On 11.09.2020 18:06, Julien Grall wrote: > --- a/xen/include/asm-x86/guest_atomics.h > +++ b/xen/include/asm-x86/guest_atomics.h > @@ -20,6 +20,7 @@ > ((void)(d), test_and_change_bit(nr, p)) > > #define guest_cmpxchg(d, ptr, o, n) ((void)(d), cmpxchg(ptr, o, n)) > +#define guest_cmpxchg64(d, ptr, o, n) ((void)(d), cmpxchg(ptr, o, n))
While them sitting side by side there's perhaps little risk of them going out of sync with one another, I still find it a little odd to open-code guest_cmpxchg() instead of using it, preferably even like #define guest_cmpxchg64 guest_cmpxchg (i.e. not using the function-like macro form). Unless of course there's a particular reason you went this route. Preferably with it adjusted this minor x86 part Acked-by: Jan Beulich <jbeul...@suse.com> Jan