[heap] Exclude the owner of the linear allocation area from evacuation.

[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165
Committed: https://crrev.com/e7aa6f91b3411f47a429e22b2e20635426ad3836
Cr-Commit-Position: refs/heads/master@{#40720}

[ulan, 2016-11-02 10:38:30]

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

[ulan, 2016-11-02 10:38:30]

PTAL. Sadly no regression test because it depends on timing of end of sweeping,
start of marking and deferred allocation with allocation folding.

[Michael Lippautz, 2016-11-02 12:37:53]

lgtm

+1 on this solution

[ulan, 2016-11-02 14:06:11]

Description was changed from

==========
[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165
==========

to

==========
[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165
==========

[Hannes Payer (out of office), 2016-11-03 10:58:47]

LGTM

With that CL, we are not selecting the currently used page for bump pointer
allocation as evacuation candidate. This may be bad for fragmentation. We should
mitigate that effect in a follow up cl.

[ulan, 2016-11-03 11:47:22]

The CQ bit was checked by ulan@chromium.org

[commit-bot: I haz the power, 2016-11-03 11:47:25]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-03 12:12:46]

Description was changed from

==========
[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165
==========

to

==========
[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165
==========

[commit-bot: I haz the power, 2016-11-03 12:12:47]

Committed patchset #1 (id:1)

[ulan, 2016-11-03 13:38:11]

On 2016/11/03 10:58:47, Hannes Payer wrote:
> LGTM
> 
> With that CL, we are not selecting the currently used page for bump pointer
> allocation as evacuation candidate. This may be bad for fragmentation. We
should
> mitigate that effect in a follow up cl.

Right. I'll check memory benchmarks. One mitigation is to allow selection of
bump pointer page when marking starts from a task.

[commit-bot: I haz the power, 2016-11-17 22:20:19]

Description was changed from

==========
[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165
==========

to

==========
[heap] Exclude the owner of the linear allocation area from evacuation.

This ensures that incremental marking step does not change the top and limit
pointers of the old space, which is needed for allocation folding.

For more info see:
https://bugs.chromium.org/p/chromium/issues/detail?id=659165#c13

BUG=chromium:659165

Committed: https://crrev.com/e7aa6f91b3411f47a429e22b2e20635426ad3836
Cr-Commit-Position: refs/heads/master@{#40720}
==========

[commit-bot: I haz the power, 2016-11-17 22:20:20]

Patchset 1 (id:??) landed as
https://crrev.com/e7aa6f91b3411f47a429e22b2e20635426ad3836
Cr-Commit-Position: refs/heads/master@{#40720}