/*
* BIRD -- The Babel protocol
*
* Copyright (c) 2015--2016 Toke Hoiland-Jorgensen
* (c) 2016--2017 Ondrej Zajicek <santiago@crfreenet.org>
* (c) 2016--2017 CZ.NIC z.s.p.o.
*
* Can be freely distributed and used under the terms of the GNU GPL.
*
* This file contains the packet and TLV handling code for the protocol.
*/
#include "babel.h"
struct babel_pkt_header {
u8 magic;
u8 version;
u16 length;
} PACKED;
struct babel_tlv {
u8 type;
u8 length;
u8 value[0];
} PACKED;
struct babel_tlv_ack_req {
u8 type;
u8 length;
u16 reserved;
u16 nonce;
u16 interval;
} PACKED;
struct babel_tlv_ack {
u8 type;
u8 length;
u16 nonce;
} PACKED;
struct babel_tlv_hello {
u8 type;
u8 length;
u16 flags;
u16 seqno;
u16 interval;
} PACKED;
struct babel_tlv_ihu {
u8 type;
u8 length;
u8 ae;
u8 reserved;
u16 rxcost;
u16 interval;
u8 addr[0];
} PACKED;
struct babel_tlv_router_id {
u8 type;
u8 length;
u16 reserved;
u64 router_id;
} PACKED;
struct babel_tlv_next_hop {
u8 type;
u8 length;
u8 ae;
u8 reserved;
u8 addr[0];
} PACKED;
struct babel_tlv_update {
u8 type;
u8 length;
u8 ae;
u8 flags;
u8 plen;
u8 omitted;
u16 interval;
u16 seqno;
u16 metric;
u8 addr[0];
} PACKED;
struct babel_tlv_route_request {
u8 type;
u8 length;
u8 ae;
u8 plen;
u8 addr[0];
} PACKED;
struct babel_tlv_seqno_request {
u8 type;
u8 length;
u8 ae;
u8 plen;
u16 seqno;
u8 hop_count;
u8 reserved;
u64 router_id;
u8 addr[0];
} PACKED;
struct babel_subtlv_source_prefix {
u8 type;
u8 length;
u8 plen;
u8 addr[0];
} PACKED;
/* Hello flags */
#define BABEL_HF_UNICAST 0x8000
/* Update flags */
#define BABEL_UF_DEF_PREFIX 0x80
#define BABEL_UF_ROUTER_ID 0x40
struct babel_parse_state {
struct babel_proto *proto;
struct babel_iface *ifa;
ip_addr saddr;
ip_addr next_hop_ip4;
ip_addr next_hop_ip6;
u64 router_id; /* Router ID used in subsequent updates */
u8 def_ip6_prefix[16]; /* Implicit IPv6 prefix in network order */
u8 def_ip4_prefix[4]; /* Implicit IPv4 prefix in network order */
u8 router_id_seen; /* router_id field is valid */
u8 def_ip6_prefix_seen; /* def_ip6_prefix is valid */
u8 def_ip4_prefix_seen; /* def_ip4_prefix is valid */
u8 current_tlv_endpos; /* End of self-terminating TLVs (offset from start) */
u8 sadr_enabled;
};
enum parse_result {
PARSE_SUCCESS,
PARSE_ERROR,
PARSE_IGNORE,
};
struct babel_write_state {
u64 router_id;
u8 router_id_seen;
ip_addr next_hop_ip4;
ip_addr next_hop_ip6;
u8 def_ip6_prefix[16]; /* Implicit IPv6 prefix in network order */
u8 def_ip6_pxlen;
};
#define DROP(DSC,VAL) do { err_dsc = DSC; err_val = VAL; goto drop; } while(0)
#define DROP1(DSC) do { err_dsc = DSC; goto drop; } while(0)
#define LOG_PKT(msg, args...) \
log_rl(&p->log_pkt_tbf, L_REMOTE "%s: " msg, p->p.name, args)
#define FIRST_TLV(p) ((struct babel_tlv *) (((struct babel_pkt_header *) p) + 1))
#define NEXT_TLV(t) ((struct babel_tlv *) (((byte *) t) + TLV_LENGTH(t)))
#define TLV_LENGTH(t) (t->type == BABEL_TLV_PAD1 ? 1 : t->length + sizeof(struct babel_tlv))
#define TLV_OPT_LENGTH(t) (t->length + sizeof(struct babel_tlv) - sizeof(*t))
#define TLV_HDR(tlv,t,l) ({ tlv->type = t; tlv->length = l - sizeof(struct babel_tlv); })
#define TLV_HDR0(tlv,t) TLV_HDR(tlv, t, tlv_data[t].min_length)
#define NET_SIZE(n) BYTES(net_pxlen(n))
static inline uint
bytes_equal(u8 *b1, u8 *b2, uint maxlen)
{
uint i;
for (i = 0; (i < maxlen) && (*b1 == *b2); i++, b1++, b2++)
;
return i;
}
static inline uint
get_time16(const void *p)
{
uint v = get_u16(p) * BABEL_TIME_UNITS;
return MAX(BABEL_MIN_INTERVAL, v);
}
static inline void
put_time16(void *p, uint v)
{
put_u16(p, v / BABEL_TIME_UNITS);
}
static inline void
read_ip4_px(net_addr *n, const void *p, uint plen)
{
ip4_addr addr = {0};
memcpy(&addr, p, BYTES(plen));
net_fill_ip4(n, ip4_ntoh(addr), plen);
}
static inline void
put_ip4_px(void *p, net_addr *n)
{
ip4_addr addr = ip4_hton(net4_prefix(n));
memcpy(p, &addr, NET_SIZE(n));
}
static inline void
read_ip6_px(net_addr *n, const void *p, uint plen)
{
ip6_addr addr = IPA_NONE;
memcpy(&addr, p, BYTES(plen));
net_fill_ip6(n, ip6_ntoh(addr), plen);
}
static inline void
put_ip6_px(void *p, net_addr *n)
{
ip6_addr addr = ip6_hton(net6_prefix(n));
memcpy(p, &addr, NET_SIZE(n));
}
static inline ip6_addr
get_ip6_ll(const void *p)
{
return ip6_build(0xfe800000, 0, get_u32(p+0), get_u32(p+4));
}
static inline void
put_ip6_ll(void *p, ip6_addr addr)
{
put_u32(p+0, _I2(addr));
put_u32(p+4, _I3(addr));
}
/*
* TLV read/write functions
*/
static int babel_read_ack_req(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_hello(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_ihu(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_router_id(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_next_hop(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_update(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_route_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_seqno_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static int babel_read_source_prefix(struct babel_tlv *hdr, union babel_msg *msg, struct babel_parse_state *state);
static uint babel_write_ack(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
static uint babel_write_hello(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
static uint babel_write_ihu(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
static uint babel_write_update(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
static uint babel_write_route_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
static uint babel_write_seqno_request(struct babel_tlv *hdr, union babel_msg *msg, struct babel_write_state *state, uint max_len);
static int babel_write_source_prefix(struct babel_tlv *hdr, net_addr *net, uint max_len);
struct babel_tlv_data {
u8 min_length;
int (*read_tlv)(struct babel_tlv *hdr, union babel_msg *m, struct babel_parse_state *state);
uint (*write_tlv)(struct babel_tlv *hdr, union babel_msg *m, struct babel_write_state *state, uint max_len);
void (*handle_tlv)(union babel_msg *m, struct babel_iface *ifa);
};
static const struct babel_tlv_data tlv_data[BABEL_TLV_MAX] = {
[BABEL_TLV_ACK_REQ] = {
sizeof(struct babel_tlv_ack_req),
babel_read_ack_req,
NULL,
babel_handle_ack_req
},
[BABEL_TLV_ACK] = {
sizeof(struct babel_tlv_ack),
NULL,
babel_write_ack,
NULL
},
[BABEL_TLV_HELLO] = {
sizeof(struct babel_tlv_hello),
babel_read_hello,
babel_write_hello,
babel_handle_hello
},
[BABEL_TLV_IHU] = {
sizeof(struct babel_tlv_ihu),
babel_read_ihu,
babel_write_ihu,
babel_handle_ihu
},
[BABEL_TLV_ROUTER_ID] = {
sizeof(struct babel_tlv_router_id),
babel_read_router_id,
NULL,
NULL
},
[BABEL_TLV_NEXT_HOP] = {
sizeof(struct babel_tlv_next_hop),
babel_read_next_hop,
NULL,
NULL
},
[BABEL_TLV_UPDATE] = {
sizeof(struct babel_tlv_update),
babel_read_update,
babel_write_update,
babel_handle_update
},
[BABEL_TLV_ROUTE_REQUEST] = {
sizeof(struct babel_tlv_route_request),
babel_read_route_request,
babel_write_route_request,
babel_handle_route_request
},
[BABEL_TLV_SEQNO_REQUEST] = {
sizeof(struct babel_tlv_seqno_request),
babel_read_seqno_request,
babel_write_seqno_request,
babel_handle_seqno_request
},
};
static int
babel_read_ack_req(struct babel_tlv *hdr, union babel_msg *m,
struct babel_parse_state *state)
{
struct babel_tlv_ack_req *tlv = (void *) hdr;
struct babel_msg_ack_req *msg = &m->ack_req;
msg->type = BABEL_TLV_ACK_REQ;
msg->nonce = get_u16(&tlv->nonce);
msg->interval = get_time16(&tlv->interval);
msg->sender = state->saddr;
if (!msg->interval)
return PARSE_ERROR;
return PARSE_SUCCESS;
}
static uint
babel_write_ack(struct babel_tlv *hdr, union babel_msg *m,
struct babel_write_state *state UNUSED, uint max_len UNUSED)
{
struct babel_tlv_ack *tlv = (void *) hdr;
struct babel_msg_ack *msg = &m->ack;
TLV_HDR0(tlv, BABEL_TLV_ACK);
put_u16(&tlv->nonce, msg->nonce);
return sizeof(struct babel_tlv_ack);
}
static int
babel_read_hello(struct babel_tlv *hdr, union babel_msg *m,
struct babel_parse_state *state)
{
struct babel_tlv_hello *tlv = (void *) hdr;
struct babel_msg_hello *msg = &m->hello;
/* We currently don't support unicast Hello */
u16 flags = get_u16(&tlv->flags);
if (flags & BABEL_HF_UNICAST)
return PARSE_IGNORE;
msg->type = BABEL_TLV_HELLO;
msg->seqno = get_u16(&tlv->seqno);
msg->interval = get_time16(&tlv->interval);
msg->sender = state->saddr;
return PARSE_SUCCESS;
}
static uint
babel_write_hello(struct babel_tlv *hdr, union babel_msg *m,
struct babel_write_state *state UNUSED, uint max_len UNUSED)
{
struct babel_tlv_hello *tlv = (void *) hdr;
struct babel_msg_hello *msg = &m->hello;
TLV_HDR0(tlv, BABEL_TLV_HELLO);
put_u16(&tlv->seqno, msg->seqno);
put_time16(&tlv->interval, msg->interval);
return sizeof(struct babel_tlv_hello);
}
static int
babel_read_ihu(struct babel_tlv *hdr, union babel_msg *m,
struct babel_parse_state *state)
{
struct babel_tlv_ihu *tlv = (void *) hdr;
struct babel_msg_ihu *msg = &m->ihu;
msg->type = BABEL_TLV_IHU;
msg->ae = tlv->ae;
msg->rxcost = get_u16(&tlv->rxcost);
msg->interval = get_time16(&tlv->interval);
msg->addr = IPA_NONE;
msg->sender = state->saddr;
if (msg->ae >= BABEL_AE_MAX)
return PARSE_IGNORE;
/*
* We only actually read link-local IPs. In every other case, the addr field
* will be 0 but validation will succeed. The handler takes care of these
* cases. We handle them here anyway because we need the length for parsing
* subtlvs.
*/
switch (msg->ae)
{
case BABEL_AE_IP4:
if (TLV_OPT_LENGTH(tlv) < 4)
return PARSE_ERROR;
state->current_tlv_endpos += 4;
break;
case BABEL_AE_IP6:
if (TLV_OPT_LENGTH(tlv) < 16)
return PARSE_ERROR;
state->current_tlv_endpos += 16;
break;
case BABEL_AE_IP6_LL:
if (TLV_OPT_LENGTH(tlv) < 8)
return PARSE_ERROR;
msg->addr = ipa_from_ip6(get_ip6_ll(&tlv->addr));
state->current_tlv_endpos += 8;
break;
}
return PARSE_SUCCESS;
}
static uint
babel_write_ihu(struct babel_tlv *hdr, union babel_msg *m,
struct babel_write_state *state UNUSED, uint max_len)
{
struct babel_tlv_ihu *tlv = (void *) hdr;
struct babel_msg_ihu *msg = &m->ihu;
if (ipa_is_link_local(msg->addr) && max_len < sizeof(struct babel_tlv_ihu) + 8)
return 0;
TLV_HDR0(tlv, BABEL_TLV_IHU);
put_u16(&tlv->rxcost, msg->rxcost);
put_time16(&tlv->interval, msg->interval);
if (!ipa_is_link_local(msg->addr))
{
tlv->ae = BABEL_AE_WILDCARD;
return sizeof(struct babel_tlv_ihu);
}
put_ip6_ll(&tlv->addr, msg->addr);
tlv->ae = BABEL_AE_IP6_LL;
hdr->length += 8;
return sizeof(struct babel_tlv_ihu) + 8;
}
static int
babel_read_router_id(struct babel_tlv *hdr, union babel_msg *m UNUSED,
struct babel_parse_state *state)
{
struct babel_tlv_router_id *tlv = (void *) hdr;
state->router_id = get_u64(&tlv->router_id);
state->router_id_seen = 1;
return PARSE_IGNORE;
}
/* This is called directly from babel_write_update() */
static uint
babel_write_router_id(struct babel_tlv *hdr, u64 router_id,
struct babel_write_state *state, uint max_len UNUSED)
{
struct babel_tlv_router_id *tlv = (void *) hdr;
/* We still assume that first min_length bytes are available and zeroed */
TLV_HDR0(tlv, BABEL_TLV_ROUTER_ID);
put_u64(&tlv->router_id, router_id);
state->router_id = router_id;
state->router_id_seen = 1;
return sizeof(struct babel_tlv_router_id);
}
static int
babel_read_next_hop(struct babel_tlv *hdr, union babel_msg *m UNUSED,
struct babel_parse_state *state)
{
struct babel_tlv_next_hop *tlv = (void *) hdr;
switch (tlv->ae)
{
case BABEL_AE_WILDCARD:
return PARSE_ERROR;
case BABEL_AE_IP4:
if (TLV_OPT_LENGTH(tlv) < sizeof(ip4_addr))
return PARSE_ERROR;
state->next_hop_ip4 = ipa_from_ip4(get_ip4(&tlv->addr));
state->current_tlv_endpos += sizeof(ip4_addr);
return PARSE_IGNORE;
case BABEL_AE_IP6:
if (TLV_OPT_LENGTH(tlv) < sizeof(ip6_addr))
return PARSE_ERROR;
state->next_hop_ip6 = ipa_from_ip6(get_ip6(&tlv->addr));
state->current_tlv_endpos += sizeof(ip6_addr);
return PARSE_IGNORE;
case BABEL_AE_IP6_LL:
if (TLV_OPT_LENGTH(tlv) < 8)
return PARSE_ERROR;
state->next_hop_ip6 = ipa_from_ip6(get_ip6_ll(&tlv->addr));
state->current_tlv_endpos += 8;
return PARSE_IGNORE;
default:
return PARSE_IGNORE;
}
return PARSE_IGNORE;
}
/* This is called directly from babel_write_update() and returns -1 if a next
hop should be written but there is not enough space. */
static int
babel_write_next_hop(struct babel_tlv *hdr, ip_addr addr,
struct babel_write_state *state, uint max_len)
{
struct babel_tlv_next_hop *tlv = (void *) hdr;
if (ipa_zero(addr))
{
/* Should not happen */
return 0;
}
else if (ipa_is_ip4(addr) && !ipa_equal(addr, state->next_hop_ip4))
{
uint len = sizeof(struct babel_tlv_next_hop) + sizeof(ip4_addr);
if (len > max_len)
return -1;
TLV_HDR(tlv, BABEL_TLV_NEXT_HOP, len);
tlv->ae = BABEL_AE_IP4;
put_ip4(&tlv->addr, ipa_to_ip4(addr));
state->next_hop_ip4 = addr;
return len;
}
else if (ipa_is_ip6(addr) && !ipa_equal(addr, state->next_hop_ip6))
{
uint len = sizeof(struct babel_tlv_next_hop) + sizeof(ip6_addr);
if (len > max_len)
return -1;
TLV_HDR(tlv, BABEL_TLV_NEXT_HOP, len);
tlv->ae = BABEL_AE_IP6;
put_ip6(&tlv->addr, ipa_to_ip6(addr));
state->next_hop_ip6 = addr;
return len;
}
return 0;
}
static int
babel_read_update(struct babel_tlv *hdr, union babel_msg *m,
struct babel_parse_state *state)
{
struct babel_tlv_update *tlv = (void *) hdr;
struct babel_msg_update *msg = &m->update;
msg->type = BABEL_TLV_UPDATE;
msg->interval = get_time16(&tlv->interval);
msg->seqno = get_u16(&tlv->seqno);
msg->metric = get_u16(&tlv->metric);
/* Length of received prefix data without omitted part */
int len = BYTES(tlv->plen) - (int) tlv->omitted;
u8 buf[16] = {};
if ((len < 0) || ((uint) len > TLV_OPT_LENGTH(tlv)))
return PARSE_ERROR;
switch (tlv->ae)
{
case BABEL_AE_WILDCARD:
if (tlv->plen > 0)
return PARSE_ERROR;
if (msg->metric != 65535)
return PARSE_ERROR;
msg->wildcard = 1;
break;
case BABEL_AE_IP4:
if (tlv->plen > IP4_MAX_PREFIX_LENGTH)
return PARSE_ERROR;
/* Cannot omit data if there is no saved prefix */
if (tlv->omitted && !state->def_ip4_prefix_seen)
return PARSE_ERROR;
/* Update must have next hop, unless it is retraction */
if (ipa_zero(state->next_hop_ip4) && (msg->metric != BABEL_INFINITY))
return PARSE_IGNORE;
/* Merge saved prefix and received prefix parts */
memcpy(buf, state->def_ip4_prefix, tlv->omitted);
memcpy(buf + tlv->omitted, tlv->addr, len);
ip4_addr prefix4 = get_ip4(buf);
net_fill_ip4(&msg->net, prefix4, tlv->plen);
if (tlv->flags & BABEL_UF_DEF_PREFIX)
{
put_ip4(state->def_ip4_prefix, prefix4);
state->def_ip4_prefix_seen = 1;
}
msg->next_hop = state->next_hop_ip4;
break;
case BABEL_AE_IP6:
if (tlv->plen > IP6_MAX_PREFIX_LENGTH)
return PARSE_ERROR;
/* Cannot omit data if there is no saved prefix */
if (tlv->omitted && !state->def_ip6_prefix_seen)
return PARSE_ERROR;
/* Merge saved prefix and received prefix parts */
memcpy(buf, state->def_ip6_prefix, tlv->omitted);
memcpy(buf + tlv->omitted, tlv->addr, len);
ip6_addr prefix6 = get_ip6(buf);
net_fill_ip6(&msg->net, prefix6, tlv->plen);
if (state->sadr_enabled)
net_make_ip6_sadr(&msg->net);
if (tlv->flags & BABEL_UF_DEF_PREFIX)
{
put_ip6(state->def_ip6_prefix, prefix6);
state->def_ip6_prefix_seen = 1;
}
if (tlv->flags & BABEL_UF_ROUTER_ID)
{
state->router_id = ((u64) _I2(prefix6)) << 32 | _I3(prefix6);
state->router_id_seen = 1;
}
msg->next_hop = state->next_hop_ip6;
break;
case BABEL_AE_IP6_LL:
/* ??? */
return PARSE_IGNORE;
default:
return PARSE_IGNORE;
}
/* Update must have Router ID, unless it is retraction */
if (!state->router_id_seen && (msg->metric != BABEL_INFINITY))
{
DBG("Babel: No router ID seen before update\n");
return PARSE_ERROR;
}
msg->router_id = state->router_id;
msg->sender = state->saddr;
state->current_tlv_endpos += len;
return PARSE_SUCCESS;
}
static uint
babel_write_update(struct babel_tlv *hdr, union babel_msg *m,
struct babel_write_state *state, uint max_len)
{
struct babel_msg_update *msg = &m->update;
uint len0 = 0;
/*
* When needed, we write Router-ID TLV before Update TLV and return size of
* both of them. There is enough space for the Router-ID TLV, because
* sizeof(struct babel_tlv_router_id) == sizeof(struct babel_tlv_update).
*
* Router ID is not used for retractions, so do not use it in such case.
*/
if ((msg->metric < BABEL_INFINITY) &&
(!state->router_id_seen || (msg->router_id != state->router_id)))
{
len0 = babel_write_router_id(hdr, msg->router_id, state, max_len);
hdr = NEXT_TLV(hdr);
}
/*
* We also may add Next Hop TLV for regular updates. It may fail for not
* enough space or it may be unnecessary as the next hop is the same as the
* last one already announced. So we handle all three cases.
*/
if (msg->metric < BABEL_INFINITY)
{
int l = babel_write_next_hop(hdr, msg->next_hop, state, max_len - len0);
if (l < 0)
return 0;
if (l)
{
len0 += l;
hdr = NEXT_TLV(hdr);
}
}
struct babel_tlv_update *tlv = (void *) hdr;
uint len = sizeof(struct babel_tlv_update) + NET_SIZE(&msg->net);
if (len0 + len > max_len)
return 0;
memset(tlv, 0, sizeof(struct babel_tlv_update));
TLV_HDR(tlv, BABEL_TLV_UPDATE, len);
if (msg->wildcard)
{
tlv->ae = BABEL_AE_WILDCARD;
tlv->plen = 0;
}
else if (msg->net.type == NET_IP4)
{
tlv->ae = BABEL_AE_IP4;
tlv->plen = net4_pxlen(&msg->net);
put_ip4_px(tlv->addr, &msg->net);
}
else
{
tlv->ae = BABEL_AE_IP6;
tlv->plen = net6_pxlen(&msg->net);
/* Address compression - omit initial matching bytes */
u8 buf[16], omit;
put_ip6(buf, net6_prefix(&msg->net));
omit = bytes_equal(buf, state->def_ip6_prefix,
MIN(tlv->plen, state->def_ip6_pxlen) / 8);
if (omit > 0)
{
memcpy(tlv->addr, buf + omit, NET_SIZE(&msg->net) - omit);
tlv->omitted = omit;
tlv->length -= omit;
len -= omit;
}
else
{
put_ip6_px(tlv->addr, &msg->net);
tlv->flags |= BABEL_UF_DEF_PREFIX;
put_ip6(state->def_ip6_prefix, net6_prefix(&msg->net));
state->def_ip6_pxlen = tlv->plen;
}
}
put_time16(&tlv->interval, msg->interval);
put_u16(&tlv->seqno, msg->seqno);
put_u16(&tlv->metric, msg->metric);
if (msg->net.type == NET_IP6_SADR)
{
int l = babel_write_source_prefix(hdr, &msg->net, max_len - (len0 + len));
if (l < 0)
return 0;
len += l;
}
return len0 + len;
}
static int
babel_read_route_request(struct babel_tlv *hdr, union babel_msg *m,
struct babel_parse_state *state)
{
struct babel_tlv_route_request *tlv = (void *) hdr;
struct babel_msg_route_request *msg = &m->route_request;
msg->type = BABEL_TLV_ROUTE_REQUEST;
switch (tlv->ae)
{
case BABEL_AE_WILDCARD:
/* Wildcard requests must have plen 0 */
if (tlv->plen > 0)
return PARSE_ERROR;
msg->full = 1;
return PARSE_SUCCESS;
case BABEL_AE_IP4:
if (tlv->plen > IP4_MAX_PREFIX_LENGTH)
return PARSE_ERROR;
if (TLV_OPT_LENGTH(tlv) < BYTES(tlv->plen))
return PARSE_ERROR;
read_ip4_px(&msg->net, tlv->addr, tlv->plen);
state->current_tlv_endpos += BYTES(tlv->plen);
return PARSE_SUCCESS;
case BABEL_AE_IP6:
if (tlv->plen > IP6_MAX_PREFIX_LENGTH)
return PARSE_ERROR;
if (TLV_OPT_LENGTH(tlv) < BYTES(tlv->plen))
return PARSE_ERROR;
read_ip6_px(&msg->net, tlv->addr, tlv->plen);
state->current_tlv_endpos += BYTES(tlv->plen);
if (state->sadr_enabled)
net_make_ip6_sadr(&msg->net);
return PARSE_SUCCESS;
case BABEL_AE_IP6_LL:
return PARSE_ERROR;
default:
return PARSE_IGNORE;
}
return PARSE_IGNORE;
}
static uint
babel_write_route_request(struct babel_tlv *hdr, union babel_msg *m,
struct babel_write_state *state UNUSED, uint max_len)
{
struct babel_tlv_route_request *tlv = (void *) hdr;
struct babel_msg_route_request *msg = &m->route_request;
uint len = sizeof(struct babel_tlv_route_request) + NET_SIZE(&msg->net);
if (len > max_len)
return 0;
TLV_HDR(tlv, BABEL_TLV_ROUTE_REQUEST, len);
if (msg->full)
{
tlv->ae = BABEL_AE_WILDCARD;
tlv->plen = 0;
}
else if (msg->net.type == NET_IP4)
{
tlv->ae = BABEL_AE_IP4;
tlv->plen = net4_pxlen(&msg->net);
put_ip4_px(tlv->addr, &msg->net);
}
else
{
tlv->ae = BABEL_AE_IP6;
tlv->plen = net6_pxlen(&msg->net);
put_ip6_px(tlv->addr, &msg->net);
}
if (msg->net.type == NET_IP6_SADR)
{
int l = babel_write_source_prefix(hdr, &msg->net, max_len - len);
if (l < 0)
return 0;
len += l;
}
return len;
}
static int
babel_read_seqno_request(struct babel_tlv *hdr, union babel_msg *m,
struct babel_parse_state *state)
{
struct babel_tlv_seqno_request *tlv = (void *) hdr;
struct babel_msg_seqno_request *msg = &m->seqno_request;
msg->type = BABEL_TLV_SEQNO_REQUEST;
msg->seqno = get_u16(&tlv->seqno);
msg->hop_count = tlv->hop_count;
msg->router_id = get_u64(&tlv->router_id);
msg->sender = state->saddr;
if (tlv->hop_count == 0)
return PARSE_ERROR;
switch (tlv->ae)
{
case BABEL_AE_WILDCARD:
return PARSE_ERROR;
case BABEL_AE_IP4:
if (tlv->plen > IP4_MAX_PREFIX_LENGTH)
return PARSE_ERROR;
if (TLV_OPT_LENGTH(tlv) < BYTES(tlv->plen))
return PARSE_ERROR;
read_ip4_px(&msg->net, tlv->addr, tlv->plen);
state->current_tlv_endpos += BYTES(tlv->plen);
return PARSE_SUCCESS;
case BABEL_AE_IP6:
if (tlv->plen > IP6_MAX_PREFIX_LENGTH)
return PARSE_ERROR;
if (TLV_OPT_LENGTH(tlv) < BYTES(tlv->plen))
return PARSE_ERROR;
read_ip6_px(&msg->net, tlv->addr, tlv->plen);
state->current_tlv_endpos += BYTES(tlv->plen);
if (state->sadr_enabled)
net_make_ip6_sadr(&msg->net);
return PARSE_SUCCESS;
case BABEL_AE_IP6_LL:
return PARSE_ERROR;
default:
return PARSE_IGNORE;
}
return PARSE_IGNORE;
}
static uint
babel_write_seqno_request(struct babel_tlv *hdr, union babel_msg *m,
struct babel_write_state *state UNUSED, uint max_len)
{
struct babel_tlv_seqno_request *tlv = (void *) hdr;
struct babel_msg_seqno_request *msg = &m->seqno_request;
uint len = sizeof(struct babel_tlv_seqno_request) + NET_SIZE(&msg->net);
if (len > max_len)
return 0;
TLV_HDR(tlv, BABEL_TLV_SEQNO_REQUEST, len);
if (msg->net.type == NET_IP4)
{
tlv->ae = BABEL_AE_IP4;
tlv->plen = net4_pxlen(&msg->net);
put_ip4_px(tlv->addr, &msg->net);
}
else
{
tlv->ae = BABEL_AE_IP6;
tlv->plen = net6_pxlen(&msg->net);
put_ip6_px(tlv->addr, &msg->net);
}
put_u16(&tlv->seqno, msg->seqno);
tlv->hop_count = msg->hop_count;
put_u64(&tlv->router_id, msg->router_id);
if (msg->net.type == NET_IP6_SADR)
{
int l = babel_write_source_prefix(hdr, &msg->net, max_len - len);
if (l < 0)
return 0;
len += l;
}
return len;
}
static int
babel_read_source_prefix(struct babel_tlv *hdr, union babel_msg *msg,
struct babel_parse_state *state UNUSED)
{
struct babel_subtlv_source_prefix *tlv = (void *) hdr;
net_addr_ip6_sadr *net;
/*
* We would like to skip the sub-TLV if SADR is not enabled, but we do not
* know AF of the enclosing TLV yet. We will do that later.
