>From 6093c6b4c447c8592e4a593590d7a31e4fdaedd2 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 29 Jul 2011 19:00:53 +0000 Subject: [PATCH 09/13] net: fix NULL dereferences in check_peer_redir() UPSTREAM: f667de2e66ea9b8f99353ff01221bb355faf3f3c backported Gergely Kalman reported crashes in check_peer_redir(). It appears commit f39925dbde778 (ipv4: Cache learned redirect information in inetpeer.) added a race, leading to possible NULL ptr dereference. Since we can now change dst neighbour, we should make sure a reader can safely use a neighbour. Add RCU protection to dst neighbour, and make sure check_peer_redir() can be called safely by different cpus in parallel. As neighbours are already freed after one RCU grace period, this patch should not add typical RCU penalty (cache cold effects) Many thanks to Gergely for providing a pretty report pointing to the bug. Reported-by: Gergely Kalman Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/dst.h | 17 +++++++++++++---- net/ipv4/ip_output.c | 17 +++++++++++++---- net/ipv4/route.c | 14 ++++++++------ net/ipv6/addrconf.c | 2 +- net/ipv6/ip6_fib.c | 2 +- net/ipv6/ip6_output.c | 20 ++++++++++++++++---- net/ipv6/route.c | 35 +++++++++++++++++++++++++---------- 7 files changed, 77 insertions(+), 30 deletions(-) diff --git a/include/net/dst.h b/include/net/dst.h index 3962b65..3723f1f 100644 --- a/include/net/dst.h +++ b/include/net/dst.h @@ -37,7 +37,7 @@ struct dst_entry { unsigned long _metrics; unsigned long expires; struct dst_entry *path; - struct neighbour *_neighbour; + struct neighbour __rcu *_neighbour; struct hh_cache *hh; #ifdef CONFIG_XFRM struct xfrm_state *xfrm; @@ -88,12 +88,17 @@ struct dst_entry { static inline struct neighbour *dst_get_neighbour(struct dst_entry *dst) { - return dst->_neighbour; + return rcu_dereference(dst->_neighbour); +} + +static inline struct neighbour *dst_get_neighbour_raw(struct dst_entry *dst) +{ + return rcu_dereference_raw(dst->_neighbour); } static inline void dst_set_neighbour(struct dst_entry *dst, struct neighbour *neigh) { - dst->_neighbour = neigh; + rcu_assign_pointer(dst->_neighbour, neigh); } extern u32 *dst_cow_metrics_generic(struct dst_entry *dst, unsigned long old); @@ -382,8 +387,12 @@ static inline void dst_rcu_free(struct rcu_head *head) static inline void dst_confirm(struct dst_entry *dst) { if (dst) { - struct neighbour *n = dst_get_neighbour(dst); + struct neighbour *n; + + rcu_read_lock(); + n = dst_get_neighbour(dst); neigh_confirm(n); + rcu_read_unlock(); } } diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index 3f994ca..808d0e2 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -204,11 +204,20 @@ static inline int ip_finish_output2(struct sk_buff *skb) skb = skb2; } + rcu_read_lock(); neigh = dst_get_neighbour(dst); - if (dst->hh) - return neigh_hh_output(dst->hh, skb); - else if (neigh) - return neigh->output(skb); + if (dst->hh) { + int res = neigh_hh_output(dst->hh, skb); + + rcu_read_unlock(); + return res; + } else if (neigh) { + int res = neigh->output(skb); + + rcu_read_unlock(); + return res; + } + rcu_read_unlock(); if (net_ratelimit()) printk(KERN_DEBUG "ip_finish_output2: No header cache and no neighbour!