Re: [cocci] Searching for special function implementations with SmPL

[Julia Lawall <julia.lawall@inria.fr>]

On Tue, 31 Jan 2023, Luca Coelho wrote:

> On Tue, 2023-01-31 at 17:08 +0100, Julia Lawall wrote:
> >
> > On Tue, 31 Jan 2023, Luca Coelho wrote:
> >
> > > On Mon, 2023-01-30 at 12:59 +0200, Luca Coelho wrote:
> > > > On Mon, 2023-01-30 at 11:47 +0100, Julia Lawall wrote:
> > > > > It doesn't find the file with the definition.  You may need to give more
> > > > > paths.
> > > > >
> > > > > I have the impression that the result is improved with
> > > > >
> > > > > -I drivers/gpu/drm/i915
> > > > >
> > > > > Maybe -I drivers/gpu/drm/i915/display is necessary also.  If you add the
> > > > > argument --relax-include-path then just -I drivers/gpu/drm/i915 may be
> > > > > sufficient.  With that option, it will just hunt around in and under the
> > > > > various include paths for a unique file that has the required name.
> > > >
> > > > Alright! That _does_ improve things! Now I can see lots of DPLL matches
> > > > in that file!
> > >
> > > These files I'm trying to parse are so complex that I keep bumping into
> > > other parsing errors.  Here's one of them:
> > >
> > > parse error
> > >  = File "./drivers/gpu/drm/i915/i915_drv.h", line 385, column 1, charpos = 9644
> > >   around = 'for',
> > >   whole content =       for ((id__) = 0; \
> > > badcount: 7
> > > bad: }
> > > bad:
> > > bad: /* Simple iterator over all initialised engines */
> > > bad: #define for_each_engine(engine__, dev_priv__, id__) \
> > > BAD:!!!!!       for ((id__) = 0; \
> > > bad:         (id__) < I915_NUM_ENGINES; \
> > > bad:         (id__)++) \
> > > bad:            for_each_if ((engine__) = (dev_priv__)->engine[(id__)])
> > > ERROR-RECOV: found sync end of #define, line 395
> > >
> > >
> > > I believe the parser is getting confused by an iterator inside an
> > > iterator's definition... I already tried to add them both with the
> > > "iterator name" directive, but it didn't help...
> > >
> > > Any idea what I can do about this?
> >
> > If you don't care about finding things in this code, you can just ignore
> > it.
>
> I don't, but this seems to be included from intel_display_power.c and
> I'm not getting the matches I expect in this C file... Maybe there are
> other hints in the debugging output?

If you know a function that should be matched, then you can check if there
are bad/BAD comments in front of that function.  bad/BAD comments
elsewhere don't matter.

You can also use the --debug option and see if the macro name that you
expect is getting collected.

You can use the --verbose-includes option to see if the file that contains
the definition of the macro of interest is actually found and included.

julia