bonding: Fix deadlock in bonding driver when using netpoll

PATCH Applied Ding Tianhong 2 participants 2 messages Feb 12, 2014 – Feb 13, 2014

Subsystems (3)

Networking Network drivers Bonding driver

lore

Share & tools

lore

Ding Tianhong

author

2014-02-12 04:07 UTC #0 lore raw

PATCH 1/1 bonding: Fix deadlock in bonding driver when using netpoll

The bonding driver take write locks and spin locks that are shared by the tx path in enslave processing and notification processing, If the netconsole is in use, the bonding can call printk which puts us in the netpoll tx path, if the netconsole is attached to the bonding driver, result in deadlock.

So add protection for these place, by checking the netpoll_block_tx state, we can defer the sending of the netconsole frames until a later time using the retransmit feature of netpoll_send_skb that is triggered on the return code NETDEV_TX_BUSY.

Cc Jay Vosburgh <[email protected]>

Cc Veaceslav Falico <[email protected]>

Cc Andy Gospodarek <[email protected]>

Signed-off-by Ding Tianhong <[email protected]>

---
 drivers/net/bonding/bond_main.c | 7 +++++++
 1 file changed, 7 insertions(+)

drivers/net/bonding/bond_main.c +7 -0

index 71ba18e..8676649 100644
@@ -1543,9 +1543,11 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
 	bond_set_carrier(bond);
 
 	if (USES_PRIMARY(bond->params.mode)) {
+		block_netpoll_tx();
 		write_lock_bh(&bond->curr_slave_lock);
 		bond_select_active_slave(bond);
 		write_unlock_bh(&bond->curr_slave_lock);
+		unblock_netpoll_tx();
 	}
 
 	pr_info("%s: enslaving %s as a%s interface with a%s link.\n",

@@ -1571,10 +1573,12 @@ err_detach:
 	if (bond->primary_slave == new_slave)
 		bond->primary_slave = NULL;
 	if (bond->curr_active_slave == new_slave) {
+		block_netpoll_tx();
 		write_lock_bh(&bond->curr_slave_lock);
 		bond_change_active_slave(bond, NULL);
 		bond_select_active_slave(bond);
 		write_unlock_bh(&bond->curr_slave_lock);
+		unblock_netpoll_tx();
 	}
 	slave_disable_netpoll(new_slave);

@@ -2864,9 +2868,12 @@ static int bond_slave_netdev_event(unsigned long event,
 		pr_info("%s: Primary slave changed to %s, reselecting active slave.\n",
 			bond->dev->name, bond->primary_slave ? slave_dev->name :
 							       "none");
+
+		block_netpoll_tx();
 		write_lock_bh(&bond->curr_slave_lock);
 		bond_select_active_slave(bond);
 		write_unlock_bh(&bond->curr_slave_lock);
+		unblock_netpoll_tx();
 		break;
 	case NETDEV_FEAT_CHANGE:
 		bond_compute_features(bond);

Signature · 1 line

1.8.0

David Miller

maintainer Applied

2014-02-13 22:13 UTC #1 lore raw

From: Ding Tianhong <[email protected]>
Date: Wed, 12 Feb 2014 12:06:40 +0800

Quoted 11 lines

The bonding driver take write locks and spin locks that are shared by the tx path in enslave processing and notification processing, If the netconsole is in use, the bonding can call printk which puts us in the netpoll tx path, if the netconsole is attached to the bonding driver, result in deadlock.
So add protection for these place, by checking the netpoll_block_tx state, we can defer the sending of the netconsole frames until a later time using the retransmit feature of netpoll_send_skb that is triggered on the return code NETDEV_TX_BUSY.

Quoted 4 lines

Cc Jay Vosburgh <[email protected]>
Cc Veaceslav Falico <[email protected]>
Cc Andy Gospodarek <[email protected]>
Signed-off-by Ding Tianhong <[email protected]>

Applied.