BMC/Intel-BMC/linux.git/net/bluetooth, branch dev-5.14-intel

Bluetooth: Fix race condition in handling NOP command

2021-09-18T11:43:51+00:00

[ Upstream commit ecb71f2566673553bc067e5b0036756871d0b9d3 ] For NOP command, need to cancel work scheduled on cmd_timer, on receiving command status or commmand complete event. Below use case might lead to race condition multiple when NOP commands are queued sequentially: hci_cmd_work() { if (atomic_read(&hdev->cmd_cnt) { . . . atomic_dec(&hdev->cmd_cnt); hci_send_frame(hdev,...); schedule_delayed_work(&hdev->cmd_timer,...); } } On receiving event for first NOP, the work scheduled on hdev->cmd_timer is not cancelled and second NOP is dequeued and sent to controller. While waiting for an event for second NOP command, work scheduled on cmd_timer for the first NOP can get scheduled, resulting in sending third NOP command (sending back to back NOP commands). This might cause issues at controller side (like memory overrun, controller going unresponsive) resulting in hci tx timeouts, hardware errors etc. The fix to this issue is to cancel the delayed work scheduled on cmd_timer on receiving command status or command complete event for NOP command (this patch handles NOP command same as any other SIG command). Signed-off-by: Kiran K Reviewed-by: Chethan T N Reviewed-by: Srivatsa Ravishankar Acked-by: Manish Mandlik Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: Fix handling of LE Enhanced Connection Complete

2021-09-18T11:43:51+00:00

[ Upstream commit cafae4cd625502f65d1798659c1aa9b62d38cc56 ] LE Enhanced Connection Complete contains the Local RPA used in the connection which must be used when set otherwise there could problems when pairing since the address used by the remote stack could be the Local RPA: BLUETOOTH CORE SPECIFICATION Version 5.2 | Vol 4, Part E page 2396 'Resolvable Private Address being used by the local device for this connection. This is only valid when the Own_Address_Type (from the HCI_LE_Create_Connection, HCI_LE_Set_Advertising_Parameters, HCI_LE_Set_Extended_Advertising_Parameters, or HCI_LE_Extended_Create_Connection commands) is set to 0x02 or 0x03, and the Controller generated a resolvable private address for the local device using a non-zero local IRK. For other Own_Address_Type values, the Controller shall return all zeros.' Signed-off-by: Luiz Augusto von Dentz Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: avoid circular locks in sco_sock_connect

