Commit | Line | Data |
---|---|---|
2a1fcdf0 HV |
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 | ||
38 | Structure of a driver | |
39 | --------------------- | |
40 | ||
41 | All drivers have the following structure: | |
42 | ||
43 | 1) A struct for each device instance containing the device state. | |
44 | ||
45 | 2) A way of initializing and commanding sub-devices (if any). | |
46 | ||
47 | 3) Creating V4L2 device nodes (/dev/videoX, /dev/vbiX, /dev/radioX and | |
48 | /dev/vtxX) and keeping track of device-node specific data. | |
49 | ||
44061c05 MCC |
50 | 4) Filehandle-specific structs containing per-filehandle data; |
51 | ||
52 | 5) video buffer handling. | |
2a1fcdf0 HV |
53 | |
54 | This is a rough schematic of how it all relates: | |
55 | ||
56 | device instances | |
57 | | | |
58 | +-sub-device instances | |
59 | | | |
60 | \-V4L2 device nodes | |
61 | | | |
62 | \-filehandle instances | |
63 | ||
64 | ||
65 | Structure of the framework | |
66 | -------------------------- | |
67 | ||
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). | |
73 | ||
74 | ||
75 | struct v4l2_device | |
76 | ------------------ | |
77 | ||
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. | |
81 | ||
82 | You must register the device instance: | |
83 | ||
84 | v4l2_device_register(struct device *dev, struct v4l2_device *v4l2_dev); | |
85 | ||
86 | Registration will initialize the v4l2_device struct and link dev->driver_data | |
3a63e449 HV |
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. | |
2a1fcdf0 | 91 | |
102e7813 HV |
92 | You can use v4l2_device_set_name() to set the name based on a driver name and |
93 | a driver-global atomic_t instance. This will generate names like ivtv0, ivtv1, | |
94 | etc. If the name ends with a digit, then it will insert a dash: cx18-0, | |
95 | cx18-1, etc. This function returns the instance number. | |
96 | ||
a47ddf14 | 97 | The first 'dev' argument is normally the struct device pointer of a pci_dev, |
073d696d | 98 | usb_interface or platform_device. It is rare for dev to be NULL, but it happens |
00575961 HV |
99 | with ISA devices or when one device creates multiple PCI devices, thus making |
100 | it impossible to associate v4l2_dev with a particular parent. | |
a47ddf14 | 101 | |
98ec6339 HV |
102 | You can also supply a notify() callback that can be called by sub-devices to |
103 | notify you of events. Whether you need to set this depends on the sub-device. | |
104 | Any notifications a sub-device supports must be defined in a header in | |
105 | include/media/<subdevice>.h. | |
106 | ||
2a1fcdf0 HV |
107 | You unregister with: |
108 | ||
109 | v4l2_device_unregister(struct v4l2_device *v4l2_dev); | |
110 | ||
111 | Unregistering will also automatically unregister all subdevs from the device. | |
112 | ||
ae6cfaac HV |
113 | If you have a hotpluggable device (e.g. a USB device), then when a disconnect |
114 | happens the parent device becomes invalid. Since v4l2_device has a pointer to | |
115 | that parent device it has to be cleared as well to mark that the parent is | |
116 | gone. To do this call: | |
117 | ||
118 | v4l2_device_disconnect(struct v4l2_device *v4l2_dev); | |
119 | ||
120 | This does *not* unregister the subdevs, so you still need to call the | |
121 | v4l2_device_unregister() function for that. If your driver is not hotpluggable, | |
122 | then there is no need to call v4l2_device_disconnect(). | |
123 | ||
2a1fcdf0 HV |
124 | Sometimes you need to iterate over all devices registered by a specific |
125 | driver. This is usually the case if multiple device drivers use the same | |
126 | hardware. E.g. the ivtvfb driver is a framebuffer driver that uses the ivtv | |
127 | hardware. The same is true for alsa drivers for example. | |
128 | ||
129 | You can iterate over all registered devices as follows: | |
130 | ||
131 | static int callback(struct device *dev, void *p) | |
