mbox series

[0/4] media: ti-vpe: cal: multiplexed streams support

Message ID 20210427132028.1005757-1-tomi.valkeinen@ideasonboard.com
Headers show
Series media: ti-vpe: cal: multiplexed streams support | expand

Message

Tomi Valkeinen April 27, 2021, 1:20 p.m. UTC
Hi,

This series adds embedded data and multiplexed streams support to TI CAL
driver. These patches depend on the subdev routing series and CAL preparation
series:

https://lore.kernel.org/linux-media/20210427124523.990938-1-tomi.valkeinen@ideasonboard.com/

https://lore.kernel.org/linux-media/20210420120433.902394-1-tomi.valkeinen@ideasonboard.com/

One detail I'm not too happy about is adding a new kernel module parameter
'cal_streams_api'. Setting this parameter enables the multiplexed streams
support for the CAL sink pads. I'd like the driver to be able to figure this
out itself, but there's no simple answer to it.

A pad in a subdev has to be either in non-multiplexed or multiplexed mode. If
the subdev cannot support multiplexed streams, then the answer is clear. But if
it can, then the question becomes "what's connected to the pad, and what does
it want". Afaik, the idea with subdev configuration is that each subdev can be
configured independently, so the driver trying to figure out the answer to that
question breaks this idea. Also, two connected subdevs that both can support
both modes wouldn't even give the answer, but the answer might be found further
away, probably from the sensor subdev.

So, maybe it should be the userspace that can set a subdev pad to either of the
modes. There's currently no such uAPI, but I think this option makes the most
sense.

Another detail to note is the embedded data support. I don't have hardware with
real embedded data, so I have not added any embedded data formats. Also, the
embedded data is captured with the video capture ioctls, not the metadata
capture ioctls. The reason for this is that in the case of CSI-2 embedded data
(at least for this hardware) the embedded data is very much like pixel data:
the data is sent in lines just before or after the pixel content and each
subdev needs a set_fmt call with the width, height, and mbus-code.

Using the metadata ioctls is possible, but it only results in much more complex
driver code, and, I think, more confusing userspace.

 Tomi

Tomi Valkeinen (4):
  media: ti-vpe: cal: add embedded data support
  media: ti-vpe: cal: use frame desc to get vc and dt
  media: ti-vpe: cal: add routing support
  media: ti-vpe: cal: add multiplexed streams support

 drivers/media/platform/ti-vpe/cal-camerarx.c | 234 ++++++++++++++++---
 drivers/media/platform/ti-vpe/cal-video.c    |  95 +++++++-
 drivers/media/platform/ti-vpe/cal.c          | 102 ++++++--
 drivers/media/platform/ti-vpe/cal.h          |  10 +-
 4 files changed, 383 insertions(+), 58 deletions(-)

Comments

Laurent Pinchart April 29, 2021, 3:03 a.m. UTC | #1
Hi Tomi,

Thank you for the patch.

On Tue, Apr 27, 2021 at 04:20:25PM +0300, Tomi Valkeinen wrote:
> Add support for capturing embedded data from the sensor. The only

> difference with capturing pixel data and embedded data is that we need

> to ensure the PIX PROC is disabled for embedded data so that CAL doesn't

> repack the data.

> 

> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>


Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>


> ---

>  drivers/media/platform/ti-vpe/cal.c | 17 ++++++++++-------

>  drivers/media/platform/ti-vpe/cal.h |  1 +

>  2 files changed, 11 insertions(+), 7 deletions(-)

> 

> diff --git a/drivers/media/platform/ti-vpe/cal.c b/drivers/media/platform/ti-vpe/cal.c

> index 62c45add4efe..fcc81024ae18 100644

> --- a/drivers/media/platform/ti-vpe/cal.c

> +++ b/drivers/media/platform/ti-vpe/cal.c

> @@ -473,14 +473,17 @@ int cal_ctx_prepare(struct cal_ctx *ctx)

