It very well could be, I suppose I could attempt to install RH
4.2 with kernel 2.0.31 to test...
[snip]
>
> > I can't imagine building a smaller monolithic kernel. I
> > simply did _not_ add in support for any device not in the machine.
> > It
>
> You may even kill support for devices you *have* in machine, if you
> want small kernel. Also, you might want to boot something smaller than
> common redhat - I do not know how many bugs are in there. Is it
> possible that bootup scripts fail, rather than kernel fails?
>
No. The kernel freezes _at startup_ while it's initializing
devices. It never makes it to loading init (therefore never even
touches rc* scripts). I'm pulling out extraneous devices, so right
now I have a mono herc card, the adaptec 1540 card, and a NIC - that's
_it_. I know what the IO and IRQ values are for each card, and I'm
moving the same hardware between the 386 and 486... the 486 boots just
fine, the 386 freezes. I've noticed this problem creep up since about
2.0.18 where RAMdisk support seemed to cause a freeze, but with option
twiddling I've been able to get 2.0.29 with libc 5.3.12 to boot on the
386 (realize that the libc is never loaded as a shared object, this is
just what's linked with the kernel when I compile).
Thanks!
J. Maynard Gelinas