BSP for MIPS32 4KC

[gregory.menke at gsfc.nasa.gov]

Hi Mike,

I know its been a while, but I finally got the opportunity to get your
patches worked out.  Its likely this patch will show up in a 4.6
snapshot.

- The __mips==32 #ifdef business looks reasonable, so I put it in as
is.  If the situation gets more complex it might be necessary to
re-evaluate how we're doing this.

- I discovered why you were having trouble with the
CPU_HARDWARE_FP==TRUE test, it turns out cpu.h which defines it TRUE
or FALSE was lacking the #defines for TRUE and FALSE, so
CPU_HARDWARE_FP was always set to undefined, and since the test was
always == undefined, it was always true.  #define'ing TRUE and FALSE
got the fpu support conditionally compiling as expected.  There is
some question about how fpu support should be included on a per-bsp or
per-MIPS variant cpu basis, but for now the cpu.h FPU options at least
work.

I've tested these mods on the JMR3904 simulator which is mips1 and
everything seems OK.  Testing on real hardware awaits the completion
of the move to our new lab, which should be done sometime between the
end of next week or the one after.  Before I give Joel the diffs, I
would like to have you test it against your hardware.  I think just
running ticker would be enough to ensure I didn't mess anything up on
the mips32 front.  I don't understand enough of your board to give you
a canned executable, so I think we have to communicate a bit to figure
that out.  If nothing else, I could make a tar.gz of my source
available.


Gregm