19852 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-03-12 14:28:59 +00:00
ishell@chromium.org
6bb57517c0 Restore of compare-objeq-elim test accidentally removed in r19229.
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/162903005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19354 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-02-13 12:22:24 +00:00
ulan@chromium.org
e95bc7eec8 Merge experimental/a64 to bleeding_edge.
BUG=v8:3113
LOG=Y
R=jochen@chromium.org, rmcilroy@chromium.org, rodolph.perfetta@arm.com

Review URL: https://codereview.chromium.org/148293020

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19311 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-02-12 09:19:30 +00:00
ishell@chromium.org
994f0f6dda Fix for a smi stores optimization on x64 with a test case.
BUG=338425
LOG=N
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/152923006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19288 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-02-11 16:02:18 +00:00
ishell@chromium.org
f46da9d43b Reland of r19102: Check elimination improvement: propagation of state through phis is supported, CheckMap narrowing implemented with tests.
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/146623006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19229 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-02-10 15:32:54 +00:00
ishell@chromium.org
2aa17c6e62 Load elimination fix: load should not be replaced with another load if the former is not dominated by the latter.
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/151333003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18985 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-31 12:03:32 +00:00
jarin@chromium.org
99ce5a2484 The current
version is passing all the existing test + a bunch of new tests
(packaged in the change list, too).

The patch extends the SlotRef object to describe captured and duplicated
objects. Since the SlotRefs are not independent of each other anymore,
there is a new SlotRefValueBuilder class that stores the SlotRefs and
later materializes the objects from the SlotRefs.

Note that unlike the previous implementation of SlotRefs, we now build
the SlotRef entries for the entire frame, not just the particular
function.  This is because duplicate objects might refer to previous
captured objects (that might live inside other inlined function's part
of the frame).

We also need to store the materialized objects between other potential
invocations of the same arguments object so that we materialize each
captured object at most once.  The materialized objects of frames live
in the new MaterielizedObjectStore object (contained in Isolate),
indexed by the frame's FP address.  Each argument materialization (and
deoptimization) tries to lookup its captured objects in the store before
building new ones.  Deoptimization also removes the materialized objects
from the store. We also schedule a lazy deopt to be sure that we always
get rid of the materialized objects and that the optmized function
adopts the materialized objects (instead of happily computing with its
captured representations).

Concerns:

- Is the FP address the right key for a frame? (Note that deoptimizer's
representation of frame is different from the argument object
materializer's one - it is not easy to find common ground.)

- Performance is suboptimal in several places, but a quick local run of
benchmarks does not seem to show a perf hit. Examples of possible
improvements: smarter generation of SlotRefs (build other functions'
SlotRefs only for captured objects and only if necessary), smarter
lookup of stored materialized objects.

- Ideally, we would like to share the code for argument materialization
with deoptimizer's materializer.  However, the supporting data structures
(mainly the frame descriptor) are quite different in each case, so it
looks more like a separate project.

Thanks for any feedback.

R=danno@chromium.org, mstarzinger@chromium.org
LOG=N
BUG=

Committed: https://code.google.com/p/v8/source/detail?r=18918

Review URL: https://codereview.chromium.org/103243005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18936 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-30 10:33:53 +00:00
jarin@chromium.org
ec51f26b9e Revert "Captured arguments object materialization"
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/130803009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18923 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-29 15:49:48 +00:00
jarin@chromium.org
868ad01ecb This is a preview of the captured arguments object materialization,
mostly to make sure that it is going in the right direction. The current
version is passing all the existing test + a bunch of new tests
(packaged in the change list, too).

The patch extends the SlotRef object to describe captured and duplicated
objects. Since the SlotRefs are not independent of each other anymore,
there is a new SlotRefValueBuilder class that stores the SlotRefs and
later materializes the objects from the SlotRefs.

Note that unlike the previous implementation of SlotRefs, we now build
the SlotRef entries for the entire frame, not just the particular
function.  This is because duplicate objects might refer to previous
captured objects (that might live inside other inlined function's part
of the frame).

We also need to store the materialized objects between other potential
invocations of the same arguments object so that we materialize each
captured object at most once.  The materialized objects of frames live
in the new MaterielizedObjectStore object (contained in Isolate),
indexed by the frame's FP address.  Each argument materialization (and
deoptimization) tries to lookup its captured objects in the store before
building new ones.  Deoptimization also removes the materialized objects
from the store. We also schedule a lazy deopt to be sure that we always
get rid of the materialized objects and that the optmized function
adopts the materialized objects (instead of happily computing with its
captured representations).

Concerns:

- Is there a simpler/more correct way to store the already-materialized
objects? (At the moment there is a custom root reference to JSArray
containing frames' FixedArrays with their captured objects.)

- Is the FP address the right key for a frame? (Note that deoptimizer's
representation of frame is different from the argument object
materializer's one - it is not easy to find common ground.)

- Performance is suboptimal in several places, but a quick local run of
benchmarks does not seem to show a perf hit. Examples of possible
improvements: smarter generation of SlotRefs (build other functions'
SlotRefs only for captured objects and only if necessary), smarter
lookup of stored materialized objects.

- Ideally, we would like to share the code for argument materialization
with deoptimizer's materializer.  However, the supporting data structures
(mainly the frame descriptor) are quite different in each case, so it
looks more like a separate project.

Thanks for any feedback.

R=mstarzinger@chromium.org, danno@chromium.org
LOG=N
BUG=

Review URL: https://codereview.chromium.org/103243005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18918 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-29 15:14:15 +00:00
ishell@chromium.org
d330d4801d Load elimination fix with a test case.
R=titzer@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/143413019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18884 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-28 16:45:04 +00:00
machenbach@chromium.org
cde3ed1fe3 Speed up some mjsunit test cases and clean up test expectations for arm and mips.
Many skipped test cases already run very fast. Removing the corresponding expectations.

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

Review URL: https://codereview.chromium.org/138503008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18812 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-24 11:36:45 +00:00
jarin@chromium.org
acf24331e3 Fixed Lithium environment generation bug for captured objects (created
by escape analysis). Added several tests that expose the bug.

Summary:
LCodegen::AddToTranslation assumes that Lithium environments are
generated by depth-first traversal, but LChunkBuilder::CreateEnvironment
was generating them in breadth-first fashion. This fixes the
CreateEnvironment to traverse the captured objects depth-first.

Note:
It might be worth considering representing LEnvironment by a list
with the same order as the serialized translation representation
rather than having two lists with a subtle relationship between
them (and then serialize in a slightly different order again).

