From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [PATCH] ARCH: Fix 32-bit x86 MSI-X allocation leakage Date: Tue, 29 Apr 2008 17:14:03 +0200 Message-ID: <20080429151403.GA19410@elte.hu> References: <20080426005850.7098.77479.stgit@scrappy.jf.intel.com> <20080429102107.GH23198@elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: torvalds@linux-foundation.org, linux-kernel@vger.kernel.org, netdev@vger.kernel.org, Rusty Russell To: "Waskiewicz Jr, Peter P" Return-path: Received: from mx3.mail.elte.hu ([157.181.1.138]:36895 "EHLO mx3.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751565AbYD2POX (ORCPT ); Tue, 29 Apr 2008 11:14:23 -0400 Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: * Waskiewicz Jr, Peter P wrote: > > * PJ Waskiewicz wrote: > > > > > This bug was introduced in the 2.6.24 i386/x86_64 tree merge, [...] > > > > that's wrong, that code was there in v2.6.23 too - and it wasnt > > touched in the unification. Why do you think it was introduced in > > the i386/x86_64 tree merge? > > 2.6.23.17 didn't have the used_vectors bitmap. The first time I could > find it used was in 2.6.24. Where was the code existing in 2.6.23? > The addition may have not been part of the merge, but it was > definitely introduced in the same timeframe, meaning the 2.6.24 > release timeframe. that change came in via lguest (commit dbeb2be2), not the 32-bit/64-bit x86 unification. Lguest is a new feature in 2.6.24 and this side-effect of it was not known until you fixed it - thanks for that! So i was just splitting hairs over your statement that it was due to the unification :-) Ingo