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 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 /* Alias RTE_LIBRTE_ETHDEV_DEBUG for backward compatibility. */
152 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
153 #define RTE_ETHDEV_DEBUG_RX
154 #define RTE_ETHDEV_DEBUG_TX
157 #include <rte_compat.h>
159 #include <rte_interrupts.h>
161 #include <rte_devargs.h>
162 #include <rte_bitops.h>
163 #include <rte_errno.h>
164 #include <rte_common.h>
165 #include <rte_config.h>
166 #include <rte_ether.h>
167 #include <rte_power_intrinsics.h>
169 #include "rte_ethdev_trace_fp.h"
170 #include "rte_dev_info.h"
172 extern int rte_eth_dev_logtype;
174 #define RTE_ETHDEV_LOG(level, ...) \
175 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__)
180 * Initializes a device iterator.
182 * This iterator allows accessing a list of devices matching some devargs.
185 * Device iterator handle initialized by the function.
186 * The fields bus_str and cls_str might be dynamically allocated,
187 * and could be freed by calling rte_eth_iterator_cleanup().
190 * Device description string.
193 * 0 on successful initialization, negative otherwise.
195 int rte_eth_iterator_init(struct rte_dev_iterator *iter, const char *devargs);
198 * Iterates on devices with devargs filter.
199 * The ownership is not checked.
201 * The next port ID is returned, and the iterator is updated.
204 * Device iterator handle initialized by rte_eth_iterator_init().
205 * Some fields bus_str and cls_str might be freed when no more port is found,
206 * by calling rte_eth_iterator_cleanup().
209 * A port ID if found, RTE_MAX_ETHPORTS otherwise.
211 uint16_t rte_eth_iterator_next(struct rte_dev_iterator *iter);
214 * Free some allocated fields of the iterator.
216 * This function is automatically called by rte_eth_iterator_next()
217 * on the last iteration (i.e. when no more matching port is found).
219 * It is safe to call this function twice; it will do nothing more.
222 * Device iterator handle initialized by rte_eth_iterator_init().
223 * The fields bus_str and cls_str are freed if needed.
225 void rte_eth_iterator_cleanup(struct rte_dev_iterator *iter);
228 * Macro to iterate over all ethdev ports matching some devargs.
230 * If a break is done before the end of the loop,
231 * the function rte_eth_iterator_cleanup() must be called.
234 * Iterated port ID of type uint16_t.
236 * Device parameters input as string of type char*.
238 * Iterator handle of type struct rte_dev_iterator, used internally.
240 #define RTE_ETH_FOREACH_MATCHING_DEV(id, devargs, iter) \
241 for (rte_eth_iterator_init(iter, devargs), \
242 id = rte_eth_iterator_next(iter); \
243 id != RTE_MAX_ETHPORTS; \
244 id = rte_eth_iterator_next(iter))
247 * A structure used to retrieve statistics for an Ethernet port.
248 * Not all statistics fields in struct rte_eth_stats are supported
249 * by any type of network interface card (NIC). If any statistics
250 * field is not supported, its value is 0.
251 * All byte-related statistics do not include Ethernet FCS regardless
252 * of whether these bytes have been delivered to the application
253 * (see RTE_ETH_RX_OFFLOAD_KEEP_CRC).
255 struct rte_eth_stats {
256 uint64_t ipackets; /**< Total number of successfully received packets. */
257 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
258 uint64_t ibytes; /**< Total number of successfully received bytes. */
259 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
261 * Total of Rx packets dropped by the HW,
262 * because there are no available buffer (i.e. Rx queues are full).
265 uint64_t ierrors; /**< Total number of erroneous received packets. */
266 uint64_t oerrors; /**< Total number of failed transmitted packets. */
267 uint64_t rx_nombuf; /**< Total number of Rx mbuf allocation failures. */
268 /* Queue stats are limited to max 256 queues */
269 /** Total number of queue Rx packets. */
270 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
271 /** Total number of queue Tx packets. */
272 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
273 /** Total number of successfully received queue bytes. */
274 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
275 /** Total number of successfully transmitted queue bytes. */
276 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
277 /** Total number of queue packets received that are dropped. */
278 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS];
281 /**@{@name Link speed capabilities
282 * Device supported speeds bitmap flags
284 #define RTE_ETH_LINK_SPEED_AUTONEG 0 /**< Autonegotiate (all speeds) */
285 #define ETH_LINK_SPEED_AUTONEG RTE_ETH_LINK_SPEED_AUTONEG
286 #define RTE_ETH_LINK_SPEED_FIXED RTE_BIT32(0) /**< Disable autoneg (fixed speed) */
287 #define ETH_LINK_SPEED_FIXED RTE_ETH_LINK_SPEED_FIXED
288 #define RTE_ETH_LINK_SPEED_10M_HD RTE_BIT32(1) /**< 10 Mbps half-duplex */
289 #define ETH_LINK_SPEED_10M_HD RTE_ETH_LINK_SPEED_10M_HD
290 #define RTE_ETH_LINK_SPEED_10M RTE_BIT32(2) /**< 10 Mbps full-duplex */
291 #define ETH_LINK_SPEED_10M RTE_ETH_LINK_SPEED_10M
292 #define RTE_ETH_LINK_SPEED_100M_HD RTE_BIT32(3) /**< 100 Mbps half-duplex */
293 #define ETH_LINK_SPEED_100M_HD RTE_ETH_LINK_SPEED_100M_HD
294 #define RTE_ETH_LINK_SPEED_100M RTE_BIT32(4) /**< 100 Mbps full-duplex */
295 #define ETH_LINK_SPEED_100M RTE_ETH_LINK_SPEED_100M
296 #define RTE_ETH_LINK_SPEED_1G RTE_BIT32(5) /**< 1 Gbps */
297 #define ETH_LINK_SPEED_1G RTE_ETH_LINK_SPEED_1G
298 #define RTE_ETH_LINK_SPEED_2_5G RTE_BIT32(6) /**< 2.5 Gbps */
299 #define ETH_LINK_SPEED_2_5G RTE_ETH_LINK_SPEED_2_5G
300 #define RTE_ETH_LINK_SPEED_5G RTE_BIT32(7) /**< 5 Gbps */
301 #define ETH_LINK_SPEED_5G RTE_ETH_LINK_SPEED_5G
302 #define RTE_ETH_LINK_SPEED_10G RTE_BIT32(8) /**< 10 Gbps */
303 #define ETH_LINK_SPEED_10G RTE_ETH_LINK_SPEED_10G
304 #define RTE_ETH_LINK_SPEED_20G RTE_BIT32(9) /**< 20 Gbps */
305 #define ETH_LINK_SPEED_20G RTE_ETH_LINK_SPEED_20G
306 #define RTE_ETH_LINK_SPEED_25G RTE_BIT32(10) /**< 25 Gbps */
307 #define ETH_LINK_SPEED_25G RTE_ETH_LINK_SPEED_25G
308 #define RTE_ETH_LINK_SPEED_40G RTE_BIT32(11) /**< 40 Gbps */
309 #define ETH_LINK_SPEED_40G RTE_ETH_LINK_SPEED_40G
310 #define RTE_ETH_LINK_SPEED_50G RTE_BIT32(12) /**< 50 Gbps */
311 #define ETH_LINK_SPEED_50G RTE_ETH_LINK_SPEED_50G
312 #define RTE_ETH_LINK_SPEED_56G RTE_BIT32(13) /**< 56 Gbps */
313 #define ETH_LINK_SPEED_56G RTE_ETH_LINK_SPEED_56G
314 #define RTE_ETH_LINK_SPEED_100G RTE_BIT32(14) /**< 100 Gbps */
315 #define ETH_LINK_SPEED_100G RTE_ETH_LINK_SPEED_100G
316 #define RTE_ETH_LINK_SPEED_200G RTE_BIT32(15) /**< 200 Gbps */
317 #define ETH_LINK_SPEED_200G RTE_ETH_LINK_SPEED_200G
320 /**@{@name Link speed
321 * Ethernet numeric link speeds in Mbps
323 #define RTE_ETH_SPEED_NUM_NONE 0 /**< Not defined */
324 #define ETH_SPEED_NUM_NONE RTE_ETH_SPEED_NUM_NONE
325 #define RTE_ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
326 #define ETH_SPEED_NUM_10M RTE_ETH_SPEED_NUM_10M
327 #define RTE_ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
328 #define ETH_SPEED_NUM_100M RTE_ETH_SPEED_NUM_100M
329 #define RTE_ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
330 #define ETH_SPEED_NUM_1G RTE_ETH_SPEED_NUM_1G
331 #define RTE_ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
332 #define ETH_SPEED_NUM_2_5G RTE_ETH_SPEED_NUM_2_5G
333 #define RTE_ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
334 #define ETH_SPEED_NUM_5G RTE_ETH_SPEED_NUM_5G
335 #define RTE_ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
336 #define ETH_SPEED_NUM_10G RTE_ETH_SPEED_NUM_10G
337 #define RTE_ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
338 #define ETH_SPEED_NUM_20G RTE_ETH_SPEED_NUM_20G
339 #define RTE_ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
340 #define ETH_SPEED_NUM_25G RTE_ETH_SPEED_NUM_25G
341 #define RTE_ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
342 #define ETH_SPEED_NUM_40G RTE_ETH_SPEED_NUM_40G
343 #define RTE_ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
344 #define ETH_SPEED_NUM_50G RTE_ETH_SPEED_NUM_50G
345 #define RTE_ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
346 #define ETH_SPEED_NUM_56G RTE_ETH_SPEED_NUM_56G
347 #define RTE_ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
348 #define ETH_SPEED_NUM_100G RTE_ETH_SPEED_NUM_100G
349 #define RTE_ETH_SPEED_NUM_200G 200000 /**< 200 Gbps */
350 #define ETH_SPEED_NUM_200G RTE_ETH_SPEED_NUM_200G
351 #define RTE_ETH_SPEED_NUM_UNKNOWN UINT32_MAX /**< Unknown */
352 #define ETH_SPEED_NUM_UNKNOWN RTE_ETH_SPEED_NUM_UNKNOWN
356 * A structure used to retrieve link-level information of an Ethernet port.
359 struct rte_eth_link {
360 uint32_t link_speed; /**< RTE_ETH_SPEED_NUM_ */
361 uint16_t link_duplex : 1; /**< RTE_ETH_LINK_[HALF/FULL]_DUPLEX */
362 uint16_t link_autoneg : 1; /**< RTE_ETH_LINK_[AUTONEG/FIXED] */
363 uint16_t link_status : 1; /**< RTE_ETH_LINK_[DOWN/UP] */
364 } __rte_aligned(8); /**< aligned for atomic64 read/write */
366 /**@{@name Link negotiation
367 * Constants used in link management.
369 #define RTE_ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
370 #define ETH_LINK_HALF_DUPLEX RTE_ETH_LINK_HALF_DUPLEX
371 #define RTE_ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
372 #define ETH_LINK_FULL_DUPLEX RTE_ETH_LINK_FULL_DUPLEX
373 #define RTE_ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
374 #define ETH_LINK_DOWN RTE_ETH_LINK_DOWN
375 #define RTE_ETH_LINK_UP 1 /**< Link is up (see link_status). */
376 #define ETH_LINK_UP RTE_ETH_LINK_UP
377 #define RTE_ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
378 #define ETH_LINK_FIXED RTE_ETH_LINK_FIXED
379 #define RTE_ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
380 #define ETH_LINK_AUTONEG RTE_ETH_LINK_AUTONEG
381 #define RTE_ETH_LINK_MAX_STR_LEN 40 /**< Max length of default link string. */
385 * A structure used to configure the ring threshold registers of an Rx/Tx
386 * queue for an Ethernet port.
388 struct rte_eth_thresh {
389 uint8_t pthresh; /**< Ring prefetch threshold. */
390 uint8_t hthresh; /**< Ring host threshold. */
391 uint8_t wthresh; /**< Ring writeback threshold. */
394 /**@{@name Multi-queue mode
395 * @see rte_eth_conf.rxmode.mq_mode.
397 #define RTE_ETH_MQ_RX_RSS_FLAG RTE_BIT32(0) /**< Enable RSS. @see rte_eth_rss_conf */
398 #define ETH_MQ_RX_RSS_FLAG RTE_ETH_MQ_RX_RSS_FLAG
399 #define RTE_ETH_MQ_RX_DCB_FLAG RTE_BIT32(1) /**< Enable DCB. */
400 #define ETH_MQ_RX_DCB_FLAG RTE_ETH_MQ_RX_DCB_FLAG
401 #define RTE_ETH_MQ_RX_VMDQ_FLAG RTE_BIT32(2) /**< Enable VMDq. */
402 #define ETH_MQ_RX_VMDQ_FLAG RTE_ETH_MQ_RX_VMDQ_FLAG
406 * A set of values to identify what method is to be used to route
407 * packets to multiple queues.
409 enum rte_eth_rx_mq_mode {
410 /** None of DCB, RSS or VMDq mode */
411 RTE_ETH_MQ_RX_NONE = 0,
413 /** For Rx side, only RSS is on */
414 RTE_ETH_MQ_RX_RSS = RTE_ETH_MQ_RX_RSS_FLAG,
415 /** For Rx side,only DCB is on. */
416 RTE_ETH_MQ_RX_DCB = RTE_ETH_MQ_RX_DCB_FLAG,
417 /** Both DCB and RSS enable */
418 RTE_ETH_MQ_RX_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG,
420 /** Only VMDq, no RSS nor DCB */
421 RTE_ETH_MQ_RX_VMDQ_ONLY = RTE_ETH_MQ_RX_VMDQ_FLAG,
422 /** RSS mode with VMDq */
423 RTE_ETH_MQ_RX_VMDQ_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_VMDQ_FLAG,
424 /** Use VMDq+DCB to route traffic to queues */
425 RTE_ETH_MQ_RX_VMDQ_DCB = RTE_ETH_MQ_RX_VMDQ_FLAG | RTE_ETH_MQ_RX_DCB_FLAG,
426 /** Enable both VMDq and DCB in VMDq */
427 RTE_ETH_MQ_RX_VMDQ_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG |
428 RTE_ETH_MQ_RX_VMDQ_FLAG,
431 #define ETH_MQ_RX_NONE RTE_ETH_MQ_RX_NONE
432 #define ETH_MQ_RX_RSS RTE_ETH_MQ_RX_RSS
433 #define ETH_MQ_RX_DCB RTE_ETH_MQ_RX_DCB
434 #define ETH_MQ_RX_DCB_RSS RTE_ETH_MQ_RX_DCB_RSS
435 #define ETH_MQ_RX_VMDQ_ONLY RTE_ETH_MQ_RX_VMDQ_ONLY
436 #define ETH_MQ_RX_VMDQ_RSS RTE_ETH_MQ_RX_VMDQ_RSS
437 #define ETH_MQ_RX_VMDQ_DCB RTE_ETH_MQ_RX_VMDQ_DCB
438 #define ETH_MQ_RX_VMDQ_DCB_RSS RTE_ETH_MQ_RX_VMDQ_DCB_RSS
441 * A set of values to identify what method is to be used to transmit
442 * packets using multi-TCs.
444 enum rte_eth_tx_mq_mode {
445 RTE_ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
446 RTE_ETH_MQ_TX_DCB, /**< For Tx side,only DCB is on. */
447 RTE_ETH_MQ_TX_VMDQ_DCB, /**< For Tx side,both DCB and VT is on. */
448 RTE_ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
450 #define ETH_MQ_TX_NONE RTE_ETH_MQ_TX_NONE
451 #define ETH_MQ_TX_DCB RTE_ETH_MQ_TX_DCB
452 #define ETH_MQ_TX_VMDQ_DCB RTE_ETH_MQ_TX_VMDQ_DCB
453 #define ETH_MQ_TX_VMDQ_ONLY RTE_ETH_MQ_TX_VMDQ_ONLY
456 * A structure used to configure the Rx features of an Ethernet port.
458 struct rte_eth_rxmode {
459 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
460 enum rte_eth_rx_mq_mode mq_mode;
461 uint32_t mtu; /**< Requested MTU. */
462 /** Maximum allowed size of LRO aggregated packet. */
463 uint32_t max_lro_pkt_size;
464 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
466 * Per-port Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags.
467 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
468 * structure are allowed to be set.
472 uint64_t reserved_64s[2]; /**< Reserved for future fields */
473 void *reserved_ptrs[2]; /**< Reserved for future fields */
477 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
478 * Note that single VLAN is treated the same as inner VLAN.
481 RTE_ETH_VLAN_TYPE_UNKNOWN = 0,
482 RTE_ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
483 RTE_ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
484 RTE_ETH_VLAN_TYPE_MAX,
487 #define ETH_VLAN_TYPE_UNKNOWN RTE_ETH_VLAN_TYPE_UNKNOWN
488 #define ETH_VLAN_TYPE_INNER RTE_ETH_VLAN_TYPE_INNER
489 #define ETH_VLAN_TYPE_OUTER RTE_ETH_VLAN_TYPE_OUTER
490 #define ETH_VLAN_TYPE_MAX RTE_ETH_VLAN_TYPE_MAX
493 * A structure used to describe a VLAN filter.
494 * If the bit corresponding to a VID is set, such VID is on.
496 struct rte_vlan_filter_conf {
501 * A structure used to configure the Receive Side Scaling (RSS) feature
502 * of an Ethernet port.
503 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
504 * to an array holding the RSS key to use for hashing specific header
505 * fields of received packets. The length of this array should be indicated
506 * by *rss_key_len* below. Otherwise, a default random hash key is used by
509 * The *rss_key_len* field of the *rss_conf* structure indicates the length
510 * in bytes of the array pointed by *rss_key*. To be compatible, this length
511 * will be checked in i40e only. Others assume 40 bytes to be used as before.
513 * The *rss_hf* field of the *rss_conf* structure indicates the different
514 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
515 * Supplying an *rss_hf* equal to zero disables the RSS feature.
517 struct rte_eth_rss_conf {
518 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
519 uint8_t rss_key_len; /**< hash key length in bytes. */
520 uint64_t rss_hf; /**< Hash functions to apply - see below. */
524 * A packet can be identified by hardware as different flow types. Different
525 * NIC hardware may support different flow types.
526 * Basically, the NIC hardware identifies the flow type as deep protocol as
527 * possible, and exclusively. For example, if a packet is identified as
528 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types,
529 * though it is an actual IPV4 packet.
531 #define RTE_ETH_FLOW_UNKNOWN 0
532 #define RTE_ETH_FLOW_RAW 1
533 #define RTE_ETH_FLOW_IPV4 2
534 #define RTE_ETH_FLOW_FRAG_IPV4 3
535 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4
536 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5
537 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6
538 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7
539 #define RTE_ETH_FLOW_IPV6 8
540 #define RTE_ETH_FLOW_FRAG_IPV6 9
541 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10
542 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11
543 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12
544 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13
545 #define RTE_ETH_FLOW_L2_PAYLOAD 14
546 #define RTE_ETH_FLOW_IPV6_EX 15
547 #define RTE_ETH_FLOW_IPV6_TCP_EX 16
548 #define RTE_ETH_FLOW_IPV6_UDP_EX 17
549 /** Consider device port number as a flow differentiator */
550 #define RTE_ETH_FLOW_PORT 18
551 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */
552 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */
553 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */
554 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */
555 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */
556 #define RTE_ETH_FLOW_MAX 24
559 * Below macros are defined for RSS offload types, they can be used to
560 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types.
562 #define RTE_ETH_RSS_IPV4 RTE_BIT64(2)
563 #define ETH_RSS_IPV4 RTE_ETH_RSS_IPV4
564 #define RTE_ETH_RSS_FRAG_IPV4 RTE_BIT64(3)
565 #define ETH_RSS_FRAG_IPV4 RTE_ETH_RSS_FRAG_IPV4
566 #define RTE_ETH_RSS_NONFRAG_IPV4_TCP RTE_BIT64(4)
567 #define ETH_RSS_NONFRAG_IPV4_TCP RTE_ETH_RSS_NONFRAG_IPV4_TCP
568 #define RTE_ETH_RSS_NONFRAG_IPV4_UDP RTE_BIT64(5)
569 #define ETH_RSS_NONFRAG_IPV4_UDP RTE_ETH_RSS_NONFRAG_IPV4_UDP
570 #define RTE_ETH_RSS_NONFRAG_IPV4_SCTP RTE_BIT64(6)
571 #define ETH_RSS_NONFRAG_IPV4_SCTP RTE_ETH_RSS_NONFRAG_IPV4_SCTP
572 #define RTE_ETH_RSS_NONFRAG_IPV4_OTHER RTE_BIT64(7)
573 #define ETH_RSS_NONFRAG_IPV4_OTHER RTE_ETH_RSS_NONFRAG_IPV4_OTHER
574 #define RTE_ETH_RSS_IPV6 RTE_BIT64(8)
575 #define ETH_RSS_IPV6 RTE_ETH_RSS_IPV6
576 #define RTE_ETH_RSS_FRAG_IPV6 RTE_BIT64(9)
577 #define ETH_RSS_FRAG_IPV6 RTE_ETH_RSS_FRAG_IPV6
578 #define RTE_ETH_RSS_NONFRAG_IPV6_TCP RTE_BIT64(10)
579 #define ETH_RSS_NONFRAG_IPV6_TCP RTE_ETH_RSS_NONFRAG_IPV6_TCP
580 #define RTE_ETH_RSS_NONFRAG_IPV6_UDP RTE_BIT64(11)
581 #define ETH_RSS_NONFRAG_IPV6_UDP RTE_ETH_RSS_NONFRAG_IPV6_UDP
582 #define RTE_ETH_RSS_NONFRAG_IPV6_SCTP RTE_BIT64(12)
583 #define ETH_RSS_NONFRAG_IPV6_SCTP RTE_ETH_RSS_NONFRAG_IPV6_SCTP
584 #define RTE_ETH_RSS_NONFRAG_IPV6_OTHER RTE_BIT64(13)
585 #define ETH_RSS_NONFRAG_IPV6_OTHER RTE_ETH_RSS_NONFRAG_IPV6_OTHER
586 #define RTE_ETH_RSS_L2_PAYLOAD RTE_BIT64(14)
587 #define ETH_RSS_L2_PAYLOAD RTE_ETH_RSS_L2_PAYLOAD
588 #define RTE_ETH_RSS_IPV6_EX RTE_BIT64(15)
589 #define ETH_RSS_IPV6_EX RTE_ETH_RSS_IPV6_EX
590 #define RTE_ETH_RSS_IPV6_TCP_EX RTE_BIT64(16)
591 #define ETH_RSS_IPV6_TCP_EX RTE_ETH_RSS_IPV6_TCP_EX
592 #define RTE_ETH_RSS_IPV6_UDP_EX RTE_BIT64(17)
593 #define ETH_RSS_IPV6_UDP_EX RTE_ETH_RSS_IPV6_UDP_EX
594 #define RTE_ETH_RSS_PORT RTE_BIT64(18)
595 #define ETH_RSS_PORT RTE_ETH_RSS_PORT
596 #define RTE_ETH_RSS_VXLAN RTE_BIT64(19)
597 #define ETH_RSS_VXLAN RTE_ETH_RSS_VXLAN
598 #define RTE_ETH_RSS_GENEVE RTE_BIT64(20)
599 #define ETH_RSS_GENEVE RTE_ETH_RSS_GENEVE
600 #define RTE_ETH_RSS_NVGRE RTE_BIT64(21)
601 #define ETH_RSS_NVGRE RTE_ETH_RSS_NVGRE
602 #define RTE_ETH_RSS_GTPU RTE_BIT64(23)
603 #define ETH_RSS_GTPU RTE_ETH_RSS_GTPU
604 #define RTE_ETH_RSS_ETH RTE_BIT64(24)
605 #define ETH_RSS_ETH RTE_ETH_RSS_ETH
606 #define RTE_ETH_RSS_S_VLAN RTE_BIT64(25)
607 #define ETH_RSS_S_VLAN RTE_ETH_RSS_S_VLAN
608 #define RTE_ETH_RSS_C_VLAN RTE_BIT64(26)
609 #define ETH_RSS_C_VLAN RTE_ETH_RSS_C_VLAN
610 #define RTE_ETH_RSS_ESP RTE_BIT64(27)
611 #define ETH_RSS_ESP RTE_ETH_RSS_ESP
612 #define RTE_ETH_RSS_AH RTE_BIT64(28)
613 #define ETH_RSS_AH RTE_ETH_RSS_AH
614 #define RTE_ETH_RSS_L2TPV3 RTE_BIT64(29)
615 #define ETH_RSS_L2TPV3 RTE_ETH_RSS_L2TPV3
616 #define RTE_ETH_RSS_PFCP RTE_BIT64(30)
617 #define ETH_RSS_PFCP RTE_ETH_RSS_PFCP
618 #define RTE_ETH_RSS_PPPOE RTE_BIT64(31)
619 #define ETH_RSS_PPPOE RTE_ETH_RSS_PPPOE
620 #define RTE_ETH_RSS_ECPRI RTE_BIT64(32)
621 #define ETH_RSS_ECPRI RTE_ETH_RSS_ECPRI
622 #define RTE_ETH_RSS_MPLS RTE_BIT64(33)
623 #define ETH_RSS_MPLS RTE_ETH_RSS_MPLS
624 #define RTE_ETH_RSS_IPV4_CHKSUM RTE_BIT64(34)
625 #define ETH_RSS_IPV4_CHKSUM RTE_ETH_RSS_IPV4_CHKSUM
628 * The ETH_RSS_L4_CHKSUM works on checksum field of any L4 header.