\n"); diff --git a/net/ipv4/route.c b/net/ipv4/route.c index b56e304..32eb4a7 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -1630,16 +1630,18 @@ static int check_peer_redir(struct dst_entry *dst, struct inet_peer *peer) { struct rtable *rt = (struct rtable *) dst; __be32 orig_gw = rt->rt_gateway; - struct neighbour *n; + struct neighbour *n, *old_n; dst_confirm(&rt->dst); - neigh_release(dst_get_neighbour(&rt->dst)); - dst_set_neighbour(&rt->dst, NULL); - rt->rt_gateway = peer->redirect_learned.a4; - rt_bind_neighbour(rt); - n = dst_get_neighbour(&rt->dst); + + n = ipv4_neigh_lookup(&rt->dst, &rt->rt_gateway); + if (IS_ERR(n)) + return PTR_ERR(n); + old_n = xchg(&rt->dst._neighbour, n); + if (old_n) + neigh_release(old_n); if (!n || !(n->nud_state & NUD_VALID)) { if (n) neigh_event_send(n, NULL); diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 1fa96a0..0f335c6 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -656,7 +656,7 @@ ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen, * layer address of our nexhop router */ - if (dst_get_neighbour(&rt->dst) == NULL) + if (dst_get_neighbour_raw(&rt->dst) == NULL) ifa->flags &= ~IFA_F_OPTIMISTIC; ifa->idev = idev; diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c index fb1eb5e..0f9b37a 100644 --- a/net/ipv6/ip6_fib.c +++ b/net/ipv6/ip6_fib.c @@ -1455,7 +1455,7 @@ static int fib6_age(struct rt6_info *rt, void *arg) RT6_TRACE("aging clone %p\n", rt); return -1; } else if ((rt->rt6i_flags & RTF_GATEWAY) && - (!(dst_get_neighbour(&rt->dst)->flags & NTF_ROUTER))) { + (!(dst_get_neighbour_raw(&rt->dst)->flags & NTF_ROUTER))) { RT6_TRACE("purging route %p via non-router but gateway\n", rt); return -1; diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 228d317..6abe5b7 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -135,12 +135,20 @@ static int ip6_finish_output2(struct sk_buff *skb) skb->len); } + rcu_read_lock(); neigh = dst_get_neighbour(dst); - if (dst->hh) - return neigh_hh_output(dst->hh, skb); - else if (neigh) - return neigh->output(skb); + if (dst->hh) { + int res = neigh_hh_output(dst->hh, skb); + rcu_read_unlock(); + return res; + } else if (neigh) { + int res = neigh->output(skb); + + rcu_read_unlock(); + return res; + } + rcu_read_unlock(); IP6_INC_STATS_BH(dev_net(dst->dev), ip6_dst_idev(dst), IPSTATS_MIB_OUTNOROUTES); kfree_skb(skb); @@ -981,12 +989,14 @@ static int ip6_dst_lookup_tail(struct sock *sk, * dst entry and replace it instead with the * dst entry of the nexthop router */ + rcu_read_lock(); n = dst_get_neighbour(*dst); if (n && !(n->nud_state & NUD_VALID)) { struct inet6_ifaddr *ifp; struct flowi6 fl_gw6; int redirect; + rcu_read_unlock(); ifp = ipv6_get_ifaddr(net, &fl6->saddr, (*dst)->dev, 1); @@ -1006,6 +1016,8 @@ static int ip6_dst_lookup_tail(struct sock *sk, if ((err = (*dst)->error)) goto out_err_release; } + } else { + rcu_read_unlock(); } #endif diff --git a/net/ipv6/route.c b/net/ipv6/route.c index c277429..2b083dd 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -363,7 +363,7 @@ out: #ifdef CONFIG_IPV6_ROUTER_PREF static void rt6_probe(struct rt6_info *rt) { - struct neighbour *neigh = rt ? dst_get_neighbour(&rt->dst) : NULL; + struct neighbour *neigh; /* * Okay, this does not seem to be appropriate * for now, however, we need to check if it @@ -372,8 +372,10 @@ static void rt6_probe(struct rt6_info *rt) * Router Reachability Probe MUST be rate-limited * to no more than one per minute. */ + rcu_read_lock(); + neigh = rt ? dst_get_neighbour(&rt->dst) : NULL; if (!neigh || (neigh->nud_state & NUD_VALID)) - return; + goto out; read_lock_bh(&neigh->lock); if (!