Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 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 | |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 50 | 4) Filehandle-specific structs containing per-filehandle data; |
| 51 | |
| 52 | 5) video buffer handling. |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 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 |
Hans Verkuil | 3a63e449 | 2009-02-14 11:54:23 -0300 | [diff] [blame] | 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. |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 91 | |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 92 | The first 'dev' argument is normally the struct device pointer of a pci_dev, |
Hans Verkuil | 3a63e449 | 2009-02-14 11:54:23 -0300 | [diff] [blame] | 93 | usb_device or platform_device. It is rare for dev to be NULL, but it happens |
Hans Verkuil | 0057596 | 2009-03-13 10:03:04 -0300 | [diff] [blame] | 94 | with ISA devices or when one device creates multiple PCI devices, thus making |
| 95 | it impossible to associate v4l2_dev with a particular parent. |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 96 | |
Hans Verkuil | 98ec633 | 2009-03-08 17:02:10 -0300 | [diff] [blame] | 97 | You can also supply a notify() callback that can be called by sub-devices to |
| 98 | notify you of events. Whether you need to set this depends on the sub-device. |
| 99 | Any notifications a sub-device supports must be defined in a header in |
| 100 | include/media/<subdevice>.h. |
| 101 | |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 102 | You unregister with: |
| 103 | |
| 104 | v4l2_device_unregister(struct v4l2_device *v4l2_dev); |
| 105 | |
| 106 | Unregistering will also automatically unregister all subdevs from the device. |
| 107 | |
| 108 | Sometimes you need to iterate over all devices registered by a specific |
| 109 | driver. This is usually the case if multiple device drivers use the same |
| 110 | hardware. E.g. the ivtvfb driver is a framebuffer driver that uses the ivtv |
| 111 | hardware. The same is true for alsa drivers for example. |
| 112 | |
| 113 | You can iterate over all registered devices as follows: |
| 114 | |
| 115 | static int callback(struct device *dev, void *p) |
| 116 | { |
| 117 | struct v4l2_device *v4l2_dev = dev_get_drvdata(dev); |
| 118 | |
| 119 | /* test if this device was inited */ |
| 120 | if (v4l2_dev == NULL) |
| 121 | return 0; |
| 122 | ... |
| 123 | return 0; |
| 124 | } |
| 125 | |
| 126 | int iterate(void *p) |
| 127 | { |
| 128 | struct device_driver *drv; |
| 129 | int err; |
| 130 | |
| 131 | /* Find driver 'ivtv' on the PCI bus. |
| 132 | pci_bus_type is a global. For USB busses use usb_bus_type. */ |
| 133 | drv = driver_find("ivtv", &pci_bus_type); |
| 134 | /* iterate over all ivtv device instances */ |
| 135 | err = driver_for_each_device(drv, NULL, p, callback); |
| 136 | put_driver(drv); |
| 137 | return err; |
| 138 | } |
| 139 | |
| 140 | Sometimes you need to keep a running counter of the device instance. This is |
| 141 | commonly used to map a device instance to an index of a module option array. |
| 142 | |
| 143 | The recommended approach is as follows: |
| 144 | |
| 145 | static atomic_t drv_instance = ATOMIC_INIT(0); |
| 146 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 147 | static int __devinit drv_probe(struct pci_dev *pdev, |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 148 | const struct pci_device_id *pci_id) |
| 149 | { |
| 150 | ... |
| 151 | state->instance = atomic_inc_return(&drv_instance) - 1; |
| 152 | } |
| 153 | |
| 154 | |
| 155 | struct v4l2_subdev |
| 156 | ------------------ |
| 157 | |
| 158 | Many drivers need to communicate with sub-devices. These devices can do all |
| 159 | sort of tasks, but most commonly they handle audio and/or video muxing, |
