1022212 - Enable WebRTC load adaptation by default

Status: RESOLVED FIXED

(Core :: WebRTC: Audio/Video, defect)

Description

[Randell Jesup [:jesup] (needinfo me)]

Per private conversation, the load adaptation code is ready to enable by default, and works across all platforms (and even BSD).

Comment 1

[Randell Jesup [:jesup] (needinfo me)]

Attachment: enable WebRTC load adaptation by default rs=jesup,gcp

Comment 2

[Randell Jesup [:jesup] (needinfo me)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/650ea68d7ef2

Comment 3

[Ryan VanderMeulen [:RyanVM]]

I'm seeing a lot of webrtc-related crashes on inbound since this landed. Hard to tell for sure given the high rate of crashes we were already seeing before (*sigh*), but it still does appear to have ticked upwards a bit.

Comment 4

[Ryan VanderMeulen [:RyanVM]]

Backed out for causing bug 1022235.

https://hg.mozilla.org/integration/mozilla-inbound/rev/ff91eb765497

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/6cc9ac1f57fa
No longer will leak threads.  Timing changes could affect the root cause for the crashtest bugs still.

Comment 6

[Ryan VanderMeulen [:RyanVM]]

So we're still expecting the crashes? Good, I guess...
https://tbpl.mozilla.org/php/getParsedLog.php?id=41324327&tree=Mozilla-Inbound

Same as before, about to run out the door. Will backout when I return if you don't beat me to it.

Comment 7

[Randell Jesup [:jesup] (needinfo me)]

(In reply to Ryan VanderMeulen [:RyanVM UTC-4] from comment #6)
> So we're still expecting the crashes? Good, I guess...
> https://tbpl.mozilla.org/php/getParsedLog.php?id=41324327&tree=Mozilla-
> Inbound
> 
> Same as before, about to run out the door. Will backout when I return if you
> don't beat me to it.

Sorry, I think I wasn't clear.  The underlying cause of a lot of crashes is an ASAN problem that is not caused by LoadManager (and apparently recent).  The bug in load manager that leaked threads seemed to provoke the ASAN bug and cause more-frequent hits esp on Android.  Removing that bug hopefully will cause this to no longer provoke the ASAN bug more than normal; but in any case the ASAN bug is splattering all over the tree, with or without load manager.

I plan to dig into the ASAN bug later tonight.

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/6cc9ac1f57fa

Comment 9

[Ed Morley [:emorley]]

This push backed out for Android crashes during 408431-1.html:
https://tbpl.mozilla.org/?tree=Mozilla-Inbound&jobname=Android%204.0%20Panda%20mozilla-inbound%20opt%20test%20crashtest&fromchange=c66224db3600&tochange=6cc9ac1f57fa

I saw bug 1022235 comment 8 - however the existing 408431-1.html crashes were on other platforms - Android 4.0 debug m-6 was green before, with these two changesets, it isn't.

remote:   https://hg.mozilla.org/integration/mozilla-inbound/rev/dc93584a0948

Comment 10

[Randell Jesup [:jesup] (needinfo me)]

Relanding with agreement in IRC from edmorely -- see bug 1022235
https://hg.mozilla.org/integration/mozilla-inbound/rev/fc05ea905ce5

Comment 11

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/fc05ea905ce5