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). */
326 #define RTE_ETH_LINK_MAX_STR_LEN 40 /**< Max length of default link string. */
329 * A structure used to configure the ring threshold registers of an RX/TX
330 * queue for an Ethernet port.
332 struct rte_eth_thresh {
333 uint8_t pthresh; /**< Ring prefetch threshold. */
334 uint8_t hthresh; /**< Ring host threshold. */
335 uint8_t wthresh; /**< Ring writeback threshold. */
339 * Simple flags are used for rte_eth_conf.rxmode.mq_mode.
341 #define ETH_MQ_RX_RSS_FLAG 0x1
342 #define ETH_MQ_RX_DCB_FLAG 0x2
343 #define ETH_MQ_RX_VMDQ_FLAG 0x4
346 * A set of values to identify what method is to be used to route
347 * packets to multiple queues.
349 enum rte_eth_rx_mq_mode {
350 /** None of DCB,RSS or VMDQ mode */
353 /** For RX side, only RSS is on */
354 ETH_MQ_RX_RSS = ETH_MQ_RX_RSS_FLAG,
355 /** For RX side,only DCB is on. */
356 ETH_MQ_RX_DCB = ETH_MQ_RX_DCB_FLAG,
357 /** Both DCB and RSS enable */
358 ETH_MQ_RX_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG,
360 /** Only VMDQ, no RSS nor DCB */
361 ETH_MQ_RX_VMDQ_ONLY = ETH_MQ_RX_VMDQ_FLAG,
362 /** RSS mode with VMDQ */
363 ETH_MQ_RX_VMDQ_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_VMDQ_FLAG,
364 /** Use VMDQ+DCB to route traffic to queues */
365 ETH_MQ_RX_VMDQ_DCB = ETH_MQ_RX_VMDQ_FLAG | ETH_MQ_RX_DCB_FLAG,
366 /** Enable both VMDQ and DCB in VMDq */
367 ETH_MQ_RX_VMDQ_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG |
372 * for rx mq mode backward compatible
374 #define ETH_RSS ETH_MQ_RX_RSS
375 #define VMDQ_DCB ETH_MQ_RX_VMDQ_DCB
376 #define ETH_DCB_RX ETH_MQ_RX_DCB
379 * A set of values to identify what method is to be used to transmit
380 * packets using multi-TCs.
382 enum rte_eth_tx_mq_mode {
383 ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
384 ETH_MQ_TX_DCB, /**< For TX side,only DCB is on. */
385 ETH_MQ_TX_VMDQ_DCB, /**< For TX side,both DCB and VT is on. */
386 ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
390 * for tx mq mode backward compatible
392 #define ETH_DCB_NONE ETH_MQ_TX_NONE
393 #define ETH_VMDQ_DCB_TX ETH_MQ_TX_VMDQ_DCB
394 #define ETH_DCB_TX ETH_MQ_TX_DCB
397 * A structure used to configure the RX features of an Ethernet port.
399 struct rte_eth_rxmode {
400 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
401 enum rte_eth_rx_mq_mode mq_mode;
402 uint32_t max_rx_pkt_len; /**< Only used if JUMBO_FRAME enabled. */
403 /** Maximum allowed size of LRO aggregated packet. */
404 uint32_t max_lro_pkt_size;
405 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
407 * Per-port Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
408 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
409 * structure are allowed to be set.
413 uint64_t reserved_64s[2]; /**< Reserved for future fields */
414 void *reserved_ptrs[2]; /**< Reserved for future fields */
418 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
419 * Note that single VLAN is treated the same as inner VLAN.
422 ETH_VLAN_TYPE_UNKNOWN = 0,
423 ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
424 ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
429 * A structure used to describe a vlan filter.
430 * If the bit corresponding to a VID is set, such VID is on.
432 struct rte_vlan_filter_conf {
437 * A structure used to configure the Receive Side Scaling (RSS) feature
438 * of an Ethernet port.
439 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
440 * to an array holding the RSS key to use for hashing specific header
441 * fields of received packets. The length of this array should be indicated
442 * by *rss_key_len* below. Otherwise, a default random hash key is used by
445 * The *rss_key_len* field of the *rss_conf* structure indicates the length
446 * in bytes of the array pointed by *rss_key*. To be compatible, this length
447 * will be checked in i40e only. Others assume 40 bytes to be used as before.
449 * The *rss_hf* field of the *rss_conf* structure indicates the different
450 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
451 * Supplying an *rss_hf* equal to zero disables the RSS feature.
453 struct rte_eth_rss_conf {
454 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
455 uint8_t rss_key_len; /**< hash key length in bytes. */
456 uint64_t rss_hf; /**< Hash functions to apply - see below. */
460 * A packet can be identified by hardware as different flow types. Different
461 * NIC hardware may support different flow types.
462 * Basically, the NIC hardware identifies the flow type as deep protocol as
463 * possible, and exclusively. For example, if a packet is identified as
464 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types,
465 * though it is an actual IPV4 packet.
467 #define RTE_ETH_FLOW_UNKNOWN 0
468 #define RTE_ETH_FLOW_RAW 1
469 #define RTE_ETH_FLOW_IPV4 2
470 #define RTE_ETH_FLOW_FRAG_IPV4 3
471 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4
472 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5
473 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6
474 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7
475 #define RTE_ETH_FLOW_IPV6 8
476 #define RTE_ETH_FLOW_FRAG_IPV6 9
477 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10
478 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11
479 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12
480 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13
481 #define RTE_ETH_FLOW_L2_PAYLOAD 14
482 #define RTE_ETH_FLOW_IPV6_EX 15
483 #define RTE_ETH_FLOW_IPV6_TCP_EX 16
484 #define RTE_ETH_FLOW_IPV6_UDP_EX 17
485 #define RTE_ETH_FLOW_PORT 18
486 /**< Consider device port number as a flow differentiator */
487 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */
488 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */
489 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */
490 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */
491 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */
492 #define RTE_ETH_FLOW_MAX 24
495 * Below macros are defined for RSS offload types, they can be used to
496 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types.
498 #define ETH_RSS_IPV4 (1ULL << 2)
499 #define ETH_RSS_FRAG_IPV4 (1ULL << 3)
500 #define ETH_RSS_NONFRAG_IPV4_TCP (1ULL << 4)
501 #define ETH_RSS_NONFRAG_IPV4_UDP (1ULL << 5)
502 #define ETH_RSS_NONFRAG_IPV4_SCTP (1ULL << 6)
503 #define ETH_RSS_NONFRAG_IPV4_OTHER (1ULL << 7)
504 #define ETH_RSS_IPV6 (1ULL << 8)
505 #define ETH_RSS_FRAG_IPV6 (1ULL << 9)
506 #define ETH_RSS_NONFRAG_IPV6_TCP (1ULL << 10)
507 #define ETH_RSS_NONFRAG_IPV6_UDP (1ULL << 11)
508 #define ETH_RSS_NONFRAG_IPV6_SCTP (1ULL << 12)
509 #define ETH_RSS_NONFRAG_IPV6_OTHER (1ULL << 13)
510 #define ETH_RSS_L2_PAYLOAD (1ULL << 14)
511 #define ETH_RSS_IPV6_EX (1ULL << 15)
512 #define ETH_RSS_IPV6_TCP_EX (1ULL << 16)
513 #define ETH_RSS_IPV6_UDP_EX (1ULL << 17)
514 #define ETH_RSS_PORT (1ULL << 18)
515 #define ETH_RSS_VXLAN (1ULL << 19)
516 #define ETH_RSS_GENEVE (1ULL << 20)
517 #define ETH_RSS_NVGRE (1ULL << 21)
518 #define ETH_RSS_GTPU (1ULL << 23)
519 #define ETH_RSS_ETH (1ULL << 24)
520 #define ETH_RSS_S_VLAN (1ULL << 25)
521 #define ETH_RSS_C_VLAN (1ULL << 26)
522 #define ETH_RSS_ESP (1ULL << 27)
523 #define ETH_RSS_AH (1ULL << 28)
524 #define ETH_RSS_L2TPV3 (1ULL << 29)
525 #define ETH_RSS_PFCP (1ULL << 30)
526 #define ETH_RSS_PPPOE (1ULL << 31)
529 * We use the following macros to combine with above ETH_RSS_* for
530 * more specific input set selection. These bits are defined starting
531 * from the high end of the 64 bits.
532 * Note: If we use above ETH_RSS_* without SRC/DST_ONLY, it represents
533 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of
534 * the same level are used simultaneously, it is the same case as none of
537 #define ETH_RSS_L3_SRC_ONLY (1ULL << 63)
538 #define ETH_RSS_L3_DST_ONLY (1ULL << 62)
539 #define ETH_RSS_L4_SRC_ONLY (1ULL << 61)
540 #define ETH_RSS_L4_DST_ONLY (1ULL << 60)
541 #define ETH_RSS_L2_SRC_ONLY (1ULL << 59)
542 #define ETH_RSS_L2_DST_ONLY (1ULL << 58)
545 * Only select IPV6 address prefix as RSS input set according to
546 * https://tools.ietf.org/html/rfc6052
547 * Must be combined with ETH_RSS_IPV6, ETH_RSS_NONFRAG_IPV6_UDP,
548 * ETH_RSS_NONFRAG_IPV6_TCP, ETH_RSS_NONFRAG_IPV6_SCTP.
550 #define RTE_ETH_RSS_L3_PRE32 (1ULL << 57)
551 #define RTE_ETH_RSS_L3_PRE40 (1ULL << 56)
552 #define RTE_ETH_RSS_L3_PRE48 (1ULL << 55)
553 #define RTE_ETH_RSS_L3_PRE56 (1ULL << 54)
554 #define RTE_ETH_RSS_L3_PRE64 (1ULL << 53)
555 #define RTE_ETH_RSS_L3_PRE96 (1ULL << 52)
558 * Use the following macros to combine with the above layers
559 * to choose inner and outer layers or both for RSS computation.
560 * Bits 50 and 51 are reserved for this.
564 * level 0, requests the default behavior.
565 * Depending on the packet type, it can mean outermost, innermost,
566 * anything in between or even no RSS.
567 * It basically stands for the innermost encapsulation level RSS
568 * can be performed on according to PMD and device capabilities.
570 #define ETH_RSS_LEVEL_PMD_DEFAULT (0ULL << 50)
573 * level 1, requests RSS to be performed on the outermost packet
574 * encapsulation level.
576 #define ETH_RSS_LEVEL_OUTERMOST (1ULL << 50)
579 * level 2, requests RSS to be performed on the specified inner packet
580 * encapsulation level, from outermost to innermost (lower to higher values).
582 #define ETH_RSS_LEVEL_INNERMOST (2ULL << 50)
583 #define ETH_RSS_LEVEL_MASK (3ULL << 50)
585 #define ETH_RSS_LEVEL(rss_hf) ((rss_hf & ETH_RSS_LEVEL_MASK) >> 50)
588 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of
589 * the same level are used simultaneously, it is the same case as
590 * none of them are added.
593 * RSS types with SRC/DST_ONLY.
597 static inline uint64_t
598 rte_eth_rss_hf_refine(uint64_t rss_hf)
600 if ((rss_hf & ETH_RSS_L3_SRC_ONLY) && (rss_hf & ETH_RSS_L3_DST_ONLY))
601 rss_hf &= ~(ETH_RSS_L3_SRC_ONLY | ETH_RSS_L3_DST_ONLY);
603 if ((rss_hf & ETH_RSS_L4_SRC_ONLY) && (rss_hf & ETH_RSS_L4_DST_ONLY))
604 rss_hf &= ~(ETH_RSS_L4_SRC_ONLY | ETH_RSS_L4_DST_ONLY);
609 #define ETH_RSS_IPV6_PRE32 ( \
611 RTE_ETH_RSS_L3_PRE32)
613 #define ETH_RSS_IPV6_PRE40 ( \
615 RTE_ETH_RSS_L3_PRE40)
617 #define ETH_RSS_IPV6_PRE48 ( \
619 RTE_ETH_RSS_L3_PRE48)
621 #define ETH_RSS_IPV6_PRE56 ( \
623 RTE_ETH_RSS_L3_PRE56)
625 #define ETH_RSS_IPV6_PRE64 ( \
627 RTE_ETH_RSS_L3_PRE64)
629 #define ETH_RSS_IPV6_PRE96 ( \
631 RTE_ETH_RSS_L3_PRE96)
633 #define ETH_RSS_IPV6_PRE32_UDP ( \
634 ETH_RSS_NONFRAG_IPV6_UDP | \
635 RTE_ETH_RSS_L3_PRE32)
637 #define ETH_RSS_IPV6_PRE40_UDP ( \
638 ETH_RSS_NONFRAG_IPV6_UDP | \
639 RTE_ETH_RSS_L3_PRE40)
641 #define ETH_RSS_IPV6_PRE48_UDP ( \
642 ETH_RSS_NONFRAG_IPV6_UDP | \
643 RTE_ETH_RSS_L3_PRE48)
645 #define ETH_RSS_IPV6_PRE56_UDP ( \
646 ETH_RSS_NONFRAG_IPV6_UDP | \
647 RTE_ETH_RSS_L3_PRE56)
649 #define ETH_RSS_IPV6_PRE64_UDP ( \
650 ETH_RSS_NONFRAG_IPV6_UDP | \
651 RTE_ETH_RSS_L3_PRE64)
653 #define ETH_RSS_IPV6_PRE96_UDP ( \
654 ETH_RSS_NONFRAG_IPV6_UDP | \
655 RTE_ETH_RSS_L3_PRE96)
657 #define ETH_RSS_IPV6_PRE32_TCP ( \
658 ETH_RSS_NONFRAG_IPV6_TCP | \
659 RTE_ETH_RSS_L3_PRE32)
661 #define ETH_RSS_IPV6_PRE40_TCP ( \
662 ETH_RSS_NONFRAG_IPV6_TCP | \
663 RTE_ETH_RSS_L3_PRE40)
665 #define ETH_RSS_IPV6_PRE48_TCP ( \
666 ETH_RSS_NONFRAG_IPV6_TCP | \
667 RTE_ETH_RSS_L3_PRE48)
669 #define ETH_RSS_IPV6_PRE56_TCP ( \
670 ETH_RSS_NONFRAG_IPV6_TCP | \
671 RTE_ETH_RSS_L3_PRE56)
673 #define ETH_RSS_IPV6_PRE64_TCP ( \
674 ETH_RSS_NONFRAG_IPV6_TCP | \
675 RTE_ETH_RSS_L3_PRE64)
677 #define ETH_RSS_IPV6_PRE96_TCP ( \
678 ETH_RSS_NONFRAG_IPV6_TCP | \
679 RTE_ETH_RSS_L3_PRE96)
681 #define ETH_RSS_IPV6_PRE32_SCTP ( \
682 ETH_RSS_NONFRAG_IPV6_SCTP | \
683 RTE_ETH_RSS_L3_PRE32)
685 #define ETH_RSS_IPV6_PRE40_SCTP ( \
686 ETH_RSS_NONFRAG_IPV6_SCTP | \
687 RTE_ETH_RSS_L3_PRE40)
689 #define ETH_RSS_IPV6_PRE48_SCTP ( \
690 ETH_RSS_NONFRAG_IPV6_SCTP | \
691 RTE_ETH_RSS_L3_PRE48)
693 #define ETH_RSS_IPV6_PRE56_SCTP ( \
694 ETH_RSS_NONFRAG_IPV6_SCTP | \
695 RTE_ETH_RSS_L3_PRE56)
697 #define ETH_RSS_IPV6_PRE64_SCTP ( \
698 ETH_RSS_NONFRAG_IPV6_SCTP | \
699 RTE_ETH_RSS_L3_PRE64)
701 #define ETH_RSS_IPV6_PRE96_SCTP ( \
702 ETH_RSS_NONFRAG_IPV6_SCTP | \
703 RTE_ETH_RSS_L3_PRE96)
705 #define ETH_RSS_IP ( \
707 ETH_RSS_FRAG_IPV4 | \
708 ETH_RSS_NONFRAG_IPV4_OTHER | \
710 ETH_RSS_FRAG_IPV6 | \
711 ETH_RSS_NONFRAG_IPV6_OTHER | \
714 #define ETH_RSS_UDP ( \
715 ETH_RSS_NONFRAG_IPV4_UDP | \
716 ETH_RSS_NONFRAG_IPV6_UDP | \
719 #define ETH_RSS_TCP ( \
720 ETH_RSS_NONFRAG_IPV4_TCP | \
721 ETH_RSS_NONFRAG_IPV6_TCP | \
724 #define ETH_RSS_SCTP ( \
725 ETH_RSS_NONFRAG_IPV4_SCTP | \
726 ETH_RSS_NONFRAG_IPV6_SCTP)
728 #define ETH_RSS_TUNNEL ( \
733 #define ETH_RSS_VLAN ( \
737 /**< Mask of valid RSS hash protocols */
738 #define ETH_RSS_PROTO_MASK ( \
740 ETH_RSS_FRAG_IPV4 | \
741 ETH_RSS_NONFRAG_IPV4_TCP | \
742 ETH_RSS_NONFRAG_IPV4_UDP | \
743 ETH_RSS_NONFRAG_IPV4_SCTP | \
744 ETH_RSS_NONFRAG_IPV4_OTHER | \
746 ETH_RSS_FRAG_IPV6 | \
747 ETH_RSS_NONFRAG_IPV6_TCP | \
748 ETH_RSS_NONFRAG_IPV6_UDP | \
749 ETH_RSS_NONFRAG_IPV6_SCTP | \
750 ETH_RSS_NONFRAG_IPV6_OTHER | \
751 ETH_RSS_L2_PAYLOAD | \
753 ETH_RSS_IPV6_TCP_EX | \
754 ETH_RSS_IPV6_UDP_EX | \
761 * Definitions used for redirection table entry size.
