Review URL: https://codereview.chromium.org/1420473002

Cr-Commit-Position: refs/heads/master@{#31481}

BUG=v8:3330
LOG=n

Review URL: https://codereview.chromium.org/1415723005

Cr-Commit-Position: refs/heads/master@{#31480}

R=littledan@chromium.org
BUG=v8:4342
LOG=Y

Committed: https://crrev.com/722719fe31fe7fd5bb50be6256b3581bb28a8169
Cr-Commit-Position: refs/heads/master@{#31390}

Review URL: https://codereview.chromium.org/1412133002

Cr-Commit-Position: refs/heads/master@{#31479}

Review URL: https://codereview.chromium.org/1416273003

Cr-Commit-Position: refs/heads/master@{#31478}

TBR=machenbach@chromium.org
BUG=v8:4470
LOG=n

Review URL: https://codereview.chromium.org/1414963004 .

Cr-Commit-Position: refs/heads/master@{#31477}

    register configurations currently. This CL provides a mechanism so that
    optimizing compilers can select different Register Configuration.

BUG=

Review URL: https://codereview.chromium.org/1405673003

Cr-Commit-Position: refs/heads/master@{#31476}

This slightly generalizes the current mechnanism for property loads to
also deal with property stores (both monomorphic and polymorphic). It's
still somewhat adhoc, but it already handles a lot of the interesting
cases. We might want to wait for keyed accesses before we start
refactoring to a more general solution.

R=jarin@chromium.org
BUG=v8:4470
LOG=n

Review URL: https://codereview.chromium.org/1422573002

Cr-Commit-Position: refs/heads/master@{#31475}

BUG=v8:4507
LOG=y
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1409353005

Cr-Commit-Position: refs/heads/master@{#31474}

This reverts commit 581ead5c.

Review URL: https://codereview.chromium.org/1413123004

Cr-Commit-Position: refs/heads/master@{#31473}

R=bmeurer@chromium.org
TEST=webkit/fast/js/toString-number

Review URL: https://codereview.chromium.org/1410543007

Cr-Commit-Position: refs/heads/master@{#31472}

This switches inlining back to use a temporary zone for parsing and
analyzing inlinees. The inlinee graph however is still built in the
same zone as the parent graph.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1422503005

Cr-Commit-Position: refs/heads/master@{#31471}

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/1417203003

Cr-Commit-Position: refs/heads/master@{#31470}

This moves the bailout point in the JSInliner up to a point where it is
still allowed to decide not to inline. Once the inlining decision has
been recorded with CompilationInfo::AddInlinedFunction, we should not
abort anymore.

R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1410023006

Cr-Commit-Position: refs/heads/master@{#31469}

The patch in https://crrev.com/ddb5c2d999c5ee6e31c4a9599bb3ddb293cc3f49
moved all fractional digit settings to default values due to a coding
error. These were not even correct default values, and users observed
errors where percentages were written as "23.0%" instead of "23%".

This patch fixes the setting propagation when appropriate and it changes
the default max fractional digits of a percentage to 0, per spec.

BUG=chromium:544122
R=mnita,jochen
CC=hichris123,adamk
LOG=Y

Review URL: https://codereview.chromium.org/1420883002

Cr-Commit-Position: refs/heads/master@{#31468}

R=cbruni@chromium.org

Review URL: https://codereview.chromium.org/1413173003

Cr-Commit-Position: refs/heads/master@{#31467}

port 81ee94b6 (r31410).

   contributed by zhengxing.li@intel.com

   original commit message:

   additional comment:
   The original r31410 patch needs some additional changes for x87
    1. The frames-x87.h is under src/x87 instead of src/crankshaft/x87

R=weiliang.lin@intel.com
BUG=

Review URL: https://codereview.chromium.org/1417193002

Cr-Commit-Position: refs/heads/master@{#31466}

Proxies are not properly supported yet.

R=rossberg
BUG=v8:3931
LOG=n

Review URL: https://codereview.chromium.org/1417243002

Cr-Commit-Position: refs/heads/master@{#31465}

This will allow us to remove background idle notification calls in Chrome.

BUG=chromium:490559
LOG=NO

Review URL: https://codereview.chromium.org/1406303008

Cr-Commit-Position: refs/heads/master@{#31464}

BUG=

Review URL: https://codereview.chromium.org/1416003003

Cr-Commit-Position: refs/heads/master@{#31463}

Both the JSTypeFeedbackSpecializer and the JSTypeFeedbackLowering is
dead code by now, since the more general JSNativeContextSpecialization
deals with the property/global load/store type feedback in a way that
also interacts properly with inlining.

BUG=v8:4470
LOG=n
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/1407913003 .

Cr-Commit-Position: refs/heads/master@{#31462}

Incremental GC now finish quickly in background tabs too because they
do not rely on idle notifications.

BUG=

Review URL: https://codereview.chromium.org/1423453003

Cr-Commit-Position: refs/heads/master@{#31461}

BUG=chromium:524425
LOG=N

Review URL: https://codereview.chromium.org/1420803002

Cr-Commit-Position: refs/heads/master@{#31460}

BUG=

Review URL: https://codereview.chromium.org/1419833002

Cr-Commit-Position: refs/heads/master@{#31459}

Remove support for "loads and stores to global vars through property cell shortcuts installed into parent script context" from all compilers.

