Before I toasted the machine trying to upgrade the bios while not getting
enough sleep...  I had an HP Vectra (Dual PPro/256k each) that had this
issue.  Of course, the reason I was trying to upgrade the bios was to try t=
o
get the MP kernel to boot.

If I come across another one of these boxes, I'll slap it up on the
operating table.  In the mean time I've sworn off boxes that didn=B9t come
natively with USB ports.

...just another datapoint

On 6/8/04 9:33 PM, "Geoffrey Lee" <glee@bluesat.unsw.edu.au> wrote:

> On Tue, Jun 08, 2004 at 08:20:13AM +0200, Frank van der Linden wrote:
>> On Tue, Jun 08, 2004 at 01:46:19PM +1000, Geoffrey Lee wrote:
>>> I think I should have been a bit clearer in my first message.  Yes, you=
 are
>>> right, I suspect that it is an interrupt issue too.  The ahc driver wil=
l
>>> send a WDTR message, to which the card replies.  In uniprocessor mode t=
his
>>> works, but in multiprocessor after the WDTR message is sent, then it
>>> believes
>>> it times out.
>>>=20
>>> The only difference that I can see is that the ioapic is enabled in MP
>>> mode ..
>>=20
>> Yeah, it's an interrupt problem.. can you try a kernel with MPVERBOSE
>> and send the boot messages?
>>=20
>=20
>=20
> Sure.  I'll do that in a private mail.
>=20
>=20
>> If it's a more recent machine, acpi and MPACPI might work, but this look=
s
>> like an older one.
>>=20
>=20
> I tried that too before, actually, but unfortunately it does not have
> ACPI.  Sigh.
>=20
> - g.