blob: 854808b67faed03f209291ae2947be5f051e8af6 [file] [log] [blame]
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -03001Overview of the V4L2 driver framework
2=====================================
3
4This text documents the various structures provided by the V4L2 framework and
5their relationships.
6
7
8Introduction
9------------
10
11The V4L2 drivers tend to be very complex due to the complexity of the
12hardware: 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
16Especially the fact that V4L2 drivers have to setup supporting ICs to
17do audio/video muxing/encoding/decoding makes it more complex than most.
18Usually these ICs are connected to the main bridge driver through one or
19more I2C busses, but other busses can also be used. Such devices are
20called 'sub-devices'.
21
22For a long time the framework was limited to the video_device struct for
23creating V4L device nodes and video_buf for handling the video buffers
24(note that this document does not discuss the video_buf framework).
25
26This meant that all drivers had to do the setup of device instances and
27connecting to sub-devices themselves. Some of this is quite complicated
28to do right and many drivers never did do it correctly.
29
30There is also a lot of common code that could never be refactored due to
31the lack of a framework.
32
33So this framework sets up the basic building blocks that all drivers
34need and this same framework should make it much easier to refactor
35common code into utility functions shared by all drivers.
36
37
38Structure of a driver
39---------------------
40
41All drivers have the following structure:
42
431) A struct for each device instance containing the device state.
44
452) A way of initializing and commanding sub-devices (if any).
46
473) 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 Chehab44061c02009-02-14 07:29:07 -0300504) Filehandle-specific structs containing per-filehandle data;
51
525) video buffer handling.
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -030053
54This 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
65Structure of the framework
66--------------------------
67
68The framework closely resembles the driver structure: it has a v4l2_device
69struct for the device instance data, a v4l2_subdev struct to refer to
70sub-device instances, the video_device struct stores V4L2 device node data
71and in the future a v4l2_fh struct will keep track of filehandle instances
72(this is not yet implemented).
73
74
75struct v4l2_device
76------------------
77
78Each device instance is represented by a struct v4l2_device (v4l2-device.h).
79Very simple devices can just allocate this struct, but most of the time you
80would embed this struct inside a larger struct.
81
82You must register the device instance:
83
84 v4l2_device_register(struct device *dev, struct v4l2_device *v4l2_dev);
85
86Registration will initialize the v4l2_device struct and link dev->driver_data
Hans Verkuil3a63e4492009-02-14 11:54:23 -030087to v4l2_dev. If v4l2_dev->name is empty then it will be set to a value derived
88from dev (driver name followed by the bus_id, to be precise). If you set it
89up before calling v4l2_device_register then it will be untouched. If dev is
90NULL, then you *must* setup v4l2_dev->name before calling v4l2_device_register.
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -030091
Hans Verkuila47ddf12008-12-19 10:20:22 -030092The first 'dev' argument is normally the struct device pointer of a pci_dev,
Janne Grunau073d6962009-04-01 08:30:06 -030093usb_interface or platform_device. It is rare for dev to be NULL, but it happens
Hans Verkuil00575962009-03-13 10:03:04 -030094with ISA devices or when one device creates multiple PCI devices, thus making
95it impossible to associate v4l2_dev with a particular parent.
Hans Verkuila47ddf12008-12-19 10:20:22 -030096
Hans Verkuil98ec6332009-03-08 17:02:10 -030097You can also supply a notify() callback that can be called by sub-devices to
98notify you of events. Whether you need to set this depends on the sub-device.
99Any notifications a sub-device supports must be defined in a header in
100include/media/<subdevice>.h.
101
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300102You unregister with:
103
104 v4l2_device_unregister(struct v4l2_device *v4l2_dev);
105
106Unregistering will also automatically unregister all subdevs from the device.
