提交 35d48903 编写于 作者: J Jiri Pirko 提交者: David S. Miller

bonding: fix rx_handler locking

This prevents possible race between bond_enslave and bond_handle_frame
as reported by Nicolas by moving rx_handler register/unregister.
slave->bond is added to hold pointer to master bonding sructure. That
way dev->master is no longer used in bond_handler_frame.
Also, this removes "BUG: scheduling while atomic" message
Reported-by: NNicolas de Pesloüan <nicolas.2p.debian@gmail.com>
Signed-off-by: NJiri Pirko <jpirko@redhat.com>
Signed-off-by: NAndy Gospodarek <andy@greyhouse.net>
Tested-by: NNicolas de Pesloüan <nicolas.2p.debian@free.fr>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
上级 cda6587c
...@@ -1482,21 +1482,16 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb) ...@@ -1482,21 +1482,16 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb)
{ {
struct sk_buff *skb = *pskb; struct sk_buff *skb = *pskb;
struct slave *slave; struct slave *slave;
struct net_device *bond_dev;
struct bonding *bond; struct bonding *bond;
slave = bond_slave_get_rcu(skb->dev);
bond_dev = ACCESS_ONCE(slave->dev->master);
if (unlikely(!bond_dev))
return RX_HANDLER_PASS;
skb = skb_share_check(skb, GFP_ATOMIC); skb = skb_share_check(skb, GFP_ATOMIC);
if (unlikely(!skb)) if (unlikely(!skb))
return RX_HANDLER_CONSUMED; return RX_HANDLER_CONSUMED;
*pskb = skb; *pskb = skb;
bond = netdev_priv(bond_dev); slave = bond_slave_get_rcu(skb->dev);
bond = slave->bond;
if (bond->params.arp_interval) if (bond->params.arp_interval)
slave->dev->last_rx = jiffies; slave->dev->last_rx = jiffies;
...@@ -1505,10 +1500,10 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb) ...@@ -1505,10 +1500,10 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb)
return RX_HANDLER_EXACT; return RX_HANDLER_EXACT;
} }
skb->dev = bond_dev; skb->dev = bond->dev;
if (bond->params.mode == BOND_MODE_ALB && if (bond->params.mode == BOND_MODE_ALB &&
bond_dev->priv_flags & IFF_BRIDGE_PORT && bond->dev->priv_flags & IFF_BRIDGE_PORT &&
skb->pkt_type == PACKET_HOST) { skb->pkt_type == PACKET_HOST) {
if (unlikely(skb_cow_head(skb, if (unlikely(skb_cow_head(skb,
...@@ -1516,7 +1511,7 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb) ...@@ -1516,7 +1511,7 @@ static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb)
kfree_skb(skb); kfree_skb(skb);
return RX_HANDLER_CONSUMED; return RX_HANDLER_CONSUMED;
} }
memcpy(eth_hdr(skb)->h_dest, bond_dev->dev_addr, ETH_ALEN); memcpy(eth_hdr(skb)->h_dest, bond->dev->dev_addr, ETH_ALEN);
} }
return RX_HANDLER_ANOTHER; return RX_HANDLER_ANOTHER;
...@@ -1698,20 +1693,15 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) ...@@ -1698,20 +1693,15 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
pr_debug("Error %d calling netdev_set_bond_master\n", res); pr_debug("Error %d calling netdev_set_bond_master\n", res);
goto err_restore_mac; goto err_restore_mac;
} }
res = netdev_rx_handler_register(slave_dev, bond_handle_frame,
new_slave);
if (res) {
pr_debug("Error %d calling netdev_rx_handler_register\n", res);
goto err_unset_master;
}
/* open the slave since the application closed it */ /* open the slave since the application closed it */
res = dev_open(slave_dev); res = dev_open(slave_dev);
if (res) { if (res) {
pr_debug("Opening slave %s failed\n", slave_dev->name); pr_debug("Opening slave %s failed\n", slave_dev->name);
goto err_unreg_rxhandler; goto err_unset_master;
} }
new_slave->bond = bond;
new_slave->dev = slave_dev; new_slave->dev = slave_dev;
slave_dev->priv_flags |= IFF_BONDING; slave_dev->priv_flags |= IFF_BONDING;
...@@ -1907,6 +1897,13 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) ...@@ -1907,6 +1897,13 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