R=titzer@chromium.org, mstarzinger@chromium.org
LOG=N
BUG=

Review URL: https://codereview.chromium.org/93803003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18470 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2014-01-07 14:36:26 +00:00
yangguo@chromium.org
2a4be7067c Refactor the compiling pipeline.
Goals:
 - easier to read, more suitable identifiers.
 - better distinction between compiling optimized/unoptimized code
 - compiler does not install code on the function.
 - easier to add features (e.g. caching optimized code for osr).
 - remove unnecessary code.

R=titzer@chromium.org

Review URL: https://codereview.chromium.org/110203002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18409 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-12-23 14:30:35 +00:00
titzer@chromium.org
be32761a67 Improve load elimination handling of transitioning stores.
BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/106973005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18388 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-12-20 12:12:41 +00:00
titzer@chromium.org
1f679a58f7 Improve check elimination with branch sensitivity on HCompareObjectEqAndBranch.
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/106733002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18377 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-12-19 17:42:21 +00:00
titzer@chromium.org
1d6710c933 Add some test cases with dead loops.
BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/98323004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18242 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-12-03 18:04:45 +00:00
titzer@chromium.org
16c4c14fac Check elimination: Learn from if(CompareMap(x)) on true branch.
BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/99043002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18210 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-12-02 18:34:33 +00:00
titzer@chromium.org
bbdd21ebb0 Fix load elimination: can only .Equals() GVN-able instructions.
BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/95193002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18133 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-28 15:27:42 +00:00
yangguo@chromium.org
afd8e5a305 Speed up long-running test cases.
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/85163003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18070 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-26 11:32:39 +00:00
yangguo@chromium.org
4716b292db Make some ARM test cases faster.
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/85473004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18069 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-26 10:43:44 +00:00
yangguo@chromium.org
aa3518a0f3 Make sure files end with exactly one new line and police this in presubmit.
The changes are (excluding presubmit.py) mechanical. I added the following
lines after the check and iterated the presubmit script until all errors
went away:

f = open(name, "w");
if contents.endswith('\n\n'):
  f.write(contents[0:-1])
else:
  f.write(contents + '\n')

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/82803005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-22 13:50:39 +00:00
yangguo@chromium.org
df9665032e Introduce %_IsMinusZero.
R=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/63423004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17639 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-11-12 11:53:13 +00:00
svenpanne@chromium.org
acb06df0e9 Tune mjsunit/compiler/expression-trees.
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/43703002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17404 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-28 07:24:19 +00:00
mstarzinger@chromium.org
be3ed75ff3 Fix materialization of captured objects with field tracking.
R=titzer@chromium.org
BUG=chromium:298990
TEST=mjsunit/compiler/escape-analysis-representation

Review URL: https://codereview.chromium.org/35133003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17321 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-22 13:48:54 +00:00
titzer@chromium.org
4d0f2cdd3b Implement global load elimination based on flow engine.
BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/27148004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-18 14:28:24 +00:00
yangguo@chromium.org
71ba8c5fb4 Retire concurrent recompilation delay for non-stress testing.
Instead, we block concurrent recompilation until unblocked. This makes
affected tests more predictable and run shorter.

R=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/26758003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@17199 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-10-14 14:15:22 +00:00
mstarzinger@chromium.org
9c3ffc4f58 Fix replaying of HCapturedObject for nested objects.
R=titzer@chromium.org
TEST=mjsunit/compiler/property-refs,mjsunit/compiler/escape-analysis

Review URL: https://codereview.chromium.org/24561002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16969 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-26 15:28:46 +00:00
titzer@chromium.org
957a6da2a2 Dynamically align OSR frames on ia32.
BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/23619076

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16798 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-18 16:05:39 +00:00
titzer@chromium.org
05797e77fd Implement local load/store elimination on basic blocks.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/24117004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16776 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-17 15:32:21 +00:00
svenpanne@chromium.org
8ed9b08e31 In the case of shift amounts with two constants and if their sum is equal 32, then shift can also be replaced with bit rotate.
R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/24095002

Patch from Bangfu Tao <bangfu.tao@samsung.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16735 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-16 13:51:14 +00:00
titzer@chromium.org
49d9555a97 Generate a custom OSR entrypoint for OSR compiles on all platforms, and transition to optimized code using the special entrypoint, instead of through the deoptimizer. Do not install the OSR compiled code as _the_ optimized code for a function.
Remove OSR-related stuff from deoptimizer.
BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/21340002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16599 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-09 16:34:40 +00:00
mstarzinger@chromium.org
ba48f3bd8c Consider out-of-bounds accesses as escaping uses.
R=titzer@chromium.org
TEST=mjsunit/compiler/escape-analysis

Review URL: https://codereview.chromium.org/23892007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16589 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-09 09:53:58 +00:00
yangguo@chromium.org
070e3b0af4 Introduce concurrent on-stack replacement.
Currently disabled behind --concurrent-osr.

R=titzer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/23710014

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16527 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-04 12:55:59 +00:00
mstarzinger@chromium.org
fa8a182208 Fix OSR to ignore phis without merge index in loop entry.
This fixes a corner case introduced by escape analysis where phis are
introduced in OSR loop entry blocks that don't have a merge index and
hence cannot contain OSR values.

R=titzer@chromium.org
TEST=mjsunit/compiler/escape-analysis

Review URL: https://codereview.chromium.org/23503025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16484 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-09-02 16:51:44 +00:00
jkummerow@chromium.org
9efb5cd23b Make VisitStatements() consistent among all AstVisitor implementations
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/23441018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16443 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-30 10:51:37 +00:00
mstarzinger@chromium.org
57ac971a78 Implement proper map checks of captured objects.
R=verwaest@chromium.org
TEST=mjsunit/compiler/escape-analysis

Review URL: https://codereview.chromium.org/23697002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16403 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-28 14:16:57 +00:00
jkummerow@chromium.org
da037f9872 H-BuildIncrement should make use of available type feedback
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/22611009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16353 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-27 11:44:45 +00:00
mstarzinger@chromium.org
e146b6e148 Fix replaying of captured objects during chunk building.
R=titzer@chromium.org

Review URL: https://codereview.chromium.org/22819011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16334 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-26 16:43:19 +00:00
yangguo@chromium.org
be48c5ae26 Rename "parallel recompilation" to "concurrent recompilation".
Also introduced macros for flag aliases for temporary backwards compatibility.

