Re: [PATCH] drivers: base: Add prototype declaration in memory.c

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Wed, Dec 18, 2013 at 11:10:24AM -0800, josh@joshtriplett.org wrote:
> On Wed, Dec 18, 2013 at 11:03:44AM -0800, Greg Kroah-Hartman wrote:
> > On Wed, Dec 18, 2013 at 11:01:08AM -0800, josh@joshtriplett.org wrote:
> > > On Wed, Dec 18, 2013 at 07:34:51AM -0800, Greg Kroah-Hartman wrote:
> > > > On Wed, Dec 18, 2013 at 12:09:53PM +0530, Rashika Kheria wrote:
> > > > > Add the prototype declaration of function memory_block_size_bytes() in
> > > > > memory.c.
> > > > > 
> > > > > This eliminates the following warning in memory.c:
> > > > > drivers/base/memory.c:87:1: warning: no previous prototype for ‘memory_block_size_bytes’ [-Wmissing-prototypes]
> > > > > 
> > > > > Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
> > > > > Reviewed-by: Josh Triplett <josh@joshtriplett.org>
> > > > > ---
> > > > >  drivers/base/memory.c |    1 +
> > > > >  1 file changed, 1 insertion(+)
> > > > > 
> > > > > diff --git a/drivers/base/memory.c b/drivers/base/memory.c
> > > > > index bece691..cfa03de 100644
> > > > > --- a/drivers/base/memory.c
> > > > > +++ b/drivers/base/memory.c
> > > > > @@ -83,6 +83,7 @@ static void memory_block_release(struct device *dev)
> > > > >  	kfree(mem);
> > > > >  }
> > > > >  
> > > > > +unsigned long __weak memory_block_size_bytes(void);
> > > > >  unsigned long __weak memory_block_size_bytes(void)
> > > > 
> > > > If this patch is the "solution" then something else is really wrong
> > > > here, please use your judgement to realize this...
> > > 
> > > It's a weak symbol provided and consumed within the same file, and then
> > > overridden on a few random architectures.  It isn't currently prototyped
> > > in any header file.  The other potential solution would be to add it to
> > > an appropriate existing header or create a new one.  Any suggestions for
> > > an appropriate existing header to prototype it in instead?
> > 
> > device.h?
> 
> include/linux/device.h?  That would work, though it seems slightly
> out of place there.
> 
> Would include/linux/memory.h work?

Sure, that's fine, I don't really care...