In message <199809062040.WAA22539@animal.cs.chalmers.se>  Lennart Augustsson wrote:
> 
> > I hate deamon processes for this and that. 
> Me too, but as long as we don't have kernel threads it is the only
> option, IMHO.
> 
> > With 0.5 seconds timeout driven state machine is a better solution.
> It's not a 0.5 second timeout.  It always waits for 0.5 seconds, because
> the PCMCIA specs says you have to.  I don't find a busy-wait for 0.5 seconds
> in the kernel acceptable.  And even less so when it is called from an
> interrupt.

I know it no timeout yet. I would fix the delay() using timeout instead of
a daemon processes thats all. 

> 
> > The problems with daemons is that they may get killed ...
> Yes, it can happen, but I'd rather take that risk.
> You could imagine creating a system process like the swapper et al,
> but I guess it could die too.

Did you ever workend in support?? Anything that the user can mess up is
bad thing (and they will mess up every single bit ...). 

> 
> > Whats the problem?
> "pcic_wait_ready never happened"
> 
> > If the Card doesn't get ready try increasing the timeouts,
> Tried that.

Have you increased the delays for reset, the one after the
power-enable and  the one after the powerdisable  too? 
I think that the problem is somewhere in that aerea as it sometimes work.
(Yes more long delays ....)

> 
> > if I remember correctly you've more than 15 MB and AFAIK the Card does some
> > kind of inventory during powerup.
> It happens both with the 8MByte and 40MByte cards.

I have 2MB and 15MB card and both work fine on  a Libretto and worked in
ISA/PCMCIA adpater on compaq DeskPro (but I had to hack a the wd-driver
because of stupid hardware).

Stefan

> 
>    -- Lennart

--
Stefan Grefen                                Tandem Computers Europe Inc.
grefen@hprc.tandem.com                       High Performance Research Center
 --- Hacking's just another word for nothing left to kludge. ---