net-gro: Prepare GRO stack for the upcoming tunneling support

This patch modifies the GRO stack to avoid the use of "network_header"
and associated macros like ip_hdr() and ipv6_hdr() in order to allow
an arbitary number of IP hdrs (v4 or v6) to be used in the
encapsulation chain. This lays the foundation for various IP
tunneling support (IP-in-IP, GRE, VXLAN, SIT,...) to be added later.

With this patch, the GRO stack traversing now is mostly based on
skb_gro_offset rather than special hdr offsets saved in skb (e.g.,
skb->network_header). As a result all but the top layer (i.e., the
the transport layer) must have hdrs of the same length in order for
a pkt to be considered for aggregation. Therefore when adding a new
encap layer (e.g., for tunneling), one must check and skip flows
(e.g., by setting NAPI_GRO_CB(p)->same_flow to 0) that have a
different hdr length.

Note that unlike the network header, the transport header can and
will continue to be set by the GRO code since there will be at
most one "transport layer" in the encap chain.

Signed-off-by: H.K. Jerry Chu <hkchu@google.com>
Suggested-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

5 files changed, 96 insertions, 74 deletions

diff --git a/net/core/dev.c b/net/core/dev.c
index 355df36360b4..c95d664b2b42 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3752,7 +3752,7 @@ static int napi_gro_complete(struct sk_buff *skb)
 		if (ptype->type != type || !ptype->callbacks.gro_complete)
 			continue;
 
-		err = ptype->callbacks.gro_complete(skb);
+		err = ptype->callbacks.gro_complete(skb, 0);
 		break;
 	}
 	rcu_read_unlock();
@@ -3818,6 +3818,23 @@ static void gro_list_prepare(struct napi_struct *napi, struct sk_buff *skb)
 	}
 }
 
+static void skb_gro_reset_offset(struct sk_buff *skb)
+{
+	const struct skb_shared_info *pinfo = skb_shinfo(skb);
+	const skb_frag_t *frag0 = &pinfo->frags[0];
+
+	NAPI_GRO_CB(skb)->data_offset = 0;
+	NAPI_GRO_CB(skb)->frag0 = NULL;
+	NAPI_GRO_CB(skb)->frag0_len = 0;
+
+	if (skb_mac_header(skb) == skb_tail_pointer(skb) &&
+	    pinfo->nr_frags &&
+	    !PageHighMem(skb_frag_page(frag0))) {
+		NAPI_GRO_CB(skb)->frag0 = skb_frag_address(frag0);
+		NAPI_GRO_CB(skb)->frag0_len = skb_frag_size(frag0);
+	}
+}
+
 static enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
 {
 	struct sk_buff **pp = NULL;
@@ -3833,6 +3850,7 @@ static enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff
 	if (skb_is_gso(skb) || skb_has_frag_list(skb))
 		goto normal;
 
+	skb_gro_reset_offset(skb);
 	gro_list_prepare(napi, skb);
 
 	rcu_read_lock();
@@ -3938,27 +3956,8 @@ static gro_result_t napi_skb_finish(gro_result_t ret, struct sk_buff *skb)
 	return ret;
 }
 
-static void skb_gro_reset_offset(struct sk_buff *skb)
-{
-	const struct skb_shared_info *pinfo = skb_shinfo(skb);
-	const skb_frag_t *frag0 = &pinfo->frags[0];
-
-	NAPI_GRO_CB(skb)->data_offset = 0;
-	NAPI_GRO_CB(skb)->frag0 = NULL;
-	NAPI_GRO_CB(skb)->frag0_len = 0;
-
-	if (skb_mac_header(skb) == skb_tail_pointer(skb) &&
-	    pinfo->nr_frags &&
-	    !PageHighMem(skb_frag_page(frag0))) {
-		NAPI_GRO_CB(skb)->frag0 = skb_frag_address(frag0);
-		NAPI_GRO_CB(skb)->frag0_len = skb_frag_size(frag0);
-	}
-}
-
 gro_result_t napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
 {
-	skb_gro_reset_offset(skb);
-
 	return napi_skb_finish(dev_gro_receive(napi, skb), skb);
 }
 EXPORT_SYMBOL(napi_gro_receive);
@@ -3992,12 +3991,7 @@ static gro_result_t napi_frags_finish(struct napi_struct *napi, struct sk_buff *
 {
 	switch (ret) {
 	case GRO_NORMAL:
-	case GRO_HELD:
-		skb->protocol = eth_type_trans(skb, skb->dev);
-
-		if (ret == GRO_HELD)
-			skb_gro_pull(skb, -ETH_HLEN);
-		else if (netif_receive_skb(skb))
+		if (netif_receive_skb(skb))
 			ret = GRO_DROP;
 		break;
 
