| Message ID | 1379541668-23085-3-git-send-email-sakari.ailus@iki.fi (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Sakari, Thank you for the patch. On Thursday 19 September 2013 01:01:06 Sakari Ailus wrote: > Do not allow streaming if a pad with MEDIA_PAD_FL_MUST_CONNECT flag is not > connected by an active link. > > This patch makes it possible to avoid drivers having to check for the most > common case of link state validation: a sink pad that must be connected. > > Signed-off-by: Sakari Ailus <sakari.ailus@iki.fi> > Tested-by: Sylwester Nawrocki <sylvester.nawrocki@gmail.com> > --- > drivers/media/media-entity.c | 41 ++++++++++++++++++++++++++++++++------- > 1 file changed, 34 insertions(+), 7 deletions(-) > > diff --git a/drivers/media/media-entity.c b/drivers/media/media-entity.c > index 2c286c3..a996e0a 100644 > --- a/drivers/media/media-entity.c > +++ b/drivers/media/media-entity.c > @@ -235,6 +235,8 @@ __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))) { > + DECLARE_BITMAP(active, entity->num_pads); > + DECLARE_BITMAP(has_no_links, entity->num_pads); > unsigned int i; > > entity->stream_count++; > @@ -248,21 +250,46 @@ __must_check int media_entity_pipeline_start(struct > media_entity *entity, if (!entity->ops || !entity->ops->link_validate) > continue; > > + bitmap_zero(active, entity->num_pads); > + bitmap_fill(has_no_links, entity->num_pads); > + > for (i = 0; i < entity->num_links; i++) { > struct media_link *link = &entity->links[i]; > - > - /* 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) > + struct media_pad *pad = link->sink->entity == entity > + ? link->sink : link->source; What about aligning the ? to the = ? With that change, Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> > + > + /* Mark that a pad is connected by a link. */ > + bitmap_clear(has_no_links, pad->index, 1); > + > + /* > + * Pads that either do not need to connect or > + * are connected through an enabled link are > + * fine. > + */ > + if (!(pad->flags & MEDIA_PAD_FL_MUST_CONNECT) || > + link->flags & MEDIA_LNK_FL_ENABLED) > + bitmap_set(active, pad->index, 1); > + > + /* > + * Link validation will only take place for > + * sink ends of the link that are enabled. > + */ > + if (link->sink != pad || > + !(link->flags & MEDIA_LNK_FL_ENABLED)) > continue; > > ret = entity->ops->link_validate(link); > if (ret < 0 && ret != -ENOIOCTLCMD) > goto error; > } > + > + /* Either no links or validated links are fine. */ > + bitmap_or(active, active, has_no_links, entity->num_pads); > + > + if (!bitmap_full(active, entity->num_pads)) { > + ret = -EPIPE; > + goto error; > + } > } > > mutex_unlock(&mdev->graph_mutex);
On Fri, Sep 20, 2013 at 10:54:22PM +0200, Laurent Pinchart wrote: > > @@ -248,21 +250,46 @@ __must_check int media_entity_pipeline_start(struct > > media_entity *entity, if (!entity->ops || !entity->ops->link_validate) > > continue; > > > > + bitmap_zero(active, entity->num_pads); > > + bitmap_fill(has_no_links, entity->num_pads); > > + > > for (i = 0; i < entity->num_links; i++) { > > struct media_link *link = &entity->links[i]; > > - > > - /* 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) > > + struct media_pad *pad = link->sink->entity == entity > > + ? link->sink : link->source; > > What about aligning the ? to the = ? With that change, How about to the beginning of the next operand rather than "="? (Think of adding parentheses around the rvalue of "=".) I think it's fine as it was, but if it's to be changed then it should be aligned to link->sink->entity IMHO. :-)
