Re: [PATCH v3 5/8] media: v4l: subdev: Make stream argument optional in state access functions

[Laurent Pinchart <laurent.pinchart@ideasonboard.com>]

Hi Sakari,

On Wed, Oct 25, 2023 at 10:19:15AM +0000, Sakari Ailus wrote:
> On Wed, Oct 25, 2023 at 11:53:35AM +0200, Hans Verkuil wrote:
> > On 23/10/2023 19:44, Sakari Ailus wrote:
> > > The sub-device state access functions take three arguments: sub-device
> > > state, pad and stream. The stream is not relevant for the majority of
> > > drivers and having to specify 0 for the stream is considered a nuisance.
> > > 
> > > Provide a two-argument macros for these state access functions to cover
> > > the needs of stream-unaware users.
> > > 
> > > Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
> > > ---
> > >  drivers/media/v4l2-core/v4l2-subdev.c | 18 ++++++-------
> > >  include/media/v4l2-subdev.h           | 39 ++++++++++++++++++++-------
> > >  2 files changed, 39 insertions(+), 18 deletions(-)
> > > 
> > > diff --git a/drivers/media/v4l2-core/v4l2-subdev.c b/drivers/media/v4l2-core/v4l2-subdev.c
> > > index f0f0af48730f..e35226587244 100644
> > > --- a/drivers/media/v4l2-core/v4l2-subdev.c
> > > +++ b/drivers/media/v4l2-core/v4l2-subdev.c
> > > @@ -1671,8 +1671,8 @@ int v4l2_subdev_set_routing_with_fmt(struct v4l2_subdev *sd,
> > >  EXPORT_SYMBOL_GPL(v4l2_subdev_set_routing_with_fmt);
> > >  
> > >  struct v4l2_mbus_framefmt *
> > > -v4l2_subdev_state_get_format(struct v4l2_subdev_state *state, unsigned int pad,
> > > -			     u32 stream)
> > > +__v4l2_subdev_state_get_format(struct v4l2_subdev_state *state,
> > > +			       unsigned int pad, u32 stream)
> > >  {
> > >  	struct v4l2_subdev_stream_configs *stream_configs;
> > >  	unsigned int i;
> > > @@ -1702,11 +1702,11 @@ v4l2_subdev_state_get_format(struct v4l2_subdev_state *state, unsigned int pad,
> > >  
> > >  	return NULL;
> > >  }
> > > -EXPORT_SYMBOL_GPL(v4l2_subdev_state_get_format);
> > > +EXPORT_SYMBOL_GPL(__v4l2_subdev_state_get_format);
> > >  
> > >  struct v4l2_rect *
> > > -v4l2_subdev_state_get_crop(struct v4l2_subdev_state *state, unsigned int pad,
> > > -			   u32 stream)
> > > +__v4l2_subdev_state_get_crop(struct v4l2_subdev_state *state, unsigned int pad,
> > > +			     u32 stream)
> > >  {
> > >  	struct v4l2_subdev_stream_configs *stream_configs;
> > >  	unsigned int i;
> > > @@ -1736,11 +1736,11 @@ v4l2_subdev_state_get_crop(struct v4l2_subdev_state *state, unsigned int pad,
> > >  
> > >  	return NULL;
> > >  }
> > > -EXPORT_SYMBOL_GPL(v4l2_subdev_state_get_crop);
> > > +EXPORT_SYMBOL_GPL(__v4l2_subdev_state_get_crop);
> > >  
> > >  struct v4l2_rect *
> > > -v4l2_subdev_state_get_compose(struct v4l2_subdev_state *state, unsigned int pad,
> > > -			      u32 stream)
> > > +__v4l2_subdev_state_get_compose(struct v4l2_subdev_state *state,
> > > +				unsigned int pad, u32 stream)
> > >  {
> > >  	struct v4l2_subdev_stream_configs *stream_configs;
> > >  	unsigned int i;
> > > @@ -1770,7 +1770,7 @@ v4l2_subdev_state_get_compose(struct v4l2_subdev_state *state, unsigned int pad,
> > >  
> > >  	return NULL;
> > >  }
> > > -EXPORT_SYMBOL_GPL(v4l2_subdev_state_get_compose);
> > > +EXPORT_SYMBOL_GPL(__v4l2_subdev_state_get_compose);
> > >  
> > >  int v4l2_subdev_routing_find_opposite_end(const struct v4l2_subdev_krouting *routing,
> > >  					  u32 pad, u32 stream, u32 *other_pad,
