Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1761581Ab3IEMCL (ORCPT ); Thu, 5 Sep 2013 08:02:11 -0400 Received: from mail-ea0-f174.google.com ([209.85.215.174]:38824 "EHLO mail-ea0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754244Ab3IEMCI (ORCPT ); Thu, 5 Sep 2013 08:02:08 -0400 Date: Thu, 5 Sep 2013 14:02:30 +0200 From: Miklos Szeredi To: Al Viro Cc: Linux-Fsdevel , Kernel Mailing List , "mszeredi@suse.cz" , David Howells , Steven Whitehouse , Trond Myklebust , Greg Kroah-Hartman Subject: Re: [PATCH 04/11] vfs: check unlinked ancestors before mount Message-ID: <20130905120230.GA21170@tucsk.piliscsaba.szeredi.hu> References: <1378374284-1484-1-git-send-email-miklos@szeredi.hu> <1378374284-1484-5-git-send-email-miklos@szeredi.hu> <20130905111852.GP13318@ZenIV.linux.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4077 Lines: 138 On Thu, Sep 05, 2013 at 01:32:10PM +0200, Miklos Szeredi wrote: > On Thu, Sep 5, 2013 at 1:18 PM, Al Viro wrote: > > Something's really odd with locking here. You are take d_lock, do one > > check, set flag, drop d_lock, grab rename_lock, do another check (taking > > and dropping d_lock in process), and, in case that check fails, grab > > d_lock again to clear the flag. > > > > At the very least it's a massive overkill. Just grab rename_lock, then > > d_lock, then do the damn check and set the flag only on success. Moreover, > > with rename_lock held, do you need d_lock on ancestors to mess with in > > has_unlinked_ancestor()? > > Yes, we need hard exclusion for the __d_drop() part. rename_lock can > provide one if we always take it for write in > check_submounts_and_drop(). But if we only take it for read then > that's not enough. > > And we do in fact also need DCACHE_MOUNTED set *before* checking > ancestors. Otherwise check_submounts_and_drop() could succeed and > has_unlinked_ancestor() return false, resulting in a dropped dentry > and a mount below it. Though this is mostly theoretical at this > point. Maybe something like this. Has less ugly locking. Untested. Thanks, Miklos --- fs/dcache.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++++ fs/internal.h | 1 + fs/namespace.c | 11 +++++------ 3 files changed, 55 insertions(+), 6 deletions(-) --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1159,6 +1159,55 @@ int have_submounts(struct dentry *parent } EXPORT_SYMBOL(have_submounts); +static bool __has_unlinked_ancestor(struct dentry *dentry) +{ + struct dentry *this; + + for (this = dentry; !IS_ROOT(this); this = this->d_parent) { + int is_unhashed; + + /* Need exclusion wrt. check_submounts_and_drop() */ + spin_lock(&this->d_lock); + is_unhashed = d_unhashed(this); + spin_unlock(&this->d_lock); + + if (is_unhashed) + return true; + } + return false; +} + +/* + * Called by mount code to check if the mountpoint is reachable (e.g. NFS can + * unhash a directory dentry and then the complete subtree can become + * unreachable). + */ +int d_set_mounted(struct dentry *dentry) +{ + int ret = 0; + + write_seqlock(&rename_lock); + spin_lock(&dentry->d_lock); + dentry->d_flags |= DCACHE_MOUNTED; + if (!IS_ROOT(dentry)) { + ret = -ENOENT; + if (d_unhashed(dentry)) { + dentry->d_flags &= ~DCACHE_MOUNTED; + goto out; + } + spin_unlock(&dentry->d_lock); + if (__has_unlinked_ancestor(dentry->d_parent)) { + spin_lock(&dentry->d_lock); + dentry->d_flags &= ~DCACHE_MOUNTED; + spin_unlock(&dentry->d_lock); + } + ret = 0; + } +out: + write_sequnlock(&rename_lock); + return ret; +} + /* * Search the dentry child list of the specified parent, * and move any unused dentries to the end of the unused --- a/fs/internal.h +++ b/fs/internal.h @@ -126,6 +126,7 @@ extern int invalidate_inodes(struct supe * dcache.c */ extern struct dentry *__d_alloc(struct super_block *, const struct qstr *); +extern int d_set_mounted(struct dentry *dentry); /* * read_write.c --- a/fs/namespace.c +++ b/fs/namespace.c @@ -611,6 +611,7 @@ static struct mountpoint *new_mountpoint { struct list_head *chain = mountpoint_hashtable + hash(NULL, dentry); struct mountpoint *mp; + int ret; list_for_each_entry(mp, chain, m_hash) { if (mp->m_dentry == dentry) { @@ -626,14 +627,12 @@ static struct mountpoint *new_mountpoint if (!mp) return ERR_PTR(-ENOMEM); - spin_lock(&dentry->d_lock); - if (d_unlinked(dentry)) { - spin_unlock(&dentry->d_lock); + ret = d_set_mounted(dentry); + if (ret) { kfree(mp); - return ERR_PTR(-ENOENT); + return ERR_PTR(ret); } - dentry->d_flags |= DCACHE_MOUNTED; - spin_unlock(&dentry->d_lock); + mp->m_dentry = dentry; mp->m_count = 1; list_add(&mp->m_hash, chain); -- 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/