linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To: Jacopo Mondi <jacopo@jmondi.org>
Cc: slongerbeam@gmail.com, sakari.ailus@iki.fi,
	hverkuil-cisco@xs4all.nl, mirela.rabulea@nxp.com,
	xavier.roumegue@oss.nxp.com, tomi.valkeinen@ideasonboard.com,
	hugues.fruchet@st.com, prabhakar.mahadev-lad.rj@bp.renesas.com,
	aford173@gmail.com, festevam@gmail.com,
	eugen.hristev@microchip.com, jbrunet@baylibre.com,
	mchehab@kernel.org, linux-media@vger.kernel.org
Subject: Re: [PATCH 14/21] media: ov5640: Implement get_selection
Date: Thu, 3 Feb 2022 00:29:47 +0200	[thread overview]
Message-ID: <YfsF247w8mPNH5e4@pendragon.ideasonboard.com> (raw)
In-Reply-To: <20220131144444.129036-3-jacopo@jmondi.org>

Hi Jacopo,

Thank you for the patch.

On Mon, Jan 31, 2022 at 03:44:42PM +0100, Jacopo Mondi wrote:
> Implement the get_selection pad operation for the OV5640 sensor driver.
> 
> The supported targets report the sensor's native size, the active pixel
> array size and the analog crop rectangle from which the image is
> produced.
> 
> Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
> ---
>  drivers/media/i2c/ov5640.c | 61 ++++++++++++++++++++++++++++++++++++++
>  1 file changed, 61 insertions(+)
> 
> diff --git a/drivers/media/i2c/ov5640.c b/drivers/media/i2c/ov5640.c
> index 762bdca83aec..ae22300b9655 100644
> --- a/drivers/media/i2c/ov5640.c
> +++ b/drivers/media/i2c/ov5640.c
> @@ -35,6 +35,13 @@
>  #define OV5640_MIN_VBLANK	24
>  #define OV5640_MAX_VTS		1968
>  
> +#define OV5640_NATIVE_WIDTH		2624
> +#define OV5640_NATIVE_HEIGHT		1964
> +#define OV5640_PIXEL_ARRAY_TOP		8
> +#define OV5640_PIXEL_ARRAY_LEFT		16
> +#define OV5640_PIXEL_ARRAY_WIDTH	2592
> +#define OV5640_PIXEL_ARRAY_HEIGHT	1944

According to the datasheet, the sensor has 8 black lines, 6 dummy lines,
1944 active lines and 6 dummy lines. Horizontally, it has 16 dummy
columns, 2592 active columns, and 16 dummy columns. If "pixel array" is
meant to refer to the active area (I dislike the "active" name here, as
the dummy lines and columns are typically "active" too, but that's a
digression), then top should be 14.

> +
>  #define OV5640_DEFAULT_SLAVE_ID 0x3c
>  
>  #define OV5640_REG_SYS_RESET02		0x3002
> @@ -2667,6 +2674,52 @@ static int ov5640_set_fmt(struct v4l2_subdev *sd,
>  	return ret;
>  }
>  
> +static int ov5640_get_selection(struct v4l2_subdev *sd,
> +				struct v4l2_subdev_state *sd_state,
> +				struct v4l2_subdev_selection *sel)
> +{
> +	struct ov5640_dev *sensor = to_ov5640_dev(sd);
> +	const struct ov5640_mode_info *mode = sensor->current_mode;
> +	const struct v4l2_rect *analog_crop = &mode->analog_crop;
> +	const struct v4l2_rect *crop = &mode->crop;
> +
> +	switch (sel->target) {
> +	case V4L2_SEL_TGT_CROP: {
> +		mutex_lock(&sensor->lock);
> +
> +		sel->r.top = analog_crop->top + OV5640_PIXEL_ARRAY_TOP;
> +		sel->r.left = analog_crop->left + OV5640_PIXEL_ARRAY_LEFT;
> +		sel->r.width = analog_crop->width
> +			     - analog_crop->left - crop->left;

Why do you subtract the left coordinates here ?

