diff mbox series

media: v4l: Add packed YUV444 24bpp pixel format

Message ID 1562166911-27454-1-git-send-email-mirela.rabulea@nxp.com (mailing list archive)
State Superseded, archived
Headers
Series media: v4l: Add packed YUV444 24bpp pixel format |

Commit Message

Mirela Rabulea July 3, 2019, 3:15 p.m. UTC 
  The added format is V4L2_PIX_FMT_YUV24, this is a packed
YUV 4:4:4 format, with 8 bits for each component, 24 bits
per sample.

This format is used by the i.MX 8QuadMax and i.MX 8DualXPlus/8QuadXPlus
JPEG encoder/decoder.

Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
---
 Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst | 37 +++++++++++++++++++++-
 drivers/media/v4l2-core/v4l2-ioctl.c               |  1 +
 include/uapi/linux/videodev2.h                     |  1 +
 3 files changed, 38 insertions(+), 1 deletion(-)

Comments

Paul Kocialkowski July 11, 2019, 8:18 a.m. UTC | #1 
Hi,

On Wed 03 Jul 19, 18:15, Mirela Rabulea wrote:
> The added format is V4L2_PIX_FMT_YUV24, this is a packed
> YUV 4:4:4 format, with 8 bits for each component, 24 bits
> per sample.
> 
> This format is used by the i.MX 8QuadMax and i.MX 8DualXPlus/8QuadXPlus
> JPEG encoder/decoder.

So this format is not aligned to 32-bit words at all and we can expect
to see cases where a single 32-bit word contains data for two pixels?

Nothing wrong with that, just checking whether I understood this right :)

