From 66c46d741e2e60f0e8b625b80edb0ab820c46d7a Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Sat, 29 Jan 2011 20:44:54 -0800 Subject: gro: Reset dev pointer on reuse On older kernels the VLAN code may zero skb->dev before dropping it and causing it to be reused by GRO. Unfortunately we didn't reset skb->dev in that case which causes the next GRO user to get a bogus skb->dev pointer. This particular problem no longer happens with the current upstream kernel due to changes in VLAN processing. However, for correctness we should still reset the skb->dev pointer in the GRO reuse function in case a future user does the same thing. Signed-off-by: Herbert Xu 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 24ea2d71e7e..93e44dbf679 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3424,6 +3424,7 @@ static void napi_reuse_skb(struct napi_struct *napi, struct sk_buff *skb) __skb_pull(skb, skb_headlen(skb)); skb_reserve(skb, NET_IP_ALIGN - skb_headroom(skb)); skb->vlan_tci = 0; + skb->dev = napi->dev; napi->skb = skb; } -- cgit v1.2.3-70-g09d2 From 8587523640441a9ff2564ebc6efeb39497ad6709 Mon Sep 17 00:00:00 2001 From: Tom Herbert Date: Mon, 31 Jan 2011 16:23:42 -0800 Subject: net: Check rps_flow_table when RPS map length is 1 In get_rps_cpu, add check that the rps_flow_table for the device is NULL when trying to take fast path when RPS map length is one. Without this, RFS is effectively disabled if map length is one which is not correct. Signed-off-by: Tom Herbert Signed-off-by: David S. Miller --- net/core/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 93e44dbf679..4c907895876 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2563,7 +2563,8 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb, map = rcu_dereference(rxqueue->rps_map); if (map) { - if (map->len == 1) { + if (map->len == 1 && + !rcu_dereference_raw(rxqueue->rps_flow_table)) { tcpu = map->cpus[0]; if (cpu_online(tcpu)) cpu = tcpu; -- cgit v1.2.3-70-g09d2 From 6d152e23ad1a7a5b40fef1f42e017d66e6115159 Mon Sep 17 00:00:00 2001 From: Andy Gospodarek Date: Wed, 2 Feb 2011 14:53:25 -0800 Subject: gro: reset skb_iif on reuse Like Herbert's change from a few days ago: 66c46d741e2e60f0e8b625b80edb0ab820c46d7a gro: Reset dev pointer on reuse this may not be necessary at this point, but we should still clean up the skb->skb_iif. If not we may end up with an invalid valid for skb->skb_iif when the skb is reused and the check is done in __netif_receive_skb. Signed-off-by: Andy Gospodarek 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 4c907895876..b6d0bf875a8 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3426,6 +3426,7 @@ static void napi_reuse_skb(struct napi_struct *napi, struct sk_buff *skb) skb_reserve(skb, NET_IP_ALIGN - skb_headroom(skb)); skb->vlan_tci = 0; skb->dev = napi->dev; + skb->skb_iif = 0; napi->skb = skb; } -- cgit v1.2.3-70-g09d2 From 8d3bdbd55a7e2a3f2c148a4830aa26dd682b21c4 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Tue, 8 Feb 2011 15:02:50 -0800 Subject: net: Fix lockdep regression caused by initializing netdev queues too early. In commit aa9421041128abb4d269ee1dc502ff65fb3b7d69 ("net: init ingress queue") we moved the allocation and lock initialization of the queues into alloc_netdev_mq() since register_netdevice() is way too late. The problem is that dev->type is not setup until the setup() callback is invoked by alloc_netdev_mq(), and the dev->type is what determines the lockdep class to use for the locks in the queues. Fix this by doing the queue allocation after the setup() callback runs. This is safe because the setup() callback is not allowed to make any state changes that need to be undone on error (memory allocations, etc.). It may, however, make state changes that are undone by free_netdev() (such as netif_napi_add(), which is done by the ipoib driver's setup routine). The previous code also leaked a reference to the &init_net namespace object on RX/TX queue allocation failures. Signed-off-by: David S. Miller --- net/core/dev.c | 27 ++++++++++++++++----------- 1 file changed, 16 insertions(+), 11 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index b6d0bf875a8..8e726cb47ed 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5660,30 +5660,35 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, dev_net_set(dev, &init_net); + dev->gso_max_size = GSO_MAX_SIZE; + + INIT_LIST_HEAD(&dev->ethtool_ntuple_list.list); + dev->ethtool_ntuple_list.count = 0; + INIT_LIST_HEAD(&dev->napi_list); + INIT_LIST_HEAD(&dev->unreg_list); + INIT_LIST_HEAD(&dev->link_watch_list); + dev->priv_flags = IFF_XMIT_DST_RELEASE; + setup(dev); + dev->num_tx_queues = txqs; dev->real_num_tx_queues = txqs; if (netif_alloc_netdev_queues(dev)) - goto free_pcpu; + goto free_all; #ifdef CONFIG_RPS dev->num_rx_queues = rxqs; dev->real_num_rx_queues = rxqs; if (netif_alloc_rx_queues(dev)) - goto free_pcpu; + goto free_all; #endif - dev->gso_max_size = GSO_MAX_SIZE; - - INIT_LIST_HEAD(&dev->ethtool_ntuple_list.list); - dev->ethtool_ntuple_list.count = 0; - INIT_LIST_HEAD(&dev->napi_list); - INIT_LIST_HEAD(&dev->unreg_list); - INIT_LIST_HEAD(&dev->link_watch_list); - dev->priv_flags = IFF_XMIT_DST_RELEASE; - setup(dev); strcpy(dev->name, name); return dev; +free_all: + free_netdev(dev); + return NULL; + free_pcpu: free_percpu(dev->pcpu_refcnt); kfree(dev->_tx); -- cgit v1.2.3-70-g09d2 From f87e6f47933e3ebeced9bb12615e830a72cedce4 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Thu, 17 Feb 2011 22:54:38 +0000 Subject: net: dont leave active on stack LIST_HEAD Eric W. Biderman and Michal Hocko reported various memory corruptions that we suspected to be related to a LIST head located on stack, that was manipulated after thread left function frame (and eventually exited, so its stack was freed and reused). Eric Dumazet suggested the problem was probably coming from commit 443457242beb (net: factorize sync-rcu call in unregister_netdevice_many) This patch fixes __dev_close() and dev_close() to properly deinit their respective LIST_HEAD(single) before exiting. References: https://lkml.org/lkml/2011/2/16/304 References: https://lkml.org/lkml/2011/2/14/223 Reported-by: Michal Hocko Tested-by: Michal Hocko Reported-by: Eric W. Biderman Tested-by: Eric W. Biderman Signed-off-by: Linus Torvalds Signed-off-by: Eric Dumazet CC: Ingo Molnar CC: Octavian Purdila Signed-off-by: David S. Miller --- net/core/dev.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 8e726cb47ed..a18c1643ea9 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1280,10 +1280,13 @@ static int __dev_close_many(struct list_head *head) static int __dev_close(struct net_device *dev) { + int retval; LIST_HEAD(single); list_add(&dev->unreg_list, &single); - return __dev_close_many(&single); + retval = __dev_close_many(&single); + list_del(&single); + return retval; } int dev_close_many(struct list_head *head) @@ -1325,7 +1328,7 @@ int dev_close(struct net_device *dev) list_add(&dev->unreg_list, &single); dev_close_many(&single); - + list_del(&single); return 0; } EXPORT_SYMBOL(dev_close); -- cgit v1.2.3-70-g09d2 From ceaaec98ad99859ac90ac6863ad0a6cd075d8e0e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 17 Feb 2011 22:59:19 +0000 Subject: net: deinit automatic LIST_HEAD commit 9b5e383c11b08784 (net: Introduce unregister_netdevice_many()) left an active LIST_HEAD() in rollback_registered(), with possible memory corruption. Even if device is freed without touching its unreg_list (and therefore touching the previous memory location holding LISTE_HEAD(single), better close the bug for good, since its really subtle. (Same fix for default_device_exit_batch() for completeness) Reported-by: Michal Hocko Tested-by: Michal Hocko Reported-by: Eric W. Biderman Tested-by: Eric W. Biderman Signed-off-by: Linus Torvalds Signed-off-by: Eric Dumazet CC: Ingo Molnar CC: Octavian Purdila CC: stable [.33+] Signed-off-by: David S. Miller --- net/core/dev.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index a18c1643ea9..8ae6631abcc 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5066,6 +5066,7 @@ static void rollback_registered(struct net_device *dev) list_add(&dev->unreg_list, &single); rollback_registered_many(&single); + list_del(&single); } unsigned long netdev_fix_features(unsigned long features, const char *name) @@ -6219,6 +6220,7 @@ static void __net_exit default_device_exit_batch(struct list_head *net_list) } } unregister_netdevice_many(&dev_kill_list); + list_del(&dev_kill_list); rtnl_unlock(); } -- cgit v1.2.3-70-g09d2