Hi everyone, These two patches make the smiapp driver independent of the s_power() core sub-device callback, and instead use runtime PM to control the device's power state. This should be the model for future sensor drivers: the s_power() callback is really something we want to get rid of. A cleaner implementation would naturally be nicer. The underlying problem is that the set_ctrl() callback is called through the control handler so there's no way to figure out in the set_ctrl() callback itself where the call came from: via the device's power_on() callback or from the user calling VIDIOC_S_CTRL. Another option would be to postpone writing control values until s_stream() callback but that introduces an additional delay for streaming which is always better to avoid. Sakari Ailus (2): smiapp: Use __v4l2_ctrl_handler_setup() smiapp: Rely on runtime PM drivers/media/i2c/smiapp/smiapp-core.c | 97 +++++++++++++--------------------- drivers/media/i2c/smiapp/smiapp-regs.c | 3 ++ drivers/media/i2c/smiapp/smiapp.h | 1 + 3 files changed, 40 insertions(+), 61 deletions(-) -- 2.11.0
