[PATCH 2/2] NFS: Cache invalidation fixup

[PATCH 2/2] NFS: Cache invalidation fixup

[Trond Myklebust]

From: Trond Myklebust <Trond.Myklebust@netapp.com>

If someone has renamed a directory on the server, triggering the d_move
code in d_materialise_unique(), then we need to invalidate the
cached directory information in the source parent directory.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---

 fs/nfs/dir.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 2cfa414..fa71936 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -936,8 +936,14 @@ static struct dentry *nfs_lookup(struct 
 no_entry:
 	res = d_materialise_unique(dentry, inode);
 	if (res != NULL) {
+		struct dentry *parent;
 		if (IS_ERR(res))
 			goto out_unlock;
+		/* Was a directory renamed! */
+		parent = dget_parent(res);
+		if (!IS_ROOT(parent))
+			nfs_mark_for_revalidate(parent->d_inode);
+		dput(parent);
 		dentry = res;
 	}
 	nfs_renew_times(dentry);

[PATCH 1/2] VFS: Make d_materialise_unique() enforce directory uniqueness

[Trond Myklebust]

From: Trond Myklebust <Trond.Myklebust@netapp.com>

If the caller tries to instantiate a directory using an inode that already
has a dentry alias, then we attempt to rename the existing dentry instead
of instantiating a new one. Fail with an ELOOP error if the rename would
affect one of our parent directories.

This behaviour is needed in order to avoid issues such as

  http://bugzilla.kernel.org/show_bug.cgi?id=7178

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---

 fs/dcache.c  |  102 +++++++++++++++++++++++++++++++++++++---------------------
 fs/nfs/dir.c |    7 +++-
 2 files changed, 71 insertions(+), 38 deletions(-)

diff --git a/fs/dcache.c b/fs/dcache.c
index 2bac4ba..f700afd 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -1469,23 +1469,21 @@ static void switch_names(struct dentry *
  * deleted it.
  */
  
-/**
- * d_move - move a dentry
+/*
+ * d_move_locked - move a dentry
  * @dentry: entry to move
  * @target: new dentry
  *
  * Update the dcache to reflect the move of a file name. Negative
  * dcache entries should not be moved in this way.
  */
-
-void d_move(struct dentry * dentry, struct dentry * target)
+static void d_move_locked(struct dentry * dentry, struct dentry * target)
 {
 	struct hlist_head *list;
 
 	if (!dentry->d_inode)
 		printk(KERN_WARNING "VFS: moving negative dcache entry\n");
 
-	spin_lock(&dcache_lock);
 	write_seqlock(&rename_lock);
 	/*
 	 * XXXX: do we really need to take target->d_lock?
@@ -1536,10 +1534,39 @@ already_unhashed:
 	fsnotify_d_move(dentry);
 	spin_unlock(&dentry->d_lock);
 	write_sequnlock(&rename_lock);
+}
+
+/**
+ * d_move - move a dentry
+ * @dentry: entry to move
+ * @target: new dentry
+ *
+ * Update the dcache to reflect the move of a file name. Negative
+ * dcache entries should not be moved in this way.
+ */
+
+void d_move(struct dentry * dentry, struct dentry * target)
+{
+	spin_lock(&dcache_lock);
+	d_move_locked(dentry, target);
 	spin_unlock(&dcache_lock);
 }
 
 /*
+ * Reject any lookup loops due to malicious remote servers
+ */
+static int d_detect_loops(struct dentry *dentry, struct inode *inode)
+{
+	struct dentry *p;
+
+	for (p = dentry; p->d_parent != p; p = p->d_parent) {
+		if (p->d_parent->d_inode == inode)
+			return -ELOOP;
+	}
+	return 0;
+}
+
+/*
  * Prepare an anonymous dentry for life in the superblock's dentry tree as a
  * named dentry in place of the dentry to be replaced.
  */
@@ -1581,7 +1608,7 @@ static void __d_materialise_dentry(struc
  */
 struct dentry *d_materialise_unique(struct dentry *dentry, struct inode *inode)
 {
-	struct dentry *alias, *actual;
+	struct dentry *actual;
 
 	BUG_ON(!d_unhashed(dentry));
 
@@ -1593,26 +1620,36 @@ struct dentry *d_materialise_unique(stru
 		goto found_lock;
 	}
 
