Adding ElementsAccessor::Unshift

Adding ElementsAccessor::Unshift
Move BackingStore specific implementation from builtins.cc tp ElementsAccessor

BUG=

Committed: https://crrev.com/ec2518eedb200532d3402785dc86cd691ef4d776
Cr-Commit-Position: refs/heads/master@{#30526}

[Camillo Bruni, 2015-08-31 09:04:03]

The CQ bit was checked by cbruni@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-08-31 09:04:11]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1322803002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1322803002/1

[commit-bot: I haz the power, 2015-08-31 09:29:42]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-08-31 09:29:43]

Dry run: This issue passed the CQ dry run.

[Camillo Bruni, 2015-08-31 12:21:42]

cbruni@chromium.org changed reviewers:
+ ishell@chromium.org

[Camillo Bruni, 2015-09-01 08:45:13]

PTAL

https://codereview.chromium.org/1322803002/diff/100001/src/elements.cc
File src/elements.cc (right):

https://codereview.chromium.org/1322803002/diff/100001/src/elements.cc#newcod...
src/elements.cc:698: uint32_t dst_index, int copy_size) {
Added this ConvertElementsWithCapacity to avoid code duplication and copying
elements twice (src_index and dst_index are new arguments)

https://codereview.chromium.org/1322803002/diff/100001/src/elements.cc#newcod...
src/elements.cc:1355: WriteBarrierMode mode =
raw_backing_store->GetWriteBarrierMode(no_gc);
Introduced the second SetImpl so we can pass along a single WriteBarrierrMode ->
less overhead.

https://codereview.chromium.org/1322803002/diff/100001/src/elements.cc#newcod...
src/elements.cc:1512: FixedArray::cast(backing_store)->set(entry, value);
If I don't use the more uniform SetValue method (which has the same sign. on
FixedArray, FixedDoubleArray and FixedTypeArray) I get a 65% speedup on an
argument heavy [].unshift(1,...1000); Probably an inlining issue?

https://codereview.chromium.org/1322803002/diff/100001/src/elements.cc#newcod...
src/elements.cc:1640: Object* value) {
If I leave out this version (for FixedDoubleArray) I get a 20% speedup on
[].unshift(1,......1000).

[Camillo Bruni, 2015-09-01 08:45:20]

The CQ bit was checked by cbruni@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-09-01 08:45:25]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1322803002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1322803002/100001

[commit-bot: I haz the power, 2015-09-01 09:15:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-09-01 09:15:07]

Dry run: This issue passed the CQ dry run.

[Camillo Bruni, 2015-09-01 09:39:47]

Making SetValue inlineable.
Perf difference is still visible:
1. [].unshift(1) is ~-5%
2. [].unshift(1,....,1000) is ~-15%

[Camillo Bruni, 2015-09-01 12:04:51]

PTAl

Removed SetValue(..., WriteBarrierMode), added direct SetImpl on subclasses.

https://codereview.chromium.org/1322803002/diff/200001/src/elements.cc
File src/elements.cc (right):

https://codereview.chromium.org/1322803002/diff/200001/src/elements.cc#newcod...
src/elements.cc:572: Object* value, WriteBarrierMode mode) {
I am still keeping this version as it makes a small perf difference without too
much added code.

[Igor Sheludko, 2015-09-01 16:05:50]

lgtm with nits:

https://codereview.chromium.org/1322803002/diff/200001/src/objects.h
File src/objects.h (right):

https://codereview.chromium.org/1322803002/diff/200001/src/objects.h#newcode2394
src/objects.h:2394: inline void SetValue(uint32_t index, Object* value);
Do we still need this method? I guess not.

https://codereview.chromium.org/1322803002/diff/200001/src/objects.h#newcode2507
src/objects.h:2507: inline void SetValue(uint32_t index, Object* value);
Do we still need this method?

[Camillo Bruni, 2015-09-01 17:32:50]

The CQ bit was checked by cbruni@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-09-01 17:32:57]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1322803002/240001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1322803002/240001

[commit-bot: I haz the power, 2015-09-01 18:29:21]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-09-01 18:29:22]

Dry run: This issue passed the CQ dry run.

[Camillo Bruni, 2015-09-01 21:18:08]

The CQ bit was checked by cbruni@chromium.org

[Camillo Bruni, 2015-09-01 21:18:09]

The patchset sent to the CQ was uploaded after l-g-t-m from ishell@chromium.org
Link to the patchset: https://codereview.chromium.org/1322803002/#ps240001
(title: "space")

[commit-bot: I haz the power, 2015-09-01 21:18:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1322803002/240001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1322803002/240001

[commit-bot: I haz the power, 2015-09-01 21:19:52]

Committed patchset #13 (id:240001)

[commit-bot: I haz the power, 2015-09-01 21:20:19]

Patchset 13 (id:??) landed as
https://crrev.com/ec2518eedb200532d3402785dc86cd691ef4d776
Cr-Commit-Position: refs/heads/master@{#30526}