1 Overview of the V4L2 driver framework
2 =====================================
4 This text documents the various structures provided by the V4L2 framework and
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
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
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).
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.
30 There is also a lot of common code that could never be refactored due to
31 the lack of a framework.
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.
41 All drivers have the following structure:
43 1) A struct for each device instance containing the device state.
45 2) A way of initializing and commanding sub-devices (if any).
47 3) Creating V4L2 device nodes (/dev/videoX, /dev/vbiX and /dev/radioX)
48 and keeping track of device-node specific data.
50 4) Filehandle-specific structs containing per-filehandle data;
52 5) video buffer handling.
54 This is a rough schematic of how it all relates:
58 +-sub-device instances
62 \-filehandle instances
65 Structure of the framework
66 --------------------------
68 The framework closely resembles the driver structure: it has a v4l2_device
69 struct for the device instance data, a v4l2_subdev struct to refer to
70 sub-device instances, the video_device struct stores V4L2 device node data
71 and the v4l2_fh struct keeps track of filehandle instances.
73 The V4L2 framework also optionally integrates with the media framework. If a
74 driver sets the struct v4l2_device mdev field, sub-devices and video nodes
75 will automatically appear in the media framework as entities.
81 Each device instance is represented by a struct v4l2_device (v4l2-device.h).
82 Very simple devices can just allocate this struct, but most of the time you
83 would embed this struct inside a larger struct.
85 You must register the device instance:
87 v4l2_device_register(struct device *dev, struct v4l2_device *v4l2_dev);
89 Registration will initialize the v4l2_device struct. If the dev->driver_data
90 field is NULL, it will be linked to v4l2_dev.
92 Drivers that want integration with the media device framework need to set
93 dev->driver_data manually to point to the driver-specific device structure
94 that embed the struct v4l2_device instance. This is achieved by a
95 dev_set_drvdata() call before registering the V4L2 device instance. They must
96 also set the struct v4l2_device mdev field to point to a properly initialized
97 and registered media_device instance.
99 If v4l2_dev->name is empty then it will be set to a value derived from dev
100 (driver name followed by the bus_id, to be precise). If you set it up before
101 calling v4l2_device_register then it will be untouched. If dev is NULL, then
102 you *must* setup v4l2_dev->name before calling v4l2_device_register.
104 You can use v4l2_device_set_name() to set the name based on a driver name and
105 a driver-global atomic_t instance. This will generate names like ivtv0, ivtv1,
106 etc. If the name ends with a digit, then it will insert a dash: cx18-0,
107 cx18-1, etc. This function returns the instance number.
109 The first 'dev' argument is normally the struct device pointer of a pci_dev,
110 usb_interface or platform_device. It is rare for dev to be NULL, but it happens
111 with ISA devices or when one device creates multiple PCI devices, thus making
112 it impossible to associate v4l2_dev with a particular parent.
114 You can also supply a notify() callback that can be called by sub-devices to
115 notify you of events. Whether you need to set this depends on the sub-device.
116 Any notifications a sub-device supports must be defined in a header in
117 include/media/<subdevice>.h.
121 v4l2_device_unregister(struct v4l2_device *v4l2_dev);
123 If the dev->driver_data field points to v4l2_dev, it will be reset to NULL.
124 Unregistering will also automatically unregister all subdevs from the device.
126 If you have a hotpluggable device (e.g. a USB device), then when a disconnect
127 happens the parent device becomes invalid. Since v4l2_device has a pointer to
128 that parent device it has to be cleared as well to mark that the parent is
129 gone. To do this call:
131 v4l2_device_disconnect(struct v4l2_device *v4l2_dev);
133 This does *not* unregister the subdevs, so you still need to call the
134 v4l2_device_unregister() function for that. If your driver is not hotpluggable,
135 then there is no need to call v4l2_device_disconnect().
137 Sometimes you need to iterate over all devices registered by a specific
138 driver. This is usually the case if multiple device drivers use the same
139 hardware. E.g. the ivtvfb driver is a framebuffer driver that uses the ivtv
140 hardware. The same is true for alsa drivers for example.
142 You can iterate over all registered devices as follows:
144 static int callback(struct device *dev, void *p)
146 struct v4l2_device *v4l2_dev = dev_get_drvdata(dev);
148 /* test if this device was inited */
149 if (v4l2_dev == NULL)
157 struct device_driver *drv;
160 /* Find driver 'ivtv' on the PCI bus.
161 pci_bus_type is a global. For USB busses use usb_bus_type. */
162 drv = driver_find("ivtv", &pci_bus_type);
163 /* iterate over all ivtv device instances */
164 err = driver_for_each_device(drv, NULL, p, callback);
169 Sometimes you need to keep a running counter of the device instance. This is
170 commonly used to map a device instance to an index of a module option array.
