From mboxrd@z Thu Jan 1 00:00:00 1970 From: Carsten Emde Subject: Re: rt for newbie Date: Fri, 28 Oct 2011 18:50:04 +0200 Message-ID: <4EAADD3C.3090501@osadl.org> References: <4EA9D980.5080903@osadl.org> <4EAA5BC2.4020908@googlemail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: sashka@mail.ru, linux-rt-users@vger.kernel.org To: Dennis Borgmann Return-path: Received: from toro.web-alm.net ([62.245.132.31]:50066 "EHLO toro.web-alm.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753917Ab1J1RCI (ORCPT ); Fri, 28 Oct 2011 13:02:08 -0400 In-Reply-To: <4EAA5BC2.4020908@googlemail.com> Sender: linux-rt-users-owner@vger.kernel.org List-ID: Hi Dennis, >> [..] >> 2. Configure the kernel for RT >> # make menuconfig >> Processor type and features ---> >> [ ] Tickless System (Dynamic Ticks) >> Preemption Model (Fully Preemptible Kernel (RT)) ---> >> Kernel hacking ---> >> [ ] Check for stack overflows > I think, your small "howto" is a little confusing. You are describing > the kernel options needed for RT-capabilities, but there is no asterisk > within the square brackets in order to show, that they should be > enabled. The way it is written in your email, it seems like they should > be disabled. Or did you really mean to disable the "Tickless System" > option and the "Check for stack overflows" option? Thank you for pointing this out. My description may, indeed, be misleading. I simply wanted to show how the line of the configuration menu should look like, i.e. not crossed -> not enabled. Checking for stack overflows almost certainly leads to a longer worst-case latency and, while tickless operating is good for the financial budget, it may be bad for the real-time behavior. The latter applies to a configuration that allows the processor to enter a sleep state when idle and to a situation when rarely occurring external events trigger a real-time process so the processor has time to go to sleep. Thanks, -Carsten.