xfs
[Top] [All Lists]

RE: SGI 1200 and the scsi/apic problem

To: Matthijs van der Klip <matthijs.van.der.klip@xxxxxx>, "'Keith Owens'" <kaos@xxxxxxxxxxxxxxxxx>
Subject: RE: SGI 1200 and the scsi/apic problem
From: Seth Mos <knuffie@xxxxxxxxx>
Date: Fri, 23 Nov 2001 14:28:55 +0100
Cc: Matthijs van der Klip <matthijs.van.der.klip@xxxxxx>, Linux XFS Mailing List <linux-xfs@xxxxxxxxxxx>
In-reply-to: <35F87783B600D411A1430060943F469A589720@EXCHANGE>
Sender: owner-linux-xfs@xxxxxxxxxxx
At 14:22 23-11-2001 +0100, Matthijs van der Klip wrote:
Keith Owens <kaos@xxxxxxxxxxxxxxxxx> wrote:
>No, this is a generic problem with the 440GX motherboards, 1200 uses
>that m/b.  The PCI routing does not work correctly on a uni-processor
>kernel, booting with an SMP kernel or UP but using the apic table
>works.  Blame Intel for not providing info on the 440GX bugs.  RH 7.2
>install has this problem on a 1200, installing with apic worked.  I
>don;t know iof the XFS 1.0.2 installer supports the apic option or not,
>AFAIK it was a RH patch.

Hi,

Still trying. Results so far:

. Kernel-2.4.9-6SGI_XFS_PR3 either with or without the 'apic' option
  results in scsi timeouts. This kernel originates from the 'testing'
  area. I have transferred/cloned a full production system to a 1200 to
  test it.

Better use the 2.4.9-13 kernel for Red Hat Linux 7.2 since it also plugs a few security holes.

. The installer for Linux-XFS 1.0.2a _does_ boot with the 'apic'. I
  reported it didn't before but I must have done something wrong.

Automagic.

Problem is I don't want to do a full install from scratch on this system,
so I'll try to update my kernel-2.4.9-6SGI_XFS_PR3 install to the newest
kernels from Release 1.0.2a and then transfer it to the 1200 again.

If the system boots with an older kernel you should be able to do a rpm -ivh kernel-2.4.9-13*.rpm to get it installed.

PS. Sorry for posting this with Outlook, but I do not have any choice at
    the moment. :(

It's plain text, that is good enough.

Cheers

--
Seth
Every program has two purposes one for which
it was written and another for which it wasn't
I use the last kind.


<Prev in Thread] Current Thread [Next in Thread>