From: Ferenc Wagner Subject: Re: [PATCH v2 1/2] Input: gpio-keys - allow platform to specify exact irq flags Date: Mon, 23 Nov 2009 19:50:20 +0100 Message-ID: <87tywlcbyr.fsf@tac.ki.iif.hu> References: <20091120084054.GC3009@core.coreip.homeip.net> <8237c9a4afd1fa883120b5793fcd91a970d2ffb7.1258976669.git.ext-mika.1.westerberg@nokia.com> <87y6lxchwf.fsf@tac.ki.iif.hu> <20091123172451.GD17813@core.coreip.homeip.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from tac.ki.iif.hu ([193.6.222.43]:48095 "EHLO tac.ki.iif.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751128AbZKWSuQ (ORCPT ); Mon, 23 Nov 2009 13:50:16 -0500 In-Reply-To: <20091123172451.GD17813@core.coreip.homeip.net> (Dmitry Torokhov's message of "Mon, 23 Nov 2009 09:24:52 -0800") Sender: linux-input-owner@vger.kernel.org List-Id: linux-input@vger.kernel.org To: Dmitry Torokhov Cc: Mika Westerberg , linux-input@vger.kernel.org Dmitry Torokhov writes: > On Mon, Nov 23, 2009 at 05:42:08PM +0100, Ferenc Wagner wrote: >> Mika Westerberg writes: >> >>> Added new field to struct gpio_keys_button: irqflags which can >>> be used to specify exact irqflags the platform wants. If not specified, >>> it defaults to: IRQF_SHARED | IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING. >>> >>> --- a/drivers/input/keyboard/gpio_keys.c >>> +++ b/drivers/input/keyboard/gpio_keys.c >>> @@ -106,10 +107,14 @@ static int __devinit gpio_keys_setup_key(struct device *dev, >>> goto fail3; >>> } >>> >>> - error = request_irq(irq, gpio_keys_isr, >>> - IRQF_SHARED | >>> - IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, >>> - desc, bdata); >>> + if (button->irqflags == 0) { >>> + irqflags = IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING | >>> + IRQF_SHARED; >>> + } else { >>> + irqflags = button->irqflags; >>> + } >>> + >>> + error = request_irq(irq, gpio_keys_isr, irqflags, desc, bdata); >>> if (error) { >>> dev_err(dev, "Unable to claim irq %d; error %d\n", >>> irq, error); >> >> linux/interrupt.h says: >> >> When requesting an interrupt without specifying a IRQF_TRIGGER, the >> setting should be assumed to be "as already configured", which may >> be as per machine or firmware initialisation. >> >> So I'm not sure it's a good idea to exclude 0 from the possible irqflag >> values. > > Hmm... OTOH we can't just require the new irq flags to be always > specified, that would break all current users. Do we need a flag > indicating that the button will be using custom IRQ flags? Or let for example -1 mean 0, which is ugly, to say the least. Anyway, I'd much prefer an implementation which doesn't work against having multiple buttons on a single IRQ line, which is the case for example in arch/mips/bcm47xx/gpio.c. Then the IRQ could still be disabled if all its buttons are disabled, unless it's shared with another driver (and IRQF_SHARED would control this aspect sharing). -- Regards, Feri.