May 2017 - Bird-users - Mailing Lists

[PATCH 1/4] Core changes to support SADR
by Dean Luga 23 May '17

23 May '17

From: dean <dluga93(a)gmail.com> This patch adds a new network of type NET_SADR_IP6, including new structures, constants, and switch cases for the new network type. Some existing functions are duplicates to handle the new network type, and netlink can now handle SADR routes. The net_route_sadr_ip6 function is a bit of a hack. Routing in SADR is supposed to match the destination address first, then out of the entries with that dst, it should choose the most specific matching src address. For the destination-only part, I use a net_addr_ip6 as the destination prefix. Its type is changed to NET_SADR_IP6 to satisfy the assertion in fib_find. The fib_find function checks the length of the prefix to distinguish between destination-only or full SADR searches. --- lib/net.c | 34 ++++++++++++++++++++++++++ lib/net.h | 66 +++++++++++++++++++++++++++++++++++++++++++++++--- nest/rt-fib.c | 16 ++++++++++++ nest/rt-table.c | 35 +++++++++++++++++++++++++- sysdep/linux/netlink.c | 26 ++++++++++++++++++-- sysdep/unix/krt.c | 4 ++- 6 files changed, 174 insertions(+), 7 deletions(-) diff --git a/lib/net.c b/lib/net.c index a00ff27..41084fe 100644 --- a/lib/net.c +++ b/lib/net.c @@ -15,6 +15,7 @@ const char * const net_label[] = { [NET_FLOW4] = "flow4", [NET_FLOW6] = "flow6", [NET_MPLS] = "mpls", + [NET_SADR_IP6] = "sadr6", }; const u16 net_addr_length[] = { @@ -27,6 +28,7 @@ const u16 net_addr_length[] = { [NET_FLOW4] = 0, [NET_FLOW6] = 0, [NET_MPLS] = sizeof(net_addr_mpls), + [NET_SADR_IP6]= sizeof(net_addr_sadr_ip6), }; const u8 net_max_prefix_length[] = { @@ -39,6 +41,7 @@ const u8 net_max_prefix_length[] = { [NET_FLOW4] = IP4_MAX_PREFIX_LENGTH, [NET_FLOW6] = IP6_MAX_PREFIX_LENGTH, [NET_MPLS] = 0, + [NET_SADR_IP6]= IP6_MAX_PREFIX_LENGTH, }; const u16 net_max_text_length[] = { @@ -51,6 +54,7 @@ const u16 net_max_text_length[] = { [NET_FLOW4] = 0, /* "flow4 { ... }" */ [NET_FLOW6] = 0, /* "flow6 { ... }" */ [NET_MPLS] = 7, /* "1048575" */ + [NET_SADR_IP6]= 92, /* "ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128 from ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff/128" */ }; @@ -104,6 +108,8 @@ net_format(const net_addr *N, char *buf, int buflen) return flow6_net_format(buf, buflen, &n->flow6); case NET_MPLS: return bsnprintf(buf, buflen, "%u", n->mpls.label); + case NET_SADR_IP6: + return bsnprintf(buf, buflen, "%I6/%d from %I6/%d", n->sadr_ip6.dst_prefix, n->sadr_ip6.dst_pxlen, n->sadr_ip6.src_prefix, n->sadr_ip6.src_pxlen); } bug("unknown network type"); @@ -126,6 +132,9 @@ net_pxmask(const net_addr *a) case NET_FLOW6: return ipa_from_ip6(ip6_mkmask(net6_pxlen(a))); + case NET_SADR_IP6: + return ipa_from_ip6(ip6_mkmask(net6_sadr_dst_pxlen(a))); + case NET_MPLS: default: return IPA_NONE; @@ -133,6 +142,16 @@ net_pxmask(const net_addr *a) } int +net_compare_sadr_ip6(const net_addr_sadr_ip6 *a, const net_addr_sadr_ip6 *b) +{ + int dst_cmp = ip6_compare(a->dst_prefix, b->dst_prefix) ?: uint_cmp(a->dst_pxlen, b->dst_pxlen); + if (dst_cmp) + return dst_cmp; + else + return ip6_compare(a->src_prefix, b->src_prefix) ?: uint_cmp(a->src_pxlen, b->src_pxlen); +} + +int net_compare(const net_addr *a, const net_addr *b) { if (a->type != b->type) @@ -158,6 +177,8 @@ net_compare(const net_addr *a, const net_addr *b) return net_compare_flow6((const net_addr_flow6 *) a, (const net_addr_flow6 *) b); case NET_MPLS: return net_compare_mpls((const net_addr_mpls *) a, (const net_addr_mpls *) b); + case NET_SADR_IP6: + return net_compare_sadr_ip6((const net_addr_sadr_ip6 *) a, (const net_addr_sadr_ip6 *) b); } return 0; } @@ -178,6 +199,7 @@ net_hash(const net_addr *n) case NET_FLOW4: return NET_HASH(n, flow4); case NET_FLOW6: return NET_HASH(n, flow6); case NET_MPLS: return NET_HASH(n, mpls); + case NET_SADR_IP6: return NET_HASH(n, ip6); default: bug("invalid type"); } } @@ -199,6 +221,7 @@ net_validate(const net_addr *n) case NET_FLOW4: return NET_VALIDATE(n, flow4); case NET_FLOW6: return NET_VALIDATE(n, flow6); case NET_MPLS: return NET_VALIDATE(n, mpls); + case NET_SADR_IP6: return NET_VALIDATE(n, sadr_ip6); default: return 0; } } @@ -224,6 +247,9 @@ net_normalize(net_addr *N) case NET_MPLS: return; + + case NET_SADR_IP6: + return net_normalize_sadr_ip6(&n->sadr_ip6); } } @@ -248,6 +274,9 @@ net_classify(const net_addr *N) case NET_MPLS: return IADDR_HOST | SCOPE_UNIVERSE; + + case NET_SADR_IP6: + return ip6_zero(n->sadr_ip6.dst_prefix) ? (IADDR_HOST | SCOPE_UNIVERSE) : ip6_classify(&n->sadr_ip6.dst_prefix); } return IADDR_INVALID; @@ -274,6 +303,11 @@ ipa_in_netX(const ip_addr a, const net_addr *n) return ip6_zero(ip6_and(ip6_xor(ipa_to_ip6(a), net6_prefix(n)), ip6_mkmask(net6_pxlen(n)))); + case NET_SADR_IP6: + if (ipa_is_ip4(a)) return 0; + return ip6_zero(ip6_and(ip6_xor(ipa_to_ip6(a), net6_sadr_dst_prefix(n)), + ip6_mkmask(net6_sadr_dst_pxlen(n)))); + case NET_MPLS: default: return 0; diff --git a/lib/net.h b/lib/net.h index 332f4c9..22eee60 100644 --- a/lib/net.h +++ b/lib/net.h @@ -12,7 +12,6 @@ #include "lib/ip.h" - #define NET_IP4 1 #define NET_IP6 2 #define NET_VPN4 3 @@ -22,7 +21,8 @@ #define NET_FLOW4 7 #define NET_FLOW6 8 #define NET_MPLS 9 -#define NET_MAX 10 +#define NET_SADR_IP6 10 +#define NET_MAX 11 #define NB_IP4 (1 << NET_IP4) #define NB_IP6 (1 << NET_IP6) @@ -33,8 +33,9 @@ #define NB_FLOW4 (1 << NET_FLOW4) #define NB_FLOW6 (1 << NET_FLOW6) #define NB_MPLS (1 << NET_MPLS) +#define NB_SADR (1 << NET_SADR_IP6) -#define NB_IP (NB_IP4 | NB_IP6) +#define NB_IP (NB_IP4 | NB_IP6 | NB_SADR) #define NB_VPN (NB_VPN4 | NB_VPN6) #define NB_FLOW (NB_FLOW4 | NB_FLOW6) #define NB_DEST (NB_IP | NB_VPN | NB_MPLS) @@ -120,6 +121,15 @@ typedef struct net_addr_mpls { u32 label; } net_addr_mpls; +typedef struct net_addr_sadr_ip6 { + u8 type; + u8 dst_pxlen; + u16 length; + ip6_addr dst_prefix; + u32 src_pxlen; // if u8, NET_ADDR_SADR_IP6 would leave non-zero padding + ip6_addr src_prefix; +} net_addr_sadr_ip6; + typedef union net_addr_union { net_addr n; net_addr_ip4 ip4; @@ -131,6 +141,7 @@ typedef union net_addr_union { net_addr_flow4 flow4; net_addr_flow6 flow6; net_addr_mpls mpls; + net_addr_sadr_ip6 sadr_ip6; } net_addr_union; @@ -169,6 +180,9 @@ extern const u16 net_max_text_length[]; #define NET_ADDR_MPLS(label) \ ((net_addr_mpls) { NET_MPLS, 20, sizeof(net_addr_mpls), label }) +#define NET_ADDR_SADR_IP6(dst_prefix,dst_pxlen,src_prefix,src_pxlen) \ + ((net_addr_sadr_ip6) { NET_SADR_IP6, dst_pxlen, sizeof(net_addr_sadr_ip6), dst_prefix, src_pxlen, src_prefix }) + static inline void net_fill_ip4(net_addr *a, ip4_addr prefix, uint pxlen) { *(net_addr_ip4 *)a = NET_ADDR_IP4(prefix, pxlen); } @@ -191,6 +205,9 @@ static inline void net_fill_roa6(net_addr *a, ip6_addr prefix, uint pxlen, uint static inline void net_fill_mpls(net_addr *a, u32 label) { *(net_addr_mpls *)a = NET_ADDR_MPLS(label); } +static inline void net_fill_sadr_ip6(net_addr *a, ip6_addr dst_prefix, uint dst_pxlen, ip6_addr src_prefix, uint src_pxlen) +{ *(net_addr_sadr_ip6 *)a = NET_ADDR_SADR_IP6(dst_prefix, dst_pxlen, src_prefix, src_pxlen); } + static inline void net_fill_ipa(net_addr *a, ip_addr prefix, uint pxlen) { if (ipa_is_ip4(prefix)) @@ -236,6 +253,9 @@ static inline int net_is_roa(const net_addr *a) static inline int net_is_vpn(const net_addr *a) { return (a->type == NET_VPN4) || (a->type == NET_VPN6); } +static inline int net_is_sadr(const net_addr *a) +{ return (a->type == NET_SADR_IP6); } + static inline ip4_addr net4_prefix(const net_addr *a) { return ((net_addr_ip4 *) a)->prefix; } @@ -243,6 +263,12 @@ static inline ip4_addr net4_prefix(const net_addr *a) static inline ip6_addr net6_prefix(const net_addr *a) { return ((net_addr_ip6 *) a)->prefix; } +static inline ip6_addr net6_sadr_dst_prefix(const net_addr *a) +{ return ((net_addr_sadr_ip6 *) a)->dst_prefix; } + +static inline ip6_addr net6_sadr_src_prefix(const net_addr *a) +{ return ((net_addr_sadr_ip6 *) a)->src_prefix; } + static inline ip_addr net_prefix(const net_addr *a) { switch (a->type) @@ -259,6 +285,9 @@ static inline ip_addr net_prefix(const net_addr *a) case NET_FLOW6: return ipa_from_ip6(net6_prefix(a)); + case NET_SADR_IP6: + return ipa_from_ip6(net6_sadr_dst_prefix(a)); + case NET_MPLS: default: return IPA_NONE; @@ -279,6 +308,12 @@ static inline uint net4_pxlen(const net_addr *a) static inline uint net6_pxlen(const net_addr *a) { return a->pxlen; } +static inline uint net6_sadr_dst_pxlen(const net_addr *a) +{ return ((net_addr_sadr_ip6 *) a)->dst_pxlen; } + +static inline uint net6_sadr_src_pxlen(const net_addr *a) +{ return ((net_addr_sadr_ip6 *) a)->src_pxlen; } + static inline uint net_pxlen(const net_addr *a) { return a->pxlen; } @@ -327,6 +362,13 @@ static inline int net_equal_flow6(const net_addr_flow6 *a, const net_addr_flow6 static inline int net_equal_mpls(const net_addr_mpls *a, const net_addr_mpls *b) { return !memcmp(a, b, sizeof(net_addr_mpls)); } +typedef net_addr_ip6 net_addr_sadr_dst; +static inline int net_equal_sadr_dst(const net_addr_ip6 *a, const net_addr_ip6 *b) +{ return net_equal_ip6(a, b); } + +static inline int net_equal_sadr_ip6(const net_addr_sadr_ip6 *a, const net_addr_sadr_ip6 *b) +{ return !memcmp(a, b, sizeof(net_addr_sadr_ip6)); } + static inline int net_equal_prefix_roa4(const net_addr_roa4 *a, const net_addr_roa4 *b) { return ip4_equal(a->prefix, b->prefix) && (a->pxlen == b->pxlen); } @@ -390,6 +432,8 @@ static inline int net_compare_flow6(const net_addr_flow6 *a, const net_addr_flow static inline int net_compare_mpls(const net_addr_mpls *a, const net_addr_mpls *b) { return uint_cmp(a->label, b->label); } +int net_compare_sadr_ip6(const net_addr_sadr_ip6 *a, const net_addr_sadr_ip6 *b); + int net_compare(const net_addr *a, const net_addr *b); @@ -423,6 +467,9 @@ static inline void net_copy_flow6(net_addr_flow6 *dst, const net_addr_flow6 *src static inline void net_copy_mpls(net_addr_mpls *dst, const net_addr_mpls *src) { memcpy(dst, src, sizeof(net_addr_mpls)); } +static inline void net_copy_sadr_ip6(net_addr_sadr_ip6 *dst, const net_addr_sadr_ip6 *src) +{ memcpy(dst, src, sizeof(net_addr_sadr_ip6)); } + /* XXXX */ static inline u32 u64_hash(u64 a) @@ -455,6 +502,12 @@ static inline u32 net_hash_flow6(const