/* $NetBSD: ip_fil.c,v 1.81 2002/06/09 16:33:39 itojun Exp $ */ /* * Copyright (C) 1993-2001 by Darren Reed. * * See the IPFILTER.LICENCE file for details on licencing. */ #ifndef SOLARIS #define SOLARIS (defined(sun) && (defined(__svr4__) || defined(__SVR4))) #endif #if defined(KERNEL) && !defined(_KERNEL) # define _KERNEL #endif #if defined(_KERNEL) && defined(__FreeBSD_version) && \ (__FreeBSD_version >= 400000) && !defined(KLD_MODULE) #include "opt_inet6.h" #endif #include #if defined(__NetBSD__) && (NetBSD >= 199905) && !defined(IPFILTER_LKM) && \ defined(_KERNEL) && !defined(_LKM) # include "opt_ipfilter_log.h" #endif #if defined(__FreeBSD__) && !defined(__FreeBSD_version) # if !defined(_KERNEL) || defined(IPFILTER_LKM) # include # endif #endif #ifdef __sgi # define _KMEMUSER # include #endif #ifndef _KERNEL # include # include # include # include # include #endif #include #include #include #if __FreeBSD_version >= 220000 && defined(_KERNEL) # include # include #else # include #endif #include #ifdef _KERNEL # include #endif #if !SOLARIS # if (NetBSD > 199609) || (OpenBSD > 199603) || (__FreeBSD_version >= 300000) # include # else # include # endif # include #else # include #endif #include #include #include #ifdef sun # include #endif #if __FreeBSD_version >= 300000 # include # if defined(_KERNEL) && !defined(IPFILTER_LKM) # include "opt_ipfilter.h" # endif #endif #ifdef __sgi #include # ifdef IFF_DRVRLOCK /* IRIX6 */ #include # endif #endif #include #include #if !(defined(__sgi) && !defined(IFF_DRVRLOCK)) /* IRIX < 6 */ # include #endif #include #include #include #include #include #include #include #ifndef _KERNEL # include # include #endif #include "netinet/ip_compat.h" #ifdef USE_INET6 # include # if !SOLARIS # include # include # endif #endif #include "netinet/ip_fil.h" #include "netinet/ip_nat.h" #include "netinet/ip_frag.h" #include "netinet/ip_state.h" #include "netinet/ip_proxy.h" #include "netinet/ip_auth.h" #if defined(__FreeBSD_version) && (__FreeBSD_version >= 300000) # include #endif #ifndef MIN # define MIN(a,b) (((a)<(b))?(a):(b)) #endif #if !SOLARIS && defined(_KERNEL) && !defined(__sgi) # include extern int ip_optcopy __P((struct ip *, struct ip *)); #endif #if !defined(lint) #if defined(__NetBSD__) #include __KERNEL_RCSID(0, "$NetBSD: ip_fil.c,v 1.81 2002/06/09 16:33:39 itojun Exp $"); #else static const char sccsid[] = "@(#)ip_fil.c 2.41 6/5/96 (C) 1993-2000 Darren Reed"; static const char rcsid[] = "@(#)Id: ip_fil.c,v 2.42.2.55 2002/03/26 15:54:39 darrenr Exp"; #endif #endif extern struct protosw inetsw[]; #ifndef _KERNEL # include "ipt.h" static struct ifnet **ifneta = NULL; static int nifs = 0; #else # if (BSD < 199306) || defined(__sgi) extern int tcp_ttl; # endif #endif #ifdef ICMP_UNREACH_FILTER_PROHIB int ipl_unreach = ICMP_UNREACH_FILTER_PROHIB; #else int ipl_unreach = ICMP_UNREACH_FILTER; #endif u_long ipl_frouteok[2] = {0, 0}; static int frzerostats __P((caddr_t)); #if defined(__NetBSD__) || defined(__OpenBSD__) || (__FreeBSD_version >= 300003) static int frrequest __P((int, u_long, caddr_t, int)); #else static int frrequest __P((int, int, caddr_t, int)); #endif #ifdef _KERNEL static int (*fr_savep) __P((ip_t *, int, void *, int, struct mbuf **)); static int send_ip __P((ip_t *, fr_info_t *, struct mbuf **)); # ifdef USE_INET6 static int ipfr_fastroute6 __P((struct mbuf *, struct mbuf **, fr_info_t *, frdest_t *)); # endif # ifdef __sgi extern kmutex_t ipf_rw; extern KRWLOCK_T ipf_mutex; # endif #else void init_ifp __P((void)); # if defined(__sgi) && (IRIX < 605) static int no_output __P((struct ifnet *, struct mbuf *, struct sockaddr *)); static int write_output __P((struct ifnet *, struct mbuf *, struct sockaddr *)); # else static int no_output __P((struct ifnet *, struct mbuf *, struct sockaddr *, struct rtentry *)); static int write_output __P((struct ifnet *, struct mbuf *, struct sockaddr *, struct rtentry *)); # endif #endif int fr_running = 0; #if (__FreeBSD_version >= 300000) && defined(_KERNEL) struct callout_handle ipfr_slowtimer_ch; #endif #if defined(__NetBSD__) && (__NetBSD_Version__ >= 104230000) # include struct callout ipfr_slowtimer_ch; #endif #if defined(__OpenBSD__) # include struct timeout ipfr_slowtimer_ch; #endif #if defined(__sgi) && defined(_KERNEL) toid_t ipfr_slowtimer_ch; #endif #if (_BSDI_VERSION >= 199510) && defined(_KERNEL) # include # include struct cfdriver iplcd = { NULL, "ipl", NULL, NULL, DV_DULL, 0 }; struct devsw iplsw = { &iplcd, iplopen, iplclose, iplread, nowrite, iplioctl, noselect, nommap, nostrat, nodump, nopsize, 0, nostop }; #endif /* _BSDI_VERSION >= 199510 && _KERNEL */ #if defined(__NetBSD__) || defined(__OpenBSD__) || (_BSDI_VERSION >= 199701) # include # if defined(NETBSD_PF) # include /* * We provide the fr_checkp name just to minimize changes later. */ int (*fr_checkp) __P((ip_t *ip, int hlen, void *ifp, int out, mb_t **mp)); # endif /* NETBSD_PF */ #endif /* __NetBSD__ */ #if defined(__NetBSD_Version__) && (__NetBSD_Version__ >= 105110000) && \ defined(_KERNEL) # include static int fr_check_wrapper(void *, struct mbuf **, struct ifnet *, int ); static int fr_check_wrapper(arg, mp, ifp, dir) void *arg; struct mbuf **mp; struct ifnet *ifp; int dir; { struct ip *ip = mtod(*mp, struct ip *); int rv, hlen = ip->ip_hl << 2; #if defined(M_CSUM_TCPv4) /* * If the packet is out-bound, we can't delay checksums * here. For in-bound, the checksum has already been * validated. */ if (dir == PFIL_OUT) { if ((*mp)->m_pkthdr.csum_flags & (M_CSUM_TCPv4|M_CSUM_UDPv4)) { in_delayed_cksum(*mp); (*mp)->m_pkthdr.csum_flags &= ~(M_CSUM_TCPv4|M_CSUM_UDPv4); } } #endif /* M_CSUM_TCPv4 */ /* * We get the packet with all fields in network byte * order. We expect ip_len and ip_off to be in host * order. We frob them, call the filter, then frob * them back. * * Note, we don't need to update the checksum, because * it has already been verified. */ NTOHS(ip->ip_len); NTOHS(ip->ip_off); rv = fr_check(ip, hlen, ifp, (dir == PFIL_OUT), mp); if (rv == 0 && *mp != NULL) { ip = mtod(*mp, struct ip *); HTONS(ip->ip_len); HTONS(ip->ip_off); } return (rv); } # ifdef USE_INET6 # include static int fr_check_wrapper6(void *, struct mbuf **, struct ifnet *, int ); static int fr_check_wrapper6(arg, mp, ifp, dir) void *arg; struct mbuf **mp; struct ifnet *ifp; int dir; { return (fr_check(mtod(*mp, struct ip *), sizeof(struct ip6_hdr), ifp, (dir == PFIL_OUT), mp)); } # endif #endif /* __NetBSD_Version >= 105110000 && _KERNEL */ #ifdef _KERNEL # if defined(IPFILTER_LKM) && !defined(__sgi) int iplidentify(s) char *s; { if (strcmp(s, "ipl") == 0) return 1; return 0; } # endif /* IPFILTER_LKM */ /* * Try to detect the case when compiling for NetBSD with pseudo-device */ # if defined(__NetBSD__) && defined(PFIL_HOOKS) void ipfilterattach(count) int count; { /* * Do nothing here, really. The filter will be enabled * by the SIOCFRENB ioctl. */ } # endif # if defined(__NetBSD__) int ipl_enable() # else int iplattach() # endif { char *defpass; int s; # if defined(__sgi) || (defined(NETBSD_PF) && (__NetBSD_Version__ >= 104200000)) int error = 0; # endif #if defined(__NetBSD_Version__) && (__NetBSD_Version__ >= 105110000) struct pfil_head *ph_inet; # ifdef USE_INET6 struct pfil_head *ph_inet6; # endif #endif SPL_NET(s); if (fr_running || (fr_checkp == fr_check)) { printf("IP Filter: already initialized\n"); SPL_X(s); return EBUSY; } # ifdef IPFILTER_LOG ipflog_init(); # endif if (nat_init() == -1) { SPL_X(s); return EIO; } if (fr_stateinit() == -1) { SPL_X(s); return EIO; } if (appr_init() == -1) { SPL_X(s); return EIO; } # ifdef NETBSD_PF # if (__NetBSD_Version__ >= 104200000) || (__FreeBSD_version >= 500011) # if __NetBSD_Version__ >= 105110000 ph_inet = pfil_head_get(PFIL_TYPE_AF, AF_INET); # ifdef USE_INET6 ph_inet6 = pfil_head_get(PFIL_TYPE_AF, AF_INET6); # endif if (ph_inet == NULL # ifdef USE_INET6 && ph_inet6 == NULL # endif ) return ENODEV; if (ph_inet != NULL) error = pfil_add_hook((void *)fr_check_wrapper, NULL, PFIL_IN|PFIL_OUT, ph_inet); else error = 0; # else error = pfil_add_hook((void *)fr_check, PFIL_IN|PFIL_OUT, &inetsw[ip_protox[IPPROTO_IP]].pr_pfh); # endif if (error) { # ifdef USE_INET6 goto pfil_error; # else SPL_X(s); appr_unload(); ip_natunload(); fr_stateunload(); return error; # endif } # else pfil_add_hook((void *)fr_check, PFIL_IN|PFIL_OUT); # endif # ifdef USE_INET6 # if __NetBSD_Version__ >= 105110000 if (ph_inet6 != NULL) error = pfil_add_hook((void *)fr_check_wrapper6, NULL, PFIL_IN|PFIL_OUT, ph_inet6); else error = 0; if (error) { pfil_remove_hook((void *)fr_check_wrapper6, NULL, PFIL_IN|PFIL_OUT, ph_inet6); # else error = pfil_add_hook((void *)fr_check, PFIL_IN|PFIL_OUT, &inet6sw[ip6_protox[IPPROTO_IPV6]].pr_pfh); if (error) { pfil_remove_hook((void *)fr_check, PFIL_IN|PFIL_OUT, &inetsw[ip_protox[IPPROTO_IP]].pr_pfh); # endif pfil_error: SPL_X(s); appr_unload(); ip_natunload(); fr_stateunload(); return error; } # endif # endif # ifdef __sgi error = ipfilter_sgi_attach(); if (error) { SPL_X(s); appr_unload(); ip_natunload(); fr_stateunload(); return error; } # endif bzero((char *)frcache, sizeof(frcache)); fr_savep = fr_checkp; fr_checkp = fr_check; fr_running = 1; SPL_X(s); if (fr_pass & FR_PASS) defpass = "pass"; else if (fr_pass & FR_BLOCK) defpass = "block"; else defpass = "no-match -> block"; printf("%s initialized. Default = %s all, Logging = %s\n", ipfilter_version, defpass, # ifdef IPFILTER_LOG "enabled"); # else "disabled"); # endif #ifdef _KERNEL # if defined(__NetBSD__) && (__NetBSD_Version__ >= 104230000) callout_init(&ipfr_slowtimer_ch); callout_reset(&ipfr_slowtimer_ch, hz / 2, ipfr_slowtimer, NULL); # else # if defined(__OpenBSD__) timeout_set(&ipfr_slowtimer_ch, ipfr_slowtimer, NULL); timeout_add(&ipfr_slowtimer_ch, hz/2); # else # if (__FreeBSD_version >= 300000) || defined(__sgi) ipfr_slowtimer_ch = timeout(ipfr_slowtimer, NULL, hz/2); # else timeout(ipfr_slowtimer, NULL, hz/2); # endif # endif # endif #endif return 0; } /* * Disable the filter by removing the hooks from the IP input/output * stream. */ # if defined(__NetBSD__) int ipl_disable() # else int ipldetach() # endif { int s, i = FR_INQUE|FR_OUTQUE; #if defined(NETBSD_PF) && (__NetBSD_Version__ >= 104200000) int error = 0; # if __NetBSD_Version__ >= 105150000 struct pfil_head *ph_inet = pfil_head_get(PFIL_TYPE_AF, AF_INET); # ifdef USE_INET6 struct pfil_head *ph_inet6 = pfil_head_get(PFIL_TYPE_AF, AF_INET6); # endif # endif #endif #ifdef _KERNEL # if defined(__NetBSD__) && (__NetBSD_Version__ >= 104230000) callout_stop(&ipfr_slowtimer_ch); # else # if (__FreeBSD_version >= 300000) untimeout(ipfr_slowtimer, NULL, ipfr_slowtimer_ch); # else # ifdef __sgi untimeout(ipfr_slowtimer_ch); # else # if defined(__OpenBSD__) timeout_del(&ipfr_slowtimer_ch); # else untimeout(ipfr_slowtimer, NULL); # endif /* OpenBSD */ # endif /* __sgi */ # endif /* FreeBSD */ # endif /* NetBSD */ #endif