cam_periph_find

[Alexander Motin]

Matthew Jacob wrote:
> There is currently nothing that keeps a periph from disappearing between
> the time you find it with cam_periph_find and the try to 'acquire' it
> later (if you indeed do so).
> 
> What do people feel about changing the semantic of cam_periph_find to up
> the refcount (requiring the caller to release or release_locked)?

My general feeling is that everything that every reference should be
counted. But with quick look on cam_periph_find() usage I would say that
most (all?) cam_periph_find() calls protected with respective SIM lock.

-- 
Alexander Motin