Merge branch 'ip_tunnel-next'

wenxu says:

====================
ip_tunnel: Refactor ip_gre collect metadata xmit to ip_md_tunnel_xmit

This patchset add tunnel_dst_cache and tnl_update_pmtu feature for
ip_md_tunnel_xmit also bugfix. Then Refactor collect metatdata mode
tunnel xmit to ip_md_tunnel_xmit
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/ip_tunnels.h

@@ -267,7 +267,7 @@ void ip_tunnel_delete_nets(struct list_head *list_net, unsigned int id,
 void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
 		    const struct iphdr *tnl_params, const u8 protocol);
 void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
-		       const u8 proto);
+		       const u8 proto, int tunnel_hlen);
 int ip_tunnel_ioctl(struct net_device *dev, struct ip_tunnel_parm *p, int cmd);
 int __ip_tunnel_change_mtu(struct net_device *dev, int new_mtu, bool strict);
 int ip_tunnel_change_mtu(struct net_device *dev, int new_mtu);

net/ipv4/ip_gre.c

@@ -458,81 +458,14 @@ static int gre_handle_offloads(struct sk_buff *skb, bool csum)
 	return iptunnel_handle_offloads(skb, csum ? SKB_GSO_GRE_CSUM : SKB_GSO_GRE);
 }
 
-static struct rtable *gre_get_rt(struct sk_buff *skb,
-				 struct net_device *dev,
-				 struct flowi4 *fl,
-				 const struct ip_tunnel_key *key)
-{
-	struct net *net = dev_net(dev);
-
-	memset(fl, 0, sizeof(*fl));
-	fl->daddr = key->u.ipv4.dst;
-	fl->saddr = key->u.ipv4.src;
-	fl->flowi4_tos = RT_TOS(key->tos);
-	fl->flowi4_mark = skb->mark;
-	fl->flowi4_proto = IPPROTO_GRE;
-
-	return ip_route_output_key(net, fl);
-}
-
-static struct rtable *prepare_fb_xmit(struct sk_buff *skb,
-				      struct net_device *dev,
-				      struct flowi4 *fl,
-				      int tunnel_hlen)
-{
-	struct ip_tunnel_info *tun_info;
-	const struct ip_tunnel_key *key;
-	struct rtable *rt = NULL;
-	int min_headroom;
-	bool use_cache;
-	int err;
-
-	tun_info = skb_tunnel_info(skb);
-	key = &tun_info->key;
-	use_cache = ip_tunnel_dst_cache_usable(skb, tun_info);
-
-	if (use_cache)
-		rt = dst_cache_get_ip4(&tun_info->dst_cache, &fl->saddr);
-	if (!rt) {
-		rt = gre_get_rt(skb, dev, fl, key);
-		if (IS_ERR(rt))
-			goto err_free_skb;
-		if (use_cache)
-			dst_cache_set_ip4(&tun_info->dst_cache, &rt->dst,
-					  fl->saddr);
-	}
-
-	min_headroom = LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len
-			+ tunnel_hlen + sizeof(struct iphdr);
-	if (skb_headroom(skb) < min_headroom || skb_header_cloned(skb)) {
-		int head_delta = SKB_DATA_ALIGN(min_headroom -
-						skb_headroom(skb) +
-						16);
-		err = pskb_expand_head(skb, max_t(int, head_delta, 0),
-				       0, GFP_ATOMIC);
-		if (unlikely(err))
-			goto err_free_rt;
-	}
-	return rt;
-
-err_free_rt:
-	ip_rt_put(rt);
-err_free_skb:
-	kfree_skb(skb);
-	dev->stats.tx_dropped++;
-	return NULL;
-}
-
 static void gre_fb_xmit(struct sk_buff *skb, struct net_device *dev,
 			__be16 proto)
 {
 	struct ip_tunnel *tunnel = netdev_priv(dev);
 	struct ip_tunnel_info *tun_info;
 	const struct ip_tunnel_key *key;
-	struct rtable *rt = NULL;
-	struct flowi4 fl;
 	int tunnel_hlen;
-	__be16 df, flags;
+	__be16 flags;
 
 	tun_info = skb_tunnel_info(skb);
 	if (unlikely(!tun_info || !(tun_info->mode & IP_TUNNEL_INFO_TX) ||
@@ -542,13 +475,12 @@ static void gre_fb_xmit(struct sk_buff *skb, struct net_device *dev,
 	key = &tun_info->key;
 	tunnel_hlen = gre_calc_hlen(key->tun_flags);
 
