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, /dev/radioX and
48 /dev/vtxX) 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 in the future a v4l2_fh struct will keep track of filehandle instances
72 (this is not yet implemented).
78 Each device instance is represented by a struct v4l2_device (v4l2-device.h).
79 Very simple devices can just allocate this struct, but most of the time you
80 would embed this struct inside a larger struct.
82 You must register the device instance:
84 v4l2_device_register(struct device *dev, struct v4l2_device *v4l2_dev);
86 Registration will initialize the v4l2_device struct and link dev->driver_data
87 to v4l2_dev. If v4l2_dev->name is empty then it will be set to a value derived
88 from dev (driver name followed by the bus_id, to be precise). If you set it
89 up before calling v4l2_device_register then it will be untouched. If dev is
90 NULL, then you *must* setup v4l2_dev->name before calling v4l2_device_register.
92 The first 'dev' argument is normally the struct device pointer of a pci_dev,
93 usb_device or platform_device. It is rare for dev to be NULL, but it happens
94 with ISA devices, for example.
98 v4l2_device_unregister(struct v4l2_device *v4l2_dev);
100 Unregistering will also automatically unregister all subdevs from the device.
102 Sometimes you need to iterate over all devices registered by a specific
103 driver. This is usually the case if multiple device drivers use the same
104 hardware. E.g. the ivtvfb driver is a framebuffer driver that uses the ivtv
105 hardware. The same is true for alsa drivers for example.
107 You can iterate over all registered devices as follows:
109 static int callback(struct device *dev, void *p)
111 struct v4l2_device *v4l2_dev = dev_get_drvdata(dev);
113 /* test if this device was inited */
114 if (v4l2_dev == NULL)
122 struct device_driver *drv;
125 /* Find driver 'ivtv' on the PCI bus.
126 pci_bus_type is a global. For USB busses use usb_bus_type. */
127 drv = driver_find("ivtv", &pci_bus_type);
128 /* iterate over all ivtv device instances */
129 err = driver_for_each_device(drv, NULL, p, callback);
134 Sometimes you need to keep a running counter of the device instance. This is
135 commonly used to map a device instance to an index of a module option array.
137 The recommended approach is as follows:
139 static atomic_t drv_instance = ATOMIC_INIT(0);
141 static int __devinit drv_probe(struct pci_dev *pdev,
142 const struct pci_device_id *pci_id)
145 state->instance = atomic_inc_return(&drv_instance) - 1;
152 Many drivers need to communicate with sub-devices. These devices can do all
153 sort of tasks, but most commonly they handle audio and/or video muxing,
154 encoding or decoding. For webcams common sub-devices are sensors and camera
157 Usually these are I2C devices, but not necessarily. In order to provide the
158 driver with a consistent interface to these sub-devices the v4l2_subdev struct
159 (v4l2-subdev.h) was created.
161 Each sub-device driver must have a v4l2_subdev struct. This struct can be
162 stand-alone for simple sub-devices or it might be embedded in a larger struct
163 if more state information needs to be stored. Usually there is a low-level
164 device struct (e.g. i2c_client) that contains the device data as setup
165 by the kernel. It is recommended to store that pointer in the private
166 data of v4l2_subdev using v4l2_set_subdevdata(). That makes it easy to go
167 from a v4l2_subdev to the actual low-level bus-specific device data.
169 You also need a way to go from the low-level struct to v4l2_subdev. For the
170 common i2c_client struct the i2c_set_clientdata() call is used to store a
171 v4l2_subdev pointer, for other busses you may have to use other methods.
173 From the bridge driver perspective you load the sub-device module and somehow
174 obtain the v4l2_subdev pointer. For i2c devices this is easy: you call
175 i2c_get_clientdata(). For other busses something similar needs to be done.
176 Helper functions exists for sub-devices on an I2C bus that do most of this
179 Each v4l2_subdev contains function pointers that sub-device drivers can
180 implement (or leave NULL if it is not applicable). Since sub-devices can do
181 so many different things and you do not want to end up with a huge ops struct
182 of which only a handful of ops are commonly implemented, the function pointers
183 are sorted according to category and each category has its own ops struct.
