qlcnic: fix deadlock bug

The following soft lockup was caught. This is a deadlock caused by
recusive locking.

Process kworker/u40:1:28016 was holding spin lock "mbx->queue_lock" in
qlcnic_83xx_mailbox_worker(), while a softirq came in and ask the same spin
lock in qlcnic_83xx_enqueue_mbx_cmd(). This lock should be hold by disable
bh..

[161846.962125] NMI watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [kworker/u40:1:28016]
[161846.962367] Modules linked in: tun ocfs2 xen_netback xen_blkback xen_gntalloc xen_gntdev xen_evtchn xenfs xen_privcmd autofs4 ocfs2_dlmfs ocfs2_stack_o2cb ocfs2_dlm ocfs2_nodemanager ocfs2_stackglue configfs bnx2fc fcoe libfcoe libfc sunrpc 8021q mrp garp bridge stp llc bonding dm_round_robin dm_multipath iTCO_wdt iTCO_vendor_support pcspkr sb_edac edac_core i2c_i801 shpchp lpc_ich mfd_core ioatdma ipmi_devintf ipmi_si ipmi_msghandler sg ext4 jbd2 mbcache2 sr_mod cdrom sd_mod igb i2c_algo_bit i2c_core ahci libahci megaraid_sas ixgbe dca ptp pps_core vxlan udp_tunnel ip6_udp_tunnel qla2xxx scsi_transport_fc qlcnic crc32c_intel be2iscsi bnx2i cnic uio cxgb4i cxgb4 cxgb3i libcxgbi ipv6 cxgb3 mdio libiscsi_tcp qla4xxx iscsi_boot_sysfs libiscsi scsi_transport_iscsi dm_mirror dm_region_hash dm_log dm_mod
[161846.962454]
[161846.962460] CPU: 1 PID: 28016 Comm: kworker/u40:1 Not tainted 4.1.12-94.5.9.el6uek.x86_64 #2
[161846.962463] Hardware name: Oracle Corporation SUN SERVER X4-2L      /ASSY,MB,X4-2L         , BIOS 26050100 09/19/2017
[161846.962489] Workqueue: qlcnic_mailbox qlcnic_83xx_mailbox_worker [qlcnic]
[161846.962493] task: ffff8801f2e34600 ti: ffff88004ca5c000 task.ti: ffff88004ca5c000
[161846.962496] RIP: e030:[<ffffffff810013aa>]  [<ffffffff810013aa>] xen_hypercall_sched_op+0xa/0x20
[161846.962506] RSP: e02b:ffff880202e43388  EFLAGS: 00000206
[161846.962509] RAX: 0000000000000000 RBX: ffff8801f6996b70 RCX: ffffffff810013aa
[161846.962511] RDX: ffff880202e433cc RSI: ffff880202e433b0 RDI: 0000000000000003
[161846.962513] RBP: ffff880202e433d0 R08: 0000000000000000 R09: ffff8801fe893200
[161846.962516] R10: ffff8801fe400538 R11: 0000000000000206 R12: ffff880202e4b000
[161846.962518] R13: 0000000000000050 R14: 0000000000000001 R15: 000000000000020d
[161846.962528] FS:  0000000000000000(0000) GS:ffff880202e40000(0000) knlGS:ffff880202e40000
[161846.962531] CS:  e033 DS: 0000 ES: 0000 CR0: 0000000080050033
[161846.962533] CR2: 0000000002612640 CR3: 00000001bb796000 CR4: 0000000000042660
[161846.962536] Stack:
[161846.962538]  ffff880202e43608 0000000000000000 ffffffff813f0442 ffff880202e433b0
[161846.962543]  0000000000000000 ffff880202e433cc ffffffff00000001 0000000000000000
[161846.962547]  00000009813f03d6 ffff880202e433e0 ffffffff813f0460 ffff880202e43440
[161846.962552] Call Trace:
[161846.962555]  <IRQ>
[161846.962565]  [<ffffffff813f0442>] ? xen_poll_irq_timeout+0x42/0x50
[161846.962570]  [<ffffffff813f0460>] xen_poll_irq+0x10/0x20
[161846.962578]  [<ffffffff81014222>] xen_lock_spinning+0xe2/0x110
[161846.962583]  [<ffffffff81013f01>] __raw_callee_save_xen_lock_spinning+0x11/0x20
[161846.962592]  [<ffffffff816e5c57>] ? _raw_spin_lock+0x57/0x80
[161846.962609]  [<ffffffffa028acfc>] qlcnic_83xx_enqueue_mbx_cmd+0x7c/0xe0 [qlcnic]
[161846.962623]  [<ffffffffa028e008>] qlcnic_83xx_issue_cmd+0x58/0x210 [qlcnic]
[161846.962636]  [<ffffffffa028caf2>] qlcnic_83xx_sre_macaddr_change+0x162/0x1d0 [qlcnic]
[161846.962649]  [<ffffffffa028cb8b>] qlcnic_83xx_change_l2_filter+0x2b/0x30 [qlcnic]
[161846.962657]  [<ffffffff8160248b>] ? __skb_flow_dissect+0x18b/0x650
[161846.962670]  [<ffffffffa02856e5>] qlcnic_send_filter+0x205/0x250 [qlcnic]
[161846.962682]  [<ffffffffa0285c77>] qlcnic_xmit_frame+0x547/0x7b0 [qlcnic]
[161846.962691]  [<ffffffff8160ac22>] xmit_one+0x82/0x1a0
[161846.962696]  [<ffffffff8160ad90>] dev_hard_start_xmit+0x50/0xa0
[161846.962701]  [<ffffffff81630112>] sch_direct_xmit+0x112/0x220
[161846.962706]  [<ffffffff8160b80f>] __dev_queue_xmit+0x1df/0x5e0
[161846.962710]  [<ffffffff8160bc33>] dev_queue_xmit_sk+0x13/0x20
[161846.962721]  [<ffffffffa0575bd5>] bond_dev_queue_xmit+0x35/0x80 [bonding]
[161846.962729]  [<ffffffffa05769fb>] __bond_start_xmit+0x1cb/0x210 [bonding]
[161846.962736]  [<ffffffffa0576a71>] bond_start_xmit+0x31/0x60 [bonding]
[161846.962740]  [<ffffffff8160ac22>] xmit_one+0x82/0x1a0
[161846.962745]  [<ffffffff8160ad90>] dev_hard_start_xmit+0x50/0xa0
[161846.962749]  [<ffffffff8160bb1e>] __dev_queue_xmit+0x4ee/0x5e0
[161846.962754]  [<ffffffff8160bc33>] dev_queue_xmit_sk+0x13/0x20
[161846.962760]  [<ffffffffa05cfa72>] vlan_dev_hard_start_xmit+0xb2/0x150 [8021q]
[161846.962764]  [<ffffffff8160ac22>] xmit_one+0x82/0x1a0
[161846.962769]  [<ffffffff8160ad90>] dev_hard_start_xmit+0x50/0xa0
[161846.962773]  [<ffffffff8160bb1e>] __dev_queue_xmit+0x4ee/0x5e0
[161846.962777]  [<ffffffff8160bc33>] dev_queue_xmit_sk+0x13/0x20
[161846.962789]  [<ffffffffa05adf74>] br_dev_queue_push_xmit+0x54/0xa0 [bridge]
[161846.962797]  [<ffffffffa05ae4ff>] br_forward_finish+0x2f/0x90 [bridge]
[161846.962807]  [<ffffffff810b0dad>] ? ttwu_do_wakeup+0x1d/0x100
[161846.962811]  [<ffffffff815f929b>] ? __alloc_skb+0x8b/0x1f0
[161846.962818]  [<ffffffffa05ae04d>] __br_forward+0x8d/0x120 [bridge]
[161846.962822]  [<ffffffff815f613b>] ? __kmalloc_reserve+0x3b/0xa0