107
Hans Verkuilae6cfaa2009-03-14 08:28:45 -0300108If you have a hotpluggable device (e.g. a USB device), then when a disconnect
109happens the parent device becomes invalid. Since v4l2_device has a pointer to
110that parent device it has to be cleared as well to mark that the parent is
111gone. To do this call:
112
113 v4l2_device_disconnect(struct v4l2_device *v4l2_dev);
114
115This does *not* unregister the subdevs, so you still need to call the
116v4l2_device_unregister() function for that. If your driver is not hotpluggable,
117then there is no need to call v4l2_device_disconnect().
118
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300119Sometimes you need to iterate over all devices registered by a specific
120driver. This is usually the case if multiple device drivers use the same
121hardware. E.g. the ivtvfb driver is a framebuffer driver that uses the ivtv
122hardware. The same is true for alsa drivers for example.
123
124You can iterate over all registered devices as follows:
125
126static int callback(struct device *dev, void *p)
127{
128 struct v4l2_device *v4l2_dev = dev_get_drvdata(dev);
129
130 /* test if this device was inited */
131 if (v4l2_dev == NULL)
132 return 0;
133 ...
134 return 0;
135}
136
137int iterate(void *p)
138{
139 struct device_driver *drv;
140 int err;
141
142 /* Find driver 'ivtv' on the PCI bus.
143 pci_bus_type is a global. For USB busses use usb_bus_type. */
144 drv = driver_find("ivtv", &pci_bus_type);
145 /* iterate over all ivtv device instances */
146 err = driver_for_each_device(drv, NULL, p, callback);
147 put_driver(drv);
148 return err;
149}
150
151Sometimes you need to keep a running counter of the device instance. This is
152commonly used to map a device instance to an index of a module option array.
153
154The recommended approach is as follows:
155
156static atomic_t drv_instance = ATOMIC_INIT(0);
157
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300158static int __devinit drv_probe(struct pci_dev *pdev,
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300159 const struct pci_device_id *pci_id)
160{
161 ...
162 state->instance = atomic_inc_return(&drv_instance) - 1;
163}
164
165
166struct v4l2_subdev
167------------------
168
169Many drivers need to communicate with sub-devices. These devices can do all
170sort of tasks, but most commonly they handle audio and/or video muxing,
171encoding or decoding. For webcams common sub-devices are sensors and camera
172controllers.
173
174Usually these are I2C devices, but not necessarily. In order to provide the
175driver with a consistent interface to these sub-devices the v4l2_subdev struct
176(v4l2-subdev.h) was created.
177
178Each sub-device driver must have a v4l2_subdev struct. This struct can be
179stand-alone for simple sub-devices or it might be embedded in a larger struct
180if more state information needs to be stored. Usually there is a low-level
181device struct (e.g. i2c_client) that contains the device data as setup
182by the kernel. It is recommended to store that pointer in the private
183data of v4l2_subdev using v4l2_set_subdevdata(). That makes it easy to go
184from a v4l2_subdev to the actual low-level bus-specific device data.
185
186You also need a way to go from the low-level struct to v4l2_subdev. For the
187common i2c_client struct the i2c_set_clientdata() call is used to store a
188v4l2_subdev pointer, for other busses you may have to use other methods.
189
190From the bridge driver perspective you load the sub-device module and somehow
191obtain the v4l2_subdev pointer. For i2c devices this is easy: you call
192i2c_get_clientdata(). For other busses something similar needs to be done.
193Helper functions exists for sub-devices on an I2C bus that do most of this
194tricky work for you.
195
196Each v4l2_subdev contains function pointers that sub-device drivers can
197implement (or leave NULL if it is not applicable). Since sub-devices can do
198so many different things and you do not want to end up with a huge ops struct
199of which only a handful of ops are commonly implemented, the function pointers
200are sorted according to category and each category has its own ops struct.
201
202The top-level ops struct contains pointers to the category ops structs, which
203may be NULL if the subdev driver does not support anything from that category.
