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 */
287 #define ETH_LINK_SPEED_200G (1 << 15) /**< 200 Gbps */
290 * Ethernet numeric link speeds in Mbps
292 #define ETH_SPEED_NUM_NONE 0 /**< Not defined */
293 #define ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
294 #define ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
295 #define ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
296 #define ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
297 #define ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
298 #define ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
299 #define ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
300 #define ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
301 #define ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
302 #define ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
303 #define ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
304 #define ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
305 #define ETH_SPEED_NUM_200G 200000 /**< 200 Gbps */
306 #define ETH_SPEED_NUM_UNKNOWN UINT32_MAX /**< Unknown */
309 * A structure used to retrieve link-level information of an Ethernet port.
312 struct rte_eth_link {
313 uint32_t link_speed; /**< ETH_SPEED_NUM_ */
314 uint16_t link_duplex : 1; /**< ETH_LINK_[HALF/FULL]_DUPLEX */
315 uint16_t link_autoneg : 1; /**< ETH_LINK_[AUTONEG/FIXED] */
316 uint16_t link_status : 1; /**< ETH_LINK_[DOWN/UP] */
317 } __rte_aligned(8); /**< aligned for atomic64 read/write */
319 /* Utility constants */
320 #define ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
321 #define ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
322 #define ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
323 #define ETH_LINK_UP 1 /**< Link is up (see link_status). */
324 #define ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
325 #define ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
328 * A structure used to configure the ring threshold registers of an RX/TX
329 * queue for an Ethernet port.
331 struct rte_eth_thresh {
332 uint8_t pthresh; /**< Ring prefetch threshold. */
333 uint8_t hthresh; /**< Ring host threshold. */
334 uint8_t wthresh; /**< Ring writeback threshold. */
338 * Simple flags are used for rte_eth_conf.rxmode.mq_mode.
340 #define ETH_MQ_RX_RSS_FLAG 0x1
341 #define ETH_MQ_RX_DCB_FLAG 0x2
342 #define ETH_MQ_RX_VMDQ_FLAG 0x4
345 * A set of values to identify what method is to be used to route
346 * packets to multiple queues.
348 enum rte_eth_rx_mq_mode {
349 /** None of DCB,RSS or VMDQ mode */
352 /** For RX side, only RSS is on */
353 ETH_MQ_RX_RSS = ETH_MQ_RX_RSS_FLAG,
354 /** For RX side,only DCB is on. */
355 ETH_MQ_RX_DCB = ETH_MQ_RX_DCB_FLAG,
356 /** Both DCB and RSS enable */
357 ETH_MQ_RX_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG,
359 /** Only VMDQ, no RSS nor DCB */
360 ETH_MQ_RX_VMDQ_ONLY = ETH_MQ_RX_VMDQ_FLAG,
361 /** RSS mode with VMDQ */
362 ETH_MQ_RX_VMDQ_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_VMDQ_FLAG,
363 /** Use VMDQ+DCB to route traffic to queues */
364 ETH_MQ_RX_VMDQ_DCB = ETH_MQ_RX_VMDQ_FLAG | ETH_MQ_RX_DCB_FLAG,
365 /** Enable both VMDQ and DCB in VMDq */
366 ETH_MQ_RX_VMDQ_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG |
371 * for rx mq mode backward compatible
373 #define ETH_RSS ETH_MQ_RX_RSS
374 #define VMDQ_DCB ETH_MQ_RX_VMDQ_DCB
375 #define ETH_DCB_RX ETH_MQ_RX_DCB
378 * A set of values to identify what method is to be used to transmit
379 * packets using multi-TCs.
381 enum rte_eth_tx_mq_mode {
382 ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
383 ETH_MQ_TX_DCB, /**< For TX side,only DCB is on. */
384 ETH_MQ_TX_VMDQ_DCB, /**< For TX side,both DCB and VT is on. */
385 ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
389 * for tx mq mode backward compatible
391 #define ETH_DCB_NONE ETH_MQ_TX_NONE
392 #define ETH_VMDQ_DCB_TX ETH_MQ_TX_VMDQ_DCB
393 #define ETH_DCB_TX ETH_MQ_TX_DCB
396 * A structure used to configure the RX features of an Ethernet port.
398 struct rte_eth_rxmode {
399 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
400 enum rte_eth_rx_mq_mode mq_mode;
401 uint32_t max_rx_pkt_len; /**< Only used if JUMBO_FRAME enabled. */
402 /** Maximum allowed size of LRO aggregated packet. */
403 uint32_t max_lro_pkt_size;
404 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
406 * Per-port Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
407 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
408 * structure are allowed to be set.
412 uint64_t reserved_64s[2]; /**< Reserved for future fields */
413 void *reserved_ptrs[2]; /**< Reserved for future fields */
417 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
418 * Note that single VLAN is treated the same as inner VLAN.
421 ETH_VLAN_TYPE_UNKNOWN = 0,
422 ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
423 ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
428 * A structure used to describe a vlan filter.
429 * If the bit corresponding to a VID is set, such VID is on.
431 struct rte_vlan_filter_conf {
436 * A structure used to configure the Receive Side Scaling (RSS) feature
437 * of an Ethernet port.
438 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
439 * to an array holding the RSS key to use for hashing specific header
440 * fields of received packets. The length of this array should be indicated
441 * by *rss_key_len* below. Otherwise, a default random hash key is used by
444 * The *rss_key_len* field of the *rss_conf* structure indicates the length
445 * in bytes of the array pointed by *rss_key*. To be compatible, this length
446 * will be checked in i40e only. Others assume 40 bytes to be used as before.
448 * The *rss_hf* field of the *rss_conf* structure indicates the different
449 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
450 * Supplying an *rss_hf* equal to zero disables the RSS feature.
452 struct rte_eth_rss_conf {
453 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
454 uint8_t rss_key_len; /**< hash key length in bytes. */
455 uint64_t rss_hf; /**< Hash functions to apply - see below. */
459 * A packet can be identified by hardware as different flow types. Different
460 * NIC hardware may support different flow types.
461 * Basically, the NIC hardware identifies the flow type as deep protocol as
462 * possible, and exclusively. For example, if a packet is identified as
463 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types,
464 * though it is an actual IPV4 packet.
466 #define RTE_ETH_FLOW_UNKNOWN 0
467 #define RTE_ETH_FLOW_RAW 1
468 #define RTE_ETH_FLOW_IPV4 2
469 #define RTE_ETH_FLOW_FRAG_IPV4 3
470 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4
471 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5
472 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6
473 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7
474 #define RTE_ETH_FLOW_IPV6 8
475 #define RTE_ETH_FLOW_FRAG_IPV6 9
476 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10
477 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11
478 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12
479 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13
480 #define RTE_ETH_FLOW_L2_PAYLOAD 14
481 #define RTE_ETH_FLOW_IPV6_EX 15
482 #define RTE_ETH_FLOW_IPV6_TCP_EX 16
483 #define RTE_ETH_FLOW_IPV6_UDP_EX 17
484 #define RTE_ETH_FLOW_PORT 18
485 /**< Consider device port number as a flow differentiator */
486 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */
487 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */
488 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */
489 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */
490 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */
491 #define RTE_ETH_FLOW_MAX 24
494 * Below macros are defined for RSS offload types, they can be used to
495 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types.
497 #define ETH_RSS_IPV4 (1ULL << 2)
498 #define ETH_RSS_FRAG_IPV4 (1ULL << 3)
499 #define ETH_RSS_NONFRAG_IPV4_TCP (1ULL << 4)
500 #define ETH_RSS_NONFRAG_IPV4_UDP (1ULL << 5)
501 #define ETH_RSS_NONFRAG_IPV4_SCTP (1ULL << 6)
502 #define ETH_RSS_NONFRAG_IPV4_OTHER (1ULL << 7)
503 #define ETH_RSS_IPV6 (1ULL << 8)
504 #define ETH_RSS_FRAG_IPV6 (1ULL << 9)
505 #define ETH_RSS_NONFRAG_IPV6_TCP (1ULL << 10)
506 #define ETH_RSS_NONFRAG_IPV6_UDP (1ULL << 11)
507 #define ETH_RSS_NONFRAG_IPV6_SCTP (1ULL << 12)
508 #define ETH_RSS_NONFRAG_IPV6_OTHER (1ULL << 13)
509 #define ETH_RSS_L2_PAYLOAD (1ULL << 14)
510 #define ETH_RSS_IPV6_EX (1ULL << 15)
511 #define ETH_RSS_IPV6_TCP_EX (1ULL << 16)
512 #define ETH_RSS_IPV6_UDP_EX (1ULL << 17)
513 #define ETH_RSS_PORT (1ULL << 18)
514 #define ETH_RSS_VXLAN (1ULL << 19)
515 #define ETH_RSS_GENEVE (1ULL << 20)
516 #define ETH_RSS_NVGRE (1ULL << 21)
517 #define ETH_RSS_GTPU (1ULL << 23)
518 #define ETH_RSS_ETH (1ULL << 24)
519 #define ETH_RSS_S_VLAN (1ULL << 25)
520 #define ETH_RSS_C_VLAN (1ULL << 26)
521 #define ETH_RSS_ESP (1ULL << 27)
522 #define ETH_RSS_AH (1ULL << 28)
523 #define ETH_RSS_L2TPV3 (1ULL << 29)
524 #define ETH_RSS_PFCP (1ULL << 30)
525 #define ETH_RSS_PPPOE (1ULL << 31)
528 * We use the following macros to combine with above ETH_RSS_* for
529 * more specific input set selection. These bits are defined starting
530 * from the high end of the 64 bits.
531 * Note: If we use above ETH_RSS_* without SRC/DST_ONLY, it represents
532 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of
533 * the same level are used simultaneously, it is the same case as none of
536 #define ETH_RSS_L3_SRC_ONLY (1ULL << 63)
537 #define ETH_RSS_L3_DST_ONLY (1ULL << 62)
538 #define ETH_RSS_L4_SRC_ONLY (1ULL << 61)
539 #define ETH_RSS_L4_DST_ONLY (1ULL << 60)
540 #define ETH_RSS_L2_SRC_ONLY (1ULL << 59)
541 #define ETH_RSS_L2_DST_ONLY (1ULL << 58)
544 * Only select IPV6 address prefix as RSS input set according to
545 * https://tools.ietf.org/html/rfc6052
546 * Must be combined with ETH_RSS_IPV6, ETH_RSS_NONFRAG_IPV6_UDP,
547 * ETH_RSS_NONFRAG_IPV6_TCP, ETH_RSS_NONFRAG_IPV6_SCTP.
549 #define RTE_ETH_RSS_L3_PRE32 (1ULL << 57)
550 #define RTE_ETH_RSS_L3_PRE40 (1ULL << 56)
551 #define RTE_ETH_RSS_L3_PRE48 (1ULL << 55)
552 #define RTE_ETH_RSS_L3_PRE56 (1ULL << 54)
553 #define RTE_ETH_RSS_L3_PRE64 (1ULL << 53)
554 #define RTE_ETH_RSS_L3_PRE96 (1ULL << 52)
557 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of
558 * the same level are used simultaneously, it is the same case as
559 * none of them are added.
562 * RSS types with SRC/DST_ONLY.
566 static inline uint64_t
567 rte_eth_rss_hf_refine(uint64_t rss_hf)
569 if ((rss_hf & ETH_RSS_L3_SRC_ONLY) && (rss_hf & ETH_RSS_L3_DST_ONLY))
570 rss_hf &= ~(ETH_RSS_L3_SRC_ONLY | ETH_RSS_L3_DST_ONLY);
572 if ((rss_hf & ETH_RSS_L4_SRC_ONLY) && (rss_hf & ETH_RSS_L4_DST_ONLY))
573 rss_hf &= ~(ETH_RSS_L4_SRC_ONLY | ETH_RSS_L4_DST_ONLY);
578 #define ETH_RSS_IPV6_PRE32 ( \
580 RTE_ETH_RSS_L3_PRE32)
582 #define ETH_RSS_IPV6_PRE40 ( \
584 RTE_ETH_RSS_L3_PRE40)
586 #define ETH_RSS_IPV6_PRE48 ( \
588 RTE_ETH_RSS_L3_PRE48)
590 #define ETH_RSS_IPV6_PRE56 ( \
592 RTE_ETH_RSS_L3_PRE56)
594 #define ETH_RSS_IPV6_PRE64 ( \
596 RTE_ETH_RSS_L3_PRE64)
598 #define ETH_RSS_IPV6_PRE96 ( \
600 RTE_ETH_RSS_L3_PRE96)
602 #define ETH_RSS_IPV6_PRE32_UDP ( \
603 ETH_RSS_NONFRAG_IPV6_UDP | \
604 RTE_ETH_RSS_L3_PRE32)
606 #define ETH_RSS_IPV6_PRE40_UDP ( \
607 ETH_RSS_NONFRAG_IPV6_UDP | \
608 RTE_ETH_RSS_L3_PRE40)
610 #define ETH_RSS_IPV6_PRE48_UDP ( \
611 ETH_RSS_NONFRAG_IPV6_UDP | \
612 RTE_ETH_RSS_L3_PRE48)
614 #define ETH_RSS_IPV6_PRE56_UDP ( \
615 ETH_RSS_NONFRAG_IPV6_UDP | \
616 RTE_ETH_RSS_L3_PRE56)
618 #define ETH_RSS_IPV6_PRE64_UDP ( \
619 ETH_RSS_NONFRAG_IPV6_UDP | \
620 RTE_ETH_RSS_L3_PRE64)
622 #define ETH_RSS_IPV6_PRE96_UDP ( \
623 ETH_RSS_NONFRAG_IPV6_UDP | \
624 RTE_ETH_RSS_L3_PRE96)
626 #define ETH_RSS_IPV6_PRE32_TCP ( \
627 ETH_RSS_NONFRAG_IPV6_TCP | \
628 RTE_ETH_RSS_L3_PRE32)
630 #define ETH_RSS_IPV6_PRE40_TCP ( \
631 ETH_RSS_NONFRAG_IPV6_TCP | \
632 RTE_ETH_RSS_L3_PRE40)
634 #define ETH_RSS_IPV6_PRE48_TCP ( \
635 ETH_RSS_NONFRAG_IPV6_TCP | \
636 RTE_ETH_RSS_L3_PRE48)
638 #define ETH_RSS_IPV6_PRE56_TCP ( \
639 ETH_RSS_NONFRAG_IPV6_TCP | \
640 RTE_ETH_RSS_L3_PRE56)
642 #define ETH_RSS_IPV6_PRE64_TCP ( \
643 ETH_RSS_NONFRAG_IPV6_TCP | \
644 RTE_ETH_RSS_L3_PRE64)
646 #define ETH_RSS_IPV6_PRE96_TCP ( \
647 ETH_RSS_NONFRAG_IPV6_TCP | \
648 RTE_ETH_RSS_L3_PRE96)
650 #define ETH_RSS_IPV6_PRE32_SCTP ( \
651 ETH_RSS_NONFRAG_IPV6_SCTP | \
652 RTE_ETH_RSS_L3_PRE32)
654 #define ETH_RSS_IPV6_PRE40_SCTP ( \
655 ETH_RSS_NONFRAG_IPV6_SCTP | \
656 RTE_ETH_RSS_L3_PRE40)
658 #define ETH_RSS_IPV6_PRE48_SCTP ( \
659 ETH_RSS_NONFRAG_IPV6_SCTP | \
660 RTE_ETH_RSS_L3_PRE48)
662 #define ETH_RSS_IPV6_PRE56_SCTP ( \
663 ETH_RSS_NONFRAG_IPV6_SCTP | \
664 RTE_ETH_RSS_L3_PRE56)
666 #define ETH_RSS_IPV6_PRE64_SCTP ( \
667 ETH_RSS_NONFRAG_IPV6_SCTP | \
668 RTE_ETH_RSS_L3_PRE64)
670 #define ETH_RSS_IPV6_PRE96_SCTP ( \
671 ETH_RSS_NONFRAG_IPV6_SCTP | \
672 RTE_ETH_RSS_L3_PRE96)
674 #define ETH_RSS_IP ( \
676 ETH_RSS_FRAG_IPV4 | \
677 ETH_RSS_NONFRAG_IPV4_OTHER | \
679 ETH_RSS_FRAG_IPV6 | \
680 ETH_RSS_NONFRAG_IPV6_OTHER | \
683 #define ETH_RSS_UDP ( \
684 ETH_RSS_NONFRAG_IPV4_UDP | \
685 ETH_RSS_NONFRAG_IPV6_UDP | \
688 #define ETH_RSS_TCP ( \
689 ETH_RSS_NONFRAG_IPV4_TCP | \
690 ETH_RSS_NONFRAG_IPV6_TCP | \
693 #define ETH_RSS_SCTP ( \
694 ETH_RSS_NONFRAG_IPV4_SCTP | \
695 ETH_RSS_NONFRAG_IPV6_SCTP)
697 #define ETH_RSS_TUNNEL ( \
702 #define ETH_RSS_VLAN ( \
706 /**< Mask of valid RSS hash protocols */
707 #define ETH_RSS_PROTO_MASK ( \
709 ETH_RSS_FRAG_IPV4 | \
710 ETH_RSS_NONFRAG_IPV4_TCP | \
711 ETH_RSS_NONFRAG_IPV4_UDP | \
712 ETH_RSS_NONFRAG_IPV4_SCTP | \
713 ETH_RSS_NONFRAG_IPV4_OTHER | \
715 ETH_RSS_FRAG_IPV6 | \
716 ETH_RSS_NONFRAG_IPV6_TCP | \
717 ETH_RSS_NONFRAG_IPV6_UDP | \
718 ETH_RSS_NONFRAG_IPV6_SCTP | \
719 ETH_RSS_NONFRAG_IPV6_OTHER | \
720 ETH_RSS_L2_PAYLOAD | \
722 ETH_RSS_IPV6_TCP_EX | \
723 ETH_RSS_IPV6_UDP_EX | \
730 * Definitions used for redirection table entry size.
