From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx3-rdu2.redhat.com ([66.187.233.73]:59354 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752017AbeENUam (ORCPT ); Mon, 14 May 2018 16:30:42 -0400 Date: Mon, 14 May 2018 16:30:40 -0400 From: Vivek Goyal To: Amir Goldstein Cc: Miklos Szeredi , Al Viro , linux-unionfs@vger.kernel.org, linux-fsdevel@vger.kernel.org Subject: Re: [PATCH v2 1/3] ovl: use d_instantiate_new() to instantiate a new dentry Message-ID: <20180514203040.GA6717@redhat.com> References: <1526116631-24186-1-git-send-email-amir73il@gmail.com> <1526116631-24186-2-git-send-email-amir73il@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1526116631-24186-2-git-send-email-amir73il@gmail.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Sat, May 12, 2018 at 12:17:09PM +0300, Amir Goldstein wrote: > Currently, there is a small window where ovl_obtain_alias() can > race with ovl_instantiate() and create two aliases for an overlay > directory inode, see Al's explanation in this post: > https://marc.info/?l=linux-fsdevel&m=152599914515224&w=2 > > This patch fixes the race, by using the d_instantiate_new() helper. > > Another logic change by this patch is that if there is an > inconcsistency and a new created upper inode apears to already > exist in icache (hashed by the same real upper inode), we will > export this error to user instead of silently not hashing the new > inode. > > Backporting only makes sense for v4.16 where NFS export was introduced. Hi Amir, So for ovlerlay, this race does not exist for directories as you fall back to lookup path and don't call ovl_obtain_alias(). Does that mean it exists for non-dir only and that's what you are trying to fix. I would be nice to add few more lines describing exact race. It makes it easier and one does not have to find another email thread and dig out a specific email to figure out what was the race exactly and how does apply to overlayfs. Thanks Vivek > > Cc: Al Viro > Cc: #v4.16 > Signed-off-by: Amir Goldstein > --- > fs/overlayfs/dir.c | 27 +++++++++++++++++---------- > fs/overlayfs/inode.c | 6 ++++++ > fs/overlayfs/overlayfs.h | 1 + > 3 files changed, 24 insertions(+), 10 deletions(-) > > diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c > index 47dc980e8b33..c41825a5ed5f 100644 > --- a/fs/overlayfs/dir.c > +++ b/fs/overlayfs/dir.c > @@ -183,23 +183,30 @@ static int ovl_set_opaque(struct dentry *dentry, struct dentry *upperdentry) > } > > /* Common operations required to be done after creation of file on upper */ > -static void ovl_instantiate(struct dentry *dentry, struct inode *inode, > - struct dentry *newdentry, bool hardlink) > +static int ovl_instantiate(struct dentry *dentry, struct inode *inode, > + struct dentry *newdentry, bool hardlink) > { > ovl_dir_modified(dentry->d_parent, false); > - ovl_copyattr(d_inode(newdentry), inode); > ovl_dentry_set_upper_alias(dentry); > if (!hardlink) { > - ovl_inode_update(inode, newdentry); > + int err; > + > + ovl_inode_init(inode, newdentry, NULL); > + err = ovl_insert_inode_locked(inode, d_inode(newdentry)); > + if (err) > + return err; > + > + d_instantiate_new(dentry, inode); > } else { > WARN_ON(ovl_inode_real(inode) != d_inode(newdentry)); > dput(newdentry); > inc_nlink(inode); > - } > - d_instantiate(dentry, inode); > - /* Force lookup of new upper hardlink to find its lower */ > - if (hardlink) > + d_instantiate(dentry, inode); > + /* Force lookup of new upper hardlink to find its lower */ > d_drop(dentry); > + } > + > + return 0; > } > > static bool ovl_type_merge(struct dentry *dentry) > @@ -238,7 +245,7 @@ static int ovl_create_upper(struct dentry *dentry, struct inode *inode, > ovl_set_opaque(dentry, newdentry); > } > > - ovl_instantiate(dentry, inode, newdentry, !!hardlink); > + err = ovl_instantiate(dentry, inode, newdentry, !!hardlink); > newdentry = NULL; > out_dput: > dput(newdentry); > @@ -439,7 +446,7 @@ static int ovl_create_over_whiteout(struct dentry *dentry, struct inode *inode, > if (err) > goto out_cleanup; > } > - ovl_instantiate(dentry, inode, newdentry, !!hardlink); > + err = ovl_instantiate(dentry, inode, newdentry, !!hardlink); > newdentry = NULL; > out_dput2: > dput(upper); > diff --git a/fs/overlayfs/inode.c b/fs/overlayfs/inode.c > index 7abcf96e94fc..70c966b1bb5a 100644 > --- a/fs/overlayfs/inode.c > +++ b/fs/overlayfs/inode.c > @@ -741,6 +741,12 @@ static bool ovl_verify_inode(struct inode *inode, struct dentry *lowerdentry, > return true; > } > > +int ovl_insert_inode_locked(struct inode *inode, struct inode *realinode) > +{ > + return insert_inode_locked4(inode, (unsigned long) realinode, > + ovl_inode_test, realinode); > +} > + > struct inode *ovl_lookup_inode(struct super_block *sb, struct dentry *real, > bool is_upper) > { > diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h > index caaa47cea2aa..dd320be86600 100644 > --- a/fs/overlayfs/overlayfs.h > +++ b/fs/overlayfs/overlayfs.h > @@ -343,6 +343,7 @@ int ovl_update_time(struct inode *inode, struct timespec *ts, int flags); > bool ovl_is_private_xattr(const char *name); > > struct inode *ovl_new_inode(struct super_block *sb, umode_t mode, dev_t rdev); > +int ovl_insert_inode_locked(struct inode *inode, struct inode *realinode); > struct inode *ovl_lookup_inode(struct super_block *sb, struct dentry *real, > bool is_upper); > struct inode *ovl_get_inode(struct super_block *sb, struct dentry *upperdentry, > -- > 2.7.4 > > -- > To unsubscribe from this list: send the line "unsubscribe linux-unionfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html