On Wed, 20 Mar 2024, Jan Beulich wrote:
> On 20.03.2024 09:50, Simone Ballarin wrote:
> > MISRA C:2012 Rule 17.1 states:
> > The features of `<stdarg.h>' shall not be used
> >
> > The Xen community wants to avoid using variadic functions except for
> > specific circumstances where it feels appropriate by strict code review.
> >
> > Add deviation for functions related to console output (printk and similar).
> >
> > Signed-off-by: Simone Ballarin <simone.balla...@bugseng.com>
> > ---
> > .../eclair_analysis/ECLAIR/deviations.ecl | 26 +++++++++++++++++++
> > docs/misra/deviations.rst | 5 ++++
> > 2 files changed, 31 insertions(+)
> >
> > diff --git a/automation/eclair_analysis/ECLAIR/deviations.ecl
> > b/automation/eclair_analysis/ECLAIR/deviations.ecl
> > index 9ac3ee4dfd..7c3559a3a0 100644
> > --- a/automation/eclair_analysis/ECLAIR/deviations.ecl
> > +++ b/automation/eclair_analysis/ECLAIR/deviations.ecl
> > @@ -382,6 +382,32 @@ explicit comment indicating the fallthrough intention
> > is present."
> > -config=MC3R1.R16.6,switch_clauses+={deliberate, "default(0)"}
> > -doc_end
> >
> > +#
> > +# Series 17.
> > +#
> > +
> > +-doc_begin="Functions related to console output are allowed to use the
> > variadic features provided by stdarg.h."
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(debugtrace_printk)&&kind(function))))"}
>
> This isn't concole output related.
Should we say "related to console output or tracing" ?
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(dt_dprintk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(printk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(guest_printk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(panic)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(dprintk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(gdprintk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(snprintf)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(scnprintf)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(xasprintf)&&kind(function))))"}
>
> These three and ...
>
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(mm_printk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(dbgp_printk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(elf_call_log_callback)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(vsnprintf)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(vscnprintf)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(xvasprintf)&&kind(function))))"}
>
> ... these three aren't either.
Maybe it is better to write it as:
"Functions related printk, logging and tracing are allowed..."
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(vprintk_common)&&kind(function))))"}
> > +-config=MC3R1.R17.1,reports+={deliberate,"any_area(^.*va_list.*$&&context(ancestor_or_self(name(guest_printk)&&kind(function))))"}
> > +-config=MC3R1.R17.1,macros+={hide , "^va_(arg|start|copy|end)$"}
> > +-doc_end
>
> Further, is there a reason xyzprintk() are all listed individually?
> Surely if we'd introduce a new flavor, we'd want that excluded too.
>
> Finally, {,g}dprintk() in neither of their incarnation use any of the
> va_* items listed. Why do they need mentioning here?
