Date: Mon, 1 Sep 2014 14:04:51 +0000 (UTC) From: Gleb Smirnoff <glebius@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r270929 - in head: share/man/man4 sys/netinet sys/sys usr.sbin/traceroute Message-ID: <201409011404.s81E4pJg051138@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: glebius Date: Mon Sep 1 14:04:51 2014 New Revision: 270929 URL: http://svnweb.freebsd.org/changeset/base/270929 Log: Make SOCK_RAW sockets to be truly raw, not modifying received and sent packets at all. Swapping byte order on SOCK_RAW was actually a bug, an artifact from the BSD network stack, that used to convert a packet to native byte order once it is received by kernel. Other operating systems didn't follow this, and later other BSD descendants fixed this, leaving us alone with the bug. Now it is clear that we should fix the bug. In collaboration with: Olivier Cochard-Labbé <olivier cochard.me> See also: https://wiki.freebsd.org/SOCK_RAW Sponsored by: Nginx, Inc. Modified: head/share/man/man4/ip.4 head/sys/netinet/raw_ip.c head/sys/sys/param.h head/usr.sbin/traceroute/Makefile Modified: head/share/man/man4/ip.4 ============================================================================== --- head/share/man/man4/ip.4 Mon Sep 1 13:00:45 2014 (r270928) +++ head/share/man/man4/ip.4 Mon Sep 1 14:04:51 2014 (r270929) @@ -28,7 +28,7 @@ .\" @(#)ip.4 8.2 (Berkeley) 11/30/93 .\" $FreeBSD$ .\" -.Dd October 12, 2012 +.Dd September 1, 2014 .Dt IP 4 .Os .Sh NAME @@ -755,13 +755,11 @@ number the socket is created with), unless the .Dv IP_HDRINCL option has been set. -Incoming packets are received with +Unlike in previous +.Bx +releases, incoming packets are received with .Tn IP -header and options intact, except for -.Va ip_len -and -.Va ip_off -fields converted to host byte order. +header and options intact, leaving all fields in network byte order. .Pp .Dv IP_HDRINCL indicates the complete IP header is included with the data @@ -784,17 +782,16 @@ the fields of the IP header, including t ip->ip_v = IPVERSION; ip->ip_hl = hlen >> 2; ip->ip_id = 0; /* 0 means kernel set appropriate value */ -ip->ip_off = offset; +ip->ip_off = htons(offset); +ip->ip_len = htons(len); .Ed .Pp -The +The packet should be provided as is to be sent over wire. +This implies all fields, including .Va ip_len and .Va ip_off -fields -.Em must -be provided in host byte order. -All other fields must be provided in network byte order. +to be in network byte order. See .Xr byteorder 3 for more information on network byte order. @@ -891,3 +888,16 @@ packets received on raw IP sockets had t subtracted from the .Va ip_len field. +.Pp +Before +.Fx 11.0 +packets received on raw IP sockets had the +.Va ip_len +and +.Va ip_off +fields converted to host byte order. +Packets written to raw IP sockets were expected to have +.Va ip_len +and +.Va ip_off +in host byte order. Modified: head/sys/netinet/raw_ip.c ============================================================================== --- head/sys/netinet/raw_ip.c Mon Sep 1 13:00:45 2014 (r270928) +++ head/sys/netinet/raw_ip.c Mon Sep 1 14:04:51 2014 (r270929) @@ -290,11 +290,6 @@ rip_input(struct mbuf **mp, int *offp, i last = NULL; ifp = m->m_pkthdr.rcvif; - /* - * Applications on raw sockets expect host byte order. - */ - ip->ip_len = ntohs(ip->ip_len); - ip->ip_off = ntohs(ip->ip_off); hash = INP_PCBHASH_RAW(proto, ip->ip_src.s_addr, ip->ip_dst.s_addr, V_ripcbinfo.ipi_hashmask); @@ -504,8 +499,8 @@ rip_output(struct mbuf *m, struct socket * and don't allow packet length sizes that will crash. */ if (((ip->ip_hl != (sizeof (*ip) >> 2)) && inp->inp_options) - || (ip->ip_len > m->m_pkthdr.len) - || (ip->ip_len < (ip->ip_hl << 2))) { + || (ntohs(ip->ip_len) > m->m_pkthdr.len) + || (ntohs(ip->ip_len) < (ip->ip_hl << 2))) { INP_RUNLOCK(inp); m_freem(m); return (EINVAL); @@ -514,13 +509,6 @@ rip_output(struct mbuf *m, struct socket ip->ip_id = ip_newid(); /* - * Applications on raw sockets pass us packets - * in host byte order. - */ - ip->ip_len = htons(ip->ip_len); - ip->ip_off = htons(ip->ip_off); - - /* * XXX prevent ip_output from overwriting header fields. */ flags |= IP_RAWOUTPUT; Modified: head/sys/sys/param.h ============================================================================== --- head/sys/sys/param.h Mon Sep 1 13:00:45 2014 (r270928) +++ head/sys/sys/param.h Mon Sep 1 14:04:51 2014 (r270929) @@ -58,7 +58,7 @@ * in the range 5 to 9. */ #undef __FreeBSD_version -#define __FreeBSD_version 1100029 /* Master, propagated to newvers */ +#define __FreeBSD_version 1100030 /* Master, propagated to newvers */ /* * __FreeBSD_kernel__ indicates that this system uses the kernel of FreeBSD, Modified: head/usr.sbin/traceroute/Makefile ============================================================================== --- head/usr.sbin/traceroute/Makefile Mon Sep 1 13:00:45 2014 (r270928) +++ head/usr.sbin/traceroute/Makefile Mon Sep 1 14:04:51 2014 (r270929) @@ -13,7 +13,7 @@ CLEANFILES= version.c CFLAGS+= -DHAVE_SYS_SELECT_H=1 -DHAVE_SYS_SOCKIO_H=1 \ -DHAVE_NET_ROUTE_H=1 -DHAVE_NET_IF_DL_H=1 \ -DHAVE_STRERROR=1 -DHAVE_USLEEP=1 \ - -DHAVE_SYS_SYSCTL_H=1 \ + -DHAVE_SYS_SYSCTL_H=1 -DBYTESWAP_IP_HDR=1 \ -DHAVE_SETLINEBUF=1 -DHAVE_RAW_OPTIONS=1 \ -DHAVE_SOCKADDR_SA_LEN=1 -DHAVE_ICMP_NEXTMTU=1 .if !defined(TRACEROUTE_NO_IPSEC)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201409011404.s81E4pJg051138>