xref: /dpdk/doc/guides/sample_app_ug/l2_forward_event.rst (revision 1f41d98c207aee8982ced709864c96c463d4503a)
1..  SPDX-License-Identifier: BSD-3-Clause
2    Copyright(c) 2010-2014 Intel Corporation.
3
4.. _l2_fwd_event_app:
5
6L2 Forwarding Eventdev Sample Application
7=========================================
8
9The L2 Forwarding eventdev sample application is a simple example of packet
10processing using the Data Plane Development Kit (DPDK) to demonstrate usage of
11poll and event mode packet I/O mechanism.
12
13Overview
14--------
15
16The L2 Forwarding eventdev sample application, performs L2 forwarding for each
17packet that is received on an RX_PORT. The destination port is the adjacent port
18from the enabled portmask, that is, if the first four ports are enabled (portmask=0x0f),
19ports 1 and 2 forward into each other, and ports 3 and 4 forward into each other.
20Also, if MAC addresses updating is enabled, the MAC addresses are affected as follows:
21
22*   The source MAC address is replaced by the TX_PORT MAC address
23
24*   The destination MAC address is replaced by  02:00:00:00:00:TX_PORT_ID
25
26Application receives packets from RX_PORT using below mentioned methods:
27
28*   Poll mode
29
30*   Eventdev mode (default)
31
32This application can be used to benchmark performance using a traffic-generator,
33as shown in the :numref:`figure_l2fwd_event_benchmark_setup`.
34
35.. _figure_l2fwd_event_benchmark_setup:
36
37.. figure:: img/l2_fwd_benchmark_setup.*
38
39   Performance Benchmark Setup (Basic Environment)
40
41Compiling the Application
42-------------------------
43
44To compile the sample application see :doc:`compiling`.
45
46The application is located in the ``l2fwd-event`` sub-directory.
47
48Running the Application
49-----------------------
50
51The application requires a number of command line options:
52
53.. code-block:: console
54
55    ./build/l2fwd-event [EAL options] -- -p PORTMASK [-q NQ] --[no-]mac-updating --mode=MODE --eventq-sched=SCHED_MODE
56
57where,
58
59*   p PORTMASK: A hexadecimal bitmask of the ports to configure
60
61*   q NQ: A number of queues (=ports) per lcore (default is 1)
62
63*   --[no-]mac-updating: Enable or disable MAC addresses updating (enabled by default).
64
65*   --mode=MODE: Packet transfer mode for I/O, poll or eventdev. Eventdev by default.
66
67*   --eventq-sched=SCHED_MODE: Event queue schedule mode, Ordered, Atomic or Parallel. Atomic by default.
68
69Sample usage commands are given below to run the application into different mode:
70
71Poll mode with 4 lcores, 16 ports and 8 RX queues per lcore and MAC address updating enabled,
72issue the command:
73
74.. code-block:: console
75
76    ./build/l2fwd-event -l 0-3 -n 4 -- -q 8 -p ffff --mode=poll
77
78Eventdev mode with 4 lcores, 16 ports , sched method ordered and MAC address updating enabled,
79issue the command:
80
81.. code-block:: console
82
83    ./build/l2fwd-event -l 0-3 -n 4 -- -p ffff --eventq-sched=ordered
84
85or
86
87.. code-block:: console
88
89    ./build/l2fwd-event -l 0-3 -n 4 -- -q 8 -p ffff --mode=eventdev --eventq-sched=ordered
90
91Refer to the *DPDK Getting Started Guide* for general information on running
92applications and the Environment Abstraction Layer (EAL) options.
93
94To run application with S/W scheduler, it uses following DPDK services:
95
96*   Software scheduler
97*   Rx adapter service function
98*   Tx adapter service function
99
100Application needs service cores to run above mentioned services. Service cores
101must be provided as EAL parameters along with the --vdev=event_sw0 to enable S/W
102scheduler. Following is the sample command:
103
104.. code-block:: console
105
106    ./build/l2fwd-event -l 0-7 -s 0-3 -n 4 ---vdev event_sw0 --q 8 -p ffff --mode=eventdev --eventq-sched=ordered
107
108Explanation
109-----------
110
111The following sections provide some explanation of the code.
112
113.. _l2_fwd_event_app_cmd_arguments:
114
115Command Line Arguments
116~~~~~~~~~~~~~~~~~~~~~~
117
118The L2 Forwarding eventdev sample application takes specific parameters,
119in addition to Environment Abstraction Layer (EAL) arguments.
120The preferred way to parse parameters is to use the getopt() function,
121since it is part of a well-defined and portable library.
122
123The parsing of arguments is done in the **l2fwd_parse_args()** function for non
124eventdev parameters and in **parse_eventdev_args()** for eventdev parameters.
125The method of argument parsing is not described here. Refer to the
126*glibc getopt(3)* man page for details.
127
128EAL arguments are parsed first, then application-specific arguments.
129This is done at the beginning of the main() function and eventdev parameters
130are parsed in eventdev_resource_setup() function during eventdev setup:
131
132.. code-block:: c
133
134    /* init EAL */
135
136    ret = rte_eal_init(argc, argv);
137    if (ret < 0)
138        rte_panic("Invalid EAL arguments\n");
139
140    argc -= ret;
141    argv += ret;
142
143    /* parse application arguments (after the EAL ones) */
144
145    ret = l2fwd_parse_args(argc, argv);
146    if (ret < 0)
147        rte_panic("Invalid L2FWD arguments\n");
148    .
149    .
150    .
151
152    /* Parse eventdev command line options */
153    ret = parse_eventdev_args(argc, argv);
154    if (ret < 0)
155        return ret;
156
157
158
159
160.. _l2_fwd_event_app_mbuf_init:
161
162Mbuf Pool Initialization
163~~~~~~~~~~~~~~~~~~~~~~~~
164
165Once the arguments are parsed, the mbuf pool is created.
166The mbuf pool contains a set of mbuf objects that will be used by the driver
167and the application to store network packet data:
168
169.. code-block:: c
170
171    /* create the mbuf pool */
172
173    l2fwd_pktmbuf_pool = rte_pktmbuf_pool_create("mbuf_pool", NB_MBUF,
174                                                 MEMPOOL_CACHE_SIZE, 0,
175                                                 RTE_MBUF_DEFAULT_BUF_SIZE,
176                                                 rte_socket_id());
177    if (l2fwd_pktmbuf_pool == NULL)
178        rte_panic("Cannot init mbuf pool\n");
179
180The rte_mempool is a generic structure used to handle pools of objects.
181In this case, it is necessary to create a pool that will be used by the driver.
182The number of allocated pkt mbufs is NB_MBUF, with a data room size of
183RTE_MBUF_DEFAULT_BUF_SIZE each.
184A per-lcore cache of 32 mbufs is kept.
185The memory is allocated in NUMA socket 0,
186but it is possible to extend this code to allocate one mbuf pool per socket.
187
188The rte_pktmbuf_pool_create() function uses the default mbuf pool and mbuf
189initializers, respectively rte_pktmbuf_pool_init() and rte_pktmbuf_init().
190An advanced application may want to use the mempool API to create the
191mbuf pool with more control.
192
193.. _l2_fwd_event_app_drv_init:
194
195Driver Initialization
196~~~~~~~~~~~~~~~~~~~~~
197
198The main part of the code in the main() function relates to the initialization
199of the driver. To fully understand this code, it is recommended to study the
200chapters that related to the Poll Mode and Event mode Driver in the
201*DPDK Programmer's Guide* - Rel 1.4 EAR and the *DPDK API Reference*.
202
203.. code-block:: c
204
205    if (rte_pci_probe() < 0)
206        rte_panic("Cannot probe PCI\n");
207
208    /* reset l2fwd_dst_ports */
209
210    for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++)
211        l2fwd_dst_ports[portid] = 0;
212
213    last_port = 0;
214
215    /*
216     * Each logical core is assigned a dedicated TX queue on each port.
217     */
218
219    RTE_ETH_FOREACH_DEV(portid) {
220        /* skip ports that are not enabled */
221
222        if ((l2fwd_enabled_port_mask & (1 << portid)) == 0)
223           continue;
224
225        if (nb_ports_in_mask % 2) {
226            l2fwd_dst_ports[portid] = last_port;
227            l2fwd_dst_ports[last_port] = portid;
228        }
229        else
230           last_port = portid;
231
232        nb_ports_in_mask++;
233
234        rte_eth_dev_info_get((uint8_t) portid, &dev_info);
235    }
236
237Observe that:
238
239*   rte_pci_probe() parses the devices on the PCI bus and initializes recognized
240    devices.
241
242The next step is to configure the RX and TX queues. For each port, there is only
243one RX queue (only one lcore is able to poll a given port). The number of TX
244queues depends on the number of available lcores. The rte_eth_dev_configure()
245function is used to configure the number of queues for a port:
246
247.. code-block:: c
248
249    ret = rte_eth_dev_configure((uint8_t)portid, 1, 1, &port_conf);
250    if (ret < 0)
251        rte_panic("Cannot configure device: err=%d, port=%u\n",
252                  ret, portid);
253
254.. _l2_fwd_event_app_rx_init:
255
256RX Queue Initialization
257~~~~~~~~~~~~~~~~~~~~~~~
258
259The application uses one lcore to poll one or several ports, depending on the -q
260option, which specifies the number of queues per lcore.
261
262For example, if the user specifies -q 4, the application is able to poll four
263ports with one lcore. If there are 16 ports on the target (and if the portmask
264argument is -p ffff ), the application will need four lcores to poll all the
265ports.
266
267.. code-block:: c
268
269    ret = rte_eth_rx_queue_setup((uint8_t) portid, 0, nb_rxd, SOCKET0,
270                                 &rx_conf, l2fwd_pktmbuf_pool);
271    if (ret < 0)
272
273        rte_panic("rte_eth_rx_queue_setup: err=%d, port=%u\n",
274                  ret, portid);
275
276The list of queues that must be polled for a given lcore is stored in a private
277structure called struct lcore_queue_conf.
278
279.. code-block:: c
280
281    struct lcore_queue_conf {
282        unsigned n_rx_port;
283        unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE];
284        struct mbuf_table tx_mbufs[L2FWD_MAX_PORTS];
285    } rte_cache_aligned;
286
287    struct lcore_queue_conf lcore_queue_conf[RTE_MAX_LCORE];
288
289The values n_rx_port and rx_port_list[] are used in the main packet processing
290loop (see :ref:`l2_fwd_event_app_rx_tx_packets`).
291
292.. _l2_fwd_event_app_tx_init:
293
294TX Queue Initialization
295~~~~~~~~~~~~~~~~~~~~~~~
296
297Each lcore should be able to transmit on any port. For every port, a single TX
298queue is initialized.
299
300.. code-block:: c
301
302    /* init one TX queue on each port */
303
304    fflush(stdout);
305
306    ret = rte_eth_tx_queue_setup((uint8_t) portid, 0, nb_txd,
307                                 rte_eth_dev_socket_id(portid), &tx_conf);
308    if (ret < 0)
309        rte_panic("rte_eth_tx_queue_setup:err=%d, port=%u\n",
310                  ret, (unsigned) portid);
311
312To configure eventdev support, application setups following components:
313
314*   Event dev
315*   Event queue
316*   Event Port
317*   Rx/Tx adapters
318*   Ethernet ports
319
320.. _l2_fwd_event_app_event_dev_init:
321
322Event device Initialization
323~~~~~~~~~~~~~~~~~~~~~~~~~~~
324Application can use either H/W or S/W based event device scheduler
325implementation and supports single instance of event device. It configures event
326device as per below configuration
327
328.. code-block:: c
329
330   struct rte_event_dev_config event_d_conf = {
331        .nb_event_queues = ethdev_count, /* Dedicated to each Ethernet port */
332        .nb_event_ports = num_workers, /* Dedicated to each lcore */
333        .nb_events_limit  = 4096,
334        .nb_event_queue_flows = 1024,
335        .nb_event_port_dequeue_depth = 128,
336        .nb_event_port_enqueue_depth = 128
337   };
338
339   ret = rte_event_dev_configure(event_d_id, &event_d_conf);
340   if (ret < 0)
341        rte_panic("Error in configuring event device\n");
342
343In case of S/W scheduler, application runs eventdev scheduler service on service
344core. Application retrieves service id and finds the best possible service core to
345run S/W scheduler.
346
347.. code-block:: c
348
349        rte_event_dev_info_get(evt_rsrc->event_d_id, &evdev_info);
350        if (evdev_info.event_dev_cap  & RTE_EVENT_DEV_CAP_DISTRIBUTED_SCHED) {
351                ret = rte_event_dev_service_id_get(evt_rsrc->event_d_id,
352                                &service_id);
353                if (ret != -ESRCH && ret != 0)
354                        rte_panic("Error in starting eventdev service\n");
355                l2fwd_event_service_enable(service_id);
356        }
357
358.. _l2_fwd_app_event_queue_init:
359
360Event queue Initialization
361~~~~~~~~~~~~~~~~~~~~~~~~~~
362Each Ethernet device is assigned a dedicated event queue which will be linked
363to all available event ports i.e. each lcore can dequeue packets from any of the
364Ethernet ports.
365
366.. code-block:: c
367
368   struct rte_event_queue_conf event_q_conf = {
369        .nb_atomic_flows = 1024,
370        .nb_atomic_order_sequences = 1024,
371        .event_queue_cfg = 0,
372        .schedule_type = RTE_SCHED_TYPE_ATOMIC,
373        .priority = RTE_EVENT_DEV_PRIORITY_HIGHEST
374   };
375
376   /* User requested sched mode */
377   event_q_conf.schedule_type = eventq_sched_mode;
378   for (event_q_id = 0; event_q_id < ethdev_count; event_q_id++) {
379        ret = rte_event_queue_setup(event_d_id, event_q_id,
380                                            &event_q_conf);
381        if (ret < 0)
382              rte_panic("Error in configuring event queue\n");
383   }
384
385In case of S/W scheduler, an extra event queue is created which will be used for
386Tx adapter service function for enqueue operation.
387
388.. _l2_fwd_app_event_port_init:
389
390Event port Initialization
391~~~~~~~~~~~~~~~~~~~~~~~~~
392Each worker thread is assigned a dedicated event port for enq/deq operations
393to/from an event device. All event ports are linked with all available event
394queues.
395
396.. code-block:: c
397
398   struct rte_event_port_conf event_p_conf = {
399        .dequeue_depth = 32,
400        .enqueue_depth = 32,
401        .new_event_threshold = 4096
402   };
403
404   for (event_p_id = 0; event_p_id < num_workers; event_p_id++) {
405        ret = rte_event_port_setup(event_d_id, event_p_id,
406                                   &event_p_conf);
407        if (ret < 0)
408              rte_panic("Error in configuring event port %d\n", event_p_id);
409
410        ret = rte_event_port_link(event_d_id, event_p_id, NULL,
411                                  NULL, 0);
412        if (ret < 0)
413              rte_panic("Error in linking event port %d to queue\n",
414                        event_p_id);
415   }
416
417In case of S/W scheduler, an extra event port is created by DPDK library which
418is retrieved  by the application and same will be used by Tx adapter service.
419
420.. code-block:: c
421
422        ret = rte_event_eth_tx_adapter_event_port_get(tx_adptr_id, &tx_port_id);
423        if (ret)
424                rte_panic("Failed to get Tx adapter port id: %d\n", ret);
425
426        ret = rte_event_port_link(event_d_id, tx_port_id,
427                                  &evt_rsrc.evq.event_q_id[
428                                        evt_rsrc.evq.nb_queues - 1],
429                                  NULL, 1);
430        if (ret != 1)
431                rte_panic("Unable to link Tx adapter port to Tx queue:err=%d\n",
432                          ret);
433
434.. _l2_fwd_event_app_adapter_init:
435
436Rx/Tx adapter Initialization
437~~~~~~~~~~~~~~~~~~~~~~~~~~~~
438Each Ethernet port is assigned a dedicated Rx/Tx adapter for H/W scheduler. Each
439Ethernet port's Rx queues are connected to its respective event queue at
440priority 0 via Rx adapter configuration and Ethernet port's tx queues are
441connected via Tx adapter.
442
443.. code-block:: c
444
445	RTE_ETH_FOREACH_DEV(port_id) {
446		if ((rsrc->enabled_port_mask & (1 << port_id)) == 0)
447			continue;
448		ret = rte_event_eth_rx_adapter_create(adapter_id, event_d_id,
449						&evt_rsrc->def_p_conf);
450		if (ret)
451			rte_panic("Failed to create rx adapter[%d]\n",
452                                  adapter_id);
453
454		/* Configure user requested sched type*/
455		eth_q_conf.ev.sched_type = rsrc->sched_type;
456		eth_q_conf.ev.queue_id = evt_rsrc->evq.event_q_id[q_id];
457		ret = rte_event_eth_rx_adapter_queue_add(adapter_id, port_id,
458							 -1, &eth_q_conf);
459		if (ret)
460			rte_panic("Failed to add queues to Rx adapter\n");
461
462		ret = rte_event_eth_rx_adapter_start(adapter_id);
463		if (ret)
464			rte_panic("Rx adapter[%d] start Failed\n", adapter_id);
465
466		evt_rsrc->rx_adptr.rx_adptr[adapter_id] = adapter_id;
467		adapter_id++;
468		if (q_id < evt_rsrc->evq.nb_queues)
469			q_id++;
470	}
471
472	adapter_id = 0;
473	RTE_ETH_FOREACH_DEV(port_id) {
474		if ((rsrc->enabled_port_mask & (1 << port_id)) == 0)
475			continue;
476		ret = rte_event_eth_tx_adapter_create(adapter_id, event_d_id,
477						&evt_rsrc->def_p_conf);
478		if (ret)
479			rte_panic("Failed to create tx adapter[%d]\n",
480                                  adapter_id);
481
482		ret = rte_event_eth_tx_adapter_queue_add(adapter_id, port_id,
483							 -1);
484		if (ret)
485			rte_panic("Failed to add queues to Tx adapter\n");
486
487		ret = rte_event_eth_tx_adapter_start(adapter_id);
488		if (ret)
489			rte_panic("Tx adapter[%d] start Failed\n", adapter_id);
490
491		evt_rsrc->tx_adptr.tx_adptr[adapter_id] = adapter_id;
492		adapter_id++;
493	}
494
495For S/W scheduler instead of dedicated adapters, common Rx/Tx adapters are
496configured which will be shared among all the Ethernet ports. Also DPDK library
497need service cores to run internal services for Rx/Tx adapters. Application gets
498service id for Rx/Tx adapters and after successful setup it runs the services
499on dedicated service cores.
500
501.. code-block:: c
502
503	for (i = 0; i < evt_rsrc->rx_adptr.nb_rx_adptr; i++) {
504		ret = rte_event_eth_rx_adapter_caps_get(evt_rsrc->event_d_id,
505				evt_rsrc->rx_adptr.rx_adptr[i], &caps);
506		if (ret < 0)
507			rte_panic("Failed to get Rx adapter[%d] caps\n",
508                                  evt_rsrc->rx_adptr.rx_adptr[i]);
509		ret = rte_event_eth_rx_adapter_service_id_get(
510                                                evt_rsrc->event_d_id,
511                                                &service_id);
512		if (ret != -ESRCH && ret != 0)
513			rte_panic("Error in starting Rx adapter[%d] service\n",
514                                  evt_rsrc->rx_adptr.rx_adptr[i]);
515		l2fwd_event_service_enable(service_id);
516	}
517
518	for (i = 0; i < evt_rsrc->tx_adptr.nb_tx_adptr; i++) {
519		ret = rte_event_eth_tx_adapter_caps_get(evt_rsrc->event_d_id,
520				evt_rsrc->tx_adptr.tx_adptr[i], &caps);
521		if (ret < 0)
522			rte_panic("Failed to get Rx adapter[%d] caps\n",
523                                  evt_rsrc->tx_adptr.tx_adptr[i]);
524		ret = rte_event_eth_tx_adapter_service_id_get(
525				evt_rsrc->event_d_id,
526				&service_id);
527		if (ret != -ESRCH && ret != 0)
528			rte_panic("Error in starting Rx adapter[%d] service\n",
529                                  evt_rsrc->tx_adptr.tx_adptr[i]);
530		l2fwd_event_service_enable(service_id);
531	}
532
533.. _l2_fwd_event_app_rx_tx_packets:
534
535Receive, Process and Transmit Packets
536~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
537
538In the **l2fwd_main_loop()** function, the main task is to read ingress packets from
539the RX queues. This is done using the following code:
540
541.. code-block:: c
542
543    /*
544     * Read packet from RX queues
545     */
546
547    for (i = 0; i < qconf->n_rx_port; i++) {
548        portid = qconf->rx_port_list[i];
549        nb_rx = rte_eth_rx_burst((uint8_t) portid, 0,  pkts_burst,
550                                 MAX_PKT_BURST);
551
552        for (j = 0; j < nb_rx; j++) {
553            m = pkts_burst[j];
554            rte_prefetch0(rte_pktmbuf_mtod(m, void *));
555            l2fwd_simple_forward(m, portid);
556        }
557    }
558
559Packets are read in a burst of size MAX_PKT_BURST. The rte_eth_rx_burst()
560function writes the mbuf pointers in a local table and returns the number of
561available mbufs in the table.
562
563Then, each mbuf in the table is processed by the l2fwd_simple_forward()
564function. The processing is very simple: process the TX port from the RX port,
565then replace the source and destination MAC addresses if MAC addresses updating
566is enabled.
567
568During the initialization process, a static array of destination ports
569(l2fwd_dst_ports[]) is filled such that for each source port, a destination port
570is assigned that is either the next or previous enabled port from the portmask.
571If number of ports are odd in portmask then packet from last port will be
572forwarded to first port i.e. if portmask=0x07, then forwarding will take place
573like p0--->p1, p1--->p2, p2--->p0.
574
575Also to optimize enqueue operation, l2fwd_simple_forward() stores incoming mbufs
576up to MAX_PKT_BURST. Once it reaches up to limit, all packets are transmitted to
577destination ports.
578
579.. code-block:: c
580
581   static void
582   l2fwd_simple_forward(struct rte_mbuf *m, uint32_t portid)
583   {
584       uint32_t dst_port;
585       int32_t sent;
586       struct rte_eth_dev_tx_buffer *buffer;
587
588       dst_port = l2fwd_dst_ports[portid];
589
590       if (mac_updating)
591           l2fwd_mac_updating(m, dst_port);
592
593       buffer = tx_buffer[dst_port];
594       sent = rte_eth_tx_buffer(dst_port, 0, buffer, m);
595       if (sent)
596       port_statistics[dst_port].tx += sent;
597   }
598
599For this test application, the processing is exactly the same for all packets
600arriving on the same RX port. Therefore, it would have been possible to call
601the rte_eth_tx_buffer() function directly from the main loop to send all the
602received packets on the same TX port, using the burst-oriented send function,
603which is more efficient.
604
605However, in real-life applications (such as, L3 routing),
606packet N is not necessarily forwarded on the same port as packet N-1.
607The application is implemented to illustrate that, so the same approach can be
608reused in a more complex application.
609
610To ensure that no packets remain in the tables, each lcore does a draining of TX
611queue in its main loop. This technique introduces some latency when there are
612not many packets to send, however it improves performance:
613
614.. code-block:: c
615
616        cur_tsc = rte_rdtsc();
617
618        /*
619        * TX burst queue drain
620        */
621        diff_tsc = cur_tsc - prev_tsc;
622        if (unlikely(diff_tsc > drain_tsc)) {
623                for (i = 0; i < qconf->n_rx_port; i++) {
624                        portid = l2fwd_dst_ports[qconf->rx_port_list[i]];
625                        buffer = tx_buffer[portid];
626                        sent = rte_eth_tx_buffer_flush(portid, 0,
627                                                       buffer);
628                        if (sent)
629                                port_statistics[portid].tx += sent;
630                }
631
632                /* if timer is enabled */
633                if (timer_period > 0) {
634                        /* advance the timer */
635                        timer_tsc += diff_tsc;
636
637                        /* if timer has reached its timeout */
638                        if (unlikely(timer_tsc >= timer_period)) {
639                                /* do this only on master core */
640                                if (lcore_id == rte_get_master_lcore()) {
641                                        print_stats();
642                                        /* reset the timer */
643                                        timer_tsc = 0;
644                                }
645                        }
646                }
647
648                prev_tsc = cur_tsc;
649        }
650
651In the **l2fwd_event_loop()** function, the main task is to read ingress
652packets from the event ports. This is done using the following code:
653
654.. code-block:: c
655
656        /* Read packet from eventdev */
657        nb_rx = rte_event_dequeue_burst(event_d_id, event_p_id,
658                                        events, deq_len, 0);
659        if (nb_rx == 0) {
660                rte_pause();
661                continue;
662        }
663
664        for (i = 0; i < nb_rx; i++) {
665                mbuf[i] = events[i].mbuf;
666                rte_prefetch0(rte_pktmbuf_mtod(mbuf[i], void *));
667        }
668
669
670Before reading packets, deq_len is fetched to ensure correct allowed deq length
671by the eventdev.
672The rte_event_dequeue_burst() function writes the mbuf pointers in a local table
673and returns the number of available mbufs in the table.
674
675Then, each mbuf in the table is processed by the l2fwd_eventdev_forward()
676function. The processing is very simple: process the TX port from the RX port,
677then replace the source and destination MAC addresses if MAC addresses updating
678is enabled.
679
680During the initialization process, a static array of destination ports
681(l2fwd_dst_ports[]) is filled such that for each source port, a destination port
682is assigned that is either the next or previous enabled port from the portmask.
683If number of ports are odd in portmask then packet from last port will be
684forwarded to first port i.e. if portmask=0x07, then forwarding will take place
685like p0--->p1, p1--->p2, p2--->p0.
686
687l2fwd_eventdev_forward() does not stores incoming mbufs. Packet will forwarded
688be to destination ports via Tx adapter or generic event dev enqueue API
689depending H/W or S/W scheduler is used.
690
691.. code-block:: c
692
693	nb_tx = rte_event_eth_tx_adapter_enqueue(event_d_id, port_id, ev,
694						 nb_rx);
695	while (nb_tx < nb_rx && !rsrc->force_quit)
696		nb_tx += rte_event_eth_tx_adapter_enqueue(
697				event_d_id, port_id,
698				ev + nb_tx, nb_rx - nb_tx);
699