/* * Copyright (c) 1980, 1986, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by the University of * California, Berkeley and its contributors. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)route.c 8.2 (Berkeley) 11/15/93 * $Id: route.c,v 1.27 1995/10/29 15:32:14 phk Exp $ */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define SA(p) ((struct sockaddr *)(p)) struct route_cb route_cb; struct rtstat rtstat; struct radix_node_head *rt_tables[AF_MAX+1]; int rttrash; /* routes not in table but not freed */ struct sockaddr wildcard; /* zero valued cookie for wildcard searches */ void rtable_init(table) void **table; { struct domain *dom; for (dom = domains; dom; dom = dom->dom_next) if (dom->dom_rtattach) dom->dom_rtattach(&table[dom->dom_family], dom->dom_rtoffset); } void route_init() { rn_init(); /* initialize all zeroes, all ones, mask table */ rtable_init((void **)rt_tables); } /* * Packet routing routines. */ void rtalloc(ro) register struct route *ro; { if (ro->ro_rt && ro->ro_rt->rt_ifp && (ro->ro_rt->rt_flags & RTF_UP)) return; /* XXX */ ro->ro_rt = rtalloc1(&ro->ro_dst, 1, 0UL); } void rtalloc_ign(ro, ignore) register struct route *ro; u_long ignore; { if (ro->ro_rt && ro->ro_rt->rt_ifp && (ro->ro_rt->rt_flags & RTF_UP)) return; /* XXX */ ro->ro_rt = rtalloc1(&ro->ro_dst, 1, ignore); } struct rtentry * rtalloc1(dst, report, ignflags) register struct sockaddr *dst; int report; u_long ignflags; { register struct radix_node_head *rnh = rt_tables[dst->sa_family]; register struct rtentry *rt; register struct radix_node *rn; struct rtentry *newrt = 0; struct rt_addrinfo info; u_long nflags; int s = splnet(), err = 0, msgtype = RTM_MISS; if (rnh && (rn = rnh->rnh_matchaddr((caddr_t)dst, rnh)) && ((rn->rn_flags & RNF_ROOT) == 0)) { newrt = rt = (struct rtentry *)rn; nflags = rt->rt_flags & ~ignflags; if (report && (nflags & (RTF_CLONING | RTF_PRCLONING))) { err = rtrequest(RTM_RESOLVE, dst, SA(0), SA(0), 0, &newrt); if (err) { newrt = rt; rt->rt_refcnt++; goto miss; } if ((rt = newrt) && (rt->rt_flags & RTF_XRESOLVE)) { msgtype = RTM_RESOLVE; goto miss; } } else rt->rt_refcnt++; } else { rtstat.rts_unreach++; miss: if (report) { bzero((caddr_t)&info, sizeof(info)); info.rti_info[RTAX_DST] = dst; rt_missmsg(msgtype, &info, 0, err); } } splx(s); return (newrt); } void rtfree(rt) register struct rtentry *rt; { register struct radix_node_head *rnh = rt_tables[rt_key(rt)->sa_family]; register struct ifaddr *ifa; if (rt == 0 || rnh == 0) panic("rtfree"); rt->rt_refcnt--; if(rnh->rnh_close && rt->rt_refcnt == 0) { rnh->rnh_close((struct radix_node *)rt, rnh); } if (rt->rt_refcnt <= 0 && (rt->rt_flags & RTF_UP) == 0) { if (rt->rt_nodes->rn_flags & (RNF_ACTIVE | RNF_ROOT)) panic ("rtfree 2"); rttrash--; if (rt->rt_refcnt < 0) { printf("rtfree: %p not freed (neg refs)\n", rt); return; } ifa = rt->rt_ifa; IFAFREE(ifa); if (rt->rt_parent) { RTFREE(rt->rt_parent); } Free(rt_key(rt)); Free(rt); } } void ifafree(ifa) register struct ifaddr *ifa; { if (ifa == NULL) panic("ifafree"); if (ifa->ifa_refcnt == 0) free(ifa, M_IFADDR); else ifa->ifa_refcnt--; } /* * Force a routing table entry to the specified * destination to go through the given gateway. * Normally called as a result of a routing redirect * message from the network layer. * * N.B.: must be called at splnet * */ void rtredirect(dst, gateway, netmask, flags, src, rtp) struct sockaddr *dst, *gateway, *netmask, *src; int flags; struct rtentry **rtp; { register