731 * Some RSS RETA sizes may not be supported by some drivers, check the
732 * documentation or the description of relevant functions for more details.
734 #define ETH_RSS_RETA_SIZE_64 64
735 #define ETH_RSS_RETA_SIZE_128 128
736 #define ETH_RSS_RETA_SIZE_256 256
737 #define ETH_RSS_RETA_SIZE_512 512
738 #define RTE_RETA_GROUP_SIZE 64
740 /* Definitions used for VMDQ and DCB functionality */
741 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */
742 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
743 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */
744 #define ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
746 /* DCB capability defines */
747 #define ETH_DCB_PG_SUPPORT 0x00000001 /**< Priority Group(ETS) support. */
748 #define ETH_DCB_PFC_SUPPORT 0x00000002 /**< Priority Flow Control support. */
750 /* Definitions used for VLAN Offload functionality */
751 #define ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
752 #define ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
753 #define ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
754 #define ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */
756 /* Definitions used for mask VLAN setting */
757 #define ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
758 #define ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
759 #define ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
760 #define ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */
761 #define ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
763 /* Definitions used for receive MAC address */
764 #define ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
766 /* Definitions used for unicast hash */
767 #define ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
769 /* Definitions used for VMDQ pool rx mode setting */
770 #define ETH_VMDQ_ACCEPT_UNTAG 0x0001 /**< accept untagged packets. */
771 #define ETH_VMDQ_ACCEPT_HASH_MC 0x0002 /**< accept packets in multicast table . */
772 #define ETH_VMDQ_ACCEPT_HASH_UC 0x0004 /**< accept packets in unicast table. */
773 #define ETH_VMDQ_ACCEPT_BROADCAST 0x0008 /**< accept broadcast packets. */
774 #define ETH_VMDQ_ACCEPT_MULTICAST 0x0010 /**< multicast promiscuous. */
776 /** Maximum nb. of vlan per mirror rule */
777 #define ETH_MIRROR_MAX_VLANS 64
779 #define ETH_MIRROR_VIRTUAL_POOL_UP 0x01 /**< Virtual Pool uplink Mirroring. */
780 #define ETH_MIRROR_UPLINK_PORT 0x02 /**< Uplink Port Mirroring. */
781 #define ETH_MIRROR_DOWNLINK_PORT 0x04 /**< Downlink Port Mirroring. */
782 #define ETH_MIRROR_VLAN 0x08 /**< VLAN Mirroring. */
783 #define ETH_MIRROR_VIRTUAL_POOL_DOWN 0x10 /**< Virtual Pool downlink Mirroring. */
786 * A structure used to configure VLAN traffic mirror of an Ethernet port.
788 struct rte_eth_vlan_mirror {
789 uint64_t vlan_mask; /**< mask for valid VLAN ID. */
790 /** VLAN ID list for vlan mirroring. */
791 uint16_t vlan_id[ETH_MIRROR_MAX_VLANS];
795 * A structure used to configure traffic mirror of an Ethernet port.
797 struct rte_eth_mirror_conf {
798 uint8_t rule_type; /**< Mirroring rule type */
799 uint8_t dst_pool; /**< Destination pool for this mirror rule. */
800 uint64_t pool_mask; /**< Bitmap of pool for pool mirroring */
801 /** VLAN ID setting for VLAN mirroring. */
802 struct rte_eth_vlan_mirror vlan;
806 * A structure used to configure 64 entries of Redirection Table of the
807 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
808 * more than 64 entries supported by hardware, an array of this structure
811 struct rte_eth_rss_reta_entry64 {
813 /**< Mask bits indicate which entries need to be updated/queried. */
814 uint16_t reta[RTE_RETA_GROUP_SIZE];
815 /**< Group of 64 redirection table entries. */
819 * This enum indicates the possible number of traffic classes
820 * in DCB configurations
822 enum rte_eth_nb_tcs {
823 ETH_4_TCS = 4, /**< 4 TCs with DCB. */
824 ETH_8_TCS = 8 /**< 8 TCs with DCB. */
828 * This enum indicates the possible number of queue pools
829 * in VMDQ configurations.
831 enum rte_eth_nb_pools {
832 ETH_8_POOLS = 8, /**< 8 VMDq pools. */
833 ETH_16_POOLS = 16, /**< 16 VMDq pools. */
834 ETH_32_POOLS = 32, /**< 32 VMDq pools. */
835 ETH_64_POOLS = 64 /**< 64 VMDq pools. */
838 /* This structure may be extended in future. */
839 struct rte_eth_dcb_rx_conf {
840 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
841 /** Traffic class each UP mapped to. */
842 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
845 struct rte_eth_vmdq_dcb_tx_conf {
846 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
847 /** Traffic class each UP mapped to. */
848 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
851 struct rte_eth_dcb_tx_conf {
852 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
853 /** Traffic class each UP mapped to. */
854 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
857 struct rte_eth_vmdq_tx_conf {
858 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
862 * A structure used to configure the VMDQ+DCB feature
863 * of an Ethernet port.
865 * Using this feature, packets are routed to a pool of queues, based
866 * on the vlan id in the vlan tag, and then to a specific queue within
867 * that pool, using the user priority vlan tag field.
869 * A default pool may be used, if desired, to route all traffic which
870 * does not match the vlan filter rules.
872 struct rte_eth_vmdq_dcb_conf {
873 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
874 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
875 uint8_t default_pool; /**< The default pool, if applicable */
876 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
878 uint16_t vlan_id; /**< The vlan id of the received frame */
879 uint64_t pools; /**< Bitmask of pools for packet rx */
880 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
881 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
882 /**< Selects a queue in a pool */
886 * A structure used to configure the VMDQ feature of an Ethernet port when
887 * not combined with the DCB feature.
889 * Using this feature, packets are routed to a pool of queues. By default,
890 * the pool selection is based on the MAC address, the vlan id in the
891 * vlan tag as specified in the pool_map array.
892 * Passing the ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
893 * selection using only the MAC address. MAC address to pool mapping is done
894 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
895 * corresponding to the pool id.
897 * Queue selection within the selected pool will be done using RSS when
898 * it is enabled or revert to the first queue of the pool if not.
900 * A default pool may be used, if desired, to route all traffic which
901 * does not match the vlan filter rules or any pool MAC address.
903 struct rte_eth_vmdq_rx_conf {
904 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
905 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
906 uint8_t default_pool; /**< The default pool, if applicable */
907 uint8_t enable_loop_back; /**< Enable VT loop back */
908 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
909 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
911 uint16_t vlan_id; /**< The vlan id of the received frame */
912 uint64_t pools; /**< Bitmask of pools for packet rx */
913 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
917 * A structure used to configure the TX features of an Ethernet port.
919 struct rte_eth_txmode {
920 enum rte_eth_tx_mq_mode mq_mode; /**< TX multi-queues mode. */
922 * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
923 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
924 * structure are allowed to be set.
930 uint8_t hw_vlan_reject_tagged : 1,
931 /**< If set, reject sending out tagged pkts */
932 hw_vlan_reject_untagged : 1,
933 /**< If set, reject sending out untagged pkts */
934 hw_vlan_insert_pvid : 1;
935 /**< If set, enable port based VLAN insertion */
937 uint64_t reserved_64s[2]; /**< Reserved for future fields */
938 void *reserved_ptrs[2]; /**< Reserved for future fields */
942 * A structure used to configure an RX ring of an Ethernet port.
944 struct rte_eth_rxconf {
945 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */
946 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */
947 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
948 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
950 * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
951 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
952 * fields on rte_eth_dev_info structure are allowed to be set.
956 uint64_t reserved_64s[2]; /**< Reserved for future fields */
957 void *reserved_ptrs[2]; /**< Reserved for future fields */
961 * A structure used to configure a TX ring of an Ethernet port.
963 struct rte_eth_txconf {
964 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */
965 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
966 uint16_t tx_free_thresh; /**< Start freeing TX buffers if there are
967 less free descriptors than this value. */
969 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
971 * Per-queue Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
972 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
973 * fields on rte_eth_dev_info structure are allowed to be set.
977 uint64_t reserved_64s[2]; /**< Reserved for future fields */
978 void *reserved_ptrs[2]; /**< Reserved for future fields */
983 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
985 * A structure used to return the hairpin capabilities that are supported.
987 struct rte_eth_hairpin_cap {
988 /** The max number of hairpin queues (different bindings). */
989 uint16_t max_nb_queues;
990 /** Max number of Rx queues to be connected to one Tx queue. */
991 uint16_t max_rx_2_tx;
992 /** Max number of Tx queues to be connected to one Rx queue. */
993 uint16_t max_tx_2_rx;
994 uint16_t max_nb_desc; /**< The max num of descriptors. */
997 #define RTE_ETH_MAX_HAIRPIN_PEERS 32
1001 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1003 * A structure used to hold hairpin peer data.
1005 struct rte_eth_hairpin_peer {
1006 uint16_t port; /**< Peer port. */
1007 uint16_t queue; /**< Peer queue. */
1012 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1014 * A structure used to configure hairpin binding.
1016 struct rte_eth_hairpin_conf {
1017 uint16_t peer_count; /**< The number of peers. */
1018 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
1022 * A structure contains information about HW descriptor ring limitations.
1024 struct rte_eth_desc_lim {
1025 uint16_t nb_max; /**< Max allowed number of descriptors. */
1026 uint16_t nb_min; /**< Min allowed number of descriptors. */
1027 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
1030 * Max allowed number of segments per whole packet.
1032 * - For TSO packet this is the total number of data descriptors allowed
1035 * @see nb_mtu_seg_max
1037 uint16_t nb_seg_max;
1040 * Max number of segments per one MTU.
1042 * - For non-TSO packet, this is the maximum allowed number of segments
1043 * in a single transmit packet.
1045 * - For TSO packet each segment within the TSO may span up to this
1050 uint16_t nb_mtu_seg_max;
1054 * This enum indicates the flow control mode
1056 enum rte_eth_fc_mode {
1057 RTE_FC_NONE = 0, /**< Disable flow control. */
1058 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */
1059 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */
1060 RTE_FC_FULL /**< Enable flow control on both side. */
1064 * A structure used to configure Ethernet flow control parameter.
1065 * These parameters will be configured into the register of the NIC.
1066 * Please refer to the corresponding data sheet for proper value.
1068 struct rte_eth_fc_conf {
1069 uint32_t high_water; /**< High threshold value to trigger XOFF */
1070 uint32_t low_water; /**< Low threshold value to trigger XON */
1071 uint16_t pause_time; /**< Pause quota in the Pause frame */
1072 uint16_t send_xon; /**< Is XON frame need be sent */
1073 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1074 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
1075 uint8_t autoneg; /**< Use Pause autoneg */
1079 * A structure used to configure Ethernet priority flow control parameter.
1080 * These parameters will be configured into the register of the NIC.
1081 * Please refer to the corresponding data sheet for proper value.
1083 struct rte_eth_pfc_conf {
1084 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
1085 uint8_t priority; /**< VLAN User Priority. */
1091 enum rte_eth_tunnel_type {
1092 RTE_TUNNEL_TYPE_NONE = 0,
1093 RTE_TUNNEL_TYPE_VXLAN,
1094 RTE_TUNNEL_TYPE_GENEVE,
1095 RTE_TUNNEL_TYPE_TEREDO,
1096 RTE_TUNNEL_TYPE_NVGRE,
1097 RTE_TUNNEL_TYPE_IP_IN_GRE,
1098 RTE_L2_TUNNEL_TYPE_E_TAG,
1099 RTE_TUNNEL_TYPE_VXLAN_GPE,
1100 RTE_TUNNEL_TYPE_MAX,
1103 /* Deprecated API file for rte_eth_dev_filter_* functions */
1104 #include "rte_eth_ctrl.h"
1107 * Memory space that can be configured to store Flow Director filters
1108 * in the board memory.
1110 enum rte_fdir_pballoc_type {
1111 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */
1112 RTE_FDIR_PBALLOC_128K, /**< 128k. */
1113 RTE_FDIR_PBALLOC_256K, /**< 256k. */
1117 * Select report mode of FDIR hash information in RX descriptors.
1119 enum rte_fdir_status_mode {
1120 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
1121 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
1122 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
1126 * A structure used to configure the Flow Director (FDIR) feature
1127 * of an Ethernet port.
1129 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored.
1131 struct rte_fdir_conf {
1132 enum rte_fdir_mode mode; /**< Flow Director mode. */
1133 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
1134 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
1135 /** RX queue of packets matching a "drop" filter in perfect mode. */
1137 struct rte_eth_fdir_masks mask;
1138 struct rte_eth_fdir_flex_conf flex_conf;
1139 /**< Flex payload configuration. */
1143 * UDP tunneling configuration.
1144 * Used to config the UDP port for a type of tunnel.
1145 * NICs need the UDP port to identify the tunnel type.
1146 * Normally a type of tunnel has a default UDP port, this structure can be used
1147 * in case if the users want to change or support more UDP port.
1149 struct rte_eth_udp_tunnel {
1150 uint16_t udp_port; /**< UDP port used for the tunnel. */
1151 uint8_t prot_type; /**< Tunnel type. Defined in rte_eth_tunnel_type. */
1155 * A structure used to enable/disable specific device interrupts.
1157 struct rte_intr_conf {
1158 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
1160 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
1162 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
1167 * A structure used to configure an Ethernet port.
1168 * Depending upon the RX multi-queue mode, extra advanced
1169 * configuration settings may be needed.
1171 struct rte_eth_conf {
1172 uint32_t link_speeds; /**< bitmap of ETH_LINK_SPEED_XXX of speeds to be
1173 used. ETH_LINK_SPEED_FIXED disables link
1174 autonegotiation, and a unique speed shall be
1175 set. Otherwise, the bitmap defines the set of
1176 speeds to be advertised. If the special value
1177 ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
1178 supported are advertised. */
1179 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */
1180 struct rte_eth_txmode txmode; /**< Port TX configuration. */
1181 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
1182 is 0, meaning the loopback mode is disabled.
1183 Read the datasheet of given ethernet controller
1184 for details. The possible values of this field
1185 are defined in implementation of each driver. */
1187 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
1188 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
1189 /**< Port vmdq+dcb configuration. */
1190 struct rte_eth_dcb_rx_conf dcb_rx_conf;
1191 /**< Port dcb RX configuration. */
1192 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
1193 /**< Port vmdq RX configuration. */
1194 } rx_adv_conf; /**< Port RX filtering configuration. */
1196 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
1197 /**< Port vmdq+dcb TX configuration. */
1198 struct rte_eth_dcb_tx_conf dcb_tx_conf;
1199 /**< Port dcb TX configuration. */
1200 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
1201 /**< Port vmdq TX configuration. */
1202 } tx_adv_conf; /**< Port TX DCB configuration (union). */
1203 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
1204 is needed,and the variable must be set ETH_DCB_PFC_SUPPORT. */
1205 uint32_t dcb_capability_en;
1206 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */
1207 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */
1211 * RX offload capabilities of a device.
1213 #define DEV_RX_OFFLOAD_VLAN_STRIP 0x00000001
1214 #define DEV_RX_OFFLOAD_IPV4_CKSUM 0x00000002
1215 #define DEV_RX_OFFLOAD_UDP_CKSUM 0x00000004
1216 #define DEV_RX_OFFLOAD_TCP_CKSUM 0x00000008
1217 #define DEV_RX_OFFLOAD_TCP_LRO 0x00000010
1218 #define DEV_RX_OFFLOAD_QINQ_STRIP 0x00000020
1219 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000040
1220 #define DEV_RX_OFFLOAD_MACSEC_STRIP 0x00000080
1221 #define DEV_RX_OFFLOAD_HEADER_SPLIT 0x00000100
1222 #define DEV_RX_OFFLOAD_VLAN_FILTER 0x00000200
1223 #define DEV_RX_OFFLOAD_VLAN_EXTEND 0x00000400
1224 #define DEV_RX_OFFLOAD_JUMBO_FRAME 0x00000800
1225 #define DEV_RX_OFFLOAD_SCATTER 0x00002000
1226 #define DEV_RX_OFFLOAD_TIMESTAMP 0x00004000
1227 #define DEV_RX_OFFLOAD_SECURITY 0x00008000
1228 #define DEV_RX_OFFLOAD_KEEP_CRC 0x00010000
1229 #define DEV_RX_OFFLOAD_SCTP_CKSUM 0x00020000
1230 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM 0x00040000
1231 #define DEV_RX_OFFLOAD_RSS_HASH 0x00080000
1233 #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \
1234 DEV_RX_OFFLOAD_UDP_CKSUM | \
1235 DEV_RX_OFFLOAD_TCP_CKSUM)
1236 #define DEV_RX_OFFLOAD_VLAN (DEV_RX_OFFLOAD_VLAN_STRIP | \
1237 DEV_RX_OFFLOAD_VLAN_FILTER | \
1238 DEV_RX_OFFLOAD_VLAN_EXTEND | \
1239 DEV_RX_OFFLOAD_QINQ_STRIP)
1242 * If new Rx offload capabilities are defined, they also must be
1243 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
1247 * TX offload capabilities of a device.
