Merge branch 'gso_send_check'

Tom Herbert says:

====================
net: Eliminate gso_send_check

gso_send_check presents a lot of complexity for what it is being used
for. It seems that there are only two cases where it might be effective:
TCP and UFO paths. In these cases, the gso_send_check function
initializes the TCP or UDP checksum respectively to the pseudo header
checksum so that the checksum computation is appropriately offloaded or
computed in the gso_segment functions. The gso_send_check functions
are only called from dev.c in skb_mac_gso_segment when ip_summed !=
CHECKSUM_PARTIAL (which seems very unlikely in TCP case). We can move
the logic of this into the respective gso_segment functions where the
checksum is initialized if ip_summed != CHECKSUM_PARTIAL.

With the above cases handled, gso_send_check is no longer needed, so
we can remove all uses of it and the fields in the offload callbacks.
With this change, ip_summed in the skb should be preserved though all
the layers of gso_segment calls.

In follow-on patches, we may be able to remove the check setup code in
tcp_gso_segment if we can guarantee that ip_summed will always be
CHECKSUM_PARTIAL (verify all paths and probably add an assert in
tcp_gro_segment).

Tested these patches by:
  - netperf TCP_STREAM test with GSO enabled
  - Forced ip_summed != CHECKSUM_PARTIAL with above
  - Ran UDP_RR with 10000 request size over GRE tunnel. This exercised
    UFO path.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

10 files changed, 83 insertions, 186 deletions

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 4354b4307e37..9f5d293a0281 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1911,7 +1911,6 @@ struct packet_type {
 struct offload_callbacks {
 	struct sk_buff		*(*gso_segment)(struct sk_buff *skb,
 						netdev_features_t features);
-	int			(*gso_send_check)(struct sk_buff *skb);
 	struct sk_buff		**(*gro_receive)(struct sk_buff **head,
 					       struct sk_buff *skb);
 	int			(*gro_complete)(struct sk_buff *skb, int nhoff);
diff --git a/net/core/dev.c b/net/core/dev.c
index db0388607329..e2ced01c01e4 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2422,16 +2422,6 @@ struct sk_buff *skb_mac_gso_segment(struct sk_buff *skb,
 	rcu_read_lock();
 	list_for_each_entry_rcu(ptype, &offload_base, list) {
 		if (ptype->type == type && ptype->callbacks.gso_segment) {
-			if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
-				int err;
-
-				err = ptype->callbacks.gso_send_check(skb);
-				segs = ERR_PTR(err);
-				if (err || skb_gso_ok(skb, features))
-					break;
-				__skb_push(skb, (skb->data -
-						 skb_network_header(skb)));
-			}
 			segs = ptype->callbacks.gso_segment(skb, features);
 			break;
 		}
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 72011cc4c13b..28e589c5f32d 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1197,40 +1197,6 @@ int inet_sk_rebuild_header(struct sock *sk)
 }
 EXPORT_SYMBOL(inet_sk_rebuild_header);
 
-static int inet_gso_send_check(struct sk_buff *skb)
-{
-	const struct net_offload *ops;
-	const struct iphdr *iph;
-	int proto;
-	int ihl;
-	int err = -EINVAL;
-
-	if (unlikely(!pskb_may_pull(skb, sizeof(*iph))))
-		goto out;
-
-	iph = ip_hdr(skb);
-	ihl = iph->ihl * 4;
-	if (ihl < sizeof(*iph))
-		goto out;
-
-	proto = iph->protocol;
-
-	/* Warning: after this point, iph might be no longer valid */
-	if (unlikely(!pskb_may_pull(skb, ihl)))
-		goto out;
-	__skb_pull(skb, ihl);
-
-	skb_reset_transport_header(skb);
-	err = -EPROTONOSUPPORT;
-
-	ops = rcu_dereference(inet_offloads[proto]);
-	if (likely(ops && ops->callbacks.gso_send_check))
-		err = ops->callbacks.gso_send_check(skb);
-
-out:
-	return err;
-}
-
 static struct sk_buff *inet_gso_segment(struct sk_buff *skb,
 					netdev_features_t features)
 {
@@ -1655,7 +1621,6 @@ static int ipv4_proc_init(void);
 static struct packet_offload ip_packet_offload __read_mostly = {
 	.type = cpu_to_be16(ETH_P_IP),
 	.callbacks = {
-		.gso_send_check = inet_gso_send_check,
 		.gso_segment = inet_gso_segment,
 		.gro_receive = inet_gro_receive,
 		.gro_complete = inet_gro_complete,
@@ -1664,7 +1629,6 @@ static struct packet_offload ip_packet_offload __read_mostly = {
 