R=hpayer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/23014007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16280 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-22 16:14:37 +00:00
titzer@chromium.org
6f3169e571 Fix deoptimization bug, where recursive call can frighten and confuse the unwitting, simple, poor caveman that is Runtime_NotifyDeoptimized.
BUG=274164
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/23201016

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16273 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-22 13:03:40 +00:00
mstarzinger@chromium.org
63defee477 First implementation of allocation elimination in Hydrogen.
This change implements a simple data-flow analysis pass over captured
objects to the existing escape analysis. It tracks the state of values
in the Hydrogen graph through CapturedObject marker instructions that
are used to construct an appropriate translation for the deoptimizer to
be able to materialize these objects again.

This can be considered a combination of scalar replacement of loads and
stores on captured objects and sinking of unused allocations.

R=titzer@chromium.org
TEST=mjsunit/compiler/escape-analysis

Review URL: https://codereview.chromium.org/21055011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@16098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-08-07 11:24:14 +00:00
titzer@chromium.org
37ee4a0369 Fix IsDeletable() for HStringAdd, HStringCharCodeAt, HStringCharFromCode.
BUG=
R=mstarzinger@chromium.org, svenpanne@chromium.org

Review URL: https://codereview.chromium.org/20241005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15934 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-29 12:35:43 +00:00
yangguo@chromium.org
cd41cb9b6d Turn on parallel recompilation for tests that assert optimization status.
R=mvstanton@chromium.org
BUG=

Review URL: https://codereview.chromium.org/19807002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15793 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-22 09:16:33 +00:00
mstarzinger@chromium.org
3eded2c06c Fix %NeverOptimizeFunction runtime call.
The current usage of this runtime function is broken as it does not
prevent inlining of the affected function but rather bails out from the
whole unit of compilation after trying to inline affected functions.
This simplifies said runtime function to avoid accidental misuse.

R=titzer@chromium.org
TEST=mjsunit/never-optimize

Review URL: https://codereview.chromium.org/19776006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15762 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-19 08:25:44 +00:00
titzer@chromium.org
9e7819fac4 Added %NeverOptimize runtime call that can disable optimizations for a method for tests.
BUG=

Review URL: https://codereview.chromium.org/18214005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15632 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-11 14:17:56 +00:00
jkummerow@chromium.org
ed6d2d5c44 Add a test case for Phi representations
BUG=chromium:167394
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/18838002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15553 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-08 13:43:43 +00:00
jkummerow@chromium.org
f0811f4e6f Fix and cleanup can_be_minus_zero computation
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/18434004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15546 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-07-08 11:15:24 +00:00
titzer@chromium.org
bfa9fe95dc Change PC for OSR entries to point to something more sensible (i.e. the first UnknownOsrValue), removing the need to record spilled OSR values and the need for duplicate deopt entries.
Review URL: https://codereview.chromium.org/16381006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15331 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-26 08:43:27 +00:00
yangguo@chromium.org
928cbcdc8d Skip parallel recompilation tests if parallel recompilation is disabled.
Parallel recompilation is usually disabled on single-core systems.

R=jkummerow@chromium.org
BUG=v8:2733

Review URL: https://codereview.chromium.org/17261021

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15231 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-20 11:01:33 +00:00
mstarzinger@chromium.org
93ab1864ea Allow the deoptimizer translation to track de-materialized objects.
This allows the deoptimizer to materialize objects (e.g. the arguments
object) while deopting without having a consective stack area holding
the object values. The LEnvironment explicitly tracks locations for
these values and preserves them in the translation.

R=svenpanne@chromium.org
TEST=mjsunit/compiler/inline-arguments

Review URL: https://codereview.chromium.org/16779004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15087 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-12 14:22:49 +00:00
yangguo@chromium.org
74556569d1 Reland "Enable map dependency to in-flight compilation info."
BUG=248076
R=ulan@chromium.org

Review URL: https://chromiumcodereview.appspot.com/16782004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15077 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-12 09:43:22 +00:00
yangguo@chromium.org
6da97b1d4a Revert "Enable map dependency to in-flight compilation info."
This includes r15032, r15030 and r15005.

R=ulan@chromium.org
BUG=248076

Review URL: https://chromiumcodereview.appspot.com/16482004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15061 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-11 11:55:56 +00:00
yangguo@chromium.org
17cfe68015 Enable map dependency to in-flight compilation info.
R=ulan@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/16542003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@15005 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-06-07 13:27:03 +00:00
mstarzinger@chromium.org
d696f7b3c1 Use explicit type feedback clearing in some tests.
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/15711004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14744 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-22 09:17:27 +00:00
titzer@chromium.org
68eb1e50ca Improve dead code elimination by transitively marking live code and removing all dead code. Replace unreachable phi removal algorithm with the new dead code elimination pass, which is more thorough.
Review URL: https://codereview.chromium.org/14676011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@14661 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-05-14 13:10:52 +00:00
mstarzinger@chromium.org
9aa25ad1a0 Allow inlining of functions containing function literals.
R=yangguo@chromium.org
BUG=v8:1322
TEST=mjsunit/compiler/inline-literals

Review URL: https://codereview.chromium.org/10702036

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13945 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-03-14 14:29:10 +00:00
yangguo@chromium.org
479e39a058 Parallel recompilation: remove interrupt for code generation.
R=jkummerow@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/12488006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13917 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-03-12 18:03:18 +00:00
svenpanne@chromium.org
e44d3b7a87 Fixed register allocation corner case.
The predicate CanBeSpilled had a bug, prohibiting the necessary spilling and
correct splitting of live ranges. Removed a redundant assertion immediately done
by the callee anyway.

Thanks to Slava for help with that issue and the entertaining historical
background of the whole story... ;-)

BUG=177883

Review URL: https://codereview.chromium.org/12631012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13891 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-03-11 09:49:00 +00:00
ulan@chromium.org
87265114c4 Emit VMLS for multiply-subtract on ARM.
BUG=none

Review URL: https://chromiumcodereview.appspot.com/12319113
Patch from Hans Wennborg <hans@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13748 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-02-27 10:24:40 +00:00
mstarzinger@chromium.org
300413b5a9 Fix f.apply() optimization when declared arguments are mutated.
R=verwaest@chromium.org
BUG=v8:2539
TEST=mjsunit/regress/regress-2539

Review URL: https://codereview.chromium.org/12255033

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13673 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-02-14 15:12:49 +00:00
mstarzinger@chromium.org
05e95eee0e Allow full inlining of f.apply(this, arguments) calls.
This allows Crankshaft to completely inline a f.apply() dispatch if the
exact number of arguments is known and the function is constant. The
deoptimizer doesn't generate the f.apply() frame during deoptimization,
so the materialized frames look like f.apply() did a tailcall.

