Hi Kevin!

On Thu, 1 Mar 2012 11:19:26 -0700, Kevin Buettner <kevinb@redhat.com> wrote:
> I think it is still okay because [...]

I agree with your reasoning.


> On Thu, 01 Mar 2012 11:30:13 +0100
> Thomas Schwinge <thomas@codesourcery.com> wrote:
> > Still learning about GDB's code layout -- is there any benefit in
> > invoking gdbarch_register_reggroup_p (as you're doing) in contrast to
> > directly going for sh_register_reggroup_p?
> 
> Well, the way I'm doing it is slower.
> 
> But, if at some point someone were to add another
> `register_reggroup_p' method for some other sh variant (as is the case
> for `register_name'), the code that I wrote will still work.  If I
> called sh_register_reggroup_p directly, then presumably the wrong
> function would be called for that new architecture variant.

Sure, but a person adding a new variant of sh_register_reggroup_p would
surely be checking all places where the current only implementation is
invoked, and decide which is now applicable, the old or the new -- or
simply defer that decision to gdbarch_register_reggroup_p then; so we
might as well do that right now.


> 	* sh-tdep.c (sh_frame_cache): Don't fetch the FPSCR register
> 	unless it exists for this architecture.

Looks good; before committing, you may want to unify spaces/tabs usage
used for indenting the comment.


Grüße,
 Thomas