 static const struct net_offload ipip_offload = {
 	.callbacks = {
-		.gso_send_check = inet_gso_send_check,
 		.gso_segment	= inet_gso_segment,
 		.gro_receive	= inet_gro_receive,
 		.gro_complete	= inet_gro_complete,
diff --git a/net/ipv4/gre_offload.c b/net/ipv4/gre_offload.c
index d3fe2ac05167..a77729503071 100644
--- a/net/ipv4/gre_offload.c
+++ b/net/ipv4/gre_offload.c
@@ -15,13 +15,6 @@
 #include <net/protocol.h>
 #include <net/gre.h>
 
-static int gre_gso_send_check(struct sk_buff *skb)
-{
-	if (!skb->encapsulation)
-		return -EINVAL;
-	return 0;
-}
-
 static struct sk_buff *gre_gso_segment(struct sk_buff *skb,
 				       netdev_features_t features)
 {
@@ -46,6 +39,9 @@ static struct sk_buff *gre_gso_segment(struct sk_buff *skb,
 				  SKB_GSO_IPIP)))
 		goto out;
 
+	if (!skb->encapsulation)
+		goto out;
+
 	if (unlikely(!pskb_may_pull(skb, sizeof(*greh))))
 		goto out;
 
@@ -256,7 +252,6 @@ static int gre_gro_complete(struct sk_buff *skb, int nhoff)
 
 static const struct net_offload gre_offload = {
 	.callbacks = {
-		.gso_send_check = gre_gso_send_check,
 		.gso_segment = gre_gso_segment,
 		.gro_receive = gre_gro_receive,
 		.gro_complete = gre_gro_complete,
diff --git a/net/ipv4/tcp_offload.c b/net/ipv4/tcp_offload.c
index 72912533a191..5b90f2f447a5 100644
--- a/net/ipv4/tcp_offload.c
+++ b/net/ipv4/tcp_offload.c
@@ -29,6 +29,28 @@ static void tcp_gso_tstamp(struct sk_buff *skb, unsigned int ts_seq,
 	}
 }
 
+struct sk_buff *tcp4_gso_segment(struct sk_buff *skb,
+				 netdev_features_t features)
+{
+	if (!pskb_may_pull(skb, sizeof(struct tcphdr)))
+		return ERR_PTR(-EINVAL);
+
+	if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
+		const struct iphdr *iph = ip_hdr(skb);
+		struct tcphdr *th = tcp_hdr(skb);
+
+		/* Set up checksum pseudo header, usually expect stack to
+		 * have done this already.
+		 */
+
+		th->check = 0;
+		skb->ip_summed = CHECKSUM_PARTIAL;
+		__tcp_v4_send_check(skb, iph->saddr, iph->daddr);
+	}
+
+	return tcp_gso_segment(skb, features);
+}
+
 struct sk_buff *tcp_gso_segment(struct sk_buff *skb,
 				netdev_features_t features)
 {
@@ -44,9 +66,6 @@ struct sk_buff *tcp_gso_segment(struct sk_buff *skb,
 	__sum16 newcheck;
 	bool ooo_okay, copy_destructor;
 
-	if (!pskb_may_pull(skb, sizeof(*th)))
-		goto out;
-
 	th = tcp_hdr(skb);
 	thlen = th->doff * 4;
 	if (thlen < sizeof(*th))
@@ -269,23 +288,6 @@ int tcp_gro_complete(struct sk_buff *skb)
 }
 EXPORT_SYMBOL(tcp_gro_complete);
 
