blob: b173bb2ad013a71bf395438ac28001ddb857e002 [file] [log] [blame]
/*
* Copyright (c) 2015 Cisco and/or its affiliates.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "map.h"
#include "../ip/ip_frag.h"
#include <vnet/ip/ip6_to_ip4.h>
#include <vnet/ip/ip4_to_ip6.h>
#define IP6_MAP_T_DUAL_LOOP
typedef enum
{
IP6_MAPT_NEXT_MAPT_TCP_UDP,
IP6_MAPT_NEXT_MAPT_ICMP,
IP6_MAPT_NEXT_MAPT_FRAGMENTED,
IP6_MAPT_NEXT_DROP,
IP6_MAPT_N_NEXT
} ip6_mapt_next_t;
typedef enum
{
IP6_MAPT_ICMP_NEXT_IP4_LOOKUP,
IP6_MAPT_ICMP_NEXT_IP4_FRAG,
IP6_MAPT_ICMP_NEXT_DROP,
IP6_MAPT_ICMP_N_NEXT
} ip6_mapt_icmp_next_t;
typedef enum
{
IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP,
IP6_MAPT_TCP_UDP_NEXT_IP4_FRAG,
IP6_MAPT_TCP_UDP_NEXT_DROP,
IP6_MAPT_TCP_UDP_N_NEXT
} ip6_mapt_tcp_udp_next_t;
typedef enum
{
IP6_MAPT_FRAGMENTED_NEXT_IP4_LOOKUP,
IP6_MAPT_FRAGMENTED_NEXT_IP4_FRAG,
IP6_MAPT_FRAGMENTED_NEXT_DROP,
IP6_MAPT_FRAGMENTED_N_NEXT
} ip6_mapt_fragmented_next_t;
static_always_inline int
ip6_map_fragment_cache (ip6_header_t * ip6, ip6_frag_hdr_t * frag,
map_domain_t * d, u16 port)
{
u32 *ignore = NULL;
map_ip4_reass_lock ();
map_ip4_reass_t *r = map_ip4_reass_get (map_get_ip4 (&ip6->src_address),
ip6_map_t_embedded_address (d,
&ip6->
dst_address),
frag_id_6to4 (frag->identification),
(ip6->protocol ==
IP_PROTOCOL_ICMP6) ?
IP_PROTOCOL_ICMP : ip6->protocol,
&ignore);
if (r)
r->port = port;
map_ip4_reass_unlock ();
return !r;
}
/* Returns the associated port or -1 */
static_always_inline i32
ip6_map_fragment_get (ip6_header_t * ip6, ip6_frag_hdr_t * frag,
map_domain_t * d)
{
u32 *ignore = NULL;
map_ip4_reass_lock ();
map_ip4_reass_t *r = map_ip4_reass_get (map_get_ip4 (&ip6->src_address),
ip6_map_t_embedded_address (d,
&ip6->
dst_address),
frag_id_6to4 (frag->identification),
(ip6->protocol ==
IP_PROTOCOL_ICMP6) ?
IP_PROTOCOL_ICMP : ip6->protocol,
&ignore);
i32 ret = r ? r->port : -1;
map_ip4_reass_unlock ();
return ret;
}
typedef struct
{
map_domain_t *d;
u16 sender_port;
} icmp6_to_icmp_ctx_t;
static int
ip6_to_ip4_set_icmp_cb (ip6_header_t * ip6, ip4_header_t * ip4, void *arg)
{
icmp6_to_icmp_ctx_t *ctx = arg;
u32 ip4_sadr;
//Security check
//Note that this prevents an intermediate IPv6 router from answering the request
ip4_sadr = map_get_ip4 (&ip6->src_address);
if (ip6->src_address.as_u64[0] !=
map_get_pfx_net (ctx->d, ip4_sadr, ctx->sender_port)
|| ip6->src_address.as_u64[1] != map_get_sfx_net (ctx->d, ip4_sadr,
ctx->sender_port))
return -1;
ip4->dst_address.as_u32 =
ip6_map_t_embedded_address (ctx->d, &ip6->dst_address);
ip4->src_address.as_u32 = ip4_sadr;
return 0;
}
static int
ip6_to_ip4_set_inner_icmp_cb (ip6_header_t * ip6, ip4_header_t * ip4,
void *arg)
{
icmp6_to_icmp_ctx_t *ctx = arg;
u32 inner_ip4_dadr;
//Security check of inner packet
inner_ip4_dadr = map_get_ip4 (&ip6->dst_address);
if (ip6->dst_address.as_u64[0] !=
map_get_pfx_net (ctx->d, inner_ip4_dadr, ctx->sender_port)
|| ip6->dst_address.as_u64[1] != map_get_sfx_net (ctx->d,
inner_ip4_dadr,
ctx->sender_port))
return -1;
ip4->dst_address.as_u32 = inner_ip4_dadr;
ip4->src_address.as_u32 =
ip6_map_t_embedded_address (ctx->d, &ip6->src_address);
return 0;
}
static uword
ip6_map_t_icmp (vlib_main_t * vm,
vlib_node_runtime_t * node, vlib_frame_t * frame)
{
u32 n_left_from, *from, next_index, *to_next, n_left_to_next;
