Hi Tomi, Laurent,
On Tue, Mar 01, 2022 at 06:11:54PM +0200, Tomi Valkeinen wrote:
> From: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
>
> Add a helper function to translate streams between two pads of a subdev,
> using the subdev's internal routing table.
>
Thanks for the patch! Please see my comments inline.
> Signed-off-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> Signed-off-by: Tomi Valkeinen <tomi.valkeinen@xxxxxxxxxxxxxxxx>
> ---
> drivers/media/v4l2-core/v4l2-subdev.c | 25 +++++++++++++++++++++++++
> include/media/v4l2-subdev.h | 23 +++++++++++++++++++++++
> 2 files changed, 48 insertions(+)
>
> diff --git a/drivers/media/v4l2-core/v4l2-subdev.c b/drivers/media/v4l2-core/v4l2-subdev.c
> index e30338a53088..6a9fc62dacbf 100644
> --- a/drivers/media/v4l2-core/v4l2-subdev.c
> +++ b/drivers/media/v4l2-core/v4l2-subdev.c
> @@ -1529,6 +1529,31 @@ v4l2_subdev_state_get_opposite_stream_format(struct v4l2_subdev_state *state,
> }
> EXPORT_SYMBOL_GPL(v4l2_subdev_state_get_opposite_stream_format);
>
> +u64 v4l2_subdev_state_xlate_streams(const struct v4l2_subdev_state *state,
> + u32 pad0, u32 pad1, u64 *streams)
> +{
> + const struct v4l2_subdev_krouting *routing = &state->routing;
> + struct v4l2_subdev_route *route;
> + u64 streams0 = 0;
> + u64 streams1 = 0;
> +
> + for_each_active_route(routing, route) {
> + if (route->sink_pad == pad0 && route->source_pad == pad1 &&
> + (*streams & BIT(route->sink_stream))) {
> + streams0 |= BIT(route->sink_stream);
> + streams1 |= BIT(route->source_stream);
> + }
> + if (route->source_pad == pad0 && route->sink_pad == pad1 &&
> + (*streams & BIT(route->source_stream))) {
> + streams0 |= BIT(route->source_stream);
> + streams1 |= BIT(route->sink_stream);
> + }
> + }
> +
> + *streams = streams0;
> + return streams1;
> +}
> +
> int v4l2_subdev_get_fmt(struct v4l2_subdev *sd, struct v4l2_subdev_state *state,
> struct v4l2_subdev_format *format)
> {
> diff --git a/include/media/v4l2-subdev.h b/include/media/v4l2-subdev.h
> index 1eced0f47296..992debe116ac 100644
> --- a/include/media/v4l2-subdev.h
> +++ b/include/media/v4l2-subdev.h
> @@ -1497,6 +1497,29 @@ struct v4l2_mbus_framefmt *
> v4l2_subdev_state_get_opposite_stream_format(struct v4l2_subdev_state *state,
> u32 pad, u32 stream);
>
> +/**
> + * v4l2_subdev_state_xlate_streams() - Translate streams from one pad to another
> + *
> + * @state: Subdevice state
> + * @pad0: The first pad
> + * @pad1: The second pad
> + * @streams: Streams bitmask on the first pad
> + *
> + * Streams on sink pads of a subdev are routed to source pads as expressed in
> + * the subdev state routing table. Stream numbers don't necessarily match on
> + * the sink and source side of a route. This function translates stream numbers
> + * on @pad0, expressed as a bitmask in @streams, to the corresponding streams
> + * on @pad1 using the routing table from the @state. It returns the stream mask
> + * on @pad1, and updates @streams with the streams that have been found in the
> + * routing table.
> + *
> + * @pad0 and @pad1 must be a sink and a source, in any order.
> + *
> + * Return: The bitmask of streams of @pad1 that are routed to @streams on @pad0.
It might be just me, but somehow I associate "translate" with turning a
name from one namespace into a corresponding name from another
namespace. In this case I initially assumed that this function is
supposed to accept stream IDs from pad0 and return corresponding
stream IDs from pad1. However, it looks like this is move like - "tell
me which streams on pad1 are connected to the given pad0 streams". Is
this correct?
If yes, maybe v4l2_subdev_state_get_routed_streams() be a better name?
Best regards,
Tomasz
