From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: Early kernel messages are overflowing the static log buffer Date: Wed, 22 Dec 2010 22:21:12 +0100 Message-ID: <20101222212112.GC3139@elte.hu> References: <4D06B317.2090608@sgi.com> <20101222130946.GC13412@elte.hu> <20101222131719.GA8061@htj.dyndns.org> <4D1249ED.9040008@sgi.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mx2.mail.elte.hu ([157.181.151.9]:53340 "EHLO mx2.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751841Ab0LVVV0 (ORCPT ); Wed, 22 Dec 2010 16:21:26 -0500 Content-Disposition: inline In-Reply-To: <4D1249ED.9040008@sgi.com> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Mike Travis Cc: Tejun Heo , "H. Peter Anvin" , Andrew Morton , David Rientjes , Len Brown , Jack Steiner , Lori Gilbertson , "linux-acpi@vger.kernel.org" , LKML * Mike Travis wrote: > Thanks. > > I have a patchset almost ready for submission. We have a giant > system (6 racks with some huge amount of nodes) for a few more > days, and the buffer still overflowed with my changes. So I'm > making one more tweak to cut down the amount of characters generated > without losing information. Also feel free to make judgement calls to cut information - people _think_ they need all that boot time noise but it's rarely useful. Memory layout is - but a lot of other details are not. So dont try to preserve things at any cost. > This doesn't really help the log buffer overflow problem as KERN_DEBUG > messages still end up in the buffer. The pr_debug() macro does compile > out the messages if the system does not have KERNEL debug set, though > distros also have this set as a default. Make the whole printout dependent on a percpu-debug boot parameter? (and default the switch to off?) Thanks, Ingo