Chuck Cranor wrote:

> actually, i am also sitting on the following diff that gets
> rid of the 64 bit instructions that cause the IDT32364 to 
> choke.   since we don't actually do 64 bit, they shouldn't
> harm anything.
> 
> i could commit as is (as Toru Nishimura suggested to me in
> an email), or i could abstract a bit ala REG_L/REG_S.  is
> anyone actively working on 64 bit mips3 support?   if that
> isn't currently a hot topic i'm inclined to follow Toru's
> advice and commit as-is.

Perhaps a comment at the top of locore_mips3.S somewhere to say
why m{f,t}c0 is currently being used instead of dm{f,t}c0?

Simon.