-	rt = prepare_fb_xmit(skb, dev, &fl, tunnel_hlen);
-	if (!rt)
-		return;
+	if (skb_cow_head(skb, dev->needed_headroom))
+		goto err_free_skb;
 
 	/* Push Tunnel header. */
 	if (gre_handle_offloads(skb, !!(tun_info->key.tun_flags & TUNNEL_CSUM)))
-		goto err_free_rt;
+		goto err_free_skb;
 
 	flags = tun_info->key.tun_flags &
 		(TUNNEL_CSUM | TUNNEL_KEY | TUNNEL_SEQ);
@@ -556,14 +488,10 @@ static void gre_fb_xmit(struct sk_buff *skb, struct net_device *dev,
 			 tunnel_id_to_key32(tun_info->key.tun_id),
 			 (flags & TUNNEL_SEQ) ? htonl(tunnel->o_seqno++) : 0);
 
-	df = key->tun_flags & TUNNEL_DONT_FRAGMENT ?  htons(IP_DF) : 0;
+	ip_md_tunnel_xmit(skb, dev, IPPROTO_GRE, tunnel_hlen);
 
-	iptunnel_xmit(skb->sk, rt, skb, fl.saddr, key->u.ipv4.dst, IPPROTO_GRE,
-		      key->tos, key->ttl, df, false);
 	return;
 
-err_free_rt:
-	ip_rt_put(rt);
 err_free_skb:
 	kfree_skb(skb);
 	dev->stats.tx_dropped++;
@@ -575,10 +503,8 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct ip_tunnel_info *tun_info;
 	const struct ip_tunnel_key *key;
 	struct erspan_metadata *md;
-	struct rtable *rt = NULL;
 	bool truncate = false;
-	__be16 df, proto;
-	struct flowi4 fl;
+	__be16 proto;
 	int tunnel_hlen;
 	int version;
 	int nhoff;
@@ -591,21 +517,20 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	key = &tun_info->key;
 	if (!(tun_info->key.tun_flags & TUNNEL_ERSPAN_OPT))
-		goto err_free_rt;
+		goto err_free_skb;
 	md = ip_tunnel_info_opts(tun_info);
 	if (!md)
-		goto err_free_rt;
+		goto err_free_skb;
 
 	/* ERSPAN has fixed 8 byte GRE header */
 	version = md->version;
 	tunnel_hlen = 8 + erspan_hdr_len(version);
 
-	rt = prepare_fb_xmit(skb, dev, &fl, tunnel_hlen);
-	if (!rt)
-		return;
+	if (skb_cow_head(skb, dev->needed_headroom))
+		goto err_free_skb;
 
 	if (gre_handle_offloads(skb, false))
-		goto err_free_rt;
+		goto err_free_skb;
 
 	if (skb->len > dev->mtu + dev->hard_header_len) {
 		pskb_trim(skb, dev->mtu + dev->hard_header_len);
@@ -634,20 +559,16 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev)
 				       truncate, true);
 		proto = htons(ETH_P_ERSPAN2);
 	} else {
-		goto err_free_rt;
+		goto err_free_skb;
 	}
 
 	gre_build_header(skb, 8, TUNNEL_SEQ,
 			 proto, 0, htonl(tunnel->o_seqno++));
 
-	df = key->tun_flags & TUNNEL_DONT_FRAGMENT ?  htons(IP_DF) : 0;
+	ip_md_tunnel_xmit(skb, dev, IPPROTO_GRE, tunnel_hlen);
 
