| Message ID | 20240730152210.25153-1-aha310510@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net] rtnetlink: fix possible deadlock in team_port_change_check | expand |
On Wed, 31 Jul 2024 00:22:10 +0900 Jeongjun Park wrote: > do_setlink() changes the flag of the device and then enslaves it. However, > in this case, if the IFF_UP flag is set, the enslavement process calls > team_add_slave() to acquire 'team->lock', but when dev_open() opens the > newly enslaved device, the NETDEV_UP event occurs, and as a result, > a deadlock occurs when team_port_change_check() tries to acquire > 'team->lock' again. You can't change behavior like this, see ec4ffd100ffb396ec
> 2024. 7. 31. 오전 8:59, Jakub Kicinski <kuba@kernel.org> 작성: > > On Wed, 31 Jul 2024 00:22:10 +0900 Jeongjun Park wrote: >> do_setlink() changes the flag of the device and then enslaves it. However, >> in this case, if the IFF_UP flag is set, the enslavement process calls >> team_add_slave() to acquire 'team->lock', but when dev_open() opens the >> newly enslaved device, the NETDEV_UP event occurs, and as a result, >> a deadlock occurs when team_port_change_check() tries to acquire >> 'team->lock' again. > > You can't change behavior like this, see ec4ffd100ffb396ec Okay. In that case, I'll write a patch that modifies the code of the team driver and send it back to you. Regards, Jeongjun Park
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 87e67194f240..178f5b85fd87 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -2896,13 +2896,6 @@ static int do_setlink(const struct sk_buff *skb, call_netdevice_notifiers(NETDEV_CHANGEADDR, dev); } - if (ifm->ifi_flags || ifm->ifi_change) { - err = dev_change_flags(dev, rtnl_dev_combine_flags(dev, ifm), - extack); - if (err < 0) - goto errout; - } - if (tb[IFLA_MASTER]) { err = do_set_master(dev, nla_get_u32(tb[IFLA_MASTER]), extack); if (err) @@ -2910,6 +2903,13 @@ static int do_setlink(const struct sk_buff *skb, status |= DO_SETLINK_MODIFIED; } + if (ifm->ifi_flags || ifm->ifi_change) { + err = dev_change_flags(dev, rtnl_dev_combine_flags(dev, ifm), + extack); + if (err < 0) + goto errout; + } + if (tb[IFLA_CARRIER]) { err = dev_change_carrier(dev, nla_get_u8(tb[IFLA_CARRIER])); if (err)
do_setlink() changes the flag of the device and then enslaves it. However, in this case, if the IFF_UP flag is set, the enslavement process calls team_add_slave() to acquire 'team->lock', but when dev_open() opens the newly enslaved device, the NETDEV_UP event occurs, and as a result, a deadlock occurs when team_port_change_check() tries to acquire 'team->lock' again. To solve this, you need to enslave it before changing the flag of the device. ============================================ WARNING: possible recursive locking detected 6.10.0-syzkaller-12562-g1722389b0d86 #0 Not tainted -------------------------------------------- syz-executor122/5360 is trying to acquire lock: ffff88802c258d40 (team->team_lock_key){+.+.}-{3:3}, at: team_port_change_check drivers/net/team/team_core.c:2950 [inline] ffff88802c258d40 (team->team_lock_key){+.+.}-{3:3}, at: team_device_event+0x2c7/0x770 drivers/net/team/team_core.c:2973 but task is already holding lock: ffff88802c258d40 (team->team_lock_key){+.+.}-{3:3}, at: team_add_slave+0x9c/0x20e0 drivers/net/team/team_core.c:1975 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(team->team_lock_key); lock(team->team_lock_key); *** DEADLOCK *** May be due to missing lock nesting notation 2 locks held by syz-executor122/5360: #0: ffffffff8fa1e9a8 (rtnl_mutex){+.