2 Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11 * Redistributions in binary form must reproduce the above copyright
12 notice, this list of conditions and the following disclaimer in
13 the documentation and/or other materials provided with the
15 * Neither the name of Intel Corporation nor the names of its
16 contributors may be used to endorse or promote products derived
17 from this software without specific prior written permission.
19 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 L2 Forwarding Sample Application (in Real and Virtualized Environments)
32 =======================================================================
34 The L2 Forwarding sample application is a simple example of packet processing using
35 the Intel® Data Plane Development Kit (Intel® DPDK) which
36 also takes advantage of Single Root I/O Virtualization (SR-IOV) features in a virtualized environment.
40 Please note that previously a separate L2 Forwarding in Virtualized Environments sample application was used,
41 however, in later Intel® DPDK versions these sample applications have been merged.
46 The L2 Forwarding sample application, which can operate in real and virtualized environments,
47 performs L2 forwarding for each packet that is received on an RX_PORT.
48 The destination port is the adjacent port from the enabled portmask, that is,
49 if the first four ports are enabled (portmask 0xf),
50 ports 1 and 2 forward into each other, and ports 3 and 4 forward into each other.
51 Also, the MAC addresses are affected as follows:
53 * The source MAC address is replaced by the TX_PORT MAC address
55 * The destination MAC address is replaced by 02:00:00:00:00:TX_PORT_ID
57 This application can be used to benchmark performance using a traffic-generator, as shown in the Figure 3.
59 The application can also be used in a virtualized environment as shown in Figure 4.
61 The L2 Forwarding application can also be used as a starting point for developing a new application based on the Intel® DPDK.
65 **Figure 3. Performance Benchmark Setup (Basic Environment)**
67 .. image4_png has been replaced
69 |l2_fwd_benchmark_setup|
73 **Figure 4. Performance Benchmark Setup (Virtualized Environment)**
75 .. image5_png has been renamed
77 |l2_fwd_virtenv_benchmark_setup|
79 Virtual Function Setup Instructions
80 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
82 This application can use the virtual function available in the system and
83 therefore can be used in a virtual machine without passing through
84 the whole Network Device into a guest machine in a virtualized scenario.
85 The virtual functions can be enabled in the host machine or the hypervisor with the respective physical function driver.
87 For example, in a Linux* host machine, it is possible to enable a virtual function using the following command:
89 .. code-block:: console
91 modprobe ixgbe max_vfs=2,2
93 This command enables two Virtual Functions on each of Physical Function of the NIC,
94 with two physical ports in the PCI configuration space.
95 It is important to note that enabled Virtual Function 0 and 2 would belong to Physical Function 0
96 and Virtual Function 1 and 3 would belong to Physical Function 1,
97 in this case enabling a total of four Virtual Functions.
99 Compiling the Application
100 -------------------------
102 #. Go to the example directory:
104 .. code-block:: console
106 export RTE_SDK=/path/to/rte_sdk cd ${RTE_SDK}/examples/l2fwd
108 #. Set the target (a default target is used if not specified). For example:
110 .. code-block:: console
112 export RTE_TARGET=x86_64-native-linuxapp-gcc
114 *See the Intel® DPDK Getting Started Guide* for possible RTE_TARGET values.
116 #. Build the application:
118 .. code-block:: console
122 Running the Application
123 -----------------------
125 The application requires a number of command line options:
127 .. code-block:: console
129 ./build/l2fwd [EAL options] -- -p PORTMASK [-q NQ]
133 * p PORTMASK: A hexadecimal bitmask of the ports to configure
135 * q NQ: A number of queues (=ports) per lcore (default is 1)
137 To run the application in linuxapp environment with 4 lcores, 16 ports and 8 RX queues per lcore, issue the command:
139 .. code-block:: console
141 $ ./build/l2fwd -c f -n 4 -- -q 8 -p ffff
143 Refer to the *Intel® *DPDK Getting Started Guide* for general information on running applications
144 and the Environment Abstraction Layer (EAL) options.
149 The following sections provide some explanation of the code.
151 Command Line Arguments
152 ~~~~~~~~~~~~~~~~~~~~~~
154 The L2 Forwarding sample application takes specific parameters,
155 in addition to Environment Abstraction Layer (EAL) arguments (see Section 9.3).
156 The preferred way to parse parameters is to use the getopt() function,
157 since it is part of a well-defined and portable library.
159 The parsing of arguments is done in the l2fwd_parse_args() function.
160 The method of argument parsing is not described here.
161 Refer to the *glibc getopt(3)* man page for details.
163 EAL arguments are parsed first, then application-specific arguments.
164 This is done at the beginning of the main() function:
170 ret = rte_eal_init(argc, argv);
172 rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n");
177 /* parse application arguments (after the EAL ones) */
179 ret = l2fwd_parse_args(argc, argv);
181 rte_exit(EXIT_FAILURE, "Invalid L2FWD arguments\n");
183 Mbuf Pool Initialization
184 ~~~~~~~~~~~~~~~~~~~~~~~~
186 Once the arguments are parsed, the mbuf pool is created.
187 The mbuf pool contains a set of mbuf objects that will be used by the driver
188 and the application to store network packet data:
192 /* create the mbuf pool */
194 l2fwd_pktmbuf_pool = rte_mempool_create("mbuf_pool", NB_MBUF, MBUF_SIZE, 32, sizeof(struct rte_pktmbuf_pool_private),
195 rte_pktmbuf_pool_init, NULL, rte_pktmbuf_init, NULL, SOCKET0, 0);
197 if (l2fwd_pktmbuf_pool == NULL)
198 rte_panic("Cannot init mbuf pool\n");
200 The rte_mempool is a generic structure used to handle pools of objects.
201 In this case, it is necessary to create a pool that will be used by the driver,
202 which expects to have some reserved space in the mempool structure,
203 sizeof(struct rte_pktmbuf_pool_private) bytes.
204 The number of allocated pkt mbufs is NB_MBUF, with a size of MBUF_SIZE each.
205 A per-lcore cache of 32 mbufs is kept.
206 The memory is allocated in NUMA socket 0,
207 but it is possible to extend this code to allocate one mbuf pool per socket.
209 Two callback pointers are also given to the rte_mempool_create() function:
211 * The first callback pointer is to rte_pktmbuf_pool_init() and is used
212 to initialize the private data of the mempool, which is needed by the driver.
213 This function is provided by the mbuf API, but can be copied and extended by the developer.
215 * The second callback pointer given to rte_mempool_create() is the mbuf initializer.
216 The default is used, that is, rte_pktmbuf_init(), which is provided in the rte_mbuf library.
217 If a more complex application wants to extend the rte_pktmbuf structure for its own needs,
218 a new function derived from rte_pktmbuf_init( ) can be created.
220 Driver Initialization
221 ~~~~~~~~~~~~~~~~~~~~~
223 The main part of the code in the main() function relates to the initialization of the driver.
224 To fully understand this code, it is recommended to study the chapters that related to the Poll Mode Driver
225 in the *Intel® DPDK Programmer's Guide* - Rel 1.4 EAR and the *Intel® DPDK API Reference*.
229 if (rte_eal_pci_probe() < 0)
230 rte_exit(EXIT_FAILURE, "Cannot probe PCI\n");
232 nb_ports = rte_eth_dev_count();
235 rte_exit(EXIT_FAILURE, "No Ethernet ports - bye\n");
237 if (nb_ports > RTE_MAX_ETHPORTS)
238 nb_ports = RTE_MAX_ETHPORTS;
240 /* reset l2fwd_dst_ports */
242 for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++)
243 l2fwd_dst_ports[portid] = 0;
248 * Each logical core is assigned a dedicated TX queue on each port.
251 for (portid = 0; portid < nb_ports; portid++) {
252 /* skip ports that are not enabled */
254 if ((l2fwd_enabled_port_mask & (1 << portid)) == 0)
257 if (nb_ports_in_mask % 2) {
258 l2fwd_dst_ports[portid] = last_port;
259 l2fwd_dst_ports[last_port] = portid;
266 rte_eth_dev_info_get((uint8_t) portid, &dev_info);
271 * rte_igb_pmd_init_all() simultaneously registers the driver as a PCI driver and as an Ethernet* Poll Mode Driver.
273 * rte_eal_pci_probe() parses the devices on the PCI bus and initializes recognized devices.
275 The next step is to configure the RX and TX queues.
276 For each port, there is only one RX queue (only one lcore is able to poll a given port).
277 The number of TX queues depends on the number of available lcores.
278 The rte_eth_dev_configure() function is used to configure the number of queues for a port:
282 ret = rte_eth_dev_configure((uint8_t)portid, 1, 1, &port_conf);
284 rte_exit(EXIT_FAILURE, "Cannot configure device: "
288 The global configuration is stored in a static structure:
292 static const struct rte_eth_conf port_conf = {
295 .header_split = 0, /**< Header Split disabled */
296 .hw_ip_checksum = 0, /**< IP checksum offload disabled */
297 .hw_vlan_filter = 0, /**< VLAN filtering disabled */
298 .jumbo_frame = 0, /**< Jumbo Frame Support disabled */
299 .hw_strip_crc= 0, /**< CRC stripped by hardware */
303 .mq_mode = ETH_DCB_NONE
307 RX Queue Initialization
308 ~~~~~~~~~~~~~~~~~~~~~~~
310 The application uses one lcore to poll one or several ports, depending on the -q option,
311 which specifies the number of queues per lcore.
313 For example, if the user specifies -q 4, the application is able to poll four ports with one lcore.
314 If there are 16 ports on the target (and if the portmask argument is -p ffff ),
315 the application will need four lcores to poll all the ports.
319 ret = rte_eth_rx_queue_setup((uint8_t) portid, 0, nb_rxd, SOCKET0, &rx_conf, l2fwd_pktmbuf_pool);
322 rte_exit(EXIT_FAILURE, "rte_eth_rx_queue_setup: "
326 The list of queues that must be polled for a given lcore is stored in a private structure called struct lcore_queue_conf.
330 struct lcore_queue_conf {
332 unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE];
333 struct mbuf_table tx_mbufs[L2FWD_MAX_PORTS];
336 struct lcore_queue_conf lcore_queue_conf[RTE_MAX_LCORE];
338 The values n_rx_port and rx_port_list[] are used in the main packet processing loop
339 (see Section 9.4.6 "Receive, Process and Transmit Packets" later in this chapter).
341 The global configuration for the RX queues is stored in a static structure:
345 static const struct rte_eth_rxconf rx_conf = {
347 .pthresh = RX_PTHRESH,
348 .hthresh = RX_HTHRESH,
349 .wthresh = RX_WTHRESH,
353 TX Queue Initialization
354 ~~~~~~~~~~~~~~~~~~~~~~~
356 Each lcore should be able to transmit on any port. For every port, a single TX queue is initialized.
360 /* init one TX queue on each port */
364 ret = rte_eth_tx_queue_setup((uint8_t) portid, 0, nb_txd, rte_eth_dev_socket_id(portid), &tx_conf);
366 rte_exit(EXIT_FAILURE, "rte_eth_tx_queue_setup:err=%d, port=%u\n", ret, (unsigned) portid);
368 The global configuration for TX queues is stored in a static structure:
372 static const struct rte_eth_txconf tx_conf = {
374 .pthresh = TX_PTHRESH,
375 .hthresh = TX_HTHRESH,
376 .wthresh = TX_WTHRESH,
378 .tx_free_thresh = RTE_TEST_TX_DESC_DEFAULT + 1, /* disable feature */
381 Receive, Process and Transmit Packets
382 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
384 In the l2fwd_main_loop() function, the main task is to read ingress packets from the RX queues.
385 This is done using the following code:
390 * Read packet from RX queues
393 for (i = 0; i < qconf->n_rx_port; i++) {
394 portid = qconf->rx_port_list[i];
395 nb_rx = rte_eth_rx_burst((uint8_t) portid, 0, pkts_burst, MAX_PKT_BURST);
397 for (j = 0; j < nb_rx; j++) {
399 rte_prefetch0[rte_pktmbuf_mtod(m, void *)); l2fwd_simple_forward(m, portid);
403 Packets are read in a burst of size MAX_PKT_BURST.
404 The rte_eth_rx_burst() function writes the mbuf pointers in a local table and returns the number of available mbufs in the table.
406 Then, each mbuf in the table is processed by the l2fwd_simple_forward() function.
407 The processing is very simple: process the TX port from the RX port, then replace the source and destination MAC addresses.
411 In the following code, one line for getting the output port requires some explanation.
413 During the initialization process, a static array of destination ports (l2fwd_dst_ports[]) is filled such that for each source port,
414 a destination port is assigned that is either the next or previous enabled port from the portmask.
415 Naturally, the number of ports in the portmask must be even, otherwise, the application exits.
420 l2fwd_simple_forward(struct rte_mbuf *m, unsigned portid)
422 struct ether_hdr *eth;
426 dst_port = l2fwd_dst_ports[portid];
428 eth = rte_pktmbuf_mtod(m, struct ether_hdr *);
430 /* 02:00:00:00:00:xx */
432 tmp = ð->d_addr.addr_bytes[0];
434 *((uint64_t *)tmp) = 0x000000000002 + ((uint64_t) dst_port << 40);
438 ether_addr_copy(&l2fwd_ports_eth_addr[dst_port], ð->s_addr);
440 l2fwd_send_packet(m, (uint8_t) dst_port);
443 Then, the packet is sent using the l2fwd_send_packet (m, dst_port) function.
444 For this test application, the processing is exactly the same for all packets arriving on the same RX port.
445 Therefore, it would have been possible to call the l2fwd_send_burst() function directly from the main loop
446 to send all the received packets on the same TX port,
447 using the burst-oriented send function, which is more efficient.
449 However, in real-life applications (such as, L3 routing),
450 packet N is not necessarily forwarded on the same port as packet N-1.
451 The application is implemented to illustrate that, so the same approach can be reused in a more complex application.
453 The l2fwd_send_packet() function stores the packet in a per-lcore and per-txport table.
454 If the table is full, the whole packets table is transmitted using the l2fwd_send_burst() function:
458 /* Send the packet on an output interface */
461 l2fwd_send_packet(struct rte_mbuf *m, uint8_t port)
463 unsigned lcore_id, len;
464 struct lcore_queue_conf \*qconf;
466 lcore_id = rte_lcore_id();
467 qconf = &lcore_queue_conf[lcore_id];
468 len = qconf->tx_mbufs[port].len;
469 qconf->tx_mbufs[port].m_table[len] = m;
472 /* enough pkts to be sent */
474 if (unlikely(len == MAX_PKT_BURST)) {
475 l2fwd_send_burst(qconf, MAX_PKT_BURST, port);
479 qconf->tx_mbufs[port].len = len; return 0;
482 To ensure that no packets remain in the tables, each lcore does a draining of TX queue in its main loop.
483 This technique introduces some latency when there are not many packets to send,
484 however it improves performance:
488 cur_tsc = rte_rdtsc();
491 * TX burst queue drain
494 diff_tsc = cur_tsc - prev_tsc;
496 if (unlikely(diff_tsc > drain_tsc)) {
497 for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) {
498 if (qconf->tx_mbufs[portid].len == 0)
501 l2fwd_send_burst(&lcore_queue_conf[lcore_id], qconf->tx_mbufs[portid].len, (uint8_t) portid);
503 qconf->tx_mbufs[portid].len = 0;
506 /* if timer is enabled */
508 if (timer_period > 0) {
509 /* advance the timer */
511 timer_tsc += diff_tsc;
513 /* if timer has reached its timeout */
515 if (unlikely(timer_tsc >= (uint64_t) timer_period)) {
516 /* do this only on master core */
518 if (lcore_id == rte_get_master_lcore()) {
521 /* reset the timer */
530 .. |l2_fwd_benchmark_setup| image:: img/l2_fwd_benchmark_setup.svg
532 .. |l2_fwd_virtenv_benchmark_setup| image:: img/l2_fwd_virtenv_benchmark_setup.png