kernel/linux.git/net, branch v5.4.7

can: j1939: j1939_sk_bind(): take priv after lock is held

2019-12-31T15:45:56+00:00

commit 00d4e14d2e4caf5f7254a505fee5eeca8cd37bd4 upstream. syzbot reproduced following crash: =============================================================================== kasan: CONFIG_KASAN_INLINE enabled kasan: GPF could be caused by NULL-ptr deref or user memory access general protection fault: 0000 [#1] PREEMPT SMP KASAN CPU: 0 PID: 9844 Comm: syz-executor.0 Not tainted 5.4.0-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:__lock_acquire+0x1254/0x4a00 kernel/locking/lockdep.c:3828 Code: 00 0f 85 96 24 00 00 48 81 c4 f0 00 00 00 5b 41 5c 41 5d 41 5e 41 5f 5d c3 48 b8 00 00 00 00 00 fc ff df 4c 89 f2 48 c1 ea 03 <80> 3c 02 00 0f 85 0b 28 00 00 49 81 3e 20 19 78 8a 0f 84 5f ee ff RSP: 0018:ffff888099c3fb48 EFLAGS: 00010006 RAX: dffffc0000000000 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000218 RSI: 0000000000000000 RDI: 0000000000000001 RBP: ffff888099c3fc60 R08: 0000000000000001 R09: 0000000000000001 R10: fffffbfff146e1d0 R11: ffff888098720400 R12: 00000000000010c0 R13: 0000000000000000 R14: 00000000000010c0 R15: 0000000000000000 FS: 00007f0559e98700(0000) GS:ffff8880ae800000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fe4d89e0000 CR3: 0000000099606000 CR4: 00000000001406f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: lock_acquire+0x190/0x410 kernel/locking/lockdep.c:4485 __raw_spin_lock_bh include/linux/spinlock_api_smp.h:135 [inline] _raw_spin_lock_bh+0x33/0x50 kernel/locking/spinlock.c:175 spin_lock_bh include/linux/spinlock.h:343 [inline] j1939_jsk_del+0x32/0x210 net/can/j1939/socket.c:89 j1939_sk_bind+0x2ea/0x8f0 net/can/j1939/socket.c:448 __sys_bind+0x239/0x290 net/socket.c:1648 __do_sys_bind net/socket.c:1659 [inline] __se_sys_bind net/socket.c:1657 [inline] __x64_sys_bind+0x73/0xb0 net/socket.c:1657 do_syscall_64+0xfa/0x790 arch/x86/entry/common.c:294 entry_SYSCALL_64_after_hwframe+0x49/0xbe RIP: 0033:0x45a679 Code: ad b6 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 7b b6 fb ff c3 66 2e 0f 1f 84 00 00 00 00 RSP: 002b:00007f0559e97c78 EFLAGS: 00000246 ORIG_RAX: 0000000000000031 RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 000000000045a679 RDX: 0000000000000018 RSI: 0000000020000240 RDI: 0000000000000003 RBP: 000000000075bf20 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 00007f0559e986d4 R13: 00000000004c09e9 R14: 00000000004d37d0 R15: 00000000ffffffff Modules linked in: ------------[ cut here ]------------ WARNING: CPU: 0 PID: 9844 at kernel/locking/mutex.c:1419 mutex_trylock+0x279/0x2f0 kernel/locking/mutex.c:1427 =============================================================================== This issues was caused by null pointer deference. Where j1939_sk_bind() was using currently not existing priv. Possible scenario may look as following: cpu0 cpu1 bind() bind() j1939_sk_bind() j1939_sk_bind() priv = jsk->priv; priv = jsk->priv; lock_sock(sock->sk); priv = j1939_netdev_start(ndev); j1939_jsk_add(priv, jsk); jsk->priv = priv; relase_sock(sock->sk); lock_sock(sock->sk); j1939_jsk_del(priv, jsk); ..... ooops ...... With this patch we move "priv = jsk->priv;" after the lock, to avoid assigning of wrong priv pointer. Reported-by: syzbot+99e9e1b200a1e363237d@syzkaller.appspotmail.com Fixes: 9d71dd0c7009 ("can: add support of SAE J1939 protocol") Signed-off-by: Oleksij Rempel Cc: linux-stable # >= v5.4 Signed-off-by: Marc Kleine-Budde Signed-off-by: Greg Kroah-Hartman

mac80211: consider QoS Null frames for STA_NULLFUNC_ACKED

2019-12-31T15:45:44+00:00

[ Upstream commit 08a5bdde3812993cb8eb7aa9124703df0de28e4b ] Commit 7b6ddeaf27ec ("mac80211: use QoS NDP for AP probing") let STAs send QoS Null frames as PS triggers if the AP was a QoS STA. However, the mac80211 PS stack relies on an interface flag IEEE80211_STA_NULLFUNC_ACKED for determining trigger frame ACK, which was not being set for acked non-QoS Null frames. The effect is an inability to trigger hardware sleep via IEEE80211_CONF_PS since the QoS Null frame was seemingly never acked. This bug only applies to drivers which set both IEEE80211_HW_REPORTS_TX_ACK_STATUS and IEEE80211_HW_PS_NULLFUNC_STACK. Detect the acked QoS Null frame to restore STA power save. Fixes: 7b6ddeaf27ec ("mac80211: use QoS NDP for AP probing") Signed-off-by: Thomas Pedersen Link: https://lore.kernel.org/r/20191119053538.25979-4-thomas@adapt-ip.com Signed-off-by: Johannes Berg Signed-off-by: Sasha Levin

