Re: [PATCH v2 1/5] memblock test: fix implicit declaration of function 'virt_to_phys'

[Mike Rapoport <rppt@kernel.org>]

On Thu, Aug 01, 2024 at 12:07:39AM +0000, Wei Yang wrote:
> On Mon, Jul 29, 2024 at 08:37:23AM +0300, Mike Rapoport wrote:
> >On Mon, Jul 29, 2024 at 01:47:21AM +0000, Wei Yang wrote:
> >> Commit 94ff46de4a73 ("memblock: Move late alloc warning down to phys
> >> alloc") introduce the usage of virt_to_phys(), which is not defined in
> >> memblock tests.
> >> 
> >> Define it in mm.h to fix the build error.
> >> 
> >> Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
> >> 
> >> ---
> >> v2: move definition to mm.h
> >> ---
> >>  tools/include/linux/mm.h | 2 ++
> >>  1 file changed, 2 insertions(+)
> >> 
> >> diff --git a/tools/include/linux/mm.h b/tools/include/linux/mm.h
> >> index cad4f2927983..9db45c5d00fc 100644
> >> --- a/tools/include/linux/mm.h
> >> +++ b/tools/include/linux/mm.h
> >> @@ -20,6 +20,8 @@
> >>  #define pfn_to_page(pfn) ((void *)((pfn) * PAGE_SIZE))
> >>  
> >>  #define phys_to_virt phys_to_virt
> >> +#define virt_to_phys(p) ((unsigned long)p)
> >> +
> >
> >Please don't break #define of phys_to_virt from the static inline
> >implementation.
> >
> 
> I don't touch the definition of phys_to_virt, so you are willing me to adjust
> it?
> 
> The macro here will be expanded to code at preprocessing stage. I don't get
> how static inline would help here. Do I miss something?

static inline allows type checking of the parameters and #define with the
same name as of the static inline allows catching other defines with the
same name.

So it's common practice to use this combination, line phys_to_virt does.
 
> >And please make virt_to_phys a static inline similarly to phys_to_virt.
> >
> >>  static inline void *phys_to_virt(unsigned long address)
> >>  {
> >>  	return __va(address);
> >> -- 
> >> 2.34.1
> >> 
> >
> >-- 
> >Sincerely yours,
> >Mike.
> 
> -- 
> Wei Yang
> Help you, Help me

-- 
Sincerely yours,
Mike.