About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Documentation / video4linux / v4l2-framework.txt


Based on kernel version 4.7.2. Page generated on 2016-08-22 22:48 EST.

1	Overview of the V4L2 driver framework
2	=====================================
3	
4	This text documents the various structures provided by the V4L2 framework and
5	their relationships.
6	
7	
8	Introduction
9	------------
10	
11	The V4L2 drivers tend to be very complex due to the complexity of the
12	hardware: most devices have multiple ICs, export multiple device nodes in
13	/dev, and create also non-V4L2 devices such as DVB, ALSA, FB, I2C and input
14	(IR) devices.
15	
16	Especially the fact that V4L2 drivers have to setup supporting ICs to
17	do audio/video muxing/encoding/decoding makes it more complex than most.
18	Usually these ICs are connected to the main bridge driver through one or
19	more I2C busses, but other busses can also be used. Such devices are
20	called 'sub-devices'.
21	
22	For a long time the framework was limited to the video_device struct for
23	creating V4L device nodes and video_buf for handling the video buffers
24	(note that this document does not discuss the video_buf framework).
25	
26	This meant that all drivers had to do the setup of device instances and
27	connecting to sub-devices themselves. Some of this is quite complicated
28	to do right and many drivers never did do it correctly.
29	
30	There is also a lot of common code that could never be refactored due to
31	the lack of a framework.
32	
33	So this framework sets up the basic building blocks that all drivers
34	need and this same framework should make it much easier to refactor
35	common code into utility functions shared by all drivers.
36	
37	A good example to look at as a reference is the v4l2-pci-skeleton.c
38	source that is available in samples/v4l/. It is a skeleton driver for
39	a PCI capture card, and demonstrates how to use the V4L2 driver
40	framework. It can be used as a template for real PCI video capture driver.
41	
42	Structure of a driver
43	---------------------
44	
45	All drivers have the following structure:
46	
47	1) A struct for each device instance containing the device state.
48	
49	2) A way of initializing and commanding sub-devices (if any).
50	
51	3) Creating V4L2 device nodes (/dev/videoX, /dev/vbiX and /dev/radioX)
52	   and keeping track of device-node specific data.
53	
54	4) Filehandle-specific structs containing per-filehandle data;
55	
56	5) video buffer handling.
57	
58	This is a rough schematic of how it all relates:
59	
60	    device instances
61	      |
62	      +-sub-device instances
63	      |
64	      \-V4L2 device nodes
65		  |
66		  \-filehandle instances
67	
68	
69	Structure of the framework
70	--------------------------
71	
72	The framework closely resembles the driver structure: it has a v4l2_device
73	struct for the device instance data, a v4l2_subdev struct to refer to
74	sub-device instances, the video_device struct stores V4L2 device node data
75	and the v4l2_fh struct keeps track of filehandle instances.
76	
77	The V4L2 framework also optionally integrates with the media framework. If a
78	driver sets the struct v4l2_device mdev field, sub-devices and video nodes
79	will automatically appear in the media framework as entities.
80	
81	
82	struct v4l2_device
83	------------------
84	
85	Each device instance is represented by a struct v4l2_device (v4l2-device.h).
86	Very simple devices can just allocate this struct, but most of the time you
87	would embed this struct inside a larger struct.
88	
89	You must register the device instance:
90	
91		v4l2_device_register(struct device *dev, struct v4l2_device *v4l2_dev);
92	
93	Registration will initialize the v4l2_device struct. If the dev->driver_data
94	field is NULL, it will be linked to v4l2_dev.
95	
96	Drivers that want integration with the media device framework need to set
97	dev->driver_data manually to point to the driver-specific device structure
98	that embed the struct v4l2_device instance. This is achieved by a
99	dev_set_drvdata() call before registering the V4L2 device instance. They must
100	also set the struct v4l2_device mdev field to point to a properly initialized
101	and registered media_device instance.
102	
103	If v4l2_dev->name is empty then it will be set to a value derived from dev
104	(driver name followed by the bus_id, to be precise). If you set it up before
105	calling v4l2_device_register then it will be untouched. If dev is NULL, then
106	you *must* setup v4l2_dev->name before calling v4l2_device_register.
107	
108	You can use v4l2_device_set_name() to set the name based on a driver name and
109	a driver-global atomic_t instance. This will generate names like ivtv0, ivtv1,
110	etc. If the name ends with a digit, then it will insert a dash: cx18-0,
111	cx18-1, etc. This function returns the instance number.
112	
113	The first 'dev' argument is normally the struct device pointer of a pci_dev,
114	usb_interface or platform_device. It is rare for dev to be NULL, but it happens
115	with ISA devices or when one device creates multiple PCI devices, thus making
116	it impossible to associate v4l2_dev with a particular parent.
117	
118	You can also supply a notify() callback that can be called by sub-devices to
119	notify you of events. Whether you need to set this depends on the sub-device.
120	Any notifications a sub-device supports must be defined in a header in
121	include/media/<subdevice>.h.
122	
123	You unregister with:
124	
125		v4l2_device_unregister(struct v4l2_device *v4l2_dev);
126	
127	If the dev->driver_data field points to v4l2_dev, it will be reset to NULL.
128	Unregistering will also automatically unregister all subdevs from the device.
129	
130	If you have a hotpluggable device (e.g. a USB device), then when a disconnect
131	happens the parent device becomes invalid. Since v4l2_device has a pointer to
132	that parent device it has to be cleared as well to mark that the parent is
133	gone. To do this call:
134	
135		v4l2_device_disconnect(struct v4l2_device *v4l2_dev);
136	
137	This does *not* unregister the subdevs, so you still need to call the
138	v4l2_device_unregister() function for that. If your driver is not hotpluggable,
139	then there is no need to call v4l2_device_disconnect().
140	
141	Sometimes you need to iterate over all devices registered by a specific
142	driver. This is usually the case if multiple device drivers use the same
143	hardware. E.g. the ivtvfb driver is a framebuffer driver that uses the ivtv
144	hardware. The same is true for alsa drivers for example.
