1 /* SPDX-License-Identifier: BSD-3-Clause
2 * Copyright(c) 2010-2017 Intel Corporation
11 * RTE Ethernet Device API
13 * The Ethernet Device API is composed of two parts:
15 * - The application-oriented Ethernet API that includes functions to setup
16 * an Ethernet device (configure it, setup its RX and TX queues and start it),
17 * to get its MAC address, the speed and the status of its physical link,
18 * to receive and to transmit packets, and so on.
20 * - The driver-oriented Ethernet API that exports functions allowing
21 * an Ethernet Poll Mode Driver (PMD) to allocate an Ethernet device instance,
22 * create memzone for HW rings and process registered callbacks, and so on.
23 * PMDs should include rte_ethdev_driver.h instead of this header.
25 * By default, all the functions of the Ethernet Device API exported by a PMD
26 * are lock-free functions which assume to not be invoked in parallel on
27 * different logical cores to work on the same target object. For instance,
28 * the receive function of a PMD cannot be invoked in parallel on two logical
29 * cores to poll the same RX queue [of the same port]. Of course, this function
30 * can be invoked in parallel by different logical cores on different RX queues.
31 * It is the responsibility of the upper level application to enforce this rule.
33 * If needed, parallel accesses by multiple logical cores to shared queues
34 * shall be explicitly protected by dedicated inline lock-aware functions
35 * built on top of their corresponding lock-free functions of the PMD API.
37 * In all functions of the Ethernet API, the Ethernet device is
38 * designated by an integer >= 0 named the device port identifier.
40 * At the Ethernet driver level, Ethernet devices are represented by a generic
41 * data structure of type *rte_eth_dev*.
43 * Ethernet devices are dynamically registered during the PCI probing phase
44 * performed at EAL initialization time.
45 * When an Ethernet device is being probed, an *rte_eth_dev* structure and
46 * a new port identifier are allocated for that device. Then, the eth_dev_init()
47 * function supplied by the Ethernet driver matching the probed PCI
48 * device is invoked to properly initialize the device.
50 * The role of the device init function consists of resetting the hardware,
51 * checking access to Non-volatile Memory (NVM), reading the MAC address
54 * If the device init operation is successful, the correspondence between
55 * the port identifier assigned to the new device and its associated
56 * *rte_eth_dev* structure is effectively registered.
57 * Otherwise, both the *rte_eth_dev* structure and the port identifier are
60 * The functions exported by the application Ethernet API to setup a device
61 * designated by its port identifier must be invoked in the following order:
62 * - rte_eth_dev_configure()
63 * - rte_eth_tx_queue_setup()
64 * - rte_eth_rx_queue_setup()
65 * - rte_eth_dev_start()
67 * Then, the network application can invoke, in any order, the functions
68 * exported by the Ethernet API to get the MAC address of a given device, to
69 * get the speed and the status of a device physical link, to receive/transmit
70 * [burst of] packets, and so on.
72 * If the application wants to change the configuration (i.e. call
73 * rte_eth_dev_configure(), rte_eth_tx_queue_setup(), or
74 * rte_eth_rx_queue_setup()), it must call rte_eth_dev_stop() first to stop the
75 * device and then do the reconfiguration before calling rte_eth_dev_start()
76 * again. The transmit and receive functions should not be invoked when the
79 * Please note that some configuration is not stored between calls to
80 * rte_eth_dev_stop()/rte_eth_dev_start(). The following configuration will
84 * - flow control settings
85 * - receive mode configuration (promiscuous mode, all-multicast mode,
86 * hardware checksum mode, RSS/VMDQ settings etc.)
87 * - VLAN filtering configuration
88 * - default MAC address
89 * - MAC addresses supplied to MAC address array
90 * - flow director filtering mode (but not filtering rules)
91 * - NIC queue statistics mappings
93 * Any other configuration will not be stored and will need to be re-entered
94 * before a call to rte_eth_dev_start().
96 * Finally, a network application can close an Ethernet device by invoking the
97 * rte_eth_dev_close() function.
99 * Each function of the application Ethernet API invokes a specific function
100 * of the PMD that controls the target device designated by its port
102 * For this purpose, all device-specific functions of an Ethernet driver are
103 * supplied through a set of pointers contained in a generic structure of type
105 * The address of the *eth_dev_ops* structure is stored in the *rte_eth_dev*
106 * structure by the device init function of the Ethernet driver, which is
107 * invoked during the PCI probing phase, as explained earlier.
109 * In other words, each function of the Ethernet API simply retrieves the
110 * *rte_eth_dev* structure associated with the device port identifier and
111 * performs an indirect invocation of the corresponding driver function
112 * supplied in the *eth_dev_ops* structure of the *rte_eth_dev* structure.
114 * For performance reasons, the address of the burst-oriented RX and TX
115 * functions of the Ethernet driver are not contained in the *eth_dev_ops*
116 * structure. Instead, they are directly stored at the beginning of the
117 * *rte_eth_dev* structure to avoid an extra indirect memory access during
120 * RTE ethernet device drivers do not use interrupts for transmitting or
121 * receiving. Instead, Ethernet drivers export Poll-Mode receive and transmit
122 * functions to applications.
123 * Both receive and transmit functions are packet-burst oriented to minimize
124 * their cost per packet through the following optimizations:
126 * - Sharing among multiple packets the incompressible cost of the
127 * invocation of receive/transmit functions.
129 * - Enabling receive/transmit functions to take advantage of burst-oriented
130 * hardware features (L1 cache, prefetch instructions, NIC head/tail
131 * registers) to minimize the number of CPU cycles per packet, for instance,
132 * by avoiding useless read memory accesses to ring descriptors, or by
133 * systematically using arrays of pointers that exactly fit L1 cache line
134 * boundaries and sizes.
136 * The burst-oriented receive function does not provide any error notification,
137 * to avoid the corresponding overhead. As a hint, the upper-level application
138 * might check the status of the device link once being systematically returned
139 * a 0 value by the receive function of the driver for a given number of tries.
148 /* Use this macro to check if LRO API is supported */
149 #define RTE_ETHDEV_HAS_LRO_SUPPORT
151 #include <rte_compat.h>
153 #include <rte_interrupts.h>
155 #include <rte_devargs.h>
156 #include <rte_errno.h>
157 #include <rte_common.h>
158 #include <rte_config.h>
159 #include <rte_ether.h>
161 #include "rte_ethdev_trace_fp.h"
162 #include "rte_dev_info.h"
164 extern int rte_eth_dev_logtype;
166 #define RTE_ETHDEV_LOG(level, ...) \
167 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__)
172 * Initializes a device iterator.
174 * This iterator allows accessing a list of devices matching some devargs.
177 * Device iterator handle initialized by the function.
178 * The fields bus_str and cls_str might be dynamically allocated,
179 * and could be freed by calling rte_eth_iterator_cleanup().
182 * Device description string.
185 * 0 on successful initialization, negative otherwise.
187 int rte_eth_iterator_init(struct rte_dev_iterator *iter, const char *devargs);
190 * Iterates on devices with devargs filter.
191 * The ownership is not checked.
193 * The next port id is returned, and the iterator is updated.
196 * Device iterator handle initialized by rte_eth_iterator_init().
197 * Some fields bus_str and cls_str might be freed when no more port is found,
198 * by calling rte_eth_iterator_cleanup().
201 * A port id if found, RTE_MAX_ETHPORTS otherwise.
203 uint16_t rte_eth_iterator_next(struct rte_dev_iterator *iter);
206 * Free some allocated fields of the iterator.
208 * This function is automatically called by rte_eth_iterator_next()
209 * on the last iteration (i.e. when no more matching port is found).
211 * It is safe to call this function twice; it will do nothing more.
214 * Device iterator handle initialized by rte_eth_iterator_init().
215 * The fields bus_str and cls_str are freed if needed.
217 void rte_eth_iterator_cleanup(struct rte_dev_iterator *iter);
220 * Macro to iterate over all ethdev ports matching some devargs.
222 * If a break is done before the end of the loop,
223 * the function rte_eth_iterator_cleanup() must be called.
226 * Iterated port id of type uint16_t.
228 * Device parameters input as string of type char*.
230 * Iterator handle of type struct rte_dev_iterator, used internally.
232 #define RTE_ETH_FOREACH_MATCHING_DEV(id, devargs, iter) \
233 for (rte_eth_iterator_init(iter, devargs), \
234 id = rte_eth_iterator_next(iter); \
235 id != RTE_MAX_ETHPORTS; \
236 id = rte_eth_iterator_next(iter))
239 * A structure used to retrieve statistics for an Ethernet port.
240 * Not all statistics fields in struct rte_eth_stats are supported
241 * by any type of network interface card (NIC). If any statistics
242 * field is not supported, its value is 0.
244 struct rte_eth_stats {
245 uint64_t ipackets; /**< Total number of successfully received packets. */
246 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
247 uint64_t ibytes; /**< Total number of successfully received bytes. */
248 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
250 /**< Total of RX packets dropped by the HW,
251 * because there are no available buffer (i.e. RX queues are full).
253 uint64_t ierrors; /**< Total number of erroneous received packets. */
254 uint64_t oerrors; /**< Total number of failed transmitted packets. */
255 uint64_t rx_nombuf; /**< Total number of RX mbuf allocation failures. */
256 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
257 /**< Total number of queue RX packets. */
258 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
259 /**< Total number of queue TX packets. */
260 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
261 /**< Total number of successfully received queue bytes. */
262 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
263 /**< Total number of successfully transmitted queue bytes. */
264 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS];
265 /**< Total number of queue packets received that are dropped. */
269 * Device supported speeds bitmap flags
271 #define ETH_LINK_SPEED_AUTONEG (0 << 0) /**< Autonegotiate (all speeds) */
272 #define ETH_LINK_SPEED_FIXED (1 << 0) /**< Disable autoneg (fixed speed) */
273 #define ETH_LINK_SPEED_10M_HD (1 << 1) /**< 10 Mbps half-duplex */
274 #define ETH_LINK_SPEED_10M (1 << 2) /**< 10 Mbps full-duplex */
275 #define ETH_LINK_SPEED_100M_HD (1 << 3) /**< 100 Mbps half-duplex */
276 #define ETH_LINK_SPEED_100M (1 << 4) /**< 100 Mbps full-duplex */
277 #define ETH_LINK_SPEED_1G (1 << 5) /**< 1 Gbps */
278 #define ETH_LINK_SPEED_2_5G (1 << 6) /**< 2.5 Gbps */
279 #define ETH_LINK_SPEED_5G (1 << 7) /**< 5 Gbps */
280 #define ETH_LINK_SPEED_10G (1 << 8) /**< 10 Gbps */
281 #define ETH_LINK_SPEED_20G (1 << 9) /**< 20 Gbps */
282 #define ETH_LINK_SPEED_25G (1 << 10) /**< 25 Gbps */
283 #define ETH_LINK_SPEED_40G (1 << 11) /**< 40 Gbps */
284 #define ETH_LINK_SPEED_50G (1 << 12) /**< 50 Gbps */
285 #define ETH_LINK_SPEED_56G (1 << 13) /**< 56 Gbps */
286 #define ETH_LINK_SPEED_100G (1 << 14) /**< 100 Gbps */
289 * Ethernet numeric link speeds in Mbps
291 #define ETH_SPEED_NUM_NONE 0 /**< Not defined */
292 #define ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
293 #define ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
294 #define ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
295 #define ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
296 #define ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
297 #define ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
298 #define ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
299 #define ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
300 #define ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
301 #define ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
302 #define ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
303 #define ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
306 * A structure used to retrieve link-level information of an Ethernet port.
309 struct rte_eth_link {
310 uint32_t link_speed; /**< ETH_SPEED_NUM_ */
311 uint16_t link_duplex : 1; /**< ETH_LINK_[HALF/FULL]_DUPLEX */
312 uint16_t link_autoneg : 1; /**< ETH_LINK_[AUTONEG/FIXED] */
313 uint16_t link_status : 1; /**< ETH_LINK_[DOWN/UP] */
314 } __rte_aligned(8); /**< aligned for atomic64 read/write */
316 /* Utility constants */
317 #define ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
318 #define ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
319 #define ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
320 #define ETH_LINK_UP 1 /**< Link is up (see link_status). */
321 #define ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
322 #define ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
325 * A structure used to configure the ring threshold registers of an RX/TX
326 * queue for an Ethernet port.
328 struct rte_eth_thresh {
329 uint8_t pthresh; /**< Ring prefetch threshold. */
330 uint8_t hthresh; /**< Ring host threshold. */
331 uint8_t wthresh; /**< Ring writeback threshold. */
335 * Simple flags are used for rte_eth_conf.rxmode.mq_mode.
337 #define ETH_MQ_RX_RSS_FLAG 0x1
338 #define ETH_MQ_RX_DCB_FLAG 0x2
339 #define ETH_MQ_RX_VMDQ_FLAG 0x4
342 * A set of values to identify what method is to be used to route
343 * packets to multiple queues.
345 enum rte_eth_rx_mq_mode {
346 /** None of DCB,RSS or VMDQ mode */
349 /** For RX side, only RSS is on */
350 ETH_MQ_RX_RSS = ETH_MQ_RX_RSS_FLAG,
351 /** For RX side,only DCB is on. */
352 ETH_MQ_RX_DCB = ETH_MQ_RX_DCB_FLAG,
353 /** Both DCB and RSS enable */
354 ETH_MQ_RX_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG,
356 /** Only VMDQ, no RSS nor DCB */
357 ETH_MQ_RX_VMDQ_ONLY = ETH_MQ_RX_VMDQ_FLAG,
358 /** RSS mode with VMDQ */
359 ETH_MQ_RX_VMDQ_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_VMDQ_FLAG,
360 /** Use VMDQ+DCB to route traffic to queues */
361 ETH_MQ_RX_VMDQ_DCB = ETH_MQ_RX_VMDQ_FLAG | ETH_MQ_RX_DCB_FLAG,
362 /** Enable both VMDQ and DCB in VMDq */
363 ETH_MQ_RX_VMDQ_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG |
368 * for rx mq mode backward compatible
370 #define ETH_RSS ETH_MQ_RX_RSS
371 #define VMDQ_DCB ETH_MQ_RX_VMDQ_DCB
372 #define ETH_DCB_RX ETH_MQ_RX_DCB
375 * A set of values to identify what method is to be used to transmit
376 * packets using multi-TCs.
378 enum rte_eth_tx_mq_mode {
379 ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
380 ETH_MQ_TX_DCB, /**< For TX side,only DCB is on. */
381 ETH_MQ_TX_VMDQ_DCB, /**< For TX side,both DCB and VT is on. */
382 ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
386 * for tx mq mode backward compatible
388 #define ETH_DCB_NONE ETH_MQ_TX_NONE
389 #define ETH_VMDQ_DCB_TX ETH_MQ_TX_VMDQ_DCB
390 #define ETH_DCB_TX ETH_MQ_TX_DCB
393 * A structure used to configure the RX features of an Ethernet port.
395 struct rte_eth_rxmode {
396 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
397 enum rte_eth_rx_mq_mode mq_mode;
398 uint32_t max_rx_pkt_len; /**< Only used if JUMBO_FRAME enabled. */
399 /** Maximum allowed size of LRO aggregated packet. */
400 uint32_t max_lro_pkt_size;
401 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
403 * Per-port Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
404 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
405 * structure are allowed to be set.
409 uint64_t reserved_64s[2]; /**< Reserved for future fields */
410 void *reserved_ptrs[2]; /**< Reserved for future fields */
414 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
415 * Note that single VLAN is treated the same as inner VLAN.
418 ETH_VLAN_TYPE_UNKNOWN = 0,
419 ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
420 ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
425 * A structure used to describe a vlan filter.
426 * If the bit corresponding to a VID is set, such VID is on.
428 struct rte_vlan_filter_conf {
433 * A structure used to configure the Receive Side Scaling (RSS) feature
434 * of an Ethernet port.
435 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
436 * to an array holding the RSS key to use for hashing specific header
437 * fields of received packets. The length of this array should be indicated
438 * by *rss_key_len* below. Otherwise, a default random hash key is used by
441 * The *rss_key_len* field of the *rss_conf* structure indicates the length
442 * in bytes of the array pointed by *rss_key*. To be compatible, this length
443 * will be checked in i40e only. Others assume 40 bytes to be used as before.
445 * The *rss_hf* field of the *rss_conf* structure indicates the different
446 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
447 * Supplying an *rss_hf* equal to zero disables the RSS feature.
449 struct rte_eth_rss_conf {
450 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
451 uint8_t rss_key_len; /**< hash key length in bytes. */
452 uint64_t rss_hf; /**< Hash functions to apply - see below. */
456 * A packet can be identified by hardware as different flow types. Different
457 * NIC hardware may support different flow types.
458 * Basically, the NIC hardware identifies the flow type as deep protocol as
459 * possible, and exclusively. For example, if a packet is identified as
460 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types,
461 * though it is an actual IPV4 packet.
463 #define RTE_ETH_FLOW_UNKNOWN 0
464 #define RTE_ETH_FLOW_RAW 1
465 #define RTE_ETH_FLOW_IPV4 2
466 #define RTE_ETH_FLOW_FRAG_IPV4 3
467 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4
468 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5
469 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6
470 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7
471 #define RTE_ETH_FLOW_IPV6 8
472 #define RTE_ETH_FLOW_FRAG_IPV6 9
473 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10
474 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11
475 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12
476 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13
477 #define RTE_ETH_FLOW_L2_PAYLOAD 14
478 #define RTE_ETH_FLOW_IPV6_EX 15
479 #define RTE_ETH_FLOW_IPV6_TCP_EX 16
480 #define RTE_ETH_FLOW_IPV6_UDP_EX 17
481 #define RTE_ETH_FLOW_PORT 18
482 /**< Consider device port number as a flow differentiator */
483 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */
484 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */
485 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */
486 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */
487 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */
488 #define RTE_ETH_FLOW_MAX 24
491 * Below macros are defined for RSS offload types, they can be used to
492 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types.