vlib_node_runtime_t *error_node =
vlib_node_get_runtime (vm, ip6_map_t_icmp_node.index);
from = vlib_frame_vector_args (frame);
n_left_from = frame->n_vectors;
next_index = node->cached_next_index;
vlib_combined_counter_main_t *cm = map_main.domain_counters;
u32 thread_index = vlib_get_thread_index ();
while (n_left_from > 0)
{
vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
while (n_left_from > 0 && n_left_to_next > 0)
{
u32 pi0;
vlib_buffer_t *p0;
u8 error0;
ip6_mapt_icmp_next_t next0;
map_domain_t *d0;
u16 len0;
icmp6_to_icmp_ctx_t ctx0;
ip6_header_t *ip60;
pi0 = to_next[0] = from[0];
from += 1;
n_left_from -= 1;
to_next += 1;
n_left_to_next -= 1;
error0 = MAP_ERROR_NONE;
next0 = IP6_MAPT_ICMP_NEXT_IP4_LOOKUP;
p0 = vlib_get_buffer (vm, pi0);
ip60 = vlib_buffer_get_current (p0);
len0 = clib_net_to_host_u16 (ip60->payload_length);
d0 =
pool_elt_at_index (map_main.domains,
vnet_buffer (p0)->map_t.map_domain_index);
ctx0.sender_port = ip6_get_port (ip60, 0, len0);
ctx0.d = d0;
if (ctx0.sender_port == 0)
{
// In case of 1:1 mapping, we don't care about the port
if (!(d0->ea_bits_len == 0 && d0->rules))
{
error0 = MAP_ERROR_ICMP;
goto err0;
}
}
if (icmp6_to_icmp
(p0, ip6_to_ip4_set_icmp_cb, d0, ip6_to_ip4_set_inner_icmp_cb,
d0))
{
error0 = MAP_ERROR_ICMP;
goto err0;
}
if (vnet_buffer (p0)->map_t.mtu < p0->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p0)->ip_frag.mtu = vnet_buffer (p0)->map_t.mtu;
vnet_buffer (p0)->ip_frag.header_offset = 0;
vnet_buffer (p0)->ip_frag.next_index = IP4_FRAG_NEXT_IP4_LOOKUP;
next0 = IP6_MAPT_ICMP_NEXT_IP4_FRAG;
}
err0:
if (PREDICT_TRUE (error0 == MAP_ERROR_NONE))
{
vlib_increment_combined_counter (cm + MAP_DOMAIN_COUNTER_RX,
thread_index,
vnet_buffer (p0)->
map_t.map_domain_index, 1,
len0);
}
else
{
next0 = IP6_MAPT_ICMP_NEXT_DROP;
}
p0->error = error_node->errors[error0];
vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
to_next, n_left_to_next, pi0,
next0);
}
vlib_put_next_frame (vm, node, next_index, n_left_to_next);
}
return frame->n_vectors;
}
static int
ip6_to_ip4_set_cb (ip6_header_t * ip6, ip4_header_t * ip4, void *ctx)
{
vlib_buffer_t *p = ctx;
ip4->dst_address.as_u32 = vnet_buffer (p)->map_t.v6.daddr;
ip4->src_address.as_u32 = vnet_buffer (p)->map_t.v6.saddr;
return 0;
}
static uword
ip6_map_t_fragmented (vlib_main_t * vm,
vlib_node_runtime_t * node, vlib_frame_t * frame)
{
u32 n_left_from, *from, next_index, *to_next, n_left_to_next;
from = vlib_frame_vector_args (frame);
n_left_from = frame->n_vectors;
next_index = node->cached_next_index;
vlib_node_runtime_t *error_node =
vlib_node_get_runtime (vm, ip6_map_t_fragmented_node.index);
while (n_left_from > 0)
{
vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
#ifdef IP6_MAP_T_DUAL_LOOP
while (n_left_from >= 4 && n_left_to_next >= 2)
{
u32 pi0, pi1;
vlib_buffer_t *p0, *p1;
u32 next0, next1;
pi0 = to_next[0] = from[0];
pi1 = to_next[1] = from[1];
from += 2;
n_left_from -= 2;
to_next += 2;
n_left_to_next -= 2;
next0 = IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP;
next1 = IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP;
p0 = vlib_get_buffer (vm, pi0);
p1 = vlib_get_buffer (vm, pi1);
if (ip6_to_ip4_fragmented (p0, ip6_to_ip4_set_cb, p0))
{
p0->error = error_node->errors[MAP_ERROR_FRAGMENT_DROPPED];
next0 = IP6_MAPT_FRAGMENTED_NEXT_DROP;
