Hi Julien, > On 28 Jul 2022, at 10:35, Julien Grall <jul...@xen.org> wrote: > > > > On 28/07/2022 08:57, Bertrand Marquis wrote: >> Hi Julien, > > Hi Bertrand, > >>> On 27 Jul 2022, at 16:46, Julien Grall <jul...@xen.org> wrote: >>> >>> Hi Xenia, >>> >>> On 27/07/2022 16:32, Xenia Ragiadakou wrote: >>>> Remove unused macro atomic_xchg(). >>>> Signed-off-by: Xenia Ragiadakou <burzalod...@gmail.com> >>>> --- >>>> xen/arch/arm/include/asm/atomic.h | 2 -- >>>> 1 file changed, 2 deletions(-) >>>> diff --git a/xen/arch/arm/include/asm/atomic.h >>>> b/xen/arch/arm/include/asm/atomic.h >>>> index f5ef744b4b..a2dc125291 100644 >>>> --- a/xen/arch/arm/include/asm/atomic.h >>>> +++ b/xen/arch/arm/include/asm/atomic.h >>>> @@ -223,8 +223,6 @@ static inline int atomic_add_unless(atomic_t *v, int >>>> a, int u) >>>> return __atomic_add_unless(v, a, u); >>>> } >>>> -#define atomic_xchg(v, new) (xchg(&((v)->counter), new)) >>>> - >>> >>> While I agree this is unused today, the wrapper is quite trivial and part >>> of the generic API (x86 also provides one). So I am not in favor of >>> removing it just to please MISRA. >>> >>> That said, if Bertrand and Stefano agrees with removing it then you should >>> also remove the x86 version to avoid inconsistency. >> I think we can keep this and maybe add a comment on top to document a known >> violation: >> /* TODO: MISRA_VIOLATION 2.5 */ > > While I am fine with this goal of the comment (i.e. indicating where Xen is > not MISRA compliant), I think this is one place where I would rather not want > one because it can get stale if someones decide to use the function.
I think the one doing that will have to update the comment otherwise we will never manage to have an analysis without findings. Having those kind of comments in the code for violation also means that they must be updated if the violation is solved. Maybe we will need a run ignoring those to identify possible violations which are not violations anymore but this might be hard to do. Cheers Bertrand > > Cheers, > > -- > Julien Grall