494 #define ETH_RSS_IPV4 (1ULL << 2)
495 #define ETH_RSS_FRAG_IPV4 (1ULL << 3)
496 #define ETH_RSS_NONFRAG_IPV4_TCP (1ULL << 4)
497 #define ETH_RSS_NONFRAG_IPV4_UDP (1ULL << 5)
498 #define ETH_RSS_NONFRAG_IPV4_SCTP (1ULL << 6)
499 #define ETH_RSS_NONFRAG_IPV4_OTHER (1ULL << 7)
500 #define ETH_RSS_IPV6 (1ULL << 8)
501 #define ETH_RSS_FRAG_IPV6 (1ULL << 9)
502 #define ETH_RSS_NONFRAG_IPV6_TCP (1ULL << 10)
503 #define ETH_RSS_NONFRAG_IPV6_UDP (1ULL << 11)
504 #define ETH_RSS_NONFRAG_IPV6_SCTP (1ULL << 12)
505 #define ETH_RSS_NONFRAG_IPV6_OTHER (1ULL << 13)
506 #define ETH_RSS_L2_PAYLOAD (1ULL << 14)
507 #define ETH_RSS_IPV6_EX (1ULL << 15)
508 #define ETH_RSS_IPV6_TCP_EX (1ULL << 16)
509 #define ETH_RSS_IPV6_UDP_EX (1ULL << 17)
510 #define ETH_RSS_PORT (1ULL << 18)
511 #define ETH_RSS_VXLAN (1ULL << 19)
512 #define ETH_RSS_GENEVE (1ULL << 20)
513 #define ETH_RSS_NVGRE (1ULL << 21)
514 #define ETH_RSS_GTPU (1ULL << 23)
515 #define ETH_RSS_ESP (1ULL << 27)
516 #define ETH_RSS_AH (1ULL << 28)
520 * We use the following macros to combine with above ETH_RSS_* for
521 * more specific input set selection. These bits are defined starting
522 * from the high end of the 64 bits.
523 * Note: If we use above ETH_RSS_* without SRC/DST_ONLY, it represents
524 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of
525 * the same level are used simultaneously, it is the same case as none of
528 #define ETH_RSS_L3_SRC_ONLY (1ULL << 63)
529 #define ETH_RSS_L3_DST_ONLY (1ULL << 62)
530 #define ETH_RSS_L4_SRC_ONLY (1ULL << 61)
531 #define ETH_RSS_L4_DST_ONLY (1ULL << 60)
534 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of
535 * the same level are used simultaneously, it is the same case as
536 * none of them are added.
539 * RSS types with SRC/DST_ONLY.
543 static inline uint64_t
544 rte_eth_rss_hf_refine(uint64_t rss_hf)
546 if ((rss_hf & ETH_RSS_L3_SRC_ONLY) && (rss_hf & ETH_RSS_L3_DST_ONLY))
547 rss_hf &= ~(ETH_RSS_L3_SRC_ONLY | ETH_RSS_L3_DST_ONLY);
549 if ((rss_hf & ETH_RSS_L4_SRC_ONLY) && (rss_hf & ETH_RSS_L4_DST_ONLY))
550 rss_hf &= ~(ETH_RSS_L4_SRC_ONLY | ETH_RSS_L4_DST_ONLY);
555 #define ETH_RSS_IP ( \
557 ETH_RSS_FRAG_IPV4 | \
558 ETH_RSS_NONFRAG_IPV4_OTHER | \
560 ETH_RSS_FRAG_IPV6 | \
561 ETH_RSS_NONFRAG_IPV6_OTHER | \
564 #define ETH_RSS_UDP ( \
565 ETH_RSS_NONFRAG_IPV4_UDP | \
566 ETH_RSS_NONFRAG_IPV6_UDP | \
569 #define ETH_RSS_TCP ( \
570 ETH_RSS_NONFRAG_IPV4_TCP | \
571 ETH_RSS_NONFRAG_IPV6_TCP | \
574 #define ETH_RSS_SCTP ( \
575 ETH_RSS_NONFRAG_IPV4_SCTP | \
576 ETH_RSS_NONFRAG_IPV6_SCTP)
578 #define ETH_RSS_TUNNEL ( \
583 /**< Mask of valid RSS hash protocols */
584 #define ETH_RSS_PROTO_MASK ( \
586 ETH_RSS_FRAG_IPV4 | \
587 ETH_RSS_NONFRAG_IPV4_TCP | \
588 ETH_RSS_NONFRAG_IPV4_UDP | \
589 ETH_RSS_NONFRAG_IPV4_SCTP | \
590 ETH_RSS_NONFRAG_IPV4_OTHER | \
592 ETH_RSS_FRAG_IPV6 | \
593 ETH_RSS_NONFRAG_IPV6_TCP | \
594 ETH_RSS_NONFRAG_IPV6_UDP | \
595 ETH_RSS_NONFRAG_IPV6_SCTP | \
596 ETH_RSS_NONFRAG_IPV6_OTHER | \
597 ETH_RSS_L2_PAYLOAD | \
599 ETH_RSS_IPV6_TCP_EX | \
600 ETH_RSS_IPV6_UDP_EX | \
607 * Definitions used for redirection table entry size.
608 * Some RSS RETA sizes may not be supported by some drivers, check the
609 * documentation or the description of relevant functions for more details.
611 #define ETH_RSS_RETA_SIZE_64 64
612 #define ETH_RSS_RETA_SIZE_128 128
613 #define ETH_RSS_RETA_SIZE_256 256
614 #define ETH_RSS_RETA_SIZE_512 512
615 #define RTE_RETA_GROUP_SIZE 64
617 /* Definitions used for VMDQ and DCB functionality */
618 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */
619 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
620 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */
621 #define ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
623 /* DCB capability defines */
624 #define ETH_DCB_PG_SUPPORT 0x00000001 /**< Priority Group(ETS) support. */
625 #define ETH_DCB_PFC_SUPPORT 0x00000002 /**< Priority Flow Control support. */
627 /* Definitions used for VLAN Offload functionality */
628 #define ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
629 #define ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
630 #define ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
631 #define ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */
633 /* Definitions used for mask VLAN setting */
634 #define ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
635 #define ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
636 #define ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
637 #define ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */
638 #define ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
640 /* Definitions used for receive MAC address */
641 #define ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
643 /* Definitions used for unicast hash */
644 #define ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
646 /* Definitions used for VMDQ pool rx mode setting */
647 #define ETH_VMDQ_ACCEPT_UNTAG 0x0001 /**< accept untagged packets. */
648 #define ETH_VMDQ_ACCEPT_HASH_MC 0x0002 /**< accept packets in multicast table . */
649 #define ETH_VMDQ_ACCEPT_HASH_UC 0x0004 /**< accept packets in unicast table. */
650 #define ETH_VMDQ_ACCEPT_BROADCAST 0x0008 /**< accept broadcast packets. */
651 #define ETH_VMDQ_ACCEPT_MULTICAST 0x0010 /**< multicast promiscuous. */
653 /** Maximum nb. of vlan per mirror rule */
654 #define ETH_MIRROR_MAX_VLANS 64
656 #define ETH_MIRROR_VIRTUAL_POOL_UP 0x01 /**< Virtual Pool uplink Mirroring. */
657 #define ETH_MIRROR_UPLINK_PORT 0x02 /**< Uplink Port Mirroring. */
658 #define ETH_MIRROR_DOWNLINK_PORT 0x04 /**< Downlink Port Mirroring. */
659 #define ETH_MIRROR_VLAN 0x08 /**< VLAN Mirroring. */
660 #define ETH_MIRROR_VIRTUAL_POOL_DOWN 0x10 /**< Virtual Pool downlink Mirroring. */
663 * A structure used to configure VLAN traffic mirror of an Ethernet port.
665 struct rte_eth_vlan_mirror {
666 uint64_t vlan_mask; /**< mask for valid VLAN ID. */
667 /** VLAN ID list for vlan mirroring. */
668 uint16_t vlan_id[ETH_MIRROR_MAX_VLANS];
672 * A structure used to configure traffic mirror of an Ethernet port.
674 struct rte_eth_mirror_conf {
675 uint8_t rule_type; /**< Mirroring rule type */
676 uint8_t dst_pool; /**< Destination pool for this mirror rule. */
677 uint64_t pool_mask; /**< Bitmap of pool for pool mirroring */
678 /** VLAN ID setting for VLAN mirroring. */
679 struct rte_eth_vlan_mirror vlan;
683 * A structure used to configure 64 entries of Redirection Table of the
684 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
685 * more than 64 entries supported by hardware, an array of this structure
688 struct rte_eth_rss_reta_entry64 {
690 /**< Mask bits indicate which entries need to be updated/queried. */
691 uint16_t reta[RTE_RETA_GROUP_SIZE];
692 /**< Group of 64 redirection table entries. */
696 * This enum indicates the possible number of traffic classes
697 * in DCB configurations
699 enum rte_eth_nb_tcs {
700 ETH_4_TCS = 4, /**< 4 TCs with DCB. */
701 ETH_8_TCS = 8 /**< 8 TCs with DCB. */
705 * This enum indicates the possible number of queue pools
706 * in VMDQ configurations.
708 enum rte_eth_nb_pools {
709 ETH_8_POOLS = 8, /**< 8 VMDq pools. */
710 ETH_16_POOLS = 16, /**< 16 VMDq pools. */
711 ETH_32_POOLS = 32, /**< 32 VMDq pools. */
712 ETH_64_POOLS = 64 /**< 64 VMDq pools. */
715 /* This structure may be extended in future. */
716 struct rte_eth_dcb_rx_conf {
717 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
718 /** Traffic class each UP mapped to. */
719 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
722 struct rte_eth_vmdq_dcb_tx_conf {
723 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
724 /** Traffic class each UP mapped to. */
725 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
728 struct rte_eth_dcb_tx_conf {
729 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
730 /** Traffic class each UP mapped to. */
731 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
734 struct rte_eth_vmdq_tx_conf {
735 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
739 * A structure used to configure the VMDQ+DCB feature
740 * of an Ethernet port.
742 * Using this feature, packets are routed to a pool of queues, based
743 * on the vlan id in the vlan tag, and then to a specific queue within
744 * that pool, using the user priority vlan tag field.
746 * A default pool may be used, if desired, to route all traffic which
747 * does not match the vlan filter rules.
749 struct rte_eth_vmdq_dcb_conf {
750 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
751 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
752 uint8_t default_pool; /**< The default pool, if applicable */
753 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
755 uint16_t vlan_id; /**< The vlan id of the received frame */
756 uint64_t pools; /**< Bitmask of pools for packet rx */
757 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
758 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
759 /**< Selects a queue in a pool */
763 * A structure used to configure the VMDQ feature of an Ethernet port when
764 * not combined with the DCB feature.
766 * Using this feature, packets are routed to a pool of queues. By default,
767 * the pool selection is based on the MAC address, the vlan id in the
768 * vlan tag as specified in the pool_map array.
769 * Passing the ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
770 * selection using only the MAC address. MAC address to pool mapping is done
771 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
772 * corresponding to the pool id.
774 * Queue selection within the selected pool will be done using RSS when
775 * it is enabled or revert to the first queue of the pool if not.
777 * A default pool may be used, if desired, to route all traffic which
778 * does not match the vlan filter rules or any pool MAC address.
780 struct rte_eth_vmdq_rx_conf {
781 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
782 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
783 uint8_t default_pool; /**< The default pool, if applicable */
784 uint8_t enable_loop_back; /**< Enable VT loop back */
785 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
786 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
788 uint16_t vlan_id; /**< The vlan id of the received frame */
789 uint64_t pools; /**< Bitmask of pools for packet rx */
790 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
794 * A structure used to configure the TX features of an Ethernet port.
796 struct rte_eth_txmode {
797 enum rte_eth_tx_mq_mode mq_mode; /**< TX multi-queues mode. */
799 * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
800 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
801 * structure are allowed to be set.
805 /* For i40e specifically */
808 uint8_t hw_vlan_reject_tagged : 1,
809 /**< If set, reject sending out tagged pkts */
810 hw_vlan_reject_untagged : 1,
811 /**< If set, reject sending out untagged pkts */
812 hw_vlan_insert_pvid : 1;
813 /**< If set, enable port based VLAN insertion */
815 uint64_t reserved_64s[2]; /**< Reserved for future fields */
816 void *reserved_ptrs[2]; /**< Reserved for future fields */
820 * A structure used to configure an RX ring of an Ethernet port.
822 struct rte_eth_rxconf {
823 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */
824 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */
825 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
826 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
828 * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
829 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
830 * fields on rte_eth_dev_info structure are allowed to be set.
834 uint64_t reserved_64s[2]; /**< Reserved for future fields */
835 void *reserved_ptrs[2]; /**< Reserved for future fields */
839 * A structure used to configure a TX ring of an Ethernet port.
841 struct rte_eth_txconf {
842 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */
843 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
844 uint16_t tx_free_thresh; /**< Start freeing TX buffers if there are
845 less free descriptors than this value. */
847 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
849 * Per-queue Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
850 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
851 * fields on rte_eth_dev_info structure are allowed to be set.
855 uint64_t reserved_64s[2]; /**< Reserved for future fields */
856 void *reserved_ptrs[2]; /**< Reserved for future fields */
861 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
863 * A structure used to return the hairpin capabilities that are supported.
865 struct rte_eth_hairpin_cap {
866 /** The max number of hairpin queues (different bindings). */
867 uint16_t max_nb_queues;
868 /** Max number of Rx queues to be connected to one Tx queue. */
869 uint16_t max_rx_2_tx;
870 /** Max number of Tx queues to be connected to one Rx queue. */
871 uint16_t max_tx_2_rx;
872 uint16_t max_nb_desc; /**< The max num of descriptors. */
875 #define RTE_ETH_MAX_HAIRPIN_PEERS 32
879 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
881 * A structure used to hold hairpin peer data.
883 struct rte_eth_hairpin_peer {
884 uint16_t port; /**< Peer port. */
885 uint16_t queue; /**< Peer queue. */
890 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
892 * A structure used to configure hairpin binding.
894 struct rte_eth_hairpin_conf {
895 uint16_t peer_count; /**< The number of peers. */
896 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
900 * A structure contains information about HW descriptor ring limitations.
902 struct rte_eth_desc_lim {
903 uint16_t nb_max; /**< Max allowed number of descriptors. */
904 uint16_t nb_min; /**< Min allowed number of descriptors. */
905 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
908 * Max allowed number of segments per whole packet.
910 * - For TSO packet this is the total number of data descriptors allowed
913 * @see nb_mtu_seg_max
918 * Max number of segments per one MTU.
920 * - For non-TSO packet, this is the maximum allowed number of segments
921 * in a single transmit packet.
923 * - For TSO packet each segment within the TSO may span up to this
928 uint16_t nb_mtu_seg_max;
932 * This enum indicates the flow control mode
934 enum rte_eth_fc_mode {
935 RTE_FC_NONE = 0, /**< Disable flow control. */
936 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */
937 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */
938 RTE_FC_FULL /**< Enable flow control on both side. */
942 * A structure used to configure Ethernet flow control parameter.
943 * These parameters will be configured into the register of the NIC.
944 * Please refer to the corresponding data sheet for proper value.
946 struct rte_eth_fc_conf {
947 uint32_t high_water; /**< High threshold value to trigger XOFF */
948 uint32_t low_water; /**< Low threshold value to trigger XON */
949 uint16_t pause_time; /**< Pause quota in the Pause frame */
950 uint16_t send_xon; /**< Is XON frame need be sent */
951 enum rte_eth_fc_mode mode; /**< Link flow control mode */
952 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
953 uint8_t autoneg; /**< Use Pause autoneg */
957 * A structure used to configure Ethernet priority flow control parameter.
958 * These parameters will be configured into the register of the NIC.
959 * Please refer to the corresponding data sheet for proper value.
961 struct rte_eth_pfc_conf {
962 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
963 uint8_t priority; /**< VLAN User Priority. */
969 enum rte_eth_tunnel_type {
970 RTE_TUNNEL_TYPE_NONE = 0,
971 RTE_TUNNEL_TYPE_VXLAN,
972 RTE_TUNNEL_TYPE_GENEVE,
973 RTE_TUNNEL_TYPE_TEREDO,
974 RTE_TUNNEL_TYPE_NVGRE,
975 RTE_TUNNEL_TYPE_IP_IN_GRE,
976 RTE_L2_TUNNEL_TYPE_E_TAG,
977 RTE_TUNNEL_TYPE_VXLAN_GPE,
981 /* Deprecated API file for rte_eth_dev_filter_* functions */
982 #include "rte_eth_ctrl.h"
985 * Memory space that can be configured to store Flow Director filters
986 * in the board memory.
988 enum rte_fdir_pballoc_type {
989 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */
990 RTE_FDIR_PBALLOC_128K, /**< 128k. */
991 RTE_FDIR_PBALLOC_256K, /**< 256k. */
995 * Select report mode of FDIR hash information in RX descriptors.
997 enum rte_fdir_status_mode {
998 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
999 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
1000 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
1004 * A structure used to configure the Flow Director (FDIR) feature
1005 * of an Ethernet port.
1007 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored.
1009 struct rte_fdir_conf {
1010 enum rte_fdir_mode mode; /**< Flow Director mode. */
1011 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
1012 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
1013 /** RX queue of packets matching a "drop" filter in perfect mode. */
1015 struct rte_eth_fdir_masks mask;
1016 struct rte_eth_fdir_flex_conf flex_conf;
1017 /**< Flex payload configuration. */
1021 * UDP tunneling configuration.
