The EverySecondTelemetryCallback_m timer fires every second until xpcom shutdown
Categories
(Core :: WebRTC: Signaling, defect)
Tracking
()
Tracking | Status | |
---|---|---|
firefox114 | --- | fixed |
People
(Reporter: florian, Assigned: bwc)
References
(Blocks 1 open bug)
Details
Attachments
(1 file)
(deleted),
text/x-phabricator-request
|
Details |
The new check I'm adding in bug 1828013 at the end of every browser chrome mochitests makes this leftover repeating timer visible when running the browser_WebrtcGlobalInformation.js
test which creates peer connections in the parent process.
From looking at the code, it seems this timer is created in PeerConnectionCtx::Initialize
, called by PeerConnectionCtx::InitializeGlobal
which creates a PeerConnectionCtx
instance stored in the global gInstance
variable, which is only cleared by PeerConnectionCtx::Destroy
, which is only called when receiving the "xpcom-will-shutdown"
observer notification.
Could this timer instead be canceled when the last peer connection of the process is closed?
Reporter | ||
Comment 1•1 year ago
|
||
EverySecondTelemetryCallback_m was already causing a performance issue in bug 1724417. At the time I didn't understand that the timer kept firing even after all peer connections were closed.
The reason I'm filing this bug today is that it blocks landing of bug 1828013 (well, if there's no quick fix I'll find a way to add an exception for this timer, but it's unfortunate).
Assignee | ||
Comment 2•1 year ago
|
||
I think it would be possible to stop/resume when all PCs are gone. It might even clear up some intermittent failures on try.
Assignee | ||
Updated•1 year ago
|
Assignee | ||
Comment 3•1 year ago
|
||
Assignee | ||
Comment 4•1 year ago
|
||
Comment 6•1 year ago
|
||
bugherder |
Description
•