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, LPM, FIB and ACL libraries in 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 hash-based, LPM-based, FIB-based or ACL-based
26 and is selected at run time.
27 When the selected lookup method is hash-based,
28 a hash object is used to emulate the flow classification stage.
29 The hash object is used in correlation with a flow table to map each input packet to its flow at runtime.
31 The hash lookup key is represented by a DiffServ 5-tuple composed of the following fields read from the input packet:
32 Source IP Address, Destination IP Address, Protocol, Source Port and Destination Port.
33 The ID of the output interface for the input packet is read from the identified flow table entry.
34 The set of flows used by the application is statically configured and loaded into the hash at initialization time.
35 When the selected lookup method is LPM or FIB based,
36 an LPM or FIB object is used to emulate the forwarding stage for IPv4 packets.
37 The LPM or FIB object is used as the routing table
38 to identify the next hop for each input packet at runtime.
40 The LPM and FIB lookup keys are represented by the destination IP address field
41 read from the input packet.
42 The ID of the output interface for the input packet is the next hop
43 returned by the LPM or FIB lookup.
44 The set of LPM and FIB rules used by the application is statically configured
45 and loaded into the LPM or FIB object at initialization time.
47 For ACL, the ACL library is used to perform both ACL and route entry lookup.
48 When packets are received from a port,
49 the application extracts the necessary information
50 from the TCP/IP header of the received packet
51 and performs a lookup in the rule database to figure out
52 whether the packets should be dropped (in the ACL range)
53 or forwarded to desired ports.
54 For ACL, the application implements packet classification
55 for the IPv4/IPv6 5-tuple syntax specifically.
56 The 5-tuple syntax consists of a source IP address, a destination IP address,
57 a source port, a destination port and a protocol identifier.
59 In the sample application, hash-based, FIB-based and ACL-based forwarding supports
61 LPM-based forwarding supports IPv4 only.
62 During the initialization phase route rules for IPv4 and IPv6 are read from rule files.
64 Compiling the Application
65 -------------------------
67 To compile the sample application see :doc:`compiling`.
69 The application is located in the ``l3fwd`` sub-directory.
71 Running the Application
72 -----------------------
74 The application has a number of command line options::
76 ./dpdk-l3fwd [EAL options] -- -p PORTMASK
80 [--lookup LOOKUP_METHOD]
81 --config(port,queue,lcore)[,(port,queue,lcore)]
82 [--eth-dest=X,MM:MM:MM:MM:MM:MM]
83 [--max-pkt-len PKTLEN]
92 [--event-vector [--event-vector-size SIZE] [--event-vector-tmo NS]]
98 * ``-p PORTMASK:`` Hexadecimal bitmask of ports to configure
100 * ``--rule_ipv4=FILE:`` specify the ipv4 rules entries file.
101 Each rule occupies one line.
103 * ``--rule_ipv6=FILE:`` specify the ipv6 rules entries file.
105 * ``-P:`` Optional, sets all ports to promiscuous mode so that packets are accepted regardless of the packet's Ethernet MAC destination address.
106 Without this option, only packets with the Ethernet MAC destination address set to the Ethernet address of the port are accepted.
108 * ``--lookup:`` Optional, select the lookup method.
110 ``em`` (Exact Match),
111 ``lpm`` (Longest Prefix Match),
112 ``fib`` (Forwarding Information Base),
113 ``acl`` (Access Control List).
116 * ``--config (port,queue,lcore)[,(port,queue,lcore)]:`` Determines which queues from which ports are mapped to which cores.
118 * ``--eth-dest=X,MM:MM:MM:MM:MM:MM:`` Optional, ethernet destination for port X.
120 * ``--max-pkt-len:`` Optional, maximum packet length in decimal (64-9600).
122 * ``--no-numa:`` Optional, disables numa awareness.
124 * ``--hash-entry-num:`` Optional, specifies the hash entry number in hexadecimal to be setup.
126 * ``--ipv6:`` Optional, set if running ipv6 packets.
128 * ``--parse-ptype:`` Optional, set to use software to analyze packet type. Without this option, hardware will check the packet type.
130 * ``--per-port-pool:`` Optional, set to use independent buffer pools per port. Without this option, single buffer pool is used for all ports.
132 * ``--mode:`` Optional, Packet transfer mode for I/O, poll or eventdev.
134 * ``--eventq-sched:`` Optional, Event queue synchronization method, Ordered, Atomic or Parallel. Only valid if --mode=eventdev.
136 * ``--event-eth-rxqs:`` Optional, Number of ethernet RX queues per device. Only valid if --mode=eventdev.
138 * ``--event-vector:`` Optional, Enable event vectorization. Only valid if --mode=eventdev.
140 * ``--event-vector-size:`` Optional, Max vector size if event vectorization is enabled.
142 * ``--event-vector-tmo:`` Optional, Max timeout to form vector in nanoseconds if event vectorization is enabled.
144 * ``--alg=<val>:`` optional, ACL classify method to use, one of:
145 ``scalar|sse|avx2|neon|altivec|avx512x16|avx512x32``
147 * ``-E:`` Optional, enable exact match,
148 legacy flag, please use ``--lookup=em`` instead.
150 * ``-L:`` Optional, enable longest prefix match,
151 legacy flag, please use ``--lookup=lpm`` instead.
154 For example, consider a dual processor socket platform with 8 physical cores, where cores 0-7 and 16-23 appear on socket 0,
155 while cores 8-15 and 24-31 appear on socket 1.
157 To enable L3 forwarding between two ports, assuming that both ports are in the same socket, using two cores, cores 1 and 2,
158 (which are in the same socket too), use the following command:
160 .. code-block:: console
162 ./<build_dir>/examples/dpdk-l3fwd -l 1,2 -n 4 -- -p 0x3 --config="(0,0,1),(1,0,2)" --rule_ipv4="rule_ipv4.cfg" --rule_ipv6="rule_ipv6.cfg"
166 * The -l option enables cores 1, 2
168 * The -p option enables ports 0 and 1
170 * The --config option enables one queue on each port and maps each (port,queue) pair to a specific core.
171 The following table shows the mapping in this example:
173 +----------+-----------+-----------+-------------------------------------+
174 | **Port** | **Queue** | **lcore** | **Description** |
176 +----------+-----------+-----------+-------------------------------------+
177 | 0 | 0 | 1 | Map queue 0 from port 0 to lcore 1. |
179 +----------+-----------+-----------+-------------------------------------+
180 | 1 | 0 | 2 | Map queue 0 from port 1 to lcore 2. |
182 +----------+-----------+-----------+-------------------------------------+
184 * The -rule_ipv4 option specifies the reading of IPv4 rules sets from the rule_ipv4.cfg file
186 * The -rule_ipv6 option specifies the reading of IPv6 rules sets from the rule_ipv6.cfg file.
188 To use eventdev mode with sync method **ordered** on above mentioned environment,
189 Following is the sample command:
191 .. code-block:: console
193 ./<build_dir>/examples/dpdk-l3fwd -l 0-3 -n 4 -a <event device> -- -p 0x3 --eventq-sched=ordered --rule_ipv4="rule_ipv4.cfg" --rule_ipv6="rule_ipv6.cfg"
197 .. code-block:: console
199 ./<build_dir>/examples/dpdk-l3fwd -l 0-3 -n 4 -a <event device> \
200 -- -p 0x03 --mode=eventdev --eventq-sched=ordered --rule_ipv4="rule_ipv4.cfg" --rule_ipv6="rule_ipv6.cfg"
204 * -a option allows the event device supported by platform.
205 The syntax used to indicate this device may vary based on platform.
207 * The --mode option defines PMD to be used for packet I/O.
209 * The --eventq-sched option enables synchronization menthod of event queue so that packets will be scheduled accordingly.
211 If application uses S/W scheduler, it uses following DPDK services:
214 * Rx adapter service function
215 * Tx adapter service function
217 Application needs service cores to run above mentioned services. Service cores
218 must be provided as EAL parameters along with the --vdev=event_sw0 to enable S/W
219 scheduler. Following is the sample command:
221 .. code-block:: console
223 ./<build_dir>/examples/dpdk-l3fwd -l 0-7 -s 0xf0000 -n 4 --vdev event_sw0 -- -p 0x3 --mode=eventdev --eventq-sched=ordered --rule_ipv4="rule_ipv4.cfg" --rule_ipv6="rule_ipv6.cfg"
225 In case of eventdev mode, *--config* option is not used for ethernet port
226 configuration. Instead each ethernet port will be configured with mentioned
231 * Each Rx queue will be connected to event queue via Rx adapter.
233 * Each Tx queue will be connected via Tx adapter.
235 Refer to the *DPDK Getting Started Guide* for general information on running applications and
236 the Environment Abstraction Layer (EAL) options.
238 .. _l3_fwd_explanation:
243 The following sections provide some explanation of the sample application code. As mentioned in the overview section,
244 the initialization and run-time paths are very similar to those of the :doc:`l2_forward_real_virtual` and :doc:`l2_forward_event`.
245 The following sections describe aspects that are specific to the L3 Forwarding sample application.
247 Parse Rules from File
248 ~~~~~~~~~~~~~~~~~~~~~
250 The application parses the rules from the file and adds them to the appropriate route table by calling the appropriate function.
251 It ignores empty and comment lines, and parses and validates the rules it reads.
252 If errors are detected, the application exits with messages to identify the errors encountered.
254 The format of the route rules differs based on which lookup method is being used.
255 Therefore, the code only decreases the priority number with each rule it parses.
256 Route rules are mandatory.
257 To read data from the specified file successfully, the application assumes the following:
259 * Each rule occupies a single line.
261 * Only the following four rule line types are valid in this application:
263 * Route rule line, which starts with a leading character 'R'
265 * Comment line, which starts with a leading character '#'
267 * ACL rule line, which starts with a leading character ‘@’
269 * Empty line, which consists of a space, form-feed ('\f'), newline ('\n'),
270 carriage return ('\r'), horizontal tab ('\t'), or vertical tab ('\v').
272 Other lines types are considered invalid.
274 * Rules are organized in descending order of priority,
275 which means rules at the head of the file always have a higher priority than those further down in the file.
277 * A typical IPv4 LPM/FIB rule line should have a format as shown below:
279 R<destination_ip>/<ip_mask_length><output_port_number>
281 * A typical IPv4 EM rule line should have a format as shown below:
283 R<destination_ip><source_ip><destination_port><source_port><protocol><output_port_number>
285 * A typical IPv4 ACL rule line should have a format as shown below:
287 .. _figure_ipv4_acl_rule:
289 .. figure:: img/ipv4_acl_rule.*
291 A typical IPv4 ACL rule
293 IPv4 addresses are specified in CIDR format as specified in RFC 4632.
294 For LPM/FIB/ACL they consist of the dot notation for the address
295 and a prefix length separated by '/'.
296 For example, 192.168.0.34/32, where the address is 192.168.0.34 and the prefix length is 32.
297 For EM they consist of just the dot notation for the address and no prefix length.
298 For example, 192.168.0.34, where the Address is 192.168.0.34.
299 EM also includes ports which are specified as a single number which represents a single port.
301 The application parses the rules from the file,
302 it ignores empty and comment lines,
303 and parses and validates the rules it reads.
304 If errors are detected, the application exits
305 with messages to identify the errors encountered.
306 The ACL rules save the index to the specific rules in the userdata field,
307 while route rules save the forwarding port number.
312 The hash object is created and loaded with the pre-configured entries read from a global array,
313 and then generate the expected 5-tuple as key to keep consistence with those of real flow
314 for the convenience to execute hash performance test on 4M/8M/16M flows.
318 The Hash initialization will setup both ipv4 and ipv6 hash table,
319 and populate the either table depending on the value of variable ipv6.
323 Value of global variable ipv6 can be specified with --ipv6 in the command line.
324 Value of global variable hash_entry_number,
325 which is used to specify the total hash entry number for all used ports in hash performance test,
326 can be specified with --hash-entry-num VALUE in command line, being its default value 4.
330 #if (APP_LOOKUP_METHOD == APP_LOOKUP_EXACT_MATCH)
333 setup_hash(int socketid)
338 /* populate the ipv4 hash */
339 populate_ipv4_flow_into_table(
340 ipv4_l3fwd_em_lookup_struct[socketid]);
342 /* populate the ipv6 hash */
343 populate_ipv6_flow_into_table(
344 ipv6_l3fwd_em_lookup_struct[socketid]);
352 The LPM object is created and loaded with the pre-configured entries read from a global array.
354 .. literalinclude:: ../../../examples/l3fwd/l3fwd_em.c
356 :start-after: Initialize exact match (hash) parameters. 8<
357 :end-before: >8 End of initialization of hash parameters.
362 The FIB object is created and loaded with the pre-configured entries
363 read from a global array.
364 The abridged code snippet below shows the FIB initialization for IPv4,
365 the full setup function including the IPv6 setup can be seen in the app code.
367 .. literalinclude:: ../../../examples/l3fwd/l3fwd_fib.c
369 :start-after: Function to setup fib. 8<
370 :end-before: >8 End of setup fib.
375 For each supported ACL rule format (IPv4 5-tuple, IPv6 6-tuple),
376 the application creates a separate context handler
377 from the ACL library for each CPU socket on the board
378 and adds parsed rules into that context.
380 Note, that for each supported rule type,
381 the application needs to calculate the expected offset of the fields
382 from the start of the packet.
383 That's why only packets with fixed IPv4/ IPv6 header are supported.
384 That allows to perform ACL classify straight over incoming packet buffer -
385 no extra protocol field retrieval need to be performed.
387 Subsequently, the application checks whether NUMA is enabled.
388 If it is, the application records the socket IDs of the CPU cores involved in the task.
390 Finally, the application creates contexts handler from the ACL library,
391 adds rules parsed from the file into the database and build an ACL trie.
392 It is important to note that the application creates an independent copy
393 of each database for each socket CPU involved in the task
394 to reduce the time for remote memory access.
396 .. literalinclude:: ../../../examples/l3fwd/l3fwd_acl.c
398 :start-after: Setup ACL context. 8<
399 :end-before: >8 End of ACL context setup.
401 Packet Forwarding for Hash-based Lookups
402 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
404 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward()
405 or simple_ipv4_fwd_4pkts() function for IPv4 packets or the simple_ipv6_fwd_4pkts() function for IPv6 packets.
406 The l3fwd_simple_forward() function provides the basic functionality for both IPv4 and IPv6 packet forwarding
407 for any number of burst packets received,
408 and the packet forwarding decision (that is, the identification of the output interface for the packet)
409 for hash-based lookups is done by the get_ipv4_dst_port() or get_ipv6_dst_port() function.
410 The get_ipv4_dst_port() function is shown below:
412 .. literalinclude:: ../../../examples/l3fwd/l3fwd_em.c
414 :start-after: Performing hash-based lookups. 8<
415 :end-before: >8 End of performing hash-based lookups.
417 The get_ipv6_dst_port() function is similar to the get_ipv4_dst_port() function.
419 The simple_ipv4_fwd_4pkts() and simple_ipv6_fwd_4pkts() function are optimized for continuous 4 valid ipv4 and ipv6 packets,
420 they leverage the multiple buffer optimization to boost the performance of forwarding packets with the exact match on hash table.
421 The key code snippet of simple_ipv4_fwd_4pkts() is shown below:
426 simple_ipv4_fwd_4pkts(struct rte_mbuf* m[4], uint16_t portid, struct lcore_conf *qconf)
430 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)));
431 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)));
432 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)));
433 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)));
435 key[0].xmm = _mm_and_si128(data[0], mask0);
436 key[1].xmm = _mm_and_si128(data[1], mask0);
437 key[2].xmm = _mm_and_si128(data[2], mask0);
438 key[3].xmm = _mm_and_si128(data[3], mask0);
440 const void *key_array[4] = {&key[0], &key[1], &key[2],&key[3]};
442 rte_hash_lookup_bulk(qconf->ipv4_lookup_struct, &key_array[0], 4, ret);
444 dst_port[0] = (ret[0] < 0)? portid:ipv4_l3fwd_out_if[ret[0]];
445 dst_port[1] = (ret[1] < 0)? portid:ipv4_l3fwd_out_if[ret[1]];
446 dst_port[2] = (ret[2] < 0)? portid:ipv4_l3fwd_out_if[ret[2]];
447 dst_port[3] = (ret[3] < 0)? portid:ipv4_l3fwd_out_if[ret[3]];
452 The simple_ipv6_fwd_4pkts() function is similar to the simple_ipv4_fwd_4pkts() function.
454 Known issue: IP packets with extensions or IP packets which are not TCP/UDP cannot work well at this mode.
456 Packet Forwarding for LPM-based Lookups
457 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
459 For each input packet, the packet forwarding operation is done by the l3fwd_simple_forward() function,
460 but the packet forwarding decision (that is, the identification of the output interface for the packet)
461 for LPM-based lookups is done by the get_ipv4_dst_port() function below:
463 .. literalinclude:: ../../../examples/l3fwd/l3fwd_lpm.c
465 :start-after: Performing LPM-based lookups. 8<
466 :end-before: >8 End of performing LPM-based lookups.
468 Packet Forwarding for FIB-based Lookups
469 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
471 The FIB library was designed to process multiple packets at once,
472 it does not have separate functions for single and bulk lookups.
473 ``rte_fib_lookup_bulk`` is used for IPv4 lookups
474 and ``rte_fib6_lookup_bulk`` for IPv6.
475 Various examples of these functions being used
476 can be found in the sample app code.
478 Eventdev Driver Initialization
479 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
480 Eventdev driver initialization is same as L2 forwarding eventdev application.
481 Refer :doc:`l2_forward_event` for more details.