204
205It looks like this:
206
207struct v4l2_subdev_core_ops {
Hans Verkuilaecde8b52008-12-30 07:14:19 -0300208 int (*g_chip_ident)(struct v4l2_subdev *sd, struct v4l2_dbg_chip_ident *chip);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300209 int (*log_status)(struct v4l2_subdev *sd);
210 int (*init)(struct v4l2_subdev *sd, u32 val);
211 ...
212};
213
214struct v4l2_subdev_tuner_ops {
215 ...
216};
217
218struct v4l2_subdev_audio_ops {
219 ...
220};
221
222struct v4l2_subdev_video_ops {
223 ...
224};
225
226struct v4l2_subdev_ops {
227 const struct v4l2_subdev_core_ops *core;
228 const struct v4l2_subdev_tuner_ops *tuner;
229 const struct v4l2_subdev_audio_ops *audio;
230 const struct v4l2_subdev_video_ops *video;
231};
232
233The core ops are common to all subdevs, the other categories are implemented
234depending on the sub-device. E.g. a video device is unlikely to support the
235audio ops and vice versa.
236
237This setup limits the number of function pointers while still making it easy
238to add new ops and categories.
239
240A sub-device driver initializes the v4l2_subdev struct using:
241
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300242 v4l2_subdev_init(sd, &ops);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300243
244Afterwards you need to initialize subdev->name with a unique name and set the
245module owner. This is done for you if you use the i2c helper functions.
246
247A device (bridge) driver needs to register the v4l2_subdev with the
248v4l2_device:
249
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300250 int err = v4l2_device_register_subdev(v4l2_dev, sd);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300251
252This can fail if the subdev module disappeared before it could be registered.
253After this function was called successfully the subdev->dev field points to
254the v4l2_device.
255
256You can unregister a sub-device using:
257
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300258 v4l2_device_unregister_subdev(sd);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300259
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300260Afterwards the subdev module can be unloaded and sd->dev == NULL.
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300261
262You can call an ops function either directly:
263
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300264 err = sd->ops->core->g_chip_ident(sd, &chip);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300265
266but it is better and easier to use this macro:
267
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300268 err = v4l2_subdev_call(sd, core, g_chip_ident, &chip);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300269
270The macro will to the right NULL pointer checks and returns -ENODEV if subdev
271is NULL, -ENOIOCTLCMD if either subdev->core or subdev->core->g_chip_ident is
272NULL, or the actual result of the subdev->ops->core->g_chip_ident ops.
273
274It is also possible to call all or a subset of the sub-devices:
275
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300276 v4l2_device_call_all(v4l2_dev, 0, core, g_chip_ident, &chip);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300277
278Any subdev that does not support this ops is skipped and error results are
279ignored. If you want to check for errors use this:
280
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300281 err = v4l2_device_call_until_err(v4l2_dev, 0, core, g_chip_ident, &chip);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300282
283Any error except -ENOIOCTLCMD will exit the loop with that error. If no
284errors (except -ENOIOCTLCMD) occured, then 0 is returned.
285
286The second argument to both calls is a group ID. If 0, then all subdevs are
287called. If non-zero, then only those whose group ID match that value will
Hans Verkuilb0167602009-02-14 12:00:53 -0300288be called. Before a bridge driver registers a subdev it can set sd->grp_id
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300289to whatever value it wants (it's 0 by default). This value is owned by the
290bridge driver and the sub-device driver will never modify or use it.
291
292The group ID gives the bridge driver more control how callbacks are called.
293For example, there may be multiple audio chips on a board, each capable of
294changing the volume. But usually only one will actually be used when the
295user want to change the volume. You can set the group ID for that subdev to
296e.g. AUDIO_CONTROLLER and specify that as the group ID value when calling
297v4l2_device_call_all(). That ensures that it will only go to the subdev
298that needs it.
299
Hans Verkuil98ec6332009-03-08 17:02:10 -0300300If the sub-device needs to notify its v4l2_device parent of an event, then
301it can call v4l2_subdev_notify(sd, notification, arg). This macro checks
302whether there is a notify() callback defined and returns -ENODEV if not.
