[heap] Fine-grained JSArrayBuffer tracking

[heap] Fine-grained JSArrayBuffer tracking

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

Committed: https://crrev.com/b2d8bfc7931eef49d527605ba485950dea41cde3
Cr-Commit-Position: refs/heads/master@{#36437}

[Michael Lippautz, 2016-05-10 15:03:53]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-10 15:03:53]

mlippautz@chromium.org changed reviewers:
+ hpayer@chromium.org

[Michael Lippautz, 2016-05-10 15:05:22]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland fixes an issue with registering a buffer on a black page, where
we would fail to track it as live.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-10 15:19:31]

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

[commit-bot: I haz the power, 2016-05-10 15:19:41]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/40001

[commit-bot: I haz the power, 2016-05-10 15:48:11]

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

[commit-bot: I haz the power, 2016-05-10 15:48:12]

Dry run: This issue passed the CQ dry run.

[Michael Lippautz, 2016-05-11 07:08:05]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland fixes an issue with registering a buffer on a black page, where
we would fail to track it as live.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-11 07:28:31]

Patchset #6 (id:100001) has been deleted

[Michael Lippautz, 2016-05-11 07:39:07]

Patchset #7 (id:140001) has been deleted

[Michael Lippautz, 2016-05-11 08:20:34]

Patchset #7 (id:160001) has been deleted

[Michael Lippautz, 2016-05-11 08:21:05]

PTAL since it's a new approach.

[Michael Lippautz, 2016-05-11 08:22:51]

https://codereview.chromium.org/1964023002/diff/180001/include/v8.h
File include/v8.h (right):

https://codereview.chromium.org/1964023002/diff/180001/include/v8.h#newcode8559
include/v8.h:8559: reinterpret_cast<intptr_t*>(
It's really unfortunate that we require this magic here but experience shows
that we should have this fast path in the header.

Note that we cast an atomic value to intptr_t and use it as a counter. Happy for
any suggestions.

[Hannes Payer (out of office), 2016-05-11 11:12:34]

I like the local array buffer trackers! I have a couple of comments.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/array-buffer-...
File src/heap/array-buffer-tracker.h (right):

https://codereview.chromium.org/1964023002/diff/200001/src/heap/array-buffer-...
src/heap/array-buffer-tracker.h:66: // For each GC round (Scavenger, or
incremental/full MC)
We do not have to talk about GC type here.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact.cc
File src/heap/mark-compact.cc (right):

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact....
src/heap/mark-compact.cc:1852:
heap_->array_buffer_tracker()->MarkLive(JSArrayBuffer::cast(object));
Why do you need to mark it as live in the evacuation visitor? This should have
been marked live by the marker.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact....
src/heap/mark-compact.cc:3580:
heap()->array_buffer_tracker()->FreeDeadInNewSpace();
Aborted once will be swept after compaction and
FreeDeadArraybuffersAndResetTracker will be called there again.

Why don't we move it to the compaction tasks? They can do the freeing in
parallel if evacuation succeeded. Otherwise the sweeper will take care of that.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact....
src/heap/mark-compact.cc:3859: p->FreeDeadArraybuffersAndResetTracker();
There will be no dead array buffers on a black page.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/scavenger.cc
File src/heap/scavenger.cc (left):

https://codereview.chromium.org/1964023002/diff/200001/src/heap/scavenger.cc#...
src/heap/scavenger.cc:298:
heap->array_buffer_tracker()->Promote(JSArrayBuffer::cast(target));
Why don't we make Promote, SemiSpaceCopy etc. a mode and call a templatized
array buffer move function?
Code like the one here will become cleaner and shorter.

https://codereview.chromium.org/1964023002/diff/200001/test/cctest/heap/utils...
File test/cctest/heap/utils-inl.h (right):

https://codereview.chromium.org/1964023002/diff/200001/test/cctest/heap/utils...
test/cctest/heap/utils-inl.h:153: static inline void GcAndSweep(Heap* heap,
AllocationSpace space) {
Why don't we make it a CollectGarbage mode that finalizes sweeping in the atomic
pause? That comes in handy for many test cases.

