In article <199511282119.OAA15048@cvs.openbsd.org>,
Niklas Hallqvist  <niklas@cvs.openbsd.org> wrote:
> + #ifdef m68k
> + 	sunos_setregs,
> + #else
>   	setregs,
> + #endif

I think it would be better to make this sunos_setregs unconditionally
and have everyone port or arch define sunos_setregs, even if all it does
is call setregs (actually though, there was a recent commit in the sparc
port to make ld.so flush caches for seperate I&D cache CPUs, so they
probably need a similar sunos_setregs anyway.