net_addr_flow6 *n) static inline u32 net_hash_mpls(const net_addr_mpls *n) { return n->label; } +static inline u32 net_hash_sadr_dst(const net_addr_sadr_dst* n) +{ return net_hash_ip6(n); } + +static inline u32 net_hash_sadr_ip6(const net_addr_sadr_ip6 *n) +{ return ip6_hash(n->dst_prefix) ^ ((u32) n->dst_pxlen << 26); } + u32 net_hash(const net_addr *a); @@ -504,6 +557,9 @@ static inline int net_validate_flow6(const net_addr_flow6 *n) static inline int net_validate_mpls(const net_addr_mpls *n) { return n->label < (1 << 20); } +static inline int net_validate_sadr_ip6(const net_addr_sadr_ip6 *n) +{ return net_validate_px6(n->dst_prefix, n->dst_pxlen) && net_validate_px6(n->src_prefix, n->src_pxlen); } + int net_validate(const net_addr *N); @@ -519,6 +575,10 @@ static inline void net_normalize_vpn4(net_addr_vpn4 *n) static inline void net_normalize_vpn6(net_addr_vpn6 *n) { net_normalize_ip6((net_addr_ip6 *) n); } +static inline void net_normalize_sadr_ip6(net_addr_sadr_ip6 *n) +{ n->dst_prefix = ip6_and(n->dst_prefix, ip6_mkmask(n->dst_pxlen)); + n->src_prefix = ip6_and(n->src_prefix, ip6_mkmask(n->src_pxlen)); } + void net_normalize(net_addr *N); diff --git a/nest/rt-fib.c b/nest/rt-fib.c index 11c31d0..5831b33 100644 --- a/nest/rt-fib.c +++ b/nest/rt-fib.c @@ -195,6 +195,7 @@ fib_hash(struct fib *f, const net_addr *a) case NET_ROA6: return FIB_HASH(f, a, roa6); case NET_FLOW4: return FIB_HASH(f, a, flow4); case NET_FLOW6: return FIB_HASH(f, a, flow6); + case NET_SADR_IP6: return FIB_HASH(f, a, sadr_ip6); default: bug("invalid type"); } } @@ -231,6 +232,20 @@ fib_find(struct fib *f, const net_addr *a) case NET_ROA6: return FIB_FIND(f, a, roa6); case NET_FLOW4: return FIB_FIND(f, a, flow4); case NET_FLOW6: return FIB_FIND(f, a, flow6); + case NET_SADR_IP6: + { + if (a->length != sizeof(net_addr_sadr_ip6)) + { + // dst only search + net_addr_ip6 a0; + net_copy((net_addr *)&a0, a); + a0.length = sizeof(net_addr_sadr_ip6); + + return FIB_FIND(f, &a0, sadr_dst); + } + else + return FIB_FIND(f, a, sadr_ip6); + } default: bug("invalid type"); } } @@ -250,6 +265,7 @@ fib_insert(struct fib *f, const net_addr *a, struct fib_node *e) case NET_ROA6: FIB_INSERT(f, a, e, roa6); return; case NET_FLOW4: FIB_INSERT(f, a, e, flow4); return; case NET_FLOW6: FIB_INSERT(f, a, e, flow6); return; + case NET_SADR_IP6: FIB_INSERT(f, a, e, sadr_ip6); return; default: bug("invalid type"); } } diff --git a/nest/rt-table.c b/nest/rt-table.c index 67ccc88..49347f8 100644 --- a/nest/rt-table.c +++ b/nest/rt-table.c @@ -95,6 +95,35 @@ net_route_ip6(rtable *t, net_addr_ip6 *n) return r; } +static inline void * +net_route_sadr_ip6(rtable *t, net_addr_sadr_ip6 *n) +{ + net *r; + + net_addr_ip6 dst = NET_ADDR_IP6(n->dst_prefix, n->dst_pxlen); + dst.type = NET_SADR_IP6; + + // search for matching dst + while (r = net_route_ip6(t, &dst)) + { + // found a matching dst, start search for entries that match both prefixes + net_addr_sadr_ip6 full_addr = + NET_ADDR_SADR_IP6(dst.prefix, dst.pxlen, n->src_prefix, n->src_pxlen); + + while (r = net_find_valid(t, (net_addr *) &full_addr), (!r) && (full_addr.src_pxlen > 0)) + { + full_addr.src_pxlen--; + ip6_clrbit(&full_addr.src_prefix, full_addr.src_pxlen); + } + + if (r) + return r; + dst.pxlen--; + } + + return NULL; +} + void * net_route(rtable *tab, const net_addr *n) { @@ -115,6 +144,9 @@ net_route(rtable *tab, const net_addr *n) case NET_ROA6: return