Re: [XEN PATCH v3 03/16] misra: add deviations for direct inclusion guards

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Mon, 11 Mar 2024, Simone Ballarin wrote:
> On 11/03/24 14:56, Jan Beulich wrote:
> > On 11.03.2024 13:00, Simone Ballarin wrote:
> > > On 11/03/24 11:08, Jan Beulich wrote:
> > > > On 11.03.2024 09:59, Simone Ballarin wrote:
> > > > > --- a/xen/arch/arm/include/asm/hypercall.h
> > > > > +++ b/xen/arch/arm/include/asm/hypercall.h
> > > > > @@ -1,3 +1,4 @@
> > > > > +/* SAF-5-safe direct inclusion guard before */
> > > > >    #ifndef __XEN_HYPERCALL_H__
> > > > >    #error "asm/hypercall.h should not be included directly - include
> > > > > xen/hypercall.h instead"
> > > > >    #endif
> > > > > --- a/xen/arch/x86/include/asm/hypercall.h
> > > > > +++ b/xen/arch/x86/include/asm/hypercall.h
> > > > > @@ -2,6 +2,7 @@
> > > > >     * asm-x86/hypercall.h
> > > > >     */
> > > > >    +/* SAF-5-safe direct inclusion guard before */
> > > > >    #ifndef __XEN_HYPERCALL_H__
> > > > >    #error "asm/hypercall.h should not be included directly - include
> > > > > xen/hypercall.h instead"
> > > > >    #endif
> > > > 
> > > > Iirc it was said that this way checking for correct guards is suppressed
> > > > altogether in Eclair, which is not what we want. Can you clarify this,
> > > > please?
> > > > 
> > > 
> > > My first change was moving this check inside the guard.
> > > You commented my patch saying that this would be an error because someone
> > > can
> > > include it directly if it has already been included indirectly.
> > > I replied telling that this was the case also before the change.
> > > You agreed with me, and we decided that the correct thing would be fixing
> > > the
> > > check and not apply my temporary change to address the finding.
> > > 
> > > Considering that the code should be amended, a SAF deviation seems to me
> > > the most appropriate way for suppressing these findings.
> > 
> > Since I don't feel your reply addresses my question, asking differently:
> > With
> > your change in place, will failure to have proper guards (later) in these
> > headers still be reported by Eclair?
> > 
> > Jan
> > 
> 
> No, if you put something between the check and the guard,
> no violation will be reported.

>From this email exchange I cannot under if Jan is OK with this patch or
not.