[Michael Lippautz, 2016-05-11 18:43:26]

PTAL

https://codereview.chromium.org/1964023002/diff/200001/src/heap/array-buffer-...
File src/heap/array-buffer-tracker.h (right):

https://codereview.chromium.org/1964023002/diff/200001/src/heap/array-buffer-...
src/heap/array-buffer-tracker.h:66: // For each GC round (Scavenger, or
incremental/full MC)
On 2016/05/11 11:12:34, Hannes Payer wrote:
> We do not have to talk about GC type here.

Done.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact.cc
File src/heap/mark-compact.cc (right):

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact....
src/heap/mark-compact.cc:1852:
heap_->array_buffer_tracker()->MarkLive(JSArrayBuffer::cast(object));
On 2016/05/11 11:12:34, Hannes Payer wrote:
> Why do you need to mark it as live in the evacuation visitor? This should have
> been marked live by the marker.

You are right, done.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact....
src/heap/mark-compact.cc:3580:
heap()->array_buffer_tracker()->FreeDeadInNewSpace();
On 2016/05/11 11:12:34, Hannes Payer wrote:
> Aborted once will be swept after compaction and
> FreeDeadArraybuffersAndResetTracker will be called there again.
> 
> Why don't we move it to the compaction tasks? They can do the freeing in
> parallel if evacuation succeeded. Otherwise the sweeper will take care of
that.

1) This call is about new space pages
2) "FreeDeadArraybuffersAndResetTracker" does 2 things:
  a) it frees dead backings stores
  b) it resets the sets for the next round

We have to be careful to not call  FreeDeadArraybuffersAndResetTracker twice
without marking inbetween. That's the reason this calls are spread out in a way
that we call it only 1x for each page.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/mark-compact....
src/heap/mark-compact.cc:3859: p->FreeDeadArraybuffersAndResetTracker();
On 2016/05/11 11:12:34, Hannes Payer wrote:
> There will be no dead array buffers on a black page.

True, but we need the "*AndResetTracker" part of the function. Otherwise we
would lag 1 GC behind with black pages, increasing floating garbage.

https://codereview.chromium.org/1964023002/diff/200001/src/heap/scavenger.cc
File src/heap/scavenger.cc (left):

https://codereview.chromium.org/1964023002/diff/200001/src/heap/scavenger.cc#...
src/heap/scavenger.cc:298:
heap->array_buffer_tracker()->Promote(JSArrayBuffer::cast(target));
On 2016/05/11 11:12:34, Hannes Payer wrote:
> Why don't we make Promote, SemiSpaceCopy etc. a mode and call a templatized
> array buffer move function?
> Code like the one here will become cleaner and shorter.

All functions do exactly the same thing -> unified. (The code is still a left
over when they did different things in a different implementation)

https://codereview.chromium.org/1964023002/diff/200001/test/cctest/heap/utils...
File test/cctest/heap/utils-inl.h (right):

https://codereview.chromium.org/1964023002/diff/200001/test/cctest/heap/utils...
test/cctest/heap/utils-inl.h:153: static inline void GcAndSweep(Heap* heap,
AllocationSpace space) {
On 2016/05/11 11:12:34, Hannes Payer wrote:
> Why don't we make it a CollectGarbage mode that finalizes sweeping in the
atomic
> pause? That comes in handy for many test cases.

Good idea! Fine to do it in a follow up? It requires some massaging of
CollectGarbage since we probably want to specify this on the CollectGarbage call
that takes a space. I suppose this will change quite some call sites.