| 160 | encoding or decoding. For webcams common sub-devices are sensors and camera |
| 161 | controllers. |
| 162 | |
| 163 | Usually these are I2C devices, but not necessarily. In order to provide the |
| 164 | driver with a consistent interface to these sub-devices the v4l2_subdev struct |
| 165 | (v4l2-subdev.h) was created. |
| 166 | |
| 167 | Each sub-device driver must have a v4l2_subdev struct. This struct can be |
| 168 | stand-alone for simple sub-devices or it might be embedded in a larger struct |
| 169 | if more state information needs to be stored. Usually there is a low-level |
| 170 | device struct (e.g. i2c_client) that contains the device data as setup |
| 171 | by the kernel. It is recommended to store that pointer in the private |
| 172 | data of v4l2_subdev using v4l2_set_subdevdata(). That makes it easy to go |
| 173 | from a v4l2_subdev to the actual low-level bus-specific device data. |
| 174 | |
| 175 | You also need a way to go from the low-level struct to v4l2_subdev. For the |
| 176 | common i2c_client struct the i2c_set_clientdata() call is used to store a |
| 177 | v4l2_subdev pointer, for other busses you may have to use other methods. |
| 178 | |
| 179 | From the bridge driver perspective you load the sub-device module and somehow |
| 180 | obtain the v4l2_subdev pointer. For i2c devices this is easy: you call |
| 181 | i2c_get_clientdata(). For other busses something similar needs to be done. |
| 182 | Helper functions exists for sub-devices on an I2C bus that do most of this |
| 183 | tricky work for you. |
| 184 | |
| 185 | Each v4l2_subdev contains function pointers that sub-device drivers can |
| 186 | implement (or leave NULL if it is not applicable). Since sub-devices can do |
| 187 | so many different things and you do not want to end up with a huge ops struct |
| 188 | of which only a handful of ops are commonly implemented, the function pointers |
| 189 | are sorted according to category and each category has its own ops struct. |
| 190 | |
| 191 | The top-level ops struct contains pointers to the category ops structs, which |
| 192 | may be NULL if the subdev driver does not support anything from that category. |
| 193 | |
| 194 | It looks like this: |
| 195 | |
| 196 | struct v4l2_subdev_core_ops { |
Hans Verkuil | aecde8b5 | 2008-12-30 07:14:19 -0300 | [diff] [blame] | 197 | int (*g_chip_ident)(struct v4l2_subdev *sd, struct v4l2_dbg_chip_ident *chip); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 198 | int (*log_status)(struct v4l2_subdev *sd); |
| 199 | int (*init)(struct v4l2_subdev *sd, u32 val); |
| 200 | ... |
| 201 | }; |
| 202 | |
| 203 | struct v4l2_subdev_tuner_ops { |
| 204 | ... |
| 205 | }; |
| 206 | |
| 207 | struct v4l2_subdev_audio_ops { |
| 208 | ... |
| 209 | }; |
| 210 | |
| 211 | struct v4l2_subdev_video_ops { |
| 212 | ... |
| 213 | }; |
| 214 | |
| 215 | struct v4l2_subdev_ops { |
| 216 | const struct v4l2_subdev_core_ops *core; |
| 217 | const struct v4l2_subdev_tuner_ops *tuner; |
| 218 | const struct v4l2_subdev_audio_ops *audio; |
| 219 | const struct v4l2_subdev_video_ops *video; |
| 220 | }; |
| 221 | |
| 222 | The core ops are common to all subdevs, the other categories are implemented |
| 223 | depending on the sub-device. E.g. a video device is unlikely to support the |
| 224 | audio ops and vice versa. |
| 225 | |
| 226 | This setup limits the number of function pointers while still making it easy |
| 227 | to add new ops and categories. |
| 228 | |
| 229 | A sub-device driver initializes the v4l2_subdev struct using: |
| 230 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 231 | v4l2_subdev_init(sd, &ops); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 232 | |
| 233 | Afterwards you need to initialize subdev->name with a unique name and set the |
| 234 | module owner. This is done for you if you use the i2c helper functions. |
