Skip to contentSkip to conversation

ceph: fix race in concurrent __ceph_remove_cap invocations

PATCH In review Luis Henriques 3 participants 5 messages Nov 12, 2020 Ceph client

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ceph: fix race in concurrent __ceph_remove_cap invocations

A NULL pointer dereference may occur in __ceph_remove_cap with some of the callbacks used in ceph_iterate_session_caps, namely trim_caps_cb and remove_session_caps_cb. These aren't protected against the concurrent execution of __ceph_remove_cap.

Since the callers of this function hold the i_ceph_lock, the fix is simply a matter of returning immediately if caps->ci is NULL.

Based on a patch from Jeff Layton.

Cc: [email protected]
URL: https://tracker.ceph.com/issues/43272

---
 fs/ceph/caps.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
fs/ceph/caps.c +9 -2
index ded4229c314a..443f164760d5 100644
@@ -1140,12 +1140,19 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
 {
 	struct ceph_mds_session *session = cap->session;
 	struct ceph_inode_info *ci = cap->ci;
-	struct ceph_mds_client *mdsc =
-		ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
+	struct ceph_mds_client *mdsc;
 	int removed = 0;
 
+	/* 'ci' being NULL means he remove have already occurred */
+	if (!ci) {
+		dout("%s: cap inode is NULL\n", __func__);
+		return;
+	}
+
 	dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);
 
+	mdsc = ceph_inode_to_client(&ci->vfs_inode)->mdsc;
+
 	/* remove from inode's cap rbtree, and clear auth cap */
 	rb_erase(&cap->ci_node, &ci->i_caps);
 	if (ci->i_auth_cap == cap) {
#1 lore raw
Quoted 43 lines · Luis Henriques

A NULL pointer dereference may occur in __ceph_remove_cap with some of the callbacks used in ceph_iterate_session_caps, namely trim_caps_cb and remove_session_caps_cb. These aren't protected against the concurrent execution of __ceph_remove_cap.

Since the callers of this function hold the i_ceph_lock, the fix is simply a matter of returning immediately if caps->ci is NULL.

Based on a patch from Jeff Layton.

Cc: [email protected]
URL: https://tracker.ceph.com/issues/43272

---
 fs/ceph/caps.c | 11 +++++++++--

 1 file changed, 9 insertions(+), 2 deletions(-)
fs/ceph/caps.c +1 -2
index ded4229c314a..443f164760d5 100644
@@ -1140,12 +1140,19 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
  {
  	struct ceph_mds_session *session = cap->session;
  	struct ceph_inode_info *ci = cap->ci;
-	struct ceph_mds_client *mdsc =
-		ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
+	struct ceph_mds_client *mdsc;
  	int removed = 0;

fs/ceph/caps.c +6 -0
+	/* 'ci' being NULL means he remove have already occurred */
+	if (!ci) {
+		dout("%s: cap inode is NULL\n", __func__);
+		return;
+	}
+

  dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);

fs/ceph/caps.c +2 -0
+	mdsc = ceph_inode_to_client(&ci->vfs_inode)->mdsc;
+
Quoted 3 lines

  /* remove from inode's cap rbtree, and clear auth cap */ rb_erase(&cap->ci_node, &ci->i_caps);
  if (ci->i_auth_cap == cap) {

Merged into testing branch (with a minor fix to the comment).

Signature · 1 line
Jeff Layton <[email protected]>
#2 lore raw
In reply to Luis Henriques
Quoted 6 lines · Luis Henriques

A NULL pointer dereference may occur in __ceph_remove_cap with some of the callbacks used in ceph_iterate_session_caps, namely trim_caps_cb and remove_session_caps_cb. These aren't protected against the concurrent execution of __ceph_remove_cap.

they are protected by session mutex, never get executed concurrently

Quoted 39 lines

Since the callers of this function hold the i_ceph_lock, the fix is simply a matter of returning immediately if caps->ci is NULL.

Based on a patch from Jeff Layton.

