On 2019-07-22, at 14:24:38 +0200, Paul Kocialkowski wrote:
> On Mon 22 Jul 19, 12:12, Jeremy Sowden wrote:
> > On 2019-07-22, at 11:36:51 +0530, Nishka Dasgupta wrote:
> > > Typecast as bool the return value of cedrus_find_format in
> > > cedrus_check_format as the return value of cedrus_check_format is
> > > always treated like a boolean value.
> > >
> > > Signed-off-by: Nishka Dasgupta <nishkadg.linux@gmail.com>
> > > ---
> > > Changes in v2:
> > > - Add !! to the returned pointer to ensure that the return value
> > >   is always either true or false, and never a non-zero value other
> > >   than true.
> > >
> > >  drivers/staging/media/sunxi/cedrus/cedrus_video.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/staging/media/sunxi/cedrus/cedrus_video.c b/drivers/staging/media/sunxi/cedrus/cedrus_video.c
> > > index e2b530b1a956..b731745f21f8 100644
> > > --- a/drivers/staging/media/sunxi/cedrus/cedrus_video.c
> > > +++ b/drivers/staging/media/sunxi/cedrus/cedrus_video.c
> > > @@ -86,7 +86,7 @@ static struct cedrus_format *cedrus_find_format(u32 pixelformat, u32 directions,
> > >  static bool cedrus_check_format(u32 pixelformat, u32 directions,
> > >  				unsigned int capabilities)
> > >  {
> > > -	return cedrus_find_format(pixelformat, directions, capabilities);
> > > +	return !!(bool)cedrus_find_format(pixelformat, directions, capabilities);
> > >  }
> >
> > I think the original was fine.  The return value of
> > cedrus_find_format will be automatically converted to bool before
> > being returned from cedrus_check_format since that is the
> > return-type of the function, and the result of converting any
> > non-zero value to bool is 1.
>
> Okay I was a bit unsure about that and wanted to play it on the safe
> side without really looking it up, but that gave me the occasion to
> verify.
>
> From what I could find (from my GNU system's
> /usr/include/unistring/stdbool.h):
>
>    Limitations of this substitute, when used in a C89 environment:
>
>        - In C99, casts and automatic conversions to '_Bool' or 'bool'
>          are performed in such a way that every nonzero value gets
>          converted to 'true', and zero gets converted to 'false'.
>          This doesn't work with this substitute.  With this
>          substitute, only the values 0 and 1 give the expected result
>          when converted to _Bool' or 'bool'.
>
> So since the kernel is built for C89 (unless I'm mistaken), I don't
> think the compiler provides any guarantee about bool values being
> converted to 1 when they are non-zero.

Ick.  I checked the C99 standard and assumed that GCC would use the same
semantics.  I've just tested with "gcc-8 -std=gnu89" and it appears to
do the right thing:

  [azazel@ulthar:/space/azazel/tmp] $ cat test.c
  #include <stdbool.h>
  #include <stdio.h>

  int
  main (void)
    {
    char *p = "test";
    bool b = p;

    fprintf (stderr, "%p, %d\n", (void *) p, b);

    return 0;
    }
  [azazel@ulthar:/space/azazel/tmp] $ gcc-8 -O2 -std=gnu89 -Wall -Wextra
  test.c -o test
  [azazel@ulthar:/space/azazel/tmp] $ ./test
  0x55d984e0e004, 1

> As a result, I think it's best to be careful.

Fair enough.

> However, I'm not sure I really see what cocinelle was unhappy about.
> You mentionned single-line functions, but I don't see how that can be
> a problem.
>
> So in the end, I think we should keep the !! and drop the (bool) cast
> if there's no particular warning about it.
>
> What do you think?

Seems sensible.

J.