R=jkummerow@chromium.org
TEST=mjsunit/compiler/inline-function-apply

Review URL: https://codereview.chromium.org/12263004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13665 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-02-14 09:55:50 +00:00
mstarzinger@chromium.org
fe8e97798e Allow inlining of multiple closures from shared function.
This allows Crankshaft to allow inlining of multiple different closures
that were all derived from the same shared function info. This pattern
appears when libraries provide generic closures that are used over and
over again at different call-sites.

R=jkummerow@chromium.org
TEST=mjsunit/compiler/inline-closures

Review URL: https://codereview.chromium.org/12071002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13522 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-25 16:55:00 +00:00
mstarzinger@chromium.org
44ffa0dcd3 Allow monomorphic loads when static type is known.
This allows Crankshaft to generate monomorphic loads when the receiver
type is statically known even though the load site has polymorphic type
feedback. This applies to inlined constructor calls and literals.

R=jkummerow@chromium.org
TEST=mjsunit/compiler/property-static

Review URL: https://codereview.chromium.org/12051058

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13500 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-24 17:54:30 +00:00
mstarzinger@chromium.org
f8d5260af7 Allow loading constant function from proto chain.
This enables Crankshaft to use HConstant for loading constant functions
on the prototype chain when building a monomorphic load. This pattern
appears in several JavaScript frameworks.

R=svenpanne@chromium.org
TEST=mjsunit/compiler/proto-chain-constant

Review URL: https://codereview.chromium.org/12052008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13463 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-22 12:03:35 +00:00
yangguo@chromium.org
7f331f6280 Make HCheckPrototypeMaps compatible with parallel recompilation.
HCheckPrototypeMaps currently records the prototype and the holder of the
prototype chain (both ends of the chain) and assumes that the chain elements
and their maps did not change in during the entirety of Crankshaft. The actual
traversal of the prototype chain happens in Lithium at code generation.
With parallel compilation, this assumption is not longer correct.

R=mstarzinger@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/11864013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13454 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2013-01-21 15:49:00 +00:00
ulan@chromium.org
74492ab2d4 Emit VMLA for multiply-add on ARM
Review URL: https://chromiumcodereview.appspot.com/11293061
Patch from Hans Wennborg <hans@chromium.org>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12958 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-14 11:01:18 +00:00
ulan@chromium.org
f106c9c9f7 Add rotate-right instruction to hydrogen and use it instead of bitwise operations
of the form ((x >>> i) | (x << (32 - i))).

This CL is based on https://chromiumcodereview.appspot.com/10984057/
by Jay Conrod <dconrod@codeaurora.org>.

R=danno@chromium.org,mstarzinger@chromium.org,dconrod@codeaurora.org

Review URL: https://chromiumcodereview.appspot.com/11033005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12855 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-05 13:28:10 +00:00
svenpanne@chromium.org
9de1d40d28 Improve handling of property loads on the proto chain.
Previously Crankshaft emitted a generic load for these, now we emit a load of a
named field, guarded by a proto chain check.

LCheckPrototypeMaps now returns the holder, which is for free, because it
already had to check its map as the last step, anyway. This is in sync with what
StubCompiler::CheckPrototype does.

Review URL: https://codereview.chromium.org/11338030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12847 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-11-05 08:53:54 +00:00
jkummerow@chromium.org
1e1470fca0 Speed up test/mjsunit/compiler/regress-or
Review URL: https://codereview.chromium.org/10969063

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12593 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-24 10:08:01 +00:00
jkummerow@chromium.org
6dc2af06dc Speed up test/mjsunit/compiler/regress-gvn
Review URL: https://codereview.chromium.org/10956059

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12592 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-24 10:07:09 +00:00
mstarzinger@chromium.org
f37f504de5 Fix arguments object materialization during deopt.
This fixes materialization of arguments objects for strict mode functions during
deoptimization. We materialize arguments from the stack area where optimized
code pushes the arguments when entering the inlined environment. For adapted
invocations we use the arguments adaptor frame for materialization.

R=svenpanne@chromium.org
BUG=v8:2261
TEST=mjsunit/regress/regress-2261,mjsunit/compiler/inline-arguments

Review URL: https://chromiumcodereview.appspot.com/10908194

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12489 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-09-12 12:28:42 +00:00
rossberg@chromium.org
b0067e9cd5 Disable test that triggers known bug.
R=verwaest@chromium.org
BUG=v8:2261

Review URL: https://chromiumcodereview.appspot.com/10896005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12399 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-28 14:17:55 +00:00
vegorov@chromium.org
106a83252d Fix DoDeferredNumberTagU to keep the value in xmm1 instead of xmm0 on x64.
xmm0 is not saved across runtime call on x64 because MacroAssembler::EnterExitFrameEpilogue preserves only allocatable XMM registers unlike on ia32 where it preserves all registers.

Cleanup handling of shifts: SHR can deoptimize only when its a shift by 0, all other shift never deoptimize.

Fix type inference for i-to-t change instruction. On X64 this ensures that write-barrier is generated correctly.

R=danno@chromium.org

Review URL: https://chromiumcodereview.appspot.com/10868032

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12373 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-23 16:14:01 +00:00
vegorov@chromium.org
f476d4d431 Allow uint32 value on optimized frames if they are consumed by safe operations.
Safe operations are those that either do not observe unsignedness or have special support for uint32 values:

- all binary bitwise operations: they perform ToInt32 on inputs;
- >> and << shifts: they perform ToInt32 on left hand side and ToUint32 on right hand side;
- >>> shift: it performs ToUint32 on both inputs;
- stores to integer external arrays (not pixel, float or double ones): these stores are "bitwise";
- HChange: special support added for conversions of uint32 values to double and tagged values;
- HSimulate: special support added for deoptimization with uint32 values in registers and stack slots;
- HPhi: phis that have only safe uses and only uint32 operands are uint32 themselves.

BUG=v8:2097
TEST=test/mjsunit/compiler/uint32.js

Review URL: https://chromiumcodereview.appspot.com/10778029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12367 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-22 15:44:17 +00:00
svenpanne@chromium.org
b10d5d7f08 Deoptimization support for accessors.
Highlights of this CL:

 * Introduced a new opcode in the deoptimizer for a setter stub frame.

