On Mon, Dec 12, 2022 at 05:08:17PM +0100, Johan Hovold wrote: > The fact that USB core and driver core clears the pointer is an > implementation detail which in principle could change. With the new kerneldoc, the fact that the USB core clears the pointer becomes part of the API. It is no longer an implementation detail but rather a guarantee that drivers can depend on. (I.e., if it does change, all the USB drivers will need to be audited.) > The important > part is that drivers generally should not touch the struct device or > containing structure after unbind (and must do so with care otherwise). While that certainly is an important point, it is not the point of usb_set_intfdata() or its kerneldoc. The main reason for adding the kerneldoc -- Vincent's original motivation -- was to inform coders that drivers generally don't need to worry about clearing the data pointer. A subtext was that clearing the pointer while the driver is still using it can lead to problems, but I don't think we need to mention this. It's pretty obvious. Alan Stern