145	
146	You can iterate over all registered devices as follows:
147	
148	static int callback(struct device *dev, void *p)
149	{
150		struct v4l2_device *v4l2_dev = dev_get_drvdata(dev);
151	
152		/* test if this device was inited */
153		if (v4l2_dev == NULL)
154			return 0;
155		...
156		return 0;
157	}
158	
159	int iterate(void *p)
160	{
161		struct device_driver *drv;
162		int err;
163	
164		/* Find driver 'ivtv' on the PCI bus.
165		   pci_bus_type is a global. For USB busses use usb_bus_type. */
166		drv = driver_find("ivtv", &pci_bus_type);
167		/* iterate over all ivtv device instances */
168		err = driver_for_each_device(drv, NULL, p, callback);
169		put_driver(drv);
170		return err;
171	}
172	
173	Sometimes you need to keep a running counter of the device instance. This is
174	commonly used to map a device instance to an index of a module option array.
175	
176	The recommended approach is as follows:
177	
178	static atomic_t drv_instance = ATOMIC_INIT(0);
179	
180	static int drv_probe(struct pci_dev *pdev, const struct pci_device_id *pci_id)
181	{
182		...
183		state->instance = atomic_inc_return(&drv_instance) - 1;
184	}
185	
186	If you have multiple device nodes then it can be difficult to know when it is
187	safe to unregister v4l2_device for hotpluggable devices. For this purpose
188	v4l2_device has refcounting support. The refcount is increased whenever
189	video_register_device is called and it is decreased whenever that device node
190	is released. When the refcount reaches zero, then the v4l2_device release()
191	callback is called. You can do your final cleanup there.
192	
193	If other device nodes (e.g. ALSA) are created, then you can increase and
194	decrease the refcount manually as well by calling:
195	
196	void v4l2_device_get(struct v4l2_device *v4l2_dev);
197	
198	or:
199	
200	int v4l2_device_put(struct v4l2_device *v4l2_dev);
201	
202	Since the initial refcount is 1 you also need to call v4l2_device_put in the
203	disconnect() callback (for USB devices) or in the remove() callback (for e.g.
204	PCI devices), otherwise the refcount will never reach 0.
205	
206	struct v4l2_subdev
207	------------------
208	
209	Many drivers need to communicate with sub-devices. These devices can do all
210	sort of tasks, but most commonly they handle audio and/or video muxing,
211	encoding or decoding. For webcams common sub-devices are sensors and camera
212	controllers.
213	
214	Usually these are I2C devices, but not necessarily. In order to provide the
215	driver with a consistent interface to these sub-devices the v4l2_subdev struct
216	(v4l2-subdev.h) was created.
217	
218	Each sub-device driver must have a v4l2_subdev struct. This struct can be
219	stand-alone for simple sub-devices or it might be embedded in a larger struct
220	if more state information needs to be stored. Usually there is a low-level
221	device struct (e.g. i2c_client) that contains the device data as setup
222	by the kernel. It is recommended to store that pointer in the private
223	data of v4l2_subdev using v4l2_set_subdevdata(). That makes it easy to go
224	from a v4l2_subdev to the actual low-level bus-specific device data.
225	
226	You also need a way to go from the low-level struct to v4l2_subdev. For the
227	common i2c_client struct the i2c_set_clientdata() call is used to store a
228	v4l2_subdev pointer, for other busses you may have to use other methods.
229	
230	Bridges might also need to store per-subdev private data, such as a pointer to
231	bridge-specific per-subdev private data. The v4l2_subdev structure provides
232	host private data for that purpose that can be accessed with
233	v4l2_get_subdev_hostdata() and v4l2_set_subdev_hostdata().
234	
235	From the bridge driver perspective you load the sub-device module and somehow
236	obtain the v4l2_subdev pointer. For i2c devices this is easy: you call
237	i2c_get_clientdata(). For other busses something similar needs to be done.
238	Helper functions exists for sub-devices on an I2C bus that do most of this
239	tricky work for you.
240	
241	Each v4l2_subdev contains function pointers that sub-device drivers can
242	implement (or leave NULL if it is not applicable). Since sub-devices can do
243	so many different things and you do not want to end up with a huge ops struct
244	of which only a handful of ops are commonly implemented, the function pointers
245	are sorted according to category and each category has its own ops struct.
246	
247	The top-level ops struct contains pointers to the category ops structs, which
248	may be NULL if the subdev driver does not support anything from that category.
249	
250	It looks like this:
251	
252	struct v4l2_subdev_core_ops {
253		int (*log_status)(struct v4l2_subdev *sd);
254		int (*init)(struct v4l2_subdev *sd, u32 val);
255		...
256	};
257	
258	struct v4l2_subdev_tuner_ops {
259		...
260	};
261	
262	struct v4l2_subdev_audio_ops {
263		...
264	};
265	
266	struct v4l2_subdev_video_ops {
267		...
268	};
269	
270	struct v4l2_subdev_pad_ops {
271		...
272	};
273	
274	struct v4l2_subdev_ops {
275		const struct v4l2_subdev_core_ops  *core;
276		const struct v4l2_subdev_tuner_ops *tuner;
277		const struct v4l2_subdev_audio_ops *audio;
278		const struct v4l2_subdev_video_ops *video;
279		const struct v4l2_subdev_pad_ops *video;
280	};
281	
282	The core ops are common to all subdevs, the other categories are implemented
283	depending on the sub-device. E.g. a video device is unlikely to support the
284	audio ops and vice versa.
285	
286	This setup limits the number of function pointers while still making it easy
287	to add new ops and categories.
288	
289	A sub-device driver initializes the v4l2_subdev struct using:
290	
291		v4l2_subdev_init(sd, &ops);
292	
293	Afterwards you need to initialize subdev->name with a unique name and set the
294	module owner. This is done for you if you use the i2c helper functions.
295	
296	If integration with the media framework is needed, you must initialize the
297	media_entity struct embedded in the v4l2_subdev struct (entity field) by
298	calling media_entity_pads_init(), if the entity has pads:
299	
300		struct media_pad *pads = &my_sd->pads;
301		int err;
302	
303		err = media_entity_pads_init(&sd->entity, npads, pads);
304	
305	The pads array must have been previously initialized. There is no need to
306	manually set the struct media_entity function and name fields, but the
307	revision field must be initialized if needed.