Cc: [email protected]
URL: https://tracker.ceph.com/issues/43272

---
 fs/ceph/caps.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
fs/ceph/caps.c +9 -2
index ded4229c314a..443f164760d5 100644
@@ -1140,12 +1140,19 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
 {
        struct ceph_mds_session *session = cap->session;
        struct ceph_inode_info *ci = cap->ci;
-       struct ceph_mds_client *mdsc =
-               ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
+       struct ceph_mds_client *mdsc;
        int removed = 0;
 
+       /* 'ci' being NULL means he remove have already occurred */
+       if (!ci) {
+               dout("%s: cap inode is NULL\n", __func__);
+               return;
+       }
+
        dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);
 
+       mdsc = ceph_inode_to_client(&ci->vfs_inode)->mdsc;
+
        /* remove from inode's cap rbtree, and clear auth cap */
        rb_erase(&cap->ci_node, &ci->i_caps);
        if (ci->i_auth_cap == cap) {
#3 lore raw
Quoted 10 lines · Zheng
Quoted 6 lines · Luis Henriques

A NULL pointer dereference may occur in __ceph_remove_cap with some of the callbacks used in ceph_iterate_session_caps, namely trim_caps_cb and remove_session_caps_cb. These aren't protected against the concurrent execution of __ceph_remove_cap.

they are protected by session mutex, never get executed concurrently

Maybe not concurrently with one another, but the s_mutex is _not_ held when __ceph_remove_caps is called from ceph_evict_inode. We can't rely on it to protect this.

Quoted 39 lines
Quoted 39 lines

Since the callers of this function hold the i_ceph_lock, the fix is simply a matter of returning immediately if caps->ci is NULL.

Based on a patch from Jeff Layton.

Cc: [email protected]
URL: https://tracker.ceph.com/issues/43272

---
 fs/ceph/caps.c | 11 +++++++++--

 1 file changed, 9 insertions(+), 2 deletions(-)
fs/ceph/caps.c +9 -2
index ded4229c314a..443f164760d5 100644
@@ -1140,12 +1140,19 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
  {
         struct ceph_mds_session *session = cap->session;
         struct ceph_inode_info *ci = cap->ci;
-       struct ceph_mds_client *mdsc =
-               ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
+       struct ceph_mds_client *mdsc;
         int removed = 0;
 
+       /* 'ci' being NULL means he remove have already occurred */
+       if (!ci) {
+               dout("%s: cap inode is NULL\n", __func__);
+               return;
+       }
+
         dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);
 
+       mdsc = ceph_inode_to_client(&ci->vfs_inode)->mdsc;
+
         /* remove from inode's cap rbtree, and clear auth cap */
         rb_erase(&cap->ci_node, &ci->i_caps);
         if (ci->i_auth_cap == cap) {
Signature · 1 line
Jeff Layton <[email protected]>
#4 lore raw
Quoted 9 lines · Jeff Layton
Quoted 8 lines · Zheng
Quoted 6 lines · Luis Henriques

A NULL pointer dereference may occur in __ceph_remove_cap with some of the callbacks used in ceph_iterate_session_caps, namely trim_caps_cb and remove_session_caps_cb. These aren't protected against the concurrent execution of __ceph_remove_cap.

Quoted 6 lines
Quoted 2 lines

they are protected by session mutex, never get executed concurrently

Maybe not concurrently with one another, but the s_mutex is _not_ held when __ceph_remove_caps is called from ceph_evict_inode. We can't rely on it to protect this.

Hmm, yeah. I guess the changelog could mention that. Thanks, Jeff.

Cheers,

Signature · 1 line
Luis
Quoted 43 lines
Quoted 39 lines
Quoted 39 lines

Since the callers of this function hold the i_ceph_lock, the fix is simply a matter of returning immediately if caps->ci is NULL.

Based on a patch from Jeff Layton.

Cc: [email protected]
URL: https://tracker.ceph.com/issues/43272

---
 fs/ceph/caps.c | 11 +++++++++--

 1 file changed, 9 insertions(+), 2 deletions(-)
fs/ceph/caps.c +9 -2
index ded4229c314a..443f164760d5 100644
@@ -1140,12 +1140,19 @@ void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release)
  {
         struct ceph_mds_session *session = cap->session;
         struct ceph_inode_info *ci = cap->ci;
-       struct ceph_mds_client *mdsc =
-               ceph_sb_to_client(ci->vfs_inode.i_sb)->mdsc;
+       struct ceph_mds_client *mdsc;
         int removed = 0;
 
+       /* 'ci' being NULL means he remove have already occurred */
+       if (!ci) {
+               dout("%s: cap inode is NULL\n", __func__);
+               return;
+       }
+
         dout("__ceph_remove_cap %p from %p\n", cap, &ci->vfs_inode);
 
+       mdsc = ceph_inode_to_client(&ci->vfs_inode)->mdsc;
+
         /* remove from inode's cap rbtree, and clear auth cap */
         rb_erase(&cap->ci_node, &ci->i_caps);
         if (ci->i_auth_cap == cap) {
Signature · 1 line
Jeff Layton <[email protected]>