SPL_NET(s); if (!fr_running) { printf("IP Filter: not initialized\n"); SPL_X(s); return 0; } printf("%s unloaded\n", ipfilter_version); fr_checkp = fr_savep; i = frflush(IPL_LOGIPF, i); fr_running = 0; # ifdef NETBSD_PF # if ((__NetBSD_Version__ >= 104200000) || (__FreeBSD_version >= 500011)) # if __NetBSD_Version__ >= 105110000 if (ph_inet != NULL) error = pfil_remove_hook((void *)fr_check_wrapper, NULL, PFIL_IN|PFIL_OUT, ph_inet); else error = 0; # else error = pfil_remove_hook((void *)fr_check, PFIL_IN|PFIL_OUT, &inetsw[ip_protox[IPPROTO_IP]].pr_pfh); # endif if (error) { SPL_X(s); return error; } # else pfil_remove_hook((void *)fr_check, PFIL_IN|PFIL_OUT); # endif # ifdef USE_INET6 # if __NetBSD_Version__ >= 105110000 if (ph_inet6 != NULL) error = pfil_remove_hook((void *)fr_check_wrapper6, NULL, PFIL_IN|PFIL_OUT, ph_inet6); else error = 0; # else error = pfil_remove_hook((void *)fr_check, PFIL_IN|PFIL_OUT, &inet6sw[ip6_protox[IPPROTO_IPV6]].pr_pfh); # endif if (error) { SPL_X(s); return error; } # endif # endif # ifdef __sgi ipfilter_sgi_detach(); # endif appr_unload(); ipfr_unload(); ip_natunload(); fr_stateunload(); fr_authunload(); SPL_X(s); return 0; } #endif /* _KERNEL */ static int frzerostats(data) caddr_t data; { friostat_t fio; int error; fr_getstat(&fio); error = IWCOPYPTR((caddr_t)&fio, data, sizeof(fio)); if (error) return EFAULT; bzero((char *)frstats, sizeof(*frstats) * 2); return 0; } /* * Filter ioctl interface. */ #ifdef __sgi int IPL_EXTERN(ioctl)(dev_t dev, int cmd, caddr_t data, int mode # ifdef _KERNEL , cred_t *cp, int *rp # endif ) #else int IPL_EXTERN(ioctl)(dev, cmd, data, mode # if (defined(_KERNEL) && ((_BSDI_VERSION >= 199510) || (BSD >= 199506) || \ (NetBSD >= 199511) || (__FreeBSD_version >= 220000) || \ defined(__OpenBSD__))) , p) struct proc *p; # else ) # endif dev_t dev; # if defined(__NetBSD__) || defined(__OpenBSD__) || \ (_BSDI_VERSION >= 199701) || (__FreeBSD_version >= 300000) u_long cmd; # else int cmd; # endif caddr_t data; int mode; #endif /* __sgi */ { #if defined(_KERNEL) && !SOLARIS int s; #endif int error = 0, unit = 0, tmp; #if (BSD >= 199306) && defined(_KERNEL) if ((securelevel >= 2) && (mode & FWRITE)) return EPERM; #endif #ifdef _KERNEL unit = GET_MINOR(dev); if ((IPL_LOGMAX < unit) || (unit < 0)) return ENXIO; #else unit = dev; #endif SPL_NET(s); if (unit == IPL_LOGNAT) { /* * If we're doing a NAT operation, implicitly enable * IP Filter if not already enabled. We do this * because filtering and NAT are really separate * operations, and it's not entirely obvious from * a user's point of view that you need to enable * the filter in order to enable NAT. * * Since the default rule is to pass all packets, * this shouldn't cause any noticeable side-effects. */ if (fr_running == 0) { #if defined(_KERNEL) error = ipl_enable(); #else error = EIO; #endif } if (error == 0) error = nat_ioctl(data, cmd, mode); SPL_X(s); return error; } if (unit == IPL_LOGSTATE) { if (fr_running) error = fr_state_ioctl(data, cmd, mode); else error = EIO; SPL_X(s); return error; } if (unit == IPL_LOGAUTH) { if (!fr_running) error = EIO; else error = fr_auth_ioctl(data, mode, cmd, NULL, NULL); SPL_X(s); return error; } switch (cmd) { case FIONREAD : #ifdef IPFILTER_LOG error = IWCOPY((caddr_t)&iplused[IPL_LOGIPF], (caddr_t)data, sizeof(iplused[IPL_LOGIPF])); #endif break; #if (!defined(IPFILTER_LKM) || defined(__NetBSD__)) && defined(_KERNEL) case SIOCFRENB : { u_int enable; if (!(mode & FWRITE)) error = EPERM; else { error = IRCOPY(data, (caddr_t)&enable, sizeof(enable)); if (error) break; if (enable) # if defined(__NetBSD__) error = ipl_enable(); # else error = iplattach(); # endif else # if defined(__NetBSD__) error = ipl_disable(); # else error = ipldetach(); # endif } break; } #endif case SIOCSETFF : if (!(mode & FWRITE)) error = EPERM; else error = IRCOPY(data, (caddr_t)&fr_flags, sizeof(fr_flags)); break; case SIOCGETFF : error = IWCOPY((caddr_t)&fr_flags, data, sizeof(fr_flags)); break; case SIOCINAFR : case SIOCRMAFR : case SIOCADAFR : case SIOCZRLST : if (!(mode & FWRITE)) error = EPERM; else error = frrequest(unit, cmd, data, fr_active); break; case SIOCINIFR : case SIOCRMIFR : case SIOCADIFR : if (!(mode & FWRITE)) error = EPERM; else error = frrequest(unit, cmd, data, 1 - fr_active); break; case SIOCSWAPA : if (!(mode & FWRITE)) error = EPERM; else { bzero((char *)frcache, sizeof(frcache[0]) * 2); *(u_int *)data = fr_active; fr_active = 1 - fr_active; } break; case SIOCGETFS : { friostat_t fio; fr_getstat(&fio); error = IWCOPYPTR((caddr_t)&fio, data, sizeof(fio)); if (error) error = EFAULT; break; } case SIOCFRZST : if (!(mode & FWRITE)) error = EPERM; else error = frzerostats(data); break; case SIOCIPFFL : if (!(mode & FWRITE)) error = EPERM; else { error = IRCOPY(data, (caddr_t)&tmp, sizeof(tmp)); if (!error) { tmp = frflush(unit, tmp); error = IWCOPY((caddr_t)&tmp, data, sizeof(tmp)); } } break; case SIOCSTLCK : error = IRCOPY(data, (caddr_t)&tmp, sizeof(tmp)); if (!error) { fr_state_lock = tmp; fr_nat_lock = tmp; fr_frag_lock = tmp; fr_auth_lock = tmp; } else error = EFAULT; break; #ifdef IPFILTER_LOG case SIOCIPFFB : if (!(mode & FWRITE)) error = EPERM; else *(int *)data = ipflog_clear(unit); break; #endif /* IPFILTER_LOG */ case SIOCGFRST : error = IWCOPYPTR((caddr_t)ipfr_fragstats(), data, sizeof(ipfrstat_t)); if (error) error = EFAULT; break; case SIOCAUTHW : case SIOCAUTHR : if (!(mode & FWRITE)) { error = EPERM; break; } case SIOCFRSYN : if (!