struct rtentry *rt; int error = 0; short *stat = 0; struct rt_addrinfo info; struct ifaddr *ifa; /* verify the gateway is directly reachable */ if ((ifa = ifa_ifwithnet(gateway)) == 0) { error = ENETUNREACH; goto out; } rt = rtalloc1(dst, 0, 0UL); /* * If the redirect isn't from our current router for this dst, * it's either old or wrong. If it redirects us to ourselves, * we have a routing loop, perhaps as a result of an interface * going down recently. */ #define equal(a1, a2) (bcmp((caddr_t)(a1), (caddr_t)(a2), (a1)->sa_len) == 0) if (!(flags & RTF_DONE) && rt && (!equal(src, rt->rt_gateway) || rt->rt_ifa != ifa)) error = EINVAL; else if (ifa_ifwithaddr(gateway)) error = EHOSTUNREACH; if (error) goto done; /* * Create a new entry if we just got back a wildcard entry * or the the lookup failed. This is necessary for hosts * which use routing redirects generated by smart gateways * to dynamically build the routing tables. */ if ((rt == 0) || (rt_mask(rt) && rt_mask(rt)->sa_len < 2)) goto create; /* * Don't listen to the redirect if it's * for a route to an interface. */ if (rt->rt_flags & RTF_GATEWAY) { if (((rt->rt_flags & RTF_HOST) == 0) && (flags & RTF_HOST)) { /* * Changing from route to net => route to host. * Create new route, rather than smashing route to net. */ create: flags |= RTF_GATEWAY | RTF_DYNAMIC; error = rtrequest((int)RTM_ADD, dst, gateway, netmask, flags, (struct rtentry **)0); stat = &rtstat.rts_dynamic; } else { /* * Smash the current notion of the gateway to * this destination. Should check about netmask!!! */ rt->rt_flags |= RTF_MODIFIED; flags |= RTF_MODIFIED; stat = &rtstat.rts_newgateway; rt_setgate(rt, rt_key(rt), gateway); } } else error = EHOSTUNREACH; done: if (rt) { if (rtp && !error) *rtp = rt; else rtfree(rt); } out: if (error) rtstat.rts_badredirect++; else if (stat != NULL) (*stat)++; bzero((caddr_t)&info, sizeof(info)); info.rti_info[RTAX_DST] = dst; info.rti_info[RTAX_GATEWAY] = gateway; info.rti_info[RTAX_NETMASK] = netmask; info.rti_info[RTAX_AUTHOR] = src; rt_missmsg(RTM_REDIRECT, &info, flags, error); } /* * Routing table ioctl interface. */ int rtioctl(req, data, p) int req; caddr_t data; struct proc *p; { #ifdef INET /* Multicast goop, grrr... */ #ifdef MROUTING return mrt_ioctl(req, data); #else return mrt_ioctl(req, data, p); #endif #else /* INET */ return ENXIO; #endif /* INET */ } struct ifaddr * ifa_ifwithroute(flags, dst, gateway) int flags; struct sockaddr *dst, *gateway; { register struct ifaddr *ifa; if ((flags & RTF_GATEWAY) == 0) { /* * If we are adding a route to an interface, * and the interface is a pt to pt link * we should search for the destination * as our clue to the interface. Otherwise * we can use the local address. */ ifa = 0; if (flags & RTF_HOST) { ifa = ifa_ifwithdstaddr(dst); } if (ifa == 0) ifa = ifa_ifwithaddr(gateway); } else { /* * If we are adding a route to a remote net * or host, the gateway may still be on the * other end of a pt to pt link. */ ifa = ifa_ifwithdstaddr(gateway); } if (ifa == 0) ifa = ifa_ifwithnet(gateway); if (ifa == 0) { struct rtentry *rt = rtalloc1(dst, 0, 0UL); if (rt == 0) return (0); rt->rt_refcnt--; if ((ifa = rt->rt_ifa) == 0) return (0); } if (ifa->ifa_addr->sa_family != dst->sa_family) { struct ifaddr *oifa = ifa; ifa = ifaof_ifpforaddr(dst, ifa->ifa_ifp); if (ifa == 0) ifa = oifa; } return (ifa); } #define ROUNDUP(a) (a>0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) static int rt_fixdelete(struct