Bluetooth: Fix advertising duplicated flags

2019-12-31T15:44:33+00:00

[ Upstream commit 6012b9346d8959194c239fd60a62dfec98d43048 ] Instances may have flags set as part of its data in which case the code should not attempt to add it again otherwise it can cause duplication: < HCI Command: LE Set Extended Advertising Data (0x08|0x0037) plen 35 Handle: 0x00 Operation: Complete extended advertising data (0x03) Fragment preference: Minimize fragmentation (0x01) Data length: 0x06 Flags: 0x04 BR/EDR Not Supported Flags: 0x06 LE General Discoverable Mode BR/EDR Not Supported Signed-off-by: Luiz Augusto von Dentz Signed-off-by: Johan Hedberg Signed-off-by: Sasha Levin

Bluetooth: hci_core: fix init for HCI_USER_CHANNEL

2019-12-31T15:44:09+00:00

[ Upstream commit eb8c101e28496888a0dcfe16ab86a1bee369e820 ] During the setup() stage, HCI device drivers expect the chip to acknowledge its setup() completion via vendor specific frames. If userspace opens() such HCI device in HCI_USER_CHANNEL [1] mode, the vendor specific frames are never tranmitted to the driver, as they are filtered in hci_rx_work(). Allow HCI devices which operate in HCI_USER_CHANNEL mode to receive frames if the HCI device is is HCI_INIT state. [1] https://www.spinics.net/lists/linux-bluetooth/msg37345.html Fixes: 23500189d7e0 ("Bluetooth: Introduce new HCI socket channel for user operation") Signed-off-by: Mattijs Korpershoek Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: Workaround directed advertising bug in Broadcom controllers

2019-12-31T15:44:08+00:00

[ Upstream commit 4c371bb95cf06ded80df0e6139fdd77cee1d9a94 ] It appears that some Broadcom controllers (eg BCM20702A0) reject LE Set Advertising Parameters command if advertising intervals provided are not within range for undirected and low duty directed advertising. Workaround this bug by populating min and max intervals with 'valid' values. < HCI Command: LE Set Advertising Parameters (0x08|0x0006) plen 15 Min advertising interval: 0.000 msec (0x0000) Max advertising interval: 0.000 msec (0x0000) Type: Connectable directed - ADV_DIRECT_IND (high duty cycle) (0x01) Own address type: Public (0x00) Direct address type: Random (0x01) Direct address: E2:F0:7B:9F:DC:F4 (Static) Channel map: 37, 38, 39 (0x07) Filter policy: Allow Scan Request from Any, Allow Connect Request from Any (0x00) > HCI Event: Command Complete (0x0e) plen 4 LE Set Advertising Parameters (0x08|0x0006) ncmd 1 Status: Invalid HCI Command Parameters (0x12) Signed-off-by: Szymon Janc Tested-by: Sören Beye Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

Bluetooth: missed cpu_to_le16 conversion in hci_init4_req

2019-12-31T15:44:08+00:00

[ Upstream commit 727ea61a5028f8ac96f75ab34cb1b56e63fd9227 ] It looks like in hci_init4_req() the request is being initialised from cpu-endian data but the packet is specified to be little-endian. This causes an warning from sparse due to __le16 to u16 conversion. Fix this by using cpu_to_le16() on the two fields in the packet. net/bluetooth/hci_core.c:845:27: warning: incorrect type in assignment (different base types) net/bluetooth/hci_core.c:845:27: expected restricted __le16 [usertype] tx_len net/bluetooth/hci_core.c:845:27: got unsigned short [usertype] le_max_tx_len net/bluetooth/hci_core.c:846:28: warning: incorrect type in assignment (different base types) net/bluetooth/hci_core.c:846:28: expected restricted __le16 [usertype] tx_time net/bluetooth/hci_core.c:846:28: got unsigned short [usertype] le_max_tx_time Signed-off-by: Ben Dooks Signed-off-by: Marcel Holtmann Signed-off-by: Sasha Levin

