On Wednesday 03 February 2010 20:18:52 Larry Finger wrote:
> On 02/03/2010 11:33 AM, Yuval Hager wrote:
> > Thank you for your efforts, I highly appreciate it.
> >
> > I pulled latest wireless-testing (git describe says v2.6.33-rc6-47373-
> > g3b4936f), and applied these two patches.
> >
> > This time there is another type of disconnection (wlan0: deauthenticating
> > from 00:22:3f:18:89:5e by local choice (reason=3)), in addition to the
> > "no probe response" I was getting previously.
> > The big change in these patches is that the system is able to get over it
> > and reconnect after the disconnection.
> 
> The short-slot patch will only affect performance; however, the patch that
> restarts the queues could certainly be the one that allows the system to
> reconnect and continue. The surprising part is that I sent it earlier and
>  was told that it did not help.
>

You are right. I retested just with the queue wake patch, and it does seem to 
reconnect as well.
I am not sure why it didn't work before - I probably applied it to the first 
version that broke (after bisecting) instead of wireless-testing, or I simply 
didn't wait enough time. Sorry about that - bisecting, compiling, rebooting 
etc. can get really confusing at times (especially late at night...).

Please let me know if you have any othes patch ideas for me to try.
 
> Does your AP have the latest firmware? At
> http://forums.wi-fiplanet.com/showthread.php?t=6536, reason 3 is listed as
>  "The access point went offline, deauthenticating the client." Other
>  explanations are likely possible. I have not checked the IEEE802.11
>  documents.
> 

I just upgraded to the latest firmware, but there is no change. I am not 
having any issues with this router while running 2.6.29 on this machine, or 
2.6.32 on another machine (using rt2x500pci). The router is Netgear WGR614V9, 
and the firmware is V1.2.24_37.0.35.
Also note that this "reason 3" only occures after the first association - 
after that it just continues with the "no probe response" I got used to.

--yuval
_______________________________________________
Bcm43xx-dev mailing list
Bcm43xx-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Reply via email to