On Tue, 21 Apr 2009, Hans Verkuil wrote: > > Well, you might look at drivers/media/video/soc_camera_platform.c for an > > example of a simple "pseudo" camera driver. Of course, with your two > > additional devices you don't want to add extra platform devices and extra > > probing. In fact, you can do this with the "old" (currently in the > > mainline) soc-camera model, where client drivers actively report > > themselves to the soc-camera core using soc_camera_device_register() / > > soc_camera_device_unregister() and the core doesn't care about the nature > > of those drivers. This is not going to be the case with the new platform / > > v4l2-subdev infrastructure, which is pretty tightly bound to i2c... So, > > we'll have to extend it too. > > Not true. v4l2-device and v4l2-subdev are bus-independent. Only the > v4l2-i2c-* helper functions in v4l2-common.c are i2c dependent. For > example, ivtv uses v4l2_subdev to control devices connected via gpio, > while cx18 uses it for a logical video decoder block on the main asic. > > Other than initialization and possibly cleanup there should be NO > bus-dependencies. Sorry, what I wrote above wasn't clear enough. By "platform / v4l2-subdev infrastructure" I meant the current (as of my patch from a couple of hours ago) soc-camera - platform stack linked to v4l2-subdev the way it is implemented there. In that patch soc-camera uses the i2c interface of v4l2-subdev directly and is thus rather i2c-centric. So, we will have to extend soc-camera to also support the bus-neutral v4l2-subdev API. Thanks Guennadi --- Guennadi Liakhovetski, Ph.D. Freelance Open-Source Software Developer http://www.open-technology.de/ -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