303Otherwise the result of the notify() call is returned.
304
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300305The advantage of using v4l2_subdev is that it is a generic struct and does
306not contain any knowledge about the underlying hardware. So a driver might
307contain several subdevs that use an I2C bus, but also a subdev that is
308controlled through GPIO pins. This distinction is only relevant when setting
309up the device, but once the subdev is registered it is completely transparent.
310
311
312I2C sub-device drivers
313----------------------
314
315Since these drivers are so common, special helper functions are available to
316ease the use of these drivers (v4l2-common.h).
317
318The recommended method of adding v4l2_subdev support to an I2C driver is to
319embed the v4l2_subdev struct into the state struct that is created for each
320I2C device instance. Very simple devices have no state struct and in that case
321you can just create a v4l2_subdev directly.
322
323A typical state struct would look like this (where 'chipname' is replaced by
324the name of the chip):
325
326struct chipname_state {
327 struct v4l2_subdev sd;
328 ... /* additional state fields */
329};
330
331Initialize the v4l2_subdev struct as follows:
332
333 v4l2_i2c_subdev_init(&state->sd, client, subdev_ops);
334
335This function will fill in all the fields of v4l2_subdev and ensure that the
336v4l2_subdev and i2c_client both point to one another.
337
338You should also add a helper inline function to go from a v4l2_subdev pointer
339to a chipname_state struct:
340
341static inline struct chipname_state *to_state(struct v4l2_subdev *sd)
342{
343 return container_of(sd, struct chipname_state, sd);
344}
345
346Use this to go from the v4l2_subdev struct to the i2c_client struct:
347
348 struct i2c_client *client = v4l2_get_subdevdata(sd);
349
350And this to go from an i2c_client to a v4l2_subdev struct:
351
352 struct v4l2_subdev *sd = i2c_get_clientdata(client);
353
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300354Make sure to call v4l2_device_unregister_subdev(sd) when the remove() callback
355is called. This will unregister the sub-device from the bridge driver. It is
356safe to call this even if the sub-device was never registered.
357
Hans Verkuilf5360bd2009-01-15 06:09:05 -0300358You need to do this because when the bridge driver destroys the i2c adapter
359the remove() callbacks are called of the i2c devices on that adapter.
360After that the corresponding v4l2_subdev structures are invalid, so they
361have to be unregistered first. Calling v4l2_device_unregister_subdev(sd)
362from the remove() callback ensures that this is always done correctly.
363
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300364
365The bridge driver also has some helper functions it can use:
366
Hans Verkuile6574f22009-04-01 03:57:53 -0300367struct v4l2_subdev *sd = v4l2_i2c_new_subdev(v4l2_dev, adapter,
368 "module_foo", "chipid", 0x36);
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300369
370This loads the given module (can be NULL if no module needs to be loaded) and
371calls i2c_new_device() with the given i2c_adapter and chip/address arguments.
Hans Verkuile6574f22009-04-01 03:57:53 -0300372If all goes well, then it registers the subdev with the v4l2_device.
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300373
374You can also use v4l2_i2c_new_probed_subdev() which is very similar to
375v4l2_i2c_new_subdev(), except that it has an array of possible I2C addresses
376that it should probe. Internally it calls i2c_new_probed_device().
377
378Both functions return NULL if something went wrong.
379
Hans Verkuil2c792522009-03-12 18:34:19 -0300380Note that the chipid you pass to v4l2_i2c_new_(probed_)subdev() is usually
381the same as the module name. It allows you to specify a chip variant, e.g.
382"saa7114" or "saa7115". In general though the i2c driver autodetects this.
383The use of chipid is something that needs to be looked at more closely at a
384later date. It differs between i2c drivers and as such can be confusing.
385To see which chip variants are supported you can look in the i2c driver code
386for the i2c_device_id table. This lists all the possibilities.
387
Hans Verkuil2a1fcdf2008-11-29 21:36:58 -0300388
389struct video_device
390-------------------
391
Hans Verkuila47ddf12008-12-19 10:20:22 -0300392The actual device nodes in the /dev directory are created using the
393video_device struct (v4l2-dev.h). This struct can either be allocated
394dynamically or embedded in a larger struct.
395
396To allocate it dynamically use:
397
398 struct video_device *vdev = video_device_alloc();
399
400 if (vdev == NULL)
401 return -ENOMEM;
402
403 vdev->release = video_device_release;
404
405If you embed it in a larger struct, then you must set the release()
406callback to your own function:
407
408 struct video_device *vdev = &my_vdev->vdev;
409
410 vdev->release = my_vdev_release;
411
412The release callback must be set and it is called when the last user
413of the video device exits.
414
415The default video_device_release() callback just calls kfree to free the
416allocated memory.
417
418You should also set these fields:
419
Hans Verkuildfa9a5a2008-12-23 12:17:23 -0300420- v4l2_dev: set to the v4l2_device parent device.
Hans Verkuila47ddf12008-12-19 10:20:22 -0300421- name: set to something descriptive and unique.
Hans Verkuilc7dd09d2008-12-23 13:42:25 -0300422- fops: set to the v4l2_file_operations struct.
Hans Verkuila47ddf12008-12-19 10:20:22 -0300423- ioctl_ops: if you use the v4l2_ioctl_ops to simplify ioctl maintenance
424 (highly recommended to use this and it might become compulsory in the
425 future!), then set this to your v4l2_ioctl_ops struct.
Hans Verkuil00575962009-03-13 10:03:04 -0300426- parent: you only set this if v4l2_device was registered with NULL as
427 the parent device struct. This only happens in cases where one hardware
428 device has multiple PCI devices that all share the same v4l2_device core.
429
430 The cx88 driver is an example of this: one core v4l2_device struct, but
431 it is used by both an raw video PCI device (cx8800) and a MPEG PCI device
432 (cx8802). Since the v4l2_device cannot be associated with a particular
433 PCI device it is setup without a parent device. But when the struct
434 video_device is setup you do know which parent PCI device to use.
Hans Verkuila47ddf12008-12-19 10:20:22 -0300435
Hans Verkuilc7dd09d2008-12-23 13:42:25 -0300436If you use v4l2_ioctl_ops, then you should set either .unlocked_ioctl or
437.ioctl to video_ioctl2 in your v4l2_file_operations struct.
438
439The v4l2_file_operations struct is a subset of file_operations. The main
440difference is that the inode argument is omitted since it is never used.
Hans Verkuila47ddf12008-12-19 10:20:22 -0300441
442
443video_device registration
444-------------------------
445
446Next you register the video device: this will create the character device
447for you.
448
449 err = video_register_device(vdev, VFL_TYPE_GRABBER, -1);
450 if (err) {
Hans Verkuil50a2a8b2008-12-22 09:13:11 -0300451 video_device_release(vdev); /* or kfree(my_vdev); */
Hans Verkuila47ddf12008-12-19 10:20:22 -0300452 return err;
453 }
454
455Which device is registered depends on the type argument. The following
456types exist:
457
458VFL_TYPE_GRABBER: videoX for video input/output devices
459VFL_TYPE_VBI: vbiX for vertical blank data (i.e. closed captions, teletext)
460VFL_TYPE_RADIO: radioX for radio tuners
461VFL_TYPE_VTX: vtxX for teletext devices (deprecated, don't use)
462
463The last argument gives you a certain amount of control over the device
464kernel number used (i.e. the X in videoX). Normally you will pass -1 to
465let the v4l2 framework pick the first free number. But if a driver creates
466many devices, then it can be useful to have different video devices in
467separate ranges. For example, video capture devices start at 0, video
468output devices start at 16.