2021-09-18T11:43:49+00:00

[ Upstream commit 734bc5ff783115aa3164f4e9dd5967ae78e0a8ab ] In a future patch, calls to bh_lock_sock in sco.c should be replaced by lock_sock now that none of the functions are run in IRQ context. However, doing so results in a circular locking dependency: ====================================================== WARNING: possible circular locking dependency detected 5.14.0-rc4-syzkaller #0 Not tainted ------------------------------------------------------ syz-executor.2/14867 is trying to acquire lock: ffff88803e3c1120 (sk_lock-AF_BLUETOOTH-BTPROTO_SCO){+.+.}-{0:0}, at: lock_sock include/net/sock.h:1613 [inline] ffff88803e3c1120 (sk_lock-AF_BLUETOOTH-BTPROTO_SCO){+.+.}-{0:0}, at: sco_conn_del+0x12a/0x2a0 net/bluetooth/sco.c:191 but task is already holding lock: ffffffff8d2dc7c8 (hci_cb_list_lock){+.+.}-{3:3}, at: hci_disconn_cfm include/net/bluetooth/hci_core.h:1497 [inline] ffffffff8d2dc7c8 (hci_cb_list_lock){+.+.}-{3:3}, at: hci_conn_hash_flush+0xda/0x260 net/bluetooth/hci_conn.c:1608 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #2 (hci_cb_list_lock){+.+.}-{3:3}: __mutex_lock_common kernel/locking/mutex.c:959 [inline] __mutex_lock+0x12a/0x10a0 kernel/locking/mutex.c:1104 hci_connect_cfm include/net/bluetooth/hci_core.h:1482 [inline] hci_remote_features_evt net/bluetooth/hci_event.c:3263 [inline] hci_event_packet+0x2f4d/0x7c50 net/bluetooth/hci_event.c:6240 hci_rx_work+0x4f8/0xd30 net/bluetooth/hci_core.c:5122 process_one_work+0x98d/0x1630 kernel/workqueue.c:2276 worker_thread+0x658/0x11f0 kernel/workqueue.c:2422 kthread+0x3e5/0x4d0 kernel/kthread.c:319 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:295 -> #1 (&hdev->lock){+.+.}-{3:3}: __mutex_lock_common kernel/locking/mutex.c:959 [inline] __mutex_lock+0x12a/0x10a0 kernel/locking/mutex.c:1104 sco_connect net/bluetooth/sco.c:245 [inline] sco_sock_connect+0x227/0xa10 net/bluetooth/sco.c:601 __sys_connect_file+0x155/0x1a0 net/socket.c:1879 __sys_connect+0x161/0x190 net/socket.c:1896 __do_sys_connect net/socket.c:1906 [inline] __se_sys_connect net/socket.c:1903 [inline] __x64_sys_connect+0x6f/0xb0 net/socket.c:1903 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae -> #0 (sk_lock-AF_BLUETOOTH-BTPROTO_SCO){+.+.}-{0:0}: check_prev_add kernel/locking/lockdep.c:3051 [inline] check_prevs_add kernel/locking/lockdep.c:3174 [inline] validate_chain kernel/locking/lockdep.c:3789 [inline] __lock_acquire+0x2a07/0x54a0 kernel/locking/lockdep.c:5015 lock_acquire kernel/locking/lockdep.c:5625 [inline] lock_acquire+0x1ab/0x510 kernel/locking/lockdep.c:5590 lock_sock_nested+0xca/0x120 net/core/sock.c:3170 lock_sock include/net/sock.h:1613 [inline] sco_conn_del+0x12a/0x2a0 net/bluetooth/sco.c:191 sco_disconn_cfm+0x71/0xb0 net/bluetooth/sco.c:1202 hci_disconn_cfm include/net/bluetooth/hci_core.h:1500 [inline] hci_conn_hash_flush+0x127/0x260 net/bluetooth/hci_conn.c:1608 hci_dev_do_close+0x528/0x1130 net/bluetooth/hci_core.c:1778 hci_unregister_dev+0x1c0/0x5a0 net/bluetooth/hci_core.c:4015 vhci_release+0x70/0xe0 drivers/bluetooth/hci_vhci.c:340 __fput+0x288/0x920 fs/file_table.c:280 task_work_run+0xdd/0x1a0 kernel/task_work.c:164 exit_task_work include/linux/task_work.h:32 [inline] do_exit+0xbd4/0x2a60 kernel/exit.c:825 do_group_exit+0x125/0x310 kernel/exit.c:922 get_signal+0x47f/0x2160 kernel/signal.c:2808 arch_do_signal_or_restart+0x2a9/0x1c40 arch/x86/kernel/signal.c:865 handle_signal_work kernel/entry/common.c:148 [inline] exit_to_user_mode_loop kernel/entry/common.c:172 [inline] exit_to_user_mode_prepare+0x17d/0x290 kernel/entry/common.c:209 __syscall_exit_to_user_mode_work kernel/entry/common.c:291 [inline] syscall_exit_to_user_mode+0x19/0x60 kernel/entry/common.c:302 ret_from_fork+0x15/0x30 arch/x86/entry/entry_64.S:288 other info that might help us debug this: Chain exists of: sk_lock-AF_BLUETOOTH-BTPROTO_SCO --> &hdev->lock --> hci_cb_list_lock Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(hci_cb_list_lock); lock(&hdev->lock); lock(hci_cb_list_lock); lock(sk_lock-AF_BLUETOOTH-BTPROTO_SCO); *** DEADLOCK *** The issue is that the lock hierarchy should go from &hdev->lock --> hci_cb_list_lock --> sk_lock-AF_BLUETOOTH-BTPROTO_SCO. For example, one such call trace is: hci_dev_do_close(): hci_dev_lock(); hci_conn_hash_flush(): hci_disconn_cfm(): mutex_lock(&hci_cb_list_lock); sco_disconn_cfm(): sco_conn_del(): lock_sock(sk); However, in sco_sock_connect, we call lock_sock before calling hci_dev_lock inside sco_connect, thus inverting the lock hierarchy. We fix this by pulling the call to hci_dev_lock out from sco_connect. Signed-off-by: Desmond Cheong Zhi Xi Signed-off-by: Luiz Augusto von Dentz Signed-off-by: Sasha Levin