radix_node *, void *); static int rt_fixchange(struct radix_node *, void *); struct rtfc_arg { struct rtentry *rt0; struct radix_node_head *rnh; }; int rtrequest(req, dst, gateway, netmask, flags, ret_nrt) int req, flags; struct sockaddr *dst, *gateway, *netmask; struct rtentry **ret_nrt; { int s = splnet(); int error = 0; register struct rtentry *rt; register struct radix_node *rn; register struct radix_node_head *rnh; struct ifaddr *ifa; struct sockaddr *ndst; #define senderr(x) { error = x ; goto bad; } if ((rnh = rt_tables[dst->sa_family]) == 0) senderr(ESRCH); if (flags & RTF_HOST) netmask = 0; switch (req) { case RTM_DELETE: if ((rn = rnh->rnh_deladdr(dst, netmask, rnh)) == 0) senderr(ESRCH); if (rn->rn_flags & (RNF_ACTIVE | RNF_ROOT)) panic ("rtrequest delete"); rt = (struct rtentry *)rn; /* * Now search what's left of the subtree for any cloned * routes which might have been formed from this node. */ if ((rt->rt_flags & RTF_PRCLONING) && netmask) { rnh->rnh_walktree_from(rnh, dst, netmask, rt_fixdelete, rt); } /* * NB: RTF_UP must be set during the search above, * because we might delete the last ref, causing * rt to get freed prematurely. */ rt->rt_flags &= ~RTF_UP; if (rt->rt_gwroute) { rt = rt->rt_gwroute; RTFREE(rt); (rt = (struct rtentry *)rn)->rt_gwroute = 0; } if ((ifa = rt->rt_ifa) && ifa->ifa_rtrequest) ifa->ifa_rtrequest(RTM_DELETE, rt, SA(0)); rttrash++; if (ret_nrt) *ret_nrt = rt; else if (rt->rt_refcnt <= 0) { rt->rt_refcnt++; rtfree(rt); } break; case RTM_RESOLVE: if (ret_nrt == 0 || (rt = *ret_nrt) == 0) senderr(EINVAL); ifa = rt->rt_ifa; flags = rt->rt_flags & ~(RTF_CLONING | RTF_PRCLONING | RTF_STATIC); flags |= RTF_WASCLONED; gateway = rt->rt_gateway; if ((netmask = rt->rt_genmask) == 0) flags |= RTF_HOST; goto makeroute; case RTM_ADD: if ((flags & RTF_GATEWAY) && !gateway) panic("rtrequest: GATEWAY but no gateway"); if ((ifa = ifa_ifwithroute(flags, dst, gateway)) == 0) senderr(ENETUNREACH); makeroute: R_Malloc(rt, struct rtentry *, sizeof(*rt)); if (rt == 0) senderr(ENOBUFS); Bzero(rt, sizeof(*rt)); rt->rt_flags = RTF_UP | flags; if (rt_setgate(rt, dst, gateway)) { Free(rt); senderr(ENOBUFS); } ndst = rt_key(rt); if (netmask) { rt_maskedcopy(dst, ndst, netmask); } else Bcopy(dst, ndst, dst->sa_len); /* * This moved from below so that rnh->rnh_addaddr() can * examine the ifa and ifp if it so desires. */ ifa->ifa_refcnt++; rt->rt_ifa = ifa; rt->rt_ifp = ifa->ifa_ifp; rn = rnh->rnh_addaddr((caddr_t)ndst, (caddr_t)netmask, rnh, rt->rt_nodes); if (rn == 0) { struct rtentry *rt2; /* * Uh-oh, we already have one of these in the tree. * We do a special hack: if the route that's already * there was generated by the protocol-cloning * mechanism, then we just blow it away and retry * the insertion of the new one. */ rt2 = rtalloc1(dst, 0, RTF_PRCLONING); if (rt2 && rt2->rt_parent) { rtrequest(RTM_DELETE, (struct sockaddr *)rt_key(rt2), rt2->rt_gateway, rt_mask(rt2), rt2->rt_flags, 0); RTFREE(rt2); rn = rnh->rnh_addaddr((caddr_t)ndst, (caddr_t)netmask, rnh, rt->rt_nodes); } } if (rn == 0) { if (rt->rt_gwroute) rtfree(rt->rt_gwroute); if (rt->rt_ifa) { IFAFREE(rt->rt_ifa); } Free(rt_key(rt)); Free(rt); senderr(EEXIST); } rt->rt_parent = 0; if (req == RTM_RESOLVE) { rt->rt_rmx = (*ret_nrt)->rt_rmx; /* copy metrics */ if ((*ret_nrt)->rt_flags & RTF_PRCLONING) { rt->rt_parent = (*ret_nrt); (*ret_nrt)->rt_refcnt++; } } if (ifa->ifa_rtrequest) ifa->ifa_rtrequest(req, rt, SA(ret_nrt ? *ret_nrt : 0)); /* * We repeat the same procedure from rt_setgate() here because * it doesn't fire when we call it there because the node * hasn't been added to the tree yet. */ if (!