On Wed, Jan 16, 2019 at 07:42:02AM +0100, Greg Kroah-Hartman wrote:
> On Tue, Jan 15, 2019 at 11:03:02PM -0700, Nathan Chancellor wrote:
> > When CONFIG_NO_AUTO_INLINE was present in linux-next (which added
> > '-fno-inline-functions' to KBUILD_CFLAGS), an allyesconfig build with
> > Clang failed at the modpost stage:
> >
> > ERROR: "is_broadcast_mac_addr" [drivers/staging/rtl8723bs/r8723bs.ko] undefined!
> > ERROR: "is_zero_mac_addr" [drivers/staging/rtl8723bs/r8723bs.ko] undefined!
> > ERROR: "is_multicast_mac_addr" [drivers/staging/rtl8723bs/r8723bs.ko] undefined!
> >
> > These functions were marked as extern inline, meaning that if inlining
> > doesn't happen, the function will be undefined, as it is above.
> >
> > This happens to work with GCC because the '-fno-inline-functions' option
> > respects the __inline attribute so all instances of these functions are
> > inlined as expected and the definition doesn't actually matter. However,
> > with Clang and '-fno-inline-functions', a function has to be marked with
> > the __always_inline attribute to be considered for inlining, which none
> > of these functions are. Clang tries to find the symbol definition
> > elsewhere as it was told and fails, which trickles down to modpost.
> >
> > To make sure that this code compiles regardless of compiler and make the
> > intention of the code clearer, use 'static __always_inline' to ensure
> > that these functions are always inlined. Some alternative solutions
> > included 'extern __always_inline' or converting these functions to
> > macros (so the preprocessor deals with them) but I would argue this is
> > the more "standard" solution.
> >
> > Signed-off-by: Nathan Chancellor <natechancellor@xxxxxxxxx>
> > ---
> > drivers/staging/rtl8723bs/include/ieee80211.h | 6 +++---
> > 1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/staging/rtl8723bs/include/ieee80211.h b/drivers/staging/rtl8723bs/include/ieee80211.h
> > index bcc8dfa8e672..959e822315b5 100644
> > --- a/drivers/staging/rtl8723bs/include/ieee80211.h
> > +++ b/drivers/staging/rtl8723bs/include/ieee80211.h
> > @@ -850,18 +850,18 @@ enum ieee80211_state {
> > #define IP_FMT "%pI4"
> > #define IP_ARG(x) (x)
> >
> > -extern __inline int is_multicast_mac_addr(const u8 *addr)
> > +static __always_inline int is_multicast_mac_addr(const u8 *addr)
>
> Ick, really? This is in a .h file, the .c file sees this, so why isn't
> clang picking it up? Worst case it just makes it a "normal" function
> and doesn't inline it, right?
>
As I understand it, the meaning of 'extern inline' is basically use this
defintion when inlining, otherwise use one from a different file or
translation unit. See commit d0a8d9378d16 ("x86/paravirt: Make
native_save_fl() extern inline"). These functions don't have any other
declaration/definition so when they aren't inlined, they don't exist.
> How about just replacing "extern" with "static", that should solve this,
> adding "__always_inline" everywhere is not going to be fun and doesn't
> make any sense.
>
Yes, that would work (and what I originally tested). My assumption with
the code is that it was intended for this function to always be inlined
but if you'd rather it just be 'static', I am happy to send a v2!
> thanks,
>
> greg k-h
Thanks for the quick reply and review,
Nathan
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