 * Added a global setter stub for returning after deoptimizing a setter.

 * We do not need special deopt support for getters, although the getter stub creates an internal frame. The normal machinery works just right for this case, although we generate a stack that can never occur during normal fullcode execution. If this hurts us one day, we can parameterize and reuse the setter deopt machinery.

Review URL: https://chromiumcodereview.appspot.com/10855098

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12328 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-17 10:43:32 +00:00
mstarzinger@chromium.org
886c0fa4f7 Extend constructor inlining test case.
This makes sure that deoptimization really happens in each hydrogen
context by not using binary operations but loads instead. This is
needed because we cannot clear BinaryOpICs explicitly.

R=svenpanne@chromium.org
TEST=mjsunit/compiler/inline-construct

Review URL: https://chromiumcodereview.appspot.com/10825382

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12315 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-16 09:13:25 +00:00
svenpanne@chromium.org
1ee6c0e30b Improved constructor inlining unit tests.
Currently we inline functions with different contexts only on ia32, so we have
to move the helper functions for the various contexts to the top level. Further
more, "new Object()" seems to prevent inlining, too, so we us a simple object
literal.

Although things get consistently inlined now, something strange seems to happen
in test/effect contexts: The DEOPT output seems to contain too few frames, and
we don't get any DEOPT ouput after the first time for those contexts. This has
to be investigated...

TBR=mstarzinger@chromium.org

Review URL: https://chromiumcodereview.appspot.com/10836258

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12312 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-15 09:39:13 +00:00
yangguo@chromium.org
cd5ee62692 Allow multiple lines of custom flags in javascript tests.
R=ulan@chromium.org
BUG=

Review URL: https://chromiumcodereview.appspot.com/10855099

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12289 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-10 12:26:33 +00:00
svenpanne@chromium.org
f9aea9fcef Inline simple setter calls.
Currently only simple setter calls are handled (i.e. no calls in count
operations or compound assignments), and deoptimization in the setter is not
handled at all. Because of the latter, we temporarily hide this feature behind
the --inline-accessors flag, just like inlining getters.

We now use an enum everywhere we depend on the handling of a return value,
passing around several boolean would be more confusing.

Made VisitReturnStatement and the final parts of TryInline more similar, so
matching them visually is a bit easier now.

Simplified the signature of AddLeaveInlined, the target of the HGoto can simply
be retrieved from the function state.

Review URL: https://chromiumcodereview.appspot.com/10836133

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@12286 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-08-10 09:05:42 +00:00
mstarzinger@chromium.org
a90c023d5e Allow inlining of functions containing array literals.
R=yangguo@chromium.org
BUG=v8:1322
TEST=mjsunit/compiler/inline-literals

Review URL: https://chromiumcodereview.appspot.com/10689005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11942 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-06-27 14:46:01 +00:00
mstarzinger@chromium.org
df35732ab2 Allow inlining of functions containing RegExp literals.
R=yangguo@chromium.org
BUG=v8:1322
TEST=mjsunit/compiler/inline-literals

Review URL: https://chromiumcodereview.appspot.com/10703005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11940 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-06-27 11:49:37 +00:00
fschneider@chromium.org
bb0a2ea766 Share optimized code for closures.
Each SharedFunctionInfo gets an optimized code map to store
one optimized code object per context. When allocating a new
closure we consult this map and check if there is optimized code
that can be shared.

This patch is based on an original patch
by Anton Muhin (http://codereview.chromium.org/6793013/).

BUG=v8:2087, v8:2094
TEST=test/mjsunit/compiler/optimized-closures.js
Review URL: https://chromiumcodereview.appspot.com/10103035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11817 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-06-14 14:06:22 +00:00
mstarzinger@chromium.org
11d24334fc Implement ClearFunctionTypeFeedback for test cases.
R=danno@chromium.org
TEST=mjsunit/compiler/inline-construct

Review URL: https://chromiumcodereview.appspot.com/10332010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11509 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-05-04 09:16:38 +00:00
svenpanne@chromium.org
cac8bbff6e Replaced the --limit-inling flag by three separate flags and bumped hard limits.
This change makes experiments with inlining limits much easier. Note that the
default values for the limits keep their old values for now. Renamed things a
bit for more consistency.

Review URL: https://chromiumcodereview.appspot.com/10162001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11397 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-04-20 10:42:12 +00:00
fschneider@chromium.org
0556f87851 Optimize ~~(expr) in optimized code.
~~ is commonly used to truncate a value to int32 (ToInt32).

This change avoid actually emitting the bitwise operations, and
just truncates the subexpression of ~~.

BUG=v8:2037
TEST=test/mjsunit/compiler/optimize-bitnot.js
Review URL: https://chromiumcodereview.appspot.com/10123007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11390 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-04-19 13:24:15 +00:00
mstarzinger@chromium.org
f7cd1e41f8 Fix illegal escape-sequences to throw syntax errors.
R=erik.corry@gmail.com
TEST=test262/S7.8.4_A6.*,test262/S7.8.4_A7.*

Review URL: https://chromiumcodereview.appspot.com/9490006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11340 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-04-16 15:54:02 +00:00
vegorov@chromium.org
983d521fe9 Reland arguments access support for inlined functions (r11109,r11118).
When pushing arguments use correct initial values instead of fetching them from the environment which can be modified.

R=fschneider@chromium.org
TEST=test/mjsunit/compiler/inline-arguments.js

Review URL: https://chromiumcodereview.appspot.com/10033028

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11274 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-04-11 13:40:55 +00:00
vegorov@chromium.org
3ccc885c78 Revert arguments access support for inlined functions (r11109,r11118).
We are inserting HPushArgument instructions after HEnterInlined based on the environment at the point of the first arguments access. Which might create use before def if there are redundant phis in the environment. 
Review URL: https://chromiumcodereview.appspot.com/9837041

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11128 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-23 15:39:34 +00:00
vegorov@chromium.org
b7dca5d5a7 Support arguments object access from inlined functions.
R=fschneider@chromium.org
TEST=test/mjsunit/compiler/inline-arguments.js

Review URL: https://chromiumcodereview.appspot.com/9837002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11109 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-22 13:53:28 +00:00
vegorov@chromium.org
790219ec55 Use correct arguments adaptation environment when inlining function containing arguments.
R=mstarzinger@google.com
BUG=V8:2014
TEST=test/mjsunit/compile/inline-arguments.js

Review URL: https://chromiumcodereview.appspot.com/9750007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11098 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-20 18:15:31 +00:00
vegorov@chromium.org
262c8bddd5 Always create HArgumentsObject on function entry.
We do not know if we are going to need it and creating it lazyly might cause us to insert it at the block that does not dominate all uses.

