Re: mis-identified cisco aironet pccard (and Re: hang with pcmcia wlan card)

[Jean Tourrilhes]

On Mon, Jul 14, 2003 at 05:52:43PM +0100, Russell King wrote:
> On Mon, Jul 14, 2003 at 09:48:52AM -0700, Jean Tourrilhes wrote:
> > On Sun, Jul 13, 2003 at 10:57:58PM +1000, Sven Dowideit wrote:
> > > Hi,
> > > i just noticed with 2.5.75 that if I boot with the cisco airo 340 wifi
> > > card in, this kernel thinks it is a memory card. when i remove it and
> > > re-insert it after boot, it then works .... see the following log :) I
> > > am running debian unstable, on an ibm t21 pIII-850 notebook
> > > 
> > > the cisco card works at boot time using 2.5.70.
> > > 
> > > as for the two patches for the pcmcia hang, this time i am running the
> > > one Russell posted, but the result is the same for the other.
> > > 
> > > as i shutdown i get a number of kernel stack dumps related to airo_stat,
> > > but the machine reboots before i can do anything about them..(what do i
> > > need to log them?)
> > > 
> > > if i replace the cisco card with a dlink orinoco card, it get recognised
> > > correctly at boot. 
> > > 
> > > to make this story more interesting, i put the thinkpad into the docking
> > > station and the cisco card into the dock's pccard (something that has
> > > locked up 2.5 every time that i tried it), and the card is recogised
> > > correctly at boot, and runs fine (there was a kernel stack dump during
> > > boot - *what do i need to do to get them logged?*)
> > > 
> > > thanks for the great work!
> > > 
> > > sven
> > 
> > 	I've seen this bug come and go in the 2.5.X serie. I believe
> > this is because of the various work happening in the Pcmcia
> > layer. Please contact Dominik Brodowski <[e-mail]>.
> 
> I think this may be my fault, but I haven't head a clear bug report for
> it yet; I've only seen vague references to something going wrong with
> aero cards on lkml yesterday, referring to what seemed to be a non-
> existent thread.

	Personally, I had this erratic problem around 2.5.6X, but it
was erratic, so at the next boot it disapeared. Since 2.5.72, I've not
seen it, so I assumed it was fixed.
	You probably know already the memory_cs is bound to any card
which doesn't have any CIS.

> Russell King

	Jean