[OLSR-users] kernel oops: NULL pointer deference on a wireless repeater

dennis khoo (spam-protected)
Fri Jun 3 05:38:40 CEST 2005


Hi all,

I finally managed to build ksymoops and piped the
kernel panic capture to it.  It seems like the null
dereference is caused by a mov instruction in the
wireless card's driver, unfortunately I can't look at
the source since it's in the binary hardware
abstraction layer of the driver.  

Could I have gotten my ksymoops trace wrong since the
kernel panic message seems to indicate the oops is
cause by the "Process olsrd (pid: 8947, stackpage..."
and not a driver?  I know I supplied it the right
System.map, previously saved /proc/ksyms and
/proc/modules before the oops.  I'm also supplying it
with --object=/sbin/olsrd but I'm not sure since the
man page seems to suggest kernel modules with -o...  
It's my first time with ksymoops so any pointers is
appreciated. 


I'm also testing the same setup and config with a
different wireless card/driver and so far so good.

thanks again
dennis



--- dennis khoo <(spam-protected)> wrote:

> 
> Hi Andreas and Bruno,
> 
> I'm using a WRAP 2C (from pcengines, a x86 geode
> processor).  I'm running linux kernel 2.4.26 (built
> from leaf-bering distribution).  I was using 0.4.8
> up
> until a few days ago and I don't think I saw this
> particular problem but I am not 100% sure.  I'll
> recheck if I have time.
> 
> I'll try to find more info through ksymoops like
> Bruno
> suggested and report back if I find anything
> interesting.  I guess it's finally time to learn how
> to build/use ksymoops:) 
> 
> 
> thanks
> dennis
> 
> 
> --- Andreas Tønnesen <(spam-protected)> wrote:
> 
> > Hi dennis,
> > 
> > Wow, I've never seen olsrd do that before...
> > The first thing to check/rule out here is HW
> > problems. What
> > kind of hardware is your second node running? What
> > kernel/distro
> > is it running?
> > 
> > - Andreas
> > 
> > 
> > dennis khoo wrote:
> > > Hi all,
> > > 
> > > I seem to run into some problem with 0.4.9, no
> > plugins
> > > were built or used.  The setup is 2 nodes. The
> > first
> > > node is a wireless AP with an ethernet
> connection
> > to
> > > the internet.  The olsrd.conf is basically
> > > olsrd.conf.default.lq with
> > > 
> > > Hna4
> > > {
> > >    0.0.0.0      0.0.0.0
> > > }
> > > 
> > > and the wireless interface block enabled.
> > > 
> > > The second node has a client radio (associated
> > with
> > > the first node) and another radio running as
> > master
> > > mode(but no one associates with it yet). So the
> > 2nd
> > > node basically functions as a repeater.
> > > 
> > > The problem is after starting olsrd for some
> > > time(ranging from 10 mins to 3 hours), I will
> > > encounter a null pointer dereference on the 2nd
> > node,
> > > always at the same EIP.  I don't get a core dump
> > since
> > > the kernel panics and restarts but here's what I
> > see:
> > > 
> > > # Unable to handle kernel NULL pointer
> dereference
> > at
> > > virtual address 00
> > > 000034
> > > *pde = 00000000
> > > Oops: 0000
> > > CPU:    0
> > > EIP:    0010:[<c481fd2c>]    Not tainted
> > > EFLAGS: 00010246
> > > eax: 00000000   ebx: c11dd810   ecx: c11dd956  
> > edx:
> > > c11dd800
> > > esi: c3c87000   edi: c11dd834   ebp: c3aafb6c  
> > esp:
> > > c3aafb54
> > > ds: 0018   es: 0018   ss: 0018
> > > Process olsrd (pid: 23626, stackpage=c3aaf000)
> > > Stack: c3c9a000 c11dde28 00000000 c11dd810
> > c11dd810
> > > 00000000 c3aafbdc c481f0b4
> > >        c3c87000 c11dd810 00000000 00000000
> > c3c50ae4
> > > c3c70162 c3aafbdc c3aafba4
> > >        00000000 0000092a 00000001 00000000
> > 00000000
> > > c3aafc7c c3c87820 c11dd810
> > > Call Trace:    [<c481f0b4>] [<c48369ec>]
> > [<c48449d4>]
> > > [<c48206f6>] [<c480a127>]
> > >   [<c480b5e3>] [<c480fffe>] [<c0233ab4>]
> > [<c480ac6c>]
> > > [<c0240cce>] [<c023727e>]
> > >   [<c024e2b3>] [<c024e220>] [<c023e840>]
> > [<c024cc77>]
> > > [<c024e220>] [<c024e200>]
> > >   [<c024e20d>] [<c023e840>] [<c024dae5>]
> > [<c024e200>]
> > > [<c0266f83>] [<c0266b00>]
> > >   [<c026d039>] [<c0230795>] [<c0231570>]
> > [<c480b94a>]
> > > [<c480765b>] [<c4807650>]
> > >   [<c0170e9c>] [<c01738d6>] [<c0170dca>]
> > [<c0170d01>]
> > > [<c0231dab>] [<c0160903>]
> > >   [<c016002b>]
> > > 
> > > Code: 8b 48 34 85 c9 8d 96 9c 01 00 00 75 23 83
> ec
> > 0c
> > > 6a 00 68 18
> > >  <0>Kernel panic: Aiee, killing interrupt
> handler!
> > > In interrupt handler - not syncing
> > > 
> > > 
> > > Can anyone recommend a quick way to find out
> where
> > in
> > > the source the problem is coming from. Any help
> is
> > > deeply appreciated.
> > > 
> > > dennis
> > > 
> > > 


		
__________________________________ 
Discover Yahoo! 
Find restaurants, movies, travel and more fun for the weekend. Check it out! 
http://discover.yahoo.com/weekend.html 




More information about the Olsr-users mailing list