185 The top-level ops struct contains pointers to the category ops structs, which
186 may be NULL if the subdev driver does not support anything from that category.
190 struct v4l2_subdev_core_ops {
191 int (*g_chip_ident)(struct v4l2_subdev *sd, struct v4l2_dbg_chip_ident *chip);
192 int (*log_status)(struct v4l2_subdev *sd);
193 int (*init)(struct v4l2_subdev *sd, u32 val);
197 struct v4l2_subdev_tuner_ops {
201 struct v4l2_subdev_audio_ops {
205 struct v4l2_subdev_video_ops {
209 struct v4l2_subdev_ops {
210 const struct v4l2_subdev_core_ops *core;
211 const struct v4l2_subdev_tuner_ops *tuner;
212 const struct v4l2_subdev_audio_ops *audio;
213 const struct v4l2_subdev_video_ops *video;
216 The core ops are common to all subdevs, the other categories are implemented
217 depending on the sub-device. E.g. a video device is unlikely to support the
218 audio ops and vice versa.
220 This setup limits the number of function pointers while still making it easy
221 to add new ops and categories.
223 A sub-device driver initializes the v4l2_subdev struct using:
225 v4l2_subdev_init(sd, &ops);
227 Afterwards you need to initialize subdev->name with a unique name and set the
228 module owner. This is done for you if you use the i2c helper functions.
230 A device (bridge) driver needs to register the v4l2_subdev with the
233 int err = v4l2_device_register_subdev(v4l2_dev, sd);
235 This can fail if the subdev module disappeared before it could be registered.
236 After this function was called successfully the subdev->dev field points to
239 You can unregister a sub-device using:
241 v4l2_device_unregister_subdev(sd);
243 Afterwards the subdev module can be unloaded and sd->dev == NULL.
245 You can call an ops function either directly:
247 err = sd->ops->core->g_chip_ident(sd, &chip);
249 but it is better and easier to use this macro:
251 err = v4l2_subdev_call(sd, core, g_chip_ident, &chip);
253 The macro will to the right NULL pointer checks and returns -ENODEV if subdev
254 is NULL, -ENOIOCTLCMD if either subdev->core or subdev->core->g_chip_ident is
255 NULL, or the actual result of the subdev->ops->core->g_chip_ident ops.
257 It is also possible to call all or a subset of the sub-devices:
259 v4l2_device_call_all(v4l2_dev, 0, core, g_chip_ident, &chip);
261 Any subdev that does not support this ops is skipped and error results are
262 ignored. If you want to check for errors use this:
264 err = v4l2_device_call_until_err(v4l2_dev, 0, core, g_chip_ident, &chip);
266 Any error except -ENOIOCTLCMD will exit the loop with that error. If no
267 errors (except -ENOIOCTLCMD) occured, then 0 is returned.
269 The second argument to both calls is a group ID. If 0, then all subdevs are
270 called. If non-zero, then only those whose group ID match that value will
271 be called. Before a bridge driver registers a subdev it can set sd->grp_id
272 to whatever value it wants (it's 0 by default). This value is owned by the
273 bridge driver and the sub-device driver will never modify or use it.
275 The group ID gives the bridge driver more control how callbacks are called.
276 For example, there may be multiple audio chips on a board, each capable of
277 changing the volume. But usually only one will actually be used when the
278 user want to change the volume. You can set the group ID for that subdev to
279 e.g. AUDIO_CONTROLLER and specify that as the group ID value when calling
280 v4l2_device_call_all(). That ensures that it will only go to the subdev
283 The advantage of using v4l2_subdev is that it is a generic struct and does
284 not contain any knowledge about the underlying hardware. So a driver might
285 contain several subdevs that use an I2C bus, but also a subdev that is
286 controlled through GPIO pins. This distinction is only relevant when setting
287 up the device, but once the subdev is registered it is completely transparent.
