Ulrich Weigand wrote:
> Hmm, this can't be quite correct: you have all those
> 'op_size * count' in there, which is broken; a REP OUTSD
> might read from a lot of memory addresses (op_size * count),
> but it writes everything to the *same* 4 bytes in I/O space.
> So it really should be 'op_size' as it used to be ...
Hmm yeah... I'll fix it.
> And b.t.w. I don't think we need to consider the case that
> the 4 I/O bytes accessed by a single OUTSD/INSD instruction
> are registered by different drivers; if this really happens,
> it must be a bug anyway.
Well I dunno... I guess that if we don't support this, we won't have
a full emulation...
I'll mail a new patch tomorrow, going to bed now.
-- Ramon
