From owner-freebsd-net@FreeBSD.ORG  Wed Oct 10 19:58:45 2012
Return-Path: <owner-freebsd-net@FreeBSD.ORG>
Delivered-To: net@FreeBSD.org
Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52])
 by hub.freebsd.org (Postfix) with ESMTP id D312DF85
 for <net@FreeBSD.org>; Wed, 10 Oct 2012 19:58:45 +0000 (UTC)
 (envelope-from glebius@FreeBSD.org)
Received: from cell.glebius.int.ru (glebius.int.ru [81.19.64.117])
 by mx1.freebsd.org (Postfix) with ESMTP id A960A8FC0A
 for <net@FreeBSD.org>; Wed, 10 Oct 2012 19:58:44 +0000 (UTC)
Received: from cell.glebius.int.ru (localhost [127.0.0.1])
 by cell.glebius.int.ru (8.14.5/8.14.5) with ESMTP id q9AJwgWE087215
 for <net@FreeBSD.org>; Wed, 10 Oct 2012 23:58:42 +0400 (MSK)
 (envelope-from glebius@FreeBSD.org)
Received: (from glebius@localhost)
 by cell.glebius.int.ru (8.14.5/8.14.5/Submit) id q9AJwgOW087214
 for net@freebsd.org; Wed, 10 Oct 2012 23:58:42 +0400 (MSK)
 (envelope-from glebius@FreeBSD.org)
X-Authentication-Warning: cell.glebius.int.ru: glebius set sender to
 glebius@FreeBSD.org using -f
Date: Wed, 10 Oct 2012 23:58:42 +0400
From: Gleb Smirnoff <glebius@FreeBSD.org>
To: net@FreeBSD.org
Subject: Re: [CFT/Review] net byte order for AF_INET
Message-ID: <20121010195842.GH34622@FreeBSD.org>
References: <20121009154128.GU34622@FreeBSD.org>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="HnQK338I3UIa/qiP"
Content-Disposition: inline
In-Reply-To: <20121009154128.GU34622@FreeBSD.org>
User-Agent: Mutt/1.5.21 (2010-09-15)
X-BeenThere: freebsd-net@freebsd.org
X-Mailman-Version: 2.1.14
Precedence: list
List-Id: Networking and TCP/IP with FreeBSD <freebsd-net.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/options/freebsd-net>,
 <mailto:freebsd-net-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-net>
List-Post: <mailto:freebsd-net@freebsd.org>
List-Help: <mailto:freebsd-net-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-net>,
 <mailto:freebsd-net-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Wed, 10 Oct 2012 19:58:45 -0000


--HnQK338I3UIa/qiP
Content-Type: text/plain; charset=koi8-r
Content-Disposition: inline

On Tue, Oct 09, 2012 at 07:41:28PM +0400, Gleb Smirnoff wrote:
T>   this is a patch that switches entire IPv4 stack to network
T> byte order. That means, that at any layer any module should
T> expect IP header in network byte order. Any host byte order
T> values can be stored in local variables only and are never stored
T> into a packet itself.

Update:

  - fresh patch against r241405, should apply to r241405
  - fixed raw sockets` as Luigi and Max Dounin noticed, also fixed
    another issue with raw sockets, now work fine - tested with ospfd
  - fixed gre(4)
  - took into account Luigi's comments about using host byte order
    in any ==, <, > comparisons.

Decided not to touch ip_reass().

-- 
Totus tuus, Glebius.

--HnQK338I3UIa/qiP
Content-Type: text/x-diff; charset=koi8-r
Content-Disposition: attachment; filename="IPv4.net-byte-order.diff"

Index: sys/netinet/tcp_input.c
===================================================================
--- sys/netinet/tcp_input.c	(revision 241405)
+++ sys/netinet/tcp_input.c	(working copy)
@@ -645,21 +645,19 @@
 		ip = mtod(m, struct ip *);
 		ipov = (struct ipovly *)ip;
 		th = (struct tcphdr *)((caddr_t)ip + off0);
-		tlen = ip->ip_len;
+		tlen = ntohs(ip->ip_len);
 
 		if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
 			if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
 				th->th_sum = m->m_pkthdr.csum_data;
 			else
 				th->th_sum = in_pseudo(ip->ip_src.s_addr,
-						ip->ip_dst.s_addr,
-						htonl(m->m_pkthdr.csum_data +
-							ip->ip_len +
-							IPPROTO_TCP));
+				    ip->ip_dst.s_addr,
+				    htonl(m->m_pkthdr.csum_data + tlen +
+				    IPPROTO_TCP));
 			th->th_sum ^= 0xffff;
 #ifdef TCPDEBUG
-			ipov->ih_len = (u_short)tlen;
-			ipov->ih_len = htons(ipov->ih_len);
+			ipov->ih_len = ip->ip_len;
 #endif
 		} else {
 			/*
@@ -667,8 +665,7 @@
 			 */
 			len = sizeof (struct ip) + tlen;
 			bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
-			ipov->ih_len = (u_short)tlen;
-			ipov->ih_len = htons(ipov->ih_len);
+			ipov->ih_len = ip->ip_len;
 			th->th_sum = in_cksum(m, len);
 		}
 		if (th->th_sum) {
Index: sys/netinet/in.h
===================================================================
--- sys/netinet/in.h	(revision 241405)
+++ sys/netinet/in.h	(working copy)
@@ -741,33 +741,6 @@
 #define	satosin(sa)	((struct sockaddr_in *)(sa))
 #define	sintosa(sin)	((struct sockaddr *)(sin))
 #define	ifatoia(ifa)	((struct in_ifaddr *)(ifa))
-
-/*
- * Historically, BSD keeps ip_len and ip_off in host format
- * when doing layer 3 processing, and this often requires
- * to translate the format back and forth.
- * To make the process explicit, we define a couple of macros
- * that also take into account the fact that at some point
- * we may want to keep those fields always in net format.
- */
-
-#if (BYTE_ORDER == BIG_ENDIAN) || defined(HAVE_NET_IPLEN)
-#define SET_NET_IPLEN(p)	do {} while (0)
-#define SET_HOST_IPLEN(p)	do {} while (0)
-#else
-#define SET_NET_IPLEN(p)	do {		\
-	struct ip *h_ip = (p);			\
-	h_ip->ip_len = htons(h_ip->ip_len);	\
-	h_ip->ip_off = htons(h_ip->ip_off);	\
-	} while (0)
-
-#define SET_HOST_IPLEN(p)	do {		\
-	struct ip *h_ip = (p);			\
-	h_ip->ip_len = ntohs(h_ip->ip_len);	\
-	h_ip->ip_off = ntohs(h_ip->ip_off);	\
-	} while (0)
-#endif /* !HAVE_NET_IPLEN */
-
 #endif /* _KERNEL */
 