net_route_ip6(tab, (net_addr_ip6 *) n0); + case NET_SADR_IP6: + return net_route_sadr_ip6(tab, (net_addr_sadr_ip6 *) n0); + default: return NULL; } @@ -1747,6 +1779,7 @@ rt_preconfig(struct config *c) rt_new_table(cf_get_symbol("master4"), NET_IP4); rt_new_table(cf_get_symbol("master6"), NET_IP6); + rt_new_table(cf_get_symbol("master_sadr6"), NET_SADR_IP6); } @@ -1980,7 +2013,7 @@ rt_new_table(struct symbol *s, uint addr_type) /* Hack that allows to 'redefine' the master table */ if ((s->class == SYM_TABLE) && (s->def == new_config->def_tables[addr_type]) && - ((addr_type == NET_IP4) || (addr_type == NET_IP6))) + ((addr_type == NET_IP4) || (addr_type == NET_IP6) || (addr_type == NET_SADR_IP6))) return s->def; struct rtable_config *c = cfg_allocz(sizeof(struct rtable_config)); diff --git a/sysdep/linux/netlink.c b/sysdep/linux/netlink.c index 083a8d9..d89ae10 100644 --- a/sysdep/linux/netlink.c +++ b/sysdep/linux/netlink.c @@ -369,6 +369,7 @@ static struct nl_want_attrs rtm_attr_want4[BIRD_RTA_MAX] = { static struct nl_want_attrs rtm_attr_want6[BIRD_RTA_MAX] = { [RTA_DST] = { 1, 1, sizeof(ip6_addr) }, + [RTA_SRC] = { 1, 1, sizeof(ip6_addr) }, [RTA_IIF] = { 1, 1, sizeof(u32) }, [RTA_OIF] = { 1, 1, sizeof(u32) }, [RTA_GATEWAY] = { 1, 1, sizeof(ip6_addr) }, @@ -1161,8 +1162,17 @@ nl_send_route(struct krt_proto *p, rte *e, struct ea_list *eattrs, int op, int d nl_add_attr_mpls(&r->h, rsize, RTA_DST, 1, &label); } else + { nl_add_attr_ipa(&r->h, rsize, RTA_DST, net_prefix(net->n.addr)); + // add source address for SADR routes + if (net->n.addr->type == NET_SADR_IP6) + { + r->r.rtm_src_len = net6_sadr_src_pxlen(net->n.addr); + nl_add_attr_ipa(&r->h, rsize, RTA_SRC, net6_sadr_src_prefix(net->n.addr)); + } + } + /* * Strange behavior for RTM_DELROUTE: * 1) rtm_family is ignored in IPv6, works for IPv4 @@ -1394,7 +1404,8 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) struct rtattr *a[BIRD_RTA_MAX]; int new = h->nlmsg_type == RTM_NEWROUTE; - net_addr dst; + net_addr dst, sadr_src; + net_fill_ip6(&sadr_src, IP6_NONE, 0); u32 oif = ~0; u32 table_id; u32 priority = 0; @@ -1420,6 +1431,9 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) if (!nl_parse_attrs(RTM_RTA(i), rtm_attr_want6, a, sizeof(a))) return; + if (a[RTA_SRC]) + net_fill_ip6(&sadr_src, rta_get_ip6(a[RTA_SRC]), i->rtm_src_len); + if (a[RTA_DST]) net_fill_ip6(&dst, rta_get_ip6(a[RTA_DST]), i->rtm_dst_len); else @@ -1496,7 +1510,15 @@ nl_parse_route(struct nl_parse_state *s, struct nlmsghdr *h) src = KRT_SRC_ALIEN; } - net *net = net_get(p->p.main_channel->table, &dst); + net *net; + if (p->p.main_channel->table->addr_type == NET_SADR_IP6) + { + net_addr_sadr_ip6 sadr_addr = + NET_ADDR_SADR_IP6(net_prefix(&dst), net_pxlen(&dst), net_prefix(&sadr_src), net_pxlen(&sadr_src)); + net = net_get(p->p.main_channel->table, (net_addr *)&sadr_addr); + } + else + 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); diff --git a/sysdep/unix/krt.c b/sysdep/unix/krt.c index c6ff627..a9ce96b 100644 --- a/sysdep/unix/krt.c +++ b/sysdep/unix/krt.c @@ -1141,7 +1141,9 @@ krt_start(struct proto *P) switch (p->p.net_type) { case NET_IP4: p->af = AF_INET; break; - case NET_IP6: p->af = AF_INET6; break; + case NET_IP6: + case NET_SADR_IP6: + p->af = AF_INET6; break; case NET_MPLS: p->af = AF_MPLS; break; default: log(L_ERR "KRT: Tried to start with strange net type: %d", p->p.net_type); return PS_START; break; } -- 2.7.4

