On Thu, 2007-08-16 at 16:25 +0100, Adam Baker wrote: > I'm starting to think that scanning from the remote device always causes a crash > if done just after the iwconfig essid and never does if left for a while first > (~ 1 minute). Scanning on the local machine I've usually done when the scan and > subsequent connect from the remote machine has succeeded. I'm not yet sure if > that is a necessary condition for the crash. To be fair, I managed to reproduce it once with b43 but not again. Maybe this weird timing behaviour was the cause of that. > I've added some printk's as you suggested, the log is attached. I > haven't yet > managed to capture this on the serial port so I can't yet guarantee > that there > aren't any calls that didn't get logged just as it crashed - I'll keep > trying. Ok. I'll look through. > Aug 16 15:53:34 gromit kernel: [ 633.972956] wmaster0: Entering ieee80211_tx_h_multicast_ps_buf(), hw.flags 1746 > Aug 16 15:53:34 gromit kernel: [ 633.972960] Leaving ieee80211_tx_h_multicast_ps_buf Hm. I thought it crashed in there. Could you put printk's into the two functions called from this, and then switch to a console and tell me which one shows up last? Elevate the printk priority to something like KERN_ERR or change your console to show all messages. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