-	/* See if a disconnected directory already exists as an anonymous root
-	 * that we should splice into the tree instead */
-	if (S_ISDIR(inode->i_mode) && (alias = __d_find_alias(inode, 1))) {
-		spin_lock(&alias->d_lock);
-
-		/* Is this a mountpoint that we could splice into our tree? */
-		if (IS_ROOT(alias))
-			goto connect_mountpoint;
-
-		if (alias->d_name.len == dentry->d_name.len &&
-		    alias->d_parent == dentry->d_parent &&
-		    memcmp(alias->d_name.name,
-			   dentry->d_name.name,
-			   dentry->d_name.len) == 0)
-			goto replace_with_alias;
-
-		spin_unlock(&alias->d_lock);
-
-		/* Doh! Seem to be aliasing directories for some reason... */
-		dput(alias);
+	if (S_ISDIR(inode->i_mode)) {
+		struct dentry *alias;
+
+		actual = ERR_PTR(d_detect_loops(dentry, inode));
+		if (IS_ERR(actual))
+			goto out_unlock;
+		/* Does an aliased dentry already exist? */
+		alias = __d_find_alias(inode, 0);
+		if (alias) {
+			actual = alias;
+			/* Is this an anonymous mountpoint that we could splice
+			 * into our tree? */
+			if (IS_ROOT(alias)) {
+				spin_lock(&alias->d_lock);
+				__d_materialise_dentry(dentry, alias);
+				__d_drop(alias);
+				goto found;
+			}
+			/* Nope, but we must(!) avoid directory aliasing */
+			if (mutex_trylock(&inode->i_sb->s_vfs_rename_mutex)) {
+				d_move_locked(alias, dentry);
+				spin_unlock(&dcache_lock);
+				mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
+				goto out_nolock;
+			}
+			spin_unlock(&dcache_lock);
+			dput(alias);
+			actual = ERR_PTR(-EBUSY);
+			goto out_nolock;
+		}
 	}
 
 	/* Add a unique reference */
@@ -1627,8 +1664,9 @@ found_lock:
 found:
 	_d_rehash(actual);
 	spin_unlock(&actual->d_lock);
+out_unlock:
 	spin_unlock(&dcache_lock);
-
+out_nolock:
 	if (actual == dentry) {
 		security_d_instantiate(dentry, inode);
 		return NULL;
@@ -1637,16 +1675,6 @@ found:
 	iput(inode);
 	return actual;
 
-	/* Convert the anonymous/root alias into an ordinary dentry */
-connect_mountpoint:
-	__d_materialise_dentry(dentry, alias);
-
-	/* Replace the candidate dentry with the alias in the tree */
-replace_with_alias:
-	__d_drop(alias);
-	actual = alias;
-	goto found;
-
 shouldnt_be_hashed:
 	spin_unlock(&dcache_lock);
 	BUG();
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index c86a1ea..2cfa414 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -935,8 +935,11 @@ static struct dentry *nfs_lookup(struct 
 
 no_entry:
 	res = d_materialise_unique(dentry, inode);
-	if (res != NULL)
+	if (res != NULL) {
+		if (IS_ERR(res))
+			goto out_unlock;
 		dentry = res;
+	}
 	nfs_renew_times(dentry);
 	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
 out_unlock:
@@ -1132,6 +1135,8 @@ static struct dentry *nfs_readdir_lookup
 	alias = d_materialise_unique(dentry, inode);
 	if (alias != NULL) {
 		dput(dentry);
+		if (IS_ERR(alias))
+			return NULL;
 		dentry = alias;
 	}
 

Re: [PATCH 1/2] VFS: Make d_materialise_unique() enforce directory uniqueness

[Miklos Szeredi]

> From: Trond Myklebust <Trond.Myklebust@netapp.com>
> 
> If the caller tries to instantiate a directory using an inode that already
> has a dentry alias, then we attempt to rename the existing dentry instead
> of instantiating a new one. Fail with an ELOOP error if the rename would
> affect one of our parent directories.
> 
> This behaviour is needed in order to avoid issues such as
> 
>   http://bugzilla.kernel.org/show_bug.cgi?id=7178
>

This looks like a stale patch.  You posted one on -fsdevel that did
i_mutex locking as well.

Miklos

Re: [PATCH 1/2] VFS: Make d_materialise_unique() enforce directory uniqueness

[Trond Myklebust]

On Thu, 2006-10-19 at 22:51 +0200, Miklos Szeredi wrote:
> > From: Trond Myklebust <Trond.Myklebust@netapp.com>
> > 
> > If the caller tries to instantiate a directory using an inode that already
> > has a dentry alias, then we attempt to rename the existing dentry instead
> > of instantiating a new one. Fail with an ELOOP error if the rename would
> > affect one of our parent directories.
> > 
> > This behaviour is needed in order to avoid issues such as
> > 
> >   http://bugzilla.kernel.org/show_bug.cgi?id=7178
> >
> 
> This looks like a stale patch.  You posted one on -fsdevel that did
> i_mutex locking as well.