1022 * Used to config the UDP port for a type of tunnel.
1023 * NICs need the UDP port to identify the tunnel type.
1024 * Normally a type of tunnel has a default UDP port, this structure can be used
1025 * in case if the users want to change or support more UDP port.
1027 struct rte_eth_udp_tunnel {
1028 uint16_t udp_port; /**< UDP port used for the tunnel. */
1029 uint8_t prot_type; /**< Tunnel type. Defined in rte_eth_tunnel_type. */
1033 * A structure used to enable/disable specific device interrupts.
1035 struct rte_intr_conf {
1036 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
1038 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
1040 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
1045 * A structure used to configure an Ethernet port.
1046 * Depending upon the RX multi-queue mode, extra advanced
1047 * configuration settings may be needed.
1049 struct rte_eth_conf {
1050 uint32_t link_speeds; /**< bitmap of ETH_LINK_SPEED_XXX of speeds to be
1051 used. ETH_LINK_SPEED_FIXED disables link
1052 autonegotiation, and a unique speed shall be
1053 set. Otherwise, the bitmap defines the set of
1054 speeds to be advertised. If the special value
1055 ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
1056 supported are advertised. */
1057 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */
1058 struct rte_eth_txmode txmode; /**< Port TX configuration. */
1059 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
1060 is 0, meaning the loopback mode is disabled.
1061 Read the datasheet of given ethernet controller
1062 for details. The possible values of this field
1063 are defined in implementation of each driver. */
1065 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
1066 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
1067 /**< Port vmdq+dcb configuration. */
1068 struct rte_eth_dcb_rx_conf dcb_rx_conf;
1069 /**< Port dcb RX configuration. */
1070 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
1071 /**< Port vmdq RX configuration. */
1072 } rx_adv_conf; /**< Port RX filtering configuration. */
1074 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
1075 /**< Port vmdq+dcb TX configuration. */
1076 struct rte_eth_dcb_tx_conf dcb_tx_conf;
1077 /**< Port dcb TX configuration. */
1078 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
1079 /**< Port vmdq TX configuration. */
1080 } tx_adv_conf; /**< Port TX DCB configuration (union). */
1081 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
1082 is needed,and the variable must be set ETH_DCB_PFC_SUPPORT. */
1083 uint32_t dcb_capability_en;
1084 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */
1085 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */
1089 * RX offload capabilities of a device.
1091 #define DEV_RX_OFFLOAD_VLAN_STRIP 0x00000001
1092 #define DEV_RX_OFFLOAD_IPV4_CKSUM 0x00000002
1093 #define DEV_RX_OFFLOAD_UDP_CKSUM 0x00000004
1094 #define DEV_RX_OFFLOAD_TCP_CKSUM 0x00000008
1095 #define DEV_RX_OFFLOAD_TCP_LRO 0x00000010
1096 #define DEV_RX_OFFLOAD_QINQ_STRIP 0x00000020
1097 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000040
1098 #define DEV_RX_OFFLOAD_MACSEC_STRIP 0x00000080
1099 #define DEV_RX_OFFLOAD_HEADER_SPLIT 0x00000100
1100 #define DEV_RX_OFFLOAD_VLAN_FILTER 0x00000200
1101 #define DEV_RX_OFFLOAD_VLAN_EXTEND 0x00000400
1102 #define DEV_RX_OFFLOAD_JUMBO_FRAME 0x00000800
1103 #define DEV_RX_OFFLOAD_SCATTER 0x00002000
1104 #define DEV_RX_OFFLOAD_TIMESTAMP 0x00004000
1105 #define DEV_RX_OFFLOAD_SECURITY 0x00008000
1106 #define DEV_RX_OFFLOAD_KEEP_CRC 0x00010000
1107 #define DEV_RX_OFFLOAD_SCTP_CKSUM 0x00020000
1108 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM 0x00040000
1109 #define DEV_RX_OFFLOAD_RSS_HASH 0x00080000
1111 #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \
1112 DEV_RX_OFFLOAD_UDP_CKSUM | \
1113 DEV_RX_OFFLOAD_TCP_CKSUM)
1114 #define DEV_RX_OFFLOAD_VLAN (DEV_RX_OFFLOAD_VLAN_STRIP | \
1115 DEV_RX_OFFLOAD_VLAN_FILTER | \
1116 DEV_RX_OFFLOAD_VLAN_EXTEND | \
1117 DEV_RX_OFFLOAD_QINQ_STRIP)
1120 * If new Rx offload capabilities are defined, they also must be
1121 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
1125 * TX offload capabilities of a device.
1127 #define DEV_TX_OFFLOAD_VLAN_INSERT 0x00000001
1128 #define DEV_TX_OFFLOAD_IPV4_CKSUM 0x00000002
1129 #define DEV_TX_OFFLOAD_UDP_CKSUM 0x00000004
1130 #define DEV_TX_OFFLOAD_TCP_CKSUM 0x00000008
1131 #define DEV_TX_OFFLOAD_SCTP_CKSUM 0x00000010
1132 #define DEV_TX_OFFLOAD_TCP_TSO 0x00000020
1133 #define DEV_TX_OFFLOAD_UDP_TSO 0x00000040
1134 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000080 /**< Used for tunneling packet. */
1135 #define DEV_TX_OFFLOAD_QINQ_INSERT 0x00000100
1136 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO 0x00000200 /**< Used for tunneling packet. */
1137 #define DEV_TX_OFFLOAD_GRE_TNL_TSO 0x00000400 /**< Used for tunneling packet. */
1138 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO 0x00000800 /**< Used for tunneling packet. */
1139 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO 0x00001000 /**< Used for tunneling packet. */
1140 #define DEV_TX_OFFLOAD_MACSEC_INSERT 0x00002000
1141 #define DEV_TX_OFFLOAD_MT_LOCKFREE 0x00004000
1142 /**< Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
1143 * tx queue without SW lock.
1145 #define DEV_TX_OFFLOAD_MULTI_SEGS 0x00008000
1146 /**< Device supports multi segment send. */
1147 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE 0x00010000
1148 /**< Device supports optimization for fast release of mbufs.
1149 * When set application must guarantee that per-queue all mbufs comes from
1150 * the same mempool and has refcnt = 1.
1152 #define DEV_TX_OFFLOAD_SECURITY 0x00020000
1154 * Device supports generic UDP tunneled packet TSO.
1155 * Application must set PKT_TX_TUNNEL_UDP and other mbuf fields required
1158 #define DEV_TX_OFFLOAD_UDP_TNL_TSO 0x00040000
1160 * Device supports generic IP tunneled packet TSO.
1161 * Application must set PKT_TX_TUNNEL_IP and other mbuf fields required
1164 #define DEV_TX_OFFLOAD_IP_TNL_TSO 0x00080000
1165 /** Device supports outer UDP checksum */
1166 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM 0x00100000
1168 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP 0x00000001
1169 /**< Device supports Rx queue setup after device started*/
1170 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP 0x00000002
1171 /**< Device supports Tx queue setup after device started*/
1174 * If new Tx offload capabilities are defined, they also must be
1175 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1179 * Fallback default preferred Rx/Tx port parameters.
1180 * These are used if an application requests default parameters
1181 * but the PMD does not provide preferred values.
1183 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1184 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1185 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1186 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1189 * Preferred Rx/Tx port parameters.
1190 * There are separate instances of this structure for transmission
1191 * and reception respectively.
1193 struct rte_eth_dev_portconf {
1194 uint16_t burst_size; /**< Device-preferred burst size */
1195 uint16_t ring_size; /**< Device-preferred size of queue rings */
1196 uint16_t nb_queues; /**< Device-preferred number of queues */
1200 * Default values for switch domain id when ethdev does not support switch
1201 * domain definitions.
1203 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX)
1206 * Ethernet device associated switch information
1208 struct rte_eth_switch_info {
1209 const char *name; /**< switch name */
1210 uint16_t domain_id; /**< switch domain id */
1213 * mapping to the devices physical switch port as enumerated from the
1214 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1215 * device this may correspond to the VF_ID of each virtual function,
1216 * but each driver should explicitly define the mapping of switch
1217 * port identifier to that physical interconnect/switch
1222 * Ethernet device information
1226 * A structure used to retrieve the contextual information of
1227 * an Ethernet device, such as the controlling driver of the
1230 struct rte_eth_dev_info {
1231 struct rte_device *device; /** Generic device information */
1232 const char *driver_name; /**< Device Driver name. */
1233 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1234 Use if_indextoname() to translate into an interface name. */
1235 uint16_t min_mtu; /**< Minimum MTU allowed */
1236 uint16_t max_mtu; /**< Maximum MTU allowed */
1237 const uint32_t *dev_flags; /**< Device flags */
1238 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */
1239 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */
1240 /** Maximum configurable size of LRO aggregated packet. */
1241 uint32_t max_lro_pkt_size;
1242 uint16_t max_rx_queues; /**< Maximum number of RX queues. */
1243 uint16_t max_tx_queues; /**< Maximum number of TX queues. */
1244 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1245 uint32_t max_hash_mac_addrs;
1246 /** Maximum number of hash MAC addresses for MTA and UTA. */
1247 uint16_t max_vfs; /**< Maximum number of VFs. */
1248 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1249 uint64_t rx_offload_capa;
1250 /**< All RX offload capabilities including all per-queue ones */
1251 uint64_t tx_offload_capa;
1252 /**< All TX offload capabilities including all per-queue ones */
1253 uint64_t rx_queue_offload_capa;
1254 /**< Device per-queue RX offload capabilities. */
1255 uint64_t tx_queue_offload_capa;
1256 /**< Device per-queue TX offload capabilities. */
1258 /**< Device redirection table size, the total number of entries. */
1259 uint8_t hash_key_size; /**< Hash key size in bytes */
1260 /** Bit mask of RSS offloads, the bit offset also means flow type */
1261 uint64_t flow_type_rss_offloads;
1262 struct rte_eth_rxconf default_rxconf; /**< Default RX configuration */
1263 struct rte_eth_txconf default_txconf; /**< Default TX configuration */
1264 uint16_t vmdq_queue_base; /**< First queue ID for VMDQ pools. */
1265 uint16_t vmdq_queue_num; /**< Queue number for VMDQ pools. */
1266 uint16_t vmdq_pool_base; /**< First ID of VMDQ pools. */
1267 struct rte_eth_desc_lim rx_desc_lim; /**< RX descriptors limits */
1268 struct rte_eth_desc_lim tx_desc_lim; /**< TX descriptors limits */
1269 uint32_t speed_capa; /**< Supported speeds bitmap (ETH_LINK_SPEED_). */
1270 /** Configured number of rx/tx queues */
1271 uint16_t nb_rx_queues; /**< Number of RX queues. */
1272 uint16_t nb_tx_queues; /**< Number of TX queues. */
1273 /** Rx parameter recommendations */
1274 struct rte_eth_dev_portconf default_rxportconf;
1275 /** Tx parameter recommendations */
1276 struct rte_eth_dev_portconf default_txportconf;
1277 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1280 * Switching information for ports on a device with a
1281 * embedded managed interconnect/switch.
1283 struct rte_eth_switch_info switch_info;
1285 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1286 void *reserved_ptrs[2]; /**< Reserved for future fields */
1290 * Ethernet device RX queue information structure.
1291 * Used to retrieve information about configured queue.
1293 struct rte_eth_rxq_info {
1294 struct rte_mempool *mp; /**< mempool used by that queue. */
1295 struct rte_eth_rxconf conf; /**< queue config parameters. */
1296 uint8_t scattered_rx; /**< scattered packets RX supported. */
1297 uint16_t nb_desc; /**< configured number of RXDs. */
1298 } __rte_cache_min_aligned;
1301 * Ethernet device TX queue information structure.
1302 * Used to retrieve information about configured queue.
1304 struct rte_eth_txq_info {
1305 struct rte_eth_txconf conf; /**< queue config parameters. */
1306 uint16_t nb_desc; /**< configured number of TXDs. */
1307 } __rte_cache_min_aligned;
1309 /* Generic Burst mode flag definition, values can be ORed. */
1312 * If the queues have different burst mode description, this bit will be set
1313 * by PMD, then the application can iterate to retrieve burst description for
1316 #define RTE_ETH_BURST_FLAG_PER_QUEUE (1ULL << 0)
1319 * Ethernet device RX/TX queue packet burst mode information structure.
1320 * Used to retrieve information about packet burst mode setting.
1322 struct rte_eth_burst_mode {
1323 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */
1325 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */
1326 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */
1329 /** Maximum name length for extended statistics counters */
1330 #define RTE_ETH_XSTATS_NAME_SIZE 64
1333 * An Ethernet device extended statistic structure
1335 * This structure is used by rte_eth_xstats_get() to provide
1336 * statistics that are not provided in the generic *rte_eth_stats*
1338 * It maps a name id, corresponding to an index in the array returned
1339 * by rte_eth_xstats_get_names(), to a statistic value.
1341 struct rte_eth_xstat {
1342 uint64_t id; /**< The index in xstats name array. */
1343 uint64_t value; /**< The statistic counter value. */
1347 * A name element for extended statistics.
1349 * An array of this structure is returned by rte_eth_xstats_get_names().
1350 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1351 * structure references these names by their array index.
1353 struct rte_eth_xstat_name {
1354 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1357 #define ETH_DCB_NUM_TCS 8
1358 #define ETH_MAX_VMDQ_POOL 64
1361 * A structure used to get the information of queue and
1362 * TC mapping on both TX and RX paths.
1364 struct rte_eth_dcb_tc_queue_mapping {
1365 /** rx queues assigned to tc per Pool */
1369 } tc_rxq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1370 /** rx queues assigned to tc per Pool */
1374 } tc_txq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1378 * A structure used to get the information of DCB.
1379 * It includes TC UP mapping and queue TC mapping.
1381 struct rte_eth_dcb_info {
1382 uint8_t nb_tcs; /**< number of TCs */
1383 uint8_t prio_tc[ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
1384 uint8_t tc_bws[ETH_DCB_NUM_TCS]; /**< TX BW percentage for each TC */
1385 /** rx queues assigned to tc */
1386 struct rte_eth_dcb_tc_queue_mapping tc_queue;
1389 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
1391 /* Macros to check for valid port */
1392 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
1393 if (!rte_eth_dev_is_valid_port(port_id)) { \
1394 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1399 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
1400 if (!rte_eth_dev_is_valid_port(port_id)) { \
1401 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1407 * l2 tunnel configuration.
1410 /**< l2 tunnel enable mask */
1411 #define ETH_L2_TUNNEL_ENABLE_MASK 0x00000001
1412 /**< l2 tunnel insertion mask */
1413 #define ETH_L2_TUNNEL_INSERTION_MASK 0x00000002
1414 /**< l2 tunnel stripping mask */
1415 #define ETH_L2_TUNNEL_STRIPPING_MASK 0x00000004
1416 /**< l2 tunnel forwarding mask */
1417 #define ETH_L2_TUNNEL_FORWARDING_MASK 0x00000008
1420 * Function type used for RX packet processing packet callbacks.
1422 * The callback function is called on RX with a burst of packets that have
1423 * been received on the given port and queue.
1426 * The Ethernet port on which RX is being performed.
1428 * The queue on the Ethernet port which is being used to receive the packets.
1430 * The burst of packets that have just been received.
1432 * The number of packets in the burst pointed to by "pkts".
1434 * The max number of packets that can be stored in the "pkts" array.
1436 * The arbitrary user parameter passed in by the application when the callback
1437 * was originally configured.
1439 * The number of packets returned to the user.
1441 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
1442 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
1446 * Function type used for TX packet processing packet callbacks.
1448 * The callback function is called on TX with a burst of packets immediately
1449 * before the packets are put onto the hardware queue for transmission.
1452 * The Ethernet port on which TX is being performed.
1454 * The queue on the Ethernet port which is being used to transmit the packets.
1456 * The burst of packets that are about to be transmitted.
1458 * The number of packets in the burst pointed to by "pkts".
1460 * The arbitrary user parameter passed in by the application when the callback
1461 * was originally configured.
1463 * The number of packets to be written to the NIC.
1465 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
1466 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
1469 * Possible states of an ethdev port.
1471 enum rte_eth_dev_state {
1472 /** Device is unused before being probed. */
1473 RTE_ETH_DEV_UNUSED = 0,
1474 /** Device is attached when allocated in probing. */
1475 RTE_ETH_DEV_ATTACHED,
1476 /** Device is in removed state when plug-out is detected. */
1477 RTE_ETH_DEV_REMOVED,
1480 struct rte_eth_dev_sriov {
1481 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
1482 uint8_t nb_q_per_pool; /**< rx queue number per pool */
1483 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
1484 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
1486 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
1488 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
1490 #define RTE_ETH_DEV_NO_OWNER 0
1492 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
1494 struct rte_eth_dev_owner {
1495 uint64_t id; /**< The owner unique identifier. */
1496 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
1500 * Port is released (i.e. totally freed and data erased) on close.
1501 * Temporary flag for PMD migration to new rte_eth_dev_close() behaviour.
1503 #define RTE_ETH_DEV_CLOSE_REMOVE 0x0001
1504 /** Device supports link state interrupt */
1505 #define RTE_ETH_DEV_INTR_LSC 0x0002
1506 /** Device is a bonded slave */
1507 #define RTE_ETH_DEV_BONDED_SLAVE 0x0004
1508 /** Device supports device removal interrupt */
1509 #define RTE_ETH_DEV_INTR_RMV 0x0008
1510 /** Device is port representor */
1511 #define RTE_ETH_DEV_REPRESENTOR 0x0010
1512 /** Device does not support MAC change after started */
1513 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR 0x0020
1516 * Iterates over valid ethdev ports owned by a specific owner.
1519 * The id of the next possible valid owned port.
1521 * The owner identifier.
