[PATCH v3 1/3] powerpc/nvram: move generic code for nvram and pstore

[Michael Ellerman]

On Wed, 2015-01-14 at 23:35 +0530, Hari Bathini wrote:
> On 01/14/2015 10:01 AM, Michael Ellerman wrote:
> 
> > On Wed, 2014-12-24 at 17:28 +0530, Hari Bathini wrote:
> > > With minor checks, we can move most of the code for nvram
> > > under pseries to a common place to be re-used by other
> > > powerpc platforms like powernv. This patch moves such
> > > common code to arch/powerpc/kernel/nvram_64.c file.
> > As I said in my reply to the previous version:
> > 
> >     ... you need to keep in mind that it is very common for us to build kernels
> >     with both POWERNV=y and PSERIES=y.
> >     
> >     So you need to make sure you're only using CONFIG_PPC_PSERIES to protect things
> >     that are optional on pseries. Not things that we *shouldn't* be doing on
> >     powernv.
> 
> we could as well do away with the PPC_PSERIES flag in a couple of
> places in
> arch/powerpc/kernel/nvram_64.c, but doing that will unnecessarily add
> few extra variables for !PPC_PSERIES case. 

Yep. I'm happy for them to be there, I just want you to explain in the
changelog that you've thought about the PSERIES=y POWERNV=y case and why the
code makes sense for that configuration.

> > Please explain in your commit message how you have dealt with that.
> 
> Sure. Will update the changelog

Thanks.
 
> > Also, you broke the build for every config that doesn't have
> > CONFIG_PPC_PSERIES, all 95 of them. This is pasemi_defconfig for example:
> 
> My bad! clobbering_unread_rtas_event should have been static inline
> while defining under !PPC_PSERIES

Correct.

Please make sure you test build at least some of the other configurations in
future. I realise it's too time consuming to build all of them, but ideally for
every config symbol you use in your patch you need to build a kernel config
where that symbol =y and =n (and =m if it's tristate).

cheers