On 10/29/22 14:35, Jonathan Cameron wrote: > On Mon, 24 Oct 2022 05:44:21 +0000 > "Vaittinen, Matti" <Matti.Vaittinen@xxxxxxxxxxxxxxxxx> wrote: > >>>> + >>>> + if (fwnode_irq_get_byname(fwnode, "INT2") > 0) >>>> + dev_warn(dev, "Only one IRQ supported\n"); >>> >>> Why? If you get the both, only the first one will be used by the driver. >>> Not really worth warning about the lack of features... >> >> My thinking regarding developing new device went along the lines: >> >> Precondition: The HW (and data-sheet) explain how there is two INT pins. >> 1. Board designer reads the data-sheet and uses both INT pins. >> 2. SW engineer finds the driver and reads the DT-binding description. >> 3. SW engineer writes the DT-description and hopes everything "just >> works". (Amount of hope is probably inversely proportional to the amount >> of experience XD). >> 4) SW engineer gives a first go to the sensor SW and notices everything >> does not just magically work. > > Ah but it does "work". We simply don't use one of the IRQs. That's not > normally a problem as there are lots of other features we don't fully > support. Not using something is not normally considered a problem. I guess you're correct... > So far I'm not seeing anything that doesn't work because we only support > one IRQ. There may be combinations of interrupts that are tricky to handle > on one IRQ line (or may not be supported at the same time on a single line) > but so far you don't support those anyway.. Adding a more informative warning > when adding those features would be reasonable > > "Feature X not supported as only a single IRQ line available". ... and I also fully agree with this. My thinking was stuck with each IRQ line having a purpose fixed by HW - which is not the case with this sensor. Yours -- Matti -- Matti Vaittinen Linux kernel developer at ROHM Semiconductors Oulu Finland ~~ When things go utterly wrong vim users can always type :help! ~~