-static int tcp_v4_gso_send_check(struct sk_buff *skb)
-{
-	const struct iphdr *iph;
-	struct tcphdr *th;
-
-	if (!pskb_may_pull(skb, sizeof(*th)))
-		return -EINVAL;
-
-	iph = ip_hdr(skb);
-	th = tcp_hdr(skb);
-
-	th->check = 0;
-	skb->ip_summed = CHECKSUM_PARTIAL;
-	__tcp_v4_send_check(skb, iph->saddr, iph->daddr);
-	return 0;
-}
-
 static struct sk_buff **tcp4_gro_receive(struct sk_buff **head, struct sk_buff *skb)
 {
 	/* Don't bother verifying checksum if we're going to flush anyway. */
@@ -313,8 +315,7 @@ static int tcp4_gro_complete(struct sk_buff *skb, int thoff)
 
 static const struct net_offload tcpv4_offload = {
 	.callbacks = {
-		.gso_send_check	=	tcp_v4_gso_send_check,
-		.gso_segment	=	tcp_gso_segment,
+		.gso_segment	=	tcp4_gso_segment,
 		.gro_receive	=	tcp4_gro_receive,
 		.gro_complete	=	tcp4_gro_complete,
 	},
diff --git a/net/ipv4/udp_offload.c b/net/ipv4/udp_offload.c
index d7c43f764c71..19ebe6a39ddc 100644
--- a/net/ipv4/udp_offload.c
+++ b/net/ipv4/udp_offload.c
@@ -25,28 +25,6 @@ struct udp_offload_priv {
 	struct udp_offload_priv __rcu *next;
 };
 
-static int udp4_ufo_send_check(struct sk_buff *skb)
-{
-	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
-		return -EINVAL;
-
-	if (likely(!skb->encapsulation)) {
-		const struct iphdr *iph;
-		struct udphdr *uh;
-
-		iph = ip_hdr(skb);
-		uh = udp_hdr(skb);
-
-		uh->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, skb->len,
-				IPPROTO_UDP, 0);
-		skb->csum_start = skb_transport_header(skb) - skb->head;
-		skb->csum_offset = offsetof(struct udphdr, check);
-		skb->ip_summed = CHECKSUM_PARTIAL;
-	}
-
-	return 0;
-}
-
 struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
 				       netdev_features_t features)
 {
@@ -128,8 +106,9 @@ static struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
 {
 	struct sk_buff *segs = ERR_PTR(-EINVAL);
 	unsigned int mss;
-	int offset;
 	__wsum csum;
+	struct udphdr *uh;
+	struct iphdr *iph;
 
 	if (skb->encapsulation &&
 	    (skb_shinfo(skb)->gso_type &
@@ -138,6 +117,9 @@ static struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
 		goto out;
 	}
 
+	if (!pskb_may_pull(skb, sizeof(struct udphdr)))
+		goto out;
+
 	mss = skb_shinfo(skb)->gso_size;
 	if (unlikely(skb->len <= mss))
 		goto out;
@@ -165,10 +147,16 @@ static struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb,
 	 * HW cannot do checksum of UDP packets sent as multiple
 	 * IP fragments.
 	 */
-	offset = skb_checksum_start_offset(skb);
-	csum = skb_checksum(skb, offset, skb->len - offset, 0);
-	offset += skb->csum_offset;
-	*(__sum16 *)(skb->data + offset) = csum_fold(csum);
+
+	uh = udp_hdr(skb);
+	iph = ip_hdr(skb);
+
+	uh->check = 0;
+	csum = skb_checksum(skb, 0, skb->len, 0);
+	uh->check = udp_v4_check(skb->len, iph->saddr, iph->daddr, csum);
+	if (uh->check == 0)
+		uh->check = CSUM_MANGLED_0;
+
 	skb->ip_summed = CHECKSUM_NONE;
 
 	/* Fragment the skb. IP headers of the fragments are updated in
@@ -353,7 +341,6 @@ static int udp4_gro_complete(struct sk_buff *skb, int nhoff)
 
 static const struct net_offload udpv4_offload = {
 	.callbacks = {
-		.gso_send_check = udp4_ufo_send_check,
 		.gso_segment = udp4_ufo_fragment,
 		.gro_receive  =	udp4_gro_receive,
 		.gro_complete =	udp4_gro_complete,
diff --git a/net/ipv6/ip6_offload.c b/net/ipv6/ip6_offload.c
index 9952f3fce30a..9034f76ae013 100644
--- a/net/ipv6/ip6_offload.c
+++ b/net/ipv6/ip6_offload.c
@@ -53,31 +53,6 @@ static int ipv6_gso_pull_exthdrs(struct sk_buff *skb, int proto)
 	return proto;
 }
 