 /* INET6 stuff */
Index: sys/netinet/tcp_subr.c
===================================================================
--- sys/netinet/tcp_subr.c	(revision 241405)
+++ sys/netinet/tcp_subr.c	(working copy)
@@ -584,10 +584,10 @@
 #ifdef INET
 	{
 		tlen += sizeof (struct tcpiphdr);
-		ip->ip_len = tlen;
+		ip->ip_len = htons(tlen);
 		ip->ip_ttl = V_ip_defttl;
 		if (V_path_mtu_discovery)
-			ip->ip_off |= IP_DF;
+			ip->ip_off |= htons(IP_DF);
 	}
 #endif
 	m->m_len = tlen;
@@ -1398,12 +1398,11 @@
 					    /*
 					     * If no alternative MTU was
 					     * proposed, try the next smaller
-					     * one.  ip->ip_len has already
-					     * been swapped in icmp_input().
+					     * one.
 					     */
 					    if (!mtu)
-						mtu = ip_next_mtu(ip->ip_len,
-						 1);
+						mtu = ip_next_mtu(
+						 ntohs(ip->ip_len), 1);
 					    if (mtu < V_tcp_minmss
 						 + sizeof(struct tcpiphdr))
 						mtu = V_tcp_minmss
Index: sys/netinet/tcp_debug.c
===================================================================
--- sys/netinet/tcp_debug.c	(revision 241405)
+++ sys/netinet/tcp_debug.c	(working copy)
@@ -175,11 +175,10 @@
 #ifdef INET6
 		    isipv6 ? ntohs(((struct ip6_hdr *)ipgen)->ip6_plen) :
 #endif
-		    ((struct ip *)ipgen)->ip_len;
+		    ntohs(((struct ip *)ipgen)->ip_len);
 		if (act == TA_OUTPUT) {
 			seq = ntohl(seq);
 			ack = ntohl(ack);
-			len = ntohs((u_short)len);
 		}
 		if (act == TA_OUTPUT)
 			len -= sizeof (struct tcphdr);
Index: sys/netinet/tcp_syncache.c
===================================================================
--- sys/netinet/tcp_syncache.c	(revision 241405)
+++ sys/netinet/tcp_syncache.c	(working copy)
@@ -1395,7 +1395,7 @@
 		ip = mtod(m, struct ip *);
 		ip->ip_v = IPVERSION;
 		ip->ip_hl = sizeof(struct ip) >> 2;
-		ip->ip_len = tlen;
+		ip->ip_len = htons(tlen);
 		ip->ip_id = 0;
 		ip->ip_off = 0;
 		ip->ip_sum = 0;
@@ -1413,7 +1413,7 @@
 		 *	2) the SCF_UNREACH flag has been set
 		 */
 		if (V_path_mtu_discovery && ((sc->sc_flags & SCF_UNREACH) == 0))
-		       ip->ip_off |= IP_DF;
+		       ip->ip_off |= htons(IP_DF);
 
 		th = (struct tcphdr *)(ip + 1);
 	}
@@ -1473,7 +1473,7 @@
 			ip6->ip6_plen = htons(ntohs(ip6->ip6_plen) + optlen);
 		else
 #endif
-			ip->ip_len += optlen;
+			ip->ip_len = htons(ntohs(ip->ip_len) + optlen);
 	} else
 		optlen = 0;
 
Index: sys/netinet/sctp_input.c
===================================================================
--- sys/netinet/sctp_input.c	(revision 241405)
+++ sys/netinet/sctp_input.c	(working copy)
@@ -6038,7 +6038,7 @@
 	dst.sin_len = sizeof(struct sockaddr_in);
 	dst.sin_port = sh->dest_port;
 	dst.sin_addr = ip->ip_dst;
-	length = ip->ip_len + iphlen;
+	length = ntohs(ip->ip_len) + iphlen;
 	/* Validate mbuf chain length with IP payload length. */
 	if (SCTP_HEADER_LEN(m) != length) {
 		SCTPDBG(SCTP_DEBUG_INPUT1,
Index: sys/netinet/udp_usrreq.c
===================================================================
--- sys/netinet/udp_usrreq.c	(revision 241405)
+++ sys/netinet/udp_usrreq.c	(working copy)
@@ -338,7 +338,7 @@
 	struct udphdr *uh;
 	struct ifnet *ifp;
 	struct inpcb *inp;
-	int len;
+	uint16_t len, ip_len;
 	struct ip save_ip;
 	struct sockaddr_in udp_in;
 #ifdef IPFIREWALL_FORWARD
@@ -392,13 +392,13 @@
 	 * reflect UDP length, drop.
 	 */
 	len = ntohs((u_short)uh->uh_ulen);
-	if (ip->ip_len != len) {
-		if (len > ip->ip_len || len < sizeof(struct udphdr)) {
+	ip_len = ntohs(ip->ip_len);
+	if (ip_len != len) {
+		if (len > ip_len || len < sizeof(struct udphdr)) {
 			UDPSTAT_INC(udps_badlen);
 			goto badunlocked;
 		}
-		m_adj(m, len - ip->ip_len);
-		/* ip->ip_len = len; */
+		m_adj(m, len - ip_len);
 	}
 
 	/*
@@ -601,7 +601,7 @@
 		if (badport_bandlim(BANDLIM_ICMP_UNREACH) < 0)
 			goto badunlocked;
 		*ip = save_ip;
-		ip->ip_len += iphlen;
+		ip->ip_len = htons(ip_len + iphlen);
 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PORT, 0, 0);
 		return;
 	}
@@ -1206,7 +1206,7 @@
 		struct ip *ip;
 
 		ip = (struct ip *)&ui->ui_i;
-		ip->ip_off |= IP_DF;
+		ip->ip_off |= htons(IP_DF);
 	}
 
 	ipflags = 0;
@@ -1233,7 +1233,7 @@
 		m->m_pkthdr.csum_data = offsetof(struct udphdr, uh_sum);
 	} else
 		ui->ui_sum = 0;
-	((struct ip *)ui)->ip_len = sizeof (struct udpiphdr) + len;
+	((struct ip *)ui)->ip_len = htons(sizeof(struct udpiphdr) + len);
 	((struct ip *)ui)->ip_ttl = inp->inp_ip_ttl;	/* XXX */
 	((struct ip *)ui)->ip_tos = tos;		/* XXX */
 	UDPSTAT_INC(udps_opackets);
@@ -1383,7 +1383,7 @@
 	m_adj(m, skip);
 
 	ip = mtod(m, struct ip *);
-	ip->ip_len -= skip;
+	ip->ip_len = htons(ntohs(ip->ip_len) - skip);
 	ip->ip_p = IPPROTO_ESP;
 
