Subject: Re: ncr0 on NetBSD/i386 AMD P133 Symbios Atlas UltraWide
To: None <current-users@NetBSD.ORG>
From: Davyd Norris <>
List: current-users
Date: 05/30/1997 12:53:28
Michael Richardson wrote:
>   with -current from late April: (transcribed)
> ncr0: SCSI phase error fixup: CCB already dequeued (0xf877ce00)
> ncr0: timeout ccb=0xf877d000 (skip)
> ncr0: timeout ccb=0xf877d200 (skip)
> ncr0: timeout ccb=0xf877d400 (skip)
> ncr0: timeout ccb=0xf877da00 (skip)
> ncr0: timeout ccb=0xf877de00 (skip)
>   machine is locked.
>   I will enclose boot message in another message, since this terminal
> doesn't have much in the way of editing facilities.
> i NCR is Symbios Ultra/Wide.

Yup, looks like you have been bitten by the old NCR bug. This is what I
get with my machine on any kernel after about December. I am still
running a 1.2BETA kernel and have upgraded as much userland as I can but
I am still stuffed with things that use the new system calls like Apache
and poll() etc.

Following on from this:

I am in the process of trying to replace my NCR with a QLogic but I
first have to get a kernel that does not give me supervisor mode panics
when I boot from it. There was something on this a while back - do I
have to upgrade any userland stuff before I can get a new kernel
running? If I do, will I still be able to boot the old kernel?

To try and get around this, I am trying to make a set of floppies with
the latest stuff on it but I am running out of space on the inst floppy
image - any ideas? Is it as simple as typing make in the floppies
directory or am I an idiot?

Davyd Norris - Systems Manager
Franklin Collins Pty. Ltd.
Melbourne, Australia.