Cheers,

Paul

> Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
> ---
>  Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst | 37 +++++++++++++++++++++-
>  drivers/media/v4l2-core/v4l2-ioctl.c               |  1 +
>  include/uapi/linux/videodev2.h                     |  1 +
>  3 files changed, 38 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst b/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
> index 41b60fa..5750ff6 100644
> --- a/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
> +++ b/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
> @@ -17,7 +17,8 @@ Description
>  ===========
>  
>  Similar to the packed RGB formats these formats store the Y, Cb and Cr
> -component of each pixel in one 16 or 32 bit word.
> +component of each pixel next to each other in memory. They occupy 16, 24 or 32
> +bits per pixel.
>  
>  
>  .. raw:: latex
> @@ -157,6 +158,40 @@ component of each pixel in one 16 or 32 bit word.
>  
>        -  :cspan:`15`
>  
> +    * .. _V4L2-PIX-FMT-YUV24:
> +
> +      - ``V4L2_PIX_FMT_YUV24``
> +      - 'YUV3'
> +
> +      - Y'\ :sub:`7`
> +      - Y'\ :sub:`6`
> +      - Y'\ :sub:`5`
> +      - Y'\ :sub:`4`
> +      - Y'\ :sub:`3`
> +      - Y'\ :sub:`2`
> +      - Y'\ :sub:`1`
> +      - Y'\ :sub:`0`
> +
> +      - Cb\ :sub:`7`
> +      - Cb\ :sub:`6`
> +      - Cb\ :sub:`5`
> +      - Cb\ :sub:`4`
> +      - Cb\ :sub:`3`
> +      - Cb\ :sub:`2`
> +      - Cb\ :sub:`1`
> +      - Cb\ :sub:`0`
> +
> +      - Cr\ :sub:`7`
> +      - Cr\ :sub:`6`
> +      - Cr\ :sub:`5`
> +      - Cr\ :sub:`4`
> +      - Cr\ :sub:`3`
> +      - Cr\ :sub:`2`
> +      - Cr\ :sub:`1`
> +      - Cr\ :sub:`0`
> +
> +      -  :cspan:`7`
> +
>      * .. _V4L2-PIX-FMT-YUV32:
>  
>        - ``V4L2_PIX_FMT_YUV32``
> diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c
> index b1f4b99..9bb6579 100644
> --- a/drivers/media/v4l2-core/v4l2-ioctl.c
> +++ b/drivers/media/v4l2-core/v4l2-ioctl.c
> @@ -1215,6 +1215,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)
>  	case V4L2_PIX_FMT_YUV444:	descr = "16-bit A/XYUV 4-4-4-4"; break;
>  	case V4L2_PIX_FMT_YUV555:	descr = "16-bit A/XYUV 1-5-5-5"; break;
>  	case V4L2_PIX_FMT_YUV565:	descr = "16-bit YUV 5-6-5"; break;
> +	case V4L2_PIX_FMT_YUV24:	descr = "24-bit YUV 4:4:4 8-8-8"; break;
>  	case V4L2_PIX_FMT_YUV32:	descr = "32-bit A/XYUV 8-8-8-8"; break;
>  	case V4L2_PIX_FMT_AYUV32:	descr = "32-bit AYUV 8-8-8-8"; break;
>  	case V4L2_PIX_FMT_XYUV32:	descr = "32-bit XYUV 8-8-8-8"; break;
> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
> index 9d9705c..752dd33 100644
> --- a/include/uapi/linux/videodev2.h
> +++ b/include/uapi/linux/videodev2.h
> @@ -577,6 +577,7 @@ struct v4l2_pix_format {
>  #define V4L2_PIX_FMT_YUV444  v4l2_fourcc('Y', '4', '4', '4') /* 16  xxxxyyyy uuuuvvvv */
>  #define V4L2_PIX_FMT_YUV555  v4l2_fourcc('Y', 'U', 'V', 'O') /* 16  YUV-5-5-5     */
>  #define V4L2_PIX_FMT_YUV565  v4l2_fourcc('Y', 'U', 'V', 'P') /* 16  YUV-5-6-5     */
> +#define V4L2_PIX_FMT_YUV24   v4l2_fourcc('Y', 'U', 'V', '3') /* 24  YUV-8-8-8     */
>  #define V4L2_PIX_FMT_YUV32   v4l2_fourcc('Y', 'U', 'V', '4') /* 32  YUV-8-8-8-8   */
>  #define V4L2_PIX_FMT_AYUV32  v4l2_fourcc('A', 'Y', 'U', 'V') /* 32  AYUV-8-8-8-8  */
>  #define V4L2_PIX_FMT_XYUV32  v4l2_fourcc('X', 'Y', 'U', 'V') /* 32  XYUV-8-8-8-8  */
> -- 
> 2.7.4
>
Mirela Rabulea July 11, 2019, 1:57 p.m. UTC | #2 
On Jo, 2019-07-11 at 10:18 +0200, Paul Kocialkowski wrote:
> Caution: EXT Email
> 
> Hi,
> 
> On Wed 03 Jul 19, 18:15, Mirela Rabulea wrote:
> > 
> > The added format is V4L2_PIX_FMT_YUV24, this is a packed
> > YUV 4:4:4 format, with 8 bits for each component, 24 bits
> > per sample.
> > 
> > This format is used by the i.MX 8QuadMax and i.MX
> > 8DualXPlus/8QuadXPlus
> > JPEG encoder/decoder.
> So this format is not aligned to 32-bit words at all and we can
> expect
> to see cases where a single 32-bit word contains data for two pixels?
> 
> Nothing wrong with that, just checking whether I understood this
> right :)
> 

Hi Paul,
yes, your understanding is correct.

Regards,
Mirela
Paul Kocialkowski July 12, 2019, 9:21 a.m. UTC | #3 
Hi,

On Thu 11 Jul 19, 13:57, Mirela Rabulea wrote:
> On Jo, 2019-07-11 at 10:18 +0200, Paul Kocialkowski wrote:
> > Caution: EXT Email
> > 
> > Hi,
> > 
> > On Wed 03 Jul 19, 18:15, Mirela Rabulea wrote:
> > > 
> > > The added format is V4L2_PIX_FMT_YUV24, this is a packed
> > > YUV 4:4:4 format, with 8 bits for each component, 24 bits
> > > per sample.
> > > 
> > > This format is used by the i.MX 8QuadMax and i.MX
> > > 8DualXPlus/8QuadXPlus
> > > JPEG encoder/decoder.
> > So this format is not aligned to 32-bit words at all and we can
> > expect
> > to see cases where a single 32-bit word contains data for two pixels?
> > 
> > Nothing wrong with that, just checking whether I understood this
> > right :)
> > 
> 
> Hi Paul,
> yes, your understanding is correct.

Out of curiosity, is the JPEG block assmiliated to (one of) the Hantro VPUs
or is it a totally different and unrelated hardware block?

