All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
To: Sakari Ailus <sakari.ailus@iki.fi>
Cc: Linux Media Mailing List <linux-media@vger.kernel.org>,
	Mauro Carvalho Chehab <mchehab@infradead.org>,
	Shuah Khan <shuahkh@osg.samsung.com>,
	javier@osg.samsung.com
Subject: Re: [PATCH 4/5] [media] media-device: use kref for media_device instance
Date: Fri, 18 Mar 2016 08:17:01 -0300	[thread overview]
Message-ID: <20160318081701.221d9d10@recife.lan> (raw)
In-Reply-To: <20160317115053.GB11084@valkosipuli.retiisi.org.uk>

Em Thu, 17 Mar 2016 13:50:53 +0200
Sakari Ailus <sakari.ailus@iki.fi> escreveu:

> Hi Mauro,
> 
> On Wed, Mar 16, 2016 at 09:04:05AM -0300, Mauro Carvalho Chehab wrote:
> > Now that the media_device can be used by multiple drivers,
> > via devres, we need to be sure that it will be dropped only
> > when all drivers stop using it.  
> 
> Not long ago we split media device initialisation and registration into two.
> The intent with that was to prevent users from seeing the media device until
> all the initialisation is finished. I suppose that with dynamic updates, or
> media device being shared with two drivers, it might be difficult to achieve
> that. At least the method has to be different.
> 
> media_device_init() and media_device_cleanup() were a pair where the latter
> undid the first. 

Yes, this patch breaks the balance of _init()/_cleanup(), but this is for
a greater good. It is very hard to remove the devnode and cleanup the
media_device struct when multiple drivers can use it. The best way to
warrant that we'll be doing that is via kref.

> This patchs remove the requirement of calling cleanup
> explitly, breaking that model. It's perhaps not a big problem, there is
> likely no single driver which would initialise the media controller device
> once but would register and unregister it multiple times. I still wonder if
> we really lost something important if we removed media_device_init() and
> media_device_cleanup() altogether and merged the functionality in
> media_device_{,un}register().
> 
> Cc Javier who wrote the patch.

That patch was written to avoid a race trouble of having the media
devnode to initialize too early. Merging them back would cause troubles.

Maybe what we could do, instead, is to rename media_device_init()
to media_device_register(), and call the function that creates the
devnode as media_device_create_devnode().

What do you think?

