From: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
To: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Cc: linux-media@vger.kernel.org, Sakari Ailus <sakari.ailus@iki.fi>,
Michal Simek <michal.simek@xilinx.com>,
Sylwester Nawrocki <s.nawrocki@samsung.com>
Subject: Re: [PATCH v1 1/5] media: mc: entity: Add pad iterator for media_pipeline
Date: Thu, 15 Dec 2022 14:48:58 +0200 [thread overview]
Message-ID: <Y5sXus6z6h1tf18b@pendragon.ideasonboard.com> (raw)
In-Reply-To: <fea9c172-65b6-8067-3957-13fbf77de6ff@ideasonboard.com>
Hi Tomi,
On Thu, Dec 15, 2022 at 02:33:43PM +0200, Tomi Valkeinen wrote:
> On 12/12/2022 16:16, Laurent Pinchart wrote:
> > Add a media_pipeline_for_each_pad() macro to iterate over pads in a
> > pipeline. This should be used by driver as a replacement of the
> > media_graph_walk API, as iterating over the media_pipeline uses the
> > cached list of pads and is thus more efficient.
> >
> > Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> > ---
> > drivers/media/mc/mc-entity.c | 18 ++++++++++++++++++
> > include/media/media-entity.h | 29 +++++++++++++++++++++++++++++
> > 2 files changed, 47 insertions(+)
> >
> > diff --git a/drivers/media/mc/mc-entity.c b/drivers/media/mc/mc-entity.c
> > index b8bcbc734eaf..70df2050951c 100644
> > --- a/drivers/media/mc/mc-entity.c
> > +++ b/drivers/media/mc/mc-entity.c
> > @@ -932,6 +932,24 @@ __must_check int media_pipeline_alloc_start(struct media_pad *pad)
> > }
> > EXPORT_SYMBOL_GPL(media_pipeline_alloc_start);
> >
> > +struct media_pad *
> > +__media_pipeline_pad_iter_next(struct media_pipeline *pipe,
> > + struct media_pipeline_pad_iter *iter,
> > + struct media_pad *pad)
> > +{
> > + if (!pad)
> > + iter->cursor = pipe->pads.next;
> > +
> > + if (iter->cursor == &pipe->pads)
> > + return NULL;
> > +
> > + pad = list_entry(iter->cursor, struct media_pipeline_pad, list)->pad;
> > + iter->cursor = iter->cursor->next;
> > +
> > + return pad;
> > +}
> > +EXPORT_SYMBOL_GPL(__media_pipeline_pad_iter_next);
> > +
> > /* -----------------------------------------------------------------------------
> > * Links management
> > */
> > diff --git a/include/media/media-entity.h b/include/media/media-entity.h
> > index 85ed08ddee9d..e881e483b550 100644
> > --- a/include/media/media-entity.h
> > +++ b/include/media/media-entity.h
> > @@ -130,6 +130,15 @@ struct media_pipeline_pad {
> > struct media_pad *pad;
> > };
> >
> > +/**
> > + * struct media_pipeline_pad_iter - Iterator for media_pipeline_for_each_pad
> > + *
> > + * @cursor: The current element
> > + */
> > +struct media_pipeline_pad_iter {
> > + struct list_head *cursor;
> > +};
> > +
>
> Is there any reason to have this iter struct in a public header, vs.
> having it in mc-entity.c?
It has to be instantiated on the stack by the user of the
media_pipeline_for_each_pad() macro. A typical usage is
struct media_pipeline_pad_iter iter;
struct media_pad *pad
media_pipeline_for_each_pad(pipe, &iter, pad) {
...
};
Note how iter is not a pointer.
> > /**
> > * struct media_link - A link object part of a media graph.
> > *
> > @@ -1163,6 +1172,26 @@ void media_pipeline_stop(struct media_pad *pad);
> > */
> > void __media_pipeline_stop(struct media_pad *pad);
> >
> > +struct media_pad *
> > +__media_pipeline_pad_iter_next(struct media_pipeline *pipe,
> > + struct media_pipeline_pad_iter *iter,
> > + struct media_pad *pad);
> > +
> > +/**
> > + * media_pipeline_for_each_pad - Iterate on all pads in a media pipeline
> > + * @pipe: The pipeline
> > + * @iter: The iterator (struct media_pipeline_pad_iter)
> > + * @pad: The iterator pad
>
> If I understand this correctly, both iter and pad are just variables the
> macro will use. In other words, they are not used to pass any values.
>
> Would it be better to declare those variables in the macro itself? Well,
> that has its downsides. But perhaps at least clarify in the doc that
> they are only variables used by the loop, and do not need to be initialized.
Now that the kernel uses C99, I suppose we could make the pad variable
locally declared within the loop:
#define media_pipeline_for_each_pad(pipe, pad) \
for (struct media_pipeline_pad *pad = __media_pipeline_pad_iter_next((pipe), iter, NULL); \
pad != NULL; \
pad = __media_pipeline_pad_iter_next((pipe), iter, pad))
Hiding the iter variable would be more difficult, as I don't think you
can declare multiple variables of different types.
I'm a bit concerned about backporting though, so I'd rather not do this
in this patch, but on top.
> > + * Iterate on all pads in a media pipeline. This is only valid after the
> > + * pipeline has been built with media_pipeline_start() and before it gets
> > + * destroyed with media_pipeline_stop().
> > + */
> > +#define media_pipeline_for_each_pad(pipe, iter, pad) \
> > + for (pad = __media_pipeline_pad_iter_next((pipe), iter, NULL); \
> > + pad != NULL; \
> > + pad = __media_pipeline_pad_iter_next((pipe), iter, pad))
> > +
> > /**
> > * media_pipeline_alloc_start - Mark a pipeline as streaming
> > * @pad: Starting pad
--
Regards,
Laurent Pinchart
next prev parent reply other threads:[~2022-12-15 12:49 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-12 14:16 [PATCH v1 0/5] media: Replace media graph walk in several drivers Laurent Pinchart
2022-12-12 14:16 ` [PATCH v1 1/5] media: mc: entity: Add pad iterator for media_pipeline Laurent Pinchart
2022-12-15 12:33 ` Tomi Valkeinen
2022-12-15 12:48 ` Laurent Pinchart [this message]
2022-12-15 12:55 ` Tomi Valkeinen
2022-12-12 14:16 ` [PATCH v1 2/5] media: mc: entity: Add entity " Laurent Pinchart
2022-12-15 12:48 ` Tomi Valkeinen
2022-12-15 12:51 ` Laurent Pinchart
2022-12-12 14:16 ` [PATCH v1 3/5] media: ti: omap3isp: Use media_pipeline_for_each_entity() Laurent Pinchart
2022-12-15 13:16 ` Tomi Valkeinen
2022-12-12 14:16 ` [PATCH v1 4/5] media: ti: omap4iss: " Laurent Pinchart
2022-12-12 14:16 ` [PATCH v1 5/5] media: xilinx: dma: Use media_pipeline_for_each_pad() Laurent Pinchart
2022-12-15 13:29 ` Tomi Valkeinen
2022-12-16 0:36 ` Laurent Pinchart
2022-12-16 6:47 ` Tomi Valkeinen
2022-12-19 22:22 ` Laurent Pinchart
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Y5sXus6z6h1tf18b@pendragon.ideasonboard.com \
--to=laurent.pinchart@ideasonboard.com \
--cc=linux-media@vger.kernel.org \
--cc=michal.simek@xilinx.com \
--cc=s.nawrocki@samsung.com \
--cc=sakari.ailus@iki.fi \
--cc=tomi.valkeinen@ideasonboard.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.