diff options
Diffstat (limited to 'sysdep')
-rw-r--r-- | sysdep/bsd/Makefile | 5 | ||||
-rw-r--r-- | sysdep/bsd/Modules | 5 | ||||
-rw-r--r-- | sysdep/bsd/krt-sock.c | 192 | ||||
-rw-r--r-- | sysdep/bsd/krt-sys.h | 4 | ||||
-rw-r--r-- | sysdep/bsd/setkey.h | 14 | ||||
-rw-r--r-- | sysdep/bsd/sysio.h | 2 | ||||
-rw-r--r-- | sysdep/cf/bsd-v6.h | 22 | ||||
-rw-r--r-- | sysdep/cf/bsd.h | 3 | ||||
-rw-r--r-- | sysdep/cf/linux-v6.h | 21 | ||||
-rw-r--r-- | sysdep/cf/linux.h | 5 | ||||
-rw-r--r-- | sysdep/linux/Makefile | 4 | ||||
-rw-r--r-- | sysdep/linux/Modules | 5 | ||||
-rw-r--r-- | sysdep/linux/netlink.c | 470 | ||||
-rw-r--r-- | sysdep/unix/Makefile | 5 | ||||
-rw-r--r-- | sysdep/unix/Modules | 12 | ||||
-rw-r--r-- | sysdep/unix/config.Y | 3 | ||||
-rw-r--r-- | sysdep/unix/io.c | 55 | ||||
-rw-r--r-- | sysdep/unix/krt.Y | 33 | ||||
-rw-r--r-- | sysdep/unix/krt.c | 169 | ||||
-rw-r--r-- | sysdep/unix/krt.h | 10 | ||||
-rw-r--r-- | sysdep/unix/log.c | 2 | ||||
-rw-r--r-- | sysdep/unix/main.c | 8 | ||||
-rw-r--r-- | sysdep/unix/unix.h | 31 |
23 files changed, 623 insertions, 457 deletions
diff --git a/sysdep/bsd/Makefile b/sysdep/bsd/Makefile new file mode 100644 index 00000000..ddc87239 --- /dev/null +++ b/sysdep/bsd/Makefile @@ -0,0 +1,5 @@ +src := krt-sock.c +obj := $(src-o-files) +$(all-daemon) +$(conf-y-targets): $(s)krt-sock.Y + diff --git a/sysdep/bsd/Modules b/sysdep/bsd/Modules deleted file mode 100644 index 39db88e9..00000000 --- a/sysdep/bsd/Modules +++ /dev/null @@ -1,5 +0,0 @@ -krt-sock.c -krt-sock.Y -krt-sys.h -sysio.h -setkey.h diff --git a/sysdep/bsd/krt-sock.c b/sysdep/bsd/krt-sock.c index f2ae81c3..3440ed63 100644 --- a/sysdep/bsd/krt-sock.c +++ b/sysdep/bsd/krt-sock.c @@ -28,9 +28,9 @@ #include "nest/route.h" #include "nest/protocol.h" #include "nest/iface.h" -#include "lib/timer.h" -#include "lib/unix.h" -#include "lib/krt.h" +#include "sysdep/unix/timer.h" +#include "sysdep/unix/unix.h" +#include "sysdep/unix/krt.h" #include "lib/string.h" #include "lib/socket.h" @@ -72,7 +72,6 @@ #endif - /* Dynamic max number of tables */ int krt_max_tables; @@ -136,7 +135,7 @@ extern int setfib(int fib); /* table_id -> krt_proto map */ #ifdef KRT_SHARED_SOCKET -static struct krt_proto *krt_table_map[KRT_MAX_TABLES]; +static struct krt_proto *krt_table_map[KRT_MAX_TABLES][2]; #endif @@ -207,7 +206,8 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e) msg.rtm.rtm_addrs = RTA_DST; msg.rtm.rtm_flags = RTF_UP | RTF_PROTO1; - if (net->n.pxlen == MAX_PREFIX_LENGTH) + /* XXXX */ + if (net_pxlen(net->n.addr) == net_max_prefix_length[net->n.addr->type]) msg.rtm.rtm_flags |= RTF_HOST; else msg.rtm.rtm_addrs |= RTA_NETMASK; @@ -245,15 +245,28 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e) gw = a->gw; -#ifdef IPV6 /* Embed interface ID to link-local address */ if (ipa_is_link_local(gw)) _I0(gw) = 0xfe800000 | (i->index & 0x0000ffff); -#endif - sockaddr_fill(&dst, BIRD_AF, net->n.prefix, NULL, 0); - sockaddr_fill(&mask, BIRD_AF, ipa_mkmask(net->n.pxlen), NULL, 0); - sockaddr_fill(&gate, BIRD_AF, gw, NULL, 0); + int af = AF_UNSPEC; + + switch (net->n.addr->type) { + case NET_IP4: + af = AF_INET; + break; + case NET_IP6: + af = AF_INET6; + break; + default: + log(L_ERR "KRT: Not sending VPN route %N to kernel", net->n.addr); + return -1; + } + + + sockaddr_fill(&dst, af, net_prefix(net->n.addr), NULL, 0); + sockaddr_fill(&mask, af, net_pxmask(net->n.addr), NULL, 0); + sockaddr_fill(&gate, af, gw, NULL, 0); switch (a->dest) { @@ -281,7 +294,7 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e) return -1; } - sockaddr_fill(&gate, BIRD_AF, i->addr->ip, NULL, 0); + sockaddr_fill(&gate, ipa_is_ip4(i->addr->ip) ? AF_INET : AF_INET6, i->addr->ip, NULL, 0); msg.rtm.rtm_addrs |= RTA_GATEWAY; } break; @@ -299,7 +312,7 @@ krt_send_route(struct krt_proto *p, int cmd, rte *e) msg.rtm.rtm_msglen = l; if ((l = write(p->sys.sk->fd, (char *)&msg, l)) < 0) { - log(L_ERR "KRT: Error sending route %I/%d to kernel: %m", net->n.prefix, net->n.pxlen); + log(L_ERR "KRT: Error sending route %N to kernel: %m", net->n.addr); return -1; } @@ -331,10 +344,12 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan) { /* p is NULL iff KRT_SHARED_SOCKET and !scan */ + int ipv6; rte *e; net *net; sockaddr dst, gate, mask; ip_addr idst, igate, imask; + net_addr ndst; void *body = (char *)msg->buf; int new = (msg->rtm.rtm_type != RTM_DELETE); char *errmsg = "KRT: Invalid route received"; @@ -352,42 +367,64 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan) if (flags & RTF_LLINFO) SKIP("link-local\n"); -#ifdef KRT_SHARED_SOCKET - if (!scan) - { - int table_id = msg->rtm.rtm_tableid; - p = (table_id < KRT_MAX_TABLES) ? krt_table_map[table_id] : NULL; - - if (!p) - SKIP("unknown table id %d\n", table_id); - } -#endif - GETADDR(&dst, RTA_DST); GETADDR(&gate, RTA_GATEWAY); GETADDR(&mask, RTA_NETMASK); - if (dst.sa.sa_family != BIRD_AF) - SKIP("invalid DST"); + switch (dst.sa.sa_family) { + case AF_INET: + ipv6 = 0; + break; + case AF_INET6: + ipv6 = 1; + break; + default: + SKIP("invalid DST"); + } + + /* We do not test family for RTA_NETMASK, because BSD sends us + some strange values, but interpreting them as IPv4/IPv6 works */ + mask.sa.sa_family = dst.sa.sa_family; idst = ipa_from_sa(&dst); imask = ipa_from_sa(&mask); - igate = (gate.sa.sa_family == BIRD_AF) ? ipa_from_sa(&gate) : IPA_NONE; + igate = (gate.sa.sa_family == dst.sa.sa_family) ? ipa_from_sa(&gate) : IPA_NONE; - /* We do not test family for RTA_NETMASK, because BSD sends us - some strange values, but interpreting them as IPv4/IPv6 works */ +#ifdef KRT_SHARED_SOCKET + if (!scan) + { + int table_id = msg->rtm.rtm_tableid; + p = (table_id < KRT_MAX_TABLES) ? krt_table_map[table_id][ipv6] : NULL; + if (!p) + SKIP("unknown table id %d\n", table_id); + } +#endif + if ((!ipv6) && (p->p.main_channel->table->addr_type != NET_IP4)) + SKIP("reading only IPv4 routes"); + if ( ipv6 && (p->p.main_channel->table->addr_type != NET_IP6)) + SKIP("reading only IPv6 routes"); int c = ipa_classify_net(idst); if ((c < 0) || !(c & IADDR_HOST) || ((c & IADDR_SCOPE_MASK) <= SCOPE_LINK)) SKIP("strange class/scope\n"); - int pxlen = (flags & RTF_HOST) ? MAX_PREFIX_LENGTH : ipa_masklen(imask); + int pxlen; + if (ipv6) + pxlen = (flags & RTF_HOST) ? IP6_MAX_PREFIX_LENGTH : ip6_masklen(&ipa_to_ip6(imask)); + else + pxlen = (flags & RTF_HOST) ? IP4_MAX_PREFIX_LENGTH : ip4_masklen(ipa_to_ip4(imask)); + if (pxlen < 0) { log(L_ERR "%s (%I) - netmask %I", errmsg, idst, imask); return; } + if (ipv6) + net_fill_ip6(&ndst, ipa_to_ip6(idst), pxlen); + else + net_fill_ip4(&ndst, ipa_to_ip4(idst), pxlen); + if ((flags & RTF_GATEWAY) && ipa_zero(igate)) - { log(L_ERR "%s (%I/%d) - missing gateway", errmsg, idst, pxlen); return; } + { log(L_ERR "%s (%N) - missing gateway", errmsg, ndst); return; } u32 self_mask = RTF_PROTO1; u32 alien_mask = RTF_STATIC | RTF_PROTO1 | RTF_GATEWAY; @@ -426,7 +463,7 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan) else src = KRT_SRC_KERNEL; - net = net_get(p->p.table, idst, pxlen); + net = net_get(p->p.main_channel->table, &ndst); rta a = { .src = p->p.main_source, @@ -455,8 +492,8 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan) a.iface = if_find_by_index(msg->rtm.rtm_index); if (!a.iface) { - log(L_ERR "KRT: Received route %I/%d with unknown ifindex %u", - net->n.prefix, net->n.pxlen, msg->rtm.rtm_index); + log(L_ERR "KRT: Received route %N with unknown ifindex %u", + net->n.addr, msg->rtm.rtm_index); return; } @@ -466,11 +503,9 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan) a.dest = RTD_ROUTER; a.gw = igate; -#ifdef IPV6 /* Clean up embedded interface ID returned in link-local address */ if (ipa_is_link_local(a.gw)) _I0(a.gw) = 0xfe800000; -#endif ng = neigh_find2(&p->p, &a.gw, a.iface, 0); if (!ng || (ng->scope == SCOPE_HOST)) @@ -480,8 +515,8 @@ krt_read_route(struct ks_msg *msg, struct krt_proto *p, int scan) if (ipa_classify(a.gw) == (IADDR_HOST | SCOPE_HOST)) return; - log(L_ERR "KRT: Received route %I/%d with strange next-hop %I", - net->n.prefix, net->n.pxlen, a.gw); + log(L_ERR "KRT: Received route %N with strange next-hop %I", + net->n.addr, a.gw); return; } } @@ -643,22 +678,24 @@ krt_read_addr(struct ks_msg *msg, int scan) GETADDR (&null, RTA_AUTHOR); GETADDR (&brd, RTA_BRD); - /* Some other family address */ - if (addr.sa.sa_family != BIRD_AF) - return; + /* Is addr family IP4 or IP6? */ + int ipv6; + switch (addr.sa.sa_family) { + case AF_INET: ipv6 = 0; break; + case AF_INET6: ipv6 = 1; break; + default: return; + } iaddr = ipa_from_sa(&addr); imask = ipa_from_sa(&mask); ibrd = ipa_from_sa(&brd); - - if ((masklen = ipa_masklen(imask)) < 0) + if ((ipv6 ? (masklen = ip6_masklen(&ipa_to_ip6(imask))) : (masklen = ip4_masklen(ipa_to_ip4(imask)))) < 0) { - log(L_ERR "KIF: Invalid masklen %I for %s", imask, iface->name); + log(L_ERR "KIF: Invalid mask %I for %s", imask, iface->name); return; } -#ifdef IPV6 /* Clean up embedded interface ID returned in link-local address */ if (ipa_is_link_local(iaddr)) @@ -666,13 +703,11 @@ krt_read_addr(struct ks_msg *msg, int scan) if (ipa_is_link_local(ibrd)) _I0(ibrd) = 0xfe800000; -#endif bzero(&ifa, sizeof(ifa)); ifa.iface = iface; ifa.ip = iaddr; - ifa.pxlen = masklen; scope = ipa_classify(ifa.ip); if (scope < 0) @@ -682,17 +717,16 @@ krt_read_addr(struct ks_msg *msg, int scan) } ifa.scope = scope & IADDR_SCOPE_MASK; - if (masklen < BITS_PER_IP_ADDRESS) + if (masklen < (ipv6 ? IP6_MAX_PREFIX_LENGTH : IP4_MAX_PREFIX_LENGTH)) { - ifa.prefix = ipa_and(ifa.ip, ipa_mkmask(masklen)); + net_fill_ipa(&ifa.prefix, ifa.ip, masklen); + net_normalize(&ifa.prefix); - if (masklen == (BITS_PER_IP_ADDRESS - 1)) + if (masklen == ((ipv6 ? IP6_MAX_PREFIX_LENGTH : IP4_MAX_PREFIX_LENGTH) - 1)) ifa.opposite = ipa_opposite_m1(ifa.ip); -#ifndef IPV6 - if (masklen == (BITS_PER_IP_ADDRESS - 2)) + if ((!ipv6) && (masklen == IP4_MAX_PREFIX_LENGTH - 2)) ifa.opposite = ipa_opposite_m2(ifa.ip); -#endif if (iface->flags & IF_BROADCAST) ifa.brd = ibrd; @@ -702,12 +736,13 @@ krt_read_addr(struct ks_msg *msg, int scan) } else if (!(iface->flags & IF_MULTIACCESS) && ipa_nonzero(ibrd)) { - ifa.prefix = ifa.opposite = ibrd; + net_fill_ipa(&ifa.prefix, ibrd, (ipv6 ? IP6_MAX_PREFIX_LENGTH : IP4_MAX_PREFIX_LENGTH)); + ifa.opposite = ibrd; ifa.flags |= IA_PEER; } else { - ifa.prefix = ifa.ip; + net_fill_ipa(&ifa.prefix, ifa.ip, (ipv6 ? IP6_MAX_PREFIX_LENGTH : IP4_MAX_PREFIX_LENGTH)); ifa.flags |= IA_HOST; } @@ -804,7 +839,7 @@ krt_sysctl_scan(struct proto *p, int cmd, int table_id) mib[0] = CTL_NET; mib[1] = PF_ROUTE; mib[2] = 0; - mib[3] = BIRD_AF; + mib[3] = 0; // Set AF to 0 for all available families mib[4] = cmd; mib[5] = 0; mcnt = 6; @@ -948,6 +983,7 @@ krt_sock_open(pool *pool, void *data, int table_id) return sk; } +static u32 krt_table_cf[(KRT_MAX_TABLES+31) / 32][2]; #ifdef KRT_SHARED_SOCKET @@ -979,7 +1015,17 @@ krt_sock_close_shared(void) int krt_sys_start(struct krt_proto *p) { - krt_table_map[KRT_CF->sys.table_id] = p; + int id = KRT_CF->sys.table_id; + + if (krt_table_cf[id/32][!!(p->af == AF_INET6)] & (1 << (id%32))) + { + log(L_ERR "%s: Multiple kernel syncers defined for table #%d", p->p.name, id); + return 0; + } + + krt_table_cf[id/32][!!(p->af == AF_INET6)] |= (1 << (id%32)); + + krt_table_map[KRT_CF->sys.table_id][!!(p->af == AF_INET6)] = p; krt_sock_open_shared(); p->sys.sk = krt_sock; @@ -990,10 +1036,12 @@ krt_sys_start(struct krt_proto *p) void krt_sys_shutdown(struct krt_proto *p) { + krt_table_cf[(KRT_CF->sys.table_id)/32][!!(p->af == AF_INET6)] &= ~(1 << ((KRT_CF->sys.table_id)%32)); + krt_sock_close_shared(); p->sys.sk = NULL; - krt_table_map[KRT_CF->sys.table_id] = NULL; + krt_table_map[KRT_CF->sys.table_id][!!(p->af == AF_INET6)] = NULL; krt_buffer_release(&p->p); } @@ -1003,6 +1051,16 @@ krt_sys_shutdown(struct krt_proto *p) int krt_sys_start(struct krt_proto *p) { + int id = KRT_CF->sys.table_id; + + if (krt_table_cf[id/32][!!(p->af == AF_INET6)] & (1 << (id%32))) + { + log(L_ERR "%s: Multiple kernel syncers defined for table #%d", p->p.name, id); + return 0; + } + + krt_table_cf[id/32][!!(p->af == AF_INET6)] |= (1 << (id%32)); + p->sys.sk = krt_sock_open(p->p.pool, p, KRT_CF->sys.table_id); return 1; } @@ -1010,6 +1068,8 @@ krt_sys_start(struct krt_proto *p) void krt_sys_shutdown(struct krt_proto *p) { + krt_table_cf[(KRT_CF->sys.table_id)/32][!!(p->af == AF_INET6)] &= ~(1 << ((KRT_CF->sys.table_id)%32)); + rfree(p->sys.sk); p->sys.sk = NULL; @@ -1021,8 +1081,6 @@ krt_sys_shutdown(struct krt_proto *p) /* KRT configuration callbacks */ -static u32 krt_table_cf[(KRT_MAX_TABLES+31) / 32]; - int krt_sys_reconfigure(struct krt_proto *p UNUSED, struct krt_config *n, struct krt_config *o) { @@ -1036,18 +1094,6 @@ krt_sys_preconfig(struct config *c UNUSED) bzero(&krt_table_cf, sizeof(krt_table_cf)); } -void -krt_sys_postconfig(struct krt_config *x) -{ - u32 *tbl = krt_table_cf; - int id = x->sys.table_id; - - if (tbl[id/32] & (1 << (id%32))) - cf_error("Multiple kernel syncers defined for table #%d", id); - - tbl[id/32] |= (1 << (id%32)); -} - void krt_sys_init_config(struct krt_config *c) { c->sys.table_id = 0; /* Default table */ @@ -1076,7 +1122,7 @@ kif_sys_shutdown(struct kif_proto *p) struct ifa * kif_get_primary_ip(struct iface *i) { -#ifndef IPV6 +#if 0 static int fd = -1; if (fd < 0) diff --git a/sysdep/bsd/krt-sys.h b/sysdep/bsd/krt-sys.h index a63f8caf..870cdf2c 100644 --- a/sysdep/bsd/krt-sys.h +++ b/sysdep/bsd/krt-sys.h @@ -44,8 +44,8 @@ struct krt_state { static inline void krt_sys_io_init(void) { } static inline void krt_sys_init(struct krt_proto *p UNUSED) { } +static inline void krt_sys_postconfig(struct krt_config *x UNUSED) { } -static inline int krt_sys_get_attr(eattr *a UNUSED, byte *buf UNUSED, int buflen UNUSED) { } - +static inline int krt_sys_get_attr(eattr *a UNUSED, byte *buf UNUSED, int buflen UNUSED) { return GA_UNKNOWN; } #endif diff --git a/sysdep/bsd/setkey.h b/sysdep/bsd/setkey.h index b417faca..3bcd8623 100644 --- a/sysdep/bsd/setkey.h +++ b/sysdep/bsd/setkey.h @@ -11,7 +11,7 @@ #include <netipsec/ipsec.h> #include "nest/bird.h" -#include "lib/unix.h" +#include "sysdep/unix/unix.h" /* @@ -63,7 +63,7 @@ setkey_send(struct sadb_msg *msg, uint len) * operations to implement replace. */ static int -setkey_md5(sockaddr *src, sockaddr *dst, char *passwd, uint type) +setkey_md5(sockaddr *src, sockaddr *dst, uint pxlen, char *passwd, uint type) { uint passwd_len = passwd ? strlen(passwd) : 0; @@ -122,7 +122,7 @@ setkey_md5(sockaddr *src, sockaddr *dst, char *passwd, uint type) saddr->sadb_address_len = PFKEY_UNIT64(len); saddr->sadb_address_exttype = SADB_EXT_ADDRESS_SRC; saddr->sadb_address_proto = IPSEC_ULPROTO_ANY; - saddr->sadb_address_prefixlen = MAX_PREFIX_LENGTH; + saddr->sadb_address_prefixlen = pxlen; memcpy(pos + sizeof(struct sadb_address), &src->sa, src->sa.sa_len); pos += len; @@ -132,7 +132,7 @@ setkey_md5(sockaddr *src, sockaddr *dst, char *passwd, uint type) daddr->sadb_address_len = PFKEY_UNIT64(len); daddr->sadb_address_exttype = SADB_EXT_ADDRESS_DST; daddr->sadb_address_proto = IPSEC_ULPROTO_ANY; - daddr->sadb_address_prefixlen = MAX_PREFIX_LENGTH; + daddr->sadb_address_prefixlen = pxlen; memcpy(pos + sizeof(struct sadb_address), &dst->sa, dst->sa.sa_len); pos += len; @@ -152,18 +152,20 @@ sk_set_md5_in_sasp_db(sock *s, ip_addr local, ip_addr remote, struct iface *ifa, sockaddr_fill(&src, s->af, local, ifa, 0); sockaddr_fill(&dst, s->af, remote, ifa, 0); + uint pxlen = (s->af == AF_INET) ? IP4_MAX_PREFIX_LENGTH : IP6_MAX_PREFIX_LENGTH; + if (passwd && *passwd) { int len = strlen(passwd); if (len > TCP_KEYLEN_MAX) ERR_MSG("The password for TCP MD5 Signature is too long"); - if (setkey_md5(&src, &dst, passwd, SADB_ADD) < 0) + if (setkey_md5(&src, &dst, pxlen, passwd, SADB_ADD) < 0) ERR_MSG("Cannot add TCP-MD5 password into the IPsec SA/SP database"); } else { - if (setkey_md5(&src, &dst, NULL, SADB_DELETE) < 0) + if (setkey_md5(&src, &dst, pxlen, NULL, SADB_DELETE) < 0) ERR_MSG("Cannot delete TCP-MD5 password from the IPsec SA/SP database"); } return 0; diff --git a/sysdep/bsd/sysio.h b/sysdep/bsd/sysio.h index 6c20733f..57c45bcf 