Updates incremental marking pass to collect object statistics.

Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics along with mark compact phase.

BUG=

Committed: https://crrev.com/31392d700289c7bfb7d8003fe8e8abc5762fc858
Cr-Commit-Position: refs/heads/master@{#36678}

[mythria, 2016-05-03 16:04:33]

mythria@chromium.org changed reviewers:
+ hpayer@chromium.org, rmcilroy@chromium.org

[mythria, 2016-05-03 16:04:34]

PTAL.

Thanks,
Mythri

https://codereview.chromium.org/1943423002/diff/1/src/heap/incremental-markin...
File src/heap/incremental-marking.cc (right):

https://codereview.chromium.org/1943423002/diff/1/src/heap/incremental-markin...
src/heap/incremental-marking.cc:181: }
Hannes, I initialize the Visitor table only with IncrementalMarking visitor.
This means that objects visited during MarkCompact phase are not accounted for
in the object statistics. I think correct fix is to account for objects visited
during both these phases. Just wanted to check with you if my understanding is
correct before I implement it. 

If I understand correctly, if an object is marked during IncrementalMarking
phase it will not be marked again during MarkCompact phase. Otherwise, we may
count objects twice.

[Hannes Payer (out of office), 2016-05-09 13:16:37]

https://codereview.chromium.org/1943423002/diff/1/src/heap/incremental-markin...
File src/heap/incremental-marking.cc (right):

https://codereview.chromium.org/1943423002/diff/1/src/heap/incremental-markin...
src/heap/incremental-marking.cc:181: }
On 2016/05/03 at 16:04:34, mythria wrote:
> Hannes, I initialize the Visitor table only with IncrementalMarking visitor.
This means that objects visited during MarkCompact phase are not accounted for
in the object statistics. I think correct fix is to account for objects visited
during both these phases. Just wanted to check with you if my understanding is
correct before I implement it. 
> 
> If I understand correctly, if an object is marked during IncrementalMarking
phase it will not be marked again during MarkCompact phase. Otherwise, we may
count objects twice.

That is correct. In case incremental marking is aborted you have to reset the
counters.

[mythria, 2016-05-10 11:25:32]

Description was changed from

==========
Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics.

BUG=
==========

to

==========
Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics along with mark compact phase.

BUG=
==========

[mythria, 2016-05-10 11:31:55]

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

[commit-bot: I haz the power, 2016-05-10 11:32:04]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1943423002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1943423002/40001

[mythria, 2016-05-10 11:32:56]

PTAL. 

Thanks,
Mythri

[commit-bot: I haz the power, 2016-05-10 12:03:36]

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

[commit-bot: I haz the power, 2016-05-10 12:03:37]

Dry run: This issue passed the CQ dry run.

[Hannes Payer (out of office), 2016-05-11 15:43:56]

https://codereview.chromium.org/1943423002/diff/40001/src/heap/mark-compact.cc
File src/heap/mark-compact.cc (right):

https://codereview.chromium.org/1943423002/diff/40001/src/heap/mark-compact.c...
src/heap/mark-compact.cc:2211: if (FLAG_track_gc_object_stats) {
I think it is cleaner to clear the object stats right after printing the trace
information.

[mythria, 2016-05-11 15:52:27]

Thanks Hannes. I thought it is required to clear the object stats there to avoid
counting the objects twice (once in incremental marking and once in mark
compact). If it is not required I will just remove it.

https://codereview.chromium.org/1943423002/diff/40001/src/heap/mark-compact.cc
File src/heap/mark-compact.cc (right):

https://codereview.chromium.org/1943423002/diff/40001/src/heap/mark-compact.c...
src/heap/mark-compact.cc:2211: if (FLAG_track_gc_object_stats) {
On 2016/05/11 15:43:56, Hannes Payer wrote:
> I think it is cleaner to clear the object stats right after printing the trace
> information.

They are already cleared after printing trace information (at the end of this
function). I thought we may have to clear here as well to avoid counting the
objects twice. I thought since we are aborting incremental marking the object
stats collected during incremental marking should be cleared. If it is just
sufficient to clear after printing trace information, I will just remove it.

[mythria, 2016-05-17 09:06:00]

Ping.

[mythria, 2016-05-23 12:31:08]

We don't need this for collecting code stats with our new plan. Though it may
still be required for collecting object stats still. PTAL when you have time.

Thanks.
Mythri

[Hannes Payer (out of office), 2016-05-24 07:30:17]

lgtm

[mythria, 2016-06-01 11:45:28]

Ross, could you PTAL.

[rmcilroy, 2016-06-01 15:36:54]

If Hannes is happy then so am I. RS-LGTM.

[mythria, 2016-06-02 09:30:00]

The CQ bit was checked by mythria@chromium.org

[commit-bot: I haz the power, 2016-06-02 09:30:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1943423002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1943423002/40001

[commit-bot: I haz the power, 2016-06-02 09:58:18]

Description was changed from

==========
Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics along with mark compact phase.

BUG=
==========

to

==========
Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics along with mark compact phase.

BUG=
==========

[commit-bot: I haz the power, 2016-06-02 09:58:19]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-06-02 10:00:06]

Description was changed from

==========
Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics along with mark compact phase.

BUG=
==========

to

==========
Updates incremental marking pass to collect object statistics.

Object statistics were collected during the mark compact phase. If
an incremental marking happened before mark compact phase then most
of the objects are already visited and hence this phase does not collect
accurate statistics. This cl updates incremental marking pass to collect
object statistics along with mark compact phase.

BUG=

Committed: https://crrev.com/31392d700289c7bfb7d8003fe8e8abc5762fc858
Cr-Commit-Position: refs/heads/master@{#36678}
==========

[commit-bot: I haz the power, 2016-06-02 10:00:08]

Patchset 3 (id:??) landed as
https://crrev.com/31392d700289c7bfb7d8003fe8e8abc5762fc858
Cr-Commit-Position: refs/heads/master@{#36678}