> > > diff --git a/include/media/v4l2-subdev.h b/include/media/v4l2-subdev.h
> > > index 71cb35133b4a..12eb0fe746d7 100644
> > > --- a/include/media/v4l2-subdev.h
> > > +++ b/include/media/v4l2-subdev.h
> > > @@ -1545,7 +1545,7 @@ int v4l2_subdev_set_routing_with_fmt(struct v4l2_subdev *sd,
> > >   * v4l2_subdev_state_get_format() - Get pointer to a stream format
> > >   * @state: subdevice state
> > >   * @pad: pad id
> > > - * @stream: stream id
> > > + * @...: stream id (optional argument)
> > >   *
> > >   * This returns a pointer to &struct v4l2_mbus_framefmt for the given pad +
> > >   * stream in the subdev state.
> > > @@ -1553,15 +1553,22 @@ int v4l2_subdev_set_routing_with_fmt(struct v4l2_subdev *sd,
> > >   * For stream-unaware drivers the format for the corresponding pad is returned.
> > >   * If the pad does not exist, NULL is returned.
> > >   */
> > > +#define v4l2_subdev_state_get_format(state, pad, ...)         \
> > > +        __v4l2_subdev_state_get_format_ ## __VA_OPT__(stream) \
> > > +        (state, pad __VA_OPT__(,) __VA_ARGS__)
> > > +#define __v4l2_subdev_state_get_format_(state, pad)     \
> > > +        __v4l2_subdev_state_get_format(state, pad, 0)
> > > +#define __v4l2_subdev_state_get_format_stream(state, pad, stream)	\
> > > +        __v4l2_subdev_state_get_format(state, pad, stream)
> > 
> > Hmm, 'git grep __VA_OPT__' finds no use of this in the kernel. Are you sure this
> > is safe to use? Does clang support it? Do all gcc versions that the kernel support
> > implement it?
> > 
> > A 10+ for creativity, but I'm a bit concerned about it.
> 
> :-)
> 
> I noticed the problem, too, as the lkp bot failed compiling this on GCC
> 7.5. Indeed only GCC 8 supports __VA_OPT__ but the oldest GCC version
> supported for kernel compilation is 7.1. So __VA_OPT__ isn't... an option.
> 
> I thought this could be done using variable arguments, making most error
> handling take place at runtime (for GCC versions not supporting
> __VA_OPT__), but then Laurent figured this can be done with just
> __VA_ARGS__ which is part of C99.
> 
> The code would look like this, which is only slightly less understandable
> than with __VA_OPT__:
> 
> #define v4l2_subdev_state_get_format(...)				\
> 	__v4l2_subdev_state_get_format_var(__VA_ARGS__, , _2)(__VA_ARGS__)

How about

 	__v4l2_subdev_state_get_format_var(__VA_ARGS__, _stream, _pad)(__VA_ARGS__)

? The internal macro and function names would then be more explicit.

> #define __v4l2_subdev_state_get_format_var(_1, _2, _3, NUM, ...)	\
> 	__v4l2_subdev_state_get_format ## NUM
> #define __v4l2_subdev_state_get_format_2(state, pad) \
> 	__v4l2_subdev_state_get_format(state, pad, 0)
> struct v4l2_mbus_framefmt *
> __v4l2_subdev_state_get_format(struct v4l2_subdev_state *state,
> 			       unsigned int pad, u32 stream);
> 
> > And if there are only a few drivers that need to give a specific stream,
> > wouldn't it be easier to just have a v4l2_subdev_state_get_format_stream
> > function and a static inline v4l2_subdev_state_get_format() that just calls
> > the _stream version with 0?
> 
> I'm trying to shorten function names here: they're overly long. Also I'd
> like to see a single set of API functions to access sub-device state
> information irrespective of how it is stored. I originally proposed just
> specifying stream zero but both Laurent and Tomi opined against it.

-- 
Regards,

Laurent Pinchart