172 The recommended approach is as follows:
174 static atomic_t drv_instance = ATOMIC_INIT(0);
176 static int drv_probe(struct pci_dev *pdev, const struct pci_device_id *pci_id)
179 state->instance = atomic_inc_return(&drv_instance) - 1;
182 If you have multiple device nodes then it can be difficult to know when it is
183 safe to unregister v4l2_device for hotpluggable devices. For this purpose
184 v4l2_device has refcounting support. The refcount is increased whenever
185 video_register_device is called and it is decreased whenever that device node
186 is released. When the refcount reaches zero, then the v4l2_device release()
187 callback is called. You can do your final cleanup there.
189 If other device nodes (e.g. ALSA) are created, then you can increase and
190 decrease the refcount manually as well by calling:
192 void v4l2_device_get(struct v4l2_device *v4l2_dev);
196 int v4l2_device_put(struct v4l2_device *v4l2_dev);
198 Since the initial refcount is 1 you also need to call v4l2_device_put in the
199 disconnect() callback (for USB devices) or in the remove() callback (for e.g.
200 PCI devices), otherwise the refcount will never reach 0.
205 Many drivers need to communicate with sub-devices. These devices can do all
206 sort of tasks, but most commonly they handle audio and/or video muxing,
207 encoding or decoding. For webcams common sub-devices are sensors and camera
210 Usually these are I2C devices, but not necessarily. In order to provide the
211 driver with a consistent interface to these sub-devices the v4l2_subdev struct
212 (v4l2-subdev.h) was created.
214 Each sub-device driver must have a v4l2_subdev struct. This struct can be
215 stand-alone for simple sub-devices or it might be embedded in a larger struct
216 if more state information needs to be stored. Usually there is a low-level
217 device struct (e.g. i2c_client) that contains the device data as setup
218 by the kernel. It is recommended to store that pointer in the private
219 data of v4l2_subdev using v4l2_set_subdevdata(). That makes it easy to go
220 from a v4l2_subdev to the actual low-level bus-specific device data.
222 You also need a way to go from the low-level struct to v4l2_subdev. For the
223 common i2c_client struct the i2c_set_clientdata() call is used to store a
224 v4l2_subdev pointer, for other busses you may have to use other methods.
226 Bridges might also need to store per-subdev private data, such as a pointer to
227 bridge-specific per-subdev private data. The v4l2_subdev structure provides
228 host private data for that purpose that can be accessed with
229 v4l2_get_subdev_hostdata() and v4l2_set_subdev_hostdata().
231 From the bridge driver perspective you load the sub-device module and somehow
232 obtain the v4l2_subdev pointer. For i2c devices this is easy: you call
233 i2c_get_clientdata(). For other busses something similar needs to be done.
234 Helper functions exists for sub-devices on an I2C bus that do most of this
237 Each v4l2_subdev contains function pointers that sub-device drivers can
238 implement (or leave NULL if it is not applicable). Since sub-devices can do
239 so many different things and you do not want to end up with a huge ops struct
240 of which only a handful of ops are commonly implemented, the function pointers
241 are sorted according to category and each category has its own ops struct.
243 The top-level ops struct contains pointers to the category ops structs, which
244 may be NULL if the subdev driver does not support anything from that category.
248 struct v4l2_subdev_core_ops {
249 int (*log_status)(struct v4l2_subdev *sd);
250 int (*init)(struct v4l2_subdev *sd, u32 val);
254 struct v4l2_subdev_tuner_ops {
258 struct v4l2_subdev_audio_ops {
262 struct v4l2_subdev_video_ops {
266 struct v4l2_subdev_pad_ops {
270 struct v4l2_subdev_ops {
271 const struct v4l2_subdev_core_ops *core;
272 const struct v4l2_subdev_tuner_ops *tuner;
273 const struct v4l2_subdev_audio_ops *audio;
274 const struct v4l2_subdev_video_ops *video;
275 const struct v4l2_subdev_pad_ops *video;
278 The core ops are common to all subdevs, the other categories are implemented
279 depending on the sub-device. E.g. a video device is unlikely to support the
280 audio ops and vice versa.
282 This setup limits the number of function pointers while still making it easy
283 to add new ops and categories.
285 A sub-device driver initializes the v4l2_subdev struct using:
287 v4l2_subdev_init(sd, &ops);
289 Afterwards you need to initialize subdev->name with a unique name and set the
290 module owner. This is done for you if you use the i2c helper functions.
292 If integration with the media framework is needed, you must initialize the
293 media_entity struct embedded in the v4l2_subdev struct (entity field) by
294 calling media_entity_init():
296 struct media_pad *pads = &my_sd->pads;
299 err = media_entity_init(&sd->entity, npads, pads, 0);
301 The pads array must have been previously initialized. There is no need to
302 manually set the struct media_entity type and name fields, but the revision
303 field must be initialized if needed.
