Christopher P. Gill wrote:
> Is this bug specific to the mac68k port?  I didn't find any reference to
> it in the FAQ document, although I could have missed it.

it is mac68k-specific (at least i'm not aware of any other ports which
have this particular problem).

> However, it seems as though it'd be just as easy for some knowledgeable
> person to patch 'df' to do the right thing.  Or am I just a little naive?

the problem isn't in 'df', it's in the kernel.  every other port has a
working 'df', so we know it's code is doing the calculations correctly.
we must simply be screwing up what data we actually pass to 'df' from the
kernel.  i've tried to dig through this before, but it's fairly easy to
get lost through all the indirection that is necessary before you finally
hit the routines in disksubr.c :-)

later.

colin