>  {

>  	int ret;

>  

> -	ret = cal_reserve_pix_proc(ctx->cal);

> -	if (ret < 0) {

> -		ctx_err(ctx, "Failed to reserve pix proc: %d\n", ret);

> -		return ret;

> -	}

> +	ctx->use_pix_proc = !ctx->fmtinfo->meta;

> +

> +	if (ctx->use_pix_proc) {

> +		ret = cal_reserve_pix_proc(ctx->cal);

> +		if (ret < 0) {

> +			ctx_err(ctx, "Failed to reserve pix proc: %d\n", ret);

> +			return ret;

> +		}

>  

> -	ctx->pix_proc = ret;

> -	ctx->use_pix_proc = true;

> +		ctx->pix_proc = ret;

> +	}

>  

>  	return 0;

>  }

> diff --git a/drivers/media/platform/ti-vpe/cal.h b/drivers/media/platform/ti-vpe/cal.h

> index 42a3f8004077..29b865d1a238 100644

> --- a/drivers/media/platform/ti-vpe/cal.h

> +++ b/drivers/media/platform/ti-vpe/cal.h

> @@ -88,6 +88,7 @@ struct cal_format_info {

>  	u32	code;

>  	/* Bits per pixel */

>  	u8	bpp;

> +	bool	meta;

>  };

>  

>  /* buffer for one video frame */


-- 
Regards,

Laurent Pinchart
Laurent Pinchart April 29, 2021, 3:27 a.m. UTC | #2
Hi Tomi,

Thank you for the patch.

On Tue, Apr 27, 2021 at 04:20:27PM +0300, Tomi Valkeinen wrote:
> Add routing support. As we still only support a single stream (i.e. a

> single route), the routing is not really used for anything yet.

> 

> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>

> ---

>  drivers/media/platform/ti-vpe/cal-camerarx.c | 64 ++++++++++++++++++++

>  drivers/media/platform/ti-vpe/cal.h          |  2 +

>  2 files changed, 66 insertions(+)

> 

> diff --git a/drivers/media/platform/ti-vpe/cal-camerarx.c b/drivers/media/platform/ti-vpe/cal-camerarx.c

> index 36103f5af6e9..3470f6dc0ec1 100644

> --- a/drivers/media/platform/ti-vpe/cal-camerarx.c

> +++ b/drivers/media/platform/ti-vpe/cal-camerarx.c

> @@ -805,6 +805,37 @@ static int cal_camerarx_sd_set_fmt(struct v4l2_subdev *sd,

>  	return 0;

>  }

>  

> +static int cal_camerarx_sd_get_routing(struct v4l2_subdev *sd,

> +				       struct v4l2_subdev_krouting *routing)

> +{

> +	struct cal_camerarx *phy = to_cal_camerarx(sd);

> +	struct v4l2_subdev_krouting *src;


const, as src isn't modified.

> +

> +	src = &phy->routing;


That could be written

	const struct v4l2_subdev_krouting *src = &phy->routing;

> +

> +	if (routing->num_routes < src->num_routes) {

> +		routing->num_routes = src->num_routes;

> +		return -ENOSPC;

> +	}

> +

> +	v4l2_subdev_cpy_routing(routing, src);

> +

> +	return 0;

> +}

> +

> +static int cal_camerarx_sd_set_routing(struct v4l2_subdev *sd,

> +				       struct v4l2_subdev_krouting *routing)

> +{

> +	struct cal_camerarx *phy = to_cal_camerarx(sd);

> +	int ret;


Shouldn't you reject !ACTIVE configs ? Same in
cal_camerarx_sd_get_routing() I suppose.

