WLAN regression before this morning

WLAN regression before this morning

[barry bouwsma]

Servus,

Last successfully built and used kernel was 2.6.26-rc8, about
the start of 29.Jun UTC or so.

Present build of GIT 2.6.26 during the merge window, updated
just before 06h UTC, today, 25.Jul.

With today's kernel, I am not even able to associate with a
WLAN less than two metres away from my ZD1211rw type USB
adapter, to say nothing of more distant WLANs with which I
could authenticate with the earlier code without problems --
I don't even come near the associate stage.

The networks all show up as expected in `iwlist wlan1 scan'
with decent strengths; while all I get from dmesg is
[ 1016.829827] wlan1: authenticate with AP 00:1X:YZ:AB:23:45
[ 1017.032769] wlan1: authenticate with AP 00:1X:YZ:AB:23:45
[ 1017.234213] wlan1: authenticate with AP 00:1X:YZ:AB:23:45
[ 1017.435402] wlan1: authentication with AP 00:1X:YZ:AB:23:45 timed out


If I've hosed myself with a new .config option, I'll be kicking
meself where it hurts most.  Else, back to -rc8 to both send
this message and to regain 'net access to do other needed so-
called `work', so sorry I haven't narrowed this down further...

And I'll feel me a fool if the failure is because I went from
a different (BSD) OS into 2.6.26-merge, while going from that
into -rc8 I promptly associate as desired.


thanks
barry bouwsma


      

Re: WLAN regression before this morning

[Larry Finger]

barry bouwsma wrote:
> Servus,
> 
> Last successfully built and used kernel was 2.6.26-rc8, about
> the start of 29.Jun UTC or so.
> 
> Present build of GIT 2.6.26 during the merge window, updated
> just before 06h UTC, today, 25.Jul.
> 
> With today's kernel, I am not even able to associate with a
> WLAN less than two metres away from my ZD1211rw type USB
> adapter, to say nothing of more distant WLANs with which I
> could authenticate with the earlier code without problems --
> I don't even come near the associate stage.
> 
> The networks all show up as expected in `iwlist wlan1 scan'
> with decent strengths; while all I get from dmesg is
> [ 1016.829827] wlan1: authenticate with AP 00:1X:YZ:AB:23:45
> [ 1017.032769] wlan1: authenticate with AP 00:1X:YZ:AB:23:45
> [ 1017.234213] wlan1: authenticate with AP 00:1X:YZ:AB:23:45
> [ 1017.435402] wlan1: authentication with AP 00:1X:YZ:AB:23:45 timed out
> 
> 
> If I've hosed myself with a new .config option, I'll be kicking
> meself where it hurts most.  Else, back to -rc8 to both send
> this message and to regain 'net access to do other needed so-
> called `work', so sorry I haven't narrowed this down further...
> 
> And I'll feel me a fool if the failure is because I went from
> a different (BSD) OS into 2.6.26-merge, while going from that
> into -rc8 I promptly associate as desired.

I too have a problem with wireless in the current 2.6.26-gitX kernels. If you 
have done the git pull, you can try bisecting to find the problem. Use 'git 
bisect good v2.6.26' to mark a known good point. If you don't wish to take the 
time bisecting, then issue a 'git checkout v2.6.26'. That should give you a 
kernel that works.

Larry

Re: WLAN regression before this morning

[Stefanik Gábor]

On Fri, Jul 25, 2008 at 5:57 PM, Larry Finger <Larry.Finger@lwfinger.net> wrote:
> I too have a problem with wireless in the current 2.6.26-gitX kernels. If
> you have done the git pull, you can try bisecting to find the problem. Use
> 'git bisect good v2.6.26' to mark a known good point. If you don't wish to
> take the time bisecting, then issue a 'git checkout v2.6.26'. That should
> give you a kernel that works.
>
> Larry
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

I am also seeing problems - although I can connect to networks (with
zd1211rw, ath5k & sometimes with b43), it disconnects after about 2
minutes with a sudden "Removed STA"/"Destroyed STA" pair (no "assume
out of range" or anything). However, if I use wpa_supplicant to
connect to wepped networks (rather than just "iwconfig wlan0 essid
<ESSID> key s:<ASCII key>"), the problem goes away, and the connection
is stable for hours.

-- 
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)