Re: [RFC PATCH v8 1/4] media: Media Device Allocator API

[shuah <shuah@kernel.org>]

Hi Hans,

On 11/20/18 4:20 AM, Hans Verkuil wrote:
> On 11/02/2018 01:31 AM, shuah@kernel.org wrote:
>> From: Shuah Khan <shuah@kernel.org>
>>
>> Media Device Allocator API to allows multiple drivers share a media device.
>> Using this API, drivers can allocate a media device with the shared struct
>> device as the key. Once the media device is allocated by a driver, other
>> drivers can get a reference to it. The media device is released when all
>> the references are released.
>>
>> Signed-off-by: Shuah Khan <shuah@kernel.org>

Thanks for catching the documentation corrections. Fixed them all and
working on the next revision of the patch.

>> ---
>>   Documentation/media/kapi/mc-core.rst |  37 ++++++++
>>   drivers/media/Makefile               |   3 +-
>>   drivers/media/media-dev-allocator.c  | 132 +++++++++++++++++++++++++++
>>   include/media/media-dev-allocator.h  |  53 +++++++++++
>>   4 files changed, 224 insertions(+), 1 deletion(-)
>>   create mode 100644 drivers/media/media-dev-allocator.c
>>   create mode 100644 include/media/media-dev-allocator.h
>>
>> diff --git a/Documentation/media/kapi/mc-core.rst b/Documentation/media/kapi/mc-core.rst
>> index 0c05503eaf1f..d6f409598065 100644
>> --- a/Documentation/media/kapi/mc-core.rst
>> +++ b/Documentation/media/kapi/mc-core.rst
>> @@ -257,8 +257,45 @@ Subsystems should facilitate link validation by providing subsystem specific
>>   helper functions to provide easy access for commonly needed information, and
>>   in the end provide a way to use driver-specific callbacks.
>>   
>> +Media Controller Device Allocator API
>> +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>> +
>> +When media device belongs to more than one driver, the shared media device
> 
> When -> When the
> 
>> +is allocated with the shared struct device as the key for look ups.
>> +
>> +Shared media device should stay in registered state until the last driver
> 
> Shared -> The shared
> 
>> +unregisters it. In addition, media device should be released when all the
> 
> media -> the media
> 
>> +references are released. Each driver gets a reference to the media device
>> +during probe, when it allocates the media device. If media device is already
>> +allocated, allocate API bumps up the refcount and return the existing media
> 
> allocate -> the allocate
> return -> returns
> 
>> +device. Driver puts the reference back from its disconnect routine when it
> 
> Driver -> The driver
> from -> in
> 
>> +calls :c:func:`media_device_delete()`.
>> +
>> +Media device is unregistered and cleaned up from the kref put handler to
> 
> Media -> The media
> from -> in
> 
>> +ensure that the media device stays in registered state until the last driver
>> +unregisters the media device.
> 
> What happens if an application still has the media device open and you forcibly
> remove the last driver? I think it should be OK since the media_devnode struct
> isn't freed until the last open filehandle closes. But it is good to test this.
> 
>> +
>> +**Driver Usage**
>> +
>> +Drivers should use the media-core routines to get register reference and
> 
> 'get register reference'? Not sure what you meant to say.
> 
>> +call :c:func:`media_device_delete()` routine to make sure the shared media
>> +device delete is handled correctly.
>> +
>> +**driver probe:**
>> +Call :c:func:`media_device_usb_allocate()` to allocate or get a reference
>> +Call :c:func:`media_device_register()`, if media devnode isn't registered
>> +
>> +**driver disconnect:**
>> +Call :c:func:`media_device_delete()` to free the media_device. Free'ing is
> 
> Free'ing -> Freeing
> 
>> +handled by the kref put handler.
>> +
>> +API Definitions
>> +^^^^^^^^^^^^^^^
>> +
>>   .. kernel-doc:: include/media/media-device.h
>>   
>>   .. kernel-doc:: include/media/media-devnode.h
>>   
>>   .. kernel-doc:: include/media/media-entity.h
>> +
>> +.. kernel-doc:: include/media/media-dev-allocator.h
>> diff --git a/drivers/media/Makefile b/drivers/media/Makefile
>> index 594b462ddf0e..8608f0a41dca 100644
>> --- a/drivers/media/Makefile
>> +++ b/drivers/media/Makefile
>> @@ -3,7 +3,8 @@
>>   # Makefile for the kernel multimedia device drivers.
>>   #
>>   
>> -media-objs	:= media-device.o media-devnode.o media-entity.o
>> +media-objs	:= media-device.o media-devnode.o media-entity.o \
>> +		   media-dev-allocator.o
> 
> Perhaps only add media-dev-allocator if CONFIG_USB is enabled?
> 
>>   
>>   #
>>   # I2C drivers should come before other drivers, otherwise they'll fail
>> diff --git a/drivers/media/media-dev-allocator.c b/drivers/media/media-dev-allocator.c
>> new file mode 100644
>> index 000000000000..262d1293dc13
>> --- /dev/null
>> +++ b/drivers/media/media-dev-allocator.c
>> @@ -0,0 +1,132 @@
>> +// SPDX-License-Identifier: GPL-2.0-or-later
>> +/*
>> + * media-dev-allocator.c - Media Controller Device Allocator API
>> + *
>> + * Copyright (c) 2018 Shuah Khan <shuah@kernel.org>
>> + *
>> + * Credits: Suggested by Laurent Pinchart <laurent.pinchart@ideasonboard.com>
>> + */
>> +
>> +/*
>> + * This file adds a global refcounted Media Controller Device Instance API.
>> + * A system wide global media device list is managed and each media device
>> + * includes a kref count. The last put on the media device releases the media
>> + * device instance.
>> + *
>> + */
>> +
>> +#include <linux/kref.h>
>> +#include <linux/module.h>
>> +#include <linux/slab.h>
>> +#include <linux/usb.h>
>> +
>> +#include <media/media-device.h>
>> +
>> +static LIST_HEAD(media_device_list);
>> +static DEFINE_MUTEX(media_device_lock);
>> +
>> +struct media_device_instance {
>> +	struct media_device mdev;
>> +	struct module *owner;
>> +	struct list_head list;
>> +	struct kref refcount;
>> +};
>> +
>> +static inline struct media_device_instance *
>> +to_media_device_instance(struct media_device *mdev)
>> +{
>> +	return container_of(mdev, struct media_device_instance, mdev);
>> +}
>> +
>> +static void media_device_instance_release(struct kref *kref)
>> +{
>> +	struct media_device_instance *mdi =
>> +		container_of(kref, struct media_device_instance, refcount);
>> +
>> +	dev_dbg(mdi->mdev.dev, "%s: mdev=%p\n", __func__, &mdi->mdev);
>> +
>> +	mutex_lock(&media_device_lock);
> 
> Can't the lock be moved down to just before list_del? Or am I missing something?
> 

The lock needs to be held while media_device_unregister() and 
media_device_cleanup() to avoid races in the unregister path if two 
drivers call media_device_delete().

>> +
>> +	media_device_unregister(&mdi->mdev);
>> +	media_device_cleanup(&mdi->mdev);
>> +
>> +	list_del(&mdi->list);
>> +	mutex_unlock(&media_device_lock);
>> +
>> +	kfree(mdi);
>> +}
>> +
>> +/* Callers should hold media_device_lock when calling this function */
>> +static struct media_device *__media_device_get(struct device *dev,
>> +					       char *module_name)
> 
> const char *

Thanks fixed all of these.

I am running final tests and will send the next version this week.

thanks,
-- Shuah