From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Greer Subject: Re: [PATCH 5/6] power: bq24190_charger: Check the interrupt status on resume Date: Mon, 16 Jan 2017 12:15:19 -0700 Message-ID: <20170116191519.GI27262@animalcreek.com> References: <20170112004154.31568-1-tony@atomide.com> <20170112004154.31568-6-tony@atomide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20170112004154.31568-6-tony@atomide.com> Sender: linux-pm-owner@vger.kernel.org To: Tony Lindgren Cc: Sebastian Reichel , linux-pm@vger.kernel.org, linux-omap@vger.kernel.org, Matt Ranostay , Liam Breck List-Id: linux-omap@vger.kernel.org On Wed, Jan 11, 2017 at 04:41:53PM -0800, Tony Lindgren wrote: > Some SoCs like omap3 can configure GPIO irqs to use Linux generic > dedicated wakeirq support. If the dedicated wakeirq is configured, > the SoC will use a always-on interrupt controller to produce wake-up > events. > > If bq24190 is configured for dedicated wakeirq, we need to check the > interrupt status on PM runtime resume. This is because the Linux > generic wakeirq will call pm_runtime_resume() on the device on a > wakeirq. And as the bq24190 interrupt is falling edge sensitive > and only active for 250 us, there will be no device interrupt seen > by the runtime SoC IRQ controller. > > Note that this can cause spurious interrupts on omap3 devices with > bq24190 connected to gpio banks 2 - 5 as there's a glitch on those > pins waking from off mode as listed in "Advisory 1.45". Devices > with this issue should not configure the optional wakeirq interrupt > in the dts file. > > Cc: Mark A. Greer > Cc: Matt Ranostay > Cc: Liam Breck > Signed-off-by: Tony Lindgren > --- Looks good except for what Liam pointed out so assuming that gets fixed: Acked-by: Mark Greer