[NET]: Detect hardware rx checksum faults correctly

Here is the patch that introduces the generic skb_checksum_complete
which also checks for hardware RX checksum faults.  If that happens,
it'll call netdev_rx_csum_fault which currently prints out a stack
trace with the device name.  In future it can turn off RX checksum.

I've converted every spot under net/ that does RX checksum checks to
use skb_checksum_complete or __skb_checksum_complete with the
exceptions of:

* Those places where checksums are done bit by bit.  These will call
netdev_rx_csum_fault directly.

* The following have not been completely checked/converted:

ipmr
ip_vs
netfilter
dccp

This patch is based on patches and suggestions from Stephen Hemminger
and David S. Miller.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 40 insertions, 42 deletions

diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 175e093ec5645..e3eceecd0496e 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -934,11 +934,11 @@ int icmp_rcv(struct sk_buff *skb)
 	case CHECKSUM_HW:
 		if (!(u16)csum_fold(skb->csum))
 			break;
-		LIMIT_NETDEBUG(KERN_DEBUG "icmp v4 hw csum failure\n");
+		/* fall through */
 	case CHECKSUM_NONE:
-		if ((u16)csum_fold(skb_checksum(skb, 0, skb->len, 0)))
+		skb->csum = 0;
+		if (__skb_checksum_complete(skb))
 			goto error;
-	default:;
 	}
 
 	if (!pskb_pull(skb, sizeof(struct icmphdr)))
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index c6247fc84060a..c04607b49212e 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -872,11 +872,18 @@ int igmp_rcv(struct sk_buff *skb)
 		return 0;
 	}
 
-	if (!pskb_may_pull(skb, sizeof(struct igmphdr)) || 
-	    (u16)csum_fold(skb_checksum(skb, 0, len, 0))) {
-		in_dev_put(in_dev);
-		kfree_skb(skb);
-		return 0;
+	if (!pskb_may_pull(skb, sizeof(struct igmphdr)))
+		goto drop;
+
+	switch (skb->ip_summed) {
+	case CHECKSUM_HW:
+		if (!(u16)csum_fold(skb->csum))
+			break;
+		/* fall through */
+	case CHECKSUM_NONE:
+		skb->csum = 0;
+		if (__skb_checksum_complete(skb))
+			goto drop;
 	}
 
 	ih = skb->h.igmph;
@@ -906,6 +913,8 @@ int igmp_rcv(struct sk_buff *skb)
 	default:
 		NETDEBUG(KERN_DEBUG "New IGMP type=%d, why we do not know about it?\n", ih->type);
 	}
+
+drop:
 	in_dev_put(in_dev);
 	kfree_skb(skb);
 	return 0;
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 896ce3f8f53ad..4e9c74b54b150 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -577,15 +577,16 @@ static int ipgre_rcv(struct sk_buff *skb)
 			goto drop_nolock;
 
 		if (flags&GRE_CSUM) {
-			if (skb->ip_summed == CHECKSUM_HW) {
+			switch (skb->ip_summed) {
+			case CHECKSUM_HW:
 				csum = (u16)csum_fold(skb->csum);
-				if (csum)
-					skb->ip_summed = CHECKSUM_NONE;
-			}
-			if (skb->ip_summed == CHECKSUM_NONE) {
-				skb->csum = skb_checksum(skb, 0, skb->len, 0);
+				if (!csum)
+					break;
+				/* fall through */
+			case CHECKSUM_NONE:
+				skb->csum = 0;
+				csum = __skb_checksum_complete(skb);
 				skb->ip_summed = CHECKSUM_HW;
-				csum = (u16)csum_fold(skb->csum);
 			}
 			offset += 4;
 		}
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_icmp.c b/net/ipv4/netfilter/ip_conntrack_proto_icmp.c
index 5198f3a1e2cd5..e4d6b268e8c4e 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_icmp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_icmp.c
@@ -13,6 +13,7 @@
 #include <linux/in.h>
 #include <linux/icmp.h>
 #include <linux/seq_file.h>
+#include <linux/skbuff.h>
 #include <net/ip.h>
 #include <net/checksum.h>
 #include <linux/netfilter.h>
@@ -230,19 +231,15 @@ icmp_error(struct sk_buff *skb, enum ip_conntrack_info *ctinfo,
 	case CHECKSUM_HW:
 		if (!(u16)csum_fold(skb->csum)) 
 			break;
-		if (LOG_INVALID(IPPROTO_ICMP))
-			nf_log_packet(PF_INET, 0, skb, NULL, NULL, NULL,
-				      "ip_ct_icmp: bad HW ICMP checksum ");
-		return -NF_ACCEPT;
+		/* fall through */
 	case CHECKSUM_NONE:
-		if ((u16)csum_fold(skb_checksum(skb, 0, skb->len, 0))) {
+		skb->csum = 0;
+		if (__skb_checksum_complete(skb)) {
 			if (LOG_INVALID(IPPROTO_ICMP))
 				nf_log_packet(PF_INET, 0, skb, NULL, NULL, NULL,
 					      "ip_ct_icmp: bad ICMP checksum ");
 			return -NF_ACCEPT;
 		}
-	default:
-		break;
 	}
 
