Subject: Re: bad system call revisited
To: Michael G. Schabert <mikeride@prez.buf.servtech.com>
From: Bill Studenmund <wrstuden@loki.stanford.edu>
List: port-mac68k
Date: 08/25/1997 13:34:22
> 
> Hi NetBSDers,
> I have a Quadra 840 AV which had been happily running GENERIC-17 with
> NetBSD 1.2 binaries since December or so. I decided to be brave & try
> Allen's new kernel, which worked fine for me. Of course, because of the
> kernel/binary synch, I couldn't use PS, which I find quite useful. So, I
> downloaded -current binaries (and GENERIC-35 just for the heck of it), &
> proceded to install them (selected them all in booter 1.1c, hit done, &
> wait awhile). I then downloaded installer 1.1e because I read that it gives
> more pty/tty's by default. I used this to build devices. I then tried
> booting using booter 1.9.4. Just after the pramtime read, I got a bad
> system call.

Did you update /etc? Also, did you include a swap partition definition in
/etc/fstab? I think the installer did by default, but that bad sys call
sounds like a swap setup problem (swap setup cahnged a few months ago, and
the utilities might not have caught up).

               On the next line was the fsck syntax line. After that was a
> request for a shell (or return for sh). I ran "fsck -f /" because "fsck /"
> said that the filesystem was still marked clean.

It really wants "fsck -p /". Your /etc script has fsck -p, which used to
work. I'm not sure what the problem is, except possably that your
/etc/fstab has "ufs" instead of "ffs" for fs type. Just a guess. I had to
make the same change too.

                                                   This gave only 2 errors,
> one asked rejoin, & I forget the other offhand, something like recover
> damaged something which was incorrect in superblock I think. This was my
> first time ever needing to run fsck in my 6 mo experience with NetBSD. I
> then tried to start from scratch by doing a shutdown -r now & going back
> into the booter. I got the same error. I thought it may have had something
> to do with Allen's kernel, even though I'd had no trouble with it the past
> couple days, so I unpacked GEN-35 (I get a big D'oh! for untarring it over
> GEN-17) & tried booting into that. I got the same error. I then renamed the
> GEN-35 kernel & reinstalled GEN-17 (luckily I don't like trashing stuff)
> but that too spit out the bad system call. Essentially, I had reinstalled
> GEN-35 as well as all 7 tarballs (renaming /etc, of course), which should
> have given me a pristine new system. This should rule out that you were
> telling Josh that it is usually a kernel/binaries sync problem.

Answers my question about /etc!

> I was wondering if anyone had any ideas before I go wiping everything &
> reinstalling GEN-17 with the 1.2 binaries & all my added binaries (dt, X,
> pine, lynx, qpopper, etc).

Check the above problems, and see if they fix things. If so, then we need
to change the install setup to keep up.

Thanks!

Bill