Anyway the change looks good to me:
Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>

Cheers,

Paul
Mirela Rabulea July 12, 2019, 1:02 p.m. UTC | #4 
On Vi, 2019-07-12 at 11:21 +0200, paul.kocialkowski@bootlin.com wrote:
> Hi,
> 
> On Thu 11 Jul 19, 13:57, Mirela Rabulea wrote:
> > 
> > On Jo, 2019-07-11 at 10:18 +0200, Paul Kocialkowski wrote:
> > > 
> > > Caution: EXT Email
> > > 
> > > Hi,
> > > 
> > > On Wed 03 Jul 19, 18:15, Mirela Rabulea wrote:
> > > > 
> > > > 
> > > > The added format is V4L2_PIX_FMT_YUV24, this is a packed
> > > > YUV 4:4:4 format, with 8 bits for each component, 24 bits
> > > > per sample.
> > > > 
> > > > This format is used by the i.MX 8QuadMax and i.MX
> > > > 8DualXPlus/8QuadXPlus
> > > > JPEG encoder/decoder.
> > > So this format is not aligned to 32-bit words at all and we can
> > > expect
> > > to see cases where a single 32-bit word contains data for two
> > > pixels?
> > > 
> > > Nothing wrong with that, just checking whether I understood this
> > > right :)
> > > 
> > Hi Paul,
> > yes, your understanding is correct.
> Out of curiosity, is the JPEG block assmiliated to (one of) the
> Hantro VPUs
> or is it a totally different and unrelated hardware block?
> 
> Anyway the change looks good to me:
> Reviewed-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
> 
> Cheers,
> 
> Paul

Hi Paul,
thanks for taking the time to review.
The JPEG decoder is a standalone core & wrapper, not assimilated to the
VPU. It is described in the reference manual here:
https://www.nxp.com/docs/en/reference-manual/IMX8DQXPRM.pdf
in chapters "15.6.3 JPEG Decoder (JPEGDEC)" and "15.6.5 JPEG Decoder
Wrapper".
The JPEG core is from a third party, and the wrapper from NXP.

Regards,
Mirela
Hans Verkuil Aug. 13, 2019, 6:54 a.m. UTC | #5 
Hi Mirela,

On 7/3/19 5:15 PM, Mirela Rabulea wrote:
> The added format is V4L2_PIX_FMT_YUV24, this is a packed
> YUV 4:4:4 format, with 8 bits for each component, 24 bits
> per sample.
> 
> This format is used by the i.MX 8QuadMax and i.MX 8DualXPlus/8QuadXPlus
> JPEG encoder/decoder.

This patch looks good, but it can only be merged if we also have a driver
that uses it. This to avoid having lots of pixelformat defines that nobody
is using in the kernel.

Are there plans to upstream this JPEG en/decoder?

