Re: Netbsd v 8.0 fails to boot on spare ultra 45

Could you try a GENERIC.UP kernel and/or disable the audio driver?

Thank you,
Jerome

On Wed, Dec 19, 2018 at 2:48 PM Michael L. Hitch <[email protected]> wrote:
>
> On Tue, 23 Oct 2018, Eduardo Horvath wrote:
>
> > On Tue, 23 Oct 2018, Glenn Mawby wrote:
> >
> >> Hi,
> >> I???ve a spare ultra 45 with one cpu, 5gig ram and a xvr-2500 graphics card.
> >>
> >> I???ve tried to boot a fresh burn of Netbsd spare 64 version 8.
> >>
> >> It starts to boot but I get Error: Software initiated reset has occurred and it reboots.  Any ideas?
> >
> > SIR is an indication the kernel got seriously stuffed up in trap
> > handling code.  And it doesn't look like OBP properly recovered since
> > you're not getting to the `ok' prompt.
> >
> > How far did it get?  Can you capture the kernel output during boot?
> >
> > Also, try disconnecting the keyboard and using the serial console.
>
>    I saw this same think back in May - I was trying to get something booted
> on my ultra 45 I had in my office.  I tried an ISO from the daily builds
> and noticed I got an SIR on boot.  I didn't pay much attention to it, as I
> had a disk at home that had a bootable system and used it.
>
>    A couple of weeks ago, I decided to turn on the ultra 45 at home, and
> since it had a build environment on it (from 2014), I thought I try some
> newer kernels.  I set up a build for netbsd-8 and -current and found that
> my netbsd-8 kernel also crashed with the SIR on boot.  I started looking
> into that and determined that it occurred when configureing the audio
> driver.  I got stuck trying to track it through the code, so now I'm
> trying builds of various dates between Sep 2015 (last good kernel I had)
> and Sep 2017.  I would presume the problem is either there when netbsd-8
> was branched, or some pullup done afterwards.  I'm hoping to find the
> exact commit with breaks it and then see how to fix it.
>
> > Not getting to the `ok' prompt makes me suspect the video card, but maybe
> > OBP just switches to the serial port after an SIR.
>
>    OBP seems to just reboot the machine with a 'boot' command.
>
> > To diagnose the problem we really need to get the pc value of the SIR
> > instruction that was executed.
> >
> > Eduardo
>
>    I don't see any PC reported when it gets the SIR, so I haven't
> determined exactly where that is occuring.
>
> Mike
>
> ---
> Michael L. Hitch                        [email protected]
> Operations Consulting,  University Information Technology
> Montana State University, Bozeman, MT     USA