R=mstarzinger@chromium.org
TEST=mjsunit/compiler/inline-arguments.js

Review URL: https://chromiumcodereview.appspot.com/9692046

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11024 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-13 14:45:03 +00:00
vegorov@chromium.org
da03f56b1f Inline functions that use arguments object in f.apply(o, arguments) pattern.
Support arguments materialization after deoptimization in all frames (not only in topmost one).

R=fschneider@chromium.org

Review URL: https://chromiumcodereview.appspot.com/9643001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11008 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-12 12:49:41 +00:00
mstarzinger@chromium.org
240e818f0c Fix inlining of strict mode constructors.
Inlined strict mode functions (that are not called as methods) will get
their receiver reset to undefined. This should not happen when inlining
constructors.

This change also simplifies the test suite to reuse the same closures
into which constructors get inlined and use gc() to force V8 to forget
collected type feedback.

R=vegorov@chromium.org
TEST=mjsunit/compiler/inline-construct

Review URL: https://chromiumcodereview.appspot.com/9597017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10920 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-05 12:46:50 +00:00
fschneider@chromium.org
1e40f7ac2c Fix a register assignment bug in typed array stores without SSE3 available.
The old code used a separate HToInt32 instruction which had a wrong register
constraint for the input register which caused wrong result when the stored value
is used after a typed array store. (UseRegister instead of UseTempRegister) when no
SSE3 is available.

This change fixes it by replacing HToInt32 with the corresponding HChange
instruction which has correct register contraints.

TEST=mjsunit/compiler/regress-toint32.js
Review URL: https://chromiumcodereview.appspot.com/9565007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10891 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-01 12:45:46 +00:00
mstarzinger@chromium.org
fd5640cf7a Implement inlined object allocation in Crankshaft.
Generates inlined code for object allocation specific to the initial map
of the given constructor function. Also forces completion of inobject
slack tracking while crankshafting to finalize instance size of these
objects.

R=vegorov@chromium.org
TEST=mjsunit/compiler/alloc-object

Review URL: https://chromiumcodereview.appspot.com/9370019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10881 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-03-01 11:10:28 +00:00
mstarzinger@chromium.org
fb8eb04bfd Implement inlining of constructor calls.
R=vegorov@chromium.org,kmillikin@chromium.org

Review URL: https://chromiumcodereview.appspot.com/9304001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10849 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-28 09:05:55 +00:00
vegorov@chromium.org
9b55ebaa3a When compiling for-in pass correct context value to the increment instruction.
Additionally force increment instruction to use int32 representation.

R=fschneider@google.com
BUG=http://crbug.com/115646
TEST=test/mjsunit/compiler/optimized-for-in.js

Review URL: https://chromiumcodereview.appspot.com/9463052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10844 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-27 14:37:27 +00:00
mstarzinger@chromium.org
30dcdb6a36 Revert r10811 because of test flakiness.
TBR=vegorov@chromium.org
BUG=v8:1322

Review URL: https://chromiumcodereview.appspot.com/9453012

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10813 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-23 14:12:44 +00:00
mstarzinger@chromium.org
c1b97fe842 Allow inlining of functions containing function literals.
R=fschneider@chromium.org,vegorov@chromium.org
BUG=v8:1322
TEST=mjsunit/compiler/inline-literals

Review URL: https://chromiumcodereview.appspot.com/9453007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10811 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-23 12:24:03 +00:00
vegorov@chromium.org
5bb6a8399d Support OSR in for-in loops.
Modify PreProcessOsrEntry to work with OSR entries that have non-empty expression stack.

Modify graph builder to take for-in state from environment instead of directly referencing emitted instructions.

Extend %OptimizeFunctionOnNextCall with an argument to force OSR to make writing OSR tests easier: %OptimizeFunctionOnNextCall(f, "osr").

R=fschneider@chromium.org
TEST=test/mjsunit/compiler/optimized-for-in.js

Review URL: https://chromiumcodereview.appspot.com/9431030

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10796 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-22 16:45:35 +00:00
vegorov@chromium.org
6703dddac4 Support fast case for-in in Crankshaft.
Only JSObject enumerables with enum cache (fast case properties, no interceptors, no enumerable properties on the prototype) are supported.

HLoadKeyedGeneric with keys produced by for-in enumeration are recognized and rewritten into direct property load by index. For this enum-cache was extended to store property indices in a separate array (see handles.cc).

New hydrogen instructions:

- HForInPrepareMap: checks for-in fast case preconditions and returns map that contains enum-cache;
- HForInCacheArray: extracts enum-cache array from the map;
- HCheckMapValue: map check with HValue map instead of immediate;
- HLoadFieldByIndex: load fast property by it's index, positive indexes denote in-object properties, negative - out of object properties;

Changed hydrogen instructions:

- HLoadKeyedFastElement: added hole check suppression for loads from internal FixedArrays that are knows to have no holes inside.

R=fschneider@chromium.org
BUG=
TEST=

Review URL: https://chromiumcodereview.appspot.com/9425045

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10794 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-22 12:47:42 +00:00
mstarzinger@chromium.org
1dd2b094a5 Implement fast literal support in Crankshaft.
This extends the current support for nested object literals we already
have in Crankshaft, to also support nested array literals and mixed
nested literals containing arrays and objects. All three types are
generated by the unified HFastLiteral instruction.

All previous upper bounds on nested literal graphs remain unchanged,
keeping the size of generated code in check.

The main intention is to boost performance of two-dimensional array
literals containing constant elements (aka. matrices).