| 235 | |
| 236 | A device (bridge) driver needs to register the v4l2_subdev with the |
| 237 | v4l2_device: |
| 238 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 239 | int err = v4l2_device_register_subdev(v4l2_dev, sd); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 240 | |
| 241 | This can fail if the subdev module disappeared before it could be registered. |
| 242 | After this function was called successfully the subdev->dev field points to |
| 243 | the v4l2_device. |
| 244 | |
| 245 | You can unregister a sub-device using: |
| 246 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 247 | v4l2_device_unregister_subdev(sd); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 248 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 249 | Afterwards the subdev module can be unloaded and sd->dev == NULL. |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 250 | |
| 251 | You can call an ops function either directly: |
| 252 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 253 | err = sd->ops->core->g_chip_ident(sd, &chip); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 254 | |
| 255 | but it is better and easier to use this macro: |
| 256 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 257 | err = v4l2_subdev_call(sd, core, g_chip_ident, &chip); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 258 | |
| 259 | The macro will to the right NULL pointer checks and returns -ENODEV if subdev |
| 260 | is NULL, -ENOIOCTLCMD if either subdev->core or subdev->core->g_chip_ident is |
| 261 | NULL, or the actual result of the subdev->ops->core->g_chip_ident ops. |
| 262 | |
| 263 | It is also possible to call all or a subset of the sub-devices: |
| 264 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 265 | v4l2_device_call_all(v4l2_dev, 0, core, g_chip_ident, &chip); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 266 | |
| 267 | Any subdev that does not support this ops is skipped and error results are |
| 268 | ignored. If you want to check for errors use this: |
| 269 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 270 | err = v4l2_device_call_until_err(v4l2_dev, 0, core, g_chip_ident, &chip); |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 271 | |
| 272 | Any error except -ENOIOCTLCMD will exit the loop with that error. If no |
| 273 | errors (except -ENOIOCTLCMD) occured, then 0 is returned. |
| 274 | |
| 275 | The second argument to both calls is a group ID. If 0, then all subdevs are |
| 276 | called. If non-zero, then only those whose group ID match that value will |
Hans Verkuil | b016760 | 2009-02-14 12:00:53 -0300 | [diff] [blame] | 277 | be called. Before a bridge driver registers a subdev it can set sd->grp_id |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 278 | to whatever value it wants (it's 0 by default). This value is owned by the |
| 279 | bridge driver and the sub-device driver will never modify or use it. |
| 280 | |
| 281 | The group ID gives the bridge driver more control how callbacks are called. |
| 282 | For example, there may be multiple audio chips on a board, each capable of |
| 283 | changing the volume. But usually only one will actually be used when the |
| 284 | user want to change the volume. You can set the group ID for that subdev to |
| 285 | e.g. AUDIO_CONTROLLER and specify that as the group ID value when calling |
| 286 | v4l2_device_call_all(). That ensures that it will only go to the subdev |
| 287 | that needs it. |
| 288 | |
Hans Verkuil | 98ec633 | 2009-03-08 17:02:10 -0300 | [diff] [blame] | 289 | If the sub-device needs to notify its v4l2_device parent of an event, then |
| 290 | it can call v4l2_subdev_notify(sd, notification, arg). This macro checks |
| 291 | whether there is a notify() callback defined and returns -ENODEV if not. |
| 292 | Otherwise the result of the notify() call is returned. |