2 7

Bird Daemon Performance and usage information
by Eloi Carbó Solé 22 May '17

22 May '17

Hello, I am working on my MSc. dissertation which is about improving existing Bird's integration with UCI configuration scheme for OpenWRT/LEDE firmwares (if someone is interested [0]). As part of my documentation I am trying to highlight Bird's performance against other well-known software as well as adoption rates in IXPs or other organisations. There is some information I already used in the past for the project that gave birth to this integration in OpenWRT (~2014) but I cannot find presentations/papers/graphics using the latest versions or news/announcements from IXPs or organisations using Bird. Right now I am referencing some presentations [1], news [2] and old data coming from years 2012-2015 but, is there any extra documents I could use as evidence of Bird's performance/adoption rate or latest presentations to the public? Best regards, -- Eloi Carbó [0] https://github.com/eloicaso/bird-openwrt [1] https://people.netfilter.org/pablo/netdev0.1/slides/ondrej- zajicek-slides-netdev01.pdf [2] https://joinup.ec.europa.eu/community/osor/case/bird- manages-routing-worlds-largest-internet-exchanges

3 3

Re: [PATCH 2/4] Static protocol supports SADR
by Dean 22 May '17

22 May '17

On 05/22/2017 01:57 PM, Ondrej Zajicek wrote: > On Sun, May 21, 2017 at 11:01:35PM +0200, Dean Luga wrote: >> From: dean <dluga93(a)gmail.com> >> >> A new channel, sadr_ip6, is used for SADR both in the kernel >> and static protocols. In the static protocol, routes can be >> inserted with the following syntax: >> >> route <dst_prefix> from <src_prefix> via "interface" >> route <dst_prefix> from <src_prefix> recursive <ip> > Hi > > You should add parsing of SADR networks to conf/confbase.Y together with > other network types in net_ nonterminal symbol. That would add support > of SADR to static procotol automatically. > > Also, SADR-based recursive routes seems like a whole new can of worms. > And we are currently rethinking recursive routes internally. Perhaps > you could postpone changes to hostentry and stick with SADR recursive > routes dependent on non-SADR IGP tables, which fits to the current > framework. Note that this was updated in 2.0-pre1 so that for other > network types than IP4/IP6, igp tables are of type IP4/IP6. > Ok. How do I go about this? Do I make the changes and submit all 4 patches again? Or another patch with the config fix and reverting the recursive SADR routes?

2 1

Source Address Dependent Routing patch
by Dean 22 May '17

22 May '17

Hello, This is another try at extending the static and OSPFv3 protocols to support Source Address Dependent Routing (SADR), also called Source Specific Routing. This basically means that routing will take into account not only the destination address, but the source address as well. IPv6 subtrees are needed to support SADR in the kernel. They should be available for Linux kernel versions newer than 3.11. If they are not enabled, the kernel needs to be compiled with the option to enable them. The patches are based on bird-2.0.0-pre0.

3 6

Problem with BIRD importing routes from neighbor to master table
by Lyubomir Millanov 22 May '17

22 May '17

Hello, I'm having this problem with BIRD refusing to import a specific route into the routing table. This is strange because this is the session with the neighbor is not new and all of his other prefixes work fine. This is a newly added prefix. What I see: bird> show route all 185.200.156.0/22 185.200.156.0/22 via 185.1.40.2 on eth1 [as41059 2017-05-11] * (100) [AS61437i Type: BGP unicast univ BGP.origin: IGP BGP.as_path: 41313 16154 61437 BGP.next_hop: 185.1.40.2 BGP.med: 100 BGP.local_pref: 70 BGP.community: (0,42) (0,3856) (0,5483) (0,6939) (0,8400) (0,13335) (0,15169) (0,15958) (0,34772) (15169,13200) (16154,10000) (41059,2) (41059,65007) (41313,2) (49401,8877) (49401,49401) BGP.ext_community: (ro, 0, 197216) bird> sh route protocol as16154 all 185.200.156.0/22 via 185.1.40.12 on eth0 [as16154 2017-05-12] * (100) [AS61437i Type: BGP unicast univ BGP.origin: IGP BGP.as_path: 16154 61437 BGP.next_hop: 185.1.40.12 BGP.local_pref: 100 BGP.community: (15169,13300) (16154,10000) (41313,6699) (65000,208) For some reason the prefix is not installed in the master routing table even though it's accepted from the neighbor. We use IN-filtering based on queries with irr tool and those queries are correct - the prefix is correctly announced. >From the above I can see that the neighbor prefix is learned after the first one (same prefix, learned from another path). I've cleared the bgp session with the neighbor without any change of the behavior. Is there a way to clear the route from the master routing table? I'd appreciate if someone can give me advice on the matter. Best regards, Lyubo

1 0

Re: Please help, local interfaces not being recognized as local
by Jerry Scharf 18 May '17

18 May '17

Ondrej, Thanks for the response. That fixed it. jerry On 5/18/17 3:10 AM, Ondrej Zajicek wrote: > On Wed, May 17, 2017 at 06:28:59PM -0700, Jerry Scharf wrote: >> Hi, >> >> So I have narrowed things down but still don't understand what is going on. >> The symptoms are as follows: >> >> I have local interfaces with either simple addresses on net or on vlans. The >> peer is on the local subnet of the interface. Some are ibgp and some are >> ebgp. If I set direct, they all fail and I don't see any listener on port >> 179. As soon as I change them to multihop, everything works fine. > Hi > > You are missing 'protocol device' which is necessary for BIRD to find out > network interfaces. > -- Soundhound Devops "What could possibly go wrong?"

1 0

Please help, local interfaces not being recognized as local
by Jerry Scharf 18 May '17

