Closed Bug 1065928 Opened 10 years ago Closed 10 years ago

plugin-container is not unloading

Categories

(Core Graveyard :: Plug-ins, defect)

31 Branch
x86
Linux
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED DUPLICATE of bug 1011339

People

(Reporter: MetamAdeptus, Unassigned)

Details

User Agent: Mozilla/5.0 (X11; Linux i686; rv:31.0) Gecko/20100101 Firefox/31.0
Build ID: 20140717132905

Steps to reproduce:

I'm using Firefox 31.0 (esr) in Ubuntu GNU/Linux 12.04.5. I'm using Caffeine applet to prevent screensaver from running, when some video content is played in Firefox, so 'plugin-container' is listed among the processes triggering this behavior. It worked fine when I was using Firefox 24.


Actual results:

After upgrade to Firefox 31 I noticed, that even after I close all tabs, containing video content, Caffeine remains active. There was a timeout to unload 'plugin-container' in Firefox 24 (approximately 3 minutes), now it stays in memory indefinitely. I can confirm this by running 'ps aux | grep plugin-container'.


Expected results:

plugin-container should be unloaded after the certain amount of time as it was in earlier versions of Firefox. I found bug reports about memory leaks, caused by plugin-container, so it seems to me that it's not good to leave it in memory more than needed.
This sounds a bit like bug 1011339 - can you try this on Nightly to see if it is fixed there?
https://nightly.mozilla.org/
Flags: needinfo?(whtyger)
That's strange. I cannot reproduce the problem in VirtualBox machine. I got Firefox 31.0esr, installed it to /opt and reused my profile from original Firefox 27, which was installed in VirtualBox. plugin-container is unloaded correctly after timeout. I tried several times, with and without closing Firefox.
Maybe this is some problem with my current profile in my desktop machine installation?
Flags: needinfo?(whtyger)
If this is like bug 1011339, you could try deleting pluginreg.dat in the profile, that should trigger it.
Yes, this really helped.
I renamed pluginreg.dat and it was recreated at new Firefox start. What's interesting - it is absolutely the same as the previous one, I compared them with 'diff'. Nevertheless, now plugin-container unloads correctly after timeout.
Thanks for your help.
Will this update apply to ESR version of Firefox, or it will be applied only to standard version?
(In reply to Whtyger from comment #4)
> I renamed pluginreg.dat and it was recreated at new Firefox start. What's
> interesting - it is absolutely the same as the previous one, I compared them
> with 'diff'.

That's expected, we just needed to trigger scanning all the plugins again to trigger the bug.

> Thanks for your help.
> Will this update apply to ESR version of Firefox, or it will be applied only
> to standard version?

This is fixed in Firefox 32, so it's not in ESR31. This doesn't seem too critical (it's only happening the first time a new plugin is found, afterwards its fine), so i'd expect the fix to go the long way to ESR38.

Thanks for confirming.
Status: UNCONFIRMED → RESOLVED
Closed: 10 years ago
Resolution: --- → DUPLICATE
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.