> +

> +	ret = v4l2_subdev_dup_routing(&phy->routing, routing);

> +	if (ret)

> +		return ret;

> +

> +	return 0;

> +}

> +

>  static int cal_camerarx_sd_init_cfg(struct v4l2_subdev *sd,

>  				    struct v4l2_subdev_pad_config *cfg)

>  {

> @@ -837,6 +868,8 @@ static const struct v4l2_subdev_pad_ops cal_camerarx_pad_ops = {

>  	.enum_frame_size = cal_camerarx_sd_enum_frame_size,

>  	.get_fmt = cal_camerarx_sd_get_fmt,

>  	.set_fmt = cal_camerarx_sd_set_fmt,

> +	.get_routing = cal_camerarx_sd_get_routing,

> +	.set_routing = cal_camerarx_sd_set_routing,

>  };

>  

>  static const struct v4l2_subdev_ops cal_camerarx_subdev_ops = {

> @@ -844,8 +877,18 @@ static const struct v4l2_subdev_ops cal_camerarx_subdev_ops = {

>  	.pad = &cal_camerarx_pad_ops,

>  };

>  

> +static bool cal_camerarx_has_route(struct media_entity *entity, unsigned int pad0,

> +			  unsigned int pad1)

> +{

> +	struct v4l2_subdev *sd = media_entity_to_v4l2_subdev(entity);

> +	struct cal_camerarx *phy = to_cal_camerarx(sd);

> +

> +	return v4l2_subdev_has_route(&phy->routing, pad0, pad1);

> +}

> +

>  static struct media_entity_operations cal_camerarx_media_ops = {

>  	.link_validate = v4l2_subdev_link_validate,

> +	.has_route = cal_camerarx_has_route,

>  };

>  

>  /* ------------------------------------------------------------------

> @@ -862,6 +905,20 @@ struct cal_camerarx *cal_camerarx_create(struct cal_dev *cal,

>  	int ret;

>  	unsigned int i;

>  

> +	struct v4l2_subdev_route routes[] = { {

> +		.sink_pad = 0,

> +		.sink_stream = 0,

> +		.source_pad = 1,

> +		.source_stream = 0,

> +		.flags = V4L2_SUBDEV_ROUTE_FL_ACTIVE,

> +	} };

> +

> +	struct v4l2_subdev_krouting routing = {

> +		.which = V4L2_SUBDEV_FORMAT_ACTIVE,

> +		.num_routes = 1,

> +		.routes = routes,

> +	};


I'd move this above the other variables. Blank lines are not customary.

> +

>  	phy = kzalloc(sizeof(*phy), GFP_KERNEL);

>  	if (!phy)

>  		return ERR_PTR(-ENOMEM);

> @@ -914,6 +971,11 @@ struct cal_camerarx *cal_camerarx_create(struct cal_dev *cal,

>  	if (ret)

>  		goto error;

>  

> +	/* Initialize routing to single route to the fist source pad */


s/fist/first/
s/pad/pad./

> +	ret = cal_camerarx_sd_set_routing(sd, &routing);

> +	if (ret)

> +		goto error;

> +

>  	ret = v4l2_device_register_subdev(&cal->v4l2_dev, sd);

>  	if (ret)

>  		goto error;

> @@ -921,6 +983,7 @@ struct cal_camerarx *cal_camerarx_create(struct cal_dev *cal,

>  	return phy;

>  

>  error:

> +	v4l2_subdev_free_routing(&phy->routing);

>  	media_entity_cleanup(&phy->subdev.entity);

>  	kfree(phy);

>  	return ERR_PTR(ret);

> @@ -932,6 +995,7 @@ void cal_camerarx_destroy(struct cal_camerarx *phy)

>  		return;

>  

>  	v4l2_device_unregister_subdev(&phy->subdev);

> +	v4l2_subdev_free_routing(&phy->routing);

>  	media_entity_cleanup(&phy->subdev.entity);

>  	of_node_put(phy->source_ep_node);

>  	of_node_put(phy->source_node);

> diff --git a/drivers/media/platform/ti-vpe/cal.h b/drivers/media/platform/ti-vpe/cal.h

