From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S967743AbeEYSfo (ORCPT ); Fri, 25 May 2018 14:35:44 -0400 Received: from mail-pg0-f67.google.com ([74.125.83.67]:39479 "EHLO mail-pg0-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S967679AbeEYSfl (ORCPT ); Fri, 25 May 2018 14:35:41 -0400 X-Google-Smtp-Source: AB8JxZo4Lk04o4aDYazoozFUCwfJnPBhEPXMwxbx/vEF7esbmrtaGsQaftfLnECbVMNvD01muyzdBg== Date: Fri, 25 May 2018 11:35:38 -0700 From: Bjorn Andersson To: Vivek Gautam Cc: Andy Gross , Manu Gautam , linux-arm-msm , linux-soc@vger.kernel.org, devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, open list Subject: Re: [PATCH] arm64: dts: qcom: msm8996: Use UFS_GDSC for UFS Message-ID: <20180525183538.GH14924@minitux> References: <20180524223122.12601-1-bjorn.andersson@linaro.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.9.2 (2017-12-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri 25 May 05:51 PDT 2018, Vivek Gautam wrote: > Hi Bjorn, > > On Fri, May 25, 2018 at 4:01 AM, Bjorn Andersson > wrote: > > The UFS host controller occationally (20%) fails to enable > > gcc_ufs_axi_clk because the UFS GDSC is not enabled. In most cases it's > > enabled through the UFS phy driver, but to make sure it's enabled let's > > enable it directly from the UFS host controller directly as well. > > > > Signed-off-by: Bjorn Andersson > > --- > > arch/arm64/boot/dts/qcom/msm8996.dtsi | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/arch/arm64/boot/dts/qcom/msm8996.dtsi b/arch/arm64/boot/dts/qcom/msm8996.dtsi > > index 380e14591686..03c7904bda14 100644 > > --- a/arch/arm64/boot/dts/qcom/msm8996.dtsi > > +++ b/arch/arm64/boot/dts/qcom/msm8996.dtsi > > @@ -674,6 +674,8 @@ > > vccq-max-microamp = <450000>; > > vccq2-max-microamp = <450000>; > > > > + power-domains = <&gcc UFS_GDSC>; > > + > > We shouldn't need power-domain with the phy. UFS_GDSC should > be attached to the controller, as the phy is powered up only after > the controller is power-up, and during collapse too, we turn off > the phy first. Afaict you're right, it should only be needed by the resources for the HCD. > Can you try testing keeping UFS_GDSC only with ufs controller and > remove it from the ufs-phy node? We are doing same on the 4.14 release > branch too for db820. I test booted this 10 times, with 100% success :) > I apologize to have missed this in your patch for ufs-related dt nodes. > Can we please fix this now? I'll send a v2, that moves the power-domain to the HCD, rather than just adding it there too. Thanks for the quick review! Regards, Bjorn