[Michael Lippautz, 2016-05-11 19:35:56]

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

[commit-bot: I haz the power, 2016-05-11 19:36:03]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/260001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/260001

[commit-bot: I haz the power, 2016-05-11 19:42:32]

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

[commit-bot: I haz the power, 2016-05-11 19:42:33]

Dry run: Try jobs failed on following builders:
  v8_linux64_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_rel_ng/builds/5628)
  v8_linux64_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_rel_ng_triggered...)
  v8_linux_nodcheck_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_nodcheck_rel_ng/bu...)
  v8_linux_nodcheck_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_nodcheck_rel_ng_tr...)
  v8_linux_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel_ng/builds/5618)
  v8_linux_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel_ng_triggered/b...)

[Michael Lippautz, 2016-05-11 19:44:20]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412
LOG=N

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_asan_rel,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-11 21:09:50]

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

[commit-bot: I haz the power, 2016-05-11 21:09:53]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/260001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/260001

[commit-bot: I haz the power, 2016-05-11 21:10:44]

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

[commit-bot: I haz the power, 2016-05-11 21:10:45]

Dry run: Try jobs failed on following builders:
  v8_linux64_asan_rel on tryserver.v8 (JOB_FAILED, no build URL)

[Michael Lippautz, 2016-05-11 21:23:37]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_asan_rel,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-11 22:05:30]

Patchset #11 (id:260001) has been deleted

[Michael Lippautz, 2016-05-11 22:05:38]

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

[commit-bot: I haz the power, 2016-05-11 22:05:43]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/240001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/240001

[commit-bot: I haz the power, 2016-05-11 22:57:33]

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

[commit-bot: I haz the power, 2016-05-11 22:57:34]

Dry run: Try jobs failed on following builders:
  v8_mac_gc_stress_dbg on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_mac_gc_stress_dbg/builds...)

[Michael Lippautz, 2016-05-12 14:06:59]

Patchset #12 (id:290001) has been deleted

[Michael Lippautz, 2016-05-12 14:31:11]

Patchset #12 (id:310001) has been deleted

[Michael Lippautz, 2016-05-12 15:29:50]

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

[commit-bot: I haz the power, 2016-05-12 15:30:00]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/330001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/330001

[Michael Lippautz, 2016-05-12 15:56:43]

Patchset #11 (id:280001) has been deleted

[commit-bot: I haz the power, 2016-05-12 16:02:54]

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

[commit-bot: I haz the power, 2016-05-12 16:02:55]

Dry run: This issue passed the CQ dry run.

[Michael Lippautz, 2016-05-13 10:32:36]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412BUG, chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-13 10:32:51]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:581412BUG, chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-13 10:56:44]

Patchset #13 (id:370001) has been deleted

[Michael Lippautz, 2016-05-13 11:51:24]

Patchset #13 (id:390001) has been deleted

[Michael Lippautz, 2016-05-17 13:41:16]

Patchset #14 (id:430001) has been deleted

[Michael Lippautz, 2016-05-17 17:19:14]

Patchset #14 (id:450001) has been deleted

[Michael Lippautz, 2016-05-17 17:19:22]

Patchset #12 (id:350001) has been deleted

[Michael Lippautz, 2016-05-17 17:19:56]

Patchset #9 (id:220001) has been deleted

[Michael Lippautz, 2016-05-17 17:20:05]

Patchset #13 (id:490001) has been deleted

[Michael Lippautz, 2016-05-17 17:20:51]

Patchset #12 (id:470001) has been deleted

[Michael Lippautz, 2016-05-17 20:56:10]

Patchset #13 (id:530001) has been deleted

[Michael Lippautz, 2016-05-17 20:56:19]

Patchset #12 (id:510001) has been deleted

[Michael Lippautz, 2016-05-18 06:03:05]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. Otherwise the sweeper
  processes them concurrently

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-18 06:46:14]

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

[commit-bot: I haz the power, 2016-05-18 06:46:18]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/590001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/590001

[commit-bot: I haz the power, 2016-05-18 07:10:11]

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

[commit-bot: I haz the power, 2016-05-18 07:10:12]

Dry run: Try jobs failed on following builders:
  v8_linux_arm_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng/builds/...)
  v8_linux_arm_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng_trigger...)
  v8_win_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_rel_ng/builds/7298)
  v8_win_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_win_rel_ng_triggered/bui...)