290 I2C sub-device drivers
291 ----------------------
293 Since these drivers are so common, special helper functions are available to
294 ease the use of these drivers (v4l2-common.h).
296 The recommended method of adding v4l2_subdev support to an I2C driver is to
297 embed the v4l2_subdev struct into the state struct that is created for each
298 I2C device instance. Very simple devices have no state struct and in that case
299 you can just create a v4l2_subdev directly.
301 A typical state struct would look like this (where 'chipname' is replaced by
302 the name of the chip):
304 struct chipname_state {
305 struct v4l2_subdev sd;
306 ... /* additional state fields */
309 Initialize the v4l2_subdev struct as follows:
311 v4l2_i2c_subdev_init(&state->sd, client, subdev_ops);
313 This function will fill in all the fields of v4l2_subdev and ensure that the
314 v4l2_subdev and i2c_client both point to one another.
316 You should also add a helper inline function to go from a v4l2_subdev pointer
317 to a chipname_state struct:
319 static inline struct chipname_state *to_state(struct v4l2_subdev *sd)
321 return container_of(sd, struct chipname_state, sd);
324 Use this to go from the v4l2_subdev struct to the i2c_client struct:
326 struct i2c_client *client = v4l2_get_subdevdata(sd);
328 And this to go from an i2c_client to a v4l2_subdev struct:
330 struct v4l2_subdev *sd = i2c_get_clientdata(client);
332 Finally you need to make a command function to make driver->command()
333 call the right subdev_ops functions:
335 static int subdev_command(struct i2c_client *client, unsigned cmd, void *arg)
337 return v4l2_subdev_command(i2c_get_clientdata(client), cmd, arg);
340 If driver->command is never used then you can leave this out. Eventually the
341 driver->command usage should be removed from v4l.
343 Make sure to call v4l2_device_unregister_subdev(sd) when the remove() callback
344 is called. This will unregister the sub-device from the bridge driver. It is
345 safe to call this even if the sub-device was never registered.
347 You need to do this because when the bridge driver destroys the i2c adapter
348 the remove() callbacks are called of the i2c devices on that adapter.
349 After that the corresponding v4l2_subdev structures are invalid, so they
350 have to be unregistered first. Calling v4l2_device_unregister_subdev(sd)
351 from the remove() callback ensures that this is always done correctly.
354 The bridge driver also has some helper functions it can use:
356 struct v4l2_subdev *sd = v4l2_i2c_new_subdev(adapter, "module_foo", "chipid", 0x36);
358 This loads the given module (can be NULL if no module needs to be loaded) and
359 calls i2c_new_device() with the given i2c_adapter and chip/address arguments.
360 If all goes well, then it registers the subdev with the v4l2_device. It gets
361 the v4l2_device by calling i2c_get_adapdata(adapter), so you should make sure
362 to call i2c_set_adapdata(adapter, v4l2_device) when you setup the i2c_adapter
365 You can also use v4l2_i2c_new_probed_subdev() which is very similar to
366 v4l2_i2c_new_subdev(), except that it has an array of possible I2C addresses
367 that it should probe. Internally it calls i2c_new_probed_device().
369 Both functions return NULL if something went wrong.
371 Note that the chipid you pass to v4l2_i2c_new_(probed_)subdev() is usually
372 the same as the module name. It allows you to specify a chip variant, e.g.
373 "saa7114" or "saa7115". In general though the i2c driver autodetects this.
374 The use of chipid is something that needs to be looked at more closely at a
375 later date. It differs between i2c drivers and as such can be confusing.
376 To see which chip variants are supported you can look in the i2c driver code
377 for the i2c_device_id table. This lists all the possibilities.
383 The actual device nodes in the /dev directory are created using the
384 video_device struct (v4l2-dev.h). This struct can either be allocated
385 dynamically or embedded in a larger struct.
387 To allocate it dynamically use:
389 struct video_device *vdev = video_device_alloc();
394 vdev->release = video_device_release;
396 If you embed it in a larger struct, then you must set the release()
397 callback to your own function:
399 struct video_device *vdev = &my_vdev->vdev;
401 vdev->release = my_vdev_release;
403 The release callback must be set and it is called when the last user
404 of the video device exits.
406 The default video_device_release() callback just calls kfree to free the
409 You should also set these fields:
411 - v4l2_dev: set to the v4l2_device parent device.
412 - name: set to something descriptive and unique.
413 - fops: set to the v4l2_file_operations struct.
414 - ioctl_ops: if you use the v4l2_ioctl_ops to simplify ioctl maintenance
415 (highly recommended to use this and it might become compulsory in the
416 future!), then set this to your v4l2_ioctl_ops struct.
418 If you use v4l2_ioctl_ops, then you should set either .unlocked_ioctl or
419 .ioctl to video_ioctl2 in your v4l2_file_operations struct.
421 The v4l2_file_operations struct is a subset of file_operations. The main
422 difference is that the inode argument is omitted since it is never used.
425 video_device registration
426 -------------------------
428 Next you register the video device: this will create the character device
431 err = video_register_device(vdev, VFL_TYPE_GRABBER, -1);
433 video_device_release(vdev); /* or kfree(my_vdev); */
437 Which device is registered depends on the type argument. The following
440 VFL_TYPE_GRABBER: videoX for video input/output devices
441 VFL_TYPE_VBI: vbiX for vertical blank data (i.e. closed captions, teletext)
442 VFL_TYPE_RADIO: radioX for radio tuners
443 VFL_TYPE_VTX: vtxX for teletext devices (deprecated, don't use)
445 The last argument gives you a certain amount of control over the device
446 kernel number used (i.e. the X in videoX). Normally you will pass -1 to
447 let the v4l2 framework pick the first free number. But if a driver creates
448 many devices, then it can be useful to have different video devices in
449 separate ranges. For example, video capture devices start at 0, video
450 output devices start at 16.
452 So you can use the last argument to specify a minimum kernel number and
453 the v4l2 framework will try to pick the first free number that is equal
454 or higher to what you passed. If that fails, then it will just pick the
457 Whenever a device node is created some attributes are also created for you.
458 If you look in /sys/class/video4linux you see the devices. Go into e.g.
459 video0 and you will see 'name' and 'index' attributes. The 'name' attribute
460 is the 'name' field of the video_device struct. The 'index' attribute is
461 a device node index that can be assigned by the driver, or that is calculated
464 If you call video_register_device(), then the index is just increased by
465 1 for each device node you register. The first video device node you register
466 always starts off with 0.
468 Alternatively you can call video_register_device_index() which is identical
469 to video_register_device(), but with an extra index argument. Here you can
470 pass a specific index value (between 0 and 31) that should be used.
472 Users can setup udev rules that utilize the index attribute to make fancy
473 device names (e.g. 'mpegX' for MPEG video capture device nodes).
475 After the device was successfully registered, then you can use these fields:
477 - vfl_type: the device type passed to video_register_device.
478 - minor: the assigned device minor number.
479 - num: the device kernel number (i.e. the X in videoX).
480 - index: the device index number (calculated or set explicitly using
481 video_register_device_index).
483 If the registration failed, then you need to call video_device_release()
484 to free the allocated video_device struct, or free your own struct if the
485 video_device was embedded in it. The vdev->release() callback will never
486 be called if the registration failed, nor should you ever attempt to
487 unregister the device if the registration failed.
493 When the video device nodes have to be removed, either during the unload
494 of the driver or because the USB device was disconnected, then you should
497 video_unregister_device(vdev);
499 This will remove the device nodes from sysfs (causing udev to remove them
502 After video_unregister_device() returns no new opens can be done.
504 However, in the case of USB devices some application might still have one
505 of these device nodes open. You should block all new accesses to read,
506 write, poll, etc. except possibly for certain ioctl operations like
509 When the last user of the video device node exits, then the vdev->release()
510 callback is called and you can do the final cleanup there.