1249 #define DEV_TX_OFFLOAD_VLAN_INSERT 0x00000001
1250 #define DEV_TX_OFFLOAD_IPV4_CKSUM 0x00000002
1251 #define DEV_TX_OFFLOAD_UDP_CKSUM 0x00000004
1252 #define DEV_TX_OFFLOAD_TCP_CKSUM 0x00000008
1253 #define DEV_TX_OFFLOAD_SCTP_CKSUM 0x00000010
1254 #define DEV_TX_OFFLOAD_TCP_TSO 0x00000020
1255 #define DEV_TX_OFFLOAD_UDP_TSO 0x00000040
1256 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000080 /**< Used for tunneling packet. */
1257 #define DEV_TX_OFFLOAD_QINQ_INSERT 0x00000100
1258 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO 0x00000200 /**< Used for tunneling packet. */
1259 #define DEV_TX_OFFLOAD_GRE_TNL_TSO 0x00000400 /**< Used for tunneling packet. */
1260 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO 0x00000800 /**< Used for tunneling packet. */
1261 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO 0x00001000 /**< Used for tunneling packet. */
1262 #define DEV_TX_OFFLOAD_MACSEC_INSERT 0x00002000
1263 #define DEV_TX_OFFLOAD_MT_LOCKFREE 0x00004000
1264 /**< Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
1265 * tx queue without SW lock.
1267 #define DEV_TX_OFFLOAD_MULTI_SEGS 0x00008000
1268 /**< Device supports multi segment send. */
1269 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE 0x00010000
1270 /**< Device supports optimization for fast release of mbufs.
1271 * When set application must guarantee that per-queue all mbufs comes from
1272 * the same mempool and has refcnt = 1.
1274 #define DEV_TX_OFFLOAD_SECURITY 0x00020000
1276 * Device supports generic UDP tunneled packet TSO.
1277 * Application must set PKT_TX_TUNNEL_UDP and other mbuf fields required
1280 #define DEV_TX_OFFLOAD_UDP_TNL_TSO 0x00040000
1282 * Device supports generic IP tunneled packet TSO.
1283 * Application must set PKT_TX_TUNNEL_IP and other mbuf fields required
1286 #define DEV_TX_OFFLOAD_IP_TNL_TSO 0x00080000
1287 /** Device supports outer UDP checksum */
1288 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM 0x00100000
1290 /** Device supports send on timestamp */
1291 #define DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP 0x00200000
1294 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP 0x00000001
1295 /**< Device supports Rx queue setup after device started*/
1296 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP 0x00000002
1297 /**< Device supports Tx queue setup after device started*/
1300 * If new Tx offload capabilities are defined, they also must be
1301 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1305 * Fallback default preferred Rx/Tx port parameters.
1306 * These are used if an application requests default parameters
1307 * but the PMD does not provide preferred values.
1309 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1310 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1311 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1312 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1315 * Preferred Rx/Tx port parameters.
1316 * There are separate instances of this structure for transmission
1317 * and reception respectively.
1319 struct rte_eth_dev_portconf {
1320 uint16_t burst_size; /**< Device-preferred burst size */
1321 uint16_t ring_size; /**< Device-preferred size of queue rings */
1322 uint16_t nb_queues; /**< Device-preferred number of queues */
1326 * Default values for switch domain id when ethdev does not support switch
1327 * domain definitions.
1329 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX)
1332 * Ethernet device associated switch information
1334 struct rte_eth_switch_info {
1335 const char *name; /**< switch name */
1336 uint16_t domain_id; /**< switch domain id */
1339 * mapping to the devices physical switch port as enumerated from the
1340 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1341 * device this may correspond to the VF_ID of each virtual function,
1342 * but each driver should explicitly define the mapping of switch
1343 * port identifier to that physical interconnect/switch
1348 * Ethernet device information
1352 * A structure used to retrieve the contextual information of
1353 * an Ethernet device, such as the controlling driver of the
1356 struct rte_eth_dev_info {
1357 struct rte_device *device; /** Generic device information */
1358 const char *driver_name; /**< Device Driver name. */
1359 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1360 Use if_indextoname() to translate into an interface name. */
1361 uint16_t min_mtu; /**< Minimum MTU allowed */
1362 uint16_t max_mtu; /**< Maximum MTU allowed */
1363 const uint32_t *dev_flags; /**< Device flags */
1364 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */
1365 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */
1366 /** Maximum configurable size of LRO aggregated packet. */
1367 uint32_t max_lro_pkt_size;
1368 uint16_t max_rx_queues; /**< Maximum number of RX queues. */
1369 uint16_t max_tx_queues; /**< Maximum number of TX queues. */
1370 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1371 uint32_t max_hash_mac_addrs;
1372 /** Maximum number of hash MAC addresses for MTA and UTA. */
1373 uint16_t max_vfs; /**< Maximum number of VFs. */
1374 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1375 uint64_t rx_offload_capa;
1376 /**< All RX offload capabilities including all per-queue ones */
1377 uint64_t tx_offload_capa;
1378 /**< All TX offload capabilities including all per-queue ones */
1379 uint64_t rx_queue_offload_capa;
1380 /**< Device per-queue RX offload capabilities. */
1381 uint64_t tx_queue_offload_capa;
1382 /**< Device per-queue TX offload capabilities. */
1384 /**< Device redirection table size, the total number of entries. */
1385 uint8_t hash_key_size; /**< Hash key size in bytes */
1386 /** Bit mask of RSS offloads, the bit offset also means flow type */
1387 uint64_t flow_type_rss_offloads;
1388 struct rte_eth_rxconf default_rxconf; /**< Default RX configuration */
1389 struct rte_eth_txconf default_txconf; /**< Default TX configuration */
1390 uint16_t vmdq_queue_base; /**< First queue ID for VMDQ pools. */
1391 uint16_t vmdq_queue_num; /**< Queue number for VMDQ pools. */
1392 uint16_t vmdq_pool_base; /**< First ID of VMDQ pools. */
1393 struct rte_eth_desc_lim rx_desc_lim; /**< RX descriptors limits */
1394 struct rte_eth_desc_lim tx_desc_lim; /**< TX descriptors limits */
1395 uint32_t speed_capa; /**< Supported speeds bitmap (ETH_LINK_SPEED_). */
1396 /** Configured number of rx/tx queues */
1397 uint16_t nb_rx_queues; /**< Number of RX queues. */
1398 uint16_t nb_tx_queues; /**< Number of TX queues. */
1399 /** Rx parameter recommendations */
1400 struct rte_eth_dev_portconf default_rxportconf;
1401 /** Tx parameter recommendations */
1402 struct rte_eth_dev_portconf default_txportconf;
1403 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1406 * Switching information for ports on a device with a
1407 * embedded managed interconnect/switch.
1409 struct rte_eth_switch_info switch_info;
1411 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1412 void *reserved_ptrs[2]; /**< Reserved for future fields */
1416 * Ethernet device RX queue information structure.
1417 * Used to retrieve information about configured queue.
1419 struct rte_eth_rxq_info {
1420 struct rte_mempool *mp; /**< mempool used by that queue. */
1421 struct rte_eth_rxconf conf; /**< queue config parameters. */
1422 uint8_t scattered_rx; /**< scattered packets RX supported. */
1423 uint16_t nb_desc; /**< configured number of RXDs. */
1424 } __rte_cache_min_aligned;
1427 * Ethernet device TX queue information structure.
1428 * Used to retrieve information about configured queue.
1430 struct rte_eth_txq_info {
1431 struct rte_eth_txconf conf; /**< queue config parameters. */
1432 uint16_t nb_desc; /**< configured number of TXDs. */
1433 } __rte_cache_min_aligned;
1435 /* Generic Burst mode flag definition, values can be ORed. */
1438 * If the queues have different burst mode description, this bit will be set
1439 * by PMD, then the application can iterate to retrieve burst description for
1442 #define RTE_ETH_BURST_FLAG_PER_QUEUE (1ULL << 0)
1445 * Ethernet device RX/TX queue packet burst mode information structure.
1446 * Used to retrieve information about packet burst mode setting.
1448 struct rte_eth_burst_mode {
1449 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */
1451 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */
1452 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */
1455 /** Maximum name length for extended statistics counters */
1456 #define RTE_ETH_XSTATS_NAME_SIZE 64
1459 * An Ethernet device extended statistic structure
1461 * This structure is used by rte_eth_xstats_get() to provide
1462 * statistics that are not provided in the generic *rte_eth_stats*
1464 * It maps a name id, corresponding to an index in the array returned
1465 * by rte_eth_xstats_get_names(), to a statistic value.
1467 struct rte_eth_xstat {
1468 uint64_t id; /**< The index in xstats name array. */
1469 uint64_t value; /**< The statistic counter value. */
1473 * A name element for extended statistics.
1475 * An array of this structure is returned by rte_eth_xstats_get_names().
1476 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1477 * structure references these names by their array index.
1479 struct rte_eth_xstat_name {
1480 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1483 #define ETH_DCB_NUM_TCS 8
1484 #define ETH_MAX_VMDQ_POOL 64
1487 * A structure used to get the information of queue and
1488 * TC mapping on both TX and RX paths.
1490 struct rte_eth_dcb_tc_queue_mapping {
1491 /** rx queues assigned to tc per Pool */
1495 } tc_rxq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1496 /** rx queues assigned to tc per Pool */
1500 } tc_txq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1504 * A structure used to get the information of DCB.
1505 * It includes TC UP mapping and queue TC mapping.
1507 struct rte_eth_dcb_info {
1508 uint8_t nb_tcs; /**< number of TCs */
1509 uint8_t prio_tc[ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
1510 uint8_t tc_bws[ETH_DCB_NUM_TCS]; /**< TX BW percentage for each TC */
1511 /** rx queues assigned to tc */
1512 struct rte_eth_dcb_tc_queue_mapping tc_queue;
1515 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
1517 /* Macros to check for valid port */
1518 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
1519 if (!rte_eth_dev_is_valid_port(port_id)) { \
1520 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1525 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
1526 if (!rte_eth_dev_is_valid_port(port_id)) { \
1527 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1533 * l2 tunnel configuration.
1536 /**< l2 tunnel enable mask */
1537 #define ETH_L2_TUNNEL_ENABLE_MASK 0x00000001
1538 /**< l2 tunnel insertion mask */
1539 #define ETH_L2_TUNNEL_INSERTION_MASK 0x00000002
1540 /**< l2 tunnel stripping mask */
1541 #define ETH_L2_TUNNEL_STRIPPING_MASK 0x00000004
1542 /**< l2 tunnel forwarding mask */
1543 #define ETH_L2_TUNNEL_FORWARDING_MASK 0x00000008
1546 * Function type used for RX packet processing packet callbacks.
1548 * The callback function is called on RX with a burst of packets that have
1549 * been received on the given port and queue.
1552 * The Ethernet port on which RX is being performed.
1554 * The queue on the Ethernet port which is being used to receive the packets.
1556 * The burst of packets that have just been received.
1558 * The number of packets in the burst pointed to by "pkts".
1560 * The max number of packets that can be stored in the "pkts" array.
1562 * The arbitrary user parameter passed in by the application when the callback
1563 * was originally configured.
1565 * The number of packets returned to the user.
1567 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
1568 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
1572 * Function type used for TX packet processing packet callbacks.
1574 * The callback function is called on TX with a burst of packets immediately
1575 * before the packets are put onto the hardware queue for transmission.
1578 * The Ethernet port on which TX is being performed.
1580 * The queue on the Ethernet port which is being used to transmit the packets.
1582 * The burst of packets that are about to be transmitted.
1584 * The number of packets in the burst pointed to by "pkts".
1586 * The arbitrary user parameter passed in by the application when the callback
1587 * was originally configured.
1589 * The number of packets to be written to the NIC.
1591 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
1592 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
1595 * Possible states of an ethdev port.
1597 enum rte_eth_dev_state {
1598 /** Device is unused before being probed. */
1599 RTE_ETH_DEV_UNUSED = 0,
1600 /** Device is attached when allocated in probing. */
1601 RTE_ETH_DEV_ATTACHED,
1602 /** Device is in removed state when plug-out is detected. */
1603 RTE_ETH_DEV_REMOVED,
1606 struct rte_eth_dev_sriov {
1607 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
1608 uint8_t nb_q_per_pool; /**< rx queue number per pool */
1609 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
1610 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
1612 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
1614 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
1616 #define RTE_ETH_DEV_NO_OWNER 0
1618 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
1620 struct rte_eth_dev_owner {
1621 uint64_t id; /**< The owner unique identifier. */
1622 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
1626 * Port is released (i.e. totally freed and data erased) on close.
1627 * Temporary flag for PMD migration to new rte_eth_dev_close() behaviour.
1629 #define RTE_ETH_DEV_CLOSE_REMOVE 0x0001
1630 /** Device supports link state interrupt */
1631 #define RTE_ETH_DEV_INTR_LSC 0x0002
1632 /** Device is a bonded slave */
1633 #define RTE_ETH_DEV_BONDED_SLAVE 0x0004
1634 /** Device supports device removal interrupt */
1635 #define RTE_ETH_DEV_INTR_RMV 0x0008
1636 /** Device is port representor */
1637 #define RTE_ETH_DEV_REPRESENTOR 0x0010
1638 /** Device does not support MAC change after started */
1639 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR 0x0020
1642 * Iterates over valid ethdev ports owned by a specific owner.
1645 * The id of the next possible valid owned port.
1647 * The owner identifier.
1648 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
1650 * Next valid port id owned by owner_id, RTE_MAX_ETHPORTS if there is none.
1652 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
1653 const uint64_t owner_id);
1656 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
1658 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
1659 for (p = rte_eth_find_next_owned_by(0, o); \
1660 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
1661 p = rte_eth_find_next_owned_by(p + 1, o))
1664 * Iterates over valid ethdev ports.
1667 * The id of the next possible valid port.
1669 * Next valid port id, RTE_MAX_ETHPORTS if there is none.
1671 uint16_t rte_eth_find_next(uint16_t port_id);
1674 * Macro to iterate over all enabled and ownerless ethdev ports.
1676 #define RTE_ETH_FOREACH_DEV(p) \
1677 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
1681 * @b EXPERIMENTAL: this API may change without prior notice.
1683 * Iterates over ethdev ports of a specified device.
1685 * @param port_id_start
1686 * The id of the next possible valid port.
1688 * The generic device behind the ports to iterate.
1690 * Next port id of the device, possibly port_id_start,
1691 * RTE_MAX_ETHPORTS if there is none.
1695 rte_eth_find_next_of(uint16_t port_id_start,
1696 const struct rte_device *parent);
1699 * Macro to iterate over all ethdev ports of a specified device.
1702 * The id of the matching port being iterated.
1704 * The rte_device pointer matching the iterated ports.
1706 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \
1707 for (port_id = rte_eth_find_next_of(0, parent); \
1708 port_id < RTE_MAX_ETHPORTS; \
1709 port_id = rte_eth_find_next_of(port_id + 1, parent))
1713 * @b EXPERIMENTAL: this API may change without prior notice.
1715 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device).
1717 * @param port_id_start
1718 * The id of the next possible valid sibling port.
1719 * @param ref_port_id
1720 * The id of a reference port to compare rte_device with.
1722 * Next sibling port id, possibly port_id_start or ref_port_id itself,
1723 * RTE_MAX_ETHPORTS if there is none.
1727 rte_eth_find_next_sibling(uint16_t port_id_start,
1728 uint16_t ref_port_id);
1731 * Macro to iterate over all ethdev ports sharing the same rte_device
1732 * as the specified port.
1733 * Note: the specified reference port is part of the loop iterations.
1736 * The id of the matching port being iterated.
1737 * @param ref_port_id
1738 * The id of the port being compared.
1740 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \
1741 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \
1742 port_id < RTE_MAX_ETHPORTS; \
1743 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id))
1747 * @b EXPERIMENTAL: this API may change without prior notice.
1749 * Get a new unique owner identifier.
1750 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
1751 * to avoid multiple management of device by different entities.
1754 * Owner identifier pointer.
1756 * Negative errno value on error, 0 on success.
1759 int rte_eth_dev_owner_new(uint64_t *owner_id);
1763 * @b EXPERIMENTAL: this API may change without prior notice.
1765 * Set an Ethernet device owner.
1768 * The identifier of the port to own.
1770 * The owner pointer.
1772 * Negative errno value on error, 0 on success.
1775 int rte_eth_dev_owner_set(const uint16_t port_id,
1776 const struct rte_eth_dev_owner *owner);
1780 * @b EXPERIMENTAL: this API may change without prior notice.
1782 * Unset Ethernet device owner to make the device ownerless.
1785 * The identifier of port to make ownerless.
1787 * The owner identifier.
1789 * 0 on success, negative errno value on error.
1792 int rte_eth_dev_owner_unset(const uint16_t port_id,
1793 const uint64_t owner_id);
1797 * @b EXPERIMENTAL: this API may change without prior notice.
1799 * Remove owner from all Ethernet devices owned by a specific owner.
1802 * The owner identifier.
1804 * 0 on success, negative errno value on error.
1807 int rte_eth_dev_owner_delete(const uint64_t owner_id);
1811 * @b EXPERIMENTAL: this API may change without prior notice.
1813 * Get the owner of an Ethernet device.
1816 * The port identifier.
1818 * The owner structure pointer to fill.
1820 * 0 on success, negative errno value on error..
1823 int rte_eth_dev_owner_get(const uint16_t port_id,
1824 struct rte_eth_dev_owner *owner);
1827 * Get the number of ports which are usable for the application.
1829 * These devices must be iterated by using the macro
1830 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
1831 * to deal with non-contiguous ranges of devices.
1834 * The count of available Ethernet devices.
1836 uint16_t rte_eth_dev_count_avail(void);
1839 * Get the total number of ports which are allocated.
1841 * Some devices may not be available for the application.
1844 * The total count of Ethernet devices.
1846 uint16_t rte_eth_dev_count_total(void);
1849 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
1850 * the bitmap link_speeds of the struct rte_eth_conf
1853 * Numerical speed value in Mbps
1855 * ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
1857 * 0 if the speed cannot be mapped
1859 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
1862 * Get DEV_RX_OFFLOAD_* flag name.
1867 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1869 const char *rte_eth_dev_rx_offload_name(uint64_t offload);
1872 * Get DEV_TX_OFFLOAD_* flag name.
1877 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1879 const char *rte_eth_dev_tx_offload_name(uint64_t offload);
1882 * Configure an Ethernet device.
1883 * This function must be invoked first before any other function in the
1884 * Ethernet API. This function can also be re-invoked when a device is in the
1888 * The port identifier of the Ethernet device to configure.
1889 * @param nb_rx_queue
1890 * The number of receive queues to set up for the Ethernet device.
1891 * @param nb_tx_queue
1892 * The number of transmit queues to set up for the Ethernet device.
1894 * The pointer to the configuration data to be used for the Ethernet device.
1895 * The *rte_eth_conf* structure includes:
1896 * - the hardware offload features to activate, with dedicated fields for
1897 * each statically configurable offload hardware feature provided by
1898 * Ethernet devices, such as IP checksum or VLAN tag stripping for
1900 * The Rx offload bitfield API is obsolete and will be deprecated.
1901 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
1902 * structure and use offloads field to set per-port offloads instead.
1903 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
1904 * the [rt]x_offload_capa returned from rte_eth_dev_info_get().
1905 * Any type of device supported offloading set in the input argument
1906 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
1907 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup()
1908 * - the Receive Side Scaling (RSS) configuration when using multiple RX
1909 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf
1910 * must be within the flow_type_rss_offloads provided by drivers via
1911 * rte_eth_dev_info_get() API.
1913 * Embedding all configuration information in a single data structure
1914 * is the more flexible method that allows the addition of new features
1915 * without changing the syntax of the API.
1917 * - 0: Success, device configured.
1918 * - <0: Error code returned by the driver configuration function.
1920 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
1921 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
1925 * @b EXPERIMENTAL: this API may change without prior notice.
1927 * Check if an Ethernet device was physically removed.
1930 * The port identifier of the Ethernet device.
1932 * 1 when the Ethernet device is removed, otherwise 0.
1936 rte_eth_dev_is_removed(uint16_t port_id);
1939 * Allocate and set up a receive queue for an Ethernet device.
1941 * The function allocates a contiguous block of memory for *nb_rx_desc*
1942 * receive descriptors from a memory zone associated with *socket_id*
1943 * and initializes each receive descriptor with a network buffer allocated
1944 * from the memory pool *mb_pool*.
1947 * The port identifier of the Ethernet device.
1948 * @param rx_queue_id
1949 * The index of the receive queue to set up.
1950 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1951 * to rte_eth_dev_configure().
1953 * The number of receive descriptors to allocate for the receive ring.
1955 * The *socket_id* argument is the socket identifier in case of NUMA.
1956 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1957 * the DMA memory allocated for the receive descriptors of the ring.
1959 * The pointer to the configuration data to be used for the receive queue.
1960 * NULL value is allowed, in which case default RX configuration
1962 * The *rx_conf* structure contains an *rx_thresh* structure with the values
1963 * of the Prefetch, Host, and Write-Back threshold registers of the receive
1965 * In addition it contains the hardware offloads features to activate using
1966 * the DEV_RX_OFFLOAD_* flags.
1967 * If an offloading set in rx_conf->offloads
1968 * hasn't been set in the input argument eth_conf->rxmode.offloads
1969 * to rte_eth_dev_configure(), it is a new added offloading, it must be
1970 * per-queue type and it is enabled for the queue.
1971 * No need to repeat any bit in rx_conf->offloads which has already been
1972 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
1973 * at port level can't be disabled at queue level.
1975 * The pointer to the memory pool from which to allocate *rte_mbuf* network
1976 * memory buffers to populate each descriptor of the receive ring.
1978 * - 0: Success, receive queue correctly set up.
1979 * - -EIO: if device is removed.
1980 * - -EINVAL: The memory pool pointer is null or the size of network buffers
1981 * which can be allocated from this memory pool does not fit the various
1982 * buffer sizes allowed by the device controller.
1983 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
1984 * allocate network memory buffers from the memory pool when
1985 * initializing receive descriptors.
1987 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
1988 uint16_t nb_rx_desc, unsigned int socket_id,
1989 const struct rte_eth_rxconf *rx_conf,
1990 struct rte_mempool *mb_pool);
1994 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1996 * Allocate and set up a hairpin receive queue for an Ethernet device.
1998 * The function set up the selected queue to be used in hairpin.
2001 * The port identifier of the Ethernet device.
2002 * @param rx_queue_id
2003 * The index of the receive queue to set up.
2004 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2005 * to rte_eth_dev_configure().
2007 * The number of receive descriptors to allocate for the receive ring.
2008 * 0 means the PMD will use default value.
2010 * The pointer to the hairpin configuration.
2013 * - (0) if successful.
2014 * - (-ENOTSUP) if hardware doesn't support.
2015 * - (-EINVAL) if bad parameter.
2016 * - (-ENOMEM) if unable to allocate the resources.
2019 int rte_eth_rx_hairpin_queue_setup
2020 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc,
2021 const struct rte_eth_hairpin_conf *conf);
2024 * Allocate and set up a transmit queue for an Ethernet device.
2027 * The port identifier of the Ethernet device.
2028 * @param tx_queue_id
2029 * The index of the transmit queue to set up.
2030 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2031 * to rte_eth_dev_configure().
2033 * The number of transmit descriptors to allocate for the transmit ring.
2035 * The *socket_id* argument is the socket identifier in case of NUMA.
2036 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2037 * the DMA memory allocated for the transmit descriptors of the ring.
2039 * The pointer to the configuration data to be used for the transmit queue.
2040 * NULL value is allowed, in which case default TX configuration
2042 * The *tx_conf* structure contains the following data:
2043 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
2044 * Write-Back threshold registers of the transmit ring.
2045 * When setting Write-Back threshold to the value greater then zero,
2046 * *tx_rs_thresh* value should be explicitly set to one.
2047 * - The *tx_free_thresh* value indicates the [minimum] number of network
2048 * buffers that must be pending in the transmit ring to trigger their
2049 * [implicit] freeing by the driver transmit function.
2050 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
2051 * descriptors that must be pending in the transmit ring before setting the
2052 * RS bit on a descriptor by the driver transmit function.
2053 * The *tx_rs_thresh* value should be less or equal then
2054 * *tx_free_thresh* value, and both of them should be less then
2056 * - The *offloads* member contains Tx offloads to be enabled.
2057 * If an offloading set in tx_conf->offloads
2058 * hasn't been set in the input argument eth_conf->txmode.offloads
2059 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2060 * per-queue type and it is enabled for the queue.
2061 * No need to repeat any bit in tx_conf->offloads which has already been
2062 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2063 * at port level can't be disabled at queue level.
2065 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
2066 * the transmit function to use default values.
2068 * - 0: Success, the transmit queue is correctly set up.
2069 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
2071 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
2072 uint16_t nb_tx_desc, unsigned int socket_id,
2073 const struct rte_eth_txconf *tx_conf);
2077 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2079 * Allocate and set up a transmit hairpin queue for an Ethernet device.
2082 * The port identifier of the Ethernet device.
2083 * @param tx_queue_id
2084 * The index of the transmit queue to set up.
2085 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2086 * to rte_eth_dev_configure().
2088 * The number of transmit descriptors to allocate for the transmit ring.
2089 * 0 to set default PMD value.
2091 * The hairpin configuration.
2094 * - (0) if successful.
2095 * - (-ENOTSUP) if hardware doesn't support.
2096 * - (-EINVAL) if bad parameter.
2097 * - (-ENOMEM) if unable to allocate the resources.
2100 int rte_eth_tx_hairpin_queue_setup
2101 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc,
2102 const struct rte_eth_hairpin_conf *conf);
2105 * Return the NUMA socket to which an Ethernet device is connected
2108 * The port identifier of the Ethernet device
2110 * The NUMA socket id to which the Ethernet device is connected or
2111 * a default of zero if the socket could not be determined.
2112 * -1 is returned is the port_id value is out of range.
2114 int rte_eth_dev_socket_id(uint16_t port_id);
2117 * Check if port_id of device is attached
2120 * The port identifier of the Ethernet device
2122 * - 0 if port is out of range or not attached
2123 * - 1 if device is attached
2125 int rte_eth_dev_is_valid_port(uint16_t port_id);
2128 * Start specified RX queue of a port. It is used when rx_deferred_start
2129 * flag of the specified queue is true.
2132 * The port identifier of the Ethernet device
2133 * @param rx_queue_id
2134 * The index of the rx queue to update the ring.
2135 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2136 * to rte_eth_dev_configure().
2138 * - 0: Success, the receive queue is started.
2139 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2140 * - -EIO: if device is removed.
2141 * - -ENOTSUP: The function not supported in PMD driver.
2143 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2146 * Stop specified RX queue of a port
2149 * The port identifier of the Ethernet device
2150 * @param rx_queue_id
2151 * The index of the rx queue to update the ring.
2152 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2153 * to rte_eth_dev_configure().
2155 * - 0: Success, the receive queue is stopped.
2156 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2157 * - -EIO: if device is removed.
2158 * - -ENOTSUP: The function not supported in PMD driver.
2160 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2163 * Start TX for specified queue of a port. It is used when tx_deferred_start
2164 * flag of the specified queue is true.
2167 * The port identifier of the Ethernet device
2168 * @param tx_queue_id
2169 * The index of the tx queue to update the ring.
2170 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2171 * to rte_eth_dev_configure().
2173 * - 0: Success, the transmit queue is started.
2174 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2175 * - -EIO: if device is removed.
2176 * - -ENOTSUP: The function not supported in PMD driver.
2178 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2181 * Stop specified TX queue of a port
2184 * The port identifier of the Ethernet device
2185 * @param tx_queue_id
2186 * The index of the tx queue to update the ring.
2187 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2188 * to rte_eth_dev_configure().
2190 * - 0: Success, the transmit queue is stopped.
2191 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2192 * - -EIO: if device is removed.
2193 * - -ENOTSUP: The function not supported in PMD driver.
2195 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2198 * Start an Ethernet device.
2200 * The device start step is the last one and consists of setting the configured
2201 * offload features and in starting the transmit and the receive units of the
2204 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before
2205 * PMD port start callback function is invoked.
2207 * On success, all basic functions exported by the Ethernet API (link status,
2208 * receive/transmit, and so on) can be invoked.
2211 * The port identifier of the Ethernet device.
2213 * - 0: Success, Ethernet device started.
2214 * - <0: Error code of the driver device start function.
2216 int rte_eth_dev_start(uint16_t port_id);
2219 * Stop an Ethernet device. The device can be restarted with a call to
2220 * rte_eth_dev_start()
2223 * The port identifier of the Ethernet device.
2225 void rte_eth_dev_stop(uint16_t port_id);
2228 * Link up an Ethernet device.
2230 * Set device link up will re-enable the device rx/tx
2231 * functionality after it is previously set device linked down.
2234 * The port identifier of the Ethernet device.
2236 * - 0: Success, Ethernet device linked up.
2237 * - <0: Error code of the driver device link up function.
2239 int rte_eth_dev_set_link_up(uint16_t port_id);
2242 * Link down an Ethernet device.
2243 * The device rx/tx functionality will be disabled if success,
2244 * and it can be re-enabled with a call to
2245 * rte_eth_dev_set_link_up()
2248 * The port identifier of the Ethernet device.
2250 int rte_eth_dev_set_link_down(uint16_t port_id);
2253 * Close a stopped Ethernet device. The device cannot be restarted!
2254 * The function frees all port resources if the driver supports
2255 * the flag RTE_ETH_DEV_CLOSE_REMOVE.
2258 * The port identifier of the Ethernet device.
2260 void rte_eth_dev_close(uint16_t port_id);
2263 * Reset a Ethernet device and keep its port id.
2265 * When a port has to be reset passively, the DPDK application can invoke
2266 * this function. For example when a PF is reset, all its VFs should also
2267 * be reset. Normally a DPDK application can invoke this function when
2268 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2269 * a port reset in other circumstances.
2271 * When this function is called, it first stops the port and then calls the
2272 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2273 * state, in which no Tx and Rx queues are setup, as if the port has been
2274 * reset and not started. The port keeps the port id it had before the
2277 * After calling rte_eth_dev_reset( ), the application should use
2278 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2279 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2280 * to reconfigure the device as appropriate.
2282 * Note: To avoid unexpected behavior, the application should stop calling
2283 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2284 * safety, all these controlling functions should be called from the same
2288 * The port identifier of the Ethernet device.
2291 * - (0) if successful.
2292 * - (-EINVAL) if port identifier is invalid.
2293 * - (-ENOTSUP) if hardware doesn't support this function.
2294 * - (-EPERM) if not ran from the primary process.
2295 * - (-EIO) if re-initialisation failed or device is removed.
2296 * - (-ENOMEM) if the reset failed due to OOM.
2297 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2299 int rte_eth_dev_reset(uint16_t port_id);
2302 * Enable receipt in promiscuous mode for an Ethernet device.
2305 * The port identifier of the Ethernet device.
2307 * - (0) if successful.
2308 * - (-ENOTSUP) if support for promiscuous_enable() does not exist
2310 * - (-ENODEV) if *port_id* invalid.
2312 int rte_eth_promiscuous_enable(uint16_t port_id);
2315 * Disable receipt in promiscuous mode for an Ethernet device.
2318 * The port identifier of the Ethernet device.
2320 * - (0) if successful.
2321 * - (-ENOTSUP) if support for promiscuous_disable() does not exist
2323 * - (-ENODEV) if *port_id* invalid.
2325 int rte_eth_promiscuous_disable(uint16_t port_id);
2328 * Return the value of promiscuous mode for an Ethernet device.
2331 * The port identifier of the Ethernet device.
2333 * - (1) if promiscuous is enabled
2334 * - (0) if promiscuous is disabled.
2337 int rte_eth_promiscuous_get(uint16_t port_id);
2340 * Enable the receipt of any multicast frame by an Ethernet device.
2343 * The port identifier of the Ethernet device.
2345 * - (0) if successful.
2346 * - (-ENOTSUP) if support for allmulticast_enable() does not exist
2348 * - (-ENODEV) if *port_id* invalid.
2350 int rte_eth_allmulticast_enable(uint16_t port_id);
2353 * Disable the receipt of all multicast frames by an Ethernet device.
2356 * The port identifier of the Ethernet device.
2358 * - (0) if successful.
2359 * - (-ENOTSUP) if support for allmulticast_disable() does not exist
2361 * - (-ENODEV) if *port_id* invalid.
2363 int rte_eth_allmulticast_disable(uint16_t port_id);
2366 * Return the value of allmulticast mode for an Ethernet device.
2369 * The port identifier of the Ethernet device.
2371 * - (1) if allmulticast is enabled
2372 * - (0) if allmulticast is disabled.
2375 int rte_eth_allmulticast_get(uint16_t port_id);
2378 * Retrieve the link status (up/down), the duplex mode (half/full),
2379 * the negotiation (auto/fixed), and if available, the speed (Mbps).
2381 * It might need to wait up to 9 seconds.
2382 * @see rte_eth_link_get_nowait.
2385 * The port identifier of the Ethernet device.
2387 * Link information written back.
2389 * - (0) if successful.
2390 * - (-ENOTSUP) if the function is not supported in PMD driver.
2391 * - (-ENODEV) if *port_id* invalid.
2393 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
2396 * Retrieve the link status (up/down), the duplex mode (half/full),
2397 * the negotiation (auto/fixed), and if available, the speed (Mbps).
2400 * The port identifier of the Ethernet device.
2402 * Link information written back.
2404 * - (0) if successful.
2405 * - (-ENOTSUP) if the function is not supported in PMD driver.
2406 * - (-ENODEV) if *port_id* invalid.
2408 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
2411 * Retrieve the general I/O statistics of an Ethernet device.
2414 * The port identifier of the Ethernet device.
2416 * A pointer to a structure of type *rte_eth_stats* to be filled with
2417 * the values of device counters for the following set of statistics:
2418 * - *ipackets* with the total of successfully received packets.
2419 * - *opackets* with the total of successfully transmitted packets.
2420 * - *ibytes* with the total of successfully received bytes.
2421 * - *obytes* with the total of successfully transmitted bytes.
2422 * - *ierrors* with the total of erroneous received packets.
2423 * - *oerrors* with the total of failed transmitted packets.
2425 * Zero if successful. Non-zero otherwise.
2427 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
2430 * Reset the general I/O statistics of an Ethernet device.
2433 * The port identifier of the Ethernet device.
2435 * - (0) if device notified to reset stats.
2436 * - (-ENOTSUP) if hardware doesn't support.
2437 * - (-ENODEV) if *port_id* invalid.
2438 * - (<0): Error code of the driver stats reset function.
2440 int rte_eth_stats_reset(uint16_t port_id);
2443 * Retrieve names of extended statistics of an Ethernet device.
2445 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
2447 * xstats_names[i].name => xstats[i].value
2449 * And the array index is same with id field of 'struct rte_eth_xstat':
2452 * This assumption makes key-value pair matching less flexible but simpler.
2455 * The port identifier of the Ethernet device.
2456 * @param xstats_names
2457 * An rte_eth_xstat_name array of at least *size* elements to
2458 * be filled. If set to NULL, the function returns the required number
2461 * The size of the xstats_names array (number of elements).
2463 * - A positive value lower or equal to size: success. The return value
2464 * is the number of entries filled in the stats table.
2465 * - A positive value higher than size: error, the given statistics table
2466 * is too small. The return value corresponds to the size that should
2467 * be given to succeed. The entries in the table are not valid and
2468 * shall not be used by the caller.
2469 * - A negative value on error (invalid port id).
2471 int rte_eth_xstats_get_names(uint16_t port_id,
2472 struct rte_eth_xstat_name *xstats_names,
2476 * Retrieve extended statistics of an Ethernet device.
2478 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
2480 * xstats_names[i].name => xstats[i].value
2482 * And the array index is same with id field of 'struct rte_eth_xstat':
2485 * This assumption makes key-value pair matching less flexible but simpler.
2488 * The port identifier of the Ethernet device.
2490 * A pointer to a table of structure of type *rte_eth_xstat*
2491 * to be filled with device statistics ids and values.
2492 * This parameter can be set to NULL if n is 0.
2494 * The size of the xstats array (number of elements).
2496 * - A positive value lower or equal to n: success. The return value
2497 * is the number of entries filled in the stats table.
2498 * - A positive value higher than n: error, the given statistics table
2499 * is too small. The return value corresponds to the size that should
2500 * be given to succeed. The entries in the table are not valid and
2501 * shall not be used by the caller.
2502 * - A negative value on error (invalid port id).
2504 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
2508 * Retrieve names of extended statistics of an Ethernet device.
2511 * The port identifier of the Ethernet device.
2512 * @param xstats_names
2513 * An rte_eth_xstat_name array of at least *size* elements to
2514 * be filled. If set to NULL, the function returns the required number
2517 * IDs array given by app to retrieve specific statistics
2519 * The size of the xstats_names array (number of elements).
2521 * - A positive value lower or equal to size: success. The return value
2522 * is the number of entries filled in the stats table.
2523 * - A positive value higher than size: error, the given statistics table
2524 * is too small. The return value corresponds to the size that should
2525 * be given to succeed. The entries in the table are not valid and
2526 * shall not be used by the caller.
2527 * - A negative value on error (invalid port id).
2530 rte_eth_xstats_get_names_by_id(uint16_t port_id,
2531 struct rte_eth_xstat_name *xstats_names, unsigned int size,
2535 * Retrieve extended statistics of an Ethernet device.
2538 * The port identifier of the Ethernet device.
2540 * A pointer to an ids array passed by application. This tells which
2541 * statistics values function should retrieve. This parameter
2542 * can be set to NULL if size is 0. In this case function will retrieve
2543 * all available statistics.
2545 * A pointer to a table to be filled with device statistics values.
2547 * The size of the ids array (number of elements).
2549 * - A positive value lower or equal to size: success. The return value
2550 * is the number of entries filled in the stats table.
2551 * - A positive value higher than size: error, the given statistics table
2552 * is too small. The return value corresponds to the size that should
2553 * be given to succeed. The entries in the table are not valid and
2554 * shall not be used by the caller.
2555 * - A negative value on error (invalid port id).
2557 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
2558 uint64_t *values, unsigned int size);
2561 * Gets the ID of a statistic from its name.
2563 * This function searches for the statistics using string compares, and
2564 * as such should not be used on the fast-path. For fast-path retrieval of
2565 * specific statistics, store the ID as provided in *id* from this function,
2566 * and pass the ID to rte_eth_xstats_get()
2568 * @param port_id The port to look up statistics from
2569 * @param xstat_name The name of the statistic to return
2570 * @param[out] id A pointer to an app-supplied uint64_t which should be
2571 * set to the ID of the stat if the stat exists.
2574 * -ENODEV for invalid port_id,
2575 * -EIO if device is removed,
2576 * -EINVAL if the xstat_name doesn't exist in port_id
2578 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
2582 * Reset extended statistics of an Ethernet device.
2585 * The port identifier of the Ethernet device.
2587 * - (0) if device notified to reset extended stats.
2588 * - (-ENOTSUP) if pmd doesn't support both
2589 * extended stats and basic stats reset.
2590 * - (-ENODEV) if *port_id* invalid.
2591 * - (<0): Error code of the driver xstats reset function.
2593 int rte_eth_xstats_reset(uint16_t port_id);
2596 * Set a mapping for the specified transmit queue to the specified per-queue
2597 * statistics counter.
2600 * The port identifier of the Ethernet device.
2601 * @param tx_queue_id
2602 * The index of the transmit queue for which a queue stats mapping is required.
2603 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2604 * to rte_eth_dev_configure().
2606 * The per-queue packet statistics functionality number that the transmit
2607 * queue is to be assigned.
2608 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
2610 * Zero if successful. Non-zero otherwise.
2612 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
2613 uint16_t tx_queue_id, uint8_t stat_idx);
2616 * Set a mapping for the specified receive queue to the specified per-queue
2617 * statistics counter.
2620 * The port identifier of the Ethernet device.
2621 * @param rx_queue_id
2622 * The index of the receive queue for which a queue stats mapping is required.
2623 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2624 * to rte_eth_dev_configure().
2626 * The per-queue packet statistics functionality number that the receive
2627 * queue is to be assigned.
2628 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
2630 * Zero if successful. Non-zero otherwise.
2632 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
2633 uint16_t rx_queue_id,
2637 * Retrieve the Ethernet address of an Ethernet device.
2640 * The port identifier of the Ethernet device.
2642 * A pointer to a structure of type *ether_addr* to be filled with
2643 * the Ethernet address of the Ethernet device.
2645 * - (0) if successful
2646 * - (-ENODEV) if *port_id* invalid.
2648 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr);
2651 * Retrieve the contextual information of an Ethernet device.
2653 * As part of this function, a number of of fields in dev_info will be
2654 * initialized as follows:
2659 * Where lim is defined within the rte_eth_dev_info_get as
2661 * const struct rte_eth_desc_lim lim = {
2662 * .nb_max = UINT16_MAX,
2665 * .nb_seg_max = UINT16_MAX,
2666 * .nb_mtu_seg_max = UINT16_MAX,
2669 * device = dev->device
2670 * min_mtu = RTE_ETHER_MIN_MTU
2671 * max_mtu = UINT16_MAX
2673 * The following fields will be populated if support for dev_infos_get()
2674 * exists for the device and the rte_eth_dev 'dev' has been populated
2675 * successfully with a call to it:
2677 * driver_name = dev->device->driver->name
2678 * nb_rx_queues = dev->data->nb_rx_queues
2679 * nb_tx_queues = dev->data->nb_tx_queues
2680 * dev_flags = &dev->data->dev_flags
2683 * The port identifier of the Ethernet device.
2685 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
2686 * the contextual information of the Ethernet device.
2688 * - (0) if successful.
2689 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device.
2690 * - (-ENODEV) if *port_id* invalid.
2692 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
2695 * Retrieve the firmware version of a device.
2698 * The port identifier of the device.
2700 * A pointer to a string array storing the firmware version of a device,
2701 * the string includes terminating null. This pointer is allocated by caller.
2703 * The size of the string array pointed by fw_version, which should be
2704 * large enough to store firmware version of the device.
2706 * - (0) if successful.
2707 * - (-ENOTSUP) if operation is not supported.
2708 * - (-ENODEV) if *port_id* invalid.
2709 * - (-EIO) if device is removed.
2710 * - (>0) if *fw_size* is not enough to store firmware version, return
2711 * the size of the non truncated string.
2713 int rte_eth_dev_fw_version_get(uint16_t port_id,
2714 char *fw_version, size_t fw_size);
2717 * Retrieve the supported packet types of an Ethernet device.
2719 * When a packet type is announced as supported, it *must* be recognized by
2720 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
2721 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
2722 * packet types for these packets:
2723 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
2724 * - Ether/Vlan/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
2725 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
2726 * - Ether/Vlan/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
2728 * When a packet is received by a PMD, the most precise type must be
2729 * returned among the ones supported. However a PMD is allowed to set
2730 * packet type that is not in the supported list, at the condition that it
2731 * is more precise. Therefore, a PMD announcing no supported packet types
2732 * can still set a matching packet type in a received packet.
2735 * Better to invoke this API after the device is already started or rx burst
2736 * function is decided, to obtain correct supported ptypes.
2738 * if a given PMD does not report what ptypes it supports, then the supported
2739 * ptype count is reported as 0.
2741 * The port identifier of the Ethernet device.
2743 * A hint of what kind of packet type which the caller is interested in.
2745 * An array pointer to store adequate packet types, allocated by caller.
2747 * Size of the array pointed by param ptypes.
2749 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
2750 * only num entries will be filled into the ptypes array, but the full
2751 * count of supported ptypes will be returned.
2752 * - (-ENODEV) if *port_id* invalid.
2754 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
2755 uint32_t *ptypes, int num);
2758 * @b EXPERIMENTAL: this API may change without prior notice.
2760 * Inform Ethernet device about reduced range of packet types to handle.
2762 * Application can use this function to set only specific ptypes that it's
2763 * interested. This information can be used by the PMD to optimize Rx path.
2765 * The function accepts an array `set_ptypes` allocated by the caller to
2766 * store the packet types set by the driver, the last element of the array
2767 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be
2768 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled
2772 * The port identifier of the Ethernet device.
2774 * The ptype family that application is interested in should be bitwise OR of
2775 * RTE_PTYPE_*_MASK or 0.
2777 * An array pointer to store set packet types, allocated by caller. The
2778 * function marks the end of array with RTE_PTYPE_UNKNOWN.
2780 * Size of the array pointed by param ptypes.
2781 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the
2785 * - (-ENODEV) if *port_id* invalid.
2786 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and
2790 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask,
2791 uint32_t *set_ptypes, unsigned int num);
2794 * Retrieve the MTU of an Ethernet device.
2797 * The port identifier of the Ethernet device.
2799 * A pointer to a uint16_t where the retrieved MTU is to be stored.
2801 * - (0) if successful.
2802 * - (-ENODEV) if *port_id* invalid.
2804 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
2807 * Change the MTU of an Ethernet device.
2810 * The port identifier of the Ethernet device.
2812 * A uint16_t for the MTU to be applied.
2814 * - (0) if successful.
2815 * - (-ENOTSUP) if operation is not supported.
2816 * - (-ENODEV) if *port_id* invalid.
2817 * - (-EIO) if device is removed.
2818 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within
2819 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or
2820 * when the mtu is set using dev->dev_ops->mtu_set.
2821 * - (-EBUSY) if operation is not allowed when the port is running
2823 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
2826 * Enable/Disable hardware filtering by an Ethernet device of received
2827 * VLAN packets tagged with a given VLAN Tag Identifier.
2830 * The port identifier of the Ethernet device.
2832 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
2834 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
2835 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
2837 * - (0) if successful.
2838 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
2839 * - (-ENODEV) if *port_id* invalid.
2840 * - (-EIO) if device is removed.
2841 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
2842 * - (-EINVAL) if *vlan_id* > 4095.
2844 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
2847 * Enable/Disable hardware VLAN Strip by a rx queue of an Ethernet device.
2850 * The port identifier of the Ethernet device.
2851 * @param rx_queue_id
2852 * The index of the receive queue for which a queue stats mapping is required.
2853 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2854 * to rte_eth_dev_configure().
2856 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
2857 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
2859 * - (0) if successful.
2860 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured.
2861 * - (-ENODEV) if *port_id* invalid.
2862 * - (-EINVAL) if *rx_queue_id* invalid.
2864 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
2868 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
2872 * The port identifier of the Ethernet device.
2876 * The Tag Protocol ID
2878 * - (0) if successful.
2879 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported.
2880 * - (-ENODEV) if *port_id* invalid.
2881 * - (-EIO) if device is removed.
2883 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
2884 enum rte_vlan_type vlan_type,
2888 * Set VLAN offload configuration on an Ethernet device.
2891 * The port identifier of the Ethernet device.
2892 * @param offload_mask
2893 * The VLAN Offload bit mask can be mixed use with "OR"
2894 * ETH_VLAN_STRIP_OFFLOAD
2895 * ETH_VLAN_FILTER_OFFLOAD
2896 * ETH_VLAN_EXTEND_OFFLOAD
2897 * ETH_QINQ_STRIP_OFFLOAD
2899 * - (0) if successful.
2900 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
2901 * - (-ENODEV) if *port_id* invalid.
2902 * - (-EIO) if device is removed.
2904 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
2907 * Read VLAN Offload configuration from an Ethernet device
2910 * The port identifier of the Ethernet device.
2912 * - (>0) if successful. Bit mask to indicate
2913 * ETH_VLAN_STRIP_OFFLOAD
2914 * ETH_VLAN_FILTER_OFFLOAD
2915 * ETH_VLAN_EXTEND_OFFLOAD
2916 * ETH_QINQ_STRIP_OFFLOAD
2917 * - (-ENODEV) if *port_id* invalid.
2919 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
2922 * Set port based TX VLAN insertion on or off.
2925 * The port identifier of the Ethernet device.
2927 * Port based TX VLAN identifier together with user priority.
2929 * Turn on or off the port based TX VLAN insertion.
2932 * - (0) if successful.
2933 * - negative if failed.
2935 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
2937 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
2941 * Structure used to buffer packets for future TX
2942 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
2944 struct rte_eth_dev_tx_buffer {
2945 buffer_tx_error_fn error_callback;
2946 void *error_userdata;
2947 uint16_t size; /**< Size of buffer for buffered tx */
2948 uint16_t length; /**< Number of packets in the array */
2949 struct rte_mbuf *pkts[];
2950 /**< Pending packets to be sent on explicit flush or when full */
2954 * Calculate the size of the tx buffer.
2957 * Number of stored packets.
2959 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
2960 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
2963 * Initialize default values for buffered transmitting
2966 * Tx buffer to be initialized.
2973 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
2976 * Configure a callback for buffered packets which cannot be sent
2978 * Register a specific callback to be called when an attempt is made to send
2979 * all packets buffered on an ethernet port, but not all packets can
2980 * successfully be sent. The callback registered here will be called only
2981 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
2982 * The default callback configured for each queue by default just frees the
2983 * packets back to the calling mempool. If additional behaviour is required,
2984 * for example, to count dropped packets, or to retry transmission of packets
2985 * which cannot be sent, this function should be used to register a suitable
2986 * callback function to implement the desired behaviour.
2987 * The example callback "rte_eth_count_unsent_packet_callback()" is also
2988 * provided as reference.
2991 * The port identifier of the Ethernet device.
2993 * The function to be used as the callback.
2995 * Arbitrary parameter to be passed to the callback function
2997 * 0 on success, or -1 on error with rte_errno set appropriately
3000 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
3001 buffer_tx_error_fn callback, void *userdata);
3004 * Callback function for silently dropping unsent buffered packets.
3006 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3007 * adjust the default behavior when buffered packets cannot be sent. This
3008 * function drops any unsent packets silently and is used by tx buffered
3009 * operations as default behavior.
3011 * NOTE: this function should not be called directly, instead it should be used
3012 * as a callback for packet buffering.