| 293 | |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 294 | The advantage of using v4l2_subdev is that it is a generic struct and does |
| 295 | not contain any knowledge about the underlying hardware. So a driver might |
| 296 | contain several subdevs that use an I2C bus, but also a subdev that is |
| 297 | controlled through GPIO pins. This distinction is only relevant when setting |
| 298 | up the device, but once the subdev is registered it is completely transparent. |
| 299 | |
| 300 | |
| 301 | I2C sub-device drivers |
| 302 | ---------------------- |
| 303 | |
| 304 | Since these drivers are so common, special helper functions are available to |
| 305 | ease the use of these drivers (v4l2-common.h). |
| 306 | |
| 307 | The recommended method of adding v4l2_subdev support to an I2C driver is to |
| 308 | embed the v4l2_subdev struct into the state struct that is created for each |
| 309 | I2C device instance. Very simple devices have no state struct and in that case |
| 310 | you can just create a v4l2_subdev directly. |
| 311 | |
| 312 | A typical state struct would look like this (where 'chipname' is replaced by |
| 313 | the name of the chip): |
| 314 | |
| 315 | struct chipname_state { |
| 316 | struct v4l2_subdev sd; |
| 317 | ... /* additional state fields */ |
| 318 | }; |
| 319 | |
| 320 | Initialize the v4l2_subdev struct as follows: |
| 321 | |
| 322 | v4l2_i2c_subdev_init(&state->sd, client, subdev_ops); |
| 323 | |
| 324 | This function will fill in all the fields of v4l2_subdev and ensure that the |
| 325 | v4l2_subdev and i2c_client both point to one another. |
| 326 | |
| 327 | You should also add a helper inline function to go from a v4l2_subdev pointer |
| 328 | to a chipname_state struct: |
| 329 | |
| 330 | static inline struct chipname_state *to_state(struct v4l2_subdev *sd) |
| 331 | { |
| 332 | return container_of(sd, struct chipname_state, sd); |
| 333 | } |
| 334 | |
| 335 | Use this to go from the v4l2_subdev struct to the i2c_client struct: |
| 336 | |
| 337 | struct i2c_client *client = v4l2_get_subdevdata(sd); |
| 338 | |
| 339 | And this to go from an i2c_client to a v4l2_subdev struct: |
| 340 | |
| 341 | struct v4l2_subdev *sd = i2c_get_clientdata(client); |
| 342 | |
| 343 | Finally you need to make a command function to make driver->command() |
| 344 | call the right subdev_ops functions: |
| 345 | |
| 346 | static int subdev_command(struct i2c_client *client, unsigned cmd, void *arg) |
| 347 | { |
| 348 | return v4l2_subdev_command(i2c_get_clientdata(client), cmd, arg); |
| 349 | } |
| 350 | |
| 351 | If driver->command is never used then you can leave this out. Eventually the |
| 352 | driver->command usage should be removed from v4l. |
| 353 | |
| 354 | Make sure to call v4l2_device_unregister_subdev(sd) when the remove() callback |
| 355 | is called. This will unregister the sub-device from the bridge driver. It is |
| 356 | safe to call this even if the sub-device was never registered. |
| 357 | |
Hans Verkuil | f5360bd | 2009-01-15 06:09:05 -0300 | [diff] [blame] | 358 | You need to do this because when the bridge driver destroys the i2c adapter |
| 359 | the remove() callbacks are called of the i2c devices on that adapter. |
| 360 | After that the corresponding v4l2_subdev structures are invalid, so they |
| 361 | have to be unregistered first. Calling v4l2_device_unregister_subdev(sd) |
| 362 | from the remove() callback ensures that this is always done correctly. |
| 363 | |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 364 | |
| 365 | The bridge driver also has some helper functions it can use: |
| 366 | |
| 367 | struct v4l2_subdev *sd = v4l2_i2c_new_subdev(adapter, "module_foo", "chipid", 0x36); |
| 368 | |
| 369 | This loads the given module (can be NULL if no module needs to be loaded) and |