513 video_device helper functions
514 -----------------------------
516 There are a few useful helper functions:
518 You can set/get driver private data in the video_device struct using:
520 void *video_get_drvdata(struct video_device *vdev);
521 void video_set_drvdata(struct video_device *vdev, void *data);
523 Note that you can safely call video_set_drvdata() before calling
524 video_register_device().
528 struct video_device *video_devdata(struct file *file);
530 returns the video_device belonging to the file struct.
532 The final helper function combines video_get_drvdata with
535 void *video_drvdata(struct file *file);
537 You can go from a video_device struct to the v4l2_device struct using:
539 struct v4l2_device *v4l2_dev = vdev->v4l2_dev;
541 video buffer helper functions
542 -----------------------------
544 The v4l2 core API provides a standard method for dealing with video
545 buffers. Those methods allow a driver to implement read(), mmap() and
546 overlay() on a consistent way.
548 There are currently methods for using video buffers on devices that
549 supports DMA with scatter/gather method (videobuf-dma-sg), DMA with
550 linear access (videobuf-dma-contig), and vmalloced buffers, mostly
551 used on USB drivers (videobuf-vmalloc).
553 Any driver using videobuf should provide operations (callbacks) for
556 ops->buf_setup - calculates the size of the video buffers and avoid they
557 to waste more than some maximum limit of RAM;
558 ops->buf_prepare - fills the video buffer structs and calls
559 videobuf_iolock() to alloc and prepare mmaped memory;
560 ops->buf_queue - advices the driver that another buffer were
561 requested (by read() or by QBUF);
562 ops->buf_release - frees any buffer that were allocated.
564 In order to use it, the driver need to have a code (generally called at
565 interrupt context) that will properly handle the buffer request lists,
566 announcing that a new buffer were filled.
568 The irq handling code should handle the videobuf task lists, in order
569 to advice videobuf that a new frame were filled, in order to honor to a
570 request. The code is generally like this one:
571 if (list_empty(&dma_q->active))
574 buf = list_entry(dma_q->active.next, struct vbuffer, vb.queue);
576 if (!waitqueue_active(&buf->vb.done))
579 /* Some logic to handle the buf may be needed here */
581 list_del(&buf->vb.queue);
582 do_gettimeofday(&buf->vb.ts);
583 wake_up(&buf->vb.done);
585 Those are the videobuffer functions used on drivers, implemented on
588 - Videobuf init functions
589 videobuf_queue_sg_init()
590 Initializes the videobuf infrastructure. This function should be
591 called before any other videobuf function on drivers that uses DMA
592 Scatter/Gather buffers.
594 videobuf_queue_dma_contig_init
595 Initializes the videobuf infrastructure. This function should be
596 called before any other videobuf function on drivers that need DMA
599 videobuf_queue_vmalloc_init()
600 Initializes the videobuf infrastructure. This function should be
601 called before any other videobuf function on USB (and other drivers)
602 that need a vmalloced type of videobuf.
605 Prepares the videobuf memory for the proper method (read, mmap, overlay).
607 - videobuf_queue_is_busy()
608 Checks if a videobuf is streaming.
610 - videobuf_queue_cancel()
611 Stops video handling.
613 - videobuf_mmap_free()
617 Stops video handling, ends mmap and frees mmap and other buffers.
619 - V4L2 api functions. Those functions correspond to VIDIOC_foo ioctls:
620 videobuf_reqbufs(), videobuf_querybuf(), videobuf_qbuf(),
621 videobuf_dqbuf(), videobuf_streamon(), videobuf_streamoff().
623 - V4L1 api function (corresponds to VIDIOCMBUF ioctl):
625 This function is used to provide backward compatibility with V4L1
628 - Some help functions for read()/poll() operations:
629 videobuf_read_stream()
630 For continuous stream read()
633 videobuf_poll_stream()
634 polling help function
636 The better way to understand it is to take a look at vivi driver. One
637 of the main reasons for vivi is to be a videobuf usage example. the
638 vivi_thread_tick() does the task that the IRQ callback would do on PCI
639 drivers (or the irq callback on USB).