(neigh->nud_state & NUD_VALID) && time_after(jiffies, neigh->updated + rt->rt6i_idev->cnf.rtr_probe_interval)) { @@ -386,8 +388,11 @@ static void rt6_probe(struct rt6_info *rt) target = (struct in6_addr *)&neigh->primary_key; addrconf_addr_solict_mult(target, &mcaddr); ndisc_send_ns(rt->rt6i_dev, NULL, target, &mcaddr, NULL); - } else + } else { read_unlock_bh(&neigh->lock); + } +out: + rcu_read_unlock(); } #else static inline void rt6_probe(struct rt6_info *rt) @@ -411,8 +416,11 @@ static inline int rt6_check_dev(struct rt6_info *rt, int oif) static inline int rt6_check_neigh(struct rt6_info *rt) { - struct neighbour *neigh = dst_get_neighbour(&rt->dst); + struct neighbour *neigh; int m; + + rcu_read_lock(); + neigh = dst_get_neighbour(&rt->dst); if (rt->rt6i_flags & RTF_NONEXTHOP || !(rt->rt6i_flags & RTF_GATEWAY)) m = 1; @@ -429,6 +437,7 @@ static inline int rt6_check_neigh(struct rt6_info *rt) read_unlock_bh(&neigh->lock); } else m = 0; + rcu_read_unlock(); return m; } @@ -767,7 +776,7 @@ static struct rt6_info *rt6_alloc_clone(struct rt6_info *ort, const struct in6_a rt->rt6i_dst.plen = 128; rt->rt6i_flags |= RTF_CACHE; rt->dst.flags |= DST_HOST; - dst_set_neighbour(&rt->dst, neigh_clone(dst_get_neighbour(&ort->dst))); + dst_set_neighbour(&rt->dst, neigh_clone(dst_get_neighbour_raw(&ort->dst))); } return rt; } @@ -801,7 +810,7 @@ restart: dst_hold(&rt->dst); read_unlock_bh(&table->tb6_lock); - if (!dst_get_neighbour(&rt->dst) && !(rt->rt6i_flags & RTF_NONEXTHOP)) + if (!dst_get_neighbour_raw(&rt->dst) && !(rt->rt6i_flags & RTF_NONEXTHOP)) nrt = rt6_alloc_cow(rt, &fl6->daddr, &fl6->saddr); else if (!(rt->dst.flags & DST_HOST)) nrt = rt6_alloc_clone(rt, &fl6->daddr); @@ -1581,7 +1590,7 @@ void rt6_redirect(const struct in6_addr *dest, const struct in6_addr *src, dst_confirm(&rt->dst); /* Duplicate redirect: silently ignore. */ - if (neigh == dst_get_neighbour(&rt->dst)) + if (neigh == dst_get_neighbour_raw(&rt->dst)) goto out; nrt = ip6_rt_copy(rt); @@ -1677,7 +1686,7 @@ again: 1. It is connected route. Action: COW 2. It is gatewayed route or NONEXTHOP route. Action: clone it. */ - if (!dst_get_neighbour(&rt->dst) && !(rt->rt6i_flags & RTF_NONEXTHOP)) + if (!dst_get_neighbour_raw(&rt->dst) && !(rt->rt6i_flags & RTF_NONEXTHOP)) nrt = rt6_alloc_cow(rt, daddr, saddr); else nrt = rt6_alloc_clone(rt, daddr); @@ -2319,6 +2328,7 @@ static int rt6_fill_node(struct net *net, struct nlmsghdr *nlh; long expires; u32 table; + struct neighbour *n; if (prefix) { /* user wants prefix routes only */ if (!(rt->rt6i_flags & RTF_PREFIX_RT)) { @@ -2407,8 +2417,11 @@ static int rt6_fill_node(struct net *net, if (rtnetlink_put_metrics(skb, dst_metrics_ptr(&rt->dst)) < 0) goto nla_put_failure; - if (dst_get_neighbour(&rt->dst)) - NLA_PUT(skb, RTA_GATEWAY, 16, &dst_get_neighbour(&rt->dst)->primary_key); + rcu_read_lock(); + n = dst_get_neighbour(&rt->dst); + if (n) + NLA_PUT(skb, RTA_GATEWAY, 16, &n->primary_key); + rcu_read_unlock(); if (rt->dst.dev) NLA_PUT_U32(skb, RTA_OIF, rt->rt6i_dev->ifindex); @@ -2601,12 +2614,14 @@ static int rt6_info_route(struct rt6_info *rt, void *p_arg) #else seq_puts(m, "00000000000000000000000000000000 00 "); #endif + rcu_read_lock(); n = dst_get_neighbour(&rt->dst); if (n) { seq_printf(m, "%pi6", n->primary_key); } else { seq_puts(m, "00000000000000000000000000000000"); } + rcu_read_unlock(); seq_printf(m, " %08x %08x %08x %08x %8s\n", rt->rt6i_metric, atomic_read(&rt->dst.__refcnt), rt->dst.__use, rt->rt6i_flags, -- 1.7.1