Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 2 Nov 1997 20:03:57 +0100 (MET)
From:      Guido van Rooij <guido@gvr.org>
To:        nate@mt.sri.com (Nate Williams)
Cc:        nate@mt.sri.com, freebsd-mobile@freebsd.org
Subject:   Re: card removal problems
Message-ID:  <199711021903.UAA05422@gvr.gvr.org>
In-Reply-To: <199711021712.KAA02976@rocky.mt.sri.com> from Nate Williams at "Nov 2, 97 10:12:39 am"

next in thread | previous in thread | raw e-mail | index | archive | help
Nate Williams wrote:
> > > > Just like with the PAO release I experience problems when removing a
> > > > card from a slot. Sometimes this results in a total lockup of the
> > > > system, while it keeps beeping.
> > > 
> > > There is a race where it's possible that the driver in in use when you
> > > remove it that is un-avoidable.  But, this seems to happen pretty
> > > rarely.
> > 
> > Actually, it seems that it almost always locks up for me once a driver
> > has ben allocated. I inserted that modem card I was talking about
> > after applying my patch (so pcic would use not irq 3). It was not being
> > used but still the machine locked up.
> 
> I'm confused.  You say it would lockup the machine *only* after a driver
> was allocated, or it didn't matter.  If it doesn't matter, then
> something is wrong with your PCIC controller, since all we do is apply
> power (or not) to your controller.

No, this is after I changed pccard.c to go downwards with irq's. So
now I can allocate a driver.

> 
> > I am now going to implement the power patches to see if that would
> > solve my problem.
> 
> I'd like to think there is a better solution that doesn't require user
> interaction.
> 
> So, if there's no entry for the card in question in /etc/pccard.conf,
> after inserting it, waiting for a driver to be NOT found, and then
> removing the card your box will lockup?  If so, then the power patches

No. It does certianly not lock up then/

> won't make any difference unless they 'power' down the card differently
> than I do in removal, which means that I can power the card down the
> same way and you won't need the 'power' patches.
> 

Is there anything I can do, except low level kernel debugging to
daignose this?

-Guido



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199711021903.UAA05422>