Subject: Re: [RFC] Running reverse-endian binaries
To: None <jonathan@DSG.Stanford.EDU>
From: M. Warner Losh <imp@bsdimp.com>
List: tech-kern
Date: 01/05/2003 00:05:01
In message: <200212262149.NAA20962@Pescadero.DSG.Stanford.EDU>
            Jonathan Stone <jonathan@DSG.Stanford.EDU> writes:
: For mips, note that you want an r4k or better; even then,
: IIRC, the per-process other-endian bit got dropped in mips32/mips64.
: (cgd? anyone else know for sure?)

It looks like this was deleted for the Vr4100 implementation.  The
Vr4100 is little endian only (at least according to the Vr4102 pdf).

>>RE: Enables/disables reversing of the endian setting in User mode
>>    (0->disabled, 1->Enabled).  This bit must be 0 since the Vr4102
>>    supports the little-endian order only.

I didn't check the other members of this family, but suspect that
there's no difference in this detail.

So there's no NetBSD/sgi binaries for NetBSD/hpcmips (unless the TX
parts support this, I don't have datasheets for them to check).

Also, one thing to consider is that there are different instructions
that may be used in the 'other endian' binary that might not be
implemented by your processor and those need to be emulated in the
kernel.

Warner