Argh! Thanks for spotting that Miklos... I had forgotten to update my
'fixes' stgit branch with the newer version. Please discard the previous
version of this patch, Linus. I'll resend...

Cheers,
  Trond

[PATCH 1/2] VFS: Make d_materialise_unique() enforce directory uniqueness

[Trond Myklebust]

From: Trond Myklebust <Trond.Myklebust@netapp.com>

If the caller tries to instantiate a directory using an inode that already
has a dentry alias, then we attempt to rename the existing dentry instead
of instantiating a new one. Fail with an ELOOP error if the rename would
affect one of our parent directories.

This behaviour is needed in order to avoid issues such as

  http://bugzilla.kernel.org/show_bug.cgi?id=7178

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---

 fs/dcache.c  |  137 ++++++++++++++++++++++++++++++++++++++++++----------------
 fs/nfs/dir.c |    7 +++
 2 files changed, 106 insertions(+), 38 deletions(-)

diff --git a/fs/dcache.c b/fs/dcache.c
index 2bac4ba..a1ff91e 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -1469,23 +1469,21 @@ static void switch_names(struct dentry *
  * deleted it.
  */
  
-/**
- * d_move - move a dentry
+/*
+ * d_move_locked - move a dentry
  * @dentry: entry to move
  * @target: new dentry
  *
  * Update the dcache to reflect the move of a file name. Negative
  * dcache entries should not be moved in this way.
  */
-
-void d_move(struct dentry * dentry, struct dentry * target)
+static void d_move_locked(struct dentry * dentry, struct dentry * target)
 {
 	struct hlist_head *list;
 
 	if (!dentry->d_inode)
 		printk(KERN_WARNING "VFS: moving negative dcache entry\n");
 
-	spin_lock(&dcache_lock);
 	write_seqlock(&rename_lock);
 	/*
 	 * XXXX: do we really need to take target->d_lock?
@@ -1536,10 +1534,84 @@ already_unhashed:
 	fsnotify_d_move(dentry);
 	spin_unlock(&dentry->d_lock);
 	write_sequnlock(&rename_lock);
+}
+
+/**
+ * d_move - move a dentry
+ * @dentry: entry to move
+ * @target: new dentry
+ *
+ * Update the dcache to reflect the move of a file name. Negative
+ * dcache entries should not be moved in this way.
+ */
+
+void d_move(struct dentry * dentry, struct dentry * target)
+{
+	spin_lock(&dcache_lock);
+	d_move_locked(dentry, target);
 	spin_unlock(&dcache_lock);
 }
 
 /*
+ * Helper that returns 1 if p1 is a parent of p2, else 0
+ */
+static int d_isparent(struct dentry *p1, struct dentry *p2)
+{
+	struct dentry *p;
+
+	for (p = p2; p->d_parent != p; p = p->d_parent) {
+		if (p->d_parent == p1)
+			return 1;
+	}
+	return 0;
+}
+
+/*
+ * This helper attempts to cope with remotely renamed directories
+ *
+ * It assumes that the caller is already holding
+ * dentry->d_parent->d_inode->i_mutex and the dcache_lock
+ *
+ * Note: If ever the locking in lock_rename() changes, then please
+ * remember to update this too...
+ *
+ * On return, dcache_lock will have been unlocked.
+ */
+static struct dentry *__d_unalias(struct dentry *dentry, struct dentry *alias)
+{
+	struct mutex *m1 = NULL, *m2 = NULL;
+	struct dentry *ret;
+
+	/* If alias and dentry share a parent, then no extra locks required */
+	if (alias->d_parent == dentry->d_parent)
+		goto out_unalias;
+
+	/* Check for loops */
+	ret = ERR_PTR(-ELOOP);
+	if (d_isparent(alias, dentry))
+		goto out_err;
+
+	/* See lock_rename() */
+	ret = ERR_PTR(-EBUSY);
+	if (!mutex_trylock(&dentry->d_sb->s_vfs_rename_mutex))
+		goto out_err;
+	m1 = &dentry->d_sb->s_vfs_rename_mutex;
+	if (!mutex_trylock(&alias->d_parent->d_inode->i_mutex))
+		goto out_err;
+	m2 = &alias->d_parent->d_inode->i_mutex;
+out_unalias:
+	d_move_locked(alias, dentry);
+	ret = alias;
+out_err:
+	spin_unlock(&dcache_lock);
+	if (m2)
+		mutex_unlock(m2);
+	if (m1)
+		mutex_unlock(m1);
+	return ret;
+}
+
+/*
  * Prepare an anonymous dentry for life in the superblock's dentry tree as a
  * named dentry in place of the dentry to be replaced.
  */
@@ -1581,7 +1653,7 @@ static void __d_materialise_dentry(struc
  */
 struct dentry *d_materialise_unique(struct dentry *dentry, struct inode *inode)
 {
-	struct dentry *alias, *actual;
+	struct dentry *actual;
 