*/
/* Check internal consistency */
if ((tlv->length < 1) ||
(tlv->plen > IP6_MAX_PREFIX_LENGTH) ||
(tlv->length < (1 + BYTES(tlv->plen))))
return PARSE_ERROR;
/* Plen MUST NOT be 0 */
if (tlv->plen == 0)
return PARSE_ERROR;
switch(msg->type)
{
case BABEL_TLV_UPDATE:
/* Wildcard updates with source prefix MUST be silently ignored */
if (msg->update.wildcard)
return PARSE_IGNORE;
net = (void *) &msg->update.net;
break;
case BABEL_TLV_ROUTE_REQUEST:
/* Wildcard requests with source addresses MUST be silently ignored */
if (msg->route_request.full)
return PARSE_IGNORE;
net = (void *) &msg->route_request.net;
break;
case BABEL_TLV_SEQNO_REQUEST:
net = (void *) &msg->seqno_request.net;
break;
default:
return PARSE_ERROR;
}
/* If SADR is active, the net has appropriate type */
if (net->type != NET_IP6_SADR)
return PARSE_IGNORE;
/* Duplicate Source Prefix sub-TLV; SHOULD ignore whole TLV */
if (net->src_pxlen > 0)
return PARSE_IGNORE;
net_addr_ip6 src;
read_ip6_px((void *) &src, tlv->addr, tlv->plen);
net->src_prefix = src.prefix;
net->src_pxlen = src.pxlen;
return PARSE_SUCCESS;
}
static int
babel_write_source_prefix(struct babel_tlv *hdr, net_addr *n, uint max_len)
{
struct babel_subtlv_source_prefix *tlv = (void *) NEXT_TLV(hdr);
net_addr_ip6_sadr *net = (void *) n;
/* Do not use this sub-TLV for default prefix */
if (net->src_pxlen == 0)
return 0;
uint len = sizeof(*tlv) + BYTES(net->src_pxlen);
if (len > max_len)
return -1;
TLV_HDR(tlv, BABEL_SUBTLV_SOURCE_PREFIX, len);
hdr->length += len;
net_addr_ip6 src = NET_ADDR_IP6(net->src_prefix, net->src_pxlen);
tlv->plen = src.pxlen;
put_ip6_px(tlv->addr, (void *) &src);
return len;
}
static inline int
babel_read_subtlvs(struct babel_tlv *hdr,
union babel_msg *msg,
struct babel_parse_state *state)
{
struct babel_tlv *tlv;
byte *pos, *end = (byte *) hdr + TLV_LENGTH(hdr);
int res;
for (tlv = (void *) hdr + state->current_tlv_endpos;
(byte *) tlv < end;
tlv = NEXT_TLV(tlv))
{
/* Ugly special case */
if (tlv->type == BABEL_TLV_PAD1)
continue;
/* The end of the common TLV header */
pos = (byte *)tlv + sizeof(struct babel_tlv);
if ((pos > end) || (pos + tlv->length > end))
return PARSE_ERROR;
/*
* The subtlv type space is non-contiguous (due to the mandatory bit), so
* use a switch for dispatch instead of the mapping array we use for TLVs
*/
switch (tlv->type)
{
case BABEL_SUBTLV_SOURCE_PREFIX:
res = babel_read_source_prefix(tlv, msg, state);
if (res != PARSE_SUCCESS)
return res;
break;
case BABEL_SUBTLV_PADN:
default:
/* Unknown mandatory subtlv; PARSE_IGNORE ignores the whole TLV */
if (tlv->type >= 128)
return PARSE_IGNORE;
break;
}
}
return PARSE_SUCCESS;
}
static inline int
babel_read_tlv(struct babel_tlv *hdr,
union babel_msg *msg,
struct babel_parse_state *state)
{
if ((hdr->type <= BABEL_TLV_PADN) ||
(hdr->type >= BABEL_TLV_MAX) ||
!tlv_data[hdr->type].read_tlv)
return PARSE_IGNORE;
if (TLV_LENGTH(hdr) < tlv_data[hdr->type].min_length)
return PARSE_ERROR;
state->current_tlv_endpos = tlv_data[hdr->type].min_length;
memset(msg, 0, sizeof(*msg));
int res = tlv_data[hdr->type].read_tlv(hdr, msg, state);
if (res != PARSE_SUCCESS)
return res;
return babel_read_subtlvs(hdr, msg, state);
}
static uint
babel_write_tlv(struct babel_tlv *hdr,
union babel_msg *msg,
struct babel_write_state *state,
uint max_len)
{
if ((msg->type <= BABEL_TLV_PADN) ||
(msg->type >= BABEL_TLV_MAX) ||
!tlv_data[msg->type].write_tlv)
return 0;
if (tlv_data[msg->type].min_length > max_len)
return 0;
memset(hdr, 0, tlv_data[msg->type].min_length);
return tlv_data[msg->type].write_tlv(hdr, msg, state, max_len);
}
/*
* Packet RX/TX functions
*/
static int
babel_send_to(struct babel_iface *ifa, ip_addr dest)
{
sock *sk = ifa->sk;
struct babel_pkt_header *hdr = (void *) sk->tbuf;
int len = get_u16(&hdr->length) + sizeof(struct babel_pkt_header);
DBG("Babel: Sending %d bytes to %I\n", len, dest);
return sk_send_to(sk, len, dest, 0);
}
/**
* babel_write_queue - Write a TLV queue to a transmission buffer
