Skip to contentSkip to conversation

btrfs: use refcount_inc_not_zero in kill_all_nodes

PATCH Superseded Josef Bacik 2 participants 2 messages Sep 26, 2019 Btrfs file system

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 btrfs: use refcount_inc_not_zero in kill_all_nodes

We hit the following warning while running down a different problem

[ 6197.175850] ------------[ cut here ]------------
[ 6197.185082] refcount_t: underflow; use-after-free.
[ 6197.194704] WARNING: CPU: 47 PID: 966 at lib/refcount.c:190 refcount_sub_and_test_checked+0x53/0x60
[ 6197.521792] Call Trace:
[ 6197.526687]  __btrfs_release_delayed_node+0x76/0x1c0
[ 6197.536615]  btrfs_kill_all_delayed_nodes+0xec/0x130
[ 6197.546532]  ? __btrfs_btree_balance_dirty+0x60/0x60
[ 6197.556482]  btrfs_clean_one_deleted_snapshot+0x71/0xd0
[ 6197.566910]  cleaner_kthread+0xfa/0x120
[ 6197.574573]  kthread+0x111/0x130
[ 6197.581022]  ? kthread_create_on_node+0x60/0x60
[ 6197.590086]  ret_from_fork+0x1f/0x30
[ 6197.597228] ---[ end trace 424bb7ae00509f56 ]---

This is because the free side drops the ref without the lock, and then takes the lock if our refcount is 0. So you can have nodes on the tree that have a refcount of 0. Fix this by zero'ing out that element in our temporary array so we don't try to kill it again.

---
v1->v2:
- I'm an idiot.

 fs/btrfs/delayed-inode.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
fs/btrfs/delayed-inode.c +7 -4
index 1f7f39b10bd0..81b2fd46886f 100644
@@ -1948,13 +1948,16 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root)
 			break;
 		}
 
-		inode_id = delayed_nodes[n - 1]->inode_id + 1;
-
-		for (i = 0; i < n; i++)
-			refcount_inc(&delayed_nodes[i]->refs);
+		for (i = 0; i < n; i++) {
+			inode_id = delayed_nodes[i]->inode_id + 1;
+			if (!refcount_inc_not_zero(&delayed_nodes[i]->refs))
+				delayed_nodes[i] = NULL;
+		}
 		spin_unlock(&root->inode_lock);
 
 		for (i = 0; i < n; i++) {
+			if (!delayed_nodes[i])
+				continue;
 			__btrfs_kill_delayed_node(delayed_nodes[i]);
 			btrfs_release_delayed_node(delayed_nodes[i]);
 		}
Signature · 1 line
2.21.0
Reviewed-by on v2 1/1
#1 lore raw
Quoted 16 lines · Josef Bacik

We hit the following warning while running down a different problem

[ 6197.175850] ------------[ cut here ]------------
[ 6197.185082] refcount_t: underflow; use-after-free.
[ 6197.194704] WARNING: CPU: 47 PID: 966 at lib/refcount.c:190 refcount_sub_and_test_checked+0x53/0x60
[ 6197.521792] Call Trace:
[ 6197.526687]  __btrfs_release_delayed_node+0x76/0x1c0
[ 6197.536615]  btrfs_kill_all_delayed_nodes+0xec/0x130
[ 6197.546532]  ? __btrfs_btree_balance_dirty+0x60/0x60
[ 6197.556482]  btrfs_clean_one_deleted_snapshot+0x71/0xd0
[ 6197.566910]  cleaner_kthread+0xfa/0x120
[ 6197.574573]  kthread+0x111/0x130
[ 6197.581022]  ? kthread_create_on_node+0x60/0x60
[ 6197.590086]  ret_from_fork+0x1f/0x30
[ 6197.597228] ---[ end trace 424bb7ae00509f56 ]---
Quoted 6 lines

This is because the free side drops the ref without the lock, and then takes the lock if our refcount is 0. So you can have nodes on the tree that have a refcount of 0. Fix this by zero'ing out that element in our temporary array so we don't try to kill it again.

Looks good, one minor nit below though and you can add:

Quoted 12 lines

---
v1->v2:
- I'm an idiot.

 fs/btrfs/delayed-inode.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
fs/btrfs/delayed-inode.c +0 -0
index 1f7f39b10bd0..81b2fd46886f 100644

fs/btrfs/delayed-inode.c void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) · line 1952

Quoted 9 lines
fs/btrfs/delayed-inode.c +2 -4
@@ -1948,13 +1948,16 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root)
 			break;
 		}
 
-		inode_id = delayed_nodes[n - 1]->inode_id + 1;
-
-		for (i = 0; i < n; i++)
-			refcount_inc(&delayed_nodes[i]->refs);
+		for (i = 0; i < n; i++) {
+			inode_id = delayed_nodes[i]->inode_id + 1;

Since you no longer are doing a break in the loop there is no point in assigning inode_id in the loop. You can retain the old code that does:
inode_id = delayed_nodes[n - 1]->inode_id + 1;

Since with the current structure it's always guaranteed that inode_id will be 1 higher than the inode_Id of the last inode in delayed_nodes .

Quoted 12 lines
fs/btrfs/delayed-inode.c +5 -0
+			if (!refcount_inc_not_zero(&delayed_nodes[i]->refs))
+				delayed_nodes[i] = NULL;
+		}
 		spin_unlock(&root->inode_lock);
 
 		for (i = 0; i < n; i++) {
+			if (!delayed_nodes[i])
+				continue;
 			__btrfs_kill_delayed_node(delayed_nodes[i]);
 			btrfs_release_delayed_node(delayed_nodes[i]);
 		}