132 | { | |
133 | struct v4l2_device *v4l2_dev = dev_get_drvdata(dev); | |
134 | ||
135 | /* test if this device was inited */ | |
136 | if (v4l2_dev == NULL) | |
137 | return 0; | |
138 | ... | |
139 | return 0; | |
140 | } | |
141 | ||
142 | int iterate(void *p) | |
143 | { | |
144 | struct device_driver *drv; | |
145 | int err; | |
146 | ||
147 | /* Find driver 'ivtv' on the PCI bus. | |
148 | pci_bus_type is a global. For USB busses use usb_bus_type. */ | |
149 | drv = driver_find("ivtv", &pci_bus_type); | |
150 | /* iterate over all ivtv device instances */ | |
151 | err = driver_for_each_device(drv, NULL, p, callback); | |
152 | put_driver(drv); | |
153 | return err; | |
154 | } | |
155 | ||
156 | Sometimes you need to keep a running counter of the device instance. This is | |
157 | commonly used to map a device instance to an index of a module option array. | |
158 | ||
159 | The recommended approach is as follows: | |
160 | ||
161 | static atomic_t drv_instance = ATOMIC_INIT(0); | |
162 | ||
89aec3e1 | 163 | static int __devinit drv_probe(struct pci_dev *pdev, |
2a1fcdf0 HV |
164 | const struct pci_device_id *pci_id) |
165 | { | |
166 | ... | |
167 | state->instance = atomic_inc_return(&drv_instance) - 1; | |
168 | } | |
169 | ||
170 | ||
171 | struct v4l2_subdev | |
172 | ------------------ | |
173 | ||
174 | Many drivers need to communicate with sub-devices. These devices can do all | |
175 | sort of tasks, but most commonly they handle audio and/or video muxing, | |
176 | encoding or decoding. For webcams common sub-devices are sensors and camera | |
177 | controllers. | |
178 | ||
179 | Usually these are I2C devices, but not necessarily. In order to provide the | |
180 | driver with a consistent interface to these sub-devices the v4l2_subdev struct | |
181 | (v4l2-subdev.h) was created. | |
182 | ||
183 | Each sub-device driver must have a v4l2_subdev struct. This struct can be | |
184 | stand-alone for simple sub-devices or it might be embedded in a larger struct | |
185 | if more state information needs to be stored. Usually there is a low-level | |
186 | device struct (e.g. i2c_client) that contains the device data as setup | |
187 | by the kernel. It is recommended to store that pointer in the private | |
188 | data of v4l2_subdev using v4l2_set_subdevdata(). That makes it easy to go | |
189 | from a v4l2_subdev to the actual low-level bus-specific device data. | |
190 | ||
191 | You also need a way to go from the low-level struct to v4l2_subdev. For the | |
192 | common i2c_client struct the i2c_set_clientdata() call is used to store a | |
193 | v4l2_subdev pointer, for other busses you may have to use other methods. | |
194 | ||
195 | From the bridge driver perspective you load the sub-device module and somehow | |
196 | obtain the v4l2_subdev pointer. For i2c devices this is easy: you call | |
197 | i2c_get_clientdata(). For other busses something similar needs to be done. | |
198 | Helper functions exists for sub-devices on an I2C bus that do most of this | |
199 | tricky work for you. | |
200 | ||
201 | Each v4l2_subdev contains function pointers that sub-device drivers can | |
202 | implement (or leave NULL if it is not applicable). Since sub-devices can do | |
203 | so many different things and you do not want to end up with a huge ops struct | |
204 | of which only a handful of ops are commonly implemented, the function pointers | |
205 | are sorted according to category and each category has its own ops struct. | |
206 | ||
207 | The top-level ops struct contains pointers to the category ops structs, which | |
208 | may be NULL if the subdev driver does not support anything from that category. | |
209 | ||
210 | It looks like this: | |
211 | ||
212 | struct v4l2_subdev_core_ops { | |
aecde8b5 | 213 | int (*g_chip_ident)(struct v4l2_subdev *sd, struct v4l2_dbg_chip_ident *chip); |
2a1fcdf0 HV |
214 | int (*log_status)(struct v4l2_subdev *sd); |
215 | int (*init)(struct v4l2_subdev *sd, u32 val); | |
216 | ... | |
217 | }; | |
218 | ||
219 | struct v4l2_subdev_tuner_ops { | |
220 | ... | |
221 | }; | |
222 | ||
223 | struct v4l2_subdev_audio_ops { | |
224 | ... | |
225 | }; | |
226 | ||
227 | struct v4l2_subdev_video_ops { | |
228 | ... | |
229 | }; | |
230 | ||
231 | struct v4l2_subdev_ops { | |
232 | const struct v4l2_subdev_core_ops *core; | |
233 | const struct v4l2_subdev_tuner_ops *tuner; | |
234 | const struct v4l2_subdev_audio_ops *audio; | |
235 | const struct v4l2_subdev_video_ops *video; | |
236 | }; | |
237 | ||
238 | The core ops are common to all subdevs, the other categories are implemented | |
239 | depending on the sub-device. E.g. a video device is unlikely to support the | |
240 | audio ops and vice versa. | |
241 | ||
242 | This setup limits the number of function pointers while still making it easy | |
243 | to add new ops and categories. | |
244 | ||
245 | A sub-device driver initializes the v4l2_subdev struct using: | |
246 | ||
89aec3e1 | 247 | v4l2_subdev_init(sd, &ops); |
2a1fcdf0 HV |
248 | |
249 | Afterwards you need to initialize subdev->name with a unique name and set the | |
250 | module owner. This is done for you if you use the i2c helper functions. | |
251 | ||
252 | A device (bridge) driver needs to register the v4l2_subdev with the | |
253 | v4l2_device: | |
254 | ||
89aec3e1 | 255 | int err = v4l2_device_register_subdev(v4l2_dev, sd); |
2a1fcdf0 HV |
256 | |
257 | This can fail if the subdev module disappeared before it could be registered. | |
258 | After this function was called successfully the subdev->dev field points to | |
259 | the v4l2_device. | |
260 | ||
261 | You can unregister a sub-device using: | |
262 | ||
89aec3e1 | 263 | v4l2_device_unregister_subdev(sd); |
2a1fcdf0 | 264 | |
89aec3e1 | 265 | Afterwards the subdev module can be unloaded and sd->dev == NULL. |
2a1fcdf0 HV |
266 | |
267 | You can call an ops function either directly: | |
268 | ||
89aec3e1 | 269 | err = sd->ops->core->g_chip_ident(sd, &chip); |
2a1fcdf0 HV |
270 | |
271 | but it is better and easier to use this macro: | |
272 | ||
89aec3e1 | 273 | err = v4l2_subdev_call(sd, core, g_chip_ident, &chip); |
2a1fcdf0 HV |
274 | |
275 | The macro will to the right NULL pointer checks and returns -ENODEV if subdev | |
276 | is NULL, -ENOIOCTLCMD if either subdev->core or subdev->core->g_chip_ident is | |
277 | NULL, or the actual result of the subdev->ops->core->g_chip_ident ops. | |
278 | ||
279 | It is also possible to call all or a subset of the sub-devices: | |
280 | ||
89aec3e1 | 281 | v4l2_device_call_all(v4l2_dev, 0, core, g_chip_ident, &chip); |
2a1fcdf0 HV |
282 | |
283 | Any subdev that does not support this ops is skipped and error results are | |
284 | ignored. If you want to check for errors use this: | |
285 | ||
89aec3e1 | 286 | err = v4l2_device_call_until_err(v4l2_dev, 0, core, g_chip_ident, &chip); |
2a1fcdf0 HV |
287 | |
288 | Any error except -ENOIOCTLCMD will exit the loop with that error. If no | |
289 | errors (except -ENOIOCTLCMD) occured, then 0 is returned. | |
290 | ||
291 | The second argument to both calls is a group ID. If 0, then all subdevs are | |
292 | called. If non-zero, then only those whose group ID match that value will | |
b0167600 | 293 | be called. Before a bridge driver registers a subdev it can set sd->grp_id |
2a1fcdf0 HV |
294 | to whatever value it wants (it's 0 by default). This value is owned by the |
295 | bridge driver and the sub-device driver will never modify or use it. | |
296 | ||
297 | The group ID gives the bridge driver more control how callbacks are called. | |
298 | For example, there may be multiple audio chips on a board, each capable of | |
299 | changing the volume. But usually only one will actually be used when the | |
300 | user want to change the volume. You can set the group ID for that subdev to | |
