From e0aac52e17a3db68fe2ceae281780a70fc69957f Mon Sep 17 00:00:00 2001 From: Simon Horman Date: Fri, 27 Jan 2012 10:45:27 +0900 Subject: [PATCH 1/4] ipvs: fix matching of fwmark templates during scheduling Commit f11017ec2d1859c661f4e2b12c4a8d250e1f47cf (2.6.37) moved the fwmark variable in subcontext that is invalidated before reaching the ip_vs_ct_in_get call. As vaddr is provided as pointer in the param structure make sure the fwmark variable is in same context. As the fwmark templates can not be matched, more and more template connections are created and the controlled connections can not go to single real server. Signed-off-by: Julian Anastasov Cc: stable@vger.kernel.org Signed-off-by: Simon Horman Signed-off-by: Pablo Neira Ayuso --- net/netfilter/ipvs/ip_vs_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c index 611c3359b94d..2555816e7788 100644 --- a/net/netfilter/ipvs/ip_vs_core.c +++ b/net/netfilter/ipvs/ip_vs_core.c @@ -232,6 +232,7 @@ ip_vs_sched_persist(struct ip_vs_service *svc, __be16 dport = 0; /* destination port to forward */ unsigned int flags; struct ip_vs_conn_param param; + const union nf_inet_addr fwmark = { .ip = htonl(svc->fwmark) }; union nf_inet_addr snet; /* source network of the client, after masking */ @@ -267,7 +268,6 @@ ip_vs_sched_persist(struct ip_vs_service *svc, { int protocol = iph.protocol; const union nf_inet_addr *vaddr = &iph.daddr; - const union nf_inet_addr fwmark = { .ip = htonl(svc->fwmark) }; __be16 vport = 0; if (dst_port == svc->port) { From a8db7b2d197a0d624baab83f0c810b0edbc4ffd0 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 6 Feb 2012 13:23:10 +0100 Subject: [PATCH 2/4] netfilter: nf_queue: fix queueing of bridged gro skbs When trying to nf_queue GRO/GSO skbs, nf_queue uses skb_gso_segment to split the skb. However, if nf_queue is called via bridge netfilter, the mac header won't be preserved -- packets will thus contain a bogus mac header. Fix this by setting skb->data to the mac header when skb->nf_bridge is set and restoring skb->data afterwards for all segments. Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_queue.c | 40 ++++++++++++++++++++++++++++++++-------- 1 file changed, 32 insertions(+), 8 deletions(-) diff --git a/net/netfilter/nf_queue.c b/net/netfilter/nf_queue.c index b3a7db678b8d..ce60cf0f6c11 100644 --- a/net/netfilter/nf_queue.c +++ b/net/netfilter/nf_queue.c @@ -203,6 +203,27 @@ err: return status; } +#ifdef CONFIG_BRIDGE_NETFILTER +/* When called from bridge netfilter, skb->data must point to MAC header + * before calling skb_gso_segment(). Else, original MAC header is lost + * and segmented skbs will be sent to wrong destination. + */ +static void nf_bridge_adjust_skb_data(struct sk_buff *skb) +{ + if (skb->nf_bridge) + __skb_push(skb, skb->network_header - skb->mac_header); +} + +static void nf_bridge_adjust_segmented_data(struct sk_buff *skb) +{ + if (skb->nf_bridge) + __skb_pull(skb, skb->network_header - skb->mac_header); +} +#else +#define nf_bridge_adjust_skb_data(s) do {} while (0) +#define nf_bridge_adjust_segmented_data(s) do {} while (0) +#endif + int nf_queue(struct sk_buff *skb, struct list_head *elem, u_int8_t pf, unsigned int hook, @@ -212,7 +233,7 @@ int nf_queue(struct sk_buff *skb, unsigned int queuenum) { struct sk_buff *segs; - int err; + int err = -EINVAL; unsigned int queued; if (!skb_is_gso(skb)) @@ -228,23 +249,25 @@ int nf_queue(struct sk_buff *skb, break; } + nf_bridge_adjust_skb_data(skb); segs = skb_gso_segment(skb, 0); /* Does not use PTR_ERR to limit the number of error codes that can be * returned by nf_queue. For instance, callers rely on -ECANCELED to mean * 'ignore this hook'. */ if (IS_ERR(segs)) - return -EINVAL; - + goto out_err; queued = 0; err = 0; do { struct sk_buff *nskb = segs->next; segs->next = NULL; - if (err == 0) + if (err == 0) { + nf_bridge_adjust_segmented_data(segs); err = __nf_queue(segs, elem, pf, hook, indev, outdev, okfn, queuenum); + } if (err == 0) queued++; else @@ -252,11 +275,12 @@ int nf_queue(struct sk_buff *skb, segs = nskb; } while (segs); - /* also free orig skb if only some segments were queued */ - if (unlikely(err && queued)) - err = 0; - if (err == 0) + if (queued) { kfree_skb(skb); + return 0; + } + out_err: + nf_bridge_adjust_segmented_data(skb); return err; } From 88ba136d6635b262f77cc418d536115fb8e4d4ab Mon Sep 17 00:00:00 2001 From: Joerg Willmann Date: Tue, 21 Feb 2012 13:26:14 +0100 Subject: [PATCH 3/4] netfilter: ebtables: fix alignment problem in ppc ebt_among extension of ebtables uses __alignof__(_xt_align) while the corresponding kernel module uses __alignof__(ebt_replace) to determine the alignment in EBT_ALIGN(). These are the results of these values on different platforms: x86 x86_64 ppc __alignof__(_xt_align) 4 8 8 __alignof__(ebt_replace) 4 8 4 ebtables fails to add rules which use the among extension. I'm using kernel 2.6.33 and ebtables 2.0.10-4 According to Bart De Schuymer, userspace alignment was changed to _xt_align to fix an alignment issue on a userspace32-kernel64 system (he thinks it was for an ARM device). So userspace must be right. The kernel alignment macro needs to change so it also uses _xt_align instead of ebt_replace. The userspace changes date back from June 29, 2009. Signed-off-by: Joerg Willmann Signed-off by: Bart De Schuymer Signed-off-by: Pablo Neira Ayuso --- include/linux/netfilter_bridge/ebtables.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/linux/netfilter_bridge/ebtables.h b/include/linux/netfilter_bridge/ebtables.h index 8797ed16feb2..4dd5bd6994a8 100644 --- a/include/linux/netfilter_bridge/ebtables.h +++ b/include/linux/netfilter_bridge/ebtables.h @@ -285,8 +285,8 @@ struct ebt_table { struct module *me; }; -#define EBT_ALIGN(s) (((s) + (__alignof__(struct ebt_replace)-1)) & \ - ~(__alignof__(struct ebt_replace)-1)) +#define EBT_ALIGN(s) (((s) + (__alignof__(struct _xt_align)-1)) & \ + ~(__alignof__(struct _xt_align)-1)) extern struct ebt_table *ebt_register_table(struct net *net, const struct ebt_table *table); extern void ebt_unregister_table(struct net *net, struct ebt_table *table); From af14cca162ddcdea017b648c21b9b091e4bf1fa4 Mon Sep 17 00:00:00 2001 From: Jozsef Kadlecsik Date: Tue, 21 Feb 2012 14:53:57 +0100 Subject: [PATCH 4/4] netfilter: ctnetlink: fix soft lockup when netlink adds new entries Marcell Zambo and Janos Farago noticed and reported that when new conntrack entries are added via netlink and the conntrack table gets full, soft lockup happens. This is because the nf_conntrack_lock is held while nf_conntrack_alloc is called, which is in turn wants to lock nf_conntrack_lock while evicting entries from the full table. The patch fixes the soft lockup with limiting the holding of the nf_conntrack_lock to the minimum, where it's absolutely required. Signed-off-by: Jozsef Kadlecsik Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_conntrack_netlink.c | 43 +++++++++++----------------- 1 file changed, 16 insertions(+), 27 deletions(-) diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 9307b033c0c9..cc705175765c 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -1367,15 +1367,12 @@ ctnetlink_create_conntrack(struct net *net, u16 zone, nf_ct_protonum(ct)); if (helper == NULL) { rcu_read_unlock(); - spin_unlock_bh(&nf_conntrack_lock); #ifdef CONFIG_MODULES if (request_module("nfct-helper-%s", helpname) < 0) { - spin_lock_bh(&nf_conntrack_lock); err = -EOPNOTSUPP; goto err1; } - spin_lock_bh(&nf_conntrack_lock); rcu_read_lock(); helper = __nf_conntrack_helper_find(helpname, nf_ct_l3num(ct), @@ -1469,7 +1466,10 @@ ctnetlink_create_conntrack(struct net *net, u16 zone, tstamp->start = ktime_to_ns(ktime_get_real()); add_timer(&ct->timeout); + spin_lock_bh(&nf_conntrack_lock); nf_conntrack_hash_insert(ct); + nf_conntrack_get(&ct->ct_general); + spin_unlock_bh(&nf_conntrack_lock); rcu_read_unlock(); return ct; @@ -1490,6 +1490,7 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb, struct nf_conntrack_tuple otuple, rtuple; struct nf_conntrack_tuple_hash *h = NULL; struct nfgenmsg *nfmsg = nlmsg_data(nlh); + struct nf_conn *ct; u_int8_t u3 = nfmsg->nfgen_family; u16 zone; int err; @@ -1512,25 +1513,22 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb, spin_lock_bh(&nf_conntrack_lock); if (cda[CTA_TUPLE_ORIG]) - h = __nf_conntrack_find(net, zone, &otuple); + h = nf_conntrack_find_get(net, zone, &otuple); else if (cda[CTA_TUPLE_REPLY]) - h = __nf_conntrack_find(net, zone, &rtuple); + h = nf_conntrack_find_get(net, zone, &rtuple); + spin_unlock_bh(&nf_conntrack_lock); if (h == NULL) { err = -ENOENT; if (nlh->nlmsg_flags & NLM_F_CREATE) { - struct nf_conn *ct; enum ip_conntrack_events events; ct = ctnetlink_create_conntrack(net, zone, cda, &otuple, &rtuple, u3); - if (IS_ERR(ct)) { - err = PTR_ERR(ct); - goto out_unlock; - } + if (IS_ERR(ct)) + return PTR_ERR(ct); + err = 0; - nf_conntrack_get(&ct->ct_general); - spin_unlock_bh(&nf_conntrack_lock); if (test_bit(IPS_EXPECTED_BIT, &ct->status)) events = IPCT_RELATED; else @@ -1545,23 +1543,19 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb, ct, NETLINK_CB(skb).pid, nlmsg_report(nlh)); nf_ct_put(ct); - } else - spin_unlock_bh(&nf_conntrack_lock); + } return err; } /* implicit 'else' */ - /* We manipulate the conntrack inside the global conntrack table lock, - * so there's no need to increase the refcount */ err = -EEXIST; + ct = nf_ct_tuplehash_to_ctrack(h); if (!(nlh->nlmsg_flags & NLM_F_EXCL)) { - struct nf_conn *ct = nf_ct_tuplehash_to_ctrack(h); - + spin_lock_bh(&nf_conntrack_lock); err = ctnetlink_change_conntrack(ct, cda); + spin_unlock_bh(&nf_conntrack_lock); if (err == 0) { - nf_conntrack_get(&ct->ct_general); - spin_unlock_bh(&nf_conntrack_lock); nf_conntrack_eventmask_report((1 << IPCT_REPLY) | (1 << IPCT_ASSURED) | (1 << IPCT_HELPER) | @@ -1570,15 +1564,10 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb, (1 << IPCT_MARK), ct, NETLINK_CB(skb).pid, nlmsg_report(nlh)); - nf_ct_put(ct); - } else - spin_unlock_bh(&nf_conntrack_lock); - - return err; + } } -out_unlock: - spin_unlock_bh(&nf_conntrack_lock); + nf_ct_put(ct); return err; }