 	/*
Index: sys/netinet/ip_carp.c
===================================================================
--- sys/netinet/ip_carp.c	(revision 241405)
+++ sys/netinet/ip_carp.c	(working copy)
@@ -783,9 +783,9 @@
 		ip->ip_v = IPVERSION;
 		ip->ip_hl = sizeof(*ip) >> 2;
 		ip->ip_tos = IPTOS_LOWDELAY;
-		ip->ip_len = len;
+		ip->ip_len = htons(len);
 		ip->ip_id = ip_newid();
-		ip->ip_off = IP_DF;
+		ip->ip_off = htons(IP_DF);
 		ip->ip_ttl = CARP_DFLTTL;
 		ip->ip_p = IPPROTO_CARP;
 		ip->ip_sum = 0;
Index: sys/netinet/raw_ip.c
===================================================================
--- sys/netinet/raw_ip.c	(revision 241405)
+++ sys/netinet/raw_ip.c	(working copy)
@@ -292,7 +292,7 @@
 	 * not modify the packet except for some
 	 * byte order swaps.
 	 */
-	ip->ip_len += off;
+	ip->ip_len = htons(ip->ip_len) + off;
 
 	hash = INP_PCBHASH_RAW(proto, ip->ip_src.s_addr,
 	    ip->ip_dst.s_addr, V_ripcbinfo.ipi_hashmask);
@@ -449,11 +449,11 @@
 		ip = mtod(m, struct ip *);
 		ip->ip_tos = inp->inp_ip_tos;
 		if (inp->inp_flags & INP_DONTFRAG)
-			ip->ip_off = IP_DF;
+			ip->ip_off = htons(IP_DF);
 		else
 			ip->ip_off = 0;
 		ip->ip_p = inp->inp_ip_p;
-		ip->ip_len = m->m_pkthdr.len;
+		ip->ip_len = htons(m->m_pkthdr.len);
 		ip->ip_src = inp->inp_laddr;
 		if (jailed(inp->inp_cred)) {
 			/*
@@ -504,6 +504,9 @@
 		if (ip->ip_id == 0)
 			ip->ip_id = ip_newid();
 
+		ip->ip_len = htons(ip->ip_len);
+		ip->ip_off = htons(ip->ip_off);
+
 		/*
 		 * XXX prevent ip_output from overwriting header fields.
 		 */
Index: sys/netinet/ip_divert.c
===================================================================
--- sys/netinet/ip_divert.c	(revision 241405)
+++ sys/netinet/ip_divert.c	(working copy)
@@ -386,10 +386,6 @@
 				INP_RUNLOCK(inp);
 				goto cantsend;
 			}
-
-			/* Convert fields to host order for ip_output() */
-			ip->ip_len = ntohs(ip->ip_len);
-			ip->ip_off = ntohs(ip->ip_off);
 			break;
 #ifdef INET6
 		case IPV6_VERSION >> 4:
Index: sys/netinet/sctp_usrreq.c
===================================================================
--- sys/netinet/sctp_usrreq.c	(revision 241405)
+++ sys/netinet/sctp_usrreq.c	(working copy)
@@ -180,7 +180,7 @@
 		SCTP_TCB_UNLOCK(stcb);
 		return;
 	}
-	totsz = ip->ip_len;
+	totsz = ntohs(ip->ip_len);
 
 	nxtsz = ntohs(icmph->icmp_nextmtu);
 	if (nxtsz == 0) {
Index: sys/netinet/ip_output.c
===================================================================
--- sys/netinet/ip_output.c	(revision 241405)
+++ sys/netinet/ip_output.c	(working copy)
@@ -102,7 +102,6 @@
 /*
  * IP output.  The packet in mbuf chain m contains a skeletal IP
  * header (with len, off, ttl, proto, tos, src, dst).
- * ip_len and ip_off are in host format.
  * The mbuf chain containing the packet will be freed.
  * The mbuf opt, if present, will not be freed.
  * If route ro is present and has ro_rt initialized, route lookup would be
@@ -175,6 +174,8 @@
 			hlen = len; /* ip->ip_hl is updated above */
 	}
 	ip = mtod(m, struct ip *);
+	ip_len = ntohs(ip->ip_len);
+	ip_off = ntohs(ip->ip_off);
 
 	/*
 	 * Fill in IP header.  If we are not allowing fragmentation,
@@ -442,7 +443,7 @@
 	 * packet or packet fragments, unless ALTQ is enabled on the given
 	 * interface in which case packetdrop should be done by queueing.
 	 */
-	n = ip->ip_len / mtu + 1; /* how many fragments ? */
+	n = ip_len / mtu + 1; /* how many fragments ? */
 	if (
 #ifdef ALTQ
 	    (!ALTQ_IS_ENABLED(&ifp->if_snd)) &&
@@ -469,7 +470,7 @@
 			goto bad;
 		}
 		/* don't allow broadcast messages to be fragmented */
-		if (ip->ip_len > mtu) {
+		if (ip_len > mtu) {
 			error = EMSGSIZE;
 			goto bad;
 		}
@@ -502,12 +503,6 @@
 	hlen = ip->ip_hl << 2;
 #endif /* IPSEC */
 
-	/*
-	 * To network byte order. pfil(9) hooks and ip_fragment() expect this.
-	 */
-	ip->ip_len = htons(ip->ip_len);
-	ip->ip_off = htons(ip->ip_off);
-
 	/* Jump over all PFIL processing if hooks are not active. */
 	if (!PFIL_HOOKED(&V_inet_pfil_hook))
 		goto passout;
@@ -544,8 +539,6 @@
 		} else {
 			if (ia != NULL)
 				ifa_free(&ia->ia_ifa);
-			ip->ip_len = ntohs(ip->ip_len);
-			ip->ip_off = ntohs(ip->ip_off);
 			goto again;	/* Redo the routing table lookup. */
 		}
 	}
@@ -579,16 +572,11 @@
 		m_tag_delete(m, fwd_tag);
 		if (ia != NULL)
 			ifa_free(&ia->ia_ifa);
-		ip->ip_len = ntohs(ip->ip_len);
-		ip->ip_off = ntohs(ip->ip_off);
 		goto again;
 	}
 #endif /* IPFIREWALL_FORWARD */
 
 passout:
-	ip_len = ntohs(ip->ip_len);
-	ip_off = ntohs(ip->ip_off);
-
 	/* 127/8 must not appear on wire - RFC1122. */
 	if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
 	    (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
@@ -1295,8 +1283,6 @@
  * calls the output routine of the loopback "driver", but with an interface
  * pointer that might NOT be a loopback interface -- evil, but easier than
  * replicating that code here.
- *
- * IP header in host byte order.
  */
 static void
 ip_mloopback(struct ifnet *ifp, struct mbuf *m, struct sockaddr_in *dst,
@@ -1326,8 +1312,6 @@
 		 * than the interface's MTU.  Can this possibly matter?
 		 */
 		ip = mtod(copym, struct ip *);
-		ip->ip_len = htons(ip->ip_len);
-		ip->ip_off = htons(ip->ip_off);
 		ip->ip_sum = 0;
 		ip->ip_sum = in_cksum(copym, hlen);
 #if 1 /* XXX */
Index: sys/netinet/in_gif.c
===================================================================
--- sys/netinet/in_gif.c	(revision 241405)
+++ sys/netinet/in_gif.c	(working copy)
@@ -192,7 +192,7 @@
 	iphdr.ip_p = proto;
 	/* version will be set in ip_output() */
 	iphdr.ip_ttl = V_ip_gif_ttl;
-	iphdr.ip_len = m->m_pkthdr.len + sizeof(struct ip);
+	iphdr.ip_len = htons(m->m_pkthdr.len + sizeof(struct ip));
 	ip_ecn_ingress((ifp->if_flags & IFF_LINK1) ? ECN_ALLOWED : ECN_NOCARE,
 		       &iphdr.ip_tos, &tos);
 