 	BUG_ON(!d_unhashed(dentry));
 
@@ -1593,26 +1665,27 @@ struct dentry *d_materialise_unique(stru
 		goto found_lock;
 	}
 
-	/* See if a disconnected directory already exists as an anonymous root
-	 * that we should splice into the tree instead */
-	if (S_ISDIR(inode->i_mode) && (alias = __d_find_alias(inode, 1))) {
-		spin_lock(&alias->d_lock);
-
-		/* Is this a mountpoint that we could splice into our tree? */
-		if (IS_ROOT(alias))
-			goto connect_mountpoint;
-
-		if (alias->d_name.len == dentry->d_name.len &&
-		    alias->d_parent == dentry->d_parent &&
-		    memcmp(alias->d_name.name,
-			   dentry->d_name.name,
-			   dentry->d_name.len) == 0)
-			goto replace_with_alias;
-
-		spin_unlock(&alias->d_lock);
-
-		/* Doh! Seem to be aliasing directories for some reason... */
-		dput(alias);
+	if (S_ISDIR(inode->i_mode)) {
+		struct dentry *alias;
+
+		/* Does an aliased dentry already exist? */
+		alias = __d_find_alias(inode, 0);
+		if (alias) {
+			actual = alias;
+			/* Is this an anonymous mountpoint that we could splice
+			 * into our tree? */
+			if (IS_ROOT(alias)) {
+				spin_lock(&alias->d_lock);
+				__d_materialise_dentry(dentry, alias);
+				__d_drop(alias);
+				goto found;
+			}
+			/* Nope, but we must(!) avoid directory aliasing */
+			actual = __d_unalias(dentry, alias);
+			if (IS_ERR(actual))
+				dput(alias);
+			goto out_nolock;
+		}
 	}
 
 	/* Add a unique reference */
@@ -1628,7 +1701,7 @@ found:
 	_d_rehash(actual);
 	spin_unlock(&actual->d_lock);
 	spin_unlock(&dcache_lock);
-
+out_nolock:
 	if (actual == dentry) {
 		security_d_instantiate(dentry, inode);
 		return NULL;
@@ -1637,16 +1710,6 @@ found:
 	iput(inode);
 	return actual;
 
-	/* Convert the anonymous/root alias into an ordinary dentry */
-connect_mountpoint:
-	__d_materialise_dentry(dentry, alias);
-
-	/* Replace the candidate dentry with the alias in the tree */
-replace_with_alias:
-	__d_drop(alias);
-	actual = alias;
-	goto found;
-
 shouldnt_be_hashed:
 	spin_unlock(&dcache_lock);
 	BUG();
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index c86a1ea..2cfa414 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -935,8 +935,11 @@ static struct dentry *nfs_lookup(struct 
 
 no_entry:
 	res = d_materialise_unique(dentry, inode);
-	if (res != NULL)
+	if (res != NULL) {
+		if (IS_ERR(res))
+			goto out_unlock;
 		dentry = res;
+	}
 	nfs_renew_times(dentry);
 	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
 out_unlock:
@@ -1132,6 +1135,8 @@ static struct dentry *nfs_readdir_lookup
 	alias = d_materialise_unique(dentry, inode);
 	if (alias != NULL) {
 		dput(dentry);
+		if (IS_ERR(alias))
+			return NULL;
 		dentry = alias;
 	}
 

[PATCH 2/2] NFS: Cache invalidation fixup

[Trond Myklebust]