| 370 | calls i2c_new_device() with the given i2c_adapter and chip/address arguments. |
| 371 | If all goes well, then it registers the subdev with the v4l2_device. It gets |
| 372 | the v4l2_device by calling i2c_get_adapdata(adapter), so you should make sure |
Hans Verkuil | 2c79252 | 2009-03-12 18:34:19 -0300 | [diff] [blame] | 373 | to call i2c_set_adapdata(adapter, v4l2_device) when you setup the i2c_adapter |
| 374 | in your driver. |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 375 | |
| 376 | You can also use v4l2_i2c_new_probed_subdev() which is very similar to |
| 377 | v4l2_i2c_new_subdev(), except that it has an array of possible I2C addresses |
| 378 | that it should probe. Internally it calls i2c_new_probed_device(). |
| 379 | |
| 380 | Both functions return NULL if something went wrong. |
| 381 | |
Hans Verkuil | 2c79252 | 2009-03-12 18:34:19 -0300 | [diff] [blame] | 382 | Note that the chipid you pass to v4l2_i2c_new_(probed_)subdev() is usually |
| 383 | the same as the module name. It allows you to specify a chip variant, e.g. |
| 384 | "saa7114" or "saa7115". In general though the i2c driver autodetects this. |
| 385 | The use of chipid is something that needs to be looked at more closely at a |
| 386 | later date. It differs between i2c drivers and as such can be confusing. |
| 387 | To see which chip variants are supported you can look in the i2c driver code |
| 388 | for the i2c_device_id table. This lists all the possibilities. |
| 389 | |
Hans Verkuil | 2a1fcdf | 2008-11-29 21:36:58 -0300 | [diff] [blame] | 390 | |
| 391 | struct video_device |
| 392 | ------------------- |
| 393 | |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 394 | The actual device nodes in the /dev directory are created using the |
| 395 | video_device struct (v4l2-dev.h). This struct can either be allocated |
| 396 | dynamically or embedded in a larger struct. |
| 397 | |
| 398 | To allocate it dynamically use: |
| 399 | |
| 400 | struct video_device *vdev = video_device_alloc(); |
| 401 | |
| 402 | if (vdev == NULL) |
| 403 | return -ENOMEM; |
| 404 | |
| 405 | vdev->release = video_device_release; |
| 406 | |
| 407 | If you embed it in a larger struct, then you must set the release() |
| 408 | callback to your own function: |
| 409 | |
| 410 | struct video_device *vdev = &my_vdev->vdev; |
| 411 | |
| 412 | vdev->release = my_vdev_release; |
| 413 | |
| 414 | The release callback must be set and it is called when the last user |
| 415 | of the video device exits. |
| 416 | |
| 417 | The default video_device_release() callback just calls kfree to free the |
| 418 | allocated memory. |
| 419 | |
| 420 | You should also set these fields: |
| 421 | |
Hans Verkuil | dfa9a5a | 2008-12-23 12:17:23 -0300 | [diff] [blame] | 422 | - v4l2_dev: set to the v4l2_device parent device. |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 423 | - name: set to something descriptive and unique. |
Hans Verkuil | c7dd09d | 2008-12-23 13:42:25 -0300 | [diff] [blame] | 424 | - fops: set to the v4l2_file_operations struct. |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 425 | - ioctl_ops: if you use the v4l2_ioctl_ops to simplify ioctl maintenance |
| 426 | (highly recommended to use this and it might become compulsory in the |
| 427 | future!), then set this to your v4l2_ioctl_ops struct. |
Hans Verkuil | 0057596 | 2009-03-13 10:03:04 -0300 | [diff] [blame] | 428 | - parent: you only set this if v4l2_device was registered with NULL as |
| 429 | the parent device struct. This only happens in cases where one hardware |
| 430 | device has multiple PCI devices that all share the same v4l2_device core. |
| 431 | |
| 432 | The cx88 driver is an example of this: one core v4l2_device struct, but |
| 433 | it is used by both an raw video PCI device (cx8800) and a MPEG PCI device |