> index 3aea444f8bf8..c96364eb0930 100644

> --- a/drivers/media/platform/ti-vpe/cal.h

> +++ b/drivers/media/platform/ti-vpe/cal.h

> @@ -184,6 +184,8 @@ struct cal_camerarx {

>  	struct mutex		mutex;

>  

>  	unsigned int enable_count;

> +

> +	struct v4l2_subdev_krouting routing;


Looking forward to storing this in v4l2_subdev_config, and embedding an
instance of v4l2_subdev_config in v4l2_subdev :-)
cal_camerarx_has_route() could then be replaced by a generic V4L2 subdev
helper, and so could cal_camerarx_sd_get_routing().

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>


>  };

>  

>  struct cal_dev {


-- 
Regards,

Laurent Pinchart
Laurent Pinchart April 29, 2021, 3:49 a.m. UTC | #3
Hi Tomi,

One more comment.

On Thu, Apr 29, 2021 at 06:27:01AM +0300, Laurent Pinchart wrote:
> On Tue, Apr 27, 2021 at 04:20:27PM +0300, Tomi Valkeinen wrote:

> > Add routing support. As we still only support a single stream (i.e. a

> > single route), the routing is not really used for anything yet.

> > 

> > Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>

> > ---

> >  drivers/media/platform/ti-vpe/cal-camerarx.c | 64 ++++++++++++++++++++

> >  drivers/media/platform/ti-vpe/cal.h          |  2 +

> >  2 files changed, 66 insertions(+)

> > 

> > diff --git a/drivers/media/platform/ti-vpe/cal-camerarx.c b/drivers/media/platform/ti-vpe/cal-camerarx.c

> > index 36103f5af6e9..3470f6dc0ec1 100644

> > --- a/drivers/media/platform/ti-vpe/cal-camerarx.c

> > +++ b/drivers/media/platform/ti-vpe/cal-camerarx.c

> > @@ -805,6 +805,37 @@ static int cal_camerarx_sd_set_fmt(struct v4l2_subdev *sd,

> >  	return 0;

> >  }

> >  

> > +static int cal_camerarx_sd_get_routing(struct v4l2_subdev *sd,

> > +				       struct v4l2_subdev_krouting *routing)

> > +{

> > +	struct cal_camerarx *phy = to_cal_camerarx(sd);

> > +	struct v4l2_subdev_krouting *src;

> 

> const, as src isn't modified.

> 

> > +

> > +	src = &phy->routing;

> 

> That could be written

> 

> 	const struct v4l2_subdev_krouting *src = &phy->routing;

> 

> > +

> > +	if (routing->num_routes < src->num_routes) {

> > +		routing->num_routes = src->num_routes;

> > +		return -ENOSPC;

> > +	}

> > +

> > +	v4l2_subdev_cpy_routing(routing, src);

> > +

> > +	return 0;

> > +}

> > +

> > +static int cal_camerarx_sd_set_routing(struct v4l2_subdev *sd,

> > +				       struct v4l2_subdev_krouting *routing)

> > +{

> > +	struct cal_camerarx *phy = to_cal_camerarx(sd);

> > +	int ret;

> 

> Shouldn't you reject !ACTIVE configs ? Same in

> cal_camerarx_sd_get_routing() I suppose.

> 

> > +

> > +	ret = v4l2_subdev_dup_routing(&phy->routing, routing);

> > +	if (ret)

> > +		return ret;


Do we need locking here ? The get and set should be serialized, and I
also think we don't want to allow routing configuration during
streaming, at least to start with.

