[linux-dvb] ifconfig locks on down - B2C2 flexcop.

stonewall at inebraska.com stonewall at inebraska.com
Wed Jul 13 03:26:46 CEST 2005


Johannes - also on this - it is consistent.  It does it every time.  I'm 
actually quite surprised I can Ctl-C out of ifconfig when I do it at the 
command line . . . and interestingly enough, all dvb0_? devices are down.  So 
it succeeds in bringing the device down on the last one - just hangs the 
ifconfig.

The ifconfig I am using is part of the net-tools package from 

http://sites.inka.de/lina/linux/NetTools/

Version 1.60-r11

Stone

On Tuesday 12 July 2005 23:39, Johannes Stezenbach wrote:
> stonewall at inebraska.com wrote:
> > Johannes - here is the dmesg output before I did a Ctrl-C to break out of
> > the errant ifconfig:
> >
> > SysRq : Show State
> >
> >  ifconfig      S 00000008     0  8226   8216                     (NOTLB)
> > f4f0fdbc 00000082 f4f0fdac 00000008 00000001 f7f10000 00000208 c03d1c00
> >        00000000 00004240 00000000 00001fff 00004240 00000000 20a35fff
> > c2000520 00000000 000034eb e61ca35d 00000268 00000040 f7f1b020 f7f1b144
> > f8e76d9d Call Trace:
> >  [<c0370b28>] __down_interruptible+0xa8/0x128
> >  [<c0370bc2>] __down_failed_interruptible+0xa/0x10
> >  [<f8f1d801>] .text.lock.dvb_demux+0x1fb/0x31a [dvb_core]
> >  [<f8f1cfbb>] dmx_section_feed_release_filter+0xab/0xd0 [dvb_core]
> >  [<f8f23afa>] dvb_net_feed_stop+0x7a/0x1e0 [dvb_core]
> >  [<c0304821>] dev_close+0xb1/0xc0
> >  [<c0305f98>] dev_change_flags+0x58/0x130
> >  [<c0348cc4>] devinet_ioctl+0x254/0x5c0
> >  [<c034b194>] inet_ioctl+0x64/0xb0
> >  [<c02faf98>] sock_ioctl+0xd8/0x210
> >  [<c0171b31>] do_ioctl+0x81/0xa0
> >  [<c0171d02>] vfs_ioctl+0x62/0x1d0
> >  [<c0171eb1>] sys_ioctl+0x41/0x70
> >  [<c01033b9>] syscall_call+0x7/0xb
>
> Hm, now the question is who's holding the dvb_demux mutex.
>
> What other tasks are in SysRq-T output? If you grep for
> .text.lock.dvb_demux, does anything show up? Or at least
> any other process with "dvb_" or "dmx_" in the call trace?
>
> If not, I have no clue. I just looked through dvb_demux.c
> and I can't see any case where a function could exit without
> releasing the mutex. (Memory corruption could cause this
> kind of lockup, but that's more difficult to check.)
>
> Johannes




More information about the linux-dvb mailing list