 checksum_skipped:
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 634dabb558fd6..ac1fcf5b4ebc2 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1110,24 +1110,18 @@ static struct sock *tcp_v4_hnd_req(struct sock *sk, struct sk_buff *skb)
 static int tcp_v4_checksum_init(struct sk_buff *skb)
 {
 	if (skb->ip_summed == CHECKSUM_HW) {
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
 		if (!tcp_v4_check(skb->h.th, skb->len, skb->nh.iph->saddr,
-				  skb->nh.iph->daddr, skb->csum))
+				  skb->nh.iph->daddr, skb->csum)) {
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
 			return 0;
-
-		LIMIT_NETDEBUG(KERN_DEBUG "hw tcp v4 csum failed\n");
-		skb->ip_summed = CHECKSUM_NONE;
+		}
 	}
+
+	skb->csum = csum_tcpudp_nofold(skb->nh.iph->saddr, skb->nh.iph->daddr,
+				       skb->len, IPPROTO_TCP, 0);
+
 	if (skb->len <= 76) {
-		if (tcp_v4_check(skb->h.th, skb->len, skb->nh.iph->saddr,
-				 skb->nh.iph->daddr,
-				 skb_checksum(skb, 0, skb->len, 0)))
-			return -1;
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
-	} else {
-		skb->csum = ~tcp_v4_check(skb->h.th, skb->len,
-					  skb->nh.iph->saddr,
-					  skb->nh.iph->daddr, 0);
+		return __skb_checksum_complete(skb);
 	}
 	return 0;
 }
@@ -1219,7 +1213,7 @@ int tcp_v4_rcv(struct sk_buff *skb)
 	 * provided case of th->doff==0 is elimineted.
 	 * So, we defer the checks. */
 	if ((skb->ip_summed != CHECKSUM_UNNECESSARY &&
-	     tcp_v4_checksum_init(skb) < 0))
+	     tcp_v4_checksum_init(skb)))
 		goto bad_packet;
 
 	th = skb->h.th;
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index e0bd1013cb0d0..2422a5f7195d0 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -761,7 +761,7 @@ int udp_ioctl(struct sock *sk, int cmd, unsigned long arg)
 
 static __inline__ int __udp_checksum_complete(struct sk_buff *skb)
 {
-	return (unsigned short)csum_fold(skb_checksum(skb, 0, skb->len, skb->csum));
+	return __skb_checksum_complete(skb);
 }
 
 static __inline__ int udp_checksum_complete(struct sk_buff *skb)
@@ -1100,11 +1100,8 @@ static int udp_checksum_init(struct sk_buff *skb, struct udphdr *uh,
 	if (uh->check == 0) {
 		skb->ip_summed = CHECKSUM_UNNECESSARY;
 	} else if (skb->ip_summed == CHECKSUM_HW) {
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
 		if (!udp_check(uh, ulen, saddr, daddr, skb->csum))
-			return 0;
-		LIMIT_NETDEBUG(KERN_DEBUG "udp v4 hw csum failure.\n");
-		skb->ip_summed = CHECKSUM_NONE;
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
 	}
 	if (skb->ip_summed != CHECKSUM_UNNECESSARY)
 		skb->csum = csum_tcpudp_nofold(saddr, daddr, ulen, IPPROTO_UDP, 0);