> 
> commit 9832e155f1ed3030fdfaa19e72c06472dc2ecb1d
> Author: Javier Martinez Canillas <javier@osg.samsung.com>
> Date:   Fri Dec 11 20:57:08 2015 -0200
> 
>     [media] media-device: split media initialization and registration
> 
> For system-wide media device, I think we'd need to introduce a new concept,
> graph, that would define an interconnected set of entities. This would
> mainly be needed for callbacks, e.g. the power on / off sequences of the
> entities could be specific to V4L as discussed earlier. With this approach
> hacks wouldn't be needed to be made in the first place to support two drivers
> sharing a media device.
> 
> What was the original reason you needed to share the media device btw.?
> 
> > 
> > Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
> > ---
> >  drivers/media/media-device.c | 48 +++++++++++++++++++++++++++++++-------------
> >  include/media/media-device.h |  3 +++
> >  2 files changed, 37 insertions(+), 14 deletions(-)
> > 
> > diff --git a/drivers/media/media-device.c b/drivers/media/media-device.c
> > index c32fa15cc76e..38e6c319fe6e 100644
> > --- a/drivers/media/media-device.c
> > +++ b/drivers/media/media-device.c
> > @@ -721,6 +721,15 @@ int __must_check __media_device_register(struct media_device *mdev,
> >  {
> >  	int ret;
> >  
> > +	/* Check if mdev was ever registered at all */
> > +	mutex_lock(&mdev->graph_mutex);
> > +	if (media_devnode_is_registered(&mdev->devnode)) {
> > +		kref_get(&mdev->kref);
> > +		mutex_unlock(&mdev->graph_mutex);
> > +		return 0;
> > +	}
> > +	kref_init(&mdev->kref);
> > +
> >  	/* Register the device node. */
> >  	mdev->devnode.fops = &media_device_fops;
> >  	mdev->devnode.parent = mdev->dev;
> > @@ -730,8 +739,10 @@ int __must_check __media_device_register(struct media_device *mdev,
> >  	mdev->topology_version = 0;
> >  
> >  	ret = media_devnode_register(&mdev->devnode, owner);
> > -	if (ret < 0)
> > +	if (ret < 0) {
> > +		media_devnode_unregister(&mdev->devnode);
> >  		return ret;
> > +	}
> >  
> >  	ret = device_create_file(&mdev->devnode.dev, &dev_attr_model);
> >  	if (ret < 0) {
> > @@ -739,6 +750,7 @@ int __must_check __media_device_register(struct media_device *mdev,
> >  		return ret;
> >  	}
> >  
> > +	mutex_unlock(&mdev->graph_mutex);
> >  	dev_dbg(mdev->dev, "Media device registered\n");
> >  
> >  	return 0;
> > @@ -773,23 +785,15 @@ void media_device_unregister_entity_notify(struct media_device *mdev,
> >  }
> >  EXPORT_SYMBOL_GPL(media_device_unregister_entity_notify);
> >  
> > -void media_device_unregister(struct media_device *mdev)
> > +static void do_media_device_unregister(struct kref *kref)
> >  {
> > +	struct media_device *mdev;
> >  	struct media_entity *entity;
> >  	struct media_entity *next;
> >  	struct media_interface *intf, *tmp_intf;
> >  	struct media_entity_notify *notify, *nextp;
> >  
> > -	if (mdev == NULL)
> > -		return;
> > -
> > -	mutex_lock(&mdev->graph_mutex);
> > -
> > -	/* Check if mdev was ever registered at all */
> > -	if (!media_devnode_is_registered(&mdev->devnode)) {
> > -		mutex_unlock(&mdev->graph_mutex);
> > -		return;
> > -	}
> > +	mdev = container_of(kref, struct media_device, kref);
> >  
> >  	/* Remove all entities from the media device */
> >  	list_for_each_entry_safe(entity, next, &mdev->entities, graph_obj.list)
> > @@ -807,13 +811,26 @@ void media_device_unregister(struct media_device *mdev)
> >  		kfree(intf);
> >  	}
> >  
> > -	mutex_unlock(&mdev->graph_mutex);
> > +	/* Check if mdev devnode was registered */
> > +	if (!media_devnode_is_registered(&mdev->devnode))
> > +		return;
> >  
> >  	device_remove_file(&mdev->devnode.dev, &dev_attr_model);
> >  	media_devnode_unregister(&mdev->devnode);
> >  
> >  	dev_dbg(mdev->dev, "Media device unregistered\n");
> >  }
> > +
> > +void media_device_unregister(struct media_device *mdev)
> > +{
> > +	if (mdev == NULL)
> > +		return;
> > +
> > +	mutex_lock(&mdev->graph_mutex);
> > +	kref_put(&mdev->kref, do_media_device_unregister);
> > +	mutex_unlock(&mdev->graph_mutex);
> > +
> > +}
> >  EXPORT_SYMBOL_GPL(media_device_unregister);
> >  
> >  static void media_device_release_devres(struct device *dev, void *res)
> > @@ -825,13 +842,16 @@ struct media_device *media_device_get_devres(struct device *dev)
> >  	struct media_device *mdev;
> >  
> >  	mdev = devres_find(dev, media_device_release_devres, NULL, NULL);
> > -	if (mdev)
> > +	if (mdev) {
> > +		kref_get(&mdev->kref);
> >  		return mdev;
> > +	}
> >  
> >  	mdev = devres_alloc(media_device_release_devres,
> >  				sizeof(struct media_device), GFP_KERNEL);
> >  	if (!mdev)
> >  		return NULL;
> > +
> >  	return devres_get(dev, mdev, NULL, NULL);
> >  }
> >  EXPORT_SYMBOL_GPL(media_device_get_devres);
> > diff --git a/include/media/media-device.h b/include/media/media-device.h
> > index ca3871b853ba..73c16e6e6b6b 100644
> > --- a/include/media/media-device.h
> > +++ b/include/media/media-device.h
> > @@ -23,6 +23,7 @@
> >  #ifndef _MEDIA_DEVICE_H
> >  #define _MEDIA_DEVICE_H
> >  
> > +#include <linux/kref.h>
> >  #include <linux/list.h>
> >  #include <linux/mutex.h>
> >  
> > @@ -283,6 +284,7 @@ struct media_entity_notify {
> >   * struct media_device - Media device
> >   * @dev:	Parent device
> >   * @devnode:	Media device node
> > + * @kref:	Object refcount
> >   * @driver_name: Optional device driver name. If not set, calls to
> >   *		%MEDIA_IOC_DEVICE_INFO will return dev->driver->name.
> >   *		This is needed for USB drivers for example, as otherwise
> > @@ -347,6 +349,7 @@ struct media_device {
> >  	/* dev->driver_data points to this struct. */
> >  	struct device *dev;
> >  	struct media_devnode devnode;
> > +	struct kref kref;
> >  
> >  	char model[32];
> >  	char driver_name[32];
> > -- 
> > 2.5.0
> >   
> 


-- 
Thanks,
Mauro

  reply	other threads:[~2016-03-18 11:17 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-16 12:04 [PATCH 1/5] [media] media-device: get rid of the spinlock Mauro Carvalho Chehab
2016-03-16 12:04 ` [PATCH 2/5] [media] media-device: Fix a comment Mauro Carvalho Chehab
2016-03-16 12:58   ` Javier Martinez Canillas
2016-03-16 12:04 ` [PATCH 3/5] [media] au0828: Unregister notifiers Mauro Carvalho Chehab
2016-03-16 13:11   ` Javier Martinez Canillas
2016-03-22 19:55   ` Shuah Khan
2016-03-16 12:04 ` [PATCH 4/5] [media] media-device: use kref for media_device instance Mauro Carvalho Chehab
2016-03-16 13:23   ` Javier Martinez Canillas
2016-03-16 14:05   ` Shuah Khan
2016-03-16 14:25     ` Mauro Carvalho Chehab
2016-03-16 14:32       ` Shuah Khan
2016-03-17 11:50   ` Sakari Ailus
2016-03-18 11:17     ` Mauro Carvalho Chehab [this message]
2016-03-18 13:10   ` Sakari Ailus
2016-03-22 19:56   ` Shuah Khan
2016-03-22 20:07     ` Shuah Khan
2016-03-16 12:04 ` [PATCH 5/5] [media] media-device: make media_device_cleanup() static Mauro Carvalho Chehab
2016-03-16 12:04   ` Mauro Carvalho Chehab
2016-03-16 12:04   ` Mauro Carvalho Chehab
2016-03-16 14:03   ` Javier Martinez Canillas
2016-03-16 14:03     ` Javier Martinez Canillas
2016-03-16 14:03     ` Javier Martinez Canillas
2016-03-16 14:36     ` Mauro Carvalho Chehab
2016-03-16 14:36       ` Mauro Carvalho Chehab
2016-03-16 14:36       ` Mauro Carvalho Chehab
2016-03-16 14:38       ` Javier Martinez Canillas
2016-03-16 14:38         ` Javier Martinez Canillas
2016-03-16 14:38         ` Javier Martinez Canillas
2016-03-22 19:57   ` Shuah Khan
2016-03-22 19:57     ` Shuah Khan
2016-03-22 19:57     ` Shuah Khan
2016-03-16 12:53 ` [PATCH 1/5] [media] media-device: get rid of the spinlock Javier Martinez Canillas
2016-03-16 13:10   ` Mauro Carvalho Chehab
2016-03-16 14:10 ` Sakari Ailus
2016-03-22 19:52 ` Shuah Khan
2016-03-24 22:11 ` 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=20160318081701.221d9d10@recife.lan \
    --to=mchehab@osg.samsung.com \
    --cc=javier@osg.samsung.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@infradead.org \
    --cc=sakari.ailus@iki.fi \
    --cc=shuahkh@osg.samsung.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.