@@ -4006,6 +4000,7 @@ static gro_result_t napi_frags_finish(struct napi_struct *napi, struct sk_buff *
 		napi_reuse_skb(napi, skb);
 		break;
 
+	case GRO_HELD:
 	case GRO_MERGED:
 		break;
 	}
@@ -4016,36 +4011,15 @@ static gro_result_t napi_frags_finish(struct napi_struct *napi, struct sk_buff *
 static struct sk_buff *napi_frags_skb(struct napi_struct *napi)
 {
 	struct sk_buff *skb = napi->skb;
-	struct ethhdr *eth;
-	unsigned int hlen;
-	unsigned int off;
 
 	napi->skb = NULL;
 
-	skb_reset_mac_header(skb);
-	skb_gro_reset_offset(skb);
-
-	off = skb_gro_offset(skb);
-	hlen = off + sizeof(*eth);
-	eth = skb_gro_header_fast(skb, off);
-	if (skb_gro_header_hard(skb, hlen)) {
-		eth = skb_gro_header_slow(skb, hlen, off);
-		if (unlikely(!eth)) {
-			napi_reuse_skb(napi, skb);
-			skb = NULL;
-			goto out;
-		}
+	if (unlikely(!pskb_may_pull(skb, sizeof(struct ethhdr)))) {
+		napi_reuse_skb(napi, skb);
+		return NULL;
 	}
+	skb->protocol = eth_type_trans(skb, skb->dev);
 
-	skb_gro_pull(skb, sizeof(*eth));
-
-	/*
-	 * This works because the only protocols we care about don't require
-	 * special handling.  We'll fix it up properly at the end.
-	 */
-	skb->protocol = eth->h_proto;
-
-out:
 	return skb;
 }
 
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 70011e029ac1..ef4f9df6d698 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1377,8 +1377,12 @@ static struct sk_buff **inet_gro_receive(struct sk_buff **head,
 		if (!NAPI_GRO_CB(p)->same_flow)
 			continue;
 
-		iph2 = ip_hdr(p);
-
+		iph2 = (struct iphdr *)(p->data + off);
+		/* The above works because, with the exception of the top
+		 * (inner most) layer, we only aggregate pkts with the same
+		 * hdr length so all the hdrs we'll need to verify will start
+		 * at the same offset.
+		 */
 		if ((iph->protocol ^ iph2->protocol) |
 		    ((__force u32)iph->saddr ^ (__force u32)iph2->saddr) |
 		    ((__force u32)iph->daddr ^ (__force u32)iph2->daddr)) {
@@ -1397,6 +1401,11 @@ static struct sk_buff **inet_gro_receive(struct sk_buff **head,
 	}
 
 	NAPI_GRO_CB(skb)->flush |= flush;
+	skb_set_network_header(skb, off);
+	/* The above will be needed by the transport layer if there is one
+	 * immediately following this IP hdr.
+	 */
+
 	skb_gro_pull(skb, sizeof(*iph));
 	skb_set_transport_header(skb, skb_gro_offset(skb));
 