Regards,

	Hans

> 
> Signed-off-by: Mirela Rabulea <mirela.rabulea@nxp.com>
> ---
>  Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst | 37 +++++++++++++++++++++-
>  drivers/media/v4l2-core/v4l2-ioctl.c               |  1 +
>  include/uapi/linux/videodev2.h                     |  1 +
>  3 files changed, 38 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst b/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
> index 41b60fa..5750ff6 100644
> --- a/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
> +++ b/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
> @@ -17,7 +17,8 @@ Description
>  ===========
>  
>  Similar to the packed RGB formats these formats store the Y, Cb and Cr
> -component of each pixel in one 16 or 32 bit word.
> +component of each pixel next to each other in memory. They occupy 16, 24 or 32
> +bits per pixel.
>  
>  
>  .. raw:: latex
> @@ -157,6 +158,40 @@ component of each pixel in one 16 or 32 bit word.
>  
>        -  :cspan:`15`
>  
> +    * .. _V4L2-PIX-FMT-YUV24:
> +
> +      - ``V4L2_PIX_FMT_YUV24``
> +      - 'YUV3'
> +
> +      - Y'\ :sub:`7`
> +      - Y'\ :sub:`6`
> +      - Y'\ :sub:`5`
> +      - Y'\ :sub:`4`
> +      - Y'\ :sub:`3`
> +      - Y'\ :sub:`2`
> +      - Y'\ :sub:`1`
> +      - Y'\ :sub:`0`
> +
> +      - Cb\ :sub:`7`
> +      - Cb\ :sub:`6`
> +      - Cb\ :sub:`5`
> +      - Cb\ :sub:`4`
> +      - Cb\ :sub:`3`
> +      - Cb\ :sub:`2`
> +      - Cb\ :sub:`1`
> +      - Cb\ :sub:`0`
> +
> +      - Cr\ :sub:`7`
> +      - Cr\ :sub:`6`
> +      - Cr\ :sub:`5`
> +      - Cr\ :sub:`4`
> +      - Cr\ :sub:`3`
> +      - Cr\ :sub:`2`
> +      - Cr\ :sub:`1`
> +      - Cr\ :sub:`0`
> +
> +      -  :cspan:`7`
> +
>      * .. _V4L2-PIX-FMT-YUV32:
>  
>        - ``V4L2_PIX_FMT_YUV32``
> diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c
> index b1f4b99..9bb6579 100644
> --- a/drivers/media/v4l2-core/v4l2-ioctl.c
> +++ b/drivers/media/v4l2-core/v4l2-ioctl.c
> @@ -1215,6 +1215,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)
>  	case V4L2_PIX_FMT_YUV444:	descr = "16-bit A/XYUV 4-4-4-4"; break;
>  	case V4L2_PIX_FMT_YUV555:	descr = "16-bit A/XYUV 1-5-5-5"; break;
>  	case V4L2_PIX_FMT_YUV565:	descr = "16-bit YUV 5-6-5"; break;
> +	case V4L2_PIX_FMT_YUV24:	descr = "24-bit YUV 4:4:4 8-8-8"; break;
>  	case V4L2_PIX_FMT_YUV32:	descr = "32-bit A/XYUV 8-8-8-8"; break;
>  	case V4L2_PIX_FMT_AYUV32:	descr = "32-bit AYUV 8-8-8-8"; break;
>  	case V4L2_PIX_FMT_XYUV32:	descr = "32-bit XYUV 8-8-8-8"; break;
> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
> index 9d9705c..752dd33 100644
> --- a/include/uapi/linux/videodev2.h
> +++ b/include/uapi/linux/videodev2.h
> @@ -577,6 +577,7 @@ struct v4l2_pix_format {
>  #define V4L2_PIX_FMT_YUV444  v4l2_fourcc('Y', '4', '4', '4') /* 16  xxxxyyyy uuuuvvvv */
>  #define V4L2_PIX_FMT_YUV555  v4l2_fourcc('Y', 'U', 'V', 'O') /* 16  YUV-5-5-5     */
>  #define V4L2_PIX_FMT_YUV565  v4l2_fourcc('Y', 'U', 'V', 'P') /* 16  YUV-5-6-5     */
> +#define V4L2_PIX_FMT_YUV24   v4l2_fourcc('Y', 'U', 'V', '3') /* 24  YUV-8-8-8     */
>  #define V4L2_PIX_FMT_YUV32   v4l2_fourcc('Y', 'U', 'V', '4') /* 32  YUV-8-8-8-8   */
>  #define V4L2_PIX_FMT_AYUV32  v4l2_fourcc('A', 'Y', 'U', 'V') /* 32  AYUV-8-8-8-8  */
>  #define V4L2_PIX_FMT_XYUV32  v4l2_fourcc('X', 'Y', 'U', 'V') /* 32  XYUV-8-8-8-8  */
>
Mirela Rabulea Aug. 22, 2019, 3 p.m. UTC | #6 
Hi Hans,

On Ma, 2019-08-13 at 08:54 +0200, Hans Verkuil wrote:
> 
> This patch looks good, but it can only be merged if we also have a
> driver
> that uses it. This to avoid having lots of pixelformat defines that
> nobody
> is using in the kernel.
> 
> Are there plans to upstream this JPEG en/decoder?
> 

Apologies for the late response, I was OOO.
Yes, I have plans to send the JPEG encoder/decoder driver upstream, but
I still have failures from v4l2-compliance test. It will take a while
until I will manage to fix those. When I will be ready for upstream, It
would help if this patch would be upstream already, but if it won't, I
can remove the RGB24 support from the driver, and send it later,
together with this patch.