> +		sel->r.height = analog_crop->height
> +			      - analog_crop->top - crop->top;
> +
> +		mutex_unlock(&sensor->lock);
> +
> +		return 0;
> +	}
> +
> +	case V4L2_SEL_TGT_NATIVE_SIZE:
> +		sel->r.top = 0;
> +		sel->r.left = 0;
> +		sel->r.width = OV5640_NATIVE_WIDTH;
> +		sel->r.height = OV5640_NATIVE_HEIGHT;
> +
> +		return 0;
> +
> +	case V4L2_SEL_TGT_CROP_DEFAULT:
> +	case V4L2_SEL_TGT_CROP_BOUNDS:
> +		sel->r.top = OV5640_PIXEL_ARRAY_TOP;
> +		sel->r.left = OV5640_PIXEL_ARRAY_LEFT;
> +		sel->r.width = OV5640_PIXEL_ARRAY_WIDTH;
> +		sel->r.height = OV5640_PIXEL_ARRAY_HEIGHT;

In libcamera we use V4L2_SEL_TGT_CROP_BOUNDS to set PixelArraySize,
ignoring the left and top coordinates, and V4L2_SEL_TGT_CROP_DEFAULT to
to set the PixelArrayActiveAreas property relative to PixelArraySize.
This means that non-zero values for the left and top coordinates of
V4L2_SEL_TGT_CROP_BOUNDS will cause issues. Is this an issue in
libcamera, or should V4L2_SEL_TGT_CROP_BOUNDS be changed here ?

The related question is, can we read the optical black lines and the
dummy lines and columns ?

> +
> +		return 0;
> +	}
> +
> +	return -EINVAL;
> +}
> +
>  static int ov5640_set_framefmt(struct ov5640_dev *sensor,
>  			       struct v4l2_mbus_framefmt *format)
>  {
> @@ -3369,6 +3422,7 @@ static const struct v4l2_subdev_pad_ops ov5640_pad_ops = {
>  	.enum_mbus_code = ov5640_enum_mbus_code,
>  	.get_fmt = ov5640_get_fmt,
>  	.set_fmt = ov5640_set_fmt,
> +	.get_selection = ov5640_get_selection,
>  	.enum_frame_size = ov5640_enum_frame_size,
>  	.enum_frame_interval = ov5640_enum_frame_interval,
>  };
> @@ -3383,9 +3437,16 @@ static int ov5640_open(struct v4l2_subdev *sd, struct v4l2_subdev_fh *fh)
>  {
>  	struct v4l2_mbus_framefmt *try_fmt =
>  		v4l2_subdev_get_try_format(sd, fh->state, 0);
> +	struct v4l2_rect *try_crop =
> +		v4l2_subdev_get_try_crop(sd, fh->state, 0);
>  
>  	*try_fmt = ov5640_default_fmt;
>  
> +	try_crop->left = OV5640_PIXEL_ARRAY_LEFT;
> +	try_crop->top = OV5640_PIXEL_ARRAY_TOP;
> +	try_crop->width = OV5640_PIXEL_ARRAY_WIDTH;
> +	try_crop->height = OV5640_PIXEL_ARRAY_HEIGHT;
> +
>  	return 0;
>  }
>  

