1426801 - Intermittent leakcheck | tab process: 40 bytes leaked (FirstIdleRunnable)

Status: RESOLVED FIXED

(Core :: DOM: Content Processes, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: aiakab [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=153055170&repo=autoland

https://queue.taskcluster.net/v1/task/SdAwKOBDRHWdiaur1MwMxg/runs/0/artifacts/public/logs/live_backing.log

Comment 1

[Andrew McCreight [:mccr8]]

This probably used to be bug 1416927 before I landed bug 1410209.

Comment 2

[Intermittent Failures Robot]

5 failures in 590 pushes (0.008 failures/push) were associated with this bug in the last 7 days.    

Repository breakdown:
* autoland: 4
* mozilla-central: 1

Platform breakdown:
* linux32: 2
* osx-10-10: 1
* linux64: 1
* linux32-stylo-disabled: 1

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1426801&startday=2017-12-18&endday=2017-12-24&tree=all

Comment 3

[Intermittent Failures Robot]

8 failures in 147 pushes (0.054 failures/push) were associated with this bug in the last 7 days. 

This is the #45 most frequent failure this week.   

Repository breakdown:
* mozilla-inbound: 3
* mozilla-central: 3
* autoland: 2

Platform breakdown:
* linux32: 4
* linux32-stylo-disabled: 2
* osx-10-10: 1
* linux64: 1

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1426801&startday=2017-12-25&endday=2017-12-31&tree=all

Comment 4

[Andrew McCreight [:mccr8]]

I looked at two logs where this is happening, and it seemed like a very short-lived process. gFirstIdleTask is cleared in FirstIdle.

    RefPtr<CancelableRunnable> firstIdleTask = NewCancelableRunnableFunction("FirstIdleRunnable",
                                                                             FirstIdle);
    gFirstIdleTask = firstIdleTask;
    NS_IdleDispatchToCurrentThread(firstIdleTask.forget());

If the call to NS_IdleDispatchToCurrentThread fails, we'd end up leaking gFirstIdleTask. I can try clearing gFirst if the dispatch fails.

Comment 5

[Andrew McCreight [:mccr8]]

Another possibility is that when ContentChild::ActorDestroy() is called when gFirstIdleTask is defined, the runnable is cancelled, but gFirstIdleTask is not cleared. That seems like a more likely cause of this.

Comment 6

[Andrew McCreight [:mccr8]]

Attachment: Bug 1426801 - Clear gFirstIdleTask if we fail to dispatch the runnable.

Review commit: https://reviewboard.mozilla.org/r/210310/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/210310/

Comment 7

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 8940024 [details]
Bug 1426801 - Clear gFirstIdleTask if we fail to dispatch the runnable.

https://reviewboard.mozilla.org/r/210310/#review216036

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/1eb22728208e
Clear gFirstIdleTask if we fail to dispatch the runnable. r=mrbkap

Comment 9

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/1eb22728208e

Comment 10

[Intermittent Failures Robot]

9 failures in 462 pushes (0.019 failures/push) were associated with this bug in the last 7 days.    

Repository breakdown:
* autoland: 4
* mozilla-inbound: 3
* try: 1
* mozilla-central: 1

Platform breakdown:
* linux64: 3
* linux32-stylo-disabled: 3
* linux32: 3

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1426801&startday=2018-01-01&endday=2018-01-07&tree=all

Comment 11

[Ryan VanderMeulen [:RyanVM]]

Please request Beta approval on this when you get a chance. It grafts cleanly.

Comment 12

[Andrew McCreight [:mccr8]]

Comment on attachment 8940024 [details]
Bug 1426801 - Clear gFirstIdleTask if we fail to dispatch the runnable.

Approval Request Comment
[Feature/Bug causing the regression]: probably bug 1324428
[User impact if declined]: None. This just fixes an intermittent orange on Treeherder.
[Is this code covered by automated tests?]: yes
[Has the fix been verified in Nightly?]: yes
[Needs manual test from QE? If yes, steps to reproduce]: no
[List of other uplifts needed for the feature/fix]: none
[Is the change risky?]: no
[Why is the change risky/not risky?]: this only does a little bit of extra cleanup in a rare situation during shutdown
[String changes made/needed]: none

Comment 13

[Gerry Chang [:gchang]]

Comment on attachment 8940024 [details]
Bug 1426801 - Clear gFirstIdleTask if we fail to dispatch the runnable.

Fix an intermittent failure. Beta58+.

Comment 14

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-beta/rev/a1d9a99f12a5