> Is it
> because we're using ELF and then converting to a.out that this
> happens to ps:
>  USER       PID %CPU %MEM   VSZ  RSS TT  STAT STARTED    TIME COMMAND
>  root       179  1.0  0.1   308    0 p0  R+   12:15AM    0:00.20 (ps)

No, it's because ps has the wrong idea about how to find a process's
arguments.  If you have some time, you might poke at ps and see what
it takes to fix this - it shouldn't be hard, but I haven't had time
for it yet.   There shouldn't be any need for kernel mods to fix this
problem.

>     Also, I know the Linux people have been using ELF for a time -
> could we get some help from there on the toolchain? I know NetBSD
> 'should' be GPL-free but we don't have an alternate toolchain anyway.

As far as I know, Linux doesn't run on the pmax or on any mips box.
In any case, what do you think would be fixed by this?  We're all
using gcc, and indeed neither the Alpha port nor the mips port use the
NetBSD toolchain, which is quite old (and which, BTW, is also mostly
covered by the GPL, since the compiler and assembler and, I think, the
loader, are all descendents of GNU tools).

Thanks for catching the bug in cpu_exec.c.   The files.pmax omission
is because I haven't checked in the new version of files.c yet.
Sorry about that.

			       _MelloN_