-- 
Regards,

Laurent Pinchart

  reply	other threads:[~2022-02-02 22:30 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-31 14:32 [PATCH 00/17] media: ov5640: Rework the clock tree programming for MIPI Jacopo Mondi
2022-01-31 14:32 ` [PATCH 01/21] media: ov5640: Add pixel rate to modes Jacopo Mondi
2022-02-01 14:52   ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 02/21] media: ov5604: Re-arrange modes definition Jacopo Mondi
2022-02-01 14:20   ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 03/21] media: ov5640: Add is_mipi() function Jacopo Mondi
2022-02-01 14:25   ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 04/21] media: ov5640: Associate bpp with formats Jacopo Mondi
2022-02-01 14:27   ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 05/21] media: ov5640: Update pixel_rate and link_freq Jacopo Mondi
2022-02-01 16:52   ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 06/21] media: ov5640: Rework CSI-2 clock tree Jacopo Mondi
2022-02-01 17:27   ` Laurent Pinchart
2022-02-07 14:07     ` Jacopo Mondi
2022-01-31 14:32 ` [PATCH 07/21] media: ov5640: Rework timings programming Jacopo Mondi
2022-02-01 19:03   ` Laurent Pinchart
2022-02-07 14:37     ` Jacopo Mondi
2022-02-07 16:39       ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 08/21] media: ov5640: Re-sort per-mode register tables Jacopo Mondi
2022-02-01 19:05   ` Laurent Pinchart
2022-02-07 14:42     ` Jacopo Mondi
2022-01-31 14:32 ` [PATCH 09/21] media: ov5640: Remove ov5640_mode_init_data Jacopo Mondi
2022-02-01 19:07   ` Laurent Pinchart
2022-02-07 14:45     ` Jacopo Mondi
2022-01-31 14:32 ` [PATCH 10/21] media: ov5640: Add HBLANK control Jacopo Mondi
2022-02-02 21:20   ` Laurent Pinchart
2022-01-31 14:32 ` [PATCH 11/21] media: ov5640: Add VBLANK control Jacopo Mondi
2022-02-02 21:35   ` Laurent Pinchart
2022-02-07 15:09     ` Jacopo Mondi
2022-02-07 17:22       ` Laurent Pinchart
2022-02-03  7:58   ` Xavier Roumegue (OSS)
2022-02-07 15:12     ` Jacopo Mondi
2022-01-31 14:44 ` [PATCH 12/21] media: ov5640: Fix durations to comply with FPS Jacopo Mondi
2022-01-31 14:44   ` [PATCH 13/21] media: ov5640: Initialize try format Jacopo Mondi
2022-02-02 21:51     ` Laurent Pinchart
2022-01-31 14:44   ` [PATCH 14/21] media: ov5640: Implement get_selection Jacopo Mondi
2022-02-02 22:29     ` Laurent Pinchart [this message]
2022-02-07 15:47       ` Jacopo Mondi
2022-02-07 17:53         ` Laurent Pinchart
2022-02-08 14:18           ` Jacopo Mondi
2022-02-08 14:41             ` Laurent Pinchart
2022-01-31 14:44   ` [PATCH 15/21] media: ov5640: Limit format to FPS in DVP mode only Jacopo Mondi
2022-02-02 22:38     ` Laurent Pinchart
2022-02-07 15:49       ` Jacopo Mondi
2022-01-31 14:44   ` [PATCH 16/21] media: ov5640: Disable s_frame_interval in MIPI mode Jacopo Mondi
2022-02-02 22:40     ` Laurent Pinchart
2022-02-02 21:48   ` [PATCH 12/21] media: ov5640: Fix durations to comply with FPS Laurent Pinchart
2022-02-07 15:58     ` Jacopo Mondi
2022-01-31 14:45 ` [PATCH 17/21] media: ov5640: Register device properties Jacopo Mondi
2022-01-31 14:45   ` [PATCH 18/21] media: ov5640: Add RGB565_1X16 format Jacopo Mondi
2022-02-02 22:48     ` Laurent Pinchart
2022-01-31 14:45   ` [PATCH 19/21] media: ov5640: Add RGB888/BGR888 formats Jacopo Mondi
2022-02-02 22:49     ` Laurent Pinchart
2022-02-02 22:44   ` [PATCH 17/21] media: ov5640: Register device properties Laurent Pinchart
2022-01-31 14:45 ` [PATCH 20/21] media: ov5640: Restrict sizes to mbus code Jacopo Mondi
2022-01-31 14:45   ` [PATCH 21/21] media: ov5640: Adjust format to bpp in s_fmt Jacopo Mondi
2022-02-02 23:03     ` Laurent Pinchart
2022-02-07 16:07       ` Jacopo Mondi
2022-02-07 16:46         ` Laurent Pinchart
2022-02-02 22:57   ` [PATCH 20/21] media: ov5640: Restrict sizes to mbus code Laurent Pinchart

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=YfsF247w8mPNH5e4@pendragon.ideasonboard.com \
    --to=laurent.pinchart@ideasonboard.com \
    --cc=aford173@gmail.com \
    --cc=eugen.hristev@microchip.com \
    --cc=festevam@gmail.com \
    --cc=hugues.fruchet@st.com \
    --cc=hverkuil-cisco@xs4all.nl \
    --cc=jacopo@jmondi.org \
    --cc=jbrunet@baylibre.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=mirela.rabulea@nxp.com \
    --cc=prabhakar.mahadev-lad.rj@bp.renesas.com \
    --cc=sakari.ailus@iki.fi \
    --cc=slongerbeam@gmail.com \
    --cc=tomi.valkeinen@ideasonboard.com \
    --cc=xavier.roumegue@oss.nxp.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).