629 * It is similar to ETH_RSS_PORT that they don't specify the specific type of
630 * L4 header. This macro is defined to replace some specific L4 (TCP/UDP/SCTP)
631 * checksum type for constructing the use of RSS offload bits.
633 * Due to above reason, some old APIs (and configuration) don't support
634 * RTE_ETH_RSS_L4_CHKSUM. The rte_flow RSS API supports it.
636 * For the case that checksum is not used in an UDP header,
637 * it takes the reserved value 0 as input for the hash function.
639 #define RTE_ETH_RSS_L4_CHKSUM RTE_BIT64(35)
640 #define ETH_RSS_L4_CHKSUM RTE_ETH_RSS_L4_CHKSUM
643 * We use the following macros to combine with above RTE_ETH_RSS_* for
644 * more specific input set selection. These bits are defined starting
645 * from the high end of the 64 bits.
646 * Note: If we use above RTE_ETH_RSS_* without SRC/DST_ONLY, it represents
647 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of
648 * the same level are used simultaneously, it is the same case as none of
651 #define RTE_ETH_RSS_L3_SRC_ONLY RTE_BIT64(63)
652 #define ETH_RSS_L3_SRC_ONLY RTE_ETH_RSS_L3_SRC_ONLY
653 #define RTE_ETH_RSS_L3_DST_ONLY RTE_BIT64(62)
654 #define ETH_RSS_L3_DST_ONLY RTE_ETH_RSS_L3_DST_ONLY
655 #define RTE_ETH_RSS_L4_SRC_ONLY RTE_BIT64(61)
656 #define ETH_RSS_L4_SRC_ONLY RTE_ETH_RSS_L4_SRC_ONLY
657 #define RTE_ETH_RSS_L4_DST_ONLY RTE_BIT64(60)
658 #define ETH_RSS_L4_DST_ONLY RTE_ETH_RSS_L4_DST_ONLY
659 #define RTE_ETH_RSS_L2_SRC_ONLY RTE_BIT64(59)
660 #define ETH_RSS_L2_SRC_ONLY RTE_ETH_RSS_L2_SRC_ONLY
661 #define RTE_ETH_RSS_L2_DST_ONLY RTE_BIT64(58)
662 #define ETH_RSS_L2_DST_ONLY RTE_ETH_RSS_L2_DST_ONLY
665 * Only select IPV6 address prefix as RSS input set according to
666 * https:tools.ietf.org/html/rfc6052
667 * Must be combined with RTE_ETH_RSS_IPV6, RTE_ETH_RSS_NONFRAG_IPV6_UDP,
668 * RTE_ETH_RSS_NONFRAG_IPV6_TCP, RTE_ETH_RSS_NONFRAG_IPV6_SCTP.
670 #define RTE_ETH_RSS_L3_PRE32 RTE_BIT64(57)
671 #define RTE_ETH_RSS_L3_PRE40 RTE_BIT64(56)
672 #define RTE_ETH_RSS_L3_PRE48 RTE_BIT64(55)
673 #define RTE_ETH_RSS_L3_PRE56 RTE_BIT64(54)
674 #define RTE_ETH_RSS_L3_PRE64 RTE_BIT64(53)
675 #define RTE_ETH_RSS_L3_PRE96 RTE_BIT64(52)
678 * Use the following macros to combine with the above layers
679 * to choose inner and outer layers or both for RSS computation.
680 * Bits 50 and 51 are reserved for this.
684 * level 0, requests the default behavior.
685 * Depending on the packet type, it can mean outermost, innermost,
686 * anything in between or even no RSS.
687 * It basically stands for the innermost encapsulation level RSS
688 * can be performed on according to PMD and device capabilities.
690 #define RTE_ETH_RSS_LEVEL_PMD_DEFAULT (UINT64_C(0) << 50)
691 #define ETH_RSS_LEVEL_PMD_DEFAULT RTE_ETH_RSS_LEVEL_PMD_DEFAULT
694 * level 1, requests RSS to be performed on the outermost packet
695 * encapsulation level.
697 #define RTE_ETH_RSS_LEVEL_OUTERMOST (UINT64_C(1) << 50)
698 #define ETH_RSS_LEVEL_OUTERMOST RTE_ETH_RSS_LEVEL_OUTERMOST
701 * level 2, requests RSS to be performed on the specified inner packet
702 * encapsulation level, from outermost to innermost (lower to higher values).
704 #define RTE_ETH_RSS_LEVEL_INNERMOST (UINT64_C(2) << 50)
705 #define ETH_RSS_LEVEL_INNERMOST RTE_ETH_RSS_LEVEL_INNERMOST
706 #define RTE_ETH_RSS_LEVEL_MASK (UINT64_C(3) << 50)
707 #define ETH_RSS_LEVEL_MASK RTE_ETH_RSS_LEVEL_MASK
709 #define RTE_ETH_RSS_LEVEL(rss_hf) ((rss_hf & RTE_ETH_RSS_LEVEL_MASK) >> 50)
710 #define ETH_RSS_LEVEL(rss_hf) RTE_ETH_RSS_LEVEL(rss_hf)
713 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of
714 * the same level are used simultaneously, it is the same case as
715 * none of them are added.
718 * RSS types with SRC/DST_ONLY.
722 static inline uint64_t
723 rte_eth_rss_hf_refine(uint64_t rss_hf)
725 if ((rss_hf & RTE_ETH_RSS_L3_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L3_DST_ONLY))
726 rss_hf &= ~(RTE_ETH_RSS_L3_SRC_ONLY | RTE_ETH_RSS_L3_DST_ONLY);
728 if ((rss_hf & RTE_ETH_RSS_L4_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L4_DST_ONLY))
729 rss_hf &= ~(RTE_ETH_RSS_L4_SRC_ONLY | RTE_ETH_RSS_L4_DST_ONLY);
734 #define RTE_ETH_RSS_IPV6_PRE32 ( \
736 RTE_ETH_RSS_L3_PRE32)
737 #define ETH_RSS_IPV6_PRE32 RTE_ETH_RSS_IPV6_PRE32
739 #define RTE_ETH_RSS_IPV6_PRE40 ( \
741 RTE_ETH_RSS_L3_PRE40)
742 #define ETH_RSS_IPV6_PRE40 RTE_ETH_RSS_IPV6_PRE40
744 #define RTE_ETH_RSS_IPV6_PRE48 ( \
746 RTE_ETH_RSS_L3_PRE48)
747 #define ETH_RSS_IPV6_PRE48 RTE_ETH_RSS_IPV6_PRE48
749 #define RTE_ETH_RSS_IPV6_PRE56 ( \
751 RTE_ETH_RSS_L3_PRE56)
752 #define ETH_RSS_IPV6_PRE56 RTE_ETH_RSS_IPV6_PRE56
754 #define RTE_ETH_RSS_IPV6_PRE64 ( \
756 RTE_ETH_RSS_L3_PRE64)
757 #define ETH_RSS_IPV6_PRE64 RTE_ETH_RSS_IPV6_PRE64
759 #define RTE_ETH_RSS_IPV6_PRE96 ( \
761 RTE_ETH_RSS_L3_PRE96)
762 #define ETH_RSS_IPV6_PRE96 RTE_ETH_RSS_IPV6_PRE96
764 #define RTE_ETH_RSS_IPV6_PRE32_UDP ( \
765 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
766 RTE_ETH_RSS_L3_PRE32)
767 #define ETH_RSS_IPV6_PRE32_UDP RTE_ETH_RSS_IPV6_PRE32_UDP
769 #define RTE_ETH_RSS_IPV6_PRE40_UDP ( \
770 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
771 RTE_ETH_RSS_L3_PRE40)
772 #define ETH_RSS_IPV6_PRE40_UDP RTE_ETH_RSS_IPV6_PRE40_UDP
774 #define RTE_ETH_RSS_IPV6_PRE48_UDP ( \
775 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
776 RTE_ETH_RSS_L3_PRE48)
777 #define ETH_RSS_IPV6_PRE48_UDP RTE_ETH_RSS_IPV6_PRE48_UDP
779 #define RTE_ETH_RSS_IPV6_PRE56_UDP ( \
780 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
781 RTE_ETH_RSS_L3_PRE56)
782 #define ETH_RSS_IPV6_PRE56_UDP RTE_ETH_RSS_IPV6_PRE56_UDP
784 #define RTE_ETH_RSS_IPV6_PRE64_UDP ( \
785 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
786 RTE_ETH_RSS_L3_PRE64)
787 #define ETH_RSS_IPV6_PRE64_UDP RTE_ETH_RSS_IPV6_PRE64_UDP
789 #define RTE_ETH_RSS_IPV6_PRE96_UDP ( \
790 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
791 RTE_ETH_RSS_L3_PRE96)
792 #define ETH_RSS_IPV6_PRE96_UDP RTE_ETH_RSS_IPV6_PRE96_UDP
794 #define RTE_ETH_RSS_IPV6_PRE32_TCP ( \
795 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
796 RTE_ETH_RSS_L3_PRE32)
797 #define ETH_RSS_IPV6_PRE32_TCP RTE_ETH_RSS_IPV6_PRE32_TCP
799 #define RTE_ETH_RSS_IPV6_PRE40_TCP ( \
800 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
801 RTE_ETH_RSS_L3_PRE40)
802 #define ETH_RSS_IPV6_PRE40_TCP RTE_ETH_RSS_IPV6_PRE40_TCP
804 #define RTE_ETH_RSS_IPV6_PRE48_TCP ( \
805 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
806 RTE_ETH_RSS_L3_PRE48)
807 #define ETH_RSS_IPV6_PRE48_TCP RTE_ETH_RSS_IPV6_PRE48_TCP
809 #define RTE_ETH_RSS_IPV6_PRE56_TCP ( \
810 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
811 RTE_ETH_RSS_L3_PRE56)
812 #define ETH_RSS_IPV6_PRE56_TCP RTE_ETH_RSS_IPV6_PRE56_TCP
814 #define RTE_ETH_RSS_IPV6_PRE64_TCP ( \
815 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
816 RTE_ETH_RSS_L3_PRE64)
817 #define ETH_RSS_IPV6_PRE64_TCP RTE_ETH_RSS_IPV6_PRE64_TCP
819 #define RTE_ETH_RSS_IPV6_PRE96_TCP ( \
820 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
821 RTE_ETH_RSS_L3_PRE96)
822 #define ETH_RSS_IPV6_PRE96_TCP RTE_ETH_RSS_IPV6_PRE96_TCP
824 #define RTE_ETH_RSS_IPV6_PRE32_SCTP ( \
825 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
826 RTE_ETH_RSS_L3_PRE32)
827 #define ETH_RSS_IPV6_PRE32_SCTP RTE_ETH_RSS_IPV6_PRE32_SCTP
829 #define RTE_ETH_RSS_IPV6_PRE40_SCTP ( \
830 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
831 RTE_ETH_RSS_L3_PRE40)
832 #define ETH_RSS_IPV6_PRE40_SCTP RTE_ETH_RSS_IPV6_PRE40_SCTP
834 #define RTE_ETH_RSS_IPV6_PRE48_SCTP ( \
835 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
836 RTE_ETH_RSS_L3_PRE48)
837 #define ETH_RSS_IPV6_PRE48_SCTP RTE_ETH_RSS_IPV6_PRE48_SCTP
839 #define RTE_ETH_RSS_IPV6_PRE56_SCTP ( \
840 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
841 RTE_ETH_RSS_L3_PRE56)
842 #define ETH_RSS_IPV6_PRE56_SCTP RTE_ETH_RSS_IPV6_PRE56_SCTP
844 #define RTE_ETH_RSS_IPV6_PRE64_SCTP ( \
845 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
846 RTE_ETH_RSS_L3_PRE64)
847 #define ETH_RSS_IPV6_PRE64_SCTP RTE_ETH_RSS_IPV6_PRE64_SCTP
849 #define RTE_ETH_RSS_IPV6_PRE96_SCTP ( \
850 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
851 RTE_ETH_RSS_L3_PRE96)
852 #define ETH_RSS_IPV6_PRE96_SCTP RTE_ETH_RSS_IPV6_PRE96_SCTP
854 #define RTE_ETH_RSS_IP ( \
856 RTE_ETH_RSS_FRAG_IPV4 | \
857 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \
859 RTE_ETH_RSS_FRAG_IPV6 | \
860 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \
862 #define ETH_RSS_IP RTE_ETH_RSS_IP
864 #define RTE_ETH_RSS_UDP ( \
865 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \
866 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
867 RTE_ETH_RSS_IPV6_UDP_EX)
868 #define ETH_RSS_UDP RTE_ETH_RSS_UDP
870 #define RTE_ETH_RSS_TCP ( \
871 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \
872 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
873 RTE_ETH_RSS_IPV6_TCP_EX)
874 #define ETH_RSS_TCP RTE_ETH_RSS_TCP
876 #define RTE_ETH_RSS_SCTP ( \
877 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \
878 RTE_ETH_RSS_NONFRAG_IPV6_SCTP)
879 #define ETH_RSS_SCTP RTE_ETH_RSS_SCTP
881 #define RTE_ETH_RSS_TUNNEL ( \
882 RTE_ETH_RSS_VXLAN | \
883 RTE_ETH_RSS_GENEVE | \
885 #define ETH_RSS_TUNNEL RTE_ETH_RSS_TUNNEL
887 #define RTE_ETH_RSS_VLAN ( \
888 RTE_ETH_RSS_S_VLAN | \
890 #define ETH_RSS_VLAN RTE_ETH_RSS_VLAN
892 /** Mask of valid RSS hash protocols */
893 #define RTE_ETH_RSS_PROTO_MASK ( \
895 RTE_ETH_RSS_FRAG_IPV4 | \
896 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \
897 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \
898 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \
899 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \
901 RTE_ETH_RSS_FRAG_IPV6 | \
902 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
903 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
904 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
905 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \
906 RTE_ETH_RSS_L2_PAYLOAD | \
907 RTE_ETH_RSS_IPV6_EX | \
908 RTE_ETH_RSS_IPV6_TCP_EX | \
909 RTE_ETH_RSS_IPV6_UDP_EX | \
911 RTE_ETH_RSS_VXLAN | \
912 RTE_ETH_RSS_GENEVE | \
913 RTE_ETH_RSS_NVGRE | \
915 #define ETH_RSS_PROTO_MASK RTE_ETH_RSS_PROTO_MASK
918 * Definitions used for redirection table entry size.
919 * Some RSS RETA sizes may not be supported by some drivers, check the
920 * documentation or the description of relevant functions for more details.
922 #define RTE_ETH_RSS_RETA_SIZE_64 64
923 #define ETH_RSS_RETA_SIZE_64 RTE_ETH_RSS_RETA_SIZE_64
924 #define RTE_ETH_RSS_RETA_SIZE_128 128
925 #define ETH_RSS_RETA_SIZE_128 RTE_ETH_RSS_RETA_SIZE_128
926 #define RTE_ETH_RSS_RETA_SIZE_256 256
927 #define ETH_RSS_RETA_SIZE_256 RTE_ETH_RSS_RETA_SIZE_256
928 #define RTE_ETH_RSS_RETA_SIZE_512 512
929 #define ETH_RSS_RETA_SIZE_512 RTE_ETH_RSS_RETA_SIZE_512
930 #define RTE_ETH_RETA_GROUP_SIZE 64
931 #define RTE_RETA_GROUP_SIZE RTE_ETH_RETA_GROUP_SIZE
933 /**@{@name VMDq and DCB maximums */
934 #define RTE_ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDq VLAN filters. */
935 #define ETH_VMDQ_MAX_VLAN_FILTERS RTE_ETH_VMDQ_MAX_VLAN_FILTERS
936 #define RTE_ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
937 #define ETH_DCB_NUM_USER_PRIORITIES RTE_ETH_DCB_NUM_USER_PRIORITIES
938 #define RTE_ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDq DCB queues. */
939 #define ETH_VMDQ_DCB_NUM_QUEUES RTE_ETH_VMDQ_DCB_NUM_QUEUES
940 #define RTE_ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
941 #define ETH_DCB_NUM_QUEUES RTE_ETH_DCB_NUM_QUEUES
944 /**@{@name DCB capabilities */
945 #define RTE_ETH_DCB_PG_SUPPORT RTE_BIT32(0) /**< Priority Group(ETS) support. */
946 #define ETH_DCB_PG_SUPPORT RTE_ETH_DCB_PG_SUPPORT
947 #define RTE_ETH_DCB_PFC_SUPPORT RTE_BIT32(1) /**< Priority Flow Control support. */
948 #define ETH_DCB_PFC_SUPPORT RTE_ETH_DCB_PFC_SUPPORT
951 /**@{@name VLAN offload bits */
952 #define RTE_ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
953 #define ETH_VLAN_STRIP_OFFLOAD RTE_ETH_VLAN_STRIP_OFFLOAD
954 #define RTE_ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
955 #define ETH_VLAN_FILTER_OFFLOAD RTE_ETH_VLAN_FILTER_OFFLOAD
956 #define RTE_ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
957 #define ETH_VLAN_EXTEND_OFFLOAD RTE_ETH_VLAN_EXTEND_OFFLOAD
958 #define RTE_ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */
959 #define ETH_QINQ_STRIP_OFFLOAD RTE_ETH_QINQ_STRIP_OFFLOAD
961 #define RTE_ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
962 #define ETH_VLAN_STRIP_MASK RTE_ETH_VLAN_STRIP_MASK
963 #define RTE_ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
964 #define ETH_VLAN_FILTER_MASK RTE_ETH_VLAN_FILTER_MASK
965 #define RTE_ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
966 #define ETH_VLAN_EXTEND_MASK RTE_ETH_VLAN_EXTEND_MASK
967 #define RTE_ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */
968 #define ETH_QINQ_STRIP_MASK RTE_ETH_QINQ_STRIP_MASK
969 #define RTE_ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
970 #define ETH_VLAN_ID_MAX RTE_ETH_VLAN_ID_MAX
973 /* Definitions used for receive MAC address */
974 #define RTE_ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
975 #define ETH_NUM_RECEIVE_MAC_ADDR RTE_ETH_NUM_RECEIVE_MAC_ADDR
977 /* Definitions used for unicast hash */
978 #define RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
979 #define ETH_VMDQ_NUM_UC_HASH_ARRAY RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY
981 /**@{@name VMDq Rx mode
982 * @see rte_eth_vmdq_rx_conf.rx_mode
984 /** Accept untagged packets. */
985 #define RTE_ETH_VMDQ_ACCEPT_UNTAG RTE_BIT32(0)
986 #define ETH_VMDQ_ACCEPT_UNTAG RTE_ETH_VMDQ_ACCEPT_UNTAG
987 /** Accept packets in multicast table. */
988 #define RTE_ETH_VMDQ_ACCEPT_HASH_MC RTE_BIT32(1)
989 #define ETH_VMDQ_ACCEPT_HASH_MC RTE_ETH_VMDQ_ACCEPT_HASH_MC
990 /** Accept packets in unicast table. */
991 #define RTE_ETH_VMDQ_ACCEPT_HASH_UC RTE_BIT32(2)
992 #define ETH_VMDQ_ACCEPT_HASH_UC RTE_ETH_VMDQ_ACCEPT_HASH_UC
993 /** Accept broadcast packets. */
994 #define RTE_ETH_VMDQ_ACCEPT_BROADCAST RTE_BIT32(3)
995 #define ETH_VMDQ_ACCEPT_BROADCAST RTE_ETH_VMDQ_ACCEPT_BROADCAST
996 /** Multicast promiscuous. */
997 #define RTE_ETH_VMDQ_ACCEPT_MULTICAST RTE_BIT32(4)
998 #define ETH_VMDQ_ACCEPT_MULTICAST RTE_ETH_VMDQ_ACCEPT_MULTICAST
1002 * A structure used to configure 64 entries of Redirection Table of the
1003 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
1004 * more than 64 entries supported by hardware, an array of this structure
1007 struct rte_eth_rss_reta_entry64 {
1008 /** Mask bits indicate which entries need to be updated/queried. */
1010 /** Group of 64 redirection table entries. */
1011 uint16_t reta[RTE_ETH_RETA_GROUP_SIZE];
1015 * This enum indicates the possible number of traffic classes
1016 * in DCB configurations
1018 enum rte_eth_nb_tcs {
1019 RTE_ETH_4_TCS = 4, /**< 4 TCs with DCB. */
1020 RTE_ETH_8_TCS = 8 /**< 8 TCs with DCB. */
1022 #define ETH_4_TCS RTE_ETH_4_TCS
1023 #define ETH_8_TCS RTE_ETH_8_TCS
1026 * This enum indicates the possible number of queue pools
1027 * in VMDq configurations.
1029 enum rte_eth_nb_pools {
1030 RTE_ETH_8_POOLS = 8, /**< 8 VMDq pools. */
1031 RTE_ETH_16_POOLS = 16, /**< 16 VMDq pools. */
1032 RTE_ETH_32_POOLS = 32, /**< 32 VMDq pools. */
1033 RTE_ETH_64_POOLS = 64 /**< 64 VMDq pools. */
1035 #define ETH_8_POOLS RTE_ETH_8_POOLS
1036 #define ETH_16_POOLS RTE_ETH_16_POOLS
1037 #define ETH_32_POOLS RTE_ETH_32_POOLS
1038 #define ETH_64_POOLS RTE_ETH_64_POOLS
1040 /* This structure may be extended in future. */
1041 struct rte_eth_dcb_rx_conf {
1042 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
1043 /** Traffic class each UP mapped to. */
1044 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1047 struct rte_eth_vmdq_dcb_tx_conf {
1048 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
1049 /** Traffic class each UP mapped to. */
1050 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1053 struct rte_eth_dcb_tx_conf {
1054 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
1055 /** Traffic class each UP mapped to. */
1056 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1059 struct rte_eth_vmdq_tx_conf {
1060 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
1064 * A structure used to configure the VMDq+DCB feature
1065 * of an Ethernet port.
1067 * Using this feature, packets are routed to a pool of queues, based
1068 * on the VLAN ID in the VLAN tag, and then to a specific queue within
1069 * that pool, using the user priority VLAN tag field.
1071 * A default pool may be used, if desired, to route all traffic which
1072 * does not match the VLAN filter rules.
1074 struct rte_eth_vmdq_dcb_conf {
1075 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
1076 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
1077 uint8_t default_pool; /**< The default pool, if applicable */
1078 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
1080 uint16_t vlan_id; /**< The VLAN ID of the received frame */
1081 uint64_t pools; /**< Bitmask of pools for packet Rx */
1082 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */
1083 /** Selects a queue in a pool */
1084 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1088 * A structure used to configure the VMDq feature of an Ethernet port when
1089 * not combined with the DCB feature.
1091 * Using this feature, packets are routed to a pool of queues. By default,
1092 * the pool selection is based on the MAC address, the VLAN ID in the
1093 * VLAN tag as specified in the pool_map array.
1094 * Passing the RTE_ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
1095 * selection using only the MAC address. MAC address to pool mapping is done
1096 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
1097 * corresponding to the pool ID.
1099 * Queue selection within the selected pool will be done using RSS when
1100 * it is enabled or revert to the first queue of the pool if not.
1102 * A default pool may be used, if desired, to route all traffic which
1103 * does not match the VLAN filter rules or any pool MAC address.
1105 struct rte_eth_vmdq_rx_conf {
1106 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
1107 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
1108 uint8_t default_pool; /**< The default pool, if applicable */
1109 uint8_t enable_loop_back; /**< Enable VT loop back */
1110 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
1111 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
1113 uint16_t vlan_id; /**< The VLAN ID of the received frame */
1114 uint64_t pools; /**< Bitmask of pools for packet Rx */
1115 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */
1119 * A structure used to configure the Tx features of an Ethernet port.
1121 struct rte_eth_txmode {
1122 enum rte_eth_tx_mq_mode mq_mode; /**< Tx multi-queues mode. */
1124 * Per-port Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags.
1125 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
1126 * structure are allowed to be set.
1132 uint8_t /** If set, reject sending out tagged pkts */
1133 hw_vlan_reject_tagged : 1,
1134 /** If set, reject sending out untagged pkts */
1135 hw_vlan_reject_untagged : 1,
1136 /** If set, enable port based VLAN insertion */
1137 hw_vlan_insert_pvid : 1;
1139 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1140 void *reserved_ptrs[2]; /**< Reserved for future fields */
1145 * @b EXPERIMENTAL: this structure may change without prior notice.
1147 * A structure used to configure an Rx packet segment to split.
1149 * If RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT flag is set in offloads field,
1150 * the PMD will split the received packets into multiple segments
1151 * according to the specification in the description array:
1153 * - The first network buffer will be allocated from the memory pool,
1154 * specified in the first array element, the second buffer, from the
1155 * pool in the second element, and so on.
1157 * - The offsets from the segment description elements specify
1158 * the data offset from the buffer beginning except the first mbuf.
1159 * The first segment offset is added with RTE_PKTMBUF_HEADROOM.
1161 * - The lengths in the elements define the maximal data amount
1162 * being received to each segment. The receiving starts with filling
1163 * up the first mbuf data buffer up to specified length. If the
1164 * there are data remaining (packet is longer than buffer in the first
1165 * mbuf) the following data will be pushed to the next segment
1166 * up to its own length, and so on.
1168 * - If the length in the segment description element is zero
1169 * the actual buffer size will be deduced from the appropriate
1170 * memory pool properties.
1172 * - If there is not enough elements to describe the buffer for entire
1173 * packet of maximal length the following parameters will be used
1174 * for the all remaining segments:
1175 * - pool from the last valid element
1176 * - the buffer size from this pool
1179 struct rte_eth_rxseg_split {
1180 struct rte_mempool *mp; /**< Memory pool to allocate segment from. */
1181 uint16_t length; /**< Segment data length, configures split point. */
1182 uint16_t offset; /**< Data offset from beginning of mbuf data buffer. */
1183 uint32_t reserved; /**< Reserved field. */
1188 * @b EXPERIMENTAL: this structure may change without prior notice.
1190 * A common structure used to describe Rx packet segment properties.
1192 union rte_eth_rxseg {
1193 /* The settings for buffer split offload. */
1194 struct rte_eth_rxseg_split split;
1195 /* The other features settings should be added here. */
1199 * A structure used to configure an Rx ring of an Ethernet port.
1201 struct rte_eth_rxconf {
1202 struct rte_eth_thresh rx_thresh; /**< Rx ring threshold registers. */
1203 uint16_t rx_free_thresh; /**< Drives the freeing of Rx descriptors. */
1204 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
1205 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1206 uint16_t rx_nseg; /**< Number of descriptions in rx_seg array. */
1208 * Share group index in Rx domain and switch domain.
1209 * Non-zero value to enable Rx queue share, zero value disable share.
1210 * PMD is responsible for Rx queue consistency checks to avoid member
1211 * port's configuration contradict to each other.
1213 uint16_t share_group;
1214 uint16_t share_qid; /**< Shared Rx queue ID in group */
1216 * Per-queue Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags.
1217 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
1218 * fields on rte_eth_dev_info structure are allowed to be set.
1222 * Points to the array of segment descriptions for an entire packet.
1223 * Array elements are properties for consecutive Rx segments.
1225 * The supported capabilities of receiving segmentation is reported
1226 * in rte_eth_dev_info.rx_seg_capa field.
1228 union rte_eth_rxseg *rx_seg;
1230 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1231 void *reserved_ptrs[2]; /**< Reserved for future fields */
1235 * A structure used to configure a Tx ring of an Ethernet port.
1237 struct rte_eth_txconf {
1238 struct rte_eth_thresh tx_thresh; /**< Tx ring threshold registers. */
1239 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
1240 uint16_t tx_free_thresh; /**< Start freeing Tx buffers if there are
1241 less free descriptors than this value. */
1243 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1245 * Per-queue Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags.
1246 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
1247 * fields on rte_eth_dev_info structure are allowed to be set.
1251 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1252 void *reserved_ptrs[2]; /**< Reserved for future fields */
1257 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1259 * A structure used to return the hairpin capabilities that are supported.
1261 struct rte_eth_hairpin_cap {
1262 /** The max number of hairpin queues (different bindings). */
1263 uint16_t max_nb_queues;
1264 /** Max number of Rx queues to be connected to one Tx queue. */
1265 uint16_t max_rx_2_tx;
1266 /** Max number of Tx queues to be connected to one Rx queue. */
1267 uint16_t max_tx_2_rx;
1268 uint16_t max_nb_desc; /**< The max num of descriptors. */
1271 #define RTE_ETH_MAX_HAIRPIN_PEERS 32
1275 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1277 * A structure used to hold hairpin peer data.
1279 struct rte_eth_hairpin_peer {
1280 uint16_t port; /**< Peer port. */
1281 uint16_t queue; /**< Peer queue. */
1286 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1288 * A structure used to configure hairpin binding.
1290 struct rte_eth_hairpin_conf {
1291 uint32_t peer_count:16; /**< The number of peers. */
1294 * Explicit Tx flow rule mode.
1295 * One hairpin pair of queues should have the same attribute.
1297 * - When set, the user should be responsible for inserting the hairpin
1298 * Tx part flows and removing them.
1299 * - When clear, the PMD will try to handle the Tx part of the flows,
1300 * e.g., by splitting one flow into two parts.
1302 uint32_t tx_explicit:1;
1305 * Manually bind hairpin queues.
1306 * One hairpin pair of queues should have the same attribute.
1308 * - When set, to enable hairpin, the user should call the hairpin bind
1309 * function after all the queues are set up properly and the ports are
1310 * started. Also, the hairpin unbind function should be called
1311 * accordingly before stopping a port that with hairpin configured.
1312 * - When clear, the PMD will try to enable the hairpin with the queues
1313 * configured automatically during port start.
1315 uint32_t manual_bind:1;
1316 uint32_t reserved:14; /**< Reserved bits. */
1317 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
1321 * A structure contains information about HW descriptor ring limitations.
1323 struct rte_eth_desc_lim {
1324 uint16_t nb_max; /**< Max allowed number of descriptors. */
1325 uint16_t nb_min; /**< Min allowed number of descriptors. */
1326 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
1329 * Max allowed number of segments per whole packet.
1331 * - For TSO packet this is the total number of data descriptors allowed
1334 * @see nb_mtu_seg_max
1336 uint16_t nb_seg_max;
1339 * Max number of segments per one MTU.
1341 * - For non-TSO packet, this is the maximum allowed number of segments
1342 * in a single transmit packet.
1344 * - For TSO packet each segment within the TSO may span up to this
1349 uint16_t nb_mtu_seg_max;
1353 * This enum indicates the flow control mode
1355 enum rte_eth_fc_mode {
1356 RTE_ETH_FC_NONE = 0, /**< Disable flow control. */
1357 RTE_ETH_FC_RX_PAUSE, /**< Rx pause frame, enable flowctrl on Tx side. */
1358 RTE_ETH_FC_TX_PAUSE, /**< Tx pause frame, enable flowctrl on Rx side. */
1359 RTE_ETH_FC_FULL /**< Enable flow control on both side. */
1362 #define RTE_FC_NONE RTE_ETH_FC_NONE
1363 #define RTE_FC_RX_PAUSE RTE_ETH_FC_RX_PAUSE
1364 #define RTE_FC_TX_PAUSE RTE_ETH_FC_TX_PAUSE
1365 #define RTE_FC_FULL RTE_ETH_FC_FULL
1368 * A structure used to configure Ethernet flow control parameter.
1369 * These parameters will be configured into the register of the NIC.
1370 * Please refer to the corresponding data sheet for proper value.
1372 struct rte_eth_fc_conf {
1373 uint32_t high_water; /**< High threshold value to trigger XOFF */
1374 uint32_t low_water; /**< Low threshold value to trigger XON */
1375 uint16_t pause_time; /**< Pause quota in the Pause frame */
1376 uint16_t send_xon; /**< Is XON frame need be sent */
1377 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1378 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
1379 uint8_t autoneg; /**< Use Pause autoneg */
1383 * A structure used to configure Ethernet priority flow control parameter.
1384 * These parameters will be configured into the register of the NIC.
1385 * Please refer to the corresponding data sheet for proper value.
1387 struct rte_eth_pfc_conf {
1388 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
1389 uint8_t priority; /**< VLAN User Priority. */
1393 * Tunnel type for device-specific classifier configuration.
1394 * @see rte_eth_udp_tunnel
1396 enum rte_eth_tunnel_type {
1397 RTE_ETH_TUNNEL_TYPE_NONE = 0,
1398 RTE_ETH_TUNNEL_TYPE_VXLAN,
1399 RTE_ETH_TUNNEL_TYPE_GENEVE,
1400 RTE_ETH_TUNNEL_TYPE_TEREDO,
1401 RTE_ETH_TUNNEL_TYPE_NVGRE,
1402 RTE_ETH_TUNNEL_TYPE_IP_IN_GRE,
1403 RTE_ETH_L2_TUNNEL_TYPE_E_TAG,
1404 RTE_ETH_TUNNEL_TYPE_VXLAN_GPE,
1405 RTE_ETH_TUNNEL_TYPE_ECPRI,
1406 RTE_ETH_TUNNEL_TYPE_MAX,
1409 #define RTE_TUNNEL_TYPE_NONE RTE_ETH_TUNNEL_TYPE_NONE
1410 #define RTE_TUNNEL_TYPE_VXLAN RTE_ETH_TUNNEL_TYPE_VXLAN
1411 #define RTE_TUNNEL_TYPE_GENEVE RTE_ETH_TUNNEL_TYPE_GENEVE
1412 #define RTE_TUNNEL_TYPE_TEREDO RTE_ETH_TUNNEL_TYPE_TEREDO
1413 #define RTE_TUNNEL_TYPE_NVGRE RTE_ETH_TUNNEL_TYPE_NVGRE
1414 #define RTE_TUNNEL_TYPE_IP_IN_GRE RTE_ETH_TUNNEL_TYPE_IP_IN_GRE
1415 #define RTE_L2_TUNNEL_TYPE_E_TAG RTE_ETH_L2_TUNNEL_TYPE_E_TAG
1416 #define RTE_TUNNEL_TYPE_VXLAN_GPE RTE_ETH_TUNNEL_TYPE_VXLAN_GPE
1417 #define RTE_TUNNEL_TYPE_ECPRI RTE_ETH_TUNNEL_TYPE_ECPRI
1418 #define RTE_TUNNEL_TYPE_MAX RTE_ETH_TUNNEL_TYPE_MAX
1420 /* Deprecated API file for rte_eth_dev_filter_* functions */
1421 #include "rte_eth_ctrl.h"
1424 * Memory space that can be configured to store Flow Director filters
1425 * in the board memory.
1427 enum rte_eth_fdir_pballoc_type {
1428 RTE_ETH_FDIR_PBALLOC_64K = 0, /**< 64k. */
1429 RTE_ETH_FDIR_PBALLOC_128K, /**< 128k. */
1430 RTE_ETH_FDIR_PBALLOC_256K, /**< 256k. */
1432 #define rte_fdir_pballoc_type rte_eth_fdir_pballoc_type
1434 #define RTE_FDIR_PBALLOC_64K RTE_ETH_FDIR_PBALLOC_64K
1435 #define RTE_FDIR_PBALLOC_128K RTE_ETH_FDIR_PBALLOC_128K
1436 #define RTE_FDIR_PBALLOC_256K RTE_ETH_FDIR_PBALLOC_256K
1439 * Select report mode of FDIR hash information in Rx descriptors.
1441 enum rte_fdir_status_mode {
1442 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
1443 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
1444 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
1448 * A structure used to configure the Flow Director (FDIR) feature
1449 * of an Ethernet port.
1451 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored.
1453 struct rte_eth_fdir_conf {
1454 enum rte_fdir_mode mode; /**< Flow Director mode. */
1455 enum rte_eth_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
1456 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
1457 /** Rx queue of packets matching a "drop" filter in perfect mode. */
1459 struct rte_eth_fdir_masks mask;
1460 /** Flex payload configuration. */
1461 struct rte_eth_fdir_flex_conf flex_conf;
1464 #define rte_fdir_conf rte_eth_fdir_conf
1467 * UDP tunneling configuration.
1469 * Used to configure the classifier of a device,
1470 * associating an UDP port with a type of tunnel.
1472 * Some NICs may need such configuration to properly parse a tunnel
1473 * with any standard or custom UDP port.
1475 struct rte_eth_udp_tunnel {
1476 uint16_t udp_port; /**< UDP port used for the tunnel. */
1477 uint8_t prot_type; /**< Tunnel type. @see rte_eth_tunnel_type */
1481 * A structure used to enable/disable specific device interrupts.
1483 struct rte_eth_intr_conf {
1484 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
1486 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
1488 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
1492 #define rte_intr_conf rte_eth_intr_conf
1495 * A structure used to configure an Ethernet port.
1496 * Depending upon the Rx multi-queue mode, extra advanced
1497 * configuration settings may be needed.
1499 struct rte_eth_conf {
1500 uint32_t link_speeds; /**< bitmap of RTE_ETH_LINK_SPEED_XXX of speeds to be
1501 used. RTE_ETH_LINK_SPEED_FIXED disables link
1502 autonegotiation, and a unique speed shall be
1503 set. Otherwise, the bitmap defines the set of
1504 speeds to be advertised. If the special value
1505 RTE_ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
1506 supported are advertised. */
1507 struct rte_eth_rxmode rxmode; /**< Port Rx configuration. */
1508 struct rte_eth_txmode txmode; /**< Port Tx configuration. */
1509 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
1510 is 0, meaning the loopback mode is disabled.
1511 Read the datasheet of given Ethernet controller
1512 for details. The possible values of this field
1513 are defined in implementation of each driver. */
1515 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
1516 /** Port VMDq+DCB configuration. */
1517 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
1518 /** Port DCB Rx configuration. */
1519 struct rte_eth_dcb_rx_conf dcb_rx_conf;
1520 /** Port VMDq Rx configuration. */
1521 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
1522 } rx_adv_conf; /**< Port Rx filtering configuration. */
1524 /** Port VMDq+DCB Tx configuration. */
1525 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
1526 /** Port DCB Tx configuration. */
1527 struct rte_eth_dcb_tx_conf dcb_tx_conf;
1528 /** Port VMDq Tx configuration. */
1529 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
1530 } tx_adv_conf; /**< Port Tx DCB configuration (union). */
1531 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
1532 is needed,and the variable must be set RTE_ETH_DCB_PFC_SUPPORT. */
1533 uint32_t dcb_capability_en;
1534 struct rte_eth_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */
1535 struct rte_eth_intr_conf intr_conf; /**< Interrupt mode configuration. */
1539 * Rx offload capabilities of a device.
1541 #define RTE_ETH_RX_OFFLOAD_VLAN_STRIP RTE_BIT64(0)
1542 #define DEV_RX_OFFLOAD_VLAN_STRIP RTE_ETH_RX_OFFLOAD_VLAN_STRIP
1543 #define RTE_ETH_RX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1)
1544 #define DEV_RX_OFFLOAD_IPV4_CKSUM RTE_ETH_RX_OFFLOAD_IPV4_CKSUM
1545 #define RTE_ETH_RX_OFFLOAD_UDP_CKSUM RTE_BIT64(2)
1546 #define DEV_RX_OFFLOAD_UDP_CKSUM RTE_ETH_RX_OFFLOAD_UDP_CKSUM
1547 #define RTE_ETH_RX_OFFLOAD_TCP_CKSUM RTE_BIT64(3)
1548 #define DEV_RX_OFFLOAD_TCP_CKSUM RTE_ETH_RX_OFFLOAD_TCP_CKSUM
1549 #define RTE_ETH_RX_OFFLOAD_TCP_LRO RTE_BIT64(4)
1550 #define DEV_RX_OFFLOAD_TCP_LRO RTE_ETH_RX_OFFLOAD_TCP_LRO
1551 #define RTE_ETH_RX_OFFLOAD_QINQ_STRIP RTE_BIT64(5)
1552 #define DEV_RX_OFFLOAD_QINQ_STRIP RTE_ETH_RX_OFFLOAD_QINQ_STRIP
1553 #define RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(6)
1554 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM
1555 #define RTE_ETH_RX_OFFLOAD_MACSEC_STRIP RTE_BIT64(7)
1556 #define DEV_RX_OFFLOAD_MACSEC_STRIP RTE_ETH_RX_OFFLOAD_MACSEC_STRIP
1557 #define RTE_ETH_RX_OFFLOAD_HEADER_SPLIT RTE_BIT64(8)
1558 #define DEV_RX_OFFLOAD_HEADER_SPLIT RTE_ETH_RX_OFFLOAD_HEADER_SPLIT
1559 #define RTE_ETH_RX_OFFLOAD_VLAN_FILTER RTE_BIT64(9)
1560 #define DEV_RX_OFFLOAD_VLAN_FILTER RTE_ETH_RX_OFFLOAD_VLAN_FILTER
1561 #define RTE_ETH_RX_OFFLOAD_VLAN_EXTEND RTE_BIT64(10)
1562 #define DEV_RX_OFFLOAD_VLAN_EXTEND RTE_ETH_RX_OFFLOAD_VLAN_EXTEND
1563 #define RTE_ETH_RX_OFFLOAD_SCATTER RTE_BIT64(13)
1564 #define DEV_RX_OFFLOAD_SCATTER RTE_ETH_RX_OFFLOAD_SCATTER
1566 * Timestamp is set by the driver in RTE_MBUF_DYNFIELD_TIMESTAMP_NAME
1567 * and RTE_MBUF_DYNFLAG_RX_TIMESTAMP_NAME is set in ol_flags.
1568 * The mbuf field and flag are registered when the offload is configured.
