On Tue, Sep 19, 2023 at 10:24:32AM +0000, Sakari Ailus wrote:
> On Mon, Sep 18, 2023 at 04:48:02PM +0300, Laurent Pinchart wrote:
> > On Mon, Sep 18, 2023 at 03:51:38PM +0300, Sakari Ailus wrote:
> > > Check the validity of pad flags on entity init. Exactly one of the flags
> > > must be set.
> > >
> > > Signed-off-by: Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx>
> > > ---
> > > drivers/media/mc/mc-entity.c | 20 ++++++++++++++++++--
> > > 1 file changed, 18 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/media/mc/mc-entity.c b/drivers/media/mc/mc-entity.c
> > > index 83468d4a440b..38d5bbae33d7 100644
> > > --- a/drivers/media/mc/mc-entity.c
> > > +++ b/drivers/media/mc/mc-entity.c
> > > @@ -195,8 +195,9 @@ int media_entity_pads_init(struct media_entity *entity, u16 num_pads,
> > > struct media_pad *pads)
> > > {
> > > struct media_device *mdev = entity->graph_obj.mdev;
> > > - struct media_pad *iter;
> > > + struct media_pad *iter, *iter2;
> > > unsigned int i = 0;
> > > + int ret = 0;
> > >
> > > if (num_pads >= MEDIA_ENTITY_MAX_PADS)
> > > return -E2BIG;
> > > @@ -210,15 +211,30 @@ int media_entity_pads_init(struct media_entity *entity, u16 num_pads,
> > > media_entity_for_each_pad(entity, iter) {
> > > iter->entity = entity;
> > > iter->index = i++;
> > > +
> > > + if (hweight32(iter->flags & (MEDIA_PAD_FL_SINK |
> > > + MEDIA_PAD_FL_SOURCE)) != 1) {
> > > + ret = -EINVAL;
> > > + break;
> > > + }
> > > +
> > > if (mdev)
> > > media_gobj_create(mdev, MEDIA_GRAPH_PAD,
> > > &iter->graph_obj);
> > > }
> > >
> > > + if (ret && mdev) {
> > > + media_entity_for_each_pad(entity, iter2) {
> > > + if (iter2 == iter)
> > > + break;
> > > + media_gobj_destroy(&iter->graph_obj);
> >
> > Wrong iterator.
> >
> > Instead of using a second iterator, which can be error-prone as shown
> > here, how about breaking when !iter->graph_obj.mdev, which indicates an
> > uninitialized object ?
>
> media_gobj_destroy() is actually a no-op if it hasn't been initialised
> (created).
I'm fine calling it unconditionally too.
> It'd be still cleaner to make the pad flag checks first and only then call
> media_gobj_create() on them. But if someone needs something more in the
> future, then that doesn't mix as well anymore.
>
> Both are fine IMO.
>
> > > + }
> > > + }
> > > +
> > > if (mdev)
> > > mutex_unlock(&mdev->graph_mutex);
> > >
> > > - return 0;
> > > + return ret;
> > > }
> > > EXPORT_SYMBOL_GPL(media_entity_pads_init);
> > >
--
Regards,
Laurent Pinchart