Index: sys/netinet/ip_options.c
===================================================================
--- sys/netinet/ip_options.c	(revision 241405)
+++ sys/netinet/ip_options.c	(working copy)
@@ -494,7 +494,7 @@
 	unsigned optlen;
 
 	optlen = opt->m_len - sizeof(p->ipopt_dst);
-	if (optlen + ip->ip_len > IP_MAXPACKET) {
+	if (optlen + ntohs(ip->ip_len) > IP_MAXPACKET) {
 		*phlen = 0;
 		return (m);		/* XXX should fail */
 	}
@@ -527,7 +527,7 @@
 	*phlen = sizeof(struct ip) + optlen;
 	ip->ip_v = IPVERSION;
 	ip->ip_hl = *phlen >> 2;
-	ip->ip_len += optlen;
+	ip->ip_len = htons(ntohs(ip->ip_len) + optlen);
 	return (m);
 }
 
Index: sys/netinet/tcp_timewait.c
===================================================================
--- sys/netinet/tcp_timewait.c	(revision 241405)
+++ sys/netinet/tcp_timewait.c	(working copy)
@@ -593,9 +593,9 @@
 		m->m_pkthdr.csum_flags = CSUM_TCP;
 		th->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr,
 		    htons(sizeof(struct tcphdr) + optlen + IPPROTO_TCP));
-		ip->ip_len = m->m_pkthdr.len;
+		ip->ip_len = htons(m->m_pkthdr.len);
 		if (V_path_mtu_discovery)
-			ip->ip_off |= IP_DF;
+			ip->ip_off |= htons(IP_DF);
 		error = ip_output(m, inp->inp_options, NULL,
 		    ((tw->tw_so_options & SO_DONTROUTE) ? IP_ROUTETOIF : 0),
 		    NULL, inp);
Index: sys/netinet/tcp_output.c
===================================================================
--- sys/netinet/tcp_output.c	(revision 241405)
+++ sys/netinet/tcp_output.c	(working copy)
@@ -1239,7 +1239,7 @@
 	struct route ro;
 
 	bzero(&ro, sizeof(ro));
-	ip->ip_len = m->m_pkthdr.len;
+	ip->ip_len = htons(m->m_pkthdr.len);
 #ifdef INET6
 	if (tp->t_inpcb->inp_vflag & INP_IPV6PROTO)
 		ip->ip_ttl = in6_selecthlim(tp->t_inpcb, NULL);
@@ -1253,7 +1253,7 @@
 	 * NB: Don't set DF on small MTU/MSS to have a safe fallback.
 	 */
 	if (V_path_mtu_discovery && tp->t_maxopd > V_tcp_minmss)
-		ip->ip_off |= IP_DF;
+		ip->ip_off |= htons(IP_DF);
 
 	error = ip_output(m, tp->t_inpcb->inp_options, &ro,
 	    ((so->so_options & SO_DONTROUTE) ? IP_ROUTETOIF : 0), 0,
Index: sys/netinet/sctputil.c
===================================================================
--- sys/netinet/sctputil.c	(revision 241405)
+++ sys/netinet/sctputil.c	(working copy)
@@ -6821,7 +6821,7 @@
 	switch (iph->ip_v) {
 #ifdef INET
 	case IPVERSION:
-		iph->ip_len -= sizeof(struct udphdr);
+		iph->ip_len = ntohs(htons(iph->ip_len) - sizeof(struct udphdr));
 		sctp_input_with_port(m, off, port);
 		break;
 #endif
Index: sys/netinet/ip_input.c
===================================================================
--- sys/netinet/ip_input.c	(revision 241405)
+++ sys/netinet/ip_input.c	(working copy)
@@ -390,9 +390,8 @@
 		m->m_flags &= ~M_FASTFWD_OURS;
 		/* Set up some basics that will be used later. */
 		ip = mtod(m, struct ip *);
-		ip->ip_len = ntohs(ip->ip_len);
-		ip->ip_off = ntohs(ip->ip_off);
 		hlen = ip->ip_hl << 2;
+		ip_len = ntohs(ip->ip_len);
 		goto ours;
 	}
 
@@ -513,8 +512,6 @@
 #ifdef IPFIREWALL_FORWARD
 	if (m->m_flags & M_FASTFWD_OURS) {
 		m->m_flags &= ~M_FASTFWD_OURS;
-		ip->ip_len = ntohs(ip->ip_len);
-		ip->ip_off = ntohs(ip->ip_off);
 		goto ours;
 	}
 	if ((dchg = (m_tag_find(m, PACKET_TAG_IPFORWARD, NULL) != NULL)) != 0) {
@@ -523,20 +520,12 @@
 		 * packets originally destined to us to some other directly
 		 * connected host.
 		 */
-		ip->ip_len = ntohs(ip->ip_len);
-		ip->ip_off = ntohs(ip->ip_off);
 		ip_forward(m, dchg);
 		return;
 	}
 #endif /* IPFIREWALL_FORWARD */
 
 passin:
-	/*
-	 *  From now and up to output pfil(9) processing in ip_output()
-	 *  the header is in host byte order.
-	 */
-	ip->ip_len = ntohs(ip->ip_len);
-	ip->ip_off = ntohs(ip->ip_off);
 
 	/*
 	 * Process options and, if not destined for us,
@@ -732,20 +721,21 @@
 	 * Attempt reassembly; if it succeeds, proceed.
 	 * ip_reass() will return a different mbuf.
 	 */
-	if (ip->ip_off & (IP_MF | IP_OFFMASK)) {
+	if (ip->ip_off & htons(IP_MF | IP_OFFMASK)) {
 		m = ip_reass(m);
 		if (m == NULL)
 			return;
 		ip = mtod(m, struct ip *);
 		/* Get the header length of the reassembled packet */
 		hlen = ip->ip_hl << 2;
+		ip_len = ntohs(ip->ip_len);
 	}
 
 	/*
 	 * Further protocols expect the packet length to be w/o the
 	 * IP header.
 	 */
-	ip->ip_len -= hlen;
+	ip->ip_len = htons(ip_len - hlen);
 
 #ifdef IPSEC
 	/*
@@ -914,22 +904,21 @@
 	 * Adjust ip_len to not reflect header,
 	 * convert offset of this to bytes.
 	 */
-	ip->ip_len -= hlen;
-	if (ip->ip_off & IP_MF) {
+	ip->ip_len = htons(ntohs(ip->ip_len) - hlen);
+	if (ip->ip_off & htons(IP_MF)) {
 		/*
 		 * Make sure that fragments have a data length
 		 * that's a non-zero multiple of 8 bytes.
 		 */
-		if (ip->ip_len == 0 || (ip->ip_len & 0x7) != 0) {
+		if (ntohs(ip->ip_len) == 0 || (ntohs(ip->ip_len & 0x7) != 0)) {
 			IPSTAT_INC(ips_toosmall); /* XXX */
 			goto dropfrag;
 		}
 		m->m_flags |= M_FRAG;
 	} else
 		m->m_flags &= ~M_FRAG;
-	ip->ip_off <<= 3;
+	ip->ip_off = htons(ntohs(ip->ip_off) << 3);
 
-
 	/*
 	 * Attempt reassembly; if it succeeds, proceed.
 	 * ip_reass() will return a different mbuf.
@@ -1000,7 +989,7 @@
 	 * Find a segment which begins after this one does.
 	 */
 	for (p = NULL, q = fp->ipq_frags; q; p = q, q = q->m_nextpkt)
-		if (GETIP(q)->ip_off > ip->ip_off)
+		if (ntohs(GETIP(q)->ip_off) > ntohs(ip->ip_off))
 			break;
 