301 | e.g. AUDIO_CONTROLLER and specify that as the group ID value when calling | |
302 | v4l2_device_call_all(). That ensures that it will only go to the subdev | |
303 | that needs it. | |
304 | ||
98ec6339 HV |
305 | If the sub-device needs to notify its v4l2_device parent of an event, then |
306 | it can call v4l2_subdev_notify(sd, notification, arg). This macro checks | |
307 | whether there is a notify() callback defined and returns -ENODEV if not. | |
308 | Otherwise the result of the notify() call is returned. | |
309 | ||
2a1fcdf0 HV |
310 | The advantage of using v4l2_subdev is that it is a generic struct and does |
311 | not contain any knowledge about the underlying hardware. So a driver might | |
312 | contain several subdevs that use an I2C bus, but also a subdev that is | |
313 | controlled through GPIO pins. This distinction is only relevant when setting | |
314 | up the device, but once the subdev is registered it is completely transparent. | |
315 | ||
316 | ||
317 | I2C sub-device drivers | |
318 | ---------------------- | |
319 | ||
320 | Since these drivers are so common, special helper functions are available to | |
321 | ease the use of these drivers (v4l2-common.h). | |
322 | ||
323 | The recommended method of adding v4l2_subdev support to an I2C driver is to | |
324 | embed the v4l2_subdev struct into the state struct that is created for each | |
325 | I2C device instance. Very simple devices have no state struct and in that case | |
326 | you can just create a v4l2_subdev directly. | |
327 | ||
328 | A typical state struct would look like this (where 'chipname' is replaced by | |
329 | the name of the chip): | |
330 | ||
331 | struct chipname_state { | |
332 | struct v4l2_subdev sd; | |
333 | ... /* additional state fields */ | |
334 | }; | |
335 | ||
336 | Initialize the v4l2_subdev struct as follows: | |
337 | ||
338 | v4l2_i2c_subdev_init(&state->sd, client, subdev_ops); | |
339 | ||
340 | This function will fill in all the fields of v4l2_subdev and ensure that the | |
341 | v4l2_subdev and i2c_client both point to one another. | |
342 | ||
343 | You should also add a helper inline function to go from a v4l2_subdev pointer | |
344 | to a chipname_state struct: | |
345 | ||
346 | static inline struct chipname_state *to_state(struct v4l2_subdev *sd) | |
347 | { | |
348 | return container_of(sd, struct chipname_state, sd); | |
349 | } | |
350 | ||
351 | Use this to go from the v4l2_subdev struct to the i2c_client struct: | |
352 | ||
353 | struct i2c_client *client = v4l2_get_subdevdata(sd); | |
354 | ||
355 | And this to go from an i2c_client to a v4l2_subdev struct: | |
356 | ||
357 | struct v4l2_subdev *sd = i2c_get_clientdata(client); | |
358 | ||
2a1fcdf0 HV |
359 | Make sure to call v4l2_device_unregister_subdev(sd) when the remove() callback |
360 | is called. This will unregister the sub-device from the bridge driver. It is | |
361 | safe to call this even if the sub-device was never registered. | |
362 | ||
f5360bdc HV |
363 | You need to do this because when the bridge driver destroys the i2c adapter |
364 | the remove() callbacks are called of the i2c devices on that adapter. | |
365 | After that the corresponding v4l2_subdev structures are invalid, so they | |
366 | have to be unregistered first. Calling v4l2_device_unregister_subdev(sd) | |
367 | from the remove() callback ensures that this is always done correctly. | |
368 | ||
2a1fcdf0 HV |
369 | |
370 | The bridge driver also has some helper functions it can use: | |
371 | ||
e6574f2f | 372 | struct v4l2_subdev *sd = v4l2_i2c_new_subdev(v4l2_dev, adapter, |
53dacb15 | 373 | "module_foo", "chipid", 0x36, NULL); |
2a1fcdf0 HV |
374 | |
375 | This loads the given module (can be NULL if no module needs to be loaded) and | |
376 | calls i2c_new_device() with the given i2c_adapter and chip/address arguments. | |