From: Trond Myklebust <Trond.Myklebust@netapp.com>

If someone has renamed a directory on the server, triggering the d_move
code in d_materialise_unique(), then we need to invalidate the
cached directory information in the source parent directory.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---

 fs/nfs/dir.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 2cfa414..fa71936 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -936,8 +936,14 @@ static struct dentry *nfs_lookup(struct 
 no_entry:
 	res = d_materialise_unique(dentry, inode);
 	if (res != NULL) {
+		struct dentry *parent;
 		if (IS_ERR(res))
 			goto out_unlock;
+		/* Was a directory renamed! */
+		parent = dget_parent(res);
+		if (!IS_ROOT(parent))
+			nfs_mark_for_revalidate(parent->d_inode);
+		dput(parent);
 		dentry = res;
 	}
 	nfs_renew_times(dentry);

Re: [PATCH 1/2] VFS: Make d_materialise_unique() enforce directory uniqueness

[Andrew Morton]

On Thu, 19 Oct 2006 17:03:58 -0400
Trond Myklebust <Trond.Myklebust@netapp.com> wrote:

> +static struct dentry *__d_unalias(struct dentry *dentry, struct dentry *alias)
> +{
> +	struct mutex *m1 = NULL, *m2 = NULL;
> +	struct dentry *ret;
> +
> +	/* If alias and dentry share a parent, then no extra locks required */
> +	if (alias->d_parent == dentry->d_parent)
> +		goto out_unalias;
> +
> +	/* Check for loops */
> +	ret = ERR_PTR(-ELOOP);
> +	if (d_isparent(alias, dentry))
> +		goto out_err;
> +
> +	/* See lock_rename() */
> +	ret = ERR_PTR(-EBUSY);
> +	if (!mutex_trylock(&dentry->d_sb->s_vfs_rename_mutex))
> +		goto out_err;
> +	m1 = &dentry->d_sb->s_vfs_rename_mutex;
> +	if (!mutex_trylock(&alias->d_parent->d_inode->i_mutex))
> +		goto out_err;
> +	m2 = &alias->d_parent->d_inode->i_mutex;
> +out_unalias:
> +	d_move_locked(alias, dentry);
> +	ret = alias;
> +out_err:
> +	spin_unlock(&dcache_lock);
> +	if (m2)
> +		mutex_unlock(m2);
> +	if (m1)
> +		mutex_unlock(m1);
> +	return ret;
> +}

The locking in there is, of course, gruesome.  There is no way in which it
can be made reliable?

Re: [PATCH 1/2] VFS: Make d_materialise_unique() enforce directory uniqueness

[Trond Myklebust]

On Thu, 2006-10-19 at 21:40 -0700, Andrew Morton wrote:
> On Thu, 19 Oct 2006 17:03:58 -0400
> Trond Myklebust <Trond.Myklebust@netapp.com> wrote:
> 
> > +static struct dentry *__d_unalias(struct dentry *dentry, struct dentry *alias)
> > +{
> > +	struct mutex *m1 = NULL, *m2 = NULL;
> > +	struct dentry *ret;
> > +
> > +	/* If alias and dentry share a parent, then no extra locks required */
> > +	if (alias->d_parent == dentry->d_parent)
> > +		goto out_unalias;
> > +
> > +	/* Check for loops */
> > +	ret = ERR_PTR(-ELOOP);
> > +	if (d_isparent(alias, dentry))
> > +		goto out_err;
> > +
> > +	/* See lock_rename() */
> > +	ret = ERR_PTR(-EBUSY);
> > +	if (!mutex_trylock(&dentry->d_sb->s_vfs_rename_mutex))
> > +		goto out_err;
> > +	m1 = &dentry->d_sb->s_vfs_rename_mutex;
> > +	if (!mutex_trylock(&alias->d_parent->d_inode->i_mutex))
> > +		goto out_err;
> > +	m2 = &alias->d_parent->d_inode->i_mutex;
> > +out_unalias:
> > +	d_move_locked(alias, dentry);
> > +	ret = alias;
> > +out_err:
> > +	spin_unlock(&dcache_lock);
> > +	if (m2)
> > +		mutex_unlock(m2);
> > +	if (m1)
> > +		mutex_unlock(m1);
> > +	return ret;
> > +}
> 
> The locking in there is, of course, gruesome.  There is no way in which it
> can be made reliable?

The generic lookup() code will grab the dir->i_mutex for the parent
directory before we get anywhere near __d_alias(). That pretty much
limits us to using mutex_trylock() since otherwise we break the nesting
rules for lock_rename(), and I can't see that we can release the
problematic dir->i_mutex without causing worse races.

Cheers,
  Trond