 	/*
@@ -1013,14 +1002,15 @@
 	 * segment, then it's checksum is invalidated.
 	 */
 	if (p) {
-		i = GETIP(p)->ip_off + GETIP(p)->ip_len - ip->ip_off;
+		i = ntohs(GETIP(p)->ip_off) + ntohs(GETIP(p)->ip_len) -
+		    ntohs(ip->ip_off);
 		if (i > 0) {
-			if (i >= ip->ip_len)
+			if (i >= ntohs(ip->ip_len))
 				goto dropfrag;
 			m_adj(m, i);
 			m->m_pkthdr.csum_flags = 0;
-			ip->ip_off += i;
-			ip->ip_len -= i;
+			ip->ip_off = htons(ntohs(ip->ip_off) + i);
+			ip->ip_len = htons(ntohs(ip->ip_len) - i);
 		}
 		m->m_nextpkt = p->m_nextpkt;
 		p->m_nextpkt = m;
@@ -1033,12 +1023,13 @@
 	 * While we overlap succeeding segments trim them or,
 	 * if they are completely covered, dequeue them.
 	 */
-	for (; q != NULL && ip->ip_off + ip->ip_len > GETIP(q)->ip_off;
-	     q = nq) {
-		i = (ip->ip_off + ip->ip_len) - GETIP(q)->ip_off;
-		if (i < GETIP(q)->ip_len) {
-			GETIP(q)->ip_len -= i;
-			GETIP(q)->ip_off += i;
+	for (; q != NULL && ntohs(ip->ip_off) + ntohs(ip->ip_len) >
+	    ntohs(GETIP(q)->ip_off); q = nq) {
+		i = (ntohs(ip->ip_off) + ntohs(ip->ip_len)) -
+		    ntohs(GETIP(q)->ip_off);
+		if (i < ntohs(GETIP(q)->ip_len)) {
+			GETIP(q)->ip_len = htons(ntohs(GETIP(q)->ip_len) - i);
+			GETIP(q)->ip_off = htons(ntohs(GETIP(q)->ip_off) + i);
 			m_adj(q, i);
 			q->m_pkthdr.csum_flags = 0;
 			break;
@@ -1062,14 +1053,14 @@
 	 */
 	next = 0;
 	for (p = NULL, q = fp->ipq_frags; q; p = q, q = q->m_nextpkt) {
-		if (GETIP(q)->ip_off != next) {
+		if (ntohs(GETIP(q)->ip_off) != next) {
 			if (fp->ipq_nfrags > V_maxfragsperpacket) {
 				IPSTAT_ADD(ips_fragdropped, fp->ipq_nfrags);
 				ip_freef(head, fp);
 			}
 			goto done;
 		}
-		next += GETIP(q)->ip_len;
+		next += ntohs(GETIP(q)->ip_len);
 	}
 	/* Make sure the last packet didn't have the IP_MF flag */
 	if (p->m_flags & M_FRAG) {
@@ -1125,7 +1116,7 @@
 	 * packet;  dequeue and discard fragment reassembly header.
 	 * Make header visible.
 	 */
-	ip->ip_len = (ip->ip_hl << 2) + next;
+	ip->ip_len = htons((ip->ip_hl << 2) + next);
 	ip->ip_src = fp->ipq_src;
 	ip->ip_dst = fp->ipq_dst;
 	TAILQ_REMOVE(head, fp, ipq_list);
@@ -1365,8 +1356,6 @@
  *
  * The srcrt parameter indicates whether the packet is being forwarded
  * via a source route.
- *
- * IP header in host byte order.
  */
 void
 ip_forward(struct mbuf *m, int srcrt)
@@ -1436,7 +1425,7 @@
 		mcopy = NULL;
 	}
 	if (mcopy != NULL) {
-		mcopy->m_len = min(ip->ip_len, M_TRAILINGSPACE(mcopy));
+		mcopy->m_len = min(ntohs(ip->ip_len), M_TRAILINGSPACE(mcopy));
 		mcopy->m_pkthdr.len = mcopy->m_len;
 		m_copydata(m, 0, mcopy->m_len, mtod(mcopy, caddr_t));
 	}
@@ -1564,7 +1553,7 @@
 			if (ia != NULL)
 				mtu = ia->ia_ifp->if_mtu;
 			else
-				mtu = ip_next_mtu(ip->ip_len, 0);
+				mtu = ip_next_mtu(ntohs(ip->ip_len), 0);
 		}
 		IPSTAT_INC(ips_cantfrag);
 		break;
Index: sys/netinet/siftr.c
===================================================================
--- sys/netinet/siftr.c	(revision 241405)
+++ sys/netinet/siftr.c	(working copy)
@@ -952,7 +952,8 @@
 			 * the mbuf cluster "at" at offset "offset" bytes from
 			 * the beginning of the "at" mbuf's data pointer.
 			 */
-			th->th_sum = in_cksum_skip(*m, ip->ip_len, ip_hl);
+			th->th_sum = in_cksum_skip(*m, ntohs(ip->ip_len),
+			    ip_hl);
 		}
 
 		/*
Index: sys/netinet/ip_icmp.c
===================================================================
--- sys/netinet/ip_icmp.c	(revision 241405)
+++ sys/netinet/ip_icmp.c	(working copy)
@@ -200,7 +200,7 @@
 	 */
 	if (n->m_flags & M_DECRYPTED)
 		goto freeit;
-	if (oip->ip_off & ~(IP_MF|IP_DF))
+	if (oip->ip_off & htons(~(IP_MF|IP_DF)))
 		goto freeit;
 	if (n->m_flags & (M_BCAST|M_MCAST))
 		goto freeit;
@@ -234,16 +234,17 @@
 		tcphlen = th->th_off << 2;
 		if (tcphlen < sizeof(struct tcphdr))
 			goto freeit;
-		if (oip->ip_len < oiphlen + tcphlen)
+		if (ntohs(oip->ip_len) < oiphlen + tcphlen)
 			goto freeit;
 		if (oiphlen + tcphlen > n->m_len && n->m_next == NULL)
 			goto stdreply;
 		if (n->m_len < oiphlen + tcphlen && 
 		    ((n = m_pullup(n, oiphlen + tcphlen)) == NULL))
 			goto freeit;
-		icmpelen = max(tcphlen, min(V_icmp_quotelen, oip->ip_len - oiphlen));
+		icmpelen = max(tcphlen, min(V_icmp_quotelen,
+		    ntohs(oip->ip_len) - oiphlen));
 	} else
-stdreply:	icmpelen = max(8, min(V_icmp_quotelen, oip->ip_len - oiphlen));
+stdreply:	icmpelen = max(8, min(V_icmp_quotelen, ntohs(oip->ip_len) - oiphlen));
 
 	icmplen = min(oiphlen + icmpelen, nlen);
 	if (icmplen < sizeof(struct ip))
@@ -293,8 +294,6 @@
 	 */
 	m_copydata(n, 0, icmplen, (caddr_t)&icp->icmp_ip);
 	nip = &icp->icmp_ip;
-	nip->ip_len = htons(nip->ip_len);
-	nip->ip_off = htons(nip->ip_off);
 
