#775 PCManFM 99% CPU when ejecting SD mounted using udisks

[Camden Lindsay]

– version of pcmanfm you are using (you can see it in "About" dialog)
PCManFM 1.1.2

– version of your operating system
[dylan@zenbook ~]$ uname -a
Linux zenbook 3.10.7-1-ARCH #1 SMP PREEMPT Thu Aug 15 11:55:34 CEST 2013 x86_64 GNU/Linux

[dylan@zenbook ~]$ cat /etc/lsb-release
LSB_VERSION=1.4
DISTRIB_ID=arch
DISTRIB_RELEASE=rolling
DISTRIB_DESCRIPTION="Arch Linux"

[dylan@zenbook ~]$ fluxbox --version
Fluxbox 1.3.5 : (c) 2001-2011 Fluxbox Team

[dylan@zenbook ~]$ sudo pacman -S udisks
warning: udisks-1.0.4-8 is up to date -- reinstalling

– steps how we can reproduce your problem
Udisks automounts SD card
Use 'eject' button in PCManFM to eject device
CPU hangs at 99%
11079 dylan 20 0 445664 15808 10664 R 99.8 0.4 0:09.32 pcmanfm

Manually umounting device does not cause this, only using eject button

Bug report FS#36686 with Arch
https://bugs.archlinux.org/task/36686

[Lonely Stranger]

I don't know what is the glib-debug but debug info should be enabled at build time as I wrote you below, you should rebuild package glib2 for that.

How to show threads in top is told in its manual - just press 'H'. :)

[Lonely Stranger]

Thank you very much for the testing and feedback provided!

[Lonely Stranger]

To view all threads backtrace you can use command in gdb:

thread apply all backtrace

Also sometimes it may help to run pcmanfm under strace utility.

[Camden Lindsay]

I came to something here... by accident

I had the following packages installed from the repositories:
glib2
glib2-debug

I re-read all of the comments and saw that you wanted be to build glib2 from AUR glib2-aur with debug enabled.

I did that, removed the glib2 and glib2-debug, and the problem no longer exists.

to recreate, i uninstalled glib2-aur, reinstalled only glib2 from repository.

Using the repository version of glib2, the problem will occur every time (with or without package glib2-debug installed).

There must be some changelist between the repository version and the aur version of glib2 that has fixed the problem! :)

[PICCORO Lenz McKAY]

Hey Camden could u investigate the changes between the glib2 and glib2-aur? i have similar problem, but only when i let connected removable drives and power on the machine with them inserted!

[Camden Lindsay]

@piccoro
I'm running
[dylan@zenbook glib2-git]$ sudo pacman -Q | grep glib2
glib2-git 2.39.91.23.g6220657-1

The current repository version in standard repository is
[dylan@zenbook glib2-git]$ sudo pacman -S -i glib2 | grep Version
Version : 2.40.0-1

I'll run a test quick like to see if i can install the standard version now and if I have the same problem or not.

[Camden Lindsay]

I can verify now that the pcmanfm and libfm in the Arch repositories now work together happily with respect to this bug.

[dylan@zenbook ~]$ pacman -Qi pcmanfm | grep Version
Version : 1.2.0-2

[dylan@zenbook ~]$ pacman -Qi libfm | grep Version
Version : 1.2.0-5

However there is a different behavior when i click the 'eject' button for the removable media; not sure if it is on purpose or not.

Before clicking 'eject' would unmount as well as eject the device.

Now it only unmounts, and the drive stays on the sidebar and on the bus.

Is this considered normal? (if anyone who knows is listening)

[Lonely Stranger]

I'm sorry for very much late answer but yes, that is normal for SD card because strictly speaking, it is not ejectable so it can be mounted again until you disconnect it.

I'm not sure what to do with this report because I still could not find a workaround and you're not able to reproduce it anymore as well.

[Camden Lindsay]

I think you can mark as 'fixed/unable to recreate' or something similar.
I think some update has fixed the initial problem :)

Many thanks!