Bluetooth: schedule SCO timeouts with delayed_work

2021-09-18T11:43:49+00:00

[ Upstream commit ba316be1b6a00db7126ed9a39f9bee434a508043 ] struct sock.sk_timer should be used as a sock cleanup timer. However, SCO uses it to implement sock timeouts. This causes issues because struct sock.sk_timer's callback is run in an IRQ context, and the timer callback function sco_sock_timeout takes a spin lock on the socket. However, other functions such as sco_conn_del and sco_conn_ready take the spin lock with interrupts enabled. This inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} lock usage could lead to deadlocks as reported by Syzbot [1]: CPU0 ---- lock(slock-AF_BLUETOOTH-BTPROTO_SCO); lock(slock-AF_BLUETOOTH-BTPROTO_SCO); To fix this, we use delayed work to implement SCO sock timouts instead. This allows us to avoid taking the spin lock on the socket in an IRQ context, and corrects the misuse of struct sock.sk_timer. As a note, cancel_delayed_work is used instead of cancel_delayed_work_sync in sco_sock_set_timer and sco_sock_clear_timer to avoid a deadlock. In the future, the call to bh_lock_sock inside sco_sock_timeout should be changed to lock_sock to synchronize with other functions using lock_sock. However, since sco_sock_set_timer and sco_sock_clear_timer are sometimes called under the locked socket (in sco_connect and __sco_sock_close), cancel_delayed_work_sync might cause them to sleep until an sco_sock_timeout that has started finishes running. But sco_sock_timeout would also sleep until it can grab the lock_sock. Using cancel_delayed_work is fine because sco_sock_timeout does not change from run to run, hence there is no functional difference between: 1. waiting for a timeout to finish running before scheduling another timeout 2. scheduling another timeout while a timeout is running. Link: https://syzkaller.appspot.com/bug?id=9089d89de0502e120f234ca0fc8a703f7368b31e [1] Reported-by: syzbot+2f6d7c28bb4bf7e82060@syzkaller.appspotmail.com Tested-by: syzbot+2f6d7c28bb4bf7e82060@syzkaller.appspotmail.com Signed-off-by: Desmond Cheong Zhi Xi Signed-off-by: Luiz Augusto von Dentz Signed-off-by: Sasha Levin

Bluetooth: Fix not generating RPA when required

2021-09-18T11:43:45+00:00

[ Upstream commit c45074d68a9b1e893d86520af71fab37693c3d7e ] Code was checking if random_addr and hdev->rpa match without first checking if the RPA has not been set (BDADDR_ANY), furthermore it was clearing HCI_RPA_EXPIRED before the command completes and the RPA is actually programmed which in case of failure would leave the expired RPA still set. Since advertising instance have a similar problem the clearing of HCI_RPA_EXPIRED has been moved to hci_event.c after checking the random address is in fact the hdev->rap and then proceed to set the expire timeout. Signed-off-by: Luiz Augusto von Dentz Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: skip invalid hci_sync_conn_complete_evt

2021-09-18T11:43:44+00:00

[ Upstream commit 92fe24a7db751b80925214ede43f8d2be792ea7b ] Syzbot reported a corrupted list in kobject_add_internal [1]. This happens when multiple HCI_EV_SYNC_CONN_COMPLETE event packets with status 0 are sent for the same HCI connection. This causes us to register the device more than once which corrupts the kset list. As this is forbidden behavior, we add a check for whether we're trying to process the same HCI_EV_SYNC_CONN_COMPLETE event multiple times for one connection. If that's the case, the event is invalid, so we report an error that the device is misbehaving, and ignore the packet. Link: https://syzkaller.appspot.com/bug?extid=66264bf2fd0476be7e6c [1] Reported-by: syzbot+66264bf2fd0476be7e6c@syzkaller.appspotmail.com Tested-by: syzbot+66264bf2fd0476be7e6c@syzkaller.appspotmail.com Signed-off-by: Desmond Cheong Zhi Xi Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: add timeout sanity check to hci_inquiry

