Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1162366Ab3DERyj (ORCPT ); Fri, 5 Apr 2013 13:54:39 -0400 Received: from mail-vc0-f169.google.com ([209.85.220.169]:62194 "EHLO mail-vc0-f169.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1162346Ab3DERyh (ORCPT ); Fri, 5 Apr 2013 13:54:37 -0400 MIME-Version: 1.0 In-Reply-To: References: From: Lucas De Marchi Date: Fri, 5 Apr 2013 14:54:16 -0300 Message-ID: Subject: Re: control on cpu/core To: christophe malvasio Cc: linux-modules , LKML Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1300 Lines: 30 On Fri, Apr 5, 2013 at 1:30 PM, christophe malvasio wrote: > hi all > i want to give the control of one core to a module > that is the kernel have to not schedule/interrupt execution on this > core and also my module have to be the only thing to send execution on > this core > i'm not sure if i can make it only from my module or if i must hack > the kernel too ? > fake core unplug or more coding ? > give me your opinion ;) You can isolate a core during boot so no process is scheduled to that core. See "isolcpus=" in Documentation/kernel-parameters.txt. I don't know what your module is doing so it might not be possible in all scenarios. The only way I see is if your module created a kernel thread. Then it can be moved to the isolated core by setting the cpu mask. See for example how kworker, watchdog, migration do. I'm CCing LKML to broaden your audience. linux-modules@vger.kernel.org is really low traffic and mostly intended for the module loading/removal process, not really about writing modules. Lucas De Marchi -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/