From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754665Ab2FMVvY (ORCPT ); Wed, 13 Jun 2012 17:51:24 -0400 Received: from mga02.intel.com ([134.134.136.20]:15166 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751373Ab2FMVvX (ORCPT ); Wed, 13 Jun 2012 17:51:23 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.67,351,1309762800"; d="scan'208";a="157291395" Date: Wed, 13 Jun 2012 14:51:22 -0700 From: Andi Kleen To: Peter Zijlstra Cc: Andi Kleen , x86@kernel.org, linux-kernel@vger.kernel.org, eranian@google.com Subject: Re: [PATCH 2/4] x86: Track minimum microcode revision globally v2 Message-ID: <20120613215122.GF32604@tassilo.jf.intel.com> References: <1339618842-26636-1-git-send-email-andi@firstfloor.org> <1339618842-26636-3-git-send-email-andi@firstfloor.org> <1339623321.8980.66.camel@twins> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1339623321.8980.66.camel@twins> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > This would really much better live in common code where it can be > preemptible in some sites. There really is no need for this to be > non-preemptible. Okay I can look. But you're really optimizing the wrong things here. Even on a 4096 CPU system such a loop is miniscule compared to the actual cost of the microcode update, which stops the complete CPU for quite some time. Basically microcode updates and low latency are incompatible Don't do it when it hurts. Unfortunately i'm changing all the callers two patches further, so it'll be messy too. -Andi -- ak@linux.intel.com -- Speaking for myself only