netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Pablo Neira Ayuso <pablo@netfilter.org>
To: netfilter-devel@vger.kernel.org
Cc: davem@davemloft.net, netdev@vger.kernel.org
Subject: [PATCH 23/37] netfilter: nf_nat_sip: fix RTP/RTCP source port translations
Date: Fri, 21 Dec 2018 01:28:27 +0100	[thread overview]
Message-ID: <20181221002841.28587-24-pablo@netfilter.org> (raw)
In-Reply-To: <20181221002841.28587-1-pablo@netfilter.org>

From: Alin Nastac <alin.nastac@gmail.com>

Each media stream negotiation between 2 SIP peers will trigger creation
of 4 different expectations (2 RTP and 2 RTCP):
 - INVITE will create expectations for the media packets sent by the
   called peer
 - reply to the INVITE will create expectations for media packets sent
   by the caller

The dport used by these expectations usually match the ones selected
by the SIP peers, but they might get translated due to conflicts with
another expectation. When such event occur, it is important to do
this translation in both directions, dport translation on the receiving
path and sport translation on the sending path.

This commit fixes the sport translation when the peer requiring it is
also the one that starts the media stream. In this scenario, first media
stream packet is forwarded from LAN to WAN and will rely on
nf_nat_sip_expected() to do the necessary sport translation. However, the
expectation matched by this packet does not contain the necessary information
for doing SNAT, this data being stored in the paired expectation created by
the sender's SIP message (INVITE or reply to it).

Signed-off-by: Alin Nastac <alin.nastac@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/netfilter/nf_nat_sip.c | 39 +++++++++++++++++++++++++++++++++++----
 1 file changed, 35 insertions(+), 4 deletions(-)

diff --git a/net/netfilter/nf_nat_sip.c b/net/netfilter/nf_nat_sip.c
index 1f3086074981..aa1be643d7a0 100644
--- a/net/netfilter/nf_nat_sip.c
+++ b/net/netfilter/nf_nat_sip.c
@@ -18,6 +18,7 @@
 
 #include <net/netfilter/nf_nat.h>
 #include <net/netfilter/nf_nat_helper.h>
+#include <net/netfilter/nf_conntrack_core.h>
 #include <net/netfilter/nf_conntrack_helper.h>
 #include <net/netfilter/nf_conntrack_expect.h>
 #include <net/netfilter/nf_conntrack_seqadj.h>
@@ -316,6 +317,9 @@ static void nf_nat_sip_seq_adjust(struct sk_buff *skb, unsigned int protoff,
 static void nf_nat_sip_expected(struct nf_conn *ct,
 				struct nf_conntrack_expect *exp)
 {
+	struct nf_conn_help *help = nfct_help(ct->master);
+	struct nf_conntrack_expect *pair_exp;
+	int range_set_for_snat = 0;
 	struct nf_nat_range2 range;
 
 	/* This must be a fresh one. */
@@ -327,15 +331,42 @@ static void nf_nat_sip_expected(struct nf_conn *ct,
 	range.min_addr = range.max_addr = exp->saved_addr;
 	nf_nat_setup_info(ct, &range, NF_NAT_MANIP_DST);
 
-	/* Change src to where master sends to, but only if the connection
-	 * actually came from the same source. */
-	if (nf_inet_addr_cmp(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3,
+	/* Do media streams SRC manip according with the parameters
+	 * found in the paired expectation.
+	 */
+	if (exp->class != SIP_EXPECT_SIGNALLING) {
+		spin_lock_bh(&nf_conntrack_expect_lock);
+		hlist_for_each_entry(pair_exp, &help->expectations, lnode) {
+			if (pair_exp->tuple.src.l3num == nf_ct_l3num(ct) &&
+			    pair_exp->tuple.dst.protonum == ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.protonum &&
+			    nf_inet_addr_cmp(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3, &pair_exp->saved_addr) &&
+			    ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u.all == pair_exp->saved_proto.all) {
+				range.flags = (NF_NAT_RANGE_MAP_IPS | NF_NAT_RANGE_PROTO_SPECIFIED);
+				range.min_proto.all = range.max_proto.all = pair_exp->tuple.dst.u.all;
+				range.min_addr = range.max_addr = pair_exp->tuple.dst.u3;
+				range_set_for_snat = 1;
+				break;
+			}
+		}
+		spin_unlock_bh(&nf_conntrack_expect_lock);
+	}
+
+	/* When no paired expectation has been found, change src to
+	 * where master sends to, but only if the connection actually came
+	 * from the same source.
+	 */
+	if (!range_set_for_snat &&
+	    nf_inet_addr_cmp(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3,
 			     &ct->master->tuplehash[exp->dir].tuple.src.u3)) {
 		range.flags = NF_NAT_RANGE_MAP_IPS;
 		range.min_addr = range.max_addr
 			= ct->master->tuplehash[!exp->dir].tuple.dst.u3;
-		nf_nat_setup_info(ct, &range, NF_NAT_MANIP_SRC);
+		range_set_for_snat = 1;
 	}
+
+	/* Perform SRC manip. */
+	if (range_set_for_snat)
+		nf_nat_setup_info(ct, &range, NF_NAT_MANIP_SRC);
 }
 