308	
309	A reference to the entity will be automatically acquired/released when the
310	subdev device node (if any) is opened/closed.
311	
312	Don't forget to cleanup the media entity before the sub-device is destroyed:
313	
314		media_entity_cleanup(&sd->entity);
315	
316	If the subdev driver intends to process video and integrate with the media
317	framework, it must implement format related functionality using
318	v4l2_subdev_pad_ops instead of v4l2_subdev_video_ops.
319	
320	In that case, the subdev driver may set the link_validate field to provide
321	its own link validation function. The link validation function is called for
322	every link in the pipeline where both of the ends of the links are V4L2
323	sub-devices. The driver is still responsible for validating the correctness
324	of the format configuration between sub-devices and video nodes.
325	
326	If link_validate op is not set, the default function
327	v4l2_subdev_link_validate_default() is used instead. This function ensures
328	that width, height and the media bus pixel code are equal on both source and
329	sink of the link. Subdev drivers are also free to use this function to
330	perform the checks mentioned above in addition to their own checks.
331	
332	There are currently two ways to register subdevices with the V4L2 core. The
333	first (traditional) possibility is to have subdevices registered by bridge
334	drivers. This can be done when the bridge driver has the complete information
335	about subdevices connected to it and knows exactly when to register them. This
336	is typically the case for internal subdevices, like video data processing units
337	within SoCs or complex PCI(e) boards, camera sensors in USB cameras or connected
338	to SoCs, which pass information about them to bridge drivers, usually in their
339	platform data.
340	
341	There are however also situations where subdevices have to be registered
342	asynchronously to bridge devices. An example of such a configuration is a Device
343	Tree based system where information about subdevices is made available to the
344	system independently from the bridge devices, e.g. when subdevices are defined
345	in DT as I2C device nodes. The API used in this second case is described further
346	below.
347	
348	Using one or the other registration method only affects the probing process, the
349	run-time bridge-subdevice interaction is in both cases the same.
350	
351	In the synchronous case a device (bridge) driver needs to register the
352	v4l2_subdev with the v4l2_device:
353	
354		int err = v4l2_device_register_subdev(v4l2_dev, sd);
355	
356	This can fail if the subdev module disappeared before it could be registered.
357	After this function was called successfully the subdev->dev field points to
358	the v4l2_device.
359	
360	If the v4l2_device parent device has a non-NULL mdev field, the sub-device
361	entity will be automatically registered with the media device.
362	
363	You can unregister a sub-device using:
364	
365		v4l2_device_unregister_subdev(sd);
366	
367	Afterwards the subdev module can be unloaded and sd->dev == NULL.
368	
369	You can call an ops function either directly:
370	
371		err = sd->ops->core->g_std(sd, &norm);
372	
373	but it is better and easier to use this macro:
374	
375		err = v4l2_subdev_call(sd, core, g_std, &norm);
376	
377	The macro will to the right NULL pointer checks and returns -ENODEV if subdev
378	is NULL, -ENOIOCTLCMD if either subdev->core or subdev->core->g_std is
379	NULL, or the actual result of the subdev->ops->core->g_std ops.
380	
381	It is also possible to call all or a subset of the sub-devices:
382	
383		v4l2_device_call_all(v4l2_dev, 0, core, g_std, &norm);
384	
385	Any subdev that does not support this ops is skipped and error results are
386	ignored. If you want to check for errors use this:
387	
388		err = v4l2_device_call_until_err(v4l2_dev, 0, core, g_std, &norm);
389	
390	Any error except -ENOIOCTLCMD will exit the loop with that error. If no
391	errors (except -ENOIOCTLCMD) occurred, then 0 is returned.
392	
393	The second argument to both calls is a group ID. If 0, then all subdevs are
394	called. If non-zero, then only those whose group ID match that value will
395	be called. Before a bridge driver registers a subdev it can set sd->grp_id
396	to whatever value it wants (it's 0 by default). This value is owned by the
397	bridge driver and the sub-device driver will never modify or use it.
398	
399	The group ID gives the bridge driver more control how callbacks are called.
400	For example, there may be multiple audio chips on a board, each capable of
401	changing the volume. But usually only one will actually be used when the
402	user want to change the volume. You can set the group ID for that subdev to
403	e.g. AUDIO_CONTROLLER and specify that as the group ID value when calling
404	v4l2_device_call_all(). That ensures that it will only go to the subdev
405	that needs it.
406	
407	If the sub-device needs to notify its v4l2_device parent of an event, then
408	it can call v4l2_subdev_notify(sd, notification, arg). This macro checks
409	whether there is a notify() callback defined and returns -ENODEV if not.
410	Otherwise the result of the notify() call is returned.
411	
412	The advantage of using v4l2_subdev is that it is a generic struct and does
413	not contain any knowledge about the underlying hardware. So a driver might
414	contain several subdevs that use an I2C bus, but also a subdev that is
415	controlled through GPIO pins. This distinction is only relevant when setting
416	up the device, but once the subdev is registered it is completely transparent.
417	
418	
419	In the asynchronous case subdevice probing can be invoked independently of the
420	bridge driver availability. The subdevice driver then has to verify whether all
421	the requirements for a successful probing are satisfied. This can include a
422	check for a master clock availability. If any of the conditions aren't satisfied
423	the driver might decide to return -EPROBE_DEFER to request further reprobing
424	attempts. Once all conditions are met the subdevice shall be registered using
425	the v4l2_async_register_subdev() function. Unregistration is performed using
426	the v4l2_async_unregister_subdev() call. Subdevices registered this way are
427	stored in a global list of subdevices, ready to be picked up by bridge drivers.
428	
429	Bridge drivers in turn have to register a notifier object with an array of
430	subdevice descriptors that the bridge device needs for its operation. This is
431	performed using the v4l2_async_notifier_register() call. To unregister the
432	notifier the driver has to call v4l2_async_notifier_unregister(). The former of
433	the two functions takes two arguments: a pointer to struct v4l2_device and a
434	pointer to struct v4l2_async_notifier. The latter contains a pointer to an array
435	of pointers to subdevice descriptors of type struct v4l2_async_subdev type. The
436	V4L2 core will then use these descriptors to match asynchronously registered
437	subdevices to them. If a match is detected the .bound() notifier callback is
438	called. After all subdevices have been located the .complete() callback is
439	called. When a subdevice is removed from the system the .unbind() method is
440	called. All three callbacks are optional.
