Bug 419139

Summary:	Process runs at 99-100% of core until shutdown and reoccurs again later
Product:	[Plasma] plasma-browser-integration	Reporter:	BeowulfRubix <r.ramyar>
Component:	Firefox	Assignee:	Kai Uwe Broulik <kde>
Status:	CONFIRMED ---
Severity:	normal	CC:	fabian, goo, schro.sb
Priority:	NOR
Version First Reported In:	unspecified
Target Milestone:	---
Platform:	Neon
OS:	Linux
Latest Commit:		Version Fixed/Implemented In:
Sentry Crash Report:
Attachments:	image.png

Description BeowulfRubix 2020-03-23 12:45:12 UTC

SUMMARY


STEPS TO REPRODUCE
1. Fan runs like crazy, triggering me to check process loads
2. Kill process, fan stops
3. Problem reoccurs later 
4. I have noticed this has been happening for some months, but don't know how long it has being for

OBSERVED RESULT


EXPECTED RESULT


SOFTWARE/OS VERSIONS
Windows: 
macOS: 
Linux/KDE Plasma: 
(available in About System)
KDE Plasma Version: 5.18.3
KDE Frameworks Version: 5.68.0
Qt Version: 5.14.1

ADDITIONAL INFORMATION

Seems related to https://bugs.kde.org/show_bug.cgi?id=413286, which was Arch and chrome

Comment 1 BeowulfRubix 2020-03-23 12:46:55 UTC

Also related to this https://bbs.archlinux.org/viewtopic.php?id=252083

Comment 2 BeowulfRubix 2020-03-23 12:55:38 UTC

Just realised that firejail might have something to do with it. Just found: 

https://github.com/netblue30/firejail/wiki/Frequently-Asked-Questions#how-do-i-enable-plasma-browser-integration-in-firefox

Will see if that fixes it

Comment 3 Kai Uwe Broulik 2020-03-23 13:53:52 UTC

Can you provide a backtrace or valgrind when it does that? I don't even see it starting the host when run through firejail.

Comment 4 BeowulfRubix 2020-03-23 16:15:45 UTC

Created attachment 126970 [details]
image.png

Definitely starting. I actually just noticed that I might be running a
nested firejail of firejail for firefox(!), but that shouldn't matter. Just
showing that it launches in this grab:

[image: image.png]

I've never backtraced a running process, so I just installed dbg package
for plasma-browser-integration. Let's see

On Mon, 23 Mar 2020 at 13:53, Kai Uwe Broulik <bugzilla_noreply@kde.org>
wrote:

> https://bugs.kde.org/show_bug.cgi?id=419139
>
> --- Comment #3 from Kai Uwe Broulik <kde@privat.broulik.de> ---
> Can you provide a backtrace or valgrind when it does that? I don't even
> see it
> starting the host when run through firejail.
>
> --
> You are receiving this mail because:
> You reported the bug.

Comment 5 Kai Uwe Broulik 2020-03-23 16:53:36 UTC

To debug a running process:
gdb --pid `pidof plasma-browser-integration-host`
and then perhaps type bt to see what it was doing, or Ctrl+C to interrupt it first.

Comment 6 Sebastian Schmidt 2020-04-25 22:20:45 UTC

I had the same issue with chromium for quite some while, but never had the time to investigate. The plasma-browser-integration-host process would always run at 100% usage of one core as soon as a video starts playing and the media widget didn't detect the video.
However, this never happened with firefox.

For me this issue was caused by DBUS_SESSION_BUS_ADDRESS not being set.
I changed the session to "Exec=dbus-run-session /usr/bin/startplasma-x11" instead of "Exec=/usr/bin/startplasma-x11". I'm not sure if this has any negative consequences, but the wayland session also uses dbus-run-session

Comment 7 goo 2023-06-05 17:38:46 UTC

Still suffering from this issue in 2023 and I cant' even remember when this bug first appeared on my desktop.
On the core running plasma-browser-integration-host I get 100% CPU usage at login time until I send an HUP signal to plasma-browser-integration-host. This happens on multiple machines running updated Fedora Linux.
Might the problem be the DBUS_SESSION_BUS_ADDRESS variable not yet being set when plasma-browser-integration-host first starts on session init?

Comment 8 goo 2024-12-19 23:56:30 UTC

plasma-browser-integration high CPU utilization is no longer a problem since Plasma 6.2.
The issue seems to be resolved for me on both my desktops.