Fwd: NetBSD 6.1.2 install report (A1200, Bliz 1230mkIV)

[Roc Vallès <vallesroc%gmail.com@localhost>]

On Tue, Dec 3, 2013 at 7:53 AM, Radoslaw Kujawa
<radoslaw.kujawa%c0ff33.net@localhost> wrote:
>
> Notice that we are using completely separate amidisplaycc driver for 
> OCS/ECS/AGA wscons support:
>
> amidisplaycc0   at mainbus0             # wscons interface to custom chips
> wsdisplay0      at amidisplaycc0 console ?
>
> wskbd0          at kbd0 console ?
> wsmouse*        at ms?
>
> I tested it few weeks ago and it worked just fine.

Noticed the WSCONS config file only when I installed Current. Will try
configuring and building another kernel sometime later, comparing my
conf with WSCONS via vimdiff.

The WSCONS kernel from the latest Current works. White on black looks
horrible via composite (what I can use with my current setup...).
Getting the default colors to be ite-like would be nice... as
commented out default kernel config entries if need be, or perhaps in
userspace.

With WSCONS, I should get virtual terminals, right? I have no clue how
to switch between them. Tried some mods + Fx but didn't get it to
switch. There's also only one "login" running. Perhaps there's some
step missing to bring the other consoles up?

> I don’t think so. For some time there was wsmouse enabled in the default 
> kernel even if no wseverything and X still worked.
> Please try it. I didn’t test X recently, but a few months ago it was working 
> fine on my A600.

On Current now, freshly installed, X crashes the kernel with an mmu
fault. Both the GENERIC kernel and the WSCONS kernel. I can provide
cores if needed.

> Ok. I’m interested in any problems with that.

For some reason, the clock went a year away (2013 vs 2012) at some
point (install or one of the first two boots), while keeping time
correct (wtf?).

I made sure clock was correct in AmigaOS, setclock saved, setclock
loaded, checked time again, then booted GENERIC and it seems to be
fine... no clue how that happened. Might be a PEBCAK on my end but
I'll keep an eye out for it happening again.

>>> That doesn’t surprise me much, but can you give more details about that 
>>> crash?
>>
>> I've uploaded the core on a server of mine. I suspect it'll be smooth
>> once I install Current. We'll see.
>
> Ok, please give it a look.

mkdb didn't crash during the first boot on Current.

> No. FPU_EMULATE is only for FPU-less machines.

Ok.

> All binaries that require FPU will run on 68882 without additional support 
> code (as far as I remember, 68882 has most extensive list of instructions 
> compared to other FPUs like in 68040 or 68060, where separate FPSP option is 
> needed).

Good to know. I knew 040+ had differences but didn't know about this.

>> Q: what to enable in config for kernel backtraces / kernel debugger to
>> show symbols (just DEBUG?)
>
> If you are using loadbsd you just need to pass the -S parameter.

Good. As I finish the email, netbsd is still dumping the second kernel
core (from starting X with GENERIC). Takes forever for some reason.
I'll re-read the loadbsd manpage afterwards.

Note that the loadbsd in Current doesn't work for Current kernels
either. Had to use the known-good one.