e6574f2f | 377 | If all goes well, then it registers the subdev with the v4l2_device. |
2a1fcdf0 | 378 | |
53dacb15 HV |
379 | You can also use the last argument of v4l2_i2c_new_subdev() to pass an array |
380 | of possible I2C addresses that it should probe. These probe addresses are | |
381 | only used if the previous argument is 0. A non-zero argument means that you | |
382 | know the exact i2c address so in that case no probing will take place. | |
2a1fcdf0 HV |
383 | |
384 | Both functions return NULL if something went wrong. | |
385 | ||
53dacb15 | 386 | Note that the chipid you pass to v4l2_i2c_new_subdev() is usually |
2c792523 HV |
387 | the same as the module name. It allows you to specify a chip variant, e.g. |
388 | "saa7114" or "saa7115". In general though the i2c driver autodetects this. | |
389 | The use of chipid is something that needs to be looked at more closely at a | |
390 | later date. It differs between i2c drivers and as such can be confusing. | |
391 | To see which chip variants are supported you can look in the i2c driver code | |
392 | for the i2c_device_id table. This lists all the possibilities. | |
393 | ||
2c0b19ac HV |
394 | There are two more helper functions: |
395 | ||
396 | v4l2_i2c_new_subdev_cfg: this function adds new irq and platform_data | |
397 | arguments and has both 'addr' and 'probed_addrs' arguments: if addr is not | |
398 | 0 then that will be used (non-probing variant), otherwise the probed_addrs | |
399 | are probed. | |
400 | ||
401 | For example: this will probe for address 0x10: | |
402 | ||
403 | struct v4l2_subdev *sd = v4l2_i2c_new_subdev_cfg(v4l2_dev, adapter, | |
404 | "module_foo", "chipid", 0, NULL, 0, I2C_ADDRS(0x10)); | |
405 | ||
406 | v4l2_i2c_new_subdev_board uses an i2c_board_info struct which is passed | |
407 | to the i2c driver and replaces the irq, platform_data and addr arguments. | |
408 | ||
409 | If the subdev supports the s_config core ops, then that op is called with | |
410 | the irq and platform_data arguments after the subdev was setup. The older | |
411 | v4l2_i2c_new_(probed_)subdev functions will call s_config as well, but with | |
412 | irq set to 0 and platform_data set to NULL. | |
413 | ||
2a1fcdf0 HV |
414 | struct video_device |
415 | ------------------- | |
416 | ||
a47ddf14 HV |
417 | The actual device nodes in the /dev directory are created using the |
418 | video_device struct (v4l2-dev.h). This struct can either be allocated | |
419 | dynamically or embedded in a larger struct. | |
420 | ||
421 | To allocate it dynamically use: | |
422 | ||
423 | struct video_device *vdev = video_device_alloc(); | |
424 | ||
425 | if (vdev == NULL) | |
426 | return -ENOMEM; | |
427 | ||
428 | vdev->release = video_device_release; | |
429 | ||
430 | If you embed it in a larger struct, then you must set the release() | |
431 | callback to your own function: | |
432 | ||
433 | struct video_device *vdev = &my_vdev->vdev; | |
434 | ||
435 | vdev->release = my_vdev_release; | |
436 | ||
437 | The release callback must be set and it is called when the last user | |
438 | of the video device exits. | |
439 | ||
440 | The default video_device_release() callback just calls kfree to free the | |
441 | allocated memory. | |
442 | ||
443 | You should also set these fields: | |
444 | ||
dfa9a5ae | 445 | - v4l2_dev: set to the v4l2_device parent device. |
a47ddf14 | 446 | - name: set to something descriptive and unique. |
c7dd09da | 447 | - fops: set to the v4l2_file_operations struct. |
a47ddf14 HV |
448 | - ioctl_ops: if you use the v4l2_ioctl_ops to simplify ioctl maintenance |
449 | (highly recommended to use this and it might become compulsory in the | |
450 | future!), then set this to your v4l2_ioctl_ops struct. | |
00575961 HV |
451 | - parent: you only set this if v4l2_device was registered with NULL as |
452 | the parent device struct. This only happens in cases where one hardware | |
