Commits · 73fc21c40d9d708a37850b01c677c602facc85b7 · Linshizhi / V8

25 Jun, 2014 33 commits

MIPS: Cleanup CallInterfaceDescriptor. · 73fc21c4

kilvadyb@homejinni.com authored Jun 25, 2014

Port r22014 (234866c)

Original commit message:
Make CallInterfaceDescriptor work like CodeStubInterfaceDescriptor, owning it's
register and representation arrays. This also eliminates a host of TSAN warnings
for static arrays.

This CL depends on https://codereview.chromium.org/352583002 landing first
(a conceptual dependency at least).

BUG=
R=palfia@homejinni.com

Review URL: https://codereview.chromium.org/356773003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22023 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

73fc21c4

Revert "Allow inlining of functions containing %_ArgumentsLength." · 70891ddf

mstarzinger@chromium.org authored Jun 25, 2014

This cuased layout test failures.

TBR=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/349923009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22022 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

70891ddf

MIPS: The IC exposes a register definition. · 55126885

kilvadyb@homejinni.com authored Jun 25, 2014

Port r22011 (be22370)

Original commit message:
Centralize a register definition in an IC that provides:
1) symbolic names for the register (like, edx == receiver)
2) defines ordering when passed on the stack

Code that implements or uses the IC should use this definition instead of "knowing" what the registers are. Or at least have the definition to validate it's assumptions.

As a side effect of avoiding runtime static initializers (enforced by tools/check-static-initializers.sh, neat), I gave ownership of the registers array to CodeStubInterfaceDescriptor. This prompted a cleanup of that struct.

BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/354863002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22021 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

55126885

Allow inlining of functions containing %_ArgumentsLength. · ace97593

mstarzinger@chromium.org authored Jun 25, 2014

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/355833002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22020 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ace97593

Collect garbage with kReduceMemoryFootprintMask in IdleNotification. · c266cd91

ulan@chromium.org authored Jun 25, 2014

BUG=350720
LOG=Y
TBR=hpayer@chromium.org

Review URL: https://codereview.chromium.org/345263004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22019 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

c266cd91

Remove distinction between hidden and normal runtime functions · eaca750b

danno@chromium.org authored Jun 25, 2014

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

Review URL: https://codereview.chromium.org/346413004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22018 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

eaca750b

ARM: Use the shifter operand to merge in previous shift instructions. · a7408e51

alexandre.rames@arm.com authored Jun 25, 2014

When possible, we transform sequences of code of the form
    lsl r1, r2, #imm
    add r0, r5, r1
into
    add r0, r5, r2 LSL #imm

This is an adaptation of r21161.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/312173002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

a7408e51

Whitespace change to trigger bots. · bd7652f1

machenbach@chromium.org authored Jun 25, 2014

TBR=jkummerow@chromium.org

BUG=

Review URL: https://codereview.chromium.org/348843010

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22016 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

bd7652f1

Fix tools/whitespace.txt. Huge perf improvements. · 79b1f09d

machenbach@chromium.org authored Jun 25, 2014

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/330913006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22015 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

79b1f09d

Cleanup CallInterfaceDescriptor · 8f905ee9

mvstanton@chromium.org authored Jun 25, 2014

Make CallInterfaceDescriptor work like CodeStubInterfaceDescriptor, owning it's
register and representation arrays. This also eliminates a host of TSAN warnings
for static arrays.

This CL depends on https://codereview.chromium.org/352583002 landing first
(a conceptual dependency at least).

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/350293003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22014 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

8f905ee9

Reorder full code for while loops to better reflect statement positions. · 7e9fc63d

yangguo@chromium.org authored Jun 25, 2014

R=jkummerow@chromium.org
BUG=v8:2047
LOG=N

Review URL: https://codereview.chromium.org/353823002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22013 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

7e9fc63d

Better fix for --trace-ic source positions: Use HOptimizedGraphBuilderWithPositions · 0bf430c1

jkummerow@chromium.org authored Jun 25, 2014

This reverts the changes made in r22008.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/355823003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22012 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

0bf430c1

The IC exposes a register definition. · addd35df

mvstanton@chromium.org authored Jun 25, 2014

Centralize a register definition in an IC that provides:
1) symbolic names for the register (like, edx == receiver)
2) defines ordering when passed on the stack

Code that implements or uses the IC should use this definition instead of "knowing" what the registers are. Or at least have the definition to validate it's assumptions.

R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/352583002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22011 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

addd35df

Blacklist failing test from GC stress runs. · 6f236862

yangguo@chromium.org authored Jun 25, 2014

The test expects a function not to deopt. However GC stress causes
the code to be marked for deopt for a reason we are not testing against.

R=marja@chromium.org

Review URL: https://codereview.chromium.org/354853002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22010 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

6f236862

Drop obsolete dont_optimize compiler hint. · 8a35e9ce

mstarzinger@chromium.org authored Jun 25, 2014

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/354813003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22009 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

8a35e9ce

Fix --trace-ic position information for LoadICs and StoreICs in optimized code · bb9d85e7

jkummerow@chromium.org authored Jun 25, 2014

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/352663002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22008 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

bb9d85e7

Wait for sweeper threads when expansion of old generation fails. · 51b189a2

hpayer@chromium.org authored Jun 25, 2014

BUG=
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/351013002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22007 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

51b189a2

Whitespace change to trigger bots. · 70ef1c24

machenbach@chromium.org authored Jun 25, 2014

TBR=jkummerow@chromium.org
BUG=

Review URL: https://codereview.chromium.org/352103003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22006 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

70ef1c24

Test that trailing commas in object literals are allowed · 3454a45f

wingo@igalia.com authored Jun 25, 2014

ES6 will allow trailing commas in object literals.  It turns out that V8
already allowed it, too, as does JSC and SpiderMonkey.

R=marja@chromium.org
BUG=

Review URL: https://codereview.chromium.org/350353002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22005 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

3454a45f

Set OS=android via GYP_DEFINES · a05abda4

jochen@chromium.org authored Jun 25, 2014

Otherwise, the variable isn't available in all scopes. Instead of playing around
with nested scopes, we just set it in GYP_DEFINES (same as chromium does)

BUG=v8:3405
R=ulan@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/357643002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22004 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

a05abda4

Grow heap slower if GC freed many global handles. · 6a5ab4b2

ulan@chromium.org authored Jun 25, 2014

BUG=263503
LOG=Y
R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/352763002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22003 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

6a5ab4b2

Promotion is the backup strategy when semi-space copy fails and vice versa. · a2f85d6c

hpayer@chromium.org authored Jun 25, 2014

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/356613004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22002 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

a2f85d6c

Do not unnecessarily expose execution state in debug event data. · 3247a71c

yangguo@chromium.org authored Jun 25, 2014

When we fire a debug event, we create duplicate execution state objects,
one as argument for the debug event listener, one as property on the
debug event data object. The latter is never used by chrome.

R=yurys@chromium.org

Review URL: https://codereview.chromium.org/355793002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22001 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

3247a71c

Suppress ASAN for non-compatible test case. · 50cc3a5b

yangguo@chromium.org authored Jun 25, 2014

This test calls quit() to exit d8 without proper tear down.

R=machenbach@chromium.org

Review URL: https://codereview.chromium.org/356623003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22000 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

50cc3a5b

Add @@iterator, .entries(), .values(), .keys() support to typed arrays · bf8e802f

wingo@igalia.com authored Jun 25, 2014

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

Review URL: https://codereview.chromium.org/336403002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21999 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

bf8e802f

Fix clusterfuzz check in auto-roll. · 52ef087d

machenbach@chromium.org authored Jun 25, 2014

Minor fixes: Strip new lines from api key. Missing import. Use string query parameters.

BUG=
TBR=jarin@chromium.org

Review URL: https://codereview.chromium.org/353813002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21998 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

52ef087d

Remove dependency from platform.h implementations on isolate · c0dd51ef

jochen@chromium.org authored Jun 25, 2014

BUG=none
R=bmeurer@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/347223004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21997 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

c0dd51ef

Add clusterfuzz check to v8 auto-roll script and use CQ. · 7833a132

machenbach@chromium.org authored Jun 25, 2014

This adds a check step that makes sure there are no new issues when rolling into chromium.

It also checks the cq bit on roll upload.

BUG=
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/315133003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21996 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

7833a132

Fix mjsunit's assertOptimized and assertUnoptimized. · 4fa6a27a

yangguo@chromium.org authored Jun 25, 2014

This was broken some time ago by a refactor.

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/349423003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21995 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

4fa6a27a

Add @@iterator support for strings · 699bc8f7

wingo@igalia.com authored Jun 25, 2014

R=rossberg@chromium.org
BUG=

Review URL: https://codereview.chromium.org/335423002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21994 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

699bc8f7

Add @@iterator to Array.prototype · f6dfa63c

wingo@igalia.com authored Jun 25, 2014

R=rossberg@chromium.org
BUG=

Review URL: https://codereview.chromium.org/338323003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21993 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

f6dfa63c

Revert "Update survival statistics correctly in the Scavenger." · 31ab363c

marja@chromium.org authored Jun 25, 2014

This reverts r21991.

Reason: lots of test failures.

BUG=
TBR=hpayer@chromium.org

Review URL: https://codereview.chromium.org/352083002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21992 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

31ab363c

Update survival statistics correctly in the Scavenger. · 661630f5

hpayer@chromium.org authored Jun 25, 2014

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/351893003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21991 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

661630f5

24 Jun, 2014 7 commits

MIPS: optimize Lithium Allocate. · 1c462184

kilvadyb@homejinni.com authored Jun 24, 2014

Port r21968 (f6c757c)

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21988 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

1c462184

ARM64: fix operator precedence issue. · 83ee3477

rodolph.perfetta@arm.com authored Jun 24, 2014

As a result the constant pool marker was always 0.

BUG=
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/350923003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21987 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

83ee3477

CPU profiler: increase the max number of captured frames. · 65e585fa

alph@chromium.org authored Jun 24, 2014

LOG=N
R=jkummerow@chromium.org, loislo@chromium.org, yurys@chromium.org

Review URL: https://codereview.chromium.org/357443003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21986 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

65e585fa

More Map methods moved to private part. · 80e1bf42

ishell@chromium.org authored Jun 24, 2014

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/350023002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21985 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

80e1bf42

Making MigrateToMap() a single bottleneck for all migrations except slow-to-fast case. · ae1aaf7c

ishell@chromium.org authored Jun 24, 2014

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/346743005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21984 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ae1aaf7c

More set_map() calls replaced with MigrateToMap(). · 220fa8ea

ishell@chromium.org authored Jun 24, 2014

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/338793004

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21983 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

220fa8ea

Promote objects if semi-space copy fails while scavenging. · 776a6568

hpayer@chromium.org authored Jun 24, 2014

BUG=388188
LOG=N
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/351853006

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21980 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

776a6568