[Click] Re: e1000 related kernel panic

[Barry Gowan]

I believe the bug I was seeing was this one:
http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=75680
I implemented the workaround suggested by Jim Garlick
and it stops the crash. 

Had no-one else seen this crash when pulling a cable
on an e1000 controlled port?
Was it caused by the click polling changes to the
driver?

 --- Barry Gowan <barry_gowan at yahoo.co.uk> wrote: > 
> Hi
> I have click working with intel-based 2 port fiber
> GigE cards. 
> I get the following panic when the cables are
> unplugged during a test where, say, 30k packets per
> second are getting bounced between two boxes. 
> 
> >>EIP; e08bec20 <[e1000]e1000_smbus_lock+b10/9e9b>  
> <=====
> Trace; e08bb293 <[e1000].text.start+233/30b0>
> Trace; c0108dfd <free_irq+8d/b0>
> Trace; e08bb205 <[e1000].text.start+1a5/30b0>
> Trace; c011d546 <del_timer_sync+406/e00>
> Trace; e08bcf39 <[e1000].text.start+1ed9/30b0>
> Trace; c0115bbb <panic+74b/7d0>
> Trace; c01bb8c0 <qdisc_restart+170/260>
> Trace; c01bb92a <qdisc_restart+1da/260>
> Trace; c011a21b <__tasklet_hi_schedule+13b/170>
> Trace; c0119fbb <do_softirq+7b/e0>
> Trace; c0108cbd <enable_irq+18d/1a0>
> Trace; c010b088 <disable_irq_nosync+1c48/3010>
> Code;  e08bec20 <[e1000]e1000_smbus_lock+b10/9e9b>
> 00000000 <_EIP>:
> Code;  e08bec20 <[e1000]e1000_smbus_lock+b10/9e9b>  
> <=====
>    0:   0f 0b                     ud2a      <=====
> Code;  e08bec22 <[e1000]e1000_smbus_lock+b12/9e9b>
>    2:   eb 3d                     jmp    41
> <_EIP+0x41> e08bec61 
> <[e1000]e1000_smbus_lock+b51/9e9b>
> Code;  e08bec24 <[e1000]e1000_smbus_lock+b14/9e9b>
>    4:   8b 06                     mov    (%esi),%eax
> Code;  e08bec26 <[e1000]e1000_smbus_lock+b16/9e9b>
>    6:   8b 40 08                  mov   
> 0x8(%eax),%eax
> Code;  e08bec29 <[e1000]e1000_smbus_lock+b19/9e9b>
>    9:   c7 86 98 00 00 00 00      movl  
> $0x0,0x98(%esi)
> Code;  e08bec30 <[e1000]e1000_smbus_lock+b20/9e9b>
>   10:   00 00 00
> Code;  e08bec33 <[e1000]e1000_smbus_lock+b23/9e9b>
>   13:   b9 00 00 00 00            mov    $0x0,%ecx
> 
> One side of the test had just the following loopback
> configuration:
> 
> p1:: PollDevice(eth1)-> c1:: Counter-> q1::
> SimpleQueue -> td1:: ToDevice(eth2)
> p2:: PollDevice(eth2)-> c2:: Counter-> q2::
> SimpleQueue -> td2:: ToDevice(eth1)
> 
> - and I got a very similar panic when the cable was
> pulled.  
> Can anyone tell me what might be happening?
> 
> cheers, 
> BG
> 
> 
>
________________________________________________________________________
> Want to chat instantly with your online friends? 
> Get the FREE Yahoo!
> Messenger http://mail.messenger.yahoo.co.uk
>  

________________________________________________________________________
Want to chat instantly with your online friends?  Get the FREE Yahoo!
Messenger http://mail.messenger.yahoo.co.uk