diff options
Diffstat (limited to 'vnet/vnet/vxlan-gpe')
| -rw-r--r-- | vnet/vnet/vxlan-gpe/decap.c | 450 | ||||
| -rw-r--r-- | vnet/vnet/vxlan-gpe/encap.c | 425 | ||||
| -rw-r--r-- | vnet/vnet/vxlan-gpe/vxlan-gpe-rfc.txt | 868 | ||||
| -rw-r--r-- | vnet/vnet/vxlan-gpe/vxlan_gpe.c | 467 | ||||
| -rw-r--r-- | vnet/vnet/vxlan-gpe/vxlan_gpe.h | 139 | ||||
| -rw-r--r-- | vnet/vnet/vxlan-gpe/vxlan_gpe_error.def | 16 | ||||
| -rw-r--r-- | vnet/vnet/vxlan-gpe/vxlan_gpe_packet.h | 82 |
7 files changed, 2447 insertions, 0 deletions
diff --git a/vnet/vnet/vxlan-gpe/decap.c b/vnet/vnet/vxlan-gpe/decap.c new file mode 100644 index 00000000000..aed5857d0b1 --- /dev/null +++ b/vnet/vnet/vxlan-gpe/decap.c @@ -0,0 +1,450 @@ +/* + * decap.c - decapsulate VXLAN GPE + * + * Copyright (c) 2013 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 <vlib/vlib.h> +#include <vnet/pg/pg.h> +#include <vnet/vxlan-gpe/vxlan_gpe.h> + +typedef struct { + u32 next_index; + u32 tunnel_index; + u32 error; +} vxlan_gpe_rx_trace_t; + +static u8 * format_vxlan_gpe_rx_trace (u8 * s, va_list * args) +{ + CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *); + CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *); + vxlan_gpe_rx_trace_t * t = va_arg (*args, vxlan_gpe_rx_trace_t *); + + if (t->tunnel_index != ~0) + { + s = format (s, "VXLAN-GPE: tunnel %d next %d error %d", t->tunnel_index, + t->next_index, t->error); + } + else + { + s = format (s, "VXLAN-GPE: no tunnel next %d error %d\n", t->next_index, + t->error); + } + return s; +} + + +static u8 * format_vxlan_gpe_with_length (u8 * s, va_list * args) +{ + CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *); + CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *); + + + return s; +} + +static uword +vxlan_gpe_input (vlib_main_t * vm, + vlib_node_runtime_t * node, + vlib_frame_t * from_frame) +{ + u32 n_left_from, next_index, * from, * to_next; + vxlan_gpe_main_t * ngm = &vxlan_gpe_main; + vnet_main_t * vnm = ngm->vnet_main; + vnet_interface_main_t * im = &vnm->interface_main; + u32 last_tunnel_index = ~0; + vxlan_gpe_tunnel_key_t last_key; + u32 pkts_decapsulated = 0; + u32 cpu_index = os_get_cpu_number(); + u32 stats_sw_if_index, stats_n_packets, stats_n_bytes; + + memset (&last_key, 0xff, sizeof (last_key)); + + from = vlib_frame_vector_args (from_frame); + n_left_from = from_frame->n_vectors; + + next_index = node->cached_next_index; + stats_sw_if_index = node->runtime_data[0]; + stats_n_packets = stats_n_bytes = 0; + + while (n_left_from > 0) + { + u32 n_left_to_next; + + vlib_get_next_frame (vm, node, next_index, + to_next, n_left_to_next); + + while (n_left_from >= 4 && n_left_to_next >= 2) + { + u32 bi0, bi1; + vlib_buffer_t * b0, * b1; + u32 next0, next1; + ip4_vxlan_gpe_header_t * iuvn0, * iuvn1; + uword * p0, * p1; + u32 tunnel_index0, tunnel_index1; + vxlan_gpe_tunnel_t * t0, * t1; + vxlan_gpe_tunnel_key_t key0, key1; + u32 error0, error1; + u32 sw_if_index0, sw_if_index1, len0, len1; + + /* Prefetch next iteration. */ + { + vlib_buffer_t * p2, * p3; + + p2 = vlib_get_buffer (vm, from[2]); + p3 = vlib_get_buffer (vm, from[3]); + + vlib_prefetch_buffer_header (p2, LOAD); + vlib_prefetch_buffer_header (p3, LOAD); + + CLIB_PREFETCH (p2->data, 2*CLIB_CACHE_LINE_BYTES, LOAD); + CLIB_PREFETCH (p3->data, 2*CLIB_CACHE_LINE_BYTES, LOAD); + } + + bi0 = from[0]; + bi1 = from[1]; + to_next[0] = bi0; + to_next[1] = bi1; + from += 2; + to_next += 2; + n_left_to_next -= 2; + n_left_from -= 2; + + b0 = vlib_get_buffer (vm, bi0); + b1 = vlib_get_buffer (vm, bi1); + + /* udp leaves current_data pointing at the vxlan header */ + vlib_buffer_advance + (b0, -(word)(sizeof(udp_header_t)+sizeof(ip4_header_t))); + vlib_buffer_advance + (b1, -(word)(sizeof(udp_header_t)+sizeof(ip4_header_t))); + + iuvn0 = vlib_buffer_get_current (b0); + iuvn1 = vlib_buffer_get_current (b1); + + /* pop (ip, udp, vxlan) */ + vlib_buffer_advance (b0, sizeof (*iuvn0)); + vlib_buffer_advance (b1, sizeof (*iuvn1)); + + tunnel_index0 = ~0; + tunnel_index1 = ~0; + error0 = 0; + error1 = 0; + + next0 = (iuvn0->vxlan.protocol < VXLAN_GPE_INPUT_N_NEXT) ? iuvn0->vxlan.protocol : VXLAN_GPE_INPUT_NEXT_DROP; + next1 = (iuvn1->vxlan.protocol < VXLAN_GPE_INPUT_N_NEXT) ? iuvn1->vxlan.protocol : VXLAN_GPE_INPUT_NEXT_DROP; + + + + + key0.local = iuvn0->ip4.dst_address.as_u32; + key1.local = iuvn1->ip4.dst_address.as_u32; + + key0.remote = iuvn0->ip4.src_address.as_u32; + key1.remote = iuvn1->ip4.src_address.as_u32; + + key0.vni = iuvn0->vxlan.vni_res; + key1.vni = iuvn1->vxlan.vni_res; + + key0.pad = 0; + key1.pad = 0; + + /* Processing for key0 */ + if (PREDICT_FALSE ((key0.as_u64[0] != last_key.as_u64[0]) + || (key0.as_u64[1] != last_key.as_u64[1]))) + { + p0 = hash_get_mem (ngm->vxlan_gpe_tunnel_by_key, &key0); + + if (p0 == 0) + { + error0 = VXLAN_GPE_ERROR_NO_SUCH_TUNNEL; + goto trace0; + } + + last_key.as_u64[0] = key0.as_u64[0]; + last_key.as_u64[1] = key0.as_u64[1]; + tunnel_index0 = last_tunnel_index = p0[0]; + } + else + tunnel_index0 = last_tunnel_index; + + t0 = pool_elt_at_index (ngm->tunnels, tunnel_index0); + + next0 = t0->protocol; + + sw_if_index0 = t0->sw_if_index; + len0 = vlib_buffer_length_in_chain(vm, b0); + + /* Required to make the l2 tag push / pop code work on l2 subifs */ + vnet_update_l2_len (b0); + + /* + * ip[46] lookup in the configured FIB + */ + vnet_buffer(b0)->sw_if_index[VLIB_TX] = t0->decap_fib_index; + + pkts_decapsulated++; + stats_n_packets += 1; + stats_n_bytes += len0; + + if (PREDICT_FALSE(sw_if_index0 != stats_sw_if_index)) + { + stats_n_packets -= 1; + stats_n_bytes -= len0; + if (stats_n_packets) + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_RX, + cpu_index, stats_sw_if_index, stats_n_packets, stats_n_bytes); + stats_n_packets = 1; + stats_n_bytes = len0; + stats_sw_if_index = sw_if_index0; + } + + trace0: + b0->error = error0 ? node->errors[error0] : 0; + + if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED)) + { + vxlan_gpe_rx_trace_t *tr + = vlib_add_trace (vm, node, b0, sizeof (*tr)); + tr->next_index = next0; + tr->error = error0; + tr->tunnel_index = tunnel_index0; + } + + + /* Processing for key1 */ + if (PREDICT_FALSE ((key1.as_u64[0] != last_key.as_u64[0]) + || (key1.as_u64[1] != last_key.as_u64[1]))) + { + p1 = hash_get_mem (ngm->vxlan_gpe_tunnel_by_key, &key1); + + if (p1 == 0) + { + error1 = VXLAN_GPE_ERROR_NO_SUCH_TUNNEL; + goto trace1; + } + + last_key.as_u64[0] = key1.as_u64[0]; + last_key.as_u64[1] = key1.as_u64[1]; + tunnel_index1 = last_tunnel_index = p1[0]; + } + else + tunnel_index1 = last_tunnel_index; + + t1 = pool_elt_at_index (ngm->tunnels, tunnel_index1); + + next1 = t1->protocol; + sw_if_index1 = t1->sw_if_index; + len1 = vlib_buffer_length_in_chain(vm, b1); + + /* Required to make the l2 tag push / pop code work on l2 subifs */ + vnet_update_l2_len (b1); + + /* + * ip[46] lookup in the configured FIB + */ + vnet_buffer(b1)->sw_if_index[VLIB_TX] = t1->decap_fib_index; + + + pkts_decapsulated++; + stats_n_packets += 1; + stats_n_bytes += len1; + + /* Batch stats increment on the same vxlan tunnel so counter + is not incremented per packet */ + if (PREDICT_FALSE(sw_if_index1 != stats_sw_if_index)) + { + stats_n_packets -= 1; + stats_n_bytes -= len1; + if (stats_n_packets) + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_RX, + cpu_index, stats_sw_if_index, stats_n_packets, stats_n_bytes); + stats_n_packets = 1; + stats_n_bytes = len1; + stats_sw_if_index = sw_if_index1; + } + vnet_buffer(b1)->sw_if_index[VLIB_TX] = t1->decap_fib_index; + + trace1: + b1->error = error1 ? node->errors[error1] : 0; + + if (PREDICT_FALSE(b1->flags & VLIB_BUFFER_IS_TRACED)) + { + vxlan_gpe_rx_trace_t *tr + = vlib_add_trace (vm, node, b1, sizeof (*tr)); + tr->next_index = next1; + tr->error = error1; + tr->tunnel_index = tunnel_index1; + } + + vlib_validate_buffer_enqueue_x2 (vm, node, next_index, + to_next, n_left_to_next, + bi0, bi1, next0, next1); + } + + while (n_left_from > 0 && n_left_to_next > 0) + { + u32 bi0; + vlib_buffer_t * b0; + u32 next0; + ip4_vxlan_gpe_header_t * iuvn0; + uword * p0; + u32 tunnel_index0; + vxlan_gpe_tunnel_t * t0; + vxlan_gpe_tunnel_key_t key0; + u32 error0; + u32 sw_if_index0, len0; + + bi0 = from[0]; + to_next[0] = bi0; + from += 1; + to_next += 1; + n_left_from -= 1; + n_left_to_next -= 1; + + b0 = vlib_get_buffer (vm, bi0); + + /* udp leaves current_data pointing at the vxlan header */ + vlib_buffer_advance + (b0, -(word)(sizeof(udp_header_t)+sizeof(ip4_header_t))); + + iuvn0 = vlib_buffer_get_current (b0); + + /* pop (ip, udp, vxlan) */ + vlib_buffer_advance (b0, sizeof (*iuvn0)); + + tunnel_index0 = ~0; + error0 = 0; + next0 = (iuvn0->vxlan.protocol < VXLAN_GPE_INPUT_N_NEXT) ? iuvn0->vxlan.protocol : VXLAN_GPE_INPUT_NEXT_DROP; + + key0.local = iuvn0->ip4.dst_address.as_u32; + key0.remote = iuvn0->ip4.src_address.as_u32; + key0.vni = iuvn0->vxlan.vni_res; + key0.pad = 0; + + if (PREDICT_FALSE ((key0.as_u64[0] != last_key.as_u64[0]) + || (key0.as_u64[1] != last_key.as_u64[1]))) + { + p0 = hash_get_mem (ngm->vxlan_gpe_tunnel_by_key, &key0); + + if (p0 == 0) + { + error0 = VXLAN_GPE_ERROR_NO_SUCH_TUNNEL; + goto trace00; + } + + last_key.as_u64[0] = key0.as_u64[0]; + last_key.as_u64[1] = key0.as_u64[1]; + tunnel_index0 = last_tunnel_index = p0[0]; + } + else + tunnel_index0 = last_tunnel_index; + + t0 = pool_elt_at_index (ngm->tunnels, tunnel_index0); + + next0 = t0->protocol; + + sw_if_index0 = t0->sw_if_index; + len0 = vlib_buffer_length_in_chain(vm, b0); + + /* Required to make the l2 tag push / pop code work on l2 subifs */ + vnet_update_l2_len (b0); + + /* + * ip[46] lookup in the configured FIB + */ + vnet_buffer(b0)->sw_if_index[VLIB_TX] = t0->decap_fib_index; + + pkts_decapsulated ++; + stats_n_packets += 1; + stats_n_bytes += len0; + + /* Batch stats increment on the same vxlan-gpe tunnel so counter + is not incremented per packet */ + if (PREDICT_FALSE(sw_if_index0 != stats_sw_if_index)) + { + stats_n_packets -= 1; + stats_n_bytes -= len0; + if (stats_n_packets) + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_RX, + cpu_index, stats_sw_if_index, stats_n_packets, stats_n_bytes); + stats_n_packets = 1; + stats_n_bytes = len0; + stats_sw_if_index = sw_if_index0; + } + + trace00: + b0->error = error0 ? node->errors[error0] : 0; + + if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED)) + { + vxlan_gpe_rx_trace_t *tr + = vlib_add_trace (vm, node, b0, sizeof (*tr)); + tr->next_index = next0; + tr->error = error0; + tr->tunnel_index = tunnel_index0; + } + vlib_validate_buffer_enqueue_x1 (vm, node, next_index, + to_next, n_left_to_next, + bi0, next0); + } + + vlib_put_next_frame (vm, node, next_index, n_left_to_next); + } + vlib_node_increment_counter (vm, vxlan_gpe_input_node.index, + VXLAN_GPE_ERROR_DECAPSULATED, + pkts_decapsulated); + /* Increment any remaining batch stats */ + if (stats_n_packets) + { + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_RX, cpu_index, + stats_sw_if_index, stats_n_packets, stats_n_bytes); + node->runtime_data[0] = stats_sw_if_index; + } + return from_frame->n_vectors; +} + +static char * vxlan_gpe_error_strings[] = { +#define vxlan_gpe_error(n,s) s, +#include <vnet/vxlan-gpe/vxlan_gpe_error.def> +#undef vxlan_gpe_error +#undef _ +}; + +VLIB_REGISTER_NODE (vxlan_gpe_input_node) = { + .function = vxlan_gpe_input, + .name = "vxlan-gpe-input", + /* Takes a vector of packets. */ + .vector_size = sizeof (u32), + .type = VLIB_NODE_TYPE_INTERNAL, + .n_errors = ARRAY_LEN(vxlan_gpe_error_strings), + .error_strings = vxlan_gpe_error_strings, + + .n_next_nodes = VXLAN_GPE_INPUT_N_NEXT, + .next_nodes = { +#define _(s,n) [VXLAN_GPE_INPUT_NEXT_##s] = n, + foreach_vxlan_gpe_input_next +#undef _ + }, + + .format_buffer = format_vxlan_gpe_with_length, + .format_trace = format_vxlan_gpe_rx_trace, + // $$$$ .unformat_buffer = unformat_vxlan_gpe_header, +}; + + diff --git a/vnet/vnet/vxlan-gpe/encap.c b/vnet/vnet/vxlan-gpe/encap.c new file mode 100644 index 00000000000..3ffe2a62db2 --- /dev/null +++ b/vnet/vnet/vxlan-gpe/encap.c @@ -0,0 +1,425 @@ +/* + * 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 <vppinfra/error.h> +#include <vppinfra/hash.h> +#include <vnet/vnet.h> +#include <vnet/ip/ip.h> +#include <vnet/ethernet/ethernet.h> +#include <vnet/vxlan-gpe/vxlan_gpe.h> + +/* Statistics (not really errors) */ +#define foreach_vxlan_gpe_encap_error \ +_(ENCAPSULATED, "good packets encapsulated") + +static char * vxlan_gpe_encap_error_strings[] = { +#define _(sym,string) string, + foreach_vxlan_gpe_encap_error +#undef _ +}; + +typedef enum { +#define _(sym,str) VXLAN_GPE_ENCAP_ERROR_##sym, + foreach_vxlan_gpe_encap_error +#undef _ + VXLAN_GPE_ENCAP_N_ERROR, +} vxlan_gpe_encap_error_t; + +typedef enum { + VXLAN_GPE_ENCAP_NEXT_IP4_LOOKUP, + VXLAN_GPE_ENCAP_NEXT_DROP, + VXLAN_GPE_ENCAP_N_NEXT +} vxlan_gpe_encap_next_t; + +typedef struct { + u32 tunnel_index; +} vxlan_gpe_encap_trace_t; + + +u8 * format_vxlan_gpe_encap_trace (u8 * s, va_list * args) +{ + CLIB_UNUSED (vlib_main_t * vm) = va_arg (*args, vlib_main_t *); + CLIB_UNUSED (vlib_node_t * node) = va_arg (*args, vlib_node_t *); + vxlan_gpe_encap_trace_t * t + = va_arg (*args, vxlan_gpe_encap_trace_t *); + + s = format (s, "VXLAN-GPE-ENCAP: tunnel %d", t->tunnel_index); + return s; +} + +#define foreach_fixed_header_offset \ +_(0) _(1) _(2) _(3) _(4) _(5) _(6) + +static uword +vxlan_gpe_encap (vlib_main_t * vm, + vlib_node_runtime_t * node, + vlib_frame_t * from_frame) +{ + u32 n_left_from, next_index, * from, * to_next; + vxlan_gpe_main_t * ngm = &vxlan_gpe_main; + vnet_main_t * vnm = ngm->vnet_main; + vnet_interface_main_t * im = &vnm->interface_main; + u32 pkts_encapsulated = 0; + u16 old_l0 = 0, old_l1 = 0; + u32 cpu_index = os_get_cpu_number(); + u32 stats_sw_if_index, stats_n_packets, stats_n_bytes; + + from = vlib_frame_vector_args (from_frame); + n_left_from = from_frame->n_vectors; + + next_index = node->cached_next_index; + stats_sw_if_index = node->runtime_data[0]; + stats_n_packets = stats_n_bytes = 0; + + while (n_left_from > 0) + { + u32 n_left_to_next; + + vlib_get_next_frame (vm, node, next_index, + to_next, n_left_to_next); + + while (n_left_from >= 4 && n_left_to_next >= 2) + { + u32 bi0, bi1; + vlib_buffer_t * b0, * b1; + u32 next0 = VXLAN_GPE_ENCAP_NEXT_IP4_LOOKUP; + u32 next1 = VXLAN_GPE_ENCAP_NEXT_IP4_LOOKUP; + u32 sw_if_index0, sw_if_index1, len0, len1; + vnet_hw_interface_t * hi0, * hi1; + ip4_header_t * ip0, * ip1; + udp_header_t * udp0, * udp1; + u64 * copy_src0, * copy_dst0; + u64 * copy_src1, * copy_dst1; + u32 * copy_src_last0, * copy_dst_last0; + u32 * copy_src_last1, * copy_dst_last1; + vxlan_gpe_tunnel_t * t0, * t1; + u16 new_l0, new_l1; + ip_csum_t sum0, sum1; + + /* Prefetch next iteration. */ + { + vlib_buffer_t * p2, * p3; + + p2 = vlib_get_buffer (vm, from[2]); + p3 = vlib_get_buffer (vm, from[3]); + + vlib_prefetch_buffer_header (p2, LOAD); + vlib_prefetch_buffer_header (p3, LOAD); + + CLIB_PREFETCH (p2->data, 2*CLIB_CACHE_LINE_BYTES, LOAD); + CLIB_PREFETCH (p3->data, 2*CLIB_CACHE_LINE_BYTES, LOAD); + } + + bi0 = from[0]; + bi1 = from[1]; + to_next[0] = bi0; + to_next[1] = bi1; + from += 2; + to_next += 2; + n_left_to_next -= 2; + n_left_from -= 2; + + b0 = vlib_get_buffer (vm, bi0); + b1 = vlib_get_buffer (vm, bi1); + + /* 1-wide cache? */ + sw_if_index0 = vnet_buffer(b0)->sw_if_index[VLIB_TX]; + sw_if_index1 = vnet_buffer(b1)->sw_if_index[VLIB_TX]; + hi0 = vnet_get_sup_hw_interface + (vnm, vnet_buffer(b0)->sw_if_index[VLIB_TX]); + hi1 = vnet_get_sup_hw_interface + (vnm, vnet_buffer(b1)->sw_if_index[VLIB_TX]); + + t0 = pool_elt_at_index (ngm->tunnels, hi0->dev_instance); + t1 = pool_elt_at_index (ngm->tunnels, hi1->dev_instance); + + ASSERT(vec_len(t0->rewrite) >= 24); + ASSERT(vec_len(t1->rewrite) >= 24); + + /* Apply the rewrite string. $$$$ vnet_rewrite? */ + vlib_buffer_advance (b0, -(word)_vec_len(t0->rewrite)); + vlib_buffer_advance (b1, -(word)_vec_len(t1->rewrite)); + + ip0 = vlib_buffer_get_current(b0); + ip1 = vlib_buffer_get_current(b1); + /* Copy the fixed header */ + copy_dst0 = (u64 *) ip0; + copy_src0 = (u64 *) t0->rewrite; + copy_dst1 = (u64 *) ip1; + copy_src1 = (u64 *) t1->rewrite; + + ASSERT (sizeof (ip4_vxlan_gpe_header_t) == 36); + + /* Copy first 36 octets 8-bytes at a time */ +#define _(offs) copy_dst0[offs] = copy_src0[offs]; + foreach_fixed_header_offset; +#undef _ +#define _(offs) copy_dst1[offs] = copy_src1[offs]; + foreach_fixed_header_offset; +#undef _ + + /* Last 4 octets. Hopefully gcc will be our friend */ + copy_dst_last0 = (u32 *)(©_dst0[7]); + copy_src_last0 = (u32 *)(©_src0[7]); + copy_dst_last1 = (u32 *)(©_dst1[7]); + copy_src_last1 = (u32 *)(©_src1[7]); + + copy_dst_last0[0] = copy_src_last0[0]; + copy_dst_last1[0] = copy_src_last1[0]; + + /* If there are TLVs to copy, do so */ + if (PREDICT_FALSE (_vec_len(t0->rewrite) > 64)) + clib_memcpy (©_dst0[3], t0->rewrite + 64 , + _vec_len (t0->rewrite)-64); + + if (PREDICT_FALSE (_vec_len(t1->rewrite) > 64)) + clib_memcpy (©_dst0[3], t1->rewrite + 64 , + _vec_len (t1->rewrite)-64); + + /* fix the <bleep>ing outer-IP checksum */ + sum0 = ip0->checksum; + /* old_l0 always 0, see the rewrite setup */ + new_l0 = + clib_host_to_net_u16 (vlib_buffer_length_in_chain (vm, b0)); + + sum0 = ip_csum_update (sum0, old_l0, new_l0, ip4_header_t, + length /* changed member */); + ip0->checksum = ip_csum_fold (sum0); + ip0->length = new_l0; + + sum1 = ip1->checksum; + /* old_l1 always 0, see the rewrite setup */ + new_l1 = + clib_host_to_net_u16 (vlib_buffer_length_in_chain (vm, b1)); + + sum1 = ip_csum_update (sum1, old_l1, new_l1, ip4_header_t, + length /* changed member */); + ip1->checksum = ip_csum_fold (sum1); + ip1->length = new_l1; + + /* Fix UDP length */ + udp0 = (udp_header_t *)(ip0+1); + new_l0 = clib_host_to_net_u16 (vlib_buffer_length_in_chain (vm, b0) + - sizeof (*ip0)); + udp1 = (udp_header_t *)(ip1+1); + new_l1 = clib_host_to_net_u16 (vlib_buffer_length_in_chain (vm, b1) + - sizeof (*ip1)); + + udp0->length = new_l0; + udp1->length = new_l1; + + /* Reset to look up tunnel partner in the configured FIB */ + vnet_buffer(b0)->sw_if_index[VLIB_TX] = t0->encap_fib_index; + vnet_buffer(b1)->sw_if_index[VLIB_TX] = t1->encap_fib_index; + vnet_buffer(b0)->sw_if_index[VLIB_RX] = sw_if_index0; + vnet_buffer(b1)->sw_if_index[VLIB_RX] = sw_if_index1; + pkts_encapsulated += 2; + + len0 = vlib_buffer_length_in_chain(vm, b0); + len1 = vlib_buffer_length_in_chain(vm, b0); + stats_n_packets += 2; + stats_n_bytes += len0 + len1; + + /* Batch stats increment on the same vxlan tunnel so counter is not + incremented per packet. Note stats are still incremented for deleted + and admin-down tunnel where packets are dropped. It is not worthwhile + to check for this rare case and affect normal path performance. */ + if (PREDICT_FALSE( + (sw_if_index0 != stats_sw_if_index) + || (sw_if_index1 != stats_sw_if_index))) { + stats_n_packets -= 2; + stats_n_bytes -= len0 + len1; + if (sw_if_index0 == sw_if_index1) { + if (stats_n_packets) + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_TX, + cpu_index, stats_sw_if_index, stats_n_packets, stats_n_bytes); + stats_sw_if_index = sw_if_index0; + stats_n_packets = 2; + stats_n_bytes = len0 + len1; + } else { + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_TX, + cpu_index, sw_if_index0, 1, len0); + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_TX, + cpu_index, sw_if_index1, 1, len1); + } + } + + if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED)) + { + vxlan_gpe_encap_trace_t *tr = + vlib_add_trace (vm, node, b0, sizeof (*tr)); + tr->tunnel_index = t0 - ngm->tunnels; + } + + if (PREDICT_FALSE(b1->flags & VLIB_BUFFER_IS_TRACED)) + { + vxlan_gpe_encap_trace_t *tr = + vlib_add_trace (vm, node, b1, sizeof (*tr)); + tr->tunnel_index = t1 - ngm->tunnels; + } + + vlib_validate_buffer_enqueue_x2 (vm, node, next_index, + to_next, n_left_to_next, + bi0, bi1, next0, next1); + } + + while (n_left_from > 0 && n_left_to_next > 0) + { + u32 bi0; + vlib_buffer_t * b0; + u32 next0 = VXLAN_GPE_ENCAP_NEXT_IP4_LOOKUP; + u32 sw_if_index0, len0; + vnet_hw_interface_t * hi0; + ip4_header_t * ip0; + udp_header_t * udp0; + u64 * copy_src0, * copy_dst0; + u32 * copy_src_last0, * copy_dst_last0; + vxlan_gpe_tunnel_t * t0; + u16 new_l0; + ip_csum_t sum0; + + bi0 = from[0]; + to_next[0] = bi0; + from += 1; + to_next += 1; + n_left_from -= 1; + n_left_to_next -= 1; + + b0 = vlib_get_buffer (vm, bi0); + + /* 1-wide cache? */ + sw_if_index0 = vnet_buffer(b0)->sw_if_index[VLIB_TX]; + hi0 = vnet_get_sup_hw_interface + (vnm, vnet_buffer(b0)->sw_if_index[VLIB_TX]); + + t0 = pool_elt_at_index (ngm->tunnels, hi0->dev_instance); + + ASSERT(vec_len(t0->rewrite) >= 24); + + /* Apply the rewrite string. $$$$ vnet_rewrite? */ + vlib_buffer_advance (b0, -(word)_vec_len(t0->rewrite)); + + ip0 = vlib_buffer_get_current(b0); + /* Copy the fixed header */ + copy_dst0 = (u64 *) ip0; + copy_src0 = (u64 *) t0->rewrite; + + ASSERT (sizeof (ip4_vxlan_gpe_header_t) == 36); + + /* Copy first 36 octets 8-bytes at a time */ +#define _(offs) copy_dst0[offs] = copy_src0[offs]; + foreach_fixed_header_offset; +#undef _ + /* Last 4 octets. Hopefully gcc will be our friend */ + copy_dst_last0 = (u32 *)(©_dst0[7]); + copy_src_last0 = (u32 *)(©_src0[7]); + + copy_dst_last0[0] = copy_src_last0[0]; + + /* If there are TLVs to copy, do so */ + if (PREDICT_FALSE (_vec_len(t0->rewrite) > 64)) + clib_memcpy (©_dst0[3], t0->rewrite + 64 , + _vec_len (t0->rewrite)-64); + + /* fix the <bleep>ing outer-IP checksum */ + sum0 = ip0->checksum; + /* old_l0 always 0, see the rewrite setup */ + new_l0 = + clib_host_to_net_u16 (vlib_buffer_length_in_chain (vm, b0)); + + sum0 = ip_csum_update (sum0, old_l0, new_l0, ip4_header_t, + length /* changed member */); + ip0->checksum = ip_csum_fold (sum0); + ip0->length = new_l0; + + /* Fix UDP length */ + udp0 = (udp_header_t *)(ip0+1); + new_l0 = clib_host_to_net_u16 (vlib_buffer_length_in_chain (vm, b0) + - sizeof (*ip0)); + + udp0->length = new_l0; + + /* Reset to look up tunnel partner in the configured FIB */ + vnet_buffer(b0)->sw_if_index[VLIB_TX] = t0->encap_fib_index; + vnet_buffer(b0)->sw_if_index[VLIB_RX] = sw_if_index0; + pkts_encapsulated ++; + + len0 = vlib_buffer_length_in_chain(vm, b0); + stats_n_packets += 1; + stats_n_bytes += len0; + + /* Batch stats increment on the same vxlan tunnel so counter is not + * incremented per packet. Note stats are still incremented for deleted + * and admin-down tunnel where packets are dropped. It is not worthwhile + * to check for this rare case and affect normal path performance. */ + if (PREDICT_FALSE(sw_if_index0 != stats_sw_if_index)) + { + stats_n_packets -= 1; + stats_n_bytes -= len0; + if (stats_n_packets) + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_TX, + cpu_index, stats_sw_if_index, stats_n_packets, stats_n_bytes); + stats_n_packets = 1; + stats_n_bytes = len0; + stats_sw_if_index = sw_if_index0; + } + if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED)) + { + vxlan_gpe_encap_trace_t *tr = + vlib_add_trace (vm, node, b0, sizeof (*tr)); + tr->tunnel_index = t0 - ngm->tunnels; + } + vlib_validate_buffer_enqueue_x1 (vm, node, next_index, + to_next, n_left_to_next, + bi0, next0); + } + + vlib_put_next_frame (vm, node, next_index, n_left_to_next); + } + vlib_node_increment_counter (vm, node->node_index, + VXLAN_GPE_ENCAP_ERROR_ENCAPSULATED, + pkts_encapsulated); + /* Increment any remaining batch stats */ + if (stats_n_packets) { + vlib_increment_combined_counter( + im->combined_sw_if_counters + VNET_INTERFACE_COUNTER_TX, cpu_index, + stats_sw_if_index, stats_n_packets, stats_n_bytes); + node->runtime_data[0] = stats_sw_if_index; + } + + return from_frame->n_vectors; +} + +VLIB_REGISTER_NODE (vxlan_gpe_encap_node) = { + .function = vxlan_gpe_encap, + .name = "vxlan-gpe-encap", + .vector_size = sizeof (u32), + .format_trace = format_vxlan_gpe_encap_trace, + .type = VLIB_NODE_TYPE_INTERNAL, + + .n_errors = ARRAY_LEN(vxlan_gpe_encap_error_strings), + .error_strings = vxlan_gpe_encap_error_strings, + + .n_next_nodes = VXLAN_GPE_ENCAP_N_NEXT, + + .next_nodes = { + [VXLAN_GPE_ENCAP_NEXT_IP4_LOOKUP] = "ip4-lookup", + [VXLAN_GPE_ENCAP_NEXT_DROP] = "error-drop", + }, +}; + diff --git a/vnet/vnet/vxlan-gpe/vxlan-gpe-rfc.txt b/vnet/vnet/vxlan-gpe/vxlan-gpe-rfc.txt new file mode 100644 index 00000000000..35cee50f573 --- /dev/null +++ b/vnet/vnet/vxlan-gpe/vxlan-gpe-rfc.txt @@ -0,0 +1,868 @@ +Network Working Group P. Quinn +Internet-Draft Cisco Systems, Inc. +Intended status: Experimental P. Agarwal +Expires: January 4, 2015 Broadcom + R. Fernando + L. Kreeger + D. Lewis + F. Maino + M. Smith + N. Yadav + Cisco Systems, Inc. + L. Yong + Huawei USA + X. Xu + Huawei Technologies + U. Elzur + Intel + P. Garg + Microsoft + July 3, 2014 + + + Generic Protocol Extension for VXLAN + draft-quinn-vxlan-gpe-03.txt + +Abstract + + This draft describes extending Virtual eXtensible Local Area Network + (VXLAN), via changes to the VXLAN header, with three new + capabilities: support for multi-protocol encapsulation, operations, + administration and management (OAM) signaling and explicit + versioning. + +Status of this Memo + + This Internet-Draft is submitted in full conformance with the + provisions of BCP 78 and BCP 79. + + Internet-Drafts are working documents of the Internet Engineering + Task Force (IETF). Note that other groups may also distribute + working documents as Internet-Drafts. The list of current Internet- + Drafts is at http://datatracker.ietf.org/drafts/current/. + + Internet-Drafts are draft documents valid for a maximum of six months + and may be updated, replaced, or obsoleted by other documents at any + time. It is inappropriate to use Internet-Drafts as reference + material or to cite them other than as "work in progress." + + + + +Quinn, et al. Expires January 4, 2015 [Page 1] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + + This Internet-Draft will expire on January 4, 2015. + +Copyright Notice + + Copyright (c) 2014 IETF Trust and the persons identified as the + document authors. All rights reserved. + + This document is subject to BCP 78 and the IETF Trust's Legal + Provisions Relating to IETF Documents + (http://trustee.ietf.org/license-info) in effect on the date of + publication of this document. Please review these documents + carefully, as they describe your rights and restrictions with respect + to this document. Code Components extracted from this document must + include Simplified BSD License text as described in Section 4.e of + the Trust Legal Provisions and are provided without warranty as + described in the Simplified BSD License. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 2] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +Table of Contents + + 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 + 2. VXLAN Without Protocol Extension . . . . . . . . . . . . . . . 5 + 3. Generic Protocol Extension VXLAN (VXLAN-gpe) . . . . . . . . . 6 + 3.1. Multi Protocol Support . . . . . . . . . . . . . . . . . . 6 + 3.2. OAM Support . . . . . . . . . . . . . . . . . . . . . . . 7 + 3.3. Version Bits . . . . . . . . . . . . . . . . . . . . . . . 7 + 4. Backward Compatibility . . . . . . . . . . . . . . . . . . . . 8 + 4.1. VXLAN VTEP to VXLAN-gpe VTEP . . . . . . . . . . . . . . . 8 + 4.2. VXLAN-gpe VTEP to VXLAN VTEP . . . . . . . . . . . . . . . 8 + 4.3. VXLAN-gpe UDP Ports . . . . . . . . . . . . . . . . . . . 8 + 4.4. VXLAN-gpe and Encapsulated IP Header Fields . . . . . . . 8 + 5. VXLAN-gpe Examples . . . . . . . . . . . . . . . . . . . . . . 9 + 6. Security Considerations . . . . . . . . . . . . . . . . . . . 11 + 7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 12 + 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 + 8.1. UDP Port . . . . . . . . . . . . . . . . . . . . . . . . . 13 + 8.2. VXLAN-gpe Next Protocol . . . . . . . . . . . . . . . . . 13 + 8.3. VXLAN-gpe Reserved Bits . . . . . . . . . . . . . . . . . 13 + 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 14 + 9.1. Normative References . . . . . . . . . . . . . . . . . . . 14 + 9.2. Informative References . . . . . . . . . . . . . . . . . . 14 + Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 15 + + + + + + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 3] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +1. Introduction + + Virtual eXtensible Local Area Network [VXLAN] defines an + encapsulation format that encapsulates Ethernet frames in an outer + UDP/IP transport. As data centers evolve, the need to carry other + protocols encapsulated in an IP packet is required, as well as the + need to provide increased visibility and diagnostic capabilities + within the overlay. The VXLAN header does not specify the protocol + being encapsulated and therefore is currently limited to + encapsulating only Ethernet frame payload, nor does it provide the + ability to define OAM protocols. Rather than defining yet another + encapsulation, VXLAN is extended to provide protocol typing and OAM + capabilities. + + This document describes extending VXLAN via the following changes: + + Next Protocol Bit (P bit): A reserved flag bit is allocated, and set + in the VXLAN-gpe header to indicate that a next protocol field is + present. + + OAM Flag Bit (O bit): A reserved flag bit is allocated, and set in + the VXLAN-gpe header, to indicate that the packet is an OAM + packet. + + Version: Two reserved bits are allocated, and set in the VXLAN-gpe + header, to indicate VXLAN-gpe protocol version. + + Next Protocol: A 8 bit next protocol field is present in the VXLAN- + gpe header. + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 4] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +2. VXLAN Without Protocol Extension + + As described in the introduction, the VXLAN header has no protocol + identifier that indicates the type of payload being carried by VXLAN. + Because of this, VXLAN is limited to an Ethernet payload. + Furthermore, the VXLAN header has no mechanism to signal OAM packets. + + The VXLAN header defines bits 0-7 as flags (some defined, some + reserved), the VXLAN network identifier (VNI) field and several + reserved bits. The flags provide flexibility to define how the + reserved bits can be used to change the definition of the VXLAN + header. + + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|R|R|R| Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + Figure 1: VXLAN Header + + + + + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 5] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +3. Generic Protocol Extension VXLAN (VXLAN-gpe) + +3.1. Multi Protocol Support + + This draft defines the following two changes to the VXLAN header in + order to support multi-protocol encapsulation: + + P Bit: Flag bit 5 is defined as the Next Protocol bit. The P bit + MUST be set to 1 to indicate the presence of the 8 bit next + protocol field. + + P = 0 indicates that the payload MUST conform to VXLAN as defined + in [VXLAN]. + + Flag bit 5 was chosen as the P bit because this flag bit is + currently reserved in VXLAN. + + Next Protocol Field: The lower 8 bits of the first word are used to + carry a next protocol. This next protocol field contains the + protocol of the encapsulated payload packet. A new protocol + registry will be requested from IANA. + + This draft defines the following Next Protocol values: + + 0x1 : IPv4 + 0x2 : IPv6 + 0x3 : Ethernet + 0x4 : Network Service Header [NSH] + + + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|P|R|R| Reserved |Next Protocol | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + + Figure 2: VXLAN-gpe Next Protocol + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 6] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +3.2. OAM Support + + Flag bit 7 is defined as the O bit. When the O bit is set to 1, the + packet is an OAM packet and OAM processing MUST occur. The OAM + protocol details are out of scope for this document. As with the + P-bit, bit 7 is currently a reserved flag in VXLAN. + + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|P|R|O| Reserved |Next Protocol | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + + Figure 3: VXLAN-gpe OAM Bit + +3.3. Version Bits + + VXLAN-gpe bits 8 and 9 are defined as version bits. These bits are + reserved in VXLAN. The version field is used to ensure backward + compatibility going forward with future VXLAN-gpe updates. + + The initial version for VXLAN-gpe is 0. + + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|P|R|O|Ver| Reserved |Next Protocol | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + + + + Figure 4: VXLAN-gpe Version Bits + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 7] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +4. Backward Compatibility + +4.1. VXLAN VTEP to VXLAN-gpe VTEP + + As per VXLAN, reserved bits 5 and 7, VXLAN-gpe P and O-bits + respectively must be set to zero. The remaining reserved bits must + be zero, including the VXLAN-gpe version field, bits 8 and 9. The + encapsulated payload MUST be Ethernet. + +4.2. VXLAN-gpe VTEP to VXLAN VTEP + + A VXLAN-gpe VTEP MUST NOT encapsulate non-Ethernet frames to a VXLAN + VTEP. When encapsulating Ethernet frames to a VXLAN VTEP, the VXLAN- + gpe VTEP will set the P bit to 0, the Next Protocol to 0 and use UDP + destination port 4789. A VXLAN-gpe VTEP MUST also set O = 0 and Ver + = 0 when encapsulating Ethernet frames to VXLAN VTEP. The receiving + VXLAN VTEP will threat this packet as a VXLAN packet. + + A method for determining the capabilities of a VXLAN VTEP (gpe or + non-gpe) is out of the scope of this draft. + +4.3. VXLAN-gpe UDP Ports + + VXLAN-gpe uses a new UDP destination port (to be assigned by IANA) + when sending traffic to VXLAN-gpe VTEPs. + +4.4. VXLAN-gpe and Encapsulated IP Header Fields + + When encapsulating and decapsulating IPv4 and IPv6 packets, certain + fields, such as IPv4 Time to Live (TTL) from the inner IP header need + to be considered. VXLAN-gpe IP encapsulation and decapsulation + utilizes the techniques described in [RFC6830], section 5.3. + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 8] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +5. VXLAN-gpe Examples + + This section provides three examples of protocols encapsulated using + the Generic Protocol Extension for VXLAN described in this document. + + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|1|R|0|0|0| Reserved | NP = IPv4 | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | Original IPv4 Packet | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + + Figure 5: IPv4 and VXLAN-gpe + + + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|1|R|0|0|0| Reserved | NP = IPv6 | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | Original IPv6 Packet | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + + Figure 6: IPv6 and VXLAN-gpe + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 9] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + + 0 1 2 3 + 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + |R|R|R|R|I|1|R|0|0|0| Reserved |NP = Ethernet | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | VXLAN Network Identifier (VNI) | Reserved | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | Original Ethernet Frame | + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + + + + Figure 7: Ethernet and VXLAN-gpe + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 10] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +6. Security Considerations + + VXLAN's security is focused on issues around L2 encapsulation into + L3. With VXLAN-gpe, issues such as spoofing, flooding, and traffic + redirection are dependent on the particular protocol payload + encapsulated. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 11] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +7. Acknowledgments + + A special thank you goes to Dino Farinacci for his guidance and + detailed review. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 12] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +8. IANA Considerations + +8.1. UDP Port + + A new UDP port will be requested from IANA. + +8.2. VXLAN-gpe Next Protocol + + IANA is requested to set up a registry of "Next Protocol". These are + 8-bit values. Next Protocol values 0, 1, 2, 3 and 4 are defined in + this draft. New values are assigned via Standards Action [RFC5226]. + + +---------------+-------------+---------------+ + | Next Protocol | Description | Reference | + +---------------+-------------+---------------+ + | 0 | Reserved | This document | + | | | | + | 1 | IPv4 | This document | + | | | | + | 2 | IPv6 | This document | + | | | | + | 3 | Ethernet | This document | + | | | | + | 4 | NSH | This document | + | | | | + | 5..253 | Unassigned | | + +---------------+-------------+---------------+ + + Table 1 + +8.3. VXLAN-gpe Reserved Bits + + There are ten bits at the beginning of the VXLAN-gpe header. New + bits are assigned via Standards Action [RFC5226]. + + Bits 0-3 - Reserved + Bit 4 - Instance ID (I bit) + Bit 5 - Next Protocol (P bit) + Bit 6 - Reserved + Bit 7 - OAM (O bit) + Bits 8-9 - Version + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 13] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +9. References + +9.1. Normative References + + [RFC0768] Postel, J., "User Datagram Protocol", STD 6, RFC 768, + August 1980. + + [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, + September 1981. + + [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate + Requirement Levels", BCP 14, RFC 2119, March 1997. + + [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an + IANA Considerations Section in RFCs", BCP 26, RFC 5226, + May 2008. + +9.2. Informative References + + [NSH] Quinn, P. and et al. , "Network Service Header", 2014. + + [RFC1700] Reynolds, J. and J. Postel, "Assigned Numbers", RFC 1700, + October 1994. + + [RFC6830] Farinacci, D., Fuller, V., Meyer, D., and D. Lewis, "The + Locator/ID Separation Protocol (LISP)", RFC 6830, + January 2013. + + [VXLAN] Dutt, D., Mahalingam, M., Duda, K., Agarwal, P., Kreeger, + L., Sridhar, T., Bursell, M., and C. Wright, "VXLAN: A + Framework for Overlaying Virtualized Layer 2 Networks over + Layer 3 Networks", 2013. + + + + + + + + + + + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 14] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + +Authors' Addresses + + Paul Quinn + Cisco Systems, Inc. + + Email: paulq@cisco.com + + + Puneet Agarwal + Broadcom + + Email: pagarwal@broadcom.com + + + Rex Fernando + Cisco Systems, Inc. + + Email: rex@cisco.com + + + Larry Kreeger + Cisco Systems, Inc. + + Email: kreeger@cisco.com + + + Darrel Lewis + Cisco Systems, Inc. + + Email: darlewis@cisco.com + + + Fabio Maino + Cisco Systems, Inc. + + Email: kreeger@cisco.com + + + Michael Smith + Cisco Systems, Inc. + + Email: michsmit@cisco.com + + + + + + + + + +Quinn, et al. Expires January 4, 2015 [Page 15] + +Internet-Draft Generic Protocol Extension for VXLAN July 2014 + + + Navindra Yadav + Cisco Systems, Inc. + + Email: nyadav@cisco.com + + + Lucy Yong + Huawei USA + + Email: lucy.yong@huawei.com + + + Xiaohu Xu + Huawei Technologies + + Email: xuxiaohu@huawei.com + + + Uri Elzur + Intel + + Email: uri.elzur@intel.com + + + Pankaj Garg + Microsoft + + Email: Garg.Pankaj@microsoft.com diff --git a/vnet/vnet/vxlan-gpe/vxlan_gpe.c b/vnet/vnet/vxlan-gpe/vxlan_gpe.c new file mode 100644 index 00000000000..ef242d0bb8f --- /dev/null +++ b/vnet/vnet/vxlan-gpe/vxlan_gpe.c @@ -0,0 +1,467 @@ +/* + * 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 <vnet/vxlan-gpe/vxlan_gpe.h> + +vxlan_gpe_main_t vxlan_gpe_main; + +u8 * format_vxlan_gpe_tunnel (u8 * s, va_list * args) +{ + vxlan_gpe_tunnel_t * t = va_arg (*args, vxlan_gpe_tunnel_t *); + vxlan_gpe_main_t * gm = &vxlan_gpe_main; + + s = format (s, "[%d] local: %U remote: %U ", + t - gm->tunnels, + format_ip4_address, &t->local, + format_ip4_address, &t->remote); + + switch (t->protocol) + { + case VXLAN_GPE_PROTOCOL_IP4: + s = format (s, "next-protocol ip4"); + case VXLAN_GPE_PROTOCOL_IP6: + s = format (s, "next-protocol ip6"); + case VXLAN_GPE_PROTOCOL_ETHERNET: + s = format (s, "next-protocol ethernet"); + case VXLAN_GPE_PROTOCOL_NSH: + s = format (s, "next-protocol nsh"); + default: + s = format (s, "next-protocol unknown %d", t->protocol); + } + + s = format (s, " fibs: (encap %d, decap %d)", + t->encap_fib_index, + t->decap_fib_index); + + s = format (s, " vxlan VNI %d ", t->vni); + + return s; +} + +static u8 * format_vxlan_gpe_name (u8 * s, va_list * args) +{ + vxlan_gpe_main_t * gm = &vxlan_gpe_main; + u32 i = va_arg (*args, u32); + u32 show_dev_instance = ~0; + + if (i < vec_len (gm->dev_inst_by_real)) + show_dev_instance = gm->dev_inst_by_real[i]; + + if (show_dev_instance != ~0) + i = show_dev_instance; + + return format (s, "vxlan_gpe_tunnel%d", i); +} + +static int vxlan_gpe_name_renumber (vnet_hw_interface_t * hi, + u32 new_dev_instance) +{ + vxlan_gpe_main_t * gm = &vxlan_gpe_main; + + vec_validate_init_empty (gm->dev_inst_by_real, hi->dev_instance, ~0); + + gm->dev_inst_by_real [hi->dev_instance] = new_dev_instance; + + return 0; +} + +static uword dummy_interface_tx (vlib_main_t * vm, + vlib_node_runtime_t * node, + vlib_frame_t * frame) +{ + clib_warning ("you shouldn't be here, leaking buffers..."); + return frame->n_vectors; +} + +static uword dummy_set_rewrite (vnet_main_t * vnm, + u32 sw_if_index, + u32 l3_type, + void * dst_address, + void * rewrite, + uword max_rewrite_bytes) +{ + return 0; +} + +VNET_DEVICE_CLASS (vxlan_gpe_device_class,static) = { + .name = "VXLAN_GPE", + .format_device_name = format_vxlan_gpe_name, + .format_tx_trace = format_vxlan_gpe_encap_trace, + .tx_function = dummy_interface_tx, + .name_renumber = vxlan_gpe_name_renumber, +}; + +static u8 * format_vxlan_gpe_header_with_length (u8 * s, va_list * args) +{ + u32 dev_instance = va_arg (*args, u32); + s = format (s, "unimplemented dev %u", dev_instance); + return s; +} + +VNET_HW_INTERFACE_CLASS (vxlan_gpe_hw_class) = { + .name = "VXLAN_GPE", + .format_header = format_vxlan_gpe_header_with_length, + .set_rewrite = dummy_set_rewrite, +}; + + +#define foreach_gpe_copy_field \ +_(local.as_u32) \ +_(remote.as_u32) \ +_(vni) \ +_(protocol) \ +_(encap_fib_index) \ +_(decap_fib_index) + +#define foreach_copy_field \ +_(src.as_u32) \ +_(dst.as_u32) \ +_(vni) \ +_(encap_fib_index) \ +_(decap_fib_index) \ +_(decap_next_index) + + + +static int vxlan_gpe_rewrite (vxlan_gpe_tunnel_t * t) +{ + u8 *rw = 0; + ip4_header_t * ip0; + ip4_vxlan_gpe_header_t * h0; + int len; + + len = sizeof (*h0); + + vec_validate_aligned (rw, len-1, CLIB_CACHE_LINE_BYTES); + + h0 = (ip4_vxlan_gpe_header_t *) rw; + + /* Fixed portion of the (outer) ip4 header */ + ip0 = &h0->ip4; + ip0->ip_version_and_header_length = 0x45; + ip0->ttl = 254; + ip0->protocol = IP_PROTOCOL_UDP; + + /* we fix up the ip4 header length and checksum after-the-fact */ + ip0->src_address.as_u32 = t->local.as_u32; + ip0->dst_address.as_u32 = t->remote.as_u32; + ip0->checksum = ip4_header_checksum (ip0); + + /* UDP header, randomize src port on something, maybe? */ + h0->udp.src_port = clib_host_to_net_u16 (4790); + h0->udp.dst_port = clib_host_to_net_u16 (UDP_DST_PORT_vxlan_gpe); + + /* VXLAN header. Are we having fun yet? */ + h0->vxlan.flags = VXLAN_GPE_FLAGS_I | VXLAN_GPE_FLAGS_P; + h0->vxlan.ver_res = VXLAN_GPE_VERSION; + h0->vxlan.protocol = VXLAN_GPE_PROTOCOL_IP4; + h0->vxlan.vni_res = clib_host_to_net_u32 (t->vni<<8); + + t->rewrite = rw; + return (0); +} + +int vnet_vxlan_gpe_add_del_tunnel +(vnet_vxlan_gpe_add_del_tunnel_args_t *a, u32 * sw_if_indexp) +{ + vxlan_gpe_main_t * gm = &vxlan_gpe_main; + vxlan_gpe_tunnel_t *t = 0; + vnet_main_t * vnm = gm->vnet_main; + vnet_hw_interface_t * hi; + uword * p; + u32 hw_if_index = ~0; + u32 sw_if_index = ~0; + int rv; + vxlan_gpe_tunnel_key_t key, *key_copy; + hash_pair_t *hp; + + key.local = a->local.as_u32; + key.remote = a->remote.as_u32; + key.vni = clib_host_to_net_u32 (a->vni << 8); + key.pad = 0; + + p = hash_get_mem (gm->vxlan_gpe_tunnel_by_key, &key); + + if (a->is_add) + { + /* adding a tunnel: tunnel must not already exist */ + if (p) + return VNET_API_ERROR_INVALID_VALUE; + + if (a->decap_next_index >= VXLAN_GPE_INPUT_N_NEXT) + return VNET_API_ERROR_INVALID_DECAP_NEXT; + + pool_get_aligned (gm->tunnels, t, CLIB_CACHE_LINE_BYTES); + memset (t, 0, sizeof (*t)); + + /* copy from arg structure */ +#define _(x) t->x = a->x; + foreach_gpe_copy_field; +#undef _ + + rv = vxlan_gpe_rewrite (t); + + if (rv) + { + pool_put (gm->tunnels, t); + return rv; + } + + key_copy = clib_mem_alloc (sizeof (*key_copy)); + clib_memcpy (key_copy, &key, sizeof (*key_copy)); + + hash_set_mem (gm->vxlan_gpe_tunnel_by_key, key_copy, + t - gm->tunnels); + + if (vec_len (gm->free_vxlan_gpe_tunnel_hw_if_indices) > 0) + { + hw_if_index = gm->free_vxlan_gpe_tunnel_hw_if_indices + [vec_len (gm->free_vxlan_gpe_tunnel_hw_if_indices)-1]; + _vec_len (gm->free_vxlan_gpe_tunnel_hw_if_indices) -= 1; + + hi = vnet_get_hw_interface (vnm, hw_if_index); + hi->dev_instance = t - gm->tunnels; + hi->hw_instance = hi->dev_instance; + } + else + { + hw_if_index = vnet_register_interface + (vnm, vxlan_gpe_device_class.index, t - gm->tunnels, + vxlan_gpe_hw_class.index, t - gm->tunnels); + hi = vnet_get_hw_interface (vnm, hw_if_index); + hi->output_node_index = vxlan_gpe_encap_node.index; + } + + t->hw_if_index = hw_if_index; + t->sw_if_index = sw_if_index = hi->sw_if_index; + + vnet_sw_interface_set_flags (vnm, hi->sw_if_index, + VNET_SW_INTERFACE_FLAG_ADMIN_UP); + } + else + { + /* deleting a tunnel: tunnel must exist */ + if (!p) + return VNET_API_ERROR_NO_SUCH_ENTRY; + + t = pool_elt_at_index (gm->tunnels, p[0]); + + vnet_sw_interface_set_flags (vnm, t->sw_if_index, 0 /* down */); + vec_add1 (gm->free_vxlan_gpe_tunnel_hw_if_indices, t->hw_if_index); + + hp = hash_get_pair (gm->vxlan_gpe_tunnel_by_key, &key); + key_copy = (void *)(hp->key); + hash_unset_mem (gm->vxlan_gpe_tunnel_by_key, &key); + clib_mem_free (key_copy); + + vec_free (t->rewrite); + pool_put (gm->tunnels, t); + } + + if (sw_if_indexp) + *sw_if_indexp = sw_if_index; + + return 0; +} + +static u32 fib_index_from_fib_id (u32 fib_id) +{ + ip4_main_t * im = &ip4_main; + uword * p; + + p = hash_get (im->fib_index_by_table_id, fib_id); + if (!p) + return ~0; + + return p[0]; +} + +static uword unformat_gpe_decap_next (unformat_input_t * input, va_list * args) +{ + u32 * result = va_arg (*args, u32 *); + u32 tmp; + + if (unformat (input, "drop")) + *result = VXLAN_GPE_INPUT_NEXT_DROP; + else if (unformat (input, "ip4")) + *result = VXLAN_GPE_INPUT_NEXT_IP4_INPUT; + else if (unformat (input, "ip6")) + *result = VXLAN_GPE_INPUT_NEXT_IP6_INPUT; + else if (unformat (input, "ethernet")) + *result = VXLAN_GPE_INPUT_NEXT_ETHERNET_INPUT; + else if (unformat (input, "nsh")) + *result = VXLAN_GPE_INPUT_NEXT_NSH_INPUT; + else if (unformat (input, "%d", &tmp)) + *result = tmp; + else + return 0; + return 1; +} + +static clib_error_t * +vxlan_gpe_add_del_tunnel_command_fn (vlib_main_t * vm, + unformat_input_t * input, + vlib_cli_command_t * cmd) +{ + unformat_input_t _line_input, * line_input = &_line_input; + u8 is_add = 1; + ip4_address_t local, remote; + u8 local_set = 0; + u8 remote_set = 0; + u32 encap_fib_index = 0; + u32 decap_fib_index = 0; + u8 protocol = VXLAN_GPE_PROTOCOL_IP4; + u32 decap_next_index = VXLAN_GPE_INPUT_NEXT_IP4_INPUT; + u32 vni; + u8 vni_set = 0; + int rv; + u32 tmp; + vnet_vxlan_gpe_add_del_tunnel_args_t _a, * a = &_a; + + /* Get a line of input. */ + if (! unformat_user (input, unformat_line_input, line_input)) + return 0; + + while (unformat_check_input (line_input) != UNFORMAT_END_OF_INPUT) { + if (unformat (line_input, "del")) + is_add = 0; + else if (unformat (line_input, "local %U", + unformat_ip4_address, &local)) + local_set = 1; + else if (unformat (line_input, "remote %U", + unformat_ip4_address, &remote)) + remote_set = 1; + else if (unformat (line_input, "encap-vrf-id %d", &tmp)) + { + encap_fib_index = fib_index_from_fib_id (tmp); + if (encap_fib_index == ~0) + return clib_error_return (0, "nonexistent encap fib id %d", tmp); + } + else if (unformat (line_input, "decap-vrf-id %d", &tmp)) + { + decap_fib_index = fib_index_from_fib_id (tmp); + if (decap_fib_index == ~0) + return clib_error_return (0, "nonexistent decap fib id %d", tmp); + } + else if (unformat (line_input, "decap-next %U", unformat_gpe_decap_next, + &decap_next_index)) + ; + else if (unformat (line_input, "vni %d", &vni)) + vni_set = 1; + else if (unformat(line_input, "next-ip4")) + protocol = VXLAN_GPE_PROTOCOL_IP4; + else if (unformat(line_input, "next-ip6")) + protocol = VXLAN_GPE_PROTOCOL_IP6; + else if (unformat(line_input, "next-ethernet")) + protocol = VXLAN_GPE_PROTOCOL_ETHERNET; + else if (unformat(line_input, "next-nsh")) + protocol = VXLAN_GPE_PROTOCOL_NSH; + else + return clib_error_return (0, "parse error: '%U'", + format_unformat_error, line_input); + } + + unformat_free (line_input); + + if (local_set == 0) + return clib_error_return (0, "tunnel local address not specified"); + + if (remote_set == 0) + return clib_error_return (0, "tunnel remote address not specified"); + + if (vni_set == 0) + return clib_error_return (0, "vni not specified"); + + memset (a, 0, sizeof (*a)); + + a->is_add = is_add; + +#define _(x) a->x = x; + foreach_gpe_copy_field; +#undef _ + + rv = vnet_vxlan_gpe_add_del_tunnel (a, 0 /* hw_if_indexp */); + + switch(rv) + { + case 0: + break; + case VNET_API_ERROR_INVALID_DECAP_NEXT: + return clib_error_return (0, "invalid decap-next..."); + + case VNET_API_ERROR_TUNNEL_EXIST: + return clib_error_return (0, "tunnel already exists..."); + + case VNET_API_ERROR_NO_SUCH_ENTRY: + return clib_error_return (0, "tunnel does not exist..."); + + default: + return clib_error_return + (0, "vnet_vxlan_gpe_add_del_tunnel returned %d", rv); + } + + return 0; +} + +VLIB_CLI_COMMAND (create_vxlan_gpe_tunnel_command, static) = { + .path = "create vxlan-gpe tunnel", + .short_help = + "create vxlan-gpe tunnel local <ip4-addr> remote <ip4-addr>" + " vni <nn> [next-ip4][next-ip6][next-ethernet][next-nsh]" + " [encap-vrf-id <nn>] [decap-vrf-id <nn>]" + " [del]\n", + .function = vxlan_gpe_add_del_tunnel_command_fn, +}; + +static clib_error_t * +show_vxlan_gpe_tunnel_command_fn (vlib_main_t * vm, + unformat_input_t * input, + vlib_cli_command_t * cmd) +{ + vxlan_gpe_main_t * gm = &vxlan_gpe_main; + vxlan_gpe_tunnel_t * t; + + if (pool_elts (gm->tunnels) == 0) + vlib_cli_output (vm, "No vxlan-gpe tunnels configured."); + + pool_foreach (t, gm->tunnels, + ({ + vlib_cli_output (vm, "%U", format_vxlan_gpe_tunnel, t); + })); + + return 0; +} + +VLIB_CLI_COMMAND (show_vxlan_gpe_tunnel_command, static) = { + .path = "show vxlan-gpe", + .function = show_vxlan_gpe_tunnel_command_fn, +}; + +clib_error_t *vxlan_gpe_init (vlib_main_t *vm) +{ + vxlan_gpe_main_t *gm = &vxlan_gpe_main; + + gm->vnet_main = vnet_get_main(); + gm->vlib_main = vm; + + gm->vxlan_gpe_tunnel_by_key + = hash_create_mem (0, sizeof(vxlan_gpe_tunnel_key_t), sizeof (uword)); + + udp_register_dst_port (vm, UDP_DST_PORT_vxlan_gpe, + vxlan_gpe_input_node.index, 1 /* is_ip4 */); + return 0; +} + +VLIB_INIT_FUNCTION(vxlan_gpe_init); + diff --git a/vnet/vnet/vxlan-gpe/vxlan_gpe.h b/vnet/vnet/vxlan-gpe/vxlan_gpe.h new file mode 100644 index 00000000000..4c2ac444e34 --- /dev/null +++ b/vnet/vnet/vxlan-gpe/vxlan_gpe.h @@ -0,0 +1,139 @@ +/* + * 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. + */ +#ifndef included_vnet_vxlan_gpe_h +#define included_vnet_vxlan_gpe_h + +#include <vppinfra/error.h> +#include <vppinfra/hash.h> +#include <vnet/vnet.h> +#include <vnet/ip/ip.h> +#include <vnet/l2/l2_input.h> +#include <vnet/ethernet/ethernet.h> +#include <vnet/vxlan-gpe/vxlan_gpe_packet.h> +#include <vnet/ip/ip4_packet.h> +#include <vnet/ip/udp.h> + + +typedef CLIB_PACKED (struct { + ip4_header_t ip4; /* 20 bytes */ + udp_header_t udp; /* 8 bytes */ + vxlan_gpe_header_t vxlan; /* 8 bytes */ +}) ip4_vxlan_gpe_header_t; + +typedef CLIB_PACKED(struct { + /* + * Key fields: local remote, vni + * all fields in NET byte order + */ + union { + struct { + u32 local; + u32 remote; + u32 vni; /* shifted 8 bits */ + u32 pad; + }; + u64 as_u64[2]; + }; +}) vxlan_gpe_tunnel_key_t; + +typedef struct { + /* Rewrite string. $$$$ embed vnet_rewrite header */ + u8 * rewrite; + + /* encapsulated protocol */ + u8 protocol; + + /* tunnel src and dst addresses */ + ip4_address_t local; + ip4_address_t remote; + + /* FIB indices */ + u32 encap_fib_index; /* tunnel partner lookup here */ + u32 decap_fib_index; /* inner IP lookup here */ + + /* vxlan VNI in HOST byte order, shifted left 8 bits */ + u32 vni; + + /* vnet intfc hw/sw_if_index */ + u32 hw_if_index; + u32 sw_if_index; + +} vxlan_gpe_tunnel_t; + +#define foreach_vxlan_gpe_input_next \ +_(DROP, "error-drop") \ +_(IP4_INPUT, "ip4-input") \ +_(IP6_INPUT, "ip6-input") \ +_(ETHERNET_INPUT, "ethernet-input") \ +_(NSH_INPUT, "nsh-input") + +typedef enum { +#define _(s,n) VXLAN_GPE_INPUT_NEXT_##s, + foreach_vxlan_gpe_input_next +#undef _ + VXLAN_GPE_INPUT_N_NEXT, +} vxlan_gpe_input_next_t; + +typedef enum { +#define vxlan_gpe_error(n,s) VXLAN_GPE_ERROR_##n, +#include <vnet/vxlan-gpe/vxlan_gpe_error.def> +#undef vxlan_gpe_error + VXLAN_GPE_N_ERROR, +} vxlan_gpe_input_error_t; + +typedef struct { + /* vector of encap tunnel instances */ + vxlan_gpe_tunnel_t *tunnels; + + /* lookup tunnel by key */ + uword * vxlan_gpe_tunnel_by_key; + + /* Free vlib hw_if_indices */ + u32 * free_vxlan_gpe_tunnel_hw_if_indices; + + /* show device instance by real device instance */ + u32 * dev_inst_by_real; + + /* convenience */ + vlib_main_t * vlib_main; + vnet_main_t * vnet_main; +} vxlan_gpe_main_t; + +vxlan_gpe_main_t vxlan_gpe_main; + +extern vlib_node_registration_t vxlan_gpe_encap_node; +extern vlib_node_registration_t vxlan_gpe_input_node; + +u8 * format_vxlan_gpe_encap_trace (u8 * s, va_list * args); + +typedef struct { + u8 is_add; + ip4_address_t local, remote; + u8 protocol; + u32 encap_fib_index; + u32 decap_fib_index; + u32 decap_next_index; + u32 vni; +} vnet_vxlan_gpe_add_del_tunnel_args_t; + + +int vnet_vxlan_gpe_add_del_tunnel +(vnet_vxlan_gpe_add_del_tunnel_args_t *a, u32 * sw_if_indexp); + + + + + +#endif /* included_vnet_vxlan_gpe_h */ diff --git a/vnet/vnet/vxlan-gpe/vxlan_gpe_error.def b/vnet/vnet/vxlan-gpe/vxlan_gpe_error.def new file mode 100644 index 00000000000..9cf1b1cb656 --- /dev/null +++ b/vnet/vnet/vxlan-gpe/vxlan_gpe_error.def @@ -0,0 +1,16 @@ +/* + * 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. + */ +vxlan_gpe_error (DECAPSULATED, "good packets decapsulated") +vxlan_gpe_error (NO_SUCH_TUNNEL, "no such tunnel packets") diff --git a/vnet/vnet/vxlan-gpe/vxlan_gpe_packet.h b/vnet/vnet/vxlan-gpe/vxlan_gpe_packet.h new file mode 100644 index 00000000000..3403cc9ebad --- /dev/null +++ b/vnet/vnet/vxlan-gpe/vxlan_gpe_packet.h @@ -0,0 +1,82 @@ +/* + * 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. + */ +#ifndef included_vxlan_gpe_packet_h +#define included_vxlan_gpe_packet_h + +/* + * From draft-quinn-vxlan-gpe-03.txt + * + * 0 1 2 3 + * 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * |R|R|R|R|I|P|R|O|Ver| Reserved |Next Protocol | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | VXLAN Network Identifier (VNI) | Reserved | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * + * I Bit: Flag bit 4 indicates that the VNI is valid. + * + * P Bit: Flag bit 5 is defined as the Next Protocol bit. The P bit + * MUST be set to 1 to indicate the presence of the 8 bit next + * protocol field. + * + * O Bit: Flag bit 7 is defined as the O bit. When the O bit is set to 1, + * + * the packet is an OAM packet and OAM processing MUST occur. The OAM + * protocol details are out of scope for this document. As with the + * P-bit, bit 7 is currently a reserved flag in VXLAN. + * + * VXLAN-gpe bits 8 and 9 are defined as version bits. These bits are + * reserved in VXLAN. The version field is used to ensure backward + * compatibility going forward with future VXLAN-gpe updates. + * + * The initial version for VXLAN-gpe is 0. + * + * This draft defines the following Next Protocol values: + * + * 0x1 : IPv4 + * 0x2 : IPv6 + * 0x3 : Ethernet + * 0x4 : Network Service Header [NSH] + */ + +#define foreach_vxlan_gpe_protocol \ +_ (0x01, IP4) \ +_ (0x02, IP6) \ +_ (0x03, ETHERNET) \ +_ (0x04, NSH) + + +typedef enum { +#define _(n,f) VXLAN_GPE_PROTOCOL_##f = n, + foreach_vxlan_gpe_protocol +#undef _ +} vxlan_gpe_protocol_t; + +typedef struct { + u8 flags; + u8 ver_res; + u8 res; + /* see vxlan_gpe_protocol_t */ + u8 protocol; + u32 vni_res; +} vxlan_gpe_header_t; + +#define VXLAN_GPE_FLAGS_I 0x08 +#define VXLAN_GPE_FLAGS_P 0x04 +#define VXLAN_GPE_FLAGS_O 0x01 +#define VXLAN_GPE_VERSION 0x0 + +#endif /* included_vxlan_gpe_packet_h */ |