On Fri, 7 Dec 2012, Will Schmidt wrote:
Date: Fri, 07 Dec 2012 11:05:37 -0600
From: Will Schmidt <will_schmidt(a)vnet.ibm.com>
To: George E Sollish <gsollish(a)autogear.net>
Cc: ppc(a)lists.fedoraproject.org
Subject: Re: IBM Intellistation 275
On Fri, 2012-12-07 at 04:51 -0500, George E Sollish wrote:
> Hi,
>
> Thanks for explaining why it doesn't work, and why crux ppc64 does. I'm
> running IBM's last firmware rev on the 9114-275 (...d79e22...) so that
> avenue's already closed. I took a quick look at the fairly intimidating
> webpage Dan suggested --- does the donor fedora system need to be a
> ppc32/64, or can it be one of my several intel installs? I like the idea
> of a smaller boot image, but that's way beyond me. Is there a chance that
> could be made available?
> Previously, I've run Crux ppc64 on this machine; currently its running
> AIX 5.3.
Have you tinkered at all with real-base values on the 275 ?
Historically when one switches between AIX (or AIX based system
diagnostics CD) and Linux, the real-base values get updated and the
updated values have potential to confuse the snot out of firmware.
I've no idea if the later bootloaders correctly handle the scenario, or
if it's actually a problem that exists with the firmware on the 275, but
may be worth poking at.
See
https://bugzilla.redhat.com/show_bug.cgi?id=527506#c5 and later for
some comments on that subject.
Thanks,
-Will
Hi Will,
When I acquired the box I tried installing fedora, failed, and then moved
over to crux 2.6 ppc64, but finally tired of the inverted color map for
the 4500P graphics. At that point I switched back to AIX 5.3. I've also
run the 9114-275 diagnostics several times, so you could be right about
the real-base values. I read the thread you mention, and I'm willing to
reset the real-base to see if that is, indeed, the problem; how do I
proceed?
Thanks.
George E Sollish
Chief Engineer
AUTO GEAR COMPANY, INC
SYRACUSE -- since 1885, the first
name in automotive gearing