>>>> Try compiling the kernel (or maybe just subr_autoconf.c) with -O
>>>> instead of -O2, and see if that helps.
>>> Also, if you upgrade the compiler to the latest (2.7.2) then that
>>> problem should disappear.
>> I get the optimization problem with 2.7.2 from -current.  1.2 kernel
>> won't even build under 2.4.5 (or whatever the 1.1 GCC version was)
>> anymore.. one of the nfs server files kills it.
>
>My experience with optimization under NetBSD/sun3 is not good.
>
>Specifically, under 1.1's gcc, you have to turn off optimization for
>the kernel - or at least for that file.  But under 1.2, you have to
>turn optimization _on_, or what happens, I think the compiler silently
>bugchecks (cc1 dies with SIGABRT, IIRC).
>
>I haven't sent-pr this because the sun3 systems are in a state of
>sufficient flux that I don't have a system on which to build the latest
>sup and see if that fixes it, and a sup over a month old is a bit stale
>to report PRs relative to.
>
>					der Mouse
>
>			    mouse@collatz.mcrcim.mcgill.edu
>		    01 EE 31 F6 BB 0C 34 36  00 F3 7C 5A C1 A0 67 1D


When booting a 3/50 via ether, I'v had simular experiences as above
[core dumps and corruption in many forms, most repeatable on demand].

However, after I got a disk for the machine, things got much more
stable [compare 'day and night' sort of behavior].






Brad Spencer - brad@anduin.eldar.org   http://anduin.eldar.org