net/smc: increase device refcount for added link group

2019-12-31T15:43:53+00:00

[ Upstream commit b3cb53c05f20c5b4026a36a7bbd3010d1f3e0a55 ] SMCD link groups belong to certain ISM-devices and SMCR link group links belong to certain IB-devices. Increase the refcount for these devices, as long as corresponding link groups exist. Signed-off-by: Ursula Braun Signed-off-by: Karsten Graul Signed-off-by: Jakub Kicinski Signed-off-by: Sasha Levin

neighbour: remove neigh_cleanup() method

2019-12-31T15:41:37+00:00

[ Upstream commit f394722fb0d0f701119368959d7cd0ecbc46363a ] neigh_cleanup() has not been used for seven years, and was a wrong design. Messing with shared pointer in bond_neigh_init() without proper memory barriers would at least trigger syzbot complains eventually. It is time to remove this stuff. Fixes: b63b70d87741 ("IPoIB: Use a private hash table for path lookup in xmit path") Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

sctp: fully initialize v4 addr in some functions

2019-12-31T15:41:35+00:00

[ Upstream commit b6f3320b1d5267e7b583a6d0c88dda518101740c ] Syzbot found a crash: BUG: KMSAN: uninit-value in crc32_body lib/crc32.c:112 [inline] BUG: KMSAN: uninit-value in crc32_le_generic lib/crc32.c:179 [inline] BUG: KMSAN: uninit-value in __crc32c_le_base+0x4fa/0xd30 lib/crc32.c:202 Call Trace: crc32_body lib/crc32.c:112 [inline] crc32_le_generic lib/crc32.c:179 [inline] __crc32c_le_base+0x4fa/0xd30 lib/crc32.c:202 chksum_update+0xb2/0x110 crypto/crc32c_generic.c:90 crypto_shash_update+0x4c5/0x530 crypto/shash.c:107 crc32c+0x150/0x220 lib/libcrc32c.c:47 sctp_csum_update+0x89/0xa0 include/net/sctp/checksum.h:36 __skb_checksum+0x1297/0x12a0 net/core/skbuff.c:2640 sctp_compute_cksum include/net/sctp/checksum.h:59 [inline] sctp_packet_pack net/sctp/output.c:528 [inline] sctp_packet_transmit+0x40fb/0x4250 net/sctp/output.c:597 sctp_outq_flush_transports net/sctp/outqueue.c:1146 [inline] sctp_outq_flush+0x1823/0x5d80 net/sctp/outqueue.c:1194 sctp_outq_uncork+0xd0/0xf0 net/sctp/outqueue.c:757 sctp_cmd_interpreter net/sctp/sm_sideeffect.c:1781 [inline] sctp_side_effects net/sctp/sm_sideeffect.c:1184 [inline] sctp_do_sm+0x8fe1/0x9720 net/sctp/sm_sideeffect.c:1155 sctp_primitive_REQUESTHEARTBEAT+0x175/0x1a0 net/sctp/primitive.c:185 sctp_apply_peer_addr_params+0x212/0x1d40 net/sctp/socket.c:2433 sctp_setsockopt_peer_addr_params net/sctp/socket.c:2686 [inline] sctp_setsockopt+0x189bb/0x19090 net/sctp/socket.c:4672 The issue was caused by transport->ipaddr set with uninit addr param, which was passed by: sctp_transport_init net/sctp/transport.c:47 [inline] sctp_transport_new+0x248/0xa00 net/sctp/transport.c:100 sctp_assoc_add_peer+0x5ba/0x2030 net/sctp/associola.c:611 sctp_process_param net/sctp/sm_make_chunk.c:2524 [inline] where 'addr' is set by sctp_v4_from_addr_param(), and it doesn't initialize the padding of addr->v4. Later when calling sctp_make_heartbeat(), hbinfo.daddr(=transport->ipaddr) will become the part of skb, and the issue occurs. This patch is to fix it by initializing the padding of addr->v4 in sctp_v4_from_addr_param(), as well as other functions that do the similar thing, and these functions shouldn't trust that the caller initializes the memory, as Marcelo suggested. Reported-by: syzbot+6dcbfea81cd3d4dd0b02@syzkaller.appspotmail.com Signed-off-by: Xin Long Acked-by: Neil Horman Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

sctp: fix memleak on err handling of stream initialization

2019-12-31T15:41:34+00:00

[ Upstream commit 951c6db954a1adefab492f6da805decacabbd1a7 ] syzbot reported a memory leak when an allocation fails within genradix_prealloc() for output streams. That's because genradix_prealloc() leaves initialized members initialized when the issue happens and SCTP stack will abort the current initialization but without cleaning up such members. The fix here is to always call genradix_free() when genradix_prealloc() fails, for output and also input streams, as it suffers from the same issue. Reported-by: syzbot+772d9e36c490b18d51d1@syzkaller.appspotmail.com Fixes: 2075e50caf5e ("sctp: convert to genradix") Signed-off-by: Marcelo Ricardo Leitner Tested-by: Xin Long Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman