On Mon, 2002-07-08 at 20:03, Todd Lyons wrote:

> See that usage count of "1" for ide-cs?  Until you get that to go to
> "0", you are never going to be able to unload that module. 

yup. did some module research here. i think it has to do with the
ide-floppy driver being counted, but cannot be removed due to the
ide-floppy being compiled into the kernel.  ide-cs.c is probabely
lacking something here.

 Sadly, I
> don't know of a method other than a reboot to clear that type of error.

ditto.

> The key here is what sequence of events occurred to make it such that it
> has a usage count of 1 but nothing claiming to be depending on it.  Even
> stranger is that the usage count of ds *did* get decremented to zero,
> but it still shows that it's dependent on ide-cs.  I'd suspect something
> funky happened with supermount, but that's purely a guess.  Trying to
> reproduce the problem will go a long way towards identifing the root
> cause.

nah. i manually mounted the /dev/hdc1 /mnt/disk. ide-floppy, nor any
other related modules to the pcmcia bus are mounted.  i believe that
your first guess what right on the money.

i'll have another look over ide-cs.c to see what's up.

oh, this brings back memories with the ehci-hcd.o module as well.
similar scenario that it was being unloaded automatically via the pcmcia
script but it *could be* manually. turned-out, that the ehci-hcd code is
correct but it resorted to a *known bug* in pcmcia/cardbus code.

eh, just purely guessing, but this may all fall back to the
pcmcia/cardbus developement. 

ok. close thread. i'm started to ramble and speculate again ;-)

-- 
Roger
-----
Verify my pgp/gnupg signature on my HomePage:
http://www.alltel.net/~rogerx/about/index.html

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to