Skip to contentSkip to conversation

Cisco Aironet 340 oops with 2.4.20

DISCUSSION Thomas Hood 4 participants 5 messages Mar 15, 2003 – Mar 18, 2003

lore
Share & tools
lore
#0 lore raw
Quoted 3 lines · Javier Achirica

I have updated the CVS (airo-linux.sf.net) with a version that correctly implementes locking (there was a bug there). Please test it and tell me if it still panics.

The 2.4.19 - 2.4.20 airo driver was giving me the infamous skb_free() related oops every few days, usually after I downloaded something big.

2.4.21-pre5 with the airo driver from airo-linux.sf.net has been running perfectly for four days now. Looks good.

Signature · 7 lines
Thomas Hood <[email protected]>

__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com
#1 lore raw

I've got a caveat to my previously posted 'works for me'. Before I'd notice that my net connection was no longer moving and then see the oops
in my log if I looked. With the cvs driver I still tend to see that my
net connection dies sometimes, but no oops. I've found hotplugging the pcmcia card brings it back to life.

It seems like I hit this with either driver primarily if I've put my laptop to sleep earlier in its current uptime. Is there possibly a deadlock somewhere around the previously oopsing code and/or the airo_cs interaction with apm?

t.

Signature · 4 lines
*********************************************************
*  tpepper@vato dot org             * Venimus, Vidimus, *
*  http://www.vato.org/~tpepper     * Dolavimus         *
*********************************************************
#2 lore raw

I'm aware of this "lock up" problem. I was hoping it was a locking issue and the last changes would solve it. Looks like not. The biggest problem I'm having with that issue is that I'm not able to reproduce it. Any suggestion?

Javier Achirica

Quoted 20 lines · Tim Pepper

I've got a caveat to my previously posted 'works for me'. Before I'd notice that my net connection was no longer moving and then see the oops
in my log if I looked. With the cvs driver I still tend to see that my
net connection dies sometimes, but no oops. I've found hotplugging the pcmcia card brings it back to life.

It seems like I hit this with either driver primarily if I've put my laptop to sleep earlier in its current uptime. Is there possibly a deadlock somewhere around the previously oopsing code and/or the airo_cs interaction with apm?

t.
--
*********************************************************

*  tpepper@vato dot org             * Venimus, Vidimus, *
*  http://www.vato.org/~tpepper     * Dolavimus         *

*********************************************************

#3 lore raw

Hi Javier,

next oops happened today with the latest cvs version

Device: wifi0

Mar 17 15:15:15 vaio kernel: Short packet 0
Mar 17 15:15:15 vaio kernel: Warning: kfree_skb passed an skb still on a list (from c0121fca).
Mar 17 15:15:15 vaio kernel: kernel BUG at skbuff.c:315!
Mar 17 15:15:15 vaio kernel: invalid operand: 0000
Mar 17 15:15:15 vaio kernel: CPU:    0
Mar 17 15:15:15 vaio kernel: EIP:    0010:[__kfree_skb+324/352]    Not tainted
Mar 17 15:15:15 vaio kernel: EIP:    0010:[<c026cd54>]    Not tainted
Mar 17 15:15:15 vaio kernel: EFLAGS: 00010286
Mar 17 15:15:15 vaio kernel: eax: 00000045   ebx: c3bcb9e0   ecx: cb392000   edx: cb393f7c
Mar 17 15:15:15 vaio kernel: esi: c1339f84   edi: 00000000   ebp: c1338000   esp: c1339f6c
Mar 17 15:15:15 vaio kernel: ds: 0018   es: 0018   ss: 0018
Mar 17 15:15:15 vaio kernel: Process keventd (pid: 2, stackpage=c1339000)
Mar 17 15:15:15 vaio kernel: Stack: c0314840 c0121fca 00000000 c1339f84 c0121fca c3bcb9e0 cd0942f8 cd0942f8
Mar 17 15:15:15 vaio kernel:        00000000 00000000 c012ac83 c032abd0 c1339fb0 00000000 c1338560 c1338570
Mar 17 15:15:15 vaio kernel:        c1338000 00000001 00000000 cffe5f90 00010000 00000000 00000700 c012ab50
Mar 17 15:15:15 vaio kernel: Call Trace:    [__run_task_queue+90/112] [__run_task_queue+90/112] [context_thread+307/448] [c
ontext_thread+0/448] [rest_init+0/64]
Mar 17 15:15:15 vaio kernel: Call Trace:    [<c0121fca>] [<c0121fca>] [<c012ac83>] [<c012ab50>] [<c0105000>]
Mar 17 15:15:15 vaio kernel:   [kernel_thread+46/64] [context_thread+0/448]
Mar 17 15:15:15 vaio kernel:   [<c010749e>] [<c012ab50>]
Mar 17 15:15:15 vaio kernel:
Mar 17 15:15:15 vaio kernel: Code: 0f 0b 3b 01 cf 2d 31 c0 8b 5c 24 14 e9 be fe ff ff 90 8d 76
Mar 17 15:16:34 vaio kernel:  <3>airo: BAP setup error too many retries
Mar 17 15:16:59 vaio kernel: mtrr: no MTRR for fd000000,400000 found
Mar 17 15:17:10 vaio kernel: SysRq : HELP : loglevel0-8 reBoot tErm kIll saK showMem Off showPc unRaw Sync showTasks Unmoun

t

Quoted 7 lines · On Mon, 17 Mar 2003 02:28:58 +0100 (MET) Javier Achirica…

I'm aware of this "lock up" problem. I was hoping it was a locking issue and the last changes would solve it. Looks like not. The biggest problem I'm having with that issue is that I'm not able to reproduce it. Any suggestion?

Javier Achirica

#4 lore raw
In reply to Javier Achirica

It seems like the following does it:

- run for a while
- suspend
- wake..run for a while
- maybe suspend and wake a few more times?
- cause a heavish burst of net traffic

If it would matter I'm running wep. It may happen more often when I've got an ipsec tunnel running over the connection as well (tunnel never left open during suspend nor ipsec module left loaded). When it's happened cpu utilisation was pretty low.

t.

Signature · 4 lines
*********************************************************
*  tpepper@vato dot org             * Venimus, Vidimus, *
*  http://www.vato.org/~tpepper     * Dolavimus         *
*********************************************************