18 May '17

Hi, So I have narrowed things down but still don't understand what is going on. The symptoms are as follows: I have local interfaces with either simple addresses on net or on vlans. The peer is on the local subnet of the interface. Some are ibgp and some are ebgp. If I set direct, they all fail and I don't see any listener on port 179. As soon as I change them to multihop, everything works fine. Also, I have defined a bunch of local aliases on lo. I then create a static block that lists all the addresses as via "lo". When I show static, they all show as dormant and they are not being redistributed by the bgp speaker. This is on a centos 7 system currently stock configured runing 1.4.5 from epel. I had been doing things with systemd-networkd and alternate kernels, but I did a clean install now. I have also tried 1.6.0 bird/centos7 build with the same issues. If someone can help me debug this, I would really appreciate it. If not, I may have to abandon using bird in favor of another routing daemon. That is certainly not my first choice. thanks, jerry here's my current config (with multihop specified and bfd disabled. There are currently no other files being included.) # this is automatically generated from borderbird4.j2, version 0.3 protocol kernel { # learn; # Learn all alien routes from the kernel persist; # Don't remove routes on bird shutdown scan time 20; # Scan kernel routing table every 20 seconds # import none; # Default is import all export all; # Default is export none # kernel table 5; # Kernel table to synchronize with (default: mai n) } protocol direct { interface "lo", "*"; preference 32000; # make sure bgp routes don't overwrite direct } log "/var/log/birdlog" all; router id 10.100.16.10; protocol static { preference 16000; route 198.49.100.45/32 via "lo"; route 8.25.217.18/32 via "lo"; route 198.49.100.18/32 via "lo"; route 8.25.217.45/32 via "lo"; route 8.6.161.18/32 via "lo"; route 8.5.161.45/32 via "lo"; } function issecondary() { return net ~ [198.49.100.45,8.25.217.18,198.49.100.18,8.25.217.45,8.6.161.18 ,8.5.161.45]; } function isprimary() { return net ~ [0.0.0.4/30]; } function net_outside() { return net ~ [172.18.10.0/24+,8.5.161.0/24+,198.49.100.0/24+]; } function is_internal_local() { return net ~ [10.100.32.10/32,10.100.32.0/20,10.100.16.10/32,10.100.16.0/20] ; } filter inside_export { if ((net = 0.0.0.0/0) || (net ~ 10.0.0.0/8)) then accept; else reject; } filter inside_import { if is_internal_local() then reject; if net ~ [10.0.0.0/8+, 172.16.0.0/12+] then accept; else reject; } filter load_default { if net = 0.0.0.0/0 then accept; else reject; } filter outside_only { if isprimary() then bgp_local_pref=100; else if issecondary() then bgp_local_pref=80; else if net_outside() then bgp_local_pref=25; if net_outside() then accept; else reject; } template bgp tocore { description "bgp to data center as"; local as 65302; multihop; export filter inside_export; import filter inside_import; # bfd on; } protocol bgp inside_netA from tocore { neighbor 10.100.16.3 as 65302; source address 10.100.16.10; } protocol bgp inside_netB from tocore { neighbor 10.100.32.3 as 65302; source address 10.100.32.10; } template bgp toedge { description "bgp to edge/ISP as"; local as 64102; multihop; export filter outside_only; import filter load_default; allow local as; # bfd on; } protocol bgp outside_edge_A from toedge { neighbor 172.18.10.1 as 18741; source address 172.18.10.4; } protocol bgp outside_edge_B from toedge { neighbor 172.18.10.2 as 18741; source address 172.18.10.5; } protocol bfd { interface "eth*" { interval 300 ms; multiplier 3; }; } include "/soundhound/etc/bird/*4.part.cfg"; -- Jerry Scharf, Soundhound DevOps "What could possibly go wrong?"

2 1

Redistribute loopback in OSPF
by Quentin Ritoul 18 May '17

18 May '17

Hello, I'm starting to use Bird, and I try to redistribute loopback in ospf. (So it seems easy...) rt01 : interface ens18 > 10.0.4.5/24 interface loopback > 10.99.99.1/32 (and 127.0.0.1/8) rt02 : interface ens18 > 10.0.4.6/24 interface loopback > 10.99.99.100/32 (and 127.0.0.1/8) Below the configuration of rt01 and rt02 : rt01 : filter loopbacks { if (net = 10.99.99.1/32 ) then accept; else reject; } protocol ospf { import all; export filter loopbacks; area 0 { interface "ens18" { cost 5; type broadcast; hello 5; retransmit 2; wait 10; dead 20; authentication cryptographic; password "toto"; }; }; } rt02 : filter loopbacks { if (net = 10.99.99.100/32 ) then accept; else reject; } protocol ospf { import all; export filter loopbacks; area 0 { interface "ens18" { cost 5; type broadcast; hello 5; retransmit 2; wait 10; dead 20; authentication cryptographic; password "toto"; }; }; } But it doesn't work. Each router don't receive the route 10.99.99.X bird> sh route 0.0.0.0/0 via 10.0.4.1 on ens18 [kernel1 22:01:40] * (10) 10.0.4.0/24 dev ens18 [direct1 22:01:40] * (240) dev ens18 [ospf1 22:01:48] I (150/5) [10.0.4.5] Could you help me please ? Regards, Quentin

3 2

OSPFv3 flooding unknown LSAs
by Dean 17 May '17

17 May '17

Hello, I am using bird 2.0.0 (from the int-new branch) and I noticed that OSPF is flooding unknown LSAs. According to RFC 5340, unknown LSAs shouldn't be flooded if the U-bit isn't set. But the lsa_flooding_allowed function doesn't check the U-bit or whether the LSA has an unknown type. Is that a mistake or did I miss something? Regards, Dean Luga.

1 1

are my posts working?
by Jerry Scharf 17 May '17

17 May '17

Hi, Sorry to spam the list and people, but I have tried to send a couple messages to the list with nothing showing up. jerry -- Jerry Scharf, Soundhound DevOps "What could possibly go wrong?"

2 1