1522 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
1524 * Next valid port id owned by owner_id, RTE_MAX_ETHPORTS if there is none.
1526 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
1527 const uint64_t owner_id);
1530 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
1532 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
1533 for (p = rte_eth_find_next_owned_by(0, o); \
1534 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
1535 p = rte_eth_find_next_owned_by(p + 1, o))
1538 * Iterates over valid ethdev ports.
1541 * The id of the next possible valid port.
1543 * Next valid port id, RTE_MAX_ETHPORTS if there is none.
1545 uint16_t rte_eth_find_next(uint16_t port_id);
1548 * Macro to iterate over all enabled and ownerless ethdev ports.
1550 #define RTE_ETH_FOREACH_DEV(p) \
1551 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
1555 * @b EXPERIMENTAL: this API may change without prior notice.
1557 * Iterates over ethdev ports of a specified device.
1559 * @param port_id_start
1560 * The id of the next possible valid port.
1562 * The generic device behind the ports to iterate.
1564 * Next port id of the device, possibly port_id_start,
1565 * RTE_MAX_ETHPORTS if there is none.
1569 rte_eth_find_next_of(uint16_t port_id_start,
1570 const struct rte_device *parent);
1573 * Macro to iterate over all ethdev ports of a specified device.
1576 * The id of the matching port being iterated.
1578 * The rte_device pointer matching the iterated ports.
1580 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \
1581 for (port_id = rte_eth_find_next_of(0, parent); \
1582 port_id < RTE_MAX_ETHPORTS; \
1583 port_id = rte_eth_find_next_of(port_id + 1, parent))
1587 * @b EXPERIMENTAL: this API may change without prior notice.
1589 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device).
1591 * @param port_id_start
1592 * The id of the next possible valid sibling port.
1593 * @param ref_port_id
1594 * The id of a reference port to compare rte_device with.
1596 * Next sibling port id, possibly port_id_start or ref_port_id itself,
1597 * RTE_MAX_ETHPORTS if there is none.
1601 rte_eth_find_next_sibling(uint16_t port_id_start,
1602 uint16_t ref_port_id);
1605 * Macro to iterate over all ethdev ports sharing the same rte_device
1606 * as the specified port.
1607 * Note: the specified reference port is part of the loop iterations.
1610 * The id of the matching port being iterated.
1611 * @param ref_port_id
1612 * The id of the port being compared.
1614 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \
1615 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \
1616 port_id < RTE_MAX_ETHPORTS; \
1617 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id))
1621 * @b EXPERIMENTAL: this API may change without prior notice.
1623 * Get a new unique owner identifier.
1624 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
1625 * to avoid multiple management of device by different entities.
1628 * Owner identifier pointer.
1630 * Negative errno value on error, 0 on success.
1633 int rte_eth_dev_owner_new(uint64_t *owner_id);
1637 * @b EXPERIMENTAL: this API may change without prior notice.
1639 * Set an Ethernet device owner.
1642 * The identifier of the port to own.
1644 * The owner pointer.
1646 * Negative errno value on error, 0 on success.
1649 int rte_eth_dev_owner_set(const uint16_t port_id,
1650 const struct rte_eth_dev_owner *owner);
1654 * @b EXPERIMENTAL: this API may change without prior notice.
1656 * Unset Ethernet device owner to make the device ownerless.
1659 * The identifier of port to make ownerless.
1661 * The owner identifier.
1663 * 0 on success, negative errno value on error.
1666 int rte_eth_dev_owner_unset(const uint16_t port_id,
1667 const uint64_t owner_id);
1671 * @b EXPERIMENTAL: this API may change without prior notice.
1673 * Remove owner from all Ethernet devices owned by a specific owner.
1676 * The owner identifier.
1678 * 0 on success, negative errno value on error.
1681 int rte_eth_dev_owner_delete(const uint64_t owner_id);
1685 * @b EXPERIMENTAL: this API may change without prior notice.
1687 * Get the owner of an Ethernet device.
1690 * The port identifier.
1692 * The owner structure pointer to fill.
1694 * 0 on success, negative errno value on error..
1697 int rte_eth_dev_owner_get(const uint16_t port_id,
1698 struct rte_eth_dev_owner *owner);
1701 * Get the number of ports which are usable for the application.
1703 * These devices must be iterated by using the macro
1704 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
1705 * to deal with non-contiguous ranges of devices.
1708 * The count of available Ethernet devices.
1710 uint16_t rte_eth_dev_count_avail(void);
1713 * Get the total number of ports which are allocated.
1715 * Some devices may not be available for the application.
1718 * The total count of Ethernet devices.
1720 uint16_t rte_eth_dev_count_total(void);
1723 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
1724 * the bitmap link_speeds of the struct rte_eth_conf
1727 * Numerical speed value in Mbps
1729 * ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
1731 * 0 if the speed cannot be mapped
1733 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
1736 * Get DEV_RX_OFFLOAD_* flag name.
1741 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1743 const char *rte_eth_dev_rx_offload_name(uint64_t offload);
1746 * Get DEV_TX_OFFLOAD_* flag name.
1751 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1753 const char *rte_eth_dev_tx_offload_name(uint64_t offload);
1756 * Configure an Ethernet device.
1757 * This function must be invoked first before any other function in the
1758 * Ethernet API. This function can also be re-invoked when a device is in the
1762 * The port identifier of the Ethernet device to configure.
1763 * @param nb_rx_queue
1764 * The number of receive queues to set up for the Ethernet device.
1765 * @param nb_tx_queue
1766 * The number of transmit queues to set up for the Ethernet device.
1768 * The pointer to the configuration data to be used for the Ethernet device.
1769 * The *rte_eth_conf* structure includes:
1770 * - the hardware offload features to activate, with dedicated fields for
1771 * each statically configurable offload hardware feature provided by
1772 * Ethernet devices, such as IP checksum or VLAN tag stripping for
1774 * The Rx offload bitfield API is obsolete and will be deprecated.
1775 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
1776 * structure and use offloads field to set per-port offloads instead.
1777 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
1778 * the [rt]x_offload_capa returned from rte_eth_dev_info_get().
1779 * Any type of device supported offloading set in the input argument
1780 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
1781 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup()
1782 * - the Receive Side Scaling (RSS) configuration when using multiple RX
1783 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf
1784 * must be within the flow_type_rss_offloads provided by drivers via
1785 * rte_eth_dev_info_get() API.
1787 * Embedding all configuration information in a single data structure
1788 * is the more flexible method that allows the addition of new features
1789 * without changing the syntax of the API.
1791 * - 0: Success, device configured.
1792 * - <0: Error code returned by the driver configuration function.
1794 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
1795 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
1799 * @b EXPERIMENTAL: this API may change without prior notice.
1801 * Check if an Ethernet device was physically removed.
1804 * The port identifier of the Ethernet device.
1806 * 1 when the Ethernet device is removed, otherwise 0.
1810 rte_eth_dev_is_removed(uint16_t port_id);
1813 * Allocate and set up a receive queue for an Ethernet device.
1815 * The function allocates a contiguous block of memory for *nb_rx_desc*
1816 * receive descriptors from a memory zone associated with *socket_id*
1817 * and initializes each receive descriptor with a network buffer allocated
1818 * from the memory pool *mb_pool*.
1821 * The port identifier of the Ethernet device.
1822 * @param rx_queue_id
1823 * The index of the receive queue to set up.
1824 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1825 * to rte_eth_dev_configure().
1827 * The number of receive descriptors to allocate for the receive ring.
1829 * The *socket_id* argument is the socket identifier in case of NUMA.
1830 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1831 * the DMA memory allocated for the receive descriptors of the ring.
1833 * The pointer to the configuration data to be used for the receive queue.
1834 * NULL value is allowed, in which case default RX configuration
1836 * The *rx_conf* structure contains an *rx_thresh* structure with the values
1837 * of the Prefetch, Host, and Write-Back threshold registers of the receive
1839 * In addition it contains the hardware offloads features to activate using
1840 * the DEV_RX_OFFLOAD_* flags.
1841 * If an offloading set in rx_conf->offloads
1842 * hasn't been set in the input argument eth_conf->rxmode.offloads
1843 * to rte_eth_dev_configure(), it is a new added offloading, it must be
1844 * per-queue type and it is enabled for the queue.
1845 * No need to repeat any bit in rx_conf->offloads which has already been
1846 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
1847 * at port level can't be disabled at queue level.
1849 * The pointer to the memory pool from which to allocate *rte_mbuf* network
1850 * memory buffers to populate each descriptor of the receive ring.
1852 * - 0: Success, receive queue correctly set up.
1853 * - -EIO: if device is removed.
1854 * - -EINVAL: The memory pool pointer is null or the size of network buffers
1855 * which can be allocated from this memory pool does not fit the various
1856 * buffer sizes allowed by the device controller.
1857 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
1858 * allocate network memory buffers from the memory pool when
1859 * initializing receive descriptors.
1861 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
1862 uint16_t nb_rx_desc, unsigned int socket_id,
1863 const struct rte_eth_rxconf *rx_conf,
1864 struct rte_mempool *mb_pool);
1868 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1870 * Allocate and set up a hairpin receive queue for an Ethernet device.
1872 * The function set up the selected queue to be used in hairpin.
1875 * The port identifier of the Ethernet device.
1876 * @param rx_queue_id
1877 * The index of the receive queue to set up.
1878 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1879 * to rte_eth_dev_configure().
1881 * The number of receive descriptors to allocate for the receive ring.
1882 * 0 means the PMD will use default value.
1884 * The pointer to the hairpin configuration.
1887 * - (0) if successful.
1888 * - (-ENOTSUP) if hardware doesn't support.
1889 * - (-EINVAL) if bad parameter.
1890 * - (-ENOMEM) if unable to allocate the resources.
1893 int rte_eth_rx_hairpin_queue_setup
1894 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc,
1895 const struct rte_eth_hairpin_conf *conf);
1898 * Allocate and set up a transmit queue for an Ethernet device.
1901 * The port identifier of the Ethernet device.
1902 * @param tx_queue_id
1903 * The index of the transmit queue to set up.
1904 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1905 * to rte_eth_dev_configure().
1907 * The number of transmit descriptors to allocate for the transmit ring.
1909 * The *socket_id* argument is the socket identifier in case of NUMA.
1910 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1911 * the DMA memory allocated for the transmit descriptors of the ring.
1913 * The pointer to the configuration data to be used for the transmit queue.
1914 * NULL value is allowed, in which case default TX configuration
1916 * The *tx_conf* structure contains the following data:
1917 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
1918 * Write-Back threshold registers of the transmit ring.
1919 * When setting Write-Back threshold to the value greater then zero,
1920 * *tx_rs_thresh* value should be explicitly set to one.
1921 * - The *tx_free_thresh* value indicates the [minimum] number of network
1922 * buffers that must be pending in the transmit ring to trigger their
1923 * [implicit] freeing by the driver transmit function.
1924 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
1925 * descriptors that must be pending in the transmit ring before setting the
1926 * RS bit on a descriptor by the driver transmit function.
1927 * The *tx_rs_thresh* value should be less or equal then
1928 * *tx_free_thresh* value, and both of them should be less then
1930 * - The *offloads* member contains Tx offloads to be enabled.
1931 * If an offloading set in tx_conf->offloads
1932 * hasn't been set in the input argument eth_conf->txmode.offloads
1933 * to rte_eth_dev_configure(), it is a new added offloading, it must be
1934 * per-queue type and it is enabled for the queue.
1935 * No need to repeat any bit in tx_conf->offloads which has already been
1936 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
1937 * at port level can't be disabled at queue level.
1939 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
1940 * the transmit function to use default values.
1942 * - 0: Success, the transmit queue is correctly set up.
1943 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
1945 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
1946 uint16_t nb_tx_desc, unsigned int socket_id,
1947 const struct rte_eth_txconf *tx_conf);
1951 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1953 * Allocate and set up a transmit hairpin queue for an Ethernet device.
1956 * The port identifier of the Ethernet device.
1957 * @param tx_queue_id
1958 * The index of the transmit queue to set up.
1959 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1960 * to rte_eth_dev_configure().
1962 * The number of transmit descriptors to allocate for the transmit ring.
1963 * 0 to set default PMD value.
1965 * The hairpin configuration.
1968 * - (0) if successful.
1969 * - (-ENOTSUP) if hardware doesn't support.
1970 * - (-EINVAL) if bad parameter.
1971 * - (-ENOMEM) if unable to allocate the resources.
1974 int rte_eth_tx_hairpin_queue_setup
1975 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc,
1976 const struct rte_eth_hairpin_conf *conf);
1979 * Return the NUMA socket to which an Ethernet device is connected
1982 * The port identifier of the Ethernet device
1984 * The NUMA socket id to which the Ethernet device is connected or
1985 * a default of zero if the socket could not be determined.
1986 * -1 is returned is the port_id value is out of range.
1988 int rte_eth_dev_socket_id(uint16_t port_id);
1991 * Check if port_id of device is attached
1994 * The port identifier of the Ethernet device
1996 * - 0 if port is out of range or not attached
1997 * - 1 if device is attached
1999 int rte_eth_dev_is_valid_port(uint16_t port_id);
2002 * Start specified RX queue of a port. It is used when rx_deferred_start
2003 * flag of the specified queue is true.
2006 * The port identifier of the Ethernet device
2007 * @param rx_queue_id
2008 * The index of the rx queue to update the ring.
2009 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2010 * to rte_eth_dev_configure().
2012 * - 0: Success, the receive queue is started.
2013 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2014 * - -EIO: if device is removed.
2015 * - -ENOTSUP: The function not supported in PMD driver.
2017 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2020 * Stop specified RX queue of a port
2023 * The port identifier of the Ethernet device
2024 * @param rx_queue_id
2025 * The index of the rx queue to update the ring.
2026 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2027 * to rte_eth_dev_configure().
2029 * - 0: Success, the receive queue is stopped.
2030 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2031 * - -EIO: if device is removed.
2032 * - -ENOTSUP: The function not supported in PMD driver.
2034 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2037 * Start TX for specified queue of a port. It is used when tx_deferred_start
2038 * flag of the specified queue is true.
2041 * The port identifier of the Ethernet device
2042 * @param tx_queue_id
2043 * The index of the tx queue to update the ring.
2044 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2045 * to rte_eth_dev_configure().
2047 * - 0: Success, the transmit queue is started.
2048 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2049 * - -EIO: if device is removed.
2050 * - -ENOTSUP: The function not supported in PMD driver.
2052 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2055 * Stop specified TX queue of a port
2058 * The port identifier of the Ethernet device
2059 * @param tx_queue_id
2060 * The index of the tx queue to update the ring.
2061 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2062 * to rte_eth_dev_configure().
2064 * - 0: Success, the transmit queue is stopped.
2065 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2066 * - -EIO: if device is removed.
2067 * - -ENOTSUP: The function not supported in PMD driver.
2069 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2072 * Start an Ethernet device.
2074 * The device start step is the last one and consists of setting the configured
2075 * offload features and in starting the transmit and the receive units of the
2078 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before
2079 * PMD port start callback function is invoked.
2081 * On success, all basic functions exported by the Ethernet API (link status,
2082 * receive/transmit, and so on) can be invoked.
2085 * The port identifier of the Ethernet device.
2087 * - 0: Success, Ethernet device started.
2088 * - <0: Error code of the driver device start function.
2090 int rte_eth_dev_start(uint16_t port_id);
2093 * Stop an Ethernet device. The device can be restarted with a call to
2094 * rte_eth_dev_start()
2097 * The port identifier of the Ethernet device.
2099 void rte_eth_dev_stop(uint16_t port_id);
2102 * Link up an Ethernet device.
2104 * Set device link up will re-enable the device rx/tx
2105 * functionality after it is previously set device linked down.
2108 * The port identifier of the Ethernet device.
2110 * - 0: Success, Ethernet device linked up.
2111 * - <0: Error code of the driver device link up function.
2113 int rte_eth_dev_set_link_up(uint16_t port_id);
2116 * Link down an Ethernet device.
2117 * The device rx/tx functionality will be disabled if success,
2118 * and it can be re-enabled with a call to
2119 * rte_eth_dev_set_link_up()
2122 * The port identifier of the Ethernet device.
2124 int rte_eth_dev_set_link_down(uint16_t port_id);
2127 * Close a stopped Ethernet device. The device cannot be restarted!
2128 * The function frees all port resources if the driver supports
2129 * the flag RTE_ETH_DEV_CLOSE_REMOVE.
2132 * The port identifier of the Ethernet device.
2134 void rte_eth_dev_close(uint16_t port_id);
2137 * Reset a Ethernet device and keep its port id.
2139 * When a port has to be reset passively, the DPDK application can invoke
2140 * this function. For example when a PF is reset, all its VFs should also
2141 * be reset. Normally a DPDK application can invoke this function when
2142 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2143 * a port reset in other circumstances.
2145 * When this function is called, it first stops the port and then calls the
2146 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2147 * state, in which no Tx and Rx queues are setup, as if the port has been
2148 * reset and not started. The port keeps the port id it had before the
2151 * After calling rte_eth_dev_reset( ), the application should use
2152 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2153 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2154 * to reconfigure the device as appropriate.
2156 * Note: To avoid unexpected behavior, the application should stop calling
2157 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2158 * safety, all these controlling functions should be called from the same
2162 * The port identifier of the Ethernet device.
2165 * - (0) if successful.
2166 * - (-EINVAL) if port identifier is invalid.
2167 * - (-ENOTSUP) if hardware doesn't support this function.
2168 * - (-EPERM) if not ran from the primary process.
2169 * - (-EIO) if re-initialisation failed or device is removed.
2170 * - (-ENOMEM) if the reset failed due to OOM.
2171 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2173 int rte_eth_dev_reset(uint16_t port_id);
2176 * Enable receipt in promiscuous mode for an Ethernet device.
