03:43 <palmer> mmind00: looks like it's just a QEMU regression

03:44 <palmer> oops, I moved computers and looks like they're on different channels

07:53 <wens> there's also the strict mode in pinctrl drivers which prevents you from requesting a GPIO when the pin is already claimed and used for another pin function. I think the opposite is also true.

08:58 <geertu> wens: correct

15:03 <scosu[m]> I am currently looking for a generic driver that maps two irqs (button push and release) to input events. I checked gpio keys but it seems that driver releases the key after a certain amount of time on its own. Any ideas?

15:24 <geertu> scosu[m]: If you use gpio-keys with interrupts, it indeed behaves like that. If you use it with gpios, and your GPIO controller supports that, it reports the real push and release events.

15:26 <scosu[m]> geertu: Thanks, unfortunately these aren't normal GPIOs, these are real interrupts coming from a PMIC

15:30 <geertu> scosu[m]: And you really have two separate interruots?

15:32 <geertu> That sounds like a great extension for gpio-keys to implement.

15:32 <scosu[m]> geertu: yes, I do have two, one for falling edge and one for rising, and both are not maskable as well

15:33 <geertu> scosu[m]: A single interrupt line that generates the same interrupt for falling and reising edge, or two different interrupts, one for falling edge and a second for rising edge?

15:36 <scosu[m]> geertu: two different interrupts one for falling and one for rising

15:37 <geertu> scosu[m]: OK, I guess that can be done, by adding a second interrupt to the interrupts property, and extending the driver.

15:38 <geertu> For the single interrupt both edges case, I posted https://lore.kernel.org/all/356b31ade897af77a06d6567601f038f56b3b2a2.1638538079.git.geert+renesas@glider.be a while ago. but no one commented.

15:44 <scosu[m]> geertu: Yes I think so, I just thought there is certainly a driver for that as it is such a generic use case. Well I better start then :D thanks anyways

15:46 <scosu[m]> Are you going to follow up on your patch? Should I base on top of that? I can't really test it as I don't have that setup right now, I can only read it.

15:55 <geertu> scosu[m]: I have no follow-up plans for now.

15:56 <geertu> The case with two separate interrupts is different, so I wouldn't depend on my patch.

16:06 <scosu[m]> geertu: Ok, I just asked as there will probably be an overlap and I didn't want anyone to do any unnecessary rebases :)