From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alex Lyashkov Subject: Re: [RFC] possible badness in prune_dcache() Date: Fri, 04 Apr 2008 18:28:46 +0300 Message-ID: <1207322926.12346.30.camel@bear.shadowland> References: <1207309200.12346.25.camel@bear.shadowland> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7BIT Cc: linux-fsdevel@vger.kernel.org, Andrew.Perepechko@Sun.COM, dhowells@redhat.com To: Miklos Szeredi Return-path: Received: from gmp-eb-inf-2.sun.com ([192.18.6.24]:49416 "EHLO gmp-eb-inf-2.sun.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755511AbYDDP2w (ORCPT ); Fri, 4 Apr 2008 11:28:52 -0400 Received: from fe-emea-09.sun.com (gmp-eb-lb-2-fe1.eu.sun.com [192.18.6.10]) by gmp-eb-inf-2.sun.com (8.13.7+Sun/8.12.9) with ESMTP id m34FSjXe019045 for ; Fri, 4 Apr 2008 15:28:46 GMT Received: from conversion-daemon.fe-emea-09.sun.com by fe-emea-09.sun.com (Sun Java System Messaging Server 6.2-8.04 (built Feb 28 2007)) id <0JYS00801YFGG400@fe-emea-09.sun.com> (original mail from Alexey.Lyashkov@Sun.COM) for linux-fsdevel@vger.kernel.org; Fri, 04 Apr 2008 16:28:45 +0100 (BST) In-reply-to: Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Thanks for answer, Yes this commit - fix issue with umount, but not resolve main question: why prune_dcache kill dentry with diffrent super block? this produce long loop with skip loop instead of early exit and move dentry from select_parent again to end of list. On Fri, 2008-04-04 at 14:42 +0200, Miklos Szeredi wrote: > > When investigation livelock in RHEL3, i found possible badness in > > prune_dcache, which exist in 2.6.24 also. > > Have you actually tested with 2.6.24? shrink_dcache_parent() isn't > called from umount anymore. I think this is the commit responsible > for fixing the issue: > > commit c636ebdb186bf37f98d3839f69293597723edb36 > Author: David Howells > Date: Wed Oct 11 01:22:19 2006 -0700 > > [PATCH] VFS: Destroy the dentries contributed by a superblock on unmounting > > Miklos >