On Tuesday, September 27, 2011 14:19:54 Guennadi Liakhovetski wrote:
> On Tue, 27 Sep 2011, Hans Verkuil wrote:
>
> > On Tuesday, September 27, 2011 13:00:24 Guennadi Liakhovetski wrote:
> > > Hi Hans
> > >
> > > On Tue, 27 Sep 2011, Hans Verkuil wrote:
> > >
> > > > On Thursday, September 08, 2011 09:45:15 Guennadi Liakhovetski wrote:
> > > > > A possibility to preallocate and initialise buffers of different sizes
> > > > > in V4L2 is required for an efficient implementation of a snapshot
> > > > > mode. This patch adds two new ioctl()s: VIDIOC_CREATE_BUFS and
> > > > > VIDIOC_PREPARE_BUF and defines respective data structures.
> > > > >
> > > > > Signed-off-by: Guennadi Liakhovetski <g.liakhovet...@gmx.de>
> > > > > ---
> > > > >
> > > > > v7: added the "experimental" comment, as suggested by Sakari - thanks.
> > > > >
> > > > > drivers/media/video/v4l2-compat-ioctl32.c | 67
> > > > > +++++++++++++++++++++++++---
> > > > > drivers/media/video/v4l2-ioctl.c | 29 ++++++++++++
> > > > > include/linux/videodev2.h | 17 +++++++
> > > > > include/media/v4l2-ioctl.h | 2 +
> > > > > 4 files changed, 107 insertions(+), 8 deletions(-)
> > >
> > > [snip]
> > >
> > > > > diff --git a/include/linux/videodev2.h b/include/linux/videodev2.h
> > > > > index a5359c6..6e87ea9 100644
> > > > > --- a/include/linux/videodev2.h
> > > > > +++ b/include/linux/videodev2.h
> > > > > @@ -653,6 +653,9 @@ struct v4l2_buffer {
> > > > > #define V4L2_BUF_FLAG_ERROR 0x0040
> > > > > #define V4L2_BUF_FLAG_TIMECODE 0x0100 /* timecode field is
> > > > > valid */
> > > > > #define V4L2_BUF_FLAG_INPUT 0x0200 /* input field is valid */
> > > > > +/* Cache handling flags */
> > > > > +#define V4L2_BUF_FLAG_NO_CACHE_INVALIDATE 0x0400
> > > > > +#define V4L2_BUF_FLAG_NO_CACHE_CLEAN 0x0800
> > > > >
> > > > > /*
> > > > > * O V E R L A Y P R E V I E W
> > > > > @@ -2098,6 +2101,15 @@ struct v4l2_dbg_chip_ident {
> > > > > __u32 revision; /* chip revision, chip specific */
> > > > > } __attribute__ ((packed));
> > > > >
> > > > > +/* VIDIOC_CREATE_BUFS */
> > > > > +struct v4l2_create_buffers {
> > > > > + __u32 index; /* output: buffers
> > > > > index...index + count - 1 have been created */
> > > > > + __u32 count;
> > > > > + enum v4l2_memory memory;
> > > > > + struct v4l2_format format; /* "type" is used
> > > > > always, the rest if sizeimage == 0 */
> > > > > + __u32 reserved[8];
> > > > > +};
> > > > > +
> > > > > /*
> > > > > * I O C T L C O D E S F O R V I D E O D E V I C E S
> > > > > *
> > > > > @@ -2188,6 +2200,11 @@ struct v4l2_dbg_chip_ident {
> > > > > #define VIDIOC_SUBSCRIBE_EVENT _IOW('V', 90, struct
> > > > > v4l2_event_subscription)
> > > > > #define VIDIOC_UNSUBSCRIBE_EVENT _IOW('V', 91, struct
> > > > > v4l2_event_subscription)
> > > > >
> > > > > +/* Experimental, the below two ioctls may change over the next
> > > > > couple of kernel
> > > > > + versions */
> > > > > +#define VIDIOC_CREATE_BUFS _IOWR('V', 92, struct
> > > > > v4l2_create_buffers)
> > > > > +#define VIDIOC_PREPARE_BUF _IOW('V', 93, struct v4l2_buffer)
> > > >
> > > > I think I would prefer _IOWR here. QBUF etc. also use IOWR and you
> > > > never know
> > > > what you might return in the future. At the very least using IOWR
> > > > allows us
> > > > to update the state field, which would be a perfectly reasonable thing
> > > > to do.
> > >
> > > Sorry, which state field do you mean? We have already marked these
> > > ioctl()
> > > as experimental, isn't this enough?
> >
> > The state field in the v4l2_buffer argument.
>
> Is this a new field? I seem to remember some discussion to replace one of
> existing unused fields in a user-exposed struct (ioctl() argument) with a
> different one. Is this what you're referring to? Can you point me out to
> the patch?
Sorry, my fault. I confused vb2_buffer 'state' with v4l2_buffer 'flags'.
What happens in the case of QBUF is that v4l2_buffer is returned and the flags
field is set to FLAG_QUEUED. Something similar can be done if you have a IOWR
PREPARE_BUF: let it set FLAG_QUEUED. That makes it consistent with the QBUF
behavior as well (I always like consistent APIs).
I'm not sure whether we want to add a FLAG_PREPARED as well. I don't think it
has much value, although it would be a trivial addition.
Regards,
Hans
>
> > The experimental tag allows for changes, that's true. So this is my proposed
> > change :-)
>
> Thanks
> Guennadi
> ---
> Guennadi Liakhovetski, Ph.D.
> Freelance Open-Source Software Developer
> http://www.open-technology.de/
>
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