(mode & FWRITE)) error = EPERM; else { #if defined(_KERNEL) && defined(__sgi) ipfsync(); #endif frsync(); } break; default : error = EINVAL; break; } SPL_X(s); return error; } void fr_forgetifp(ifp) void *ifp; { register frentry_t *f; WRITE_ENTER(&ipf_mutex); for (f = ipacct[0][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; for (f = ipacct[1][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; for (f = ipfilter[0][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; for (f = ipfilter[1][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; #ifdef USE_INET6 for (f = ipacct6[0][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; for (f = ipacct6[1][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; for (f = ipfilter6[0][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; for (f = ipfilter6[1][fr_active]; (f != NULL); f = f->fr_next) if (f->fr_ifa == ifp) f->fr_ifa = (void *)-1; #endif RWLOCK_EXIT(&ipf_mutex); ip_natsync(ifp); } static int frrequest(unit, req, data, set) int unit; #if defined(__NetBSD__) || defined(__OpenBSD__) || (__FreeBSD_version >= 300003) u_long req; #else int req; #endif int set; caddr_t data; { register frentry_t *fp, *f, **fprev; register frentry_t **ftail; frgroup_t *fg = NULL; int error = 0, in, i; u_int *p, *pp; frentry_t frd; frdest_t *fdp; u_int group; fp = &frd; error = IRCOPYPTR(data, (caddr_t)fp, sizeof(*fp)); if (error) return EFAULT; fp->fr_ref = 0; #if (BSD >= 199306) && defined(_KERNEL) if ((securelevel > 0) && (fp->fr_func != NULL)) return EPERM; #endif /* * Check that the group number does exist and that if a head group * has been specified, doesn't exist. */ if ((req != SIOCZRLST) && fp->fr_grhead && fr_findgroup((u_int)fp->fr_grhead, fp->fr_flags, unit, set, NULL)) return EEXIST; if ((req != SIOCZRLST) && fp->fr_group && !fr_findgroup((u_int)fp->fr_group, fp->fr_flags, unit, set, NULL)) return ESRCH; in = (fp->fr_flags & FR_INQUE) ? 0 : 1; if (unit == IPL_LOGAUTH) ftail = fprev = &ipauth; else if ((fp->fr_flags & FR_ACCOUNT) && (fp->fr_v == 4)) ftail = fprev = &ipacct[in][set]; else if ((fp->fr_flags & (FR_OUTQUE|FR_INQUE)) && (fp->fr_v == 4)) ftail = fprev = &ipfilter[in][set]; #ifdef USE_INET6 else if ((fp->fr_flags & FR_ACCOUNT) && (fp->fr_v == 6)) ftail = fprev = &ipacct6[in][set]; else if ((fp->fr_flags & (FR_OUTQUE|FR_INQUE)) && (fp->fr_v == 6)) ftail = fprev = &ipfilter6[in][set]; #endif else return ESRCH; if ((group = fp->fr_group)) { if (!(fg = fr_findgroup(group, fp->fr_flags, unit, set, NULL))) return ESRCH; ftail = fprev = fg->fg_start; } bzero((char *)frcache, sizeof(frcache[0]) * 2); for (i = 0; i < 4; i++) { if ((fp->fr_ifnames[i][1] == '\0') && ((fp->fr_ifnames[i][0] == '-') || (fp->fr_ifnames[i][0] == '*'))) { fp->fr_ifas[i] = NULL; } else if (*fp->fr_ifnames[i]) { fp->fr_ifas[i] = GETUNIT(fp->fr_ifnames[i], fp->fr_v); if (!fp->fr_ifas[i]) fp->fr_ifas[i] = (void *)-1; } } fdp = &fp->fr_dif; fp->fr_flags &= ~FR_DUP; if (*fdp->fd_ifname) { fdp->fd_ifp = GETUNIT(fdp->fd_ifname, fp->fr_v); if (!fdp->fd_ifp) fdp->fd_ifp = (struct ifnet *)-1; else fp->fr_flags |= FR_DUP; } fdp = &fp->fr_tif; if (*fdp->fd_ifname) { fdp->fd_ifp = GETUNIT(fdp->fd_ifname, fp->fr_v); if (!fdp->fd_ifp) fdp->fd_ifp = (struct ifnet *)-1; } /* * Look for a matching filter rule, but don't include the next or * interface pointer in the comparison (fr_next, fr_ifa). */ for (fp->fr_cksum = 0, p = (u_int *)&fp->fr_ip, pp = &fp->fr_cksum; p < pp; p++) fp->fr_cksum += *p; for (; (f = *ftail); ftail = &f->fr_next) if ((fp->fr_cksum == f->fr_cksum) && !bcmp((char *)&f->fr_ip, (char *)&fp->fr_ip, FR_CMPSIZ)) break; /* * If zero'ing statistics, copy current to caller and zero. */ if (req == SIOCZRLST) { if (!f) return ESRCH; error = IWCOPYPTR((caddr_t)f, data, sizeof(*f)); if (error) return EFAULT; f->fr_hits = 0; f->fr_bytes = 0; return 0; } if (!f) { if (req != SIOCINAFR && req != SIOCINIFR) while ((f = *ftail)) ftail = &f->fr_next; else { if (fp->fr_hits) { ftail = fprev; while (--fp->fr_hits && (f = *ftail)) ftail = &f->fr_next; } f = NULL; } } if (req == SIOCRMAFR || req == SIOCRMIFR) { if (!f) error = ESRCH; else { /* * Only return EBUSY if there is a group list, else * it's probably just state information referencing * the rule. */ if ((f->fr_ref > 1) && f->fr_grp) return EBUSY; if (fg && fg->fg_head) fg->fg_head->fr_ref--; if (f->fr_grhead) fr_delgroup((u_int)f->fr_grhead, fp->fr_flags, unit, set); fixskip(fprev, f, -1); *ftail = f->fr_next; f->fr_next = NULL; f->fr_ref--; if (f->fr_ref == 0) KFREE(f); } } else { if (f) error = EEXIST; else { KMALLOC(f, frentry_t *); if (f != NULL) { if (fg && fg->fg_head) fg->fg_head->fr_ref++; bcopy((char *)fp, (char *)f, sizeof(*f)); f->fr_ref = 1; f->fr_hits = 0; f->fr_next = *ftail; *ftail = f; if (req == SIOCINIFR || req == SIOCINAFR) fixskip(fprev, f, 1); f->fr_grp = NULL; if ((group = f->fr_grhead)) fg = fr_addgroup(group, f, unit, set); } else error = ENOMEM; } } return (error); } #ifdef _KERNEL /* * routines below for saving IP headers to buffer */ # ifdef __sgi # ifdef _KERNEL int IPL_EXTERN(open)(dev_t *pdev, int flags, int devtype, cred_t *cp) # else int IPL_EXTERN(open)(dev_t dev, int flags) # endif # else int IPL_EXTERN(open)(dev, flags # if ((_BSDI_VERSION >= 199510) || (BSD >= 199506) || (NetBSD >= 199511) || \ (__FreeBSD_version >= 220000) || defined(__OpenBSD__)) && defined(_KERNEL) , devtype, p) int devtype; struct proc *p; # else ) # endif dev_t dev; int flags; # endif /* __sgi */ { # if defined(__sgi) && defined(_KERNEL) u_int min = geteminor(*pdev); # else u_int min = GET_MINOR(dev); # endif if (IPL_LOGMAX < min) min = ENXIO; else min = 0; return min; } # ifdef __sgi int IPL_EXTERN(close)(dev_t dev, int flags, int devtype, cred_t *cp) #else int IPL_EXTERN(close)(dev, flags # if ((_BSDI_VERSION >= 199510) || (BSD >= 199506) || (NetBSD >= 199511) || \ (__FreeBSD_version >= 220000) || defined(__OpenBSD__)) && defined(_KERNEL) , devtype, p) int devtype; struct proc *p; # else ) # endif dev_t dev; int flags; # endif /* __sgi */ { u_int min = GET_MINOR(dev); if (IPL_LOGMAX < min) min = ENXIO; else min = 0; return min; } /* * iplread/ipllog * both of these must operate with at least splnet() lest they be * called during packet processing and cause an inconsistancy to appear in * the filter lists. */ # ifdef __sgi int IPL_EXTERN(read)(dev_t dev, uio_t *uio, cred_t *crp) # else # if BSD >= 199306 int IPL_EXTERN(read)(dev, uio, ioflag) int ioflag; # else int IPL_EXTERN(read)(dev, uio) # endif dev_t dev; register struct uio *uio; # endif /* __sgi */ { # ifdef IPFILTER_LOG return ipflog_read(GET_MINOR(dev), uio); # else return ENXIO; # endif } /* * send_reset - this could conceivably be a call to tcp_respond(), but that * requires a large amount of setting up and isn't any more efficient. */ int send_reset(oip, fin) struct ip *oip; fr_info_t *fin; { struct tcphdr *tcp, *tcp2; int tlen = 0, hlen; struct mbuf *m; #ifdef USE_INET6 ip6_t *ip6, *oip6 = (ip6_t *)oip; #endif ip_t *ip; tcp = (struct tcphdr *)fin->fin_dp; if (tcp->th_flags & TH_RST) return -1; /* feedback loop */ # if (BSD < 199306) || defined(__sgi) m = m_get(M_DONTWAIT, MT_HEADER); # else m = m_gethdr(M_DONTWAIT, MT_HEADER); # endif if (m == NULL) return ENOBUFS; if (m == NULL) return -1; tlen = fin->fin_dlen - (tcp->th_off << 2) + ((tcp->th_flags & TH_SYN) ? 1 : 0) + ((tcp->th_flags & TH_FIN) ? 1 : 0); #ifdef USE_INET6 hlen = (fin->fin_v == 6) ? sizeof(ip6_t) : sizeof(ip_t); #else hlen = sizeof(ip_t); #endif m->m_len = sizeof(*tcp2) + hlen; # if BSD >= 199306 m->m_data += max_linkhdr; m->m_pkthdr.len = m->m_len; m->m_pkthdr.rcvif = (struct ifnet *)0; # endif ip = mtod(m, struct ip *); # ifdef USE_INET6 ip6 = (ip6_t *)ip; # endif bzero((char *)ip, sizeof(*tcp2) + hlen); tcp2 = (struct tcphdr *)((char *)ip + hlen); tcp2->th_sport = tcp->th_dport; tcp2->th_dport = tcp->th_sport; if (tcp->th_flags & TH_ACK) { tcp2->th_seq = tcp->th_ack; tcp2->th_flags = TH_RST; } else { tcp2->th_ack = ntohl(tcp->th_seq); tcp2->th_ack += tlen; tcp2->th_ack = htonl(tcp2->th_ack); tcp2->th_flags = TH_RST|TH_ACK; } tcp2->th_off = sizeof(*tcp2) >> 2; # ifdef USE_INET6 if (fin->fin_v == 6) { ip6->ip6_plen = htons(sizeof(struct tcphdr)); ip6->ip6_nxt = IPPROTO_TCP; ip6->ip6_src = oip6->ip6_dst; ip6->ip6_dst = oip6->ip6_src; tcp2->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(*ip6), sizeof(*tcp2)); return send_ip(oip, fin, &m); } # endif ip->ip_p = IPPROTO_TCP; ip->ip_len = htons(sizeof(struct tcphdr)); ip->ip_src.s_addr = oip->ip_dst.s_addr; ip->ip_dst.s_addr = oip->ip_src.s_addr; tcp2->th_sum = in_cksum(m, hlen + sizeof(*tcp2)); ip->ip_len = hlen + sizeof(*tcp2); return send_ip(oip, fin, &m); } /* * Send an IP(v4/v6) datagram out into the network */ static int send_ip(oip, fin, mp) ip_t *oip; fr_info_t *fin; struct mbuf **mp; { struct mbuf *m = *mp; char *dpsave; int error; ip_t *ip; dpsave = fin->fin_dp; ip = mtod(m, ip_t *); ip->ip_v = fin->fin_v; if (ip->ip_v == 4) { ip->ip_hl = (sizeof(*oip) >> 2); ip->ip_v = IPVERSION; ip->ip_tos = oip->ip_tos; ip->ip_id = oip->ip_id; ip->ip_off = 0; # if (BSD < 199306) || defined(__sgi) ip->ip_ttl = tcp_ttl; # else ip->ip_ttl = ip_defttl; # endif ip->ip_sum = 0; fin->fin_dp = (char *)(ip + 1); } # ifdef USE_INET6 else if (ip->ip_v == 6) { ip6_t *ip6 = (ip6_t *)ip; ip6->ip6_hlim = 127; fin->fin_dp = (char *)(ip6 + 1); } # endif # ifdef IPSEC m->m_pkthdr.rcvif = NULL; # endif error = ipfr_fastroute(m, mp, fin, NULL); fin->fin_dp = dpsave; return error; } int send_icmp_err(oip, type, fin, dst) ip_t *oip; int type; fr_info_t *fin; int dst; { int err, hlen = 0, xtra = 0, iclen, ohlen = 0, avail, code; u_short shlen, slen = 0, soff = 0; struct in_addr dst4; struct icmp *icmp; struct mbuf *m; void *ifp; #ifdef USE_INET6 ip6_t *ip6, *oip6 = (ip6_t *)oip; struct in6_addr dst6; #endif ip_t *ip; if ((type < 0) || (type > ICMP_MAXTYPE)) return -1; code = fin->fin_icode; #ifdef USE_INET6 if ((code < 0) || (code > sizeof(icmptoicmp6unreach)/sizeof(int))) return -1; #endif avail = 0; m = NULL; ifp = fin->fin_ifp; if (fin->fin_v == 4) { if ((oip->ip_p == IPPROTO_ICMP) && !