469
470So you can use the last argument to specify a minimum kernel number and
471the v4l2 framework will try to pick the first free number that is equal
472or higher to what you passed. If that fails, then it will just pick the
473first free number.
474
475Whenever a device node is created some attributes are also created for you.
476If you look in /sys/class/video4linux you see the devices. Go into e.g.
477video0 and you will see 'name' and 'index' attributes. The 'name' attribute
478is the 'name' field of the video_device struct. The 'index' attribute is
479a device node index that can be assigned by the driver, or that is calculated
480for you.
481
482If you call video_register_device(), then the index is just increased by
4831 for each device node you register. The first video device node you register
484always starts off with 0.
485
486Alternatively you can call video_register_device_index() which is identical
487to video_register_device(), but with an extra index argument. Here you can
488pass a specific index value (between 0 and 31) that should be used.
489
490Users can setup udev rules that utilize the index attribute to make fancy
491device names (e.g. 'mpegX' for MPEG video capture device nodes).
492
493After the device was successfully registered, then you can use these fields:
494
495- vfl_type: the device type passed to video_register_device.
496- minor: the assigned device minor number.
497- num: the device kernel number (i.e. the X in videoX).
498- index: the device index number (calculated or set explicitly using
499 video_register_device_index).
500
501If the registration failed, then you need to call video_device_release()
502to free the allocated video_device struct, or free your own struct if the
503video_device was embedded in it. The vdev->release() callback will never
504be called if the registration failed, nor should you ever attempt to
505unregister the device if the registration failed.
506
507
508video_device cleanup
509--------------------
510
511When the video device nodes have to be removed, either during the unload
512of the driver or because the USB device was disconnected, then you should
513unregister them:
514
515 video_unregister_device(vdev);
516
517This will remove the device nodes from sysfs (causing udev to remove them
518from /dev).
519
520After video_unregister_device() returns no new opens can be done.
521
522However, in the case of USB devices some application might still have one
523of these device nodes open. You should block all new accesses to read,
524write, poll, etc. except possibly for certain ioctl operations like
525queueing buffers.
526
527When the last user of the video device node exits, then the vdev->release()
528callback is called and you can do the final cleanup there.
529
530
531video_device helper functions
532-----------------------------
533
534There are a few useful helper functions:
535
536You can set/get driver private data in the video_device struct using:
537
Hans Verkuil89aec3e2009-02-07 07:07:04 -0300538void *video_get_drvdata(struct video_device *vdev);
539void video_set_drvdata(struct video_device *vdev, void *data);
Hans Verkuila47ddf12008-12-19 10:20:22 -0300540
541Note that you can safely call video_set_drvdata() before calling
542video_register_device().
543
544And this function:
545
546struct video_device *video_devdata(struct file *file);
547
548returns the video_device belonging to the file struct.
549
550The final helper function combines video_get_drvdata with
551video_devdata:
552
553void *video_drvdata(struct file *file);
554
555You can go from a video_device struct to the v4l2_device struct using:
556
Hans Verkuildfa9a5a2008-12-23 12:17:23 -0300557struct v4l2_device *v4l2_dev = vdev->v4l2_dev;
Mauro Carvalho Chehab44061c02009-02-14 07:29:07 -0300558
559video buffer helper functions
560-----------------------------
561
562The v4l2 core API provides a standard method for dealing with video
563buffers. Those methods allow a driver to implement read(), mmap() and
564overlay() on a consistent way.
565
566There are currently methods for using video buffers on devices that
567supports DMA with scatter/gather method (videobuf-dma-sg), DMA with
568linear access (videobuf-dma-contig), and vmalloced buffers, mostly
569used on USB drivers (videobuf-vmalloc).