100644 --- a/sysdep/bsd/sysio.h +++ b/sysdep/bsd/sysio.h @@ -196,7 +196,7 @@ sk_prepare_ip_header(sock *s, void *hdr, int dlen) #if defined(__FreeBSD__) #define USE_MD5SIG_SETKEY -#include "lib/setkey.h" +#include "sysdep/bsd/setkey.h" #endif int diff --git a/sysdep/cf/bsd-v6.h b/sysdep/cf/bsd-v6.h deleted file mode 100644 index 745dfba3..00000000 --- a/sysdep/cf/bsd-v6.h +++ /dev/null @@ -1,22 +0,0 @@ -/* - * Configuration for *BSD based systems (tested on FreeBSD and NetBSD) - * - * (c) 2004 Ondrej Filip <feela@network.cz> - * - * Can be freely distributed and used under the terms of the GNU GPL. - */ - -#define IPV6 - -#define CONFIG_AUTO_ROUTES -#define CONFIG_SELF_CONSCIOUS -#define CONFIG_MULTIPLE_TABLES -#define CONFIG_SINGLE_ROUTE - -#define CONFIG_SKIP_MC_BIND -#define CONFIG_NO_IFACE_BIND - -/* -Link: sysdep/unix -Link: sysdep/bsd - */ diff --git a/sysdep/cf/bsd.h b/sysdep/cf/bsd.h index 51beb42b..22c54277 100644 --- a/sysdep/cf/bsd.h +++ b/sysdep/cf/bsd.h @@ -15,6 +15,9 @@ #define CONFIG_NO_IFACE_BIND #define CONFIG_USE_HDRINCL +#define CONFIG_INCLUDE_SYSIO_H "sysdep/bsd/sysio.h" +#define CONFIG_INCLUDE_KRTSYS_H "sysdep/bsd/krt-sys.h" + /* Link: sysdep/unix Link: sysdep/bsd diff --git a/sysdep/cf/linux-v6.h b/sysdep/cf/linux-v6.h deleted file mode 100644 index 09f60377..00000000 --- a/sysdep/cf/linux-v6.h +++ /dev/null @@ -1,21 +0,0 @@ -/* - * Configuration for Linux based systems running IPv6 - * - * (c) 1998--1999 Martin Mares <mj@ucw.cz> - * - * Can be freely distributed and used under the terms of the GNU GPL. - */ - -#define IPV6 - -#define CONFIG_AUTO_ROUTES -#define CONFIG_SELF_CONSCIOUS -#define CONFIG_MULTIPLE_TABLES -#define CONFIG_ALL_TABLES_AT_ONCE - -#define CONFIG_RESTRICTED_PRIVILEGES - -/* -Link: sysdep/linux -Link: sysdep/unix - */ diff --git a/sysdep/cf/linux.h b/sysdep/cf/linux.h index 9e34f869..cec9499c 100644 --- a/sysdep/cf/linux.h +++ b/sysdep/cf/linux.h @@ -14,7 +14,12 @@ #define CONFIG_MC_PROPER_SRC #define CONFIG_UNIX_DONTROUTE +#define CONFIG_INCLUDE_SYSIO_H "sysdep/linux/sysio.h" +#define CONFIG_INCLUDE_KRTSYS_H "sysdep/linux/krt-sys.h" + #define CONFIG_RESTRICTED_PRIVILEGES +#define CONFIG_INCLUDE_SYSPRIV_H "sysdep/linux/syspriv.h" + /* Link: sysdep/linux diff --git a/sysdep/linux/Makefile b/sysdep/linux/Makefile new file mode 100644 index 00000000..23cf1d9d --- /dev/null +++ b/sysdep/linux/Makefile @@ -0,0 +1,4 @@ +src := netlink.c +obj := $(src-o-files) +$(all-daemon) +$(conf-y-targets): $(s)netlink.Y diff --git a/sysdep/linux/Modules b/sysdep/linux/Modules deleted file mode 100644 index 940660b6..00000000 --- a/sysdep/linux/Modules +++ /dev/null @@ -1,5 +0,0 @@ -krt-sys.h -netlink.c -netlink.Y -sysio.h -syspriv.h diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c index 79dd1405..7af575a7 100644 --- a/sysdep/linux/netlink.c +++ b/sysdep/linux/netlink.c @@ -6,6 +6,7 @@ * Can be freely distributed and used under the terms of the GNU GPL. */ +#include <alloca.h> #include <stdio.h> #include <unistd.h> #include <fcntl.h> @@ -19,9 +20,9 @@ #include "nest/route.h" #include "nest/protocol.h" #include "nest/iface.h" -#include "lib/timer.h" -#include "lib/unix.h" -#include "lib/krt.h" +#include "sysdep/unix/timer.h" +#include "sysdep/unix/unix.h" +#include "sysdep/unix/krt.h" #include "lib/socket.h" #include "lib/string.h" #include "lib/hash.h" @@ -50,11 +51,7 @@ #endif -#ifdef IPV6 -#define krt_ecmp6(X) 1 -#else -#define krt_ecmp6(X) 0 -#endif +#define krt_ecmp6(p) ((p)->af == AF_INET6) /* * Structure nl_parse_state keeps state of received route processing. Ideally, @@ -302,20 +299,18 @@ static struct nl_want_attrs ifla_attr_want[BIRD_IFLA_MAX] = { #define BIRD_IFA_MAX (IFA_FLAGS+1) -#ifndef IPV6 static struct nl_want_attrs ifa_attr_want4[BIRD_IFA_MAX] = { [IFA_ADDRESS] = { 1, 1, sizeof(ip4_addr) }, [IFA_LOCAL] = { 1, 1, sizeof(ip4_addr) }, [IFA_BROADCAST] = { 1, 1, sizeof(ip4_addr) }, [IFA_FLAGS] = { 1, 1, sizeof(u32) }, }; -#else + static struct nl_want_attrs ifa_attr_want6[BIRD_IFA_MAX] = { [IFA_ADDRESS] = { 1, 1, sizeof(ip6_addr) }, [IFA_LOCAL] = { 1, 1, sizeof(ip6_addr) }, [IFA_FLAGS] = { 1, 1, sizeof(u32) }, }; -#endif #define BIRD_RTA_MAX (RTA_TABLE+1) @@ -324,7 +319,6 @@ static struct nl_want_attrs mpnh_attr_want4[BIRD_RTA_MAX] = { [RTA_GATEWAY] = { 1, 1, sizeof(ip4_addr) }, }; -#ifndef IPV6 static struct nl_want_attrs rtm_attr_want4[BIRD_RTA_MAX] = { [RTA_DST] = { 1, 1, sizeof(ip4_addr) }, [RTA_OIF] = { 1, 1, sizeof(u32) }, @@ -336,7 +330,7 @@ static struct nl_want_attrs rtm_attr_want4[BIRD_RTA_MAX] = { [RTA_FLOW] = { 1, 1, sizeof(u32) }, [RTA_TABLE] = { 1, 1, sizeof(u32) }, }; -#else + static struct nl_want_attrs rtm_attr_want6[BIRD_RTA_MAX] = { [RTA_DST] = { 1, 1, sizeof(ip6_addr) }, [RTA_IIF] = { 1, 1, sizeof(u32) }, @@ -348,7 +342,6 @@ static struct nl_want_attrs rtm_attr_want6[BIRD_RTA_MAX] = { [RTA_FLOW] = { 1, 1, sizeof(u32) }, [RTA_TABLE] = { 1, 1, sizeof(u32) }, }; -#endif static int @@ -364,7 +357,7 @@ nl_parse_attrs(struct rtattr *a, struct nl_want_attrs *want, struct rtattr **k, if (want[a->rta_type].checksize && (RTA_PAYLOAD(a) != want[a->rta_type].size)) { - log(L_ERR "nl_parse_attrs: Malformed message received"); + log(L_ERR "nl_parse_attrs: Malformed attribute received"); return 0; } @@ -389,6 +382,13 @@ static inline ip4_addr rta_get_ip4(struct rtattr *a) static inline ip6_addr rta_get_ip6(struct rtattr *a) { return ip6_ntoh(*(ip6_addr *) RTA_DATA(a)); } +static inline ip_addr rta_get_ipa(struct rtattr *a) +{ + if (RTA_PAYLOAD(a) == sizeof(ip4_addr)) + return ipa_from_ip4(rta_get_ip4(a)); + else + return ipa_from_ip6(rta_get_ip6(a)); +} struct rtattr * nl_add_attr(struct nlmsghdr *h, uint bufsize, uint code, const void *data, uint dlen) @@ -411,16 +411,32 @@ nl_add_attr(struct nlmsghdr *h, uint bufsize, uint code, const void *data, uint } static inline void -nl_add_attr_u32(struct nlmsghdr *h, unsigned bufsize, int code, u32 data) +nl_add_attr_u32(struct nlmsghdr *h, uint bufsize, int code, u32 data) { nl_add_attr(h, bufsize, code, &data, 4); } static inline void -nl_add_attr_ipa(struct nlmsghdr *h, unsigned bufsize, int code, ip_addr ipa) +nl_add_attr_ip4(struct nlmsghdr *h, uint bufsize, int code, ip4_addr ip4) +{ + ip4 = ip4_hton(ip4); + nl_add_attr(h, bufsize, code, &ip4, sizeof(ip4)); +} + +static inline void +nl_add_attr_ip6(struct nlmsghdr *h, uint bufsize, int code, ip6_addr ip6) { - ipa_hton(ipa); - nl_add_attr(h, bufsize, code, &ipa, sizeof(ipa)); + ip6 = ip6_hton(ip6); + nl_add_attr(h, bufsize, code, &ip6, sizeof(ip6)); +} + +static inline void +nl_add_attr_ipa(struct nlmsghdr *h, uint bufsize, int code, ip_addr ipa) +{ + if (ipa_is_ip4(ipa)) + nl_add_attr_ip4(h, bufsize, code, ipa_to_ip4(ipa)); + else + nl_add_attr_ip6(h, bufsize, code, ipa_to_ip6(ipa)); } static inline struct rtattr * @@ -456,7 +472,7 @@ nl_close_nexthop(struct nlmsghdr *h, struct rtnexthop *nh) } static void -nl_add_multipath(struct nlmsghdr *h, unsigned bufsize, struct mpnh *nh) +nl_add_multipath(struct nlmsghdr *h, uint bufsize, struct mpnh *nh) { struct rtattr *a = nl_open_attr(h, bufsize, RTA_MULTIPATH); @@ -518,12 +534,12 @@ nl_parse_multipath(struct krt_proto *p, struct rtattr *ra) nl_parse_attrs(RTNH_DATA(nh), mpnh_attr_want4, a, sizeof(a)); if (a[RTA_GATEWAY]) { - memcpy(&rv->gw, RTA_DATA(a[RTA_GATEWAY]), sizeof(ip_addr)); - ipa_ntoh(rv->gw); + rv->gw = rta_get_ipa(a[RTA_GATEWAY]); - neighbor *ng = neigh_find2(&p->p, &rv->gw, rv->iface, - (nh->rtnh_flags & RTNH_F_ONLINK) ? NEF_ONLINK : 0); - if (!ng || (ng->scope == SCOPE_HOST)) + neighbor *nbr; + nbr = neigh_find2(&p->p, &rv->gw, rv->iface, + (nh->rtnh_flags & RTNH_F_ONLINK) ? NEF_ONLINK : 0); + if (!nbr || (nbr->scope == SCOPE_HOST)) return NULL; } else @@ -658,52 +674,131 @@ nl_parse_link(struct nlmsghdr *h, int scan) } static void -nl_parse_addr(struct nlmsghdr *h, int scan) +nl_parse_addr4(struct ifaddrmsg *i, int scan, int new) { - struct ifaddrmsg *i; struct rtattr *a[BIRD_IFA_MAX]; - int new = h->nlmsg_type == RTM_NEWADDR; - struct ifa ifa; struct iface *ifi; - int scope; u32 ifa_flags; + int scope; - if (!