-static int ipv6_gso_send_check(struct sk_buff *skb)
-{
-	const struct ipv6hdr *ipv6h;
-	const struct net_offload *ops;
-	int err = -EINVAL;
-
-	if (unlikely(!pskb_may_pull(skb, sizeof(*ipv6h))))
-		goto out;
-
-	ipv6h = ipv6_hdr(skb);
-	__skb_pull(skb, sizeof(*ipv6h));
-	err = -EPROTONOSUPPORT;
-
-	ops = rcu_dereference(inet6_offloads[
-		ipv6_gso_pull_exthdrs(skb, ipv6h->nexthdr)]);
-
-	if (likely(ops && ops->callbacks.gso_send_check)) {
-		skb_reset_transport_header(skb);
-		err = ops->callbacks.gso_send_check(skb);
-	}
-
-out:
-	return err;
-}
-
 static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,
 	netdev_features_t features)
 {
@@ -306,7 +281,6 @@ out_unlock:
 static struct packet_offload ipv6_packet_offload __read_mostly = {
 	.type = cpu_to_be16(ETH_P_IPV6),
 	.callbacks = {
-		.gso_send_check = ipv6_gso_send_check,
 		.gso_segment = ipv6_gso_segment,
 		.gro_receive = ipv6_gro_receive,
 		.gro_complete = ipv6_gro_complete,
@@ -315,7 +289,6 @@ static struct packet_offload ipv6_packet_offload __read_mostly = {
 
 static const struct net_offload sit_offload = {
 	.callbacks = {
-		.gso_send_check = ipv6_gso_send_check,
 		.gso_segment	= ipv6_gso_segment,
 		.gro_receive	= ipv6_gro_receive,
 		.gro_complete	= ipv6_gro_complete,
diff --git a/net/ipv6/tcpv6_offload.c b/net/ipv6/tcpv6_offload.c
index dbb3d9262bf6..c1ab77105b4c 100644
--- a/net/ipv6/tcpv6_offload.c
+++ b/net/ipv6/tcpv6_offload.c
@@ -15,23 +15,6 @@
 #include <net/ip6_checksum.h>
 #include "ip6_offload.h"
 
-static int tcp_v6_gso_send_check(struct sk_buff *skb)
-{
-	const struct ipv6hdr *ipv6h;
-	struct tcphdr *th;
-
-	if (!pskb_may_pull(skb, sizeof(*th)))
-		return -EINVAL;
-
-	ipv6h = ipv6_hdr(skb);
-	th = tcp_hdr(skb);
-
-	th->check = 0;
-	skb->ip_summed = CHECKSUM_PARTIAL;
-	__tcp_v6_send_check(skb, &ipv6h->saddr, &ipv6h->daddr);
-	return 0;
-}
-
 static struct sk_buff **tcp6_gro_receive(struct sk_buff **head,
 					 struct sk_buff *skb)
 {
@@ -58,10 +41,32 @@ static int tcp6_gro_complete(struct sk_buff *skb, int thoff)
 	return tcp_gro_complete(skb);
 }
 
+struct sk_buff *tcp6_gso_segment(struct sk_buff *skb,
+				 netdev_features_t features)
+{
+	struct tcphdr *th;
+
+	if (!pskb_may_pull(skb, sizeof(*th)))
+		return ERR_PTR(-EINVAL);
+
+	if (unlikely(skb->ip_summed != CHECKSUM_PARTIAL)) {
+		const struct ipv6hdr *ipv6h = ipv6_hdr(skb);
+		struct tcphdr *th = tcp_hdr(skb);
+
+		/* Set up pseudo header, usually expect stack to have done
+		 * this.
+		 */
+
+		th->check = 0;
+		skb->ip_summed = CHECKSUM_PARTIAL;
+		__tcp_v6_send_check(skb, &ipv6h->saddr, &ipv6h->daddr);
+	}
+
+	return tcp_gso_segment(skb, features);
+}
 static const struct net_offload tcpv6_offload = {
 	.callbacks = {
-		.gso_send_check	=	tcp_v6_gso_send_check,
-		.gso_segment	=	tcp_gso_segment,
+		.gso_segment	=	tcp6_gso_segment,
 		.gro_receive	=	tcp6_gro_receive,
 		.gro_complete	=	tcp6_gro_complete,
 	},
diff --git a/net/ipv6/udp_offload.c b/net/ipv6/udp_offload.c
index de85f809bf29..212ebfc7973f 100644
--- a/net/ipv6/udp_offload.c
+++ b/net/ipv6/udp_offload.c
@@ -17,28 +17,6 @@
 #include <net/ip6_checksum.h>
 #include "ip6_offload.h"
 
