Re: IBM Workpad z50 dmesg collection

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

On Sat, 8 May 2021 at 19:55, Alexander Bochmann <ab%lists.gxis.de@localhost> wrote:
>
> Hi,
>
> I sent a couple of dmesg outputs to dmesgd, but only got up to 8.1 for
> now (8.2 hangs after "Updating motd" during init, and I'm running out
> of time). 9.0 and 9.1 don't find a boot device,

Excellent - thanks (and sorry for reply latency on this)

> Contrary to what I wrote before, PCMCIA still is operable in 8.x on the
> z50, so maybe I mixed that up with that happened in 9...
>
> So I have:
>
> 7.0.2: https://dmesgd.nycbug.org/index.cgi?do=view&id=6072
> This release is rock solid on the z50, except for a few spurious kernel
> messages (mostly "vrpiu:AD: " and three numbers, and some with
> "giu data: " plus ASCII art).
> I've been running it for weeks building software locally on the system.
>
> 7.1.2: https://dmesgd.nycbug.org/index.cgi?do=view&id=6073
> Adds some "ne0: NIC memory corrupt" kernel messages, but networking is
> operable in principle (I downloaded the other kernels I tried to boot
> with this version, and the 7.0.2 userland).
>
> 7.2: doesn't boot, so I don't have a dmesg
> Hangs after the "pcic0: WARNING: powerhook_establish is deprecated"
> message
>
> 8.0: https://dmesgd.nycbug.org/index.cgi?do=view&id=6075
> Userland starts failing with this release. ntpdate segfaults during
> bootup (still on 7.0.2 binaries otherwise), and I collected about a
> dozen core files in 5 minutes on running this kernel.
>
> 8.1: https://dmesgd.nycbug.org/index.cgi?do=view&id=6076
> Init hangs at "virecover" (until aborted using Ctrl-C), same with the
> segfaults.
>
> 8.2: more of the same, but didn't complete booting,
> so I don't have a dmesg right now.
>
> 9.0 and 9.1 don't find a boot device, and while the ne2k card is
> detected, I get the "ne0: where did the card go?" message immediately
> after. Since I didn't find the serial cable I didn't have an easy
> method to save a dmesg without storage and network.
>
> I do see messages for the storage device too - it finds an atabus0
> and a wdc1, but there is no wd0 device in the end. Doesn't matter if
> I use the CF slot or a PCMCIA/CF adapter card.
> I could post a few photos of the boot screen that I took.
>
> So I guess there's two topics - what happened to the userland between
> 7.1.2 and 8.0, and what's up with the PCMCIA devices from 9.0 on.

Iff (and its a big iff) 9.x has fixed the segfault issues, then it may
only be the pcmcia issues which need to be fixed.

More thoughts
- Does the latest
http://nyftp.netbsd.org/pub/NetBSD-daily/HEAD/latest/hpcmips/ kernel
boot? In the unlikely event of the pcmcia issue being fixed since 9.x
its probably worth checking
- For the 8.x boots, I wonder if it could be floating point or shared
library related. There are static linked binaries in /rescue,
including 'sleep', 'expr' & 'ps' which can use floating point. It may
be worth a small amount of testing to see if an error can be triggered
with a /bin/ vs /rescue/ binary
- What would you require as a minimum for testing - just a built
kernel? (Thinking about bisecting the device issue between 8 & 9)

Thanks

David