453 | device has multiple PCI devices that all share the same v4l2_device core. | |
454 | ||
455 | The cx88 driver is an example of this: one core v4l2_device struct, but | |
456 | it is used by both an raw video PCI device (cx8800) and a MPEG PCI device | |
457 | (cx8802). Since the v4l2_device cannot be associated with a particular | |
458 | PCI device it is setup without a parent device. But when the struct | |
459 | video_device is setup you do know which parent PCI device to use. | |
a47ddf14 | 460 | |
c7dd09da HV |
461 | If you use v4l2_ioctl_ops, then you should set either .unlocked_ioctl or |
462 | .ioctl to video_ioctl2 in your v4l2_file_operations struct. | |
463 | ||
464 | The v4l2_file_operations struct is a subset of file_operations. The main | |
465 | difference is that the inode argument is omitted since it is never used. | |
a47ddf14 HV |
466 | |
467 | ||
468 | video_device registration | |
469 | ------------------------- | |
470 | ||
471 | Next you register the video device: this will create the character device | |
472 | for you. | |
473 | ||
474 | err = video_register_device(vdev, VFL_TYPE_GRABBER, -1); | |
475 | if (err) { | |
50a2a8b3 | 476 | video_device_release(vdev); /* or kfree(my_vdev); */ |
a47ddf14 HV |
477 | return err; |
478 | } | |
479 | ||
480 | Which device is registered depends on the type argument. The following | |
481 | types exist: | |
482 | ||
483 | VFL_TYPE_GRABBER: videoX for video input/output devices | |
484 | VFL_TYPE_VBI: vbiX for vertical blank data (i.e. closed captions, teletext) | |
485 | VFL_TYPE_RADIO: radioX for radio tuners | |
486 | VFL_TYPE_VTX: vtxX for teletext devices (deprecated, don't use) | |
487 | ||
488 | The last argument gives you a certain amount of control over the device | |
6b5270d2 HV |
489 | device node number used (i.e. the X in videoX). Normally you will pass -1 |
490 | to let the v4l2 framework pick the first free number. But sometimes users | |
491 | want to select a specific node number. It is common that drivers allow | |
492 | the user to select a specific device node number through a driver module | |
493 | option. That number is then passed to this function and video_register_device | |
494 | will attempt to select that device node number. If that number was already | |
495 | in use, then the next free device node number will be selected and it | |
496 | will send a warning to the kernel log. | |
497 | ||
498 | Another use-case is if a driver creates many devices. In that case it can | |
499 | be useful to place different video devices in separate ranges. For example, | |
500 | video capture devices start at 0, video output devices start at 16. | |
22e22125 HV |
501 | So you can use the last argument to specify a minimum device node number |
502 | and the v4l2 framework will try to pick the first free number that is equal | |
a47ddf14 HV |
503 | or higher to what you passed. If that fails, then it will just pick the |
504 | first free number. | |
505 | ||
6b5270d2 HV |
506 | Since in this case you do not care about a warning about not being able |
507 | to select the specified device node number, you can call the function | |
508 | video_register_device_no_warn() instead. | |
509 | ||
a47ddf14 HV |
510 | Whenever a device node is created some attributes are also created for you. |
511 | If you look in /sys/class/video4linux you see the devices. Go into e.g. | |
512 | video0 and you will see 'name' and 'index' attributes. The 'name' attribute | |
7ae0cd9b | 513 | is the 'name' field of the video_device struct. |
a47ddf14 | 514 | |
7ae0cd9b HV |
515 | The 'index' attribute is the index of the device node: for each call to |
516 | video_register_device() the index is just increased by 1. The first video | |
517 | device node you register always starts with index 0. | |
a47ddf14 HV |
518 | |
519 | Users can setup udev rules that utilize the index attribute to make fancy | |