441	
442	
443	V4L2 sub-device userspace API
444	-----------------------------
445	
446	Beside exposing a kernel API through the v4l2_subdev_ops structure, V4L2
447	sub-devices can also be controlled directly by userspace applications.
448	
449	Device nodes named v4l-subdevX can be created in /dev to access sub-devices
450	directly. If a sub-device supports direct userspace configuration it must set
451	the V4L2_SUBDEV_FL_HAS_DEVNODE flag before being registered.
452	
453	After registering sub-devices, the v4l2_device driver can create device nodes
454	for all registered sub-devices marked with V4L2_SUBDEV_FL_HAS_DEVNODE by calling
455	v4l2_device_register_subdev_nodes(). Those device nodes will be automatically
456	removed when sub-devices are unregistered.
457	
458	The device node handles a subset of the V4L2 API.
459	
460	VIDIOC_QUERYCTRL
461	VIDIOC_QUERYMENU
462	VIDIOC_G_CTRL
463	VIDIOC_S_CTRL
464	VIDIOC_G_EXT_CTRLS
465	VIDIOC_S_EXT_CTRLS
466	VIDIOC_TRY_EXT_CTRLS
467	
468		The controls ioctls are identical to the ones defined in V4L2. They
469		behave identically, with the only exception that they deal only with
470		controls implemented in the sub-device. Depending on the driver, those
471		controls can be also be accessed through one (or several) V4L2 device
472		nodes.
473	
474	VIDIOC_DQEVENT
475	VIDIOC_SUBSCRIBE_EVENT
476	VIDIOC_UNSUBSCRIBE_EVENT
477	
478		The events ioctls are identical to the ones defined in V4L2. They
479		behave identically, with the only exception that they deal only with
480		events generated by the sub-device. Depending on the driver, those
481		events can also be reported by one (or several) V4L2 device nodes.
482	
483		Sub-device drivers that want to use events need to set the
484		V4L2_SUBDEV_USES_EVENTS v4l2_subdev::flags and initialize
485		v4l2_subdev::nevents to events queue depth before registering the
486		sub-device. After registration events can be queued as usual on the
487		v4l2_subdev::devnode device node.
488	
489		To properly support events, the poll() file operation is also
490		implemented.
491	
492	Private ioctls
493	
494		All ioctls not in the above list are passed directly to the sub-device
495		driver through the core::ioctl operation.
496	
497	
498	I2C sub-device drivers
499	----------------------
500	
501	Since these drivers are so common, special helper functions are available to
502	ease the use of these drivers (v4l2-common.h).
503	
504	The recommended method of adding v4l2_subdev support to an I2C driver is to
505	embed the v4l2_subdev struct into the state struct that is created for each
506	I2C device instance. Very simple devices have no state struct and in that case
507	you can just create a v4l2_subdev directly.
508	
509	A typical state struct would look like this (where 'chipname' is replaced by
510	the name of the chip):
511	
512	struct chipname_state {
513		struct v4l2_subdev sd;
514		...  /* additional state fields */
515	};
516	
517	Initialize the v4l2_subdev struct as follows:
518	
519		v4l2_i2c_subdev_init(&state->sd, client, subdev_ops);
520	
521	This function will fill in all the fields of v4l2_subdev and ensure that the
522	v4l2_subdev and i2c_client both point to one another.
523	
524	You should also add a helper inline function to go from a v4l2_subdev pointer
525	to a chipname_state struct:
526	
527	static inline struct chipname_state *to_state(struct v4l2_subdev *sd)
528	{
529		return container_of(sd, struct chipname_state, sd);
530	}
531	
532	Use this to go from the v4l2_subdev struct to the i2c_client struct:
533	
534		struct i2c_client *client = v4l2_get_subdevdata(sd);
535	
536	And this to go from an i2c_client to a v4l2_subdev struct:
537	
538		struct v4l2_subdev *sd = i2c_get_clientdata(client);
539	
540	Make sure to call v4l2_device_unregister_subdev(sd) when the remove() callback
541	is called. This will unregister the sub-device from the bridge driver. It is
542	safe to call this even if the sub-device was never registered.
543	
544	You need to do this because when the bridge driver destroys the i2c adapter
545	the remove() callbacks are called of the i2c devices on that adapter.
546	After that the corresponding v4l2_subdev structures are invalid, so they
547	have to be unregistered first. Calling v4l2_device_unregister_subdev(sd)
548	from the remove() callback ensures that this is always done correctly.
549	
550	
551	The bridge driver also has some helper functions it can use:
552	
553	struct v4l2_subdev *sd = v4l2_i2c_new_subdev(v4l2_dev, adapter,
554		       "module_foo", "chipid", 0x36, NULL);
555	
556	This loads the given module (can be NULL if no module needs to be loaded) and
557	calls i2c_new_device() with the given i2c_adapter and chip/address arguments.
558	If all goes well, then it registers the subdev with the v4l2_device.
559	
560	You can also use the last argument of v4l2_i2c_new_subdev() to pass an array
561	of possible I2C addresses that it should probe. These probe addresses are
562	only used if the previous argument is 0. A non-zero argument means that you
563	know the exact i2c address so in that case no probing will take place.
564	
565	Both functions return NULL if something went wrong.
566	
567	Note that the chipid you pass to v4l2_i2c_new_subdev() is usually
568	the same as the module name. It allows you to specify a chip variant, e.g.
569	"saa7114" or "saa7115". In general though the i2c driver autodetects this.
570	The use of chipid is something that needs to be looked at more closely at a
571	later date. It differs between i2c drivers and as such can be confusing.
572	To see which chip variants are supported you can look in the i2c driver code
573	for the i2c_device_id table. This lists all the possibilities.
