> From: osymh@gemini.oscs.montana.edu (Michael L. Hitch)
> 
> On Dec  6, 12:04am, "Stephen J. Roznowski" wrote:
> > > From: William Coldwell <billc@warp.cryogenic.com>
> > > 
> > > I diddled with conf.c and fd.c to fix a conflict with kern_descrip.c in the
> > > latest sup, but ended up breaking my kernel with a jump to 0 MMU fault.
> > > 
> > > *** /ftp/src/sys/arch/amiga/dev/fd.c	Mon Dec  5 20:32:56 1994
> > > --- fd.c	Mon Dec  5 18:41:55 1994
> > > ***************
> > > *** 434,440 ****
> > >   
> > >   /*ARGSUSED*/
> > >   int
> > > ! fdclose(dev, flags, devtype, p)
> > 
> > A much simplier fix is to make fdclose static inside fd.c. I've
> > done this and am currently running a kernel built this way.
> 
>   If fdclose() is static in fd.c, then how can it ever get called?  It
> has to be external so that the tables in conf.c are initialized properly.

Arrgh!!! You can tell how often I use my floppy drive. I assume my kernel
will bite the dust next time I try to use it! :-)

-SR