| 434 | (cx8802). Since the v4l2_device cannot be associated with a particular |
| 435 | PCI device it is setup without a parent device. But when the struct |
| 436 | video_device is setup you do know which parent PCI device to use. |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 437 | |
Hans Verkuil | c7dd09d | 2008-12-23 13:42:25 -0300 | [diff] [blame] | 438 | If you use v4l2_ioctl_ops, then you should set either .unlocked_ioctl or |
| 439 | .ioctl to video_ioctl2 in your v4l2_file_operations struct. |
| 440 | |
| 441 | The v4l2_file_operations struct is a subset of file_operations. The main |
| 442 | difference is that the inode argument is omitted since it is never used. |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 443 | |
| 444 | |
| 445 | video_device registration |
| 446 | ------------------------- |
| 447 | |
| 448 | Next you register the video device: this will create the character device |
| 449 | for you. |
| 450 | |
| 451 | err = video_register_device(vdev, VFL_TYPE_GRABBER, -1); |
| 452 | if (err) { |
Hans Verkuil | 50a2a8b | 2008-12-22 09:13:11 -0300 | [diff] [blame] | 453 | video_device_release(vdev); /* or kfree(my_vdev); */ |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 454 | return err; |
| 455 | } |
| 456 | |
| 457 | Which device is registered depends on the type argument. The following |
| 458 | types exist: |
| 459 | |
| 460 | VFL_TYPE_GRABBER: videoX for video input/output devices |
| 461 | VFL_TYPE_VBI: vbiX for vertical blank data (i.e. closed captions, teletext) |
| 462 | VFL_TYPE_RADIO: radioX for radio tuners |
| 463 | VFL_TYPE_VTX: vtxX for teletext devices (deprecated, don't use) |
| 464 | |
| 465 | The last argument gives you a certain amount of control over the device |
| 466 | kernel number used (i.e. the X in videoX). Normally you will pass -1 to |
| 467 | let the v4l2 framework pick the first free number. But if a driver creates |
| 468 | many devices, then it can be useful to have different video devices in |
| 469 | separate ranges. For example, video capture devices start at 0, video |
| 470 | output devices start at 16. |
| 471 | |
| 472 | So you can use the last argument to specify a minimum kernel number and |
| 473 | the v4l2 framework will try to pick the first free number that is equal |
| 474 | or higher to what you passed. If that fails, then it will just pick the |
| 475 | first free number. |
| 476 | |
| 477 | Whenever a device node is created some attributes are also created for you. |
| 478 | If you look in /sys/class/video4linux you see the devices. Go into e.g. |
| 479 | video0 and you will see 'name' and 'index' attributes. The 'name' attribute |
| 480 | is the 'name' field of the video_device struct. The 'index' attribute is |
| 481 | a device node index that can be assigned by the driver, or that is calculated |
| 482 | for you. |
| 483 | |
| 484 | If you call video_register_device(), then the index is just increased by |
| 485 | 1 for each device node you register. The first video device node you register |
| 486 | always starts off with 0. |
| 487 | |
| 488 | Alternatively you can call video_register_device_index() which is identical |
| 489 | to video_register_device(), but with an extra index argument. Here you can |
| 490 | pass a specific index value (between 0 and 31) that should be used. |
| 491 | |
| 492 | Users can setup udev rules that utilize the index attribute to make fancy |
| 493 | device names (e.g. 'mpegX' for MPEG video capture device nodes). |
| 494 | |
| 495 | After the device was successfully registered, then you can use these fields: |
| 496 | |
| 497 | - vfl_type: the device type passed to video_register_device. |
| 498 | - minor: the assigned device minor number. |
| 499 | - num: the device kernel number (i.e. the X in videoX). |
| 500 | - index: the device index number (calculated or set explicitly using |
