Some sensors (like the one in Nokia N900) provide metadata in the first
couple of lines. Make that information information available to the
pipeline.
Signed-off-by: Ivaylo Dimitrov <[email protected]>
---
drivers/media/i2c/smiapp/smiapp-core.c | 12 ++++++++++++
drivers/media/i2c/smiapp/smiapp.h | 1 +
2 files changed, 13 insertions(+)
diff --git a/drivers/media/i2c/smiapp/smiapp-core.c b/drivers/media/i2c/smiapp/smiapp-core.c
index a215efe..3dfe387 100644
--- a/drivers/media/i2c/smiapp/smiapp-core.c
+++ b/drivers/media/i2c/smiapp/smiapp-core.c
@@ -188,6 +188,8 @@ static int smiapp_read_frame_fmt(struct smiapp_sensor *sensor)
embedded_end = 0;
}
+ sensor->image_start = image_start;
+
dev_dbg(&client->dev, "embedded data from lines %d to %d\n",
embedded_start, embedded_end);
dev_dbg(&client->dev, "image data starts at line %d\n", image_start);
@@ -2280,6 +2282,15 @@ static int smiapp_get_skip_frames(struct v4l2_subdev *subdev, u32 *frames)
return 0;
}
+static int smiapp_get_skip_top_lines(struct v4l2_subdev *subdev, u32 *lines)
+{
+ struct smiapp_sensor *sensor = to_smiapp_sensor(subdev);
+
+ *lines = sensor->image_start;
+
+ return 0;
+}
+
/* -----------------------------------------------------------------------------
* sysfs attributes
*/
@@ -2890,6 +2901,7 @@ static const struct v4l2_subdev_pad_ops smiapp_pad_ops = {
static const struct v4l2_subdev_sensor_ops smiapp_sensor_ops = {
.g_skip_frames = smiapp_get_skip_frames,
+ .g_skip_top_lines = smiapp_get_skip_top_lines,
};
static const struct v4l2_subdev_ops smiapp_ops = {
diff --git a/drivers/media/i2c/smiapp/smiapp.h b/drivers/media/i2c/smiapp/smiapp.h
index f6af0cc..c8b4ca0 100644
--- a/drivers/media/i2c/smiapp/smiapp.h
+++ b/drivers/media/i2c/smiapp/smiapp.h
@@ -217,6 +217,7 @@ struct smiapp_sensor {
u8 hvflip_inv_mask; /* H/VFLIP inversion due to sensor orientation */
u8 frame_skip;
+ u32 image_start; /* Offset to first line after metadata lines */
int power_count;
--
1.9.1
Hi Ivaylo,
On Sat, Apr 16, 2016 at 11:12:20AM +0300, Ivaylo Dimitrov wrote:
> Some sensors (like the one in Nokia N900) provide metadata in the first
> couple of lines. Make that information information available to the
> pipeline.
>
> Signed-off-by: Ivaylo Dimitrov <[email protected]>
> ---
> drivers/media/i2c/smiapp/smiapp-core.c | 12 ++++++++++++
> drivers/media/i2c/smiapp/smiapp.h | 1 +
> 2 files changed, 13 insertions(+)
>
> diff --git a/drivers/media/i2c/smiapp/smiapp-core.c b/drivers/media/i2c/smiapp/smiapp-core.c
> index a215efe..3dfe387 100644
> --- a/drivers/media/i2c/smiapp/smiapp-core.c
> +++ b/drivers/media/i2c/smiapp/smiapp-core.c
> @@ -188,6 +188,8 @@ static int smiapp_read_frame_fmt(struct smiapp_sensor *sensor)
> embedded_end = 0;
> }
>
> + sensor->image_start = image_start;
> +
> dev_dbg(&client->dev, "embedded data from lines %d to %d\n",
> embedded_start, embedded_end);
> dev_dbg(&client->dev, "image data starts at line %d\n", image_start);
> @@ -2280,6 +2282,15 @@ static int smiapp_get_skip_frames(struct v4l2_subdev *subdev, u32 *frames)
> return 0;
> }
>
> +static int smiapp_get_skip_top_lines(struct v4l2_subdev *subdev, u32 *lines)
> +{
> + struct smiapp_sensor *sensor = to_smiapp_sensor(subdev);
> +
> + *lines = sensor->image_start;
> +
> + return 0;
> +}
> +
> /* -----------------------------------------------------------------------------
> * sysfs attributes
> */
> @@ -2890,6 +2901,7 @@ static const struct v4l2_subdev_pad_ops smiapp_pad_ops = {
>
> static const struct v4l2_subdev_sensor_ops smiapp_sensor_ops = {
> .g_skip_frames = smiapp_get_skip_frames,
> + .g_skip_top_lines = smiapp_get_skip_top_lines,
> };
>
> static const struct v4l2_subdev_ops smiapp_ops = {
> diff --git a/drivers/media/i2c/smiapp/smiapp.h b/drivers/media/i2c/smiapp/smiapp.h
> index f6af0cc..c8b4ca0 100644
> --- a/drivers/media/i2c/smiapp/smiapp.h
> +++ b/drivers/media/i2c/smiapp/smiapp.h
> @@ -217,6 +217,7 @@ struct smiapp_sensor {
>
> u8 hvflip_inv_mask; /* H/VFLIP inversion due to sensor orientation */
> u8 frame_skip;
> + u32 image_start; /* Offset to first line after metadata lines */
>
> int power_count;
>
I'm afraid I think this is not exactly the best way to approach the issue.
It'd work, somehow, yes, but ---
1. A compliant sensor (at least in theory) is able to tell this information
itself. The number of metadata lines is present in the sensor frame format
descriptors.