2021-09-15T08:02:25+00:00

[ Upstream commit f41a4b2b5eb7872109723dab8ae1603bdd9d9ec1 ] Syzbot hit "task hung" bug in hci_req_sync(). The problem was in unreasonable huge inquiry timeout passed from userspace. Fix it by adding sanity check for timeout value to hci_inquiry(). Since hci_inquiry() is the only user of hci_req_sync() with user controlled timeout value, it makes sense to check timeout value in hci_inquiry() and don't touch hci_req_sync(). Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-and-tested-by: syzbot+be2baed593ea56c6a84c@syzkaller.appspotmail.com Signed-off-by: Pavel Skripkin Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: Move shutdown callback before flushing tx and rx queue

2021-09-15T08:02:23+00:00

[ Upstream commit 0ea53674d07fb6db2dd7a7ec2fdc85a12eb246c2 ] Commit 0ea9fd001a14 ("Bluetooth: Shutdown controller after workqueues are flushed or cancelled") introduced a regression that makes mtkbtsdio driver stops working: [ 36.593956] Bluetooth: hci0: Firmware already downloaded [ 46.814613] Bluetooth: hci0: Execution of wmt command timed out [ 46.814619] Bluetooth: hci0: Failed to send wmt func ctrl (-110) The shutdown callback depends on the result of hdev->rx_work, so we should call it before flushing rx_work: -> btmtksdio_shutdown() -> mtk_hci_wmt_sync() -> __hci_cmd_send() -> wait for BTMTKSDIO_TX_WAIT_VND_EVT gets cleared -> btmtksdio_recv_event() -> hci_recv_frame() -> queue_work(hdev->workqueue, &hdev->rx_work) -> clears BTMTKSDIO_TX_WAIT_VND_EVT So move the shutdown callback before flushing TX/RX queue to resolve the issue. Reported-and-tested-by: Mattijs Korpershoek Tested-by: Hsin-Yi Wang Cc: Guenter Roeck Fixes: 0ea9fd001a14 ("Bluetooth: Shutdown controller after workqueues are flushed or cancelled") Signed-off-by: Kai-Heng Feng Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: fix repeated calls to sco_sock_kill

2021-09-15T08:02:21+00:00

[ Upstream commit e1dee2c1de2b4dd00eb44004a4bda6326ed07b59 ] In commit 4e1a720d0312 ("Bluetooth: avoid killing an already killed socket"), a check was added to sco_sock_kill to skip killing a socket if the SOCK_DEAD flag was set. This was done after a trace for a use-after-free bug showed that the same sock pointer was being killed twice. Unfortunately, this check prevents sco_sock_kill from running on any socket. sco_sock_kill kills a socket only if it's zapped and orphaned, however sock_orphan announces that the socket is dead before detaching it. i.e., orphaned sockets have the SOCK_DEAD flag set. To fix this, we remove the check for SOCK_DEAD, and avoid repeated calls to sco_sock_kill by removing incorrect calls in: 1. sco_sock_timeout. The socket should not be killed on timeout as further processing is expected to be done. For example, sco_sock_connect sets the timer then waits for the socket to be connected or for an error to be returned. 2. sco_conn_del. This function should clean up resources for the connection, but the socket itself should be cleaned up in sco_sock_release. 3. sco_sock_close. Calls to sco_sock_close in sco_sock_cleanup_listen and sco_sock_release are followed by sco_sock_kill. Hence the duplicated call should be removed. Fixes: 4e1a720d0312 ("Bluetooth: avoid killing an already killed socket") Signed-off-by: Desmond Cheong Zhi Xi Signed-off-by: Luiz Augusto von Dentz Signed-off-by: Sasha Levin

Bluetooth: increase BTNAMSIZ to 21 chars to fix potential buffer overflow

2021-09-15T08:02:19+00:00

[ Upstream commit 713baf3dae8f45dc8ada4ed2f5fdcbf94a5c274d ] An earlier commit replaced using batostr to using %pMR sprintf for the construction of session->name. Static analysis detected that this new method can use a total of 21 characters (including the trailing '\0') so we need to increase the BTNAMSIZ from 18 to 21 to fix potential buffer overflows. Addresses-Coverity: ("Out-of-bounds write") Fixes: fcb73338ed53 ("Bluetooth: Use %pMR in sprintf/seq_printf instead of batostr") Signed-off-by: Colin Ian King Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin