============================================
WARNING: possible recursive locking detected
syzkaller #0 Not tainted
--------------------------------------------
kworker/0:8/1895 is trying to acquire lock:
ffff88818ad65558 (&qdisc_xmit_lock_key#4){+.-.}-{3:3}, at: sch_direct_xmit+0x15f/0x4c0

but task is already holding lock:
ffff88802d68b558 (&qdisc_xmit_lock_key#4){+.-.}-{3:3}, at: sch_direct_xmit+0x15f/0x4c0

other info that might help us debug this:
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(&qdisc_xmit_lock_key#4);
  lock(&qdisc_xmit_lock_key#4);

 *** DEADLOCK ***

 May be due to missing lock nesting notation

15 locks held by kworker/0:8/1895:
 #0: ffff88816d8b0548 ((wq_completion)mld){+.+.}-{0:0}, at: process_scheduled_works+0x9ea/0x1830
 #1: ffffc90021bf7c40 ((work_completion)(&(&idev->mc_ifc_work)->work)){+.+.}-{0:0}, at: process_scheduled_works+0xa25/0x1830
 #2: ffff88819409d538 (&idev->mc_lock){+.+.}-{4:4}, at: mld_ifc_work+0x2d/0xe70
 #3: ffffffff8e760320 (rcu_read_lock){....}-{1:3}, at: mld_sendpack+0x213/0xe40
 #4: ffffffff8e760320 (rcu_read_lock){....}-{1:3}, at: ip6_output+0x126/0x550
 #5: ffffffff8e760380 (rcu_read_lock_bh){....}-{1:3}, at: __dev_queue_xmit+0x274/0x38a0
 #6: ffffffff8e760320 (rcu_read_lock){....}-{1:3}, at: bond_start_xmit+0xc9/0x1960
 #7: ffffffff8e760380 (rcu_read_lock_bh){....}-{1:3}, at: __dev_queue_xmit+0x274/0x38a0
 #8: ffff8880219b6228 (dev->qdisc_tx_busylock ?: &qdisc_tx_busylock#5){+...}-{3:3}, at: __dev_queue_xmit+0x1002/0x38a0
 #9: ffff88802d68b558 (&qdisc_xmit_lock_key#4){+.-.}-{3:3}, at: sch_direct_xmit+0x15f/0x4c0
 #10: ffffffff8e760320 (rcu_read_lock){....}-{1:3}, at: ip_output+0x5b/0x450
 #11: ffffffff8e760320 (rcu_read_lock){....}-{1:3}, at: ip_finish_output2+0x3c2/0x1070
 #12: ffffffff8e760320 (rcu_read_lock){....}-{1:3}, at: arp_xmit+0x23/0x270
 #13: ffffffff8e760380 (rcu_read_lock_bh){....}-{1:3}, at: __dev_queue_xmit+0x274/0x38a0
 #14: ffff88818edb3228 (dev->qdisc_tx_busylock ?: &qdisc_tx_busylock#5){+...}-{3:3}, at: __dev_queue_xmit+0x1002/0x38a0

stack backtrace:
CPU: 0 UID: 0 PID: 1895 Comm: kworker/0:8 Not tainted syzkaller #0 PREEMPT(full) 
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
Workqueue: mld mld_ifc_work
Call Trace:
 <TASK>
 dump_stack_lvl+0xe8/0x150
 print_deadlock_bug+0x279/0x290
 __lock_acquire+0x253f/0x2cf0
 lock_acquire+0xf0/0x2e0
 _raw_spin_lock+0x2e/0x40
 sch_direct_xmit+0x15f/0x4c0
 __dev_queue_xmit+0x1538/0x38a0
 NF_HOOK+0x33a/0x3c0
 arp_xmit+0x16c/0x270
 arp_solicit+0xbe0/0xe40
 __neigh_event_send+0xf05/0x1530
 neigh_resolve_output+0x198/0x750
 ip_finish_output2+0xca9/0x1070
 ip_output+0x29f/0x450
 iptunnel_xmit+0x59f/0xa00
 ip_tunnel_xmit+0x193a/0x1f20
 __gre_xmit+0x19e/0x240
 erspan_xmit+0xa81/0x1670
 dev_hard_start_xmit+0x2d8/0x870
 sch_direct_xmit+0x251/0x4c0
 __dev_queue_xmit+0x1538/0x38a0
 bond_start_xmit+0x77d/0x1960
 dev_hard_start_xmit+0x2d8/0x870
 __dev_queue_xmit+0x168f/0x38a0
 ip6_output+0x340/0x550
 NF_HOOK+0xa2/0x3a0
 mld_sendpack+0x8b4/0xe40
 mld_ifc_work+0x835/0xe70
 process_scheduled_works+0xb02/0x1830
 worker_thread+0xa50/0xfc0
 kthread+0x388/0x470
 ret_from_fork+0x51e/0xb90
 ret_from_fork_asm+0x1a/0x30
 </TASK>
