On Wed, 10 Apr 2019, Joe Lawrence wrote:

> On Wed, Apr 10, 2019 at 11:50:53AM -0400, Joe Lawrence wrote:
> > From: Josh Poimboeuf <jpoim...@redhat.com>
> > 
> > Define macros KLP_MODULE_RELOC and KLP_SYMPOS in
> > include/linux/livepatch.h to improve user-friendliness of the
> > livepatch annotation process.
> > 
> > Signed-off-by: Josh Poimboeuf <jpoim...@redhat.com>
> > Signed-off-by: Joao Moreira <jmore...@suse.de>
> > Signed-off-by: Joe Lawrence <joe.lawre...@redhat.com>
> > ---
> >  include/linux/livepatch.h | 12 ++++++++++++
> >  1 file changed, 12 insertions(+)
> > 
> > diff --git a/include/linux/livepatch.h b/include/linux/livepatch.h
> > index 16b48e8b29a2..947cfc2d1980 100644
> > --- a/include/linux/livepatch.h
> > +++ b/include/linux/livepatch.h
> > @@ -236,6 +236,18 @@ void *klp_shadow_get_or_alloc(void *obj, unsigned long 
> > id,
> >  void klp_shadow_free(void *obj, unsigned long id, klp_shadow_dtor_t dtor);
> >  void klp_shadow_free_all(unsigned long id, klp_shadow_dtor_t dtor);
> >  
> > +/* Used to annotate symbol relocations in livepatches */
> > +#define KLP_MODULE_RELOC(obj)                                              
> > \
> > +   struct klp_module_reloc                                         \
> > +   __attribute__((__section__(".klp.module_relocs." #obj)))        \
> > +   __attribute__((aligned (4)))
> > +
> > +#define KLP_SYMPOS(symbol, pos)                                            
> > \
> > +   {                                                               \
> > +           .sym = &symbol,                                         \
> > +           .sympos = pos,                                          \
> > +   },
>         ^^
> nit: if we dropped the trailing array comma delimiter from KLP_SYMPOS
> macro, the invocations would look more intuitively like an array.  For
> example:
> 
>   KLP_MODULE_RELOC(test_klp_convert_mod) test_klp_convert_mod_relocs_a[] = {
>       KLP_SYMPOS(driver_name, 0),
>       KLP_SYMPOS(homonym_string, 2),
>       KLP_SYMPOS(get_homonym_string, 2),
>   };
> 
> But I could not figure out a good regex to reference if other such
> kernel preprocessor macros include or exclude the delimiter.  Are there
> reasons to include it?

No, I don't think so. We should change it. Good catch.

Miroslav

Reply via email to