762 * Some RSS RETA sizes may not be supported by some drivers, check the
763 * documentation or the description of relevant functions for more details.
765 #define ETH_RSS_RETA_SIZE_64 64
766 #define ETH_RSS_RETA_SIZE_128 128
767 #define ETH_RSS_RETA_SIZE_256 256
768 #define ETH_RSS_RETA_SIZE_512 512
769 #define RTE_RETA_GROUP_SIZE 64
771 /* Definitions used for VMDQ and DCB functionality */
772 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */
773 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
774 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */
775 #define ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
777 /* DCB capability defines */
778 #define ETH_DCB_PG_SUPPORT 0x00000001 /**< Priority Group(ETS) support. */
779 #define ETH_DCB_PFC_SUPPORT 0x00000002 /**< Priority Flow Control support. */
781 /* Definitions used for VLAN Offload functionality */
782 #define ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
783 #define ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
784 #define ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
785 #define ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */
787 /* Definitions used for mask VLAN setting */
788 #define ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
789 #define ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
790 #define ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
791 #define ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */
792 #define ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
794 /* Definitions used for receive MAC address */
795 #define ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
797 /* Definitions used for unicast hash */
798 #define ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
800 /* Definitions used for VMDQ pool rx mode setting */
801 #define ETH_VMDQ_ACCEPT_UNTAG 0x0001 /**< accept untagged packets. */
802 #define ETH_VMDQ_ACCEPT_HASH_MC 0x0002 /**< accept packets in multicast table . */
803 #define ETH_VMDQ_ACCEPT_HASH_UC 0x0004 /**< accept packets in unicast table. */
804 #define ETH_VMDQ_ACCEPT_BROADCAST 0x0008 /**< accept broadcast packets. */
805 #define ETH_VMDQ_ACCEPT_MULTICAST 0x0010 /**< multicast promiscuous. */
807 /** Maximum nb. of vlan per mirror rule */
808 #define ETH_MIRROR_MAX_VLANS 64
810 #define ETH_MIRROR_VIRTUAL_POOL_UP 0x01 /**< Virtual Pool uplink Mirroring. */
811 #define ETH_MIRROR_UPLINK_PORT 0x02 /**< Uplink Port Mirroring. */
812 #define ETH_MIRROR_DOWNLINK_PORT 0x04 /**< Downlink Port Mirroring. */
813 #define ETH_MIRROR_VLAN 0x08 /**< VLAN Mirroring. */
814 #define ETH_MIRROR_VIRTUAL_POOL_DOWN 0x10 /**< Virtual Pool downlink Mirroring. */
817 * A structure used to configure VLAN traffic mirror of an Ethernet port.
819 struct rte_eth_vlan_mirror {
820 uint64_t vlan_mask; /**< mask for valid VLAN ID. */
821 /** VLAN ID list for vlan mirroring. */
822 uint16_t vlan_id[ETH_MIRROR_MAX_VLANS];
826 * A structure used to configure traffic mirror of an Ethernet port.
828 struct rte_eth_mirror_conf {
829 uint8_t rule_type; /**< Mirroring rule type */
830 uint8_t dst_pool; /**< Destination pool for this mirror rule. */
831 uint64_t pool_mask; /**< Bitmap of pool for pool mirroring */
832 /** VLAN ID setting for VLAN mirroring. */
833 struct rte_eth_vlan_mirror vlan;
837 * A structure used to configure 64 entries of Redirection Table of the
838 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
839 * more than 64 entries supported by hardware, an array of this structure
842 struct rte_eth_rss_reta_entry64 {
844 /**< Mask bits indicate which entries need to be updated/queried. */
845 uint16_t reta[RTE_RETA_GROUP_SIZE];
846 /**< Group of 64 redirection table entries. */
850 * This enum indicates the possible number of traffic classes
851 * in DCB configurations
853 enum rte_eth_nb_tcs {
854 ETH_4_TCS = 4, /**< 4 TCs with DCB. */
855 ETH_8_TCS = 8 /**< 8 TCs with DCB. */
859 * This enum indicates the possible number of queue pools
860 * in VMDQ configurations.
862 enum rte_eth_nb_pools {
863 ETH_8_POOLS = 8, /**< 8 VMDq pools. */
864 ETH_16_POOLS = 16, /**< 16 VMDq pools. */
865 ETH_32_POOLS = 32, /**< 32 VMDq pools. */
866 ETH_64_POOLS = 64 /**< 64 VMDq pools. */
869 /* This structure may be extended in future. */
870 struct rte_eth_dcb_rx_conf {
871 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
872 /** Traffic class each UP mapped to. */
873 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
876 struct rte_eth_vmdq_dcb_tx_conf {
877 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
878 /** Traffic class each UP mapped to. */
879 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
882 struct rte_eth_dcb_tx_conf {
883 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
884 /** Traffic class each UP mapped to. */
885 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
888 struct rte_eth_vmdq_tx_conf {
889 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
893 * A structure used to configure the VMDQ+DCB feature
894 * of an Ethernet port.
896 * Using this feature, packets are routed to a pool of queues, based
897 * on the vlan id in the vlan tag, and then to a specific queue within
898 * that pool, using the user priority vlan tag field.
900 * A default pool may be used, if desired, to route all traffic which
901 * does not match the vlan filter rules.
903 struct rte_eth_vmdq_dcb_conf {
904 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 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 nb_pool_maps; /**< We can have up to 64 filters/mappings */
909 uint16_t vlan_id; /**< The vlan id of the received frame */
910 uint64_t pools; /**< Bitmask of pools for packet rx */
911 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
912 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
913 /**< Selects a queue in a pool */
917 * A structure used to configure the VMDQ feature of an Ethernet port when
918 * not combined with the DCB feature.
920 * Using this feature, packets are routed to a pool of queues. By default,
921 * the pool selection is based on the MAC address, the vlan id in the
922 * vlan tag as specified in the pool_map array.
923 * Passing the ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
924 * selection using only the MAC address. MAC address to pool mapping is done
925 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
926 * corresponding to the pool id.
928 * Queue selection within the selected pool will be done using RSS when
929 * it is enabled or revert to the first queue of the pool if not.
931 * A default pool may be used, if desired, to route all traffic which
932 * does not match the vlan filter rules or any pool MAC address.
934 struct rte_eth_vmdq_rx_conf {
935 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
936 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
937 uint8_t default_pool; /**< The default pool, if applicable */
938 uint8_t enable_loop_back; /**< Enable VT loop back */
939 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
940 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
942 uint16_t vlan_id; /**< The vlan id of the received frame */
943 uint64_t pools; /**< Bitmask of pools for packet rx */
944 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
948 * A structure used to configure the TX features of an Ethernet port.
950 struct rte_eth_txmode {
951 enum rte_eth_tx_mq_mode mq_mode; /**< TX multi-queues mode. */
953 * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
954 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
955 * structure are allowed to be set.
961 uint8_t hw_vlan_reject_tagged : 1,
962 /**< If set, reject sending out tagged pkts */
963 hw_vlan_reject_untagged : 1,
964 /**< If set, reject sending out untagged pkts */
965 hw_vlan_insert_pvid : 1;
966 /**< If set, enable port based VLAN insertion */
968 uint64_t reserved_64s[2]; /**< Reserved for future fields */
969 void *reserved_ptrs[2]; /**< Reserved for future fields */
973 * A structure used to configure an RX ring of an Ethernet port.
975 struct rte_eth_rxconf {
976 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */
977 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */
978 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
979 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
981 * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
982 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
983 * fields on rte_eth_dev_info structure are allowed to be set.
987 uint64_t reserved_64s[2]; /**< Reserved for future fields */
988 void *reserved_ptrs[2]; /**< Reserved for future fields */
992 * A structure used to configure a TX ring of an Ethernet port.
994 struct rte_eth_txconf {
995 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */
996 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
997 uint16_t tx_free_thresh; /**< Start freeing TX buffers if there are
998 less free descriptors than this value. */
1000 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1002 * Per-queue Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
1003 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
1004 * fields on rte_eth_dev_info structure are allowed to be set.
1008 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1009 void *reserved_ptrs[2]; /**< Reserved for future fields */
1014 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1016 * A structure used to return the hairpin capabilities that are supported.
1018 struct rte_eth_hairpin_cap {
1019 /** The max number of hairpin queues (different bindings). */
1020 uint16_t max_nb_queues;
1021 /** Max number of Rx queues to be connected to one Tx queue. */
1022 uint16_t max_rx_2_tx;
1023 /** Max number of Tx queues to be connected to one Rx queue. */
1024 uint16_t max_tx_2_rx;
1025 uint16_t max_nb_desc; /**< The max num of descriptors. */
1028 #define RTE_ETH_MAX_HAIRPIN_PEERS 32
1032 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1034 * A structure used to hold hairpin peer data.
1036 struct rte_eth_hairpin_peer {
1037 uint16_t port; /**< Peer port. */
1038 uint16_t queue; /**< Peer queue. */
1043 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1045 * A structure used to configure hairpin binding.
1047 struct rte_eth_hairpin_conf {
1048 uint16_t peer_count; /**< The number of peers. */
1049 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
1053 * A structure contains information about HW descriptor ring limitations.
1055 struct rte_eth_desc_lim {
1056 uint16_t nb_max; /**< Max allowed number of descriptors. */
1057 uint16_t nb_min; /**< Min allowed number of descriptors. */
1058 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
1061 * Max allowed number of segments per whole packet.
1063 * - For TSO packet this is the total number of data descriptors allowed
1066 * @see nb_mtu_seg_max
1068 uint16_t nb_seg_max;
1071 * Max number of segments per one MTU.
1073 * - For non-TSO packet, this is the maximum allowed number of segments
1074 * in a single transmit packet.
1076 * - For TSO packet each segment within the TSO may span up to this
1081 uint16_t nb_mtu_seg_max;
1085 * This enum indicates the flow control mode
1087 enum rte_eth_fc_mode {
1088 RTE_FC_NONE = 0, /**< Disable flow control. */
1089 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */
1090 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */
1091 RTE_FC_FULL /**< Enable flow control on both side. */
1095 * A structure used to configure Ethernet flow control parameter.
1096 * These parameters will be configured into the register of the NIC.
1097 * Please refer to the corresponding data sheet for proper value.
1099 struct rte_eth_fc_conf {
1100 uint32_t high_water; /**< High threshold value to trigger XOFF */
1101 uint32_t low_water; /**< Low threshold value to trigger XON */
1102 uint16_t pause_time; /**< Pause quota in the Pause frame */
1103 uint16_t send_xon; /**< Is XON frame need be sent */
1104 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1105 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
1106 uint8_t autoneg; /**< Use Pause autoneg */
1110 * A structure used to configure Ethernet priority flow control parameter.
1111 * These parameters will be configured into the register of the NIC.
1112 * Please refer to the corresponding data sheet for proper value.
1114 struct rte_eth_pfc_conf {
1115 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
1116 uint8_t priority; /**< VLAN User Priority. */
1122 enum rte_eth_tunnel_type {
1123 RTE_TUNNEL_TYPE_NONE = 0,
1124 RTE_TUNNEL_TYPE_VXLAN,
1125 RTE_TUNNEL_TYPE_GENEVE,
1126 RTE_TUNNEL_TYPE_TEREDO,
1127 RTE_TUNNEL_TYPE_NVGRE,
1128 RTE_TUNNEL_TYPE_IP_IN_GRE,
1129 RTE_L2_TUNNEL_TYPE_E_TAG,
1130 RTE_TUNNEL_TYPE_VXLAN_GPE,
1131 RTE_TUNNEL_TYPE_MAX,
1134 /* Deprecated API file for rte_eth_dev_filter_* functions */
1135 #include "rte_eth_ctrl.h"
1138 * Memory space that can be configured to store Flow Director filters
1139 * in the board memory.
1141 enum rte_fdir_pballoc_type {
1142 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */
1143 RTE_FDIR_PBALLOC_128K, /**< 128k. */
1144 RTE_FDIR_PBALLOC_256K, /**< 256k. */
1148 * Select report mode of FDIR hash information in RX descriptors.
1150 enum rte_fdir_status_mode {
1151 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
1152 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
1153 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
1157 * A structure used to configure the Flow Director (FDIR) feature
1158 * of an Ethernet port.
1160 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored.
1162 struct rte_fdir_conf {
1163 enum rte_fdir_mode mode; /**< Flow Director mode. */
1164 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
1165 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
1166 /** RX queue of packets matching a "drop" filter in perfect mode. */
1168 struct rte_eth_fdir_masks mask;
1169 struct rte_eth_fdir_flex_conf flex_conf;
1170 /**< Flex payload configuration. */
1174 * UDP tunneling configuration.
1175 * Used to config the UDP port for a type of tunnel.
1176 * NICs need the UDP port to identify the tunnel type.
1177 * Normally a type of tunnel has a default UDP port, this structure can be used
1178 * in case if the users want to change or support more UDP port.
1180 struct rte_eth_udp_tunnel {
1181 uint16_t udp_port; /**< UDP port used for the tunnel. */
1182 uint8_t prot_type; /**< Tunnel type. Defined in rte_eth_tunnel_type. */
1186 * A structure used to enable/disable specific device interrupts.
1188 struct rte_intr_conf {
1189 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
1191 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
1193 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
1198 * A structure used to configure an Ethernet port.
1199 * Depending upon the RX multi-queue mode, extra advanced
1200 * configuration settings may be needed.
1202 struct rte_eth_conf {
1203 uint32_t link_speeds; /**< bitmap of ETH_LINK_SPEED_XXX of speeds to be
1204 used. ETH_LINK_SPEED_FIXED disables link
1205 autonegotiation, and a unique speed shall be
1206 set. Otherwise, the bitmap defines the set of
1207 speeds to be advertised. If the special value
1208 ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
1209 supported are advertised. */
1210 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */
1211 struct rte_eth_txmode txmode; /**< Port TX configuration. */
1212 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
1213 is 0, meaning the loopback mode is disabled.
1214 Read the datasheet of given ethernet controller
1215 for details. The possible values of this field
1216 are defined in implementation of each driver. */
1218 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
1219 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
1220 /**< Port vmdq+dcb configuration. */
1221 struct rte_eth_dcb_rx_conf dcb_rx_conf;
1222 /**< Port dcb RX configuration. */
1223 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
1224 /**< Port vmdq RX configuration. */
1225 } rx_adv_conf; /**< Port RX filtering configuration. */
1227 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
1228 /**< Port vmdq+dcb TX configuration. */
1229 struct rte_eth_dcb_tx_conf dcb_tx_conf;
1230 /**< Port dcb TX configuration. */
1231 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
1232 /**< Port vmdq TX configuration. */
1233 } tx_adv_conf; /**< Port TX DCB configuration (union). */
1234 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
1235 is needed,and the variable must be set ETH_DCB_PFC_SUPPORT. */
1236 uint32_t dcb_capability_en;
1237 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */
1238 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */
1242 * RX offload capabilities of a device.
1244 #define DEV_RX_OFFLOAD_VLAN_STRIP 0x00000001
1245 #define DEV_RX_OFFLOAD_IPV4_CKSUM 0x00000002
1246 #define DEV_RX_OFFLOAD_UDP_CKSUM 0x00000004
1247 #define DEV_RX_OFFLOAD_TCP_CKSUM 0x00000008
1248 #define DEV_RX_OFFLOAD_TCP_LRO 0x00000010
1249 #define DEV_RX_OFFLOAD_QINQ_STRIP 0x00000020
1250 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000040
1251 #define DEV_RX_OFFLOAD_MACSEC_STRIP 0x00000080
1252 #define DEV_RX_OFFLOAD_HEADER_SPLIT 0x00000100
1253 #define DEV_RX_OFFLOAD_VLAN_FILTER 0x00000200
1254 #define DEV_RX_OFFLOAD_VLAN_EXTEND 0x00000400
1255 #define DEV_RX_OFFLOAD_JUMBO_FRAME 0x00000800
1256 #define DEV_RX_OFFLOAD_SCATTER 0x00002000
1257 #define DEV_RX_OFFLOAD_TIMESTAMP 0x00004000
1258 #define DEV_RX_OFFLOAD_SECURITY 0x00008000
1259 #define DEV_RX_OFFLOAD_KEEP_CRC 0x00010000
1260 #define DEV_RX_OFFLOAD_SCTP_CKSUM 0x00020000
1261 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM 0x00040000
1262 #define DEV_RX_OFFLOAD_RSS_HASH 0x00080000
1264 #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \
1265 DEV_RX_OFFLOAD_UDP_CKSUM | \
1266 DEV_RX_OFFLOAD_TCP_CKSUM)
1267 #define DEV_RX_OFFLOAD_VLAN (DEV_RX_OFFLOAD_VLAN_STRIP | \
1268 DEV_RX_OFFLOAD_VLAN_FILTER | \
1269 DEV_RX_OFFLOAD_VLAN_EXTEND | \
1270 DEV_RX_OFFLOAD_QINQ_STRIP)
1273 * If new Rx offload capabilities are defined, they also must be
1274 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
1278 * TX offload capabilities of a device.