[161846.962829]  [<ffffffff810be55e>] ? update_rq_runnable_avg+0xee/0x230
[161846.962836]  [<ffffffffa05ae176>] br_forward+0x96/0xb0 [bridge]
[161846.962845]  [<ffffffffa05af85e>] br_handle_frame_finish+0x1ae/0x420 [bridge]
[161846.962853]  [<ffffffffa05afc4f>] br_handle_frame+0x17f/0x260 [bridge]
[161846.962862]  [<ffffffffa05afad0>] ? br_handle_frame_finish+0x420/0x420 [bridge]
[161846.962867]  [<ffffffff8160d057>] __netif_receive_skb_core+0x1f7/0x870
[161846.962872]  [<ffffffff8160d6f2>] __netif_receive_skb+0x22/0x70
[161846.962877]  [<ffffffff8160d913>] netif_receive_skb_internal+0x23/0x90
[161846.962884]  [<ffffffffa07512ea>] ? xenvif_idx_release+0xea/0x100 [xen_netback]
[161846.962889]  [<ffffffff816e5a10>] ? _raw_spin_unlock_irqrestore+0x20/0x50
[161846.962893]  [<ffffffff8160e624>] netif_receive_skb_sk+0x24/0x90
[161846.962899]  [<ffffffffa075269a>] xenvif_tx_submit+0x2ca/0x3f0 [xen_netback]
[161846.962906]  [<ffffffffa0753f0c>] xenvif_tx_action+0x9c/0xd0 [xen_netback]
[161846.962915]  [<ffffffffa07567f5>] xenvif_poll+0x35/0x70 [xen_netback]
[161846.962920]  [<ffffffff8160e01b>] napi_poll+0xcb/0x1e0
[161846.962925]  [<ffffffff8160e1c0>] net_rx_action+0x90/0x1c0
[161846.962931]  [<ffffffff8108aaba>] __do_softirq+0x10a/0x350
[161846.962938]  [<ffffffff8108ae75>] irq_exit+0x125/0x130
[161846.962943]  [<ffffffff813f03a9>] xen_evtchn_do_upcall+0x39/0x50
[161846.962950]  [<ffffffff816e7ffe>] xen_do_hypervisor_callback+0x1e/0x40
[161846.962952]  <EOI>
[161846.962959]  [<ffffffff816e5c4a>] ? _raw_spin_lock+0x4a/0x80
[161846.962964]  [<ffffffff816e5b1e>] ? _raw_spin_lock_irqsave+0x1e/0xa0
[161846.962978]  [<ffffffffa028e279>] ? qlcnic_83xx_mailbox_worker+0xb9/0x2a0 [qlcnic]
[161846.962991]  [<ffffffff810a14e1>] ? process_one_work+0x151/0x4b0
[161846.962995]  [<ffffffff8100c3f2>] ? check_events+0x12/0x20
[161846.963001]  [<ffffffff810a1960>] ? worker_thread+0x120/0x480
[161846.963005]  [<ffffffff816e187b>] ? __schedule+0x30b/0x890
[161846.963010]  [<ffffffff810a1840>] ? process_one_work+0x4b0/0x4b0
[161846.963015]  [<ffffffff810a1840>] ? process_one_work+0x4b0/0x4b0
[161846.963021]  [<ffffffff810a6b3e>] ? kthread+0xce/0xf0
[161846.963025]  [<ffffffff810a6a70>] ? kthread_freezable_should_stop+0x70/0x70
[161846.963031]  [<ffffffff816e6522>] ? ret_from_fork+0x42/0x70
[161846.963035]  [<ffffffff810a6a70>] ? kthread_freezable_should_stop+0x70/0x70
[161846.963037] Code: cc 51 41 53 b8 1c 00 00 00 0f 05 41 5b 59 c3 cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc 51 41 53 b8 1d 00 00 00 0f 05 <41> 5b 59 c3 cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc cc

Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c
index f7080d0ab8746263c6955163640964eac2fd2cb0..46b0372dd032695394b742c17951ca1e25c7ca0a 100644 (file)
--- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c
+++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_83xx_hw.c
@@ -3891,7 +3891,7 @@ static void qlcnic_83xx_flush_mbx_queue(struct qlcnic_adapter *adapter)
        struct list_head *head = &mbx->cmd_q;
        struct qlcnic_cmd_args *cmd = NULL;
 
-       spin_lock(&mbx->queue_lock);
+       spin_lock_bh(&mbx->queue_lock);
 
        while (!list_empty(head)) {
                cmd = list_entry(head->next, struct qlcnic_cmd_args, list);
@@ -3902,7 +3902,7 @@ static void qlcnic_83xx_flush_mbx_queue(struct qlcnic_adapter *adapter)
                qlcnic_83xx_notify_cmd_completion(adapter, cmd);
        }
 
-       spin_unlock(&mbx->queue_lock);
+       spin_unlock_bh(&mbx->queue_lock);
 }
 
 static int qlcnic_83xx_check_mbx_status(struct qlcnic_adapter *adapter)
@@ -3938,12 +3938,12 @@ static void qlcnic_83xx_dequeue_mbx_cmd(struct qlcnic_adapter *adapter,
 {
        struct qlcnic_mailbox *mbx = adapter->ahw->mailbox;
 
-       spin_lock(&mbx->queue_lock);
+       spin_lock_bh(&mbx->queue_lock);
 
        list_del(&cmd->list);
        mbx->num_cmds--;
 
-       spin_unlock(&mbx->queue_lock);
+       spin_unlock_bh(&mbx->queue_lock);
 
        qlcnic_83xx_notify_cmd_completion(adapter, cmd);
 }
@@ -4008,7 +4008,7 @@ static int qlcnic_83xx_enqueue_mbx_cmd(struct qlcnic_adapter *adapter,
                init_completion(&cmd->completion);
                cmd->rsp_opcode = QLC_83XX_MBX_RESPONSE_UNKNOWN;
 
-               spin_lock(&mbx->queue_lock);
+               spin_lock_bh(&mbx->queue_lock);
 
                list_add_tail(&cmd->list, &mbx->cmd_q);
                mbx->num_cmds++;
@@ -4016,7 +4016,7 @@ static int qlcnic_83xx_enqueue_mbx_cmd(struct qlcnic_adapter *adapter,
                *timeout = cmd->total_cmds * QLC_83XX_MBX_TIMEOUT;
                queue_work(mbx->work_q, &mbx->work);
 
-               spin_unlock(&mbx->queue_lock);
+               spin_unlock_bh(&mbx->queue_lock);
 
                return 0;
        }
@@ -4112,15 +4112,15 @@ static void qlcnic_83xx_mailbox_worker(struct work_struct *work)
                mbx->rsp_status = QLC_83XX_MBX_RESPONSE_WAIT;
                spin_unlock_irqrestore(&mbx->aen_lock, flags);
 
-               spin_lock(&mbx->queue_lock);
+               spin_lock_bh(&mbx->queue_lock);
 
                if (list_empty(head)) {
-                       spin_unlock(&mbx->queue_lock);
+                       spin_unlock_bh(&mbx->queue_lock);
                        return;
                }
                cmd = list_entry(head->next, struct qlcnic_cmd_args, list);
 
-               spin_unlock(&mbx->queue_lock);
+               spin_unlock_bh(&mbx->queue_lock);
 
                mbx_ops->encode_cmd(adapter, cmd);
                mbx_ops->nofity_fw(adapter, QLC_83XX_MBX_REQUEST);