574	
575	There are two more helper functions:
576	
577	v4l2_i2c_new_subdev_cfg: this function adds new irq and platform_data
578	arguments and has both 'addr' and 'probed_addrs' arguments: if addr is not
579	0 then that will be used (non-probing variant), otherwise the probed_addrs
580	are probed.
581	
582	For example: this will probe for address 0x10:
583	
584	struct v4l2_subdev *sd = v4l2_i2c_new_subdev_cfg(v4l2_dev, adapter,
585		       "module_foo", "chipid", 0, NULL, 0, I2C_ADDRS(0x10));
586	
587	v4l2_i2c_new_subdev_board uses an i2c_board_info struct which is passed
588	to the i2c driver and replaces the irq, platform_data and addr arguments.
589	
590	If the subdev supports the s_config core ops, then that op is called with
591	the irq and platform_data arguments after the subdev was setup. The older
592	v4l2_i2c_new_(probed_)subdev functions will call s_config as well, but with
593	irq set to 0 and platform_data set to NULL.
594	
595	struct video_device
596	-------------------
597	
598	The actual device nodes in the /dev directory are created using the
599	video_device struct (v4l2-dev.h). This struct can either be allocated
600	dynamically or embedded in a larger struct.
601	
602	To allocate it dynamically use:
603	
604		struct video_device *vdev = video_device_alloc();
605	
606		if (vdev == NULL)
607			return -ENOMEM;
608	
609		vdev->release = video_device_release;
610	
611	If you embed it in a larger struct, then you must set the release()
612	callback to your own function:
613	
614		struct video_device *vdev = &my_vdev->vdev;
615	
616		vdev->release = my_vdev_release;
617	
618	The release callback must be set and it is called when the last user
619	of the video device exits.
620	
621	The default video_device_release() callback just calls kfree to free the
622	allocated memory.
623	
624	There is also a video_device_release_empty() function that does nothing
625	(is empty) and can be used if the struct is embedded and there is nothing
626	to do when it is released.
627	
628	You should also set these fields:
629	
630	- v4l2_dev: must be set to the v4l2_device parent device.
631	
632	- name: set to something descriptive and unique.
633	
634	- vfl_dir: set this to VFL_DIR_RX for capture devices (VFL_DIR_RX has value 0,
635	  so this is normally already the default), set to VFL_DIR_TX for output
636	  devices and VFL_DIR_M2M for mem2mem (codec) devices.
637	
638	- fops: set to the v4l2_file_operations struct.
639	
640	- ioctl_ops: if you use the v4l2_ioctl_ops to simplify ioctl maintenance
641	  (highly recommended to use this and it might become compulsory in the
642	  future!), then set this to your v4l2_ioctl_ops struct. The vfl_type and
643	  vfl_dir fields are used to disable ops that do not match the type/dir
644	  combination. E.g. VBI ops are disabled for non-VBI nodes, and output ops
645	  are disabled for a capture device. This makes it possible to provide
646	  just one v4l2_ioctl_ops struct for both vbi and video nodes.
647	
648	- lock: leave to NULL if you want to do all the locking in the driver.
649	  Otherwise you give it a pointer to a struct mutex_lock and before the
650	  unlocked_ioctl file operation is called this lock will be taken by the
651	  core and released afterwards. See the next section for more details.
652	
653	- queue: a pointer to the struct vb2_queue associated with this device node.
654	  If queue is non-NULL, and queue->lock is non-NULL, then queue->lock is
655	  used for the queuing ioctls (VIDIOC_REQBUFS, CREATE_BUFS, QBUF, DQBUF,
656	  QUERYBUF, PREPARE_BUF, STREAMON and STREAMOFF) instead of the lock above.
657	  That way the vb2 queuing framework does not have to wait for other ioctls.
658	  This queue pointer is also used by the vb2 helper functions to check for
659	  queuing ownership (i.e. is the filehandle calling it allowed to do the
660	  operation).
661	
662	- prio: keeps track of the priorities. Used to implement VIDIOC_G/S_PRIORITY.
663	  If left to NULL, then it will use the struct v4l2_prio_state in v4l2_device.
664	  If you want to have a separate priority state per (group of) device node(s),
665	  then you can point it to your own struct v4l2_prio_state.
666	
667	- dev_parent: you only set this if v4l2_device was registered with NULL as
668	  the parent device struct. This only happens in cases where one hardware
669	  device has multiple PCI devices that all share the same v4l2_device core.
670	
671	  The cx88 driver is an example of this: one core v4l2_device struct, but
672	  it is used by both a raw video PCI device (cx8800) and a MPEG PCI device
673	  (cx8802). Since the v4l2_device cannot be associated with two PCI devices
674	  at the same time it is setup without a parent device. But when the struct
675	  video_device is initialized you *do* know which parent PCI device to use and
676	  so you set dev_device to the correct PCI device.
677	
678	If you use v4l2_ioctl_ops, then you should set .unlocked_ioctl to video_ioctl2
679	in your v4l2_file_operations struct.
680	
681	Do not use .ioctl! This is deprecated and will go away in the future.
682	
683	In some cases you want to tell the core that a function you had specified in
684	your v4l2_ioctl_ops should be ignored. You can mark such ioctls by calling this
685	function before video_device_register is called:
686	
687	void v4l2_disable_ioctl(struct video_device *vdev, unsigned int cmd);
688	
689	This tends to be needed if based on external factors (e.g. which card is
690	being used) you want to turns off certain features in v4l2_ioctl_ops without
691	having to make a new struct.
692	
693	The v4l2_file_operations struct is a subset of file_operations. The main
694	difference is that the inode argument is omitted since it is never used.
695	
696	If integration with the media framework is needed, you must initialize the
697	media_entity struct embedded in the video_device struct (entity field) by
698	calling media_entity_pads_init():
699	
700		struct media_pad *pad = &my_vdev->pad;
701		int err;
702	
703		err = media_entity_pads_init(&vdev->entity, 1, pad);
704	
705	The pads array must have been previously initialized. There is no need to
706	manually set the struct media_entity type and name fields.
707	
708	A reference to the entity will be automatically acquired/released when the
709	video device is opened/closed.
