datapath: Fix Tunnel options TOS

Use DSCP bits from ToS set on tunnel.

Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Bug #8822

diff --git a/datapath/tunnel.c b/datapath/tunnel.c
index c2133bb4722a15b8f378f539d49d4022a51eecf2..d651c11833885a573cc743d1a08bc6d2a40e001e 100644 (file)
--- a/datapath/tunnel.c
+++ b/datapath/tunnel.c
@@ -1002,12 +1002,15 @@ 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(port_key_get_net(&mutable->key), &rt, &fl)))
@@ -1017,7 +1020,7 @@ 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(port_key_get_net(&mutable->key), &fl);
@@ -1034,7 +1037,7 @@ 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;
@@ -1045,7 +1048,7 @@ 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;
@@ -1219,8 +1222,6 @@ 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))
@@ -1228,6 +1229,8 @@ 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);
@@ -1402,7 +1405,8 @@ static int tnl_set_config(struct net *net, struct nlattr *options,
 
        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 94760ccf81ce077ee45cf05b170aee8d08e348e1..0cd9b30cb7c60be65db3bef9ec07377c8fced151 100644 (file)
--- a/vswitchd/vswitch.xml
+++ b/vswitchd/vswitch.xml
@@ -1267,7 +1267,8 @@
 
       <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.