[ovs-dev] [1.4 backports 6/6] datapath: Fix Tunnel options TOS

[Ben Pfaff]

From: Pravin B Shelar <pshelar at nicira.com>

Use DSCP bits from ToS set on tunnel.

This is a crossport of commit 749ae9504293dbb695dd67402acbd47acbcbeb83
from master.

Bug #8822.
Signed-off-by: Pravin B Shelar <pshelar at nicira.com>
Acked-by: Jesse Gross <jesse at nicira.com>
---
 datapath/tunnel.c    |   20 ++++++++++++--------
 vswitchd/vswitch.xml |    3 ++-
 2 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/datapath/tunnel.c b/datapath/tunnel.c
index 4ce830f..1ebbc77 100644
--- a/datapath/tunnel.c
+++ b/datapath/tunnel.c
@@ -991,12 +991,15 @@ unlock:
 static struct rtable *__find_route(const struct tnl_mutable_config *mutable,
 				   u8 ipproto, u8 tos)
 {
+	/* Tunnel configuration keeps DSCP part of TOS bits, But Linux
+	 * router expect RT_TOS bits only. */
+
 #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,39)
 	struct flowi fl = { .nl_u = { .ip4_u = {
 					.daddr = mutable->key.daddr,
 					.saddr = mutable->key.saddr,
-					.tos = tos } },
-			    .proto = ipproto };
+					.tos   = RT_TOS(tos) } },
+					.proto = ipproto };
 	struct rtable *rt;
 
 	if (unlikely(ip_route_output_key(&init_net, &rt, &fl)))
@@ -1006,7 +1009,7 @@ static struct rtable *__find_route(const struct tnl_mutable_config *mutable,
 #else
 	struct flowi4 fl = { .daddr = mutable->key.daddr,
 			     .saddr = mutable->key.saddr,
-			     .flowi4_tos = tos,
+			     .flowi4_tos = RT_TOS(tos),
 			     .flowi4_proto = ipproto };
 
 	return ip_route_output_key(&init_net, &fl);
@@ -1023,7 +1026,7 @@ static struct rtable *find_route(struct vport *vport,
 	*cache = NULL;
 	tos = RT_TOS(tos);
 
-	if (likely(tos == mutable->tos &&
+	if (likely(tos == RT_TOS(mutable->tos) &&
 	    check_cache_valid(cur_cache, mutable))) {
 		*cache = cur_cache;
 		return cur_cache->rt;
@@ -1034,7 +1037,7 @@ static struct rtable *find_route(struct vport *vport,
 		if (IS_ERR(rt))
 			return NULL;
 
-		if (likely(tos == mutable->tos))
+		if (likely(tos == RT_TOS(mutable->tos)))
 			*cache = build_cache(vport, mutable, rt);
 
 		return rt;
@@ -1208,8 +1211,6 @@ int ovs_tnl_send(struct vport *vport, struct sk_buff *skb)
 	else
 		tos = mutable->tos;
 
-	tos = INET_ECN_encapsulate(tos, inner_tos);
-
 	/* Route lookup */
 	rt = find_route(vport, mutable, tos, &cache);
 	if (unlikely(!rt))
@@ -1217,6 +1218,8 @@ int ovs_tnl_send(struct vport *vport, struct sk_buff *skb)
 	if (unlikely(!cache))
 		unattached_dst = &rt_dst(rt);
 
+	tos = INET_ECN_encapsulate(tos, inner_tos);
+
 	/* Reset SKB */
 	nf_reset(skb);
 	secpath_reset(skb);
@@ -1389,7 +1392,8 @@ static int tnl_set_config(struct nlattr *options, const struct tnl_ops *tnl_ops,
 
 	if (a[OVS_TUNNEL_ATTR_TOS]) {
 		mutable->tos = nla_get_u8(a[OVS_TUNNEL_ATTR_TOS]);
-		if (mutable->tos != RT_TOS(mutable->tos))
+		/* Reject ToS config with ECN bits set. */
+		if (mutable->tos & INET_ECN_MASK)
 			return -EINVAL;
 	}
 
diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
index ebbcbe7..303a98d 100644
--- a/vswitchd/vswitch.xml
+++ b/vswitchd/vswitch.xml
@@ -1259,7 +1259,8 @@
 
       <column name="options" key="tos">
         Optional.  The value of the ToS bits to be set on the encapsulating
-        packet.  It may also be the word <code>inherit</code>, in which case
+        packet.  ToS is interpreted as DSCP and ECN bits, ECN part must be
+        zero.  It may also be the word <code>inherit</code>, in which case
         the ToS will be copied from the inner packet if it is IPv4 or IPv6
         (otherwise it will be 0).  The ECN fields are always inherited.
         Default is 0.
-- 
1.7.2.5