(i = nl_checkin(h, sizeof(*i)))) + if (!nl_parse_attrs(IFA_RTA(i), ifa_attr_want4, a, sizeof(a))) return; - switch (i->ifa_family) + if (!a[IFA_LOCAL]) { -#ifndef IPV6 - case AF_INET: - if (!nl_parse_attrs(IFA_RTA(i), ifa_attr_want4, a, sizeof(a))) - return; - if (!a[IFA_LOCAL]) - { - log(L_ERR "KIF: Malformed message received (missing IFA_LOCAL)"); - return; - } - break; -#else - case AF_INET6: - if (!nl_parse_attrs(IFA_RTA(i), ifa_attr_want6, a, sizeof(a))) - return; - break; -#endif - default: - return; + log(L_ERR "KIF: Malformed message received (missing IFA_LOCAL)"); + return; } - if (!a[IFA_ADDRESS]) { log(L_ERR "KIF: Malformed message received (missing IFA_ADDRESS)"); return; } + ifi = if_find_by_index(i->ifa_index); + if (!ifi) + { + log(L_ERR "KIF: Received address message for unknown interface %d", i->ifa_index); + return; + } + if (a[IFA_FLAGS]) ifa_flags = rta_get_u32(a[IFA_FLAGS]); else ifa_flags = i->ifa_flags; + struct ifa ifa; + bzero(&ifa, sizeof(ifa)); + ifa.iface = ifi; + if (ifa_flags & IFA_F_SECONDARY) + ifa.flags |= IA_SECONDARY; + + ifa.ip = rta_get_ipa(a[IFA_LOCAL]); + + if (i->ifa_prefixlen > IP4_MAX_PREFIX_LENGTH) + { + log(L_ERR "KIF: Invalid prefix length for interface %s: %d", ifi->name, i->ifa_prefixlen); + new = 0; + } + if (i->ifa_prefixlen == IP4_MAX_PREFIX_LENGTH) + { + ifa.brd = rta_get_ipa(a[IFA_ADDRESS]); + net_fill_ip4(&ifa.prefix, rta_get_ip4(a[IFA_ADDRESS]), i->ifa_prefixlen); + + /* It is either a host address or a peer address */ + if (ipa_equal(ifa.ip, ifa.brd)) + ifa.flags |= IA_HOST; + else + { + ifa.flags |= IA_PEER; + ifa.opposite = ifa.brd; + } + } + else + { + net_fill_ip4(&ifa.prefix, ipa_to_ip4(ifa.ip), i->ifa_prefixlen); + net_normalize(&ifa.prefix); + + if (i->ifa_prefixlen == IP4_MAX_PREFIX_LENGTH - 1) + ifa.opposite = ipa_opposite_m1(ifa.ip); + + if (i->ifa_prefixlen == IP4_MAX_PREFIX_LENGTH - 2) + ifa.opposite = ipa_opposite_m2(ifa.ip); + + if ((ifi->flags & IF_BROADCAST) && a[IFA_BROADCAST]) + { + ip4_addr xbrd = rta_get_ip4(a[IFA_BROADCAST]); + ip4_addr ybrd = ip4_or(ipa_to_ip4(ifa.ip), ip4_not(ip4_mkmask(i->ifa_prefixlen))); + + if (ip4_equal(xbrd, net4_prefix(&ifa.prefix)) || ip4_equal(xbrd, ybrd)) + ifa.brd = ipa_from_ip4(xbrd); + else if (ifi->flags & IF_TMP_DOWN) /* Complain only during the first scan */ + { + log(L_ERR "KIF: Invalid broadcast address %I4 for %s", xbrd, ifi->name); + ifa.brd = ipa_from_ip4(ybrd); + } + } + } + + scope = ipa_classify(ifa.ip); + if (scope < 0) + { + log(L_ERR "KIF: Invalid interface address %I for %s", ifa.ip, ifi->name); + return; + } + ifa.scope = scope & IADDR_SCOPE_MASK; + + DBG("KIF: IF%d(%s): %s IPA %I, flg %x, net %N, brd %I, opp %I\n", + ifi->index, ifi->name, + new ? "added" : "removed", + ifa.ip, ifa.flags, ifa.prefix, ifa.brd, ifa.opposite); + + if (new) + ifa_update(&ifa); + else + ifa_delete(&ifa); + + if (!scan) + if_end_partial_update(ifi); +} + +static void +nl_parse_addr6(struct ifaddrmsg *i, int scan, int new) +{ + struct rtattr *a[BIRD_IFA_MAX]; + struct iface *ifi; + u32 ifa_flags; + int scope; + + if (!nl_parse_attrs(IFA_RTA(i), ifa_attr_want6, a, sizeof(a))) + return; + + if (!a[IFA_ADDRESS]) + { + log(L_ERR "KIF: Malformed message received (missing IFA_ADDRESS)"); + return; + } + ifi = if_find_by_index(i->ifa_index); if (!ifi) { @@ -711,65 +806,50 @@ nl_parse_addr(struct nlmsghdr *h, int scan) return; } + if (a[IFA_FLAGS]) + ifa_flags = rta_get_u32(a[IFA_FLAGS]); + else + ifa_flags = i->ifa_flags; + + struct ifa ifa; bzero(&ifa, sizeof(ifa)); ifa.iface = ifi; if (ifa_flags & IFA_F_SECONDARY) ifa.flags |= IA_SECONDARY; -#ifdef IPV6 /* Ignore tentative addresses silently */ if (ifa_flags & IFA_F_TENTATIVE) return; -#endif /* IFA_LOCAL can be unset for IPv6 interfaces */ - memcpy(&ifa.ip, RTA_DATA(a[IFA_LOCAL] ? : a[IFA_ADDRESS]), sizeof(ifa.ip)); - ipa_ntoh(ifa.ip); - ifa.pxlen = i->ifa_prefixlen; - if (i->ifa_prefixlen > BITS_PER_IP_ADDRESS) + ifa.ip = rta_get_ipa(a[IFA_LOCAL] ? : a[IFA_ADDRESS]); + + if (i->ifa_prefixlen > IP6_MAX_PREFIX_LENGTH) { log(L_ERR "KIF: Invalid prefix length for interface %s: %d", ifi->name, i->ifa_prefixlen); new = 0; } - if (i->ifa_prefixlen == BITS_PER_IP_ADDRESS) + if (i->ifa_prefixlen == IP6_MAX_PREFIX_LENGTH) { - ip_addr addr; - memcpy(&addr, RTA_DATA(a[IFA_ADDRESS]), sizeof(addr)); - ipa_ntoh(addr); - ifa.prefix = ifa.brd = addr; + ifa.brd = rta_get_ipa(a[IFA_ADDRESS]); + net_fill_ip6(&ifa.prefix, rta_get_ip6(a[IFA_ADDRESS]), i->ifa_prefixlen); /* It is either a host address or a peer address */ - if (ipa_equal(ifa.ip, addr)) + if (ipa_equal(ifa.ip, ifa.brd)) ifa.flags |= IA_HOST; else { ifa.flags |= IA_PEER; - ifa.opposite = addr; + ifa.opposite = ifa.brd; } } else { - ip_addr netmask = ipa_mkmask(ifa.pxlen); - ifa.prefix = ipa_and(ifa.ip, netmask); - ifa.brd = ipa_or(ifa.ip, ipa_not(netmask)); - if (i->ifa_prefixlen == BITS_PER_IP_ADDRESS - 1) - ifa.opposite = ipa_opposite_m1(ifa.ip); + net_fill_ip6(&ifa.prefix, ipa_to_ip6(ifa.ip), i->ifa_prefixlen); + net_normalize(&ifa.prefix); -#ifndef IPV6 - if (i->ifa_prefixlen == BITS_PER_IP_ADDRESS - 2) - ifa.opposite = ipa_opposite_m2(ifa.ip); - - if ((ifi->flags & IF_BROADCAST) && a[IFA_BROADCAST]) - { - ip_addr xbrd; - memcpy(&xbrd, RTA_DATA(a[IFA_BROADCAST]), sizeof(xbrd)); - ipa_ntoh(xbrd); - if (ipa_equal(xbrd, ifa.prefix) || ipa_equal(xbrd, ifa.brd)) - ifa.brd = xbrd; - else if (ifi->flags & IF_TMP_DOWN) /* Complain only during the first scan */ - log(L_ERR "KIF: Invalid broadcast address %I for %s", xbrd, ifi->name); - } -#endif + if (i->ifa_prefixlen == IP6_MAX_PREFIX_LENGTH - 1) + ifa.opposite = ipa_opposite_m1(ifa.ip); } scope = ipa_classify(ifa.ip); @@ -780,10 +860,10 @@ nl_parse_addr(struct nlmsghdr *h, int scan) } ifa.scope = scope & IADDR_SCOPE_MASK; - DBG("KIF: IF%d(%s): %s IPA %I, flg %x, net %I/%d, brd %I, opp %I\n", + DBG("KIF: IF%d(%s): %s IPA %I, flg %x, net %N, brd %I, opp %I\n", ifi->index, ifi->name, new ? "added" : "removed", - ifa.ip, ifa.flags, ifa.prefix, ifa.pxlen, ifa.brd, ifa.opposite); + ifa.ip, ifa.flags, ifa.prefix, ifa.brd, ifa.opposite); if (new) ifa_update(&ifa); @@ -794,6 +874,26 @@ nl_parse_addr(struct nlmsghdr *h, int scan) if_end_partial_update(ifi); } +static void +nl_parse_addr(struct nlmsghdr *h, int scan) +{ + struct ifaddrmsg *i; + + if (!(i = nl_checkin(h, sizeof(*i)))) + return; + + int new = (h->nlmsg_type == RTM_NEWADDR); + + switch (i->ifa_family) + { + case AF_INET: + return nl_parse_addr4(i, scan, new); + + case AF_INET6: + return nl_parse_addr6(i, scan, new); + } +} + void kif_do_scan(struct kif_proto *p UNUSED) { @@ -808,7 +908,14 @@ kif_do_scan(struct kif_proto *p UNUSED) else log(L_DEBUG "nl_scan_ifaces: Unknown packet received (type=%d)", h->nlmsg_type); - nl_request_dump(BIRD_AF, RTM_GETADDR); + nl_request_dump(AF_INET, RTM_GETADDR); + while (h = nl_get_scan()) + if (h->nlmsg_type == RTM_NEWADDR || h->nlmsg_type == RTM_DELADDR) + nl_parse_addr(h, 1); + else + log(L_DEBUG "nl_scan_ifaces: Unknown packet received (type=%d)", h->nlmsg_type); + + nl_request_dump(AF_INET6, RTM_GETADDR); while (h = nl_get_scan()) if (h->nlmsg_type == RTM_NEWADDR || h->nlmsg_type == RTM_DELADDR) nl_parse_addr(h, 1); @@ -830,10 +937,10 @@ krt_table_id(struct krt_proto *p) static HASH(struct krt_proto) nl_table_map; -#define RTH_FN(k) u32_hash(k) -#define RTH_EQ(k1,k2) k1 == k2 -#define RTH_KEY(p) krt_table_id(p) -#define RTH_NEXT(p) p->sys.hash_next +#define RTH_KEY(p) p->af, krt_table_id(p) +#define RTH_NEXT(p) p->sys.hash_next +#define RTH_EQ(a1,i1,a2,i2) a1 == a2 && i1 == i2 +#define RTH_FN(a,i) a ^ u32_hash(i) #define RTH_REHASH rth_rehash #define RTH_PARAMS /8, *2, 2, 2, 6, 20 @@ -880,27 +987,31 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d eattr *ea; net *net = e->net; rta *a = e->attrs; + int bufsize = 128 + KRT_METRICS_MAX*8 + nh_bufsize(a->nexthops); u32 priority = 0; struct { struct nlmsghdr h; struct rtmsg r; - char buf[128 + KRT_METRICS_MAX*8 + nh_bufsize(a->nexthops)]; - } r; + char buf[0]; + } *r; + + int rsize = sizeof(*r) + bufsize; + r = alloca(rsize); - DBG("nl_send_route(%I/%d,op=%x)\n", net->n.prefix, net->n.pxlen, op); + DBG("nl_send_route(%N,op=%x)\n", net->n.addr, op); - bzero(&r.h, sizeof(r.h)); - bzero(&r.r, sizeof(r.r)); - r.h.nlmsg_type = op ? RTM_NEWROUTE : RTM_DELROUTE; - r.h.