+.}-{3:3}, at: rtnl_lock net/core/rtnetlink.c:79 [inline] #0: ffffffff8fa1e9a8 (rtnl_mutex){+.+.}-{3:3}, at: rtnetlink_rcv_msg+0x372/0xea0 net/core/rtnetlink.c:6644 #1: ffff88802c258d40 (team->team_lock_key){+.+.}-{3:3}, at: team_add_slave+0x9c/0x20e0 drivers/net/team/team_core.c:1975 stack backtrace: CPU: 0 UID: 0 PID: 5360 Comm: syz-executor122 Not tainted 6.10.0-syzkaller-12562-g1722389b0d86 #0 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014 Call Trace: <TASK> __dump_stack lib/dump_stack.c:93 [inline] dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:119 check_deadlock kernel/locking/lockdep.c:3061 [inline] validate_chain kernel/locking/lockdep.c:3855 [inline] __lock_acquire+0x2167/0x3cb0 kernel/locking/lockdep.c:5142 lock_acquire kernel/locking/lockdep.c:5759 [inline] lock_acquire+0x1b1/0x560 kernel/locking/lockdep.c:5724 __mutex_lock_common kernel/locking/mutex.c:608 [inline] __mutex_lock+0x175/0x9c0 kernel/locking/mutex.c:752 team_port_change_check drivers/net/team/team_core.c:2950 [inline] team_device_event+0x2c7/0x770 drivers/net/team/team_core.c:2973 notifier_call_chain+0xb9/0x410 kernel/notifier.c:93 call_netdevice_notifiers_info+0xbe/0x140 net/core/dev.c:1994 call_netdevice_notifiers_extack net/core/dev.c:2032 [inline] call_netdevice_notifiers net/core/dev.c:2046 [inline] __dev_notify_flags+0x12d/0x2e0 net/core/dev.c:8876 dev_change_flags+0x10c/0x160 net/core/dev.c:8914 vlan_device_event+0xdfc/0x2120 net/8021q/vlan.c:468 notifier_call_chain+0xb9/0x410 kernel/notifier.c:93 call_netdevice_notifiers_info+0xbe/0x140 net/core/dev.c:1994 call_netdevice_notifiers_extack net/core/dev.c:2032 [inline] call_netdevice_notifiers net/core/dev.c:2046 [inline] dev_open net/core/dev.c:1515 [inline] dev_open+0x144/0x160 net/core/dev.c:1503 team_port_add drivers/net/team/team_core.c:1216 [inline] team_add_slave+0xacd/0x20e0 drivers/net/team/team_core.c:1976 do_set_master+0x1bc/0x230 net/core/rtnetlink.c:2701 do_setlink+0xcaf/0x3ff0 net/core/rtnetlink.c:2907 __rtnl_newlink+0xc35/0x1960 net/core/rtnetlink.c:3696 rtnl_newlink+0x67/0xa0 net/core/rtnetlink.c:3743 rtnetlink_rcv_msg+0x3c7/0xea0 net/core/rtnetlink.c:6647 netlink_rcv_skb+0x16b/0x440 net/netlink/af_netlink.c:2550 netlink_unicast_kernel net/netlink/af_netlink.c:1331 [inline] netlink_unicast+0x544/0x830 net/netlink/af_netlink.c:1357 netlink_sendmsg+0x8b8/0xd70 net/netlink/af_netlink.c:1901 sock_sendmsg_nosec net/socket.c:730 [inline] __sock_sendmsg net/socket.c:745 [inline] ____sys_sendmsg+0xab5/0xc90 net/socket.c:2597 ___sys_sendmsg+0x135/0x1e0 net/socket.c:2651 __sys_sendmsg+0x117/0x1f0 net/socket.c:2680 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7f424ca7e7b9 Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 31 1a 00 00 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 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffd8c496978 EFLAGS: 00000246 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 00007f424ca7e7b9 RDX: 0000000000000000 RSI: 0000000020000600 RDI: 0000000000000012 RBP: 0000000000000000 R08: 0000000000000001 R09: 0000000000000001 R10: 0000000000000001 R11: 0000000000000246 R12: 0000000000000000 R13: 0000000000000000 R14: 0000000000000000 R15: 00007ffd8c4969a0 Reported-by: syzbot+b668da2bc4cb9670bf58@syzkaller.appspotmail.com Fixes: ec4ffd100ffb ("Revert "net: rtnetlink: Enslave device before bringing it up"") Signed-off-by: Jeongjun Park <aha310510@gmail.com> --- net/core/rtnetlink.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) --