On Thu Dec 19, 2024 at 8:35 PM CET, Dmitry Torokhov wrote: > Hi Mathieu, > > On Thu, Dec 19, 2024 at 05:21:23PM +0100, Mathieu Dubois-Briand wrote: > > + > > + input_set_capability(input, EV_MSC, MSC_SCAN); > > + if (!max7360_keypad->no_autorepeat) > > + __set_bit(EV_REP, input->evbit); > > + > > + input_set_drvdata(input, max7360_keypad); > > + > > + flags = IRQF_TRIGGER_LOW | IRQF_ONESHOT | IRQF_SHARED; > > Shared? Why? And why do you need a temp variable? > Ok, this is probably wrong. I have a board using the MAX7360, were both interrupt lines of the chipset (INTI, interrupt for GPIO and rotary encoder and INTK, interrupt for the keypad) have been both wired to the same pin on the CPU side. So on this board, the interrupt line is indeed shared between the three corresponding drivers. Yet this is probably very specific to my board, as I have seen no data about MAXIM suggesting this design. So having IRQF_SHARED is probably more a hack than a valid implementation. I will drop it in my next series. > > Thanks. Thanks a lot for you review, I am preparing a new version of this series that should address all your comments. -- Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