710	
711	ioctls and locking
712	------------------
713	
714	The V4L core provides optional locking services. The main service is the
715	lock field in struct video_device, which is a pointer to a mutex. If you set
716	this pointer, then that will be used by unlocked_ioctl to serialize all ioctls.
717	
718	If you are using the videobuf2 framework, then there is a second lock that you
719	can set: video_device->queue->lock. If set, then this lock will be used instead
720	of video_device->lock to serialize all queuing ioctls (see the previous section
721	for the full list of those ioctls).
722	
723	The advantage of using a different lock for the queuing ioctls is that for some
724	drivers (particularly USB drivers) certain commands such as setting controls
725	can take a long time, so you want to use a separate lock for the buffer queuing
726	ioctls. That way your VIDIOC_DQBUF doesn't stall because the driver is busy
727	changing the e.g. exposure of the webcam.
728	
729	Of course, you can always do all the locking yourself by leaving both lock
730	pointers at NULL.
731	
732	If you use the old videobuf then you must pass the video_device lock to the
733	videobuf queue initialize function: if videobuf has to wait for a frame to
734	arrive, then it will temporarily unlock the lock and relock it afterwards. If
735	your driver also waits in the code, then you should do the same to allow other
736	processes to access the device node while the first process is waiting for
737	something.
738	
739	In the case of videobuf2 you will need to implement the wait_prepare and
740	wait_finish callbacks to unlock/lock if applicable. If you use the queue->lock
741	pointer, then you can use the helper functions vb2_ops_wait_prepare/finish.
742	
743	The implementation of a hotplug disconnect should also take the lock from
744	video_device before calling v4l2_device_disconnect. If you are also using
745	video_device->queue->lock, then you have to first lock video_device->queue->lock
746	followed by video_device->lock. That way you can be sure no ioctl is running
747	when you call v4l2_device_disconnect.
748	
749	video_device registration
750	-------------------------
751	
752	Next you register the video device: this will create the character device
753	for you.
754	
755		err = video_register_device(vdev, VFL_TYPE_GRABBER, -1);
756		if (err) {
757			video_device_release(vdev); /* or kfree(my_vdev); */
758			return err;
759		}
760	
761	If the v4l2_device parent device has a non-NULL mdev field, the video device
762	entity will be automatically registered with the media device.
763	
764	Which device is registered depends on the type argument. The following
765	types exist:
766	
767	VFL_TYPE_GRABBER: videoX for video input/output devices
768	VFL_TYPE_VBI: vbiX for vertical blank data (i.e. closed captions, teletext)
769	VFL_TYPE_RADIO: radioX for radio tuners
770	VFL_TYPE_SDR: swradioX for Software Defined Radio tuners
771	
772	The last argument gives you a certain amount of control over the device
773	device node number used (i.e. the X in videoX). Normally you will pass -1
774	to let the v4l2 framework pick the first free number. But sometimes users
775	want to select a specific node number. It is common that drivers allow
776	the user to select a specific device node number through a driver module
777	option. That number is then passed to this function and video_register_device
778	will attempt to select that device node number. If that number was already
779	in use, then the next free device node number will be selected and it
780	will send a warning to the kernel log.
781	
782	Another use-case is if a driver creates many devices. In that case it can
783	be useful to place different video devices in separate ranges. For example,
784	video capture devices start at 0, video output devices start at 16.
785	So you can use the last argument to specify a minimum device node number
786	and the v4l2 framework will try to pick the first free number that is equal
787	or higher to what you passed. If that fails, then it will just pick the
788	first free number.
789	
790	Since in this case you do not care about a warning about not being able
791	to select the specified device node number, you can call the function
792	video_register_device_no_warn() instead.
793	
794	Whenever a device node is created some attributes are also created for you.
795	If you look in /sys/class/video4linux you see the devices. Go into e.g.
796	video0 and you will see 'name', 'dev_debug' and 'index' attributes. The 'name'
797	attribute is the 'name' field of the video_device struct. The 'dev_debug' attribute
798	can be used to enable core debugging. See the next section for more detailed
799	information on this.
800	
801	The 'index' attribute is the index of the device node: for each call to
802	video_register_device() the index is just increased by 1. The first video
803	device node you register always starts with index 0.
804	
805	Users can setup udev rules that utilize the index attribute to make fancy
806	device names (e.g. 'mpegX' for MPEG video capture device nodes).
807	
808	After the device was successfully registered, then you can use these fields:
809	
810	- vfl_type: the device type passed to video_register_device.
811	- minor: the assigned device minor number.
812	- num: the device node number (i.e. the X in videoX).
813	- index: the device index number.
814	
815	If the registration failed, then you need to call video_device_release()
816	to free the allocated video_device struct, or free your own struct if the
817	video_device was embedded in it. The vdev->release() callback will never
818	be called if the registration failed, nor should you ever attempt to
819	unregister the device if the registration failed.
820	
821	video device debugging
822	----------------------
823	
824	The 'dev_debug' attribute that is created for each video, vbi, radio or swradio
825	device in /sys/class/video4linux/<devX>/ allows you to enable logging of
826	file operations.
827	
828	It is a bitmask and the following bits can be set:
829	
830	0x01: Log the ioctl name and error code. VIDIOC_(D)QBUF ioctls are only logged
831	      if bit 0x08 is also set.
832	0x02: Log the ioctl name arguments and error code. VIDIOC_(D)QBUF ioctls are
833	      only logged if bit 0x08 is also set.
834	0x04: Log the file operations open, release, read, write, mmap and
835	      get_unmapped_area. The read and write operations are only logged if
836	      bit 0x08 is also set.
837	0x08: Log the read and write file operations and the VIDIOC_QBUF and
838	      VIDIOC_DQBUF ioctls.
839	0x10: Log the poll file operation.
840	
841	video_device cleanup
842	--------------------
843	
844	When the video device nodes have to be removed, either during the unload
845	of the driver or because the USB device was disconnected, then you should
846	unregister them:
847	
848		video_unregister_device(vdev);
849	
850	This will remove the device nodes from sysfs (causing udev to remove them
851	from /dev).