305 A reference to the entity will be automatically acquired/released when the
306 subdev device node (if any) is opened/closed.
308 Don't forget to cleanup the media entity before the sub-device is destroyed:
310 media_entity_cleanup(&sd->entity);
312 If the subdev driver intends to process video and integrate with the media
313 framework, it must implement format related functionality using
314 v4l2_subdev_pad_ops instead of v4l2_subdev_video_ops.
316 In that case, the subdev driver may set the link_validate field to provide
317 its own link validation function. The link validation function is called for
318 every link in the pipeline where both of the ends of the links are V4L2
319 sub-devices. The driver is still responsible for validating the correctness
320 of the format configuration between sub-devices and video nodes.
322 If link_validate op is not set, the default function
323 v4l2_subdev_link_validate_default() is used instead. This function ensures
324 that width, height and the media bus pixel code are equal on both source and
325 sink of the link. Subdev drivers are also free to use this function to
326 perform the checks mentioned above in addition to their own checks.
328 A device (bridge) driver needs to register the v4l2_subdev with the
331 int err = v4l2_device_register_subdev(v4l2_dev, sd);
333 This can fail if the subdev module disappeared before it could be registered.
334 After this function was called successfully the subdev->dev field points to
337 If the v4l2_device parent device has a non-NULL mdev field, the sub-device
338 entity will be automatically registered with the media device.
340 You can unregister a sub-device using:
342 v4l2_device_unregister_subdev(sd);
344 Afterwards the subdev module can be unloaded and sd->dev == NULL.
346 You can call an ops function either directly:
348 err = sd->ops->core->g_std(sd, &norm);
350 but it is better and easier to use this macro:
352 err = v4l2_subdev_call(sd, core, g_std, &norm);
354 The macro will to the right NULL pointer checks and returns -ENODEV if subdev
355 is NULL, -ENOIOCTLCMD if either subdev->core or subdev->core->g_std is
356 NULL, or the actual result of the subdev->ops->core->g_std ops.
358 It is also possible to call all or a subset of the sub-devices:
360 v4l2_device_call_all(v4l2_dev, 0, core, g_std, &norm);
362 Any subdev that does not support this ops is skipped and error results are
363 ignored. If you want to check for errors use this:
365 err = v4l2_device_call_until_err(v4l2_dev, 0, core, g_std, &norm);
367 Any error except -ENOIOCTLCMD will exit the loop with that error. If no
368 errors (except -ENOIOCTLCMD) occurred, then 0 is returned.
370 The second argument to both calls is a group ID. If 0, then all subdevs are
371 called. If non-zero, then only those whose group ID match that value will
372 be called. Before a bridge driver registers a subdev it can set sd->grp_id
373 to whatever value it wants (it's 0 by default). This value is owned by the
374 bridge driver and the sub-device driver will never modify or use it.
376 The group ID gives the bridge driver more control how callbacks are called.
377 For example, there may be multiple audio chips on a board, each capable of
378 changing the volume. But usually only one will actually be used when the
379 user want to change the volume. You can set the group ID for that subdev to
380 e.g. AUDIO_CONTROLLER and specify that as the group ID value when calling
381 v4l2_device_call_all(). That ensures that it will only go to the subdev
384 If the sub-device needs to notify its v4l2_device parent of an event, then
385 it can call v4l2_subdev_notify(sd, notification, arg). This macro checks
386 whether there is a notify() callback defined and returns -ENODEV if not.
387 Otherwise the result of the notify() call is returned.
389 The advantage of using v4l2_subdev is that it is a generic struct and does
390 not contain any knowledge about the underlying hardware. So a driver might
391 contain several subdevs that use an I2C bus, but also a subdev that is
392 controlled through GPIO pins. This distinction is only relevant when setting
393 up the device, but once the subdev is registered it is completely transparent.
396 V4L2 sub-device userspace API
397 -----------------------------
399 Beside exposing a kernel API through the v4l2_subdev_ops structure, V4L2
400 sub-devices can also be controlled directly by userspace applications.
402 Device nodes named v4l-subdevX can be created in /dev to access sub-devices
403 directly. If a sub-device supports direct userspace configuration it must set
404 the V4L2_SUBDEV_FL_HAS_DEVNODE flag before being registered.
406 After registering sub-devices, the v4l2_device driver can create device nodes
407 for all registered sub-devices marked with V4L2_SUBDEV_FL_HAS_DEVNODE by calling
408 v4l2_device_register_subdev_nodes(). Those device nodes will be automatically
409 removed when sub-devices are unregistered.