 	/*
 	 * Set up ICMP message mbuf and copy old IP header (without options
@@ -309,7 +308,7 @@
 	m->m_pkthdr.rcvif = n->m_pkthdr.rcvif;
 	nip = mtod(m, struct ip *);
 	bcopy((caddr_t)oip, (caddr_t)nip, sizeof(struct ip));
-	nip->ip_len = m->m_len;
+	nip->ip_len = htons(m->m_len);
 	nip->ip_v = IPVERSION;
 	nip->ip_hl = 5;
 	nip->ip_p = IPPROTO_ICMP;
@@ -331,7 +330,7 @@
 	struct ip *ip = mtod(m, struct ip *);
 	struct sockaddr_in icmpsrc, icmpdst, icmpgw;
 	int hlen = off;
-	int icmplen = ip->ip_len;
+	int icmplen = ntohs(ip->ip_len);
 	int i, code;
 	void (*ctlfunc)(int, struct sockaddr *, void *);
 	int fibnum;
@@ -472,7 +471,6 @@
 			ICMPSTAT_INC(icps_badlen);
 			goto freeit;
 		}
-		icp->icmp_ip.ip_len = ntohs(icp->icmp_ip.ip_len);
 		/* Discard ICMP's in response to multicast packets */
 		if (IN_MULTICAST(ntohl(icp->icmp_ip.ip_dst.s_addr)))
 			goto badcode;
@@ -565,7 +563,8 @@
 		}
 		ifa_free(&ia->ia_ifa);
 reflect:
-		ip->ip_len += hlen;	/* since ip_input deducts this */
+		/* Since ip_input() deducts this. */
+		ip->ip_len = htons(ntohs(ip->ip_len) + hlen);
 		ICMPSTAT_INC(icps_reflect);
 		ICMPSTAT_INC(icps_outhist[icp->icmp_type]);
 		icmp_reflect(m);
@@ -835,7 +834,7 @@
 		 * Now strip out original options by copying rest of first
 		 * mbuf's data back, and adjust the IP length.
 		 */
-		ip->ip_len -= optlen;
+		ip->ip_len = htons(ntohs(ip->ip_len) - optlen);
 		ip->ip_v = IPVERSION;
 		ip->ip_hl = 5;
 		m->m_len -= optlen;
@@ -869,7 +868,7 @@
 	m->m_len -= hlen;
 	icp = mtod(m, struct icmp *);
 	icp->icmp_cksum = 0;
-	icp->icmp_cksum = in_cksum(m, ip->ip_len - hlen);
+	icp->icmp_cksum = in_cksum(m, ntohs(ip->ip_len) - hlen);
 	m->m_data -= hlen;
 	m->m_len += hlen;
 	m->m_pkthdr.rcvif = (struct ifnet *)0;
Index: sys/netinet/ip_gre.c
===================================================================
--- sys/netinet/ip_gre.c	(revision 241405)
+++ sys/netinet/ip_gre.c	(working copy)
@@ -274,12 +274,10 @@
 
 	/*
 	 * On FreeBSD, rip_input() supplies us with ip->ip_len
-	 * already converted into host byteorder and also decreases
-	 * it by the lengh of IP header, however, ip_input() expects
-	 * that this field is in the original format (network byteorder
-	 * and full size of IP packet), so that adjust accordingly.
+	 * decreased by the lengh of IP header, however, ip_input()
+	 * expects it to be full size of IP packet, so adjust accordingly.
 	 */
-	ip->ip_len = htons(ip->ip_len + sizeof(struct ip) - msiz);
+	ip->ip_len = htons(ntohs(ip->ip_len) + sizeof(struct ip) - msiz);
 
 	ip->ip_sum = 0;
 	ip->ip_sum = in_cksum(m, (ip->ip_hl << 2));
Index: sys/netinet/ip_ipsec.c
===================================================================
--- sys/netinet/ip_ipsec.c	(revision 241405)
+++ sys/netinet/ip_ipsec.c	(working copy)
@@ -264,7 +264,6 @@
 {
 #ifdef IPSEC
 	struct secpolicy *sp = NULL;
-	struct ip *ip = mtod(*m, struct ip *);
 	struct tdb_ident *tdbi;
 	struct m_tag *mtag;
 	int s;
@@ -332,9 +331,6 @@
 			}
 		}
 
-		ip->ip_len = htons(ip->ip_len);
-		ip->ip_off = htons(ip->ip_off);
-
 		/*
 		 * Do delayed checksums now because we send before
 		 * this is done in the normal processing path.
@@ -345,6 +341,8 @@
 		}
 #ifdef SCTP
 		if ((*m)->m_pkthdr.csum_flags & CSUM_SCTP) {
+			struct ip *ip = mtod(*m, struct ip *);
+
 			sctp_delayed_cksum(*m, (uint32_t)(ip->ip_hl << 2));
 			(*m)->m_pkthdr.csum_flags &= ~CSUM_SCTP;
 		}
@@ -359,8 +357,6 @@
 			 * IPsec processing and return without error.
 			 */
 			*error = 0;
-			ip->ip_len = ntohs(ip->ip_len);
-			ip->ip_off = ntohs(ip->ip_off);
 			goto done;
 		}
 		/*
Index: sys/netinet/ip_mroute.c
===================================================================
--- sys/netinet/ip_mroute.c	(revision 241405)
+++ sys/netinet/ip_mroute.c	(working copy)
@@ -1493,7 +1493,7 @@
 {
     struct ip  *ip = mtod(m, struct ip *);
     vifi_t vifi;
-    int plen = ip->ip_len;
+    int plen = ntohs(ip->ip_len);
 
     VIF_LOCK_ASSERT();
 
@@ -2376,10 +2376,7 @@
 
     /* Take care of delayed checksums */
     if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA) {
-	/* XXX: in_delayed_cksum() expects net byte order */
-	ip->ip_len = htons(ip->ip_len);
 	in_delayed_cksum(m);
-	ip->ip_len = ntohs(ip->ip_len);
 	m->m_pkthdr.csum_flags &= ~CSUM_DELAY_DATA;
     }
 
@@ -2401,10 +2398,8 @@
     /* Compute the MTU after the PIM Register encapsulation */
     mtu = 0xffff - sizeof(pim_encap_iphdr) - sizeof(pim_encap_pimhdr);
 
-    if (ip->ip_len <= mtu) {
+    if (ntohs(ip->ip_len) <= mtu) {
 	/* Turn the IP header into a valid one */
-	ip->ip_len = htons(ip->ip_len);
-	ip->ip_off = htons(ip->ip_off);
 	ip->ip_sum = 0;
 	ip->ip_sum = in_cksum(mb_copy, ip->ip_hl << 2);
     } else {
@@ -2509,7 +2504,8 @@
     ip_outer = mtod(mb_first, struct ip *);
     *ip_outer = pim_encap_iphdr;
     ip_outer->ip_id = ip_newid();
-    ip_outer->ip_len = len + sizeof(pim_encap_iphdr) + sizeof(pim_encap_pimhdr);
+    ip_outer->ip_len = htons(len + sizeof(pim_encap_iphdr) +
+	sizeof(pim_encap_pimhdr));
     ip_outer->ip_src = V_viftable[vifi].v_lcl_addr;
     ip_outer->ip_dst = rt->mfc_rp;
     /*
@@ -2517,8 +2513,8 @@
      * IP_DF bit.
      */
     ip_outer->ip_tos = ip->ip_tos;
-    if (ntohs(ip->ip_off) & IP_DF)
-	ip_outer->ip_off |= IP_DF;
+    if (ip->ip_off & htons(IP_DF))
+	ip_outer->ip_off |= htons(IP_DF);
     pimhdr = (struct pim_encap_pimhdr *)((caddr_t)ip_outer
 					 + sizeof(pim_encap_iphdr));
     *pimhdr = pim_encap_pimhdr;
@@ -2571,7 +2567,7 @@
     struct ip *ip = mtod(m, struct ip *);
     struct pim *pim;
     int minlen;
-    int datalen = ip->ip_len;
+    int datalen = ntohs(ip->ip_len);
     int ip_tos;
     int iphlen = off;
 