2179 * The port identifier of the Ethernet device.
2181 * - (0) if successful.
2182 * - (-ENOTSUP) if support for promiscuous_enable() does not exist
2184 * - (-ENODEV) if *port_id* invalid.
2186 int rte_eth_promiscuous_enable(uint16_t port_id);
2189 * Disable receipt in promiscuous mode for an Ethernet device.
2192 * The port identifier of the Ethernet device.
2194 * - (0) if successful.
2195 * - (-ENOTSUP) if support for promiscuous_disable() does not exist
2197 * - (-ENODEV) if *port_id* invalid.
2199 int rte_eth_promiscuous_disable(uint16_t port_id);
2202 * Return the value of promiscuous mode for an Ethernet device.
2205 * The port identifier of the Ethernet device.
2207 * - (1) if promiscuous is enabled
2208 * - (0) if promiscuous is disabled.
2211 int rte_eth_promiscuous_get(uint16_t port_id);
2214 * Enable the receipt of any multicast frame by an Ethernet device.
2217 * The port identifier of the Ethernet device.
2219 * - (0) if successful.
2220 * - (-ENOTSUP) if support for allmulticast_enable() does not exist
2222 * - (-ENODEV) if *port_id* invalid.
2224 int rte_eth_allmulticast_enable(uint16_t port_id);
2227 * Disable the receipt of all multicast frames by an Ethernet device.
2230 * The port identifier of the Ethernet device.
2232 * - (0) if successful.
2233 * - (-ENOTSUP) if support for allmulticast_disable() does not exist
2235 * - (-ENODEV) if *port_id* invalid.
2237 int rte_eth_allmulticast_disable(uint16_t port_id);
2240 * Return the value of allmulticast mode for an Ethernet device.
2243 * The port identifier of the Ethernet device.
2245 * - (1) if allmulticast is enabled
2246 * - (0) if allmulticast is disabled.
2249 int rte_eth_allmulticast_get(uint16_t port_id);
2252 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
2253 * or FULL-DUPLEX) of the physical link of an Ethernet device. It might need
2254 * to wait up to 9 seconds in it.
2257 * The port identifier of the Ethernet device.
2259 * A pointer to an *rte_eth_link* structure to be filled with
2260 * the status, the speed and the mode of the Ethernet device link.
2262 * - (0) if successful.
2263 * - (-ENOTSUP) if the function is not supported in PMD driver.
2264 * - (-ENODEV) if *port_id* invalid.
2266 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
2269 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
2270 * or FULL-DUPLEX) of the physical link of an Ethernet device. It is a no-wait
2271 * version of rte_eth_link_get().
2274 * The port identifier of the Ethernet device.
2276 * A pointer to an *rte_eth_link* structure to be filled with
2277 * the status, the speed and the mode of the Ethernet device link.
2279 * - (0) if successful.
2280 * - (-ENOTSUP) if the function is not supported in PMD driver.
2281 * - (-ENODEV) if *port_id* invalid.
2283 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
2286 * Retrieve the general I/O statistics of an Ethernet device.
2289 * The port identifier of the Ethernet device.
2291 * A pointer to a structure of type *rte_eth_stats* to be filled with
2292 * the values of device counters for the following set of statistics:
2293 * - *ipackets* with the total of successfully received packets.
2294 * - *opackets* with the total of successfully transmitted packets.
2295 * - *ibytes* with the total of successfully received bytes.
2296 * - *obytes* with the total of successfully transmitted bytes.
2297 * - *ierrors* with the total of erroneous received packets.
2298 * - *oerrors* with the total of failed transmitted packets.
2300 * Zero if successful. Non-zero otherwise.
2302 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
2305 * Reset the general I/O statistics of an Ethernet device.
2308 * The port identifier of the Ethernet device.
2310 * - (0) if device notified to reset stats.
2311 * - (-ENOTSUP) if hardware doesn't support.
2312 * - (-ENODEV) if *port_id* invalid.
2313 * - (<0): Error code of the driver stats reset function.
2315 int rte_eth_stats_reset(uint16_t port_id);
2318 * Retrieve names of extended statistics of an Ethernet device.
2320 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
2322 * xstats_names[i].name => xstats[i].value
2324 * And the array index is same with id field of 'struct rte_eth_xstat':
2327 * This assumption makes key-value pair matching less flexible but simpler.
2330 * The port identifier of the Ethernet device.
2331 * @param xstats_names
2332 * An rte_eth_xstat_name array of at least *size* elements to
2333 * be filled. If set to NULL, the function returns the required number
2336 * The size of the xstats_names array (number of elements).
2338 * - A positive value lower or equal to size: success. The return value
2339 * is the number of entries filled in the stats table.
2340 * - A positive value higher than size: error, the given statistics table
2341 * is too small. The return value corresponds to the size that should
2342 * be given to succeed. The entries in the table are not valid and
2343 * shall not be used by the caller.
2344 * - A negative value on error (invalid port id).
2346 int rte_eth_xstats_get_names(uint16_t port_id,
2347 struct rte_eth_xstat_name *xstats_names,
2351 * Retrieve extended statistics of an Ethernet device.
2353 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
2355 * xstats_names[i].name => xstats[i].value
2357 * And the array index is same with id field of 'struct rte_eth_xstat':
2360 * This assumption makes key-value pair matching less flexible but simpler.
2363 * The port identifier of the Ethernet device.
2365 * A pointer to a table of structure of type *rte_eth_xstat*
2366 * to be filled with device statistics ids and values.
2367 * This parameter can be set to NULL if n is 0.
2369 * The size of the xstats array (number of elements).
2371 * - A positive value lower or equal to n: success. The return value
2372 * is the number of entries filled in the stats table.
2373 * - A positive value higher than n: error, the given statistics table
2374 * is too small. The return value corresponds to the size that should
2375 * be given to succeed. The entries in the table are not valid and
2376 * shall not be used by the caller.
2377 * - A negative value on error (invalid port id).
2379 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
2383 * Retrieve names of extended statistics of an Ethernet device.
2386 * The port identifier of the Ethernet device.
2387 * @param xstats_names
2388 * An rte_eth_xstat_name array of at least *size* elements to
2389 * be filled. If set to NULL, the function returns the required number
2392 * IDs array given by app to retrieve specific statistics
2394 * The size of the xstats_names array (number of elements).
2396 * - A positive value lower or equal to size: success. The return value
2397 * is the number of entries filled in the stats table.
2398 * - A positive value higher than size: error, the given statistics table
2399 * is too small. The return value corresponds to the size that should
2400 * be given to succeed. The entries in the table are not valid and
2401 * shall not be used by the caller.
2402 * - A negative value on error (invalid port id).
2405 rte_eth_xstats_get_names_by_id(uint16_t port_id,
2406 struct rte_eth_xstat_name *xstats_names, unsigned int size,
2410 * Retrieve extended statistics of an Ethernet device.
2413 * The port identifier of the Ethernet device.
2415 * A pointer to an ids array passed by application. This tells which
2416 * statistics values function should retrieve. This parameter
2417 * can be set to NULL if size is 0. In this case function will retrieve
2418 * all available statistics.
2420 * A pointer to a table to be filled with device statistics values.
2422 * The size of the ids array (number of elements).
2424 * - A positive value lower or equal to size: success. The return value
2425 * is the number of entries filled in the stats table.
2426 * - A positive value higher than size: error, the given statistics table
2427 * is too small. The return value corresponds to the size that should
2428 * be given to succeed. The entries in the table are not valid and
2429 * shall not be used by the caller.
2430 * - A negative value on error (invalid port id).
2432 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
2433 uint64_t *values, unsigned int size);
2436 * Gets the ID of a statistic from its name.
2438 * This function searches for the statistics using string compares, and
2439 * as such should not be used on the fast-path. For fast-path retrieval of
2440 * specific statistics, store the ID as provided in *id* from this function,
2441 * and pass the ID to rte_eth_xstats_get()
2443 * @param port_id The port to look up statistics from
2444 * @param xstat_name The name of the statistic to return
2445 * @param[out] id A pointer to an app-supplied uint64_t which should be
2446 * set to the ID of the stat if the stat exists.
2449 * -ENODEV for invalid port_id,
2450 * -EIO if device is removed,
2451 * -EINVAL if the xstat_name doesn't exist in port_id
2453 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
2457 * Reset extended statistics of an Ethernet device.
2460 * The port identifier of the Ethernet device.
2462 * - (0) if device notified to reset extended stats.
2463 * - (-ENOTSUP) if pmd doesn't support both
2464 * extended stats and basic stats reset.
2465 * - (-ENODEV) if *port_id* invalid.
2466 * - (<0): Error code of the driver xstats reset function.
2468 int rte_eth_xstats_reset(uint16_t port_id);
2471 * Set a mapping for the specified transmit queue to the specified per-queue
2472 * statistics counter.
2475 * The port identifier of the Ethernet device.
2476 * @param tx_queue_id
2477 * The index of the transmit queue for which a queue stats mapping is required.
2478 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2479 * to rte_eth_dev_configure().
2481 * The per-queue packet statistics functionality number that the transmit
2482 * queue is to be assigned.
2483 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
2485 * Zero if successful. Non-zero otherwise.
2487 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
2488 uint16_t tx_queue_id, uint8_t stat_idx);
2491 * Set a mapping for the specified receive queue to the specified per-queue
2492 * statistics counter.
2495 * The port identifier of the Ethernet device.
2496 * @param rx_queue_id
2497 * The index of the receive queue for which a queue stats mapping is required.
2498 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2499 * to rte_eth_dev_configure().
2501 * The per-queue packet statistics functionality number that the receive
2502 * queue is to be assigned.
2503 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
2505 * Zero if successful. Non-zero otherwise.
2507 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
2508 uint16_t rx_queue_id,
2512 * Retrieve the Ethernet address of an Ethernet device.
2515 * The port identifier of the Ethernet device.
2517 * A pointer to a structure of type *ether_addr* to be filled with
2518 * the Ethernet address of the Ethernet device.
2520 * - (0) if successful
2521 * - (-ENODEV) if *port_id* invalid.
2523 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr);
2526 * Retrieve the contextual information of an Ethernet device.
2528 * As part of this function, a number of of fields in dev_info will be
2529 * initialized as follows:
2534 * Where lim is defined within the rte_eth_dev_info_get as
2536 * const struct rte_eth_desc_lim lim = {
2537 * .nb_max = UINT16_MAX,
2540 * .nb_seg_max = UINT16_MAX,
2541 * .nb_mtu_seg_max = UINT16_MAX,
2544 * device = dev->device
2545 * min_mtu = RTE_ETHER_MIN_MTU
2546 * max_mtu = UINT16_MAX
2548 * The following fields will be populated if support for dev_infos_get()
2549 * exists for the device and the rte_eth_dev 'dev' has been populated
2550 * successfully with a call to it:
2552 * driver_name = dev->device->driver->name
2553 * nb_rx_queues = dev->data->nb_rx_queues
2554 * nb_tx_queues = dev->data->nb_tx_queues
2555 * dev_flags = &dev->data->dev_flags
2558 * The port identifier of the Ethernet device.
2560 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
2561 * the contextual information of the Ethernet device.
2563 * - (0) if successful.
2564 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device.
2565 * - (-ENODEV) if *port_id* invalid.
2567 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
2570 * Retrieve the firmware version of a device.
2573 * The port identifier of the device.
2575 * A pointer to a string array storing the firmware version of a device,
2576 * the string includes terminating null. This pointer is allocated by caller.
2578 * The size of the string array pointed by fw_version, which should be
2579 * large enough to store firmware version of the device.
2581 * - (0) if successful.
2582 * - (-ENOTSUP) if operation is not supported.
2583 * - (-ENODEV) if *port_id* invalid.
2584 * - (-EIO) if device is removed.
2585 * - (>0) if *fw_size* is not enough to store firmware version, return
2586 * the size of the non truncated string.
2588 int rte_eth_dev_fw_version_get(uint16_t port_id,
2589 char *fw_version, size_t fw_size);
2592 * Retrieve the supported packet types of an Ethernet device.
2594 * When a packet type is announced as supported, it *must* be recognized by
2595 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
2596 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
2597 * packet types for these packets:
2598 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
2599 * - Ether/Vlan/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
2600 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
2601 * - Ether/Vlan/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
2603 * When a packet is received by a PMD, the most precise type must be
2604 * returned among the ones supported. However a PMD is allowed to set
2605 * packet type that is not in the supported list, at the condition that it
2606 * is more precise. Therefore, a PMD announcing no supported packet types
2607 * can still set a matching packet type in a received packet.
2610 * Better to invoke this API after the device is already started or rx burst
2611 * function is decided, to obtain correct supported ptypes.
2613 * if a given PMD does not report what ptypes it supports, then the supported
2614 * ptype count is reported as 0.
2616 * The port identifier of the Ethernet device.
2618 * A hint of what kind of packet type which the caller is interested in.
2620 * An array pointer to store adequate packet types, allocated by caller.
2622 * Size of the array pointed by param ptypes.
2624 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
2625 * only num entries will be filled into the ptypes array, but the full
2626 * count of supported ptypes will be returned.
2627 * - (-ENODEV) if *port_id* invalid.
2629 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
2630 uint32_t *ptypes, int num);
2633 * @b EXPERIMENTAL: this API may change without prior notice.
2635 * Inform Ethernet device about reduced range of packet types to handle.
2637 * Application can use this function to set only specific ptypes that it's
2638 * interested. This information can be used by the PMD to optimize Rx path.
2640 * The function accepts an array `set_ptypes` allocated by the caller to
2641 * store the packet types set by the driver, the last element of the array
2642 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be
2643 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled
2647 * The port identifier of the Ethernet device.
2649 * The ptype family that application is interested in should be bitwise OR of
2650 * RTE_PTYPE_*_MASK or 0.
2652 * An array pointer to store set packet types, allocated by caller. The
2653 * function marks the end of array with RTE_PTYPE_UNKNOWN.
2655 * Size of the array pointed by param ptypes.
2656 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the
2660 * - (-ENODEV) if *port_id* invalid.
2661 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and
2665 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask,
2666 uint32_t *set_ptypes, unsigned int num);
2669 * Retrieve the MTU of an Ethernet device.
2672 * The port identifier of the Ethernet device.
2674 * A pointer to a uint16_t where the retrieved MTU is to be stored.
2676 * - (0) if successful.
2677 * - (-ENODEV) if *port_id* invalid.
2679 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
2682 * Change the MTU of an Ethernet device.
2685 * The port identifier of the Ethernet device.
2687 * A uint16_t for the MTU to be applied.
2689 * - (0) if successful.
2690 * - (-ENOTSUP) if operation is not supported.
2691 * - (-ENODEV) if *port_id* invalid.
2692 * - (-EIO) if device is removed.
2693 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within
2694 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or
2695 * when the mtu is set using dev->dev_ops->mtu_set.
2696 * - (-EBUSY) if operation is not allowed when the port is running
2698 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
2701 * Enable/Disable hardware filtering by an Ethernet device of received
2702 * VLAN packets tagged with a given VLAN Tag Identifier.
2705 * The port identifier of the Ethernet device.
2707 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
2709 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
2710 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
2712 * - (0) if successful.
2713 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
2714 * - (-ENODEV) if *port_id* invalid.
2715 * - (-EIO) if device is removed.
2716 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
2717 * - (-EINVAL) if *vlan_id* > 4095.
2719 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
2722 * Enable/Disable hardware VLAN Strip by a rx queue of an Ethernet device.
2723 * 82599/X540/X550 can support VLAN stripping at the rx queue level
2726 * The port identifier of the Ethernet device.
2727 * @param rx_queue_id
2728 * The index of the receive queue for which a queue stats mapping is required.
2729 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2730 * to rte_eth_dev_configure().
2732 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
2733 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
2735 * - (0) if successful.
2736 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured.
2737 * - (-ENODEV) if *port_id* invalid.
2738 * - (-EINVAL) if *rx_queue_id* invalid.
2740 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
2744 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
2745 * the VLAN Header. This is a register setup available on some Intel NIC, not
2746 * but all, please check the data sheet for availability.
2749 * The port identifier of the Ethernet device.
2753 * The Tag Protocol ID
2755 * - (0) if successful.
2756 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported.
2757 * - (-ENODEV) if *port_id* invalid.
2758 * - (-EIO) if device is removed.
2760 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
2761 enum rte_vlan_type vlan_type,
2765 * Set VLAN offload configuration on an Ethernet device
2766 * Enable/Disable Extended VLAN by an Ethernet device, This is a register setup
2767 * available on some Intel NIC, not but all, please check the data sheet for
2769 * Enable/Disable VLAN Strip can be done on rx queue for certain NIC, but here
2770 * the configuration is applied on the port level.
2773 * The port identifier of the Ethernet device.
2774 * @param offload_mask
2775 * The VLAN Offload bit mask can be mixed use with "OR"
2776 * ETH_VLAN_STRIP_OFFLOAD
2777 * ETH_VLAN_FILTER_OFFLOAD
2778 * ETH_VLAN_EXTEND_OFFLOAD
2779 * ETH_QINQ_STRIP_OFFLOAD
2781 * - (0) if successful.
2782 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
2783 * - (-ENODEV) if *port_id* invalid.
2784 * - (-EIO) if device is removed.
2786 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
2789 * Read VLAN Offload configuration from an Ethernet device
2792 * The port identifier of the Ethernet device.
2794 * - (>0) if successful. Bit mask to indicate
2795 * ETH_VLAN_STRIP_OFFLOAD
2796 * ETH_VLAN_FILTER_OFFLOAD
2797 * ETH_VLAN_EXTEND_OFFLOAD
2798 * ETH_QINQ_STRIP_OFFLOAD
2799 * - (-ENODEV) if *port_id* invalid.
2801 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
2804 * Set port based TX VLAN insertion on or off.
2807 * The port identifier of the Ethernet device.
