The dup() function returns a duplicate of the file descriptor. Wrapping it in a FileDescriptor isn't wrong as such, but it prevents from using it in contexts where a UniqueFD is needed. As the duplicate is guaranteed to have a single owner when created, return it as a UniqueFD instead. A FileDescriptor can easily be created from the UniqueFD if desired. Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Reviewed-by: Hirokazu Honda <hiroh@chromium.org> Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>