1224022 - Intermittent test_WebAudioMemoryReporting.html | Non-zero usage for explicit/webaudio/audio-node/ScriptProcessorNode/dom-nodes

Status: RESOLVED FIXED

(Core :: Web Audio, defect, P2)

Description

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://treeherder.mozilla.org/logviewer.html#?job_id=17139504&repo=mozilla-inbound

Comment 1

[Karl Tomlinson (:karlt)]

Attachment: produce memory report after processing main thread messages

which may add new streams.

Comment 2

[Intermittent Failures Robot]

13 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-beta: 8
* mozilla-inbound: 3
* b2g-inbound: 2

Platform breakdown:
* windows7-32: 5
* linux64: 2
* windows8-64: 1
* osx-10-6: 1
* osx-10-10: 1
* linux32: 1
* android-4-3-armv7-api11: 1
* android-2-3-armv7-api9: 1

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1224022&startday=2015-11-09&endday=2015-11-15&tree=all

Comment 3

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/16bd0e0af41c

Comment 4

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/16bd0e0af41c

Comment 5

[Intermittent Failures Robot]

7 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-inbound: 4
* mozilla-beta: 1
* mozilla-aurora: 1
* b2g-inbound: 1

Platform breakdown:
* windows7-32: 6
* osx-10-6: 1

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1224022&startday=2015-11-16&endday=2015-11-22&tree=all

Comment 6

[Intermittent Failures Robot]

6 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* b2g-inbound: 3
* mozilla-inbound: 1
* mozilla-beta: 1
* fx-team: 1

Platform breakdown:
* windows7-32: 6

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1224022&startday=2016-01-11&endday=2016-01-17&tree=all

Comment 7

[Karl Tomlinson (:karlt)]

Attachment: bug 1224022 document mLifecycleState multithread access management

Review commit: https://reviewboard.mozilla.org/r/62062/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/62062/

Comment 8

[Karl Tomlinson (:karlt)]

Attachment: bug 1224022 don't try to wake up an empty MSG that is shutting down to report memory

Review commit: https://reviewboard.mozilla.org/r/62064/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/62064/

Comment 9

[Karl Tomlinson (:karlt)]

Attachment: bug 1224022 use pointers instead of copies of string literals in AudioNode memory reporting

Review commit: https://reviewboard.mozilla.org/r/62066/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/62066/

Comment 10

[Karl Tomlinson (:karlt)]

Attachment: bug 1224022 store a node type string pointer on engines and use that for memory reporting

removing some off-main-thread usage of AudioNodeEngine::mNode.

Review commit: https://reviewboard.mozilla.org/r/62068/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/62068/

Comment 11

[Karl Tomlinson (:karlt)]

Attachment: bug 1224022 collect memory reports of AudioNode dom objects on the main thread

This will permit allowing the main thread to run while collecting
reports from graph thread objects.

Review commit: https://reviewboard.mozilla.org/r/62070/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/62070/

Comment 12

[Karl Tomlinson (:karlt)]

Attachment: bug 1224022 record MSG memory sizes asynchronously

This removes the one second timeout for MSG collection, extending the timeout
period to the 50 second timeout of nsMemoryReporterManager.

Also removed:

* The condition variable logic that can stop waiting without checking the
  condition set when memory reports are complete.

* Races with mAudioStreamSizes modification on two threads after wait timeout.

Memory from streams in offline graphs that are not yet running is now also
included.

Review commit: https://reviewboard.mozilla.org/r/62072/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/62072/

Comment 13

[Karl Tomlinson (:karlt)]

Comment on attachment 8687031 [details] [diff] [review]
produce memory report after processing main thread messages

Sorry.  Wrong bug.

Comment 14

[Paul Adenot (:padenot)]

Comment on attachment 8767602 [details]
bug 1224022 document mLifecycleState multithread access management

https://reviewboard.mozilla.org/r/62062/#review58768

Comment 15

[Paul Adenot (:padenot)]

Comment on attachment 8767603 [details]
bug 1224022 don't try to wake up an empty MSG that is shutting down to report memory

https://reviewboard.mozilla.org/r/62064/#review58770

Comment 16

[Paul Adenot (:padenot)]

Comment on attachment 8767604 [details]
bug 1224022 use pointers instead of copies of string literals in AudioNode memory reporting

https://reviewboard.mozilla.org/r/62066/#review58772

Comment 17

[Paul Adenot (:padenot)]

Comment on attachment 8767605 [details]
bug 1224022 store a node type string pointer on engines and use that for memory reporting

https://reviewboard.mozilla.org/r/62068/#review58774

Comment 18

[Paul Adenot (:padenot)]

Comment on attachment 8767606 [details]
bug 1224022 collect memory reports of AudioNode dom objects on the main thread

https://reviewboard.mozilla.org/r/62070/#review58778