2809 * Port based TX VLAN identifier together with user priority.
2811 * Turn on or off the port based TX VLAN insertion.
2814 * - (0) if successful.
2815 * - negative if failed.
2817 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
2819 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
2823 * Structure used to buffer packets for future TX
2824 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
2826 struct rte_eth_dev_tx_buffer {
2827 buffer_tx_error_fn error_callback;
2828 void *error_userdata;
2829 uint16_t size; /**< Size of buffer for buffered tx */
2830 uint16_t length; /**< Number of packets in the array */
2831 struct rte_mbuf *pkts[];
2832 /**< Pending packets to be sent on explicit flush or when full */
2836 * Calculate the size of the tx buffer.
2839 * Number of stored packets.
2841 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
2842 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
2845 * Initialize default values for buffered transmitting
2848 * Tx buffer to be initialized.
2855 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
2858 * Configure a callback for buffered packets which cannot be sent
2860 * Register a specific callback to be called when an attempt is made to send
2861 * all packets buffered on an ethernet port, but not all packets can
2862 * successfully be sent. The callback registered here will be called only
2863 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
2864 * The default callback configured for each queue by default just frees the
2865 * packets back to the calling mempool. If additional behaviour is required,
2866 * for example, to count dropped packets, or to retry transmission of packets
2867 * which cannot be sent, this function should be used to register a suitable
2868 * callback function to implement the desired behaviour.
2869 * The example callback "rte_eth_count_unsent_packet_callback()" is also
2870 * provided as reference.
2873 * The port identifier of the Ethernet device.
2875 * The function to be used as the callback.
2877 * Arbitrary parameter to be passed to the callback function
2879 * 0 on success, or -1 on error with rte_errno set appropriately
2882 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
2883 buffer_tx_error_fn callback, void *userdata);
2886 * Callback function for silently dropping unsent buffered packets.
2888 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
2889 * adjust the default behavior when buffered packets cannot be sent. This
2890 * function drops any unsent packets silently and is used by tx buffered
2891 * operations as default behavior.
2893 * NOTE: this function should not be called directly, instead it should be used
2894 * as a callback for packet buffering.
2896 * NOTE: when configuring this function as a callback with
2897 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
2898 * should point to an uint64_t value.
2901 * The previously buffered packets which could not be sent
2903 * The number of unsent packets in the pkts array
2908 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
2912 * Callback function for tracking unsent buffered packets.
2914 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
2915 * adjust the default behavior when buffered packets cannot be sent. This
2916 * function drops any unsent packets, but also updates a user-supplied counter
2917 * to track the overall number of packets dropped. The counter should be an
2918 * uint64_t variable.
2920 * NOTE: this function should not be called directly, instead it should be used
2921 * as a callback for packet buffering.
2923 * NOTE: when configuring this function as a callback with
2924 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
2925 * should point to an uint64_t value.
2928 * The previously buffered packets which could not be sent
2930 * The number of unsent packets in the pkts array
2932 * Pointer to an uint64_t value, which will be incremented by unsent
2935 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
2939 * Request the driver to free mbufs currently cached by the driver. The
2940 * driver will only free the mbuf if it is no longer in use. It is the
2941 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is
2945 * The port identifier of the Ethernet device.
2947 * The index of the transmit queue through which output packets must be
2949 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2950 * to rte_eth_dev_configure().
2952 * Maximum number of packets to free. Use 0 to indicate all possible packets
2953 * should be freed. Note that a packet may be using multiple mbufs.
2956 * -ENODEV: Invalid interface
2957 * -EIO: device is removed
2958 * -ENOTSUP: Driver does not support function
2960 * 0-n: Number of packets freed. More packets may still remain in ring that
2964 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
2967 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
2970 enum rte_eth_event_ipsec_subtype {
2971 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
2972 /**< Unknown event type */
2973 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
2974 /**< Sequence number overflow */
2975 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
2976 /**< Soft time expiry of SA */
2977 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
2978 /**< Soft byte expiry of SA */
2979 RTE_ETH_EVENT_IPSEC_MAX
2980 /**< Max value of this enum */
2984 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
2985 * information of the IPsec offload event.
2987 struct rte_eth_event_ipsec_desc {
2988 enum rte_eth_event_ipsec_subtype subtype;
2989 /**< Type of RTE_ETH_EVENT_IPSEC_* event */
2991 /**< Event specific metadata
2993 * For the following events, *userdata* registered
2994 * with the *rte_security_session* would be returned
2997 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
2998 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
2999 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
3001 * @see struct rte_security_session_conf
3007 * The eth device event type for interrupt, and maybe others in the future.
3009 enum rte_eth_event_type {
3010 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3011 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3012 RTE_ETH_EVENT_QUEUE_STATE,
3013 /**< queue state event (enabled/disabled) */
3014 RTE_ETH_EVENT_INTR_RESET,
3015 /**< reset interrupt event, sent to VF on PF reset */
3016 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3017 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3018 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3019 RTE_ETH_EVENT_NEW, /**< port is probed */
3020 RTE_ETH_EVENT_DESTROY, /**< port is released */
3021 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
3022 RTE_ETH_EVENT_MAX /**< max value of this enum */
3025 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3026 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3027 /**< user application callback to be registered for interrupts */
3030 * Register a callback function for port event.
3034 * RTE_ETH_ALL means register the event for all port ids.
3038 * User supplied callback function to be called.
3040 * Pointer to the parameters for the registered callback.
3043 * - On success, zero.
3044 * - On failure, a negative value.
3046 int rte_eth_dev_callback_register(uint16_t port_id,
3047 enum rte_eth_event_type event,
3048 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3051 * Unregister a callback function for port event.
3055 * RTE_ETH_ALL means unregister the event for all port ids.
3059 * User supplied callback function to be called.
3061 * Pointer to the parameters for the registered callback. -1 means to
3062 * remove all for the same callback address and same event.
3065 * - On success, zero.
3066 * - On failure, a negative value.
3068 int rte_eth_dev_callback_unregister(uint16_t port_id,
3069 enum rte_eth_event_type event,
3070 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3073 * When there is no rx packet coming in Rx Queue for a long time, we can
3074 * sleep lcore related to RX Queue for power saving, and enable rx interrupt
3075 * to be triggered when Rx packet arrives.
3077 * The rte_eth_dev_rx_intr_enable() function enables rx queue
3078 * interrupt on specific rx queue of a port.
3081 * The port identifier of the Ethernet device.
3083 * The index of the receive queue from which to retrieve input packets.
3084 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3085 * to rte_eth_dev_configure().
3087 * - (0) if successful.
3088 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3090 * - (-ENODEV) if *port_id* invalid.
3091 * - (-EIO) if device is removed.
3093 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
3096 * When lcore wakes up from rx interrupt indicating packet coming, disable rx
3097 * interrupt and returns to polling mode.
3099 * The rte_eth_dev_rx_intr_disable() function disables rx queue
3100 * interrupt on specific rx queue of a port.
3103 * The port identifier of the Ethernet device.
3105 * The index of the receive queue from which to retrieve input packets.
3106 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3107 * to rte_eth_dev_configure().
3109 * - (0) if successful.
3110 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3112 * - (-ENODEV) if *port_id* invalid.
3113 * - (-EIO) if device is removed.
3115 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
3118 * RX Interrupt control per port.
3121 * The port identifier of the Ethernet device.
3123 * Epoll instance fd which the intr vector associated to.
3124 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3126 * The operation be performed for the vector.
3127 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3131 * - On success, zero.
3132 * - On failure, a negative value.
3134 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
3137 * RX Interrupt control per queue.
3140 * The port identifier of the Ethernet device.
3142 * The index of the receive queue from which to retrieve input packets.
3143 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3144 * to rte_eth_dev_configure().
3146 * Epoll instance fd which the intr vector associated to.
3147 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3149 * The operation be performed for the vector.
3150 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3154 * - On success, zero.
3155 * - On failure, a negative value.
3157 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
3158 int epfd, int op, void *data);
3162 * @b EXPERIMENTAL: this API may change without prior notice.
3164 * Get interrupt fd per Rx queue.
3167 * The port identifier of the Ethernet device.
3169 * The index of the receive queue from which to retrieve input packets.
3170 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3171 * to rte_eth_dev_configure().
3173 * - (>=0) the interrupt fd associated to the requested Rx queue if
3179 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id);
3182 * Turn on the LED on the Ethernet device.
3183 * This function turns on the LED on the Ethernet device.
3186 * The port identifier of the Ethernet device.
3188 * - (0) if successful.
3189 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3191 * - (-ENODEV) if *port_id* invalid.
3192 * - (-EIO) if device is removed.
3194 int rte_eth_led_on(uint16_t port_id);
3197 * Turn off the LED on the Ethernet device.
3198 * This function turns off the LED on the Ethernet device.
3201 * The port identifier of the Ethernet device.
3203 * - (0) if successful.
3204 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3206 * - (-ENODEV) if *port_id* invalid.
3207 * - (-EIO) if device is removed.
3209 int rte_eth_led_off(uint16_t port_id);
3212 * Get current status of the Ethernet link flow control for Ethernet device
3215 * The port identifier of the Ethernet device.
3217 * The pointer to the structure where to store the flow control parameters.
3219 * - (0) if successful.
3220 * - (-ENOTSUP) if hardware doesn't support flow control.
3221 * - (-ENODEV) if *port_id* invalid.
3222 * - (-EIO) if device is removed.
3224 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
3225 struct rte_eth_fc_conf *fc_conf);
3228 * Configure the Ethernet link flow control for Ethernet device
3231 * The port identifier of the Ethernet device.
3233 * The pointer to the structure of the flow control parameters.
3235 * - (0) if successful.
3236 * - (-ENOTSUP) if hardware doesn't support flow control mode.
3237 * - (-ENODEV) if *port_id* invalid.
3238 * - (-EINVAL) if bad parameter
3239 * - (-EIO) if flow control setup failure or device is removed.
3241 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
3242 struct rte_eth_fc_conf *fc_conf);
3245 * Configure the Ethernet priority flow control under DCB environment
3246 * for Ethernet device.
3249 * The port identifier of the Ethernet device.
3251 * The pointer to the structure of the priority flow control parameters.
3253 * - (0) if successful.
3254 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
3255 * - (-ENODEV) if *port_id* invalid.
3256 * - (-EINVAL) if bad parameter
3257 * - (-EIO) if flow control setup failure or device is removed.
3259 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
3260 struct rte_eth_pfc_conf *pfc_conf);
3263 * Add a MAC address to an internal array of addresses used to enable whitelist
3264 * filtering to accept packets only if the destination MAC address matches.
3267 * The port identifier of the Ethernet device.
3269 * The MAC address to add.
3271 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
3272 * not enabled, this should be set to 0.
3274 * - (0) if successfully added or *mac_addr* was already added.
3275 * - (-ENOTSUP) if hardware doesn't support this feature.
3276 * - (-ENODEV) if *port* is invalid.
3277 * - (-EIO) if device is removed.
3278 * - (-ENOSPC) if no more MAC addresses can be added.
3279 * - (-EINVAL) if MAC address is invalid.
3281 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr,
3285 * Remove a MAC address from the internal array of addresses.
3288 * The port identifier of the Ethernet device.
3290 * MAC address to remove.
3292 * - (0) if successful, or *mac_addr* didn't exist.
3293 * - (-ENOTSUP) if hardware doesn't support.
3294 * - (-ENODEV) if *port* invalid.
3295 * - (-EADDRINUSE) if attempting to remove the default MAC address
3297 int rte_eth_dev_mac_addr_remove(uint16_t port_id,
3298 struct rte_ether_addr *mac_addr);
3301 * Set the default MAC address.
3304 * The port identifier of the Ethernet device.
3306 * New default MAC address.
3308 * - (0) if successful, or *mac_addr* didn't exist.
3309 * - (-ENOTSUP) if hardware doesn't support.
3310 * - (-ENODEV) if *port* invalid.
3311 * - (-EINVAL) if MAC address is invalid.
3313 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
3314 struct rte_ether_addr *mac_addr);
3317 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3320 * The port identifier of the Ethernet device.
3324 * Redirection table size. The table size can be queried by
3325 * rte_eth_dev_info_get().
3327 * - (0) if successful.
3328 * - (-ENOTSUP) if hardware doesn't support.
3329 * - (-EINVAL) if bad parameter.
3330 * - (-EIO) if device is removed.
3332 int rte_eth_dev_rss_reta_update(uint16_t port_id,
3333 struct rte_eth_rss_reta_entry64 *reta_conf,
3334 uint16_t reta_size);
3337 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3340 * The port identifier of the Ethernet device.
3342 * RETA to query. For each requested reta entry, corresponding bit
3343 * in mask must be set.
3345 * Redirection table size. The table size can be queried by
3346 * rte_eth_dev_info_get().
3348 * - (0) if successful.
3349 * - (-ENOTSUP) if hardware doesn't support.
3350 * - (-EINVAL) if bad parameter.
3351 * - (-EIO) if device is removed.
3353 int rte_eth_dev_rss_reta_query(uint16_t port_id,
3354 struct rte_eth_rss_reta_entry64 *reta_conf,
3355 uint16_t reta_size);
3358 * Updates unicast hash table for receiving packet with the given destination
3359 * MAC address, and the packet is routed to all VFs for which the RX mode is
3360 * accept packets that match the unicast hash table.
3363 * The port identifier of the Ethernet device.
3365 * Unicast MAC address.
3367 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
3368 * 0 - Clear an unicast hash bit.
3370 * - (0) if successful.
3371 * - (-ENOTSUP) if hardware doesn't support.
3372 * - (-ENODEV) if *port_id* invalid.
3373 * - (-EIO) if device is removed.
3374 * - (-EINVAL) if bad parameter.
3376 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr,
3380 * Updates all unicast hash bitmaps for receiving packet with any Unicast
3381 * Ethernet MAC addresses,the packet is routed to all VFs for which the RX
3382 * mode is accept packets that match the unicast hash table.
3385 * The port identifier of the Ethernet device.
3387 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
3389 * 0 - Clear all unicast hash bitmaps
3391 * - (0) if successful.
3392 * - (-ENOTSUP) if hardware doesn't support.
3393 * - (-ENODEV) if *port_id* invalid.
3394 * - (-EIO) if device is removed.
3395 * - (-EINVAL) if bad parameter.
3397 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
3400 * Set a traffic mirroring rule on an Ethernet device
3403 * The port identifier of the Ethernet device.
3404 * @param mirror_conf
3405 * The pointer to the traffic mirroring structure describing the mirroring rule.
3406 * The *rte_eth_vm_mirror_conf* structure includes the type of mirroring rule,
3407 * destination pool and the value of rule if enable vlan or pool mirroring.
3410 * The index of traffic mirroring rule, we support four separated rules.
3412 * 1 - Enable a mirroring rule.
3413 * 0 - Disable a mirroring rule.
3415 * - (0) if successful.
3416 * - (-ENOTSUP) if hardware doesn't support this feature.
3417 * - (-ENODEV) if *port_id* invalid.
3418 * - (-EIO) if device is removed.
3419 * - (-EINVAL) if the mr_conf information is not correct.
3421 int rte_eth_mirror_rule_set(uint16_t port_id,
3422 struct rte_eth_mirror_conf *mirror_conf,
3427 * Reset a traffic mirroring rule on an Ethernet device.
3430 * The port identifier of the Ethernet device.
3432 * The index of traffic mirroring rule, we support four separated rules.
3434 * - (0) if successful.
3435 * - (-ENOTSUP) if hardware doesn't support this feature.
3436 * - (-ENODEV) if *port_id* invalid.
3437 * - (-EIO) if device is removed.
3438 * - (-EINVAL) if bad parameter.
3440 int rte_eth_mirror_rule_reset(uint16_t port_id,
3444 * Set the rate limitation for a queue on an Ethernet device.
3447 * The port identifier of the Ethernet device.
3451 * The tx rate in Mbps. Allocated from the total port link speed.
3453 * - (0) if successful.
3454 * - (-ENOTSUP) if hardware doesn't support this feature.
3455 * - (-ENODEV) if *port_id* invalid.
3456 * - (-EIO) if device is removed.
3457 * - (-EINVAL) if bad parameter.
3459 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
3463 * Configuration of Receive Side Scaling hash computation of Ethernet device.
3466 * The port identifier of the Ethernet device.
3468 * The new configuration to use for RSS hash computation on the port.
3470 * - (0) if successful.
3471 * - (-ENODEV) if port identifier is invalid.
3472 * - (-EIO) if device is removed.
3473 * - (-ENOTSUP) if hardware doesn't support.
3474 * - (-EINVAL) if bad parameter.
3476 int rte_eth_dev_rss_hash_update(uint16_t port_id,
3477 struct rte_eth_rss_conf *rss_conf);
3480 * Retrieve current configuration of Receive Side Scaling hash computation
3481 * of Ethernet device.
3484 * The port identifier of the Ethernet device.
3486 * Where to store the current RSS hash configuration of the Ethernet device.
3488 * - (0) if successful.
3489 * - (-ENODEV) if port identifier is invalid.
3490 * - (-EIO) if device is removed.
3491 * - (-ENOTSUP) if hardware doesn't support RSS.
3494 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
3495 struct rte_eth_rss_conf *rss_conf);
3498 * Add UDP tunneling port for a specific type of tunnel.
3499 * The packets with this UDP port will be identified as this type of tunnel.
3500 * Before enabling any offloading function for a tunnel, users can call this API
3501 * to change or add more UDP port for the tunnel. So the offloading function
3502 * can take effect on the packets with the specific UDP port.
3505 * The port identifier of the Ethernet device.
3507 * UDP tunneling configuration.
3510 * - (0) if successful.
3511 * - (-ENODEV) if port identifier is invalid.
3512 * - (-EIO) if device is removed.
