Subject: Re: PCMCIA modems and ppp difficulties
To: Jim Rees <rees@umich.edu>
From: Stefan Grefen <grefen@hprc.tandem.com>
List: port-i386
Date: 10/09/1996 21:03:44
In message <19961009120607.24120.qmail@mail.NetBSD.ORG>  Jim Rees wrote:
[..]

> 
> Before NetBSD we used Mach, and its pcmcia support was rock-solid.  It had
> lots of nice features, like not disconnecting the modem when you reboot
> (some people claim this is impossible; I invite them to my office for a
> demo).  I could power down the entire machine except for the modem, for days
> at a time, and when I powered back up everything worked.
> 
> Mach took a completely different approach.  Instead of treating the pcmcia
> bus as a bus, it provided an interface to it from user land.  That's it.  A
> user program talked to the card and mapped it into isa bus space.  From
> there the regular com driver took care of everything.  That may not fit the
> NetBSD architecture, but it certainly worked much better.

As the author of the original PCMCIA-framework I want to explain why the
architecture ended up that way. My reason not to rely on user-land stuff was
that I wantetd to do diskless boots and to boot from pcmcia-harddisks. 
The later was never done as I did get one of the little HP-palmtops.

I had never problems swapping cards and I don't know why you want your
keep running when turn of your machine. 
The reason I never tried to check if it can survive a reboot was, that I had 
to reinitialze my pcmcia chip at boot time because the BIOS on my laptop 
messed it up. Your power trick works only on a limited set of hardware I guess.

Stefan





--
Stefan Grefen                                Tandem Computers Europe Inc.
grefen@hprc.tandem.com                       High Performance Research Center
You should never bet against anything in science at odds of more than
about 10^12 to 1.
                -- Ernest Rutherford