[v3,38/46] media: rkisp1: isp: Disallow multiple active sources

The ISP supports multiple source subdevs, but can only capture from a
single one at a time. The source is selected through link setup. The
driver finds the active source in its .s_stream() handler using the
media_entity_remote_pad() function. This fails to reject invalid
configurations with multiple active sources. Fix it by using the
media_pad_remote_pad_unique() helper instead, and inline
rkisp1_isp_get_source() in rkisp1_isp_s_stream() as the function is
small and has a single caller.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
---
Changes since v2:

- Update media_pad_remote_pad_unique() function name in commit message
---
 .../platform/rockchip/rkisp1/rkisp1-isp.c     | 30 ++++++++-----------
 1 file changed, 13 insertions(+), 17 deletions(-)

On 11.07.2022 15:42, Laurent Pinchart wrote:
>The ISP supports multiple source subdevs, but can only capture from a
>single one at a time. The source is selected through link setup. The
>driver finds the active source in its .s_stream() handler using the
>media_entity_remote_pad() function. This fails to reject invalid
>configurations with multiple active sources. Fix it by using the
>media_pad_remote_pad_unique() helper instead, and inline
>rkisp1_isp_get_source() in rkisp1_isp_s_stream() as the function is
>small and has a single caller.
>
>Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
>---
>Changes since v2:
>
>- Update media_pad_remote_pad_unique() function name in commit message
>---
> .../platform/rockchip/rkisp1/rkisp1-isp.c     | 30 ++++++++-----------
> 1 file changed, 13 insertions(+), 17 deletions(-)
>
>diff --git a/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c b/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c
>index 37623b73b1d9..d7e2802d11f5 100644
>--- a/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c
>+++ b/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c
>@@ -58,20 +58,6 @@
>  * Helpers
>  */
>
>-static struct v4l2_subdev *rkisp1_isp_get_source(struct v4l2_subdev *sd)
>-{
>-	struct media_pad *local, *remote;
>-	struct media_entity *sensor_me;
>-
>-	local = &sd->entity.pads[RKISP1_ISP_PAD_SINK_VIDEO];
>-	remote = media_pad_remote_pad_first(local);
>-	if (!remote)
>-		return NULL;
>-
>-	sensor_me = remote->entity;
>-	return media_entity_to_v4l2_subdev(sensor_me);
>-}
>-
> static struct v4l2_mbus_framefmt *
> rkisp1_isp_get_pad_fmt(struct rkisp1_isp *isp,
> 		       struct v4l2_subdev_state *sd_state,
>@@ -743,6 +729,8 @@ static int rkisp1_isp_s_stream(struct v4l2_subdev *sd, int enable)
> 	struct rkisp1_isp *isp = to_rkisp1_isp(sd);
> 	struct rkisp1_device *rkisp1 = isp->rkisp1;
> 	const struct rkisp1_sensor_async *asd;
>+	struct media_pad *source_pad;
>+	struct media_pad *sink_pad;
> 	int ret;
>
> 	if (!enable) {
>@@ -754,10 +742,18 @@ static int rkisp1_isp_s_stream(struct v4l2_subdev *sd, int enable)
> 		return 0;
> 	}
>
>-	rkisp1->source = rkisp1_isp_get_source(sd);

so 'rkisp1->source' is the source of the isp right? and in addition you later
add csi->source for the csi source. Maybe move rkisp1->source field to rkisp1_isp
to make it clear it is the isp source.
Just a suggestion. If you don't feel like, then:

Reviewed-by: Dafna Hirschfeld <dafna@fastmail.com>

>+	sink_pad = &isp->pads[RKISP1_ISP_PAD_SINK_VIDEO];
>+	source_pad = media_pad_remote_pad_unique(sink_pad);
>+	if (IS_ERR(source_pad)) {
>+		dev_dbg(rkisp1->dev, "Failed to get source for ISP: %ld\n",
>+			PTR_ERR(source_pad));
>+		return -EPIPE;
>+	}
>+
>+	rkisp1->source = media_entity_to_v4l2_subdev(source_pad->entity);
> 	if (!rkisp1->source) {
>-		dev_warn(rkisp1->dev, "No link between isp and source\n");
>-		return -ENODEV;
>+		/* This should really not happen, so is not worth a message. */
>+		return -EPIPE;
> 	}
>
> 	asd = container_of(rkisp1->source->asd, struct rkisp1_sensor_async,
>-- 
>Regards,
>
>Laurent Pinchart
>

