Hi Vinod, On Tuesday, 29 May 2018 11:29:32 EEST Vinod wrote: > Hi folks, > > I am writing a driver for camera control inteface which is an i2c > controller. So looking up the code I think it can be a v4l subdev, > right? Can it be an independent i2c master and not v4l subdev? What do you mean by "camera control interface" here ? A hardware device handling communication with camera sensors ? I assume the communication bus is I2C ? Is that "camera control interface" plain I2C or does it have additional features ? If we're talking about an I2C controller a V4L2 subdev is not only unneeded, but it wouldn't help. You need an I2C master. > Second the control sports GPIOs. It can support a set of > synchronization primitives so it's possible to drive I2C clients and > GPIOs with hardware controlled timing to allow for sync control of > sensors hooked and also for fancy strobe. How would we represent these > gpios in v4l2 and allow the control, any ideas on that. Even if your main use case it related to camera, synchronization of I2C and GPIO doesn't seem to be a V4L2 feature to me. It sounds that you need to implement that int he I2C and GPIO subsystems. -- Regards, Laurent Pinchart