852	
853	After video_unregister_device() returns no new opens can be done. However,
854	in the case of USB devices some application might still have one of these
855	device nodes open. So after the unregister all file operations (except
856	release, of course) will return an error as well.
857	
858	When the last user of the video device node exits, then the vdev->release()
859	callback is called and you can do the final cleanup there.
860	
861	Don't forget to cleanup the media entity associated with the video device if
862	it has been initialized:
863	
864		media_entity_cleanup(&vdev->entity);
865	
866	This can be done from the release callback.
867	
868	
869	video_device helper functions
870	-----------------------------
871	
872	There are a few useful helper functions:
873	
874	- file/video_device private data
875	
876	You can set/get driver private data in the video_device struct using:
877	
878	void *video_get_drvdata(struct video_device *vdev);
879	void video_set_drvdata(struct video_device *vdev, void *data);
880	
881	Note that you can safely call video_set_drvdata() before calling
882	video_register_device().
883	
884	And this function:
885	
886	struct video_device *video_devdata(struct file *file);
887	
888	returns the video_device belonging to the file struct.
889	
890	The video_drvdata function combines video_get_drvdata with video_devdata:
891	
892	void *video_drvdata(struct file *file);
893	
894	You can go from a video_device struct to the v4l2_device struct using:
895	
896	struct v4l2_device *v4l2_dev = vdev->v4l2_dev;
897	
898	- Device node name
899	
900	The video_device node kernel name can be retrieved using
901	
902	const char *video_device_node_name(struct video_device *vdev);
903	
904	The name is used as a hint by userspace tools such as udev. The function
905	should be used where possible instead of accessing the video_device::num and
906	video_device::minor fields.
907	
908	
909	video buffer helper functions
910	-----------------------------
911	
912	The v4l2 core API provides a set of standard methods (called "videobuf")
913	for dealing with video buffers. Those methods allow a driver to implement
914	read(), mmap() and overlay() in a consistent way.  There are currently
915	methods for using video buffers on devices that supports DMA with
916	scatter/gather method (videobuf-dma-sg), DMA with linear access
917	(videobuf-dma-contig), and vmalloced buffers, mostly used on USB drivers
918	(videobuf-vmalloc).
919	
920	Please see Documentation/video4linux/videobuf for more information on how
921	to use the videobuf layer.
922	
923	struct v4l2_fh
924	--------------
925	
926	struct v4l2_fh provides a way to easily keep file handle specific data
927	that is used by the V4L2 framework. New drivers must use struct v4l2_fh
928	since it is also used to implement priority handling (VIDIOC_G/S_PRIORITY).
929	
930	The users of v4l2_fh (in the V4L2 framework, not the driver) know
931	whether a driver uses v4l2_fh as its file->private_data pointer by
932	testing the V4L2_FL_USES_V4L2_FH bit in video_device->flags. This bit is
933	set whenever v4l2_fh_init() is called.
934	
935	struct v4l2_fh is allocated as a part of the driver's own file handle
936	structure and file->private_data is set to it in the driver's open
937	function by the driver.
938	
939	In many cases the struct v4l2_fh will be embedded in a larger structure.
940	In that case you should call v4l2_fh_init+v4l2_fh_add in open() and
941	v4l2_fh_del+v4l2_fh_exit in release().
942	
943	Drivers can extract their own file handle structure by using the container_of
944	macro. Example:
945	
946	struct my_fh {
947		int blah;
948		struct v4l2_fh fh;
949	};
950	
951	...
952	
953	int my_open(struct file *file)
954	{
955		struct my_fh *my_fh;
956		struct video_device *vfd;
957		int ret;
958	
959		...
960	
961		my_fh = kzalloc(sizeof(*my_fh), GFP_KERNEL);
962	
963		...
964	
965		v4l2_fh_init(&my_fh->fh, vfd);
966	
967		...
968	
969		file->private_data = &my_fh->fh;
970		v4l2_fh_add(&my_fh->fh);
971		return 0;
972	}
973	
974	int my_release(struct file *file)
975	{
976		struct v4l2_fh *fh = file->private_data;
977		struct my_fh *my_fh = container_of(fh, struct my_fh, fh);
978	
979		...
980		v4l2_fh_del(&my_fh->fh);
981		v4l2_fh_exit(&my_fh->fh);
982		kfree(my_fh);
983		return 0;
984	}
985	
986	Below is a short description of the v4l2_fh functions used:
987	
988	void v4l2_fh_init(struct v4l2_fh *fh, struct video_device *vdev)
989	
990	  Initialise the file handle. This *MUST* be performed in the driver's
991	  v4l2_file_operations->open() handler.
992	
993	void v4l2_fh_add(struct v4l2_fh *fh)
994	
995	  Add a v4l2_fh to video_device file handle list. Must be called once the
996	  file handle is completely initialized.
997	
998	void v4l2_fh_del(struct v4l2_fh *fh)
999	
1000	  Unassociate the file handle from video_device(). The file handle
1001	  exit function may now be called.
1002	
1003	void v4l2_fh_exit(struct v4l2_fh *fh)
1004	
1005	  Uninitialise the file handle. After uninitialisation the v4l2_fh
1006	  memory can be freed.
1007	
1008	
1009	If struct v4l2_fh is not embedded, then you can use these helper functions:
1010	
1011	int v4l2_fh_open(struct file *filp)
1012	
1013	  This allocates a struct v4l2_fh, initializes it and adds it to the struct
1014	  video_device associated with the file struct.
1015	
1016	int v4l2_fh_release(struct file *filp)
1017	
1018	  This deletes it from the struct video_device associated with the file
1019	  struct, uninitialised the v4l2_fh and frees it.
1020	
1021	These two functions can be plugged into the v4l2_file_operation's open() and
1022	release() ops.
