From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751843Ab2LTWPG (ORCPT ); Thu, 20 Dec 2012 17:15:06 -0500 Received: from hqemgate03.nvidia.com ([216.228.121.140]:16437 "EHLO hqemgate03.nvidia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750983Ab2LTWPA (ORCPT ); Thu, 20 Dec 2012 17:15:00 -0500 X-PGP-Universal: processed; by hqnvupgp05.nvidia.com on Thu, 20 Dec 2012 14:13:51 -0800 Date: Fri, 21 Dec 2012 00:13:48 +0200 From: Peter De Schrijver To: Marc Zyngier CC: Hiroshi Doyu , "linux-tegra@vger.kernel.org" , "andrew@lunn.ch" , "linux@arm.linux.org.uk" , "jason@lakedaemon.net" , "johnstul@us.ibm.com" , "devicetree-discuss@lists.ozlabs.org" , "linux-doc@vger.kernel.org" , "linux-kernel@vger.kernel.org" , "rob.herring@calxeda.com" , "tglx@linutronix.de" , "linux-arm-kernel@lists.infradead.org" Subject: Re: [PATCH 5/9] clocksource: tegra: Enable ARM arch_timer with TSC Message-ID: <20121220221348.GB19766@tbergstrom-lnx.Nvidia.com> References: <50D305A6.2080904@arm.com> <20121220125507.GB6819@tbergstrom-lnx.Nvidia.com> <50D31365.9000009@arm.com> <20121220.164230.292625215885249791.hdoyu@nvidia.com> <50D34630.9020708@arm.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Disposition: inline In-Reply-To: <50D34630.9020708@arm.com> X-NVConfidentiality: public User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > > > > "tegra114.dtsi" has the folloiwng "tsc" entry. So can we consider that > > if dts has this entry, CNTFRQ is not set, which implies it's in secure > > mode. kernel should set it up by itself? Otherwise, skip this setup > > and use it. For example: > > > > tsc { > > compatible = "nvidia,tegra114-tsc"; > > reg = <0x700f0000 0x20000>; > > + setup-cntfrq; > > }; > > > > Is this what you explained in the above? > > At least, kernel can survive without bootloader/firmware support, ATM. > > No. The DT should only describe the hardware, and not something that is > Linux specific. > > Just use the "clock-frequency" attribute in the timer arch-timer node, > and get rid of this CNTFRQ setting. The driver already knows how to deal > with this situation if this attribute is set. The frequency is probed at boottime by the kernel though. It will not necessarily be the same for every board. Cheers, Peter.