411 The device node handles a subset of the V4L2 API.
421 The controls ioctls are identical to the ones defined in V4L2. They
422 behave identically, with the only exception that they deal only with
423 controls implemented in the sub-device. Depending on the driver, those
424 controls can be also be accessed through one (or several) V4L2 device
428 VIDIOC_SUBSCRIBE_EVENT
429 VIDIOC_UNSUBSCRIBE_EVENT
431 The events ioctls are identical to the ones defined in V4L2. They
432 behave identically, with the only exception that they deal only with
433 events generated by the sub-device. Depending on the driver, those
434 events can also be reported by one (or several) V4L2 device nodes.
436 Sub-device drivers that want to use events need to set the
437 V4L2_SUBDEV_USES_EVENTS v4l2_subdev::flags and initialize
438 v4l2_subdev::nevents to events queue depth before registering the
439 sub-device. After registration events can be queued as usual on the
440 v4l2_subdev::devnode device node.
442 To properly support events, the poll() file operation is also
447 All ioctls not in the above list are passed directly to the sub-device
448 driver through the core::ioctl operation.
451 I2C sub-device drivers
452 ----------------------
454 Since these drivers are so common, special helper functions are available to
455 ease the use of these drivers (v4l2-common.h).
457 The recommended method of adding v4l2_subdev support to an I2C driver is to
458 embed the v4l2_subdev struct into the state struct that is created for each
459 I2C device instance. Very simple devices have no state struct and in that case
460 you can just create a v4l2_subdev directly.
462 A typical state struct would look like this (where 'chipname' is replaced by
463 the name of the chip):
465 struct chipname_state {
466 struct v4l2_subdev sd;
467 ... /* additional state fields */
470 Initialize the v4l2_subdev struct as follows:
472 v4l2_i2c_subdev_init(&state->sd, client, subdev_ops);
474 This function will fill in all the fields of v4l2_subdev and ensure that the
475 v4l2_subdev and i2c_client both point to one another.
477 You should also add a helper inline function to go from a v4l2_subdev pointer
478 to a chipname_state struct:
480 static inline struct chipname_state *to_state(struct v4l2_subdev *sd)
482 return container_of(sd, struct chipname_state, sd);
485 Use this to go from the v4l2_subdev struct to the i2c_client struct:
487 struct i2c_client *client = v4l2_get_subdevdata(sd);
489 And this to go from an i2c_client to a v4l2_subdev struct:
491 struct v4l2_subdev *sd = i2c_get_clientdata(client);
493 Make sure to call v4l2_device_unregister_subdev(sd) when the remove() callback
494 is called. This will unregister the sub-device from the bridge driver. It is
495 safe to call this even if the sub-device was never registered.
497 You need to do this because when the bridge driver destroys the i2c adapter
498 the remove() callbacks are called of the i2c devices on that adapter.
499 After that the corresponding v4l2_subdev structures are invalid, so they
500 have to be unregistered first. Calling v4l2_device_unregister_subdev(sd)
501 from the remove() callback ensures that this is always done correctly.
504 The bridge driver also has some helper functions it can use:
506 struct v4l2_subdev *sd = v4l2_i2c_new_subdev(v4l2_dev, adapter,
507 "module_foo", "chipid", 0x36, NULL);
509 This loads the given module (can be NULL if no module needs to be loaded) and
510 calls i2c_new_device() with the given i2c_adapter and chip/address arguments.
511 If all goes well, then it registers the subdev with the v4l2_device.
513 You can also use the last argument of v4l2_i2c_new_subdev() to pass an array
514 of possible I2C addresses that it should probe. These probe addresses are
515 only used if the previous argument is 0. A non-zero argument means that you
516 know the exact i2c address so in that case no probing will take place.
518 Both functions return NULL if something went wrong.
520 Note that the chipid you pass to v4l2_i2c_new_subdev() is usually
521 the same as the module name. It allows you to specify a chip variant, e.g.
522 "saa7114" or "saa7115". In general though the i2c driver autodetects this.
523 The use of chipid is something that needs to be looked at more closely at a
524 later date. It differs between i2c drivers and as such can be confusing.
525 To see which chip variants are supported you can look in the i2c driver code
526 for the i2c_device_id table. This lists all the possibilities.
528 There are two more helper functions:
530 v4l2_i2c_new_subdev_cfg: this function adds new irq and platform_data
531 arguments and has both 'addr' and 'probed_addrs' arguments: if addr is not
532 0 then that will be used (non-probing variant), otherwise the probed_addrs
535 For example: this will probe for address 0x10:
537 struct v4l2_subdev *sd = v4l2_i2c_new_subdev_cfg(v4l2_dev, adapter,
538 "module_foo", "chipid", 0, NULL, 0, I2C_ADDRS(0x10));
540 v4l2_i2c_new_subdev_board uses an i2c_board_info struct which is passed
541 to the i2c driver and replaces the irq, platform_data and addr arguments.