570
571Any driver using videobuf should provide operations (callbacks) for
572four handlers:
573
574ops->buf_setup - calculates the size of the video buffers and avoid they
575 to waste more than some maximum limit of RAM;
576ops->buf_prepare - fills the video buffer structs and calls
577 videobuf_iolock() to alloc and prepare mmaped memory;
578ops->buf_queue - advices the driver that another buffer were
579 requested (by read() or by QBUF);
580ops->buf_release - frees any buffer that were allocated.
581
582In order to use it, the driver need to have a code (generally called at
583interrupt context) that will properly handle the buffer request lists,
584announcing that a new buffer were filled.
585
586The irq handling code should handle the videobuf task lists, in order
587to advice videobuf that a new frame were filled, in order to honor to a
588request. The code is generally like this one:
Mauro Carvalho Chehaba7a1c0e2009-02-14 07:51:28 -0300589 if (list_empty(&dma_q->active))
Mauro Carvalho Chehab44061c02009-02-14 07:29:07 -0300590 return;
591
Mauro Carvalho Chehaba7a1c0e2009-02-14 07:51:28 -0300592 buf = list_entry(dma_q->active.next, struct vbuffer, vb.queue);
Mauro Carvalho Chehab44061c02009-02-14 07:29:07 -0300593
Mauro Carvalho Chehaba7a1c0e2009-02-14 07:51:28 -0300594 if (!waitqueue_active(&buf->vb.done))
Mauro Carvalho Chehab44061c02009-02-14 07:29:07 -0300595 return;
596
597 /* Some logic to handle the buf may be needed here */
598
Mauro Carvalho Chehaba7a1c0e2009-02-14 07:51:28 -0300599 list_del(&buf->vb.queue);
600 do_gettimeofday(&buf->vb.ts);
601 wake_up(&buf->vb.done);
Mauro Carvalho Chehab44061c02009-02-14 07:29:07 -0300602
603Those are the videobuffer functions used on drivers, implemented on
604videobuf-core:
605
Mauro Carvalho Chehaba7a1c0e2009-02-14 07:51:28 -0300606- Videobuf init functions
607 videobuf_queue_sg_init()
608 Initializes the videobuf infrastructure. This function should be
609 called before any other videobuf function on drivers that uses DMA
610 Scatter/Gather buffers.
611
612 videobuf_queue_dma_contig_init
613 Initializes the videobuf infrastructure. This function should be
614 called before any other videobuf function on drivers that need DMA
615 contiguous buffers.
616
617 videobuf_queue_vmalloc_init()
618 Initializes the videobuf infrastructure. This function should be
619 called before any other videobuf function on USB (and other drivers)
620 that need a vmalloced type of videobuf.
Mauro Carvalho Chehab44061c02009-02-14 07:29:07 -0300621
622- videobuf_iolock()
623 Prepares the videobuf memory for the proper method (read, mmap, overlay).
624
625- videobuf_queue_is_busy()
626 Checks if a videobuf is streaming.
627
628- videobuf_queue_cancel()
629 Stops video handling.
630
631- videobuf_mmap_free()
632 frees mmap buffers.
633
634- videobuf_stop()
635 Stops video handling, ends mmap and frees mmap and other buffers.
636
637- V4L2 api functions. Those functions correspond to VIDIOC_foo ioctls:
638 videobuf_reqbufs(), videobuf_querybuf(), videobuf_qbuf(),
639 videobuf_dqbuf(), videobuf_streamon(), videobuf_streamoff().
640
641- V4L1 api function (corresponds to VIDIOCMBUF ioctl):
642 videobuf_cgmbuf()
643 This function is used to provide backward compatibility with V4L1
644 API.
645
646- Some help functions for read()/poll() operations:
647 videobuf_read_stream()
648 For continuous stream read()
649 videobuf_read_one()
650 For snapshot read()
651 videobuf_poll_stream()
652 polling help function
653
654The better way to understand it is to take a look at vivi driver. One
655of the main reasons for vivi is to be a videobuf usage example. the
656vivi_thread_tick() does the task that the IRQ callback would do on PCI
657drivers (or the irq callback on USB).