-static int udp6_ufo_send_check(struct sk_buff *skb)
-{
-	const struct ipv6hdr *ipv6h;
-	struct udphdr *uh;
-
-	if (!pskb_may_pull(skb, sizeof(*uh)))
-		return -EINVAL;
-
-	if (likely(!skb->encapsulation)) {
-		ipv6h = ipv6_hdr(skb);
-		uh = udp_hdr(skb);
-
-		uh->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr, skb->len,
-					     IPPROTO_UDP, 0);
-		skb->csum_start = skb_transport_header(skb) - skb->head;
-		skb->csum_offset = offsetof(struct udphdr, check);
-		skb->ip_summed = CHECKSUM_PARTIAL;
-	}
-
-	return 0;
-}
-
 static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb,
 					 netdev_features_t features)
 {
@@ -49,7 +27,6 @@ static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb,
 	u8 *packet_start, *prevhdr;
 	u8 nexthdr;
 	u8 frag_hdr_sz = sizeof(struct frag_hdr);
-	int offset;
 	__wsum csum;
 	int tnl_hlen;
 
@@ -83,13 +60,27 @@ static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb,
 	    (SKB_GSO_UDP_TUNNEL|SKB_GSO_UDP_TUNNEL_CSUM))
 		segs = skb_udp_tunnel_segment(skb, features);
 	else {
+		const struct ipv6hdr *ipv6h;
+		struct udphdr *uh;
+
+		if (!pskb_may_pull(skb, sizeof(struct udphdr)))
+			goto out;
+
 		/* Do software UFO. Complete and fill in the UDP checksum as HW cannot
 		 * do checksum of UDP packets sent as multiple IP fragments.
 		 */
-		offset = skb_checksum_start_offset(skb);
-		csum = skb_checksum(skb, offset, skb->len - offset, 0);
-		offset += skb->csum_offset;
-		*(__sum16 *)(skb->data + offset) = csum_fold(csum);
+
+		uh = udp_hdr(skb);
+		ipv6h = ipv6_hdr(skb);
+
+		uh->check = 0;
+		csum = skb_checksum(skb, 0, skb->len, 0);
+		uh->check = udp_v6_check(skb->len, &ipv6h->saddr,
+					  &ipv6h->daddr, csum);
+
+		if (uh->check == 0)
+			uh->check = CSUM_MANGLED_0;
+
 		skb->ip_summed = CHECKSUM_NONE;
 
 		/* Check if there is enough headroom to insert fragment header. */
@@ -170,7 +161,6 @@ static int udp6_gro_complete(struct sk_buff *skb, int nhoff)
 
 static const struct net_offload udpv6_offload = {
 	.callbacks = {
-		.gso_send_check =	udp6_ufo_send_check,
 		.gso_segment	=	udp6_ufo_fragment,
 		.gro_receive	=	udp6_gro_receive,
 		.gro_complete	=	udp6_gro_complete,
diff --git a/net/mpls/mpls_gso.c b/net/mpls/mpls_gso.c
index 6b38d083e1c9..e28ed2ef5b06 100644
--- a/net/mpls/mpls_gso.c
+++ b/net/mpls/mpls_gso.c
@@ -65,15 +65,9 @@ out:
 	return segs;
 }
 
-static int mpls_gso_send_check(struct sk_buff *skb)
-{
-	return 0;
-}
-
 static struct packet_offload mpls_mc_offload = {
 	.type = cpu_to_be16(ETH_P_MPLS_MC),
 	.callbacks = {
-		.gso_send_check =	mpls_gso_send_check,
 		.gso_segment    =	mpls_gso_segment,
 	},
 };
@@ -81,7 +75,6 @@ static struct packet_offload mpls_mc_offload = {
 static struct packet_offload mpls_uc_offload = {
 	.type = cpu_to_be16(ETH_P_MPLS_UC),
 	.callbacks = {
-		.gso_send_check =	mpls_gso_send_check,
 		.gso_segment    =	mpls_gso_segment,
 	},
 };