543 If the subdev supports the s_config core ops, then that op is called with
544 the irq and platform_data arguments after the subdev was setup. The older
545 v4l2_i2c_new_(probed_)subdev functions will call s_config as well, but with
546 irq set to 0 and platform_data set to NULL.
551 The actual device nodes in the /dev directory are created using the
552 video_device struct (v4l2-dev.h). This struct can either be allocated
553 dynamically or embedded in a larger struct.
555 To allocate it dynamically use:
557 struct video_device *vdev = video_device_alloc();
562 vdev->release = video_device_release;
564 If you embed it in a larger struct, then you must set the release()
565 callback to your own function:
567 struct video_device *vdev = &my_vdev->vdev;
569 vdev->release = my_vdev_release;
571 The release callback must be set and it is called when the last user
572 of the video device exits.
574 The default video_device_release() callback just calls kfree to free the
577 There is also a video_device_release_empty() function that does nothing
578 (is empty) and can be used if the struct is embedded and there is nothing
579 to do when it is released.
581 You should also set these fields:
583 - v4l2_dev: must be set to the v4l2_device parent device.
585 - name: set to something descriptive and unique.
587 - vfl_dir: set this to VFL_DIR_RX for capture devices (VFL_DIR_RX has value 0,
588 so this is normally already the default), set to VFL_DIR_TX for output
589 devices and VFL_DIR_M2M for mem2mem (codec) devices.
591 - fops: set to the v4l2_file_operations struct.
593 - ioctl_ops: if you use the v4l2_ioctl_ops to simplify ioctl maintenance
594 (highly recommended to use this and it might become compulsory in the
595 future!), then set this to your v4l2_ioctl_ops struct. The vfl_type and
596 vfl_dir fields are used to disable ops that do not match the type/dir
597 combination. E.g. VBI ops are disabled for non-VBI nodes, and output ops
598 are disabled for a capture device. This makes it possible to provide
599 just one v4l2_ioctl_ops struct for both vbi and video nodes.
601 - lock: leave to NULL if you want to do all the locking in the driver.
602 Otherwise you give it a pointer to a struct mutex_lock and before the
603 unlocked_ioctl file operation is called this lock will be taken by the
604 core and released afterwards. See the next section for more details.
606 - queue: a pointer to the struct vb2_queue associated with this device node.
607 If queue is non-NULL, and queue->lock is non-NULL, then queue->lock is
608 used for the queuing ioctls (VIDIOC_REQBUFS, CREATE_BUFS, QBUF, DQBUF,
609 QUERYBUF, PREPARE_BUF, STREAMON and STREAMOFF) instead of the lock above.
610 That way the vb2 queuing framework does not have to wait for other ioctls.
611 This queue pointer is also used by the vb2 helper functions to check for
612 queuing ownership (i.e. is the filehandle calling it allowed to do the
615 - prio: keeps track of the priorities. Used to implement VIDIOC_G/S_PRIORITY.
616 If left to NULL, then it will use the struct v4l2_prio_state in v4l2_device.
617 If you want to have a separate priority state per (group of) device node(s),
618 then you can point it to your own struct v4l2_prio_state.
620 - dev_parent: you only set this if v4l2_device was registered with NULL as
621 the parent device struct. This only happens in cases where one hardware
622 device has multiple PCI devices that all share the same v4l2_device core.
624 The cx88 driver is an example of this: one core v4l2_device struct, but
625 it is used by both a raw video PCI device (cx8800) and a MPEG PCI device
626 (cx8802). Since the v4l2_device cannot be associated with two PCI devices
627 at the same time it is setup without a parent device. But when the struct
628 video_device is initialized you *do* know which parent PCI device to use and
629 so you set dev_device to the correct PCI device.
631 - flags: optional. Set to V4L2_FL_USE_FH_PRIO if you want to let the framework
632 handle the VIDIOC_G/S_PRIORITY ioctls. This requires that you use struct
633 v4l2_fh. Eventually this flag will disappear once all drivers use the core
634 priority handling. But for now it has to be set explicitly.
636 If you use v4l2_ioctl_ops, then you should set .unlocked_ioctl to video_ioctl2
637 in your v4l2_file_operations struct.
639 Do not use .ioctl! This is deprecated and will go away in the future.
641 In some cases you want to tell the core that a function you had specified in
642 your v4l2_ioctl_ops should be ignored. You can mark such ioctls by calling this
643 function before video_device_register is called:
645 void v4l2_disable_ioctl(struct video_device *vdev, unsigned int cmd);
647 This tends to be needed if based on external factors (e.g. which card is
648 being used) you want to turns off certain features in v4l2_ioctl_ops without
649 having to make a new struct.
