Since l2cap_sock_teardown_cb doesn't acquire the channel lock before
setting the socket as zapped, it could potentially race with
l2cap_sock_release which frees the socket. Thus, wait until the cleanup
is complete before marking the socket as zapped.
This race was reproduced on a JBL GO speaker after the remote device
rejected L2CAP connection due to resource unavailability.
Here is a dmesg log with debug logs from a repro of this bug:
[ 3465.424086] Bluetooth: hci_core.c:hci_acldata_packet() hci0 len 16 handle 0x0003 flags 0x0002
[ 3465.424090] Bluetooth: hci_conn.c:hci_conn_enter_active_mode() hcon 00000000cfedd07d mode 0
[ 3465.424094] Bluetooth: l2cap_core.c:l2cap_recv_acldata() conn 000000007eae8952 len 16 flags 0x2
[ 3465.424098] Bluetooth: l2cap_core.c:l2cap_recv_frame() len 12, cid 0x0001
[ 3465.424102] Bluetooth: l2cap_core.c:l2cap_raw_recv() conn 000000007eae8952
[ 3465.424175] Bluetooth: l2cap_core.c:l2cap_sig_channel() code 0x03 len 8 id 0x0c
[ 3465.424180] Bluetooth: l2cap_core.c:l2cap_connect_create_rsp() dcid 0x0045 scid 0x0000 result 0x02 status 0x00
[ 3465.424189] Bluetooth: l2cap_core.c:l2cap_chan_put() chan 000000006acf9bff orig refcnt 4
[ 3465.424196] Bluetooth: l2cap_core.c:l2cap_chan_del() chan 000000006acf9bff, conn 000000007eae8952, err 111, state BT_CONNECT
[ 3465.424203] Bluetooth: l2cap_sock.c:l2cap_sock_teardown_cb() chan 000000006acf9bff state BT_CONNECT
[ 3465.424221] Bluetooth: l2cap_core.c:l2cap_chan_put() chan 000000006acf9bff orig refcnt 3
[ 3465.424226] Bluetooth: hci_core.h:hci_conn_drop() hcon 00000000cfedd07d orig refcnt 6
[ 3465.424234] BUG: spinlock bad magic on CPU#2, kworker/u17:0/159
[ 3465.425626] Bluetooth: hci_sock.c:hci_sock_sendmsg() sock 000000002bb0cb64 sk 00000000a7964053
[ 3465.430330] lock: 0xffffff804410aac0, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
[ 3465.430332] Causing a watchdog bite!
Signed-off-by: Abhishek Pandit-Subedi <abhishekpandit@xxxxxxxxxxxx>
Reported-by: Balakrishna Godavarthi <bgodavar@xxxxxxxxxxxxxx>
Reviewed-by: Manish Mandlik <mmandlik@xxxxxxxxxxxx>
---
We had some more data available (outside of dmesg and oops) that led us
to suspect a race between l2cap_sock_teardown_cb and l2cap_sock_release.
I've left this out of the commit message since it's not an oops or dmesg
logs.
Crash stack from CPU4:
--
-24 |spin_bug(
| [X19] lock = 0xFFFFFF810BDB1EC0,
| [X20] msg = 0xFFFFFFD143FD7960)
-25 |debug_spin_lock_before(inline)
| [X19] lock = 0xFFFFFF810BDB1EC0
-25 |do_raw_spin_lock(
| [X19] lock = 0xFFFFFF810BDB1EC0)
-26 |raw_spin_lock_irqsave(
| [X19] lock = 0xFFFFFF810BDB1EC0)
-27 |skb_peek(inline)
-27 |__skb_dequeue(inline)
-27 |skb_dequeue(
| [X20] list = 0xFFFFFF810BDB1EA8)
| [locdesc] flags = 12297829382473034410
-28 |skb_queue_purge(
| [X19] list = 0xFFFFFF810BDB1EA8 -> (
| [D:0xFFFFFF810BDB1EA8] next = 0x0,
| [D:0xFFFFFF810BDB1EB0] prev = 0x0,
| [D:0xFFFFFF810BDB1EB8] qlen = 0,
| [D:0xFFFFFF810BDB1EC0] lock = ([D:0xFFFFFF810BDB1EC0] rlock = ([D:0xFFFFFF810BDB1EC0] raw_lock
| [X0] skb = ???
