|
Bugzilla – Full Text Bug Listing |
| Summary: | unregister_netdevice: waiting for modemB to become free. Usage count = 1 | ||
|---|---|---|---|
| Product: | [openSUSE] openSUSE 10.3 | Reporter: | Forgotten User ZhJd0F0L3x <forgotten_ZhJd0F0L3x> |
| Component: | Kernel | Assignee: | Karsten Keil <karsten.keil> |
| Status: | RESOLVED WORKSFORME | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Normal | ||
| Priority: | P5 - None | ||
| Version: | RC 2 | ||
| Target Milestone: | --- | ||
| Hardware: | Other | ||
| OS: | Other | ||
| Whiteboard: | |||
| Found By: | Component Test | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Bug Depends on: | |||
| Bug Blocks: | 140732 | ||
| Attachments: |
/var/log/messages, including sysrq-t
"ps auxwww" ifcfg-modemB ifcfg-modemK (Kabel) ifcfg-modemU wlan LAN /var/log/messages excerpt, including the events leading up to the problem and a sysrq-t |
||
|
Description
Forgotten User ZhJd0F0L3x
2006-02-28 10:36:05 UTC
Created attachment 70598 [details]
/var/log/messages, including sysrq-t
Created attachment 70599 [details]
"ps auxwww"
Note that i had to reboot to get the machine running again.
And no, i have not seen something like that since ~suse 9.1, so it is probably not easily reproducible :-)
Can you try to reproduce it ? modemB (why modemB ?) is the ppp device I guess, please provide all ifcfg-* files. (In reply to comment #3) > Can you try to reproduce it ? give me a testcace :-) > modemB (why modemB ?) is the ppp device I guess, please provide all ifcfg-* > files. modemB == bluetooth. Easier to remember "modemB", "modemU" etc instead of modem0, modem1, ... I will attach them all. Created attachment 70608 [details]
ifcfg-modemB
Created attachment 70609 [details]
ifcfg-modemK (Kabel)
Created attachment 70610 [details]
ifcfg-modemU
Created attachment 70611 [details]
wlan
Created attachment 70612 [details]
LAN
So "give me a testcace" means, you already tried several times to do the same as you describe in "Following situation" with Beta5 with out hitting this problem ? yes. I use dialin and RAS every day and i think (from the ps output) that it already hung when killing rasconnect, before plugging into the suse network. And i have not seen those "waiting for ... to become free" for a long time. It maybe related to this mainline bug report: http://bugzilla.kernel.org/show_bug.cgi?id=6197 The mainline report is about pppoe, but maybe other ppp addons have the same problem. I'll monitor this bug and the ongoing discussion on netdevel. I did not saw any of these problems with newer version, so I think it was solved. If the problem still exist, please reopen it. Still happened today.
What i did:
- At home, i was in the local LAN via interface "cable" (e1000) or "air" (ipw2200)
- While in the LAN, i dialled into the 3G network via a option data card (pppd), default route pointed to the dialup
- started VPN (rasconnect.pl)
- moved from home to office, ppp connection still alive, WLAN probably disconnected
- plugged machine into the office LAN
- called "scpm switch", to switch the networking setup for the office
- scpm switch restarts networking
- this includes stopping smpppd, which did not stop the active pppd.
(no smpppd enabled in the office profile)
- i noticed pppd was still active, so i simply pulled the card out while the scpm switch was running (i thought that the default route pointing to dialup would be a problem, so i wanted to get rid of it).
Now i got the infamous "kernel: unregister_netdevice: waiting for modemX to become free. Usage count = 1" message and all i could do was reboot. Everything that was trying to access /proc/net/ was apparently hanging, also "ifconfig" and "ip".
I did a sysrq-T trace which i will attach.
This is with 10.3RC2
Created attachment 175085 [details]
/var/log/messages excerpt, including the events leading up to the problem and a sysrq-t
Newer kernel did resolve some network usage bugs so I hope that this is resolved with 11.0. If you run into this again with 11.0, please open a new bug. |