(rt->rt_flags & RTF_HOST)) { struct rtfc_arg arg; arg.rnh = rnh; arg.rt0 = rt; rnh->rnh_walktree_from(rnh, rt_key(rt), rt_mask(rt), rt_fixchange, &arg); } if (ret_nrt) { *ret_nrt = rt; rt->rt_refcnt++; } break; } bad: splx(s); return (error); } /* * Called from rtrequest(RTM_DELETE, ...) to fix up the route's ``family'' * (i.e., the routes related to it by the operation of cloning). This * routine is iterated over all potential former-child-routes by way of * rnh->rnh_walktree_from() above, and those that actually are children of * the late parent (passed in as VP here) are themselves deleted. */ static int rt_fixdelete(struct radix_node *rn, void *vp) { struct rtentry *rt = (struct rtentry *)rn; struct rtentry *rt0 = vp; if (rt->rt_parent == rt0 && !(rt->rt_flags & RTF_PINNED)) { return rtrequest(RTM_DELETE, rt_key(rt), (struct sockaddr *)0, rt_mask(rt), rt->rt_flags, (struct rtentry **)0); } return 0; } /* * This routine is called from rt_setgate() to do the analogous thing for * adds and changes. There is the added complication in this case of a * middle insert; i.e., insertion of a new network route between an older * network route and (cloned) host routes. For this reason, a simple check * of rt->rt_parent is insufficient; each candidate route must be tested * against the (mask, value) of the new route (passed as before in vp) * to see if the new route matches it. Unfortunately, this has the obnoxious * property of also triggering for insertion /above/ a pre-existing network * route and clones. Sigh. This may be fixed some day. * * XXX - it may be possible to do fixdelete() for changes and reserve this * routine just for adds. I'm not sure why I thought it was necessary to do * changes this way. */ #ifdef DEBUG int rtfcdebug = 0; #endif static int rt_fixchange(struct radix_node *rn, void *vp) { struct rtentry *rt = (struct rtentry *)rn; struct rtfc_arg *ap = vp; struct rtentry *rt0 = ap->rt0; struct radix_node_head *rnh = ap->rnh; u_char *xk1, *xm1, *xk2; int i, len; #ifdef DEBUG if (rtfcdebug) printf("rt_fixchange: rt %p, rt0 %p\n", rt, rt0); #endif if (!rt->rt_parent || (rt->rt_flags & RTF_PINNED)) { #ifdef DEBUG if(rtfcdebug) printf("no parent or pinned\n"); #endif return 0; } if (rt->rt_parent == rt0) { #ifdef DEBUG if(rtfcdebug) printf("parent match\n"); #endif return rtrequest(RTM_DELETE, rt_key(rt), (struct sockaddr *)0, rt_mask(rt), rt->rt_flags, (struct rtentry **)0); } /* * There probably is a function somewhere which does this... * if not, there should be. */ len = imin(((struct sockaddr *)rt_key(rt0))->sa_len, ((struct sockaddr *)rt_key(rt))->sa_len); xk1 = (u_char *)rt_key(rt0); xm1 = (u_char *)rt_mask(rt0); xk2 = (u_char *)rt_key(rt); for (i = rnh->rnh_treetop->rn_off; i < len; i++) { if ((xk2[i] & xm1[i]) != xk1[i]) { #ifdef DEBUG if(rtfcdebug) printf("no match\n"); #endif return 0; } } /* * OK, this node is a clone, and matches the node currently being * changed/added under the node's mask. So, get rid of it. */ #ifdef DEBUG if(rtfcdebug) printf("deleting\n"); #endif return rtrequest(RTM_DELETE, rt_key(rt), (struct sockaddr *)0, rt_mask(rt), rt->rt_flags, (struct rtentry **)0); } int rt_setgate(rt0, dst, gate) struct rtentry *rt0; struct sockaddr *dst, *gate; { caddr_t new, old; int dlen = ROUNDUP(dst->sa_len), glen = ROUNDUP(gate->sa_len); register struct rtentry *rt = rt0; struct radix_node_head *rnh = rt_tables[dst->sa_family]; if (rt->rt_gateway == 0 || glen > ROUNDUP(rt->rt_gateway->sa_len)) { old = (caddr_t)rt_key(rt); R_Malloc(new, caddr_t, dlen + glen); if (new == 0) return 1; rt->rt_nodes->rn_key = new; } else { new = rt->rt_nodes->rn_key; old = 0; } Bcopy(gate, (rt->rt_gateway = (struct sockaddr *)(new + dlen)), glen); if (old) { Bcopy(dst, new, dlen); Free(old); } if (rt->rt_gwroute) { rt = rt->rt_gwroute; RTFREE(rt); rt = rt0; rt->rt_gwroute = 0; } /* * Cloning loop avoidance: * In the presence of protocol-cloning and bad configuration, * it is possible to get stuck in bottomless mutual recursion * (rtrequest rt_setgate rtalloc1). We avoid this by not allowing * protocol-cloning to operate for gateways (which is probably the * correct choice anyway), and avoid the resulting reference loops * by disallowing any route to run through itself as a gateway. * This is obviuosly mandatory when we get rt->rt_output(). */ if (rt->rt_flags & RTF_GATEWAY) { rt->rt_gwroute = rtalloc1(gate, 1, RTF_PRCLONING); if (rt->rt_gwroute == rt) { RTFREE(rt->rt_gwroute); rt->rt_gwroute = 0; return 1; /* failure */ } } /* * This isn't going to do anything useful for host routes, so * don't bother. Also make sure we have a reasonable mask * (we don't yet have one during adds). */ if (!(rt->rt_flags & RTF_HOST) && rt_mask(rt) != 0) { struct rtfc_arg arg; arg.rnh = rnh; arg.rt0 = rt; rnh->rnh_walktree_from(rnh, rt_key(rt), rt_mask(rt), rt_fixchange, &arg); } return 0; } void rt_maskedcopy(src, dst, netmask) struct sockaddr *src, *dst, *netmask; { register u_char *cp1 = (u_char *)src; register u_char *cp2 = (u_char *)dst; register u_char *cp3 = (u_char *)netmask; u_char *cplim = cp2 + *cp3; u_char *cplim2 = cp2 + *cp1; *cp2++ = *cp1++; *cp2++ = *cp1++; /* copies sa_len & sa_family */ cp3 += 2; if (cplim > cplim2) cplim = cplim2; while (cp2 < cplim) *cp2++ = *cp1++ & *cp3++; if (cp2 < cplim2) bzero((caddr_t)cp2, (unsigned)(cplim2 - cp2)); } /* * Set up a routing table entry, normally * for an interface. */ int rtinit(ifa, cmd, flags) register struct ifaddr *ifa; int cmd, flags; { register struct rtentry *rt; register struct sockaddr *dst; register struct sockaddr *deldst; struct mbuf *m = 0; struct rtentry *nrt = 0; int error; dst = flags & RTF_HOST ? ifa->ifa_dstaddr : ifa->ifa_addr; if (cmd == RTM_DELETE) { if ((flags & RTF_HOST) == 0 && ifa->ifa_netmask) { m = m_get(M_WAIT, MT_SONAME); deldst = mtod(m, struct sockaddr *); rt_maskedcopy(dst, deldst, ifa->ifa_netmask); dst = deldst; } rt = rtalloc1(dst, 0, 0UL); if (rt) { rt->rt_refcnt--; if (rt->rt_ifa != ifa) { if (m) (void) m_free(m); return (flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH); } } } error = rtrequest(cmd, dst, ifa->ifa_addr, ifa->ifa_netmask, flags | ifa->ifa_flags, &nrt); if (m) (void) m_free(m); if (cmd == RTM_DELETE && error == 0 && (rt = nrt)) { rt_newaddrmsg(cmd, ifa, error, nrt); if (rt->rt_refcnt <= 0) { rt->rt_refcnt++; rtfree(rt); } } if (cmd == RTM_ADD && error == 0 && (rt = nrt)) { rt->rt_refcnt--; if (rt->rt_ifa != ifa) { printf("rtinit: wrong ifa (%p) was (%p)\n", ifa, rt->rt_ifa); if (rt->rt_ifa->ifa_rtrequest) rt->rt_ifa->ifa_rtrequest(RTM_DELETE, rt, SA(0)); IFAFREE(rt->rt_ifa); rt->rt_ifa = ifa; rt->rt_ifp = ifa->ifa_ifp; ifa->ifa_refcnt++; if (ifa->ifa_rtrequest) ifa->ifa_rtrequest(RTM_ADD, rt, SA(0)); } rt_newaddrmsg(cmd, ifa, error, nrt); } return (error); }