520 | device names (e.g. 'mpegX' for MPEG video capture device nodes). | |
521 | ||
522 | After the device was successfully registered, then you can use these fields: | |
523 | ||
524 | - vfl_type: the device type passed to video_register_device. | |
525 | - minor: the assigned device minor number. | |
22e22125 | 526 | - num: the device node number (i.e. the X in videoX). |
7ae0cd9b | 527 | - index: the device index number. |
a47ddf14 HV |
528 | |
529 | If the registration failed, then you need to call video_device_release() | |
530 | to free the allocated video_device struct, or free your own struct if the | |
531 | video_device was embedded in it. The vdev->release() callback will never | |
532 | be called if the registration failed, nor should you ever attempt to | |
533 | unregister the device if the registration failed. | |
534 | ||
535 | ||
536 | video_device cleanup | |
537 | -------------------- | |
538 | ||
539 | When the video device nodes have to be removed, either during the unload | |
540 | of the driver or because the USB device was disconnected, then you should | |
541 | unregister them: | |
542 | ||
543 | video_unregister_device(vdev); | |
544 | ||
545 | This will remove the device nodes from sysfs (causing udev to remove them | |
546 | from /dev). | |
547 | ||
548 | After video_unregister_device() returns no new opens can be done. | |
549 | ||
550 | However, in the case of USB devices some application might still have one | |
551 | of these device nodes open. You should block all new accesses to read, | |
552 | write, poll, etc. except possibly for certain ioctl operations like | |
553 | queueing buffers. | |
554 | ||
555 | When the last user of the video device node exits, then the vdev->release() | |
556 | callback is called and you can do the final cleanup there. | |
557 | ||
558 | ||
559 | video_device helper functions | |
560 | ----------------------------- | |
561 | ||
562 | There are a few useful helper functions: | |
563 | ||
eac8ea53 LP |
564 | - file/video_device private data |
565 | ||
a47ddf14 HV |
566 | You can set/get driver private data in the video_device struct using: |
567 | ||
89aec3e1 HV |
568 | void *video_get_drvdata(struct video_device *vdev); |
569 | void video_set_drvdata(struct video_device *vdev, void *data); | |
a47ddf14 HV |
570 | |
571 | Note that you can safely call video_set_drvdata() before calling | |
572 | video_register_device(). | |
573 | ||
574 | And this function: | |
575 | ||
576 | struct video_device *video_devdata(struct file *file); | |
577 | ||
578 | returns the video_device belonging to the file struct. | |
579 | ||
eac8ea53 | 580 | The video_drvdata function combines video_get_drvdata with video_devdata: |
a47ddf14 HV |
581 | |
582 | void *video_drvdata(struct file *file); | |
583 | ||
584 | You can go from a video_device struct to the v4l2_device struct using: | |
585 | ||
dfa9a5ae | 586 | struct v4l2_device *v4l2_dev = vdev->v4l2_dev; |
44061c05 | 587 | |
eac8ea53 LP |
588 | - Device node name |
589 | ||
590 | The video_device node kernel name can be retrieved using | |
591 | ||
592 | const char *video_device_node_name(struct video_device *vdev); | |
593 | ||
594 | The name is used as a hint by userspace tools such as udev. The function | |
595 | should be used where possible instead of accessing the video_device::num and | |
596 | video_device::minor fields. | |
597 | ||
598 | ||
44061c05 MCC |
599 | video buffer helper functions |
600 | ----------------------------- | |
601 | ||
602 | The v4l2 core API provides a standard method for dealing with video | |
603 | buffers. Those methods allow a driver to implement read(), mmap() and | |
604 | overlay() on a consistent way. | |
605 | ||
606 | There are currently methods for using video buffers on devices that | |
607 | supports DMA with scatter/gather method (videobuf-dma-sg), DMA with | |
608 | linear access (videobuf-dma-contig), and vmalloced buffers, mostly | |