| 501 | video_register_device_index). |
| 502 | |
| 503 | If the registration failed, then you need to call video_device_release() |
| 504 | to free the allocated video_device struct, or free your own struct if the |
| 505 | video_device was embedded in it. The vdev->release() callback will never |
| 506 | be called if the registration failed, nor should you ever attempt to |
| 507 | unregister the device if the registration failed. |
| 508 | |
| 509 | |
| 510 | video_device cleanup |
| 511 | -------------------- |
| 512 | |
| 513 | When the video device nodes have to be removed, either during the unload |
| 514 | of the driver or because the USB device was disconnected, then you should |
| 515 | unregister them: |
| 516 | |
| 517 | video_unregister_device(vdev); |
| 518 | |
| 519 | This will remove the device nodes from sysfs (causing udev to remove them |
| 520 | from /dev). |
| 521 | |
| 522 | After video_unregister_device() returns no new opens can be done. |
| 523 | |
| 524 | However, in the case of USB devices some application might still have one |
| 525 | of these device nodes open. You should block all new accesses to read, |
| 526 | write, poll, etc. except possibly for certain ioctl operations like |
| 527 | queueing buffers. |
| 528 | |
| 529 | When the last user of the video device node exits, then the vdev->release() |
| 530 | callback is called and you can do the final cleanup there. |
| 531 | |
| 532 | |
| 533 | video_device helper functions |
| 534 | ----------------------------- |
| 535 | |
| 536 | There are a few useful helper functions: |
| 537 | |
| 538 | You can set/get driver private data in the video_device struct using: |
| 539 | |
Hans Verkuil | 89aec3e | 2009-02-07 07:07:04 -0300 | [diff] [blame] | 540 | void *video_get_drvdata(struct video_device *vdev); |
| 541 | void video_set_drvdata(struct video_device *vdev, void *data); |
Hans Verkuil | a47ddf1 | 2008-12-19 10:20:22 -0300 | [diff] [blame] | 542 | |
| 543 | Note that you can safely call video_set_drvdata() before calling |
| 544 | video_register_device(). |
| 545 | |
| 546 | And this function: |
| 547 | |
| 548 | struct video_device *video_devdata(struct file *file); |
| 549 | |
| 550 | returns the video_device belonging to the file struct. |
| 551 | |
| 552 | The final helper function combines video_get_drvdata with |
| 553 | video_devdata: |
| 554 | |
| 555 | void *video_drvdata(struct file *file); |
| 556 | |
| 557 | You can go from a video_device struct to the v4l2_device struct using: |
| 558 | |
Hans Verkuil | dfa9a5a | 2008-12-23 12:17:23 -0300 | [diff] [blame] | 559 | struct v4l2_device *v4l2_dev = vdev->v4l2_dev; |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 560 | |
| 561 | video buffer helper functions |
| 562 | ----------------------------- |
| 563 | |
| 564 | The v4l2 core API provides a standard method for dealing with video |
| 565 | buffers. Those methods allow a driver to implement read(), mmap() and |
| 566 | overlay() on a consistent way. |
| 567 | |
| 568 | There are currently methods for using video buffers on devices that |
| 569 | supports DMA with scatter/gather method (videobuf-dma-sg), DMA with |
| 570 | linear access (videobuf-dma-contig), and vmalloced buffers, mostly |
| 571 | used on USB drivers (videobuf-vmalloc). |
| 572 | |
| 573 | Any driver using videobuf should provide operations (callbacks) for |
| 574 | four handlers: |
| 575 | |
| 576 | ops->buf_setup - calculates the size of the video buffers and avoid they |
| 577 | to waste more than some maximum limit of RAM; |
| 578 | ops->buf_prepare - fills the video buffer structs and calls |
| 579 | videobuf_iolock() to alloc and prepare mmaped memory; |
| 580 | ops->buf_queue - advices the driver that another buffer were |
| 581 | requested (by read() or by QBUF); |
| 582 | ops->buf_release - frees any buffer that were allocated. |
| 583 | |
