static int zero = 0;
 static int one = 1;
 static int label_limit = (1 << 20) - 1;
+static int ttl_max = 255;
 
 static void rtmsg_lfib(int event, u32 label, struct mpls_route *rt,
                       struct nlmsghdr *nlh, struct net *net, u32 portid,
                .extra1         = &zero,
                .extra2         = &one,
        },
+       {
+               .procname       = "default_ttl",
+               .data           = MPLS_NS_SYSCTL_OFFSET(mpls.default_ttl),
+               .maxlen         = sizeof(int),
+               .mode           = 0644,
+               .proc_handler   = proc_dointvec_minmax,
+               .extra1         = &one,
+               .extra2         = &ttl_max,
+       },
        { }
 };
 
        net->mpls.platform_labels = 0;
        net->mpls.platform_label = NULL;
        net->mpls.ip_ttl_propagate = 1;
+       net->mpls.default_ttl = 255;
 
        table = kmemdup(mpls_table, sizeof(mpls_table), GFP_KERNEL);
        if (table == NULL)
 
 
 static const struct nla_policy mpls_iptunnel_policy[MPLS_IPTUNNEL_MAX + 1] = {
        [MPLS_IPTUNNEL_DST]     = { .type = NLA_U32 },
+       [MPLS_IPTUNNEL_TTL]     = { .type = NLA_U8 },
 };
 
 static unsigned int mpls_encap_size(struct mpls_iptunnel_encap *en)
        struct rtable *rt = NULL;
        struct rt6_info *rt6 = NULL;
        struct mpls_dev *out_mdev;
+       struct net *net;
        int err = 0;
        bool bos;
        int i;
 
        /* Find the output device */
        out_dev = dst->dev;
-
-       /* Obtain the ttl */
-       if (dst->ops->family == AF_INET) {
-               ttl = ip_hdr(skb)->ttl;
-               rt = (struct rtable *)dst;
-       } else if (dst->ops->family == AF_INET6) {
-               ttl = ipv6_hdr(skb)->hop_limit;
-               rt6 = (struct rt6_info *)dst;
-       } else {
-               goto drop;
-       }
+       net = dev_net(out_dev);
 
        skb_orphan(skb);
 
 
        tun_encap_info = mpls_lwtunnel_encap(dst->lwtstate);
 
+       /* Obtain the ttl using the following set of rules.
+        *
+        * LWT ttl propagation setting:
+        *  - disabled => use default TTL value from LWT
+        *  - enabled  => use TTL value from IPv4/IPv6 header
+        *  - default  =>
+        *   Global ttl propagation setting:
+        *    - disabled => use default TTL value from global setting
+        *    - enabled => use TTL value from IPv4/IPv6 header
+        */
+       if (dst->ops->family == AF_INET) {
+               if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DISABLED)
+                       ttl = tun_encap_info->default_ttl;
+               else if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DEFAULT &&
+                        !net->mpls.ip_ttl_propagate)
+                       ttl = net->mpls.default_ttl;
+               else
+                       ttl = ip_hdr(skb)->ttl;
+               rt = (struct rtable *)dst;
+       } else if (dst->ops->family == AF_INET6) {
+               if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DISABLED)
+                       ttl = tun_encap_info->default_ttl;
+               else if (tun_encap_info->ttl_propagate == MPLS_TTL_PROP_DEFAULT &&
+                        !net->mpls.ip_ttl_propagate)
+                       ttl = net->mpls.default_ttl;
+               else
+                       ttl = ipv6_hdr(skb)->hop_limit;
+               rt6 = (struct rt6_info *)dst;
+       } else {
+               goto drop;
+       }
+
        /* Verify the destination can hold the packet */
        new_header_size = mpls_encap_size(tun_encap_info);
        mtu = mpls_dev_mtu(out_dev);
                             &tun_encap_info->labels, tun_encap_info->label);
        if (ret)
                goto errout;
+
+       tun_encap_info->ttl_propagate = MPLS_TTL_PROP_DEFAULT;
+
+       if (tb[MPLS_IPTUNNEL_TTL]) {
+               tun_encap_info->default_ttl = nla_get_u8(tb[MPLS_IPTUNNEL_TTL]);
+               /* TTL 0 implies propagate from IP header */
+               tun_encap_info->ttl_propagate = tun_encap_info->default_ttl ?
+                       MPLS_TTL_PROP_DISABLED :
+                       MPLS_TTL_PROP_ENABLED;
+       }
+
        newts->type = LWTUNNEL_ENCAP_MPLS;
        newts->flags |= LWTUNNEL_STATE_XMIT_REDIRECT;
        newts->headroom = mpls_encap_size(tun_encap_info);
                           tun_encap_info->label))
                goto nla_put_failure;
 
+       if (tun_encap_info->ttl_propagate != MPLS_TTL_PROP_DEFAULT &&
+           nla_put_u8(skb, MPLS_IPTUNNEL_TTL, tun_encap_info->default_ttl))
+               goto nla_put_failure;
+
        return 0;
 
 nla_put_failure:
 static int mpls_encap_nlsize(struct lwtunnel_state *lwtstate)
 {
        struct mpls_iptunnel_encap *tun_encap_info;
+       int nlsize;
 
        tun_encap_info = mpls_lwtunnel_encap(lwtstate);
 
-       return nla_total_size(tun_encap_info->labels * 4);
+       nlsize = nla_total_size(tun_encap_info->labels * 4);
+
+       if (tun_encap_info->ttl_propagate != MPLS_TTL_PROP_DEFAULT)
+               nlsize += nla_total_size(1);
+
+       return nlsize;
 }
 
 static int mpls_encap_cmp(struct lwtunnel_state *a, struct lwtunnel_state *b)
        struct mpls_iptunnel_encap *b_hdr = mpls_lwtunnel_encap(b);
        int l;
 
-       if (a_hdr->labels != b_hdr->labels)
+       if (a_hdr->labels != b_hdr->labels ||
+           a_hdr->ttl_propagate != b_hdr->ttl_propagate ||
+           a_hdr->default_ttl != b_hdr->default_ttl)
                return 1;
 
        for (l = 0; l < MAX_NEW_LABELS; l++)