Re: gcc-3.2 breaks rmap on s390x

[Davide Libenzi]

On Thu, 3 Apr 2003, Rik van Riel wrote:

> On Thu, 3 Apr 2003, Pete Zaitcev wrote:
>
> >  			cpu_relax();
> > +			barrier();
>
> Gah, now I look over the source I see that cpu_relax() is always
> used together with barrier() ...
>
> I guess the best long-term thing (2.5) would be to build a barrier
> into cpu_relax(), but for 2.4-rmap I'll just add your patch.

cpu_relax() is a barrier on Intel ( pause ). It's just a coincidence IMHO.

- Davide