1280 #define DEV_TX_OFFLOAD_VLAN_INSERT 0x00000001
1281 #define DEV_TX_OFFLOAD_IPV4_CKSUM 0x00000002
1282 #define DEV_TX_OFFLOAD_UDP_CKSUM 0x00000004
1283 #define DEV_TX_OFFLOAD_TCP_CKSUM 0x00000008
1284 #define DEV_TX_OFFLOAD_SCTP_CKSUM 0x00000010
1285 #define DEV_TX_OFFLOAD_TCP_TSO 0x00000020
1286 #define DEV_TX_OFFLOAD_UDP_TSO 0x00000040
1287 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000080 /**< Used for tunneling packet. */
1288 #define DEV_TX_OFFLOAD_QINQ_INSERT 0x00000100
1289 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO 0x00000200 /**< Used for tunneling packet. */
1290 #define DEV_TX_OFFLOAD_GRE_TNL_TSO 0x00000400 /**< Used for tunneling packet. */
1291 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO 0x00000800 /**< Used for tunneling packet. */
1292 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO 0x00001000 /**< Used for tunneling packet. */
1293 #define DEV_TX_OFFLOAD_MACSEC_INSERT 0x00002000
1294 #define DEV_TX_OFFLOAD_MT_LOCKFREE 0x00004000
1295 /**< Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
1296 * tx queue without SW lock.
1298 #define DEV_TX_OFFLOAD_MULTI_SEGS 0x00008000
1299 /**< Device supports multi segment send. */
1300 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE 0x00010000
1301 /**< Device supports optimization for fast release of mbufs.
1302 * When set application must guarantee that per-queue all mbufs comes from
1303 * the same mempool and has refcnt = 1.
1305 #define DEV_TX_OFFLOAD_SECURITY 0x00020000
1307 * Device supports generic UDP tunneled packet TSO.
1308 * Application must set PKT_TX_TUNNEL_UDP and other mbuf fields required
1311 #define DEV_TX_OFFLOAD_UDP_TNL_TSO 0x00040000
1313 * Device supports generic IP tunneled packet TSO.
1314 * Application must set PKT_TX_TUNNEL_IP and other mbuf fields required
1317 #define DEV_TX_OFFLOAD_IP_TNL_TSO 0x00080000
1318 /** Device supports outer UDP checksum */
1319 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM 0x00100000
1321 /** Device supports send on timestamp */
1322 #define DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP 0x00200000
1325 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP 0x00000001
1326 /**< Device supports Rx queue setup after device started*/
1327 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP 0x00000002
1328 /**< Device supports Tx queue setup after device started*/
1331 * If new Tx offload capabilities are defined, they also must be
1332 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1336 * Fallback default preferred Rx/Tx port parameters.
1337 * These are used if an application requests default parameters
1338 * but the PMD does not provide preferred values.
1340 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1341 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1342 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1343 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1346 * Preferred Rx/Tx port parameters.
1347 * There are separate instances of this structure for transmission
1348 * and reception respectively.
1350 struct rte_eth_dev_portconf {
1351 uint16_t burst_size; /**< Device-preferred burst size */
1352 uint16_t ring_size; /**< Device-preferred size of queue rings */
1353 uint16_t nb_queues; /**< Device-preferred number of queues */
1357 * Default values for switch domain id when ethdev does not support switch
1358 * domain definitions.
1360 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX)
1363 * Ethernet device associated switch information
1365 struct rte_eth_switch_info {
1366 const char *name; /**< switch name */
1367 uint16_t domain_id; /**< switch domain id */
1370 * mapping to the devices physical switch port as enumerated from the
1371 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1372 * device this may correspond to the VF_ID of each virtual function,
1373 * but each driver should explicitly define the mapping of switch
1374 * port identifier to that physical interconnect/switch
1379 * Ethernet device information
1383 * A structure used to retrieve the contextual information of
1384 * an Ethernet device, such as the controlling driver of the
1387 struct rte_eth_dev_info {
1388 struct rte_device *device; /** Generic device information */
1389 const char *driver_name; /**< Device Driver name. */
1390 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1391 Use if_indextoname() to translate into an interface name. */
1392 uint16_t min_mtu; /**< Minimum MTU allowed */
1393 uint16_t max_mtu; /**< Maximum MTU allowed */
1394 const uint32_t *dev_flags; /**< Device flags */
1395 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */
1396 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */
1397 /** Maximum configurable size of LRO aggregated packet. */
1398 uint32_t max_lro_pkt_size;
1399 uint16_t max_rx_queues; /**< Maximum number of RX queues. */
1400 uint16_t max_tx_queues; /**< Maximum number of TX queues. */
1401 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1402 uint32_t max_hash_mac_addrs;
1403 /** Maximum number of hash MAC addresses for MTA and UTA. */
1404 uint16_t max_vfs; /**< Maximum number of VFs. */
1405 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1406 uint64_t rx_offload_capa;
1407 /**< All RX offload capabilities including all per-queue ones */
1408 uint64_t tx_offload_capa;
1409 /**< All TX offload capabilities including all per-queue ones */
1410 uint64_t rx_queue_offload_capa;
1411 /**< Device per-queue RX offload capabilities. */
1412 uint64_t tx_queue_offload_capa;
1413 /**< Device per-queue TX offload capabilities. */
1415 /**< Device redirection table size, the total number of entries. */
1416 uint8_t hash_key_size; /**< Hash key size in bytes */
1417 /** Bit mask of RSS offloads, the bit offset also means flow type */
1418 uint64_t flow_type_rss_offloads;
1419 struct rte_eth_rxconf default_rxconf; /**< Default RX configuration */
1420 struct rte_eth_txconf default_txconf; /**< Default TX configuration */
1421 uint16_t vmdq_queue_base; /**< First queue ID for VMDQ pools. */
1422 uint16_t vmdq_queue_num; /**< Queue number for VMDQ pools. */
1423 uint16_t vmdq_pool_base; /**< First ID of VMDQ pools. */
1424 struct rte_eth_desc_lim rx_desc_lim; /**< RX descriptors limits */
1425 struct rte_eth_desc_lim tx_desc_lim; /**< TX descriptors limits */
1426 uint32_t speed_capa; /**< Supported speeds bitmap (ETH_LINK_SPEED_). */
1427 /** Configured number of rx/tx queues */
1428 uint16_t nb_rx_queues; /**< Number of RX queues. */
1429 uint16_t nb_tx_queues; /**< Number of TX queues. */
1430 /** Rx parameter recommendations */
1431 struct rte_eth_dev_portconf default_rxportconf;
1432 /** Tx parameter recommendations */
1433 struct rte_eth_dev_portconf default_txportconf;
1434 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1437 * Switching information for ports on a device with a
1438 * embedded managed interconnect/switch.
1440 struct rte_eth_switch_info switch_info;
1442 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1443 void *reserved_ptrs[2]; /**< Reserved for future fields */
1447 * Ethernet device RX queue information structure.
1448 * Used to retrieve information about configured queue.
1450 struct rte_eth_rxq_info {
1451 struct rte_mempool *mp; /**< mempool used by that queue. */
1452 struct rte_eth_rxconf conf; /**< queue config parameters. */
1453 uint8_t scattered_rx; /**< scattered packets RX supported. */
1454 uint16_t nb_desc; /**< configured number of RXDs. */
1455 uint16_t rx_buf_size; /**< hardware receive buffer size. */
1456 } __rte_cache_min_aligned;
1459 * Ethernet device TX queue information structure.
1460 * Used to retrieve information about configured queue.
1462 struct rte_eth_txq_info {
1463 struct rte_eth_txconf conf; /**< queue config parameters. */
1464 uint16_t nb_desc; /**< configured number of TXDs. */
1465 } __rte_cache_min_aligned;
1467 /* Generic Burst mode flag definition, values can be ORed. */
1470 * If the queues have different burst mode description, this bit will be set
1471 * by PMD, then the application can iterate to retrieve burst description for
1474 #define RTE_ETH_BURST_FLAG_PER_QUEUE (1ULL << 0)
1477 * Ethernet device RX/TX queue packet burst mode information structure.
1478 * Used to retrieve information about packet burst mode setting.
1480 struct rte_eth_burst_mode {
1481 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */
1483 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */
1484 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */
1487 /** Maximum name length for extended statistics counters */
1488 #define RTE_ETH_XSTATS_NAME_SIZE 64
1491 * An Ethernet device extended statistic structure
1493 * This structure is used by rte_eth_xstats_get() to provide
1494 * statistics that are not provided in the generic *rte_eth_stats*
1496 * It maps a name id, corresponding to an index in the array returned
1497 * by rte_eth_xstats_get_names(), to a statistic value.
1499 struct rte_eth_xstat {
1500 uint64_t id; /**< The index in xstats name array. */
1501 uint64_t value; /**< The statistic counter value. */
1505 * A name element for extended statistics.
1507 * An array of this structure is returned by rte_eth_xstats_get_names().
1508 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1509 * structure references these names by their array index.
1511 struct rte_eth_xstat_name {
1512 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1515 #define ETH_DCB_NUM_TCS 8
1516 #define ETH_MAX_VMDQ_POOL 64
1519 * A structure used to get the information of queue and
1520 * TC mapping on both TX and RX paths.
1522 struct rte_eth_dcb_tc_queue_mapping {
1523 /** rx queues assigned to tc per Pool */
1527 } tc_rxq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1528 /** rx queues assigned to tc per Pool */
1532 } tc_txq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1536 * A structure used to get the information of DCB.
1537 * It includes TC UP mapping and queue TC mapping.
1539 struct rte_eth_dcb_info {
1540 uint8_t nb_tcs; /**< number of TCs */
1541 uint8_t prio_tc[ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
1542 uint8_t tc_bws[ETH_DCB_NUM_TCS]; /**< TX BW percentage for each TC */
1543 /** rx queues assigned to tc */
1544 struct rte_eth_dcb_tc_queue_mapping tc_queue;
1547 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
1549 /* Macros to check for valid port */
1550 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
1551 if (!rte_eth_dev_is_valid_port(port_id)) { \
1552 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1557 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
1558 if (!rte_eth_dev_is_valid_port(port_id)) { \
1559 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1565 * l2 tunnel configuration.
1568 /**< l2 tunnel enable mask */
1569 #define ETH_L2_TUNNEL_ENABLE_MASK 0x00000001
1570 /**< l2 tunnel insertion mask */
1571 #define ETH_L2_TUNNEL_INSERTION_MASK 0x00000002
1572 /**< l2 tunnel stripping mask */
1573 #define ETH_L2_TUNNEL_STRIPPING_MASK 0x00000004
1574 /**< l2 tunnel forwarding mask */
1575 #define ETH_L2_TUNNEL_FORWARDING_MASK 0x00000008
1578 * Function type used for RX packet processing packet callbacks.
1580 * The callback function is called on RX with a burst of packets that have
1581 * been received on the given port and queue.
1584 * The Ethernet port on which RX is being performed.
1586 * The queue on the Ethernet port which is being used to receive the packets.
1588 * The burst of packets that have just been received.
1590 * The number of packets in the burst pointed to by "pkts".
1592 * The max number of packets that can be stored in the "pkts" array.
1594 * The arbitrary user parameter passed in by the application when the callback
1595 * was originally configured.
1597 * The number of packets returned to the user.
1599 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
1600 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
1604 * Function type used for TX packet processing packet callbacks.
1606 * The callback function is called on TX with a burst of packets immediately
1607 * before the packets are put onto the hardware queue for transmission.
1610 * The Ethernet port on which TX is being performed.
1612 * The queue on the Ethernet port which is being used to transmit the packets.
1614 * The burst of packets that are about to be transmitted.
1616 * The number of packets in the burst pointed to by "pkts".
1618 * The arbitrary user parameter passed in by the application when the callback
1619 * was originally configured.
1621 * The number of packets to be written to the NIC.
1623 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
1624 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
1627 * Possible states of an ethdev port.
1629 enum rte_eth_dev_state {
1630 /** Device is unused before being probed. */
1631 RTE_ETH_DEV_UNUSED = 0,
1632 /** Device is attached when allocated in probing. */
1633 RTE_ETH_DEV_ATTACHED,
1634 /** Device is in removed state when plug-out is detected. */
1635 RTE_ETH_DEV_REMOVED,
1638 struct rte_eth_dev_sriov {
1639 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
1640 uint8_t nb_q_per_pool; /**< rx queue number per pool */
1641 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
1642 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
1644 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
1646 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
1648 #define RTE_ETH_DEV_NO_OWNER 0
1650 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
1652 struct rte_eth_dev_owner {
1653 uint64_t id; /**< The owner unique identifier. */
1654 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
1657 /** Device supports link state interrupt */
1658 #define RTE_ETH_DEV_INTR_LSC 0x0002
1659 /** Device is a bonded slave */
1660 #define RTE_ETH_DEV_BONDED_SLAVE 0x0004
1661 /** Device supports device removal interrupt */
1662 #define RTE_ETH_DEV_INTR_RMV 0x0008
1663 /** Device is port representor */
1664 #define RTE_ETH_DEV_REPRESENTOR 0x0010
1665 /** Device does not support MAC change after started */
1666 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR 0x0020
1669 * Iterates over valid ethdev ports owned by a specific owner.
1672 * The id of the next possible valid owned port.
1674 * The owner identifier.
1675 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
1677 * Next valid port id owned by owner_id, RTE_MAX_ETHPORTS if there is none.
1679 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
1680 const uint64_t owner_id);
1683 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
1685 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
1686 for (p = rte_eth_find_next_owned_by(0, o); \
1687 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
1688 p = rte_eth_find_next_owned_by(p + 1, o))
1691 * Iterates over valid ethdev ports.
1694 * The id of the next possible valid port.
1696 * Next valid port id, RTE_MAX_ETHPORTS if there is none.
1698 uint16_t rte_eth_find_next(uint16_t port_id);
1701 * Macro to iterate over all enabled and ownerless ethdev ports.
1703 #define RTE_ETH_FOREACH_DEV(p) \
1704 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
1708 * @b EXPERIMENTAL: this API may change without prior notice.
1710 * Iterates over ethdev ports of a specified device.
1712 * @param port_id_start
1713 * The id of the next possible valid port.
1715 * The generic device behind the ports to iterate.
1717 * Next port id of the device, possibly port_id_start,
1718 * RTE_MAX_ETHPORTS if there is none.
1722 rte_eth_find_next_of(uint16_t port_id_start,
1723 const struct rte_device *parent);
1726 * Macro to iterate over all ethdev ports of a specified device.
1729 * The id of the matching port being iterated.
1731 * The rte_device pointer matching the iterated ports.
1733 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \
1734 for (port_id = rte_eth_find_next_of(0, parent); \
1735 port_id < RTE_MAX_ETHPORTS; \
1736 port_id = rte_eth_find_next_of(port_id + 1, parent))
1740 * @b EXPERIMENTAL: this API may change without prior notice.
1742 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device).
1744 * @param port_id_start
1745 * The id of the next possible valid sibling port.
1746 * @param ref_port_id
1747 * The id of a reference port to compare rte_device with.
1749 * Next sibling port id, possibly port_id_start or ref_port_id itself,
1750 * RTE_MAX_ETHPORTS if there is none.
1754 rte_eth_find_next_sibling(uint16_t port_id_start, uint16_t ref_port_id);
1757 * Macro to iterate over all ethdev ports sharing the same rte_device
1758 * as the specified port.
1759 * Note: the specified reference port is part of the loop iterations.
1762 * The id of the matching port being iterated.
1763 * @param ref_port_id
1764 * The id of the port being compared.
1766 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \
1767 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \
1768 port_id < RTE_MAX_ETHPORTS; \
1769 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id))
1773 * @b EXPERIMENTAL: this API may change without prior notice.
1775 * Get a new unique owner identifier.
1776 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
1777 * to avoid multiple management of device by different entities.
1780 * Owner identifier pointer.
1782 * Negative errno value on error, 0 on success.
1785 int rte_eth_dev_owner_new(uint64_t *owner_id);
1789 * @b EXPERIMENTAL: this API may change without prior notice.
1791 * Set an Ethernet device owner.
1794 * The identifier of the port to own.
1796 * The owner pointer.
1798 * Negative errno value on error, 0 on success.
1801 int rte_eth_dev_owner_set(const uint16_t port_id,
1802 const struct rte_eth_dev_owner *owner);
1806 * @b EXPERIMENTAL: this API may change without prior notice.
1808 * Unset Ethernet device owner to make the device ownerless.
1811 * The identifier of port to make ownerless.
1813 * The owner identifier.
1815 * 0 on success, negative errno value on error.
1818 int rte_eth_dev_owner_unset(const uint16_t port_id,
1819 const uint64_t owner_id);
1823 * @b EXPERIMENTAL: this API may change without prior notice.
1825 * Remove owner from all Ethernet devices owned by a specific owner.
1828 * The owner identifier.
1830 * 0 on success, negative errno value on error.
1833 int rte_eth_dev_owner_delete(const uint64_t owner_id);
1837 * @b EXPERIMENTAL: this API may change without prior notice.
1839 * Get the owner of an Ethernet device.
1842 * The port identifier.
1844 * The owner structure pointer to fill.
1846 * 0 on success, negative errno value on error..
1849 int rte_eth_dev_owner_get(const uint16_t port_id,
1850 struct rte_eth_dev_owner *owner);
1853 * Get the number of ports which are usable for the application.
1855 * These devices must be iterated by using the macro
1856 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
1857 * to deal with non-contiguous ranges of devices.
1860 * The count of available Ethernet devices.
1862 uint16_t rte_eth_dev_count_avail(void);
1865 * Get the total number of ports which are allocated.
1867 * Some devices may not be available for the application.
1870 * The total count of Ethernet devices.
1872 uint16_t rte_eth_dev_count_total(void);
1875 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
1876 * the bitmap link_speeds of the struct rte_eth_conf
1879 * Numerical speed value in Mbps
1881 * ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
1883 * 0 if the speed cannot be mapped
1885 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
1888 * Get DEV_RX_OFFLOAD_* flag name.
1893 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1895 const char *rte_eth_dev_rx_offload_name(uint64_t offload);
1898 * Get DEV_TX_OFFLOAD_* flag name.
1903 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1905 const char *rte_eth_dev_tx_offload_name(uint64_t offload);
1908 * Configure an Ethernet device.
1909 * This function must be invoked first before any other function in the
1910 * Ethernet API. This function can also be re-invoked when a device is in the
1914 * The port identifier of the Ethernet device to configure.
1915 * @param nb_rx_queue
1916 * The number of receive queues to set up for the Ethernet device.
1917 * @param nb_tx_queue
1918 * The number of transmit queues to set up for the Ethernet device.
1920 * The pointer to the configuration data to be used for the Ethernet device.
1921 * The *rte_eth_conf* structure includes:
1922 * - the hardware offload features to activate, with dedicated fields for
1923 * each statically configurable offload hardware feature provided by
1924 * Ethernet devices, such as IP checksum or VLAN tag stripping for
1926 * The Rx offload bitfield API is obsolete and will be deprecated.
1927 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
1928 * structure and use offloads field to set per-port offloads instead.
1929 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
1930 * the [rt]x_offload_capa returned from rte_eth_dev_info_get().
1931 * Any type of device supported offloading set in the input argument
1932 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
1933 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup()
1934 * - the Receive Side Scaling (RSS) configuration when using multiple RX
1935 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf
1936 * must be within the flow_type_rss_offloads provided by drivers via
1937 * rte_eth_dev_info_get() API.
1939 * Embedding all configuration information in a single data structure
1940 * is the more flexible method that allows the addition of new features
1941 * without changing the syntax of the API.
1943 * - 0: Success, device configured.
1944 * - <0: Error code returned by the driver configuration function.
1946 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
1947 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
1951 * @b EXPERIMENTAL: this API may change without prior notice.
1953 * Check if an Ethernet device was physically removed.
1956 * The port identifier of the Ethernet device.
1958 * 1 when the Ethernet device is removed, otherwise 0.
1962 rte_eth_dev_is_removed(uint16_t port_id);
1965 * Allocate and set up a receive queue for an Ethernet device.
1967 * The function allocates a contiguous block of memory for *nb_rx_desc*
1968 * receive descriptors from a memory zone associated with *socket_id*
1969 * and initializes each receive descriptor with a network buffer allocated
1970 * from the memory pool *mb_pool*.
1973 * The port identifier of the Ethernet device.
1974 * @param rx_queue_id
1975 * The index of the receive queue to set up.
1976 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1977 * to rte_eth_dev_configure().
1979 * The number of receive descriptors to allocate for the receive ring.
1981 * The *socket_id* argument is the socket identifier in case of NUMA.
1982 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1983 * the DMA memory allocated for the receive descriptors of the ring.
1985 * The pointer to the configuration data to be used for the receive queue.
1986 * NULL value is allowed, in which case default RX configuration
1988 * The *rx_conf* structure contains an *rx_thresh* structure with the values
1989 * of the Prefetch, Host, and Write-Back threshold registers of the receive
1991 * In addition it contains the hardware offloads features to activate using
1992 * the DEV_RX_OFFLOAD_* flags.
1993 * If an offloading set in rx_conf->offloads
1994 * hasn't been set in the input argument eth_conf->rxmode.offloads
1995 * to rte_eth_dev_configure(), it is a new added offloading, it must be
1996 * per-queue type and it is enabled for the queue.
1997 * No need to repeat any bit in rx_conf->offloads which has already been
1998 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
1999 * at port level can't be disabled at queue level.
2001 * The pointer to the memory pool from which to allocate *rte_mbuf* network
2002 * memory buffers to populate each descriptor of the receive ring.
2004 * - 0: Success, receive queue correctly set up.
2005 * - -EIO: if device is removed.
2006 * - -EINVAL: The memory pool pointer is null or the size of network buffers
2007 * which can be allocated from this memory pool does not fit the various
2008 * buffer sizes allowed by the device controller.
2009 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
2010 * allocate network memory buffers from the memory pool when
2011 * initializing receive descriptors.
2013 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
2014 uint16_t nb_rx_desc, unsigned int socket_id,
2015 const struct rte_eth_rxconf *rx_conf,
2016 struct rte_mempool *mb_pool);
2020 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2022 * Allocate and set up a hairpin receive queue for an Ethernet device.
2024 * The function set up the selected queue to be used in hairpin.
2027 * The port identifier of the Ethernet device.
2028 * @param rx_queue_id
2029 * The index of the receive queue to set up.
2030 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2031 * to rte_eth_dev_configure().
2033 * The number of receive descriptors to allocate for the receive ring.
2034 * 0 means the PMD will use default value.
2036 * The pointer to the hairpin configuration.
2039 * - (0) if successful.
2040 * - (-ENOTSUP) if hardware doesn't support.
2041 * - (-EINVAL) if bad parameter.
2042 * - (-ENOMEM) if unable to allocate the resources.
2045 int rte_eth_rx_hairpin_queue_setup
2046 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc,
2047 const struct rte_eth_hairpin_conf *conf);
2050 * Allocate and set up a transmit queue for an Ethernet device.
2053 * The port identifier of the Ethernet device.
2054 * @param tx_queue_id
2055 * The index of the transmit queue to set up.
2056 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2057 * to rte_eth_dev_configure().
2059 * The number of transmit descriptors to allocate for the transmit ring.
2061 * The *socket_id* argument is the socket identifier in case of NUMA.
2062 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2063 * the DMA memory allocated for the transmit descriptors of the ring.
2065 * The pointer to the configuration data to be used for the transmit queue.
2066 * NULL value is allowed, in which case default TX configuration
2068 * The *tx_conf* structure contains the following data:
2069 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
2070 * Write-Back threshold registers of the transmit ring.
2071 * When setting Write-Back threshold to the value greater then zero,
2072 * *tx_rs_thresh* value should be explicitly set to one.
2073 * - The *tx_free_thresh* value indicates the [minimum] number of network
2074 * buffers that must be pending in the transmit ring to trigger their
2075 * [implicit] freeing by the driver transmit function.
2076 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
2077 * descriptors that must be pending in the transmit ring before setting the
2078 * RS bit on a descriptor by the driver transmit function.
2079 * The *tx_rs_thresh* value should be less or equal then
2080 * *tx_free_thresh* value, and both of them should be less then
2082 * - The *offloads* member contains Tx offloads to be enabled.
2083 * If an offloading set in tx_conf->offloads
2084 * hasn't been set in the input argument eth_conf->txmode.offloads
2085 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2086 * per-queue type and it is enabled for the queue.
2087 * No need to repeat any bit in tx_conf->offloads which has already been
2088 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2089 * at port level can't be disabled at queue level.
2091 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
2092 * the transmit function to use default values.
2094 * - 0: Success, the transmit queue is correctly set up.
2095 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
2097 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
2098 uint16_t nb_tx_desc, unsigned int socket_id,
2099 const struct rte_eth_txconf *tx_conf);
2103 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2105 * Allocate and set up a transmit hairpin queue for an Ethernet device.
2108 * The port identifier of the Ethernet device.
2109 * @param tx_queue_id
2110 * The index of the transmit queue to set up.
2111 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2112 * to rte_eth_dev_configure().
2114 * The number of transmit descriptors to allocate for the transmit ring.
2115 * 0 to set default PMD value.
2117 * The hairpin configuration.
2120 * - (0) if successful.
2121 * - (-ENOTSUP) if hardware doesn't support.
2122 * - (-EINVAL) if bad parameter.
2123 * - (-ENOMEM) if unable to allocate the resources.
2126 int rte_eth_tx_hairpin_queue_setup
2127 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc,
2128 const struct rte_eth_hairpin_conf *conf);
2131 * Return the NUMA socket to which an Ethernet device is connected
2134 * The port identifier of the Ethernet device
2136 * The NUMA socket id to which the Ethernet device is connected or
2137 * a default of zero if the socket could not be determined.
2138 * -1 is returned is the port_id value is out of range.
2140 int rte_eth_dev_socket_id(uint16_t port_id);
2143 * Check if port_id of device is attached
2146 * The port identifier of the Ethernet device
2148 * - 0 if port is out of range or not attached
2149 * - 1 if device is attached
2151 int rte_eth_dev_is_valid_port(uint16_t port_id);
2154 * Start specified RX queue of a port. It is used when rx_deferred_start
2155 * flag of the specified queue is true.
2158 * The port identifier of the Ethernet device
2159 * @param rx_queue_id
2160 * The index of the rx queue to update the ring.
2161 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2162 * to rte_eth_dev_configure().
2164 * - 0: Success, the receive queue is started.
2165 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2166 * - -EIO: if device is removed.
2167 * - -ENOTSUP: The function not supported in PMD driver.
2169 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2172 * Stop specified RX queue of a port
2175 * The port identifier of the Ethernet device
2176 * @param rx_queue_id
2177 * The index of the rx queue to update the ring.
2178 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2179 * to rte_eth_dev_configure().
2181 * - 0: Success, the receive queue is stopped.
2182 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2183 * - -EIO: if device is removed.
2184 * - -ENOTSUP: The function not supported in PMD driver.
2186 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2189 * Start TX for specified queue of a port. It is used when tx_deferred_start
2190 * flag of the specified queue is true.
2193 * The port identifier of the Ethernet device
2194 * @param tx_queue_id
2195 * The index of the tx queue to update the ring.
2196 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2197 * to rte_eth_dev_configure().
2199 * - 0: Success, the transmit queue is started.
2200 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2201 * - -EIO: if device is removed.
2202 * - -ENOTSUP: The function not supported in PMD driver.
2204 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2207 * Stop specified TX queue of a port
2210 * The port identifier of the Ethernet device
2211 * @param tx_queue_id
2212 * The index of the tx queue to update the ring.
2213 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2214 * to rte_eth_dev_configure().
2216 * - 0: Success, the transmit queue is stopped.
2217 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin.
2218 * - -EIO: if device is removed.
2219 * - -ENOTSUP: The function not supported in PMD driver.
2221 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2224 * Start an Ethernet device.
2226 * The device start step is the last one and consists of setting the configured
2227 * offload features and in starting the transmit and the receive units of the
2230 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before
2231 * PMD port start callback function is invoked.
2233 * On success, all basic functions exported by the Ethernet API (link status,
2234 * receive/transmit, and so on) can be invoked.
2237 * The port identifier of the Ethernet device.
2239 * - 0: Success, Ethernet device started.
2240 * - <0: Error code of the driver device start function.
2242 int rte_eth_dev_start(uint16_t port_id);
2245 * Stop an Ethernet device. The device can be restarted with a call to
2246 * rte_eth_dev_start()
2249 * The port identifier of the Ethernet device.
2251 void rte_eth_dev_stop(uint16_t port_id);
2254 * Link up an Ethernet device.
2256 * Set device link up will re-enable the device rx/tx
2257 * functionality after it is previously set device linked down.
2260 * The port identifier of the Ethernet device.
2262 * - 0: Success, Ethernet device linked up.
2263 * - <0: Error code of the driver device link up function.
2265 int rte_eth_dev_set_link_up(uint16_t port_id);
2268 * Link down an Ethernet device.
2269 * The device rx/tx functionality will be disabled if success,
2270 * and it can be re-enabled with a call to
2271 * rte_eth_dev_set_link_up()
2274 * The port identifier of the Ethernet device.
2276 int rte_eth_dev_set_link_down(uint16_t port_id);
2279 * Close a stopped Ethernet device. The device cannot be restarted!
2280 * The function frees all port resources.
2283 * The port identifier of the Ethernet device.
2285 void rte_eth_dev_close(uint16_t port_id);
2288 * Reset a Ethernet device and keep its port id.
2290 * When a port has to be reset passively, the DPDK application can invoke
2291 * this function. For example when a PF is reset, all its VFs should also
2292 * be reset. Normally a DPDK application can invoke this function when
2293 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2294 * a port reset in other circumstances.
2296 * When this function is called, it first stops the port and then calls the
2297 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2298 * state, in which no Tx and Rx queues are setup, as if the port has been
2299 * reset and not started. The port keeps the port id it had before the
2302 * After calling rte_eth_dev_reset( ), the application should use
2303 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2304 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2305 * to reconfigure the device as appropriate.
2307 * Note: To avoid unexpected behavior, the application should stop calling
2308 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2309 * safety, all these controlling functions should be called from the same
2313 * The port identifier of the Ethernet device.
2316 * - (0) if successful.
2317 * - (-EINVAL) if port identifier is invalid.
2318 * - (-ENOTSUP) if hardware doesn't support this function.
2319 * - (-EPERM) if not ran from the primary process.
2320 * - (-EIO) if re-initialisation failed or device is removed.
2321 * - (-ENOMEM) if the reset failed due to OOM.
2322 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2324 int rte_eth_dev_reset(uint16_t port_id);
2327 * Enable receipt in promiscuous mode for an Ethernet device.
2330 * The port identifier of the Ethernet device.
2332 * - (0) if successful.
2333 * - (-ENOTSUP) if support for promiscuous_enable() does not exist
2335 * - (-ENODEV) if *port_id* invalid.
2337 int rte_eth_promiscuous_enable(uint16_t port_id);
2340 * Disable receipt in promiscuous mode for an Ethernet device.
2343 * The port identifier of the Ethernet device.
2345 * - (0) if successful.
2346 * - (-ENOTSUP) if support for promiscuous_disable() does not exist
2348 * - (-ENODEV) if *port_id* invalid.
2350 int rte_eth_promiscuous_disable(uint16_t port_id);
2353 * Return the value of promiscuous mode for an Ethernet device.
2356 * The port identifier of the Ethernet device.
2358 * - (1) if promiscuous is enabled
2359 * - (0) if promiscuous is disabled.
2362 int rte_eth_promiscuous_get(uint16_t port_id);
2365 * Enable the receipt of any multicast frame by an Ethernet device.
2368 * The port identifier of the Ethernet device.
2370 * - (0) if successful.
2371 * - (-ENOTSUP) if support for allmulticast_enable() does not exist
2373 * - (-ENODEV) if *port_id* invalid.
2375 int rte_eth_allmulticast_enable(uint16_t port_id);
2378 * Disable the receipt of all multicast frames by an Ethernet device.
2381 * The port identifier of the Ethernet device.
2383 * - (0) if successful.
2384 * - (-ENOTSUP) if support for allmulticast_disable() does not exist
2386 * - (-ENODEV) if *port_id* invalid.
2388 int rte_eth_allmulticast_disable(uint16_t port_id);
2391 * Return the value of allmulticast mode for an Ethernet device.
2394 * The port identifier of the Ethernet device.
2396 * - (1) if allmulticast is enabled
2397 * - (0) if allmulticast is disabled.
2400 int rte_eth_allmulticast_get(uint16_t port_id);
2403 * Retrieve the link status (up/down), the duplex mode (half/full),
2404 * the negotiation (auto/fixed), and if available, the speed (Mbps).
2406 * It might need to wait up to 9 seconds.
2407 * @see rte_eth_link_get_nowait.
2410 * The port identifier of the Ethernet device.
2412 * Link information written back.
2414 * - (0) if successful.
2415 * - (-ENOTSUP) if the function is not supported in PMD driver.
2416 * - (-ENODEV) if *port_id* invalid.
2418 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
2421 * Retrieve the link status (up/down), the duplex mode (half/full),
2422 * the negotiation (auto/fixed), and if available, the speed (Mbps).
2425 * The port identifier of the Ethernet device.
2427 * Link information written back.
2429 * - (0) if successful.
2430 * - (-ENOTSUP) if the function is not supported in PMD driver.
2431 * - (-ENODEV) if *port_id* invalid.
2433 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
2437 * @b EXPERIMENTAL: this API may change without prior notice.
2439 * The function converts a link_speed to a string. It handles all special
2440 * values like unknown or none speed.
2443 * link_speed of rte_eth_link struct
2445 * Link speed in textual format. It's pointer to immutable memory.
2446 * No free is required.
2449 const char *rte_eth_link_speed_to_str(uint32_t link_speed);
2453 * @b EXPERIMENTAL: this API may change without prior notice.
2455 * The function converts a rte_eth_link struct representing a link status to
2459 * A pointer to a string to be filled with textual representation of
2460 * device status. At least ETH_LINK_MAX_STR_LEN bytes should be allocated to
2461 * store default link status text.
2463 * Length of available memory at 'str' string.
2465 * Link status returned by rte_eth_link_get function
2467 * Number of bytes written to str array.
2470 int rte_eth_link_to_str(char *str, size_t len,
2471 const struct rte_eth_link *eth_link);
2474 * Retrieve the general I/O statistics of an Ethernet device.
2477 * The port identifier of the Ethernet device.
2479 * A pointer to a structure of type *rte_eth_stats* to be filled with
2480 * the values of device counters for the following set of statistics:
2481 * - *ipackets* with the total of successfully received packets.
2482 * - *opackets* with the total of successfully transmitted packets.
2483 * - *ibytes* with the total of successfully received bytes.
2484 * - *obytes* with the total of successfully transmitted bytes.
2485 * - *ierrors* with the total of erroneous received packets.
2486 * - *oerrors* with the total of failed transmitted packets.
2488 * Zero if successful. Non-zero otherwise.
2490 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
2493 * Reset the general I/O statistics of an Ethernet device.
2496 * The port identifier of the Ethernet device.
2498 * - (0) if device notified to reset stats.
2499 * - (-ENOTSUP) if hardware doesn't support.
2500 * - (-ENODEV) if *port_id* invalid.
2501 * - (<0): Error code of the driver stats reset function.
2503 int rte_eth_stats_reset(uint16_t port_id);
2506 * Retrieve names of extended statistics of an Ethernet device.
2508 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
2510 * xstats_names[i].name => xstats[i].value
2512 * And the array index is same with id field of 'struct rte_eth_xstat':
2515 * This assumption makes key-value pair matching less flexible but simpler.
2518 * The port identifier of the Ethernet device.
2519 * @param xstats_names
2520 * An rte_eth_xstat_name array of at least *size* elements to
2521 * be filled. If set to NULL, the function returns the required number
2524 * The size of the xstats_names array (number of elements).
2526 * - A positive value lower or equal to size: success. The return value
2527 * is the number of entries filled in the stats table.
2528 * - A positive value higher than size: error, the given statistics table
2529 * is too small. The return value corresponds to the size that should
2530 * be given to succeed. The entries in the table are not valid and
2531 * shall not be used by the caller.
2532 * - A negative value on error (invalid port id).
2534 int rte_eth_xstats_get_names(uint16_t port_id,
2535 struct rte_eth_xstat_name *xstats_names,
2539 * Retrieve extended statistics of an Ethernet device.
2541 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
2543 * xstats_names[i].name => xstats[i].value
2545 * And the array index is same with id field of 'struct rte_eth_xstat':
2548 * This assumption makes key-value pair matching less flexible but simpler.
2551 * The port identifier of the Ethernet device.
2553 * A pointer to a table of structure of type *rte_eth_xstat*
2554 * to be filled with device statistics ids and values.
2555 * This parameter can be set to NULL if n is 0.
2557 * The size of the xstats array (number of elements).
2559 * - A positive value lower or equal to n: success. The return value
2560 * is the number of entries filled in the stats table.
2561 * - A positive value higher than n: error, the given statistics table
2562 * is too small. The return value corresponds to the size that should
2563 * be given to succeed. The entries in the table are not valid and
2564 * shall not be used by the caller.
2565 * - A negative value on error (invalid port id).
2567 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
2571 * Retrieve names of extended statistics of an Ethernet device.
2574 * The port identifier of the Ethernet device.
2575 * @param xstats_names
2576 * An rte_eth_xstat_name array of at least *size* elements to
2577 * be filled. If set to NULL, the function returns the required number
2580 * IDs array given by app to retrieve specific statistics
2582 * The size of the xstats_names array (number of elements).
2584 * - A positive value lower or equal to size: success. The return value
2585 * is the number of entries filled in the stats table.
2586 * - A positive value higher than size: error, the given statistics table
2587 * is too small. The return value corresponds to the size that should
2588 * be given to succeed. The entries in the table are not valid and
2589 * shall not be used by the caller.
2590 * - A negative value on error (invalid port id).
2593 rte_eth_xstats_get_names_by_id(uint16_t port_id,
2594 struct rte_eth_xstat_name *xstats_names, unsigned int size,
2598 * Retrieve extended statistics of an Ethernet device.
2601 * The port identifier of the Ethernet device.
2603 * A pointer to an ids array passed by application. This tells which
2604 * statistics values function should retrieve. This parameter
2605 * can be set to NULL if size is 0. In this case function will retrieve
2606 * all available statistics.
2608 * A pointer to a table to be filled with device statistics values.
2610 * The size of the ids array (number of elements).
2612 * - A positive value lower or equal to size: success. The return value
2613 * is the number of entries filled in the stats table.
2614 * - A positive value higher than size: error, the given statistics table
2615 * is too small. The return value corresponds to the size that should
2616 * be given to succeed. The entries in the table are not valid and
2617 * shall not be used by the caller.
2618 * - A negative value on error (invalid port id).
2620 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
2621 uint64_t *values, unsigned int size);
2624 * Gets the ID of a statistic from its name.
2626 * This function searches for the statistics using string compares, and
2627 * as such should not be used on the fast-path. For fast-path retrieval of
2628 * specific statistics, store the ID as provided in *id* from this function,
2629 * and pass the ID to rte_eth_xstats_get()
2631 * @param port_id The port to look up statistics from
2632 * @param xstat_name The name of the statistic to return
2633 * @param[out] id A pointer to an app-supplied uint64_t which should be
2634 * set to the ID of the stat if the stat exists.
2637 * -ENODEV for invalid port_id,
2638 * -EIO if device is removed,
2639 * -EINVAL if the xstat_name doesn't exist in port_id
2641 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
2645 * Reset extended statistics of an Ethernet device.
2648 * The port identifier of the Ethernet device.
2650 * - (0) if device notified to reset extended stats.
2651 * - (-ENOTSUP) if pmd doesn't support both
2652 * extended stats and basic stats reset.
2653 * - (-ENODEV) if *port_id* invalid.
2654 * - (<0): Error code of the driver xstats reset function.
2656 int rte_eth_xstats_reset(uint16_t port_id);
2659 * Set a mapping for the specified transmit queue to the specified per-queue
2660 * statistics counter.
2663 * The port identifier of the Ethernet device.
2664 * @param tx_queue_id
2665 * The index of the transmit queue for which a queue stats mapping is required.
2666 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2667 * to rte_eth_dev_configure().
2669 * The per-queue packet statistics functionality number that the transmit
2670 * queue is to be assigned.
2671 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
2673 * Zero if successful. Non-zero otherwise.
2675 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
2676 uint16_t tx_queue_id, uint8_t stat_idx);
2679 * Set a mapping for the specified receive queue to the specified per-queue
2680 * statistics counter.
2683 * The port identifier of the Ethernet device.
2684 * @param rx_queue_id
2685 * The index of the receive queue for which a queue stats mapping is required.
2686 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2687 * to rte_eth_dev_configure().
2689 * The per-queue packet statistics functionality number that the receive
2690 * queue is to be assigned.
2691 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
2693 * Zero if successful. Non-zero otherwise.
2695 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
2696 uint16_t rx_queue_id,
2700 * Retrieve the Ethernet address of an Ethernet device.
2703 * The port identifier of the Ethernet device.
2705 * A pointer to a structure of type *ether_addr* to be filled with
2706 * the Ethernet address of the Ethernet device.
2708 * - (0) if successful
2709 * - (-ENODEV) if *port_id* invalid.
2711 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr);
2714 * Retrieve the contextual information of an Ethernet device.
2716 * As part of this function, a number of of fields in dev_info will be
2717 * initialized as follows:
2722 * Where lim is defined within the rte_eth_dev_info_get as
2724 * const struct rte_eth_desc_lim lim = {
2725 * .nb_max = UINT16_MAX,
2728 * .nb_seg_max = UINT16_MAX,
2729 * .nb_mtu_seg_max = UINT16_MAX,
2732 * device = dev->device
2733 * min_mtu = RTE_ETHER_MIN_MTU
2734 * max_mtu = UINT16_MAX
2736 * The following fields will be populated if support for dev_infos_get()
2737 * exists for the device and the rte_eth_dev 'dev' has been populated
2738 * successfully with a call to it:
2740 * driver_name = dev->device->driver->name
2741 * nb_rx_queues = dev->data->nb_rx_queues
2742 * nb_tx_queues = dev->data->nb_tx_queues
2743 * dev_flags = &dev->data->dev_flags
2746 * The port identifier of the Ethernet device.
2748 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
2749 * the contextual information of the Ethernet device.
2751 * - (0) if successful.
2752 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device.
2753 * - (-ENODEV) if *port_id* invalid.
2755 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
2758 * Retrieve the firmware version of a device.
2761 * The port identifier of the device.
2763 * A pointer to a string array storing the firmware version of a device,
2764 * the string includes terminating null. This pointer is allocated by caller.
2766 * The size of the string array pointed by fw_version, which should be
2767 * large enough to store firmware version of the device.
2769 * - (0) if successful.
2770 * - (-ENOTSUP) if operation is not supported.
2771 * - (-ENODEV) if *port_id* invalid.
2772 * - (-EIO) if device is removed.
2773 * - (>0) if *fw_size* is not enough to store firmware version, return
2774 * the size of the non truncated string.
2776 int rte_eth_dev_fw_version_get(uint16_t port_id,
2777 char *fw_version, size_t fw_size);
2780 * Retrieve the supported packet types of an Ethernet device.
2782 * When a packet type is announced as supported, it *must* be recognized by
2783 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
2784 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
2785 * packet types for these packets:
2786 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
2787 * - Ether/Vlan/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
2788 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
2789 * - Ether/Vlan/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
2791 * When a packet is received by a PMD, the most precise type must be
2792 * returned among the ones supported. However a PMD is allowed to set
2793 * packet type that is not in the supported list, at the condition that it
2794 * is more precise. Therefore, a PMD announcing no supported packet types
2795 * can still set a matching packet type in a received packet.
2798 * Better to invoke this API after the device is already started or rx burst
2799 * function is decided, to obtain correct supported ptypes.
2801 * if a given PMD does not report what ptypes it supports, then the supported
2802 * ptype count is reported as 0.
2804 * The port identifier of the Ethernet device.
2806 * A hint of what kind of packet type which the caller is interested in.
2808 * An array pointer to store adequate packet types, allocated by caller.
2810 * Size of the array pointed by param ptypes.
2812 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
2813 * only num entries will be filled into the ptypes array, but the full
2814 * count of supported ptypes will be returned.
2815 * - (-ENODEV) if *port_id* invalid.
2817 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
2818 uint32_t *ptypes, int num);
2821 * @b EXPERIMENTAL: this API may change without prior notice.
2823 * Inform Ethernet device about reduced range of packet types to handle.
2825 * Application can use this function to set only specific ptypes that it's
2826 * interested. This information can be used by the PMD to optimize Rx path.
2828 * The function accepts an array `set_ptypes` allocated by the caller to
2829 * store the packet types set by the driver, the last element of the array
2830 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be
2831 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled
2835 * The port identifier of the Ethernet device.
2837 * The ptype family that application is interested in should be bitwise OR of
2838 * RTE_PTYPE_*_MASK or 0.
2840 * An array pointer to store set packet types, allocated by caller. The
2841 * function marks the end of array with RTE_PTYPE_UNKNOWN.
2843 * Size of the array pointed by param ptypes.
2844 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the
2848 * - (-ENODEV) if *port_id* invalid.
2849 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and
2853 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask,
2854 uint32_t *set_ptypes, unsigned int num);
2857 * Retrieve the MTU of an Ethernet device.
2860 * The port identifier of the Ethernet device.
2862 * A pointer to a uint16_t where the retrieved MTU is to be stored.
2864 * - (0) if successful.
2865 * - (-ENODEV) if *port_id* invalid.
2867 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
2870 * Change the MTU of an Ethernet device.
2873 * The port identifier of the Ethernet device.
2875 * A uint16_t for the MTU to be applied.
2877 * - (0) if successful.
2878 * - (-ENOTSUP) if operation is not supported.
2879 * - (-ENODEV) if *port_id* invalid.
2880 * - (-EIO) if device is removed.
2881 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within
2882 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or
2883 * when the mtu is set using dev->dev_ops->mtu_set.
2884 * - (-EBUSY) if operation is not allowed when the port is running
2886 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
2889 * Enable/Disable hardware filtering by an Ethernet device of received
2890 * VLAN packets tagged with a given VLAN Tag Identifier.
2893 * The port identifier of the Ethernet device.
2895 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
2897 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
2898 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
2900 * - (0) if successful.
2901 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
2902 * - (-ENODEV) if *port_id* invalid.
2903 * - (-EIO) if device is removed.
2904 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
2905 * - (-EINVAL) if *vlan_id* > 4095.
2907 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
2910 * Enable/Disable hardware VLAN Strip by a rx queue of an Ethernet device.
2913 * The port identifier of the Ethernet device.
2914 * @param rx_queue_id
2915 * The index of the receive queue for which a queue stats mapping is required.
2916 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2917 * to rte_eth_dev_configure().
2919 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
2920 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
2922 * - (0) if successful.
2923 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured.
2924 * - (-ENODEV) if *port_id* invalid.
2925 * - (-EINVAL) if *rx_queue_id* invalid.
2927 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
2931 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
2935 * The port identifier of the Ethernet device.
2939 * The Tag Protocol ID
2941 * - (0) if successful.
2942 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported.
2943 * - (-ENODEV) if *port_id* invalid.
2944 * - (-EIO) if device is removed.
2946 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
2947 enum rte_vlan_type vlan_type,
2951 * Set VLAN offload configuration on an Ethernet device.
2954 * The port identifier of the Ethernet device.
2955 * @param offload_mask
2956 * The VLAN Offload bit mask can be mixed use with "OR"
2957 * ETH_VLAN_STRIP_OFFLOAD
2958 * ETH_VLAN_FILTER_OFFLOAD
2959 * ETH_VLAN_EXTEND_OFFLOAD
2960 * ETH_QINQ_STRIP_OFFLOAD
2962 * - (0) if successful.
2963 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
2964 * - (-ENODEV) if *port_id* invalid.
2965 * - (-EIO) if device is removed.
2967 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
2970 * Read VLAN Offload configuration from an Ethernet device
2973 * The port identifier of the Ethernet device.
2975 * - (>0) if successful. Bit mask to indicate
2976 * ETH_VLAN_STRIP_OFFLOAD
2977 * ETH_VLAN_FILTER_OFFLOAD
2978 * ETH_VLAN_EXTEND_OFFLOAD
2979 * ETH_QINQ_STRIP_OFFLOAD
2980 * - (-ENODEV) if *port_id* invalid.
2982 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
2985 * Set port based TX VLAN insertion on or off.
2988 * The port identifier of the Ethernet device.
2990 * Port based TX VLAN identifier together with user priority.
2992 * Turn on or off the port based TX VLAN insertion.
2995 * - (0) if successful.
2996 * - negative if failed.
2998 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
3000 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
3004 * Structure used to buffer packets for future TX
3005 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
3007 struct rte_eth_dev_tx_buffer {
3008 buffer_tx_error_fn error_callback;
3009 void *error_userdata;
3010 uint16_t size; /**< Size of buffer for buffered tx */
3011 uint16_t length; /**< Number of packets in the array */
3012 struct rte_mbuf *pkts[];
3013 /**< Pending packets to be sent on explicit flush or when full */
3017 * Calculate the size of the tx buffer.
3020 * Number of stored packets.
3022 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
3023 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
3026 * Initialize default values for buffered transmitting
3029 * Tx buffer to be initialized.
3036 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
3039 * Configure a callback for buffered packets which cannot be sent
3041 * Register a specific callback to be called when an attempt is made to send
3042 * all packets buffered on an ethernet port, but not all packets can
3043 * successfully be sent. The callback registered here will be called only
3044 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
3045 * The default callback configured for each queue by default just frees the
3046 * packets back to the calling mempool. If additional behaviour is required,
3047 * for example, to count dropped packets, or to retry transmission of packets
3048 * which cannot be sent, this function should be used to register a suitable
3049 * callback function to implement the desired behaviour.
3050 * The example callback "rte_eth_count_unsent_packet_callback()" is also
3051 * provided as reference.
3054 * The port identifier of the Ethernet device.
3056 * The function to be used as the callback.
3058 * Arbitrary parameter to be passed to the callback function
3060 * 0 on success, or -1 on error with rte_errno set appropriately
3063 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
3064 buffer_tx_error_fn callback, void *userdata);
3067 * Callback function for silently dropping unsent buffered packets.