3014 * NOTE: when configuring this function as a callback with
3015 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3016 * should point to an uint64_t value.
3019 * The previously buffered packets which could not be sent
3021 * The number of unsent packets in the pkts array
3026 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
3030 * Callback function for tracking unsent buffered packets.
3032 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3033 * adjust the default behavior when buffered packets cannot be sent. This
3034 * function drops any unsent packets, but also updates a user-supplied counter
3035 * to track the overall number of packets dropped. The counter should be an
3036 * uint64_t variable.
3038 * NOTE: this function should not be called directly, instead it should be used
3039 * as a callback for packet buffering.
3041 * NOTE: when configuring this function as a callback with
3042 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3043 * should point to an uint64_t value.
3046 * The previously buffered packets which could not be sent
3048 * The number of unsent packets in the pkts array
3050 * Pointer to an uint64_t value, which will be incremented by unsent
3053 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
3057 * Request the driver to free mbufs currently cached by the driver. The
3058 * driver will only free the mbuf if it is no longer in use. It is the
3059 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is
3063 * The port identifier of the Ethernet device.
3065 * The index of the transmit queue through which output packets must be
3067 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3068 * to rte_eth_dev_configure().
3070 * Maximum number of packets to free. Use 0 to indicate all possible packets
3071 * should be freed. Note that a packet may be using multiple mbufs.
3074 * -ENODEV: Invalid interface
3075 * -EIO: device is removed
3076 * -ENOTSUP: Driver does not support function
3078 * 0-n: Number of packets freed. More packets may still remain in ring that
3082 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
3085 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
3088 enum rte_eth_event_ipsec_subtype {
3089 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
3090 /**< Unknown event type */
3091 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
3092 /**< Sequence number overflow */
3093 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
3094 /**< Soft time expiry of SA */
3095 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
3096 /**< Soft byte expiry of SA */
3097 RTE_ETH_EVENT_IPSEC_MAX
3098 /**< Max value of this enum */
3102 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
3103 * information of the IPsec offload event.
3105 struct rte_eth_event_ipsec_desc {
3106 enum rte_eth_event_ipsec_subtype subtype;
3107 /**< Type of RTE_ETH_EVENT_IPSEC_* event */
3109 /**< Event specific metadata
3111 * For the following events, *userdata* registered
3112 * with the *rte_security_session* would be returned
3115 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
3116 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
3117 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
3119 * @see struct rte_security_session_conf
3125 * The eth device event type for interrupt, and maybe others in the future.
3127 enum rte_eth_event_type {
3128 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3129 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3130 RTE_ETH_EVENT_QUEUE_STATE,
3131 /**< queue state event (enabled/disabled) */
3132 RTE_ETH_EVENT_INTR_RESET,
3133 /**< reset interrupt event, sent to VF on PF reset */
3134 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3135 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3136 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3137 RTE_ETH_EVENT_NEW, /**< port is probed */
3138 RTE_ETH_EVENT_DESTROY, /**< port is released */
3139 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
3140 RTE_ETH_EVENT_FLOW_AGED,/**< New aged-out flows is detected */
3141 RTE_ETH_EVENT_MAX /**< max value of this enum */
3144 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3145 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3146 /**< user application callback to be registered for interrupts */
3149 * Register a callback function for port event.
3153 * RTE_ETH_ALL means register the event for all port ids.
3157 * User supplied callback function to be called.
3159 * Pointer to the parameters for the registered callback.
3162 * - On success, zero.
3163 * - On failure, a negative value.
3165 int rte_eth_dev_callback_register(uint16_t port_id,
3166 enum rte_eth_event_type event,
3167 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3170 * Unregister a callback function for port event.
3174 * RTE_ETH_ALL means unregister the event for all port ids.
3178 * User supplied callback function to be called.
3180 * Pointer to the parameters for the registered callback. -1 means to
3181 * remove all for the same callback address and same event.
3184 * - On success, zero.
3185 * - On failure, a negative value.
3187 int rte_eth_dev_callback_unregister(uint16_t port_id,
3188 enum rte_eth_event_type event,
3189 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3192 * When there is no rx packet coming in Rx Queue for a long time, we can
3193 * sleep lcore related to RX Queue for power saving, and enable rx interrupt
3194 * to be triggered when Rx packet arrives.
3196 * The rte_eth_dev_rx_intr_enable() function enables rx queue
3197 * interrupt on specific rx queue of a port.
3200 * The port identifier of the Ethernet device.
3202 * The index of the receive queue from which to retrieve input packets.
3203 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3204 * to rte_eth_dev_configure().
3206 * - (0) if successful.
3207 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3209 * - (-ENODEV) if *port_id* invalid.
3210 * - (-EIO) if device is removed.
3212 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
3215 * When lcore wakes up from rx interrupt indicating packet coming, disable rx
3216 * interrupt and returns to polling mode.
3218 * The rte_eth_dev_rx_intr_disable() function disables rx queue
3219 * interrupt on specific rx queue of a port.
3222 * The port identifier of the Ethernet device.
3224 * The index of the receive queue from which to retrieve input packets.
3225 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3226 * to rte_eth_dev_configure().
3228 * - (0) if successful.
3229 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3231 * - (-ENODEV) if *port_id* invalid.
3232 * - (-EIO) if device is removed.
3234 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
3237 * RX Interrupt control per port.
3240 * The port identifier of the Ethernet device.
3242 * Epoll instance fd which the intr vector associated to.
3243 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3245 * The operation be performed for the vector.
3246 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3250 * - On success, zero.
3251 * - On failure, a negative value.
3253 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
3256 * RX Interrupt control per queue.
3259 * The port identifier of the Ethernet device.
3261 * The index of the receive queue from which to retrieve input packets.
3262 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3263 * to rte_eth_dev_configure().
3265 * Epoll instance fd which the intr vector associated to.
3266 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3268 * The operation be performed for the vector.
3269 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3273 * - On success, zero.
3274 * - On failure, a negative value.
3276 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
3277 int epfd, int op, void *data);
3281 * @b EXPERIMENTAL: this API may change without prior notice.
3283 * Get interrupt fd per Rx queue.
3286 * The port identifier of the Ethernet device.
3288 * The index of the receive queue from which to retrieve input packets.
3289 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3290 * to rte_eth_dev_configure().
3292 * - (>=0) the interrupt fd associated to the requested Rx queue if
3298 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id);
3301 * Turn on the LED on the Ethernet device.
3302 * This function turns on the LED on the Ethernet device.
3305 * The port identifier of the Ethernet device.
3307 * - (0) if successful.
3308 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3310 * - (-ENODEV) if *port_id* invalid.
3311 * - (-EIO) if device is removed.
3313 int rte_eth_led_on(uint16_t port_id);
3316 * Turn off the LED on the Ethernet device.
3317 * This function turns off the LED on the Ethernet device.
3320 * The port identifier of the Ethernet device.
3322 * - (0) if successful.
3323 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3325 * - (-ENODEV) if *port_id* invalid.
3326 * - (-EIO) if device is removed.
3328 int rte_eth_led_off(uint16_t port_id);
3331 * Get current status of the Ethernet link flow control for Ethernet device
3334 * The port identifier of the Ethernet device.
3336 * The pointer to the structure where to store the flow control parameters.
3338 * - (0) if successful.
3339 * - (-ENOTSUP) if hardware doesn't support flow control.
3340 * - (-ENODEV) if *port_id* invalid.
3341 * - (-EIO) if device is removed.
3343 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
3344 struct rte_eth_fc_conf *fc_conf);
3347 * Configure the Ethernet link flow control for Ethernet device
3350 * The port identifier of the Ethernet device.
3352 * The pointer to the structure of the flow control parameters.
3354 * - (0) if successful.
3355 * - (-ENOTSUP) if hardware doesn't support flow control mode.
3356 * - (-ENODEV) if *port_id* invalid.
3357 * - (-EINVAL) if bad parameter
3358 * - (-EIO) if flow control setup failure or device is removed.
3360 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
3361 struct rte_eth_fc_conf *fc_conf);
3364 * Configure the Ethernet priority flow control under DCB environment
3365 * for Ethernet device.
3368 * The port identifier of the Ethernet device.
3370 * The pointer to the structure of the priority flow control parameters.
3372 * - (0) if successful.
3373 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
3374 * - (-ENODEV) if *port_id* invalid.
3375 * - (-EINVAL) if bad parameter
3376 * - (-EIO) if flow control setup failure or device is removed.
3378 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
3379 struct rte_eth_pfc_conf *pfc_conf);
3382 * Add a MAC address to the set used for filtering incoming packets.
3385 * The port identifier of the Ethernet device.
3387 * The MAC address to add.
3389 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
3390 * not enabled, this should be set to 0.
3392 * - (0) if successfully added or *mac_addr* was already added.
3393 * - (-ENOTSUP) if hardware doesn't support this feature.
3394 * - (-ENODEV) if *port* is invalid.
3395 * - (-EIO) if device is removed.
3396 * - (-ENOSPC) if no more MAC addresses can be added.
3397 * - (-EINVAL) if MAC address is invalid.
3399 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr,
3403 * Remove a MAC address from the internal array of addresses.
3406 * The port identifier of the Ethernet device.
3408 * MAC address to remove.
3410 * - (0) if successful, or *mac_addr* didn't exist.
3411 * - (-ENOTSUP) if hardware doesn't support.
3412 * - (-ENODEV) if *port* invalid.
3413 * - (-EADDRINUSE) if attempting to remove the default MAC address
3415 int rte_eth_dev_mac_addr_remove(uint16_t port_id,
3416 struct rte_ether_addr *mac_addr);
3419 * Set the default MAC address.
3422 * The port identifier of the Ethernet device.
3424 * New default MAC address.
3426 * - (0) if successful, or *mac_addr* didn't exist.
3427 * - (-ENOTSUP) if hardware doesn't support.
3428 * - (-ENODEV) if *port* invalid.
3429 * - (-EINVAL) if MAC address is invalid.
3431 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
3432 struct rte_ether_addr *mac_addr);
3435 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3438 * The port identifier of the Ethernet device.
3442 * Redirection table size. The table size can be queried by
3443 * rte_eth_dev_info_get().
3445 * - (0) if successful.
3446 * - (-ENOTSUP) if hardware doesn't support.
3447 * - (-EINVAL) if bad parameter.
3448 * - (-EIO) if device is removed.
3450 int rte_eth_dev_rss_reta_update(uint16_t port_id,
3451 struct rte_eth_rss_reta_entry64 *reta_conf,
3452 uint16_t reta_size);
3455 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3458 * The port identifier of the Ethernet device.
3460 * RETA to query. For each requested reta entry, corresponding bit
3461 * in mask must be set.
3463 * Redirection table size. The table size can be queried by
3464 * rte_eth_dev_info_get().
3466 * - (0) if successful.
3467 * - (-ENOTSUP) if hardware doesn't support.
3468 * - (-EINVAL) if bad parameter.
3469 * - (-EIO) if device is removed.
3471 int rte_eth_dev_rss_reta_query(uint16_t port_id,
3472 struct rte_eth_rss_reta_entry64 *reta_conf,
3473 uint16_t reta_size);
3476 * Updates unicast hash table for receiving packet with the given destination
3477 * MAC address, and the packet is routed to all VFs for which the RX mode is
3478 * accept packets that match the unicast hash table.
3481 * The port identifier of the Ethernet device.
3483 * Unicast MAC address.
3485 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
3486 * 0 - Clear an unicast hash bit.
3488 * - (0) if successful.
3489 * - (-ENOTSUP) if hardware doesn't support.
3490 * - (-ENODEV) if *port_id* invalid.
3491 * - (-EIO) if device is removed.
3492 * - (-EINVAL) if bad parameter.
3494 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr,
3498 * Updates all unicast hash bitmaps for receiving packet with any Unicast
3499 * Ethernet MAC addresses,the packet is routed to all VFs for which the RX
3500 * mode is accept packets that match the unicast hash table.
3503 * The port identifier of the Ethernet device.
3505 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
3507 * 0 - Clear all unicast hash bitmaps
3509 * - (0) if successful.
3510 * - (-ENOTSUP) if hardware doesn't support.
3511 * - (-ENODEV) if *port_id* invalid.
3512 * - (-EIO) if device is removed.
3513 * - (-EINVAL) if bad parameter.
3515 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
3518 * Set a traffic mirroring rule on an Ethernet device
3521 * The port identifier of the Ethernet device.
3522 * @param mirror_conf
3523 * The pointer to the traffic mirroring structure describing the mirroring rule.
3524 * The *rte_eth_vm_mirror_conf* structure includes the type of mirroring rule,
3525 * destination pool and the value of rule if enable vlan or pool mirroring.
3528 * The index of traffic mirroring rule, we support four separated rules.
3530 * 1 - Enable a mirroring rule.
3531 * 0 - Disable a mirroring rule.
3533 * - (0) if successful.
3534 * - (-ENOTSUP) if hardware doesn't support this feature.
3535 * - (-ENODEV) if *port_id* invalid.
3536 * - (-EIO) if device is removed.
3537 * - (-EINVAL) if the mr_conf information is not correct.
3539 int rte_eth_mirror_rule_set(uint16_t port_id,
3540 struct rte_eth_mirror_conf *mirror_conf,
3545 * Reset a traffic mirroring rule on an Ethernet device.
3548 * The port identifier of the Ethernet device.
3550 * The index of traffic mirroring rule, we support four separated rules.
3552 * - (0) if successful.
3553 * - (-ENOTSUP) if hardware doesn't support this feature.
3554 * - (-ENODEV) if *port_id* invalid.
3555 * - (-EIO) if device is removed.
3556 * - (-EINVAL) if bad parameter.
3558 int rte_eth_mirror_rule_reset(uint16_t port_id,
3562 * Set the rate limitation for a queue on an Ethernet device.
3565 * The port identifier of the Ethernet device.
3569 * The tx rate in Mbps. Allocated from the total port link speed.
3571 * - (0) if successful.
3572 * - (-ENOTSUP) if hardware doesn't support this feature.
3573 * - (-ENODEV) if *port_id* invalid.
3574 * - (-EIO) if device is removed.
3575 * - (-EINVAL) if bad parameter.
3577 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
3581 * Configuration of Receive Side Scaling hash computation of Ethernet device.
3584 * The port identifier of the Ethernet device.
3586 * The new configuration to use for RSS hash computation on the port.
3588 * - (0) if successful.
3589 * - (-ENODEV) if port identifier is invalid.
3590 * - (-EIO) if device is removed.
3591 * - (-ENOTSUP) if hardware doesn't support.
3592 * - (-EINVAL) if bad parameter.
3594 int rte_eth_dev_rss_hash_update(uint16_t port_id,
3595 struct rte_eth_rss_conf *rss_conf);
3598 * Retrieve current configuration of Receive Side Scaling hash computation
3599 * of Ethernet device.
3602 * The port identifier of the Ethernet device.
3604 * Where to store the current RSS hash configuration of the Ethernet device.
3606 * - (0) if successful.
3607 * - (-ENODEV) if port identifier is invalid.
3608 * - (-EIO) if device is removed.
3609 * - (-ENOTSUP) if hardware doesn't support RSS.
3612 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
3613 struct rte_eth_rss_conf *rss_conf);
3616 * Add UDP tunneling port for a specific type of tunnel.
3617 * The packets with this UDP port will be identified as this type of tunnel.
3618 * Before enabling any offloading function for a tunnel, users can call this API
3619 * to change or add more UDP port for the tunnel. So the offloading function
3620 * can take effect on the packets with the specific UDP port.
3623 * The port identifier of the Ethernet device.
3625 * UDP tunneling configuration.
3628 * - (0) if successful.
3629 * - (-ENODEV) if port identifier is invalid.
3630 * - (-EIO) if device is removed.
3631 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3634 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
3635 struct rte_eth_udp_tunnel *tunnel_udp);
3638 * Delete UDP tunneling port a specific type of tunnel.
3639 * The packets with this UDP port will not be identified as this type of tunnel
3641 * Before enabling any offloading function for a tunnel, users can call this API
3642 * to delete a UDP port for the tunnel. So the offloading function will not take
3643 * effect on the packets with the specific UDP port.
3646 * The port identifier of the Ethernet device.
3648 * UDP tunneling configuration.
3651 * - (0) if successful.
3652 * - (-ENODEV) if port identifier is invalid.
3653 * - (-EIO) if device is removed.
3654 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3657 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
3658 struct rte_eth_udp_tunnel *tunnel_udp);
3661 * Check whether the filter type is supported on an Ethernet device.
3662 * All the supported filter types are defined in 'rte_eth_ctrl.h'.
3665 * The port identifier of the Ethernet device.
3666 * @param filter_type
3669 * - (0) if successful.
3670 * - (-ENOTSUP) if hardware doesn't support this filter type.
3671 * - (-ENODEV) if *port_id* invalid.
3672 * - (-EIO) if device is removed.
3675 int rte_eth_dev_filter_supported(uint16_t port_id,
3676 enum rte_filter_type filter_type);
3679 * Take operations to assigned filter type on an Ethernet device.
3680 * All the supported operations and filter types are defined in 'rte_eth_ctrl.h'.