[Michael Lippautz, 2016-05-18 07:42:53]

Patchset #13 (id:570001) has been deleted

[Michael Lippautz, 2016-05-18 07:42:56]

Patchset #13 (id:590001) has been deleted

[Michael Lippautz, 2016-05-18 07:43:07]

Patchset #12 (id:550001) has been deleted

[Michael Lippautz, 2016-05-18 08:59:03]

Patchset #12 (id:610001) has been deleted

[Michael Lippautz, 2016-05-18 09:43:35]

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

[commit-bot: I haz the power, 2016-05-18 09:43:50]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/630001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/630001

[Michael Lippautz, 2016-05-18 09:59:15]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. Otherwise the sweeper
  processes them concurrently

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-18 09:59:40]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page.

The reland also switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-18 09:59:58]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page
contains the set of its contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

[Michael Lippautz, 2016-05-18 10:35:56]

PTAL, the description is also updated to reflect what's going on

[Michael Lippautz, 2016-05-19 15:21:55]

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

[commit-bot: I haz the power, 2016-05-19 15:22:06]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/670001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/670001

[commit-bot: I haz the power, 2016-05-19 15:23:20]

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

[commit-bot: I haz the power, 2016-05-19 15:23:21]

Dry run: Try jobs failed on following builders:
  v8_linux_arm64_gc_stress_dbg on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm64_gc_stress_db...)
  v8_linux_arm64_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm64_rel_ng/build...)
  v8_linux_gcc_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)
  v8_linux_mips64el_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_mips64el_compile_r...)
  v8_mac_gc_stress_dbg on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_mac_gc_stress_dbg/builds...)
  v8_mac_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_mac_rel_ng/builds/1900)

[Michael Lippautz, 2016-05-19 15:34:23]

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

[commit-bot: I haz the power, 2016-05-19 15:34:28]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/690001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/690001

[commit-bot: I haz the power, 2016-05-19 15:41:19]

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

[commit-bot: I haz the power, 2016-05-19 15:41:21]

Dry run: Try jobs failed on following builders:
  v8_linux_arm_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng/builds/...)
  v8_linux_arm_rel_ng_triggered on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel_ng_trigger...)

[Michael Lippautz, 2016-05-19 17:02:29]

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

[commit-bot: I haz the power, 2016-05-19 17:02:38]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/710001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/710001

[commit-bot: I haz the power, 2016-05-19 17:48:29]

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

[commit-bot: I haz the power, 2016-05-19 17:48:31]

Dry run: This issue passed the CQ dry run.

[Hannes Payer (out of office), 2016-05-20 12:21:17]

looking good already, cool stuff

https://codereview.chromium.org/1964023002/diff/710001/include/v8.h
File include/v8.h (right):

https://codereview.chromium.org/1964023002/diff/710001/include/v8.h#newcode7360
include/v8.h:7360: kApiPointerSize + kApiPointerSize;
Why the extra word?

https://codereview.chromium.org/1964023002/diff/710001/src/heap/array-buffer-...
File src/heap/array-buffer-tracker.h (right):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/array-buffer-...
src/heap/array-buffer-tracker.h:21: class LocalArrayBufferTracker {
Add a long and detailed description that explains the design.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/array-buffer-...
src/heap/array-buffer-tracker.h:38: // Resets the tracking set, i.e., not yet
discovered <- currently live.
Clearer comment please.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/heap.cc
File src/heap/heap.cc (right):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/heap.cc#newco...
src/heap/heap.cc:2017:
Page::FromAddress(buffer->address())->IsFlagSet(Page::BLACK_PAGE);
Instead of checking for black page, it would be better to check for object color
to be independent of black pages.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/incremental-m...
File src/heap/incremental-marking.cc (right):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/incremental-m...
src/heap/incremental-marking.cc:550: heap_->old_space()->ForAllPages([](Page* p)
{ p->ResetTracker(); });
Factor this line out into a function since you have two call sites.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/mark-compact.cc
File src/heap/mark-compact.cc (left):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/mark-compact....
src/heap/mark-compact.cc:1728: // If we end up needing more special cases, we
should factor this out.
Nice!