@@ -1411,10 +1420,10 @@ out:
 	return pp;
 }
 
-static int inet_gro_complete(struct sk_buff *skb)
+static int inet_gro_complete(struct sk_buff *skb, int nhoff)
 {
-	__be16 newlen = htons(skb->len - skb_network_offset(skb));
-	struct iphdr *iph = ip_hdr(skb);
+	__be16 newlen = htons(skb->len - nhoff);
+	struct iphdr *iph = (struct iphdr *)(skb->data + nhoff);
 	const struct net_offload *ops;
 	int proto = iph->protocol;
 	int err = -ENOSYS;
@@ -1427,7 +1436,11 @@ static int inet_gro_complete(struct sk_buff *skb)
 	if (WARN_ON(!ops || !ops->callbacks.gro_complete))
 		goto out_unlock;
 
-	err = ops->callbacks.gro_complete(skb);
+	/* Only need to add sizeof(*iph) to get to the next hdr below
+	 * because any hdr with option will have been flushed in
+	 * inet_gro_receive().
+	 */
+	err = ops->callbacks.gro_complete(skb, nhoff + sizeof(*iph));
 
 out_unlock:
 	rcu_read_unlock();
diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
index 05606353c7e7..2658a27f540d 100644
--- a/net/ipv4/tcp_offload.c
+++ b/net/ipv4/tcp_offload.c
@@ -240,7 +240,7 @@ int tcp_gro_complete(struct sk_buff *skb)
 {
 	struct tcphdr *th = tcp_hdr(skb);
 
-	skb->csum_start = skb_transport_header(skb) - skb->head;
+	skb->csum_start = (unsigned char *)th - skb->head;
 	skb->csum_offset = offsetof(struct tcphdr, check);
 	skb->ip_summed = CHECKSUM_PARTIAL;
 
@@ -272,6 +272,7 @@ static int tcp_v4_gso_send_check(struct sk_buff *skb)
 
 static struct sk_buff **tcp4_gro_receive(struct sk_buff **head, struct sk_buff *skb)
 {
+	/* Use the IP hdr immediately proceeding for this transport */
 	const struct iphdr *iph = skb_gro_network_header(skb);
 	__wsum wsum;
 
@@ -303,13 +304,13 @@ skip_csum:
 	return tcp_gro_receive(head, skb);
 }
 
-static int tcp4_gro_complete(struct sk_buff *skb)
+static int tcp4_gro_complete(struct sk_buff *skb, int thoff)
 {
 	const struct iphdr *iph = ip_hdr(skb);
 	struct tcphdr *th = tcp_hdr(skb);
 
-	th->check = ~tcp_v4_check(skb->len - skb_transport_offset(skb),
-				  iph->saddr, iph->daddr, 0);
+	th->check = ~tcp_v4_check(skb->len - thoff, iph->saddr,
+				  iph->daddr, 0);
 	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
 
 	return tcp_gro_complete(skb);
diff --git a/net/ipv6/ip6_offload.c b/net/ipv6/ip6_offload.c
index 4b851692b1f6..7540a0ed75ae 100644
--- a/net/ipv6/ip6_offload.c
+++ b/net/ipv6/ip6_offload.c
@@ -154,6 +154,35 @@ out:
 	return segs;
 }
 
+/* Return the total length of all the extension hdrs, following the same
+ * logic in ipv6_gso_pull_exthdrs() when parsing ext-hdrs.
+ */
+static int ipv6_exthdrs_len(struct ipv6hdr *iph,
+			    const struct net_offload **opps)
+{
+	struct ipv6_opt_hdr *opth = NULL;
+	int len = 0, proto, optlen;
+
+	proto = iph->nexthdr;
+	for (;;) {
+		if (proto != NEXTHDR_HOP) {
+			*opps = rcu_dereference(inet6_offloads[proto]);
+			if (unlikely(!(*opps)))
+				break;
+			if (!((*opps)->flags & INET6_PROTO_GSO_EXTHDR))
+				break;
+		}
+		if (opth == NULL)
+			opth = (void *)(iph+1);
+		else
+			opth = (void *)opth + optlen;
+		optlen = ipv6_optlen(opth);
+		len += optlen;
+		proto = opth->nexthdr;
+	}
+	return len;
+}
+
 static struct sk_buff **ipv6_gro_receive(struct sk_buff **head,
 					 struct sk_buff *skb)
 {
@@ -177,6 +206,7 @@ static struct sk_buff **ipv6_gro_receive(struct sk_buff **head,
 			goto out;
 	}
 