609 | used on USB drivers (videobuf-vmalloc). | |
610 | ||
611 | Any driver using videobuf should provide operations (callbacks) for | |
612 | four handlers: | |
613 | ||
614 | ops->buf_setup - calculates the size of the video buffers and avoid they | |
615 | to waste more than some maximum limit of RAM; | |
616 | ops->buf_prepare - fills the video buffer structs and calls | |
617 | videobuf_iolock() to alloc and prepare mmaped memory; | |
618 | ops->buf_queue - advices the driver that another buffer were | |
619 | requested (by read() or by QBUF); | |
620 | ops->buf_release - frees any buffer that were allocated. | |
621 | ||
622 | In order to use it, the driver need to have a code (generally called at | |
623 | interrupt context) that will properly handle the buffer request lists, | |
624 | announcing that a new buffer were filled. | |
625 | ||
626 | The irq handling code should handle the videobuf task lists, in order | |
627 | to advice videobuf that a new frame were filled, in order to honor to a | |
628 | request. The code is generally like this one: | |
a7a1c0e6 | 629 | if (list_empty(&dma_q->active)) |
44061c05 MCC |
630 | return; |
631 | ||
a7a1c0e6 | 632 | buf = list_entry(dma_q->active.next, struct vbuffer, vb.queue); |
44061c05 | 633 | |
a7a1c0e6 | 634 | if (!waitqueue_active(&buf->vb.done)) |
44061c05 MCC |
635 | return; |
636 | ||
637 | /* Some logic to handle the buf may be needed here */ | |
638 | ||
a7a1c0e6 MCC |
639 | list_del(&buf->vb.queue); |
640 | do_gettimeofday(&buf->vb.ts); | |
641 | wake_up(&buf->vb.done); | |
44061c05 MCC |
642 | |
643 | Those are the videobuffer functions used on drivers, implemented on | |
644 | videobuf-core: | |
645 | ||
a7a1c0e6 MCC |
646 | - Videobuf init functions |
647 | videobuf_queue_sg_init() | |
648 | Initializes the videobuf infrastructure. This function should be | |
649 | called before any other videobuf function on drivers that uses DMA | |
650 | Scatter/Gather buffers. | |
651 | ||
652 | videobuf_queue_dma_contig_init | |
653 | Initializes the videobuf infrastructure. This function should be | |
654 | called before any other videobuf function on drivers that need DMA | |
655 | contiguous buffers. | |
656 | ||
657 | videobuf_queue_vmalloc_init() | |
658 | Initializes the videobuf infrastructure. This function should be | |
659 | called before any other videobuf function on USB (and other drivers) | |
660 | that need a vmalloced type of videobuf. | |
44061c05 MCC |
661 | |
662 | - videobuf_iolock() | |
663 | Prepares the videobuf memory for the proper method (read, mmap, overlay). | |
664 | ||
665 | - videobuf_queue_is_busy() | |
666 | Checks if a videobuf is streaming. | |
667 | ||
668 | - videobuf_queue_cancel() | |
669 | Stops video handling. | |
670 | ||
671 | - videobuf_mmap_free() | |
672 | frees mmap buffers. | |
673 | ||
674 | - videobuf_stop() | |
675 | Stops video handling, ends mmap and frees mmap and other buffers. | |
676 | ||
677 | - V4L2 api functions. Those functions correspond to VIDIOC_foo ioctls: | |
678 | videobuf_reqbufs(), videobuf_querybuf(), videobuf_qbuf(), | |
679 | videobuf_dqbuf(), videobuf_streamon(), videobuf_streamoff(). | |
680 | ||
681 | - V4L1 api function (corresponds to VIDIOCMBUF ioctl): | |
682 | videobuf_cgmbuf() | |
683 | This function is used to provide backward compatibility with V4L1 | |
684 | API. | |
685 | ||
686 | - Some help functions for read()/poll() operations: | |
687 | videobuf_read_stream() | |
688 | For continuous stream read() | |
689 | videobuf_read_one() | |
690 | For snapshot read() | |
691 | videobuf_poll_stream() | |
692 | polling help function | |
693 | ||
694 | The better way to understand it is to take a look at vivi driver. One | |
695 | of the main reasons for vivi is to be a videobuf usage example. the | |
696 | vivi_thread_tick() does the task that the IRQ callback would do on PCI | |
697 | drivers (or the irq callback on USB). |