if (res) if (res)
goto err_close; goto err_close;
res = netdev_rx_handler_register(slave_dev, bond_handle_frame,
new_slave);
if (res) {
pr_debug("Error %d calling netdev_rx_handler_register\n", res);
goto err_dest_symlinks;
}
pr_info("%s: enslaving %s as a%s interface with a%s link.\n", pr_info("%s: enslaving %s as a%s interface with a%s link.\n",
bond_dev->name, slave_dev->name, bond_dev->name, slave_dev->name,
bond_is_active_slave(new_slave) ? "n active" : " backup", bond_is_active_slave(new_slave) ? "n active" : " backup",
...@@ -1916,13 +1913,12 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) ...@@ -1916,13 +1913,12 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
return 0; return 0;
/* Undo stages on error */ /* Undo stages on error */
err_dest_symlinks:
bond_destroy_slave_symlinks(bond_dev, slave_dev);
err_close: err_close:
dev_close(slave_dev); dev_close(slave_dev);
err_unreg_rxhandler:
netdev_rx_handler_unregister(slave_dev);
synchronize_net();
err_unset_master: err_unset_master:
netdev_set_bond_master(slave_dev, NULL); netdev_set_bond_master(slave_dev, NULL);
...@@ -1988,6 +1984,14 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev) ...@@ -1988,6 +1984,14 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev)
return -EINVAL; return -EINVAL;
} }
/* unregister rx_handler early so bond_handle_frame wouldn't be called
* for this slave anymore.
*/
netdev_rx_handler_unregister(slave_dev);
write_unlock_bh(&bond->lock);
synchronize_net();
write_lock_bh(&bond->lock);
if (!bond->params.fail_over_mac) { if (!bond->params.fail_over_mac) {
if (!compare_ether_addr(bond_dev->dev_addr, slave->perm_hwaddr) && if (!compare_ether_addr(bond_dev->dev_addr, slave->perm_hwaddr) &&
bond->slave_cnt > 1) bond->slave_cnt > 1)
...@@ -2104,8 +2108,6 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev) ...@@ -2104,8 +2108,6 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev)
netif_addr_unlock_bh(bond_dev); netif_addr_unlock_bh(bond_dev);
} }
netdev_rx_handler_unregister(slave_dev);
synchronize_net();
netdev_set_bond_master(slave_dev, NULL); netdev_set_bond_master(slave_dev, NULL);
slave_disable_netpoll(slave); slave_disable_netpoll(slave);
...@@ -2186,6 +2188,12 @@ static int bond_release_all(struct net_device *bond_dev) ...@@ -2186,6 +2188,12 @@ static int bond_release_all(struct net_device *bond_dev)
*/ */
write_unlock_bh(&bond->lock); write_unlock_bh(&bond->lock);
/* unregister rx_handler early so bond_handle_frame wouldn't
* be called for this slave anymore.
*/
netdev_rx_handler_unregister(slave_dev);
synchronize_net();
if (bond_is_lb(bond)) { if (bond_is_lb(bond)) {
/* must be called only after the slave /* must be called only after the slave
* has been detached from the list * has been detached from the list
...@@ -2217,8 +2225,6 @@ static int bond_release_all(struct net_device *bond_dev) ...@@ -2217,8 +2225,6 @@ static int bond_release_all(struct net_device *bond_dev)
netif_addr_unlock_bh(bond_dev); netif_addr_unlock_bh(bond_dev);
} }
netdev_rx_handler_unregister(slave_dev);
synchronize_net();
netdev_set_bond_master(slave_dev, NULL); netdev_set_bond_master(slave_dev, NULL);
slave_disable_netpoll(slave); slave_disable_netpoll(slave);
......
...@@ -187,6 +187,7 @@ struct slave { ...@@ -187,6 +187,7 @@ struct slave {
struct net_device *dev; /* first - useful for panic debug */ struct net_device *dev; /* first - useful for panic debug */
struct slave *next; struct slave *next;
struct slave *prev; struct slave *prev;
struct bonding *bond; /* our master */
int delay; int delay;
unsigned long jiffies; unsigned long jiffies;
unsigned long last_arp_rx; unsigned long last_arp_rx;
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册