3069 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3070 * adjust the default behavior when buffered packets cannot be sent. This
3071 * function drops any unsent packets silently and is used by tx buffered
3072 * operations as default behavior.
3074 * NOTE: this function should not be called directly, instead it should be used
3075 * as a callback for packet buffering.
3077 * NOTE: when configuring this function as a callback with
3078 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3079 * should point to an uint64_t value.
3082 * The previously buffered packets which could not be sent
3084 * The number of unsent packets in the pkts array
3089 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
3093 * Callback function for tracking unsent buffered packets.
3095 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3096 * adjust the default behavior when buffered packets cannot be sent. This
3097 * function drops any unsent packets, but also updates a user-supplied counter
3098 * to track the overall number of packets dropped. The counter should be an
3099 * uint64_t variable.
3101 * NOTE: this function should not be called directly, instead it should be used
3102 * as a callback for packet buffering.
3104 * NOTE: when configuring this function as a callback with
3105 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3106 * should point to an uint64_t value.
3109 * The previously buffered packets which could not be sent
3111 * The number of unsent packets in the pkts array
3113 * Pointer to an uint64_t value, which will be incremented by unsent
3116 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
3120 * Request the driver to free mbufs currently cached by the driver. The
3121 * driver will only free the mbuf if it is no longer in use. It is the
3122 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is
3126 * The port identifier of the Ethernet device.
3128 * The index of the transmit queue through which output packets must be
3130 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3131 * to rte_eth_dev_configure().
3133 * Maximum number of packets to free. Use 0 to indicate all possible packets
3134 * should be freed. Note that a packet may be using multiple mbufs.
3137 * -ENODEV: Invalid interface
3138 * -EIO: device is removed
3139 * -ENOTSUP: Driver does not support function
3141 * 0-n: Number of packets freed. More packets may still remain in ring that
3145 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
3148 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
3151 enum rte_eth_event_ipsec_subtype {
3152 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
3153 /**< Unknown event type */
3154 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
3155 /**< Sequence number overflow */
3156 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
3157 /**< Soft time expiry of SA */
3158 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
3159 /**< Soft byte expiry of SA */
3160 RTE_ETH_EVENT_IPSEC_MAX
3161 /**< Max value of this enum */
3165 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
3166 * information of the IPsec offload event.
3168 struct rte_eth_event_ipsec_desc {
3169 enum rte_eth_event_ipsec_subtype subtype;
3170 /**< Type of RTE_ETH_EVENT_IPSEC_* event */
3172 /**< Event specific metadata
3174 * For the following events, *userdata* registered
3175 * with the *rte_security_session* would be returned
3178 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
3179 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
3180 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
3182 * @see struct rte_security_session_conf
3188 * The eth device event type for interrupt, and maybe others in the future.
3190 enum rte_eth_event_type {
3191 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3192 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3193 RTE_ETH_EVENT_QUEUE_STATE,
3194 /**< queue state event (enabled/disabled) */
3195 RTE_ETH_EVENT_INTR_RESET,
3196 /**< reset interrupt event, sent to VF on PF reset */
3197 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3198 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3199 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3200 RTE_ETH_EVENT_NEW, /**< port is probed */
3201 RTE_ETH_EVENT_DESTROY, /**< port is released */
3202 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
3203 RTE_ETH_EVENT_FLOW_AGED,/**< New aged-out flows is detected */
3204 RTE_ETH_EVENT_MAX /**< max value of this enum */
3207 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3208 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3209 /**< user application callback to be registered for interrupts */
3212 * Register a callback function for port event.
3216 * RTE_ETH_ALL means register the event for all port ids.
3220 * User supplied callback function to be called.
3222 * Pointer to the parameters for the registered callback.
3225 * - On success, zero.
3226 * - On failure, a negative value.
3228 int rte_eth_dev_callback_register(uint16_t port_id,
3229 enum rte_eth_event_type event,
3230 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3233 * Unregister a callback function for port event.
3237 * RTE_ETH_ALL means unregister the event for all port ids.
3241 * User supplied callback function to be called.
3243 * Pointer to the parameters for the registered callback. -1 means to
3244 * remove all for the same callback address and same event.
3247 * - On success, zero.
3248 * - On failure, a negative value.
3250 int rte_eth_dev_callback_unregister(uint16_t port_id,
3251 enum rte_eth_event_type event,
3252 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3255 * When there is no rx packet coming in Rx Queue for a long time, we can
3256 * sleep lcore related to RX Queue for power saving, and enable rx interrupt
3257 * to be triggered when Rx packet arrives.
3259 * The rte_eth_dev_rx_intr_enable() function enables rx queue
3260 * interrupt on specific rx queue of a port.
3263 * The port identifier of the Ethernet device.
3265 * The index of the receive queue from which to retrieve input packets.
3266 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3267 * to rte_eth_dev_configure().
3269 * - (0) if successful.
3270 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3272 * - (-ENODEV) if *port_id* invalid.
3273 * - (-EIO) if device is removed.
3275 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
3278 * When lcore wakes up from rx interrupt indicating packet coming, disable rx
3279 * interrupt and returns to polling mode.
3281 * The rte_eth_dev_rx_intr_disable() function disables rx queue
3282 * interrupt on specific rx queue of a port.
3285 * The port identifier of the Ethernet device.
3287 * The index of the receive queue from which to retrieve input packets.
3288 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3289 * to rte_eth_dev_configure().
3291 * - (0) if successful.
3292 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3294 * - (-ENODEV) if *port_id* invalid.
3295 * - (-EIO) if device is removed.
3297 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
3300 * RX Interrupt control per port.
3303 * The port identifier of the Ethernet device.
3305 * Epoll instance fd which the intr vector associated to.
3306 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3308 * The operation be performed for the vector.
3309 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3313 * - On success, zero.
3314 * - On failure, a negative value.
3316 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
3319 * RX Interrupt control per queue.
3322 * The port identifier of the Ethernet device.
3324 * The index of the receive queue from which to retrieve input packets.
3325 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3326 * to rte_eth_dev_configure().
3328 * Epoll instance fd which the intr vector associated to.
3329 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3331 * The operation be performed for the vector.
3332 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3336 * - On success, zero.
3337 * - On failure, a negative value.
3339 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
3340 int epfd, int op, void *data);
3344 * @b EXPERIMENTAL: this API may change without prior notice.
3346 * Get interrupt fd per Rx queue.
3349 * The port identifier of the Ethernet device.
3351 * The index of the receive queue from which to retrieve input packets.
3352 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3353 * to rte_eth_dev_configure().
3355 * - (>=0) the interrupt fd associated to the requested Rx queue if
3361 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id);
3364 * Turn on the LED on the Ethernet device.
3365 * This function turns on the LED on the Ethernet device.
3368 * The port identifier of the Ethernet device.
3370 * - (0) if successful.
3371 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3373 * - (-ENODEV) if *port_id* invalid.
3374 * - (-EIO) if device is removed.
3376 int rte_eth_led_on(uint16_t port_id);
3379 * Turn off the LED on the Ethernet device.
3380 * This function turns off the LED on the Ethernet device.
3383 * The port identifier of the Ethernet device.
3385 * - (0) if successful.
3386 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3388 * - (-ENODEV) if *port_id* invalid.
3389 * - (-EIO) if device is removed.
3391 int rte_eth_led_off(uint16_t port_id);
3394 * Get current status of the Ethernet link flow control for Ethernet device
3397 * The port identifier of the Ethernet device.
3399 * The pointer to the structure where to store the flow control parameters.
3401 * - (0) if successful.
3402 * - (-ENOTSUP) if hardware doesn't support flow control.
3403 * - (-ENODEV) if *port_id* invalid.
3404 * - (-EIO) if device is removed.
3406 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
3407 struct rte_eth_fc_conf *fc_conf);
3410 * Configure the Ethernet link flow control for Ethernet device
3413 * The port identifier of the Ethernet device.
3415 * The pointer to the structure of the flow control parameters.
3417 * - (0) if successful.
3418 * - (-ENOTSUP) if hardware doesn't support flow control mode.
3419 * - (-ENODEV) if *port_id* invalid.
3420 * - (-EINVAL) if bad parameter
3421 * - (-EIO) if flow control setup failure or device is removed.
3423 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
3424 struct rte_eth_fc_conf *fc_conf);
3427 * Configure the Ethernet priority flow control under DCB environment
3428 * for Ethernet device.
3431 * The port identifier of the Ethernet device.
3433 * The pointer to the structure of the priority flow control parameters.
3435 * - (0) if successful.
3436 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
3437 * - (-ENODEV) if *port_id* invalid.
3438 * - (-EINVAL) if bad parameter
3439 * - (-EIO) if flow control setup failure or device is removed.
3441 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
3442 struct rte_eth_pfc_conf *pfc_conf);
3445 * Add a MAC address to the set used for filtering incoming packets.
3448 * The port identifier of the Ethernet device.
3450 * The MAC address to add.
3452 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
3453 * not enabled, this should be set to 0.
3455 * - (0) if successfully added or *mac_addr* was already added.
3456 * - (-ENOTSUP) if hardware doesn't support this feature.
3457 * - (-ENODEV) if *port* is invalid.
3458 * - (-EIO) if device is removed.
3459 * - (-ENOSPC) if no more MAC addresses can be added.
3460 * - (-EINVAL) if MAC address is invalid.
3462 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr,
3466 * Remove a MAC address from the internal array of addresses.
3469 * The port identifier of the Ethernet device.
3471 * MAC address to remove.
3473 * - (0) if successful, or *mac_addr* didn't exist.
3474 * - (-ENOTSUP) if hardware doesn't support.
3475 * - (-ENODEV) if *port* invalid.
3476 * - (-EADDRINUSE) if attempting to remove the default MAC address
3478 int rte_eth_dev_mac_addr_remove(uint16_t port_id,
3479 struct rte_ether_addr *mac_addr);
3482 * Set the default MAC address.
3485 * The port identifier of the Ethernet device.
3487 * New default MAC address.
3489 * - (0) if successful, or *mac_addr* didn't exist.
3490 * - (-ENOTSUP) if hardware doesn't support.
3491 * - (-ENODEV) if *port* invalid.
3492 * - (-EINVAL) if MAC address is invalid.
3494 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
3495 struct rte_ether_addr *mac_addr);
3498 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3501 * The port identifier of the Ethernet device.
3505 * Redirection table size. The table size can be queried by
3506 * rte_eth_dev_info_get().
3508 * - (0) if successful.
3509 * - (-ENOTSUP) if hardware doesn't support.
3510 * - (-EINVAL) if bad parameter.
3511 * - (-EIO) if device is removed.
3513 int rte_eth_dev_rss_reta_update(uint16_t port_id,
3514 struct rte_eth_rss_reta_entry64 *reta_conf,
3515 uint16_t reta_size);
3518 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3521 * The port identifier of the Ethernet device.
3523 * RETA to query. For each requested reta entry, corresponding bit
3524 * in mask must be set.
3526 * Redirection table size. The table size can be queried by
3527 * rte_eth_dev_info_get().
3529 * - (0) if successful.
3530 * - (-ENOTSUP) if hardware doesn't support.
3531 * - (-EINVAL) if bad parameter.
3532 * - (-EIO) if device is removed.
3534 int rte_eth_dev_rss_reta_query(uint16_t port_id,
3535 struct rte_eth_rss_reta_entry64 *reta_conf,
3536 uint16_t reta_size);
3539 * Updates unicast hash table for receiving packet with the given destination
3540 * MAC address, and the packet is routed to all VFs for which the RX mode is
3541 * accept packets that match the unicast hash table.
3544 * The port identifier of the Ethernet device.
3546 * Unicast MAC address.
3548 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
3549 * 0 - Clear an unicast hash bit.
3551 * - (0) if successful.
3552 * - (-ENOTSUP) if hardware doesn't support.
3553 * - (-ENODEV) if *port_id* invalid.
3554 * - (-EIO) if device is removed.
3555 * - (-EINVAL) if bad parameter.
3557 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr,
3561 * Updates all unicast hash bitmaps for receiving packet with any Unicast
3562 * Ethernet MAC addresses,the packet is routed to all VFs for which the RX
3563 * mode is accept packets that match the unicast hash table.
3566 * The port identifier of the Ethernet device.
3568 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
3570 * 0 - Clear all unicast hash bitmaps
3572 * - (0) if successful.
3573 * - (-ENOTSUP) if hardware doesn't support.
3574 * - (-ENODEV) if *port_id* invalid.
3575 * - (-EIO) if device is removed.
3576 * - (-EINVAL) if bad parameter.
3578 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
3581 * Set a traffic mirroring rule on an Ethernet device
3584 * The port identifier of the Ethernet device.
3585 * @param mirror_conf
3586 * The pointer to the traffic mirroring structure describing the mirroring rule.
3587 * The *rte_eth_vm_mirror_conf* structure includes the type of mirroring rule,
3588 * destination pool and the value of rule if enable vlan or pool mirroring.
3591 * The index of traffic mirroring rule, we support four separated rules.
3593 * 1 - Enable a mirroring rule.
3594 * 0 - Disable a mirroring rule.
3596 * - (0) if successful.
3597 * - (-ENOTSUP) if hardware doesn't support this feature.
3598 * - (-ENODEV) if *port_id* invalid.
3599 * - (-EIO) if device is removed.
3600 * - (-EINVAL) if the mr_conf information is not correct.
3602 int rte_eth_mirror_rule_set(uint16_t port_id,
3603 struct rte_eth_mirror_conf *mirror_conf,
3608 * Reset a traffic mirroring rule on an Ethernet device.
3611 * The port identifier of the Ethernet device.
3613 * The index of traffic mirroring rule, we support four separated rules.
3615 * - (0) if successful.
3616 * - (-ENOTSUP) if hardware doesn't support this feature.
3617 * - (-ENODEV) if *port_id* invalid.
3618 * - (-EIO) if device is removed.
3619 * - (-EINVAL) if bad parameter.
3621 int rte_eth_mirror_rule_reset(uint16_t port_id,
3625 * Set the rate limitation for a queue on an Ethernet device.
3628 * The port identifier of the Ethernet device.
3632 * The tx rate in Mbps. Allocated from the total port link speed.
3634 * - (0) if successful.
3635 * - (-ENOTSUP) if hardware doesn't support this feature.
3636 * - (-ENODEV) if *port_id* invalid.
3637 * - (-EIO) if device is removed.
3638 * - (-EINVAL) if bad parameter.
3640 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
3644 * Configuration of Receive Side Scaling hash computation of Ethernet device.
3647 * The port identifier of the Ethernet device.
3649 * The new configuration to use for RSS hash computation on the port.
3651 * - (0) if successful.
3652 * - (-ENODEV) if port identifier is invalid.
3653 * - (-EIO) if device is removed.
3654 * - (-ENOTSUP) if hardware doesn't support.
3655 * - (-EINVAL) if bad parameter.
3657 int rte_eth_dev_rss_hash_update(uint16_t port_id,
3658 struct rte_eth_rss_conf *rss_conf);
3661 * Retrieve current configuration of Receive Side Scaling hash computation
3662 * of Ethernet device.
3665 * The port identifier of the Ethernet device.
3667 * Where to store the current RSS hash configuration of the Ethernet device.
3669 * - (0) if successful.
3670 * - (-ENODEV) if port identifier is invalid.
3671 * - (-EIO) if device is removed.
3672 * - (-ENOTSUP) if hardware doesn't support RSS.
3675 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
3676 struct rte_eth_rss_conf *rss_conf);
3679 * Add UDP tunneling port for a specific type of tunnel.
3680 * The packets with this UDP port will be identified as this type of tunnel.
3681 * Before enabling any offloading function for a tunnel, users can call this API
3682 * to change or add more UDP port for the tunnel. So the offloading function
3683 * can take effect on the packets with the specific UDP port.
3686 * The port identifier of the Ethernet device.
3688 * UDP tunneling configuration.
3691 * - (0) if successful.
3692 * - (-ENODEV) if port identifier is invalid.
3693 * - (-EIO) if device is removed.
3694 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3697 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
3698 struct rte_eth_udp_tunnel *tunnel_udp);
3701 * Delete UDP tunneling port a specific type of tunnel.
3702 * The packets with this UDP port will not be identified as this type of tunnel
3704 * Before enabling any offloading function for a tunnel, users can call this API
3705 * to delete a UDP port for the tunnel. So the offloading function will not take
3706 * effect on the packets with the specific UDP port.
3709 * The port identifier of the Ethernet device.
3711 * UDP tunneling configuration.
3714 * - (0) if successful.
3715 * - (-ENODEV) if port identifier is invalid.
3716 * - (-EIO) if device is removed.
3717 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3720 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
3721 struct rte_eth_udp_tunnel *tunnel_udp);
3724 * Check whether the filter type is supported on an Ethernet device.
3725 * All the supported filter types are defined in 'rte_eth_ctrl.h'.
3728 * The port identifier of the Ethernet device.
3729 * @param filter_type
3732 * - (0) if successful.
3733 * - (-ENOTSUP) if hardware doesn't support this filter type.
3734 * - (-ENODEV) if *port_id* invalid.
3735 * - (-EIO) if device is removed.
3738 int rte_eth_dev_filter_supported(uint16_t port_id,
3739 enum rte_filter_type filter_type);
3742 * Take operations to assigned filter type on an Ethernet device.
3743 * All the supported operations and filter types are defined in 'rte_eth_ctrl.h'.
3746 * The port identifier of the Ethernet device.
3747 * @param filter_type
3750 * Type of operation.
3752 * A pointer to arguments defined specifically for the operation.
3754 * - (0) if successful.
3755 * - (-ENOTSUP) if hardware doesn't support.
3756 * - (-ENODEV) if *port_id* invalid.
3757 * - (-EIO) if device is removed.
3758 * - others depends on the specific operations implementation.
3761 int rte_eth_dev_filter_ctrl(uint16_t port_id, enum rte_filter_type filter_type,
3762 enum rte_filter_op filter_op, void *arg);
3765 * Get DCB information on an Ethernet device.
3768 * The port identifier of the Ethernet device.
3772 * - (0) if successful.
3773 * - (-ENODEV) if port identifier is invalid.
3774 * - (-EIO) if device is removed.
3775 * - (-ENOTSUP) if hardware doesn't support.
3777 int rte_eth_dev_get_dcb_info(uint16_t port_id,
3778 struct rte_eth_dcb_info *dcb_info);
3780 struct rte_eth_rxtx_callback;
3783 * Add a callback to be called on packet RX on a given port and queue.
3785 * This API configures a function to be called for each burst of
3786 * packets received on a given NIC port queue. The return value is a pointer
3787 * that can be used to later remove the callback using
3788 * rte_eth_remove_rx_callback().
3790 * Multiple functions are called in the order that they are added.
3793 * The port identifier of the Ethernet device.
3795 * The queue on the Ethernet device on which the callback is to be added.
3797 * The callback function
3799 * A generic pointer parameter which will be passed to each invocation of the
3800 * callback function on this port and queue.
3804 * On success, a pointer value which can later be used to remove the callback.
3806 const struct rte_eth_rxtx_callback *
3807 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
3808 rte_rx_callback_fn fn, void *user_param);
3811 * Add a callback that must be called first on packet RX on a given port
3814 * This API configures a first function to be called for each burst of
3815 * packets received on a given NIC port queue. The return value is a pointer
3816 * that can be used to later remove the callback using
3817 * rte_eth_remove_rx_callback().
3819 * Multiple functions are called in the order that they are added.
3822 * The port identifier of the Ethernet device.
3824 * The queue on the Ethernet device on which the callback is to be added.
3826 * The callback function
3828 * A generic pointer parameter which will be passed to each invocation of the
3829 * callback function on this port and queue.
3833 * On success, a pointer value which can later be used to remove the callback.
3835 const struct rte_eth_rxtx_callback *
3836 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
3837 rte_rx_callback_fn fn, void *user_param);
3840 * Add a callback to be called on packet TX on a given port and queue.
3842 * This API configures a function to be called for each burst of
3843 * packets sent on a given NIC port queue. The return value is a pointer
3844 * that can be used to later remove the callback using
3845 * rte_eth_remove_tx_callback().
3847 * Multiple functions are called in the order that they are added.
3850 * The port identifier of the Ethernet device.
3852 * The queue on the Ethernet device on which the callback is to be added.
3854 * The callback function
3856 * A generic pointer parameter which will be passed to each invocation of the
3857 * callback function on this port and queue.
3861 * On success, a pointer value which can later be used to remove the callback.
3863 const struct rte_eth_rxtx_callback *
3864 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
3865 rte_tx_callback_fn fn, void *user_param);
3868 * Remove an RX packet callback from a given port and queue.
3870 * This function is used to removed callbacks that were added to a NIC port
3871 * queue using rte_eth_add_rx_callback().
3873 * Note: the callback is removed from the callback list but it isn't freed
3874 * since the it may still be in use. The memory for the callback can be
3875 * subsequently freed back by the application by calling rte_free():
3877 * - Immediately - if the port is stopped, or the user knows that no
3878 * callbacks are in flight e.g. if called from the thread doing RX/TX
3881 * - After a short delay - where the delay is sufficient to allow any
3882 * in-flight callbacks to complete.
3885 * The port identifier of the Ethernet device.
3887 * The queue on the Ethernet device from which the callback is to be removed.
3889 * User supplied callback created via rte_eth_add_rx_callback().
3892 * - 0: Success. Callback was removed.
3893 * - -ENOTSUP: Callback support is not available.
3894 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3895 * is NULL or not found for the port/queue.
3897 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
3898 const struct rte_eth_rxtx_callback *user_cb);
3901 * Remove a TX packet callback from a given port and queue.
3903 * This function is used to removed callbacks that were added to a NIC port
3904 * queue using rte_eth_add_tx_callback().
3906 * Note: the callback is removed from the callback list but it isn't freed
3907 * since the it may still be in use. The memory for the callback can be
3908 * subsequently freed back by the application by calling rte_free():
3910 * - Immediately - if the port is stopped, or the user knows that no
3911 * callbacks are in flight e.g. if called from the thread doing RX/TX
3914 * - After a short delay - where the delay is sufficient to allow any
3915 * in-flight callbacks to complete.
3918 * The port identifier of the Ethernet device.
3920 * The queue on the Ethernet device from which the callback is to be removed.
3922 * User supplied callback created via rte_eth_add_tx_callback().
3925 * - 0: Success. Callback was removed.
3926 * - -ENOTSUP: Callback support is not available.
3927 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3928 * is NULL or not found for the port/queue.
3930 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
3931 const struct rte_eth_rxtx_callback *user_cb);
3934 * Retrieve information about given port's RX queue.
3937 * The port identifier of the Ethernet device.
3939 * The RX queue on the Ethernet device for which information
3940 * will be retrieved.
3942 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
3943 * the information of the Ethernet device.
3947 * - -ENOTSUP: routine is not supported by the device PMD.
3948 * - -EINVAL: The port_id or the queue_id is out of range, or the queue
3951 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3952 struct rte_eth_rxq_info *qinfo);
3955 * Retrieve information about given port's TX queue.
3958 * The port identifier of the Ethernet device.
3960 * The TX queue on the Ethernet device for which information
3961 * will be retrieved.
3963 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
3964 * the information of the Ethernet device.
3968 * - -ENOTSUP: routine is not supported by the device PMD.
3969 * - -EINVAL: The port_id or the queue_id is out of range, or the queue
3972 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3973 struct rte_eth_txq_info *qinfo);
3976 * Retrieve information about the Rx packet burst mode.
3979 * The port identifier of the Ethernet device.
3981 * The Rx queue on the Ethernet device for which information
3982 * will be retrieved.
3984 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
3985 * with the information of the packet burst mode.
3989 * - -ENOTSUP: routine is not supported by the device PMD.
3990 * - -EINVAL: The port_id or the queue_id is out of range.
3993 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
3994 struct rte_eth_burst_mode *mode);
3997 * Retrieve information about the Tx packet burst mode.
4000 * The port identifier of the Ethernet device.
4002 * The Tx queue on the Ethernet device for which information
4003 * will be retrieved.
4005 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4006 * with the information of the packet burst mode.
4010 * - -ENOTSUP: routine is not supported by the device PMD.
4011 * - -EINVAL: The port_id or the queue_id is out of range.
4014 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4015 struct rte_eth_burst_mode *mode);
4018 * Retrieve device registers and register attributes (number of registers and
4022 * The port identifier of the Ethernet device.
4024 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
4025 * NULL the function fills in the width and length fields. If non-NULL
4026 * the registers are put into the buffer pointed at by the data field.
4028 * - (0) if successful.
4029 * - (-ENOTSUP) if hardware doesn't support.
4030 * - (-ENODEV) if *port_id* invalid.
4031 * - (-EIO) if device is removed.
4032 * - others depends on the specific operations implementation.
4034 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
4037 * Retrieve size of device EEPROM
4040 * The port identifier of the Ethernet device.
4042 * - (>=0) EEPROM size if successful.
4043 * - (-ENOTSUP) if hardware doesn't support.
4044 * - (-ENODEV) if *port_id* invalid.
4045 * - (-EIO) if device is removed.
4046 * - others depends on the specific operations implementation.
4048 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
4051 * Retrieve EEPROM and EEPROM attribute
4054 * The port identifier of the Ethernet device.
4056 * The template includes buffer for return EEPROM data and
4057 * EEPROM attributes to be filled.
4059 * - (0) if successful.
4060 * - (-ENOTSUP) if hardware doesn't support.
4061 * - (-ENODEV) if *port_id* invalid.
4062 * - (-EIO) if device is removed.
4063 * - others depends on the specific operations implementation.
4065 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4068 * Program EEPROM with provided data
4071 * The port identifier of the Ethernet device.
4073 * The template includes EEPROM data for programming and
4074 * EEPROM attributes to be filled
4076 * - (0) if successful.
4077 * - (-ENOTSUP) if hardware doesn't support.
4078 * - (-ENODEV) if *port_id* invalid.
4079 * - (-EIO) if device is removed.
4080 * - others depends on the specific operations implementation.
4082 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4086 * @b EXPERIMENTAL: this API may change without prior notice.
4088 * Retrieve the type and size of plugin module EEPROM
4091 * The port identifier of the Ethernet device.
4093 * The type and size of plugin module EEPROM.
4095 * - (0) if successful.
4096 * - (-ENOTSUP) if hardware doesn't support.
4097 * - (-ENODEV) if *port_id* invalid.
4098 * - (-EIO) if device is removed.
4099 * - others depends on the specific operations implementation.
4103 rte_eth_dev_get_module_info(uint16_t port_id,
4104 struct rte_eth_dev_module_info *modinfo);
4108 * @b EXPERIMENTAL: this API may change without prior notice.
4110 * Retrieve the data of plugin module EEPROM
4113 * The port identifier of the Ethernet device.
4115 * The template includes the plugin module EEPROM attributes, and the
4116 * buffer for return plugin module EEPROM data.
4118 * - (0) if successful.
4119 * - (-ENOTSUP) if hardware doesn't support.
4120 * - (-ENODEV) if *port_id* invalid.
4121 * - (-EIO) if device is removed.
4122 * - others depends on the specific operations implementation.
4126 rte_eth_dev_get_module_eeprom(uint16_t port_id,
4127 struct rte_dev_eeprom_info *info);
4130 * Set the list of multicast addresses to filter on an Ethernet device.
4133 * The port identifier of the Ethernet device.
4134 * @param mc_addr_set
4135 * The array of multicast addresses to set. Equal to NULL when the function
4136 * is invoked to flush the set of filtered addresses.
4138 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
4139 * when the function is invoked to flush the set of filtered addresses.
4141 * - (0) if successful.
4142 * - (-ENODEV) if *port_id* invalid.
4143 * - (-EIO) if device is removed.
4144 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
4145 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
4147 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
4148 struct rte_ether_addr *mc_addr_set,
4149 uint32_t nb_mc_addr);
4152 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
4155 * The port identifier of the Ethernet device.
4159 * - -ENODEV: The port ID is invalid.
4160 * - -EIO: if device is removed.
4161 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4163 int rte_eth_timesync_enable(uint16_t port_id);
4166 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
4169 * The port identifier of the Ethernet device.
4173 * - -ENODEV: The port ID is invalid.
4174 * - -EIO: if device is removed.
4175 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4177 int rte_eth_timesync_disable(uint16_t port_id);
4180 * Read an IEEE1588/802.1AS RX timestamp from an Ethernet device.
4183 * The port identifier of the Ethernet device.
4185 * Pointer to the timestamp struct.
4187 * Device specific flags. Used to pass the RX timesync register index to
4188 * i40e. Unused in igb/ixgbe, pass 0 instead.
4192 * - -EINVAL: No timestamp is available.
4193 * - -ENODEV: The port ID is invalid.
4194 * - -EIO: if device is removed.
4195 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4197 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
4198 struct timespec *timestamp, uint32_t flags);
4201 * Read an IEEE1588/802.1AS TX timestamp from an Ethernet device.
4204 * The port identifier of the Ethernet device.
4206 * Pointer to the timestamp struct.
4210 * - -EINVAL: No timestamp is available.
4211 * - -ENODEV: The port ID is invalid.
4212 * - -EIO: if device is removed.
4213 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4215 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
4216 struct timespec *timestamp);
4219 * Adjust the timesync clock on an Ethernet device.
4221 * This is usually used in conjunction with other Ethdev timesync functions to
4222 * synchronize the device time using the IEEE1588/802.1AS protocol.
4225 * The port identifier of the Ethernet device.
4227 * The adjustment in nanoseconds.
4231 * - -ENODEV: The port ID is invalid.
4232 * - -EIO: if device is removed.
4233 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4235 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
4238 * Read the time from the timesync clock on an Ethernet device.
4240 * This is usually used in conjunction with other Ethdev timesync functions to
4241 * synchronize the device time using the IEEE1588/802.1AS protocol.
4244 * The port identifier of the Ethernet device.
4246 * Pointer to the timespec struct that holds the time.
4251 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
4254 * Set the time of the timesync clock on an Ethernet device.
4256 * This is usually used in conjunction with other Ethdev timesync functions to
4257 * synchronize the device time using the IEEE1588/802.1AS protocol.
4260 * The port identifier of the Ethernet device.
4262 * Pointer to the timespec struct that holds the time.
4266 * - -EINVAL: No timestamp is available.
4267 * - -ENODEV: The port ID is invalid.
4268 * - -EIO: if device is removed.
4269 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4271 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
4275 * @b EXPERIMENTAL: this API may change without prior notice.
4277 * Read the current clock counter of an Ethernet device
4279 * This returns the current raw clock value of an Ethernet device. It is
4280 * a raw amount of ticks, with no given time reference.
4281 * The value returned here is from the same clock than the one
4282 * filling timestamp field of Rx packets when using hardware timestamp
4283 * offload. Therefore it can be used to compute a precise conversion of
4284 * the device clock to the real time.
4286 * E.g, a simple heuristic to derivate the frequency would be:
4287 * uint64_t start, end;
4288 * rte_eth_read_clock(port, start);
4289 * rte_delay_ms(100);
4290 * rte_eth_read_clock(port, end);
4291 * double freq = (end - start) * 10;
4293 * Compute a common reference with:
4294 * uint64_t base_time_sec = current_time();
4295 * uint64_t base_clock;
4296 * rte_eth_read_clock(port, base_clock);
4298 * Then, convert the raw mbuf timestamp with:
4299 * base_time_sec + (double)(mbuf->timestamp - base_clock) / freq;
4301 * This simple example will not provide a very good accuracy. One must
4302 * at least measure multiple times the frequency and do a regression.
4303 * To avoid deviation from the system time, the common reference can
4304 * be repeated from time to time. The integer division can also be
4305 * converted by a multiplication and a shift for better performance.
4308 * The port identifier of the Ethernet device.
4310 * Pointer to the uint64_t that holds the raw clock value.
4314 * - -ENODEV: The port ID is invalid.
4315 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4319 rte_eth_read_clock(uint16_t port_id, uint64_t *clock);
4322 * Config l2 tunnel ether type of an Ethernet device for filtering specific
4323 * tunnel packets by ether type.
4326 * The port identifier of the Ethernet device.
4328 * l2 tunnel configuration.
4331 * - (0) if successful.
4332 * - (-ENODEV) if port identifier is invalid.
4333 * - (-EIO) if device is removed.
4334 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4337 rte_eth_dev_l2_tunnel_eth_type_conf(uint16_t port_id,
4338 struct rte_eth_l2_tunnel_conf *l2_tunnel);
4341 * Enable/disable l2 tunnel offload functions. Include,
4342 * 1, The ability of parsing a type of l2 tunnel of an Ethernet device.
4343 * Filtering, forwarding and offloading this type of tunnel packets depend on
4345 * 2, Stripping the l2 tunnel tag.
4346 * 3, Insertion of the l2 tunnel tag.
4347 * 4, Forwarding the packets based on the l2 tunnel tag.
4350 * The port identifier of the Ethernet device.
4352 * l2 tunnel parameters.
4354 * Indicate the offload function.
4356 * Enable or disable this function.
4359 * - (0) if successful.
4360 * - (-ENODEV) if port identifier is invalid.
4361 * - (-EIO) if device is removed.
4362 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4365 rte_eth_dev_l2_tunnel_offload_set(uint16_t port_id,
4366 struct rte_eth_l2_tunnel_conf *l2_tunnel,
4371 * Get the port id from device name. The device name should be specified
4373 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0
4374 * - SoC device name, for example- fsl-gmac0
4375 * - vdev dpdk name, for example- net_[pcap0|null0|tap0]
4378 * pci address or name of the device
4380 * pointer to port identifier of the device
4382 * - (0) if successful and port_id is filled.
4383 * - (-ENODEV or -EINVAL) on failure.
4386 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
4389 * Get the device name from port id. The device name is specified as below:
4390 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0
4391 * - SoC device name, for example- fsl-gmac0
4392 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0]
4395 * Port identifier of the device.
4397 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
4399 * - (0) if successful.
4400 * - (-EINVAL) on failure.
4403 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
4406 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
4407 * the ethernet device information, otherwise adjust them to boundaries.
4410 * The port identifier of the Ethernet device.
4412 * A pointer to a uint16_t where the number of receive
4413 * descriptors stored.
