[heap] MinorMC: Only iterate new space global handles for ptr updates

[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the
minor MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354
Review-Url: https://codereview.chromium.org/2869413002
Cr-Commit-Position: refs/heads/master@{#45248}
Committed: https://chromium.googlesource.com/v8/v8/+/4ea91a0190ab5a167791eb0e2e47d355068193a3

[Michael Lippautz, 2017-05-10 13:39:07]

Description was changed from

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

BUG=chromium:651354
==========

to

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

BUG=chromium:651354
==========

[Michael Lippautz, 2017-05-10 13:39:07]

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

[Michael Lippautz, 2017-05-10 13:39:16]

PTAL

[Michael Lippautz, 2017-05-10 13:45:42]

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

[Michael Lippautz, 2017-05-10 13:45:48]

+jochen

[Hannes Payer (out of office), 2017-05-10 13:46:56]

LGTM, I guess

[jochen (gone - plz use gerrit), 2017-05-10 13:48:02]

test maybe?

[Michael Lippautz, 2017-05-10 20:31:01]

Description was changed from

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

BUG=chromium:651354
==========

to

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the minor
MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354
==========

[Michael Lippautz, 2017-05-10 20:31:01]

Description was changed from

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

BUG=chromium:651354
==========

to

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the minor
MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354
==========

[Michael Lippautz, 2017-05-10 20:31:41]

Description was changed from

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the minor
MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354
==========

to

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the
minor MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354
==========

[Michael Lippautz, 2017-05-10 20:34:27]

Updated the description to include why we need the different mode. Picking the
wrong mode will crash on the first minor mc in layout tests.

I'd really like to land this soon to continue investigating whether the minor MC
is feasible at all.

[Michael Lippautz, 2017-05-10 20:34:36]

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

[commit-bot: I haz the power, 2017-05-10 20:34:40]

Dry run: 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, 2017-05-10 21:25:29]

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

[commit-bot: I haz the power, 2017-05-10 21:25:30]

Dry run: This issue passed the CQ dry run.

[jochen (gone - plz use gerrit), 2017-05-11 07:58:07]

lgtm

[Michael Lippautz, 2017-05-11 08:05:17]

The CQ bit was checked by mlippautz@chromium.org

[commit-bot: I haz the power, 2017-05-11 08:05:26]

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, 2017-05-11 08:07:00]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1494489917128270, "parent_rev":
"f9c4fc0d204cd3f557712b255119952e41cbc4ad", "commit_rev":
"4ea91a0190ab5a167791eb0e2e47d355068193a3"}

[commit-bot: I haz the power, 2017-05-11 08:07:08]

Description was changed from

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the
minor MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354
==========

to

==========
[heap] MinorMC: Only iterate new space global handles for ptr updates

The reason we need this mode is that IterateRoots for the Scavenger only
captures dependent weak nodes. This is also what we do for marking for the
minor MC.

Since the regular marking might also mark objects that are weakly
(non-dependently) pointed to by nodes we need to capture all of them during
pointers updating. The reason this works for the Scavenger is because we do one
pass at the end of the scavenger (combined with resetting) that captures all
those nodes.

BUG=chromium:651354

Review-Url: https://codereview.chromium.org/2869413002
Cr-Commit-Position: refs/heads/master@{#45248}
Committed:
https://chromium.googlesource.com/v8/v8/+/4ea91a0190ab5a167791eb0e2e47d355068...
==========

[commit-bot: I haz the power, 2017-05-11 08:07:09]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/v8/v8/+/4ea91a0190ab5a167791eb0e2e47d355068...