-	iptunnel_xmit(skb->sk, rt, skb, fl.saddr, key->u.ipv4.dst, IPPROTO_GRE,
-		      key->tos, key->ttl, df, false);
 	return;
 
-err_free_rt:
-	ip_rt_put(rt);
 err_free_skb:
 	kfree_skb(skb);
 	dev->stats.tx_dropped++;
@@ -656,13 +577,18 @@ static void erspan_fb_xmit(struct sk_buff *skb, struct net_device *dev)
 static int gre_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 {
 	struct ip_tunnel_info *info = skb_tunnel_info(skb);
+	const struct ip_tunnel_key *key;
 	struct rtable *rt;
 	struct flowi4 fl4;
 
 	if (ip_tunnel_info_af(info) != AF_INET)
 		return -EINVAL;
 
-	rt = gre_get_rt(skb, dev, &fl4, &info->key);
+	key = &info->key;
+	ip_tunnel_init_flow(&fl4, IPPROTO_GRE, key->u.ipv4.dst, key->u.ipv4.src,
+			    tunnel_id_to_key32(key->tun_id), key->tos, 0,
+			    skb->mark);
+	rt = ip_route_output_key(dev_net(dev), &fl4);
 	if (IS_ERR(rt))
 		return PTR_ERR(rt);
 

net/ipv4/ip_tunnel.c

@@ -501,15 +501,19 @@ EXPORT_SYMBOL_GPL(ip_tunnel_encap_setup);
 
 static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,
 			    struct rtable *rt, __be16 df,
-			    const struct iphdr *inner_iph)
+			    const struct iphdr *inner_iph,
+			    int tunnel_hlen, __be32 dst, bool md)
 {
 	struct ip_tunnel *tunnel = netdev_priv(dev);
-	int pkt_size = skb->len - tunnel->hlen - dev->hard_header_len;
+	int pkt_size;
 	int mtu;
 
+	tunnel_hlen = md ? tunnel_hlen : tunnel->hlen;
+	pkt_size = skb->len - tunnel_hlen - dev->hard_header_len;
+
 	if (df)
 		mtu = dst_mtu(&rt->dst) - dev->hard_header_len
-					- sizeof(struct iphdr) - tunnel->hlen;
+					- sizeof(struct iphdr) - tunnel_hlen;
 	else
 		mtu = skb_dst(skb) ? dst_mtu(skb_dst(skb)) : dev->mtu;
 
@@ -527,11 +531,13 @@ static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,
 #if IS_ENABLED(CONFIG_IPV6)
 	else if (skb->protocol == htons(ETH_P_IPV6)) {
 		struct rt6_info *rt6 = (struct rt6_info *)skb_dst(skb);
+		__be32 daddr;
+
+		daddr = md ? dst : tunnel->parms.iph.daddr;
 
 		if (rt6 && mtu < dst_mtu(skb_dst(skb)) &&
 			   mtu >= IPV6_MIN_MTU) {
-			if ((tunnel->parms.iph.daddr &&
-			    !ipv4_is_multicast(tunnel->parms.iph.daddr)) ||
+			if ((daddr && !ipv4_is_multicast(daddr)) ||
 			    rt6->rt6i_dst.plen == 128) {
 				rt6->rt6i_flags |= RTF_MODIFIED;
 				dst_metric_set(skb_dst(skb), RTAX_MTU, mtu);
@@ -548,17 +554,19 @@ static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,
 	return 0;
 }
 