3513 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3516 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
3517 struct rte_eth_udp_tunnel *tunnel_udp);
3520 * Delete UDP tunneling port a specific type of tunnel.
3521 * The packets with this UDP port will not be identified as this type of tunnel
3523 * Before enabling any offloading function for a tunnel, users can call this API
3524 * to delete a UDP port for the tunnel. So the offloading function will not take
3525 * effect on the packets with the specific UDP port.
3528 * The port identifier of the Ethernet device.
3530 * UDP tunneling configuration.
3533 * - (0) if successful.
3534 * - (-ENODEV) if port identifier is invalid.
3535 * - (-EIO) if device is removed.
3536 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3539 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
3540 struct rte_eth_udp_tunnel *tunnel_udp);
3543 * Check whether the filter type is supported on an Ethernet device.
3544 * All the supported filter types are defined in 'rte_eth_ctrl.h'.
3547 * The port identifier of the Ethernet device.
3548 * @param filter_type
3551 * - (0) if successful.
3552 * - (-ENOTSUP) if hardware doesn't support this filter type.
3553 * - (-ENODEV) if *port_id* invalid.
3554 * - (-EIO) if device is removed.
3557 int rte_eth_dev_filter_supported(uint16_t port_id,
3558 enum rte_filter_type filter_type);
3561 * Take operations to assigned filter type on an Ethernet device.
3562 * All the supported operations and filter types are defined in 'rte_eth_ctrl.h'.
3565 * The port identifier of the Ethernet device.
3566 * @param filter_type
3569 * Type of operation.
3571 * A pointer to arguments defined specifically for the operation.
3573 * - (0) if successful.
3574 * - (-ENOTSUP) if hardware doesn't support.
3575 * - (-ENODEV) if *port_id* invalid.
3576 * - (-EIO) if device is removed.
3577 * - others depends on the specific operations implementation.
3580 int rte_eth_dev_filter_ctrl(uint16_t port_id, enum rte_filter_type filter_type,
3581 enum rte_filter_op filter_op, void *arg);
3584 * Get DCB information on an Ethernet device.
3587 * The port identifier of the Ethernet device.
3591 * - (0) if successful.
3592 * - (-ENODEV) if port identifier is invalid.
3593 * - (-EIO) if device is removed.
3594 * - (-ENOTSUP) if hardware doesn't support.
3596 int rte_eth_dev_get_dcb_info(uint16_t port_id,
3597 struct rte_eth_dcb_info *dcb_info);
3599 struct rte_eth_rxtx_callback;
3602 * Add a callback to be called on packet RX on a given port and queue.
3604 * This API configures a function to be called for each burst of
3605 * packets received on a given NIC port queue. The return value is a pointer
3606 * that can be used to later remove the callback using
3607 * rte_eth_remove_rx_callback().
3609 * Multiple functions are called in the order that they are added.
3612 * The port identifier of the Ethernet device.
3614 * The queue on the Ethernet device on which the callback is to be added.
3616 * The callback function
3618 * A generic pointer parameter which will be passed to each invocation of the
3619 * callback function on this port and queue.
3623 * On success, a pointer value which can later be used to remove the callback.
3625 const struct rte_eth_rxtx_callback *
3626 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
3627 rte_rx_callback_fn fn, void *user_param);
3630 * Add a callback that must be called first on packet RX on a given port
3633 * This API configures a first function to be called for each burst of
3634 * packets received on a given NIC port queue. The return value is a pointer
3635 * that can be used to later remove the callback using
3636 * rte_eth_remove_rx_callback().
3638 * Multiple functions are called in the order that they are added.
3641 * The port identifier of the Ethernet device.
3643 * The queue on the Ethernet device on which the callback is to be added.
3645 * The callback function
3647 * A generic pointer parameter which will be passed to each invocation of the
3648 * callback function on this port and queue.
3652 * On success, a pointer value which can later be used to remove the callback.
3654 const struct rte_eth_rxtx_callback *
3655 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
3656 rte_rx_callback_fn fn, void *user_param);
3659 * Add a callback to be called on packet TX on a given port and queue.
3661 * This API configures a function to be called for each burst of
3662 * packets sent on a given NIC port queue. The return value is a pointer
3663 * that can be used to later remove the callback using
3664 * rte_eth_remove_tx_callback().
3666 * Multiple functions are called in the order that they are added.
3669 * The port identifier of the Ethernet device.
3671 * The queue on the Ethernet device on which the callback is to be added.
3673 * The callback function
3675 * A generic pointer parameter which will be passed to each invocation of the
3676 * callback function on this port and queue.
3680 * On success, a pointer value which can later be used to remove the callback.
3682 const struct rte_eth_rxtx_callback *
3683 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
3684 rte_tx_callback_fn fn, void *user_param);
3687 * Remove an RX packet callback from a given port and queue.
3689 * This function is used to removed callbacks that were added to a NIC port
3690 * queue using rte_eth_add_rx_callback().
3692 * Note: the callback is removed from the callback list but it isn't freed
3693 * since the it may still be in use. The memory for the callback can be
3694 * subsequently freed back by the application by calling rte_free():
3696 * - Immediately - if the port is stopped, or the user knows that no
3697 * callbacks are in flight e.g. if called from the thread doing RX/TX
3700 * - After a short delay - where the delay is sufficient to allow any
3701 * in-flight callbacks to complete.
3704 * The port identifier of the Ethernet device.
3706 * The queue on the Ethernet device from which the callback is to be removed.
3708 * User supplied callback created via rte_eth_add_rx_callback().
3711 * - 0: Success. Callback was removed.
3712 * - -ENOTSUP: Callback support is not available.
3713 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3714 * is NULL or not found for the port/queue.
3716 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
3717 const struct rte_eth_rxtx_callback *user_cb);
3720 * Remove a TX packet callback from a given port and queue.
3722 * This function is used to removed callbacks that were added to a NIC port
3723 * queue using rte_eth_add_tx_callback().
3725 * Note: the callback is removed from the callback list but it isn't freed
3726 * since the it may still be in use. The memory for the callback can be
3727 * subsequently freed back by the application by calling rte_free():
3729 * - Immediately - if the port is stopped, or the user knows that no
3730 * callbacks are in flight e.g. if called from the thread doing RX/TX
3733 * - After a short delay - where the delay is sufficient to allow any
3734 * in-flight callbacks to complete.
3737 * The port identifier of the Ethernet device.
3739 * The queue on the Ethernet device from which the callback is to be removed.
3741 * User supplied callback created via rte_eth_add_tx_callback().
3744 * - 0: Success. Callback was removed.
3745 * - -ENOTSUP: Callback support is not available.
3746 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3747 * is NULL or not found for the port/queue.
3749 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
3750 const struct rte_eth_rxtx_callback *user_cb);
3753 * Retrieve information about given port's RX queue.
3756 * The port identifier of the Ethernet device.
3758 * The RX queue on the Ethernet device for which information
3759 * will be retrieved.
3761 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
3762 * the information of the Ethernet device.
3766 * - -ENOTSUP: routine is not supported by the device PMD.
3767 * - -EINVAL: The port_id or the queue_id is out of range, or the queue
3770 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3771 struct rte_eth_rxq_info *qinfo);
3774 * Retrieve information about given port's TX queue.
3777 * The port identifier of the Ethernet device.
3779 * The TX queue on the Ethernet device for which information
3780 * will be retrieved.
3782 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
3783 * the information of the Ethernet device.
3787 * - -ENOTSUP: routine is not supported by the device PMD.
3788 * - -EINVAL: The port_id or the queue_id is out of range, or the queue
3791 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3792 struct rte_eth_txq_info *qinfo);
3795 * Retrieve information about the Rx packet burst mode.
3798 * The port identifier of the Ethernet device.
3800 * The Rx queue on the Ethernet device for which information
3801 * will be retrieved.
3803 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
3804 * with the information of the packet burst mode.
3808 * - -ENOTSUP: routine is not supported by the device PMD.
3809 * - -EINVAL: The port_id or the queue_id is out of range.
3812 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
3813 struct rte_eth_burst_mode *mode);
3816 * Retrieve information about the Tx packet burst mode.
3819 * The port identifier of the Ethernet device.
3821 * The Tx queue on the Ethernet device for which information
3822 * will be retrieved.
3824 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
3825 * with the information of the packet burst mode.
3829 * - -ENOTSUP: routine is not supported by the device PMD.
3830 * - -EINVAL: The port_id or the queue_id is out of range.
3833 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
3834 struct rte_eth_burst_mode *mode);
3837 * Retrieve device registers and register attributes (number of registers and
3841 * The port identifier of the Ethernet device.
3843 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
3844 * NULL the function fills in the width and length fields. If non-NULL
3845 * the registers are put into the buffer pointed at by the data field.
3847 * - (0) if successful.
3848 * - (-ENOTSUP) if hardware doesn't support.
3849 * - (-ENODEV) if *port_id* invalid.
3850 * - (-EIO) if device is removed.
3851 * - others depends on the specific operations implementation.
3853 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
3856 * Retrieve size of device EEPROM
3859 * The port identifier of the Ethernet device.
3861 * - (>=0) EEPROM size if successful.
3862 * - (-ENOTSUP) if hardware doesn't support.
3863 * - (-ENODEV) if *port_id* invalid.
3864 * - (-EIO) if device is removed.
3865 * - others depends on the specific operations implementation.
3867 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
3870 * Retrieve EEPROM and EEPROM attribute
3873 * The port identifier of the Ethernet device.
3875 * The template includes buffer for return EEPROM data and
3876 * EEPROM attributes to be filled.
3878 * - (0) if successful.
3879 * - (-ENOTSUP) if hardware doesn't support.
3880 * - (-ENODEV) if *port_id* invalid.
3881 * - (-EIO) if device is removed.
3882 * - others depends on the specific operations implementation.
3884 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
3887 * Program EEPROM with provided data
3890 * The port identifier of the Ethernet device.
3892 * The template includes EEPROM data for programming and
3893 * EEPROM attributes to be filled
3895 * - (0) if successful.
3896 * - (-ENOTSUP) if hardware doesn't support.
3897 * - (-ENODEV) if *port_id* invalid.
3898 * - (-EIO) if device is removed.
3899 * - others depends on the specific operations implementation.
3901 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
3905 * @b EXPERIMENTAL: this API may change without prior notice.
3907 * Retrieve the type and size of plugin module EEPROM
3910 * The port identifier of the Ethernet device.
3912 * The type and size of plugin module EEPROM.
3914 * - (0) if successful.
3915 * - (-ENOTSUP) if hardware doesn't support.
3916 * - (-ENODEV) if *port_id* invalid.
3917 * - (-EIO) if device is removed.
3918 * - others depends on the specific operations implementation.
3922 rte_eth_dev_get_module_info(uint16_t port_id,
3923 struct rte_eth_dev_module_info *modinfo);
3927 * @b EXPERIMENTAL: this API may change without prior notice.
3929 * Retrieve the data of plugin module EEPROM
3932 * The port identifier of the Ethernet device.
3934 * The template includes the plugin module EEPROM attributes, and the
3935 * buffer for return plugin module EEPROM data.
3937 * - (0) if successful.
3938 * - (-ENOTSUP) if hardware doesn't support.
3939 * - (-ENODEV) if *port_id* invalid.
3940 * - (-EIO) if device is removed.
3941 * - others depends on the specific operations implementation.
3945 rte_eth_dev_get_module_eeprom(uint16_t port_id,
3946 struct rte_dev_eeprom_info *info);
3949 * Set the list of multicast addresses to filter on an Ethernet device.
3952 * The port identifier of the Ethernet device.
3953 * @param mc_addr_set
3954 * The array of multicast addresses to set. Equal to NULL when the function
3955 * is invoked to flush the set of filtered addresses.
3957 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
3958 * when the function is invoked to flush the set of filtered addresses.
3960 * - (0) if successful.
3961 * - (-ENODEV) if *port_id* invalid.
3962 * - (-EIO) if device is removed.
3963 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
3964 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
3966 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
3967 struct rte_ether_addr *mc_addr_set,
3968 uint32_t nb_mc_addr);
3971 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
3974 * The port identifier of the Ethernet device.
3978 * - -ENODEV: The port ID is invalid.
3979 * - -EIO: if device is removed.
3980 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3982 int rte_eth_timesync_enable(uint16_t port_id);
3985 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
3988 * The port identifier of the Ethernet device.
3992 * - -ENODEV: The port ID is invalid.
3993 * - -EIO: if device is removed.
3994 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3996 int rte_eth_timesync_disable(uint16_t port_id);
3999 * Read an IEEE1588/802.1AS RX timestamp from an Ethernet device.
4002 * The port identifier of the Ethernet device.
4004 * Pointer to the timestamp struct.
4006 * Device specific flags. Used to pass the RX timesync register index to
4007 * i40e. Unused in igb/ixgbe, pass 0 instead.
4011 * - -EINVAL: No timestamp is available.
4012 * - -ENODEV: The port ID is invalid.
4013 * - -EIO: if device is removed.
4014 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4016 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
4017 struct timespec *timestamp, uint32_t flags);
4020 * Read an IEEE1588/802.1AS TX timestamp from an Ethernet device.
4023 * The port identifier of the Ethernet device.
4025 * Pointer to the timestamp struct.
4029 * - -EINVAL: No timestamp is available.
4030 * - -ENODEV: The port ID is invalid.
4031 * - -EIO: if device is removed.
4032 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4034 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
4035 struct timespec *timestamp);
4038 * Adjust the timesync clock on an Ethernet device.
4040 * This is usually used in conjunction with other Ethdev timesync functions to
4041 * synchronize the device time using the IEEE1588/802.1AS protocol.
4044 * The port identifier of the Ethernet device.
4046 * The adjustment in nanoseconds.
4050 * - -ENODEV: The port ID is invalid.
4051 * - -EIO: if device is removed.
4052 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4054 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
4057 * Read the time from the timesync clock on an Ethernet device.
4059 * This is usually used in conjunction with other Ethdev timesync functions to
4060 * synchronize the device time using the IEEE1588/802.1AS protocol.
4063 * The port identifier of the Ethernet device.
4065 * Pointer to the timespec struct that holds the time.
4070 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
4073 * Set the time of the timesync clock on an Ethernet device.
4075 * This is usually used in conjunction with other Ethdev timesync functions to
4076 * synchronize the device time using the IEEE1588/802.1AS protocol.
4079 * The port identifier of the Ethernet device.
4081 * Pointer to the timespec struct that holds the time.
4085 * - -EINVAL: No timestamp is available.
4086 * - -ENODEV: The port ID is invalid.
4087 * - -EIO: if device is removed.
4088 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4090 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
4094 * @b EXPERIMENTAL: this API may change without prior notice.
4096 * Read the current clock counter of an Ethernet device
4098 * This returns the current raw clock value of an Ethernet device. It is
4099 * a raw amount of ticks, with no given time reference.
4100 * The value returned here is from the same clock than the one
4101 * filling timestamp field of Rx packets when using hardware timestamp
4102 * offload. Therefore it can be used to compute a precise conversion of
4103 * the device clock to the real time.
4105 * E.g, a simple heuristic to derivate the frequency would be:
4106 * uint64_t start, end;
4107 * rte_eth_read_clock(port, start);
4108 * rte_delay_ms(100);
4109 * rte_eth_read_clock(port, end);
4110 * double freq = (end - start) * 10;
4112 * Compute a common reference with:
4113 * uint64_t base_time_sec = current_time();
4114 * uint64_t base_clock;
4115 * rte_eth_read_clock(port, base_clock);
4117 * Then, convert the raw mbuf timestamp with:
4118 * base_time_sec + (double)(mbuf->timestamp - base_clock) / freq;
4120 * This simple example will not provide a very good accuracy. One must
4121 * at least measure multiple times the frequency and do a regression.
4122 * To avoid deviation from the system time, the common reference can
4123 * be repeated from time to time. The integer division can also be
4124 * converted by a multiplication and a shift for better performance.
4127 * The port identifier of the Ethernet device.
4129 * Pointer to the uint64_t that holds the raw clock value.
4133 * - -ENODEV: The port ID is invalid.
4134 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4138 rte_eth_read_clock(uint16_t port_id, uint64_t *clock);
4141 * Config l2 tunnel ether type of an Ethernet device for filtering specific
4142 * tunnel packets by ether type.
4145 * The port identifier of the Ethernet device.
4147 * l2 tunnel configuration.
4150 * - (0) if successful.
4151 * - (-ENODEV) if port identifier is invalid.
4152 * - (-EIO) if device is removed.
4153 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4156 rte_eth_dev_l2_tunnel_eth_type_conf(uint16_t port_id,
4157 struct rte_eth_l2_tunnel_conf *l2_tunnel);
4160 * Enable/disable l2 tunnel offload functions. Include,
4161 * 1, The ability of parsing a type of l2 tunnel of an Ethernet device.
4162 * Filtering, forwarding and offloading this type of tunnel packets depend on
4164 * 2, Stripping the l2 tunnel tag.
4165 * 3, Insertion of the l2 tunnel tag.
4166 * 4, Forwarding the packets based on the l2 tunnel tag.
4169 * The port identifier of the Ethernet device.
4171 * l2 tunnel parameters.
4173 * Indicate the offload function.
4175 * Enable or disable this function.
4178 * - (0) if successful.
4179 * - (-ENODEV) if port identifier is invalid.
4180 * - (-EIO) if device is removed.
4181 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4184 rte_eth_dev_l2_tunnel_offload_set(uint16_t port_id,
4185 struct rte_eth_l2_tunnel_conf *l2_tunnel,
4190 * Get the port id from device name. The device name should be specified
4192 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0
4193 * - SoC device name, for example- fsl-gmac0
4194 * - vdev dpdk name, for example- net_[pcap0|null0|tap0]
4197 * pci address or name of the device
4199 * pointer to port identifier of the device
4201 * - (0) if successful and port_id is filled.