1570 #define RTE_ETH_RX_OFFLOAD_TIMESTAMP RTE_BIT64(14)
1571 #define DEV_RX_OFFLOAD_TIMESTAMP RTE_ETH_RX_OFFLOAD_TIMESTAMP
1572 #define RTE_ETH_RX_OFFLOAD_SECURITY RTE_BIT64(15)
1573 #define DEV_RX_OFFLOAD_SECURITY RTE_ETH_RX_OFFLOAD_SECURITY
1574 #define RTE_ETH_RX_OFFLOAD_KEEP_CRC RTE_BIT64(16)
1575 #define DEV_RX_OFFLOAD_KEEP_CRC RTE_ETH_RX_OFFLOAD_KEEP_CRC
1576 #define RTE_ETH_RX_OFFLOAD_SCTP_CKSUM RTE_BIT64(17)
1577 #define DEV_RX_OFFLOAD_SCTP_CKSUM RTE_ETH_RX_OFFLOAD_SCTP_CKSUM
1578 #define RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(18)
1579 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM
1580 #define RTE_ETH_RX_OFFLOAD_RSS_HASH RTE_BIT64(19)
1581 #define DEV_RX_OFFLOAD_RSS_HASH RTE_ETH_RX_OFFLOAD_RSS_HASH
1582 #define RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT RTE_BIT64(20)
1584 #define RTE_ETH_RX_OFFLOAD_CHECKSUM (RTE_ETH_RX_OFFLOAD_IPV4_CKSUM | \
1585 RTE_ETH_RX_OFFLOAD_UDP_CKSUM | \
1586 RTE_ETH_RX_OFFLOAD_TCP_CKSUM)
1587 #define DEV_RX_OFFLOAD_CHECKSUM RTE_ETH_RX_OFFLOAD_CHECKSUM
1588 #define RTE_ETH_RX_OFFLOAD_VLAN (RTE_ETH_RX_OFFLOAD_VLAN_STRIP | \
1589 RTE_ETH_RX_OFFLOAD_VLAN_FILTER | \
1590 RTE_ETH_RX_OFFLOAD_VLAN_EXTEND | \
1591 RTE_ETH_RX_OFFLOAD_QINQ_STRIP)
1592 #define DEV_RX_OFFLOAD_VLAN RTE_ETH_RX_OFFLOAD_VLAN
1595 * If new Rx offload capabilities are defined, they also must be
1596 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
1600 * Tx offload capabilities of a device.
1602 #define RTE_ETH_TX_OFFLOAD_VLAN_INSERT RTE_BIT64(0)
1603 #define DEV_TX_OFFLOAD_VLAN_INSERT RTE_ETH_TX_OFFLOAD_VLAN_INSERT
1604 #define RTE_ETH_TX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1)
1605 #define DEV_TX_OFFLOAD_IPV4_CKSUM RTE_ETH_TX_OFFLOAD_IPV4_CKSUM
1606 #define RTE_ETH_TX_OFFLOAD_UDP_CKSUM RTE_BIT64(2)
1607 #define DEV_TX_OFFLOAD_UDP_CKSUM RTE_ETH_TX_OFFLOAD_UDP_CKSUM
1608 #define RTE_ETH_TX_OFFLOAD_TCP_CKSUM RTE_BIT64(3)
1609 #define DEV_TX_OFFLOAD_TCP_CKSUM RTE_ETH_TX_OFFLOAD_TCP_CKSUM
1610 #define RTE_ETH_TX_OFFLOAD_SCTP_CKSUM RTE_BIT64(4)
1611 #define DEV_TX_OFFLOAD_SCTP_CKSUM RTE_ETH_TX_OFFLOAD_SCTP_CKSUM
1612 #define RTE_ETH_TX_OFFLOAD_TCP_TSO RTE_BIT64(5)
1613 #define DEV_TX_OFFLOAD_TCP_TSO RTE_ETH_TX_OFFLOAD_TCP_TSO
1614 #define RTE_ETH_TX_OFFLOAD_UDP_TSO RTE_BIT64(6)
1615 #define DEV_TX_OFFLOAD_UDP_TSO RTE_ETH_TX_OFFLOAD_UDP_TSO
1616 #define RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(7) /**< Used for tunneling packet. */
1617 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
1618 #define RTE_ETH_TX_OFFLOAD_QINQ_INSERT RTE_BIT64(8)
1619 #define DEV_TX_OFFLOAD_QINQ_INSERT RTE_ETH_TX_OFFLOAD_QINQ_INSERT
1620 #define RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO RTE_BIT64(9) /**< Used for tunneling packet. */
1621 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO
1622 #define RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO RTE_BIT64(10) /**< Used for tunneling packet. */
1623 #define DEV_TX_OFFLOAD_GRE_TNL_TSO RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO
1624 #define RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO RTE_BIT64(11) /**< Used for tunneling packet. */
1625 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO
1626 #define RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO RTE_BIT64(12) /**< Used for tunneling packet. */
1627 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO
1628 #define RTE_ETH_TX_OFFLOAD_MACSEC_INSERT RTE_BIT64(13)
1629 #define DEV_TX_OFFLOAD_MACSEC_INSERT RTE_ETH_TX_OFFLOAD_MACSEC_INSERT
1631 * Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
1632 * Tx queue without SW lock.
1634 #define RTE_ETH_TX_OFFLOAD_MT_LOCKFREE RTE_BIT64(14)
1635 #define DEV_TX_OFFLOAD_MT_LOCKFREE RTE_ETH_TX_OFFLOAD_MT_LOCKFREE
1636 /** Device supports multi segment send. */
1637 #define RTE_ETH_TX_OFFLOAD_MULTI_SEGS RTE_BIT64(15)
1638 #define DEV_TX_OFFLOAD_MULTI_SEGS RTE_ETH_TX_OFFLOAD_MULTI_SEGS
1640 * Device supports optimization for fast release of mbufs.
1641 * When set application must guarantee that per-queue all mbufs comes from
1642 * the same mempool and has refcnt = 1.
1644 #define RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE RTE_BIT64(16)
1645 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE
1646 #define RTE_ETH_TX_OFFLOAD_SECURITY RTE_BIT64(17)
1647 #define DEV_TX_OFFLOAD_SECURITY RTE_ETH_TX_OFFLOAD_SECURITY
1649 * Device supports generic UDP tunneled packet TSO.
1650 * Application must set PKT_TX_TUNNEL_UDP and other mbuf fields required
1653 #define RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO RTE_BIT64(18)
1654 #define DEV_TX_OFFLOAD_UDP_TNL_TSO RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO
1656 * Device supports generic IP tunneled packet TSO.
1657 * Application must set PKT_TX_TUNNEL_IP and other mbuf fields required
1660 #define RTE_ETH_TX_OFFLOAD_IP_TNL_TSO RTE_BIT64(19)
1661 #define DEV_TX_OFFLOAD_IP_TNL_TSO RTE_ETH_TX_OFFLOAD_IP_TNL_TSO
1662 /** Device supports outer UDP checksum */
1663 #define RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(20)
1664 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM
1666 * Device sends on time read from RTE_MBUF_DYNFIELD_TIMESTAMP_NAME
1667 * if RTE_MBUF_DYNFLAG_TX_TIMESTAMP_NAME is set in ol_flags.
1668 * The mbuf field and flag are registered when the offload is configured.
1670 #define RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_BIT64(21)
1671 #define DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP
1673 * If new Tx offload capabilities are defined, they also must be
1674 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1677 /**@{@name Device capabilities
1678 * Non-offload capabilities reported in rte_eth_dev_info.dev_capa.
1680 /** Device supports Rx queue setup after device started. */
1681 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP RTE_BIT64(0)
1682 /** Device supports Tx queue setup after device started. */
1683 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP RTE_BIT64(1)
1685 * Device supports shared Rx queue among ports within Rx domain and
1686 * switch domain. Mbufs are consumed by shared Rx queue instead of
1687 * each queue. Multiple groups are supported by share_group of Rx
1688 * queue configuration. Shared Rx queue is identified by PMD using
1689 * share_qid of Rx queue configuration. Polling any port in the group
1690 * receive packets of all member ports, source port identified by
1693 #define RTE_ETH_DEV_CAPA_RXQ_SHARE RTE_BIT64(2)
1697 * Fallback default preferred Rx/Tx port parameters.
1698 * These are used if an application requests default parameters
1699 * but the PMD does not provide preferred values.
1701 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1702 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1703 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1704 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1707 * Preferred Rx/Tx port parameters.
1708 * There are separate instances of this structure for transmission
1709 * and reception respectively.
1711 struct rte_eth_dev_portconf {
1712 uint16_t burst_size; /**< Device-preferred burst size */
1713 uint16_t ring_size; /**< Device-preferred size of queue rings */
1714 uint16_t nb_queues; /**< Device-preferred number of queues */
1718 * Default values for switch domain ID when ethdev does not support switch
1719 * domain definitions.
1721 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX)
1724 * Ethernet device associated switch information
1726 struct rte_eth_switch_info {
1727 const char *name; /**< switch name */
1728 uint16_t domain_id; /**< switch domain ID */
1730 * Mapping to the devices physical switch port as enumerated from the
1731 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1732 * device this may correspond to the VF_ID of each virtual function,
1733 * but each driver should explicitly define the mapping of switch
1734 * port identifier to that physical interconnect/switch
1738 * Shared Rx queue sub-domain boundary. Only ports in same Rx domain
1739 * and switch domain can share Rx queue. Valid only if device advertised
1740 * RTE_ETH_DEV_CAPA_RXQ_SHARE capability.
1747 * @b EXPERIMENTAL: this structure may change without prior notice.
1749 * Ethernet device Rx buffer segmentation capabilities.
1751 struct rte_eth_rxseg_capa {
1753 uint32_t multi_pools:1; /**< Supports receiving to multiple pools.*/
1754 uint32_t offset_allowed:1; /**< Supports buffer offsets. */
1755 uint32_t offset_align_log2:4; /**< Required offset alignment. */
1756 uint16_t max_nseg; /**< Maximum amount of segments to split. */
1757 uint16_t reserved; /**< Reserved field. */
1761 * Ethernet device information
1765 * Ethernet device representor port type.
1767 enum rte_eth_representor_type {
1768 RTE_ETH_REPRESENTOR_NONE, /**< not a representor. */
1769 RTE_ETH_REPRESENTOR_VF, /**< representor of Virtual Function. */
1770 RTE_ETH_REPRESENTOR_SF, /**< representor of Sub Function. */
1771 RTE_ETH_REPRESENTOR_PF, /**< representor of Physical Function. */
1775 * A structure used to retrieve the contextual information of
1776 * an Ethernet device, such as the controlling driver of the
1779 struct rte_eth_dev_info {
1780 struct rte_device *device; /** Generic device information */
1781 const char *driver_name; /**< Device Driver name. */
1782 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1783 Use if_indextoname() to translate into an interface name. */
1784 uint16_t min_mtu; /**< Minimum MTU allowed */
1785 uint16_t max_mtu; /**< Maximum MTU allowed */
1786 const uint32_t *dev_flags; /**< Device flags */
1787 uint32_t min_rx_bufsize; /**< Minimum size of Rx buffer. */
1788 uint32_t max_rx_pktlen; /**< Maximum configurable length of Rx pkt. */
1789 /** Maximum configurable size of LRO aggregated packet. */
1790 uint32_t max_lro_pkt_size;
1791 uint16_t max_rx_queues; /**< Maximum number of Rx queues. */
1792 uint16_t max_tx_queues; /**< Maximum number of Tx queues. */
1793 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1794 uint32_t max_hash_mac_addrs;
1795 /** Maximum number of hash MAC addresses for MTA and UTA. */
1796 uint16_t max_vfs; /**< Maximum number of VFs. */
1797 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1798 struct rte_eth_rxseg_capa rx_seg_capa; /**< Segmentation capability.*/
1799 /** All Rx offload capabilities including all per-queue ones */
1800 uint64_t rx_offload_capa;
1801 /** All Tx offload capabilities including all per-queue ones */
1802 uint64_t tx_offload_capa;
1803 /** Device per-queue Rx offload capabilities. */
1804 uint64_t rx_queue_offload_capa;
1805 /** Device per-queue Tx offload capabilities. */
1806 uint64_t tx_queue_offload_capa;
1807 /** Device redirection table size, the total number of entries. */
1809 uint8_t hash_key_size; /**< Hash key size in bytes */
1810 /** Bit mask of RSS offloads, the bit offset also means flow type */
1811 uint64_t flow_type_rss_offloads;
1812 struct rte_eth_rxconf default_rxconf; /**< Default Rx configuration */
1813 struct rte_eth_txconf default_txconf; /**< Default Tx configuration */
1814 uint16_t vmdq_queue_base; /**< First queue ID for VMDq pools. */
1815 uint16_t vmdq_queue_num; /**< Queue number for VMDq pools. */
1816 uint16_t vmdq_pool_base; /**< First ID of VMDq pools. */
1817 struct rte_eth_desc_lim rx_desc_lim; /**< Rx descriptors limits */
1818 struct rte_eth_desc_lim tx_desc_lim; /**< Tx descriptors limits */
1819 uint32_t speed_capa; /**< Supported speeds bitmap (RTE_ETH_LINK_SPEED_). */
1820 /** Configured number of Rx/Tx queues */
1821 uint16_t nb_rx_queues; /**< Number of Rx queues. */
1822 uint16_t nb_tx_queues; /**< Number of Tx queues. */
1823 /** Rx parameter recommendations */
1824 struct rte_eth_dev_portconf default_rxportconf;
1825 /** Tx parameter recommendations */
1826 struct rte_eth_dev_portconf default_txportconf;
1827 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1830 * Switching information for ports on a device with a
1831 * embedded managed interconnect/switch.
1833 struct rte_eth_switch_info switch_info;
1835 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1836 void *reserved_ptrs[2]; /**< Reserved for future fields */
1839 /**@{@name Rx/Tx queue states */
1840 #define RTE_ETH_QUEUE_STATE_STOPPED 0 /**< Queue stopped. */
1841 #define RTE_ETH_QUEUE_STATE_STARTED 1 /**< Queue started. */
1842 #define RTE_ETH_QUEUE_STATE_HAIRPIN 2 /**< Queue used for hairpin. */
1846 * Ethernet device Rx queue information structure.
1847 * Used to retrieve information about configured queue.
1849 struct rte_eth_rxq_info {
1850 struct rte_mempool *mp; /**< mempool used by that queue. */
1851 struct rte_eth_rxconf conf; /**< queue config parameters. */
1852 uint8_t scattered_rx; /**< scattered packets Rx supported. */
1853 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */
1854 uint16_t nb_desc; /**< configured number of RXDs. */
1855 uint16_t rx_buf_size; /**< hardware receive buffer size. */
1856 } __rte_cache_min_aligned;
1859 * Ethernet device Tx queue information structure.
1860 * Used to retrieve information about configured queue.
1862 struct rte_eth_txq_info {
1863 struct rte_eth_txconf conf; /**< queue config parameters. */
1864 uint16_t nb_desc; /**< configured number of TXDs. */
1865 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */
1866 } __rte_cache_min_aligned;
1868 /* Generic Burst mode flag definition, values can be ORed. */
1871 * If the queues have different burst mode description, this bit will be set
1872 * by PMD, then the application can iterate to retrieve burst description for
1875 #define RTE_ETH_BURST_FLAG_PER_QUEUE RTE_BIT64(0)
1878 * Ethernet device Rx/Tx queue packet burst mode information structure.
1879 * Used to retrieve information about packet burst mode setting.
1881 struct rte_eth_burst_mode {
1882 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */
1884 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */
1885 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */
1888 /** Maximum name length for extended statistics counters */
1889 #define RTE_ETH_XSTATS_NAME_SIZE 64
1892 * An Ethernet device extended statistic structure
1894 * This structure is used by rte_eth_xstats_get() to provide
1895 * statistics that are not provided in the generic *rte_eth_stats*
1897 * It maps a name ID, corresponding to an index in the array returned
1898 * by rte_eth_xstats_get_names(), to a statistic value.
1900 struct rte_eth_xstat {
1901 uint64_t id; /**< The index in xstats name array. */
1902 uint64_t value; /**< The statistic counter value. */
1906 * A name element for extended statistics.
1908 * An array of this structure is returned by rte_eth_xstats_get_names().
1909 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1910 * structure references these names by their array index.
1912 * The xstats should follow a common naming scheme.
1913 * Some names are standardized in rte_stats_strings.
1915 * - rx_missed_errors
1917 * - tx_size_128_to_255_packets
1919 struct rte_eth_xstat_name {
1920 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1923 #define RTE_ETH_DCB_NUM_TCS 8
1924 #define ETH_DCB_NUM_TCS RTE_ETH_DCB_NUM_TCS
1925 #define RTE_ETH_MAX_VMDQ_POOL 64
1926 #define ETH_MAX_VMDQ_POOL RTE_ETH_MAX_VMDQ_POOL
1929 * A structure used to get the information of queue and
1930 * TC mapping on both Tx and Rx paths.
1932 struct rte_eth_dcb_tc_queue_mapping {
1933 /** Rx queues assigned to tc per Pool */
1937 } tc_rxq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS];
1938 /** Rx queues assigned to tc per Pool */
1942 } tc_txq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS];
1946 * A structure used to get the information of DCB.
1947 * It includes TC UP mapping and queue TC mapping.
1949 struct rte_eth_dcb_info {
1950 uint8_t nb_tcs; /**< number of TCs */
1951 uint8_t prio_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
1952 uint8_t tc_bws[RTE_ETH_DCB_NUM_TCS]; /**< Tx BW percentage for each TC */
1953 /** Rx queues assigned to tc */
1954 struct rte_eth_dcb_tc_queue_mapping tc_queue;
1958 * This enum indicates the possible Forward Error Correction (FEC) modes
1959 * of an ethdev port.
1961 enum rte_eth_fec_mode {
1962 RTE_ETH_FEC_NOFEC = 0, /**< FEC is off */
1963 RTE_ETH_FEC_AUTO, /**< FEC autonegotiation modes */
1964 RTE_ETH_FEC_BASER, /**< FEC using common algorithm */
1965 RTE_ETH_FEC_RS, /**< FEC using RS algorithm */
1968 /* Translate from FEC mode to FEC capa */
1969 #define RTE_ETH_FEC_MODE_TO_CAPA(x) RTE_BIT32(x)
1971 /* This macro indicates FEC capa mask */
1972 #define RTE_ETH_FEC_MODE_CAPA_MASK(x) RTE_BIT32(RTE_ETH_FEC_ ## x)
1974 /* A structure used to get capabilities per link speed */
1975 struct rte_eth_fec_capa {
1976 uint32_t speed; /**< Link speed (see RTE_ETH_SPEED_NUM_*) */
1977 uint32_t capa; /**< FEC capabilities bitmask */
1980 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
1982 /* Macros to check for valid port */
1983 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
1984 if (!rte_eth_dev_is_valid_port(port_id)) { \
1985 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1990 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
1991 if (!rte_eth_dev_is_valid_port(port_id)) { \
1992 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1998 * Function type used for Rx packet processing packet callbacks.
2000 * The callback function is called on Rx with a burst of packets that have
2001 * been received on the given port and queue.
2004 * The Ethernet port on which Rx is being performed.
2006 * The queue on the Ethernet port which is being used to receive the packets.
2008 * The burst of packets that have just been received.
2010 * The number of packets in the burst pointed to by "pkts".
2012 * The max number of packets that can be stored in the "pkts" array.
2014 * The arbitrary user parameter passed in by the application when the callback
2015 * was originally configured.
2017 * The number of packets returned to the user.
2019 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
2020 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
2024 * Function type used for Tx packet processing packet callbacks.
2026 * The callback function is called on Tx with a burst of packets immediately
2027 * before the packets are put onto the hardware queue for transmission.
2030 * The Ethernet port on which Tx is being performed.
2032 * The queue on the Ethernet port which is being used to transmit the packets.
2034 * The burst of packets that are about to be transmitted.
2036 * The number of packets in the burst pointed to by "pkts".
2038 * The arbitrary user parameter passed in by the application when the callback
2039 * was originally configured.
2041 * The number of packets to be written to the NIC.
2043 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
2044 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
2047 * Possible states of an ethdev port.
2049 enum rte_eth_dev_state {
2050 /** Device is unused before being probed. */
2051 RTE_ETH_DEV_UNUSED = 0,
2052 /** Device is attached when allocated in probing. */
2053 RTE_ETH_DEV_ATTACHED,
2054 /** Device is in removed state when plug-out is detected. */
2055 RTE_ETH_DEV_REMOVED,
2058 struct rte_eth_dev_sriov {
2059 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
2060 uint8_t nb_q_per_pool; /**< Rx queue number per pool */
2061 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
2062 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
2064 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
2066 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
2068 #define RTE_ETH_DEV_NO_OWNER 0
2070 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
2072 struct rte_eth_dev_owner {
2073 uint64_t id; /**< The owner unique identifier. */
2074 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
2077 /**@{@name Device flags
2078 * Flags internally saved in rte_eth_dev_data.dev_flags
2079 * and reported in rte_eth_dev_info.dev_flags.
2081 /** PMD supports thread-safe flow operations */
2082 #define RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE RTE_BIT32(0)
2083 /** Device supports link state interrupt */
2084 #define RTE_ETH_DEV_INTR_LSC RTE_BIT32(1)
2085 /** Device is a bonded slave */
2086 #define RTE_ETH_DEV_BONDED_SLAVE RTE_BIT32(2)
2087 /** Device supports device removal interrupt */
2088 #define RTE_ETH_DEV_INTR_RMV RTE_BIT32(3)
2089 /** Device is port representor */
2090 #define RTE_ETH_DEV_REPRESENTOR RTE_BIT32(4)
2091 /** Device does not support MAC change after started */
2092 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR RTE_BIT32(5)
2094 * Queue xstats filled automatically by ethdev layer.
2095 * PMDs filling the queue xstats themselves should not set this flag
2097 #define RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS RTE_BIT32(6)
2101 * Iterates over valid ethdev ports owned by a specific owner.
2104 * The ID of the next possible valid owned port.
2106 * The owner identifier.
2107 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
2109 * Next valid port ID owned by owner_id, RTE_MAX_ETHPORTS if there is none.
2111 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
2112 const uint64_t owner_id);
2115 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
2117 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
2118 for (p = rte_eth_find_next_owned_by(0, o); \
2119 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
2120 p = rte_eth_find_next_owned_by(p + 1, o))
2123 * Iterates over valid ethdev ports.
2126 * The ID of the next possible valid port.
2128 * Next valid port ID, RTE_MAX_ETHPORTS if there is none.
2130 uint16_t rte_eth_find_next(uint16_t port_id);
2133 * Macro to iterate over all enabled and ownerless ethdev ports.
2135 #define RTE_ETH_FOREACH_DEV(p) \
2136 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
2139 * Iterates over ethdev ports of a specified device.
2141 * @param port_id_start
2142 * The ID of the next possible valid port.
2144 * The generic device behind the ports to iterate.
2146 * Next port ID of the device, possibly port_id_start,
2147 * RTE_MAX_ETHPORTS if there is none.
2150 rte_eth_find_next_of(uint16_t port_id_start,
2151 const struct rte_device *parent);
2154 * Macro to iterate over all ethdev ports of a specified device.
2157 * The ID of the matching port being iterated.
2159 * The rte_device pointer matching the iterated ports.
2161 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \
2162 for (port_id = rte_eth_find_next_of(0, parent); \
2163 port_id < RTE_MAX_ETHPORTS; \
2164 port_id = rte_eth_find_next_of(port_id + 1, parent))
2167 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device).
2169 * @param port_id_start
2170 * The ID of the next possible valid sibling port.
2171 * @param ref_port_id
2172 * The ID of a reference port to compare rte_device with.
2174 * Next sibling port ID, possibly port_id_start or ref_port_id itself,
2175 * RTE_MAX_ETHPORTS if there is none.
2178 rte_eth_find_next_sibling(uint16_t port_id_start, uint16_t ref_port_id);
2181 * Macro to iterate over all ethdev ports sharing the same rte_device
2182 * as the specified port.
2183 * Note: the specified reference port is part of the loop iterations.
2186 * The ID of the matching port being iterated.
2187 * @param ref_port_id
2188 * The ID of the port being compared.
2190 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \
2191 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \
2192 port_id < RTE_MAX_ETHPORTS; \
2193 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id))
2197 * @b EXPERIMENTAL: this API may change without prior notice.
2199 * Get a new unique owner identifier.
2200 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
2201 * to avoid multiple management of device by different entities.
2204 * Owner identifier pointer.
2206 * Negative errno value on error, 0 on success.
2209 int rte_eth_dev_owner_new(uint64_t *owner_id);
2213 * @b EXPERIMENTAL: this API may change without prior notice.
2215 * Set an Ethernet device owner.
2218 * The identifier of the port to own.
2220 * The owner pointer.
2222 * Negative errno value on error, 0 on success.
2225 int rte_eth_dev_owner_set(const uint16_t port_id,
2226 const struct rte_eth_dev_owner *owner);
2230 * @b EXPERIMENTAL: this API may change without prior notice.
2232 * Unset Ethernet device owner to make the device ownerless.
2235 * The identifier of port to make ownerless.
2237 * The owner identifier.
2239 * 0 on success, negative errno value on error.
2242 int rte_eth_dev_owner_unset(const uint16_t port_id,
2243 const uint64_t owner_id);
2247 * @b EXPERIMENTAL: this API may change without prior notice.
2249 * Remove owner from all Ethernet devices owned by a specific owner.
2252 * The owner identifier.
2254 * 0 on success, negative errno value on error.
2257 int rte_eth_dev_owner_delete(const uint64_t owner_id);
2261 * @b EXPERIMENTAL: this API may change without prior notice.
2263 * Get the owner of an Ethernet device.
2266 * The port identifier.
2268 * The owner structure pointer to fill.
2270 * 0 on success, negative errno value on error..
2273 int rte_eth_dev_owner_get(const uint16_t port_id,
2274 struct rte_eth_dev_owner *owner);
2277 * Get the number of ports which are usable for the application.
2279 * These devices must be iterated by using the macro
2280 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
2281 * to deal with non-contiguous ranges of devices.
2284 * The count of available Ethernet devices.
2286 uint16_t rte_eth_dev_count_avail(void);
2289 * Get the total number of ports which are allocated.
2291 * Some devices may not be available for the application.
2294 * The total count of Ethernet devices.
2296 uint16_t rte_eth_dev_count_total(void);
2299 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
2300 * the bitmap link_speeds of the struct rte_eth_conf
2303 * Numerical speed value in Mbps
2305 * RTE_ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
2307 * 0 if the speed cannot be mapped
2309 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
2312 * Get RTE_ETH_RX_OFFLOAD_* flag name.
2317 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
2319 const char *rte_eth_dev_rx_offload_name(uint64_t offload);
2322 * Get RTE_ETH_TX_OFFLOAD_* flag name.
2327 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
2329 const char *rte_eth_dev_tx_offload_name(uint64_t offload);
2333 * @b EXPERIMENTAL: this API may change without prior notice.
2335 * Get RTE_ETH_DEV_CAPA_* flag name.
2340 * Capability name or 'UNKNOWN' if the flag cannot be recognized.
2343 const char *rte_eth_dev_capability_name(uint64_t capability);
2346 * Configure an Ethernet device.
2347 * This function must be invoked first before any other function in the
2348 * Ethernet API. This function can also be re-invoked when a device is in the
2352 * The port identifier of the Ethernet device to configure.
2353 * @param nb_rx_queue
2354 * The number of receive queues to set up for the Ethernet device.
2355 * @param nb_tx_queue
2356 * The number of transmit queues to set up for the Ethernet device.
2358 * The pointer to the configuration data to be used for the Ethernet device.
2359 * The *rte_eth_conf* structure includes:
2360 * - the hardware offload features to activate, with dedicated fields for
2361 * each statically configurable offload hardware feature provided by
2362 * Ethernet devices, such as IP checksum or VLAN tag stripping for
2364 * The Rx offload bitfield API is obsolete and will be deprecated.
2365 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
2366 * structure and use offloads field to set per-port offloads instead.
2367 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
2368 * the [rt]x_offload_capa returned from rte_eth_dev_info_get().
2369 * Any type of device supported offloading set in the input argument
2370 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
2371 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup()
2372 * - the Receive Side Scaling (RSS) configuration when using multiple Rx
2373 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf
2374 * must be within the flow_type_rss_offloads provided by drivers via
2375 * rte_eth_dev_info_get() API.
2377 * Embedding all configuration information in a single data structure
2378 * is the more flexible method that allows the addition of new features
2379 * without changing the syntax of the API.
2381 * - 0: Success, device configured.
2382 * - <0: Error code returned by the driver configuration function.
2384 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
2385 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
2389 * @b EXPERIMENTAL: this API may change without prior notice.
2391 * Check if an Ethernet device was physically removed.
2394 * The port identifier of the Ethernet device.
2396 * 1 when the Ethernet device is removed, otherwise 0.
2400 rte_eth_dev_is_removed(uint16_t port_id);
2403 * Allocate and set up a receive queue for an Ethernet device.
2405 * The function allocates a contiguous block of memory for *nb_rx_desc*
2406 * receive descriptors from a memory zone associated with *socket_id*
2407 * and initializes each receive descriptor with a network buffer allocated
2408 * from the memory pool *mb_pool*.
2411 * The port identifier of the Ethernet device.
2412 * @param rx_queue_id
2413 * The index of the receive queue to set up.
2414 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2415 * to rte_eth_dev_configure().
2417 * The number of receive descriptors to allocate for the receive ring.
2419 * The *socket_id* argument is the socket identifier in case of NUMA.
2420 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2421 * the DMA memory allocated for the receive descriptors of the ring.
2423 * The pointer to the configuration data to be used for the receive queue.
2424 * NULL value is allowed, in which case default Rx configuration
2426 * The *rx_conf* structure contains an *rx_thresh* structure with the values
2427 * of the Prefetch, Host, and Write-Back threshold registers of the receive
2429 * In addition it contains the hardware offloads features to activate using
2430 * the RTE_ETH_RX_OFFLOAD_* flags.
2431 * If an offloading set in rx_conf->offloads
2432 * hasn't been set in the input argument eth_conf->rxmode.offloads
2433 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2434 * per-queue type and it is enabled for the queue.
2435 * No need to repeat any bit in rx_conf->offloads which has already been
2436 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2437 * at port level can't be disabled at queue level.
2438 * The configuration structure also contains the pointer to the array
2439 * of the receiving buffer segment descriptions, see rx_seg and rx_nseg
2440 * fields, this extended configuration might be used by split offloads like
2441 * RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT. If mb_pool is not NULL,
2442 * the extended configuration fields must be set to NULL and zero.
2444 * The pointer to the memory pool from which to allocate *rte_mbuf* network
2445 * memory buffers to populate each descriptor of the receive ring. There are
2446 * two options to provide Rx buffer configuration:
2448 * mb_pool is not NULL, rx_conf.rx_nseg is 0.
2449 * - multiple segments description:
2450 * mb_pool is NULL, rx_conf.rx_seg is not NULL, rx_conf.rx_nseg is not 0.
2451 * Taken only if flag RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT is set in offloads.
2454 * - 0: Success, receive queue correctly set up.
2455 * - -EIO: if device is removed.
2456 * - -ENODEV: if *port_id* is invalid.
2457 * - -EINVAL: The memory pool pointer is null or the size of network buffers
2458 * which can be allocated from this memory pool does not fit the various
2459 * buffer sizes allowed by the device controller.
2460 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
2461 * allocate network memory buffers from the memory pool when
2462 * initializing receive descriptors.
2464 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
2465 uint16_t nb_rx_desc, unsigned int socket_id,
2466 const struct rte_eth_rxconf *rx_conf,
2467 struct rte_mempool *mb_pool);
2471 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2473 * Allocate and set up a hairpin receive queue for an Ethernet device.
2475 * The function set up the selected queue to be used in hairpin.
2478 * The port identifier of the Ethernet device.
2479 * @param rx_queue_id
2480 * The index of the receive queue to set up.
2481 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2482 * to rte_eth_dev_configure().
2484 * The number of receive descriptors to allocate for the receive ring.
2485 * 0 means the PMD will use default value.
2487 * The pointer to the hairpin configuration.
2490 * - (0) if successful.
2491 * - (-ENODEV) if *port_id* is invalid.
2492 * - (-ENOTSUP) if hardware doesn't support.
2493 * - (-EINVAL) if bad parameter.
2494 * - (-ENOMEM) if unable to allocate the resources.
2497 int rte_eth_rx_hairpin_queue_setup
2498 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc,
2499 const struct rte_eth_hairpin_conf *conf);
2502 * Allocate and set up a transmit queue for an Ethernet device.
2505 * The port identifier of the Ethernet device.
2506 * @param tx_queue_id
2507 * The index of the transmit queue to set up.
2508 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2509 * to rte_eth_dev_configure().
2511 * The number of transmit descriptors to allocate for the transmit ring.
2513 * The *socket_id* argument is the socket identifier in case of NUMA.
2514 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2515 * the DMA memory allocated for the transmit descriptors of the ring.
2517 * The pointer to the configuration data to be used for the transmit queue.
2518 * NULL value is allowed, in which case default Tx configuration
2520 * The *tx_conf* structure contains the following data:
2521 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
2522 * Write-Back threshold registers of the transmit ring.
2523 * When setting Write-Back threshold to the value greater then zero,
2524 * *tx_rs_thresh* value should be explicitly set to one.
2525 * - The *tx_free_thresh* value indicates the [minimum] number of network
2526 * buffers that must be pending in the transmit ring to trigger their
2527 * [implicit] freeing by the driver transmit function.
2528 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
2529 * descriptors that must be pending in the transmit ring before setting the
2530 * RS bit on a descriptor by the driver transmit function.
2531 * The *tx_rs_thresh* value should be less or equal then
2532 * *tx_free_thresh* value, and both of them should be less then
2534 * - The *offloads* member contains Tx offloads to be enabled.
2535 * If an offloading set in tx_conf->offloads
2536 * hasn't been set in the input argument eth_conf->txmode.offloads
2537 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2538 * per-queue type and it is enabled for the queue.
2539 * No need to repeat any bit in tx_conf->offloads which has already been
2540 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2541 * at port level can't be disabled at queue level.
2543 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
2544 * the transmit function to use default values.
2546 * - 0: Success, the transmit queue is correctly set up.
2547 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
2549 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
2550 uint16_t nb_tx_desc, unsigned int socket_id,
2551 const struct rte_eth_txconf *tx_conf);
2555 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2557 * Allocate and set up a transmit hairpin queue for an Ethernet device.
2560 * The port identifier of the Ethernet device.
2561 * @param tx_queue_id
2562 * The index of the transmit queue to set up.
2563 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2564 * to rte_eth_dev_configure().
2566 * The number of transmit descriptors to allocate for the transmit ring.
2567 * 0 to set default PMD value.
2569 * The hairpin configuration.
2572 * - (0) if successful.
2573 * - (-ENODEV) if *port_id* is invalid.
2574 * - (-ENOTSUP) if hardware doesn't support.
2575 * - (-EINVAL) if bad parameter.
2576 * - (-ENOMEM) if unable to allocate the resources.
2579 int rte_eth_tx_hairpin_queue_setup
2580 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc,
2581 const struct rte_eth_hairpin_conf *conf);
2585 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2587 * Get all the hairpin peer Rx / Tx ports of the current port.
2588 * The caller should ensure that the array is large enough to save the ports
2592 * The port identifier of the Ethernet device.
2594 * Pointer to the array to store the peer ports list.
2596 * Length of the array to store the port identifiers.
2598 * Current port to peer port direction
2599 * positive - current used as Tx to get all peer Rx ports.
2600 * zero - current used as Rx to get all peer Tx ports.
2603 * - (0 or positive) actual peer ports number.
2604 * - (-EINVAL) if bad parameter.
2605 * - (-ENODEV) if *port_id* invalid
2606 * - (-ENOTSUP) if hardware doesn't support.
2607 * - Others detailed errors from PMD drivers.
2610 int rte_eth_hairpin_get_peer_ports(uint16_t port_id, uint16_t *peer_ports,
2611 size_t len, uint32_t direction);
2615 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2617 * Bind all hairpin Tx queues of one port to the Rx queues of the peer port.
2618 * It is only allowed to call this function after all hairpin queues are
2619 * configured properly and the devices are in started state.
2622 * The identifier of the Tx port.
2624 * The identifier of peer Rx port.
2625 * RTE_MAX_ETHPORTS is allowed for the traversal of all devices.
2626 * Rx port ID could have the same value as Tx port ID.
2629 * - (0) if successful.
2630 * - (-ENODEV) if Tx port ID is invalid.
2631 * - (-EBUSY) if device is not in started state.
2632 * - (-ENOTSUP) if hardware doesn't support.
2633 * - Others detailed errors from PMD drivers.
2636 int rte_eth_hairpin_bind(uint16_t tx_port, uint16_t rx_port);
2640 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2642 * Unbind all hairpin Tx queues of one port from the Rx queues of the peer port.
2643 * This should be called before closing the Tx or Rx devices, if the bind
2644 * function is called before.
2645 * After unbinding the hairpin ports pair, it is allowed to bind them again.
2646 * Changing queues configuration should be after stopping the device(s).
2649 * The identifier of the Tx port.
2651 * The identifier of peer Rx port.
2652 * RTE_MAX_ETHPORTS is allowed for traversal of all devices.
2653 * Rx port ID could have the same value as Tx port ID.
2656 * - (0) if successful.
2657 * - (-ENODEV) if Tx port ID is invalid.
2658 * - (-EBUSY) if device is in stopped state.
2659 * - (-ENOTSUP) if hardware doesn't support.
2660 * - Others detailed errors from PMD drivers.
2663 int rte_eth_hairpin_unbind(uint16_t tx_port, uint16_t rx_port);
2666 * Return the NUMA socket to which an Ethernet device is connected
2669 * The port identifier of the Ethernet device
2671 * The NUMA socket ID to which the Ethernet device is connected or
2672 * a default of zero if the socket could not be determined.
2673 * -1 is returned is the port_id value is out of range.
2675 int rte_eth_dev_socket_id(uint16_t port_id);
2678 * Check if port_id of device is attached
2681 * The port identifier of the Ethernet device
2683 * - 0 if port is out of range or not attached
2684 * - 1 if device is attached
2686 int rte_eth_dev_is_valid_port(uint16_t port_id);
2689 * Start specified Rx queue of a port. It is used when rx_deferred_start
2690 * flag of the specified queue is true.
2693 * The port identifier of the Ethernet device
2694 * @param rx_queue_id
2695 * The index of the Rx queue to update the ring.
2696 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2697 * to rte_eth_dev_configure().
2699 * - 0: Success, the receive queue is started.
2700 * - -ENODEV: if *port_id* is invalid.
2701 * - -EINVAL: The queue_id out of range or belong to hairpin.
2702 * - -EIO: if device is removed.
2703 * - -ENOTSUP: The function not supported in PMD driver.
2705 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2708 * Stop specified Rx queue of a port
2711 * The port identifier of the Ethernet device
2712 * @param rx_queue_id
2713 * The index of the Rx queue to update the ring.
2714 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2715 * to rte_eth_dev_configure().
2717 * - 0: Success, the receive queue is stopped.
2718 * - -ENODEV: if *port_id* is invalid.
2719 * - -EINVAL: The queue_id out of range or belong to hairpin.
2720 * - -EIO: if device is removed.
2721 * - -ENOTSUP: The function not supported in PMD driver.
2723 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2726 * Start Tx for specified queue of a port. It is used when tx_deferred_start
2727 * flag of the specified queue is true.
2730 * The port identifier of the Ethernet device
2731 * @param tx_queue_id
2732 * The index of the Tx queue to update the ring.
2733 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2734 * to rte_eth_dev_configure().
2736 * - 0: Success, the transmit queue is started.
2737 * - -ENODEV: if *port_id* is invalid.
2738 * - -EINVAL: The queue_id out of range or belong to hairpin.
2739 * - -EIO: if device is removed.
2740 * - -ENOTSUP: The function not supported in PMD driver.
2742 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2745 * Stop specified Tx queue of a port
2748 * The port identifier of the Ethernet device
2749 * @param tx_queue_id
2750 * The index of the Tx queue to update the ring.
2751 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2752 * to rte_eth_dev_configure().
2754 * - 0: Success, the transmit queue is stopped.
2755 * - -ENODEV: if *port_id* is invalid.
2756 * - -EINVAL: The queue_id out of range or belong to hairpin.
2757 * - -EIO: if device is removed.
2758 * - -ENOTSUP: The function not supported in PMD driver.
2760 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2763 * Start an Ethernet device.
2765 * The device start step is the last one and consists of setting the configured
2766 * offload features and in starting the transmit and the receive units of the
2769 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before
2770 * PMD port start callback function is invoked.
2772 * On success, all basic functions exported by the Ethernet API (link status,
2773 * receive/transmit, and so on) can be invoked.
2776 * The port identifier of the Ethernet device.
2778 * - 0: Success, Ethernet device started.
2779 * - <0: Error code of the driver device start function.
2781 int rte_eth_dev_start(uint16_t port_id);
2784 * Stop an Ethernet device. The device can be restarted with a call to
2785 * rte_eth_dev_start()
2788 * The port identifier of the Ethernet device.
2790 * - 0: Success, Ethernet device stopped.
2791 * - <0: Error code of the driver device stop function.
2793 int rte_eth_dev_stop(uint16_t port_id);
2796 * Link up an Ethernet device.
2798 * Set device link up will re-enable the device Rx/Tx
2799 * functionality after it is previously set device linked down.
2802 * The port identifier of the Ethernet device.
2804 * - 0: Success, Ethernet device linked up.
2805 * - <0: Error code of the driver device link up function.
2807 int rte_eth_dev_set_link_up(uint16_t port_id);
2810 * Link down an Ethernet device.
2811 * The device Rx/Tx functionality will be disabled if success,
2812 * and it can be re-enabled with a call to
2813 * rte_eth_dev_set_link_up()
2816 * The port identifier of the Ethernet device.
2818 int rte_eth_dev_set_link_down(uint16_t port_id);
2821 * Close a stopped Ethernet device. The device cannot be restarted!
2822 * The function frees all port resources.
2825 * The port identifier of the Ethernet device.
2827 * - Zero if the port is closed successfully.
2828 * - Negative if something went wrong.
2830 int rte_eth_dev_close(uint16_t port_id);
2833 * Reset a Ethernet device and keep its port ID.
2835 * When a port has to be reset passively, the DPDK application can invoke
2836 * this function. For example when a PF is reset, all its VFs should also
2837 * be reset. Normally a DPDK application can invoke this function when
2838 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2839 * a port reset in other circumstances.
2841 * When this function is called, it first stops the port and then calls the
2842 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2843 * state, in which no Tx and Rx queues are setup, as if the port has been
2844 * reset and not started. The port keeps the port ID it had before the
2847 * After calling rte_eth_dev_reset( ), the application should use
2848 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2849 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2850 * to reconfigure the device as appropriate.
2852 * Note: To avoid unexpected behavior, the application should stop calling
2853 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2854 * safety, all these controlling functions should be called from the same
2858 * The port identifier of the Ethernet device.
2861 * - (0) if successful.
2862 * - (-ENODEV) if *port_id* is invalid.
2863 * - (-ENOTSUP) if hardware doesn't support this function.
2864 * - (-EPERM) if not ran from the primary process.
2865 * - (-EIO) if re-initialisation failed or device is removed.
2866 * - (-ENOMEM) if the reset failed due to OOM.
2867 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2869 int rte_eth_dev_reset(uint16_t port_id);
2872 * Enable receipt in promiscuous mode for an Ethernet device.
2875 * The port identifier of the Ethernet device.
2877 * - (0) if successful.
2878 * - (-ENOTSUP) if support for promiscuous_enable() does not exist
2880 * - (-ENODEV) if *port_id* invalid.
2882 int rte_eth_promiscuous_enable(uint16_t port_id);
2885 * Disable receipt in promiscuous mode for an Ethernet device.
2888 * The port identifier of the Ethernet device.
2890 * - (0) if successful.
2891 * - (-ENOTSUP) if support for promiscuous_disable() does not exist
2893 * - (-ENODEV) if *port_id* invalid.
2895 int rte_eth_promiscuous_disable(uint16_t port_id);
2898 * Return the value of promiscuous mode for an Ethernet device.
2901 * The port identifier of the Ethernet device.
2903 * - (1) if promiscuous is enabled
2904 * - (0) if promiscuous is disabled.
2907 int rte_eth_promiscuous_get(uint16_t port_id);
2910 * Enable the receipt of any multicast frame by an Ethernet device.
2913 * The port identifier of the Ethernet device.
2915 * - (0) if successful.
2916 * - (-ENOTSUP) if support for allmulticast_enable() does not exist
2918 * - (-ENODEV) if *port_id* invalid.
2920 int rte_eth_allmulticast_enable(uint16_t port_id);
2923 * Disable the receipt of all multicast frames by an Ethernet device.
2926 * The port identifier of the Ethernet device.
2928 * - (0) if successful.
2929 * - (-ENOTSUP) if support for allmulticast_disable() does not exist
2931 * - (-ENODEV) if *port_id* invalid.
2933 int rte_eth_allmulticast_disable(uint16_t port_id);
2936 * Return the value of allmulticast mode for an Ethernet device.
2939 * The port identifier of the Ethernet device.
2941 * - (1) if allmulticast is enabled
2942 * - (0) if allmulticast is disabled.
2945 int rte_eth_allmulticast_get(uint16_t port_id);
2948 * Retrieve the link status (up/down), the duplex mode (half/full),
2949 * the negotiation (auto/fixed), and if available, the speed (Mbps).