(fin->fin_fi.fi_fl & FI_SHORT)) switch (ntohs(fin->fin_data[0]) >> 8) { case ICMP_ECHO : case ICMP_TSTAMP : case ICMP_IREQ : case ICMP_MASKREQ : break; default : return 0; } # if (BSD < 199306) || defined(__sgi) avail = MLEN; m = m_get(M_DONTWAIT, MT_HEADER); # else avail = MHLEN; m = m_gethdr(M_DONTWAIT, MT_HEADER); # endif if (m == NULL) return ENOBUFS; if (dst == 0) { if (fr_ifpaddr(4, ifp, &dst4) == -1) return -1; } else dst4.s_addr = oip->ip_dst.s_addr; hlen = sizeof(ip_t); ohlen = oip->ip_hl << 2; xtra = 8; } #ifdef USE_INET6 else if (fin->fin_v == 6) { hlen = sizeof(ip6_t); ohlen = sizeof(ip6_t); type = icmptoicmp6types[type]; if (type == ICMP6_DST_UNREACH) code = icmptoicmp6unreach[code]; MGETHDR(m, M_DONTWAIT, MT_HEADER); if (!m) return ENOBUFS; MCLGET(m, M_DONTWAIT); if ((m->m_flags & M_EXT) == 0) { m_freem(m); return ENOBUFS; } # ifdef M_TRAILINGSPACE m->m_len = 0; avail = M_TRAILINGSPACE(m); # else avail = (m->m_flags & M_EXT) ? MCLBYTES : MHLEN; # endif xtra = MIN(ntohs(oip6->ip6_plen) + sizeof(ip6_t), avail - hlen - sizeof(*icmp) - max_linkhdr); if (dst == 0) { if (fr_ifpaddr(6, ifp, (struct in_addr *)&dst6) == -1) return -1; } else dst6 = oip6->ip6_dst; } #endif iclen = hlen + sizeof(*icmp); # if BSD >= 199306 avail -= (max_linkhdr + iclen); m->m_data += max_linkhdr; m->m_pkthdr.rcvif = (struct ifnet *)0; if (xtra > avail) xtra = avail; iclen += xtra; m->m_pkthdr.len = iclen; #else avail -= (m->m_off + iclen); if (xtra > avail) xtra = avail; iclen += xtra; #endif m->m_len = iclen; ip = mtod(m, ip_t *); icmp = (struct icmp *)((char *)ip + hlen); bzero((char *)ip, iclen); icmp->icmp_type = type; icmp->icmp_code = fin->fin_icode; icmp->icmp_cksum = 0; #ifdef icmp_nextmtu if (type == ICMP_UNREACH && fin->fin_icode == ICMP_UNREACH_NEEDFRAG && ifp) icmp->icmp_nextmtu = htons(((struct ifnet *) ifp)->if_mtu); #endif if (avail) { bcopy((char *)oip, (char *)&icmp->icmp_ip, MIN(ohlen, avail)); avail -= MIN(ohlen, avail); } #ifdef USE_INET6 ip6 = (ip6_t *)ip; if (fin->fin_v == 6) { ip6->ip6_flow = 0; ip6->ip6_plen = htons(iclen - hlen); ip6->ip6_nxt = IPPROTO_ICMPV6; ip6->ip6_hlim = 0; ip6->ip6_src = dst6; ip6->ip6_dst = oip6->ip6_src; if (avail) bcopy((char *)oip + ohlen, (char *)&icmp->icmp_ip + ohlen, avail); icmp->icmp_cksum = in6_cksum(m, IPPROTO_ICMPV6, sizeof(*ip6), iclen - hlen); } else #endif { slen = oip->ip_len; oip->ip_len = htons(oip->ip_len); soff = oip->ip_off; oip->ip_off = htons(ip->ip_off); ip->ip_src.s_addr = dst4.s_addr; ip->ip_dst.s_addr = oip->ip_src.s_addr; if (avail > 8) avail = 8; if (avail) bcopy((char *)oip + ohlen, (char *)&icmp->icmp_ip + ohlen, avail); icmp->icmp_cksum = ipf_cksum((u_short *)icmp, sizeof(*icmp) + 8); ip->ip_len = iclen; ip->ip_p = IPPROTO_ICMP; } shlen = fin->fin_hlen; fin->fin_hlen = hlen; err = send_ip(oip, fin, &m); fin->fin_hlen = shlen; #ifdef USE_INET6 if (fin->fin_v == 4) #endif { oip->ip_len = slen; oip->ip_off = soff; } return err; } # if !defined(IPFILTER_LKM) && !defined(__sgi) && \ (!defined(__FreeBSD_version) || (__FreeBSD_version < 300000)) # if (BSD < 199306) int iplinit __P((void)); int # else void iplinit __P((void)); void # endif iplinit() { # if defined(__NetBSD__) if (ipl_enable() != 0) # else if (iplattach() != 0) # endif { printf("IP Filter failed to attach\n"); } ip_init(); } # endif /* ! __NetBSD__ */ /* * Return the length of the entire mbuf. */ size_t mbufchainlen(m0) register struct mbuf *m0; { #if BSD >= 199306 return m0->m_pkthdr.len; #else register size_t len = 0; for (; m0; m0 = m0->m_next) len += m0->m_len; return len; #endif } int ipfr_fastroute(m0, mpp, fin, fdp) struct mbuf *m0, **mpp; fr_info_t *fin; frdest_t *fdp; { register struct ip *ip, *mhip; register struct mbuf *m = m0; register struct route *ro; int len, off, error = 0, hlen, code; struct ifnet *ifp, *sifp; struct sockaddr_in *dst; struct route iproute; frentry_t *fr; ip = NULL; ro = NULL; ifp = NULL; ro = &iproute; ro->ro_rt = NULL; #ifdef USE_INET6 if (fin->fin_v == 6) { error = ipfr_fastroute6(m0, mpp, fin, fdp); if (error != 0) goto bad; goto done; } #else if (fin->fin_v == 6) goto bad; #endif #ifdef M_WRITABLE /* * HOT FIX/KLUDGE: * * If the mbuf we're about to send is not writable (because of * a cluster reference, for example) we'll need to make a copy * of it since this routine modifies the contents. * * If you have non-crappy network hardware that can transmit data * from the mbuf, rather than making a copy, this is gonna be a * problem. */ if (M_WRITABLE(m) == 0) { if ((m0 = m_dup(m, M_DONTWAIT)) != NULL) { m_freem(*mpp); *mpp = m0; m = m0; } else { error = ENOBUFS; m_freem(*mpp); goto done; } } #endif hlen = fin->fin_hlen; ip = mtod(m0, struct ip *); #if defined(__NetBSD__) && defined(M_CSUM_IPv4) /* * Clear any in-bound checksum flags for this packet. */ # if (__NetBSD_Version__ > 105009999) m0->m_pkthdr.csum_flags = 0; # else m0->m_pkthdr.csuminfo = 0; # endif #endif /* __NetBSD__ && M_CSUM_IPv4 */ /* * Route packet. */ bzero((caddr_t)ro, sizeof (*ro)); dst = (struct sockaddr_in *)&ro->ro_dst; dst->sin_family = AF_INET; dst->sin_addr = ip->ip_dst; fr = fin->fin_fr; if (fdp != NULL) ifp = fdp->fd_ifp; else ifp = fin->fin_ifp; /* * In case we're here due to "to " being used with "keep state", * check that we're going in the correct direction. */ if ((fr != NULL) && (fin->fin_rev != 0)) { if ((ifp != NULL) && (fdp == &fr->fr_tif)) return 0; } else if (fdp != NULL) { if (fdp->fd_ip.s_addr != 0) dst->sin_addr = fdp->fd_ip; } # if BSD >= 199306 dst->sin_len = sizeof(*dst); # endif # if (BSD >= 199306) && !defined(__NetBSD__) && !defined(__bsdi__) && \ !defined(__OpenBSD__) # ifdef RTF_CLONING rtalloc_ign(ro, RTF_CLONING); # else rtalloc_ign(ro, RTF_PRCLONING); # endif # else rtalloc(ro); # endif if (!ifp) { if (!fr || !(fr->fr_flags & FR_FASTROUTE)) { error = -2; goto bad; } } if ((ifp == NULL) && (ro->ro_rt != NULL)) ifp = ro->ro_rt->rt_ifp; if ((ro->ro_rt == NULL) || (ifp == NULL)) { if (in_localaddr(ip->ip_dst)) error = EHOSTUNREACH; else error = ENETUNREACH; goto bad; } if (ro->ro_rt->rt_flags & RTF_GATEWAY) { #if BSD >= 199306 dst = (struct sockaddr_in *)ro->ro_rt->rt_gateway; #else dst = (struct sockaddr_in *)&ro->ro_rt->rt_gateway; #endif } ro->ro_rt->rt_use++; /* * For input packets which are being "fastrouted", they won't * go back through output filtering and miss their chance to get * NAT'd and counted. */ if (fin->fin_out == 0) { sifp = fin->fin_ifp; fin->fin_ifp = ifp; fin->fin_out = 1; if ((fin->fin_fr = ipacct[1][fr_active]) && (fr_scanlist(FR_NOMATCH, ip, fin, m) & FR_ACCOUNT)) { ATOMIC_INCL(frstats[1].fr_acct); } fin->fin_fr = NULL; if (!fr || !