From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754479AbcIQUKP (ORCPT ); Sat, 17 Sep 2016 16:10:15 -0400 Received: from mx1.redhat.com ([209.132.183.28]:54960 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751982AbcIQUKJ (ORCPT ); Sat, 17 Sep 2016 16:10:09 -0400 Date: Sat, 17 Sep 2016 22:10:02 +0200 From: Mateusz Guzik To: Ian Kent Cc: Andrew Morton , autofs mailing list , Kernel Mailing List , Al Viro , linux-fsdevel , Omar Sandoval , "Eric W. Biederman" Subject: Re: [PATCH 3/4] autofs - make mountpoint checks namespace aware Message-ID: <20160917201000.omswgttgyzcu7jt6@mguzik> References: <20160914061434.24714.490.stgit@pluto.themaw.net> <20160914061445.24714.68331.stgit@pluto.themaw.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20160914061445.24714.68331.stgit@pluto.themaw.net> User-Agent: Mutt/1.6.0.1 (2016-04-01) X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.38]); Sat, 17 Sep 2016 20:10:08 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Sep 14, 2016 at 02:14:45PM +0800, Ian Kent wrote: > If an automount mount is clone(2)ed into a file system that is > propagation private, when it later expires in the originating > namespace subsequent calls to autofs ->d_automount() for that > dentry in the original namespace will return ELOOP until the > mount is manually umounted in the cloned namespace. > > In the same way, if an autofs mount is triggered by automount(8) > running within a container the dentry will be seen as mounted in > the root init namespace and calls to ->d_automount() in that namespace > will return ELOOP until the mount is umounted within the container. > > Also, have_submounts() can return an incorect result when a mount > exists in a namespace other than the one being checked. > > @@ -460,7 +460,7 @@ static int autofs4_d_manage(struct dentry *dentry, bool rcu_walk) > > if (ino->flags & AUTOFS_INF_WANT_EXPIRE) > return 0; > - if (d_mountpoint(dentry)) > + if (is_local_mountpoint(dentry)) > return 0; > inode = d_inode_rcu(dentry); > if (inode && S_ISLNK(inode->i_mode)) This change is within RCU lookup. is_local_mountpoint may end up calling __is_local_mountpoint, which will optionally take the namespace_sem lock, resulting in a splat: #0: (&(&sbi->fs_lock)->rlock){+.+...}, at: [] autofs4_d_manage+0x202/0x290^M Preemption disabled at:[] autofs4_d_manage+0x202/0x290^M ^M CPU: 1 PID: 1307 Comm: iknowthis Not tainted 4.8.0-rc6-next-20160916dupa #448^M Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011^M ffff8800077378f8 ffffffff818eaffc 0000000000000001 0000000000000000^M ffff880007737930 ffffffff8110c870 ffff880007588048 ffffffff82483840^M 0000000000000015 0000000000000000 ffff880007588040 ffff880007737978^M Call Trace:^M [] dump_stack+0x85/0xc9^M [] ___might_sleep+0x1e0/0x2e0^M [] __might_sleep+0x71/0xe0^M [] ? preempt_count_sub+0x39/0x80^M [] down_read+0x2a/0xc0^M [] __is_local_mountpoint+0x66/0xe0^M [] autofs4_d_manage+0x263/0x290^M [] follow_managed+0x157/0x480^M [] lookup_fast+0x3ab/0x690^M [] ? trailing_symlink+0x370/0x370^M [] ? path_init+0x917/0xa10^M [] ? __mutex_init+0x77/0x80^M [] path_openat+0x2bc/0x13e0^M [] ? path_lookupat+0x1f0/0x1f0^M [] ? __asan_loadN+0xf/0x20^M [] ? pvclock_clocksource_read+0xd6/0x180^M [] ? kvm_clock_read+0x23/0x40^M [] do_filp_open+0x122/0x1c0^M [] ? preempt_count_sub+0x39/0x80^M [] ? may_open_dev+0x50/0x50^M [] ? preempt_count_sub.part.67+0x18/0x90^M [] ? preempt_count_sub+0x39/0x80^M [] ? _raw_spin_unlock+0x31/0x50^M [] ? __alloc_fd+0x141/0x2b0^M [] do_sys_open+0x17c/0x2c0^M [] ? filp_open+0x60/0x60^M [] ? trace_hardirqs_on_thunk+0x1a/0x1c^M [] SyS_open+0x1e/0x20^M [] entry_SYSCALL_64_fastpath+0x1f/0xc2^M [] ? trace_hardirqs_off_caller+0xc5/0x120^M I don't know this code. Perhaps it will be perfectly fine performance wise to just drop out of RCU lookup in this case. -- Mateusz Guzik