On Mon, Nov 16, 2015 at 9:57 AM, Linus Walleij <linus.walleij@xxxxxxxxxx> wrote: > On Mon, Nov 16, 2015 at 7:15 AM, Denis Ciocca <denis.ciocca@xxxxxx> wrote: > >> I'm not convinced about this one. If we declare sensors output as open drain, >> interrupt can be shared (as you also say). >> >> So, the irq management function should first check source registers if >> interrupt is really generated by sensor or not. > > Yeah I was not testing actually using two sensors at the time so far, only > setting this property in the device tree for two sensors so that it was > possible to share a line between two sensors, but I only use one at a > time, I didn't try to start generic_buffer on both similtaneously... > > OK I'll see if we can get the trigger to return IRQ_HANDLED or > IRQ_NO_IRQ. Actually this ties into the IRQ request discussion: to read the status from the sensor, we need to do some I2C traffic, and for that, we need to have a threaded IRQ handler (today we don't have that, we just request a fastpath IRQ using the request_threaded() function), leading back to the other discussion about any_context_irq() and how we should just call down to the trigger layer if we are running a real thread in the IRQ handler. It's a can of worms :D Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html