3683 * The port identifier of the Ethernet device.
3684 * @param filter_type
3687 * Type of operation.
3689 * A pointer to arguments defined specifically for the operation.
3691 * - (0) if successful.
3692 * - (-ENOTSUP) if hardware doesn't support.
3693 * - (-ENODEV) if *port_id* invalid.
3694 * - (-EIO) if device is removed.
3695 * - others depends on the specific operations implementation.
3698 int rte_eth_dev_filter_ctrl(uint16_t port_id, enum rte_filter_type filter_type,
3699 enum rte_filter_op filter_op, void *arg);
3702 * Get DCB information on an Ethernet device.
3705 * The port identifier of the Ethernet device.
3709 * - (0) if successful.
3710 * - (-ENODEV) if port identifier is invalid.
3711 * - (-EIO) if device is removed.
3712 * - (-ENOTSUP) if hardware doesn't support.
3714 int rte_eth_dev_get_dcb_info(uint16_t port_id,
3715 struct rte_eth_dcb_info *dcb_info);
3717 struct rte_eth_rxtx_callback;
3720 * Add a callback to be called on packet RX on a given port and queue.
3722 * This API configures a function to be called for each burst of
3723 * packets received on a given NIC port queue. The return value is a pointer
3724 * that can be used to later remove the callback using
3725 * rte_eth_remove_rx_callback().
3727 * Multiple functions are called in the order that they are added.
3730 * The port identifier of the Ethernet device.
3732 * The queue on the Ethernet device on which the callback is to be added.
3734 * The callback function
3736 * A generic pointer parameter which will be passed to each invocation of the
3737 * callback function on this port and queue.
3741 * On success, a pointer value which can later be used to remove the callback.
3743 const struct rte_eth_rxtx_callback *
3744 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
3745 rte_rx_callback_fn fn, void *user_param);
3748 * Add a callback that must be called first on packet RX on a given port
3751 * This API configures a first function to be called for each burst of
3752 * packets received on a given NIC port queue. The return value is a pointer
3753 * that can be used to later remove the callback using
3754 * rte_eth_remove_rx_callback().
3756 * Multiple functions are called in the order that they are added.
3759 * The port identifier of the Ethernet device.
3761 * The queue on the Ethernet device on which the callback is to be added.
3763 * The callback function
3765 * A generic pointer parameter which will be passed to each invocation of the
3766 * callback function on this port and queue.
3770 * On success, a pointer value which can later be used to remove the callback.
3772 const struct rte_eth_rxtx_callback *
3773 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
3774 rte_rx_callback_fn fn, void *user_param);
3777 * Add a callback to be called on packet TX on a given port and queue.
3779 * This API configures a function to be called for each burst of
3780 * packets sent on a given NIC port queue. The return value is a pointer
3781 * that can be used to later remove the callback using
3782 * rte_eth_remove_tx_callback().
3784 * Multiple functions are called in the order that they are added.
3787 * The port identifier of the Ethernet device.
3789 * The queue on the Ethernet device on which the callback is to be added.
3791 * The callback function
3793 * A generic pointer parameter which will be passed to each invocation of the
3794 * callback function on this port and queue.
3798 * On success, a pointer value which can later be used to remove the callback.
3800 const struct rte_eth_rxtx_callback *
3801 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
3802 rte_tx_callback_fn fn, void *user_param);
3805 * Remove an RX packet callback from a given port and queue.
3807 * This function is used to removed callbacks that were added to a NIC port
3808 * queue using rte_eth_add_rx_callback().
3810 * Note: the callback is removed from the callback list but it isn't freed
3811 * since the it may still be in use. The memory for the callback can be
3812 * subsequently freed back by the application by calling rte_free():
3814 * - Immediately - if the port is stopped, or the user knows that no
3815 * callbacks are in flight e.g. if called from the thread doing RX/TX
3818 * - After a short delay - where the delay is sufficient to allow any
3819 * in-flight callbacks to complete.
3822 * The port identifier of the Ethernet device.
3824 * The queue on the Ethernet device from which the callback is to be removed.
3826 * User supplied callback created via rte_eth_add_rx_callback().
3829 * - 0: Success. Callback was removed.
3830 * - -ENOTSUP: Callback support is not available.
3831 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3832 * is NULL or not found for the port/queue.
3834 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
3835 const struct rte_eth_rxtx_callback *user_cb);
3838 * Remove a TX packet callback from a given port and queue.
3840 * This function is used to removed callbacks that were added to a NIC port
3841 * queue using rte_eth_add_tx_callback().
3843 * Note: the callback is removed from the callback list but it isn't freed
3844 * since the it may still be in use. The memory for the callback can be
3845 * subsequently freed back by the application by calling rte_free():
3847 * - Immediately - if the port is stopped, or the user knows that no
3848 * callbacks are in flight e.g. if called from the thread doing RX/TX
3851 * - After a short delay - where the delay is sufficient to allow any
3852 * in-flight callbacks to complete.
3855 * The port identifier of the Ethernet device.
3857 * The queue on the Ethernet device from which the callback is to be removed.
3859 * User supplied callback created via rte_eth_add_tx_callback().
3862 * - 0: Success. Callback was removed.
3863 * - -ENOTSUP: Callback support is not available.
3864 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3865 * is NULL or not found for the port/queue.
3867 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
3868 const struct rte_eth_rxtx_callback *user_cb);
3871 * Retrieve information about given port's RX queue.
3874 * The port identifier of the Ethernet device.
3876 * The RX queue on the Ethernet device for which information
3877 * will be retrieved.
3879 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
3880 * the information of the Ethernet device.
3884 * - -ENOTSUP: routine is not supported by the device PMD.
3885 * - -EINVAL: The port_id or the queue_id is out of range, or the queue
3888 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3889 struct rte_eth_rxq_info *qinfo);
3892 * Retrieve information about given port's TX queue.
3895 * The port identifier of the Ethernet device.
3897 * The TX queue on the Ethernet device for which information
3898 * will be retrieved.
3900 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
3901 * the information of the Ethernet device.
3905 * - -ENOTSUP: routine is not supported by the device PMD.
3906 * - -EINVAL: The port_id or the queue_id is out of range, or the queue
3909 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3910 struct rte_eth_txq_info *qinfo);
3913 * Retrieve information about the Rx packet burst mode.
3916 * The port identifier of the Ethernet device.
3918 * The Rx queue on the Ethernet device for which information
3919 * will be retrieved.
3921 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
3922 * with the information of the packet burst mode.
3926 * - -ENOTSUP: routine is not supported by the device PMD.
3927 * - -EINVAL: The port_id or the queue_id is out of range.
3930 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
3931 struct rte_eth_burst_mode *mode);
3934 * Retrieve information about the Tx packet burst mode.
3937 * The port identifier of the Ethernet device.
3939 * The Tx queue on the Ethernet device for which information
3940 * will be retrieved.
3942 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
3943 * with the information of the packet burst mode.
3947 * - -ENOTSUP: routine is not supported by the device PMD.
3948 * - -EINVAL: The port_id or the queue_id is out of range.
3951 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
3952 struct rte_eth_burst_mode *mode);
3955 * Retrieve device registers and register attributes (number of registers and
3959 * The port identifier of the Ethernet device.
3961 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
3962 * NULL the function fills in the width and length fields. If non-NULL
3963 * the registers are put into the buffer pointed at by the data field.
3965 * - (0) if successful.
3966 * - (-ENOTSUP) if hardware doesn't support.
3967 * - (-ENODEV) if *port_id* invalid.
3968 * - (-EIO) if device is removed.
3969 * - others depends on the specific operations implementation.
3971 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
3974 * Retrieve size of device EEPROM
3977 * The port identifier of the Ethernet device.
3979 * - (>=0) EEPROM size if successful.
3980 * - (-ENOTSUP) if hardware doesn't support.
3981 * - (-ENODEV) if *port_id* invalid.
3982 * - (-EIO) if device is removed.
3983 * - others depends on the specific operations implementation.
3985 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
3988 * Retrieve EEPROM and EEPROM attribute
3991 * The port identifier of the Ethernet device.
3993 * The template includes buffer for return EEPROM data and
3994 * EEPROM attributes to be filled.
3996 * - (0) if successful.
3997 * - (-ENOTSUP) if hardware doesn't support.
3998 * - (-ENODEV) if *port_id* invalid.
3999 * - (-EIO) if device is removed.
4000 * - others depends on the specific operations implementation.
4002 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4005 * Program EEPROM with provided data
4008 * The port identifier of the Ethernet device.
4010 * The template includes EEPROM data for programming and
4011 * EEPROM attributes to be filled
4013 * - (0) if successful.
4014 * - (-ENOTSUP) if hardware doesn't support.
4015 * - (-ENODEV) if *port_id* invalid.
4016 * - (-EIO) if device is removed.
4017 * - others depends on the specific operations implementation.
4019 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4023 * @b EXPERIMENTAL: this API may change without prior notice.
4025 * Retrieve the type and size of plugin module EEPROM
4028 * The port identifier of the Ethernet device.
4030 * The type and size of plugin module EEPROM.
4032 * - (0) if successful.
4033 * - (-ENOTSUP) if hardware doesn't support.
4034 * - (-ENODEV) if *port_id* invalid.
4035 * - (-EIO) if device is removed.
4036 * - others depends on the specific operations implementation.
4040 rte_eth_dev_get_module_info(uint16_t port_id,
4041 struct rte_eth_dev_module_info *modinfo);
4045 * @b EXPERIMENTAL: this API may change without prior notice.
4047 * Retrieve the data of plugin module EEPROM
4050 * The port identifier of the Ethernet device.
4052 * The template includes the plugin module EEPROM attributes, and the
4053 * buffer for return plugin module EEPROM data.
4055 * - (0) if successful.
4056 * - (-ENOTSUP) if hardware doesn't support.
4057 * - (-ENODEV) if *port_id* invalid.
4058 * - (-EIO) if device is removed.
4059 * - others depends on the specific operations implementation.
4063 rte_eth_dev_get_module_eeprom(uint16_t port_id,
4064 struct rte_dev_eeprom_info *info);
4067 * Set the list of multicast addresses to filter on an Ethernet device.
4070 * The port identifier of the Ethernet device.
4071 * @param mc_addr_set
4072 * The array of multicast addresses to set. Equal to NULL when the function
4073 * is invoked to flush the set of filtered addresses.
4075 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
4076 * when the function is invoked to flush the set of filtered addresses.
4078 * - (0) if successful.
4079 * - (-ENODEV) if *port_id* invalid.
4080 * - (-EIO) if device is removed.
4081 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
4082 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
4084 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
4085 struct rte_ether_addr *mc_addr_set,
4086 uint32_t nb_mc_addr);
4089 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
4092 * The port identifier of the Ethernet device.
4096 * - -ENODEV: The port ID is invalid.
4097 * - -EIO: if device is removed.
4098 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4100 int rte_eth_timesync_enable(uint16_t port_id);
4103 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
4106 * The port identifier of the Ethernet device.
4110 * - -ENODEV: The port ID is invalid.
4111 * - -EIO: if device is removed.
4112 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4114 int rte_eth_timesync_disable(uint16_t port_id);
4117 * Read an IEEE1588/802.1AS RX timestamp from an Ethernet device.
4120 * The port identifier of the Ethernet device.
4122 * Pointer to the timestamp struct.
4124 * Device specific flags. Used to pass the RX timesync register index to
4125 * i40e. Unused in igb/ixgbe, pass 0 instead.
4129 * - -EINVAL: No timestamp is available.
4130 * - -ENODEV: The port ID is invalid.
4131 * - -EIO: if device is removed.
4132 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4134 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
4135 struct timespec *timestamp, uint32_t flags);
4138 * Read an IEEE1588/802.1AS TX timestamp from an Ethernet device.
4141 * The port identifier of the Ethernet device.
4143 * Pointer to the timestamp struct.
4147 * - -EINVAL: No timestamp is available.
4148 * - -ENODEV: The port ID is invalid.
4149 * - -EIO: if device is removed.
4150 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4152 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
4153 struct timespec *timestamp);
4156 * Adjust the timesync clock on an Ethernet device.
4158 * This is usually used in conjunction with other Ethdev timesync functions to
4159 * synchronize the device time using the IEEE1588/802.1AS protocol.
4162 * The port identifier of the Ethernet device.
4164 * The adjustment in nanoseconds.
4168 * - -ENODEV: The port ID is invalid.
4169 * - -EIO: if device is removed.
4170 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4172 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
4175 * Read the time from the timesync clock on an Ethernet device.
4177 * This is usually used in conjunction with other Ethdev timesync functions to
4178 * synchronize the device time using the IEEE1588/802.1AS protocol.
4181 * The port identifier of the Ethernet device.
4183 * Pointer to the timespec struct that holds the time.
4188 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
4191 * Set the time of the timesync clock on an Ethernet device.
4193 * This is usually used in conjunction with other Ethdev timesync functions to
4194 * synchronize the device time using the IEEE1588/802.1AS protocol.
4197 * The port identifier of the Ethernet device.
4199 * Pointer to the timespec struct that holds the time.
4203 * - -EINVAL: No timestamp is available.
4204 * - -ENODEV: The port ID is invalid.
4205 * - -EIO: if device is removed.
4206 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4208 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
4212 * @b EXPERIMENTAL: this API may change without prior notice.
4214 * Read the current clock counter of an Ethernet device
4216 * This returns the current raw clock value of an Ethernet device. It is
4217 * a raw amount of ticks, with no given time reference.
4218 * The value returned here is from the same clock than the one
4219 * filling timestamp field of Rx packets when using hardware timestamp
4220 * offload. Therefore it can be used to compute a precise conversion of
4221 * the device clock to the real time.
4223 * E.g, a simple heuristic to derivate the frequency would be:
4224 * uint64_t start, end;
4225 * rte_eth_read_clock(port, start);
4226 * rte_delay_ms(100);
4227 * rte_eth_read_clock(port, end);
4228 * double freq = (end - start) * 10;
4230 * Compute a common reference with:
4231 * uint64_t base_time_sec = current_time();
4232 * uint64_t base_clock;
4233 * rte_eth_read_clock(port, base_clock);
4235 * Then, convert the raw mbuf timestamp with:
4236 * base_time_sec + (double)(mbuf->timestamp - base_clock) / freq;
4238 * This simple example will not provide a very good accuracy. One must
4239 * at least measure multiple times the frequency and do a regression.
4240 * To avoid deviation from the system time, the common reference can
4241 * be repeated from time to time. The integer division can also be
4242 * converted by a multiplication and a shift for better performance.
4245 * The port identifier of the Ethernet device.
4247 * Pointer to the uint64_t that holds the raw clock value.
4251 * - -ENODEV: The port ID is invalid.
4252 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4256 rte_eth_read_clock(uint16_t port_id, uint64_t *clock);
4259 * Config l2 tunnel ether type of an Ethernet device for filtering specific
4260 * tunnel packets by ether type.
4263 * The port identifier of the Ethernet device.
4265 * l2 tunnel configuration.
4268 * - (0) if successful.
4269 * - (-ENODEV) if port identifier is invalid.
4270 * - (-EIO) if device is removed.
4271 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4274 rte_eth_dev_l2_tunnel_eth_type_conf(uint16_t port_id,
4275 struct rte_eth_l2_tunnel_conf *l2_tunnel);
4278 * Enable/disable l2 tunnel offload functions. Include,
4279 * 1, The ability of parsing a type of l2 tunnel of an Ethernet device.
4280 * Filtering, forwarding and offloading this type of tunnel packets depend on
4282 * 2, Stripping the l2 tunnel tag.
4283 * 3, Insertion of the l2 tunnel tag.
4284 * 4, Forwarding the packets based on the l2 tunnel tag.
4287 * The port identifier of the Ethernet device.
4289 * l2 tunnel parameters.
4291 * Indicate the offload function.
4293 * Enable or disable this function.
4296 * - (0) if successful.
4297 * - (-ENODEV) if port identifier is invalid.
4298 * - (-EIO) if device is removed.
4299 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4302 rte_eth_dev_l2_tunnel_offload_set(uint16_t port_id,
4303 struct rte_eth_l2_tunnel_conf *l2_tunnel,
4308 * Get the port id from device name. The device name should be specified
4310 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0
4311 * - SoC device name, for example- fsl-gmac0
4312 * - vdev dpdk name, for example- net_[pcap0|null0|tap0]
4315 * pci address or name of the device
4317 * pointer to port identifier of the device
4319 * - (0) if successful and port_id is filled.
4320 * - (-ENODEV or -EINVAL) on failure.
4323 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
4326 * Get the device name from port id. The device name is specified as below:
4327 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0
4328 * - SoC device name, for example- fsl-gmac0
4329 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0]
4332 * Port identifier of the device.
4334 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
4336 * - (0) if successful.
4337 * - (-EINVAL) on failure.
4340 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
4343 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
4344 * the ethernet device information, otherwise adjust them to boundaries.
4347 * The port identifier of the Ethernet device.
4349 * A pointer to a uint16_t where the number of receive
4350 * descriptors stored.
4352 * A pointer to a uint16_t where the number of transmit
4353 * descriptors stored.
