On Wed, 26 Jun 2024, Federico Serafini wrote:
> On 26/06/24 09:37, Oleksii wrote:
> > On Tue, 2024-06-25 at 18:59 -0700, Stefano Stabellini wrote:
> > > > +-doc_begin="The conversion from a function pointer to unsigned
> > > > long or (void *) does not lose any information, provided that the
> > > > target type has enough bits to store it."
> > > > +-config=MC3R1.R11.1,casts+={safe,
> > > > + "from(type(canonical(__function_pointer_types)))
> > > > + &&to(type(canonical(builtin(unsigned
> > > > long)||pointer(builtin(void)))))
> > > > + &&relation(definitely_preserves_value)"
> > > > +}
> > > > +-doc_end
> > >
> > > This one and the ones below are the important ones! I think we should
> > > have them in the tree as soon as possible ideally 4.19. I ask for
> > > a release-ack.
> > Just want to be sure that I understand deviations properly with this
> > example.
> >
> > If the deviation above is merged, then it would be safe from a MISRA
> > point of view to cast a function pointer to 'unsigned long' or 'void
> > *', and thereby MISRA won't complain about code with such conversions?
>
> Exactly, taking into account section 4.7 of GCC manual.
Yes. From a Xen release perspective, it will only affect the static
analysis jobs, making them report fewer violations. The reason why
those specifically are important is that they are significant changes
over the plain rule and they were already documented in rules.rst.
