xref: /dpdk/examples/distributor/main.c (revision 2d0c29a37a9c080c1cccb1ad7941aba2ccf5437e)
1 /* SPDX-License-Identifier: BSD-3-Clause
2  * Copyright(c) 2010-2017 Intel Corporation
3  */
4 
5 #include <stdint.h>
6 #include <inttypes.h>
7 #include <unistd.h>
8 #include <signal.h>
9 #include <getopt.h>
10 
11 #include <rte_eal.h>
12 #include <rte_ethdev.h>
13 #include <rte_cycles.h>
14 #include <rte_malloc.h>
15 #include <rte_debug.h>
16 #include <rte_prefetch.h>
17 #include <rte_distributor.h>
18 #include <rte_pause.h>
19 #include <rte_power.h>
20 
21 #define RX_RING_SIZE 1024
22 #define TX_RING_SIZE 1024
23 #define NUM_MBUFS ((64*1024)-1)
24 #define MBUF_CACHE_SIZE 128
25 #define BURST_SIZE 64
26 #define SCHED_RX_RING_SZ 8192
27 #define SCHED_TX_RING_SZ 65536
28 #define BURST_SIZE_TX 32
29 
30 #define RTE_LOGTYPE_DISTRAPP RTE_LOGTYPE_USER1
31 
32 #define ANSI_COLOR_RED     "\x1b[31m"
33 #define ANSI_COLOR_RESET   "\x1b[0m"
34 
35 /* mask of enabled ports */
36 static uint32_t enabled_port_mask;
37 volatile uint8_t quit_signal;
38 volatile uint8_t quit_signal_rx;
39 volatile uint8_t quit_signal_dist;
40 volatile uint8_t quit_signal_work;
41 unsigned int power_lib_initialised;
42 
43 static volatile struct app_stats {
44 	struct {
45 		uint64_t rx_pkts;
46 		uint64_t returned_pkts;
47 		uint64_t enqueued_pkts;
48 		uint64_t enqdrop_pkts;
49 	} rx __rte_cache_aligned;
50 	int pad1 __rte_cache_aligned;
51 
52 	struct {
53 		uint64_t in_pkts;
54 		uint64_t ret_pkts;
55 		uint64_t sent_pkts;
56 		uint64_t enqdrop_pkts;
57 	} dist __rte_cache_aligned;
58 	int pad2 __rte_cache_aligned;
59 
60 	struct {
61 		uint64_t dequeue_pkts;
62 		uint64_t tx_pkts;
63 		uint64_t enqdrop_pkts;
64 	} tx __rte_cache_aligned;
65 	int pad3 __rte_cache_aligned;
66 
67 	uint64_t worker_pkts[64] __rte_cache_aligned;
68 
69 	int pad4 __rte_cache_aligned;
70 
71 	uint64_t worker_bursts[64][8] __rte_cache_aligned;
72 
73 	int pad5 __rte_cache_aligned;
74 
75 	uint64_t port_rx_pkts[64] __rte_cache_aligned;
76 	uint64_t port_tx_pkts[64] __rte_cache_aligned;
77 } app_stats;
78 
79 struct app_stats prev_app_stats;
80 
81 static const struct rte_eth_conf port_conf_default = {
82 	.rxmode = {
83 		.mq_mode = ETH_MQ_RX_RSS,
84 		.max_rx_pkt_len = ETHER_MAX_LEN,
85 	},
86 	.txmode = {
87 		.mq_mode = ETH_MQ_TX_NONE,
88 	},
89 	.rx_adv_conf = {
90 		.rss_conf = {
91 			.rss_hf = ETH_RSS_IP | ETH_RSS_UDP |
92 				ETH_RSS_TCP | ETH_RSS_SCTP,
93 		}
94 	},
95 };
96 
97 struct output_buffer {
98 	unsigned count;
99 	struct rte_mbuf *mbufs[BURST_SIZE];
100 };
101 
102 static void print_stats(void);
103 
104 /*
105  * Initialises a given port using global settings and with the rx buffers
106  * coming from the mbuf_pool passed as parameter
107  */
108 static inline int
109 port_init(uint16_t port, struct rte_mempool *mbuf_pool)
110 {
111 	struct rte_eth_conf port_conf = port_conf_default;
112 	const uint16_t rxRings = 1, txRings = rte_lcore_count() - 1;
113 	int retval;
114 	uint16_t q;
115 	uint16_t nb_rxd = RX_RING_SIZE;
116 	uint16_t nb_txd = TX_RING_SIZE;
117 	struct rte_eth_dev_info dev_info;
118 	struct rte_eth_txconf txconf;
119 
120 	if (!rte_eth_dev_is_valid_port(port))
121 		return -1;
122 
123 	rte_eth_dev_info_get(port, &dev_info);
124 	if (dev_info.tx_offload_capa & DEV_TX_OFFLOAD_MBUF_FAST_FREE)
125 		port_conf.txmode.offloads |=
126 			DEV_TX_OFFLOAD_MBUF_FAST_FREE;
127 
128 	port_conf.rx_adv_conf.rss_conf.rss_hf &=
129 		dev_info.flow_type_rss_offloads;
130 	if (port_conf.rx_adv_conf.rss_conf.rss_hf !=
131 			port_conf_default.rx_adv_conf.rss_conf.rss_hf) {
132 		printf("Port %u modified RSS hash function based on hardware support,"
133 			"requested:%#"PRIx64" configured:%#"PRIx64"\n",
134 			port,
135 			port_conf_default.rx_adv_conf.rss_conf.rss_hf,
136 			port_conf.rx_adv_conf.rss_conf.rss_hf);
137 	}
138 
139 	retval = rte_eth_dev_configure(port, rxRings, txRings, &port_conf);
140 	if (retval != 0)
141 		return retval;
142 
143 	retval = rte_eth_dev_adjust_nb_rx_tx_desc(port, &nb_rxd, &nb_txd);
144 	if (retval != 0)
145 		return retval;
146 
147 	for (q = 0; q < rxRings; q++) {
148 		retval = rte_eth_rx_queue_setup(port, q, nb_rxd,
149 						rte_eth_dev_socket_id(port),
150 						NULL, mbuf_pool);
151 		if (retval < 0)
152 			return retval;
153 	}
154 
155 	txconf = dev_info.default_txconf;
156 	txconf.offloads = port_conf.txmode.offloads;
157 	for (q = 0; q < txRings; q++) {
158 		retval = rte_eth_tx_queue_setup(port, q, nb_txd,
159 						rte_eth_dev_socket_id(port),
160 						&txconf);
161 		if (retval < 0)
162 			return retval;
163 	}
164 
165 	retval = rte_eth_dev_start(port);
166 	if (retval < 0)
167 		return retval;
168 
169 	struct rte_eth_link link;
170 	rte_eth_link_get_nowait(port, &link);
171 	while (!link.link_status) {
172 		printf("Waiting for Link up on port %"PRIu16"\n", port);
173 		sleep(1);
174 		rte_eth_link_get_nowait(port, &link);
175 	}
176 
177 	if (!link.link_status) {
178 		printf("Link down on port %"PRIu16"\n", port);
179 		return 0;
180 	}
181 
182 	struct ether_addr addr;
183 	rte_eth_macaddr_get(port, &addr);
184 	printf("Port %u MAC: %02"PRIx8" %02"PRIx8" %02"PRIx8
185 			" %02"PRIx8" %02"PRIx8" %02"PRIx8"\n",
186 			port,
187 			addr.addr_bytes[0], addr.addr_bytes[1],
188 			addr.addr_bytes[2], addr.addr_bytes[3],
189 			addr.addr_bytes[4], addr.addr_bytes[5]);
190 
191 	rte_eth_promiscuous_enable(port);
192 
193 	return 0;
194 }
195 
196 struct lcore_params {
197 	unsigned worker_id;
198 	struct rte_distributor *d;
199 	struct rte_ring *rx_dist_ring;
200 	struct rte_ring *dist_tx_ring;
201 	struct rte_mempool *mem_pool;
202 };
203 
204 static int
205 lcore_rx(struct lcore_params *p)
206 {
207 	const uint16_t nb_ports = rte_eth_dev_count_avail();
208 	const int socket_id = rte_socket_id();
209 	uint16_t port;
210 	struct rte_mbuf *bufs[BURST_SIZE*2];
211 
212 	RTE_ETH_FOREACH_DEV(port) {
213 		/* skip ports that are not enabled */
214 		if ((enabled_port_mask & (1 << port)) == 0)
215 			continue;
216 
217 		if (rte_eth_dev_socket_id(port) > 0 &&
218 				rte_eth_dev_socket_id(port) != socket_id)
219 			printf("WARNING, port %u is on remote NUMA node to "
220 					"RX thread.\n\tPerformance will not "
221 					"be optimal.\n", port);
222 	}
223 
224 	printf("\nCore %u doing packet RX.\n", rte_lcore_id());
225 	port = 0;
226 	while (!quit_signal_rx) {
227 
228 		/* skip ports that are not enabled */
229 		if ((enabled_port_mask & (1 << port)) == 0) {
230 			if (++port == nb_ports)
231 				port = 0;
232 			continue;
233 		}
234 		const uint16_t nb_rx = rte_eth_rx_burst(port, 0, bufs,
235 				BURST_SIZE);
236 		if (unlikely(nb_rx == 0)) {
237 			if (++port == nb_ports)
238 				port = 0;
239 			continue;
240 		}
241 		app_stats.rx.rx_pkts += nb_rx;
242 
243 /*
244  * You can run the distributor on the rx core with this code. Returned
245  * packets are then send straight to the tx core.
246  */
247 #if 0
248 	rte_distributor_process(d, bufs, nb_rx);
249 	const uint16_t nb_ret = rte_distributor_returned_pktsd,
250 			bufs, BURST_SIZE*2);
251 
252 		app_stats.rx.returned_pkts += nb_ret;
253 		if (unlikely(nb_ret == 0)) {
254 			if (++port == nb_ports)
255 				port = 0;
256 			continue;
257 		}
258 
259 		struct rte_ring *tx_ring = p->dist_tx_ring;
260 		uint16_t sent = rte_ring_enqueue_burst(tx_ring,
261 				(void *)bufs, nb_ret, NULL);
262 #else
263 		uint16_t nb_ret = nb_rx;
264 		/*
265 		 * Swap the following two lines if you want the rx traffic
266 		 * to go directly to tx, no distribution.
267 		 */
268 		struct rte_ring *out_ring = p->rx_dist_ring;
269 		/* struct rte_ring *out_ring = p->dist_tx_ring; */
270 
271 		uint16_t sent = rte_ring_enqueue_burst(out_ring,
272 				(void *)bufs, nb_ret, NULL);
273 #endif
274 
275 		app_stats.rx.enqueued_pkts += sent;
276 		if (unlikely(sent < nb_ret)) {
277 			app_stats.rx.enqdrop_pkts +=  nb_ret - sent;
278 			RTE_LOG_DP(DEBUG, DISTRAPP,
279 				"%s:Packet loss due to full ring\n", __func__);
280 			while (sent < nb_ret)
281 				rte_pktmbuf_free(bufs[sent++]);
282 		}
283 		if (++port == nb_ports)
284 			port = 0;
285 	}
286 	if (power_lib_initialised)
287 		rte_power_exit(rte_lcore_id());
288 	/* set worker & tx threads quit flag */
289 	printf("\nCore %u exiting rx task.\n", rte_lcore_id());
290 	quit_signal = 1;
291 	return 0;
292 }
293 
294 static inline void
295 flush_one_port(struct output_buffer *outbuf, uint8_t outp)
296 {
297 	unsigned int nb_tx = rte_eth_tx_burst(outp, 0,
298 			outbuf->mbufs, outbuf->count);
299 	app_stats.tx.tx_pkts += outbuf->count;
300 
301 	if (unlikely(nb_tx < outbuf->count)) {
302 		app_stats.tx.enqdrop_pkts +=  outbuf->count - nb_tx;
303 		do {
304 			rte_pktmbuf_free(outbuf->mbufs[nb_tx]);
305 		} while (++nb_tx < outbuf->count);
306 	}
307 	outbuf->count = 0;
308 }
309 
310 static inline void
311 flush_all_ports(struct output_buffer *tx_buffers)
312 {
313 	uint16_t outp;
314 
315 	RTE_ETH_FOREACH_DEV(outp) {
316 		/* skip ports that are not enabled */
317 		if ((enabled_port_mask & (1 << outp)) == 0)
318 			continue;
319 
320 		if (tx_buffers[outp].count == 0)
321 			continue;
322 
323 		flush_one_port(&tx_buffers[outp], outp);
324 	}
325 }
326 
327 
328 
329 static int
330 lcore_distributor(struct lcore_params *p)
331 {
332 	struct rte_ring *in_r = p->rx_dist_ring;
333 	struct rte_ring *out_r = p->dist_tx_ring;
334 	struct rte_mbuf *bufs[BURST_SIZE * 4];
335 	struct rte_distributor *d = p->d;
336 
337 	printf("\nCore %u acting as distributor core.\n", rte_lcore_id());
338 	while (!quit_signal_dist) {
339 		const uint16_t nb_rx = rte_ring_dequeue_burst(in_r,
340 				(void *)bufs, BURST_SIZE*1, NULL);
341 		if (nb_rx) {
342 			app_stats.dist.in_pkts += nb_rx;
343 
344 			/* Distribute the packets */
345 			rte_distributor_process(d, bufs, nb_rx);
346 			/* Handle Returns */
347 			const uint16_t nb_ret =
348 				rte_distributor_returned_pkts(d,
349 					bufs, BURST_SIZE*2);
350 
351 			if (unlikely(nb_ret == 0))
352 				continue;
353 			app_stats.dist.ret_pkts += nb_ret;
354 
355 			uint16_t sent = rte_ring_enqueue_burst(out_r,
356 					(void *)bufs, nb_ret, NULL);
357 			app_stats.dist.sent_pkts += sent;
358 			if (unlikely(sent < nb_ret)) {
359 				app_stats.dist.enqdrop_pkts += nb_ret - sent;
360 				RTE_LOG(DEBUG, DISTRAPP,
361 					"%s:Packet loss due to full out ring\n",
362 					__func__);
363 				while (sent < nb_ret)
364 					rte_pktmbuf_free(bufs[sent++]);
365 			}
366 		}
367 	}
368 	printf("\nCore %u exiting distributor task.\n", rte_lcore_id());
369 	quit_signal_work = 1;
370 	if (power_lib_initialised)
371 		rte_power_exit(rte_lcore_id());
372 	rte_distributor_flush(d);
373 	/* Unblock any returns so workers can exit */
374 	rte_distributor_clear_returns(d);
375 	quit_signal_rx = 1;
376 	return 0;
377 }
378 
379 
380 static int
381 lcore_tx(struct rte_ring *in_r)
382 {
383 	static struct output_buffer tx_buffers[RTE_MAX_ETHPORTS];
384 	const int socket_id = rte_socket_id();
385 	uint16_t port;
386 
387 	RTE_ETH_FOREACH_DEV(port) {
388 		/* skip ports that are not enabled */
389 		if ((enabled_port_mask & (1 << port)) == 0)
390 			continue;
391 
392 		if (rte_eth_dev_socket_id(port) > 0 &&
393 				rte_eth_dev_socket_id(port) != socket_id)
394 			printf("WARNING, port %u is on remote NUMA node to "
395 					"TX thread.\n\tPerformance will not "
396 					"be optimal.\n", port);
397 	}
398 
399 	printf("\nCore %u doing packet TX.\n", rte_lcore_id());
400 	while (!quit_signal) {
401 
402 		RTE_ETH_FOREACH_DEV(port) {
403 			/* skip ports that are not enabled */
404 			if ((enabled_port_mask & (1 << port)) == 0)
405 				continue;
406 
407 			struct rte_mbuf *bufs[BURST_SIZE_TX];
408 			const uint16_t nb_rx = rte_ring_dequeue_burst(in_r,
409 					(void *)bufs, BURST_SIZE_TX, NULL);
410 			app_stats.tx.dequeue_pkts += nb_rx;
411 
412 			/* if we get no traffic, flush anything we have */
413 			if (unlikely(nb_rx == 0)) {
414 				flush_all_ports(tx_buffers);
415 				continue;
416 			}
417 
418 			/* for traffic we receive, queue it up for transmit */
419 			uint16_t i;
420 			rte_prefetch_non_temporal((void *)bufs[0]);
421 			rte_prefetch_non_temporal((void *)bufs[1]);
422 			rte_prefetch_non_temporal((void *)bufs[2]);
423 			for (i = 0; i < nb_rx; i++) {
424 				struct output_buffer *outbuf;
425 				uint8_t outp;
426 				rte_prefetch_non_temporal((void *)bufs[i + 3]);
427 				/*
428 				 * workers should update in_port to hold the
429 				 * output port value
430 				 */
431 				outp = bufs[i]->port;
432 				/* skip ports that are not enabled */
433 				if ((enabled_port_mask & (1 << outp)) == 0)
434 					continue;
435 
436 				outbuf = &tx_buffers[outp];
437 				outbuf->mbufs[outbuf->count++] = bufs[i];
438 				if (outbuf->count == BURST_SIZE_TX)
439 					flush_one_port(outbuf, outp);
440 			}
441 		}
442 	}
443 	if (power_lib_initialised)
444 		rte_power_exit(rte_lcore_id());
445 	printf("\nCore %u exiting tx task.\n", rte_lcore_id());
446 	return 0;
447 }
448 
449 static void
450 int_handler(int sig_num)
451 {
452 	printf("Exiting on signal %d\n", sig_num);
453 	/* set quit flag for rx thread to exit */
454 	quit_signal_dist = 1;
455 }
456 
457 static void
458 print_stats(void)
459 {
460 	struct rte_eth_stats eth_stats;
461 	unsigned int i, j;
462 	const unsigned int num_workers = rte_lcore_count() - 4;
463 
464 	RTE_ETH_FOREACH_DEV(i) {
465 		rte_eth_stats_get(i, &eth_stats);
466 		app_stats.port_rx_pkts[i] = eth_stats.ipackets;
467 		app_stats.port_tx_pkts[i] = eth_stats.opackets;
468 	}
469 
470 	printf("\n\nRX Thread:\n");
471 	RTE_ETH_FOREACH_DEV(i) {
472 		printf("Port %u Pktsin : %5.2f\n", i,
473 				(app_stats.port_rx_pkts[i] -
474 				prev_app_stats.port_rx_pkts[i])/1000000.0);
475 		prev_app_stats.port_rx_pkts[i] = app_stats.port_rx_pkts[i];
476 	}
477 	printf(" - Received:    %5.2f\n",
478 			(app_stats.rx.rx_pkts -
479 			prev_app_stats.rx.rx_pkts)/1000000.0);
480 	printf(" - Returned:    %5.2f\n",
481 			(app_stats.rx.returned_pkts -
482 			prev_app_stats.rx.returned_pkts)/1000000.0);
483 	printf(" - Enqueued:    %5.2f\n",
484 			(app_stats.rx.enqueued_pkts -
485 			prev_app_stats.rx.enqueued_pkts)/1000000.0);
486 	printf(" - Dropped:     %s%5.2f%s\n", ANSI_COLOR_RED,
487 			(app_stats.rx.enqdrop_pkts -
488 			prev_app_stats.rx.enqdrop_pkts)/1000000.0,
489 			ANSI_COLOR_RESET);
490 
491 	printf("Distributor thread:\n");
492 	printf(" - In:          %5.2f\n",
493 			(app_stats.dist.in_pkts -
494 			prev_app_stats.dist.in_pkts)/1000000.0);
495 	printf(" - Returned:    %5.2f\n",
496 			(app_stats.dist.ret_pkts -
497 			prev_app_stats.dist.ret_pkts)/1000000.0);
498 	printf(" - Sent:        %5.2f\n",
499 			(app_stats.dist.sent_pkts -
500 			prev_app_stats.dist.sent_pkts)/1000000.0);
501 	printf(" - Dropped      %s%5.2f%s\n", ANSI_COLOR_RED,
502 			(app_stats.dist.enqdrop_pkts -
503 			prev_app_stats.dist.enqdrop_pkts)/1000000.0,
504 			ANSI_COLOR_RESET);
505 
506 	printf("TX thread:\n");
507 	printf(" - Dequeued:    %5.2f\n",
508 			(app_stats.tx.dequeue_pkts -
509 			prev_app_stats.tx.dequeue_pkts)/1000000.0);
510 	RTE_ETH_FOREACH_DEV(i) {
511 		printf("Port %u Pktsout: %5.2f\n",
512 				i, (app_stats.port_tx_pkts[i] -
513 				prev_app_stats.port_tx_pkts[i])/1000000.0);
514 		prev_app_stats.port_tx_pkts[i] = app_stats.port_tx_pkts[i];
515 	}
516 	printf(" - Transmitted: %5.2f\n",
517 			(app_stats.tx.tx_pkts -
518 			prev_app_stats.tx.tx_pkts)/1000000.0);
519 	printf(" - Dropped:     %s%5.2f%s\n", ANSI_COLOR_RED,
520 			(app_stats.tx.enqdrop_pkts -
521 			prev_app_stats.tx.enqdrop_pkts)/1000000.0,
522 			ANSI_COLOR_RESET);
523 
524 	prev_app_stats.rx.rx_pkts = app_stats.rx.rx_pkts;
525 	prev_app_stats.rx.returned_pkts = app_stats.rx.returned_pkts;
526 	prev_app_stats.rx.enqueued_pkts = app_stats.rx.enqueued_pkts;
527 	prev_app_stats.rx.enqdrop_pkts = app_stats.rx.enqdrop_pkts;
528 	prev_app_stats.dist.in_pkts = app_stats.dist.in_pkts;
529 	prev_app_stats.dist.ret_pkts = app_stats.dist.ret_pkts;
530 	prev_app_stats.dist.sent_pkts = app_stats.dist.sent_pkts;
531 	prev_app_stats.dist.enqdrop_pkts = app_stats.dist.enqdrop_pkts;
532 	prev_app_stats.tx.dequeue_pkts = app_stats.tx.dequeue_pkts;
533 	prev_app_stats.tx.tx_pkts = app_stats.tx.tx_pkts;
534 	prev_app_stats.tx.enqdrop_pkts = app_stats.tx.enqdrop_pkts;
535 
536 	for (i = 0; i < num_workers; i++) {
537 		printf("Worker %02u Pkts: %5.2f. Bursts(1-8): ", i,
538 				(app_stats.worker_pkts[i] -
539 				prev_app_stats.worker_pkts[i])/1000000.0);
540 		for (j = 0; j < 8; j++) {
541 			printf("%"PRIu64" ", app_stats.worker_bursts[i][j]);
542 			app_stats.worker_bursts[i][j] = 0;
543 		}
544 		printf("\n");
545 		prev_app_stats.worker_pkts[i] = app_stats.worker_pkts[i];
546 	}
547 }
548 
549 static int
550 lcore_worker(struct lcore_params *p)
551 {
552 	struct rte_distributor *d = p->d;
553 	const unsigned id = p->worker_id;
554 	unsigned int num = 0;
555 	unsigned int i;
556 
557 	/*
558 	 * for single port, xor_val will be zero so we won't modify the output
559 	 * port, otherwise we send traffic from 0 to 1, 2 to 3, and vice versa
560 	 */
561 	const unsigned xor_val = (rte_eth_dev_count_avail() > 1);
562 	struct rte_mbuf *buf[8] __rte_cache_aligned;
563 
564 	for (i = 0; i < 8; i++)
565 		buf[i] = NULL;
566 
567 	app_stats.worker_pkts[p->worker_id] = 1;
568 
569 	printf("\nCore %u acting as worker core.\n", rte_lcore_id());
570 	while (!quit_signal_work) {
571 		num = rte_distributor_get_pkt(d, id, buf, buf, num);
572 		/* Do a little bit of work for each packet */
573 		for (i = 0; i < num; i++) {
574 			uint64_t t = rte_rdtsc()+100;
575 
576 			while (rte_rdtsc() < t)
577 				rte_pause();
578 			buf[i]->port ^= xor_val;
579 		}
580 
581 		app_stats.worker_pkts[p->worker_id] += num;
582 		if (num > 0)
583 			app_stats.worker_bursts[p->worker_id][num-1]++;
584 	}
585 	if (power_lib_initialised)
586 		rte_power_exit(rte_lcore_id());
587 	rte_free(p);
588 	return 0;
589 }
590 
591 static int
592 init_power_library(void)
593 {
594 	int ret = 0, lcore_id;
595 	RTE_LCORE_FOREACH_SLAVE(lcore_id) {
596 		/* init power management library */
597 		ret = rte_power_init(lcore_id);
598 		if (ret) {
599 			RTE_LOG(ERR, POWER,
600 				"Library initialization failed on core %u\n",
601 				lcore_id);
602 			/*
603 			 * Return on first failure, we'll fall back
604 			 * to non-power operation
605 			 */
606 			return ret;
607 		}
608 	}
609 	return ret;
610 }
611 
612 /* display usage */
613 static void
614 print_usage(const char *prgname)
615 {
616 	printf("%s [EAL options] -- -p PORTMASK\n"
617 			"  -p PORTMASK: hexadecimal bitmask of ports to configure\n",
618 			prgname);
619 }
620 
621 static int
622 parse_portmask(const char *portmask)
623 {
624 	char *end = NULL;
625 	unsigned long pm;
626 
627 	/* parse hexadecimal string */
628 	pm = strtoul(portmask, &end, 16);
629 	if ((portmask[0] == '\0') || (end == NULL) || (*end != '\0'))
630 		return -1;
631 
632 	if (pm == 0)
633 		return -1;
634 
635 	return pm;
636 }
637 
638 /* Parse the argument given in the command line of the application */
639 static int
640 parse_args(int argc, char **argv)
641 {
642 	int opt;
643 	char **argvopt;
644 	int option_index;
645 	char *prgname = argv[0];
646 	static struct option lgopts[] = {
647 		{NULL, 0, 0, 0}
648 	};
649 
650 	argvopt = argv;
651 
652 	while ((opt = getopt_long(argc, argvopt, "p:",
653 			lgopts, &option_index)) != EOF) {
654 
655 		switch (opt) {
656 		/* portmask */
657 		case 'p':
658 			enabled_port_mask = parse_portmask(optarg);
659 			if (enabled_port_mask == 0) {
660 				printf("invalid portmask\n");
661 				print_usage(prgname);
662 				return -1;
663 			}
664 			break;
665 
666 		default:
667 			print_usage(prgname);
668 			return -1;
669 		}
670 	}
671 
672 	if (optind <= 1) {
673 		print_usage(prgname);
674 		return -1;
675 	}
676 
677 	argv[optind-1] = prgname;
678 
679 	optind = 1; /* reset getopt lib */
680 	return 0;
681 }
682 
683 /* Main function, does initialization and calls the per-lcore functions */
684 int
685 main(int argc, char *argv[])
686 {
687 	struct rte_mempool *mbuf_pool;
688 	struct rte_distributor *d;
689 	struct rte_ring *dist_tx_ring;
690 	struct rte_ring *rx_dist_ring;
691 	struct rte_power_core_capabilities lcore_cap;
692 	unsigned int lcore_id, worker_id = 0;
693 	int distr_core_id = -1, rx_core_id = -1, tx_core_id = -1;
694 	unsigned nb_ports;
695 	uint16_t portid;
696 	uint16_t nb_ports_available;
697 	uint64_t t, freq;
698 
699 	/* catch ctrl-c so we can print on exit */
700 	signal(SIGINT, int_handler);
701 
702 	/* init EAL */
703 	int ret = rte_eal_init(argc, argv);
704 	if (ret < 0)
705 		rte_exit(EXIT_FAILURE, "Error with EAL initialization\n");
706 	argc -= ret;
707 	argv += ret;
708 
709 	/* parse application arguments (after the EAL ones) */
710 	ret = parse_args(argc, argv);
711 	if (ret < 0)
712 		rte_exit(EXIT_FAILURE, "Invalid distributor parameters\n");
713 
714 	if (rte_lcore_count() < 5)
715 		rte_exit(EXIT_FAILURE, "Error, This application needs at "
716 				"least 5 logical cores to run:\n"
717 				"1 lcore for stats (can be core 0)\n"
718 				"1 lcore for packet RX\n"
719 				"1 lcore for distribution\n"
720 				"1 lcore for packet TX\n"
721 				"and at least 1 lcore for worker threads\n");
722 
723 	if (init_power_library() == 0)
724 		power_lib_initialised = 1;
725 
726 	nb_ports = rte_eth_dev_count_avail();
727 	if (nb_ports == 0)
728 		rte_exit(EXIT_FAILURE, "Error: no ethernet ports detected\n");
729 	if (nb_ports != 1 && (nb_ports & 1))
730 		rte_exit(EXIT_FAILURE, "Error: number of ports must be even, except "
731 				"when using a single port\n");
732 
733 	mbuf_pool = rte_pktmbuf_pool_create("MBUF_POOL",
734 		NUM_MBUFS * nb_ports, MBUF_CACHE_SIZE, 0,
735 		RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id());
736 	if (mbuf_pool == NULL)
737 		rte_exit(EXIT_FAILURE, "Cannot create mbuf pool\n");
738 	nb_ports_available = nb_ports;
739 
740 	/* initialize all ports */
741 	RTE_ETH_FOREACH_DEV(portid) {
742 		/* skip ports that are not enabled */
743 		if ((enabled_port_mask & (1 << portid)) == 0) {
744 			printf("\nSkipping disabled port %d\n", portid);
745 			nb_ports_available--;
746 			continue;
747 		}
748 		/* init port */
749 		printf("Initializing port %u... done\n", portid);
750 
751 		if (port_init(portid, mbuf_pool) != 0)
752 			rte_exit(EXIT_FAILURE, "Cannot initialize port %u\n",
753 					portid);
754 	}
755 
756 	if (!nb_ports_available) {
757 		rte_exit(EXIT_FAILURE,
758 				"All available ports are disabled. Please set portmask.\n");
759 	}
760 
761 	d = rte_distributor_create("PKT_DIST", rte_socket_id(),
762 			rte_lcore_count() - 4,
763 			RTE_DIST_ALG_BURST);
764 	if (d == NULL)
765 		rte_exit(EXIT_FAILURE, "Cannot create distributor\n");
766 
767 	/*
768 	 * scheduler ring is read by the transmitter core, and written to
769 	 * by scheduler core
770 	 */
771 	dist_tx_ring = rte_ring_create("Output_ring", SCHED_TX_RING_SZ,
772 			rte_socket_id(), RING_F_SC_DEQ | RING_F_SP_ENQ);
773 	if (dist_tx_ring == NULL)
774 		rte_exit(EXIT_FAILURE, "Cannot create output ring\n");
775 
776 	rx_dist_ring = rte_ring_create("Input_ring", SCHED_RX_RING_SZ,
777 			rte_socket_id(), RING_F_SC_DEQ | RING_F_SP_ENQ);
778 	if (rx_dist_ring == NULL)
779 		rte_exit(EXIT_FAILURE, "Cannot create output ring\n");
780 
781 	if (power_lib_initialised) {
782 		/*
783 		 * Here we'll pre-assign lcore ids to the rx, tx and
784 		 * distributor workloads if there's higher frequency
785 		 * on those cores e.g. if Turbo Boost is enabled.
786 		 * It's also worth mentioning that it will assign cores in a
787 		 * specific order, so that if there's less than three
788 		 * available, the higher frequency cores will go to the
789 		 * distributor first, then rx, then tx.