+	skb_set_network_header(skb, off);
 	skb_gro_pull(skb, sizeof(*iph));
 	skb_set_transport_header(skb, skb_gro_offset(skb));
 
@@ -211,12 +241,16 @@ static struct sk_buff **ipv6_gro_receive(struct sk_buff **head,
 		if (!NAPI_GRO_CB(p)->same_flow)
 			continue;
 
-		iph2 = ipv6_hdr(p);
+		iph2 = (struct ipv6hdr *)(p->data + off);
 		first_word = *(__be32 *)iph ^ *(__be32 *)iph2 ;
 
-		/* All fields must match except length and Traffic Class. */
-		if (nlen != skb_network_header_len(p) ||
-		    (first_word & htonl(0xF00FFFFF)) ||
+		/* All fields must match except length and Traffic Class.
+		 * XXX skbs on the gro_list have all been parsed and pulled
+		 * already so we don't need to compare nlen
+		 * (nlen != (sizeof(*iph2) + ipv6_exthdrs_len(iph2, &ops)))
+		 * memcmp() alone below is suffcient, right?
+		 */
+		 if ((first_word & htonl(0xF00FFFFF)) ||
 		    memcmp(&iph->nexthdr, &iph2->nexthdr,
 			   nlen - offsetof(struct ipv6hdr, nexthdr))) {
 			NAPI_GRO_CB(p)->same_flow = 0;
@@ -245,21 +279,21 @@ out:
 	return pp;
 }
 
-static int ipv6_gro_complete(struct sk_buff *skb)
+static int ipv6_gro_complete(struct sk_buff *skb, int nhoff)
 {
 	const struct net_offload *ops;
-	struct ipv6hdr *iph = ipv6_hdr(skb);
+	struct ipv6hdr *iph = (struct ipv6hdr *)(skb->data + nhoff);
 	int err = -ENOSYS;
 
-	iph->payload_len = htons(skb->len - skb_network_offset(skb) -
-				 sizeof(*iph));
+	iph->payload_len = htons(skb->len - nhoff - sizeof(*iph));
 
 	rcu_read_lock();
-	ops = rcu_dereference(inet6_offloads[NAPI_GRO_CB(skb)->proto]);
+
+	nhoff += sizeof(*iph) + ipv6_exthdrs_len(iph, &ops);
 	if (WARN_ON(!ops || !ops->callbacks.gro_complete))
 		goto out_unlock;
 
-	err = ops->callbacks.gro_complete(skb);
+	err = ops->callbacks.gro_complete(skb, nhoff);
 
 out_unlock:
 	rcu_read_unlock();
diff --git a/net/ipv6/tcpv6_offload.c b/net/ipv6/tcpv6_offload.c
index 6d18157dc32c..0d78132ff18a 100644
--- a/net/ipv6/tcpv6_offload.c
+++ b/net/ipv6/tcpv6_offload.c
@@ -66,13 +66,13 @@ skip_csum:
 	return tcp_gro_receive(head, skb);
 }
 
-static int tcp6_gro_complete(struct sk_buff *skb)
+static int tcp6_gro_complete(struct sk_buff *skb, int thoff)
 {
 	const struct ipv6hdr *iph = ipv6_hdr(skb);
 	struct tcphdr *th = tcp_hdr(skb);
 
-	th->check = ~tcp_v6_check(skb->len - skb_transport_offset(skb),
-				  &iph->saddr, &iph->daddr, 0);
+	th->check = ~tcp_v6_check(skb->len - thoff, &iph->saddr,
+				  &iph->daddr, 0);
 	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
 
 	return tcp_gro_complete(skb);