(fr->fr_flags & FR_RETMASK)) (void) fr_checkstate(ip, fin); (void) ip_natout(ip, fin); fin->fin_ifp = sifp; } else ip->ip_sum = 0; /* * If small enough for interface, can just send directly. */ if (ip->ip_len <= ifp->if_mtu) { # ifndef sparc # if (!defined(__FreeBSD__) && !(_BSDI_VERSION >= 199510)) ip->ip_id = htons(ip->ip_id); # endif ip->ip_len = htons(ip->ip_len); ip->ip_off = htons(ip->ip_off); # endif # if defined(__NetBSD__) && defined(M_CSUM_IPv4) # if (__NetBSD_Version__ > 105009999) if (ifp->if_csum_flags_tx & IFCAP_CSUM_IPv4) m->m_pkthdr.csum_flags |= M_CSUM_IPv4; else if (ip->ip_sum == 0) ip->ip_sum = in_cksum(m, hlen); # else if (ifp->if_capabilities & IFCAP_CSUM_IPv4) m->m_pkthdr.csuminfo |= M_CSUM_IPv4; else if (ip->ip_sum == 0) ip->ip_sum = in_cksum(m, hlen); # endif # else if (!ip->ip_sum) ip->ip_sum = in_cksum(m, hlen); # endif /* __NetBSD__ && M_CSUM_IPv4 */ # if (BSD >= 199306) || (defined(IRIX) && (IRIX >= 605)) error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst, ro->ro_rt); # else error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst); # endif goto done; } /* * Too large for interface; fragment if possible. * Must be able to put at least 8 bytes per fragment. */ if (ip->ip_off & IP_DF) { error = EMSGSIZE; goto bad; } len = (ifp->if_mtu - hlen) &~ 7; if (len < 8) { error = EMSGSIZE; goto bad; } { int mhlen, firstlen = len; struct mbuf **mnext = &m->m_act; /* * Loop through length of segment after first fragment, * make new header and copy data of each part and link onto chain. */ m0 = m; mhlen = sizeof (struct ip); for (off = hlen + len; off < ip->ip_len; off += len) { # ifdef MGETHDR MGETHDR(m, M_DONTWAIT, MT_HEADER); # else MGET(m, M_DONTWAIT, MT_HEADER); # endif if (m == 0) { error = ENOBUFS; goto bad; } # if BSD >= 199306 m->m_data += max_linkhdr; # else m->m_off = MMAXOFF - hlen; # endif mhip = mtod(m, struct ip *); bcopy((char *)ip, (char *)mhip, sizeof(*ip)); if (hlen > sizeof (struct ip)) { mhlen = ip_optcopy(ip, mhip) + sizeof (struct ip); mhip->ip_hl = mhlen >> 2; } m->m_len = mhlen; mhip->ip_off = ((off - hlen) >> 3) + (ip->ip_off & ~IP_MF); if (ip->ip_off & IP_MF) mhip->ip_off |= IP_MF; if (off + len >= ip->ip_len) len = ip->ip_len - off; else mhip->ip_off |= IP_MF; mhip->ip_len = htons((u_short)(len + mhlen)); m->m_next = m_copy(m0, off, len); if (m->m_next == 0) { error = ENOBUFS; /* ??? */ goto sendorfree; } # if BSD >= 199306 m->m_pkthdr.len = mhlen + len; m->m_pkthdr.rcvif = NULL; # endif mhip->ip_off = htons((u_short)mhip->ip_off); mhip->ip_sum = 0; mhip->ip_sum = in_cksum(m, mhlen); *mnext = m; mnext = &m->m_act; } /* * Update first fragment by trimming what's been copied out * and updating header, then send each fragment (in order). */ m_adj(m0, hlen + firstlen - ip->ip_len); ip->ip_len = htons((u_short)(hlen + firstlen)); ip->ip_off = htons((u_short)(ip->ip_off | IP_MF)); ip->ip_sum = 0; ip->ip_sum = in_cksum(m0, hlen); sendorfree: for (m = m0; m; m = m0) { m0 = m->m_act; m->m_act = 0; if (error == 0) # if (BSD >= 199306) || (defined(IRIX) && (IRIX >= 605)) error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst, ro->ro_rt); # else error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst); # endif else m_freem(m); } } done: if (!error) ipl_frouteok[0]++; else ipl_frouteok[1]++; if (ro->ro_rt != NULL) { RTFREE(ro->ro_rt); } *mpp = NULL; return error; bad: if ((error == EMSGSIZE) && (fin->fin_v == 4)) { sifp = fin->fin_ifp; code = fin->fin_icode; fin->fin_icode = ICMP_UNREACH_NEEDFRAG; fin->fin_ifp = ifp; (void) send_icmp_err(ip, ICMP_UNREACH, fin, 1); fin->fin_ifp = sifp; fin->fin_icode = code; } m_freem(m); goto done; } /* * Return true or false depending on whether the route to the * given IP address uses the same interface as the one passed. */ int fr_verifysrc(ipa, ifp) struct in_addr ipa; void *ifp; { struct sockaddr_in *dst; struct route iproute; bzero((char *)&iproute, sizeof(iproute)); dst = (struct sockaddr_in *)&iproute.ro_dst; # if (BSD >= 199306) dst->sin_len = sizeof(*dst); # endif dst->sin_family = AF_INET; dst->sin_addr = ipa; # if (BSD >= 199306) && !defined(__NetBSD__) && !defined(__bsdi__) && \ !defined(__OpenBSD__) # ifdef RTF_CLONING rtalloc_ign(&iproute, RTF_CLONING); # else rtalloc_ign(&iproute, RTF_PRCLONING); # endif # else rtalloc(&iproute); # endif if (iproute.ro_rt == NULL) return 0; return (ifp == iproute.ro_rt->rt_ifp); } # ifdef USE_GETIFNAME char * get_ifname(ifp) struct ifnet *ifp; { static char workbuf[64]; sprintf(workbuf, "%s%d", ifp->if_name, ifp->if_unit); return workbuf; } # endif # if defined(USE_INET6) /* * This is the IPv6 specific fastroute code. It doesn't clean up the mbuf's * or ensure that it is an IPv6 packet that is being forwarded, those are * expected to be done by the called (ipfr_fastroute). */ static int ipfr_fastroute6(m0, mpp, fin, fdp) struct mbuf *m0, **mpp; fr_info_t *fin; frdest_t *fdp; { struct route_in6 ip6route; struct sockaddr_in6 *dst6; struct route_in6 *ro; struct ifnet *ifp; frentry_t *fr; int error; ifp = NULL; ro = &ip6route; fr = fin->fin_fr; bzero((caddr_t)ro, sizeof(*ro)); dst6 = (struct sockaddr_in6 *)&ro->ro_dst; dst6->sin6_family = AF_INET6; dst6->sin6_len = sizeof(struct sockaddr_in6); dst6->sin6_addr = fin->fin_fi.fi_src.in6; if (fdp != NULL) ifp = fdp->fd_ifp; if ((fr != NULL) && (fin->fin_rev != 0)) { if ((ifp != NULL) && (fdp == &fr->fr_tif)) return 0; } else if (fdp != NULL) { if (IP6_NOTZERO(&fdp->fd_ip6)) dst6->sin6_addr = fdp->fd_ip6.in6; } if ((ifp == NULL) && ((fr == NULL) || !