2. The more generic problem of describing the frame layout should be solved.
Sensor metadata is just a special case of this. I've proposed frame
descriptors (see an old RFC
<URL:http://www.spinics.net/lists/linux-media/msg67295.html>), but this is
just a partial solution as well; the APIs would need to be extended to
support metadata capture (I think Laurent has been working on that).
So a proper solution will require a little bit of time still.
--
Kind regards,
Sakari Ailus
e-mail: [email protected] XMPP: [email protected]
Hi,
On 18.04.2016 00:44, Sakari Ailus wrote:
> Hi Ivaylo,
>
> On Sat, Apr 16, 2016 at 11:12:20AM +0300, Ivaylo Dimitrov wrote:
>> Some sensors (like the one in Nokia N900) provide metadata in the first
>> couple of lines. Make that information information available to the
>> pipeline.
>>
>> Signed-off-by: Ivaylo Dimitrov <[email protected]>
>> ---
>> drivers/media/i2c/smiapp/smiapp-core.c | 12 ++++++++++++
>> drivers/media/i2c/smiapp/smiapp.h | 1 +
>> 2 files changed, 13 insertions(+)
>>
...
>
> I'm afraid I think this is not exactly the best way to approach the issue.
> It'd work, somehow, yes, but ---
>
> 1. A compliant sensor (at least in theory) is able to tell this information
> itself. The number of metadata lines is present in the sensor frame format
> descriptors.
>
Right. And this is where that number is taken from in the patch and made
available to whoever wants to use it. See
http://lxr.free-electrons.com/source/drivers/media/i2c/smiapp/smiapp-core.c#L177
. I don't really understand your point here. Maybe the patch description
is fuzzy? Could you elaborate?
> 2. The more generic problem of describing the frame layout should be solved.
> Sensor metadata is just a special case of this. I've proposed frame
> descriptors (see an old RFC
> <URL:http://www.spinics.net/lists/linux-media/msg67295.html>), but this is
> just a partial solution as well; the APIs would need to be extended to
> support metadata capture (I think Laurent has been working on that).
>
Could be, however what we have right now is
http://lxr.free-electrons.com/source/drivers/media/platform/omap3isp/ispccp2.c#L369.
Also, the patch is not trying to solve the problem with frame format
description(or anything in general), but a mere way to pass an already
available information in the sensor which is needed by omap3isp, by
using an already existing API. I don't see how's that related to the way
v4l API going to evolve in some (distant?) future. Not to say that once
those frame format descriptors are available, it should be relatively
easy to simply remove g_skip_top_lines form v4l2_subdev_sensor_ops and
fix the drivers to use the new API.
BTW if you have any idea on how to pass (or set) the number of lines to
be skipped at the start of the frame to omap3isp driver in some other
way, I am fine with dropping the $subject patch and sending another one
implementing your proposal.
Regards,
Ivo
Hi Ivaylo,
On Mon, Apr 18, 2016 at 09:27:53AM +0300, Ivaylo Dimitrov wrote:
> Hi,
>
> On 18.04.2016 00:44, Sakari Ailus wrote:
> >Hi Ivaylo,
> >
> >On Sat, Apr 16, 2016 at 11:12:20AM +0300, Ivaylo Dimitrov wrote:
> >>Some sensors (like the one in Nokia N900) provide metadata in the first
> >>couple of lines. Make that information information available to the
> >>pipeline.
> >>
> >>Signed-off-by: Ivaylo Dimitrov <[email protected]>
> >>---
> >> drivers/media/i2c/smiapp/smiapp-core.c | 12 ++++++++++++
> >> drivers/media/i2c/smiapp/smiapp.h | 1 +
> >> 2 files changed, 13 insertions(+)
> >>
> ...
> >
> >I'm afraid I think this is not exactly the best way to approach the issue.
> >It'd work, somehow, yes, but ---
> >
> >1. A compliant sensor (at least in theory) is able to tell this information
> >itself. The number of metadata lines is present in the sensor frame format
> >descriptors.
> >
>
> Right. And this is where that number is taken from in the patch and made
> available to whoever wants to use it. See http://lxr.free-electrons.com/source/drivers/media/i2c/smiapp/smiapp-core.c#L177
> . I don't really understand your point here. Maybe the patch description is
> fuzzy? Could you elaborate?
I missed just that part, apologies for that. I'll apply this into my tree.
>
> >2. The more generic problem of describing the frame layout should be solved.
> >Sensor metadata is just a special case of this. I've proposed frame
> >descriptors (see an old RFC
> ><URL:http://www.spinics.net/lists/linux-media/msg67295.html>), but this is
> >just a partial solution as well; the APIs would need to be extended to
> >support metadata capture (I think Laurent has been working on that).
> >
>
> Could be, however what we have right now is http://lxr.free-electrons.com/source/drivers/media/platform/omap3isp/ispccp2.c#L369.
> Also, the patch is not trying to solve the problem with frame format
> description(or anything in general), but a mere way to pass an already
> available information in the sensor which is needed by omap3isp, by using an
> already existing API. I don't see how's that related to the way v4l API
> going to evolve in some (distant?) future. Not to say that once those frame
> format descriptors are available, it should be relatively easy to simply
> remove g_skip_top_lines form v4l2_subdev_sensor_ops and fix the drivers to
> use the new API.
>
> BTW if you have any idea on how to pass (or set) the number of lines to be
> skipped at the start of the frame to omap3isp driver in some other way, I am
> fine with dropping the $subject patch and sending another one implementing
> your proposal.
Hopefully we'll have a better solution in not so distant future. Metadata is
just a special case for frame descriptors.
--
Kind regards,
Sakari Ailus
e-mail: [email protected] XMPP: [email protected]