Re: 2.6-test2 on unusual UE2

[Ben Collins <bcollins@debian.org>]

> > The thing is that the driver detects in the order that PCI probes. So
> > lower numbered busses and devices get probed first. Likewise, lower
> > ordered scsi ID's get probed first.
> 
> His problem is different Ben, it's finding the Qlogic vs. Symbios
> cards in a different order in 2.6.0
> 
> Ie. in one the qlogic's get found first, in the other the symbios
> cards are found first.
> 
> Something in the 2.6.x scsi makefile changes are causing this
> to happen.

Ah, I missed that. Only thing to do is build in the one that is for the
root disk, and make the other a module.

> > 1) The compiler should help here
> 
> I don't think so :)  Especially since he tried this with egcs64
> on 2.4.x which should work.

True, but definitely can't hurt in the 2.6 case :)

> I think they just do the SMP boot sequence or cpu numbering different
> in the OBP found on his box.
> 
> > 2) If you still get this, boot with the -p option on the OBP or SILO
> >    command line. It will get you earlyboot prom output so we can see
> >    where it stops.
> 
> Indeed.

Almost forgot, prtconf -pv output from either Solaris or Linux would be
helpful aswell.

-- 
Debian     - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/