From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Boyd Subject: Re: [PATCH RESEND RFC 1/4] drivers: pinctrl: qcom: add wakeup capability to GPIO Date: Fri, 10 Aug 2018 08:06:30 -0700 Message-ID: <153391359077.37448.3410612264580038499@swboyd.mtv.corp.google.com> References: <20180801020021.9782-1-ilina@codeaurora.org> <20180801194538.GA6422@codeaurora.org> <86wot9wb9u.wl-marc.zyngier@arm.com> <20180802065104.GA27850@codeaurora.org> <86sh3xw7m9.wl-marc.zyngier@arm.com> <20180802125827.GB27850@codeaurora.org> <153370830708.220756.4595316550560511917@swboyd.mtv.corp.google.com> <20180808072632.21f076b6@why.wild-wind.fr.eu.org> <153383585322.220756.9422019201626837843@swboyd.mtv.corp.google.com> <86wosypsvr.wl-marc.zyngier@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <86wosypsvr.wl-marc.zyngier@arm.com> Sender: linux-kernel-owner@vger.kernel.org To: Marc Zyngier Cc: Lina Iyer , evgreen@chromium.org, linus.walleij@linaro.org, bjorn.andersson@linaro.org, rplsssn@codeaurora.org, linux-kernel@vger.kernel.org, linux-arm-msm@vger.kernel.org, rnayak@codeaurora.org, devicetree@vger.kernel.org List-Id: linux-arm-msm@vger.kernel.org Quoting Marc Zyngier (2018-08-10 00:45:12) > On Thu, 09 Aug 2018 18:30:53 +0100, > Stephen Boyd wrote: > > = > > Quoting Marc Zyngier (2018-08-07 23:26:32) > > > = > > > Level interrupts should be taken care of without doing anything, by t= he > > > very nature of being a level signal. > > = > > Right. I suspect we'll still need to configure the PDC to actually wake > > up on the level triggered signal though so PDC needs to be told to > > unmask the line. > = > Surely this can be done at suspend time with the PDC driver tracking > the interrupts that are configured as a wake-up source (although it > needs to track an interrupt that is logically connected to the TLMM, > which sucks). The PDC also needs to be configured for wakeups from deep CPU idle states where the GIC and TLMM are powered down. Lina, can you confirm this? Hooking system suspend in that case won't work. Is your hope that we can avoid using hierarchical irqdomains here entirely?