1023	
1024	
1025	Several drivers need to do something when the first file handle is opened and
1026	when the last file handle closes. Two helper functions were added to check
1027	whether the v4l2_fh struct is the only open filehandle of the associated
1028	device node:
1029	
1030	int v4l2_fh_is_singular(struct v4l2_fh *fh)
1031	
1032	  Returns 1 if the file handle is the only open file handle, else 0.
1033	
1034	int v4l2_fh_is_singular_file(struct file *filp)
1035	
1036	  Same, but it calls v4l2_fh_is_singular with filp->private_data.
1037	
1038	
1039	V4L2 events
1040	-----------
1041	
1042	The V4L2 events provide a generic way to pass events to user space.
1043	The driver must use v4l2_fh to be able to support V4L2 events.
1044	
1045	Events are defined by a type and an optional ID. The ID may refer to a V4L2
1046	object such as a control ID. If unused, then the ID is 0.
1047	
1048	When the user subscribes to an event the driver will allocate a number of
1049	kevent structs for that event. So every (type, ID) event tuple will have
1050	its own set of kevent structs. This guarantees that if a driver is generating
1051	lots of events of one type in a short time, then that will not overwrite
1052	events of another type.
1053	
1054	But if you get more events of one type than the number of kevents that were
1055	reserved, then the oldest event will be dropped and the new one added.
1056	
1057	Furthermore, the internal struct v4l2_subscribed_event has merge() and
1058	replace() callbacks which drivers can set. These callbacks are called when
1059	a new event is raised and there is no more room. The replace() callback
1060	allows you to replace the payload of the old event with that of the new event,
1061	merging any relevant data from the old payload into the new payload that
1062	replaces it. It is called when this event type has only one kevent struct
1063	allocated. The merge() callback allows you to merge the oldest event payload
1064	into that of the second-oldest event payload. It is called when there are two
1065	or more kevent structs allocated.
1066	
1067	This way no status information is lost, just the intermediate steps leading
1068	up to that state.
1069	
1070	A good example of these replace/merge callbacks is in v4l2-event.c:
1071	ctrls_replace() and ctrls_merge() callbacks for the control event.
1072	
1073	Note: these callbacks can be called from interrupt context, so they must be
1074	fast.
1075	
1076	Useful functions:
1077	
1078	void v4l2_event_queue(struct video_device *vdev, const struct v4l2_event *ev)
1079	
1080	  Queue events to video device. The driver's only responsibility is to fill
1081	  in the type and the data fields. The other fields will be filled in by
1082	  V4L2.
1083	
1084	int v4l2_event_subscribe(struct v4l2_fh *fh,
1085				 struct v4l2_event_subscription *sub, unsigned elems,
1086				 const struct v4l2_subscribed_event_ops *ops)
1087	
1088	  The video_device->ioctl_ops->vidioc_subscribe_event must check the driver
1089	  is able to produce events with specified event id. Then it calls
1090	  v4l2_event_subscribe() to subscribe the event.
1091	
1092	  The elems argument is the size of the event queue for this event. If it is 0,
1093	  then the framework will fill in a default value (this depends on the event
1094	  type).
1095	
1096	  The ops argument allows the driver to specify a number of callbacks:
1097	  * add:     called when a new listener gets added (subscribing to the same
1098	             event twice will only cause this callback to get called once)
1099	  * del:     called when a listener stops listening
1100	  * replace: replace event 'old' with event 'new'.
1101	  * merge:   merge event 'old' into event 'new'.
1102	  All 4 callbacks are optional, if you don't want to specify any callbacks
1103	  the ops argument itself maybe NULL.
1104	
1105	int v4l2_event_unsubscribe(struct v4l2_fh *fh,
1106				   struct v4l2_event_subscription *sub)
1107	
1108	  vidioc_unsubscribe_event in struct v4l2_ioctl_ops. A driver may use
1109	  v4l2_event_unsubscribe() directly unless it wants to be involved in
1110	  unsubscription process.
1111	
1112	  The special type V4L2_EVENT_ALL may be used to unsubscribe all events. The
1113	  drivers may want to handle this in a special way.
1114	
1115	int v4l2_event_pending(struct v4l2_fh *fh)
1116	
1117	  Returns the number of pending events. Useful when implementing poll.
1118	
1119	Events are delivered to user space through the poll system call. The driver
1120	can use v4l2_fh->wait (a wait_queue_head_t) as the argument for poll_wait().
1121	
1122	There are standard and private events. New standard events must use the
1123	smallest available event type. The drivers must allocate their events from
1124	their own class starting from class base. Class base is
1125	V4L2_EVENT_PRIVATE_START + n * 1000 where n is the lowest available number.
1126	The first event type in the class is reserved for future use, so the first
1127	available event type is 'class base + 1'.
1128	
1129	An example on how the V4L2 events may be used can be found in the OMAP
1130	3 ISP driver (drivers/media/platform/omap3isp).
1131	
1132	A subdev can directly send an event to the v4l2_device notify function with
1133	V4L2_DEVICE_NOTIFY_EVENT. This allows the bridge to map the subdev that sends
1134	the event to the video node(s) associated with the subdev that need to be
1135	informed about such an event.
1136	
1137	V4L2 clocks
1138	-----------
1139	
1140	Many subdevices, like camera sensors, TV decoders and encoders, need a clock
1141	signal to be supplied by the system. Often this clock is supplied by the
1142	respective bridge device. The Linux kernel provides a Common Clock Framework for
1143	this purpose. However, it is not (yet) available on all architectures. Besides,
1144	the nature of the multi-functional (clock, data + synchronisation, I2C control)
1145	connection of subdevices to the system might impose special requirements on the
1146	clock API usage. E.g. V4L2 has to support clock provider driver unregistration
1147	while a subdevice driver is holding a reference to the clock. For these reasons
1148	a V4L2 clock helper API has been developed and is provided to bridge and
1149	subdevice drivers.
1150	
1151	The API consists of two parts: two functions to register and unregister a V4L2
1152	clock source: v4l2_clk_register() and v4l2_clk_unregister() and calls to control
1153	a clock object, similar to the respective generic clock API calls:
1154	v4l2_clk_get(), v4l2_clk_put(), v4l2_clk_enable(), v4l2_clk_disable(),
1155	v4l2_clk_get_rate(), and v4l2_clk_set_rate(). Clock suppliers have to provide
1156	clock operations that will be called when clock users invoke respective API
1157	methods.
1158	
1159	It is expected that once the CCF becomes available on all relevant
1160	architectures this API will be removed.
Hide Line Numbers


About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog