[lwlan-user] linux-wlan-ng 0.2.1-pre10

[David Cook]

On Mon, 2003-07-14 at 11:11, Solomon Peachy wrote: 
> ftp://ftp.linux-wlan.org/pub/linux-wlan-ng/linux-wlan-ng-0.2.1-pre10.tar.gz
> 
> It hasn't changed in almost two weeks, so I'd better just get this out 
> the door all the same.  Been busy.  :)
> 
> Other than USB changes... make various wakeup conditions volatile, as 
> apparently some arcitechures/compilers have register allocation issues 
> with shared data members. 
> 
> I want to get USB working on 2.5/2.6 before 0.2.1-final.  I don't plan 
> to fix the USB build for 2.2.x at all though -- it simply isn't worth 
> it, given the sorry state of 2.2.x USB -- and indeed, I may just rip the 
> non-working 2.2.x stuff out of the USB code altogether.
> 
> Anyway, enjoy!
> 
> -pre10
>   - Make the wakeup flag in the usbctlx structure volatile, as well as 
>     other wait condition variables.
>   - Make the rx_urb buffer dynamically allocated so we can re-post the 
>     URB as soon as possible.
>   - Get rid of completion.h
>   - More wlan_compat updates.
>   - Disable TXEX reporting on USB as well as TXOK.  should improve
>     stability a tad.
>   - Further 2.5 USB work.  Not quite there yet.
> 
> 
>  - Pizza



Solomon -

	Figured I would check out your new 'pre10' changes.  They
compile/build/install just fine.

	But, when I try to connect using this new one, the connection
does NOT happen.  Ditto when I try unplugging/replugging to exercise
my hotplug support (which worked fine under 'pre9').  When I check
'/var/log/messages', I see some failures that weren't there under
'pre9':


Jul 14 14:14:23 wireless-127 kernel: prism2sta_getcardinfo: Failed to
retrieve NICIDENTITY
Jul 14 14:14:23 wireless-127 kernel: prism2sta_getcardinfo: Failed,
result=-5
Jul 14 14:14:23 wireless-127 kernel: prism2sta_ifstate:
prism2sta_getcardinfo() failed,result=-5
Jul 14 14:14:23 wireless-127 /etc/hotplug/wlan.agent[1511]: Failed to
enable the device, resultcode= implementation_failure .


Is this because of something I'm forgetting to do, or might this be
unwanted regression because of driver changes?

Cheers...

		Dave