-void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, u8 proto)
+void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
+		       u8 proto, int tunnel_hlen)
 {
 	struct ip_tunnel *tunnel = netdev_priv(dev);
 	u32 headroom = sizeof(struct iphdr);
 	struct ip_tunnel_info *tun_info;
 	const struct ip_tunnel_key *key;
 	const struct iphdr *inner_iph;
-	struct rtable *rt;
+	struct rtable *rt = NULL;
 	struct flowi4 fl4;
 	__be16 df = 0;
 	u8 tos, ttl;
+	bool use_cache;
 
 	tun_info = skb_tunnel_info(skb);
 	if (unlikely(!tun_info || !(tun_info->mode & IP_TUNNEL_INFO_TX) ||
@@ -574,20 +582,39 @@ void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, u8 proto)
 		else if (skb->protocol == htons(ETH_P_IPV6))
 			tos = ipv6_get_dsfield((const struct ipv6hdr *)inner_iph);
 	}
-	ip_tunnel_init_flow(&fl4, proto, key->u.ipv4.dst, key->u.ipv4.src, 0,
-			    RT_TOS(tos), tunnel->parms.link, tunnel->fwmark);
+	ip_tunnel_init_flow(&fl4, proto, key->u.ipv4.dst, key->u.ipv4.src,
+			    tunnel_id_to_key32(key->tun_id), RT_TOS(tos),
+			    0, skb->mark);
 	if (tunnel->encap.type != TUNNEL_ENCAP_NONE)
 		goto tx_error;
-	rt = ip_route_output_key(tunnel->net, &fl4);
-	if (IS_ERR(rt)) {
-		dev->stats.tx_carrier_errors++;
-		goto tx_error;
+
+	use_cache = ip_tunnel_dst_cache_usable(skb, tun_info);
+	if (use_cache)
+		rt = dst_cache_get_ip4(&tun_info->dst_cache, &fl4.saddr);
+	if (!rt) {
+		rt = ip_route_output_key(tunnel->net, &fl4);
+		if (IS_ERR(rt)) {
+			dev->stats.tx_carrier_errors++;
+			goto tx_error;
+		}
+		if (use_cache)
+			dst_cache_set_ip4(&tun_info->dst_cache, &rt->dst,
+					  fl4.saddr);
 	}
 	if (rt->dst.dev == dev) {
 		ip_rt_put(rt);
 		dev->stats.collisions++;
 		goto tx_error;
 	}
+
+	if (key->tun_flags & TUNNEL_DONT_FRAGMENT)
+		df = htons(IP_DF);
+	if (tnl_update_pmtu(dev, skb, rt, df, inner_iph, tunnel_hlen,
+			    key->u.ipv4.dst, true)) {
+		ip_rt_put(rt);
+		goto tx_error;
+	}
+
 	tos = ip_tunnel_ecn_encap(tos, inner_iph, skb);
 	ttl = key->ttl;
 	if (ttl == 0) {
@@ -598,10 +625,10 @@ void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, u8 proto)
 		else
 			ttl = ip4_dst_hoplimit(&rt->dst);
 	}
-	if (key->tun_flags & TUNNEL_DONT_FRAGMENT)
-		df = htons(IP_DF);
-	else if (skb->protocol == htons(ETH_P_IP))
+
+	if (!df && skb->protocol == htons(ETH_P_IP))
 		df = inner_iph->frag_off & htons(IP_DF);
+
 	headroom += LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len;
 	if (headroom > dev->needed_headroom)
 		dev->needed_headroom = headroom;
@@ -731,7 +758,8 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
 		goto tx_error;
 	}
 
-	if (tnl_update_pmtu(dev, skb, rt, tnl_params->frag_off, inner_iph)) {
+	if (tnl_update_pmtu(dev, skb, rt, tnl_params->frag_off, inner_iph,
+			    0, 0, false)) {
 		ip_rt_put(rt);
 		goto tx_error;
 	}

net/ipv4/ipip.c

@@ -302,7 +302,7 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb,
 	skb_set_inner_ipproto(skb, ipproto);
 
 	if (tunnel->collect_md)
-		ip_md_tunnel_xmit(skb, dev, ipproto);
+		ip_md_tunnel_xmit(skb, dev, ipproto, 0);
 	else
 		ip_tunnel_xmit(skb, dev, tiph, ipproto);
 	return NETDEV_TX_OK;