Merge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6

Documentation/networking/dccp.txt

@@ -0,0 +1,56 @@
+DCCP protocol
+============
+
+Last updated: 10 November 2005
+
+Contents
+========
+
+- Introduction
+- Missing features
+- Socket options
+- Notes
+
+Introduction
+============
+
+Datagram Congestion Control Protocol (DCCP) is an unreliable, connection
+based protocol designed to solve issues present in UDP and TCP particularly
+for real time and multimedia traffic.
+
+It has a base protocol and pluggable congestion control IDs (CCIDs).
+
+It is at draft RFC status and the homepage for DCCP as a protocol is at:
+	http://www.icir.org/kohler/dcp/
+
+Missing features
+================
+
+The DCCP implementation does not currently have all the features that are in
+the draft RFC.
+
+In particular the following are missing:
+- CCID2 support
+- feature negotiation
+
+When testing against other implementations it appears that elapsed time
+options are not coded compliant to the specification.
+
+Socket options
+==============
+
+DCCP_SOCKOPT_PACKET_SIZE is used for CCID3 to set default packet size for
+calculations.
+
+DCCP_SOCKOPT_SERVICE sets the service. This is compulsory as per the
+specification. If you don't set it you will get EPROTO.
+
+Notes
+=====
+
+SELinux does not yet have support for DCCP. You will need to turn it off or
+else you will get EACCES.
+
+DCCP does not travel through NAT successfully at present. This is because
+the checksum covers the psuedo-header as per TCP and UDP. It should be
+relatively trivial to add Linux NAT support for DCCP.

MAINTAINERS

@@ -707,7 +707,7 @@ DCCP PROTOCOL
 P:	Arnaldo Carvalho de Melo
 M:	acme@mandriva.com
 L:	dccp@vger.kernel.org
-W:	http://www.wlug.org.nz/DCCP
+W:	http://linux-net.osdl.org/index.php/DCCP
 S:	Maintained
 
 DECnet NETWORK LAYER

drivers/atm/horizon.c

@@ -1511,8 +1511,8 @@ static inline short setup_idle_tx_channel (hrz_dev * dev, hrz_vcc * vcc) {
     // a.k.a. prepare the channel and remember that we have done so.
 
     tx_ch_desc * tx_desc = &memmap->tx_descs[tx_channel];
-    u16 rd_ptr;
-    u16 wr_ptr;
+    u32 rd_ptr;
+    u32 wr_ptr;
     u16 channel = vcc->channel;
 
     unsigned long flags;

drivers/net/bnx2.c

@@ -2707,7 +2707,7 @@ bnx2_init_nvram(struct bnx2 *bp)
 
 	if (j == entry_count) {
 		bp->flash_info = NULL;
-		printk(KERN_ALERT "Unknown flash/EEPROM type.\n");
+		printk(KERN_ALERT PFX "Unknown flash/EEPROM type.\n");
 		rc = -ENODEV;
 	}
 
@@ -3903,6 +3903,8 @@ bnx2_test_loopback(struct bnx2 *bp)
 
 	pkt_size = 1514;
 	skb = dev_alloc_skb(pkt_size);
+	if (!skb)
+		return -ENOMEM;
 	packet = skb_put(skb, pkt_size);
 	memcpy(packet, bp->mac_addr, 6);
 	memset(packet + 6, 0x0, 8);
@@ -4798,11 +4800,7 @@ bnx2_get_eeprom(struct net_device *dev, struct ethtool_eeprom *eeprom,
   	struct bnx2 *bp = dev->priv;
 	int rc;
 
-	if (eeprom->offset > bp->flash_info->total_size)
-		return -EINVAL;
-
-	if ((eeprom->offset + eeprom->len) > bp->flash_info->total_size)
-		eeprom->len = bp->flash_info->total_size - eeprom->offset;
+	/* parameters already validated in ethtool_get_eeprom */
 
 	rc = bnx2_nvram_read(bp, eeprom->offset, eebuf, eeprom->len);
 
@@ -4816,11 +4814,7 @@ bnx2_set_eeprom(struct net_device *dev, struct ethtool_eeprom *eeprom,
   	struct bnx2 *bp = dev->priv;
 	int rc;
 