(fr->fr_flags & FR_FASTROUTE))) return -2; rtalloc((struct route *)ro); if ((ifp == NULL) && (ro->ro_rt != NULL)) ifp = ro->ro_rt->rt_ifp; if ((ro->ro_rt == NULL) || (ifp == NULL) || (ifp != ro->ro_rt->rt_ifp)) { error = EHOSTUNREACH; } else { if (ro->ro_rt->rt_flags & RTF_GATEWAY) dst6 = (struct sockaddr_in6 *)ro->ro_rt->rt_gateway; ro->ro_rt->rt_use++; if (m0->m_pkthdr.len <= IN6_LINKMTU(ifp)) error = nd6_output(ifp, fin->fin_ifp, m0, dst6, ro->ro_rt); else error = EMSGSIZE; } if (ro->ro_rt != NULL) { RTFREE(ro->ro_rt); } return error; } # endif #else /* #ifdef _KERNEL */ # if defined(__sgi) && (IRIX < 605) static int no_output __P((struct ifnet *ifp, struct mbuf *m, struct sockaddr *s)) # else static int no_output __P((struct ifnet *ifp, struct mbuf *m, struct sockaddr *s, struct rtentry *rt)) # endif { return 0; } # ifdef __STDC__ # if defined(__sgi) && (IRIX < 605) static int write_output __P((struct ifnet *ifp, struct mbuf *m, struct sockaddr *s)) # else static int write_output __P((struct ifnet *ifp, struct mbuf *m, struct sockaddr *s, struct rtentry *rt)) # endif { ip_t *ip = (ip_t *)m; # else static int write_output(ifp, ip) struct ifnet *ifp; ip_t *ip; { # endif char fname[32]; int fd; # if (defined(NetBSD) && (NetBSD <= 1991011) && (NetBSD >= 199606)) || \ (defined(OpenBSD) && (OpenBSD >= 199603)) sprintf(fname, "%s", ifp->if_xname); # else sprintf(fname, "%s%d", ifp->if_name, ifp->if_unit); # endif fd = open(fname, O_WRONLY|O_APPEND); if (fd == -1) { perror("open"); return -1; } write(fd, (char *)ip, ntohs(ip->ip_len)); close(fd); return 0; } char *get_ifname(ifp) struct ifnet *ifp; { # if (defined(NetBSD) && (NetBSD <= 1991011) && (NetBSD >= 199606)) || \ (defined(OpenBSD) && (OpenBSD >= 199603)) return ifp->if_xname; # else static char fullifname[LIFNAMSIZ]; sprintf(fullifname, "%s%d", ifp->if_name, ifp->if_unit); return fullifname; # endif } struct ifnet *get_unit(ifname, v) char *ifname; int v; { struct ifnet *ifp, **ifa, **old_ifneta; for (ifa = ifneta; ifa && (ifp = *ifa); ifa++) { # if (defined(NetBSD) && (NetBSD <= 1991011) && (NetBSD >= 199606)) || \ (defined(OpenBSD) && (OpenBSD >= 199603)) if (!strncmp(ifname, ifp->if_xname, sizeof(ifp->if_xname))) # else char fullname[LIFNAMSIZ]; sprintf(fullname, "%s%d", ifp->if_name, ifp->if_unit); if (!strcmp(ifname, fullname)) # endif return ifp; } if (!ifneta) { ifneta = (struct ifnet **)malloc(sizeof(ifp) * 2); if (!ifneta) return NULL; ifneta[1] = NULL; ifneta[0] = (struct ifnet *)calloc(1, sizeof(*ifp)); if (!ifneta[0]) { free(ifneta); return NULL; } nifs = 1; } else { old_ifneta = ifneta; nifs++; ifneta = (struct ifnet **)realloc(ifneta, (nifs + 1) * sizeof(*ifa)); if (!ifneta) { free(old_ifneta); nifs = 0; return NULL; } ifneta[nifs] = NULL; ifneta[nifs - 1] = (struct ifnet *)malloc(sizeof(*ifp)); if (!ifneta[nifs - 1]) { nifs--; return NULL; } } ifp = ifneta[nifs - 1]; # if (defined(NetBSD) && (NetBSD <= 1991011) && (NetBSD >= 199606)) || \ (defined(OpenBSD) && (OpenBSD >= 199603)) strncpy(ifp->if_xname, ifname, sizeof(ifp->if_xname)); # else ifp->if_name = strdup(ifname); ifname = ifp->if_name; while (*ifname && !isdigit(*ifname)) ifname++; if (*ifname && isdigit(*ifname)) { ifp->if_unit = atoi(ifname); *ifname = '\0'; } else ifp->if_unit = -1; # endif ifp->if_output = no_output; return ifp; } void init_ifp() { struct ifnet *ifp, **ifa; char fname[32]; int fd; # if (defined(NetBSD) && (NetBSD <= 1991011) && (NetBSD >= 199606)) || \ (defined(OpenBSD) && (OpenBSD >= 199603)) for (ifa = ifneta; ifa && (ifp = *ifa); ifa++) { ifp->if_output = write_output; sprintf(fname, "/tmp/%s", ifp->if_xname); fd = open(fname, O_WRONLY|O_CREAT|O_EXCL|O_TRUNC, 0600); if (fd == -1) perror("open"); else close(fd); } # else for (ifa = ifneta; ifa && (ifp = *ifa); ifa++) { ifp->if_output = write_output; sprintf(fname, "/tmp/%s%d", ifp->if_name, ifp->if_unit); fd = open(fname, O_WRONLY|O_CREAT|O_EXCL|O_TRUNC, 0600); if (fd == -1) perror("open"); else close(fd); } # endif } int send_reset(ip, fin) ip_t *ip; fr_info_t *fin; { verbose("- TCP RST sent\n"); return 0; } int send_icmp_err(ip, code, fin, dst) ip_t *ip; int code; fr_info_t *fin; int dst; { verbose("- ICMP UNREACHABLE RST sent\n"); return 0; } void frsync() { return; } void m_copydata(m, off, len, cp) mb_t *m; int off, len; caddr_t cp; { bcopy((char *)m + off, cp, len); } int ipfuiomove(buf, len, rwflag, uio) caddr_t buf; int len, rwflag; struct uio *uio; { int left, ioc, num, offset; struct iovec *io; char *start; if (rwflag == UIO_READ) { left = len; ioc = 0; offset = uio->uio_offset; while ((left > 0) && (ioc < uio->uio_iovcnt)) { io = uio->uio_iov + ioc; num = io->iov_len; if (num > left) num = left; start = (char *)io->iov_base + offset; if (start > (char *)io->iov_base + io->iov_len) { offset -= io->iov_len; ioc++; continue; } bcopy(buf, start, num); uio->uio_resid -= num; uio->uio_offset += num; left -= num; if (left > 0) ioc++; } if (left > 0) return EFAULT; } return 0; } #endif /* _KERNEL */