4202 * - (-ENODEV or -EINVAL) on failure.
4205 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
4208 * Get the device name from port id. The device name is specified as below:
4209 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0
4210 * - SoC device name, for example- fsl-gmac0
4211 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0]
4214 * Port identifier of the device.
4216 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
4218 * - (0) if successful.
4219 * - (-EINVAL) on failure.
4222 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
4225 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
4226 * the ethernet device information, otherwise adjust them to boundaries.
4229 * The port identifier of the Ethernet device.
4231 * A pointer to a uint16_t where the number of receive
4232 * descriptors stored.
4234 * A pointer to a uint16_t where the number of transmit
4235 * descriptors stored.
4237 * - (0) if successful.
4238 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
4240 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
4241 uint16_t *nb_rx_desc,
4242 uint16_t *nb_tx_desc);
4245 * Test if a port supports specific mempool ops.
4248 * Port identifier of the Ethernet device.
4250 * The name of the pool operations to test.
4252 * - 0: best mempool ops choice for this port.
4253 * - 1: mempool ops are supported for this port.
4254 * - -ENOTSUP: mempool ops not supported for this port.
4255 * - -ENODEV: Invalid port Identifier.
4256 * - -EINVAL: Pool param is null.
4259 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
4262 * Get the security context for the Ethernet device.
4265 * Port identifier of the Ethernet device
4268 * - pointer to security context on success.
4271 rte_eth_dev_get_sec_ctx(uint16_t port_id);
4275 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4277 * Query the device hairpin capabilities.
4280 * The port identifier of the Ethernet device.
4282 * Pointer to a structure that will hold the hairpin capabilities.
4284 * - (0) if successful.
4285 * - (-ENOTSUP) if hardware doesn't support.
4288 int rte_eth_dev_hairpin_capability_get(uint16_t port_id,
4289 struct rte_eth_hairpin_cap *cap);
4291 #include <rte_ethdev_core.h>
4295 * Retrieve a burst of input packets from a receive queue of an Ethernet
4296 * device. The retrieved packets are stored in *rte_mbuf* structures whose
4297 * pointers are supplied in the *rx_pkts* array.
4299 * The rte_eth_rx_burst() function loops, parsing the RX ring of the
4300 * receive queue, up to *nb_pkts* packets, and for each completed RX
4301 * descriptor in the ring, it performs the following operations:
4303 * - Initialize the *rte_mbuf* data structure associated with the
4304 * RX descriptor according to the information provided by the NIC into
4305 * that RX descriptor.
4307 * - Store the *rte_mbuf* data structure into the next entry of the
4310 * - Replenish the RX descriptor with a new *rte_mbuf* buffer
4311 * allocated from the memory pool associated with the receive queue at
4312 * initialization time.
4314 * When retrieving an input packet that was scattered by the controller
4315 * into multiple receive descriptors, the rte_eth_rx_burst() function
4316 * appends the associated *rte_mbuf* buffers to the first buffer of the
4319 * The rte_eth_rx_burst() function returns the number of packets
4320 * actually retrieved, which is the number of *rte_mbuf* data structures
4321 * effectively supplied into the *rx_pkts* array.
4322 * A return value equal to *nb_pkts* indicates that the RX queue contained
4323 * at least *rx_pkts* packets, and this is likely to signify that other
4324 * received packets remain in the input queue. Applications implementing
4325 * a "retrieve as much received packets as possible" policy can check this
4326 * specific case and keep invoking the rte_eth_rx_burst() function until
4327 * a value less than *nb_pkts* is returned.
4329 * This receive method has the following advantages:
4331 * - It allows a run-to-completion network stack engine to retrieve and
4332 * to immediately process received packets in a fast burst-oriented
4333 * approach, avoiding the overhead of unnecessary intermediate packet
4334 * queue/dequeue operations.
4336 * - Conversely, it also allows an asynchronous-oriented processing
4337 * method to retrieve bursts of received packets and to immediately
4338 * queue them for further parallel processing by another logical core,
4339 * for instance. However, instead of having received packets being
4340 * individually queued by the driver, this approach allows the caller
4341 * of the rte_eth_rx_burst() function to queue a burst of retrieved
4342 * packets at a time and therefore dramatically reduce the cost of
4343 * enqueue/dequeue operations per packet.
4345 * - It allows the rte_eth_rx_burst() function of the driver to take
4346 * advantage of burst-oriented hardware features (CPU cache,
4347 * prefetch instructions, and so on) to minimize the number of CPU
4348 * cycles per packet.
4350 * To summarize, the proposed receive API enables many
4351 * burst-oriented optimizations in both synchronous and asynchronous
4352 * packet processing environments with no overhead in both cases.
4354 * The rte_eth_rx_burst() function does not provide any error
4355 * notification to avoid the corresponding overhead. As a hint, the
4356 * upper-level application might check the status of the device link once
4357 * being systematically returned a 0 value for a given number of tries.
4360 * The port identifier of the Ethernet device.
4362 * The index of the receive queue from which to retrieve input packets.
4363 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4364 * to rte_eth_dev_configure().
4366 * The address of an array of pointers to *rte_mbuf* structures that
4367 * must be large enough to store *nb_pkts* pointers in it.
4369 * The maximum number of packets to retrieve.
4371 * The number of packets actually retrieved, which is the number
4372 * of pointers to *rte_mbuf* structures effectively supplied to the
4375 static inline uint16_t
4376 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
4377 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
4379 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4382 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4383 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4384 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_pkt_burst, 0);
4386 if (queue_id >= dev->data->nb_rx_queues) {
4387 RTE_ETHDEV_LOG(ERR, "Invalid RX queue_id=%u\n", queue_id);
4391 nb_rx = (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id],
4394 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4395 if (unlikely(dev->post_rx_burst_cbs[queue_id] != NULL)) {
4396 struct rte_eth_rxtx_callback *cb =
4397 dev->post_rx_burst_cbs[queue_id];
4400 nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx,
4401 nb_pkts, cb->param);
4403 } while (cb != NULL);
4407 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx);
4412 * Get the number of used descriptors of a rx queue
4415 * The port identifier of the Ethernet device.
4417 * The queue id on the specific port.
4419 * The number of used descriptors in the specific queue, or:
4420 * (-EINVAL) if *port_id* or *queue_id* is invalid
4421 * (-ENOTSUP) if the device does not support this function
4424 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
4426 struct rte_eth_dev *dev;
4428 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
4429 dev = &rte_eth_devices[port_id];
4430 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_queue_count, -ENOTSUP);
4431 if (queue_id >= dev->data->nb_rx_queues)
4434 return (int)(*dev->dev_ops->rx_queue_count)(dev, queue_id);
4438 * Check if the DD bit of the specific RX descriptor in the queue has been set
4441 * The port identifier of the Ethernet device.
4443 * The queue id on the specific port.
4445 * The offset of the descriptor ID from tail.
4447 * - (1) if the specific DD bit is set.
4448 * - (0) if the specific DD bit is not set.
4449 * - (-ENODEV) if *port_id* invalid.
4450 * - (-ENOTSUP) if the device does not support this function
4453 rte_eth_rx_descriptor_done(uint16_t port_id, uint16_t queue_id, uint16_t offset)
4455 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4456 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4457 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_done, -ENOTSUP);
4458 return (*dev->dev_ops->rx_descriptor_done)( \
4459 dev->data->rx_queues[queue_id], offset);
4462 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
4463 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
4464 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
4467 * Check the status of a Rx descriptor in the queue
4469 * It should be called in a similar context than the Rx function:
4470 * - on a dataplane core
4471 * - not concurrently on the same queue
4473 * Since it's a dataplane function, no check is performed on port_id and
4474 * queue_id. The caller must therefore ensure that the port is enabled
4475 * and the queue is configured and running.
4477 * Note: accessing to a random descriptor in the ring may trigger cache
4478 * misses and have a performance impact.
4481 * A valid port identifier of the Ethernet device which.
4483 * A valid Rx queue identifier on this port.
4485 * The offset of the descriptor starting from tail (0 is the next
4486 * packet to be received by the driver).
4489 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
4491 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
4492 * not yet processed by the driver (i.e. in the receive queue).
4493 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
4494 * the driver and not yet returned to hw, or reserved by the hw.
4495 * - (-EINVAL) bad descriptor offset.
4496 * - (-ENOTSUP) if the device does not support this function.
4497 * - (-ENODEV) bad port or queue (only if compiled with debug).
4500 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
4503 struct rte_eth_dev *dev;
4506 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4507 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4509 dev = &rte_eth_devices[port_id];
4510 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4511 if (queue_id >= dev->data->nb_rx_queues)
4514 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_status, -ENOTSUP);
4515 rxq = dev->data->rx_queues[queue_id];
4517 return (*dev->dev_ops->rx_descriptor_status)(rxq, offset);
4520 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
4521 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
4522 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
4525 * Check the status of a Tx descriptor in the queue.
4527 * It should be called in a similar context than the Tx function:
4528 * - on a dataplane core
4529 * - not concurrently on the same queue
4531 * Since it's a dataplane function, no check is performed on port_id and
4532 * queue_id. The caller must therefore ensure that the port is enabled
4533 * and the queue is configured and running.
4535 * Note: accessing to a random descriptor in the ring may trigger cache
4536 * misses and have a performance impact.
4539 * A valid port identifier of the Ethernet device which.
4541 * A valid Tx queue identifier on this port.
4543 * The offset of the descriptor starting from tail (0 is the place where
4544 * the next packet will be send).
4547 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
4548 * in the transmit queue.
4549 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
4550 * be reused by the driver.
4551 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
4552 * driver or the hardware.
4553 * - (-EINVAL) bad descriptor offset.
4554 * - (-ENOTSUP) if the device does not support this function.
4555 * - (-ENODEV) bad port or queue (only if compiled with debug).
4557 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
4558 uint16_t queue_id, uint16_t offset)
4560 struct rte_eth_dev *dev;
4563 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4564 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4566 dev = &rte_eth_devices[port_id];
4567 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4568 if (queue_id >= dev->data->nb_tx_queues)
4571 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->tx_descriptor_status, -ENOTSUP);
4572 txq = dev->data->tx_queues[queue_id];
4574 return (*dev->dev_ops->tx_descriptor_status)(txq, offset);
4578 * Send a burst of output packets on a transmit queue of an Ethernet device.
4580 * The rte_eth_tx_burst() function is invoked to transmit output packets
4581 * on the output queue *queue_id* of the Ethernet device designated by its
4583 * The *nb_pkts* parameter is the number of packets to send which are
4584 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4585 * allocated from a pool created with rte_pktmbuf_pool_create().
4586 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
4587 * up to the number of transmit descriptors available in the TX ring of the
4589 * For each packet to send, the rte_eth_tx_burst() function performs
4590 * the following operations:
4592 * - Pick up the next available descriptor in the transmit ring.
4594 * - Free the network buffer previously sent with that descriptor, if any.
4596 * - Initialize the transmit descriptor with the information provided
4597 * in the *rte_mbuf data structure.
4599 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
4600 * the rte_eth_tx_burst() function uses several transmit descriptors
4603 * The rte_eth_tx_burst() function returns the number of packets it
4604 * actually sent. A return value equal to *nb_pkts* means that all packets
4605 * have been sent, and this is likely to signify that other output packets
4606 * could be immediately transmitted again. Applications that implement a
4607 * "send as many packets to transmit as possible" policy can check this
4608 * specific case and keep invoking the rte_eth_tx_burst() function until
4609 * a value less than *nb_pkts* is returned.
4611 * It is the responsibility of the rte_eth_tx_burst() function to
4612 * transparently free the memory buffers of packets previously sent.
4613 * This feature is driven by the *tx_free_thresh* value supplied to the
4614 * rte_eth_dev_configure() function at device configuration time.
4615 * When the number of free TX descriptors drops below this threshold, the
4616 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
4617 * of those packets whose transmission was effectively completed.
4619 * If the PMD is DEV_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
4620 * invoke this function concurrently on the same tx queue without SW lock.
4621 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads
4623 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
4627 * The port identifier of the Ethernet device.
4629 * The index of the transmit queue through which output packets must be
4631 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4632 * to rte_eth_dev_configure().
4634 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4635 * which contain the output packets.
4637 * The maximum number of packets to transmit.
4639 * The number of output packets actually stored in transmit descriptors of
4640 * the transmit ring. The return value can be less than the value of the
4641 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
4643 static inline uint16_t
4644 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
4645 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4647 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4649 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4650 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4651 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_pkt_burst, 0);
4653 if (queue_id >= dev->data->nb_tx_queues) {
4654 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4659 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4660 struct rte_eth_rxtx_callback *cb = dev->pre_tx_burst_cbs[queue_id];
4662 if (unlikely(cb != NULL)) {
4664 nb_pkts = cb->fn.tx(port_id, queue_id, tx_pkts, nb_pkts,
4667 } while (cb != NULL);
4671 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts,
4673 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts);
4677 * Process a burst of output packets on a transmit queue of an Ethernet device.
4679 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
4680 * transmitted on the output queue *queue_id* of the Ethernet device designated
4682 * The *nb_pkts* parameter is the number of packets to be prepared which are
4683 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4684 * allocated from a pool created with rte_pktmbuf_pool_create().
4685 * For each packet to send, the rte_eth_tx_prepare() function performs
4686 * the following operations:
4688 * - Check if packet meets devices requirements for tx offloads.
4690 * - Check limitations about number of segments.
4692 * - Check additional requirements when debug is enabled.
4694 * - Update and/or reset required checksums when tx offload is set for packet.
4696 * Since this function can modify packet data, provided mbufs must be safely
4697 * writable (e.g. modified data cannot be in shared segment).
4699 * The rte_eth_tx_prepare() function returns the number of packets ready to be
4700 * sent. A return value equal to *nb_pkts* means that all packets are valid and
4701 * ready to be sent, otherwise stops processing on the first invalid packet and
4702 * leaves the rest packets untouched.
4704 * When this functionality is not implemented in the driver, all packets are
4705 * are returned untouched.
4708 * The port identifier of the Ethernet device.
4709 * The value must be a valid port id.
4711 * The index of the transmit queue through which output packets must be
4713 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4714 * to rte_eth_dev_configure().
4716 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4717 * which contain the output packets.
4719 * The maximum number of packets to process.
4721 * The number of packets correct and ready to be sent. The return value can be
4722 * less than the value of the *tx_pkts* parameter when some packet doesn't
4723 * meet devices requirements with rte_errno set appropriately:
4724 * - EINVAL: offload flags are not correctly set
4725 * - ENOTSUP: the offload feature is not supported by the hardware
4729 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
4731 static inline uint16_t
4732 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
4733 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4735 struct rte_eth_dev *dev;
4737 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4738 if (!rte_eth_dev_is_valid_port(port_id)) {
4739 RTE_ETHDEV_LOG(ERR, "Invalid TX port_id=%u\n", port_id);
4745 dev = &rte_eth_devices[port_id];
4747 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4748 if (queue_id >= dev->data->nb_tx_queues) {
4749 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4755 if (!dev->tx_pkt_prepare)
4758 return (*dev->tx_pkt_prepare)(dev->data->tx_queues[queue_id],
4765 * Native NOOP operation for compilation targets which doesn't require any
4766 * preparations steps, and functional NOOP may introduce unnecessary performance
4769 * Generally this is not a good idea to turn it on globally and didn't should
4770 * be used if behavior of tx_preparation can change.
4773 static inline uint16_t
4774 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
4775 __rte_unused uint16_t queue_id,
4776 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4784 * Send any packets queued up for transmission on a port and HW queue
4786 * This causes an explicit flush of packets previously buffered via the
4787 * rte_eth_tx_buffer() function. It returns the number of packets successfully
4788 * sent to the NIC, and calls the error callback for any unsent packets. Unless
4789 * explicitly set up otherwise, the default callback simply frees the unsent
4790 * packets back to the owning mempool.
4793 * The port identifier of the Ethernet device.
4795 * The index of the transmit queue through which output packets must be
4797 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4798 * to rte_eth_dev_configure().
4800 * Buffer of packets to be transmit.
4802 * The number of packets successfully sent to the Ethernet device. The error
4803 * callback is called for any packets which could not be sent.
4805 static inline uint16_t
4806 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
4807 struct rte_eth_dev_tx_buffer *buffer)
4810 uint16_t to_send = buffer->length;
4815 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
4819 /* All packets sent, or to be dealt with by callback below */
4820 if (unlikely(sent != to_send))
4821 buffer->error_callback(&buffer->pkts[sent],
4822 (uint16_t)(to_send - sent),
4823 buffer->error_userdata);
4829 * Buffer a single packet for future transmission on a port and queue
4831 * This function takes a single mbuf/packet and buffers it for later
4832 * transmission on the particular port and queue specified. Once the buffer is
4833 * full of packets, an attempt will be made to transmit all the buffered
4834 * packets. In case of error, where not all packets can be transmitted, a
4835 * callback is called with the unsent packets as a parameter. If no callback
4836 * is explicitly set up, the unsent packets are just freed back to the owning
4837 * mempool. The function returns the number of packets actually sent i.e.
4838 * 0 if no buffer flush occurred, otherwise the number of packets successfully
4842 * The port identifier of the Ethernet device.
4844 * The index of the transmit queue through which output packets must be
4846 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4847 * to rte_eth_dev_configure().
4849 * Buffer used to collect packets to be sent.
4851 * Pointer to the packet mbuf to be sent.
4853 * 0 = packet has been buffered for later transmission
4854 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
4855 * causing N packets to be sent, and the error callback to be called for
4858 static __rte_always_inline uint16_t
4859 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
4860 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
4862 buffer->pkts[buffer->length++] = tx_pkt;
4863 if (buffer->length < buffer->size)
4866 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
4873 #endif /* _RTE_ETHDEV_H_ */