The plan is to implement the same idea using vector IC machinery.
Stubs implementations and scopes modifications are left untouched for now.

Review URL: https://codereview.chromium.org/1419823003

Cr-Commit-Position: refs/heads/master@{#31458}

NOTRY=true

Review URL: https://codereview.chromium.org/1415863003

Cr-Commit-Position: refs/heads/master@{#31457}

Use a unified NamedAccess operator parameter for both JSLoadNamed and
JSStoreNamed, and similar use PropertyAccess for both JSLoadProperty and
JSStoreProperty.

Review URL: https://codereview.chromium.org/1418993002

Cr-Commit-Position: refs/heads/master@{#31456}

R=jarin@chromium.org
BUG=v8:4506
LOG=n

Review URL: https://codereview.chromium.org/1413503006

Cr-Commit-Position: refs/heads/master@{#31455}

Cr-Commit-Position: refs/heads/master@{#31454}

It appears we haven't wired in the 32 float constant passed on the stack.

BUG=v8:4506
LOG=n

Review URL: https://codereview.chromium.org/1407003015

Cr-Commit-Position: refs/heads/master@{#31453}

BUG=v8:4406
LOG=N

Review URL: https://codereview.chromium.org/1419983002

Cr-Commit-Position: refs/heads/master@{#31452}

Rolling v8/tools/swarming_client to 8fce79620b04bbe5415ace1103db27505bdc4c06

TBR=machenbach@chromium.org,vogelheim@chromium.org,hablich@chromium.org

Review URL: https://codereview.chromium.org/1412173007

Cr-Commit-Position: refs/heads/master@{#31451}

Add mechanism for adding per-context ECMAScript intrinsics to Template
objects, which may be instantiated in multiple contexts.

BUG=
LOG=N
R=jochen@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/1409593002

Cr-Commit-Position: refs/heads/master@{#31450}

With --debug-code specified, 8 KB is not sufficient to contain
some of the builtins on PPC64.

This path is relevant only when snapshots are off.

R=rmcilroy@chromium.org, ulan@chromium.org, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1415463002

Cr-Commit-Position: refs/heads/master@{#31449}

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/1422543005

Cr-Commit-Position: refs/heads/master@{#31448}

Proxies are not properly supported yet.

R=rossberg
BUG=v8:3931
LOG=n

Review URL: https://codereview.chromium.org/1416433003

Cr-Commit-Position: refs/heads/master@{#31447}

(With a v8::Vector, the client is responsible for memory management.
 I think there can be a situation where the Vector has a char[1] backing
 store with '\0' in it, in which case the current code would leak.
 If we always Dispose() the backing store this should be avoided.
 Since dispose will delete[] the actual backing store, this should
 also work then the backing store is nullptr.)

R=jochen@chromium.org
BUG=chromium:525885
LOG=N

Review URL: https://codereview.chromium.org/1410543005

Cr-Commit-Position: refs/heads/master@{#31446}

This change adds new flavors of Visit() methods for obtaining
expression results:

- VisitForAccumulatorValue() which places result in the accumulator.
- VisitForRegisterValue() which places the result in a register.
- VisitForEffect() which evaluates the expression and discards the result.

The targets of these calls place the expression result with
result_scope()->SetResultInRegister() or
result_scope()->SetResultInAccumulator().

By being smarter about result locations, there's less temporary
register usage. However, we now have a hazard with assignments
in binary expressions that didn't exist before. This change detects and
DCHECK's when a hazard is detected. A follow on CL will address this.

There are consequential changes to test-bytecode-generator.cc and
this change also adds new bytecode macros A(x, n) and THIS(n) for
register file entries for arguments and this.

BUG=v8:4280
LOG=NO

Review URL: https://codereview.chromium.org/1392933002

Cr-Commit-Position: refs/heads/master@{#31445}

The newly introduced root makes sure that we do not flush the
optimized code while the function is being compiled.

BUG=v8:4493
LOG=n

Review URL: https://codereview.chromium.org/1415133002

Cr-Commit-Position: refs/heads/master@{#31444}

This reverts commit a31cef44.

Original message:

[heap] Divide available memory upon compaction tasks
- Fairly (round-robin) divide available memory upon compaction tasks.
- Ensure an upper limit (of memory) since dividing is O(n) for n free-space
  nodes.
- Refill from free lists managed by sweeper once a compaction space becomes
  empty.

Assumption for dividing memory: Memory in the free lists is sparse upon starting
compaction (which means that only few nodes are available), except for memory
reducer GCs, which happen in idle time though (so it's less of a problem).

BUG=chromium:524425
LOG=N

Review URL: https://codereview.chromium.org/1421583002

Cr-Commit-Position: refs/heads/master@{#31443}

Set "no recover" in diagnostic mode to keep triggering
crashes. Otherwise, the v8 test driver won't notice tests
as failing.

Also port some common sanitizer cflags to cfi - those can be
removed on the build side afterwards.

BUG=chromium:515782
LOG=n

Review URL: https://codereview.chromium.org/1414763004

Cr-Commit-Position: refs/heads/master@{#31442}