Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 3E2EEC433F5 for ; Sat, 11 Dec 2021 15:33:11 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231341AbhLKPbJ (ORCPT ); Sat, 11 Dec 2021 10:31:09 -0500 Received: from 49-237-179-185.static.tentacle.fi ([185.179.237.49]:48268 "EHLO bitmer.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229836AbhLKPbI (ORCPT ); Sat, 11 Dec 2021 10:31:08 -0500 Received: from 88-114-185-38.elisa-laajakaista.fi ([88.114.185.38] helo=[192.168.1.42]) by bitmer.com with esmtpsa (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.89) (envelope-from ) id 1mw4L1-0005wU-GF; Sat, 11 Dec 2021 17:31:03 +0200 Message-ID: Date: Sat, 11 Dec 2021 17:30:57 +0200 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101 Thunderbird/91.3.2 Subject: Re: [PATCH] ARM: dts: Fix timer regression for beagleboard revision c Content-Language: en-US To: Tony Lindgren , linux-omap@vger.kernel.org Cc: =?UTF-8?Q?Beno=c3=aet_Cousson?= , devicetree@vger.kernel.org, linux-kernel@vger.kernel.org, Daniel Lezcano , Thomas Gleixner , Rob Herring References: <20211125144834.52457-1-tony@atomide.com> From: Jarkko Nikula In-Reply-To: <20211125144834.52457-1-tony@atomide.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Tony On 11/25/21 16:48, Tony Lindgren wrote: > Commit e428e250fde6 ("ARM: dts: Configure system timers for omap3") > caused a timer regression for beagleboard revision c where the system > clockevent stops working if omap3isp module is unloaded. > > Turns out we still have beagleboard revisions a-b4 capacitor c70 quirks > applied that limit the usable timers for no good reason. This also affects > the power management as we use the system clock instead of the 32k clock > source. > > Let's fix the issue by adding a new omap3-beagle-ab4.dts for the old timer > quirks. This allows us to remove the timer quirks for later beagleboard > revisions. We also need to update the related timer quirk check for the > correct compatible property. > > Fixes: e428e250fde6 ("ARM: dts: Configure system timers for omap3") > Cc: linux-kernel@vger.kernel.org > Cc: Daniel Lezcano > Cc: Thomas Gleixner > Cc: Rob Herring > Reported-by: Jarkko Nikula > Signed-off-by: Tony Lindgren > --- > .../devicetree/bindings/arm/omap/omap.txt | 3 ++ > arch/arm/boot/dts/Makefile | 1 + > arch/arm/boot/dts/omap3-beagle-ab4.dts | 47 +++++++++++++++++++ > arch/arm/boot/dts/omap3-beagle.dts | 33 ------------- > drivers/clocksource/timer-ti-dm-systimer.c | 2 +- > 5 files changed, 52 insertions(+), 34 deletions(-) > create mode 100644 arch/arm/boot/dts/omap3-beagle-ab4.dts > I must have some error in my methodology since I cannot see the issue being fixed with your patch :-( Testing at commit 6f513529296f ("Merge tag 'for-5.16-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux") alias ma='nice make -j `getconf _NPROCESSORS_ONLN` ARCH=arm CROSS_COMPILE="ccache arm-linux-gnueabihf-"' make ARCH=arm omap2plus_defconfig ma cat arch/arm/boot/dts/omap3-beagle.dtb >>arch/arm/boot/zImage ma LOADADDR=0x80008000 uImage -> copy uImage & modules, boot # rmmod omap3_isp # sleep 1 -> This is the regression. Sleep is blocked until I hit keys over serial console rm vmlinux arch/arm/boot/zImage patch -p1 >arch/arm/boot/zImage ma LOADADDR=0x80008000 uImage -> copy uImage & modules to µSD, boot # rmmod omap3_isp # sleep 1 -> is still blocked until I hit keys over serial console When I compare the dmesg this is the only difference (along with "linux version" line) before user space starts. I.e. timer configuration is the same in both. [ 0.000000] OF: fdt: Machine model: TI OMAP3 BeagleBoard -> [ 0.000000] OF: fdt: Machine model: TI OMAP3 BeagleBoard A to B4 Jarkko