If an entity has sink pads, at least one of them must be connected to another pad with an enabled link. If a driver with multiple sink pads has more strict requirements the check should be done in the driver itself. Just requiring one sink pad is connected with an enabled link is enough API-wise: entities with sink pads with only disabled links should not be allowed to stream in the first place, but also in a different operation mode a device might require only one of its pads connected with an active link. If an entity has an ability to function as a source entity another logical entity connected to the aforementioned one should be used for the purpose. Signed-off-by: Sakari Ailus <sakari.ailus@xxxxxx> --- drivers/media/media-entity.c | 16 +++++++++++++--- 1 files changed, 13 insertions(+), 3 deletions(-) diff --git a/drivers/media/media-entity.c b/drivers/media/media-entity.c index e1cd132..8846ea7 100644 --- a/drivers/media/media-entity.c +++ b/drivers/media/media-entity.c @@ -227,6 +227,7 @@ __must_check int media_entity_pipeline_start(struct media_entity *entity, media_entity_graph_walk_start(&graph, entity); while ((entity = media_entity_graph_walk_next(&graph))) { + bool has_sink = false, active_sink = false; unsigned int i; entity->stream_count++; @@ -243,18 +244,27 @@ __must_check int media_entity_pipeline_start(struct media_entity *entity, for (i = 0; i < entity->num_links; i++) { struct media_link *link = &entity->links[i]; + /* Are we the sink or not? */ + if (link->sink->entity != entity) + continue; + + has_sink = true; + /* Is this pad part of an enabled link? */ if (!(link->flags & MEDIA_LNK_FL_ENABLED)) continue; - /* Are we the sink or not? */ - if (link->sink->entity != entity) - continue; + active_sink = true; ret = entity->ops->link_validate(link); if (ret < 0 && ret != -ENOIOCTLCMD) goto error; } + + if (has_sink && !active_sink) { + ret = -EPIPE; + goto error; + } } mutex_unlock(&mdev->graph_mutex); -- 1.7.2.5 _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel