Subject: RE: Multia PCMCIA
To: 'Bernard Gardner' <>
From: David Woyciesjes <>
List: port-alpha
Date: 04/27/2000 09:32:10
I do have WinNT4 SP5 installed on a Multia at home... Send me a note of what
you want me to look for... I'll stick a card in and see what it comes up
with... The more specific you are, the better answers I can get for you...

---   David Woyciesjes
---   C & IS Support Specialist
---   Yale University Press
---   (203) 432-0953
---   ICQ # - 905818

> -----Original Message-----
> From:	Bernard Gardner []
> Sent:	Thursday, April 27, 2000 2:58 AM
> To:	David Woyciesjes
> Cc:
> Subject:	Re: Multia PCMCIA
> On Wed, Apr 26, 2000 at 01:43:05PM -0400, David Woyciesjes wrote:
> > I just saw you posting on NetBSD/Alpha mailing list.. Have you gotten
> the
> > PCMCIA to work? I have loaded NetBSD 1.4.2 "official" release, but
> NetBSD
> > doesn't know I have a 3Com 3C589C Etherlink III LAN PC Card
> (PCMCIA2.0/2.1
> > and PC Card Standard) inserted in it.... The are some postings regarding
> > this at ---
> Hi,
> I saw the start of this thread last night, but I was feeling a little
> ill, so I figured I'd wait until I was feeling better to reply. Today
> I'm home in bed, but I'm feeling a little better now, so here goes.
> I've been trying on and off (mostly off) for the past few weeks to
> get the PCMCIA support in NetBSD to work on my Multia. To do this,
> I've built a kernel with support for the PCMCIA hardware (you need
> a config file with the lines:
> #PCMCIA controller
> pcic*   at      isa? port 0x3e0 iomem 0xd0000 iosiz 0x4000 
> pcic*   at      isa? port 0x3e2 iomem 0xd4000 iosiz 0x4000
> #PCMCIA bus support
> pcmcia* at      pcic? controller ? socket ?
> And a line for whatever PC Card(s) you wish to use.
> With this config, I've been able to get my Multia to recognise the
> Lucent WaveLAN card that's currently in it, and it can read the
> ethernet address correctly from the card. However, the problem I'm
> currently having is that the -current kernel code attempts to probe
> for the ISA interrupts that are attatched to the pcic (PC Interface
> Controller?) and on the Multia this doesn't work. The kernel probes
> the two unused interrupts (skips those that it's already got marked
> as in use), neither of them are detected as being connected to the
> pcic(s) (noise about stray interrupts, and when you turn on debugging
> in the relevant files, it's reasonably clear that it's not behaving
> as a good i386 PC should) and as a result, when you attempt to configure
> the network interface (wi0 in this case) it can't start the interface
> (kernel message is "wi0: couldn't establish interrupt handler).
> I've tried configuring an interrupt in the config file (adding irq 5
> to the end of both pcic lines in the config file), this produces the
> warning "pcic0: warning: configured irq 5 not detected as available"
> The code is a little too smart for its own good. There is some code
> in /sys/dev/isa/i82365_isasubr.c that disables the probing for the
> hpcmips platform, I may see what happens if I disable the probing,
> and follow a config more like those found on an hpcmips.
> My (pretty random) thoughts on this are
> a) It actually looks like there is no interrupt associated with the
> pcic (maybe need to do something with isacfg in the SRM console)
> b) Linux claims to support the PC Card slots on a UDB, maybe booting
> a Linux kernel on the box would reveal some magic. (Linux/axp doesn't
> support the WaveLAN cards, so I haven't bothered with this tack yet)
> c) It looks like it should be possible to have the pcic interface
> run in an interruptless (polling) mode, although this might only be
> for support of card insert/remove detect events.
> d) NT probably supported these interfaces, so if it were possible
> for me to find an NT distribution that could be installed on the
> box that would probably serve a similar fact finding purpose.
> I'm quite keen to get this working, but I can't make any promises
> about when I might manage it, I mostly don't have the time to play
> with this.  I'm watching your progress with some interest, it's
> good to see that I'm not the only person who is interested in this,
> it means that it's more likely to actually happen, it's a shame for
> this to be so close but not quite work.
> Bernard.