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg)); - r.h.nlmsg_flags = op | NLM_F_REQUEST | NLM_F_ACK; + bzero(&r->h, sizeof(r->h)); + bzero(&r->r, sizeof(r->r)); + r->h.nlmsg_type = op ? RTM_NEWROUTE : RTM_DELROUTE; + r->h.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg)); + r->h.nlmsg_flags = op | NLM_F_REQUEST | NLM_F_ACK; - r.r.rtm_family = BIRD_AF; - r.r.rtm_dst_len = net->n.pxlen; - r.r.rtm_protocol = RTPROT_BIRD; - r.r.rtm_scope = RT_SCOPE_NOWHERE; - nl_add_attr_ipa(&r.h, sizeof(r), RTA_DST, net->n.prefix); + r->r.rtm_family = p->af; + r->r.rtm_dst_len = net_pxlen(net->n.addr); + r->r.rtm_protocol = RTPROT_BIRD; + r->r.rtm_scope = RT_SCOPE_UNIVERSE; + nl_add_attr_ipa(&r->h, rsize, RTA_DST, net_prefix(net->n.addr)); /* * Strange behavior for RTM_DELROUTE: @@ -910,9 +1021,9 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d */ if (krt_table_id(p) < 256) - r.r.rtm_table = krt_table_id(p); + r->r.rtm_table = krt_table_id(p); else - nl_add_attr_u32(&r.h, sizeof(r), RTA_TABLE, krt_table_id(p)); + nl_add_attr_u32(&r->h, rsize, RTA_TABLE, krt_table_id(p)); if (a->source == RTS_DUMMY) priority = e->u.krt.metric; @@ -922,7 +1033,7 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d priority = ea->u.data; if (priority) - nl_add_attr_u32(&r.h, sizeof(r), RTA_PRIORITY, priority); + nl_add_attr_u32(&r->h, sizeof(r), RTA_PRIORITY, priority); /* For route delete, we do not specify remaining route attributes */ if (op == NL_OP_DELETE) @@ -930,15 +1041,15 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d /* Default scope is LINK for device routes, UNIVERSE otherwise */ if (ea = ea_find(eattrs, EA_KRT_SCOPE)) - r.r.rtm_scope = ea->u.data; + r->r.rtm_scope = ea->u.data; else - r.r.rtm_scope = (dest == RTD_DEVICE) ? RT_SCOPE_LINK : RT_SCOPE_UNIVERSE; + r->r.rtm_scope = (dest == RTD_DEVICE) ? RT_SCOPE_LINK : RT_SCOPE_UNIVERSE; if (ea = ea_find(eattrs, EA_KRT_PREFSRC)) - nl_add_attr_ipa(&r.h, sizeof(r), RTA_PREFSRC, *(ip_addr *)ea->u.ptr->data); + nl_add_attr_ipa(&r->h, rsize, RTA_PREFSRC, *(ip_addr *)ea->u.ptr->data); if (ea = ea_find(eattrs, EA_KRT_REALM)) - nl_add_attr_u32(&r.h, sizeof(r), RTA_FLOW, ea->u.data); + nl_add_attr_u32(&r->h, rsize, RTA_FLOW, ea->u.data); u32 metrics[KRT_METRICS_MAX]; @@ -953,7 +1064,7 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d } if (metrics[0]) - nl_add_metrics(&r.h, sizeof(r), metrics, KRT_METRICS_MAX); + nl_add_metrics(&r->h, rsize, metrics, KRT_METRICS_MAX); dest: @@ -961,26 +1072,26 @@ dest: switch (dest) { case RTD_ROUTER: - r.r.rtm_type = RTN_UNICAST; - nl_add_attr_u32(&r.h, sizeof(r), RTA_OIF, iface->index); - nl_add_attr_ipa(&r.h, sizeof(r), RTA_GATEWAY, gw); + r->r.rtm_type = RTN_UNICAST; + nl_add_attr_u32(&r->h, rsize, RTA_OIF, iface->index); + nl_add_attr_ipa(&r->h, rsize, RTA_GATEWAY, gw); break; case RTD_DEVICE: - r.r.rtm_type = RTN_UNICAST; - nl_add_attr_u32(&r.h, sizeof(r), RTA_OIF, iface->index); + r->r.rtm_type = RTN_UNICAST; + nl_add_attr_u32(&r->h, rsize, RTA_OIF, iface->index); break; case RTD_BLACKHOLE: - r.r.rtm_type = RTN_BLACKHOLE; + r->r.rtm_type = RTN_BLACKHOLE; break; case RTD_UNREACHABLE: - r.r.rtm_type = RTN_UNREACHABLE; + r->r.rtm_type = RTN_UNREACHABLE; break; case RTD_PROHIBIT: - r.r.rtm_type = RTN_PROHIBIT; + r->r.rtm_type = RTN_PROHIBIT; break; case RTD_MULTIPATH: - r.r.rtm_type = RTN_UNICAST; - nl_add_multipath(&r.h, sizeof(r), a->nexthops); + r->r.rtm_type = RTN_UNICAST; + nl_add_multipath(&r->h, rsize, a->nexthops); break; case RTD_NONE: break; @@ -989,7 +1100,7 @@ dest: } /* Ignore missing for DELETE */ - return nl_exchange(&r.h, (op == NL_OP_DELETE)); + return nl_exchange(&r->h, (op == NL_OP_DELETE)); } static inline int @@ -1137,9 +1248,9 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) struct rtattr *a[BIRD_RTA_MAX]; int new = h->nlmsg_type == RTM_NEWROUTE; - ip_addr dst = IPA_NONE; + net_addr dst; u32 oif = ~0; - u32 table; + u32 table_id; u32 priority = 0; u32 def_scope = RT_SCOPE_UNIVERSE; int src; @@ -1149,47 +1260,48 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) switch (i->rtm_family) { -#ifndef IPV6 - case AF_INET: - if (!nl_parse_attrs(RTM_RTA(i), rtm_attr_want4, a, sizeof(a))) - return; - break; -#else - case AF_INET6: - if (!nl_parse_attrs(RTM_RTA(i), rtm_attr_want6, a, sizeof(a))) - return; - break; -#endif - default: + case AF_INET: + if (!nl_parse_attrs(RTM_RTA(i), rtm_attr_want4, a, sizeof(a))) return; - } - if (a[RTA_DST]) - { - memcpy(&dst, RTA_DATA(a[RTA_DST]), sizeof(dst)); - ipa_ntoh(dst); + if (a[RTA_DST]) + net_fill_ip4(&dst, rta_get_ip4(a[RTA_DST]), i->rtm_dst_len); + else + net_fill_ip4(&dst, IP4_NONE, 0); + break; + + case AF_INET6: + if (!nl_parse_attrs(RTM_RTA(i), rtm_attr_want6, a, sizeof(a))) + return; + + if (a[RTA_DST]) + net_fill_ip6(&dst, rta_get_ip6(a[RTA_DST]), i->rtm_dst_len); + else + net_fill_ip6(&dst, IP6_NONE, 0); + break; + + default: + return; } if (a[RTA_OIF]) oif = rta_get_u32(a[RTA_OIF]); if (a[RTA_TABLE]) - table = rta_get_u32(a[RTA_TABLE]); + table_id = rta_get_u32(a[RTA_TABLE]); else - table = i->rtm_table; + table_id = i->rtm_table; - p = HASH_FIND(nl_table_map, RTH, table); /* Do we know this table? */ - DBG("KRT: Got %I/%d, type=%d, oif=%d, table=%d, prid=%d, proto=%s\n", dst, i->rtm_dst_len, i->rtm_type, oif, table, i->rtm_protocol, p ? p->p.name : "(none)"); + /* Do we know this table? */ + p = HASH_FIND(nl_table_map, RTH, i->rtm_family, table_id); if (!p) SKIP("unknown table %d\n", table); -#ifdef IPV6 if (a[RTA_IIF]) SKIP("IIF set\n"); -#else + if (i->rtm_tos != 0) /* We don't support TOS */ SKIP("TOS %02x\n", i->rtm_tos); -#endif if (s->scan && !new) SKIP("RTM_DELROUTE in scan\n"); @@ -1197,7 +1309,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) if (a[RTA_PRIORITY]) priority = rta_get_u32(a[RTA_PRIORITY]); - int c = ipa_classify_net(dst); + int c = net_classify(&dst); if ((c < 0) || !(c & IADDR_HOST) || ((c & IADDR_SCOPE_MASK) <= SCOPE_LINK)) SKIP("strange class/scope\n"); @@ -1225,7 +1337,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) src = KRT_SRC_ALIEN; } - net *net = net_get(p->p.table, dst, i->rtm_dst_len); + net *net = net_get(p->p.main_channel->table, &dst); if (s->net && !nl_mergable_route(s, net, p, priority, i->rtm_type)) nl_announce_route(s); @@ -1246,8 +1358,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) ra->nexthops = nl_parse_multipath(p, a[RTA_MULTIPATH]); if (!ra->nexthops) { - log(L_ERR "KRT: Received strange multipath route %I/%d", - net->n.prefix, net->n.pxlen); + log(L_ERR "KRT: Received strange multipath route %N", net->n.addr); return; } @@ -1257,30 +1368,26 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) ra->iface = if_find_by_index(oif); if (!ra->iface) { - log(L_ERR "KRT: Received route %I/%d with unknown ifindex %u", - net->n.prefix, net->n.pxlen, oif); + log(L_ERR "KRT: Received route %N with unknown ifindex %u", net->n.addr, oif); return; } if (a[RTA_GATEWAY]) { - neighbor *ng; ra->dest = RTD_ROUTER; - memcpy(&ra->gw, RTA_DATA(a[RTA_GATEWAY]), sizeof(ra->gw)); - ipa_ntoh(ra->gw); + ra->gw = rta_get_ipa(a[RTA_GATEWAY]); -#ifdef IPV6 /* Silently skip strange 6to4 routes */ - if (ipa_in_net(ra->gw, IPA_NONE, 96)) + const net_addr_ip6 sit = NET_ADDR_IP6(IP6_NONE, 96); + if ((i->rtm_family == AF_INET6) && ipa_in_netX(ra->gw, (net_addr *) &sit)) return; -#endif - ng = neigh_find2(&p->p, &ra->gw, ra->iface, - (i->rtm_flags & RTNH_F_ONLINK) ? NEF_ONLINK : 0); - if (!ng || (ng->scope == SCOPE_HOST)) + neighbor *nbr; + nbr = neigh_find2(&p->p, &ra->gw, ra->iface, + (i->rtm_flags & RTNH_F_ONLINK) ? NEF_ONLINK : 0); + if (!nbr || (nbr->scope == SCOPE_HOST)) { - log(L_ERR "KRT: Received route %I/%d with strange next-hop %I", - net->n.prefix, net->n.pxlen, ra->gw); + log(L_ERR "KRT: Received route %N with strange next-hop %I", net->n.addr, ra->gw); return; } } @@ -1321,9 +1428,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) if (a[RTA_PREFSRC]) { - ip_addr ps; - memcpy(&ps, RTA_DATA(a[RTA_PREFSRC]), sizeof(ps)); - ipa_ntoh(ps); + ip_addr ps = rta_get_ipa(a[RTA_PREFSRC]); ea_list *ea = lp_alloc(s->pool, sizeof(ea_list) + sizeof(eattr)); ea->next = ra->eattrs; @@ -1359,8 +1464,7 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) if (nl_parse_metrics(a[RTA_METRICS], metrics, ARRAY_SIZE(metrics)) < 0) { - log(L_ERR "KRT: Received route %I/%d with strange RTA_METRICS attribute", - net->n.prefix, net->n.pxlen); + log(L_ERR "KRT: Received route %N with strange RTA_METRICS attribute", net->n.addr); return; } @@ -1422,15 +1526,22 @@ krt_do_scan(struct krt_proto *p UNUSED) /* CONFIG_ALL_TABLES_AT_ONCE => p is NUL struct nlmsghdr *h; struct nl_parse_state s; - nl_parse_begin(&s, 1, krt_ecmp6(p)); - - nl_request_dump(BIRD_AF, RTM_GETROUTE); + nl_parse_begin(&s, 1, 0); + nl_request_dump(AF_INET, RTM_GETROUTE); while (h = nl_get_scan()) if (h->nlmsg_type == RTM_NEWROUTE || h->nlmsg_type == RTM_DELROUTE) nl_parse_route(&s, h); else log(L_DEBUG "nl_scan_fire: Unknown packet received (type=%d)", h->nlmsg_type); + nl_parse_end(&s); + nl_parse_begin(&s, 1, 1); + nl_request_dump(AF_INET6, RTM_GETROUTE); + while (h = nl_get_scan()) + if (h->nlmsg_type == RTM_NEWROUTE || h->nlmsg_type == RTM_DELROUTE) + nl_parse_route(&s, h); + else + log(L_DEBUG "nl_scan_fire: Unknown packet received (type=%d)", h->nlmsg_type); nl_parse_end(&s); } @@ -1552,11 +1663,10 @@ nl_open_async(void) bzero(&sa, sizeof(sa)); sa.nl_family = AF_NETLINK; -#ifdef IPV6 - sa.nl_groups = RTMGRP_LINK | RTMGRP_IPV6_IFADDR | RTMGRP_IPV6_ROUTE; -#else - sa.nl_groups = RTMGRP_LINK | RTMGRP_IPV4_IFADDR | RTMGRP_IPV4_ROUTE; -#endif + sa.nl_groups = RTMGRP_LINK | + RTMGRP_IPV4_IFADDR | RTMGRP_IPV4_ROUTE | + RTMGRP_IPV6_IFADDR | RTMGRP_IPV6_ROUTE; + if (bind(fd, (struct sockaddr *) &sa, sizeof(sa)) < 0) { log(L_ERR "Unable to bind asynchronous rtnetlink socket: %m"); @@ -1590,7 +1700,7 @@ krt_sys_io_init(void) int krt_sys_start(struct krt_proto *p) { - struct krt_proto *old = HASH_FIND(nl_table_map, RTH, krt_table_id(p)); + struct krt_proto *old = HASH_FIND(nl_table_map, RTH, p->af, krt_table_id(p)); if (old) { diff --git a/sysdep/unix/Makefile b/sysdep/unix/Makefile new file mode 100644 index 00000000..c5d55431 --- /dev/null +++ b/sysdep/unix/Makefile @@ -0,0 +1,5 @@ +src := io.c krt.c log.c main.c random.c +obj := $(src-o-files) +$(all-daemon) +$(cf-local) +$(conf-y-targets): $(s)krt.Y diff --git a/sysdep/unix/Modules b/sysdep/unix/Modules deleted file mode 100644 index 2c6514df..00000000 --- a/sysdep/unix/Modules +++ /dev/null @@ -1,12 +0,0 @@ -log.c -main.c -timer.h -io.c -unix.h -endian.h -config.Y -random.c - -krt.c -krt.h -krt.Y diff --git a/sysdep/unix/config.Y b/sysdep/unix/config.Y index d6ab8cab..ebadd454 100644 --- a/sysdep/unix/config.Y +++ b/sysdep/unix/config.Y @@ -8,7 +8,7 @@ CF_HDR -#include "lib/unix.h" +#include "sysdep/unix/unix.h" #include <stdio.h> CF_DECLS @@ -92,6 +92,7 @@ timeformat_which: | PROTOCOL { $$ = &new_config->tf_proto; } | BASE { $$ = &new_config->tf_base; } | LOG { $$ = &new_config->tf_log; } + ; timeformat_spec: timeformat_which TEXT { *$1 = (struct timeformat){$2, NULL, 0}; } diff --git a/sysdep/unix/io.c b/sysdep/unix/io.c index 3ceadd98..e90964c1 100644 --- a/sysdep/unix/io.c +++ b/sysdep/unix/io.c @@ -32,14 +32,14 @@ #include "nest/bird.h" #include "lib/lists.h" #include "lib/resource.h" -#include "lib/timer.h" +#include "sysdep/unix/timer.h" #include "lib/socket.h" #include "lib/event.h" #include "lib/string.h" #include "nest/iface.h" -#include "lib/unix.h" -#include "lib/sysio.h" +#include "sysdep/unix/unix.h" +#include CONFIG_INCLUDE_SYSIO_H /* Maximum number of calls of tx handler for one socket in one * poll iteration. Should be small enough to not monopolize CPU by @@ -1253,7 +1253,7 @@ sk_setup(sock *s) if (sk_is_ipv6(s)) { - if (s->flags & SKF_V6ONLY) + if (s->type != SK_IP) if (setsockopt(fd, SOL_IPV6, IPV6_V6ONLY, &y, sizeof(y)) < 0) ERR("IPV6_V6ONLY"); @@ -1319,8 +1319,8 @@ sk_passive_connected(sock *s, int type) sock *t = sk_new(s->pool); t->type = type; - t->fd = fd; t->af = s->af; + t->fd = fd; t->ttl = s->ttl; t->tos = s->tos; t->rbsize = s->rbsize; @@ -1367,13 +1367,46 @@ sk_passive_connected(sock *s, int type) int sk_open(sock *s) { - int af = BIRD_AF; + int af = AF_UNSPEC; int fd = -1; int do_bind = 0; int bind_port = 0; ip_addr bind_addr = IPA_NONE; sockaddr sa; + if (s->type <= SK_IP) + { + /* + * For TCP/IP sockets, Address family (IPv4 or IPv6) can be specified either + * explicitly (SK_IPV4 or SK_IPV6) or implicitly (based on saddr, daddr). + * But the specifications have to be consistent. + */ + + switch (s->subtype) + { + case 0: + ASSERT(ipa_zero(s->saddr) || ipa_zero(s->daddr) || + (ipa_is_ip4(s->saddr) == ipa_is_ip4(s->daddr))); + af = (ipa_is_ip4(s->saddr) || ipa_is_ip4(s->daddr)) ? AF_INET : AF_INET6; + break; + + case SK_IPV4: + ASSERT(ipa_zero(s->saddr) || ipa_is_ip4(s->saddr)); + ASSERT(ipa_zero(s->daddr) || ipa_is_ip4(s->daddr)); + af = AF_INET; + break; + + case SK_IPV6: + ASSERT(ipa_zero(s->saddr) || !ipa_is_ip4(s->saddr)); + ASSERT(ipa_zero(s->daddr) || !ipa_is_ip4(s->daddr)); + af = AF_INET6; + break; + + default: + bug("Invalid subtype %d", s->subtype); + } + } + switch (s->type) { case SK_TCP_ACTIVE: @@ -1441,7 +1474,7 @@ sk_open(sock *s) if (sk_set_high_port(s) < 0) log(L_WARN "Socket error: %s%#m", s->err); - sockaddr_fill(&sa, af, bind_addr, s->iface, bind_port); + sockaddr_fill(&sa, s->af, bind_addr, s->iface, bind_port); if (bind(fd, &sa.sa, SA_LEN(sa)) < 0) ERR2("bind"); } @@ -1453,7 +1486,7 @@ sk_open(sock *s) switch (s->type) { case SK_TCP_ACTIVE: - sockaddr_fill(&sa, af, s->daddr, s->iface, s->dport); + sockaddr_fill(&sa, s->af, s->daddr, s->iface, s->dport); if (connect(fd, &sa.sa, SA_LEN(sa)) >= 0) sk_tcp_connected(s); else if (errno != EINTR && errno != EAGAIN && errno != EINPROGRESS && @@ -1853,6 +1886,12 @@ sk_write(sock *s) } } +int sk_is_ipv4(sock *s) +{ return s->af == AF_INET; } + +int sk_is_ipv6(sock *s) +{ return s->af == AF_INET6; } + void sk_err(sock *s, int revents) { diff --git a/sysdep/unix/krt.Y b/sysdep/unix/krt.Y index 6fe39fa9..33dc4a19 100644 --- a/sysdep/unix/krt.Y +++ b/sysdep/unix/krt.Y @@ -8,13 +8,23 @@ CF_HDR -#include "lib/krt.h" +#include "sysdep/unix/krt.h" CF_DEFINES #define THIS_KRT ((struct krt_config *) this_proto) #define THIS_KIF ((struct kif_config *) this_proto) +static void +krt_set_merge_paths(struct channel_config *cc, uint merge, uint limit) +{ + if ((limit <= 0) || (limit > 255)) + cf_error("Merge paths limit must be in range 1-255"); + + cc->ra_mode = merge ? RA_MERGED : RA_OPTIMAL; + cc->merge_limit = limit; +} + CF_DECLS CF_KEYWORDS(KERNEL, PERSIST, SCAN, TIME, LEARN, DEVICE, ROUTES, GRACEFUL, RESTART, KRT_SOURCE, KRT_METRIC, MERGE, PATHS) @@ -27,11 +37,12 @@ CF_GRAMMAR CF_ADDTO(proto, kern_proto '}') -kern_proto_start: proto_start KERNEL { this_proto = krt_init_config($1); } +kern_proto_start: proto_start KERNEL { + this_proto = krt_init_config($1); +} ; CF_ADDTO(kern_proto, kern_proto_start proto_name '{') -CF_ADDTO(kern_proto, kern_proto proto_item ';') CF_ADDTO(kern_proto, kern_proto kern_item ';') kern_mp_limit: @@ -40,7 +51,9 @@ kern_mp_limit: ; kern_item: - PERSIST bool { THIS_KRT->persist = $2; } + proto_item + | proto_channel { this_proto->net_type = $1->net_type; } + | PERSIST bool { THIS_KRT->persist = $2; } | SCAN TIME expr { /* Scan time of 0 means scan on startup only */ THIS_KRT->scan_time = $3; @@ -55,7 +68,7 @@ kern_item: | DEVICE ROUTES bool { THIS_KRT->devroutes = $3; } | GRACEFUL RESTART bool { THIS_KRT->graceful_restart = $3; } | MERGE PATHS bool kern_mp_limit { - THIS_KRT->merge_paths = $3 ? $4 : 0; + krt_set_merge_paths(this_channel, $3, $4); #ifndef KRT_ALLOW_MERGE_PATHS if ($3) cf_error("Path merging not supported on this platform"); @@ -71,19 +84,17 @@ kif_proto_start: proto_start DEVICE { this_proto = kif_init_config($1); } ; CF_ADDTO(kif_proto, kif_proto_start proto_name '{') -CF_ADDTO(kif_proto, kif_proto proto_item ';') CF_ADDTO(kif_proto, kif_proto kif_item ';') kif_item: - SCAN TIME expr { + proto_item + | SCAN TIME expr { /* Scan time of 0 means scan on startup only */ THIS_KIF->scan_time = $3; } - | PRIMARY text_or_none prefix_or_ipa { + | PRIMARY opttext net_or_ipa { struct kif_primary_item *kpi = cfg_alloc(sizeof (struct kif_primary_item)); - kpi->pattern = $2; - kpi->prefix = $3.addr; - kpi->pxlen = $3.len; + kpi->addr = $3; add_tail(&THIS_KIF->primary, &kpi->n); } ; diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c index ef98cb3a..d4cb964e 100644 --- a/sysdep/unix/krt.c +++ b/sysdep/unix/krt.c @@ -56,7 +56,7 @@ #include "nest/route.h" #include "nest/protocol.h" #include "filter/filter.h" -#include "lib/timer.h" +#include "sysdep/unix/timer.h" #include "conf/conf.h" #include "lib/string.h" @@ -131,14 +131,14 @@ prefer_addr(struct ifa *a, struct ifa *b) } static inline struct ifa * -find_preferred_ifa(struct iface *i, ip_addr prefix, ip_addr mask) +find_preferred_ifa(struct iface *i, const net_addr *n) { struct ifa *a, *b = NULL; WALK_LIST(a, i->addrs) { if (!