Re: [PATCH] xen:add deviations for MISRA C 2012 Rule R5.2

[Alessandro Zucchelli <alessandro.zucchelli@xxxxxxxxxxx>]

On 2024-11-14 03:23, Stefano Stabellini wrote:
> On Wed, 13 Nov 2024, Jan Beulich wrote:
> > On 13.11.2024 11:48, Alessandro Zucchelli wrote:
> > > At this link you can see all the violations of Rule 5.2:
> > >
> > > 
> > https://saas.eclairit.com:3787/fs/var/local/eclair/xen-project.ecdf/xen-project/people/bugseng/xen/ECLAIR_normal/40_characters/X86_64/8143097084/PROJECT.ecd;/by_service/MC3R1.R5.2.html
> >
> > Thank you. From a cursory look these all appear to be a result of the 
> > 40
> > chars limit we put in place (quite arbitrarily). That's not mentioned 
> > at
> > all ...
> >
> > > By deviating the two macros CHECK_NAME_ and DEFINE_COMPAT_HANDLE all the
> > > violations are addressed.
> > >
> > > On 2024-11-13 11:31, Jan Beulich wrote:
> > >> On 13.11.2024 09:41, Alessandro Zucchelli wrote:
> > >>> This addresses violations of MISRA C:2012 Rule 5.2 which states as
> > >>> following: Identifiers declared in the same scope and name space shall
> > >>> be distinct.
> > >>>
> > >>> This deviation addresses violations of Rule 5.2 arising from
> > >>> identifiers generated through token pasting macros CHECK_NAME_ and
> > >>> DEFINE_COMPAT_HANDLE.
> >
> > ... in the description.
> >
> > Together with the 5.4 patch having the same lack of context, I wonder
> > whether we shouldn't simply up that limit.
> Yes: if we up to the limit to 64 (another arbitrary number), can be 
> mark
> both 5.4 and 5.1 as clean? If so, I think we should do it right away.
Hi,
In doc/misra/rules.rst, the limit is documented as 40, but in the 
configuration
file automation/eclair_analysis/ECLAIR/toolchain.ecl, it is set to 63.

If you intend to proceed by increasing the limit, you may ignore this 
deviation
as well as my other deviation concerning Rule 5.4. In that case, please 
update
the documentation in doc/misra/rules.rst to reflect the new limit.

> > Or else, as suggested there,
> > to instead exclude such derived identifiers. After all the derived 
> > ones
> > will be distinct as long as what they're derived from is distinct.
> >
> > Finally - please don't top-post.
> >
> > Jan
> >
> > >> For each of the two, can you provide an example of where collisions
> > >> result? At least for the latter I can't even see how that would
> > >> work without the compiler complaining (i.e. the build breaking).
--
Alessandro Zucchelli, B.Sc.

Software Engineer, BUGSENG (https://bugseng.com)