gnome-system-monitor crashes when wifi (fwlanusb) statistics are queried

Bug #222213 reported by Simon Dierl
10
Affects Status Importance Assigned to Milestone
linux-restricted-modules-2.6.24 (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

Architecture: i386
DistroRelease: Ubuntu 8.04
Versions:
gnome-system-monitor: 2.22.0-1ubuntu3
linux-restricted-modules-2.6.24-16-generic: 2.6.24.12-16.34

When opening gnome-system-monitor or displaying the "Network" graph in the respective applet, the application crashes.

When the application itself is used, a window is drawn filled with white colour, it can be force-quit (via GNOME since the application doesn't answer to clicking on "X"), but the gnome-system-monitor process will stay and cannot be removed even with kill -9.
When the applet is used (only when the "Network" graph is drawn), it immediately crashes with the appropriate error messages (reload etc.), but the malicious effects are the same as with the application (presumably parts stay in memory as well).

After the crash has been provoked, most application fail to work correctly.
sudo (even if invoked from a "real" terminal, not a terminal emulator under X (gnome-terminal doesn't work, anyway, so I don't know if it works there)), hangs after the password has been typed in, most other command line applications seem to work correctly. Most graphical applications do not work (firefox, for example, creates the "firefox" process, so apparently the script hangs, I haven't checked if other applications create a process). Panel applets do not seem to work as well, including the shutdown applet. If a restart of X is provoked (Ctrl+Alt+Backspace), X shuts down and nothing happens afterwards. A reboot via Ctrl+Altl+Del doesn't work, either (although is seems that large parts of the reboot process are completed).

If my wireless device (Fritz! WLAN USB Stick, using proprietary fwlanusb driver included in linux-restricted-modules-2.6.24-16-generic) is not present, the monitor starts normally. The bug occurs independently of network connectivity status. If the stick is plugged in while the system monitor is already running, the crash occurs. The bug is apparently related to the device/driver or to WLAN connections in general (?) since the network statistics of my LAN port are displayed without problems.

In Ubuntu 7.10, the bug did not occur, it was introduced with the upgrade to 8.04 (the applet is part of my panel).

The bug can be provoked in a Live-CD environment.

Tags: hardy
Revision history for this message
Philipp Decker (relpaed) wrote :

Experienced the same thing, exactly as described by Simon above; same configuration.

There are two scenarios, both fail:
1. Fritz Wlan-Stick is plugged in befor booting up; booting works, even the login. Right after the login, gnome-system-monitor crashes, etc...
2. if plugged in after logging in the same thing occurs. However, the fwlanusb module is loaded correctly.

If the stick is unplugged, it still does not work; since sudo does not work any more the fwlanusb module cannot be unloaded - the only solution is a hot reset, which is evil...

Revision history for this message
aleandro (aleandrodasilva) wrote :

Same by my configuration.

Scenarios:

1. the same
2. if plugged in after logging nothing happens but the window of proprietary drivers opens (three files in the usb stick). When I run the .inf the usb stick tries to connect via wireless to internet. After that is all as described in scenario 1.

Forced reboot is the only solution.
Remark: under ubuntu 7.10 worked out of box.

If you want I can provide screenshots before the total crash of the gnome system.

Revision history for this message
Simon Dierl (simon.dierl) wrote :

Another way to provoke the bug is to use the stick with NetworkManager (instead of "manual configuration"). As soon as an attempt to connect to a network is made, the system breaks down (the nm-applet reports that it is "connecting" and doesn't even get to ask for the encryption key). Additionally, the displayed data is incorrect, e.g. the signal strength is always zero.

Can also be observed on Live-CD (to avoid messing up file systems).

Revision history for this message
Simon Dierl (simon.dierl) wrote :

It seems to be possible to bring down the system without a reset, switch to a virtual terminal and press Ctrl+Alt+Del, the machine gets down to a blinking line in the top left corner, press Ctrl+Alt+Del again to sync the disks and reboot the system. Apparently, the file systems are not messed up this way.

During reboot, when killing NetworkManager (has to be terminated, it refuses to exit during regular shutdown procedures), it prints the following message:

NetworkManager: <WARN> nm_hal_deinit(): libhal shutdown failed - Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

Still "works" with:
linux-restricted-modules 2.6.24.17.19
gnome-system-monitor 2.22.1-0ubuntu2
although now you have to look at the "Resources" tab in the monitor to crash the system.

Revision history for this message
exactt (giesbert) wrote :

hi guys,

have you checked your dmesg output and compared it to the problems mentioned here: https://bugs.launchpad.net/ubuntu/+source/linux-restricted-modules-2.6.24/+bug/200589 ?

Revision history for this message
Simon Dierl (simon.dierl) wrote :

I get a similar output when provoking the bug with gnome-system-monitor, although the part where the device is suddenly removed and reinitialized a few seconds later (959.866236-982.994272) does not occur. The call stack is only slightly different from mine (see dmesg.log and traces.diff).

Since the other bug is provoked by netspeed_applet, I suspect that tuco (and most of the other posters) use NetworkManager to configure their WLANs. As mentioned above, this immediately kills the system, too (I use the "manual configuration"). Their machines boot up, try to use NetworkManager and provoke a kernel? bug that prevents most applications from being launched (including GDM), rendering the system unusable.

Although I do not get the connection issue when using gnome-system-monitor, I have already seen this output together with the URB/usb_read_write_sync messages from his syslog after some random lockups (they are probably all aspects of the same problem) of my system (parts from my kern.log are attached, they are NOT RELATED to the gnome-system-manager issue!). The last timestamp in my kern.log (3502.807284) is plainly wrong, the system was long rebooted by then.

I have compressed the logs to upload them as a single file.

Revision history for this message
collinp (collinp) wrote :

I am having a very similar bug, the difference is that i dont use wifi. But the after effects is near the same, firefox wont open, no panel apps work, barely anything works, a x restart does not solve it.

Revision history for this message
J.G. (j.gorski) wrote :

I can't open gnome-system-monitor, after a frequent use of firefox. Firefox starts to fail scrolling pages, then it freezes, and i am not able to open the monitor to kill the process. Similar to Collin problem. It's not ok, that you can't open the monitor, i think it would must be a priority.

Bryce Harrington (bryce)
tags: added: hardy
Revision history for this message
bugbot (bugbot) wrote :

This bug report was filed against an old version of Ubuntu.
Can you confirm whether this is still an issue in natty?

If you don't mind, it would be very helpful if you could update the bug
report in launchpad to 'Fix Released' if it is no longer an issue for
you, or if it is still occurring under natty, please tag the bug 'natty'
so it's easier for us to track.

Changed in linux-restricted-modules-2.6.24 (Ubuntu):
status: New → Incomplete
Revision history for this message
Simon Dierl (simon.dierl) wrote :

The Linux Kernel from Ubuntu 8.10 upwards will not load fwlanusb (GPL-violating USB driver), thus, "Fix Released".

Changed in linux-restricted-modules-2.6.24 (Ubuntu):
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.