(a->flags & IA_SECONDARY) && - ipa_equal(ipa_and(a->ip, mask), prefix) && + (!n || ipa_in_netX(a->ip, n)) && (!b || prefer_addr(a, b))) b = a; } @@ -156,21 +156,21 @@ kif_choose_primary(struct iface *i) WALK_LIST(it, cf->primary) { if (!it->pattern || patmatch(it->pattern, i->name)) - if (a = find_preferred_ifa(i, it->prefix, ipa_mkmask(it->pxlen))) + if (a = find_preferred_ifa(i, &it->addr)) return a; } if (a = kif_get_primary_ip(i)) return a; - return find_preferred_ifa(i, IPA_NONE, IPA_NONE); + return find_preferred_ifa(i, NULL); } static struct proto * kif_init(struct proto_config *c) { - struct kif_proto *p = proto_new(c, sizeof(struct kif_proto)); + struct kif_proto *p = proto_new(c); kif_sys_init(p); return &p->p; @@ -266,9 +266,6 @@ kif_copy_config(struct proto_config *dest, struct proto_config *src) struct kif_config *d = (struct kif_config *) dest; struct kif_config *s = (struct kif_config *) src; - /* Shallow copy of everything (just scan_time currently) */ - proto_copy_rest(dest, src, sizeof(struct kif_config)); - /* Copy primary addr list */ cfg_copy_list(&d->primary, &s->primary, sizeof(struct kif_primary_item)); @@ -280,7 +277,7 @@ kif_copy_config(struct proto_config *dest, struct proto_config *src) struct protocol proto_unix_iface = { .name = "Device", .template = "device%d", - .preference = DEF_PREF_DIRECT, + .proto_size = sizeof(struct kif_proto), .config_size = sizeof(struct kif_config), .preconfig = kif_preconfig, .init = kif_init, @@ -298,14 +295,14 @@ static inline void krt_trace_in(struct krt_proto *p, rte *e, char *msg) { if (p->p.debug & D_PACKETS) - log(L_TRACE "%s: %I/%d: %s", p->p.name, e->net->n.prefix, e->net->n.pxlen, msg); + log(L_TRACE "%s: %N: %s", p->p.name, e->net->n.addr, msg); } static inline void krt_trace_in_rl(struct tbf *f, struct krt_proto *p, rte *e, char *msg) { if (p->p.debug & D_PACKETS) - log_rl(f, L_TRACE "%s: %I/%d: %s", p->p.name, e->net->n.prefix, e->net->n.pxlen, msg); + log_rl(f, L_TRACE "%s: %N: %s", p->p.name, e->net->n.addr, msg); } /* @@ -348,19 +345,15 @@ krt_learn_announce_update(struct krt_proto *p, rte *e) net *n = e->net; rta *aa = rta_clone(e->attrs); rte *ee = rte_get_temp(aa); - net *nn = net_get(p->p.table, n->n.prefix, n->n.pxlen); - ee->net = nn; ee->pflags = 0; - ee->pref = p->p.preference; ee->u.krt = e->u.krt; - rte_update(&p->p, nn, ee); + rte_update(&p->p, n->n.addr, ee); } static void krt_learn_announce_delete(struct krt_proto *p, net *n) { - n = net_find(p->p.table, n->n.prefix, n->n.pxlen); - rte_update(&p->p, n, NULL); + rte_update(&p->p, n->n.addr, NULL); } /* Called when alien route is discovered during scan */ @@ -368,7 +361,7 @@ static void krt_learn_scan(struct krt_proto *p, rte *e) { net *n0 = e->net; - net *n = net_get(&p->krt_table, n0->n.prefix, n0->n.pxlen); + net *n = net_get(&p->krt_table, n0->n.addr); rte *m, **mm; e->attrs = rta_lookup(e->attrs); @@ -412,9 +405,8 @@ krt_learn_prune(struct krt_proto *p) FIB_ITERATE_INIT(&fit, fib); again: - FIB_ITERATE_START(fib, &fit, f) + FIB_ITERATE_START(fib, &fit, net, n) { - net *n = (net *) f; rte *e, **ee, *best, **pbest, *old_best; /* @@ -455,8 +447,8 @@ again: if (old_best) krt_learn_announce_delete(p, n); - FIB_ITERATE_PUT(&fit, f); - fib_delete(fib, f); + FIB_ITERATE_PUT(&fit); + fib_delete(fib, n); goto again; } @@ -473,7 +465,7 @@ again: else DBG("%I/%d: uptodate (metric=%d)\n", n->n.prefix, n->n.pxlen, best->u.krt.metric); } - FIB_ITERATE_END(f); + FIB_ITERATE_END; p->reload = 0; } @@ -482,7 +474,7 @@ static void krt_learn_async(struct krt_proto *p, rte *e, int new) { net *n0 = e->net; - net *n = net_get(&p->krt_table, n0->n.prefix, n0->n.pxlen); + net *n = net_get(&p->krt_table, n0->n.addr); rte *g, **gg, *best, **bestp, *old_best; e->attrs = rta_lookup(e->attrs); @@ -588,12 +580,11 @@ krt_dump_attrs(rte *e) static void krt_flush_routes(struct krt_proto *p) { - struct rtable *t = p->p.table; + struct rtable *t = p->p.main_channel->table; KRT_TRACE(p, D_EVENTS, "Flushing kernel routes"); - FIB_WALK(&t->fib, f) + FIB_WALK(&t->fib, net, n) { - net *n = (net *) f; rte *e = n->routes; if (rte_is_valid(e) && (n->n.flags & KRF_INSTALLED)) { @@ -608,12 +599,12 @@ krt_flush_routes(struct krt_proto *p) static struct rte * krt_export_net(struct krt_proto *p, net *net, rte **rt_free, ea_list **tmpa) { - struct announce_hook *ah = p->p.main_ahook; - struct filter *filter = ah->out_filter; + struct channel *c = p->p.main_channel; + struct filter *filter = c->out_filter; rte *rt; - if (p->p.accept_ra_types == RA_MERGED) - return rt_export_merged(ah, net, rt_free, tmpa, krt_filter_lp, 1); + if (c->ra_mode == RA_MERGED) + return rt_export_merged(c, net, rt_free, tmpa, krt_filter_lp, 1); rt = net->routes; *rt_free = NULL; @@ -760,13 +751,12 @@ krt_got_route(struct krt_proto *p, rte *e) static void krt_prune(struct krt_proto *p) { - struct rtable *t = p->p.table; + struct rtable *t = p->p.main_channel->table; KRT_TRACE(p, D_EVENTS, "Pruning table %s", t->name); - FIB_WALK(&t->fib, f) + FIB_WALK(&t->fib, net, n) { - net *n = (net *) f; - int verdict = f->flags & KRF_VERDICT_MASK; + int verdict = n->n.flags & KRF_VERDICT_MASK; rte *new, *old, *rt_free = NULL; ea_list *tmpa = NULL; @@ -795,7 +785,7 @@ krt_prune(struct krt_proto *p) switch (verdict) { case KRF_CREATE: - if (new && (f->flags & KRF_INSTALLED)) + if (new && (n->n.flags & KRF_INSTALLED)) { krt_trace_in(p, new, "reinstalling"); krt_replace_rte(p, n, new, NULL, tmpa); @@ -822,7 +812,7 @@ krt_prune(struct krt_proto *p) if (rt_free) rte_free(rt_free); lp_flush(krt_filter_lp); - f->flags &= ~KRF_VERDICT_MASK; + n->n.flags &= ~KRF_VERDICT_MASK; } FIB_WALK_END; @@ -1033,7 +1023,7 @@ krt_import_control(struct proto *P, rte **new, ea_list **attrs, struct linpool * } static void -krt_rt_notify(struct proto *P, struct rtable *table UNUSED, net *net, +krt_rt_notify(struct proto *P, struct channel *ch UNUSED, net *net, rte *new, rte *old, struct ea_list *eattrs) { struct krt_proto *p = (struct krt_proto *) P; @@ -1067,10 +1057,10 @@ krt_if_notify(struct proto *P, uint flags, struct iface *iface UNUSED) krt_scan_timer_kick(p); } -static int -krt_reload_routes(struct proto *P) +static void +krt_reload_routes(struct channel *C) { - struct krt_proto *p = (struct krt_proto *) P; + struct krt_proto *p = (void *) C->proto; /* Although we keep learned routes in krt_table, we rather schedule a scan */ @@ -1079,14 +1069,12 @@ krt_reload_routes(struct proto *P) p->reload = 1; krt_scan_timer_kick(p); } - - return 1; } static void -krt_feed_end(struct proto *P) +krt_feed_end(struct channel *C) { - struct krt_proto *p = (struct krt_proto *) P; + struct krt_proto *p = (void *) C->proto; p->ready = 1; krt_scan_timer_kick(p); @@ -1107,14 +1095,42 @@ krt_rte_same(rte *a, rte *b) struct krt_config *krt_cf; +static void +krt_preconfig(struct protocol *P UNUSED, struct config *c) +{ + krt_cf = NULL; + krt_sys_preconfig(c); +} + +static void +krt_postconfig(struct proto_config *CF) +{ + struct krt_config *cf = (void *) CF; + + if (EMPTY_LIST(CF->channels)) + cf_error("Channel not specified"); + +#ifdef CONFIG_ALL_TABLES_AT_ONCE + if (krt_cf->scan_time != cf->scan_time) + cf_error("All kernel syncers must use the same table scan interval"); +#endif + + struct rtable_config *tab = proto_cf_main_channel(CF)->table; + if (tab->krt_attached) + cf_error("Kernel syncer (%s) already attached to table %s", tab->krt_attached->name, tab->name); + tab->krt_attached = CF; + + krt_sys_postconfig(cf); +} + static struct proto * -krt_init(struct proto_config *C) +krt_init(struct proto_config *CF) { - struct krt_proto *p = proto_new(C, sizeof(struct krt_proto)); - struct krt_config *c = (struct krt_config *) C; + struct krt_proto *p = proto_new(CF); + // struct krt_config *cf = (void *) CF; + + p->p.main_channel = proto_add_channel(&p->p, proto_cf_main_channel(CF)); - p->p.accept_ra_types = c->merge_paths ? RA_MERGED : RA_OPTIMAL; - p->p.merge_limit = c->merge_paths; p->p.import_control = krt_import_control; p->p.rt_notify = krt_rt_notify; p->p.if_notify = krt_if_notify; @@ -1133,6 +1149,13 @@ krt_start(struct proto *P) { struct krt_proto *p = (struct krt_proto *) P; + switch (p->p.net_type) + { + case NET_IP4: p->af = AF_INET; break; + case NET_IP6: p->af = AF_INET6; break; + default: ASSERT(0); + } + add_tail(&krt_proto_list, &p->krt_node); #ifdef KRT_ALLOW_LEARN @@ -1147,8 +1170,8 @@ krt_start(struct proto *P) krt_scan_timer_start(p); - if (P->gr_recovery && KRT_CF->graceful_restart) - P->gr_wait = 1; + if (p->p.gr_recovery && KRT_CF->graceful_restart) + p->p.main_channel->gr_wait = 1; return PS_UP; } @@ -1177,40 +1200,19 @@ krt_shutdown(struct proto *P) } static int -krt_reconfigure(struct proto *p, struct proto_config *new) +krt_reconfigure(struct proto *p, struct proto_config *CF) { - struct krt_config *o = (struct krt_config *) p->cf; - struct krt_config *n = (struct krt_config *) new; + struct krt_config *o = (void *) p->cf; + struct krt_config *n = (void *) CF; + + if (!proto_configure_channel(p, &p->main_channel, proto_cf_main_channel(CF))) + return 0; if (!krt_sys_reconfigure((struct krt_proto *) p, n, o)) return 0; /* persist, graceful restart need not be the same */ - return o->scan_time == n->scan_time && o->learn == n->learn && - o->devroutes == n->devroutes && o->merge_paths == n->merge_paths; -} - -static void -krt_preconfig(struct protocol *P UNUSED, struct config *c) -{ - krt_cf = NULL; - krt_sys_preconfig(c); -} - -static void -krt_postconfig(struct proto_config *C) -{ - struct krt_config *c = (struct krt_config *) C; - -#ifdef CONFIG_ALL_TABLES_AT_ONCE - if (krt_cf->scan_time != c->scan_time) - cf_error("All kernel syncers must use the same table scan interval"); -#endif - - if (C->table->krt_attached) - cf_error("Kernel syncer (%s) already attached to table %s", C->table->krt_attached->name, C->table->name); - C->table->krt_attached = C; - krt_sys_postconfig(c); + return o->scan_time == n->scan_time && o->learn == n->learn && o->devroutes == n->devroutes; } struct proto_config * @@ -1234,9 +1236,6 @@ krt_copy_config(struct proto_config *dest, struct proto_config *src) struct krt_config *d = (struct krt_config *) dest; struct krt_config *s = (struct krt_config *) src; - /* Shallow copy of everything */ - proto_copy_rest(dest, src, sizeof(struct krt_config)); - /* Fix sysdep parts */ krt_sys_copy_config(d, s); } @@ -1265,6 +1264,8 @@ struct protocol proto_unix_kernel = { .template = "kernel%d", .attr_class = EAP_KRT, .preference = DEF_PREF_INHERITED, + .channel_mask = NB_IP, + .proto_size = sizeof(struct krt_proto), .config_size = sizeof(struct krt_config), .preconfig = krt_preconfig, .postconfig = krt_postconfig, diff --git a/sysdep/unix/krt.h b/sysdep/unix/krt.h index d4a8717e..cb404de3 100644 --- a/sysdep/unix/krt.h +++ b/sysdep/unix/krt.h @@ -15,7 +15,8 @@ struct krt_proto; struct kif_config; struct kif_proto; -#include "lib/krt-sys.h" +#include "sysdep/config.h" +#include CONFIG_INCLUDE_KRTSYS_H /* Flags stored in net->n.flags, rest are in nest/route.h */ @@ -49,7 +50,6 @@ struct krt_config { int learn; /* Learn routes from other sources */ int devroutes; /* Allow export of device routes */ int graceful_restart; /* Regard graceful restart recovery */ - int merge_paths; /* Exported routes are merged for ECMP */ }; struct krt_proto { @@ -65,6 +65,7 @@ struct krt_proto { #endif node krt_node; /* Node in krt_proto_list */ + byte af; /* Kernel address family (AF_*) */ byte ready; /* Initial feed has been finished */ byte initialized; /* First scan has been finished */ byte reload; /* Next scan is doing reload */ @@ -96,8 +97,7 @@ extern struct protocol proto_unix_iface; struct kif_primary_item { node n; byte *pattern; - ip_addr prefix; - int pxlen; + net_addr addr; }; struct kif_config { @@ -112,7 +112,7 @@ struct kif_proto { struct kif_state sys; /* Sysdep state */ }; -struct kif_proto *kif_proto; +extern struct kif_proto *kif_proto; #define KIF_CF ((struct kif_config *)p->p.cf) diff --git a/sysdep/unix/log.c b/sysdep/unix/log.c index 1fd64426..e5c5e74e 100644 --- a/sysdep/unix/log.c +++ b/sysdep/unix/log.c @@ -27,7 +27,7 @@ #include "nest/mrtdump.h" #include "lib/string.h" #include "lib/lists.h" -#include "lib/unix.h" +#include "sysdep/unix/unix.h" static FILE *dbgf; static list *current_log_list; diff --git a/sysdep/unix/main.c b/sysdep/unix/main.c index a7989920..9594269d 100644 --- a/sysdep/unix/main.c +++ b/sysdep/unix/main.c @@ -69,7 +69,7 @@ async_dump(void) */ #ifdef CONFIG_RESTRICTED_PRIVILEGES -#include "lib/syspriv.h" +#include CONFIG_INCLUDE_SYSPRIV_H #else static inline void @@ -569,6 +569,10 @@ sysdep_shutdown_done(void) * Signals */ +volatile int async_config_flag; +volatile int async_dump_flag; +volatile int async_shutdown_flag; + static void handle_sighup(int sig UNUSED) { @@ -818,7 +822,7 @@ main(int argc, char **argv) io_init(); rt_init(); if_init(); - roa_init(); +// roa_init(); config_init(); uid_t use_uid = get_uid(use_user); diff --git a/sysdep/unix/unix.h b/sysdep/unix/unix.h index 4e0ff841..dcaab729 100644 --- a/sysdep/unix/unix.h +++ b/sysdep/unix/unix.h @@ -47,14 +47,6 @@ typedef struct sockaddr_bird { } sockaddr; -#ifdef IPV6 -#define BIRD_AF AF_INET6 -#define ipa_from_sa(x) ipa_from_sa6(x) -#else -#define BIRD_AF AF_INET -#define ipa_from_sa(x) ipa_from_sa4(x) -#endif - /* This is sloppy hack, it should be detected by configure script */ /* Linux systems have it defined so this is definition for BSD systems */ @@ -75,17 +67,21 @@ static inline ip_addr ipa_from_sa4(sockaddr *sa) static inline ip_addr ipa_from_sa6(sockaddr *sa) { return ipa_from_in6(((struct sockaddr_in6 *) sa)->sin6_addr); } +static inline ip_addr ipa_from_sa(sockaddr *sa) +{ + switch (sa->sa.sa_family) + { + case AF_INET: return ipa_from_sa4(sa); + case AF_INET6: return ipa_from_sa6(sa); + default: return IPA_NONE; + } +} + static inline struct in_addr ipa_to_in4(ip_addr a) { return (struct in_addr) { htonl(ipa_to_u32(a)) }; } -#ifdef IPV6 static inline struct in6_addr ipa_to_in6(ip_addr a) { return (struct in6_addr) { .s6_addr32 = { htonl(_I0(a)), htonl(_I1(a)), htonl(_I2(a)), htonl(_I3(a)) } }; } -#else -/* Temporary dummy */ -static inline struct in6_addr ipa_to_in6(ip_addr a) -{ return (struct in6_addr) { .s6_addr32 = { 0, 0, 0, 0 } }; } -#endif void sockaddr_fill(sockaddr *sa, int af, ip_addr a, struct iface *ifa, uint port); int sockaddr_read(sockaddr *sa, int af, ip_addr *a, struct iface **ifa, uint *port); @@ -95,9 +91,9 @@ int sockaddr_read(sockaddr *sa, int af, ip_addr *a, struct iface **ifa, uint *po #define SUN_LEN(ptr) ((size_t) (((struct sockaddr_un *) 0)->sun_path) + strlen ((ptr)->sun_path)) #endif -volatile int async_config_flag; -volatile int async_dump_flag; -volatile int async_shutdown_flag; +extern volatile int async_config_flag; +extern volatile int async_dump_flag; +extern volatile int async_shutdown_flag; void io_init(void); void io_loop(void); @@ -106,7 +102,6 @@ int sk_open_unix(struct birdsock *s, char *name); void *tracked_fopen(struct pool *, char *name, char *mode); void test_old_bird(char *path); - /* krt.c bits */ void krt_io_init(void); |