From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752404AbbLCVkW (ORCPT ); Thu, 3 Dec 2015 16:40:22 -0500 Received: from muru.com ([72.249.23.125]:50640 "EHLO muru.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751177AbbLCVkU (ORCPT ); Thu, 3 Dec 2015 16:40:20 -0500 Date: Thu, 3 Dec 2015 13:40:14 -0800 From: Tony Lindgren To: Sudeep Holla Cc: Linus Walleij , Grygorii Strashko , "linux-gpio@vger.kernel.org" , "linux-kernel@vger.kernel.org" , linux-omap@vger.kernel.org Subject: Re: [PATCH 2/2] pinctrl: single: remove misuse of IRQF_NO_SUSPEND flag Message-ID: <20151203214013.GB23396@atomide.com> References: <1448644860-29323-1-git-send-email-sudeep.holla@arm.com> <1448644860-29323-2-git-send-email-sudeep.holla@arm.com> <20151203181337.GV23396@atomide.com> <566090FC.1020502@arm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <566090FC.1020502@arm.com> User-Agent: Mutt/1.5.24 (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Sudeep Holla [151203 11:00]: > On 03/12/15 18:13, Tony Lindgren wrote: > >At least on omaps, this controller is always powered and we never want to > >suspend it as it handles wake-up events for all the IO pins. And that > >usecase sounds exactly like what you're describing above. > > > > Understood, but I assume this is a generic driver that can be used by > any pinmux. Right no question about that, but we need to keep things working. I just pasted the output to this thread what happens coming back up from suspend. > >I don't quite follow what your suggested alternative for an interrupt > >controller is? > > Why can't we use enable_irq_wake even for parent/interrupt controller as > they can be considered as parent wakeup irq. I agree the interrupt > controller may not be powered down, but still it's part of wakeup and > the irq core needs to identify that. By just marking IRQF_NO_SUSPEND, > you are saying that you can handle interrupt in the suspend path but not > informing that it's a wakeup interrupt. > > With this change, the wakeup handler (including the parent handler) is > called when it's safe as the irq core maintains the state machine. Maybe paste a suggested patch and I can try it. I guess you mean call that from pinctrl-single.c. > >At least we need to have the alternative patched in with this chage before > >just removing IRQF_NO_SUSPEND. > > > > I have added irq_set_irq_wake(pcs_soc->irq, state) in pcs_irq_set_wake > which ensures it's marked for wakeup. Hmm well see the error I pasted in this thread, maybe that provides more clues. Regards, Tony