Re: iwlagn interrupt problem

[Zdenek Kabelac <zkabelac@redhat.com>]

Dan Williams napsal(a):
> On Fri, 2008-12-12 at 12:31 +0100, Zdenek Kabelac wrote:
>> Alexander E. Patrakov napsal(a):
>>> 2008/12/12 Zdenek Kabelac <zkabelac@redhat.com>:
>>>> I assume that this backtrace I've got with 2.6.28-rc7 is the same issue ?
>>> No, this is different. My issue is with unacceptable latency when
>>> bringing the interface down. In your dmesg, the interface was never
>>> brought up or down.
>>>
>> Hmm - it was something like an hour after resume when the problem started,
>>
>> Here is timed part of the same backtrace:
>>
>> [65627.782972] wlan0: associate with AP 00:11:d8:da:65:40
>> [65627.785569] wlan0: RX ReassocResp from 00:11:d8:da:65:40 (capab=0x401 
>> status=0 aid=8)
>> [65627.785576] wlan0: associated
>> [69541.209873] wlan0: No ProbeResp from current AP 00:11:d8:da:65:40 - assume 
>> out of range
>> [69541.810104] iwl3945: Error sending REPLY_SCAN_CMD: time out after 500ms.
>> [69542.309861] iwl3945: Error sending REPLY_RXON_ASSOC: time out after 500ms.
>>
>>
>> The begging of my trace looked similar.
>>
>> Should I open new bugzilla separately ?
> 
> If you can reproduce this, it sounds more like a firmware problem, so
> Reinette is probably going to want you to enable firmware debugging in
> the driver.  Try removing the iwl3945 module with rmmod, then "modprobe
> iwl3945 debug=0x43fff", reproduce this command timeout, and attach the
> logs to whatever bug report you file.
> 

Unfortunately I do not have a reproducing case and it happened just once on 
Wednesday after resume.

But when (if) it will repeat I'll try to reinsert module with debug - I think 
without this my bug is not resolvable right ?

Zdenek