Re: [PATCH] powerpc: Kconfig: remove BE-only platforms from LE kernel build

From: Boqun Feng
Date: Tue Sep 08 2015 - 23:26:31 EST


Hi Michael,

On Wed, Sep 09, 2015 at 12:26:44PM +1000, Michael Ellerman wrote:
> On Mon, 2015-09-07 at 07:58 +0800, Boqun Feng wrote:
> > diff --git a/arch/powerpc/platforms/cell/Kconfig b/arch/powerpc/platforms/cell/Kconfig
> > index 2f23133..808a904 100644
> > --- a/arch/powerpc/platforms/cell/Kconfig
> > +++ b/arch/powerpc/platforms/cell/Kconfig
> > @@ -25,7 +25,7 @@ config PPC_CELL_NATIVE
> >
> > config PPC_IBM_CELL_BLADE
> > bool "IBM Cell Blade"
> > - depends on PPC64 && PPC_BOOK3S
> > + depends on PPC64 && PPC_BOOK3S && CPU_BIG_ENDIAN
>
> We end up saying this five times.
>
> We already have PPC_BOOK3S_64 which captures the first two conditions, should
> we add a PPC_BOOK3S_64_BE which expresses it all?
>

I'm not sure whether this is worth.. IMO, we add a config option only if

1. we can use this config somewhere in the code, for example,
"#ifdef CONFIG_XXX"
or

2. we want to offer a option for users to choose.

PPC_BOOK3S_64_BE satisfies neither condition. Further more,
CONFIG_PPC64, CONFIG_PPC_BOOK3S and CONFIG_CPU_BIG_ENDIAN all are used
in some #ifdefs in current code. As a newbie of kernel, I'm happy to get
some knowledge like: "If I'm hacking PS3, then #ifndef CONFIG_PPC64 is
guaranteed to be false, so I can just ignore the code guarded by them".
But if we add a PPC_BOOK3S_64_BE, it will take a little more effort to
see this.

Regards,
Boqun

Attachment: signature.asc
Description: PGP signature