1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2010-2014 Intel Corporation.
4 L3 Forwarding Sample Application
5 ================================
7 The L3 Forwarding application is a simple example of packet processing using the DPDK.
8 The application performs L3 forwarding.
13 The application demonstrates the use of the hash and LPM libraries in the DPDK to implement packet forwarding.
14 The initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual`.
15 The main difference from the L2 Forwarding sample application is that the forwarding decision
16 is made based on information read from the input packet.
18 The lookup method is either hash-based or LPM-based and is selected at run time. When the selected lookup method is hash-based,
19 a hash object is used to emulate the flow classification stage.
20 The hash object is used in correlation with a flow table to map each input packet to its flow at runtime.
22 The hash lookup key is represented by a DiffServ 5-tuple composed of the following fields read from the input packet:
23 Source IP Address, Destination IP Address, Protocol, Source Port and Destination Port.
24 The ID of the output interface for the input packet is read from the identified flow table entry.
25 The set of flows used by the application is statically configured and loaded into the hash at initialization time.
26 When the selected lookup method is LPM based, an LPM object is used to emulate the forwarding stage for IPv4 packets.
27 The LPM object is used as the routing table to identify the next hop for each input packet at runtime.
29 The LPM lookup key is represented by the Destination IP Address field read from the input packet.
30 The ID of the output interface for the input packet is the next hop returned by the LPM lookup.
31 The set of LPM rules used by the application is statically configured and loaded into the LPM object at initialization time.
33 In the sample application, hash-based forwarding supports IPv4 and IPv6. LPM-based forwarding supports IPv4 only.
35 Compiling the Application
36 -------------------------
38 To compile the sample application see :doc:`compiling`.
40 The application is located in the ``l3fwd`` sub-directory.
42 Running the Application
43 -----------------------
45 The application has a number of command line options::
47 ./l3fwd [EAL options] -- -p PORTMASK
51 --config(port,queue,lcore)[,(port,queue,lcore)]
52 [--eth-dest=X,MM:MM:MM:MM:MM:MM]
53 [--enable-jumbo [--max-pkt-len PKTLEN]]
62 * ``-p PORTMASK:`` Hexadecimal bitmask of ports to configure
64 * ``-P:`` Optional, sets all ports to promiscuous mode so that packets are accepted regardless of the packet's Ethernet MAC destination address.
65 Without this option, only packets with the Ethernet MAC destination address set to the Ethernet address of the port are accepted.
67 * ``-E:`` Optional, enable exact match.
69 * ``-L:`` Optional, enable longest prefix match.
71 * ``--config (port,queue,lcore)[,(port,queue,lcore)]:`` Determines which queues from which ports are mapped to which cores.
73 * ``--eth-dest=X,MM:MM:MM:MM:MM:MM:`` Optional, ethernet destination for port X.
75 * ``--enable-jumbo:`` Optional, enables jumbo frames.
77 * ``--max-pkt-len:`` Optional, under the premise of enabling jumbo, maximum packet length in decimal (64-9600).
79 * ``--no-numa:`` Optional, disables numa awareness.
81 * ``--hash-entry-num:`` Optional, specifies the hash entry number in hexadecimal to be setup.
83 * ``--ipv6:`` Optional, set if running ipv6 packets.
85 * ``--parse-ptype:`` Optional, set to use software to analyze packet type. Without this option, hardware will check the packet type.
87 * ``--per-port-pool:`` Optional, set to use independent buffer pools per port. Without this option, single buffer pool is used for all ports.
89 For example, consider a dual processor socket platform with 8 physical cores, where cores 0-7 and 16-23 appear on socket 0,
90 while cores 8-15 and 24-31 appear on socket 1.
92 To enable L3 forwarding between two ports, assuming that both ports are in the same socket, using two cores, cores 1 and 2,
93 (which are in the same socket too), use the following command:
95 .. code-block:: console
97 ./build/l3fwd -l 1,2 -n 4 -- -p 0x3 --config="(0,0,1),(1,0,2)"
101 * The -l option enables cores 1, 2
103 * The -p option enables ports 0 and 1
105 * The --config option enables one queue on each port and maps each (port,queue) pair to a specific core.
106 The following table shows the mapping in this example:
108 +----------+-----------+-----------+-------------------------------------+
109 | **Port** | **Queue** | **lcore** | **Description** |
111 +----------+-----------+-----------+-------------------------------------+
112 | 0 | 0 | 1 | Map queue 0 from port 0 to lcore 1. |
114 +----------+-----------+-----------+-------------------------------------+
115 | 1 | 0 | 2 | Map queue 0 from port 1 to lcore 2. |
117 +----------+-----------+-----------+-------------------------------------+
119 Refer to the *DPDK Getting Started Guide* for general information on running applications and
120 the Environment Abstraction Layer (EAL) options.
122 .. _l3_fwd_explanation:
127 The following sections provide some explanation of the sample application code. As mentioned in the overview section,
128 the initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual`.
129 The following sections describe aspects that are specific to the L3 Forwarding sample application.
134 The hash object is created and loaded with the pre-configured entries read from a global array,
135 and then generate the expected 5-tuple as key to keep consistence with those of real flow
136 for the convenience to execute hash performance test on 4M/8M/16M flows.
140 The Hash initialization will setup both ipv4 and ipv6 hash table,
141 and populate the either table depending on the value of variable ipv6.
142 To support the hash performance test with up to 8M single direction flows/16M bi-direction flows,
143 populate_ipv4_many_flow_into_table() function will populate the hash table with specified hash table entry number(default 4M).
147 Value of global variable ipv6 can be specified with --ipv6 in the command line.
148 Value of global variable hash_entry_number,
149 which is used to specify the total hash entry number for all used ports in hash performance test,
150 can be specified with --hash-entry-num VALUE in command line, being its default value 4.
154 #if (APP_LOOKUP_METHOD == APP_LOOKUP_EXACT_MATCH)
157 setup_hash(int socketid)
161 if (hash_entry_number != HASH_ENTRY_NUMBER_DEFAULT) {
163 /* populate the ipv4 hash */
164 populate_ipv4_many_flow_into_table(ipv4_l3fwd_lookup_struct[socketid], hash_entry_number);
166 /* populate the ipv6 hash */
167 populate_ipv6_many_flow_into_table( ipv6_l3fwd_lookup_struct[socketid], hash_entry_number);
171 /* populate the ipv4 hash */
172 populate_ipv4_few_flow_into_table(ipv4_l3fwd_lookup_struct[socketid]);
174 /* populate the ipv6 hash */
175 populate_ipv6_few_flow_into_table(ipv6_l3fwd_lookup_struct[socketid]);
184 The LPM object is created and loaded with the pre-configured entries read from a global array.
188 #if (APP_LOOKUP_METHOD == APP_LOOKUP_LPM)
191 setup_lpm(int socketid)
197 /* create the LPM table */
199 snprintf(s, sizeof(s), "IPV4_L3FWD_LPM_%d", socketid);
201 ipv4_l3fwd_lookup_struct[socketid] = rte_lpm_create(s, socketid, IPV4_L3FWD_LPM_MAX_RULES, 0);
203 if (ipv4_l3fwd_lookup_struct[socketid] == NULL)
204 rte_exit(EXIT_FAILURE, "Unable to create the l3fwd LPM table"
205 " on socket %d\n", socketid);
207 /* populate the LPM table */
209 for (i = 0; i < IPV4_L3FWD_NUM_ROUTES; i++) {
210 /* skip unused ports */
212 if ((1 << ipv4_l3fwd_route_array[i].if_out & enabled_port_mask) == 0)
215 ret = rte_lpm_add(ipv4_l3fwd_lookup_struct[socketid], ipv4_l3fwd_route_array[i].ip,
216 ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
219 rte_exit(EXIT_FAILURE, "Unable to add entry %u to the "
220 "l3fwd LPM table on socket %d\n", i, socketid);
223 printf("LPM: Adding route 0x%08x / %d (%d)\n",
224 (unsigned)ipv4_l3fwd_route_array[i].ip, ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
229 Packet Forwarding for Hash-based Lookups
230 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
232 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward()
233 or simple_ipv4_fwd_4pkts() function for IPv4 packets or the simple_ipv6_fwd_4pkts() function for IPv6 packets.
234 The l3fwd_simple_forward() function provides the basic functionality for both IPv4 and IPv6 packet forwarding
235 for any number of burst packets received,
236 and the packet forwarding decision (that is, the identification of the output interface for the packet)
237 for hash-based lookups is done by the get_ipv4_dst_port() or get_ipv6_dst_port() function.
238 The get_ipv4_dst_port() function is shown below:
242 static inline uint8_t
243 get_ipv4_dst_port(void *ipv4_hdr, uint16_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
246 union ipv4_5tuple_host key;
248 ipv4_hdr = (uint8_t *)ipv4_hdr + offsetof(struct rte_ipv4_hdr, time_to_live);
250 m128i data = _mm_loadu_si128(( m128i*)(ipv4_hdr));
252 /* Get 5 tuple: dst port, src port, dst IP address, src IP address and protocol */
254 key.xmm = _mm_and_si128(data, mask0);
256 /* Find destination port */
258 ret = rte_hash_lookup(ipv4_l3fwd_lookup_struct, (const void *)&key);
260 return (uint8_t)((ret < 0)? portid : ipv4_l3fwd_out_if[ret]);
263 The get_ipv6_dst_port() function is similar to the get_ipv4_dst_port() function.
265 The simple_ipv4_fwd_4pkts() and simple_ipv6_fwd_4pkts() function are optimized for continuous 4 valid ipv4 and ipv6 packets,
266 they leverage the multiple buffer optimization to boost the performance of forwarding packets with the exact match on hash table.
267 The key code snippet of simple_ipv4_fwd_4pkts() is shown below:
272 simple_ipv4_fwd_4pkts(struct rte_mbuf* m[4], uint16_t portid, struct lcore_conf *qconf)
276 data[0] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[0], unsigned char *) + sizeof(struct rte_ether_hdr) + offsetof(struct rte_ipv4_hdr, time_to_live)));
277 data[1] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[1], unsigned char *) + sizeof(struct rte_ether_hdr) + offsetof(struct rte_ipv4_hdr, time_to_live)));
278 data[2] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[2], unsigned char *) + sizeof(struct rte_ether_hdr) + offsetof(struct rte_ipv4_hdr, time_to_live)));
279 data[3] = _mm_loadu_si128(( m128i*)(rte_pktmbuf_mtod(m[3], unsigned char *) + sizeof(struct rte_ether_hdr) + offsetof(struct rte_ipv4_hdr, time_to_live)));
281 key[0].xmm = _mm_and_si128(data[0], mask0);
282 key[1].xmm = _mm_and_si128(data[1], mask0);
283 key[2].xmm = _mm_and_si128(data[2], mask0);
284 key[3].xmm = _mm_and_si128(data[3], mask0);
286 const void *key_array[4] = {&key[0], &key[1], &key[2],&key[3]};
288 rte_hash_lookup_bulk(qconf->ipv4_lookup_struct, &key_array[0], 4, ret);
290 dst_port[0] = (ret[0] < 0)? portid:ipv4_l3fwd_out_if[ret[0]];
291 dst_port[1] = (ret[1] < 0)? portid:ipv4_l3fwd_out_if[ret[1]];
292 dst_port[2] = (ret[2] < 0)? portid:ipv4_l3fwd_out_if[ret[2]];
293 dst_port[3] = (ret[3] < 0)? portid:ipv4_l3fwd_out_if[ret[3]];
298 The simple_ipv6_fwd_4pkts() function is similar to the simple_ipv4_fwd_4pkts() function.
300 Known issue: IP packets with extensions or IP packets which are not TCP/UDP cannot work well at this mode.
302 Packet Forwarding for LPM-based Lookups
303 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
305 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward() function,
306 but the packet forwarding decision (that is, the identification of the output interface for the packet)
307 for LPM-based lookups is done by the get_ipv4_dst_port() function below:
311 static inline uint16_t
312 get_ipv4_dst_port(struct rte_ipv4_hdr *ipv4_hdr, uint16_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
316 return ((rte_lpm_lookup(ipv4_l3fwd_lookup_struct, rte_be_to_cpu_32(ipv4_hdr->dst_addr), &next_hop) == 0)? next_hop : portid);