Re: [PATCH 2/2] MCE, AMD: MCE decoding support for AMD Family 16h

[Borislav Petkov <bp@alien8.de>]

On Tue, Dec 18, 2012 at 02:10:59PM -0800, Joe Perches wrote:
> On Tue, 2012-12-18 at 15:06 -0600, Jacob Shin wrote:
> > Add MCE decoding logic for AMD Family 16h processors.
> 
> More trivia:
> 
> > diff --git a/drivers/edac/mce_amd.c b/drivers/edac/mce_amd.c
> []
> > @@ -64,6 +64,10 @@ EXPORT_SYMBOL_GPL(to_msgs);
> >  const char * const ii_msgs[] = { "MEM", "RESV", "IO", "GEN" };
> >  EXPORT_SYMBOL_GPL(ii_msgs);
> >  
> > +/* internal error type */
> > +const char * const uu_msgs[] = { "RESV", "RESV", "HWA", "RESV" };
> > +EXPORT_SYMBOL_GPL(uu_msgs);
> 
> Do these really need to be exported?
> These don't seem to be used outside of mce_amd.
> 
> If these really need export, are these very good names?
> Perhaps these should be prefixed with something?

Hmm, some of those were used in amd64_edac at some point but then the
whole decoding code got concentrated into mce_amd.c but the exports
remained.

PP() and thus pp_msgs is still used there, for example.

So, if you'd like, you could wait until I apply Jacob's patches and push
that git branch to kernel.org and then write a patch ontop removing all
the exports except pp_msgs (you could add them to the "amd_" namespace
by calling it amd_pp_msgs or something) test that patch and send it to
me.

Thanks.

-- 
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--