Re: Panic at boot on 4/330 with cgfour and cgsix

[David Brownlee <abs%absd.org@localhost>]

On Wed, 15 Jan 2025 at 23:17, foo bar <tokenalt%gmail.com@localhost> wrote:
>
> Hello all
>
> So as the subject says I'm having problems with these two
> framebuffers. It seems to be related to the issue reported a few
> months ago here.
>
> https://mail-index.netbsd.org/port-sparc/2024/04/15/msg002644.html
>
> I don't see the crazy interrupts but both framebuffers trigger a data
> fault szerr. For the cgsix I made a video of it but unfortunately the
> card died not long after so I can't test further.
>
> https://www.youtube.com/watch?v=iUPH3DHH8oA
>
> Luckily my cgfour is still working and I've attached a dmesg from
> current below. Any help would be appreciated.
>
> [panic message]

I missed an opportunity to pick up a 4/330 a while back, still regret it :-p

Would you be able to test boot earlier NetBSD kernels to see if this
is a regression? Start with NetBSD-9 and work back until you find a
kernel that works (or not!) You do not need to worry about userland at
this point, just enough to find out if the framebuffer attaches
without issue.

Another option might be to build a custom kernel with some printfs
scattered through cgfour.c - it looks like it gets at least as far as
https://nxr.netbsd.org/xref/src/sys/arch/sparc/dev/cgfour.c#250 when
it prints the resolution. My gut feeling is something may have been
missed ii the obio code paths when something was updated for sbus (but
that is very handwavy).

It is usually very quick and simple to setup a faster *nix to cross
build any NetBSD kernel (maybe a box being used to netboot? :) You
*can* build a NetBSD kernel on a sun4 box... but you may have time to
order and collect an Amazon delivery before it finishes...

David