651 The v4l2_file_operations struct is a subset of file_operations. The main
652 difference is that the inode argument is omitted since it is never used.
654 If integration with the media framework is needed, you must initialize the
655 media_entity struct embedded in the video_device struct (entity field) by
656 calling media_entity_init():
658 struct media_pad *pad = &my_vdev->pad;
661 err = media_entity_init(&vdev->entity, 1, pad, 0);
663 The pads array must have been previously initialized. There is no need to
664 manually set the struct media_entity type and name fields.
666 A reference to the entity will be automatically acquired/released when the
667 video device is opened/closed.
672 The V4L core provides optional locking services. The main service is the
673 lock field in struct video_device, which is a pointer to a mutex. If you set
674 this pointer, then that will be used by unlocked_ioctl to serialize all ioctls.
676 If you are using the videobuf2 framework, then there is a second lock that you
677 can set: video_device->queue->lock. If set, then this lock will be used instead
678 of video_device->lock to serialize all queuing ioctls (see the previous section
679 for the full list of those ioctls).
681 The advantage of using a different lock for the queuing ioctls is that for some
682 drivers (particularly USB drivers) certain commands such as setting controls
683 can take a long time, so you want to use a separate lock for the buffer queuing
684 ioctls. That way your VIDIOC_DQBUF doesn't stall because the driver is busy
685 changing the e.g. exposure of the webcam.
687 Of course, you can always do all the locking yourself by leaving both lock
690 If you use the old videobuf then you must pass the video_device lock to the
691 videobuf queue initialize function: if videobuf has to wait for a frame to
692 arrive, then it will temporarily unlock the lock and relock it afterwards. If
693 your driver also waits in the code, then you should do the same to allow other
694 processes to access the device node while the first process is waiting for
697 In the case of videobuf2 you will need to implement the wait_prepare and
698 wait_finish callbacks to unlock/lock if applicable. If you use the queue->lock
699 pointer, then you can use the helper functions vb2_ops_wait_prepare/finish.
701 The implementation of a hotplug disconnect should also take the lock from
702 video_device before calling v4l2_device_disconnect. If you are also using
703 video_device->queue->lock, then you have to first lock video_device->queue->lock
704 followed by video_device->lock. That way you can be sure no ioctl is running
705 when you call v4l2_device_disconnect.
707 video_device registration
708 -------------------------
710 Next you register the video device: this will create the character device
713 err = video_register_device(vdev, VFL_TYPE_GRABBER, -1);
715 video_device_release(vdev); /* or kfree(my_vdev); */
719 If the v4l2_device parent device has a non-NULL mdev field, the video device
720 entity will be automatically registered with the media device.
722 Which device is registered depends on the type argument. The following
725 VFL_TYPE_GRABBER: videoX for video input/output devices
726 VFL_TYPE_VBI: vbiX for vertical blank data (i.e. closed captions, teletext)
727 VFL_TYPE_RADIO: radioX for radio tuners
729 The last argument gives you a certain amount of control over the device
730 device node number used (i.e. the X in videoX). Normally you will pass -1
731 to let the v4l2 framework pick the first free number. But sometimes users
732 want to select a specific node number. It is common that drivers allow
733 the user to select a specific device node number through a driver module
734 option. That number is then passed to this function and video_register_device
735 will attempt to select that device node number. If that number was already
736 in use, then the next free device node number will be selected and it
737 will send a warning to the kernel log.
739 Another use-case is if a driver creates many devices. In that case it can
740 be useful to place different video devices in separate ranges. For example,
741 video capture devices start at 0, video output devices start at 16.
742 So you can use the last argument to specify a minimum device node number
743 and the v4l2 framework will try to pick the first free number that is equal
744 or higher to what you passed. If that fails, then it will just pick the
747 Since in this case you do not care about a warning about not being able
748 to select the specified device node number, you can call the function
749 video_register_device_no_warn() instead.
751 Whenever a device node is created some attributes are also created for you.
752 If you look in /sys/class/video4linux you see the devices. Go into e.g.
753 video0 and you will see 'name' and 'index' attributes. The 'name' attribute
754 is the 'name' field of the video_device struct.
756 The 'index' attribute is the index of the device node: for each call to
757 video_register_device() the index is just increased by 1. The first video
758 device node you register always starts with index 0.
760 Users can setup udev rules that utilize the index attribute to make fancy
761 device names (e.g. 'mpegX' for MPEG video capture device nodes).
763 After the device was successfully registered, then you can use these fields:
765 - vfl_type: the device type passed to video_register_device.
766 - minor: the assigned device minor number.
767 - num: the device node number (i.e. the X in videoX).
768 - index: the device index number.