}
else
{
if (vnet_buffer (p0)->map_t.mtu < p0->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p0)->ip_frag.mtu = vnet_buffer (p0)->map_t.mtu;
vnet_buffer (p0)->ip_frag.header_offset = 0;
vnet_buffer (p0)->ip_frag.next_index =
IP4_FRAG_NEXT_IP4_LOOKUP;
next0 = IP6_MAPT_FRAGMENTED_NEXT_IP4_FRAG;
}
}
if (ip6_to_ip4_fragmented (p1, ip6_to_ip4_set_cb, p1))
{
p1->error = error_node->errors[MAP_ERROR_FRAGMENT_DROPPED];
next1 = IP6_MAPT_FRAGMENTED_NEXT_DROP;
}
else
{
if (vnet_buffer (p1)->map_t.mtu < p1->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p1)->ip_frag.mtu = vnet_buffer (p1)->map_t.mtu;
vnet_buffer (p1)->ip_frag.header_offset = 0;
vnet_buffer (p1)->ip_frag.next_index =
IP4_FRAG_NEXT_IP4_LOOKUP;
next1 = IP6_MAPT_FRAGMENTED_NEXT_IP4_FRAG;
}
}
vlib_validate_buffer_enqueue_x2 (vm, node, next_index,
to_next, n_left_to_next, pi0, pi1,
next0, next1);
}
#endif
while (n_left_from > 0 && n_left_to_next > 0)
{
u32 pi0;
vlib_buffer_t *p0;
u32 next0;
pi0 = to_next[0] = from[0];
from += 1;
n_left_from -= 1;
to_next += 1;
n_left_to_next -= 1;
next0 = IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP;
p0 = vlib_get_buffer (vm, pi0);
if (ip6_to_ip4_fragmented (p0, ip6_to_ip4_set_cb, p0))
{
p0->error = error_node->errors[MAP_ERROR_FRAGMENT_DROPPED];
next0 = IP6_MAPT_FRAGMENTED_NEXT_DROP;
}
else
{
if (vnet_buffer (p0)->map_t.mtu < p0->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p0)->ip_frag.mtu = vnet_buffer (p0)->map_t.mtu;
vnet_buffer (p0)->ip_frag.header_offset = 0;
vnet_buffer (p0)->ip_frag.next_index =
IP4_FRAG_NEXT_IP4_LOOKUP;
next0 = IP6_MAPT_FRAGMENTED_NEXT_IP4_FRAG;
}
}
vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
to_next, n_left_to_next, pi0,
next0);
}
vlib_put_next_frame (vm, node, next_index, n_left_to_next);
}
return frame->n_vectors;
}
static uword
ip6_map_t_tcp_udp (vlib_main_t * vm,
vlib_node_runtime_t * node, vlib_frame_t * frame)
{
u32 n_left_from, *from, next_index, *to_next, n_left_to_next;
vlib_node_runtime_t *error_node =
vlib_node_get_runtime (vm, ip6_map_t_tcp_udp_node.index);
from = vlib_frame_vector_args (frame);
n_left_from = frame->n_vectors;
next_index = node->cached_next_index;
while (n_left_from > 0)
{
vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
#ifdef IP6_MAP_T_DUAL_LOOP
while (n_left_from >= 4 && n_left_to_next >= 2)
{
u32 pi0, pi1;
vlib_buffer_t *p0, *p1;
ip6_mapt_tcp_udp_next_t next0, next1;
pi0 = to_next[0] = from[0];
pi1 = to_next[1] = from[1];
from += 2;
n_left_from -= 2;
to_next += 2;
n_left_to_next -= 2;
next0 = IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP;
next1 = IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP;
p0 = vlib_get_buffer (vm, pi0);
p1 = vlib_get_buffer (vm, pi1);
if (ip6_to_ip4_tcp_udp (p0, ip6_to_ip4_set_cb, p0, 1))
{
p0->error = error_node->errors[MAP_ERROR_UNKNOWN];
next0 = IP6_MAPT_TCP_UDP_NEXT_DROP;
}
else
{
if (vnet_buffer (p0)->map_t.mtu < p0->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p0)->ip_frag.mtu = vnet_buffer (p0)->map_t.mtu;
vnet_buffer (p0)->ip_frag.header_offset = 0;
vnet_buffer (p0)->ip_frag.next_index =
IP4_FRAG_NEXT_IP4_LOOKUP;
next0 = IP6_MAPT_TCP_UDP_NEXT_IP4_FRAG;
}
}
if (ip6_to_ip4_tcp_udp (p1, ip6_to_ip4_set_cb, p1, 1))
{
p1->error = error_node->errors[MAP_ERROR_UNKNOWN];
next1 = IP6_MAPT_TCP_UDP_NEXT_DROP;
}
else
{
if (vnet_buffer (p1)->map_t.mtu < p1->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p1)->ip_frag.mtu = vnet_buffer (p1)->map_t.mtu;
vnet_buffer (p1)->ip_frag.header_offset = 0;
vnet_buffer (p1)->ip_frag.next_index =
IP4_FRAG_NEXT_IP4_LOOKUP;
next1 = IP6_MAPT_TCP_UDP_NEXT_IP4_FRAG;
}
}
vlib_validate_buffer_enqueue_x2 (vm, node, next_index, to_next,
n_left_to_next, pi0, pi1, next0,
next1);
}
#endif
while (n_left_from > 0 && n_left_to_next > 0)
{
u32 pi0;
vlib_buffer_t *p0;
ip6_mapt_tcp_udp_next_t next0;
pi0 = to_next[0] = from[0];
from += 1;
n_left_from -= 1;
to_next += 1;
n_left_to_next -= 1;
next0 = IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP;
p0 = vlib_get_buffer (vm, pi0);
if (ip6_to_ip4_tcp_udp (p0, ip6_to_ip4_set_cb, p0, 1))
{
p0->error = error_node->errors[MAP_ERROR_UNKNOWN];
next0 = IP6_MAPT_TCP_UDP_NEXT_DROP;
}
else
{
if (vnet_buffer (p0)->map_t.mtu < p0->current_length)
{
//Send to fragmentation node if necessary
vnet_buffer (p0)->ip_frag.mtu = vnet_buffer (p0)->map_t.mtu;
vnet_buffer (p0)->ip_frag.header_offset = 0;
vnet_buffer (p0)->ip_frag.next_index =
IP4_FRAG_NEXT_IP4_LOOKUP;
next0 = IP6_MAPT_TCP_UDP_NEXT_IP4_FRAG;
}
}
vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
to_next, n_left_to_next, pi0,
next0);
}
vlib_put_next_frame (vm, node, next_index, n_left_to_next);
}
return frame->n_vectors;
}
static_always_inline void
ip6_map_t_classify (vlib_buffer_t * p0, ip6_header_t * ip60,
map_domain_t * d0, i32 * src_port0,
u8 * error0, ip6_mapt_next_t * next0,
u32 l4_len0, ip6_frag_hdr_t * frag0)
{
if (PREDICT_FALSE (vnet_buffer (p0)->map_t.v6.frag_offset &&
ip6_frag_hdr_offset (frag0)))
{
*next0 = IP6_MAPT_NEXT_MAPT_FRAGMENTED;
if (d0->ea_bits_len == 0 && d0->rules)
{
*src_port0 = 0;
}
else
{
*src_port0 = ip6_map_fragment_get (ip60, frag0, d0);
*error0 = (*src_port0 != -1) ? *error0 : MAP_ERROR_FRAGMENT_DROPPED;
}
}
else
if (PREDICT_TRUE
(vnet_buffer (p0)->map_t.v6.l4_protocol == IP_PROTOCOL_TCP))
{
*error0 =
l4_len0 < sizeof (tcp_header_t) ? MAP_ERROR_MALFORMED : *error0;
vnet_buffer (p0)->map_t.checksum_offset =
vnet_buffer (p0)->map_t.v6.l4_offset + 16;
*next0 = IP6_MAPT_NEXT_MAPT_TCP_UDP;
*src_port0 =
(i32) *
((u16 *) u8_ptr_add (ip60, vnet_buffer (p0)->map_t.v6.l4_offset));
}
else
if (PREDICT_TRUE
(vnet_buffer (p0)->map_t.v6.l4_protocol == IP_PROTOCOL_UDP))
{
*error0 =
l4_len0 < sizeof (udp_header_t) ? MAP_ERROR_MALFORMED : *error0;
vnet_buffer (p0)->map_t.checksum_offset =
vnet_buffer (p0)->map_t.v6.l4_offset + 6;
*next0 = IP6_MAPT_NEXT_MAPT_TCP_UDP;
*src_port0 =
(i32) *
((u16 *) u8_ptr_add (ip60, vnet_buffer (p0)->map_t.v6.l4_offset));
}
else if (vnet_buffer (p0)->map_t.v6.l4_protocol == IP_PROTOCOL_ICMP6)
{
*error0 =
l4_len0 < sizeof (icmp46_header_t) ? MAP_ERROR_MALFORMED : *error0;
*next0 = IP6_MAPT_NEXT_MAPT_ICMP;
if (d0->ea_bits_len == 0 && d0->rules)
{
*src_port0 = 0;
}
else
if (((icmp46_header_t *)
u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.v6.l4_offset))->code ==
ICMP6_echo_reply
|| ((icmp46_header_t *)
u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.v6.l4_offset))->code ==
ICMP6_echo_request)
{
*src_port0 =
(i32) *
((u16 *)
u8_ptr_add (ip60, vnet_buffer (p0)->map_t.v6.l4_offset + 6));
}
}
else
{
//TODO: In case of 1:1 mapping, it might be possible to do something with those packets.
*error0 = MAP_ERROR_BAD_PROTOCOL;
}
}
static uword
ip6_map_t (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * frame)
{
u32 n_left_from, *from, next_index, *to_next, n_left_to_next;
vlib_node_runtime_t *error_node =
vlib_node_get_runtime (vm, ip6_map_t_node.index);
vlib_combined_counter_main_t *cm = map_main.domain_counters;
u32 thread_index = vlib_get_thread_index ();
from = vlib_frame_vector_args (frame);
n_left_from = frame->n_vectors;
next_index = node->cached_next_index;
while (n_left_from > 0)
{
vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
#ifdef IP6_MAP_T_DUAL_LOOP
while (n_left_from >= 4 && n_left_to_next >= 2)
{
u32 pi0, pi1;
vlib_buffer_t *p0, *p1;
ip6_header_t *ip60, *ip61;
u8 error0, error1;
ip6_mapt_next_t next0, next1;
u32 l4_len0, l4_len1;
i32 src_port0, src_port1;
map_domain_t *d0, *d1;
ip6_frag_hdr_t *frag0, *frag1;
u32 saddr0, saddr1;
next0 = next1 = 0; //Because compiler whines
pi0 = to_next[0] = from[0];
pi1 = to_next[1] = from[1];
from += 2;
n_left_from -= 2;
to_next += 2;
n_left_to_next -= 2;
error0 = MAP_ERROR_NONE;
error1 = MAP_ERROR_NONE;
p0 = vlib_get_buffer (vm, pi0);
p1 = vlib_get_buffer (vm, pi1);
ip60 = vlib_buffer_get_current (p0);
ip61 = vlib_buffer_get_current (p1);
saddr0 = map_get_ip4 (&ip60->src_address);
saddr1 = map_get_ip4 (&ip61->src_address);
d0 = ip6_map_get_domain (vnet_buffer (p0)->ip.adj_index[VLIB_TX],
(ip4_address_t *) & saddr0,
&vnet_buffer (p0)->map_t.map_domain_index,
&error0);
d1 =
ip6_map_get_domain (vnet_buffer (p1)->ip.adj_index[VLIB_TX],
(ip4_address_t *) & saddr1,
&vnet_buffer (p1)->map_t.map_domain_index,
&error1);
vnet_buffer (p0)->map_t.v6.saddr = saddr0;
vnet_buffer (p1)->map_t.v6.saddr = saddr1;
vnet_buffer (p0)->map_t.v6.daddr =
ip6_map_t_embedded_address (d0, &ip60->dst_address);
vnet_buffer (p1)->map_t.v6.daddr =
ip6_map_t_embedded_address (d1, &ip61->dst_address);
vnet_buffer (p0)->map_t.mtu = d0->mtu ? d0->mtu : ~0;
vnet_buffer (p1)->map_t.mtu = d1->mtu ? d1->mtu : ~0;
if (PREDICT_FALSE (ip6_parse (ip60, p0->current_length,
&(vnet_buffer (p0)->map_t.
v6.l4_protocol),
&(vnet_buffer (p0)->map_t.
v6.l4_offset),
&(vnet_buffer (p0)->map_t.
v6.frag_offset))))
{
error0 = MAP_ERROR_MALFORMED;
next0 = IP6_MAPT_NEXT_DROP;
}
if (PREDICT_FALSE (ip6_parse (ip61, p1->current_length,
&(vnet_buffer (p1)->map_t.
v6.l4_protocol),
&(vnet_buffer (p1)->map_t.
v6.l4_offset),
&(vnet_buffer (p1)->map_t.
v6.frag_offset))))
{
error1 = MAP_ERROR_MALFORMED;
next1 = IP6_MAPT_NEXT_DROP;
}
src_port0 = src_port1 = -1;
l4_len0 = (u32) clib_net_to_host_u16 (ip60->payload_length) +
sizeof (*ip60) - vnet_buffer (p0)->map_t.v6.l4_offset;
l4_len1 = (u32) clib_net_to_host_u16 (ip61->payload_length) +
sizeof (*ip60) - vnet_buffer (p1)->map_t.v6.l4_offset;
frag0 =
(ip6_frag_hdr_t *) u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.
v6.frag_offset);
frag1 =
(ip6_frag_hdr_t *) u8_ptr_add (ip61,
vnet_buffer (p1)->map_t.
v6.frag_offset);
ip6_map_t_classify (p0, ip60, d0, &src_port0, &error0, &next0,
l4_len0, frag0);
ip6_map_t_classify (p1, ip61, d1, &src_port1, &error1, &next1,
l4_len1, frag1);
if (PREDICT_FALSE
((src_port0 != -1)
&& (ip60->src_address.as_u64[0] !=
map_get_pfx_net (d0, vnet_buffer (p0)->map_t.v6.saddr,
src_port0)
|| ip60->src_address.as_u64[1] != map_get_sfx_net (d0,
vnet_buffer
(p0)->map_t.v6.saddr,
src_port0))))
{
error0 = MAP_ERROR_SEC_CHECK;
}
if (PREDICT_FALSE
((src_port1 != -1)
&& (ip61->src_address.as_u64[0] !=
map_get_pfx_net (d1, vnet_buffer (p1)->map_t.v6.saddr,
src_port1)
|| ip61->src_address.as_u64[1] != map_get_sfx_net (d1,
vnet_buffer
(p1)->map_t.v6.saddr,
src_port1))))
{
error1 = MAP_ERROR_SEC_CHECK;
}
if (PREDICT_FALSE (vnet_buffer (p0)->map_t.v6.frag_offset &&
!ip6_frag_hdr_offset ((ip6_frag_hdr_t *)
u8_ptr_add (ip60,
vnet_buffer
(p0)->map_t.
v6.frag_offset)))
&& (src_port0 != -1) && (d0->ea_bits_len != 0 || !d0->rules)
&& (error0 == MAP_ERROR_NONE))
{
ip6_map_fragment_cache (ip60,
(ip6_frag_hdr_t *) u8_ptr_add (ip60,
vnet_buffer
(p0)->map_t.
v6.frag_offset),
d0, src_port0);
}
if (PREDICT_FALSE (vnet_buffer (p1)->map_t.v6.frag_offset &&
!ip6_frag_hdr_offset ((ip6_frag_hdr_t *)
u8_ptr_add (ip61,
vnet_buffer
(p1)->map_t.
v6.frag_offset)))
&& (src_port1 != -1) && (d1->ea_bits_len != 0 || !d1->rules)
&& (error1 == MAP_ERROR_NONE))
{
ip6_map_fragment_cache (ip61,
(ip6_frag_hdr_t *) u8_ptr_add (ip61,
vnet_buffer
(p1)->map_t.
v6.frag_offset),
d1, src_port1);
}
if (PREDICT_TRUE
(error0 == MAP_ERROR_NONE && next0 != IP6_MAPT_NEXT_MAPT_ICMP))
{
vlib_increment_combined_counter (cm + MAP_DOMAIN_COUNTER_RX,
thread_index,
vnet_buffer (p0)->
map_t.map_domain_index, 1,
clib_net_to_host_u16
(ip60->payload_length));
}
if (PREDICT_TRUE
(error1 == MAP_ERROR_NONE && next1 != IP6_MAPT_NEXT_MAPT_ICMP))
{
vlib_increment_combined_counter (cm + MAP_DOMAIN_COUNTER_RX,
thread_index,
vnet_buffer (p1)->
map_t.map_domain_index, 1,
clib_net_to_host_u16
(ip61->payload_length));
}
next0 = (error0 != MAP_ERROR_NONE) ? IP6_MAPT_NEXT_DROP : next0;
next1 = (error1 != MAP_ERROR_NONE) ? IP6_MAPT_NEXT_DROP : next1;
p0->error = error_node->errors[error0];
p1->error = error_node->errors[error1];
vlib_validate_buffer_enqueue_x2 (vm, node, next_index, to_next,
n_left_to_next, pi0, pi1, next0,
next1);
}
#endif
while (n_left_from > 0 && n_left_to_next > 0)
{
u32 pi0;
vlib_buffer_t *p0;
ip6_header_t *ip60;
u8 error0;
u32 l4_len0;
i32 src_port0;
map_domain_t *d0;
ip6_frag_hdr_t *frag0;
ip6_mapt_next_t next0 = 0;
u32 saddr;
pi0 = to_next[0] = from[0];
from += 1;
n_left_from -= 1;
to_next += 1;
n_left_to_next -= 1;
error0 = MAP_ERROR_NONE;
p0 = vlib_get_buffer (vm, pi0);
ip60 = vlib_buffer_get_current (p0);
//Save saddr in a different variable to not overwrite ip.adj_index
saddr = map_get_ip4 (&ip60->src_address);
d0 = ip6_map_get_domain (vnet_buffer (p0)->ip.adj_index[VLIB_TX],
(ip4_address_t *) & saddr,
&vnet_buffer (p0)->map_t.map_domain_index,
&error0);
//FIXME: What if d0 is null
vnet_buffer (p0)->map_t.v6.saddr = saddr;
vnet_buffer (p0)->map_t.v6.daddr =
ip6_map_t_embedded_address (d0, &ip60->dst_address);
vnet_buffer (p0)->map_t.mtu = d0->mtu ? d0->mtu : ~0;
if (PREDICT_FALSE (ip6_parse (ip60, p0->current_length,
&(vnet_buffer (p0)->map_t.
v6.l4_protocol),
&(vnet_buffer (p0)->map_t.
v6.l4_offset),
&(vnet_buffer (p0)->map_t.
v6.frag_offset))))
{
error0 = MAP_ERROR_MALFORMED;
next0 = IP6_MAPT_NEXT_DROP;
}
src_port0 = -1;
l4_len0 = (u32) clib_net_to_host_u16 (ip60->payload_length) +
sizeof (*ip60) - vnet_buffer (p0)->map_t.v6.l4_offset;
frag0 =
(ip6_frag_hdr_t *) u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.
v6.frag_offset);
if (PREDICT_FALSE (vnet_buffer (p0)->map_t.v6.frag_offset &&
ip6_frag_hdr_offset (frag0)))
{
src_port0 = ip6_map_fragment_get (ip60, frag0, d0);
error0 = (src_port0 != -1) ? error0 : MAP_ERROR_FRAGMENT_MEMORY;
next0 = IP6_MAPT_NEXT_MAPT_FRAGMENTED;
}
else
if (PREDICT_TRUE
(vnet_buffer (p0)->map_t.v6.l4_protocol == IP_PROTOCOL_TCP))
{
error0 =
l4_len0 <
sizeof (tcp_header_t) ? MAP_ERROR_MALFORMED : error0;
vnet_buffer (p0)->map_t.checksum_offset =
vnet_buffer (p0)->map_t.v6.l4_offset + 16;
next0 = IP6_MAPT_NEXT_MAPT_TCP_UDP;
src_port0 =
(i32) *
((u16 *)
u8_ptr_add (ip60, vnet_buffer (p0)->map_t.v6.l4_offset));
}
else
if (PREDICT_TRUE
(vnet_buffer (p0)->map_t.v6.l4_protocol == IP_PROTOCOL_UDP))
{
error0 =
l4_len0 <
sizeof (udp_header_t) ? MAP_ERROR_MALFORMED : error0;
vnet_buffer (p0)->map_t.checksum_offset =
vnet_buffer (p0)->map_t.v6.l4_offset + 6;
next0 = IP6_MAPT_NEXT_MAPT_TCP_UDP;
src_port0 =
(i32) *
((u16 *)
u8_ptr_add (ip60, vnet_buffer (p0)->map_t.v6.l4_offset));
}
else if (vnet_buffer (p0)->map_t.v6.l4_protocol ==
IP_PROTOCOL_ICMP6)
{
error0 =
l4_len0 <
sizeof (icmp46_header_t) ? MAP_ERROR_MALFORMED : error0;
next0 = IP6_MAPT_NEXT_MAPT_ICMP;
if (((icmp46_header_t *)
u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.v6.l4_offset))->code ==
ICMP6_echo_reply
|| ((icmp46_header_t *)
u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.v6.
l4_offset))->code == ICMP6_echo_request)
src_port0 =
(i32) *
((u16 *)
u8_ptr_add (ip60,
vnet_buffer (p0)->map_t.v6.l4_offset + 6));
}
else
{
//TODO: In case of 1:1 mapping, it might be possible to do something with those packets.
error0 = MAP_ERROR_BAD_PROTOCOL;
}
//Security check
if (PREDICT_FALSE
((src_port0 != -1)
&& (ip60->src_address.as_u64[0] !=
map_get_pfx_net (d0, vnet_buffer (p0)->map_t.v6.saddr,
src_port0)
|| ip60->src_address.as_u64[1] != map_get_sfx_net (d0,
vnet_buffer
(p0)->map_t.v6.saddr,
src_port0))))
{
//Security check when src_port0 is not zero (non-first fragment, UDP or TCP)
error0 = MAP_ERROR_SEC_CHECK;
}
//Fragmented first packet needs to be cached for following packets
if (PREDICT_FALSE (vnet_buffer (p0)->map_t.v6.frag_offset &&
!ip6_frag_hdr_offset ((ip6_frag_hdr_t *)
u8_ptr_add (ip60,
vnet_buffer
(p0)->map_t.
v6.frag_offset)))
&& (src_port0 != -1) && (d0->ea_bits_len != 0 || !d0->rules)
&& (error0 == MAP_ERROR_NONE))
{
ip6_map_fragment_cache (ip60,
(ip6_frag_hdr_t *) u8_ptr_add (ip60,
vnet_buffer
(p0)->map_t.
v6.frag_offset),
d0, src_port0);
}
if (PREDICT_TRUE
(error0 == MAP_ERROR_NONE && next0 != IP6_MAPT_NEXT_MAPT_ICMP))
{
vlib_increment_combined_counter (cm + MAP_DOMAIN_COUNTER_RX,
thread_index,
vnet_buffer (p0)->
map_t.map_domain_index, 1,
clib_net_to_host_u16
(ip60->payload_length));
}
next0 = (error0 != MAP_ERROR_NONE) ? IP6_MAPT_NEXT_DROP : next0;
p0->error = error_node->errors[error0];
vlib_validate_buffer_enqueue_x1 (vm, node, next_index,
to_next, n_left_to_next, pi0,
next0);
}
vlib_put_next_frame (vm, node, next_index, n_left_to_next);
}
return frame->n_vectors;
}
static char *map_t_error_strings[] = {
#define _(sym,string) string,
foreach_map_error
#undef _
};
/* *INDENT-OFF* */
VLIB_REGISTER_NODE(ip6_map_t_fragmented_node) = {
.function = ip6_map_t_fragmented,
.name = "ip6-map-t-fragmented",
.vector_size = sizeof (u32),
.format_trace = format_map_trace,
.type = VLIB_NODE_TYPE_INTERNAL,
.n_errors = MAP_N_ERROR,
.error_strings = map_t_error_strings,
.n_next_nodes = IP6_MAPT_FRAGMENTED_N_NEXT,
.next_nodes = {
[IP6_MAPT_FRAGMENTED_NEXT_IP4_LOOKUP] = "ip4-lookup",
[IP6_MAPT_FRAGMENTED_NEXT_IP4_FRAG] = IP4_FRAG_NODE_NAME,
[IP6_MAPT_FRAGMENTED_NEXT_DROP] = "error-drop",
},
};
/* *INDENT-ON* */
/* *INDENT-OFF* */
VLIB_REGISTER_NODE(ip6_map_t_icmp_node) = {
.function = ip6_map_t_icmp,
.name = "ip6-map-t-icmp",
.vector_size = sizeof (u32),
.format_trace = format_map_trace,
.type = VLIB_NODE_TYPE_INTERNAL,
.n_errors = MAP_N_ERROR,
.error_strings = map_t_error_strings,
.n_next_nodes = IP6_MAPT_ICMP_N_NEXT,
.next_nodes = {
[IP6_MAPT_ICMP_NEXT_IP4_LOOKUP] = "ip4-lookup",
[IP6_MAPT_ICMP_NEXT_IP4_FRAG] = IP4_FRAG_NODE_NAME,
[IP6_MAPT_ICMP_NEXT_DROP] = "error-drop",
},
};
/* *INDENT-ON* */
/* *INDENT-OFF* */
VLIB_REGISTER_NODE(ip6_map_t_tcp_udp_node) = {
.function = ip6_map_t_tcp_udp,
.name = "ip6-map-t-tcp-udp",
.vector_size = sizeof (u32),
.format_trace = format_map_trace,
.type = VLIB_NODE_TYPE_INTERNAL,
.n_errors = MAP_N_ERROR,
.error_strings = map_t_error_strings,
.n_next_nodes = IP6_MAPT_TCP_UDP_N_NEXT,
.next_nodes = {
[IP6_MAPT_TCP_UDP_NEXT_IP4_LOOKUP] = "ip4-lookup",
[IP6_MAPT_TCP_UDP_NEXT_IP4_FRAG] = IP4_FRAG_NODE_NAME,
[IP6_MAPT_TCP_UDP_NEXT_DROP] = "error-drop",
},
};
/* *INDENT-ON* */
/* *INDENT-OFF* */
VLIB_REGISTER_NODE(ip6_map_t_node) = {
.function = ip6_map_t,
.name = "ip6-map-t",
.vector_size = sizeof(u32),
.format_trace = format_map_trace,
.type = VLIB_NODE_TYPE_INTERNAL,
.n_errors = MAP_N_ERROR,
.error_strings = map_t_error_strings,
.n_next_nodes = IP6_MAPT_N_NEXT,
.next_nodes = {
[IP6_MAPT_NEXT_MAPT_TCP_UDP] = "ip6-map-t-tcp-udp",
[IP6_MAPT_NEXT_MAPT_ICMP] = "ip6-map-t-icmp",
[IP6_MAPT_NEXT_MAPT_FRAGMENTED] = "ip6-map-t-fragmented",
[IP6_MAPT_NEXT_DROP] = "error-drop",
},
};
/* *INDENT-ON* */
/*
* fd.io coding-style-patch-verification: ON
*
* Local Variables:
* eval: (c-set-style "gnu")
* End:
*/