-29 |l2cap_seq_list_free(inline)
| [locdesc] seq_list = 0xFFFFFF810BDB1ED8 -> (
| [D:0xFFFFFF810BDB1ED8] head = 0,
| [D:0xFFFFFF810BDB1EDA] tail = 0,
| [D:0xFFFFFF810BDB1EDC] mask = 0,
| [D:0xFFFFFF810BDB1EE0] list = 0x0)
-29 |l2cap_chan_del(
| [X19] chan = 0xFFFFFF810BDB1C00,
| ?)
-30 |l2cap_chan_unlock(inline)
-30 |l2cap_connect_create_rsp(inline)
| [X20] conn = 0xFFFFFF81231F2600
| [locdesc] err = 0
| [X27] chan = 0xFFFFFF810BDB1C00
-30 |l2cap_bredr_sig_cmd(inline)
| [X20] conn = 0xFFFFFF81231F2600
| [locdesc] err = 0
-30 |l2cap_sig_channel(inline)
| [X20] conn = 0xFFFFFF81231F2600
| [X19] skb = 0xFFFFFF813DE4C040
| [X28] data = 0xFFFFFF8131582014
| [locdesc] cmd_len = 43690
-30 |l2cap_recv_frame(
| [X20] conn = 0xFFFFFF81231F2600,
| [X19] skb = 0xFFFFFF813DE4C040)
| [locdesc] psm = 43690
-31 |l2cap_recv_acldata(
| ?,
| [X19] skb = 0xFFFFFF813DE4C040,
| ?)
| [X21] len = 16
-32 |hci_rx_work(
| ?)
| [X21] hdev = 0xFFFFFF8133A02000
-33 |__read_once_size(inline)
| [locdesc] size = 4
-33 |atomic_read(inline)
| [locdesc] __u = ([locdesc] __val = -1431655766, [locdesc] __c = (170))
-33 |static_key_count(inline)
-33 |static_key_false(inline)
-33 |trace_workqueue_execute_end(inline)
| [X22] work = 0xFFFFFF8133A02838
-33 |process_one_work(
| [X19] worker = 0xFFFFFF8133FE4500,
| [X22] work = 0xFFFFFF8133A02838)
| [locdesc] work_color = -1431655766
-34 |__read_once_size(inline)
| [locdesc] size = 8
-34 |list_empty(inline)
| [locdesc] __u = ([locdesc] __val = 0xAAAAAAAAAAAAAAAA, [locdesc] __c = (170))
-34 |worker_thread(
| [X19] __worker = 0xFFFFFF8133FE4500)
| [X19] worker = 0xFFFFFF8133FE4500
-35 |kthread(
| [X20] _create = 0xFFFFFF8133FB3A00)
| [X20] create = 0xFFFFFF8133FB3A00
| [X0] ret = ???
-36 |ret_from_fork(asm)
net/bluetooth/l2cap_sock.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c
index e1a3e66b175402..e7cfe28140c39b 100644
--- a/net/bluetooth/l2cap_sock.c
+++ b/net/bluetooth/l2cap_sock.c
@@ -1521,8 +1521,6 @@ static void l2cap_sock_teardown_cb(struct l2cap_chan *chan, int err)
parent = bt_sk(sk)->parent;
- sock_set_flag(sk, SOCK_ZAPPED);
-
switch (chan->state) {
case BT_OPEN:
case BT_BOUND:
@@ -1549,8 +1547,11 @@ static void l2cap_sock_teardown_cb(struct l2cap_chan *chan, int err)
break;
}
-
release_sock(sk);
+
+ /* Only zap after cleanup to avoid use after free race */
+ sock_set_flag(sk, SOCK_ZAPPED);
+
}
static void l2cap_sock_state_change_cb(struct l2cap_chan *chan, int state,
--
2.28.0.618.gf4bc123cb7-goog
