From b4bd07c20ba0c1fa7ad09ba257e0a5cfc2bf6bb3 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 6 Feb 2009 22:06:43 -0800 Subject: net_dma: call dmaengine_get only if NET_DMA enabled Based upon a patch from Atsushi Nemoto -------------------- The commit 649274d993212e7c23c0cb734572c2311c200872 ("net_dma: acquire/release dma channels on ifup/ifdown") added unconditional call of dmaengine_get() to net_dma. The API should be called only if NET_DMA was enabled. -------------------- Signed-off-by: David S. Miller Acked-by: Dan Williams --- net/core/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 5379b0c1190a..a17e00662363 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1090,7 +1090,7 @@ int dev_open(struct net_device *dev) /* * Enable NET_DMA */ - dmaengine_get(); + net_dmaengine_get(); /* * Initialize multicasting status @@ -1172,7 +1172,7 @@ int dev_close(struct net_device *dev) /* * Shutdown NET_DMA */ - dmaengine_put(); + net_dmaengine_put(); return 0; } -- cgit From 4ead443163b798661c2a2ede5e512e116a9e41e7 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Sun, 1 Mar 2009 00:11:52 -0800 Subject: netpoll: Add drop checks to all entry points The netpoll entry checks are required to ensure that we don't receive normal packets when invoked via netpoll. Unfortunately it only ever worked for the netif_receive_skb/netif_rx entry points. The VLAN (and subsequently GRO) entry point didn't have the check and therefore can trigger all sorts of weird problems. This patch adds the netpoll check to all entry points. I'm still uneasy with receiving at all under netpoll (which apparently is only used by the out-of-tree kdump code). The reason is it is perfectly legal to receive all data including headers into highmem if netpoll is off, but if you try to do that with netpoll on and someone gets a printk in an IRQ handler you're going to get a nice BUG_ON. Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- net/8021q/vlan_core.c | 10 ++++++++++ net/core/dev.c | 6 ++++++ 2 files changed, 16 insertions(+) (limited to 'net/core/dev.c') diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c index e9db889d6222..2886d2fb9ab5 100644 --- a/net/8021q/vlan_core.c +++ b/net/8021q/vlan_core.c @@ -1,12 +1,16 @@ #include #include #include +#include #include "vlan.h" /* VLAN rx hw acceleration helper. This acts like netif_{rx,receive_skb}(). */ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp, u16 vlan_tci, int polling) { + if (netpoll_rx(skb)) + return NET_RX_DROP; + if (skb_bond_should_drop(skb)) goto drop; @@ -100,6 +104,9 @@ int vlan_gro_receive(struct napi_struct *napi, struct vlan_group *grp, { int err = NET_RX_SUCCESS; + if (netpoll_receive_skb(skb)) + return NET_RX_DROP; + switch (vlan_gro_common(napi, grp, vlan_tci, skb)) { case -1: return netif_receive_skb(skb); @@ -126,6 +133,9 @@ int vlan_gro_frags(struct napi_struct *napi, struct vlan_group *grp, if (!skb) goto out; + if (netpoll_receive_skb(skb)) + goto out; + err = NET_RX_SUCCESS; switch (vlan_gro_common(napi, grp, vlan_tci, skb)) { diff --git a/net/core/dev.c b/net/core/dev.c index a17e00662363..72b0d26fd46d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2488,6 +2488,9 @@ static int __napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb) int napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb) { + if (netpoll_receive_skb(skb)) + return NET_RX_DROP; + switch (__napi_gro_receive(napi, skb)) { case -1: return netif_receive_skb(skb); @@ -2558,6 +2561,9 @@ int napi_gro_frags(struct napi_struct *napi, struct napi_gro_fraginfo *info) if (!skb) goto out; + if (netpoll_receive_skb(skb)) + goto out; + err = NET_RX_SUCCESS; switch (__napi_gro_receive(napi, skb)) { -- cgit From 17edde520927070a6bf14a6a75027c0b843443e5 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Sun, 22 Feb 2009 00:11:09 -0800 Subject: netns: Remove net_alive It turns out that net_alive is unnecessary, and the original problem that led to it being added was simply that the icmp code thought it was a network device and wound up being unable to handle packets while there were still packets in the network namespace. Now that icmp and tcp have been fixed to properly register themselves this problem is no longer present and we have a stronger guarantee that packets will not arrive in a network namespace then that provided by net_alive in netif_receive_skb. So remove net_alive allowing packet reception run a little faster. Additionally document the strong reason why network namespace cleanup is safe so that if something happens again someone else will have a chance of figuring it out. Signed-off-by: Eric W. Biederman Signed-off-by: David S. Miller --- include/net/net_namespace.h | 27 +++++++++++++++++---------- net/core/dev.c | 6 ------ net/core/net_namespace.c | 3 --- 3 files changed, 17 insertions(+), 19 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h index 6fc13d905c5f..ded434b032a4 100644 --- a/include/net/net_namespace.h +++ b/include/net/net_namespace.h @@ -109,11 +109,6 @@ extern struct list_head net_namespace_list; #ifdef CONFIG_NET_NS extern void __put_net(struct net *net); -static inline int net_alive(struct net *net) -{ - return net && atomic_read(&net->count); -} - static inline struct net *get_net(struct net *net) { atomic_inc(&net->count); @@ -145,11 +140,6 @@ int net_eq(const struct net *net1, const struct net *net2) } #else -static inline int net_alive(struct net *net) -{ - return 1; -} - static inline struct net *get_net(struct net *net) { return net; @@ -234,6 +224,23 @@ struct pernet_operations { void (*exit)(struct net *net); }; +/* + * Use these carefully. If you implement a network device and it + * needs per network namespace operations use device pernet operations, + * otherwise use pernet subsys operations. + * + * This is critically important. Most of the network code cleanup + * runs with the assumption that dev_remove_pack has been called so no + * new packets will arrive during and after the cleanup functions have + * been called. dev_remove_pack is not per namespace so instead the + * guarantee of no more packets arriving in a network namespace is + * provided by ensuring that all network devices and all sockets have + * left the network namespace before the cleanup methods are called. + * + * For the longest time the ipv4 icmp code was registered as a pernet + * device which caused kernel oops, and panics during network + * namespace cleanup. So please don't get this wrong. + */ extern int register_pernet_subsys(struct pernet_operations *); extern void unregister_pernet_subsys(struct pernet_operations *); extern int register_pernet_gen_subsys(int *id, struct pernet_operations *); diff --git a/net/core/dev.c b/net/core/dev.c index 72b0d26fd46d..9e4afe650e7a 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2267,12 +2267,6 @@ int netif_receive_skb(struct sk_buff *skb) rcu_read_lock(); - /* Don't receive packets in an exiting network namespace */ - if (!net_alive(dev_net(skb->dev))) { - kfree_skb(skb); - goto out; - } - #ifdef CONFIG_NET_CLS_ACT if (skb->tc_verd & TC_NCLS) { skb->tc_verd = CLR_TC_NCLS(skb->tc_verd); diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c index 2adb1a7d361f..e3bebd36f053 100644 --- a/net/core/net_namespace.c +++ b/net/core/net_namespace.c @@ -157,9 +157,6 @@ static void cleanup_net(struct work_struct *work) struct pernet_operations *ops; struct net *net; - /* Be very certain incoming network packets will not find us */ - rcu_barrier(); - net = container_of(work, struct net, work); mutex_lock(&net_mutex); -- cgit From 54acd0efab072cb70e87206329d561b297f93bbb Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 4 Mar 2009 23:01:02 -0800 Subject: net: Fix missing dev->neigh_setup in register_netdevice(). Signed-off-by: David S. Miller --- net/core/dev.