 static unsigned int nf_nat_sip_expect(struct sk_buff *skb, unsigned int protoff,
-- 
2.11.0

  parent reply	other threads:[~2018-12-21  0:29 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-21  0:28 [PATCH 00/37] Netfilter updates for net-next Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 01/37] netfilter: ipset: Allow matching on destination MAC address for mac and ipmac sets Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 02/37] netfilter: ipset: Make invalid MAC address checks consistent Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 03/37] netfilter: ipset: Introduction of new commands and protocol version 7 Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 04/37] netfilter: ctnetlink: always honor CTA_MARK_MASK Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 05/37] netfilter: nf_flow_table: make nf_flow_table_iterate() static Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 06/37] netfilter: nf_flow_table: simplify nf_flow_offload_gc_step() Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 07/37] netfilter: Replace call_rcu_bh(), rcu_barrier_bh(), and synchronize_rcu_bh() Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 08/37] netfilter: remove NFC_* cache bits Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 09/37] netfilter: nat: remove l4 protocol port rovers Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 10/37] netfilter: ipset: fix ip_set_byindex function Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 11/37] netfilter: ipset: replace a strncpy() with strscpy() Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 12/37] netfilter: nat: remove unnecessary 'else if' branch Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 13/37] netfilter: nat: limit port clash resolution attempts Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 14/37] netfilter: remove unused parameters in nf_ct_l4proto_[un]register_sysctl() Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 15/37] netfilter: remove NF_NAT_RANGE_PROTO_RANDOM support Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 16/37] netfilter: nat: un-export nf_nat_l4proto_unique_tuple Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 17/37] netfilter: nat: remove l4proto->unique_tuple Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 18/37] netfilter: nat: fold in_range indirection into caller Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 19/37] netfilter: nat: remove l4proto->in_range Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 20/37] netfilter: nat: remove l4proto->nlattr_to_range Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 21/37] netfilter: nat: remove l4proto->manip_pkt Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 22/37] netfilter: nat: remove nf_nat_l4proto struct Pablo Neira Ayuso
2018-12-21  0:28 ` Pablo Neira Ayuso [this message]
2018-12-21  0:28 ` [PATCH 24/37] netfilter: nf_tables: Speed up selective rule dumps Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 25/37] netfilter: ipt_CLUSTERIP: fix deadlock in netns exit routine Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 26/37] netfilter: ipt_CLUSTERIP: remove wrong WARN_ON_ONCE " Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 27/37] netfilter: ipt_CLUSTERIP: fix sleep-in-atomic bug in clusterip_config_entry_put() Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 28/37] netfilter: ipt_CLUSTERIP: check MAC address when duplicate config is set Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 29/37] netfilter: conntrack: udp: only extend timeout to stream mode after 2s Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 30/37] netfilter: conntrack: udp: set stream timeout to 2 minutes Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 31/37] netfilter: conntrack: register sysctl table for gre Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 32/37] netfilter: conntrack: un-export seq_print_acct Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 33/37] netfilter: conntrack: add mnemonics for sysctl table Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 34/37] netfilter: conntrack: merge acct and helper sysctl table with main one Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 35/37] netfilter: conntrack: merge ecache and timestamp sysctl tables " Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 36/37] netfilter: conntrack: remove empty pernet fini stubs Pablo Neira Ayuso
2018-12-21  0:28 ` [PATCH 37/37] netfilter: netns: shrink netns_ct struct Pablo Neira Ayuso
2018-12-21  2:48 ` [PATCH 00/37] Netfilter updates for net-next David Miller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181221002841.28587-24-pablo@netfilter.org \
    --to=pablo@netfilter.org \
    --cc=davem@davemloft.net \
    --cc=netdev@vger.kernel.org \
    --cc=netfilter-devel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).