2951 * It might need to wait up to 9 seconds.
2952 * @see rte_eth_link_get_nowait.
2955 * The port identifier of the Ethernet device.
2957 * Link information written back.
2959 * - (0) if successful.
2960 * - (-ENOTSUP) if the function is not supported in PMD driver.
2961 * - (-ENODEV) if *port_id* invalid.
2962 * - (-EINVAL) if bad parameter.
2964 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
2967 * Retrieve the link status (up/down), the duplex mode (half/full),
2968 * the negotiation (auto/fixed), and if available, the speed (Mbps).
2971 * The port identifier of the Ethernet device.
2973 * Link information written back.
2975 * - (0) if successful.
2976 * - (-ENOTSUP) if the function is not supported in PMD driver.
2977 * - (-ENODEV) if *port_id* invalid.
2978 * - (-EINVAL) if bad parameter.
2980 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
2984 * @b EXPERIMENTAL: this API may change without prior notice.
2986 * The function converts a link_speed to a string. It handles all special
2987 * values like unknown or none speed.
2990 * link_speed of rte_eth_link struct
2992 * Link speed in textual format. It's pointer to immutable memory.
2993 * No free is required.
2996 const char *rte_eth_link_speed_to_str(uint32_t link_speed);
3000 * @b EXPERIMENTAL: this API may change without prior notice.
3002 * The function converts a rte_eth_link struct representing a link status to
3006 * A pointer to a string to be filled with textual representation of
3007 * device status. At least RTE_ETH_LINK_MAX_STR_LEN bytes should be allocated to
3008 * store default link status text.
3010 * Length of available memory at 'str' string.
3012 * Link status returned by rte_eth_link_get function
3014 * Number of bytes written to str array or -EINVAL if bad parameter.
3017 int rte_eth_link_to_str(char *str, size_t len,
3018 const struct rte_eth_link *eth_link);
3021 * Retrieve the general I/O statistics of an Ethernet device.
3024 * The port identifier of the Ethernet device.
3026 * A pointer to a structure of type *rte_eth_stats* to be filled with
3027 * the values of device counters for the following set of statistics:
3028 * - *ipackets* with the total of successfully received packets.
3029 * - *opackets* with the total of successfully transmitted packets.
3030 * - *ibytes* with the total of successfully received bytes.
3031 * - *obytes* with the total of successfully transmitted bytes.
3032 * - *ierrors* with the total of erroneous received packets.
3033 * - *oerrors* with the total of failed transmitted packets.
3035 * Zero if successful. Non-zero otherwise.
3037 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
3040 * Reset the general I/O statistics of an Ethernet device.
3043 * The port identifier of the Ethernet device.
3045 * - (0) if device notified to reset stats.
3046 * - (-ENOTSUP) if hardware doesn't support.
3047 * - (-ENODEV) if *port_id* invalid.
3048 * - (<0): Error code of the driver stats reset function.
3050 int rte_eth_stats_reset(uint16_t port_id);
3053 * Retrieve names of extended statistics of an Ethernet device.
3055 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
3057 * xstats_names[i].name => xstats[i].value
3059 * And the array index is same with id field of 'struct rte_eth_xstat':
3062 * This assumption makes key-value pair matching less flexible but simpler.
3065 * The port identifier of the Ethernet device.
3066 * @param xstats_names
3067 * An rte_eth_xstat_name array of at least *size* elements to
3068 * be filled. If set to NULL, the function returns the required number
3071 * The size of the xstats_names array (number of elements).
3073 * - A positive value lower or equal to size: success. The return value
3074 * is the number of entries filled in the stats table.
3075 * - A positive value higher than size: error, the given statistics table
3076 * is too small. The return value corresponds to the size that should
3077 * be given to succeed. The entries in the table are not valid and
3078 * shall not be used by the caller.
3079 * - A negative value on error (invalid port ID).
3081 int rte_eth_xstats_get_names(uint16_t port_id,
3082 struct rte_eth_xstat_name *xstats_names,
3086 * Retrieve extended statistics of an Ethernet device.
3088 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
3090 * xstats_names[i].name => xstats[i].value
3092 * And the array index is same with id field of 'struct rte_eth_xstat':
3095 * This assumption makes key-value pair matching less flexible but simpler.
3098 * The port identifier of the Ethernet device.
3100 * A pointer to a table of structure of type *rte_eth_xstat*
3101 * to be filled with device statistics ids and values.
3102 * This parameter can be set to NULL if n is 0.
3104 * The size of the xstats array (number of elements).
3106 * - A positive value lower or equal to n: success. The return value
3107 * is the number of entries filled in the stats table.
3108 * - A positive value higher than n: error, the given statistics table
3109 * is too small. The return value corresponds to the size that should
3110 * be given to succeed. The entries in the table are not valid and
3111 * shall not be used by the caller.
3112 * - A negative value on error (invalid port ID).
3114 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
3118 * Retrieve names of extended statistics of an Ethernet device.
3121 * The port identifier of the Ethernet device.
3122 * @param xstats_names
3123 * Array to be filled in with names of requested device statistics.
3124 * Must not be NULL if @p ids are specified (not NULL).
3126 * Number of elements in @p xstats_names array (if not NULL) and in
3127 * @p ids array (if not NULL). Must be 0 if both array pointers are NULL.
3129 * IDs array given by app to retrieve specific statistics. May be NULL to
3130 * retrieve names of all available statistics or, if @p xstats_names is
3131 * NULL as well, just the number of available statistics.
3133 * - A positive value lower or equal to size: success. The return value
3134 * is the number of entries filled in the stats table.
3135 * - A positive value higher than size: success. The given statistics table
3136 * is too small. The return value corresponds to the size that should
3137 * be given to succeed. The entries in the table are not valid and
3138 * shall not be used by the caller.
3139 * - A negative value on error.
3142 rte_eth_xstats_get_names_by_id(uint16_t port_id,
3143 struct rte_eth_xstat_name *xstats_names, unsigned int size,
3147 * Retrieve extended statistics of an Ethernet device.
3150 * The port identifier of the Ethernet device.
3152 * IDs array given by app to retrieve specific statistics. May be NULL to
3153 * retrieve all available statistics or, if @p values is NULL as well,
3154 * just the number of available statistics.
3156 * Array to be filled in with requested device statistics.
3157 * Must not be NULL if ids are specified (not NULL).
3159 * Number of elements in @p values array (if not NULL) and in @p ids
3160 * array (if not NULL). Must be 0 if both array pointers are NULL.
3162 * - A positive value lower or equal to size: success. The return value
3163 * is the number of entries filled in the stats table.
3164 * - A positive value higher than size: success: The given statistics table
3165 * is too small. The return value corresponds to the size that should
3166 * be given to succeed. The entries in the table are not valid and
3167 * shall not be used by the caller.
3168 * - A negative value on error.
3170 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
3171 uint64_t *values, unsigned int size);
3174 * Gets the ID of a statistic from its name.
3176 * This function searches for the statistics using string compares, and
3177 * as such should not be used on the fast-path. For fast-path retrieval of
3178 * specific statistics, store the ID as provided in *id* from this function,
3179 * and pass the ID to rte_eth_xstats_get()
3181 * @param port_id The port to look up statistics from
3182 * @param xstat_name The name of the statistic to return
3183 * @param[out] id A pointer to an app-supplied uint64_t which should be
3184 * set to the ID of the stat if the stat exists.
3187 * -ENODEV for invalid port_id,
3188 * -EIO if device is removed,
3189 * -EINVAL if the xstat_name doesn't exist in port_id
3190 * -ENOMEM if bad parameter.
3192 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
3196 * Reset extended statistics of an Ethernet device.
3199 * The port identifier of the Ethernet device.
3201 * - (0) if device notified to reset extended stats.
3202 * - (-ENOTSUP) if pmd doesn't support both
3203 * extended stats and basic stats reset.
3204 * - (-ENODEV) if *port_id* invalid.
3205 * - (<0): Error code of the driver xstats reset function.
3207 int rte_eth_xstats_reset(uint16_t port_id);
3210 * Set a mapping for the specified transmit queue to the specified per-queue
3211 * statistics counter.
3214 * The port identifier of the Ethernet device.
3215 * @param tx_queue_id
3216 * The index of the transmit queue for which a queue stats mapping is required.
3217 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3218 * to rte_eth_dev_configure().
3220 * The per-queue packet statistics functionality number that the transmit
3221 * queue is to be assigned.
3222 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
3223 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256.
3225 * Zero if successful. Non-zero otherwise.
3227 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
3228 uint16_t tx_queue_id, uint8_t stat_idx);
3231 * Set a mapping for the specified receive queue to the specified per-queue
3232 * statistics counter.
3235 * The port identifier of the Ethernet device.
3236 * @param rx_queue_id
3237 * The index of the receive queue for which a queue stats mapping is required.
3238 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3239 * to rte_eth_dev_configure().
3241 * The per-queue packet statistics functionality number that the receive
3242 * queue is to be assigned.
3243 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
3244 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256.
3246 * Zero if successful. Non-zero otherwise.
3248 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
3249 uint16_t rx_queue_id,
3253 * Retrieve the Ethernet address of an Ethernet device.
3256 * The port identifier of the Ethernet device.
3258 * A pointer to a structure of type *ether_addr* to be filled with
3259 * the Ethernet address of the Ethernet device.
3261 * - (0) if successful
3262 * - (-ENODEV) if *port_id* invalid.
3263 * - (-EINVAL) if bad parameter.
3265 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr);
3269 * @b EXPERIMENTAL: this API may change without prior notice
3271 * Retrieve the Ethernet addresses of an Ethernet device.
3274 * The port identifier of the Ethernet device.
3276 * A pointer to an array of structures of type *ether_addr* to be filled with
3277 * the Ethernet addresses of the Ethernet device.
3279 * Number of elements in the @p ma array.
3280 * Note that rte_eth_dev_info::max_mac_addrs can be used to retrieve
3281 * max number of Ethernet addresses for given port.
3283 * - number of retrieved addresses if successful
3284 * - (-ENODEV) if *port_id* invalid.
3285 * - (-EINVAL) if bad parameter.
3288 int rte_eth_macaddrs_get(uint16_t port_id, struct rte_ether_addr *ma,
3292 * Retrieve the contextual information of an Ethernet device.
3294 * As part of this function, a number of of fields in dev_info will be
3295 * initialized as follows:
3300 * Where lim is defined within the rte_eth_dev_info_get as
3302 * const struct rte_eth_desc_lim lim = {
3303 * .nb_max = UINT16_MAX,
3306 * .nb_seg_max = UINT16_MAX,
3307 * .nb_mtu_seg_max = UINT16_MAX,
3310 * device = dev->device
3311 * min_mtu = RTE_ETHER_MIN_LEN - RTE_ETHER_HDR_LEN - RTE_ETHER_CRC_LEN
3312 * max_mtu = UINT16_MAX
3314 * The following fields will be populated if support for dev_infos_get()
3315 * exists for the device and the rte_eth_dev 'dev' has been populated
3316 * successfully with a call to it:
3318 * driver_name = dev->device->driver->name
3319 * nb_rx_queues = dev->data->nb_rx_queues
3320 * nb_tx_queues = dev->data->nb_tx_queues
3321 * dev_flags = &dev->data->dev_flags
3324 * The port identifier of the Ethernet device.
3326 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
3327 * the contextual information of the Ethernet device.
3329 * - (0) if successful.
3330 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device.
3331 * - (-ENODEV) if *port_id* invalid.
3332 * - (-EINVAL) if bad parameter.
3334 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
3338 * @b EXPERIMENTAL: this API may change without prior notice.
3340 * Retrieve the configuration of an Ethernet device.
3343 * The port identifier of the Ethernet device.
3345 * Location for Ethernet device configuration to be filled in.
3347 * - (0) if successful.
3348 * - (-ENODEV) if *port_id* invalid.
3349 * - (-EINVAL) if bad parameter.
3352 int rte_eth_dev_conf_get(uint16_t port_id, struct rte_eth_conf *dev_conf);
3355 * Retrieve the firmware version of a device.
3358 * The port identifier of the device.
3360 * A pointer to a string array storing the firmware version of a device,
3361 * the string includes terminating null. This pointer is allocated by caller.
3363 * The size of the string array pointed by fw_version, which should be
3364 * large enough to store firmware version of the device.
3366 * - (0) if successful.
3367 * - (-ENOTSUP) if operation is not supported.
3368 * - (-ENODEV) if *port_id* invalid.
3369 * - (-EIO) if device is removed.
3370 * - (-EINVAL) if bad parameter.
3371 * - (>0) if *fw_size* is not enough to store firmware version, return
3372 * the size of the non truncated string.
3374 int rte_eth_dev_fw_version_get(uint16_t port_id,
3375 char *fw_version, size_t fw_size);
3378 * Retrieve the supported packet types of an Ethernet device.
3380 * When a packet type is announced as supported, it *must* be recognized by
3381 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
3382 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
3383 * packet types for these packets:
3384 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
3385 * - Ether/VLAN/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
3386 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
3387 * - Ether/VLAN/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
3389 * When a packet is received by a PMD, the most precise type must be
3390 * returned among the ones supported. However a PMD is allowed to set
3391 * packet type that is not in the supported list, at the condition that it
3392 * is more precise. Therefore, a PMD announcing no supported packet types
3393 * can still set a matching packet type in a received packet.
3396 * Better to invoke this API after the device is already started or Rx burst
3397 * function is decided, to obtain correct supported ptypes.
3399 * if a given PMD does not report what ptypes it supports, then the supported
3400 * ptype count is reported as 0.
3402 * The port identifier of the Ethernet device.
3404 * A hint of what kind of packet type which the caller is interested in.
3406 * An array pointer to store adequate packet types, allocated by caller.
3408 * Size of the array pointed by param ptypes.
3410 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
3411 * only num entries will be filled into the ptypes array, but the full
3412 * count of supported ptypes will be returned.
3413 * - (-ENODEV) if *port_id* invalid.
3414 * - (-EINVAL) if bad parameter.
3416 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
3417 uint32_t *ptypes, int num);
3419 * Inform Ethernet device about reduced range of packet types to handle.
3421 * Application can use this function to set only specific ptypes that it's
3422 * interested. This information can be used by the PMD to optimize Rx path.
3424 * The function accepts an array `set_ptypes` allocated by the caller to
3425 * store the packet types set by the driver, the last element of the array
3426 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be
3427 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled
3431 * The port identifier of the Ethernet device.
3433 * The ptype family that application is interested in should be bitwise OR of
3434 * RTE_PTYPE_*_MASK or 0.
3436 * An array pointer to store set packet types, allocated by caller. The
3437 * function marks the end of array with RTE_PTYPE_UNKNOWN.
3439 * Size of the array pointed by param ptypes.
3440 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the
3444 * - (-ENODEV) if *port_id* invalid.
3445 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and
3448 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask,
3449 uint32_t *set_ptypes, unsigned int num);
3452 * Retrieve the MTU of an Ethernet device.
3455 * The port identifier of the Ethernet device.
3457 * A pointer to a uint16_t where the retrieved MTU is to be stored.
3459 * - (0) if successful.
3460 * - (-ENODEV) if *port_id* invalid.
3461 * - (-EINVAL) if bad parameter.
3463 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
3466 * Change the MTU of an Ethernet device.
3469 * The port identifier of the Ethernet device.
3471 * A uint16_t for the MTU to be applied.
3473 * - (0) if successful.
3474 * - (-ENOTSUP) if operation is not supported.
3475 * - (-ENODEV) if *port_id* invalid.
3476 * - (-EIO) if device is removed.
3477 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within
3478 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or
3479 * when the mtu is set using dev->dev_ops->mtu_set.
3480 * - (-EBUSY) if operation is not allowed when the port is running
3482 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
3485 * Enable/Disable hardware filtering by an Ethernet device of received
3486 * VLAN packets tagged with a given VLAN Tag Identifier.
3489 * The port identifier of the Ethernet device.
3491 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
3493 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
3494 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
3496 * - (0) if successful.
3497 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
3498 * - (-ENODEV) if *port_id* invalid.
3499 * - (-EIO) if device is removed.
3500 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
3501 * - (-EINVAL) if *vlan_id* > 4095.
3503 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
3506 * Enable/Disable hardware VLAN Strip by a Rx queue of an Ethernet device.
3509 * The port identifier of the Ethernet device.
3510 * @param rx_queue_id
3511 * The index of the receive queue for which a queue stats mapping is required.
3512 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3513 * to rte_eth_dev_configure().
3515 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
3516 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
3518 * - (0) if successful.
3519 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured.
3520 * - (-ENODEV) if *port_id* invalid.
3521 * - (-EINVAL) if *rx_queue_id* invalid.
3523 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
3527 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
3531 * The port identifier of the Ethernet device.
3535 * The Tag Protocol ID
3537 * - (0) if successful.
3538 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported.
3539 * - (-ENODEV) if *port_id* invalid.
3540 * - (-EIO) if device is removed.
3542 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
3543 enum rte_vlan_type vlan_type,
3547 * Set VLAN offload configuration on an Ethernet device.
3550 * The port identifier of the Ethernet device.
3551 * @param offload_mask
3552 * The VLAN Offload bit mask can be mixed use with "OR"
3553 * RTE_ETH_VLAN_STRIP_OFFLOAD
3554 * RTE_ETH_VLAN_FILTER_OFFLOAD
3555 * RTE_ETH_VLAN_EXTEND_OFFLOAD
3556 * RTE_ETH_QINQ_STRIP_OFFLOAD
3558 * - (0) if successful.
3559 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
3560 * - (-ENODEV) if *port_id* invalid.
3561 * - (-EIO) if device is removed.
3563 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
3566 * Read VLAN Offload configuration from an Ethernet device
3569 * The port identifier of the Ethernet device.
3571 * - (>0) if successful. Bit mask to indicate
3572 * RTE_ETH_VLAN_STRIP_OFFLOAD
3573 * RTE_ETH_VLAN_FILTER_OFFLOAD
3574 * RTE_ETH_VLAN_EXTEND_OFFLOAD
3575 * RTE_ETH_QINQ_STRIP_OFFLOAD
3576 * - (-ENODEV) if *port_id* invalid.
3578 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
3581 * Set port based Tx VLAN insertion on or off.
3584 * The port identifier of the Ethernet device.
3586 * Port based Tx VLAN identifier together with user priority.
3588 * Turn on or off the port based Tx VLAN insertion.
3591 * - (0) if successful.
3592 * - negative if failed.
3594 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
3596 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
3600 * Structure used to buffer packets for future Tx
3601 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
3603 struct rte_eth_dev_tx_buffer {
3604 buffer_tx_error_fn error_callback;
3605 void *error_userdata;
3606 uint16_t size; /**< Size of buffer for buffered Tx */
3607 uint16_t length; /**< Number of packets in the array */
3608 /** Pending packets to be sent on explicit flush or when full */
3609 struct rte_mbuf *pkts[];
3613 * Calculate the size of the Tx buffer.
3616 * Number of stored packets.
3618 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
3619 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
3622 * Initialize default values for buffered transmitting
3625 * Tx buffer to be initialized.
3632 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
3635 * Configure a callback for buffered packets which cannot be sent
3637 * Register a specific callback to be called when an attempt is made to send
3638 * all packets buffered on an Ethernet port, but not all packets can
3639 * successfully be sent. The callback registered here will be called only
3640 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
3641 * The default callback configured for each queue by default just frees the
3642 * packets back to the calling mempool. If additional behaviour is required,
3643 * for example, to count dropped packets, or to retry transmission of packets
3644 * which cannot be sent, this function should be used to register a suitable
3645 * callback function to implement the desired behaviour.
3646 * The example callback "rte_eth_count_unsent_packet_callback()" is also
3647 * provided as reference.
3650 * The port identifier of the Ethernet device.
3652 * The function to be used as the callback.
3654 * Arbitrary parameter to be passed to the callback function
3656 * 0 on success, or -EINVAL if bad parameter
3659 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
3660 buffer_tx_error_fn callback, void *userdata);
3663 * Callback function for silently dropping unsent buffered packets.
3665 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3666 * adjust the default behavior when buffered packets cannot be sent. This
3667 * function drops any unsent packets silently and is used by Tx buffered
3668 * operations as default behavior.
3670 * NOTE: this function should not be called directly, instead it should be used
3671 * as a callback for packet buffering.
3673 * NOTE: when configuring this function as a callback with
3674 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3675 * should point to an uint64_t value.
3678 * The previously buffered packets which could not be sent
3680 * The number of unsent packets in the pkts array
3685 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
3689 * Callback function for tracking unsent buffered packets.
3691 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3692 * adjust the default behavior when buffered packets cannot be sent. This
3693 * function drops any unsent packets, but also updates a user-supplied counter
3694 * to track the overall number of packets dropped. The counter should be an
3695 * uint64_t variable.
3697 * NOTE: this function should not be called directly, instead it should be used
3698 * as a callback for packet buffering.
3700 * NOTE: when configuring this function as a callback with
3701 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3702 * should point to an uint64_t value.
3705 * The previously buffered packets which could not be sent
3707 * The number of unsent packets in the pkts array
3709 * Pointer to an uint64_t value, which will be incremented by unsent
3712 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
3716 * Request the driver to free mbufs currently cached by the driver. The
3717 * driver will only free the mbuf if it is no longer in use. It is the
3718 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is
3722 * The port identifier of the Ethernet device.
3724 * The index of the transmit queue through which output packets must be
3726 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3727 * to rte_eth_dev_configure().
3729 * Maximum number of packets to free. Use 0 to indicate all possible packets
3730 * should be freed. Note that a packet may be using multiple mbufs.
3733 * -ENODEV: Invalid interface
3734 * -EIO: device is removed
3735 * -ENOTSUP: Driver does not support function
3737 * 0-n: Number of packets freed. More packets may still remain in ring that
3741 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
3744 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
3747 enum rte_eth_event_ipsec_subtype {
3748 /** Unknown event type */
3749 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
3750 /** Sequence number overflow */
3751 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
3752 /** Soft time expiry of SA */
3753 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
3754 /** Soft byte expiry of SA */
3755 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
3756 /** Max value of this enum */
3757 RTE_ETH_EVENT_IPSEC_MAX
3761 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
3762 * information of the IPsec offload event.
3764 struct rte_eth_event_ipsec_desc {
3765 /** Type of RTE_ETH_EVENT_IPSEC_* event */
3766 enum rte_eth_event_ipsec_subtype subtype;
3768 * Event specific metadata.
3770 * For the following events, *userdata* registered
3771 * with the *rte_security_session* would be returned
3774 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
3775 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
3776 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
3778 * @see struct rte_security_session_conf
3785 * The eth device event type for interrupt, and maybe others in the future.
