On Tue, Jan 14, 2014 at 09:35:48PM +0400, Andrey Vagin wrote:
> ----
> Eric and Florian, could you look at this patch. When you say,
> that it looks good, I will ask the user to validate it.
> I can't reorder these actions, because it's reproduced on a real host
> with real users. Thanks.
We didn't get new reports from users for the last week, so these patches
fix the problem.
> ----
>
> nf_conntrack_free can't be called for a conntract with non-zero ref-counter,
> because it can race with nf_conntrack_find_get().
>
> A conntrack slab is created with SLAB_DESTROY_BY_RCU. Non-zero
> ref-conunter says that this conntrack is used now. So when we release a
> conntrack with non-zero counter, we break this assumption.
>
> CPU1 CPU2
> ____nf_conntrack_find()
> nf_ct_put()
> destroy_conntrack()
> ...
> init_conntrack
> __nf_conntrack_alloc (set use = 1)
> atomic_inc_not_zero(&ct->use) (use = 2)
> if (!l4proto->new(ct, skb, dataoff, timeouts))
> nf_conntrack_free(ct); (use = 2 !!!)
> ...
> __nf_conntrack_alloc (set use = 1)
> if (!nf_ct_key_equal(h, tuple, zone))
> nf_ct_put(ct); (use = 0)
> destroy_conntrack()
> /* continue to work with CT */
>
> After applying the path "[PATCH] netfilter: nf_conntrack: fix RCU race
> in nf_conntrack_find_get (v3)" another bug was triggered in
> destroy_conntrack():
> <4>[67096.759334] ------------[ cut here ]------------
> <2>[67096.759353] kernel BUG at net/netfilter/nf_conntrack_core.c:211!
> ...
> <4>[67096.759837] Pid: 498649, comm: atdd veid: 666 Tainted: G C --------------- 2.6.32-042stab084.18 #1 042stab084_18 /DQ45CB
> <4>[67096.759932] RIP: 0010:[<ffffffffa03d99ac>] [<ffffffffa03d99ac>] destroy_conntrack+0x15c/0x190 [nf_conntrack]
> <4>[67096.760255] Call Trace:
> <4>[67096.760255] [<ffffffff814844a7>] nf_conntrack_destroy+0x17/0x30
> <4>[67096.760255] [<ffffffffa03d9bb5>] nf_conntrack_find_get+0x85/0x130 [nf_conntrack]
> <4>[67096.760255] [<ffffffffa03d9fb2>] nf_conntrack_in+0x352/0xb60 [nf_conntrack]
> <4>[67096.760255] [<ffffffffa048c771>] ipv4_conntrack_local+0x51/0x60 [nf_conntrack_ipv4]
> <4>[67096.760255] [<ffffffff81484419>] nf_iterate+0x69/0xb0
> <4>[67096.760255] [<ffffffff814b5b00>] ? dst_output+0x0/0x20
> <4>[67096.760255] [<ffffffff814845d4>] nf_hook_slow+0x74/0x110
> <4>[67096.760255] [<ffffffff814b5b00>] ? dst_output+0x0/0x20
> <4>[67096.760255] [<ffffffff814b66d5>] raw_sendmsg+0x775/0x910
> <4>[67096.760255] [<ffffffff8104c5a8>] ? flush_tlb_others_ipi+0x128/0x130
> <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20
> <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20
> <4>[67096.760255] [<ffffffff814c136a>] inet_sendmsg+0x4a/0xb0
> <4>[67096.760255] [<ffffffff81444e93>] ? sock_sendmsg+0x13/0x140
> <4>[67096.760255] [<ffffffff81444f97>] sock_sendmsg+0x117/0x140
> <4>[67096.760255] [<ffffffff8102e299>] ? native_smp_send_reschedule+0x49/0x60
> <4>[67096.760255] [<ffffffff81519beb>] ? _spin_unlock_bh+0x1b/0x20
> <4>[67096.760255] [<ffffffff8109d930>] ? autoremove_wake_function+0x0/0x40
> <4>[67096.760255] [<ffffffff814960f0>] ? do_ip_setsockopt+0x90/0xd80
> <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20
> <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20
> <4>[67096.760255] [<ffffffff814457c9>] sys_sendto+0x139/0x190
> <4>[67096.760255] [<ffffffff810efa77>] ? audit_syscall_entry+0x1d7/0x200
> <4>[67096.760255] [<ffffffff810ef7c5>] ? __audit_syscall_exit+0x265/0x290
> <4>[67096.760255] [<ffffffff81474daf>] compat_sys_socketcall+0x13f/0x210
> <4>[67096.760255] [<ffffffff8104dea3>] ia32_sysret+0x0/0x5
>
> Cc: Eric Dumazet <eric.dumazet@xxxxxxxxx>
> Cc: Florian Westphal <fw@xxxxxxxxx>
> Cc: Cyrill Gorcunov <gorcunov@xxxxxxxxxx>
> Cc: Vasiliy Averin <vvs@xxxxxxxxxxxxx>
> Signed-off-by: Andrey Vagin <avagin@xxxxxxxxxx>
> ---
> include/net/netfilter/nf_conntrack.h | 1 -
> net/netfilter/nf_conntrack_core.c | 18 +++++++++++-------
> net/netfilter/nf_conntrack_netlink.c | 2 +-
> net/netfilter/nf_synproxy_core.c | 4 ++--
> net/netfilter/xt_CT.c | 2 +-
> 5 files changed, 15 insertions(+), 12 deletions(-)
>
> diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h
> index 01ea6ee..d338316 100644
> --- a/include/net/netfilter/nf_conntrack.h
> +++ b/include/net/netfilter/nf_conntrack.h
> @@ -243,7 +243,6 @@ void nf_ct_untracked_status_or(unsigned long bits);
> void nf_ct_iterate_cleanup(struct net *net,
> int (*iter)(struct nf_conn *i, void *data),
> void *data, u32 portid, int report);
> -void nf_conntrack_free(struct nf_conn *ct);
> struct nf_conn *nf_conntrack_alloc(struct net *net, u16 zone,
> const struct nf_conntrack_tuple *orig,
> const struct nf_conntrack_tuple *repl,
> diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
> index b56e53b..c38cc74 100644
> --- a/net/netfilter/nf_conntrack_core.c
> +++ b/net/netfilter/nf_conntrack_core.c
> @@ -198,6 +198,8 @@ clean_from_lists(struct nf_conn *ct)
> nf_ct_remove_expectations(ct);
> }
>
> +static void nf_conntrack_free(struct nf_conn *ct);
> +
> static void
> destroy_conntrack(struct nf_conntrack *nfct)
> {
> @@ -226,9 +228,8 @@ destroy_conntrack(struct nf_conntrack *nfct)
> * too. */
> nf_ct_remove_expectations(ct);
>
> - /* We overload first tuple to link into unconfirmed or dying list.*/
> - BUG_ON(hlist_nulls_unhashed(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnnode));
> - hlist_nulls_del_rcu(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnnode);
> + if (!hlist_nulls_unhashed(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnnode))
> + hlist_nulls_del_rcu(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnnode);
>
> NF_CT_STAT_INC(net, delete);
> spin_unlock_bh(&nf_conntrack_lock);
> @@ -772,18 +773,21 @@ struct nf_conn *nf_conntrack_alloc(struct net *net, u16 zone,
> }
> EXPORT_SYMBOL_GPL(nf_conntrack_alloc);
>
> -void nf_conntrack_free(struct nf_conn *ct)
> +static void nf_conntrack_free(struct nf_conn *ct)
> {
> struct net *net = nf_ct_net(ct);
>
> + /* A freed object has refcnt == 0, thats
> + * the golden rule for SLAB_DESTROY_BY_RCU
> + */
> + NF_CT_ASSERT(atomic_read(&ct->ct_general.use) == 0);
> +
> nf_ct_ext_destroy(ct);
> nf_ct_ext_free(ct);
> kmem_cache_free(net->ct.nf_conntrack_cachep, ct);
> smp_mb__before_atomic_dec();
> atomic_dec(&net->ct.count);
> }
> -EXPORT_SYMBOL_GPL(nf_conntrack_free);
> -
>
> /* Allocate a new conntrack: we return -ENOMEM if classification
> failed due to stress. Otherwise it really is unclassifiable. */
> @@ -835,7 +839,7 @@ init_conntrack(struct net *net, struct nf_conn *tmpl,
> }
>
> if (!l4proto->new(ct, skb, dataoff, timeouts)) {
> - nf_conntrack_free(ct);
> + nf_ct_put(ct);
> pr_debug("init conntrack: can't track with proto module\n");
> return NULL;
> }
> diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c
> index 3e91ad3..fadd0f3 100644
> --- a/net/netfilter/nf_conntrack_netlink.c
> +++ b/net/netfilter/nf_conntrack_netlink.c
> @@ -1732,7 +1732,7 @@ ctnetlink_create_conntrack(struct net *net, u16 zone,
> err2:
> rcu_read_unlock();
> err1:
> - nf_conntrack_free(ct);
> + nf_ct_put(ct);
> return ERR_PTR(err);
> }
>
> diff --git a/net/netfilter/nf_synproxy_core.c b/net/netfilter/nf_synproxy_core.c
> index 9858e3e..d12234c 100644
> --- a/net/netfilter/nf_synproxy_core.c
> +++ b/net/netfilter/nf_synproxy_core.c
> @@ -381,7 +381,7 @@ static int __net_init synproxy_net_init(struct net *net)
> err3:
> free_percpu(snet->stats);
> err2:
> - nf_conntrack_free(ct);
> + nf_ct_put(ct);
> err1:
> return err;
> }
> @@ -390,7 +390,7 @@ static void __net_exit synproxy_net_exit(struct net *net)
> {
> struct synproxy_net *snet = synproxy_pernet(net);
>
> - nf_conntrack_free(snet->tmpl);
> + nf_ct_put(snet->tmpl);
> synproxy_proc_exit(net);
> free_percpu(snet->stats);
> }
> diff --git a/net/netfilter/xt_CT.c b/net/netfilter/xt_CT.c
> index da35ac0..da4edfe 100644
> --- a/net/netfilter/xt_CT.c
> +++ b/net/netfilter/xt_CT.c
> @@ -237,7 +237,7 @@ out:
> return 0;
>
> err3:
> - nf_conntrack_free(ct);
> + nf_ct_put(ct);
> err2:
> nf_ct_l3proto_module_put(par->family);
> err1:
> --
> 1.8.4.2
>
--
To unsubscribe from this list: send the line "unsubscribe netfilter" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