> > +

> > +	return 0;

> > +}

> > +

> >  static int cal_camerarx_sd_init_cfg(struct v4l2_subdev *sd,

> >  				    struct v4l2_subdev_pad_config *cfg)

> >  {

> > @@ -837,6 +868,8 @@ static const struct v4l2_subdev_pad_ops cal_camerarx_pad_ops = {

> >  	.enum_frame_size = cal_camerarx_sd_enum_frame_size,

> >  	.get_fmt = cal_camerarx_sd_get_fmt,

> >  	.set_fmt = cal_camerarx_sd_set_fmt,

> > +	.get_routing = cal_camerarx_sd_get_routing,

> > +	.set_routing = cal_camerarx_sd_set_routing,

> >  };

> >  

> >  static const struct v4l2_subdev_ops cal_camerarx_subdev_ops = {

> > @@ -844,8 +877,18 @@ static const struct v4l2_subdev_ops cal_camerarx_subdev_ops = {

> >  	.pad = &cal_camerarx_pad_ops,

> >  };

> >  

> > +static bool cal_camerarx_has_route(struct media_entity *entity, unsigned int pad0,

> > +			  unsigned int pad1)

> > +{

> > +	struct v4l2_subdev *sd = media_entity_to_v4l2_subdev(entity);

> > +	struct cal_camerarx *phy = to_cal_camerarx(sd);

> > +

> > +	return v4l2_subdev_has_route(&phy->routing, pad0, pad1);

> > +}

> > +

> >  static struct media_entity_operations cal_camerarx_media_ops = {

> >  	.link_validate = v4l2_subdev_link_validate,

> > +	.has_route = cal_camerarx_has_route,

> >  };

> >  

> >  /* ------------------------------------------------------------------

> > @@ -862,6 +905,20 @@ struct cal_camerarx *cal_camerarx_create(struct cal_dev *cal,

> >  	int ret;

> >  	unsigned int i;

> >  

> > +	struct v4l2_subdev_route routes[] = { {

> > +		.sink_pad = 0,

> > +		.sink_stream = 0,

> > +		.source_pad = 1,

> > +		.source_stream = 0,

> > +		.flags = V4L2_SUBDEV_ROUTE_FL_ACTIVE,

> > +	} };

> > +

> > +	struct v4l2_subdev_krouting routing = {

> > +		.which = V4L2_SUBDEV_FORMAT_ACTIVE,

> > +		.num_routes = 1,

> > +		.routes = routes,

> > +	};

> 

> I'd move this above the other variables. Blank lines are not customary.

> 

> > +

> >  	phy = kzalloc(sizeof(*phy), GFP_KERNEL);

> >  	if (!phy)

> >  		return ERR_PTR(-ENOMEM);

> > @@ -914,6 +971,11 @@ struct cal_camerarx *cal_camerarx_create(struct cal_dev *cal,

> >  	if (ret)

> >  		goto error;

> >  

> > +	/* Initialize routing to single route to the fist source pad */

> 

> s/fist/first/

> s/pad/pad./

> 

> > +	ret = cal_camerarx_sd_set_routing(sd, &routing);

> > +	if (ret)

> > +		goto error;

> > +

> >  	ret = v4l2_device_register_subdev(&cal->v4l2_dev, sd);

> >  	if (ret)

> >  		goto error;

> > @@ -921,6 +983,7 @@ struct cal_camerarx *cal_camerarx_create(struct cal_dev *cal,

> >  	return phy;

> >  

> >  error:

> > +	v4l2_subdev_free_routing(&phy->routing);

> >  	media_entity_cleanup(&phy->subdev.entity);

> >  	kfree(phy);

> >  	return ERR_PTR(ret);

> > @@ -932,6 +995,7 @@ void cal_camerarx_destroy(struct cal_camerarx *phy)

> >  		return;

> >  

> >  	v4l2_device_unregister_subdev(&phy->subdev);

> > +	v4l2_subdev_free_routing(&phy->routing);

> >  	media_entity_cleanup(&phy->subdev.entity);

> >  	of_node_put(phy->source_ep_node);

> >  	of_node_put(phy->source_node);

> > diff --git a/drivers/media/platform/ti-vpe/cal.h b/drivers/media/platform/ti-vpe/cal.h

> > index 3aea444f8bf8..c96364eb0930 100644

> > --- a/drivers/media/platform/ti-vpe/cal.h

> > +++ b/drivers/media/platform/ti-vpe/cal.h

> > @@ -184,6 +184,8 @@ struct cal_camerarx {

> >  	struct mutex		mutex;

> >  

> >  	unsigned int enable_count;

> > +

> > +	struct v4l2_subdev_krouting routing;

> 

> Looking forward to storing this in v4l2_subdev_config, and embedding an

> instance of v4l2_subdev_config in v4l2_subdev :-)

> cal_camerarx_has_route() could then be replaced by a generic V4L2 subdev

> helper, and so could cal_camerarx_sd_get_routing().

> 

> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

> 

> >  };

> >  

> >  struct cal_dev {


-- 
Regards,

Laurent Pinchart
Laurent Pinchart April 29, 2021, 3:53 a.m. UTC | #4
Hi Tomi,

Thank you for the series. Nice to see the multiplexed streams API in
operation.

On Tue, Apr 27, 2021 at 04:20:24PM +0300, Tomi Valkeinen wrote:
> Hi,

> 

> This series adds embedded data and multiplexed streams support to TI CAL

> driver. These patches depend on the subdev routing series and CAL preparation

> series:

> 

> https://lore.kernel.org/linux-media/20210427124523.990938-1-tomi.valkeinen@ideasonboard.com/

> 

> https://lore.kernel.org/linux-media/20210420120433.902394-1-tomi.valkeinen@ideasonboard.com/

> 

> One detail I'm not too happy about is adding a new kernel module parameter

> 'cal_streams_api'. Setting this parameter enables the multiplexed streams

> support for the CAL sink pads. I'd like the driver to be able to figure this

> out itself, but there's no simple answer to it.

> 

> A pad in a subdev has to be either in non-multiplexed or multiplexed mode. If

> the subdev cannot support multiplexed streams, then the answer is clear. But if

> it can, then the question becomes "what's connected to the pad, and what does

> it want". Afaik, the idea with subdev configuration is that each subdev can be

> configured independently, so the driver trying to figure out the answer to that

> question breaks this idea. Also, two connected subdevs that both can support

> both modes wouldn't even give the answer, but the answer might be found further

> away, probably from the sensor subdev.

> 

> So, maybe it should be the userspace that can set a subdev pad to either of the

> modes. There's currently no such uAPI, but I think this option makes the most

> sense.


We'll need to brainstorm this a bit. I agree that the module parameter
isn't a good solution.

> Another detail to note is the embedded data support. I don't have hardware with

> real embedded data, so I have not added any embedded data formats. Also, the

> embedded data is captured with the video capture ioctls, not the metadata

> capture ioctls. The reason for this is that in the case of CSI-2 embedded data

> (at least for this hardware) the embedded data is very much like pixel data:

> the data is sent in lines just before or after the pixel content and each

> subdev needs a set_fmt call with the width, height, and mbus-code.

> 

> Using the metadata ioctls is possible, but it only results in much more complex

> driver code, and, I think, more confusing userspace.

> 

>  Tomi

> 

> Tomi Valkeinen (4):

>   media: ti-vpe: cal: add embedded data support

>   media: ti-vpe: cal: use frame desc to get vc and dt

>   media: ti-vpe: cal: add routing support

>   media: ti-vpe: cal: add multiplexed streams support

> 

>  drivers/media/platform/ti-vpe/cal-camerarx.c | 234 ++++++++++++++++---

>  drivers/media/platform/ti-vpe/cal-video.c    |  95 +++++++-

>  drivers/media/platform/ti-vpe/cal.c          | 102 ++++++--

>  drivers/media/platform/ti-vpe/cal.h          |  10 +-

>  4 files changed, 383 insertions(+), 58 deletions(-)


-- 
Regards,

Laurent Pinchart