Open Bug 1603158 Opened 6 years ago Updated 3 years ago

Use AsyncShutdown barrier for many things in nsIRemoteAgent.close()

Categories

(Remote Protocol :: Agent, defect, P3)

defect

Tracking

(Not tracked)

People

(Reporter: ato, Unassigned)

References

()

Details

nsIRemoteAgent.close() generally gets called so late in the shutdown process that many of the cleanup tasks it performs will fail because the relevant subsystems have already been shutdown.

We should employ async shutdown barriers for the things we need to clean up so that when close() is called, most of the state will already been reset.

Priority: -- → P3

A single owned barrier will be ok here. Then different clients can register, and the owner will wait until each of those have been finished their own shutdown. We should include the one for the stream registry then too.

Summary: Use async shutdown barriers for many things in nsIRemoteAgent.close() → Use AsyncShutdown barrier for many things in nsIRemoteAgent.close()
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.