790 		 */
791 		RTE_LCORE_FOREACH_SLAVE(lcore_id) {
792 
793 			rte_power_get_capabilities(lcore_id, &lcore_cap);
794 
795 			if (lcore_cap.priority != 1)
796 				continue;
797 
798 			if (distr_core_id < 0) {
799 				distr_core_id = lcore_id;
800 				printf("Distributor on priority core %d\n",
801 					lcore_id);
802 				continue;
803 			}
804 			if (rx_core_id < 0) {
805 				rx_core_id = lcore_id;
806 				printf("Rx on priority core %d\n",
807 					lcore_id);
808 				continue;
809 			}
810 			if (tx_core_id < 0) {
811 				tx_core_id = lcore_id;
812 				printf("Tx on priority core %d\n",
813 					lcore_id);
814 				continue;
815 			}
816 		}
817 	}
818 
819 	/*
820 	 * If there's any of the key workloads left without an lcore_id
821 	 * after the high performing core assignment above, pre-assign
822 	 * them here.
823 	 */
824 	RTE_LCORE_FOREACH_SLAVE(lcore_id) {
825 		if (lcore_id == (unsigned int)distr_core_id ||
826 				lcore_id == (unsigned int)rx_core_id ||
827 				lcore_id == (unsigned int)tx_core_id)
828 			continue;
829 		if (distr_core_id < 0) {
830 			distr_core_id = lcore_id;
831 			printf("Distributor on core %d\n", lcore_id);
832 			continue;
833 		}
834 		if (rx_core_id < 0) {
835 			rx_core_id = lcore_id;
836 			printf("Rx on core %d\n", lcore_id);
837 			continue;
838 		}
839 		if (tx_core_id < 0) {
840 			tx_core_id = lcore_id;
841 			printf("Tx on core %d\n", lcore_id);
842 			continue;
843 		}
844 	}
845 
846 	printf(" tx id %d, dist id %d, rx id %d\n",
847 			tx_core_id,
848 			distr_core_id,
849 			rx_core_id);
850 
851 	/*
852 	 * Kick off all the worker threads first, avoiding the pre-assigned
853 	 * lcore_ids for tx, rx and distributor workloads.
854 	 */
855 	RTE_LCORE_FOREACH_SLAVE(lcore_id) {
856 		if (lcore_id == (unsigned int)distr_core_id ||
857 				lcore_id == (unsigned int)rx_core_id ||
858 				lcore_id == (unsigned int)tx_core_id)
859 			continue;
860 		printf("Starting thread %d as worker, lcore_id %d\n",
861 				worker_id, lcore_id);
862 		struct lcore_params *p =
863 			rte_malloc(NULL, sizeof(*p), 0);
864 		if (!p)
865 			rte_panic("malloc failure\n");
866 		*p = (struct lcore_params){worker_id++, d, rx_dist_ring,
867 			dist_tx_ring, mbuf_pool};
868 
869 		rte_eal_remote_launch((lcore_function_t *)lcore_worker,
870 				p, lcore_id);
871 	}
872 
873 	/* Start tx core */
874 	rte_eal_remote_launch((lcore_function_t *)lcore_tx,
875 			dist_tx_ring, tx_core_id);
876 
877 	/* Start distributor core */
878 	struct lcore_params *pd =
879 		rte_malloc(NULL, sizeof(*pd), 0);
880 	if (!pd)
881 		rte_panic("malloc failure\n");
882 	*pd = (struct lcore_params){worker_id++, d,
883 		rx_dist_ring, dist_tx_ring, mbuf_pool};
884 	rte_eal_remote_launch(
885 			(lcore_function_t *)lcore_distributor,
886 			pd, distr_core_id);
887 
888 	/* Start rx core */
889 	struct lcore_params *pr =
890 		rte_malloc(NULL, sizeof(*pr), 0);
891 	if (!pr)
892 		rte_panic("malloc failure\n");
893 	*pr = (struct lcore_params){worker_id++, d, rx_dist_ring,
894 		dist_tx_ring, mbuf_pool};
895 	rte_eal_remote_launch((lcore_function_t *)lcore_rx,
896 			pr, rx_core_id);
897 
898 	freq = rte_get_timer_hz();
899 	t = rte_rdtsc() + freq;
900 	while (!quit_signal_dist) {
901 		if (t < rte_rdtsc()) {
902 			print_stats();
903 			t = rte_rdtsc() + freq;
904 		}
905 		usleep(1000);
906 	}
907 
908 	RTE_LCORE_FOREACH_SLAVE(lcore_id) {
909 		if (rte_eal_wait_lcore(lcore_id) < 0)
910 			return -1;
911 	}
912 
913 	print_stats();
914 
915 	rte_free(pd);
916 	rte_free(pr);
917 
918 	return 0;
919 }
920