Index: sys/netinet/sctp_output.c
===================================================================
--- sys/netinet/sctp_output.c	(revision 241405)
+++ sys/netinet/sctp_output.c	(working copy)
@@ -3980,7 +3980,7 @@
 				tos_value |= sctp_get_ect(stcb);
 			}
 			if ((nofragment_flag) && (port == 0)) {
-				ip->ip_off = IP_DF;
+				ip->ip_off = htons(IP_DF);
 			} else
 				ip->ip_off = 0;
 
@@ -3988,7 +3988,7 @@
 			ip->ip_id = ip_newid();
 
 			ip->ip_ttl = inp->ip_inp.inp.inp_ip_ttl;
-			ip->ip_len = packet_length;
+			ip->ip_len = htons(packet_length);
 			ip->ip_tos = tos_value;
 			if (port) {
 				ip->ip_p = IPPROTO_UDP;
@@ -10991,7 +10991,7 @@
 				udp->uh_sum = 0;
 			}
 		}
-		ip->ip_len = len;
+		ip->ip_len = htons(len);
 		if (port) {
 #if defined(SCTP_WITH_NO_CSUM)
 			SCTP_STAT_INCR(sctps_sendnocrc);
Index: sys/netinet/igmp.c
===================================================================
--- sys/netinet/igmp.c	(revision 241405)
+++ sys/netinet/igmp.c	(working copy)
@@ -1442,7 +1442,7 @@
 
 	ip = mtod(m, struct ip *);
 	iphlen = off;
-	igmplen = ip->ip_len;
+	igmplen = ntohs(ip->ip_len);
 
 	/*
 	 * Validate lengths.
@@ -2225,7 +2225,7 @@
 
 	ip = mtod(m, struct ip *);
 	ip->ip_tos = 0;
-	ip->ip_len = sizeof(struct ip) + sizeof(struct igmp);
+	ip->ip_len = htons(sizeof(struct ip) + sizeof(struct igmp));
 	ip->ip_off = 0;
 	ip->ip_p = IPPROTO_IGMP;
 	ip->ip_src.s_addr = INADDR_ANY;
@@ -3522,8 +3522,8 @@
 
 	ip = mtod(m, struct ip *);
 	ip->ip_tos = IPTOS_PREC_INTERNETCONTROL;
-	ip->ip_len = hdrlen + igmpreclen;
-	ip->ip_off = IP_DF;
+	ip->ip_len = htons(hdrlen + igmpreclen);
+	ip->ip_off = htons(IP_DF);
 	ip->ip_p = IPPROTO_IGMP;
 	ip->ip_sum = 0;
 
Index: sys/netinet6/ip6_ipsec.c
===================================================================
--- sys/netinet6/ip6_ipsec.c	(revision 241405)
+++ sys/netinet6/ip6_ipsec.c	(working copy)
@@ -295,15 +295,9 @@
 		 */
 #ifdef INET
 		if ((*m)->m_pkthdr.csum_flags & CSUM_DELAY_DATA) {
-			struct ip *ip;
-
 			ipseclog((LOG_DEBUG,
 			    "%s: we do not support IPv4 over IPv6", __func__));
-			/* XXX: in_delayed_cksum() expects net byte order */
-			ip = mtod(*m, struct ip *);
-			ip->ip_len = htons(ip->ip_len);
 			in_delayed_cksum(*m);
-			ip->ip_len = ntohs(ip->ip_len);
 			(*m)->m_pkthdr.csum_flags &= ~CSUM_DELAY_DATA;
 		}
 #endif
Index: sys/netpfil/pf/pf.c
===================================================================
--- sys/netpfil/pf/pf.c	(revision 241405)
+++ sys/netpfil/pf/pf.c	(working copy)
@@ -2252,8 +2252,8 @@
 		h->ip_v = 4;
 		h->ip_hl = sizeof(*h) >> 2;
 		h->ip_tos = IPTOS_LOWDELAY;
-		h->ip_off = V_path_mtu_discovery ? IP_DF : 0;
-		h->ip_len = len;
+		h->ip_off = V_path_mtu_discovery ? htons(IP_DF) : 0;
+		h->ip_len = htons(len);
 		h->ip_ttl = ttl ? ttl : V_ip_defttl;
 		h->ip_sum = 0;
 
@@ -2316,17 +2316,8 @@
 	switch (af) {
 #ifdef INET
 	case AF_INET:
-	    {
-		struct ip *ip;
-
-		/* icmp_error() expects host byte ordering */
-		ip = mtod(m0, struct ip *);
-		NTOHS(ip->ip_len);
-		NTOHS(ip->ip_off);
-
 		pfse->pfse_type = PFSE_ICMP;
 		break;
-	    }
 #endif /* INET */
 #ifdef INET6
 	case AF_INET6:
Index: sys/netpfil/pf/if_pfsync.c
===================================================================
--- sys/netpfil/pf/if_pfsync.c	(revision 241405)
+++ sys/netpfil/pf/if_pfsync.c	(working copy)
@@ -1384,7 +1384,7 @@
 		ip->ip_hl = sizeof(sc->sc_template) >> 2;
 		ip->ip_tos = IPTOS_LOWDELAY;
 		/* len and id are set later. */
-		ip->ip_off = IP_DF;
+		ip->ip_off = htons(IP_DF);
 		ip->ip_ttl = PFSYNC_DFLTTL;
 		ip->ip_p = IPPROTO_PFSYNC;
 		ip->ip_src.s_addr = INADDR_ANY;
@@ -1522,7 +1522,7 @@
 	bcopy(&sc->sc_template, ip, sizeof(*ip));
 	offset = sizeof(*ip);
 
-	ip->ip_len = m->m_pkthdr.len;
+	ip->ip_len = htons(m->m_pkthdr.len);
 	ip->ip_id = htons(ip_randomid());
 
 	/* build the pfsync header */
Index: sys/netpfil/ipfw/ip_fw_dynamic.c
===================================================================
--- sys/netpfil/ipfw/ip_fw_dynamic.c	(revision 241405)
+++ sys/netpfil/ipfw/ip_fw_dynamic.c	(working copy)
@@ -1017,8 +1017,7 @@
 		h->ip_hl = sizeof(*h) >> 2;
 		h->ip_tos = IPTOS_LOWDELAY;
 		h->ip_off = 0;
-		/* ip_len must be in host format for ip_output */
-		h->ip_len = len;
+		h->ip_len = htons(len);
 		h->ip_ttl = V_ip_defttl;
 		h->ip_sum = 0;
 		break;
Index: sys/netpfil/ipfw/ip_fw_pfil.c
===================================================================
--- sys/netpfil/ipfw/ip_fw_pfil.c	(revision 241405)
+++ sys/netpfil/ipfw/ip_fw_pfil.c	(working copy)
@@ -431,7 +431,6 @@
 		int hlen;
 		struct mbuf *reass;
 
