On Sun, 26 Nov 2000, Ben Harris wrote:

> On Sun, 26 Nov 2000, Kjetil B. Thomassen wrote:
> 
> > rtc0 at iic0 addr 0xa0: PCF8583
> 
> Hmm.  There's something wrong here.  It should report "PCF8583, 32.768 kHz
> clock".  I've been getting the same problem (and occasionally "event mode,
> stopped" on my A3020 -- I'll slap a logic probe on the I^2C bus and see
> what it's playing at.

Is there anything I can do?

> > BTW, that reminds me. Is it possible to have the ISDN route provide the
> > IP address and then mount the root and swap from my NetBSD/sparc 1.4
> > box?
> 
> To do that, you'll need to arrange the that the router provide a root path
> for the R140.  If you can get it to emit arbitrary DHCP options, this
> should be possible.  If not, you'll need to run dhcpd on your SPARC (or
> use RARP/bootparams).

What if I run dhcp on the router to give it the IP address, and then run
bootparams on my ss2. Will that work?

Anyway, I'll check my router doc and see if it supports it. If not, I
guess I just have to disable it and use the ss2.

> > Also, are there any particular things to your setup of booting NetBSD
> > over the network? I didn't find any information that was relevant to
> > either arm32 or arm26.
> 
> There's nothing arm26-specific you need to do if you're loading the kernel
> from disc.  You'll need a DHCP server and an NFS server, appropriately
> configured.

Good.

> > A couple of other things about the probe above:
> > 1) The internal ST-506 drive does not seem to have been found.
> 
> This isn't surprising -- I don't have a driver for it yet, and since it's
> in a fixed place, the kernel doesn't trip over it like it does over
> surprising podules.

Are you planning to have a driver for it? If so, when?

> > IDE Interface (000c:0022) at podulebus0 slot 2 not configured
> 
> > 2) The IDE Interface in slot 2 is an HCCS one.
> 
> I've added that to sys/arch/arm32/podulebus/podules.  Note that the arm26
                                 ^^
Shouldn't that be arm26?

> podulebus code just prints the podule's on-board description (if it can
> find it) rather than having the entire podule list compiled in.

Just like arm32 as far as I can tell.

> > podulebus0:3: ROM is not 8 bits wide; ignoring it
> > podule (a807:5929) at podulebus0 slot 3 not configured
> 
> > 3) The card in slot 3 is an AKA-30 SCSI card with a PowerROM from Partis
> >    Computing. It does not seem like it works on pre-Risc PC computers,
> >    so this is something I will ask Partis about.
> 
> Hmm.  From the messages it looks like it's starting up in an odd state,
> since the AKA-30 can't take a 16-bit ROM and the podule ID looks like
> garbage..

You're right. There is something fishy around here. I will have to look
into this. Maybe I should try my AKA-32, or try this card in my Risc PC.

The arm26 snapsnot has been extracted to an export area on my ss2, so as
soon as I get it exporting that and a swap area, I will test a proper
boot.

Then the cross-compiler seemed to install properly on my ss2, and I will
start compiling. Would it be enough to do cvs checkout -D "05 Oct 2000" to
get the sources for that snapshot?

Thanks and TIA!

Kjetil B.
mailto:kjetil@thomassen.priv.no
http://www.thomassen.priv.no/