* @ifa: Interface holding the transmission buffer
* @queue: TLV queue to write (containing internal-format TLVs)
*
* This function writes a packet to the interface transmission buffer with as
* many TLVs from the &queue as will fit in the buffer. It returns the number of
* bytes written (NOT counting the packet header). The function is called by
* babel_send_queue() and babel_send_unicast() to construct packets for
* transmission, and uses per-TLV helper functions to convert the
* internal-format TLVs to their wire representations.
*
* The TLVs in the queue are freed after they are written to the buffer.
*/
static uint
babel_write_queue(struct babel_iface *ifa, list *queue)
{
struct babel_proto *p = ifa->proto;
struct babel_write_state state = { .next_hop_ip6 = ifa->addr };
if (EMPTY_LIST(*queue))
return 0;
byte *pos = ifa->sk->tbuf;
byte *end = pos + ifa->tx_length;
struct babel_pkt_header *pkt = (void *) pos;
pkt->magic = BABEL_MAGIC;
pkt->version = BABEL_VERSION;
pkt->length = 0;
pos += sizeof(struct babel_pkt_header);
struct babel_msg_node *msg;
WALK_LIST_FIRST(msg, *queue)
{
if (pos >= end)
break;
int len = babel_write_tlv((struct babel_tlv *) pos, &msg->msg, &state, end - pos);
if (!len)
break;
pos += len;
rem_node(NODE msg);
sl_free(p->msg_slab, msg);
}
uint plen = pos - (byte *) pkt;
put_u16(&pkt->length, plen - sizeof(struct babel_pkt_header));
return plen;
}
void
babel_send_queue(void *arg)
{
struct babel_iface *ifa = arg;
while ((babel_write_queue(ifa, &ifa->msg_queue) > 0) &&
(babel_send_to(ifa, IP6_BABEL_ROUTERS) > 0));
}
static inline void
babel_kick_queue(struct babel_iface *ifa)
{
/*
* Only schedule send event if there is not already data in the socket buffer.
* Otherwise we may overwrite the data already in the buffer.
*/
if ((ifa->sk->tpos == ifa->sk->tbuf) && !ev_active(ifa->send_event))
ev_schedule(ifa->send_event);
}
/**
* babel_send_unicast - send a single TLV via unicast to a destination
* @msg: TLV to send
* @ifa: Interface to send via
* @dest: Destination of the TLV
*
* This function is used to send a single TLV via unicast to a designated
* receiver. This is used for replying to certain incoming requests, and for
* sending unicast requests to refresh routes before they expire.
*/
void
babel_send_unicast(union babel_msg *msg, struct babel_iface *ifa, ip_addr dest)
{
struct babel_proto *p = ifa->proto;
struct babel_msg_node *msgn = sl_alloc(p->msg_slab);
list queue;
msgn->msg = *msg;
init_list(&queue);
add_tail(&queue, NODE msgn);
babel_write_queue(ifa, &queue);
babel_send_to(ifa, dest);
/* We could overwrite waiting packet here, we may have to kick TX queue */
if (!EMPTY_LIST(ifa->msg_queue))
babel_kick_queue(ifa);
}
/**
* babel_enqueue - enqueue a TLV for transmission on an interface
* @msg: TLV to enqueue (in internal TLV format)
* @ifa: Interface to enqueue to
*
* This function is called to enqueue a TLV for subsequent transmission on an
* interface. The transmission event is triggered whenever a TLV is enqueued;
* this ensures that TLVs will be transmitted in a timely manner, but that TLVs
* which are enqueued in rapid succession can be transmitted together in one
* packet.
*/
void
babel_enqueue(union babel_msg *msg, struct babel_iface *ifa)
{
struct babel_proto *p = ifa->proto;
struct babel_msg_node *msgn = sl_alloc(p->msg_slab);
msgn->msg = *msg;
add_tail(&ifa->msg_queue, NODE msgn);
babel_kick_queue(ifa);
}
/**
* babel_process_packet - process incoming data packet
* @pkt: Pointer to the packet data
* @len: Length of received packet
* @saddr: Address of packet sender
* @ifa: Interface packet was received on.
*
* This function is the main processing hook of incoming Babel packets. It
* checks that the packet header is well-formed, then processes the TLVs
* contained in the packet. This is done in two passes: First all TLVs are
* parsed into the internal TLV format. If a TLV parser fails, processing of the
* rest of the packet is aborted.
