linux-kernel - Re: [PATCH v13 02/13] media:Add nv12mt_8l128 and nv12mt_10be

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 2 Dec 2021 10:39:39 +0100
From:   Hans Verkuil <hverkuil-cisco@...all.nl>
To:     Ming Qian <ming.qian@....com>, mchehab@...nel.org,
        shawnguo@...nel.org, robh+dt@...nel.org, s.hauer@...gutronix.de
Cc:     kernel@...gutronix.de, festevam@...il.com, linux-imx@....com,
        aisheng.dong@....com, linux-media@...r.kernel.org,
        linux-kernel@...r.kernel.org, devicetree@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH v13 02/13] media:Add nv12mt_8l128 and nv12mt_10be_8l128
 video format.

On 30/11/2021 10:48, Ming Qian wrote:
> nv12mt_8l128 is 8-bit tiled nv12 format used by amphion decoder.
> nv12mt_10be_8l128 is 10-bit tiled format used by amphion decoder.
> The tile size is 8x128
> 
> Signed-off-by: Ming Qian <ming.qian@....com>
> Signed-off-by: Shijie Qin <shijie.qin@....com>
> Signed-off-by: Zhou Peng <eagle.zhou@....com>
> ---
>  .../userspace-api/media/v4l/pixfmt-yuv-planar.rst | 15 +++++++++++++++
>  drivers/media/v4l2-core/v4l2-ioctl.c              |  2 ++
>  include/uapi/linux/videodev2.h                    |  2 ++
>  3 files changed, 19 insertions(+)
> 
> diff --git a/Documentation/userspace-api/media/v4l/pixfmt-yuv-planar.rst b/Documentation/userspace-api/media/v4l/pixfmt-yuv-planar.rst
> index 3a09d93d405b..fc3baa2753ab 100644
> --- a/Documentation/userspace-api/media/v4l/pixfmt-yuv-planar.rst
> +++ b/Documentation/userspace-api/media/v4l/pixfmt-yuv-planar.rst
> @@ -257,6 +257,8 @@ of the luma plane.
>  .. _V4L2-PIX-FMT-NV12-4L4:
>  .. _V4L2-PIX-FMT-NV12-16L16:
>  .. _V4L2-PIX-FMT-NV12-32L32:
> +.. _V4L2_PIX_FMT_NV12MT_8L128:
> +.. _V4L2_PIX_FMT_NV12MT_10BE_8L128:
>  
>  Tiled NV12
>  ----------
> @@ -296,6 +298,19 @@ tiles linearly in memory. The line stride and image height must be
>  aligned to a multiple of 32. The layouts of the luma and chroma planes are
>  identical.
>  
> +``V4L2_PIX_FMT_NV12MT_8L128`` is similar to ``V4L2_PIX_FMT_NV12M`` but stores
> +pixel in 2D 8x128 tiles, and stores tiles linearly in memory.

pixel -> pixels (note: also wrong in the text V4L2_PIX_FMT_NV12_4L4/16L16/32L32)

Shouldn't this be called V4L2_PIX_FMT_NV12M_8L128? The 'MT' suffix seems to be specific
to macroblock tiles and not linear tiles.

> +The image height must be aligned to a multiple of 128.
> +The layouts of the luma and chroma planes are identical.
> +
> +``V4L2_PIX_FMT_NV12MT_10BE_8L128`` is similar to ``V4L2_PIX_FMT_NV12M`` but stores
> +10 bits pixel in 2D 8x128 tiles, and stores tiles linearly in memory.
> +the data is arranged at the big end.

at the big end -> in big endian order

I assume the 10 bit pixels are packed? So 5 bytes contain 4 10-bit pixels layout like
this (for luma):

byte 0: Y0(bits 9-2)
byte 1: Y0(bits 1-0) Y1(bits 9-4)
byte 2: Y1(bits 3-0) Y2(bits 9-6)
byte 3: Y2(bits 5-0) Y3(bits 9-8)
byte 4: Y3(bits 7-0)

> +The image height must be aligned to a multiple of 128.
> +The layouts of the luma and chroma planes are identical.
> +Note the tile size is 8bytes multiplied by 128 bytes,
> +it means that the low bits and high bits of one pixel may be in differnt tiles.

differnt -> different

> +
>  .. _nv12mt:
>  
>  .. kernel-figure:: nv12mt.svg
> diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c
> index 69b74d0e8a90..400eec0157a7 100644
> --- a/drivers/media/v4l2-core/v4l2-ioctl.c
> +++ b/drivers/media/v4l2-core/v4l2-ioctl.c
> @@ -1388,6 +1388,8 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt)
>  	case V4L2_META_FMT_VIVID:       descr = "Vivid Metadata"; break;
>  	case V4L2_META_FMT_RK_ISP1_PARAMS:	descr = "Rockchip ISP1 3A Parameters"; break;
>  	case V4L2_META_FMT_RK_ISP1_STAT_3A:	descr = "Rockchip ISP1 3A Statistics"; break;
> +	case V4L2_PIX_FMT_NV12MT_8L128:	descr = "NV12M (8x128 Linear)"; break;
> +	case V4L2_PIX_FMT_NV12MT_10BE_8L128:	descr = "NV12M 10BE(8x128 Linear)"; break;

"10-bit NV12M (8x128 Linear, BE)"

>  
>  	default:
>  		/* Compressed formats */
> diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
> index f118fe7a9f58..9443c3109928 100644
> --- a/include/uapi/linux/videodev2.h
> +++ b/include/uapi/linux/videodev2.h
> @@ -632,6 +632,8 @@ struct v4l2_pix_format {
>  /* Tiled YUV formats, non contiguous planes */
>  #define V4L2_PIX_FMT_NV12MT  v4l2_fourcc('T', 'M', '1', '2') /* 12  Y/CbCr 4:2:0 64x32 tiles */
>  #define V4L2_PIX_FMT_NV12MT_16X16 v4l2_fourcc('V', 'M', '1', '2') /* 12  Y/CbCr 4:2:0 16x16 tiles */
> +#define V4L2_PIX_FMT_NV12MT_8L128      v4l2_fourcc('N', 'A', '1', '2') /* Y/CbCr 4:2:0 8x128 tiles */
> +#define V4L2_PIX_FMT_NV12MT_10BE_8L128 v4l2_fourcc('N', 'T', '1', '2') /* Y/CbCr 4:2:0 10-bit 8x128 tiles */

Use v4l2_fourcc_be to denote that this is a BE format.

>  
>  /* Bayer formats - see http://www.siliconimaging.com/RGB%20Bayer.htm */
>  #define V4L2_PIX_FMT_SBGGR8  v4l2_fourcc('B', 'A', '8', '1') /*  8  BGBG.. GRGR.. */
> 

Regards,

	Hans