Re: [XEN PATCH v2] docs/misra: add asm-offset.c to exclude-list

[Julien Grall <julien@xxxxxxx>]

Hi Jan,

On 09/02/2024 11:46, Jan Beulich wrote:
> On 09.02.2024 10:40, Julien Grall wrote:
> > Replying on the v2 as well.
> And answering here despite the respective question was raised on the
> v1 thread: I'm certainly okay with the more detailed commit message.
Ah yes. Sorry, I replied to v1 first and then realized it may have been 
easier to comment on v2.
> A few nits, though:
>
> > On 08/02/2024 15:56, Jan Beulich wrote:
> > > On 08.02.2024 16:50, Nicola Vetrini wrote:
> > > > These files contain several deliberate violations of MISRA C rules and
> > > > they are not linked in the final Xen binary, therefore they can be exempted
> > > > from MISRA compliance.
> > I'd like the commit message to be expanded a little bit to explain which
> > MISRA rules are a problem. This helped me to understand why we excluded
> > rather than fixed.
> >
> > Base on the previous discussion, I would suggest:
> >
> > These files contain several deliberate violation of MISRA C rules such as:
> violations
>
> >     * R20.12 on Arm for macros DEFINE and OFFSET, where the second
> > argument of OFFSET is a macro and is used as a normal parameter and a
> > stringification operand.
> Is this really for Arm only?
I don't exactly know. I took Nicola's comment and massage it for the 
commit message. I am assuming that this was also not exhaustive list, so 
my aim was to only provide some example.
Thinking of it, I don't see why it would only be a problem on Arm. I can 
drop the "on Arm".
> >     * R2.1 because the file is not linked That said it was decided to
> > deviate the rule itselfed to deviate that aspect).
> There look to be punctuation issues here. Also s/itselfed/itself/, and
> the duplicate "deviate" is also a little odd to read (maybe "deal with"
> or "address" in place of the 2nd instance).
Doh, indeed. This wants to be:

"R2.1 because the file is not linked. That said, it was decided to 
deviate the ruule itself to address that aspect."
> > The files are also not linked in the final Xen binary, therefore they
> > can be expempted from MISRA compliance.
> Looks to duplicate what the latter half of the 2nd bullet point has.
> If to be kept: s/expempted/exempted/.
I will remove.

> > > > --- a/docs/misra/exclude-list.json
> > > > +++ b/docs/misra/exclude-list.json
> > > > @@ -101,6 +101,10 @@
> > > >                "rel_path": "arch/x86/efi/check.c",
> > > >                "comment": "The resulting code is not included in the final Xen 
> > > > binary, ignore for now"
> > > >            },
> > > > +        {
> > > > +          "rel_path": "arch/*/*/asm-offsets.c",
> > > > +          "comment": "The resulting code is not included in the final Xen binary, 
> > > > ignore for now"
> > > > +        },
> > > >            {
> > > >                "rel_path": "common/coverage/*",
> > > >                "comment": "Files to support gcov, ignore for now"
> > > ... something looks odd with indentation; can probably be adjusted
> > > while committing.
> > I am happy to take care of both the commit message and the indentation
> > on commit.
> Okay, I'll leave that to you then.
Thanks. I will do it shortly.

Cheers,

--
Julien Grall