-		SET_HOST_IPLEN(ip); /* ip_reass wants host order */
 		reass = ip_reass(clone); /* Reassemble packet. */
 		if (reass == NULL)
 			return 0; /* not an error */
@@ -442,7 +441,6 @@
 		 */
 		ip = mtod(reass, struct ip *);
 		hlen = ip->ip_hl << 2;
-		SET_NET_IPLEN(ip);
 		ip->ip_sum = 0;
 		if (hlen == sizeof(struct ip))
 			ip->ip_sum = in_cksum_hdr(ip);
Index: sys/netpfil/ipfw/ip_dn_io.c
===================================================================
--- sys/netpfil/ipfw/ip_dn_io.c	(revision 241405)
+++ sys/netpfil/ipfw/ip_dn_io.c	(working copy)
@@ -658,13 +658,10 @@
 
 		switch (dst) {
 		case DIR_OUT:
-			SET_HOST_IPLEN(mtod(m, struct ip *));
 			ip_output(m, NULL, NULL, IP_FORWARDING, NULL, NULL);
 			break ;
 
 		case DIR_IN :
-			/* put header in network format for ip_input() */
-			//SET_NET_IPLEN(mtod(m, struct ip *));
 			netisr_dispatch(NETISR_IP, m);
 			break;
 
Index: sys/netpfil/ipfw/ip_fw2.c
===================================================================
--- sys/netpfil/ipfw/ip_fw2.c	(revision 241405)
+++ sys/netpfil/ipfw/ip_fw2.c	(working copy)
@@ -627,8 +627,6 @@
 		m_adj(m, args->L3offset);
 #endif
 	if (code != ICMP_REJECT_RST) { /* Send an ICMP unreach */
-		/* We need the IP header in host order for icmp_error(). */
-		SET_HOST_IPLEN(ip);
 		icmp_error(args->m, ICMP_UNREACH, code, 0L, 0);
 	} else if (args->f_id.proto == IPPROTO_TCP) {
 		struct tcphdr *const tcp =
@@ -2418,11 +2416,6 @@
 				/* if not fragmented, go to next rule */
 				if ((ip_off & (IP_MF | IP_OFFMASK)) == 0)
 				    break;
-				/* 
-				 * ip_reass() expects len & off in host
-				 * byte order.
-				 */
-				SET_HOST_IPLEN(ip);
 
 				args->m = m = ip_reass(m);
 
@@ -2436,7 +2429,6 @@
 
 				    ip = mtod(m, struct ip *);
 				    hlen = ip->ip_hl << 2;
-				    SET_NET_IPLEN(ip);
 				    ip->ip_sum = 0;
 				    if (hlen == sizeof(struct ip))
 					ip->ip_sum = in_cksum_hdr(ip);
Index: sys/netipsec/ipsec.c
===================================================================
--- sys/netipsec/ipsec.c	(revision 241405)
+++ sys/netipsec/ipsec.c	(working copy)
@@ -597,10 +597,9 @@
 	IPSEC_ASSERT(m != NULL, ("null mbuf"));
 	IPSEC_ASSERT(m->m_pkthdr.len >= sizeof(struct ip),("packet too short"));
 
-	/* NB: ip_input() flips it into host endian. XXX Need more checking. */
 	if (m->m_len >= sizeof (struct ip)) {
 		struct ip *ip = mtod(m, struct ip *);
-		if (ip->ip_off & (IP_MF | IP_OFFMASK))
+		if (ip->ip_off & htons(IP_MF | IP_OFFMASK))
 			goto done;
 #ifdef _IP_VHL
 		off = _IP_VHL_HL(ip->ip_vhl) << 2;
@@ -612,7 +611,7 @@
 		struct ip ih;
 
 		m_copydata(m, 0, sizeof (struct ip), (caddr_t) &ih);
-		if (ih.ip_off & (IP_MF | IP_OFFMASK))
+		if (ih.ip_off & htons(IP_MF | IP_OFFMASK))
 			goto done;
 #ifdef _IP_VHL
 		off = _IP_VHL_HL(ih.ip_vhl) << 2;
Index: sys/netgraph/ng_ipfw.c
===================================================================
--- sys/netgraph/ng_ipfw.c	(revision 241405)
+++ sys/netgraph/ng_ipfw.c	(working copy)
@@ -265,7 +265,6 @@
 		switch (ip->ip_v) {
 #ifdef INET
 		case IPVERSION:
-			SET_HOST_IPLEN(ip);
 			return (ip_output(m, NULL, NULL, IP_FORWARDING,
 			    NULL, NULL));
 #endif
Index: sys/net/if_stf.c
===================================================================
--- sys/net/if_stf.c	(revision 241405)
+++ sys/net/if_stf.c	(working copy)
@@ -523,7 +523,7 @@
 	bcopy(&in4, &ip->ip_dst, sizeof(ip->ip_dst));
 	ip->ip_p = IPPROTO_IPV6;
 	ip->ip_ttl = ip_stf_ttl;
-	ip->ip_len = m->m_pkthdr.len;	/*host order*/
+	ip->ip_len = htons(m->m_pkthdr.len);
 	if (ifp->if_flags & IFF_LINK1)
 		ip_ecn_ingress(ECN_ALLOWED, &ip->ip_tos, &tos);
 	else
Index: sys/net/if_gre.c
===================================================================
--- sys/net/if_gre.c	(revision 241405)
+++ sys/net/if_gre.c	(working copy)
@@ -356,7 +356,7 @@
 			 * RFC2004 specifies that fragmented diagrams shouldn't
 			 * be encapsulated.
 			 */
-			if (ip->ip_off & (IP_MF | IP_OFFMASK)) {
+			if (ip->ip_off & htons(IP_MF | IP_OFFMASK)) {
 				_IF_DROP(&ifp->if_snd);
 				m_freem(m);
 				error = EINVAL;    /* is there better errno? */
@@ -410,7 +410,7 @@
 			}
 			ip = mtod(m, struct ip *);
 			memcpy((caddr_t)(ip + 1), &mob_h, (unsigned)msiz);
-			ip->ip_len = ntohs(ip->ip_len) + msiz;
+			ip->ip_len = htons(ntohs(ip->ip_len) + msiz);
 		} else {  /* AF_INET */
 			_IF_DROP(&ifp->if_snd);
 			m_freem(m);
@@ -493,7 +493,7 @@
 		((struct ip*)gh)->ip_ttl = GRE_TTL;
 		((struct ip*)gh)->ip_tos = gre_ip_tos;
 		((struct ip*)gh)->ip_id = gre_ip_id;
-		gh->gi_len = m->m_pkthdr.len;
+		gh->gi_len = htons(m->m_pkthdr.len);
 	}
 
 	ifp->if_opackets++;

--HnQK338I3UIa/qiP--