On Fri, Dec 22, 2000 at 02:06:37PM +0000, Ben Harris wrote:
> On Fri, 22 Dec 2000, Nicholas Clark wrote:
> 
> > On Thu, Dec 21, 2000 at 11:16:01PM -0800, Chuck Silvers wrote:
> > > userland binaries should be in the same "port".  (that still leaves the
> > > question of chips like the MIPS that can run both big- and little-endian,
> > > but let's leave that aside for the moment.)
> > 
> > Maybe I misread your intent, but to me your exact words imply that there
> > will never be a big endian arm port.

sorry, I didn't mean to imply that either.  what exactly would be
different in the source code for a big-endian vs. little-endian ARM port,
other than the one line in endian_machdep.h (and various things which would
be defined based on that one line)?  we currently have both big- and
little-endian MIPS ports, but there's no "mipsel" or "mipseb" directories
in the source tree.  there would need to be "armel" and "armeb" directories
in the release tree for the two sets of userland binaries, and of course
the kernels would be separate, but that doesn't seem very hard to deal with.


> > I don't know if I'm repeating something already known here, but there
> > is already a big endian arm linux port for the Intel IXP1200
> 
> Hmm.  I was wondering whether there had ever been any machines that used
> the big-endian ARMs.
> 
> > I don't know if this affects NetBSD thinking.
> 
> If nothing else, it probably means that a merged ARM MACHINE_ARCH should
> probably be called "armel" rather than "arm".

for the MIPS platforms, MACHINE_ARCH is "mipsel" or "mipseb" even
though there are no directories with those names in the source tree.
most of the MIPS-specific source code is shared even between hardware
of different endianness.  the same should be true for ARM.

-Chuck