R=danno@chromium.org
TEST=mjsunit/compiler/literals-optimized

Review URL: https://chromiumcodereview.appspot.com/9403018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10734 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-16 17:32:30 +00:00
mstarzinger@chromium.org
15c368ce4c Revert r10721 because of test flakiness.
TBR=fschneider@chromium.org
BUG=v8:1322

Review URL: https://chromiumcodereview.appspot.com/9417013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10733 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-16 17:13:37 +00:00
mstarzinger@chromium.org
acb83c06f4 Allow inlining of functions containing function literals.
R=fschneider@chromium.org
BUG=v8:1322
TEST=mjsunit/compiler/inline-literals

Review URL: https://chromiumcodereview.appspot.com/9419005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10721 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-16 14:01:41 +00:00
mstarzinger@chromium.org
0db91d40e9 Allow inlining of functions containing object literals.
R=fschneider@chromium.org
BUG=v8:1322
TEST=mjsunit/compiler/inline-literals

Review URL: https://chromiumcodereview.appspot.com/9388007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10689 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-14 08:57:10 +00:00
fschneider@chromium.org
b8b50400d2 Inline builtin Math functions functions in more cases.
Until now we only could inline as specialized HIR instructions when called
as a method (e.g. Math.abs)

It is very common practice to abbreviate calls to those functions by defining
a global or local variable like:

var a = Math.abs;
var x = a(123);

This change allows inlining them when called as a function (global or local).
Review URL: https://chromiumcodereview.appspot.com/9365013

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10640 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-02-08 12:08:46 +00:00
vegorov@chromium.org
04289e8d17 Support inlining at call-sites with mismatched number of arguments.
Review URL: https://chromiumcodereview.appspot.com/9265004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10483 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2012-01-24 08:43:12 +00:00
fschneider@chromium.org
c1662a199b Fix a bug with register use in optimized Math.round.
We're not allowed to modify the input register and have to
use a temporary instead, otherwise the result of expressions
containing Math.round can be wrong.

BUG=106351
TEST=test/mjsunit/compiler/regress-106351.js
Review URL: http://codereview.chromium.org/8833007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10190 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-12-07 10:13:46 +00:00
fschneider@chromium.org
8480569467 Fix lazy deoptimization at HInvokeFunction and enable target-recording call-function stub.
Changes the way we do lazy deoptimization:

1. For side-effect instructions, we insert the lazy-deopt call at
the following LLazyBailout instruction.

     CALL
     GAP
     LAZY-BAILOUT ==> lazy-deopt-call

2. For other instructions (StackCheck) we insert it right after the
instruction since the deopt targets an earlier deoptimization environment.

   STACK-CHECK
   GAP ==> lazy-deopt-call

The pc of the lazy-deopt call that will be patched in is recorded in the
deoptimization input data. Each Lithium instruction can have 0..n safepoints.
All safepoints get the deoptimization index of the associated LAZY-BAILOUT
instruction. On lazy deoptimization we use the return-pc to find the safepoint.
The safepoint tells us the deoptimization index, which in turn finds us the
PC where to insert the lazy-deopt-call.

Additional changes:
 * RegExpLiteral marked it as having side-effects so that it 
   gets an explicitlazy-bailout instruction (instead of
   treating it specially like stack-checks)
 * Enable target recording CallFunctionStub to achieve
   more inlining on optimized code.

BUG=v8:1789
TEST=jslint and uglify run without crashing, mjsunit/compiler/regress-lazy-deopt.js
Review URL: http://codereview.chromium.org/8492004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10006 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-11-16 08:44:30 +00:00
fschneider@chromium.org
c48f928480 Speedup unit test to avoid timeout on slow ARM simulator.
This test depends on OSR being triggered. That's why I can't
use %OptimizeFunctionOnNextCall.
Review URL: http://codereview.chromium.org/8555004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9987 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-11-14 12:33:44 +00:00
keuchel@chromium.org
72dba271eb Reapply r9870 "Remove some initialization checks based on source positions.".
This reverts r9896 "Revert r9870 due to browser-test failures." See below for
the diff from the previous version for the ia32 platform. The code for other
platforms has been changed accordingly.

TEST=mjsunit/compiler/lazy-const-lookup.js

diff --git a/src/ia32/full-codegen-ia32.cc b/src/ia32/full-codegen-ia32.cc
index 2cbf518..1990f2f 100644
--- a/src/ia32/full-codegen-ia32.cc
+++ b/src/ia32/full-codegen-ia32.cc
@@ -1258,13 +1258,17 @@ void FullCodeGenerator::EmitVariableLoad(VariableProxy* proxy) {
         // binding is initialized:
         //   function() { f(); let x = 1; function f() { x = 2; } }
         //
-        // Check that we always have valid source position.
-        ASSERT(var->initializer_position() != RelocInfo::kNoPosition);
-        ASSERT(proxy->position() != RelocInfo::kNoPosition);
-        bool skip_init_check =
-            var->mode() != CONST &&
-            var->scope()->DeclarationScope() == scope()->DeclarationScope() &&
-            var->initializer_position() < proxy->position();
+        bool skip_init_check;
+        if (var->scope()->DeclarationScope() != scope()->DeclarationScope()) {
+          skip_init_check = false;
+        } else {
+          // Check that we always have valid source position.
+          ASSERT(var->initializer_position() != RelocInfo::kNoPosition);
+          ASSERT(proxy->position() != RelocInfo::kNoPosition);
+          skip_init_check = var->mode() != CONST &&
+              var->initializer_position() < proxy->position();
+        }
+
         if (!skip_init_check) {
           // Let and const need a read barrier.
           Label done;

Review URL: http://codereview.chromium.org/8479034

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9915 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-11-08 13:28:53 +00:00
fschneider@chromium.org
2d4bb1803d Fix bug in inlining call-as-function when inlining multiple levels deep.
This change fixes a off-by-one level error when dropping the
function from the environment. The function of the outermost
environment was not dropped.

BUG=v8:1785
TEST=test/mjsunit/compiler/regress-inline-callfunctionstub.js
Review URL: http://codereview.chromium.org/8341019

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9789 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-10-26 10:31:06 +00:00
fschneider@chromium.org
53e7502fa0 Fix bug in environment simulation after inlined call-as-function.
This change is based on my previous change enabling inlining calls-as-function
fixing the bugs related to deoptimization.

The function value on top of the environment was dropped too late in the old code.
As a result we could get a wrong value on top after deoptimization.

This change includes r9619. It was reverted because of test failures that are fixed
with this patch.
Review URL: http://codereview.chromium.org/8360001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9758 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-10-24 13:53:08 +00:00
keuchel@chromium.org
c6464d500b Replace boolean indications of strict mode by an enum value.
Review URL: http://codereview.chromium.org/8344082

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9746 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-10-24 07:47:22 +00:00
mstarzinger@chromium.org
ac712f13c3 Fix evaluation order of GT and LTE operators.
According to the ES5 spec all ">" and "<=" expressions should be be
evaluated left-to-right. This obsoletes old hacks for reversing the
order to be ES3 compliant.