4415 * A pointer to a uint16_t where the number of transmit
4416 * descriptors stored.
4418 * - (0) if successful.
4419 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
4421 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
4422 uint16_t *nb_rx_desc,
4423 uint16_t *nb_tx_desc);
4426 * Test if a port supports specific mempool ops.
4429 * Port identifier of the Ethernet device.
4431 * The name of the pool operations to test.
4433 * - 0: best mempool ops choice for this port.
4434 * - 1: mempool ops are supported for this port.
4435 * - -ENOTSUP: mempool ops not supported for this port.
4436 * - -ENODEV: Invalid port Identifier.
4437 * - -EINVAL: Pool param is null.
4440 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
4443 * Get the security context for the Ethernet device.
4446 * Port identifier of the Ethernet device
4449 * - pointer to security context on success.
4452 rte_eth_dev_get_sec_ctx(uint16_t port_id);
4456 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4458 * Query the device hairpin capabilities.
4461 * The port identifier of the Ethernet device.
4463 * Pointer to a structure that will hold the hairpin capabilities.
4465 * - (0) if successful.
4466 * - (-ENOTSUP) if hardware doesn't support.
4469 int rte_eth_dev_hairpin_capability_get(uint16_t port_id,
4470 struct rte_eth_hairpin_cap *cap);
4472 #include <rte_ethdev_core.h>
4476 * Retrieve a burst of input packets from a receive queue of an Ethernet
4477 * device. The retrieved packets are stored in *rte_mbuf* structures whose
4478 * pointers are supplied in the *rx_pkts* array.
4480 * The rte_eth_rx_burst() function loops, parsing the RX ring of the
4481 * receive queue, up to *nb_pkts* packets, and for each completed RX
4482 * descriptor in the ring, it performs the following operations:
4484 * - Initialize the *rte_mbuf* data structure associated with the
4485 * RX descriptor according to the information provided by the NIC into
4486 * that RX descriptor.
4488 * - Store the *rte_mbuf* data structure into the next entry of the
4491 * - Replenish the RX descriptor with a new *rte_mbuf* buffer
4492 * allocated from the memory pool associated with the receive queue at
4493 * initialization time.
4495 * When retrieving an input packet that was scattered by the controller
4496 * into multiple receive descriptors, the rte_eth_rx_burst() function
4497 * appends the associated *rte_mbuf* buffers to the first buffer of the
4500 * The rte_eth_rx_burst() function returns the number of packets
4501 * actually retrieved, which is the number of *rte_mbuf* data structures
4502 * effectively supplied into the *rx_pkts* array.
4503 * A return value equal to *nb_pkts* indicates that the RX queue contained
4504 * at least *rx_pkts* packets, and this is likely to signify that other
4505 * received packets remain in the input queue. Applications implementing
4506 * a "retrieve as much received packets as possible" policy can check this
4507 * specific case and keep invoking the rte_eth_rx_burst() function until
4508 * a value less than *nb_pkts* is returned.
4510 * This receive method has the following advantages:
4512 * - It allows a run-to-completion network stack engine to retrieve and
4513 * to immediately process received packets in a fast burst-oriented
4514 * approach, avoiding the overhead of unnecessary intermediate packet
4515 * queue/dequeue operations.
4517 * - Conversely, it also allows an asynchronous-oriented processing
4518 * method to retrieve bursts of received packets and to immediately
4519 * queue them for further parallel processing by another logical core,
4520 * for instance. However, instead of having received packets being
4521 * individually queued by the driver, this approach allows the caller
4522 * of the rte_eth_rx_burst() function to queue a burst of retrieved
4523 * packets at a time and therefore dramatically reduce the cost of
4524 * enqueue/dequeue operations per packet.
4526 * - It allows the rte_eth_rx_burst() function of the driver to take
4527 * advantage of burst-oriented hardware features (CPU cache,
4528 * prefetch instructions, and so on) to minimize the number of CPU
4529 * cycles per packet.
4531 * To summarize, the proposed receive API enables many
4532 * burst-oriented optimizations in both synchronous and asynchronous
4533 * packet processing environments with no overhead in both cases.
4535 * The rte_eth_rx_burst() function does not provide any error
4536 * notification to avoid the corresponding overhead. As a hint, the
4537 * upper-level application might check the status of the device link once
4538 * being systematically returned a 0 value for a given number of tries.
4541 * The port identifier of the Ethernet device.
4543 * The index of the receive queue from which to retrieve input packets.
4544 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4545 * to rte_eth_dev_configure().
4547 * The address of an array of pointers to *rte_mbuf* structures that
4548 * must be large enough to store *nb_pkts* pointers in it.
4550 * The maximum number of packets to retrieve.
4552 * The number of packets actually retrieved, which is the number
4553 * of pointers to *rte_mbuf* structures effectively supplied to the
4556 static inline uint16_t
4557 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
4558 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
4560 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4563 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4564 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4565 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_pkt_burst, 0);
4567 if (queue_id >= dev->data->nb_rx_queues) {
4568 RTE_ETHDEV_LOG(ERR, "Invalid RX queue_id=%u\n", queue_id);
4572 nb_rx = (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id],
4575 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4576 if (unlikely(dev->post_rx_burst_cbs[queue_id] != NULL)) {
4577 struct rte_eth_rxtx_callback *cb =
4578 dev->post_rx_burst_cbs[queue_id];
4581 nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx,
4582 nb_pkts, cb->param);
4584 } while (cb != NULL);
4588 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx);
4593 * Get the number of used descriptors of a rx queue
4596 * The port identifier of the Ethernet device.
4598 * The queue id on the specific port.
4600 * The number of used descriptors in the specific queue, or:
4601 * (-EINVAL) if *port_id* or *queue_id* is invalid
4602 * (-ENOTSUP) if the device does not support this function
4605 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
4607 struct rte_eth_dev *dev;
4609 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
4610 dev = &rte_eth_devices[port_id];
4611 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_queue_count, -ENOTSUP);
4612 if (queue_id >= dev->data->nb_rx_queues)
4615 return (int)(*dev->rx_queue_count)(dev, queue_id);
4619 * Check if the DD bit of the specific RX descriptor in the queue has been set
4622 * The port identifier of the Ethernet device.
4624 * The queue id on the specific port.
4626 * The offset of the descriptor ID from tail.
4628 * - (1) if the specific DD bit is set.
4629 * - (0) if the specific DD bit is not set.
4630 * - (-ENODEV) if *port_id* invalid.
4631 * - (-ENOTSUP) if the device does not support this function
4635 rte_eth_rx_descriptor_done(uint16_t port_id, uint16_t queue_id, uint16_t offset)
4637 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4638 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4639 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_descriptor_done, -ENOTSUP);
4640 return (*dev->rx_descriptor_done)(dev->data->rx_queues[queue_id], offset);
4643 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
4644 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
4645 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
4648 * Check the status of a Rx descriptor in the queue
4650 * It should be called in a similar context than the Rx function:
4651 * - on a dataplane core
4652 * - not concurrently on the same queue
4654 * Since it's a dataplane function, no check is performed on port_id and
4655 * queue_id. The caller must therefore ensure that the port is enabled
4656 * and the queue is configured and running.
4658 * Note: accessing to a random descriptor in the ring may trigger cache
4659 * misses and have a performance impact.
4662 * A valid port identifier of the Ethernet device which.
4664 * A valid Rx queue identifier on this port.
4666 * The offset of the descriptor starting from tail (0 is the next
4667 * packet to be received by the driver).
4670 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
4672 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
4673 * not yet processed by the driver (i.e. in the receive queue).
4674 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
4675 * the driver and not yet returned to hw, or reserved by the hw.
4676 * - (-EINVAL) bad descriptor offset.
4677 * - (-ENOTSUP) if the device does not support this function.
4678 * - (-ENODEV) bad port or queue (only if compiled with debug).
4681 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
4684 struct rte_eth_dev *dev;
4687 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4688 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4690 dev = &rte_eth_devices[port_id];
4691 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4692 if (queue_id >= dev->data->nb_rx_queues)
4695 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_descriptor_status, -ENOTSUP);
4696 rxq = dev->data->rx_queues[queue_id];
4698 return (*dev->rx_descriptor_status)(rxq, offset);
4701 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
4702 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
4703 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
4706 * Check the status of a Tx descriptor in the queue.
4708 * It should be called in a similar context than the Tx function:
4709 * - on a dataplane core
4710 * - not concurrently on the same queue
4712 * Since it's a dataplane function, no check is performed on port_id and
4713 * queue_id. The caller must therefore ensure that the port is enabled
4714 * and the queue is configured and running.
4716 * Note: accessing to a random descriptor in the ring may trigger cache
4717 * misses and have a performance impact.
4720 * A valid port identifier of the Ethernet device which.
4722 * A valid Tx queue identifier on this port.
4724 * The offset of the descriptor starting from tail (0 is the place where
4725 * the next packet will be send).
4728 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
4729 * in the transmit queue.
4730 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
4731 * be reused by the driver.
4732 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
4733 * driver or the hardware.
4734 * - (-EINVAL) bad descriptor offset.
4735 * - (-ENOTSUP) if the device does not support this function.
4736 * - (-ENODEV) bad port or queue (only if compiled with debug).
4738 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
4739 uint16_t queue_id, uint16_t offset)
4741 struct rte_eth_dev *dev;
4744 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4745 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4747 dev = &rte_eth_devices[port_id];
4748 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4749 if (queue_id >= dev->data->nb_tx_queues)
4752 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_descriptor_status, -ENOTSUP);
4753 txq = dev->data->tx_queues[queue_id];
4755 return (*dev->tx_descriptor_status)(txq, offset);
4759 * Send a burst of output packets on a transmit queue of an Ethernet device.
4761 * The rte_eth_tx_burst() function is invoked to transmit output packets
4762 * on the output queue *queue_id* of the Ethernet device designated by its
4764 * The *nb_pkts* parameter is the number of packets to send which are
4765 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4766 * allocated from a pool created with rte_pktmbuf_pool_create().
4767 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
4768 * up to the number of transmit descriptors available in the TX ring of the
4770 * For each packet to send, the rte_eth_tx_burst() function performs
4771 * the following operations:
4773 * - Pick up the next available descriptor in the transmit ring.
4775 * - Free the network buffer previously sent with that descriptor, if any.
4777 * - Initialize the transmit descriptor with the information provided
4778 * in the *rte_mbuf data structure.
4780 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
4781 * the rte_eth_tx_burst() function uses several transmit descriptors
4784 * The rte_eth_tx_burst() function returns the number of packets it
4785 * actually sent. A return value equal to *nb_pkts* means that all packets
4786 * have been sent, and this is likely to signify that other output packets
4787 * could be immediately transmitted again. Applications that implement a
4788 * "send as many packets to transmit as possible" policy can check this
4789 * specific case and keep invoking the rte_eth_tx_burst() function until
4790 * a value less than *nb_pkts* is returned.
4792 * It is the responsibility of the rte_eth_tx_burst() function to
4793 * transparently free the memory buffers of packets previously sent.
4794 * This feature is driven by the *tx_free_thresh* value supplied to the
4795 * rte_eth_dev_configure() function at device configuration time.
4796 * When the number of free TX descriptors drops below this threshold, the
4797 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
4798 * of those packets whose transmission was effectively completed.
4800 * If the PMD is DEV_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
4801 * invoke this function concurrently on the same tx queue without SW lock.
4802 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads
4804 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
4808 * The port identifier of the Ethernet device.
4810 * The index of the transmit queue through which output packets must be
4812 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4813 * to rte_eth_dev_configure().
4815 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4816 * which contain the output packets.
4818 * The maximum number of packets to transmit.
4820 * The number of output packets actually stored in transmit descriptors of
4821 * the transmit ring. The return value can be less than the value of the
4822 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
4824 static inline uint16_t
4825 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
4826 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4828 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4830 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4831 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4832 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_pkt_burst, 0);
4834 if (queue_id >= dev->data->nb_tx_queues) {
4835 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4840 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4841 struct rte_eth_rxtx_callback *cb = dev->pre_tx_burst_cbs[queue_id];
4843 if (unlikely(cb != NULL)) {
4845 nb_pkts = cb->fn.tx(port_id, queue_id, tx_pkts, nb_pkts,
4848 } while (cb != NULL);
4852 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts,
4854 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts);
4858 * Process a burst of output packets on a transmit queue of an Ethernet device.
4860 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
4861 * transmitted on the output queue *queue_id* of the Ethernet device designated
4863 * The *nb_pkts* parameter is the number of packets to be prepared which are
4864 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4865 * allocated from a pool created with rte_pktmbuf_pool_create().
4866 * For each packet to send, the rte_eth_tx_prepare() function performs
4867 * the following operations:
4869 * - Check if packet meets devices requirements for tx offloads.
4871 * - Check limitations about number of segments.
4873 * - Check additional requirements when debug is enabled.
4875 * - Update and/or reset required checksums when tx offload is set for packet.
4877 * Since this function can modify packet data, provided mbufs must be safely
4878 * writable (e.g. modified data cannot be in shared segment).
4880 * The rte_eth_tx_prepare() function returns the number of packets ready to be
4881 * sent. A return value equal to *nb_pkts* means that all packets are valid and
4882 * ready to be sent, otherwise stops processing on the first invalid packet and
4883 * leaves the rest packets untouched.
4885 * When this functionality is not implemented in the driver, all packets are
4886 * are returned untouched.
4889 * The port identifier of the Ethernet device.
4890 * The value must be a valid port id.
4892 * The index of the transmit queue through which output packets must be
4894 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4895 * to rte_eth_dev_configure().
4897 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4898 * which contain the output packets.
4900 * The maximum number of packets to process.
4902 * The number of packets correct and ready to be sent. The return value can be
4903 * less than the value of the *tx_pkts* parameter when some packet doesn't
4904 * meet devices requirements with rte_errno set appropriately:
4905 * - EINVAL: offload flags are not correctly set
4906 * - ENOTSUP: the offload feature is not supported by the hardware
4910 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
4912 static inline uint16_t
4913 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
4914 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4916 struct rte_eth_dev *dev;
4918 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4919 if (!rte_eth_dev_is_valid_port(port_id)) {
4920 RTE_ETHDEV_LOG(ERR, "Invalid TX port_id=%u\n", port_id);
4926 dev = &rte_eth_devices[port_id];
4928 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4929 if (queue_id >= dev->data->nb_tx_queues) {
4930 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4936 if (!dev->tx_pkt_prepare)
4939 return (*dev->tx_pkt_prepare)(dev->data->tx_queues[queue_id],
4946 * Native NOOP operation for compilation targets which doesn't require any
4947 * preparations steps, and functional NOOP may introduce unnecessary performance
4950 * Generally this is not a good idea to turn it on globally and didn't should
4951 * be used if behavior of tx_preparation can change.
4954 static inline uint16_t
4955 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
4956 __rte_unused uint16_t queue_id,
4957 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4965 * Send any packets queued up for transmission on a port and HW queue
4967 * This causes an explicit flush of packets previously buffered via the
4968 * rte_eth_tx_buffer() function. It returns the number of packets successfully
4969 * sent to the NIC, and calls the error callback for any unsent packets. Unless
4970 * explicitly set up otherwise, the default callback simply frees the unsent
4971 * packets back to the owning mempool.
4974 * The port identifier of the Ethernet device.
4976 * The index of the transmit queue through which output packets must be
4978 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4979 * to rte_eth_dev_configure().
4981 * Buffer of packets to be transmit.
4983 * The number of packets successfully sent to the Ethernet device. The error
4984 * callback is called for any packets which could not be sent.
4986 static inline uint16_t
4987 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
4988 struct rte_eth_dev_tx_buffer *buffer)
4991 uint16_t to_send = buffer->length;
4996 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
5000 /* All packets sent, or to be dealt with by callback below */
5001 if (unlikely(sent != to_send))
5002 buffer->error_callback(&buffer->pkts[sent],
5003 (uint16_t)(to_send - sent),
5004 buffer->error_userdata);
5010 * Buffer a single packet for future transmission on a port and queue
5012 * This function takes a single mbuf/packet and buffers it for later
5013 * transmission on the particular port and queue specified. Once the buffer is
5014 * full of packets, an attempt will be made to transmit all the buffered
5015 * packets. In case of error, where not all packets can be transmitted, a
5016 * callback is called with the unsent packets as a parameter. If no callback
5017 * is explicitly set up, the unsent packets are just freed back to the owning
5018 * mempool. The function returns the number of packets actually sent i.e.
5019 * 0 if no buffer flush occurred, otherwise the number of packets successfully
5023 * The port identifier of the Ethernet device.
5025 * The index of the transmit queue through which output packets must be
5027 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5028 * to rte_eth_dev_configure().
5030 * Buffer used to collect packets to be sent.
5032 * Pointer to the packet mbuf to be sent.
5034 * 0 = packet has been buffered for later transmission
5035 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
5036 * causing N packets to be sent, and the error callback to be called for
5039 static __rte_always_inline uint16_t
5040 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
5041 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
5043 buffer->pkts[buffer->length++] = tx_pkt;
5044 if (buffer->length < buffer->size)
5047 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
5054 #endif /* _RTE_ETHDEV_H_ */