From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758695AbZC3XLT (ORCPT ); Mon, 30 Mar 2009 19:11:19 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753704AbZC3XLH (ORCPT ); Mon, 30 Mar 2009 19:11:07 -0400 Received: from out01.mta.xmission.com ([166.70.13.231]:55099 "EHLO out01.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752540AbZC3XLG (ORCPT ); Mon, 30 Mar 2009 19:11:06 -0400 To: Alexey Dobriyan Cc: viro@ZenIV.linux.org.uk, linux-kernel@vger.kernel.org Subject: Re: [PATCH] proc 1/6: implement support for automounts in task directories References: <20090330225326.GB19892@x200.localdomain> From: ebiederm@xmission.com (Eric W. Biederman) Date: Mon, 30 Mar 2009 16:10:59 -0700 In-Reply-To: <20090330225326.GB19892@x200.localdomain> (Alexey Dobriyan's message of "Tue\, 31 Mar 2009 02\:53\:26 +0400") Message-ID: User-Agent: Gnus/5.11 (Gnus v5.11) Emacs/22.2 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-XM-SPF: eid=;;;mid=;;;hst=in02.mta.xmission.com;;;ip=67.169.126.145;;;frm=ebiederm@xmission.com;;;spf=neutral X-SA-Exim-Connect-IP: 67.169.126.145 X-SA-Exim-Rcpt-To: adobriyan@gmail.com, linux-kernel@vger.kernel.org, viro@ZenIV.linux.org.uk X-SA-Exim-Mail-From: ebiederm@xmission.com X-SA-Exim-Version: 4.2.1 (built Thu, 25 Oct 2007 00:26:12 +0000) X-SA-Exim-Scanned: No (on in02.mta.xmission.com); Unknown failure Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Alexey Dobriyan writes: > I don't have a lot of opinion on this bug. I checked patchset to fix known > reproducer, timeout is obviously taken out of air and not nice. As for timeout. The logic is that mark_mounts_for_expiry is just an aging mechanism, and if something is actually in use it is not aged out. In the general case the call to proc_shrink_automounts in kernel/exit.c will clean up everything. It is possible but unlikely in practice that someone will pin a lot of process directories open then stop pinning them, and after that point nothing will remain open, so we close that hole. I set timeout huge because it should never be needed in real life. That said I'm up for taking this as a starting point and generalizing the mechanisms in fs/namespace.c. So I don't need a timeout at all, but that is an orthogonal discussion. Eric