R=lrn@chromium.org
BUG=v8:1752

Review URL: http://codereview.chromium.org/8275035

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9641 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-10-17 07:43:40 +00:00
fschneider@chromium.org
76c869434d Fix a bug with uninitialized const variables in the optimizing compiler.
We have to check for uninitialized uses before phi-elimination. Otherwise we
may miss such a use and result in using the hole value instead. This
causes a NULL-dereference or assertion failure.

BUG=96989
TEST=mjsunit/compiler/regress-96989.js
Review URL: http://codereview.chromium.org/7974009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9337 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-09-20 10:26:01 +00:00
fschneider@chromium.org
9e4663a8d9 Enable inlining of functions that reference context slots.
Review URL: http://codereview.chromium.org/7887038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9294 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-09-15 10:57:47 +00:00
kmillikin@chromium.org
94777e213d Remove variable rewrites and the unneccesary Slot class.
R=fschneider@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/7824038

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9162 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-09-07 11:02:31 +00:00
yangguo@chromium.org
86a62d0da3 Added check for trailing whitespaces and corrected existing violations.
Review URL: http://codereview.chromium.org/7826007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@9094 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-09-01 11:28:10 +00:00
svenpanne@chromium.org
6f6c882e19 Fixed code generation for LBranch on ARM when the operand's representation is double.
The condition code for branching on the result of vcmp was wrong, effectively
swapping the true/false branches.

TEST=regress-lbranch-double.js
Review URL: http://codereview.chromium.org/7553010

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8784 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-08-02 15:14:12 +00:00
ager@chromium.org
8ec22db350 Correct the limit of local variables in a optimized functions.
The encoding constraint is that we have 128 values. We use [-64,0] for
parameters and [0,63] for locals. However, for locals we restricted to
64 and not 63.

R=kmillikin@chromium.org
TEST=mjsunit/compiler/regress-max-locals-for-osr.js

Review URL: http://codereview.chromium.org/6995108

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8242 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-06-09 14:52:58 +00:00
fschneider@chromium.org
5083dec67f Fix timeout and a small bug in a regression test.
Review URL: http://codereview.chromium.org/7085031

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8106 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-30 12:48:35 +00:00
fschneider@chromium.org
e0985887bf Simple support for const variables in Crankshaft.
The approach is to handle the common case in the optimizing
compiler and to bailout for the rare corner cases.

This is done by initializing all local const-variables with
the hole value and disallowing any use of the hole value statically.
Review URL: http://codereview.chromium.org/6026006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8104 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-30 11:31:41 +00:00
kmillikin@chromium.org
f8b01f369e Add a simple test for inlining of arguments accesses.
R=whesse@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/7062018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@8074 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-26 11:17:15 +00:00
fschneider@chromium.org
66911961cf Fix bug in optimized compiler's switch-statement.
In the case where the default-clause occurs as the first clause,
the case-blocks were not wired up correctly.

BUG=v8:1394
TEST=mjsunit/compiler/regress-1394.js
Review URL: http://codereview.chromium.org/7037023

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7927 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-18 11:06:07 +00:00
ager@chromium.org
8a0b1f5bc9 Allow closures to be optimized if outer contexts that call eval are all in strict mode.
R=kmillikin@chromium.org
BUG=
TEST=mjsunit/compiler/eval-introduced-closure.js

Review URL: http://codereview.chromium.org/6993008

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7853 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-11 11:26:11 +00:00
karlklose@chromium.org
d43066050a Replace loops by OptimizeFunctionOnNextCall in regress-1085 and regress-1210.
Review URL: http://codereview.chromium.org/6938001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7800 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-06 09:10:28 +00:00
karlklose@chromium.org
8b917d4d96 Replace long running loops by OptimizeFunctionOnNextCall in some tests that are often timing out on ARM.
Review URL: http://codereview.chromium.org/6910022

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7765 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-05-03 13:53:08 +00:00
lrn@chromium.org
7aec228dbb Cleanup of mjsunit.js code and make assertEquals more strict.
Encapsulate the helper functions in mjsunit.js.
Now only exposes the exception class and the assertXXX functions.

Make assertEquals use === instead of ==.
This prevents a lot of possiblefalse positives in tests, and avoids
having to do assertTrue(expected === actual) when you need it.

Fixed some tests that were either buggy or assuming == test.

Review URL: http://codereview.chromium.org/6869007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7628 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-04-15 11:35:36 +00:00
fschneider@chromium.org
9783526239 Enable inlining functions containing throw.
After Kevin's change to the graph builder to allow aborting graph construction
inside arbitrary expressions this just works.

BUG=v8:1143
Review URL: http://codereview.chromium.org/6839017

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7605 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-04-13 13:09:58 +00:00
jkummerow@chromium.org
ed968b1042 Introduce runtime function %OptimizeFunctionOnNextCall to manually trigger optimization.
TEST=existing unit tests still pass

Review URL: http://codereview.chromium.org/6821009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7572 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-04-11 13:24:50 +00:00
sgjesse@chromium.org
1e8079fcc9 Increase coverage of global loads in optimized code
In the cases where a global property cell cannot be used in the optimized code
use standard load ic to get the property instead of bailing out.

This is re-committing r7212 and r7215 which where reverted in r7239 with the addition of recoring the source position in the hydrogen code for the LoadGlobalCell instruction. To record that position an optional position field has been added to the variable proxy AST node.
Review URL: http://codereview.chromium.org/6758007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7474 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-04-01 11:54:04 +00:00
ricow@chromium.org
b4bae54ef2 Decrease number of runs in mjsunit/compiler/pic.js (Test still gets optimized, even without --stress-opt)
Review URL: http://codereview.chromium.org/6731049

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7414 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-29 13:20:05 +00:00
ricow@chromium.org
e0c7d3cf08 Run array-length less times to make the builder green.
We still get this method optimized even without --always-opt flag.
Review URL: http://codereview.chromium.org/6719029

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7381 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-28 08:51:08 +00:00
fschneider@chromium.org
e6cbf659d1 Fix bug that caused invalid code motion for certain loads instructions.
The dependency flags of instructions depending on a previous check have to
be a super-set of the flags of the check instructions.
Review URL: http://codereview.chromium.org/6730025

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7343 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-24 11:37:24 +00:00
vitalyr@chromium.org
e26ae48786 Remove empty test/mjsunit/compiler/global-accessors.js to make lint happy.
Review URL: http://codereview.chromium.org/6712001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@7244 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
2011-03-17 16:12:13 +00:00