gc-linux-cvs

Re: [Gc-linux-cvs] linux/arch/ppc/kernel cputable.c,1.6,1.7

[Albert H. <alb...@ya...>]

--- Arthur Othieno <a.o...@bl...> wrote:
> And in doing so, you just broke the enumeration. The
> entries need
> to be ordered by .pvr_value. That's why 750CXe
> (Gekko) came before
> "745/755".

Uhmm, when looking at the current table that's not
true for already existing entries. The table is really
not sorted by pvr_value.

But looking at arch/ppc/kernel/misc.S I see that the
identify_cpu function stops on first match. So
specific entries (with a deeper mask) must be placed
_before_ more generic ones.

So at the end you are right. Gekko must be put before
745/755.

> Please revert.

Thanks a lot for pointing that out.

Cheers,
Albert

PS: Any comments about the recently commited EXI
framework?



		
______________________________________________ 
Renovamos el Correo Yahoo!: ¡250 MB GRATIS! 
Nuevos servicios, más seguridad 
http://correo.yahoo.es

Re: [Gc-linux-cvs] linux/arch/ppc/kernel cputable.c,1.6,1.7

[<a.o...@bl...>]

On Thu, Mar 17, 2005 at 06:53:20PM +0100, Albert Herranz wrote:
> Uhmm, when looking at the current table that's not
> true for already existing entries. The table is really
> not sorted by pvr_value.
> 
> But looking at arch/ppc/kernel/misc.S I see that the
> identify_cpu function stops on first match. So
> specific entries (with a deeper mask) must be placed
> _before_ more generic ones.
> 
> So at the end you are right. Gekko must be put before
> 745/755.
> 
Correct. I couldn't have put that better :)

> PS: Any comments about the recently commited EXI
> framework?
 
Haven't really looked at it. Probably later on during the day as I
catch up with my swamp of mail :(