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 9e4afe650e7a..2dd484ed3dbb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4339,6 +4339,7 @@ int register_netdevice(struct net_device *dev) dev->do_ioctl = ops->ndo_do_ioctl; dev->set_config = ops->ndo_set_config; dev->change_mtu = ops->ndo_change_mtu; + dev->neigh_setup = ops->ndo_neigh_setup; dev->tx_timeout = ops->ndo_tx_timeout; dev->get_stats = ops->ndo_get_stats; dev->vlan_rx_register = ops->ndo_vlan_rx_register; -- cgit From 9d40bbda599def1e1d155d7f7dca14fe8744bd2b Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Wed, 4 Mar 2009 23:46:25 -0800 Subject: vlan: Fix vlan-in-vlan crashes. As analyzed by Patrick McHardy, vlan needs to reset it's netdev_ops pointer in it's ->init() function but this leaves the compat method pointers stale. Add a netdev_resync_ops() and call it from the vlan code. Any other driver which changes ->netdev_ops after register_netdevice() will need to call this new function after doing so too. With help from Patrick McHardy. Tested-by: Patrick McHardy Signed-off-by: David S. Miller --- include/linux/netdevice.h | 1 + net/8021q/vlan_dev.c | 3 ++- net/core/dev.c | 56 ++++++++++++++++++++++++++++------------------- 3 files changed, 37 insertions(+), 23 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index ec54785d34f9..659366734f3f 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1079,6 +1079,7 @@ extern void synchronize_net(void); extern int register_netdevice_notifier(struct notifier_block *nb); extern int unregister_netdevice_notifier(struct notifier_block *nb); extern int init_dummy_netdev(struct net_device *dev); +extern void netdev_resync_ops(struct net_device *dev); extern int call_netdevice_notifiers(unsigned long val, struct net_device *dev); extern struct net_device *dev_get_by_index(struct net *net, int ifindex); diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index 4a19acd3a32b..1b34135cf990 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -553,7 +553,7 @@ static int vlan_dev_neigh_setup(struct net_device *dev, struct neigh_parms *pa) int err = 0; if (netif_device_present(real_dev) && ops->ndo_neigh_setup) - err = ops->ndo_neigh_setup(dev, pa); + err = ops->ndo_neigh_setup(real_dev, pa); return err; } @@ -639,6 +639,7 @@ static int vlan_dev_init(struct net_device *dev) dev->hard_header_len = real_dev->hard_header_len + VLAN_HLEN; dev->netdev_ops = &vlan_netdev_ops; } + netdev_resync_ops(dev); if (is_vlan_dev(real_dev)) subclass = 1; diff --git a/net/core/dev.c b/net/core/dev.c index 2dd484ed3dbb..f1129706ce7b 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4282,6 +4282,39 @@ unsigned long netdev_fix_features(unsigned long features, const char *name) } EXPORT_SYMBOL(netdev_fix_features); +/* Some devices need to (re-)set their netdev_ops inside + * ->init() or similar. If that happens, we have to setup + * the compat pointers again. + */ +void netdev_resync_ops(struct net_device *dev) +{ +#ifdef CONFIG_COMPAT_NET_DEV_OPS + const struct net_device_ops *ops = dev->netdev_ops; + + dev->init = ops->ndo_init; + dev->uninit = ops->ndo_uninit; + dev->open = ops->ndo_open; + dev->change_rx_flags = ops->ndo_change_rx_flags; + dev->set_rx_mode = ops->ndo_set_rx_mode; + dev->set_multicast_list = ops->ndo_set_multicast_list; + dev->set_mac_address = ops->ndo_set_mac_address; + dev->validate_addr = ops->ndo_validate_addr; + dev->do_ioctl = ops->ndo_do_ioctl; + dev->set_config = ops->ndo_set_config; + dev->change_mtu = ops->ndo_change_mtu; + dev->neigh_setup = ops->ndo_neigh_setup; + dev->tx_timeout = ops->ndo_tx_timeout; + dev->get_stats = ops->ndo_get_stats; + dev->vlan_rx_register = ops->ndo_vlan_rx_register; + dev->vlan_rx_add_vid = ops->ndo_vlan_rx_add_vid; + dev->vlan_rx_kill_vid = ops->ndo_vlan_rx_kill_vid; +#ifdef CONFIG_NET_POLL_CONTROLLER + dev->poll_controller = ops->ndo_poll_controller; +#endif +#endif +} +EXPORT_SYMBOL(netdev_resync_ops); + /** * register_netdevice - register a network device * @dev: device to register @@ -4326,28 +4359,7 @@ int register_netdevice(struct net_device *dev) * This is temporary until all network devices are converted. */ if (dev->netdev_ops) { - const struct net_device_ops *ops = dev->netdev_ops; - - dev->init = ops->ndo_init; - dev->uninit = ops->ndo_uninit; - dev->open = ops->ndo_open; - dev->change_rx_flags = ops->ndo_change_rx_flags; - dev->set_rx_mode = ops->ndo_set_rx_mode; - dev->set_multicast_list = ops->ndo_set_multicast_list; - dev->set_mac_address = ops->ndo_set_mac_address; - dev->validate_addr = ops->ndo_validate_addr; - dev->do_ioctl = ops->ndo_do_ioctl; - dev->set_config = ops->ndo_set_config; - dev->change_mtu = ops->ndo_change_mtu; - dev->neigh_setup = ops->ndo_neigh_setup; - dev->tx_timeout = ops->ndo_tx_timeout; - dev->get_stats = ops->ndo_get_stats; - dev->vlan_rx_register = ops->ndo_vlan_rx_register; - dev->vlan_rx_add_vid = ops->ndo_vlan_rx_add_vid; - dev->vlan_rx_kill_vid = ops->ndo_vlan_rx_kill_vid; -#ifdef CONFIG_NET_POLL_CONTROLLER - dev->poll_controller = ops->ndo_poll_controller; -#endif + netdev_resync_ops(dev); } else { char drivername[64]; pr_info("%s (%s): not using net_device_ops yet\n", -- cgit From 303c6a0251852ecbdc5c15e466dcaff5971f7517 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Tue, 17 Mar 2009 13:11:29 -0700 Subject: gro: Fix legacy path napi_complete crash On the legacy netif_rx path, I incorrectly tried to optimise the napi_complete call by using __napi_complete before we reenable IRQs. This simply doesn't work since we need to flush the held GRO packets first. This patch fixes it by doing the obvious thing of reenabling IRQs first and then calling napi_complete. Reported-by: Frank Blaschka Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- net/core/dev.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index f1129706ce7b..2565f6d1d661 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2588,9 +2588,9 @@ static int process_backlog(struct napi_struct *napi, int quota) local_irq_disable(); skb = __skb_dequeue(&queue->input_pkt_queue); if (!skb) { - __napi_complete(napi); local_irq_enable(); - break; + napi_complete(napi); + goto out; } local_irq_enable(); @@ -2599,6 +2599,7 @@ static int process_backlog(struct napi_struct *napi, int quota) napi_gro_flush(napi); +out: return work; } -- cgit From e4a389a9b5c892446b5de2038bdc0cca8703c615 Mon Sep 17 00:00:00 2001 From: Roel Kluin Date: Wed, 18 Mar 2009 23:12:13 -0700 Subject: net: kfree(napi->skb) => kfree_skb struct sk_buff pointers should be freed with kfree_skb. Signed-off-by: Roel Kluin Signed-off-by: David S. Miller --- net/core/dev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 2565f6d1d661..e3fe5c705606 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2672,7 +2672,7 @@ void netif_napi_del(struct napi_struct *napi) struct sk_buff *skb, *next; list_del_init(&napi->dev_list); - kfree(napi->skb); + kfree_skb(napi->skb); for (skb = napi->gro_list; skb; skb = next) { next = skb->next; -- cgit