From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755789Ab3H3Ljc (ORCPT ); Fri, 30 Aug 2013 07:39:32 -0400 Received: from caramon.arm.linux.org.uk ([78.32.30.218]:51229 "EHLO caramon.arm.linux.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754552Ab3H3Lja (ORCPT ); Fri, 30 Aug 2013 07:39:30 -0400 Date: Fri, 30 Aug 2013 12:39:18 +0100 From: Russell King - ARM Linux To: "Jon Medhurst (Tixy)" Cc: Sebastian Hesselbarth , Arnd Bergmann , linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, Pawel Moll Subject: Re: [PATCH RFC v2 14/16] ARM: vexpress: remove custom .init_time hook Message-ID: <20130830113918.GV6617@n2100.arm.linux.org.uk> References: <1376964271-22715-1-git-send-email-sebastian.hesselbarth@gmail.com> <1377638890-371-15-git-send-email-sebastian.hesselbarth@gmail.com> <201308291535.53005.arnd@arndb.de> <521F901B.3090305@gmail.com> <1377856951.3655.34.camel@linaro1.home> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1377856951.3655.34.camel@linaro1.home> User-Agent: Mutt/1.5.19 (2009-01-05) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Aug 30, 2013 at 11:02:31AM +0100, Jon Medhurst (Tixy) wrote: > After adding of_clk_init(NULL) to time_init() things boot OK for me with > this patch. However, do we know that sched_clock is never going to get > read before time_init() has actually started the clock it reads? Are we > making things more fragile? As I've been saying for years, sched_clock gets read for the first time in sched_init() -> init_idle() which occurs before time_init(). Whether that matters or not is debatable, but the first read should return zero, and it will return zero after initialization - so I think all we end up losing is some accounting of the time taken in the early kernel boot. If we wish to talk about correctness, then arguably it isn't correct. From mboxrd@z Thu Jan 1 00:00:00 1970 From: linux@arm.linux.org.uk (Russell King - ARM Linux) Date: Fri, 30 Aug 2013 12:39:18 +0100 Subject: [PATCH RFC v2 14/16] ARM: vexpress: remove custom .init_time hook In-Reply-To: <1377856951.3655.34.camel@linaro1.home> References: <1376964271-22715-1-git-send-email-sebastian.hesselbarth@gmail.com> <1377638890-371-15-git-send-email-sebastian.hesselbarth@gmail.com> <201308291535.53005.arnd@arndb.de> <521F901B.3090305@gmail.com> <1377856951.3655.34.camel@linaro1.home> Message-ID: <20130830113918.GV6617@n2100.arm.linux.org.uk> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Fri, Aug 30, 2013 at 11:02:31AM +0100, Jon Medhurst (Tixy) wrote: > After adding of_clk_init(NULL) to time_init() things boot OK for me with > this patch. However, do we know that sched_clock is never going to get > read before time_init() has actually started the clock it reads? Are we > making things more fragile? As I've been saying for years, sched_clock gets read for the first time in sched_init() -> init_idle() which occurs before time_init(). Whether that matters or not is debatable, but the first read should return zero, and it will return zero after initialization - so I think all we end up losing is some accounting of the time taken in the early kernel boot. If we wish to talk about correctness, then arguably it isn't correct.