[patch] Arm FPA stfpls

[Richard Earnshaw]

On Wed, 2005-09-28 at 15:39, Paul Brook wrote:
> On Wednesday 28 September 2005 15:15, Richard Earnshaw wrote:
> > On Tue, 2005-09-27 at 16:42, Paul Brook wrote:
> > > Patch ping:
> > >
> > > Ambiguous FPA mnemonics in unified assembly mode:
>
> > This wasn't quite what I had in mind, since it makes only the 'pl/ls'
> > condition infix.  I think these instructions should be infix for all
> > conditions, not just pl/ls.
> 
> Should we allow conditional postfixes for these instructions? If so should 
> they trigger a warning?
> Should we warn about conditional infixes in other FPA instructions?
> Does this apply to all FPA instructions, or just stfp/ldfp?

As we discussed last night, I think we should leave all obsolete insns
as-is.  So *all* FPA insns should remain infix only.  As should teqp,
tstp, cmpp & cmnp.

This implies, for example, that cmppls is a legacy cmppl insn with infix
notation, not cmpp with postfix LS condition.

R.