On Sat, 4 Dec 2021 at 06:33, kernel test robot <lkp@xxxxxxxxx> wrote:
[...]
I had been wondering if barriers could be static __always_inline, like70 | #define smp_rmb() do { kcsan_rmb(); __smp_rmb(); } while (0)include/asm-generic/barrier.h:70:30: warning: 'kcsan_rmb' is static but used in inline function 'pte_offset_kernel' which is not static
| ^~~~~~~~~
arch/alpha/include/asm/pgtable.h:303:9: note: in expansion of macro 'smp_rmb'
303 | smp_rmb(); /* see above */
| ^~~~~~~
70 | #define smp_rmb() do { kcsan_rmb(); __smp_rmb(); } while (0)include/asm-generic/barrier.h:70:30: warning: 'kcsan_rmb' is static but used in inline function 'pmd_offset' which is not static
| ^~~~~~~~~
arch/alpha/include/asm/pgtable.h:293:9: note: in expansion of macro 'smp_rmb'
293 | smp_rmb(); /* see above */
| ^~~~~~~
we do for atomic_t operations. Apparently this answers that questions.
This should fix it:
https://lkml.kernel.org/r/20211204125703.3344454-1-elver@xxxxxxxxxx
Unrelatedly, I wonder if I can get the test robot to also check some
of my branches:
https://git.kernel.org/pub/scm/linux/kernel/git/melver/linux.git/
That'd be very useful, to catch things like this early, and avoid
disturbing Paul too much. :-)
Thanks,
-- Marco