770 If the registration failed, then you need to call video_device_release()
771 to free the allocated video_device struct, or free your own struct if the
772 video_device was embedded in it. The vdev->release() callback will never
773 be called if the registration failed, nor should you ever attempt to
774 unregister the device if the registration failed.
780 When the video device nodes have to be removed, either during the unload
781 of the driver or because the USB device was disconnected, then you should
784 video_unregister_device(vdev);
786 This will remove the device nodes from sysfs (causing udev to remove them
789 After video_unregister_device() returns no new opens can be done. However,
790 in the case of USB devices some application might still have one of these
791 device nodes open. So after the unregister all file operations (except
792 release, of course) will return an error as well.
794 When the last user of the video device node exits, then the vdev->release()
795 callback is called and you can do the final cleanup there.
797 Don't forget to cleanup the media entity associated with the video device if
798 it has been initialized:
800 media_entity_cleanup(&vdev->entity);
802 This can be done from the release callback.
805 video_device helper functions
806 -----------------------------
808 There are a few useful helper functions:
810 - file/video_device private data
812 You can set/get driver private data in the video_device struct using:
814 void *video_get_drvdata(struct video_device *vdev);
815 void video_set_drvdata(struct video_device *vdev, void *data);
817 Note that you can safely call video_set_drvdata() before calling
818 video_register_device().
822 struct video_device *video_devdata(struct file *file);
824 returns the video_device belonging to the file struct.
826 The video_drvdata function combines video_get_drvdata with video_devdata:
828 void *video_drvdata(struct file *file);
830 You can go from a video_device struct to the v4l2_device struct using:
832 struct v4l2_device *v4l2_dev = vdev->v4l2_dev;
836 The video_device node kernel name can be retrieved using
838 const char *video_device_node_name(struct video_device *vdev);
840 The name is used as a hint by userspace tools such as udev. The function
841 should be used where possible instead of accessing the video_device::num and
842 video_device::minor fields.
845 video buffer helper functions
846 -----------------------------
848 The v4l2 core API provides a set of standard methods (called "videobuf")
849 for dealing with video buffers. Those methods allow a driver to implement
850 read(), mmap() and overlay() in a consistent way. There are currently
851 methods for using video buffers on devices that supports DMA with
852 scatter/gather method (videobuf-dma-sg), DMA with linear access
853 (videobuf-dma-contig), and vmalloced buffers, mostly used on USB drivers
856 Please see Documentation/video4linux/videobuf for more information on how
857 to use the videobuf layer.
862 struct v4l2_fh provides a way to easily keep file handle specific data
863 that is used by the V4L2 framework. New drivers must use struct v4l2_fh
864 since it is also used to implement priority handling (VIDIOC_G/S_PRIORITY)
865 if the video_device flag V4L2_FL_USE_FH_PRIO is also set.
867 The users of v4l2_fh (in the V4L2 framework, not the driver) know
868 whether a driver uses v4l2_fh as its file->private_data pointer by
869 testing the V4L2_FL_USES_V4L2_FH bit in video_device->flags. This bit is
870 set whenever v4l2_fh_init() is called.
872 struct v4l2_fh is allocated as a part of the driver's own file handle
873 structure and file->private_data is set to it in the driver's open
874 function by the driver.
876 In many cases the struct v4l2_fh will be embedded in a larger structure.
877 In that case you should call v4l2_fh_init+v4l2_fh_add in open() and
878 v4l2_fh_del+v4l2_fh_exit in release().
880 Drivers can extract their own file handle structure by using the container_of
890 int my_open(struct file *file)
893 struct video_device *vfd;
898 my_fh = kzalloc(sizeof(*my_fh), GFP_KERNEL);
902 v4l2_fh_init(&my_fh->fh, vfd);
906 file->private_data = &my_fh->fh;
907 v4l2_fh_add(&my_fh->fh);
911 int my_release(struct file *file)
913 struct v4l2_fh *fh = file->private_data;
914 struct my_fh *my_fh = container_of(fh, struct my_fh, fh);
917 v4l2_fh_del(&my_fh->fh);
918 v4l2_fh_exit(&my_fh->fh);
923 Below is a short description of the v4l2_fh functions used:
925 void v4l2_fh_init(struct v4l2_fh *fh, struct video_device *vdev)
927 Initialise the file handle. This *MUST* be performed in the driver's
928 v4l2_file_operations->open() handler.
930 void v4l2_fh_add(struct v4l2_fh *fh)
932 Add a v4l2_fh to video_device file handle list. Must be called once the
933 file handle is completely initialized.
935 void v4l2_fh_del(struct v4l2_fh *fh)
937 Unassociate the file handle from video_device(). The file handle
938 exit function may now be called.