3787 enum rte_eth_event_type {
3788 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3789 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3790 /** queue state event (enabled/disabled) */
3791 RTE_ETH_EVENT_QUEUE_STATE,
3792 /** reset interrupt event, sent to VF on PF reset */
3793 RTE_ETH_EVENT_INTR_RESET,
3794 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3795 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3796 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3797 RTE_ETH_EVENT_NEW, /**< port is probed */
3798 RTE_ETH_EVENT_DESTROY, /**< port is released */
3799 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
3800 RTE_ETH_EVENT_FLOW_AGED,/**< New aged-out flows is detected */
3801 RTE_ETH_EVENT_MAX /**< max value of this enum */
3804 /** User application callback to be registered for interrupts. */
3805 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3806 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3809 * Register a callback function for port event.
3813 * RTE_ETH_ALL means register the event for all port ids.
3817 * User supplied callback function to be called.
3819 * Pointer to the parameters for the registered callback.
3822 * - On success, zero.
3823 * - On failure, a negative value.
3825 int rte_eth_dev_callback_register(uint16_t port_id,
3826 enum rte_eth_event_type event,
3827 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3830 * Unregister a callback function for port event.
3834 * RTE_ETH_ALL means unregister the event for all port ids.
3838 * User supplied callback function to be called.
3840 * Pointer to the parameters for the registered callback. -1 means to
3841 * remove all for the same callback address and same event.
3844 * - On success, zero.
3845 * - On failure, a negative value.
3847 int rte_eth_dev_callback_unregister(uint16_t port_id,
3848 enum rte_eth_event_type event,
3849 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3852 * When there is no Rx packet coming in Rx Queue for a long time, we can
3853 * sleep lcore related to Rx Queue for power saving, and enable Rx interrupt
3854 * to be triggered when Rx packet arrives.
3856 * The rte_eth_dev_rx_intr_enable() function enables Rx queue
3857 * interrupt on specific Rx queue of a port.
3860 * The port identifier of the Ethernet device.
3862 * The index of the receive queue from which to retrieve input packets.
3863 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3864 * to rte_eth_dev_configure().
3866 * - (0) if successful.
3867 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3869 * - (-ENODEV) if *port_id* invalid.
3870 * - (-EIO) if device is removed.
3872 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
3875 * When lcore wakes up from Rx interrupt indicating packet coming, disable Rx
3876 * interrupt and returns to polling mode.
3878 * The rte_eth_dev_rx_intr_disable() function disables Rx queue
3879 * interrupt on specific Rx queue of a port.
3882 * The port identifier of the Ethernet device.
3884 * The index of the receive queue from which to retrieve input packets.
3885 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3886 * to rte_eth_dev_configure().
3888 * - (0) if successful.
3889 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3891 * - (-ENODEV) if *port_id* invalid.
3892 * - (-EIO) if device is removed.
3894 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
3897 * Rx Interrupt control per port.
3900 * The port identifier of the Ethernet device.
3902 * Epoll instance fd which the intr vector associated to.
3903 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3905 * The operation be performed for the vector.
3906 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3910 * - On success, zero.
3911 * - On failure, a negative value.
3913 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
3916 * Rx Interrupt control per queue.
3919 * The port identifier of the Ethernet device.
3921 * The index of the receive queue from which to retrieve input packets.
3922 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3923 * to rte_eth_dev_configure().
3925 * Epoll instance fd which the intr vector associated to.
3926 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3928 * The operation be performed for the vector.
3929 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3933 * - On success, zero.
3934 * - On failure, a negative value.
3936 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
3937 int epfd, int op, void *data);
3940 * Get interrupt fd per Rx queue.
3943 * The port identifier of the Ethernet device.
3945 * The index of the receive queue from which to retrieve input packets.
3946 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3947 * to rte_eth_dev_configure().
3949 * - (>=0) the interrupt fd associated to the requested Rx queue if
3954 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id);
3957 * Turn on the LED on the Ethernet device.
3958 * This function turns on the LED on the Ethernet device.
3961 * The port identifier of the Ethernet device.
3963 * - (0) if successful.
3964 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3966 * - (-ENODEV) if *port_id* invalid.
3967 * - (-EIO) if device is removed.
3969 int rte_eth_led_on(uint16_t port_id);
3972 * Turn off the LED on the Ethernet device.
3973 * This function turns off the LED on the Ethernet device.
3976 * The port identifier of the Ethernet device.
3978 * - (0) if successful.
3979 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3981 * - (-ENODEV) if *port_id* invalid.
3982 * - (-EIO) if device is removed.
3984 int rte_eth_led_off(uint16_t port_id);
3988 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
3990 * Get Forward Error Correction(FEC) capability.
3993 * The port identifier of the Ethernet device.
3994 * @param speed_fec_capa
3995 * speed_fec_capa is out only with per-speed capabilities.
3996 * If set to NULL, the function returns the required number
3997 * of required array entries.
3999 * a number of elements in an speed_fec_capa array.
4002 * - A non-negative value lower or equal to num: success. The return value
4003 * is the number of entries filled in the fec capa array.
4004 * - A non-negative value higher than num: error, the given fec capa array
4005 * is too small. The return value corresponds to the num that should
4006 * be given to succeed. The entries in fec capa array are not valid and
4007 * shall not be used by the caller.
4008 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4010 * - (-EIO) if device is removed.
4011 * - (-ENODEV) if *port_id* invalid.
4012 * - (-EINVAL) if *num* or *speed_fec_capa* invalid
4015 int rte_eth_fec_get_capability(uint16_t port_id,
4016 struct rte_eth_fec_capa *speed_fec_capa,
4021 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4023 * Get current Forward Error Correction(FEC) mode.
4024 * If link is down and AUTO is enabled, AUTO is returned, otherwise,
4025 * configured FEC mode is returned.
4026 * If link is up, current FEC mode is returned.
4029 * The port identifier of the Ethernet device.
4031 * A bitmask of enabled FEC modes. If AUTO bit is set, other
4032 * bits specify FEC modes which may be negotiated. If AUTO
4033 * bit is clear, specify FEC modes to be used (only one valid
4034 * mode per speed may be set).
4036 * - (0) if successful.
4037 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4039 * - (-EIO) if device is removed.
4040 * - (-ENODEV) if *port_id* invalid.
4043 int rte_eth_fec_get(uint16_t port_id, uint32_t *fec_capa);
4047 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4049 * Set Forward Error Correction(FEC) mode.
4052 * The port identifier of the Ethernet device.
4054 * A bitmask of allowed FEC modes. If AUTO bit is set, other
4055 * bits specify FEC modes which may be negotiated. If AUTO
4056 * bit is clear, specify FEC modes to be used (only one valid
4057 * mode per speed may be set).
4059 * - (0) if successful.
4060 * - (-EINVAL) if the FEC mode is not valid.
4061 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4062 * - (-EIO) if device is removed.
4063 * - (-ENODEV) if *port_id* invalid.
4066 int rte_eth_fec_set(uint16_t port_id, uint32_t fec_capa);
4069 * Get current status of the Ethernet link flow control for Ethernet device
4072 * The port identifier of the Ethernet device.
4074 * The pointer to the structure where to store the flow control parameters.
4076 * - (0) if successful.
4077 * - (-ENOTSUP) if hardware doesn't support flow control.
4078 * - (-ENODEV) if *port_id* invalid.
4079 * - (-EIO) if device is removed.
4080 * - (-EINVAL) if bad parameter.
4082 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
4083 struct rte_eth_fc_conf *fc_conf);
4086 * Configure the Ethernet link flow control for Ethernet device
4089 * The port identifier of the Ethernet device.
4091 * The pointer to the structure of the flow control parameters.
4093 * - (0) if successful.
4094 * - (-ENOTSUP) if hardware doesn't support flow control mode.
4095 * - (-ENODEV) if *port_id* invalid.
4096 * - (-EINVAL) if bad parameter
4097 * - (-EIO) if flow control setup failure or device is removed.
4099 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
4100 struct rte_eth_fc_conf *fc_conf);
4103 * Configure the Ethernet priority flow control under DCB environment
4104 * for Ethernet device.
4107 * The port identifier of the Ethernet device.
4109 * The pointer to the structure of the priority flow control parameters.
4111 * - (0) if successful.
4112 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
4113 * - (-ENODEV) if *port_id* invalid.
4114 * - (-EINVAL) if bad parameter
4115 * - (-EIO) if flow control setup failure or device is removed.
4117 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
4118 struct rte_eth_pfc_conf *pfc_conf);
4121 * Add a MAC address to the set used for filtering incoming packets.
4124 * The port identifier of the Ethernet device.
4126 * The MAC address to add.
4128 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
4129 * not enabled, this should be set to 0.
4131 * - (0) if successfully added or *mac_addr* was already added.
4132 * - (-ENOTSUP) if hardware doesn't support this feature.
4133 * - (-ENODEV) if *port* is invalid.
4134 * - (-EIO) if device is removed.
4135 * - (-ENOSPC) if no more MAC addresses can be added.
4136 * - (-EINVAL) if MAC address is invalid.
4138 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr,
4142 * Remove a MAC address from the internal array of addresses.
4145 * The port identifier of the Ethernet device.
4147 * MAC address to remove.
4149 * - (0) if successful, or *mac_addr* didn't exist.
4150 * - (-ENOTSUP) if hardware doesn't support.
4151 * - (-ENODEV) if *port* invalid.
4152 * - (-EADDRINUSE) if attempting to remove the default MAC address.
4153 * - (-EINVAL) if MAC address is invalid.
4155 int rte_eth_dev_mac_addr_remove(uint16_t port_id,
4156 struct rte_ether_addr *mac_addr);
4159 * Set the default MAC address.
4162 * The port identifier of the Ethernet device.
4164 * New default MAC address.
4166 * - (0) if successful, or *mac_addr* didn't exist.
4167 * - (-ENOTSUP) if hardware doesn't support.
4168 * - (-ENODEV) if *port* invalid.
4169 * - (-EINVAL) if MAC address is invalid.
4171 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
4172 struct rte_ether_addr *mac_addr);
4175 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
4178 * The port identifier of the Ethernet device.
4182 * Redirection table size. The table size can be queried by
4183 * rte_eth_dev_info_get().
4185 * - (0) if successful.
4186 * - (-ENODEV) if *port_id* is invalid.
4187 * - (-ENOTSUP) if hardware doesn't support.
4188 * - (-EINVAL) if bad parameter.
4189 * - (-EIO) if device is removed.
4191 int rte_eth_dev_rss_reta_update(uint16_t port_id,
4192 struct rte_eth_rss_reta_entry64 *reta_conf,
4193 uint16_t reta_size);
4196 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
4199 * The port identifier of the Ethernet device.
4201 * RETA to query. For each requested reta entry, corresponding bit
4202 * in mask must be set.
4204 * Redirection table size. The table size can be queried by
4205 * rte_eth_dev_info_get().
4207 * - (0) if successful.
4208 * - (-ENODEV) if *port_id* is invalid.
4209 * - (-ENOTSUP) if hardware doesn't support.
4210 * - (-EINVAL) if bad parameter.
4211 * - (-EIO) if device is removed.
4213 int rte_eth_dev_rss_reta_query(uint16_t port_id,
4214 struct rte_eth_rss_reta_entry64 *reta_conf,
4215 uint16_t reta_size);
4218 * Updates unicast hash table for receiving packet with the given destination
4219 * MAC address, and the packet is routed to all VFs for which the Rx mode is
4220 * accept packets that match the unicast hash table.
4223 * The port identifier of the Ethernet device.
4225 * Unicast MAC address.
4227 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
4228 * 0 - Clear an unicast hash bit.
4230 * - (0) if successful.
4231 * - (-ENOTSUP) if hardware doesn't support.
4232 * - (-ENODEV) if *port_id* invalid.
4233 * - (-EIO) if device is removed.
4234 * - (-EINVAL) if bad parameter.
4236 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr,
4240 * Updates all unicast hash bitmaps for receiving packet with any Unicast
4241 * Ethernet MAC addresses,the packet is routed to all VFs for which the Rx
4242 * mode is accept packets that match the unicast hash table.
4245 * The port identifier of the Ethernet device.
4247 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
4249 * 0 - Clear all unicast hash bitmaps
4251 * - (0) if successful.
4252 * - (-ENOTSUP) if hardware doesn't support.
4253 * - (-ENODEV) if *port_id* invalid.
4254 * - (-EIO) if device is removed.
4255 * - (-EINVAL) if bad parameter.
4257 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
4260 * Set the rate limitation for a queue on an Ethernet device.
4263 * The port identifier of the Ethernet device.
4267 * The Tx rate in Mbps. Allocated from the total port link speed.
4269 * - (0) if successful.
4270 * - (-ENOTSUP) if hardware doesn't support this feature.
4271 * - (-ENODEV) if *port_id* invalid.
4272 * - (-EIO) if device is removed.
4273 * - (-EINVAL) if bad parameter.
4275 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
4279 * Configuration of Receive Side Scaling hash computation of Ethernet device.
4282 * The port identifier of the Ethernet device.
4284 * The new configuration to use for RSS hash computation on the port.
4286 * - (0) if successful.
4287 * - (-ENODEV) if port identifier is invalid.
4288 * - (-EIO) if device is removed.
4289 * - (-ENOTSUP) if hardware doesn't support.
4290 * - (-EINVAL) if bad parameter.
4292 int rte_eth_dev_rss_hash_update(uint16_t port_id,
4293 struct rte_eth_rss_conf *rss_conf);
4296 * Retrieve current configuration of Receive Side Scaling hash computation
4297 * of Ethernet device.
4300 * The port identifier of the Ethernet device.
4302 * Where to store the current RSS hash configuration of the Ethernet device.
4304 * - (0) if successful.
4305 * - (-ENODEV) if port identifier is invalid.
4306 * - (-EIO) if device is removed.
4307 * - (-ENOTSUP) if hardware doesn't support RSS.
4308 * - (-EINVAL) if bad parameter.
4311 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
4312 struct rte_eth_rss_conf *rss_conf);
4315 * Add UDP tunneling port for a type of tunnel.
4317 * Some NICs may require such configuration to properly parse a tunnel
4318 * with any standard or custom UDP port.
4319 * The packets with this UDP port will be parsed for this type of tunnel.
4320 * The device parser will also check the rest of the tunnel headers
4321 * before classifying the packet.
4323 * With some devices, this API will affect packet classification, i.e.:
4324 * - mbuf.packet_type reported on Rx
4325 * - rte_flow rules with tunnel items
4328 * The port identifier of the Ethernet device.
4330 * UDP tunneling configuration.
4333 * - (0) if successful.
4334 * - (-ENODEV) if port identifier is invalid.
4335 * - (-EIO) if device is removed.
4336 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4339 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
4340 struct rte_eth_udp_tunnel *tunnel_udp);
4343 * Delete UDP tunneling port for a type of tunnel.
4345 * The packets with this UDP port will not be classified as this type of tunnel
4346 * anymore if the device use such mapping for tunnel packet classification.
4348 * @see rte_eth_dev_udp_tunnel_port_add
4351 * The port identifier of the Ethernet device.
4353 * UDP tunneling configuration.
4356 * - (0) if successful.
4357 * - (-ENODEV) if port identifier is invalid.
4358 * - (-EIO) if device is removed.
4359 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4362 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
4363 struct rte_eth_udp_tunnel *tunnel_udp);
4366 * Get DCB information on an Ethernet device.
4369 * The port identifier of the Ethernet device.
4373 * - (0) if successful.
4374 * - (-ENODEV) if port identifier is invalid.
4375 * - (-EIO) if device is removed.
4376 * - (-ENOTSUP) if hardware doesn't support.
4377 * - (-EINVAL) if bad parameter.
4379 int rte_eth_dev_get_dcb_info(uint16_t port_id,
4380 struct rte_eth_dcb_info *dcb_info);
4382 struct rte_eth_rxtx_callback;
4385 * Add a callback to be called on packet Rx on a given port and queue.
4387 * This API configures a function to be called for each burst of
4388 * packets received on a given NIC port queue. The return value is a pointer
4389 * that can be used to later remove the callback using
4390 * rte_eth_remove_rx_callback().
4392 * Multiple functions are called in the order that they are added.
4395 * The port identifier of the Ethernet device.
4397 * The queue on the Ethernet device on which the callback is to be added.
4399 * The callback function
4401 * A generic pointer parameter which will be passed to each invocation of the
4402 * callback function on this port and queue. Inter-thread synchronization
4403 * of any user data changes is the responsibility of the user.
4407 * On success, a pointer value which can later be used to remove the callback.
4409 const struct rte_eth_rxtx_callback *
4410 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
4411 rte_rx_callback_fn fn, void *user_param);
4414 * Add a callback that must be called first on packet Rx on a given port
4417 * This API configures a first function to be called for each burst of
4418 * packets received on a given NIC port queue. The return value is a pointer
4419 * that can be used to later remove the callback using
4420 * rte_eth_remove_rx_callback().
4422 * Multiple functions are called in the order that they are added.
4425 * The port identifier of the Ethernet device.
4427 * The queue on the Ethernet device on which the callback is to be added.
4429 * The callback function
4431 * A generic pointer parameter which will be passed to each invocation of the
4432 * callback function on this port and queue. Inter-thread synchronization
4433 * of any user data changes is the responsibility of the user.
4437 * On success, a pointer value which can later be used to remove the callback.
4439 const struct rte_eth_rxtx_callback *
4440 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
4441 rte_rx_callback_fn fn, void *user_param);
4444 * Add a callback to be called on packet Tx on a given port and queue.
4446 * This API configures a function to be called for each burst of
4447 * packets sent on a given NIC port queue. The return value is a pointer
4448 * that can be used to later remove the callback using
4449 * rte_eth_remove_tx_callback().
4451 * Multiple functions are called in the order that they are added.
4454 * The port identifier of the Ethernet device.
4456 * The queue on the Ethernet device on which the callback is to be added.
4458 * The callback function
4460 * A generic pointer parameter which will be passed to each invocation of the
4461 * callback function on this port and queue. Inter-thread synchronization
4462 * of any user data changes is the responsibility of the user.
4466 * On success, a pointer value which can later be used to remove the callback.
4468 const struct rte_eth_rxtx_callback *
4469 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
4470 rte_tx_callback_fn fn, void *user_param);
4473 * Remove an Rx packet callback from a given port and queue.
4475 * This function is used to removed callbacks that were added to a NIC port
4476 * queue using rte_eth_add_rx_callback().
4478 * Note: the callback is removed from the callback list but it isn't freed
4479 * since the it may still be in use. The memory for the callback can be
4480 * subsequently freed back by the application by calling rte_free():
4482 * - Immediately - if the port is stopped, or the user knows that no
4483 * callbacks are in flight e.g. if called from the thread doing Rx/Tx
4486 * - After a short delay - where the delay is sufficient to allow any
4487 * in-flight callbacks to complete. Alternately, the RCU mechanism can be
4488 * used to detect when data plane threads have ceased referencing the
4492 * The port identifier of the Ethernet device.
4494 * The queue on the Ethernet device from which the callback is to be removed.
4496 * User supplied callback created via rte_eth_add_rx_callback().
4499 * - 0: Success. Callback was removed.
4500 * - -ENODEV: If *port_id* is invalid.
4501 * - -ENOTSUP: Callback support is not available.
4502 * - -EINVAL: The queue_id is out of range, or the callback
4503 * is NULL or not found for the port/queue.
4505 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
4506 const struct rte_eth_rxtx_callback *user_cb);
4509 * Remove a Tx packet callback from a given port and queue.
4511 * This function is used to removed callbacks that were added to a NIC port
4512 * queue using rte_eth_add_tx_callback().
4514 * Note: the callback is removed from the callback list but it isn't freed
4515 * since the it may still be in use. The memory for the callback can be
4516 * subsequently freed back by the application by calling rte_free():
4518 * - Immediately - if the port is stopped, or the user knows that no
4519 * callbacks are in flight e.g. if called from the thread doing Rx/Tx
4522 * - After a short delay - where the delay is sufficient to allow any
4523 * in-flight callbacks to complete. Alternately, the RCU mechanism can be
4524 * used to detect when data plane threads have ceased referencing the
4528 * The port identifier of the Ethernet device.
4530 * The queue on the Ethernet device from which the callback is to be removed.
4532 * User supplied callback created via rte_eth_add_tx_callback().
4535 * - 0: Success. Callback was removed.
4536 * - -ENODEV: If *port_id* is invalid.
4537 * - -ENOTSUP: Callback support is not available.
4538 * - -EINVAL: The queue_id is out of range, or the callback
4539 * is NULL or not found for the port/queue.
4541 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
4542 const struct rte_eth_rxtx_callback *user_cb);
4545 * Retrieve information about given port's Rx queue.
4548 * The port identifier of the Ethernet device.
4550 * The Rx queue on the Ethernet device for which information
4551 * will be retrieved.
4553 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
4554 * the information of the Ethernet device.
4558 * - -ENODEV: If *port_id* is invalid.
4559 * - -ENOTSUP: routine is not supported by the device PMD.
4560 * - -EINVAL: The queue_id is out of range, or the queue
4563 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4564 struct rte_eth_rxq_info *qinfo);
4567 * Retrieve information about given port's Tx queue.
4570 * The port identifier of the Ethernet device.
4572 * The Tx queue on the Ethernet device for which information
4573 * will be retrieved.
4575 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
4576 * the information of the Ethernet device.
4580 * - -ENODEV: If *port_id* is invalid.
4581 * - -ENOTSUP: routine is not supported by the device PMD.
4582 * - -EINVAL: The queue_id is out of range, or the queue
4585 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4586 struct rte_eth_txq_info *qinfo);
4589 * Retrieve information about the Rx packet burst mode.
4592 * The port identifier of the Ethernet device.
4594 * The Rx queue on the Ethernet device for which information
4595 * will be retrieved.
4597 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4598 * with the information of the packet burst mode.
4602 * - -ENODEV: If *port_id* is invalid.
4603 * - -ENOTSUP: routine is not supported by the device PMD.
4604 * - -EINVAL: The queue_id is out of range.
4606 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4607 struct rte_eth_burst_mode *mode);
4610 * Retrieve information about the Tx packet burst mode.
4613 * The port identifier of the Ethernet device.
4615 * The Tx queue on the Ethernet device for which information
4616 * will be retrieved.
4618 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4619 * with the information of the packet burst mode.
4623 * - -ENODEV: If *port_id* is invalid.
4624 * - -ENOTSUP: routine is not supported by the device PMD.
4625 * - -EINVAL: The queue_id is out of range.