[Michael Lippautz, 2016-05-23 08:41:01]

Thanks a lot!

https://codereview.chromium.org/1964023002/diff/710001/include/v8.h
File include/v8.h (right):

https://codereview.chromium.org/1964023002/diff/710001/include/v8.h#newcode7360
include/v8.h:7360: kApiPointerSize + kApiPointerSize;
On 2016/05/20 12:21:17, Hannes Payer wrote:
> Why the extra word?

I added an intptr_t to Heap, changing the offset of the roots.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/array-buffer-...
File src/heap/array-buffer-tracker.h (right):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/array-buffer-...
src/heap/array-buffer-tracker.h:21: class LocalArrayBufferTracker {
On 2016/05/20 12:21:17, Hannes Payer wrote:
> Add a long and detailed description that explains the design.

Done.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/array-buffer-...
src/heap/array-buffer-tracker.h:38: // Resets the tracking set, i.e., not yet
discovered <- currently live.
On 2016/05/20 12:21:17, Hannes Payer wrote:
> Clearer comment please.

Done.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/heap.cc
File src/heap/heap.cc (right):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/heap.cc#newco...
src/heap/heap.cc:2017:
Page::FromAddress(buffer->address())->IsFlagSet(Page::BLACK_PAGE);
On 2016/05/20 12:21:17, Hannes Payer wrote:
> Instead of checking for black page, it would be better to check for object
color
> to be independent of black pages.

Done.

https://codereview.chromium.org/1964023002/diff/710001/src/heap/incremental-m...
File src/heap/incremental-marking.cc (right):

https://codereview.chromium.org/1964023002/diff/710001/src/heap/incremental-m...
src/heap/incremental-marking.cc:550: heap_->old_space()->ForAllPages([](Page* p)
{ p->ResetTracker(); });
On 2016/05/20 12:21:17, Hannes Payer wrote:
> Factor this line out into a function since you have two call sites.

Done.

[Hannes Payer (out of office), 2016-05-23 08:48:18]

lgtm

[Michael Lippautz, 2016-05-23 08:48:34]

Description was changed from

==========
Reland of "[heap] Fine-grained JSArrayBuffer tracking"

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

This reverts commit c0fe26d25325c3bef47ab70e0cbcae6f4f2d782c.
==========

to

==========
[heap] Fine-grained JSArrayBuffer tracking

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel
==========

[Michael Lippautz, 2016-05-23 08:56:08]

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

[commit-bot: I haz the power, 2016-05-23 08:56:18]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/750001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/750001

[commit-bot: I haz the power, 2016-05-23 09:43:00]

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

[commit-bot: I haz the power, 2016-05-23 09:43:01]

Dry run: This issue passed the CQ dry run.

[Michael Lippautz, 2016-05-23 09:51:44]

The CQ bit was checked by mlippautz@chromium.org

[Michael Lippautz, 2016-05-23 09:51:44]

The patchset sent to the CQ was uploaded after l-g-t-m from hpayer@chromium.org
Link to the patchset: https://codereview.chromium.org/1964023002/#ps750001
(title: "Rebase (tests)")

[commit-bot: I haz the power, 2016-05-23 09:51:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/750001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/750001

[commit-bot: I haz the power, 2016-05-23 09:54:20]

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

[commit-bot: I haz the power, 2016-05-23 09:54:22]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/15748)

[Michael Lippautz, 2016-05-23 09:57:47]

The CQ bit was checked by mlippautz@chromium.org

[commit-bot: I haz the power, 2016-05-23 09:58:00]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/750001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/750001

[commit-bot: I haz the power, 2016-05-23 10:06:19]