Hi Sakari, On Monday 23 September 2013 22:57:02 Sakari Ailus wrote: > On Fri, Sep 20, 2013 at 10:54:22PM +0200, Laurent Pinchart wrote: > > > @@ -248,21 +250,46 @@ __must_check int > > > media_entity_pipeline_start(struct media_entity *entity, > > > if (!entity->ops || !entity->ops->link_validate) > > > continue; > > > > > > + bitmap_zero(active, entity->num_pads); > > > + bitmap_fill(has_no_links, entity->num_pads); > > > + > > > for (i = 0; i < entity->num_links; i++) { > > > struct media_link *link = &entity->links[i]; > > > - > > > - /* 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) > > > + struct media_pad *pad = link->sink->entity == entity > > > + ? link->sink : link->source; > > > > What about aligning the ? to the = ? With that change, > > How about to the beginning of the next operand rather than "="? > > (Think of adding parentheses around the rvalue of "=".) > > I think it's fine as it was, but if it's to be changed then it should be > aligned to link->sink->entity IMHO. :-) My preference goes for aligning the ? under the =, but I agree it's not logical from an rvalue point of view :-) I would favor aligning the ? under the l of link, but enough bikeshedding for now, please pick whichever solution you prefer :-)
Hi Laurent, On Tue, Oct 01, 2013 at 02:39:14PM +0200, Laurent Pinchart wrote: > On Monday 23 September 2013 22:57:02 Sakari Ailus wrote: > > On Fri, Sep 20, 2013 at 10:54:22PM +0200, Laurent Pinchart wrote: > > > > @@ -248,21 +250,46 @@ __must_check int > > > > media_entity_pipeline_start(struct media_entity *entity, > > > > if (!entity->ops || !entity->ops->link_validate) > > > > continue; > > > > > > > > + bitmap_zero(active, entity->num_pads); > > > > + bitmap_fill(has_no_links, entity->num_pads); > > > > + > > > > for (i = 0; i < entity->num_links; i++) { > > > > struct media_link *link = &entity->links[i]; > > > > - > > > > - /* 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) > > > > + struct media_pad *pad = link->sink->entity == entity > > > > + ? link->sink : link->source; > > > > > > What about aligning the ? to the = ? With that change, > > > > How about to the beginning of the next operand rather than "="? > > > > (Think of adding parentheses around the rvalue of "=".) > > > > I think it's fine as it was, but if it's to be changed then it should be > > aligned to link->sink->entity IMHO. :-) > > My preference goes for aligning the ? under the =, but I agree it's not > logical from an rvalue point of view :-) I would favor aligning the ? under > the l of link, but enough bikeshedding for now, please pick whichever solution > you prefer :-) ? goes under l of the link; agreed. I'll resend the set.
diff --git a/drivers/media/media-entity.c b/drivers/media/media-entity.c index 2c286c3..a996e0a 100644 --- a/drivers/media/media-entity.c +++ b/drivers/media/media-entity.c @@ -235,6 +235,8 @@ __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))) { + DECLARE_BITMAP(active, entity->num_pads); + DECLARE_BITMAP(has_no_links, entity->num_pads); unsigned int i; entity->stream_count++; @@ -248,21 +250,46 @@ __must_check int media_entity_pipeline_start(struct media_entity *entity, if (!entity->ops || !entity->ops->link_validate) continue; + bitmap_zero(active, entity->num_pads); + bitmap_fill(has_no_links, entity->num_pads); + for (i = 0; i < entity->num_links; i++) { struct media_link *link = &entity->links[i]; - - /* 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) + struct media_pad *pad = link->sink->entity == entity + ? link->sink : link->source; + + /* Mark that a pad is connected by a link. */ + bitmap_clear(has_no_links, pad->index, 1); + + /* + * Pads that either do not need to connect or + * are connected through an enabled link are + * fine. + */ + if (!(pad->flags & MEDIA_PAD_FL_MUST_CONNECT) || + link->flags & MEDIA_LNK_FL_ENABLED) + bitmap_set(active, pad->index, 1); + + /* + * Link validation will only take place for + * sink ends of the link that are enabled. + */ + if (link->sink != pad || + !(link->flags & MEDIA_LNK_FL_ENABLED)) continue; ret = entity->ops->link_validate(link); if (ret < 0 && ret != -ENOIOCTLCMD) goto error; } + + /* Either no links or validated links are fine. */ + bitmap_or(active, active, has_no_links, entity->num_pads); + + if (!bitmap_full(active, entity->num_pads)) { + ret = -EPIPE; + goto error; + } } mutex_unlock(&mdev->graph_mutex);