Hi Dafna,

On Tue, Jul 12, 2022 at 09:40:05PM +0300, Dafna Hirschfeld wrote:
> On 11.07.2022 15:42, Laurent Pinchart wrote:
> > The ISP supports multiple source subdevs, but can only capture from a
> > single one at a time. The source is selected through link setup. The
> > driver finds the active source in its .s_stream() handler using the
> > media_entity_remote_pad() function. This fails to reject invalid
> > configurations with multiple active sources. Fix it by using the
> > media_pad_remote_pad_unique() helper instead, and inline
> > rkisp1_isp_get_source() in rkisp1_isp_s_stream() as the function is
> > small and has a single caller.
> > 
> > Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
> > ---
> > Changes since v2:
> > 
> > - Update media_pad_remote_pad_unique() function name in commit message
> > ---
> >  .../platform/rockchip/rkisp1/rkisp1-isp.c     | 30 ++++++++-----------
> >  1 file changed, 13 insertions(+), 17 deletions(-)
> > 
> > diff --git a/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c b/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c
> > index 37623b73b1d9..d7e2802d11f5 100644
> > --- a/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c
> > +++ b/drivers/media/platform/rockchip/rkisp1/rkisp1-isp.c
> > @@ -58,20 +58,6 @@
> >   * Helpers
> >   */
> > 
> > -static struct v4l2_subdev *rkisp1_isp_get_source(struct v4l2_subdev *sd)
> > -{
> > -	struct media_pad *local, *remote;
> > -	struct media_entity *sensor_me;
> > -
> > -	local = &sd->entity.pads[RKISP1_ISP_PAD_SINK_VIDEO];
> > -	remote = media_pad_remote_pad_first(local);
> > -	if (!remote)
> > -		return NULL;
> > -
> > -	sensor_me = remote->entity;
> > -	return media_entity_to_v4l2_subdev(sensor_me);
> > -}
> > -
> >  static struct v4l2_mbus_framefmt *
> >  rkisp1_isp_get_pad_fmt(struct rkisp1_isp *isp,
> >  		       struct v4l2_subdev_state *sd_state,
> > @@ -743,6 +729,8 @@ static int rkisp1_isp_s_stream(struct v4l2_subdev *sd, int enable)
> >  	struct rkisp1_isp *isp = to_rkisp1_isp(sd);
> >  	struct rkisp1_device *rkisp1 = isp->rkisp1;
> >  	const struct rkisp1_sensor_async *asd;
> > +	struct media_pad *source_pad;
> > +	struct media_pad *sink_pad;
> >  	int ret;
> > 
> >  	if (!enable) {
> > @@ -754,10 +742,18 @@ static int rkisp1_isp_s_stream(struct v4l2_subdev *sd, int enable)
> >  		return 0;
> >  	}
> > 
> > -	rkisp1->source = rkisp1_isp_get_source(sd);
> 
> so 'rkisp1->source' is the source of the isp right? and in addition you later
> add csi->source for the csi source. Maybe move rkisp1->source field to rkisp1_isp
> to make it clear it is the isp source.
> Just a suggestion. If you don't feel like, then:

That's a very good idea. I'll add a patch on top to do this.

> Reviewed-by: Dafna Hirschfeld <dafna@fastmail.com>
> 
> > +	sink_pad = &isp->pads[RKISP1_ISP_PAD_SINK_VIDEO];
> > +	source_pad = media_pad_remote_pad_unique(sink_pad);
> > +	if (IS_ERR(source_pad)) {
> > +		dev_dbg(rkisp1->dev, "Failed to get source for ISP: %ld\n",
> > +			PTR_ERR(source_pad));
> > +		return -EPIPE;
> > +	}
> > +
> > +	rkisp1->source = media_entity_to_v4l2_subdev(source_pad->entity);
> >  	if (!rkisp1->source) {
> > -		dev_warn(rkisp1->dev, "No link between isp and source\n");
> > -		return -ENODEV;
> > +		/* This should really not happen, so is not worth a message. */
> > +		return -EPIPE;
> >  	}
> > 
> >  	asd = container_of(rkisp1->source->asd, struct rkisp1_sensor_async,