Le vendredi 09 février 2024 à 19:28 +0000, Dave Stevenson a écrit : > Hi Nicolas > > On Fri, 9 Feb 2024 at 17:53, Nicolas Dufresne <nicolas@xxxxxxxxxxxx> wrote: > > > > Le vendredi 09 février 2024 à 17:48 +0100, Jacopo Mondi a écrit : > > > From: Dave Stevenson <dave.stevenson@xxxxxxxxxxxxxxx> > > > > > > This is the format used by monochrome 12bit image sensors. > > > > > > Signed-off-by: Dave Stevenson <dave.stevenson@xxxxxxxxxxxxxxx> > > > Signed-off-by: Jacopo Mondi <jacopo.mondi@xxxxxxxxxxxxxxxx> > > > --- > > > .../userspace-api/media/v4l/pixfmt-y12p.rst | 38 +++++++++++++++++++ > > > .../userspace-api/media/v4l/yuv-formats.rst | 1 + > > > drivers/media/v4l2-core/v4l2-ioctl.c | 1 + > > > include/uapi/linux/videodev2.h | 1 + > > > 4 files changed, 41 insertions(+) > > > create mode 100644 Documentation/userspace-api/media/v4l/pixfmt-y12p.rst > > > > > > diff --git a/Documentation/userspace-api/media/v4l/pixfmt-y12p.rst b/Documentation/userspace-api/media/v4l/pixfmt-y12p.rst > > > new file mode 100644 > > > index 000000000000..b2eb4a72724d > > > --- /dev/null > > > +++ b/Documentation/userspace-api/media/v4l/pixfmt-y12p.rst > > > @@ -0,0 +1,38 @@ > > > +.. SPDX-License-Identifier: GFDL-1.1-no-invariants-or-later > > > + > > > +.. _V4L2-PIX-FMT-Y12P: > > > + > > > +****************************** > > > +V4L2_PIX_FMT_Y12P ('Y12P') > > > +****************************** > > > + > > > +Grey-scale image as a MIPI RAW12 packed array > > > + > > > + > > > +Description > > > +=========== > > > + > > > +This is a packed grey-scale image format with a depth of 12 bits per > > > +pixel. Two consecutive pixels are packed into 3 bytes. The first 2 bytes > > > +contain the 8 high order bits of the pixels, and the 3rd byte contains the 4 > > > +least significants bits of each pixel, in the same order. > > > > This is an interesting arrangement, which make me feel that Y12P is perhaps bit > > too generic ? Perhaps something like: > > > > V4L2_PIX_FMT_Y12_MIPI > > > > That being said, I don't mind if you reserve the nice and short name for the > > first occurrence of such format in 20 years (it might equally be the last). Will > > they do the same for Y10, by storing 4 bytes of higher 8bit of 4 pixels, and > > packing the remaining lower 2 bits packed in the fif bytes ? Cause for this one, > > we'd have confusion, since CODEC usually just place all the bits in order over 5 > > bytes here (which have its own issues of course). > > We already have V4L2_PIX_FMT_Y10P defined for the MIPI packing for > 10bit luma-only. > https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/pixfmt-yuv-luma.html > > This is only extending the definitions for the other bit depths using > the same pattern. > > Adding the P is also the same pattern as for the Bayer formats - > V4L2_PIX_FMT_SRGGB10 for the 10bit unpacked to 16bit, and > V4L2_PIX_FMT_SRGGB10P for the MIPI packed variant. > > I'm not inventing anything new here :) A bit sad, but what can we do, I keep missing patches ;-P Nicolas > > Dave > > > > + > > > +**Byte Order.** > > > +Each cell is one byte. > > > + > > > +.. tabularcolumns:: |p{2.2cm}|p{1.2cm}|p{1.2cm}|p{3.1cm}| > > > + > > > + > > > +.. flat-table:: > > > + :header-rows: 0 > > > + :stub-columns: 0 > > > + :widths: 2 1 1 1 > > > + > > > + > > > + - - start + 0: > > > + - Y'\ :sub:`00high` > > > + - Y'\ :sub:`01high` > > > + - Y'\ :sub:`01low`\ (bits 7--4) > > > + > > > + Y'\ :sub:`00low`\ (bits 3--0) > > > + > > > diff --git a/Documentation/userspace-api/media/v4l/yuv-formats.rst b/Documentation/userspace-api/media/v4l/yuv-formats.rst > > > index 24b34cdfa6fe..7c9ccfdd94cd 100644 > > > --- a/Documentation/userspace-api/media/v4l/yuv-formats.rst > > > +++ b/Documentation/userspace-api/media/v4l/yuv-formats.rst > > > @@ -267,6 +267,7 @@ image. > > > pixfmt-packed-yuv > > > pixfmt-yuv-planar > > > pixfmt-yuv-luma > > > + pixfmt-y12p > > > pixfmt-y8i > > > pixfmt-y12i > > > pixfmt-uv8 > > > diff --git a/drivers/media/v4l2-core/v4l2-ioctl.c b/drivers/media/v4l2-core/v4l2-ioctl.c > > > index 33076af4dfdb..483498c55899 100644 > > > --- a/drivers/media/v4l2-core/v4l2-ioctl.c > > > +++ b/drivers/media/v4l2-core/v4l2-ioctl.c > > > @@ -1311,6 +1311,7 @@ static void v4l_fill_fmtdesc(struct v4l2_fmtdesc *fmt) > > > case V4L2_PIX_FMT_Y16_BE: descr = "16-bit Greyscale BE"; break; > > > case V4L2_PIX_FMT_Y10BPACK: descr = "10-bit Greyscale (Packed)"; break; > > > case V4L2_PIX_FMT_Y10P: descr = "10-bit Greyscale (MIPI Packed)"; break; > > > + case V4L2_PIX_FMT_Y12P: descr = "12-bit Greyscale (MIPI Packed)"; break; > > > case V4L2_PIX_FMT_IPU3_Y10: descr = "10-bit greyscale (IPU3 Packed)"; break; > > > case V4L2_PIX_FMT_Y8I: descr = "Interleaved 8-bit Greyscale"; break; > > > case V4L2_PIX_FMT_Y12I: descr = "Interleaved 12-bit Greyscale"; break; > > > diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h > > > index a8015e5e7fa4..11ebf9b22ccb 100644 > > > --- a/include/uapi/linux/videodev2.h > > > +++ b/include/uapi/linux/videodev2.h > > > @@ -598,6 +598,7 @@ struct v4l2_pix_format { > > > /* Grey bit-packed formats */ > > > #define V4L2_PIX_FMT_Y10BPACK v4l2_fourcc('Y', '1', '0', 'B') /* 10 Greyscale bit-packed */ > > > #define V4L2_PIX_FMT_Y10P v4l2_fourcc('Y', '1', '0', 'P') /* 10 Greyscale, MIPI RAW10 packed */ > > > +#define V4L2_PIX_FMT_Y12P v4l2_fourcc('Y', '1', '2', 'P') /* 12 Greyscale, MIPI RAW12 packed */ > > > #define V4L2_PIX_FMT_IPU3_Y10 v4l2_fourcc('i', 'p', '3', 'y') /* IPU3 packed 10-bit greyscale */ > > > > > > /* Palette formats */ > >