Re: [PATCH] Fix bootstrap failure on K6

[Jan Hubicka <jh at suse dot cz>]

> 
> On Sat, 18 Oct 2003, Eric Botcazou wrote:
> > May I ask you to take a look at
> > http://gcc.gnu.org/ml/gcc-patches/2003-10/msg01386.html
> > upon which this patch sort of depends?
> 
> Instruction attributes in the back-end are at the very limits
> of my reviewing "comfort zone".  I'm happy with the patch itself
> (which obviously only affects k6 and fixes the bootstrap failure),
> but the thing I can't comment on is whether the x86/k6 maintainers
> would prefer the problem to be fixed another way, for example, by
> using an attribute other than mode.
> 
> If Richard Henderson or Jan Hubicka indicate that this is the
> appropriate way to go, the patch can go in.  Alternatively, you can

Yes, the patch is right thing to do.  The only reason why K6 machine
description checks mode directly is the fact that mode attribute has
been introduced later and I failed to update this place.

Honza
> apply the patch now, on condition that you promise to make the
> necessary changes, should either RTH or Jan prefer an alternate
> solution.
> 
> A reasonable compromise?
> 
> Roger
> --
>