On Fri, Dec 13, 2019 at 01:31:18PM +0100, guillaume La Roque wrote: > Hi Johan, > > On 12/13/19 12:17 PM, Johan Hovold wrote: > > On Fri, Dec 13, 2019 at 11:55:21AM +0100, Guillaume La Roque wrote: > >> @@ -1421,6 +1422,7 @@ static int bcm_serdev_probe(struct serdev_device *serdev) > >> #endif > >> bcmdev->serdev_hu.serdev = serdev; > >> serdev_device_set_drvdata(serdev, bcmdev); > >> + bcmdev->irq = of_irq_get(bcmdev->dev->of_node, 0); > > Shouldn't you be used using of_irq_get_byname()? > i can use it if you prefer but no other interrupt need to be defined Maybe not needed then. Was just thinking it may make it more clear that you now have two ways to specify the "host-wakeup" interrupt (and in your proposed implementation the interrupts-property happens to take priority). Perhaps that can be sorted out when you submit the binding update for review. Johan
