-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi all,
my raw sensation is that ipw2100 is more unstable in 2.6.33 than in 2.6.32.
Today I got an ipw2100 crash without doing anything illegal; see attached dmesg.
Also I find these messages in /var/log/messages.log:
- --->8-----
Feb 24 21:53:20 fester kernel: ipw2100: exit - failed to send CARD_DISABLE command
Feb 24 21:53:20 fester kernel: ipw2100: exit - failed to send CARD_DISABLE command
Feb 24 21:54:16 fester kernel: ipw2100: Fatal interrupt. Scheduling firmware restart.
Feb 24 21:56:47 fester kernel: ipw2100: exit - failed to send CARD_DISABLE command
Feb 24 21:58:51 fester kernel: ipw2100: exit - failed to send CARD_DISABLE command
Feb 24 22:07:11 fester kernel: ipw2100: exit - failed to send CARD_DISABLE command
Feb 24 22:09:57 fester kernel: ipw2100: Fatal interrupt. Scheduling firmware restart.
Feb 24 22:11:09 fester acpid: client 3208[0:100] has disconnected
Feb 24 22:11:09 fester acpid: client connected from 3208[0:100]
Feb 24 22:11:09 fester acpid: 1 client rule loaded
Feb 24 22:28:49 fester kernel: ipw2100: Fatal interrupt. Scheduling firmware restart.
- --->8-----
This is weird; something is clearly not working OK. However there is to say that the above messages are present with stable 2.6.32 as well with latest 2.6.33, but I got crash only with the linus git version.
Note: my kernel sources are patched with the BFS scheduler, I suppose that does not invalidate the results of my findings regarding this ipw2100 crash
Best regards,
- --
Daniele C.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEcBAEBAgAGBQJLhZuyAAoJEON28As9epPAyLUH/RDpleeV0KCpYbRxRAYX1C33
OUz063Dgi5Llb4uW5tqtt31oerrKqV9o/oiJ3yQ7aKxVioJ0cnuUDAM+EOgtBoql
fVTJJxG1mY70DJwOuYkhXsxmz/Eln2Qk/lEept5/lOm0SNgQHlcbqWOV/gCNbBMn
hnY5TnQy9vFG+fD2Ui07z5Zl/r6gtXDb+WQk0vyjYY35v/eWPut9UYhT4oE8avxp
/H+cp1KA/JN6jzWJ8m2/rxLSmQUrJgHfUs7oTZVKfAWXuS8Q3AW4C+WHolOMXMkp
do2cv0IX3tOOUHRf5xwk4naNoIBfSDxbLRTrLlfcDvYMxIOtUmZj8z/M8Qia//E=
=e3ql
-----END PGP SIGNATURE-----
On Thu, 2010-02-25 at 05:35 +0800, Daniele C. wrote:
> - --->8-----
> Feb 24 21:53:20 fester kernel: ipw2100: exit - failed to send
> CARD_DISABLE command
> Feb 24 21:53:20 fester kernel: ipw2100: exit - failed to send
> CARD_DISABLE command
> Feb 24 21:54:16 fester kernel: ipw2100: Fatal interrupt. Scheduling
> firmware restart.
> Feb 24 21:56:47 fester kernel: ipw2100: exit - failed to send
> CARD_DISABLE command
> Feb 24 21:58:51 fester kernel: ipw2100: exit - failed to send
> CARD_DISABLE command
> Feb 24 22:07:11 fester kernel: ipw2100: exit - failed to send
> CARD_DISABLE command
> Feb 24 22:09:57 fester kernel: ipw2100: Fatal interrupt. Scheduling
> firmware restart.
> Feb 24 22:11:09 fester acpid: client 3208[0:100] has disconnected
> Feb 24 22:11:09 fester acpid: client connected from 3208[0:100]
> Feb 24 22:11:09 fester acpid: 1 client rule loaded
> Feb 24 22:28:49 fester kernel: ipw2100: Fatal interrupt. Scheduling
> firmware restart.
> - --->8-----
This is a firmware error. Does the driver recover itself later?
Thanks,
-yi
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
- -------- Original Message --------
Subject: Re: ipw2100 crash with latest git
From: Zhu Yi <[email protected]>
To: legolas558 <[email protected]>
Date: Thu Feb 25 2010 01:52:33 GMT+0100 (CET)
> On Thu, 2010-02-25 at 05:35 +0800, Daniele C. wrote:
>> - --->8-----
>> Feb 24 21:53:20 fester kernel: ipw2100: exit - failed to send
>> CARD_DISABLE command
>> Feb 24 21:53:20 fester kernel: ipw2100: exit - failed to send
>> CARD_DISABLE command
>> Feb 24 21:54:16 fester kernel: ipw2100: Fatal interrupt. Scheduling
>> firmware restart.
>> Feb 24 21:56:47 fester kernel: ipw2100: exit - failed to send
>> CARD_DISABLE command
>> Feb 24 21:58:51 fester kernel: ipw2100: exit - failed to send
>> CARD_DISABLE command
>> Feb 24 22:07:11 fester kernel: ipw2100: exit - failed to send
>> CARD_DISABLE command
>> Feb 24 22:09:57 fester kernel: ipw2100: Fatal interrupt. Scheduling
>> firmware restart.
>> Feb 24 22:11:09 fester acpid: client 3208[0:100] has disconnected
>> Feb 24 22:11:09 fester acpid: client connected from 3208[0:100]
>> Feb 24 22:11:09 fester acpid: 1 client rule loaded
>> Feb 24 22:28:49 fester kernel: ipw2100: Fatal interrupt. Scheduling
>> firmware restart.
>> - --->8-----
>
> This is a firmware error. Does the driver recover itself later?
>
I cannot scientifically say yes, but by memory I'd say that it does after a full unload/reload of ipw2100 module. I will of course test this if it happens again.
Meanwhile I got the crash also with 2.6.32 (attached), so it's definitively a not yet fixed bug and not something due to my custom patches.
Triggering conditions are basically "hammering" the device to change APs without any ipw2100 module reload in the process.
P.S. I think the acpid event can be ignored because I disconnected an USB device at that time (totally unrelated).
Thanks,
- --
Daniele C.
> Thanks,
> -yi
>
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEcBAEBAgAGBQJLheUXAAoJEON28As9epPAZfUH/36MTt6fp5xQhqmRbs6ODjeO
RQrl98cHP/c21Xv9E8P29r++gVkEHXtBiwszENFZ3R0zpVcadB0Z6RXfYckAuahr
bCEwk6izckcTnGPSOXkqhunyo9IZD5VHJThRI3Q8Jbl5fo3/k1+tY25okMDs+qI1
X94w9CdcxltS06iiIiG2VAwrqxoAVePuA2u9mXx5fEoHwZLSyEQ/uEyju90peEut
l5wYuRguyfR/sHhv2bDAtpvqt60+mr1vV/XTufjKymg8RAvjLIiDAiJhCQ6Ldk4U
Ls+HGF3cxxKq6kSi3qMoy7HDVOzKHIwKZB+MFMHC6RPuDHcCk8uonYWTgoEcqy0=
=te4d
-----END PGP SIGNATURE-----