| 584 | In order to use it, the driver need to have a code (generally called at |
| 585 | interrupt context) that will properly handle the buffer request lists, |
| 586 | announcing that a new buffer were filled. |
| 587 | |
| 588 | The irq handling code should handle the videobuf task lists, in order |
| 589 | to advice videobuf that a new frame were filled, in order to honor to a |
| 590 | request. The code is generally like this one: |
Mauro Carvalho Chehab | a7a1c0e | 2009-02-14 07:51:28 -0300 | [diff] [blame] | 591 | if (list_empty(&dma_q->active)) |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 592 | return; |
| 593 | |
Mauro Carvalho Chehab | a7a1c0e | 2009-02-14 07:51:28 -0300 | [diff] [blame] | 594 | buf = list_entry(dma_q->active.next, struct vbuffer, vb.queue); |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 595 | |
Mauro Carvalho Chehab | a7a1c0e | 2009-02-14 07:51:28 -0300 | [diff] [blame] | 596 | if (!waitqueue_active(&buf->vb.done)) |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 597 | return; |
| 598 | |
| 599 | /* Some logic to handle the buf may be needed here */ |
| 600 | |
Mauro Carvalho Chehab | a7a1c0e | 2009-02-14 07:51:28 -0300 | [diff] [blame] | 601 | list_del(&buf->vb.queue); |
| 602 | do_gettimeofday(&buf->vb.ts); |
| 603 | wake_up(&buf->vb.done); |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 604 | |
| 605 | Those are the videobuffer functions used on drivers, implemented on |
| 606 | videobuf-core: |
| 607 | |
Mauro Carvalho Chehab | a7a1c0e | 2009-02-14 07:51:28 -0300 | [diff] [blame] | 608 | - Videobuf init functions |
| 609 | videobuf_queue_sg_init() |
| 610 | Initializes the videobuf infrastructure. This function should be |
| 611 | called before any other videobuf function on drivers that uses DMA |
| 612 | Scatter/Gather buffers. |
| 613 | |
| 614 | videobuf_queue_dma_contig_init |
| 615 | Initializes the videobuf infrastructure. This function should be |
| 616 | called before any other videobuf function on drivers that need DMA |
| 617 | contiguous buffers. |
| 618 | |
| 619 | videobuf_queue_vmalloc_init() |
| 620 | Initializes the videobuf infrastructure. This function should be |
| 621 | called before any other videobuf function on USB (and other drivers) |
| 622 | that need a vmalloced type of videobuf. |
Mauro Carvalho Chehab | 44061c0 | 2009-02-14 07:29:07 -0300 | [diff] [blame] | 623 | |
| 624 | - videobuf_iolock() |
| 625 | Prepares the videobuf memory for the proper method (read, mmap, overlay). |
| 626 | |
| 627 | - videobuf_queue_is_busy() |
| 628 | Checks if a videobuf is streaming. |
| 629 | |
| 630 | - videobuf_queue_cancel() |
| 631 | Stops video handling. |
| 632 | |
| 633 | - videobuf_mmap_free() |
| 634 | frees mmap buffers. |
| 635 | |
| 636 | - videobuf_stop() |
| 637 | Stops video handling, ends mmap and frees mmap and other buffers. |
| 638 | |
| 639 | - V4L2 api functions. Those functions correspond to VIDIOC_foo ioctls: |
| 640 | videobuf_reqbufs(), videobuf_querybuf(), videobuf_qbuf(), |
| 641 | videobuf_dqbuf(), videobuf_streamon(), videobuf_streamoff(). |
| 642 | |
| 643 | - V4L1 api function (corresponds to VIDIOCMBUF ioctl): |
| 644 | videobuf_cgmbuf() |
| 645 | This function is used to provide backward compatibility with V4L1 |
| 646 | API. |
| 647 | |
| 648 | - Some help functions for read()/poll() operations: |
| 649 | videobuf_read_stream() |
| 650 | For continuous stream read() |
| 651 | videobuf_read_one() |
| 652 | For snapshot read() |
| 653 | videobuf_poll_stream() |
| 654 | polling help function |
| 655 | |
| 656 | The better way to understand it is to take a look at vivi driver. One |
| 657 | of the main reasons for vivi is to be a videobuf usage example. the |
| 658 | vivi_thread_tick() does the task that the IRQ callback would do on PCI |
| 659 | drivers (or the irq callback on USB). |