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

[commit-bot: I haz the power, 2016-05-23 10:06:21]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/15753)

[Michael Lippautz, 2016-05-23 10:23:35]

The CQ bit was checked by mlippautz@chromium.org

[commit-bot: I haz the power, 2016-05-23 10:23:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/750001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/750001

[commit-bot: I haz the power, 2016-05-23 10:26:59]

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

[commit-bot: I haz the power, 2016-05-23 10:27:01]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/15756)

[Michael Lippautz, 2016-05-23 10:32:57]

The CQ bit was checked by mlippautz@chromium.org

[Michael Lippautz, 2016-05-23 10:32:57]

The patchset sent to the CQ was uploaded after l-g-t-m from hpayer@chromium.org
Link to the patchset: https://codereview.chromium.org/1964023002/#ps770001
(title: "Rebase again")

[commit-bot: I haz the power, 2016-05-23 10:33:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1964023002/770001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1964023002/770001

[commit-bot: I haz the power, 2016-05-23 11:20:01]

Description was changed from

==========
[heap] Fine-grained JSArrayBuffer tracking

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel
==========

to

==========
[heap] Fine-grained JSArrayBuffer tracking

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel
==========

[commit-bot: I haz the power, 2016-05-23 11:20:03]

Committed patchset #19 (id:770001)

[commit-bot: I haz the power, 2016-05-23 11:21:59]

Description was changed from

==========
[heap] Fine-grained JSArrayBuffer tracking

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel
==========

to

==========
[heap] Fine-grained JSArrayBuffer tracking

Track based on JSArrayBuffer addresses instead of the attached backing store.
This way we can later on iterate buffers on a single page. The reland also
switches to a page-based implementation where a page contains the set of its
contained (live and dead) buffers.

Details of tracking:
- Scavenge: New space pages are processes in bulk on the main thread
- MC: Unswept pages are processed in bulk in parallel. All other pages
  are processed by the sweeper concurrently.

BUG=chromium:611688
LOG=N
CQ_EXTRA_TRYBOTS=tryserver.v8:v8_linux_arm64_gc_stress_dbg,v8_linux_gc_stress_dbg,v8_mac_gc_stress_dbg,v8_linux64_tsan_rel,v8_mac64_asan_rel

Committed: https://crrev.com/b2d8bfc7931eef49d527605ba485950dea41cde3
Cr-Commit-Position: refs/heads/master@{#36437}
==========

[commit-bot: I haz the power, 2016-05-23 11:22:01]

Patchset 19 (id:??) landed as
https://crrev.com/b2d8bfc7931eef49d527605ba485950dea41cde3
Cr-Commit-Position: refs/heads/master@{#36437}

[Marijn Kruisselbrink, 2016-05-23 21:40:56]

A revert of this CL (patchset #19 id:770001) has been created in
https://codereview.chromium.org/2006923002/ by mek@chromium.org.

The reason for reverting is: Causing frequent test flakes on maps_pixel_tests on
various chromium bots.

For example in 
https://build.chromium.org/p/chromium.gpu/builders/Linux%20Release%20%28NVIDI...

it crashes in 
	 0  chrome!Process<(lambda at
../../v8/src/heap/array-buffer-tracker-inl.h:51:15)> [spaces.h : 962 + 0x0]
	 1  chrome!EvacuatePage [array-buffer-tracker-inl.h : 51 + 0x5]
	 2  chrome!RunInternal [mark-compact.cc : 3213 + 0x5]
....

[Michael Achenbach, 2016-05-24 06:45:14]

A revert of this CL (patchset #19 id:770001) has been created in
https://codereview.chromium.org/2000373003/ by machenbach@chromium.org.

The reason for reverting is: http://crbug.com/614142.

[Michael Lippautz, 2016-05-24 07:20:59]

A revert of this CL (patchset #19 id:770001) has been created in
https://codereview.chromium.org/2006183003/ by mlippautz@chromium.org.

The reason for reverting is: Revert it..