diff options
Diffstat (limited to 'src/vnet/ip/ip6.h')
-rw-r--r-- | src/vnet/ip/ip6.h | 605 |
1 files changed, 605 insertions, 0 deletions
diff --git a/src/vnet/ip/ip6.h b/src/vnet/ip/ip6.h new file mode 100644 index 00000000..8aef53a9 --- /dev/null +++ b/src/vnet/ip/ip6.h @@ -0,0 +1,605 @@ +/* + * Copyright (c) 2016 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. + */ +/* + * ip/ip6.h: ip6 main include file + * + * Copyright (c) 2008 Eliot Dresselhaus + * + * Permission is hereby granted, free of charge, to any person obtaining + * a copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sublicense, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice shall be + * included in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE + * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION + * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION + * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef included_ip_ip6_h +#define included_ip_ip6_h + +#include <vlib/mc.h> +#include <vlib/buffer.h> +#include <vnet/ethernet/packet.h> +#include <vnet/ip/ip6_packet.h> +#include <vnet/ip/ip6_hop_by_hop_packet.h> +#include <vnet/ip/lookup.h> +#include <stdbool.h> +#include <vppinfra/bihash_24_8.h> +#include <vppinfra/bihash_template.h> +#include <vnet/util/radix.h> + +/* + * Default size of the ip6 fib hash table + */ +#define IP6_FIB_DEFAULT_HASH_NUM_BUCKETS (64 * 1024) +#define IP6_FIB_DEFAULT_HASH_MEMORY_SIZE (32<<20) + +typedef struct +{ + ip6_address_t addr; + u32 dst_address_length; + u32 vrf_index; +} ip6_fib_key_t; + +typedef struct +{ + /* Table ID (hash key) for this FIB. */ + u32 table_id; + + /* Index into FIB vector. */ + u32 index; +} ip6_fib_t; + +typedef struct ip6_mfib_t +{ + /* Table ID (hash key) for this FIB. */ + u32 table_id; + + /* Index into FIB vector. */ + u32 index; + + /* + * Pointer to the top of a radix tree. + * This cannot be realloc'd, hence it cannot be inlined with this table + */ + struct radix_node_head *rhead; +} ip6_mfib_t; + +struct ip6_main_t; + +typedef void (ip6_add_del_interface_address_function_t) + (struct ip6_main_t * im, + uword opaque, + u32 sw_if_index, + ip6_address_t * address, + u32 address_length, u32 if_address_index, u32 is_del); + +typedef struct +{ + ip6_add_del_interface_address_function_t *function; + uword function_opaque; +} ip6_add_del_interface_address_callback_t; + +typedef void (ip6_table_bind_function_t) + (struct ip6_main_t * im, + uword opaque, u32 sw_if_index, u32 new_fib_index, u32 old_fib_index); + +typedef struct +{ + ip6_table_bind_function_t *function; + uword function_opaque; +} ip6_table_bind_callback_t; + +/** + * Enumeration of the FIB table instance types + */ +typedef enum ip6_fib_table_instance_type_t_ +{ + /** + * This table stores the routes that are used to forward traffic. + * The key is the prefix, the result the adjacnecy to forward on. + */ + IP6_FIB_TABLE_FWDING, + /** + * The table that stores ALL routes learned by the DP. + * Some of these routes may not be ready to install in forwarding + * at a given time. + * The key in this table is the prefix, the result is the fib_entry_t + */ + IP6_FIB_TABLE_NON_FWDING, +} ip6_fib_table_instance_type_t; + +#define IP6_FIB_NUM_TABLES (IP6_FIB_TABLE_NON_FWDING+1) + +/** + * A represenation of a single IP6 table + */ +typedef struct ip6_fib_table_instance_t_ +{ + /* The hash table */ + BVT (clib_bihash) ip6_hash; + + /* bitmap / refcounts / vector of mask widths to search */ + uword *non_empty_dst_address_length_bitmap; + u8 *prefix_lengths_in_search_order; + i32 dst_address_length_refcounts[129]; +} ip6_fib_table_instance_t; + +typedef struct ip6_main_t +{ + /** + * The two FIB tables; fwding and non-fwding + */ + ip6_fib_table_instance_t ip6_table[IP6_FIB_NUM_TABLES]; + + ip_lookup_main_t lookup_main; + + /* Pool of FIBs. */ + struct fib_table_t_ *fibs; + + /* Pool of V6 FIBs. */ + ip6_fib_t *v6_fibs; + + /** Vector of MFIBs. */ + struct mfib_table_t_ *mfibs; + + /* Network byte orders subnet mask for each prefix length */ + ip6_address_t fib_masks[129]; + + /* Table index indexed by software interface. */ + u32 *fib_index_by_sw_if_index; + + /** Table index indexed by software interface. */ + u32 *mfib_index_by_sw_if_index; + + /* IP6 enabled count by software interface */ + u8 *ip_enabled_by_sw_if_index; + + /* Hash table mapping table id to fib index. + ID space is not necessarily dense; index space is dense. */ + uword *fib_index_by_table_id; + + /** Hash table mapping table id to multicast fib index. + ID space is not necessarily dense; index space is dense. */ + uword *mfib_index_by_table_id; + + /* Hash table mapping interface rewrite adjacency index by sw if index. */ + uword *interface_route_adj_index_by_sw_if_index; + + /* Functions to call when interface address changes. */ + ip6_add_del_interface_address_callback_t + * add_del_interface_address_callbacks; + + /** Functions to call when interface to table biding changes. */ + ip6_table_bind_callback_t *table_bind_callbacks; + + /* Template used to generate IP6 neighbor solicitation packets. */ + vlib_packet_template_t discover_neighbor_packet_template; + + /* ip6 lookup table config parameters */ + u32 lookup_table_nbuckets; + uword lookup_table_size; + + /* Seed for Jenkins hash used to compute ip6 flow hash. */ + u32 flow_hash_seed; + + struct + { + /* TTL to use for host generated packets. */ + u8 ttl; + + u8 pad[3]; + } host_config; + + /* HBH processing enabled? */ + u8 hbh_enabled; +} ip6_main_t; + +/* Global ip6 main structure. */ +extern ip6_main_t ip6_main; + +/* Global ip6 input node. Errors get attached to ip6 input node. */ +extern vlib_node_registration_t ip6_input_node; +extern vlib_node_registration_t ip6_rewrite_node; +extern vlib_node_registration_t ip6_rewrite_mcast_node; +extern vlib_node_registration_t ip6_rewrite_local_node; +extern vlib_node_registration_t ip6_discover_neighbor_node; +extern vlib_node_registration_t ip6_glean_node; +extern vlib_node_registration_t ip6_midchain_node; + +always_inline uword +ip6_destination_matches_route (const ip6_main_t * im, + const ip6_address_t * key, + const ip6_address_t * dest, uword dest_length) +{ + int i; + for (i = 0; i < ARRAY_LEN (key->as_uword); i++) + { + if ((key->as_uword[i] ^ dest->as_uword[i]) & im-> + fib_masks[dest_length].as_uword[i]) + return 0; + } + return 1; +} + +always_inline uword +ip6_destination_matches_interface (ip6_main_t * im, + ip6_address_t * key, + ip_interface_address_t * ia) +{ + ip6_address_t *a = ip_interface_address_get_address (&im->lookup_main, ia); + return ip6_destination_matches_route (im, key, a, ia->address_length); +} + +/* As above but allows for unaligned destinations (e.g. works right from IP header of packet). */ +always_inline uword +ip6_unaligned_destination_matches_route (ip6_main_t * im, + ip6_address_t * key, + ip6_address_t * dest, + uword dest_length) +{ + int i; + for (i = 0; i < ARRAY_LEN (key->as_uword); i++) + { + if ((clib_mem_unaligned (&key->as_uword[i], uword) ^ dest->as_uword[i]) + & im->fib_masks[dest_length].as_uword[i]) + return 0; + } + return 1; +} + +always_inline int +ip6_src_address_for_packet (ip_lookup_main_t * lm, + u32 sw_if_index, ip6_address_t * src) +{ + u32 if_add_index = lm->if_address_pool_index_by_sw_if_index[sw_if_index]; + if (PREDICT_TRUE (if_add_index != ~0)) + { + ip_interface_address_t *if_add = + pool_elt_at_index (lm->if_address_pool, if_add_index); + ip6_address_t *if_ip = ip_interface_address_get_address (lm, if_add); + *src = *if_ip; + return (0); + } + else + { + src->as_u64[0] = 0; + src->as_u64[1] = 0; + } + return (!0); +} + +/* Find interface address which matches destination. */ +always_inline ip6_address_t * +ip6_interface_address_matching_destination (ip6_main_t * im, + ip6_address_t * dst, + u32 sw_if_index, + ip_interface_address_t ** + result_ia) +{ + ip_lookup_main_t *lm = &im->lookup_main; + ip_interface_address_t *ia; + ip6_address_t *result = 0; + + /* *INDENT-OFF* */ + foreach_ip_interface_address (lm, ia, sw_if_index, + 1 /* honor unnumbered */, + ({ + ip6_address_t * a = ip_interface_address_get_address (lm, ia); + if (ip6_destination_matches_route (im, dst, a, ia->address_length)) + { + result = a; + break; + } + })); + /* *INDENT-ON* */ + if (result_ia) + *result_ia = result ? ia : 0; + return result; +} + +clib_error_t *ip6_add_del_interface_address (vlib_main_t * vm, + u32 sw_if_index, + ip6_address_t * address, + u32 address_length, u32 is_del); +void ip6_sw_interface_enable_disable (u32 sw_if_index, u32 is_enable); + +/** + * @brie get first IPv6 interface address + */ +ip6_address_t *ip6_interface_first_address (ip6_main_t * im, u32 sw_if_index); + +int ip6_address_compare (ip6_address_t * a1, ip6_address_t * a2); + +clib_error_t *ip6_probe_neighbor (vlib_main_t * vm, ip6_address_t * dst, + u32 sw_if_index); + +uword +ip6_udp_register_listener (vlib_main_t * vm, + u16 dst_port, u32 next_node_index); + +u16 ip6_tcp_udp_icmp_compute_checksum (vlib_main_t * vm, vlib_buffer_t * p0, + ip6_header_t * ip0, + int *bogus_lengthp); + +void ip6_register_protocol (u32 protocol, u32 node_index); + +serialize_function_t serialize_vnet_ip6_main, unserialize_vnet_ip6_main; + +void ip6_ethernet_update_adjacency (vnet_main_t * vnm, + u32 sw_if_index, u32 ai); + + +void +ip6_link_local_address_from_ethernet_mac_address (ip6_address_t * ip, + u8 * mac); + +void +ip6_ethernet_mac_address_from_link_local_address (u8 * mac, + ip6_address_t * ip); + +int vnet_set_ip6_flow_hash (u32 table_id, + flow_hash_config_t flow_hash_config); + +clib_error_t *enable_ip6_interface (vlib_main_t * vm, u32 sw_if_index); + +clib_error_t *disable_ip6_interface (vlib_main_t * vm, u32 sw_if_index); + +int ip6_interface_enabled (vlib_main_t * vm, u32 sw_if_index); + +clib_error_t *set_ip6_link_local_address (vlib_main_t * vm, + u32 sw_if_index, + ip6_address_t * address); + +int vnet_add_del_ip6_nd_change_event (vnet_main_t * vnm, + void *data_callback, + u32 pid, + void *address_arg, + uword node_index, + uword type_opaque, + uword data, int is_add); + +int vnet_ip6_nd_term (vlib_main_t * vm, + vlib_node_runtime_t * node, + vlib_buffer_t * p0, + ethernet_header_t * eth, + ip6_header_t * ip, u32 sw_if_index, u16 bd_index); + +void send_ip6_na (vlib_main_t * vm, vnet_hw_interface_t * hi); + +u8 *format_ip6_forward_next_trace (u8 * s, va_list * args); + +u32 ip6_tcp_udp_icmp_validate_checksum (vlib_main_t * vm, vlib_buffer_t * p0); + +int vnet_set_ip6_classify_intfc (vlib_main_t * vm, u32 sw_if_index, + u32 table_index); +extern vlib_node_registration_t ip6_lookup_node; + +/* Compute flow hash. We'll use it to select which Sponge to use for this + flow. And other things. */ +always_inline u32 +ip6_compute_flow_hash (const ip6_header_t * ip, + flow_hash_config_t flow_hash_config) +{ + tcp_header_t *tcp; + u64 a, b, c; + u64 t1, t2; + uword is_tcp_udp = 0; + u8 protocol = ip->protocol; + + if (PREDICT_TRUE + ((ip->protocol == IP_PROTOCOL_TCP) + || (ip->protocol == IP_PROTOCOL_UDP))) + { + is_tcp_udp = 1; + tcp = (void *) (ip + 1); + } + else if (ip->protocol == IP_PROTOCOL_IP6_HOP_BY_HOP_OPTIONS) + { + ip6_hop_by_hop_header_t *hbh = (ip6_hop_by_hop_header_t *) (ip + 1); + if ((hbh->protocol == IP_PROTOCOL_TCP) || + (hbh->protocol == IP_PROTOCOL_UDP)) + { + is_tcp_udp = 1; + tcp = (tcp_header_t *) ((u8 *) hbh + ((hbh->length + 1) << 3)); + } + protocol = hbh->protocol; + } + + t1 = (ip->src_address.as_u64[0] ^ ip->src_address.as_u64[1]); + t1 = (flow_hash_config & IP_FLOW_HASH_SRC_ADDR) ? t1 : 0; + + t2 = (ip->dst_address.as_u64[0] ^ ip->dst_address.as_u64[1]); + t2 = (flow_hash_config & IP_FLOW_HASH_DST_ADDR) ? t2 : 0; + + a = (flow_hash_config & IP_FLOW_HASH_REVERSE_SRC_DST) ? t2 : t1; + b = (flow_hash_config & IP_FLOW_HASH_REVERSE_SRC_DST) ? t1 : t2; + b ^= (flow_hash_config & IP_FLOW_HASH_PROTO) ? protocol : 0; + + t1 = is_tcp_udp ? tcp->src : 0; + t2 = is_tcp_udp ? tcp->dst : 0; + + t1 = (flow_hash_config & IP_FLOW_HASH_SRC_PORT) ? t1 : 0; + t2 = (flow_hash_config & IP_FLOW_HASH_DST_PORT) ? t2 : 0; + + c = (flow_hash_config & IP_FLOW_HASH_REVERSE_SRC_DST) ? + ((t1 << 16) | t2) : ((t2 << 16) | t1); + + hash_mix64 (a, b, c); + return (u32) c; +} + +/* ip6_locate_header + * + * This function is to search for the header specified by the protocol number + * in find_hdr_type. + * This is used to locate a specific IPv6 extension header + * or to find transport layer header. + * 1. If the find_hdr_type < 0 then it finds and returns the protocol number and + * offset stored in *offset of the transport or ESP header in the chain if + * found. + * 2. If a header with find_hdr_type > 0 protocol number is found then the + * offset is stored in *offset and protocol number of the header is + * returned. + * 3. If find_hdr_type is not found or packet is malformed or + * it is a non-first fragment -1 is returned. + */ +always_inline int +ip6_locate_header (vlib_buffer_t * p0, + ip6_header_t * ip0, int find_hdr_type, u32 * offset) +{ + u8 next_proto = ip0->protocol; + u8 *next_header; + u8 done = 0; + u32 cur_offset; + u8 *temp_nxthdr = 0; + u32 exthdr_len = 0; + + next_header = ip6_next_header (ip0); + cur_offset = sizeof (ip6_header_t); + while (1) + { + done = (next_proto == find_hdr_type); + if (PREDICT_FALSE + (next_header >= + (u8 *) vlib_buffer_get_current (p0) + p0->current_length)) + { + //A malicious packet could set an extension header with a too big size + return (-1); + } + if (done) + break; + if ((!ip6_ext_hdr (next_proto)) || next_proto == IP_PROTOCOL_IP6_NONXT) + { + if (find_hdr_type < 0) + break; + return -1; + } + if (next_proto == IP_PROTOCOL_IPV6_FRAGMENTATION) + { + ip6_frag_hdr_t *frag_hdr = (ip6_frag_hdr_t *) next_header; + u16 frag_off = ip6_frag_hdr_offset (frag_hdr); + /* Non first fragment return -1 */ + if (frag_off) + return (-1); + exthdr_len = sizeof (ip6_frag_hdr_t); + temp_nxthdr = next_header + exthdr_len; + } + else if (next_proto == IP_PROTOCOL_IPSEC_AH) + { + exthdr_len = + ip6_ext_authhdr_len (((ip6_ext_header_t *) next_header)); + temp_nxthdr = next_header + exthdr_len; + } + else + { + exthdr_len = + ip6_ext_header_len (((ip6_ext_header_t *) next_header)); + temp_nxthdr = next_header + exthdr_len; + } + next_proto = ((ip6_ext_header_t *) next_header)->next_hdr; + next_header = temp_nxthdr; + cur_offset += exthdr_len; + } + + *offset = cur_offset; + return (next_proto); +} + +u8 *format_ip6_hop_by_hop_ext_hdr (u8 * s, va_list * args); +/* + * Hop-by-Hop handling + */ +typedef struct +{ + /* Array of function pointers to HBH option handling routines */ + int (*options[256]) (vlib_buffer_t * b, ip6_header_t * ip, + ip6_hop_by_hop_option_t * opt); + u8 *(*trace[256]) (u8 * s, ip6_hop_by_hop_option_t * opt); + uword next_override; +} ip6_hop_by_hop_main_t; + +extern ip6_hop_by_hop_main_t ip6_hop_by_hop_main; + +int ip6_hbh_register_option (u8 option, + int options (vlib_buffer_t * b, + ip6_header_t * ip, + ip6_hop_by_hop_option_t * opt), + u8 * trace (u8 * s, + ip6_hop_by_hop_option_t * opt)); +int ip6_hbh_unregister_option (u8 option); +void ip6_hbh_set_next_override (uword next); + +/** + * Push IPv6 header to buffer + * + * @param vm - vlib_main + * @param b - buffer to write the header to + * @param src - source IP + * @param dst - destination IP + * @param prot - payload proto + * + * @return - pointer to start of IP header + */ +always_inline void * +vlib_buffer_push_ip6 (vlib_main_t * vm, vlib_buffer_t * b, + ip6_address_t * src, ip6_address_t * dst, int proto) +{ + ip6_header_t *ip6h; + u16 payload_length; + + /* make some room */ + ip6h = vlib_buffer_push_uninit (b, sizeof (ip6_header_t)); + + ip6h->ip_version_traffic_class_and_flow_label = + clib_host_to_net_u32 (0x6 << 28); + + /* calculate ip6 payload length */ + payload_length = vlib_buffer_length_in_chain (vm, b); + payload_length -= sizeof (*ip6h); + + ip6h->payload_length = clib_host_to_net_u16 (payload_length); + + ip6h->hop_limit = 0xff; + ip6h->protocol = proto; + clib_memcpy (ip6h->src_address.as_u8, src->as_u8, + sizeof (ip6h->src_address)); + clib_memcpy (ip6h->dst_address.as_u8, dst->as_u8, + sizeof (ip6h->src_address)); + b->flags |= VNET_BUFFER_F_IS_IP6; + + return ip6h; +} + +#endif /* included_ip_ip6_h */ + +/* + * fd.io coding-style-patch-verification: ON + * + * Local Variables: + * eval: (c-set-style "gnu") + * End: + */ |