-	if (eeprom->offset > bp->flash_info->total_size)
-		return -EINVAL;
-
-	if ((eeprom->offset + eeprom->len) > bp->flash_info->total_size)
-		eeprom->len = bp->flash_info->total_size - eeprom->offset;
+	/* parameters already validated in ethtool_set_eeprom */
 
 	rc = bnx2_nvram_write(bp, eeprom->offset, eebuf, eeprom->len);
 

include/linux/if_ether.h

@@ -21,6 +21,8 @@
 #ifndef _LINUX_IF_ETHER_H
 #define _LINUX_IF_ETHER_H
 
+#include <linux/types.h>
+
 /*
  *	IEEE 802.3 Ethernet magic constants.  The frame sizes omit the preamble
  *	and FCS/CRC (frame check sequence). 
@@ -100,7 +102,7 @@
 struct ethhdr {
 	unsigned char	h_dest[ETH_ALEN];	/* destination eth addr	*/
 	unsigned char	h_source[ETH_ALEN];	/* source ether addr	*/
-	unsigned short	h_proto;		/* packet type ID field	*/
+	__be16		h_proto;		/* packet type ID field	*/
 } __attribute__((packed));
 
 #ifdef __KERNEL__

include/linux/netdevice.h

@@ -927,6 +927,13 @@ extern int		netdev_max_backlog;
 extern int		weight_p;
 extern int		netdev_set_master(struct net_device *dev, struct net_device *master);
 extern int skb_checksum_help(struct sk_buff *skb, int inward);
+#ifdef CONFIG_BUG
+extern void netdev_rx_csum_fault(struct net_device *dev);
+#else
+static inline void netdev_rx_csum_fault(struct net_device *dev)
+{
+}
+#endif
 /* rx skb timestamps */
 extern void		net_enable_timestamp(void);
 extern void		net_disable_timestamp(void);

include/linux/skbuff.h

@@ -1236,8 +1236,7 @@ extern unsigned int    datagram_poll(struct file *file, struct socket *sock,
 extern int	       skb_copy_datagram_iovec(const struct sk_buff *from,
 					       int offset, struct iovec *to,
 					       int size);
-extern int	       skb_copy_and_csum_datagram_iovec(const
-							struct sk_buff *skb,
+extern int	       skb_copy_and_csum_datagram_iovec(struct sk_buff *skb,
 							int hlen,
 							struct iovec *iov);
 extern void	       skb_free_datagram(struct sock *sk, struct sk_buff *skb);
@@ -1305,6 +1304,30 @@ static inline void skb_set_timestamp(struct sk_buff *skb, const struct timeval *
 
 extern void __net_timestamp(struct sk_buff *skb);
 
+extern unsigned int __skb_checksum_complete(struct sk_buff *skb);
+
+/**
+ *	skb_checksum_complete - Calculate checksum of an entire packet
+ *	@skb: packet to process
+ *
+ *	This function calculates the checksum over the entire packet plus
+ *	the value of skb->csum.  The latter can be used to supply the
+ *	checksum of a pseudo header as used by TCP/UDP.  It returns the
+ *	checksum.
+ *
+ *	For protocols that contain complete checksums such as ICMP/TCP/UDP,
+ *	this function can be used to verify that checksum on received
+ *	packets.  In that case the function should return zero if the
+ *	checksum is correct.  In particular, this function will return zero
+ *	if skb->ip_summed is CHECKSUM_UNNECESSARY which indicates that the
+ *	hardware has already verified the correctness of the checksum.
+ */
+static inline unsigned int skb_checksum_complete(struct sk_buff *skb)
+{
+	return skb->ip_summed != CHECKSUM_UNNECESSARY &&
+		__skb_checksum_complete(skb);
+}
+
 #ifdef CONFIG_NETFILTER
 static inline void nf_conntrack_put(struct nf_conntrack *nfct)
 {

include/net/tcp.h

@@ -27,6 +27,7 @@
 #include <linux/slab.h>
 #include <linux/cache.h>
 #include <linux/percpu.h>
+#include <linux/skbuff.h>
 
 #include <net/inet_connection_sock.h>
 #include <net/inet_timewait_sock.h>
@@ -852,7 +853,7 @@ static __inline__ u16 tcp_v4_check(struct tcphdr *th, int len,
 
 static __inline__ int __tcp_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 tcp_checksum_complete(struct sk_buff *skb)

net/core/datagram.c

@@ -350,6 +350,20 @@ static int skb_copy_and_csum_datagram(const struct sk_buff *skb, int offset,
 	return -EFAULT;
 }
 
