From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751431Ab2LJQs4 (ORCPT ); Mon, 10 Dec 2012 11:48:56 -0500 Received: from mx1.redhat.com ([209.132.183.28]:11491 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750818Ab2LJQsz (ORCPT ); Mon, 10 Dec 2012 11:48:55 -0500 Date: Mon, 10 Dec 2012 11:48:13 -0500 From: Don Zickus To: Seiji Aguchi Cc: "Luck, Tony" , "linux-kernel@vger.kernel.org" , "cbouatmailru@gmail.com" , "ccross@android.com" , "keescook@chromium.org" , "dle-develop@lists.sourceforge.net" , Satoru Moriya Subject: Re: [RFC][PATCH] pstore: Skip spinlock when just one cpu is online Message-ID: <20121210164813.GA53431@redhat.com> References: <3908561D78D1C84285E8C5FCA982C28F1C9658B1@ORSMSX108.amr.corp.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: 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 On Fri, Dec 07, 2012 at 11:43:03PM +0000, Seiji Aguchi wrote: > > Can all these things really happen (did you run into this problem on a real system?). Or is this just a theoretical problem. Ugly (but > > practical) hacks might be OK to solve real problems. > > It is a theoretical problem right now. > But it is a timing issue and there is a possibility to happen actually. > > > But do we really want them to fix problems that actually never happen? > > If we find a problem (even if it is theoretical), we can't say "It actually never happen.". > > I have some reasons to submit this patch before reproducing actually. > > 1) > It is too late if we fix a problem after it actually happened in case where we apply Linux, including pstore, > to mission critical systems, because the failure of those systems has a great impact on a whole society. > Customers in this area ask us to fix a problem as soon as possible. > On the other hand, this kind of timing issue is hard to reproduce. > So, our support service engineers often work all night to reproduce it. > It is a nightmare for us. > > If we can fix it with a small patch in adance, it is really helpful for us. As I said in my email I just sent, it may not help you without testing it. As there are probably other problems in that un-tested theoretical scenario. > > 2) > In the long term, I plan to add a kmsg_dump to a kexec path because kdump may fail in the real world. > In that case, we need another troubleshooting material like pstore to detect a root cause of failure. But you are assuming that kmsg_dump is perfect and it isn't, in which case by putting kmsg_dump in the kdump path, you actually may be blocking kdump from working. That is the biggest hold up for those guys from including it I believe. Cheers, Don