On Mon, Apr 3, 2017 at 6:05 PM, Thierry Reding <thierry.reding@xxxxxxxxx> wrote: > here's a first draft of patches to make GPIOLIB_IRQCHIP flexible enough > to support the Tegra186 use-case. Instead of just fixing the bugs that I > had pointed out earlier, this goes one step further and completely > integrates IRQ chip registration into GPIO chip registration. The idea > is that drivers will initialize a set of fields (collected in a new > struct gpio_irq_chip) before calling gpiochip_add() (or derivatives) and > gpiolib will internally set up an IRQ domain and register the IRQ chip. > > Currently this makes use of the new infrastructure only in the Tegra186 > driver, but I plan to convert a couple of others once we can agree that > this is a direction that you're comfortable with. So I like what you are doing here, I applied patch 1 which is an obvious fix. Patches 2-11 need to be applied early after v4.12-rc1 (after rebasing) and pulled into both the GPIO and pin control trees so we do not get development glitches. It gives a more elegant structure to things. I will look into patch 12 especially to see how you use the infrastructure. Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
