From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752908AbYCKWJz (ORCPT ); Tue, 11 Mar 2008 18:09:55 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750925AbYCKWJs (ORCPT ); Tue, 11 Mar 2008 18:09:48 -0400 Received: from one.firstfloor.org ([213.235.205.2]:45419 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750874AbYCKWJr (ORCPT ); Tue, 11 Mar 2008 18:09:47 -0400 Date: Tue, 11 Mar 2008 23:11:44 +0100 From: Andi Kleen To: Thomas Gleixner Cc: Andi Kleen , Ingo Molnar , linux-kernel@vger.kernel.org, akpm@osdl.org Subject: Re: [PATCH REPOST for 2.6.25] Use an own random generator for pageattr-test.c Message-ID: <20080311221144.GC428@one.firstfloor.org> References: <20080311013014.GA28682@basil.nowhere.org> <20080311104531.GB18917@one.firstfloor.org> <20080311114832.GE18917@one.firstfloor.org> <20080311214925.GA428@one.firstfloor.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.1i Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > The only bug that code ever caught aside of tons of false positives It caught bugs when I originally worked on cpa. However that had still the revert code and admittedly several were in that area (I'm still suspecting at some point you'll have to readd that -- perhaps it'll be more useful for you then) However I had originally a longer runtime length too, but later decreased it when I sent out the patch (since i didn't want to submit another "rcu torture" which runs for hours...) Perhaps the result was a little too short too, that's possible. > Finding a bug, when it was exposed by a static pattern, is trivial, > but the challenge is to make such tests useful enough with random > patterns. And there are ways to do that, e.g. by making the debug > output informative enough to provide information about the problem in > detail instead of printing some useless info "a != b". Ok I admit I don't know how to do that effectively for cpa, but perhaps you do. -Andi