(Sorry for the top posting) Linus, Auto request GPIO as an input line when used as IRQ line for DT case has been broken and we did discuss the issue on couple of threads. Javier's intermediate approach with $subject series was addressing this long standing issue. But while fixing that case, unfortunately breaks another case as highlighted by Alexander. I think the default OMAP DT files will continue to work with these patches applied and mostly doesn't break anything in default configuration. Ofcourse with the DT modification as done by Alexander will expose the issue. So either case we will something broken. I really wanted to have the auto request GPIO supported when used as IRQ line but surely not at expense of breaking the client drivers. So I am fine with whatever you decide. Regards, Santosh ________________________________________ From: Linus Walleij [linus.walleij@xxxxxxxxxx] Sent: Sunday, July 28, 2013 10:11 AM To: Alexander Holler Cc: ext Tony Lindgren; Grant Likely; Shilimkar, Santosh; Kevin Hilman; Javier Martinez Canillas; Jon Hunter; Jean-Christophe PLAGNIOL-VILLARD; Enric Balletbo Serra; Linux-OMAP; Florian Vaussard; Aaro Koskinen; Krishnamoorthy, Balaji T Subject: Re: [PATCH v4 1/2] gpio/omap: don't create an IRQ mapping for every GPIO on DT On Sun, Jul 28, 2013 at 2:59 PM, Alexander Holler <holler@xxxxxxxxxxxxx> wrote: > Am 28.07.2013 13:14, schrieb Linus Walleij: >> What shall we do with this mess now? (...) > Hmm, maybe something which adds a mapping for a IRQ when gpio_to_irq() > is called would help. I think this bug is pointing back to my initial remark that we need to flag, in the DT node for each GPIOchip that provides "auto-IRQs" exactly which IRQ lines that will be used as simple IRQ lines. This is what I called "GPIO hogs" a while back. If this had been done, gpio_request_one() would *only* have been called on the GPIOs marked as being used for IRQs simply. And we wouldn't have had this bug. I am leaning towards reverting the whole shebang right now, and using that as an argument to hammer in my opinion that we need to do it my way unless something incredibly clever comes up. Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