940 void v4l2_fh_exit(struct v4l2_fh *fh)
942 Uninitialise the file handle. After uninitialisation the v4l2_fh
946 If struct v4l2_fh is not embedded, then you can use these helper functions:
948 int v4l2_fh_open(struct file *filp)
950 This allocates a struct v4l2_fh, initializes it and adds it to the struct
951 video_device associated with the file struct.
953 int v4l2_fh_release(struct file *filp)
955 This deletes it from the struct video_device associated with the file
956 struct, uninitialised the v4l2_fh and frees it.
958 These two functions can be plugged into the v4l2_file_operation's open() and
962 Several drivers need to do something when the first file handle is opened and
963 when the last file handle closes. Two helper functions were added to check
964 whether the v4l2_fh struct is the only open filehandle of the associated
967 int v4l2_fh_is_singular(struct v4l2_fh *fh)
969 Returns 1 if the file handle is the only open file handle, else 0.
971 int v4l2_fh_is_singular_file(struct file *filp)
973 Same, but it calls v4l2_fh_is_singular with filp->private_data.
979 The V4L2 events provide a generic way to pass events to user space.
980 The driver must use v4l2_fh to be able to support V4L2 events.
982 Events are defined by a type and an optional ID. The ID may refer to a V4L2
983 object such as a control ID. If unused, then the ID is 0.
985 When the user subscribes to an event the driver will allocate a number of
986 kevent structs for that event. So every (type, ID) event tuple will have
987 its own set of kevent structs. This guarantees that if a driver is generating
988 lots of events of one type in a short time, then that will not overwrite
989 events of another type.
991 But if you get more events of one type than the number of kevents that were
992 reserved, then the oldest event will be dropped and the new one added.
994 Furthermore, the internal struct v4l2_subscribed_event has merge() and
995 replace() callbacks which drivers can set. These callbacks are called when
996 a new event is raised and there is no more room. The replace() callback
997 allows you to replace the payload of the old event with that of the new event,
998 merging any relevant data from the old payload into the new payload that
999 replaces it. It is called when this event type has only one kevent struct
1000 allocated. The merge() callback allows you to merge the oldest event payload
1001 into that of the second-oldest event payload. It is called when there are two
1002 or more kevent structs allocated.
1004 This way no status information is lost, just the intermediate steps leading
1007 A good example of these replace/merge callbacks is in v4l2-event.c:
1008 ctrls_replace() and ctrls_merge() callbacks for the control event.
1010 Note: these callbacks can be called from interrupt context, so they must be
1015 void v4l2_event_queue(struct video_device *vdev, const struct v4l2_event *ev)
1017 Queue events to video device. The driver's only responsibility is to fill
1018 in the type and the data fields. The other fields will be filled in by
1021 int v4l2_event_subscribe(struct v4l2_fh *fh,
1022 struct v4l2_event_subscription *sub, unsigned elems,
1023 const struct v4l2_subscribed_event_ops *ops)
1025 The video_device->ioctl_ops->vidioc_subscribe_event must check the driver
1026 is able to produce events with specified event id. Then it calls
1027 v4l2_event_subscribe() to subscribe the event.
1029 The elems argument is the size of the event queue for this event. If it is 0,
1030 then the framework will fill in a default value (this depends on the event
1033 The ops argument allows the driver to specify a number of callbacks:
1034 * add: called when a new listener gets added (subscribing to the same
1035 event twice will only cause this callback to get called once)
1036 * del: called when a listener stops listening
1037 * replace: replace event 'old' with event 'new'.
1038 * merge: merge event 'old' into event 'new'.
1039 All 4 callbacks are optional, if you don't want to specify any callbacks
1040 the ops argument itself maybe NULL.
1042 int v4l2_event_unsubscribe(struct v4l2_fh *fh,
1043 struct v4l2_event_subscription *sub)
1045 vidioc_unsubscribe_event in struct v4l2_ioctl_ops. A driver may use
1046 v4l2_event_unsubscribe() directly unless it wants to be involved in
1047 unsubscription process.
1049 The special type V4L2_EVENT_ALL may be used to unsubscribe all events. The
1050 drivers may want to handle this in a special way.
1052 int v4l2_event_pending(struct v4l2_fh *fh)
1054 Returns the number of pending events. Useful when implementing poll.
1056 Events are delivered to user space through the poll system call. The driver
1057 can use v4l2_fh->wait (a wait_queue_head_t) as the argument for poll_wait().
1059 There are standard and private events. New standard events must use the
1060 smallest available event type. The drivers must allocate their events from
1061 their own class starting from class base. Class base is
1062 V4L2_EVENT_PRIVATE_START + n * 1000 where n is the lowest available number.
1063 The first event type in the class is reserved for future use, so the first
1064 available event type is 'class base + 1'.
1066 An example on how the V4L2 events may be used can be found in the OMAP
1067 3 ISP driver (drivers/media/platform/omap3isp).