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
8 DPDK to demonstrate usage of poll and event mode packet I/O mechanism.
9 The application performs L3 forwarding.
14 The application demonstrates the use of the hash and LPM libraries in the DPDK
15 to implement packet forwarding using poll or event mode PMDs for packet I/O.
16 The initialization and run-time paths are very similar to those of the
17 :doc:`l2_forward_real_virtual` and :doc:`l2_forward_event`.
18 The main difference from the L2 Forwarding sample application is that optionally
19 packet can be Rx/Tx from/to eventdev instead of port directly and forwarding
20 decision is made based on information read from the input packet.
22 Eventdev can optionally use S/W or H/W (if supported by platform) scheduler
23 implementation for packet I/O based on run time parameters.
25 The lookup method is either hash-based or LPM-based and is selected at run time. When the selected lookup method is hash-based,
26 a hash object is used to emulate the flow classification stage.
27 The hash object is used in correlation with a flow table to map each input packet to its flow at runtime.
29 The hash lookup key is represented by a DiffServ 5-tuple composed of the following fields read from the input packet:
30 Source IP Address, Destination IP Address, Protocol, Source Port and Destination Port.
31 The ID of the output interface for the input packet is read from the identified flow table entry.
32 The set of flows used by the application is statically configured and loaded into the hash at initialization time.
33 When the selected lookup method is LPM based, an LPM object is used to emulate the forwarding stage for IPv4 packets.
34 The LPM object is used as the routing table to identify the next hop for each input packet at runtime.
36 The LPM lookup key is represented by the Destination IP Address field read from the input packet.
37 The ID of the output interface for the input packet is the next hop returned by the LPM lookup.
38 The set of LPM rules used by the application is statically configured and loaded into the LPM object at initialization time.
40 In the sample application, hash-based forwarding supports IPv4 and IPv6. LPM-based forwarding supports IPv4 only.
42 Compiling the Application
43 -------------------------
45 To compile the sample application see :doc:`compiling`.
47 The application is located in the ``l3fwd`` sub-directory.
49 Running the Application
50 -----------------------
52 The application has a number of command line options::
54 ./dpdk-l3fwd [EAL options] -- -p PORTMASK
56 [--lookup LOOKUP_METHOD]
57 --config(port,queue,lcore)[,(port,queue,lcore)]
58 [--eth-dest=X,MM:MM:MM:MM:MM:MM]
59 [--enable-jumbo [--max-pkt-len PKTLEN]]
73 * ``-p PORTMASK:`` Hexadecimal bitmask of ports to configure
75 * ``-P:`` Optional, sets all ports to promiscuous mode so that packets are accepted regardless of the packet's Ethernet MAC destination address.
76 Without this option, only packets with the Ethernet MAC destination address set to the Ethernet address of the port are accepted.
78 * ``--lookup:`` Optional, select the lookup method.
81 ``lpm`` (Longest Prefix Match),
82 ``fib`` (Forwarding Information Base).
85 * ``--config (port,queue,lcore)[,(port,queue,lcore)]:`` Determines which queues from which ports are mapped to which cores.
87 * ``--eth-dest=X,MM:MM:MM:MM:MM:MM:`` Optional, ethernet destination for port X.
89 * ``--enable-jumbo:`` Optional, enables jumbo frames.
91 * ``--max-pkt-len:`` Optional, under the premise of enabling jumbo, maximum packet length in decimal (64-9600).
93 * ``--no-numa:`` Optional, disables numa awareness.
95 * ``--hash-entry-num:`` Optional, specifies the hash entry number in hexadecimal to be setup.
97 * ``--ipv6:`` Optional, set if running ipv6 packets.
99 * ``--parse-ptype:`` Optional, set to use software to analyze packet type. Without this option, hardware will check the packet type.
101 * ``--per-port-pool:`` Optional, set to use independent buffer pools per port. Without this option, single buffer pool is used for all ports.
103 * ``--mode:`` Optional, Packet transfer mode for I/O, poll or eventdev.
105 * ``--eventq-sched:`` Optional, Event queue synchronization method, Ordered, Atomic or Parallel. Only valid if --mode=eventdev.
107 * ``--event-eth-rxqs:`` Optional, Number of ethernet RX queues per device. Only valid if --mode=eventdev.
109 * ``-E:`` Optional, enable exact match,
110 legacy flag, please use ``--lookup=em`` instead.
112 * ``-L:`` Optional, enable longest prefix match,
113 legacy flag, please use ``--lookup=lpm`` instead.
116 For example, consider a dual processor socket platform with 8 physical cores, where cores 0-7 and 16-23 appear on socket 0,
117 while cores 8-15 and 24-31 appear on socket 1.
119 To enable L3 forwarding between two ports, assuming that both ports are in the same socket, using two cores, cores 1 and 2,
120 (which are in the same socket too), use the following command:
122 .. code-block:: console
124 ./<build_dir>/examples/dpdk-l3fwd -l 1,2 -n 4 -- -p 0x3 --config="(0,0,1),(1,0,2)"
128 * The -l option enables cores 1, 2
130 * The -p option enables ports 0 and 1
132 * The --config option enables one queue on each port and maps each (port,queue) pair to a specific core.
133 The following table shows the mapping in this example:
135 +----------+-----------+-----------+-------------------------------------+
136 | **Port** | **Queue** | **lcore** | **Description** |
138 +----------+-----------+-----------+-------------------------------------+
139 | 0 | 0 | 1 | Map queue 0 from port 0 to lcore 1. |
141 +----------+-----------+-----------+-------------------------------------+
142 | 1 | 0 | 2 | Map queue 0 from port 1 to lcore 2. |
144 +----------+-----------+-----------+-------------------------------------+
146 To use eventdev mode with sync method **ordered** on above mentioned environment,
147 Following is the sample command:
149 .. code-block:: console
151 ./<build_dir>/examples/dpdk-l3fwd -l 0-3 -n 4 -a <event device> -- -p 0x3 --eventq-sched=ordered
155 .. code-block:: console
157 ./<build_dir>/examples/dpdk-l3fwd -l 0-3 -n 4 -a <event device> \
158 -- -p 0x03 --mode=eventdev --eventq-sched=ordered
162 * -a option allows the event device supported by platform.
163 The syntax used to indicate this device may vary based on platform.
165 * The --mode option defines PMD to be used for packet I/O.
167 * The --eventq-sched option enables synchronization menthod of event queue so that packets will be scheduled accordingly.
169 If application uses S/W scheduler, it uses following DPDK services:
172 * Rx adapter service function
173 * Tx adapter service function
175 Application needs service cores to run above mentioned services. Service cores
176 must be provided as EAL parameters along with the --vdev=event_sw0 to enable S/W
177 scheduler. Following is the sample command:
179 .. code-block:: console
181 ./<build_dir>/examples/dpdk-l3fwd -l 0-7 -s 0xf0000 -n 4 --vdev event_sw0 -- -p 0x3 --mode=eventdev --eventq-sched=ordered
183 In case of eventdev mode, *--config* option is not used for ethernet port
184 configuration. Instead each ethernet port will be configured with mentioned
189 * Each Rx queue will be connected to event queue via Rx adapter.
191 * Each Tx queue will be connected via Tx adapter.
193 Refer to the *DPDK Getting Started Guide* for general information on running applications and
194 the Environment Abstraction Layer (EAL) options.
196 .. _l3_fwd_explanation:
201 The following sections provide some explanation of the sample application code. As mentioned in the overview section,
202 the initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual` and :doc:`l2_forward_event`.
203 The following sections describe aspects that are specific to the L3 Forwarding sample application.
208 The hash object is created and loaded with the pre-configured entries read from a global array,
209 and then generate the expected 5-tuple as key to keep consistence with those of real flow
210 for the convenience to execute hash performance test on 4M/8M/16M flows.
214 The Hash initialization will setup both ipv4 and ipv6 hash table,
215 and populate the either table depending on the value of variable ipv6.
216 To support the hash performance test with up to 8M single direction flows/16M bi-direction flows,
217 populate_ipv4_many_flow_into_table() function will populate the hash table with specified hash table entry number(default 4M).
221 Value of global variable ipv6 can be specified with --ipv6 in the command line.
222 Value of global variable hash_entry_number,
223 which is used to specify the total hash entry number for all used ports in hash performance test,
224 can be specified with --hash-entry-num VALUE in command line, being its default value 4.
228 #if (APP_LOOKUP_METHOD == APP_LOOKUP_EXACT_MATCH)
231 setup_hash(int socketid)
235 if (hash_entry_number != HASH_ENTRY_NUMBER_DEFAULT) {
237 /* populate the ipv4 hash */
238 populate_ipv4_many_flow_into_table(ipv4_l3fwd_lookup_struct[socketid], hash_entry_number);
240 /* populate the ipv6 hash */
241 populate_ipv6_many_flow_into_table( ipv6_l3fwd_lookup_struct[socketid], hash_entry_number);
245 /* populate the ipv4 hash */
246 populate_ipv4_few_flow_into_table(ipv4_l3fwd_lookup_struct[socketid]);
248 /* populate the ipv6 hash */
249 populate_ipv6_few_flow_into_table(ipv6_l3fwd_lookup_struct[socketid]);
258 The LPM object is created and loaded with the pre-configured entries read from a global array.
262 #if (APP_LOOKUP_METHOD == APP_LOOKUP_LPM)
265 setup_lpm(int socketid)
271 /* create the LPM table */
273 snprintf(s, sizeof(s), "IPV4_L3FWD_LPM_%d", socketid);
275 ipv4_l3fwd_lookup_struct[socketid] = rte_lpm_create(s, socketid, IPV4_L3FWD_LPM_MAX_RULES, 0);
277 if (ipv4_l3fwd_lookup_struct[socketid] == NULL)
278 rte_exit(EXIT_FAILURE, "Unable to create the l3fwd LPM table"
279 " on socket %d\n", socketid);
281 /* populate the LPM table */
283 for (i = 0; i < IPV4_L3FWD_NUM_ROUTES; i++) {
284 /* skip unused ports */
286 if ((1 << ipv4_l3fwd_route_array[i].if_out & enabled_port_mask) == 0)
289 ret = rte_lpm_add(ipv4_l3fwd_lookup_struct[socketid], ipv4_l3fwd_route_array[i].ip,
290 ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
293 rte_exit(EXIT_FAILURE, "Unable to add entry %u to the "
294 "l3fwd LPM table on socket %d\n", i, socketid);
297 printf("LPM: Adding route 0x%08x / %d (%d)\n",
298 (unsigned)ipv4_l3fwd_route_array[i].ip, ipv4_l3fwd_route_array[i].depth, ipv4_l3fwd_route_array[i].if_out);
303 Packet Forwarding for Hash-based Lookups
304 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
306 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward()
307 or simple_ipv4_fwd_4pkts() function for IPv4 packets or the simple_ipv6_fwd_4pkts() function for IPv6 packets.
308 The l3fwd_simple_forward() function provides the basic functionality for both IPv4 and IPv6 packet forwarding
309 for any number of burst packets received,
310 and the packet forwarding decision (that is, the identification of the output interface for the packet)
311 for hash-based lookups is done by the get_ipv4_dst_port() or get_ipv6_dst_port() function.
312 The get_ipv4_dst_port() function is shown below:
316 static inline uint8_t
317 get_ipv4_dst_port(void *ipv4_hdr, uint16_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
320 union ipv4_5tuple_host key;
322 ipv4_hdr = (uint8_t *)ipv4_hdr + offsetof(struct rte_ipv4_hdr, time_to_live);
324 m128i data = _mm_loadu_si128(( m128i*)(ipv4_hdr));
326 /* Get 5 tuple: dst port, src port, dst IP address, src IP address and protocol */
328 key.xmm = _mm_and_si128(data, mask0);
330 /* Find destination port */
332 ret = rte_hash_lookup(ipv4_l3fwd_lookup_struct, (const void *)&key);
334 return (uint8_t)((ret < 0)? portid : ipv4_l3fwd_out_if[ret]);
337 The get_ipv6_dst_port() function is similar to the get_ipv4_dst_port() function.
339 The simple_ipv4_fwd_4pkts() and simple_ipv6_fwd_4pkts() function are optimized for continuous 4 valid ipv4 and ipv6 packets,
340 they leverage the multiple buffer optimization to boost the performance of forwarding packets with the exact match on hash table.
341 The key code snippet of simple_ipv4_fwd_4pkts() is shown below:
346 simple_ipv4_fwd_4pkts(struct rte_mbuf* m[4], uint16_t portid, struct lcore_conf *qconf)
350 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)));
351 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)));
352 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)));
353 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)));
355 key[0].xmm = _mm_and_si128(data[0], mask0);
356 key[1].xmm = _mm_and_si128(data[1], mask0);
357 key[2].xmm = _mm_and_si128(data[2], mask0);
358 key[3].xmm = _mm_and_si128(data[3], mask0);
360 const void *key_array[4] = {&key[0], &key[1], &key[2],&key[3]};
362 rte_hash_lookup_bulk(qconf->ipv4_lookup_struct, &key_array[0], 4, ret);
364 dst_port[0] = (ret[0] < 0)? portid:ipv4_l3fwd_out_if[ret[0]];
365 dst_port[1] = (ret[1] < 0)? portid:ipv4_l3fwd_out_if[ret[1]];
366 dst_port[2] = (ret[2] < 0)? portid:ipv4_l3fwd_out_if[ret[2]];
367 dst_port[3] = (ret[3] < 0)? portid:ipv4_l3fwd_out_if[ret[3]];
372 The simple_ipv6_fwd_4pkts() function is similar to the simple_ipv4_fwd_4pkts() function.
374 Known issue: IP packets with extensions or IP packets which are not TCP/UDP cannot work well at this mode.
376 Packet Forwarding for LPM-based Lookups
377 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
379 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward() function,
380 but the packet forwarding decision (that is, the identification of the output interface for the packet)
381 for LPM-based lookups is done by the get_ipv4_dst_port() function below:
385 static inline uint16_t
386 get_ipv4_dst_port(struct rte_ipv4_hdr *ipv4_hdr, uint16_t portid, lookup_struct_t *ipv4_l3fwd_lookup_struct)
390 return ((rte_lpm_lookup(ipv4_l3fwd_lookup_struct, rte_be_to_cpu_32(ipv4_hdr->dst_addr), &next_hop) == 0)? next_hop : portid);
393 Eventdev Driver Initialization
394 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
395 Eventdev driver initialization is same as L2 forwarding eventdev application.
396 Refer :doc:`l2_forward_event` for more details.