Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S269198AbUJFKzH (ORCPT ); Wed, 6 Oct 2004 06:55:07 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S269203AbUJFKzH (ORCPT ); Wed, 6 Oct 2004 06:55:07 -0400 Received: from bay-bridge.veritas.com ([143.127.3.10]:35741 "EHLO MTVMIME01.enterprise.veritas.com") by vger.kernel.org with ESMTP id S269198AbUJFKzC (ORCPT ); Wed, 6 Oct 2004 06:55:02 -0400 Date: Wed, 6 Oct 2004 11:54:47 +0100 (BST) From: Hugh Dickins X-X-Sender: hugh@localhost.localdomain To: Peter Zijlstra cc: Andrew Morton , Ingo Molnar , Pasi Savolainen , Subject: Re: 2.6.9-rc3-mm2 In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1190 Lines: 26 On Tue, 5 Oct 2004, Hugh Dickins wrote: > On Tue, 5 Oct 2004, Peter Zijlstra wrote: > > On Mon, 2004-10-04 at 19:13, Pasi Savolainen wrote: > > > make: *** wait: No child processes. Stop. > > Just a simple mee too! > > I've been suffering the occasional leftover zombie from multiple > kernel builds precisely since the preempt-smp.patch went in; been > hunting it unsuccessfully in spare moments, yesterday noticed that > bug, today realize it's probably what I've been hunting - I'm > about to start my own tests again, can't be sure until tomorrow. Yes, that _raw_read_trylock patch has fixed my zombie problem: since 2.6.9-rc2-mm1 my repeated kernel builds would hang after several hours, with a zombie (usually cc1) on one cpu and its waiting parent (usually gcc) on another, with the puzzle of why the parent had never got woken - because do_wait's read_lock(&tasklist_lock) went ahead even while exit_notify held the write lock. Hugh - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/