4355 * - (0) if successful.
4356 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
4358 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
4359 uint16_t *nb_rx_desc,
4360 uint16_t *nb_tx_desc);
4363 * Test if a port supports specific mempool ops.
4366 * Port identifier of the Ethernet device.
4368 * The name of the pool operations to test.
4370 * - 0: best mempool ops choice for this port.
4371 * - 1: mempool ops are supported for this port.
4372 * - -ENOTSUP: mempool ops not supported for this port.
4373 * - -ENODEV: Invalid port Identifier.
4374 * - -EINVAL: Pool param is null.
4377 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
4380 * Get the security context for the Ethernet device.
4383 * Port identifier of the Ethernet device
4386 * - pointer to security context on success.
4389 rte_eth_dev_get_sec_ctx(uint16_t port_id);
4393 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4395 * Query the device hairpin capabilities.
4398 * The port identifier of the Ethernet device.
4400 * Pointer to a structure that will hold the hairpin capabilities.
4402 * - (0) if successful.
4403 * - (-ENOTSUP) if hardware doesn't support.
4406 int rte_eth_dev_hairpin_capability_get(uint16_t port_id,
4407 struct rte_eth_hairpin_cap *cap);
4409 #include <rte_ethdev_core.h>
4413 * Retrieve a burst of input packets from a receive queue of an Ethernet
4414 * device. The retrieved packets are stored in *rte_mbuf* structures whose
4415 * pointers are supplied in the *rx_pkts* array.
4417 * The rte_eth_rx_burst() function loops, parsing the RX ring of the
4418 * receive queue, up to *nb_pkts* packets, and for each completed RX
4419 * descriptor in the ring, it performs the following operations:
4421 * - Initialize the *rte_mbuf* data structure associated with the
4422 * RX descriptor according to the information provided by the NIC into
4423 * that RX descriptor.
4425 * - Store the *rte_mbuf* data structure into the next entry of the
4428 * - Replenish the RX descriptor with a new *rte_mbuf* buffer
4429 * allocated from the memory pool associated with the receive queue at
4430 * initialization time.
4432 * When retrieving an input packet that was scattered by the controller
4433 * into multiple receive descriptors, the rte_eth_rx_burst() function
4434 * appends the associated *rte_mbuf* buffers to the first buffer of the
4437 * The rte_eth_rx_burst() function returns the number of packets
4438 * actually retrieved, which is the number of *rte_mbuf* data structures
4439 * effectively supplied into the *rx_pkts* array.
4440 * A return value equal to *nb_pkts* indicates that the RX queue contained
4441 * at least *rx_pkts* packets, and this is likely to signify that other
4442 * received packets remain in the input queue. Applications implementing
4443 * a "retrieve as much received packets as possible" policy can check this
4444 * specific case and keep invoking the rte_eth_rx_burst() function until
4445 * a value less than *nb_pkts* is returned.
4447 * This receive method has the following advantages:
4449 * - It allows a run-to-completion network stack engine to retrieve and
4450 * to immediately process received packets in a fast burst-oriented
4451 * approach, avoiding the overhead of unnecessary intermediate packet
4452 * queue/dequeue operations.
4454 * - Conversely, it also allows an asynchronous-oriented processing
4455 * method to retrieve bursts of received packets and to immediately
4456 * queue them for further parallel processing by another logical core,
4457 * for instance. However, instead of having received packets being
4458 * individually queued by the driver, this approach allows the caller
4459 * of the rte_eth_rx_burst() function to queue a burst of retrieved
4460 * packets at a time and therefore dramatically reduce the cost of
4461 * enqueue/dequeue operations per packet.
4463 * - It allows the rte_eth_rx_burst() function of the driver to take
4464 * advantage of burst-oriented hardware features (CPU cache,
4465 * prefetch instructions, and so on) to minimize the number of CPU
4466 * cycles per packet.
4468 * To summarize, the proposed receive API enables many
4469 * burst-oriented optimizations in both synchronous and asynchronous
4470 * packet processing environments with no overhead in both cases.
4472 * The rte_eth_rx_burst() function does not provide any error
4473 * notification to avoid the corresponding overhead. As a hint, the
4474 * upper-level application might check the status of the device link once
4475 * being systematically returned a 0 value for a given number of tries.
4478 * The port identifier of the Ethernet device.
4480 * The index of the receive queue from which to retrieve input packets.
4481 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4482 * to rte_eth_dev_configure().
4484 * The address of an array of pointers to *rte_mbuf* structures that
4485 * must be large enough to store *nb_pkts* pointers in it.
4487 * The maximum number of packets to retrieve.
4489 * The number of packets actually retrieved, which is the number
4490 * of pointers to *rte_mbuf* structures effectively supplied to the
4493 static inline uint16_t
4494 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
4495 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
4497 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4500 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4501 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4502 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_pkt_burst, 0);
4504 if (queue_id >= dev->data->nb_rx_queues) {
4505 RTE_ETHDEV_LOG(ERR, "Invalid RX queue_id=%u\n", queue_id);
4509 nb_rx = (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id],
4512 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4513 if (unlikely(dev->post_rx_burst_cbs[queue_id] != NULL)) {
4514 struct rte_eth_rxtx_callback *cb =
4515 dev->post_rx_burst_cbs[queue_id];
4518 nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx,
4519 nb_pkts, cb->param);
4521 } while (cb != NULL);
4525 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx);
4530 * Get the number of used descriptors of a rx queue
4533 * The port identifier of the Ethernet device.
4535 * The queue id on the specific port.
4537 * The number of used descriptors in the specific queue, or:
4538 * (-EINVAL) if *port_id* or *queue_id* is invalid
4539 * (-ENOTSUP) if the device does not support this function
4542 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
4544 struct rte_eth_dev *dev;
4546 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
4547 dev = &rte_eth_devices[port_id];
4548 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_queue_count, -ENOTSUP);
4549 if (queue_id >= dev->data->nb_rx_queues)
4552 return (int)(*dev->dev_ops->rx_queue_count)(dev, queue_id);
4556 * Check if the DD bit of the specific RX descriptor in the queue has been set
4559 * The port identifier of the Ethernet device.
4561 * The queue id on the specific port.
4563 * The offset of the descriptor ID from tail.
4565 * - (1) if the specific DD bit is set.
4566 * - (0) if the specific DD bit is not set.
4567 * - (-ENODEV) if *port_id* invalid.
4568 * - (-ENOTSUP) if the device does not support this function
4571 rte_eth_rx_descriptor_done(uint16_t port_id, uint16_t queue_id, uint16_t offset)
4573 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4574 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4575 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_done, -ENOTSUP);
4576 return (*dev->dev_ops->rx_descriptor_done)( \
4577 dev->data->rx_queues[queue_id], offset);
4580 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
4581 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
4582 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
4585 * Check the status of a Rx descriptor in the queue
4587 * It should be called in a similar context than the Rx function:
4588 * - on a dataplane core
4589 * - not concurrently on the same queue
4591 * Since it's a dataplane function, no check is performed on port_id and
4592 * queue_id. The caller must therefore ensure that the port is enabled
4593 * and the queue is configured and running.
4595 * Note: accessing to a random descriptor in the ring may trigger cache
4596 * misses and have a performance impact.
4599 * A valid port identifier of the Ethernet device which.
4601 * A valid Rx queue identifier on this port.
4603 * The offset of the descriptor starting from tail (0 is the next
4604 * packet to be received by the driver).
4607 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
4609 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
4610 * not yet processed by the driver (i.e. in the receive queue).
4611 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
4612 * the driver and not yet returned to hw, or reserved by the hw.
4613 * - (-EINVAL) bad descriptor offset.
4614 * - (-ENOTSUP) if the device does not support this function.
4615 * - (-ENODEV) bad port or queue (only if compiled with debug).
4618 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
4621 struct rte_eth_dev *dev;
4624 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4625 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4627 dev = &rte_eth_devices[port_id];
4628 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4629 if (queue_id >= dev->data->nb_rx_queues)
4632 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_status, -ENOTSUP);
4633 rxq = dev->data->rx_queues[queue_id];
4635 return (*dev->dev_ops->rx_descriptor_status)(rxq, offset);
4638 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
4639 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
4640 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
4643 * Check the status of a Tx descriptor in the queue.
4645 * It should be called in a similar context than the Tx function:
4646 * - on a dataplane core
4647 * - not concurrently on the same queue
4649 * Since it's a dataplane function, no check is performed on port_id and
4650 * queue_id. The caller must therefore ensure that the port is enabled
4651 * and the queue is configured and running.
4653 * Note: accessing to a random descriptor in the ring may trigger cache
4654 * misses and have a performance impact.
4657 * A valid port identifier of the Ethernet device which.
4659 * A valid Tx queue identifier on this port.
4661 * The offset of the descriptor starting from tail (0 is the place where
4662 * the next packet will be send).
4665 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
4666 * in the transmit queue.
4667 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
4668 * be reused by the driver.
4669 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
4670 * driver or the hardware.
4671 * - (-EINVAL) bad descriptor offset.
4672 * - (-ENOTSUP) if the device does not support this function.
4673 * - (-ENODEV) bad port or queue (only if compiled with debug).
4675 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
4676 uint16_t queue_id, uint16_t offset)
4678 struct rte_eth_dev *dev;
4681 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4682 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4684 dev = &rte_eth_devices[port_id];
4685 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4686 if (queue_id >= dev->data->nb_tx_queues)
4689 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->tx_descriptor_status, -ENOTSUP);
4690 txq = dev->data->tx_queues[queue_id];
4692 return (*dev->dev_ops->tx_descriptor_status)(txq, offset);
4696 * Send a burst of output packets on a transmit queue of an Ethernet device.
4698 * The rte_eth_tx_burst() function is invoked to transmit output packets
4699 * on the output queue *queue_id* of the Ethernet device designated by its
4701 * The *nb_pkts* parameter is the number of packets to send which are
4702 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4703 * allocated from a pool created with rte_pktmbuf_pool_create().
4704 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
4705 * up to the number of transmit descriptors available in the TX ring of the
4707 * For each packet to send, the rte_eth_tx_burst() function performs
4708 * the following operations:
4710 * - Pick up the next available descriptor in the transmit ring.
4712 * - Free the network buffer previously sent with that descriptor, if any.
4714 * - Initialize the transmit descriptor with the information provided
4715 * in the *rte_mbuf data structure.
4717 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
4718 * the rte_eth_tx_burst() function uses several transmit descriptors
4721 * The rte_eth_tx_burst() function returns the number of packets it
4722 * actually sent. A return value equal to *nb_pkts* means that all packets
4723 * have been sent, and this is likely to signify that other output packets
4724 * could be immediately transmitted again. Applications that implement a
4725 * "send as many packets to transmit as possible" policy can check this
4726 * specific case and keep invoking the rte_eth_tx_burst() function until
4727 * a value less than *nb_pkts* is returned.
4729 * It is the responsibility of the rte_eth_tx_burst() function to
4730 * transparently free the memory buffers of packets previously sent.
4731 * This feature is driven by the *tx_free_thresh* value supplied to the
4732 * rte_eth_dev_configure() function at device configuration time.
4733 * When the number of free TX descriptors drops below this threshold, the
4734 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
4735 * of those packets whose transmission was effectively completed.
4737 * If the PMD is DEV_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
4738 * invoke this function concurrently on the same tx queue without SW lock.
4739 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads
4741 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
4745 * The port identifier of the Ethernet device.
4747 * The index of the transmit queue through which output packets must be
4749 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4750 * to rte_eth_dev_configure().
4752 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4753 * which contain the output packets.
4755 * The maximum number of packets to transmit.
4757 * The number of output packets actually stored in transmit descriptors of
4758 * the transmit ring. The return value can be less than the value of the
4759 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
4761 static inline uint16_t
4762 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
4763 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4765 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4767 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4768 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4769 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_pkt_burst, 0);
4771 if (queue_id >= dev->data->nb_tx_queues) {
4772 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4777 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4778 struct rte_eth_rxtx_callback *cb = dev->pre_tx_burst_cbs[queue_id];
4780 if (unlikely(cb != NULL)) {
4782 nb_pkts = cb->fn.tx(port_id, queue_id, tx_pkts, nb_pkts,
4785 } while (cb != NULL);
4789 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts,
4791 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts);
4795 * Process a burst of output packets on a transmit queue of an Ethernet device.
4797 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
4798 * transmitted on the output queue *queue_id* of the Ethernet device designated
4800 * The *nb_pkts* parameter is the number of packets to be prepared which are
4801 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4802 * allocated from a pool created with rte_pktmbuf_pool_create().
4803 * For each packet to send, the rte_eth_tx_prepare() function performs
4804 * the following operations:
4806 * - Check if packet meets devices requirements for tx offloads.
4808 * - Check limitations about number of segments.
4810 * - Check additional requirements when debug is enabled.
4812 * - Update and/or reset required checksums when tx offload is set for packet.
4814 * Since this function can modify packet data, provided mbufs must be safely
4815 * writable (e.g. modified data cannot be in shared segment).
4817 * The rte_eth_tx_prepare() function returns the number of packets ready to be
4818 * sent. A return value equal to *nb_pkts* means that all packets are valid and
4819 * ready to be sent, otherwise stops processing on the first invalid packet and
4820 * leaves the rest packets untouched.
4822 * When this functionality is not implemented in the driver, all packets are
4823 * are returned untouched.
4826 * The port identifier of the Ethernet device.
4827 * The value must be a valid port id.
4829 * The index of the transmit queue through which output packets must be
4831 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4832 * to rte_eth_dev_configure().
4834 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4835 * which contain the output packets.
4837 * The maximum number of packets to process.
4839 * The number of packets correct and ready to be sent. The return value can be
4840 * less than the value of the *tx_pkts* parameter when some packet doesn't
4841 * meet devices requirements with rte_errno set appropriately:
4842 * - EINVAL: offload flags are not correctly set
4843 * - ENOTSUP: the offload feature is not supported by the hardware
4847 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
4849 static inline uint16_t
4850 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
4851 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4853 struct rte_eth_dev *dev;
4855 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4856 if (!rte_eth_dev_is_valid_port(port_id)) {
4857 RTE_ETHDEV_LOG(ERR, "Invalid TX port_id=%u\n", port_id);
4863 dev = &rte_eth_devices[port_id];
4865 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4866 if (queue_id >= dev->data->nb_tx_queues) {
4867 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4873 if (!dev->tx_pkt_prepare)
4876 return (*dev->tx_pkt_prepare)(dev->data->tx_queues[queue_id],
4883 * Native NOOP operation for compilation targets which doesn't require any
4884 * preparations steps, and functional NOOP may introduce unnecessary performance
4887 * Generally this is not a good idea to turn it on globally and didn't should
4888 * be used if behavior of tx_preparation can change.
4891 static inline uint16_t
4892 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
4893 __rte_unused uint16_t queue_id,
4894 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4902 * Send any packets queued up for transmission on a port and HW queue
4904 * This causes an explicit flush of packets previously buffered via the
4905 * rte_eth_tx_buffer() function. It returns the number of packets successfully
4906 * sent to the NIC, and calls the error callback for any unsent packets. Unless
4907 * explicitly set up otherwise, the default callback simply frees the unsent
4908 * packets back to the owning mempool.
4911 * The port identifier of the Ethernet device.
4913 * The index of the transmit queue through which output packets must be
4915 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4916 * to rte_eth_dev_configure().
4918 * Buffer of packets to be transmit.
4920 * The number of packets successfully sent to the Ethernet device. The error
4921 * callback is called for any packets which could not be sent.
4923 static inline uint16_t
4924 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
4925 struct rte_eth_dev_tx_buffer *buffer)
4928 uint16_t to_send = buffer->length;
4933 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
4937 /* All packets sent, or to be dealt with by callback below */
4938 if (unlikely(sent != to_send))
4939 buffer->error_callback(&buffer->pkts[sent],
4940 (uint16_t)(to_send - sent),
4941 buffer->error_userdata);
4947 * Buffer a single packet for future transmission on a port and queue
4949 * This function takes a single mbuf/packet and buffers it for later
4950 * transmission on the particular port and queue specified. Once the buffer is
4951 * full of packets, an attempt will be made to transmit all the buffered
4952 * packets. In case of error, where not all packets can be transmitted, a
4953 * callback is called with the unsent packets as a parameter. If no callback
4954 * is explicitly set up, the unsent packets are just freed back to the owning
4955 * mempool. The function returns the number of packets actually sent i.e.
4956 * 0 if no buffer flush occurred, otherwise the number of packets successfully
4960 * The port identifier of the Ethernet device.
4962 * The index of the transmit queue through which output packets must be
4964 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4965 * to rte_eth_dev_configure().
4967 * Buffer used to collect packets to be sent.
4969 * Pointer to the packet mbuf to be sent.
4971 * 0 = packet has been buffered for later transmission
4972 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
4973 * causing N packets to be sent, and the error callback to be called for
4976 static __rte_always_inline uint16_t
4977 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
4978 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
4980 buffer->pkts[buffer->length++] = tx_pkt;
4981 if (buffer->length < buffer->size)
4984 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
4991 #endif /* _RTE_ETHDEV_H_ */