With recent 2.6 kernels, the command brctl delbr br0 hangs during shutdown, and the kernel prints this message again and again: unregister_netdevice: waiting for br0 to become free. Usage count = 1 I was able to reproduce this problem with several kernels between (and including) 2.6.9-rc1 and 2.6.11-rc2-bk9. I haven't tried later versions than 2.6.11-rc2-bk9 yet. 2.6.8.1-bk2 still works fine. My .config was taken from the Debian kernel-image-2.6.10-1-686 package, adapted to the kernel.org kernels and the processor type set to Pentium Classic. I can email it on request. This script should reproduce the problem reliably: #!/bin/bash set -x mount proc -t proc /proc ifconfig lo 127.0.0.1 brctl addbr br0 modprobe e100 # also works with 3c5x9 brctl addif br0 eth0 ifconfig br0 192.168.1.1 ifconfig eth0 up sleep 5 ifconfig lo down lsmod | grep bridge brctl delif br0 eth0 ifconfig br0 down brctl delbr br0 To run it, I booted with init=/bin/bash to make sure the distribution's startup scripts don't interfere (although it might work anyway). I already reported this to the linux-kernel list, since I couldn't tell whether this is a bug in the bridging code or somewhere else. On second reading of REPORTING-BUGS, I probably should have sent it here first. For reference, these are my postings to linux-kernel: http://www.ussg.iu.edu/hypermail/linux/kernel/0501.3/2222.html http://www.ussg.iu.edu/hypermail/linux/kernel/0502.0/0139.html Please let me know if you need any further information, and please Cc me on replies, since I am not subscribed to the bridge list. Mirko
Stephen Hemminger
2007-Apr-18 17:22 UTC
[Bridge] recent 2.6 kernels hang on bridge shutdown
On Wed, 2 Feb 2005 18:56:52 +0100 Mirko Parthey <mirko.parthey@informatik.tu-chemnitz.de> wrote:> With recent 2.6 kernels, the command > brctl delbr br0 > hangs during shutdown, and the kernel prints this message again and again: > unregister_netdevice: waiting for br0 to become free. Usage count = 1Are you running IPV6? if so it is most likely related to IPV6.> I was able to reproduce this problem with several kernels between > (and including) 2.6.9-rc1 and 2.6.11-rc2-bk9. > I haven't tried later versions than 2.6.11-rc2-bk9 yet. > 2.6.8.1-bk2 still works fine. > > My .config was taken from the Debian kernel-image-2.6.10-1-686 package, > adapted to the kernel.org kernels and the processor type set to Pentium > Classic. I can email it on request. > > This script should reproduce the problem reliably: > #!/bin/bash > set -x > mount proc -t proc /proc > ifconfig lo 127.0.0.1 > brctl addbr br0 > modprobe e100 # also works with 3c5x9 > brctl addif br0 eth0 > ifconfig br0 192.168.1.1 > ifconfig eth0 up > sleep 5 > ifconfig lo down > lsmod | grep bridge > brctl delif br0 eth0 > ifconfig br0 down > brctl delbr br0 > > To run it, I booted with init=/bin/bash to make sure the distribution's > startup scripts don't interfere (although it might work anyway). > > I already reported this to the linux-kernel list, since I couldn't tell > whether this is a bug in the bridging code or somewhere else. On second > reading of REPORTING-BUGS, I probably should have sent it here first. > > For reference, these are my postings to linux-kernel: > http://www.ussg.iu.edu/hypermail/linux/kernel/0501.3/2222.html > http://www.ussg.iu.edu/hypermail/linux/kernel/0502.0/0139.html > > Please let me know if you need any further information, > and please Cc me on replies, since I am not subscribed to the bridge list. > > Mirko-- Stephen Hemminger <shemminger@osdl.org>
Reasonably Related Threads
- Fwd: Running R from terminal in ubuntu 16.04 gives Fatal Error message
- Badness in local_bh_enable at kernel/softirq.c:140
- Fwd: Running R from terminal in ubuntu 16.04 gives Fatal Error message
- FW: DNAT and masq problem with kernel 2.4.23
- yum segfault - rpmforge problem?