*
* After the parsing step, the TLV handlers are called for each parsed TLV in
* order.
*/
static void
babel_process_packet(struct babel_pkt_header *pkt, int len,
ip_addr saddr, struct babel_iface *ifa)
{
struct babel_proto *p = ifa->proto;
struct babel_tlv *tlv;
struct babel_msg_node *msg;
list msgs;
int res;
int plen = sizeof(struct babel_pkt_header) + get_u16(&pkt->length);
byte *pos;
byte *end = (byte *)pkt + plen;
struct babel_parse_state state = {
.proto = p,
.ifa = ifa,
.saddr = saddr,
.next_hop_ip6 = saddr,
.sadr_enabled = babel_sadr_enabled(p),
};
if ((pkt->magic != BABEL_MAGIC) || (pkt->version != BABEL_VERSION))
{
TRACE(D_PACKETS, "Strange packet from %I via %s - magic %d version %d",
saddr, ifa->iface->name, pkt->magic, pkt->version);
return;
}
if (plen > len)
{
LOG_PKT("Bad packet from %I via %s - %s (%u)",
saddr, ifa->iface->name, "length mismatch", plen);
return;
}
TRACE(D_PACKETS, "Packet received from %I via %s",
saddr, ifa->iface->name);
init_list(&msgs);
/* First pass through the packet TLV by TLV, parsing each into internal data
structures. */
for (tlv = FIRST_TLV(pkt);
(byte *)tlv < end;
tlv = NEXT_TLV(tlv))
{
/* Ugly special case */
if (tlv->type == BABEL_TLV_PAD1)
continue;
/* The end of the common TLV header */
pos = (byte *)tlv + sizeof(struct babel_tlv);
if ((pos > end) || (pos + tlv->length > end))
{
LOG_PKT("Bad TLV from %I via %s type %d pos %d - framing error",
saddr, ifa->iface->name, tlv->type, (byte *)tlv - (byte *)pkt);
break;
}
msg = sl_alloc(p->msg_slab);
res = babel_read_tlv(tlv, &msg->msg, &state);
if (res == PARSE_SUCCESS)
{
add_tail(&msgs, NODE msg);
}
else if (res == PARSE_IGNORE)
{
DBG("Babel: Ignoring TLV of type %d\n", tlv->type);
sl_free(p->msg_slab, msg);
}
else /* PARSE_ERROR */
{
LOG_PKT("Bad TLV from %I via %s type %d pos %d - parse error",
saddr, ifa->iface->name, tlv->type, (byte *)tlv - (byte *)pkt);
sl_free(p->msg_slab, msg);
break;
}
}
/* Parsing done, handle all parsed TLVs */
WALK_LIST_FIRST(msg, msgs)
{
if (tlv_data[msg->msg.type].handle_tlv)
tlv_data[msg->msg.type].handle_tlv(&msg->msg, ifa);
rem_node(NODE msg);
sl_free(p->msg_slab, msg);
}
}
static void
babel_err_hook(sock *sk, int err)
{
struct babel_iface *ifa = sk->data;
struct babel_proto *p = ifa->proto;
log(L_ERR "%s: Socket error on %s: %M", p->p.name, ifa->iface->name, err);
/* FIXME: Drop queued TLVs here? */
}
static void
babel_tx_hook(sock *sk)
{
struct babel_iface *ifa = sk->data;
DBG("Babel: TX hook called (iface %s, src %I, dst %I)\n",
sk->iface->name, sk->saddr, sk->daddr);
babel_send_queue(ifa);
}
static int
babel_rx_hook(sock *sk, uint len)
{
struct babel_iface *ifa = sk->data;
struct babel_proto *p = ifa->proto;
const char *err_dsc = NULL;
uint err_val = 0;
if (sk->lifindex != ifa->iface->index)
return 1;
DBG("Babel: RX hook called (iface %s, src %I, dst %I)\n",
sk->iface->name, sk->faddr, sk->laddr);
/* Silently ignore my own packets */
if (ipa_equal(sk->faddr, sk->saddr))
return 1;
if (!ipa_is_link_local(sk->faddr))
DROP1("wrong src address");
if (sk->fport != ifa->cf->port)
DROP("wrong src port", sk->fport);
if (len < sizeof(struct babel_pkt_header))
DROP("too short", len);
if (sk->flags & SKF_TRUNCATED)
DROP("truncated", len);
babel_process_packet((struct babel_pkt_header *) sk->rbuf, len, sk->faddr, ifa);
return 1;
drop:
LOG_PKT("Bad packet from %I via %s - %s (%u)",
sk->faddr, sk->iface->name, err_dsc, err_val);
return 1;
}
int
babel_open_socket(struct babel_iface *ifa)
{
struct babel_proto *p = ifa->proto;
sock *sk;
sk = sk_new(ifa->pool);
sk->type = SK_UDP;
sk->sport = ifa->cf->port;
sk->dport = ifa->cf->port;
sk->iface = ifa->iface;
sk->saddr = ifa->addr;
sk->vrf = p->p.vrf;
sk->rx_hook = babel_rx_hook;
sk->tx_hook = babel_tx_hook;
sk->err_hook = babel_err_hook;
sk->data = ifa;
sk->tos = ifa->cf->tx_tos;
sk->priority = ifa->cf->tx_priority;
sk->ttl = 1;
sk->flags = SKF_LADDR_RX;
if (sk_open(sk) < 0)
goto err;
if (sk_setup_multicast(sk) < 0)
goto err;
if (sk_join_group(sk, IP6_BABEL_ROUTERS) < 0)
goto err;
ifa->sk = sk;
return 1;
err:
sk_log_error(sk, p->p.name);
rfree(sk);
return 0;
}
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>