+unsigned int __skb_checksum_complete(struct sk_buff *skb)
+{
+	unsigned int sum;
+
+	sum = (u16)csum_fold(skb_checksum(skb, 0, skb->len, skb->csum));
+	if (likely(!sum)) {
+		if (unlikely(skb->ip_summed == CHECKSUM_HW))
+			netdev_rx_csum_fault(skb->dev);
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+	}
+	return sum;
+}
+EXPORT_SYMBOL(__skb_checksum_complete);
+
 /**
  *	skb_copy_and_csum_datagram_iovec - Copy and checkum skb to user iovec.
  *	@skb: skbuff
@@ -363,7 +377,7 @@ static int skb_copy_and_csum_datagram(const struct sk_buff *skb, int offset,
  *		 -EFAULT - fault during copy. Beware, in this case iovec
  *			   can be modified!
  */
-int skb_copy_and_csum_datagram_iovec(const struct sk_buff *skb,
+int skb_copy_and_csum_datagram_iovec(struct sk_buff *skb,
 				     int hlen, struct iovec *iov)
 {
 	unsigned int csum;
@@ -376,8 +390,7 @@ int skb_copy_and_csum_datagram_iovec(const struct sk_buff *skb,
 		iov++;
 
 	if (iov->iov_len < chunk) {
-		if ((unsigned short)csum_fold(skb_checksum(skb, 0, chunk + hlen,
-							   skb->csum)))
+		if (__skb_checksum_complete(skb))
 			goto csum_error;
 		if (skb_copy_datagram_iovec(skb, hlen, iov, chunk))
 			goto fault;
@@ -388,6 +401,8 @@ int skb_copy_and_csum_datagram_iovec(const struct sk_buff *skb,
 			goto fault;
 		if ((unsigned short)csum_fold(csum))
 			goto csum_error;
+		if (unlikely(skb->ip_summed == CHECKSUM_HW))
+			netdev_rx_csum_fault(skb->dev);
 		iov->iov_len -= chunk;
 		iov->iov_base += chunk;
 	}

net/core/dev.c

@@ -1108,6 +1108,18 @@ int skb_checksum_help(struct sk_buff *skb, int inward)
 	return ret;
 }
 
+/* Take action when hardware reception checksum errors are detected. */
+#ifdef CONFIG_BUG
+void netdev_rx_csum_fault(struct net_device *dev)
+{
+	if (net_ratelimit()) {
+		printk(KERN_ERR "%s: hw csum failure.\n", dev->name);
+		dump_stack();
+	}
+}
+EXPORT_SYMBOL(netdev_rx_csum_fault);
+#endif
+
 #ifdef CONFIG_HIGHMEM
 /* Actually, we should eliminate this check as soon as we know, that:
  * 1. IOMMU is present and allows to map all the memory.

net/core/netpoll.c

@@ -101,16 +101,20 @@ void netpoll_queue(struct sk_buff *skb)
 static int checksum_udp(struct sk_buff *skb, struct udphdr *uh,
 			     unsigned short ulen, u32 saddr, u32 daddr)
 {
-	if (uh->check == 0)
+	unsigned int psum;
+
+	if (uh->check == 0 || skb->ip_summed == CHECKSUM_UNNECESSARY)
 		return 0;
 
-	if (skb->ip_summed == CHECKSUM_HW)
-		return csum_tcpudp_magic(
-			saddr, daddr, ulen, IPPROTO_UDP, skb->csum);
+	psum = csum_tcpudp_nofold(saddr, daddr, ulen, IPPROTO_UDP, 0);
+
+	if (skb->ip_summed == CHECKSUM_HW &&
+	    !(u16)csum_fold(csum_add(psum, skb->csum)))
+		return 0;
 
-	skb->csum = csum_tcpudp_nofold(saddr, daddr, ulen, IPPROTO_UDP, 0);
+	skb->csum = psum;
 
-	return csum_fold(skb_checksum(skb, 0, skb->len, skb->csum));
+	return __skb_checksum_complete(skb);
 }
 
 /*
@@ -489,7 +493,7 @@ int __netpoll_rx(struct sk_buff *skb)
 
 	if (ulen != len)
 		goto out;
-	if (checksum_udp(skb, uh, ulen, iph->saddr, iph->daddr) < 0)
+	if (checksum_udp(skb, uh, ulen, iph->saddr, iph->daddr))
 		goto out;
 	if (np->local_ip && np->local_ip != ntohl(iph->daddr))
 		goto out;

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)))

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;

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;
 		}