On Thu, 12 Dec 2024 09:08:22 +0100 Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote: > On Wed, Dec 11, 2024 at 11:06:44PM +0000, Lothar Rubusch wrote: > > Add interrupt-names INT1 and INT2 for the two interrupt lines of the > > sensor. Only one line will be connected for incoming events. The driver > > needs to be configured accordingly. This is all driver info. This patch description just needs to say something like: " There are two interrupt lines that may be connected. As the device can route each type of interrupt to one or other of those lines, interrupt-names is necessary for two reasons. - One interrupt line is connected, the device has to be configured to send interrupts to that line. - Two interrupt lines connected. The device can route all interrupts to one line or elect to split them up. If no interrupt lines are connected, device functionality may be restricted. " Note as below, the required interrupts entry should be removed in a precursor patch. > If no interrupt line is set up, the > > sensor will fall back to FIFO bypass mode and still measure, but no > > interrupt based events are possible. > > There was interrupt before and it was required, so I do not understand > last statement. You describe case which is impossible. Binding was wrong. Interrupt isn't required for quite a bit of the functionality. I'd like to see an earlier patch removing that required entry and explaining why rather than jumping into adding the new interrupt-names part without resolving that. Its a relaxation of constraints so probably no need to backport that patch. Jonathan > > Best regards, > Krzysztof >