Regards,
Mirela
diff mbox series

Patch

diff --git a/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst b/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
index 41b60fa..5750ff6 100644
--- a/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
+++ b/Documentation/media/uapi/v4l/pixfmt-packed-yuv.rst
@@ -17,7 +17,8 @@  Description
 ===========
 
 Similar to the packed RGB formats these formats store the Y, Cb and Cr
-component of each pixel in one 16 or 32 bit word.
+component of each pixel next to each other in memory. They occupy 16, 24 or 32
+bits per pixel.
 
 
 .. raw:: latex
@@ -157,6 +158,40 @@  component of each pixel in one 16 or 32 bit word.
 
       -  :cspan:`15`
 
+    * .. _V4L2-PIX-FMT-YUV24:
+
+      - ``V4L2_PIX_FMT_YUV24``
+      - 'YUV3'
+
+      - Y'\ :sub:`7`
+      - Y'\ :sub:`6`
+      - Y'\ :sub:`5`
+      - Y'\ :sub:`4`
+      - Y'\ :sub:`3`
+      - Y'\ :sub:`2`
+      - Y'\ :sub:`1`
+      - Y'\ :sub:`0`
+
+      - Cb\ :sub:`7`
+      - Cb\ :sub:`6`
+      - Cb\ :sub:`5`
+      - Cb\ :sub:`4`
+      - Cb\ :sub:`3`
+      - Cb\ :sub:`2`
+      - Cb\ :sub:`1`
+      - Cb\ :sub:`0`
+
+      - Cr\ :sub:`7`
+      - Cr\ :sub:`6`
+      - Cr\ :sub:`5`
+      - Cr\ :sub:`4`
+      - Cr\ :sub:`3`
+      - Cr\ :sub:`2`
+      - Cr\ :sub:`1`
+      - Cr\ :sub:`0`
+
+      -  :cspan:`7`
+
     * .. _V4L2-PIX-FMT-YUV32:
 
       - ``V4L2_PIX_FMT_YUV32``
diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c
index b1f4b99..9bb6579 100644
--- a/drivers/media/v4l2-core/v4l2-ioctl.c
+++ b/drivers/media/v4l2-core/v4l2-ioctl.c
@@ -1215,6 +1215,7 @@  static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)
 	case V4L2_PIX_FMT_YUV444:	descr = "16-bit A/XYUV 4-4-4-4"; break;
 	case V4L2_PIX_FMT_YUV555:	descr = "16-bit A/XYUV 1-5-5-5"; break;
 	case V4L2_PIX_FMT_YUV565:	descr = "16-bit YUV 5-6-5"; break;
+	case V4L2_PIX_FMT_YUV24:	descr = "24-bit YUV 4:4:4 8-8-8"; break;
 	case V4L2_PIX_FMT_YUV32:	descr = "32-bit A/XYUV 8-8-8-8"; break;
 	case V4L2_PIX_FMT_AYUV32:	descr = "32-bit AYUV 8-8-8-8"; break;
 	case V4L2_PIX_FMT_XYUV32:	descr = "32-bit XYUV 8-8-8-8"; break;
diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
index 9d9705c..752dd33 100644
--- a/include/uapi/linux/videodev2.h
+++ b/include/uapi/linux/videodev2.h
@@ -577,6 +577,7 @@  struct v4l2_pix_format {
 #define V4L2_PIX_FMT_YUV444  v4l2_fourcc('Y', '4', '4', '4') /* 16  xxxxyyyy uuuuvvvv */
 #define V4L2_PIX_FMT_YUV555  v4l2_fourcc('Y', 'U', 'V', 'O') /* 16  YUV-5-5-5     */
 #define V4L2_PIX_FMT_YUV565  v4l2_fourcc('Y', 'U', 'V', 'P') /* 16  YUV-5-6-5     */
+#define V4L2_PIX_FMT_YUV24   v4l2_fourcc('Y', 'U', 'V', '3') /* 24  YUV-8-8-8     */
 #define V4L2_PIX_FMT_YUV32   v4l2_fourcc('Y', 'U', 'V', '4') /* 32  YUV-8-8-8-8   */
 #define V4L2_PIX_FMT_AYUV32  v4l2_fourcc('A', 'Y', 'U', 'V') /* 32  AYUV-8-8-8-8  */
 #define V4L2_PIX_FMT_XYUV32  v4l2_fourcc('X', 'Y', 'U', 'V') /* 32  XYUV-8-8-8-8  */