[OpenAFS-devel] latest process.i386.s not working on NetBSD

[Tracy Di Marco White]

In message <20040809120208.5615713022@citi.umich.edu>, Jim Rees writes:
>Is this a UKERNEL build?  The includes at the end of param.i386_nbsd20.h are
>obviously wrong.

I'm not sure what UKERNEL is, so I'm not sure.

>The basic problem here is that process.s needs afs/param.h, but will break
>if it also gets sys/param.h (because this is asm, not C).  The
>poorly-thought-out fix is IGNORE_STDS_H.  I think it's a mistake to include
>sys/param.h from afs/param.h, but that would be hard to fix.
>
>I don't think anyone has looked at the NetBSD build in a long time.  I would
>be surprised if it worked at all.

I've been running a NetBSD server for a couple months at work.  One of
my co-workers started with it around 1.3.52 it looks like.  I set up
a cell using OpenAFS on NetBSD over the weekend at home, using HEAD and
making sure it all compiled.  There weren't that many problems with the
server side of things.

The OpenAFS client doesn't work for NetBSD yet.  I'm hoping for that
to change, but the server was the easier part.

-Tracy