4627 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4628 struct rte_eth_burst_mode *mode);
4632 * @b EXPERIMENTAL: this API may change without prior notice.
4634 * Retrieve the monitor condition for a given receive queue.
4637 * The port identifier of the Ethernet device.
4639 * The Rx queue on the Ethernet device for which information
4640 * will be retrieved.
4642 * The pointer to power-optimized monitoring condition structure.
4646 * -ENOTSUP: Operation not supported.
4647 * -EINVAL: Invalid parameters.
4648 * -ENODEV: Invalid port ID.
4651 int rte_eth_get_monitor_addr(uint16_t port_id, uint16_t queue_id,
4652 struct rte_power_monitor_cond *pmc);
4655 * Retrieve device registers and register attributes (number of registers and
4659 * The port identifier of the Ethernet device.
4661 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
4662 * NULL the function fills in the width and length fields. If non-NULL
4663 * the registers are put into the buffer pointed at by the data field.
4665 * - (0) if successful.
4666 * - (-ENOTSUP) if hardware doesn't support.
4667 * - (-EINVAL) if bad parameter.
4668 * - (-ENODEV) if *port_id* invalid.
4669 * - (-EIO) if device is removed.
4670 * - others depends on the specific operations implementation.
4672 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
4675 * Retrieve size of device EEPROM
4678 * The port identifier of the Ethernet device.
4680 * - (>=0) EEPROM size if successful.
4681 * - (-ENOTSUP) if hardware doesn't support.
4682 * - (-ENODEV) if *port_id* invalid.
4683 * - (-EIO) if device is removed.
4684 * - others depends on the specific operations implementation.
4686 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
4689 * Retrieve EEPROM and EEPROM attribute
4692 * The port identifier of the Ethernet device.
4694 * The template includes buffer for return EEPROM data and
4695 * EEPROM attributes to be filled.
4697 * - (0) if successful.
4698 * - (-ENOTSUP) if hardware doesn't support.
4699 * - (-EINVAL) if bad parameter.
4700 * - (-ENODEV) if *port_id* invalid.
4701 * - (-EIO) if device is removed.
4702 * - others depends on the specific operations implementation.
4704 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4707 * Program EEPROM with provided data
4710 * The port identifier of the Ethernet device.
4712 * The template includes EEPROM data for programming and
4713 * EEPROM attributes to be filled
4715 * - (0) if successful.
4716 * - (-ENOTSUP) if hardware doesn't support.
4717 * - (-ENODEV) if *port_id* invalid.
4718 * - (-EINVAL) if bad parameter.
4719 * - (-EIO) if device is removed.
4720 * - others depends on the specific operations implementation.
4722 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4726 * @b EXPERIMENTAL: this API may change without prior notice.
4728 * Retrieve the type and size of plugin module EEPROM
4731 * The port identifier of the Ethernet device.
4733 * The type and size of plugin module EEPROM.
4735 * - (0) if successful.
4736 * - (-ENOTSUP) if hardware doesn't support.
4737 * - (-ENODEV) if *port_id* invalid.
4738 * - (-EINVAL) if bad parameter.
4739 * - (-EIO) if device is removed.
4740 * - others depends on the specific operations implementation.
4744 rte_eth_dev_get_module_info(uint16_t port_id,
4745 struct rte_eth_dev_module_info *modinfo);
4749 * @b EXPERIMENTAL: this API may change without prior notice.
4751 * Retrieve the data of plugin module EEPROM
4754 * The port identifier of the Ethernet device.
4756 * The template includes the plugin module EEPROM attributes, and the
4757 * buffer for return plugin module EEPROM data.
4759 * - (0) if successful.
4760 * - (-ENOTSUP) if hardware doesn't support.
4761 * - (-EINVAL) if bad parameter.
4762 * - (-ENODEV) if *port_id* invalid.
4763 * - (-EIO) if device is removed.
4764 * - others depends on the specific operations implementation.
4768 rte_eth_dev_get_module_eeprom(uint16_t port_id,
4769 struct rte_dev_eeprom_info *info);
4772 * Set the list of multicast addresses to filter on an Ethernet device.
4775 * The port identifier of the Ethernet device.
4776 * @param mc_addr_set
4777 * The array of multicast addresses to set. Equal to NULL when the function
4778 * is invoked to flush the set of filtered addresses.
4780 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
4781 * when the function is invoked to flush the set of filtered addresses.
4783 * - (0) if successful.
4784 * - (-ENODEV) if *port_id* invalid.
4785 * - (-EIO) if device is removed.
4786 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
4787 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
4788 * - (-EINVAL) if bad parameter.
4790 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
4791 struct rte_ether_addr *mc_addr_set,
4792 uint32_t nb_mc_addr);
4795 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
4798 * The port identifier of the Ethernet device.
4802 * - -ENODEV: The port ID is invalid.
4803 * - -EIO: if device is removed.
4804 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4806 int rte_eth_timesync_enable(uint16_t port_id);
4809 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
4812 * The port identifier of the Ethernet device.
4816 * - -ENODEV: The port ID is invalid.
4817 * - -EIO: if device is removed.
4818 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4820 int rte_eth_timesync_disable(uint16_t port_id);
4823 * Read an IEEE1588/802.1AS Rx timestamp from an Ethernet device.
4826 * The port identifier of the Ethernet device.
4828 * Pointer to the timestamp struct.
4830 * Device specific flags. Used to pass the Rx timesync register index to
4831 * i40e. Unused in igb/ixgbe, pass 0 instead.
4835 * - -EINVAL: No timestamp is available.
4836 * - -ENODEV: The port ID is invalid.
4837 * - -EIO: if device is removed.
4838 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4840 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
4841 struct timespec *timestamp, uint32_t flags);
4844 * Read an IEEE1588/802.1AS Tx timestamp from an Ethernet device.
4847 * The port identifier of the Ethernet device.
4849 * Pointer to the timestamp struct.
4853 * - -EINVAL: No timestamp is available.
4854 * - -ENODEV: The port ID is invalid.
4855 * - -EIO: if device is removed.
4856 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4858 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
4859 struct timespec *timestamp);
4862 * Adjust the timesync clock on an Ethernet device.
4864 * This is usually used in conjunction with other Ethdev timesync functions to
4865 * synchronize the device time using the IEEE1588/802.1AS protocol.
4868 * The port identifier of the Ethernet device.
4870 * The adjustment in nanoseconds.
4874 * - -ENODEV: The port ID is invalid.
4875 * - -EIO: if device is removed.
4876 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4878 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
4881 * Read the time from the timesync clock on an Ethernet device.
4883 * This is usually used in conjunction with other Ethdev timesync functions to
4884 * synchronize the device time using the IEEE1588/802.1AS protocol.
4887 * The port identifier of the Ethernet device.
4889 * Pointer to the timespec struct that holds the time.
4893 * - -EINVAL: Bad parameter.
4895 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
4898 * Set the time of the timesync clock on an Ethernet device.
4900 * This is usually used in conjunction with other Ethdev timesync functions to
4901 * synchronize the device time using the IEEE1588/802.1AS protocol.
4904 * The port identifier of the Ethernet device.
4906 * Pointer to the timespec struct that holds the time.
4910 * - -EINVAL: No timestamp is available.
4911 * - -ENODEV: The port ID is invalid.
4912 * - -EIO: if device is removed.
4913 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4915 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
4919 * @b EXPERIMENTAL: this API may change without prior notice.
4921 * Read the current clock counter of an Ethernet device
4923 * This returns the current raw clock value of an Ethernet device. It is
4924 * a raw amount of ticks, with no given time reference.
4925 * The value returned here is from the same clock than the one
4926 * filling timestamp field of Rx packets when using hardware timestamp
4927 * offload. Therefore it can be used to compute a precise conversion of
4928 * the device clock to the real time.
4930 * E.g, a simple heuristic to derivate the frequency would be:
4931 * uint64_t start, end;
4932 * rte_eth_read_clock(port, start);
4933 * rte_delay_ms(100);
4934 * rte_eth_read_clock(port, end);
4935 * double freq = (end - start) * 10;
4937 * Compute a common reference with:
4938 * uint64_t base_time_sec = current_time();
4939 * uint64_t base_clock;
4940 * rte_eth_read_clock(port, base_clock);
4942 * Then, convert the raw mbuf timestamp with:
4943 * base_time_sec + (double)(*timestamp_dynfield(mbuf) - base_clock) / freq;
4945 * This simple example will not provide a very good accuracy. One must
4946 * at least measure multiple times the frequency and do a regression.
4947 * To avoid deviation from the system time, the common reference can
4948 * be repeated from time to time. The integer division can also be
4949 * converted by a multiplication and a shift for better performance.
4952 * The port identifier of the Ethernet device.
4954 * Pointer to the uint64_t that holds the raw clock value.
4958 * - -ENODEV: The port ID is invalid.
4959 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4960 * - -EINVAL: if bad parameter.
4964 rte_eth_read_clock(uint16_t port_id, uint64_t *clock);
4967 * Get the port ID from device name. The device name should be specified
4969 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0
4970 * - SoC device name, for example- fsl-gmac0
4971 * - vdev dpdk name, for example- net_[pcap0|null0|tap0]
4974 * pci address or name of the device
4976 * pointer to port identifier of the device
4978 * - (0) if successful and port_id is filled.
4979 * - (-ENODEV or -EINVAL) on failure.
4982 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
4985 * Get the device name from port ID. The device name is specified as below:
4986 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0
4987 * - SoC device name, for example- fsl-gmac0
4988 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0]
4991 * Port identifier of the device.
4993 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
4995 * - (0) if successful.
4996 * - (-ENODEV) if *port_id* is invalid.
4997 * - (-EINVAL) on failure.
5000 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
5003 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
5004 * the Ethernet device information, otherwise adjust them to boundaries.
5007 * The port identifier of the Ethernet device.
5009 * A pointer to a uint16_t where the number of receive
5010 * descriptors stored.
5012 * A pointer to a uint16_t where the number of transmit
5013 * descriptors stored.
5015 * - (0) if successful.
5016 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
5018 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
5019 uint16_t *nb_rx_desc,
5020 uint16_t *nb_tx_desc);
5023 * Test if a port supports specific mempool ops.
5026 * Port identifier of the Ethernet device.
5028 * The name of the pool operations to test.
5030 * - 0: best mempool ops choice for this port.
5031 * - 1: mempool ops are supported for this port.
5032 * - -ENOTSUP: mempool ops not supported for this port.
5033 * - -ENODEV: Invalid port Identifier.
5034 * - -EINVAL: Pool param is null.
5037 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
5040 * Get the security context for the Ethernet device.
5043 * Port identifier of the Ethernet device
5046 * - pointer to security context on success.
5049 rte_eth_dev_get_sec_ctx(uint16_t port_id);
5053 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
5055 * Query the device hairpin capabilities.
5058 * The port identifier of the Ethernet device.
5060 * Pointer to a structure that will hold the hairpin capabilities.
5062 * - (0) if successful.
5063 * - (-ENOTSUP) if hardware doesn't support.
5064 * - (-EINVAL) if bad parameter.
5067 int rte_eth_dev_hairpin_capability_get(uint16_t port_id,
5068 struct rte_eth_hairpin_cap *cap);
5072 * @b EXPERIMENTAL: this structure may change without prior notice.
5074 * Ethernet device representor ID range entry
5076 struct rte_eth_representor_range {
5077 enum rte_eth_representor_type type; /**< Representor type */
5078 int controller; /**< Controller index */
5079 int pf; /**< Physical function index */
5082 int vf; /**< VF start index */
5083 int sf; /**< SF start index */
5085 uint32_t id_base; /**< Representor ID start index */
5086 uint32_t id_end; /**< Representor ID end index */
5087 char name[RTE_DEV_NAME_MAX_LEN]; /**< Representor name */
5092 * @b EXPERIMENTAL: this structure may change without prior notice.
5094 * Ethernet device representor information
5096 struct rte_eth_representor_info {
5097 uint16_t controller; /**< Controller ID of caller device. */
5098 uint16_t pf; /**< Physical function ID of caller device. */
5099 uint32_t nb_ranges_alloc; /**< Size of the ranges array. */
5100 uint32_t nb_ranges; /**< Number of initialized ranges. */
5101 struct rte_eth_representor_range ranges[];/**< Representor ID range. */
5105 * Retrieve the representor info of the device.
5107 * Get device representor info to be able to calculate a unique
5108 * representor ID. @see rte_eth_representor_id_get helper.
5111 * The port identifier of the device.
5113 * A pointer to a representor info structure.
5114 * NULL to return number of range entries and allocate memory
5115 * for next call to store detail.
5116 * The number of ranges that were written into this structure
5117 * will be placed into its nb_ranges field. This number cannot be
5118 * larger than the nb_ranges_alloc that by the user before calling
5119 * this function. It can be smaller than the value returned by the
5120 * function, however.
5122 * - (-ENOTSUP) if operation is not supported.
5123 * - (-ENODEV) if *port_id* invalid.
5124 * - (-EIO) if device is removed.
5125 * - (>=0) number of available representor range entries.
5128 int rte_eth_representor_info_get(uint16_t port_id,
5129 struct rte_eth_representor_info *info);
5131 /** The NIC is able to deliver flag (if set) with packets to the PMD. */
5132 #define RTE_ETH_RX_METADATA_USER_FLAG RTE_BIT64(0)
5134 /** The NIC is able to deliver mark ID with packets to the PMD. */
5135 #define RTE_ETH_RX_METADATA_USER_MARK RTE_BIT64(1)
5137 /** The NIC is able to deliver tunnel ID with packets to the PMD. */
5138 #define RTE_ETH_RX_METADATA_TUNNEL_ID RTE_BIT64(2)
5142 * @b EXPERIMENTAL: this API may change without prior notice
5144 * Negotiate the NIC's ability to deliver specific kinds of metadata to the PMD.
5146 * Invoke this API before the first rte_eth_dev_configure() invocation
5147 * to let the PMD make preparations that are inconvenient to do later.
5149 * The negotiation process is as follows:
5151 * - the application requests features intending to use at least some of them;
5152 * - the PMD responds with the guaranteed subset of the requested feature set;
5153 * - the application can retry negotiation with another set of features;
5154 * - the application can pass zero to clear the negotiation result;
5155 * - the last negotiated result takes effect upon
5156 * the ethdev configure and start.
5159 * The PMD is supposed to first consider enabling the requested feature set
5160 * in its entirety. Only if it fails to do so, does it have the right to
5161 * respond with a smaller set of the originally requested features.
5164 * Return code (-ENOTSUP) does not necessarily mean that the requested
5165 * features are unsupported. In this case, the application should just
5166 * assume that these features can be used without prior negotiations.
5169 * Port (ethdev) identifier
5171 * @param[inout] features
5172 * Feature selection buffer
5175 * - (-EBUSY) if the port can't handle this in its current state;
5176 * - (-ENOTSUP) if the method itself is not supported by the PMD;
5177 * - (-ENODEV) if *port_id* is invalid;
5178 * - (-EINVAL) if *features* is NULL;
5179 * - (-EIO) if the device is removed;
5183 int rte_eth_rx_metadata_negotiate(uint16_t port_id, uint64_t *features);
5185 #include <rte_ethdev_core.h>
5189 * Helper routine for rte_eth_rx_burst().
5190 * Should be called at exit from PMD's rte_eth_rx_bulk implementation.
5191 * Does necessary post-processing - invokes Rx callbacks if any, etc.
5194 * The port identifier of the Ethernet device.
5196 * The index of the receive queue from which to retrieve input packets.
5198 * The address of an array of pointers to *rte_mbuf* structures that
5199 * have been retrieved from the device.
5201 * The number of packets that were retrieved from the device.
5203 * The number of elements in @p rx_pkts array.
5205 * Opaque pointer of Rx queue callback related data.
5208 * The number of packets effectively supplied to the @p rx_pkts array.
5210 uint16_t rte_eth_call_rx_callbacks(uint16_t port_id, uint16_t queue_id,
5211 struct rte_mbuf **rx_pkts, uint16_t nb_rx, uint16_t nb_pkts,
5216 * Retrieve a burst of input packets from a receive queue of an Ethernet
5217 * device. The retrieved packets are stored in *rte_mbuf* structures whose
5218 * pointers are supplied in the *rx_pkts* array.
5220 * The rte_eth_rx_burst() function loops, parsing the Rx ring of the
5221 * receive queue, up to *nb_pkts* packets, and for each completed Rx
5222 * descriptor in the ring, it performs the following operations:
5224 * - Initialize the *rte_mbuf* data structure associated with the
5225 * Rx descriptor according to the information provided by the NIC into
5226 * that Rx descriptor.
5228 * - Store the *rte_mbuf* data structure into the next entry of the
5231 * - Replenish the Rx descriptor with a new *rte_mbuf* buffer
5232 * allocated from the memory pool associated with the receive queue at
5233 * initialization time.
5235 * When retrieving an input packet that was scattered by the controller
5236 * into multiple receive descriptors, the rte_eth_rx_burst() function
5237 * appends the associated *rte_mbuf* buffers to the first buffer of the
5240 * The rte_eth_rx_burst() function returns the number of packets
5241 * actually retrieved, which is the number of *rte_mbuf* data structures
5242 * effectively supplied into the *rx_pkts* array.
5243 * A return value equal to *nb_pkts* indicates that the Rx queue contained
5244 * at least *rx_pkts* packets, and this is likely to signify that other
5245 * received packets remain in the input queue. Applications implementing
5246 * a "retrieve as much received packets as possible" policy can check this
5247 * specific case and keep invoking the rte_eth_rx_burst() function until
5248 * a value less than *nb_pkts* is returned.
5250 * This receive method has the following advantages:
5252 * - It allows a run-to-completion network stack engine to retrieve and
5253 * to immediately process received packets in a fast burst-oriented
5254 * approach, avoiding the overhead of unnecessary intermediate packet
5255 * queue/dequeue operations.
5257 * - Conversely, it also allows an asynchronous-oriented processing
5258 * method to retrieve bursts of received packets and to immediately
5259 * queue them for further parallel processing by another logical core,
5260 * for instance. However, instead of having received packets being
5261 * individually queued by the driver, this approach allows the caller
5262 * of the rte_eth_rx_burst() function to queue a burst of retrieved
5263 * packets at a time and therefore dramatically reduce the cost of
5264 * enqueue/dequeue operations per packet.
5266 * - It allows the rte_eth_rx_burst() function of the driver to take
5267 * advantage of burst-oriented hardware features (CPU cache,
5268 * prefetch instructions, and so on) to minimize the number of CPU
5269 * cycles per packet.
5271 * To summarize, the proposed receive API enables many
5272 * burst-oriented optimizations in both synchronous and asynchronous
5273 * packet processing environments with no overhead in both cases.
5276 * Some drivers using vector instructions require that *nb_pkts* is
5277 * divisible by 4 or 8, depending on the driver implementation.
5279 * The rte_eth_rx_burst() function does not provide any error
5280 * notification to avoid the corresponding overhead. As a hint, the
5281 * upper-level application might check the status of the device link once
5282 * being systematically returned a 0 value for a given number of tries.
5285 * The port identifier of the Ethernet device.
5287 * The index of the receive queue from which to retrieve input packets.
5288 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
5289 * to rte_eth_dev_configure().
5291 * The address of an array of pointers to *rte_mbuf* structures that
5292 * must be large enough to store *nb_pkts* pointers in it.
5294 * The maximum number of packets to retrieve.
5295 * The value must be divisible by 8 in order to work with any driver.
5297 * The number of packets actually retrieved, which is the number
5298 * of pointers to *rte_mbuf* structures effectively supplied to the
5301 static inline uint16_t
5302 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
5303 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
5306 struct rte_eth_fp_ops *p;
5309 #ifdef RTE_ETHDEV_DEBUG_RX
5310 if (port_id >= RTE_MAX_ETHPORTS ||
5311 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5313 "Invalid port_id=%u or queue_id=%u\n",
5319 /* fetch pointer to queue data */
5320 p = &rte_eth_fp_ops[port_id];
5321 qd = p->rxq.data[queue_id];
5323 #ifdef RTE_ETHDEV_DEBUG_RX
5324 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
5327 RTE_ETHDEV_LOG(ERR, "Invalid Rx queue_id=%u for port_id=%u\n",
5333 nb_rx = p->rx_pkt_burst(qd, rx_pkts, nb_pkts);
5335 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
5339 /* __ATOMIC_RELEASE memory order was used when the
5340 * call back was inserted into the list.
5341 * Since there is a clear dependency between loading
5342 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
5345 cb = __atomic_load_n((void **)&p->rxq.clbk[queue_id],
5347 if (unlikely(cb != NULL))
5348 nb_rx = rte_eth_call_rx_callbacks(port_id, queue_id,
5349 rx_pkts, nb_rx, nb_pkts, cb);
5353 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx);
5358 * Get the number of used descriptors of a Rx queue
5361 * The port identifier of the Ethernet device.
5363 * The queue ID on the specific port.
5365 * The number of used descriptors in the specific queue, or:
5366 * - (-ENODEV) if *port_id* is invalid.
5367 * (-EINVAL) if *queue_id* is invalid
5368 * (-ENOTSUP) if the device does not support this function
5371 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
5373 struct rte_eth_fp_ops *p;
5376 if (port_id >= RTE_MAX_ETHPORTS ||
5377 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5379 "Invalid port_id=%u or queue_id=%u\n",
5384 /* fetch pointer to queue data */
5385 p = &rte_eth_fp_ops[port_id];
5386 qd = p->rxq.data[queue_id];
5388 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5389 RTE_FUNC_PTR_OR_ERR_RET(*p->rx_queue_count, -ENOTSUP);
5393 return (int)(*p->rx_queue_count)(qd);
5396 /**@{@name Rx hardware descriptor states
5397 * @see rte_eth_rx_descriptor_status
5399 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
5400 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
5401 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
5405 * Check the status of a Rx descriptor in the queue
5407 * It should be called in a similar context than the Rx function:
5408 * - on a dataplane core
5409 * - not concurrently on the same queue
5411 * Since it's a dataplane function, no check is performed on port_id and
5412 * queue_id. The caller must therefore ensure that the port is enabled
5413 * and the queue is configured and running.
5415 * Note: accessing to a random descriptor in the ring may trigger cache
5416 * misses and have a performance impact.
5419 * A valid port identifier of the Ethernet device which.
5421 * A valid Rx queue identifier on this port.
5423 * The offset of the descriptor starting from tail (0 is the next
5424 * packet to be received by the driver).
5427 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
5429 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
5430 * not yet processed by the driver (i.e. in the receive queue).
5431 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
5432 * the driver and not yet returned to hw, or reserved by the hw.
5433 * - (-EINVAL) bad descriptor offset.
5434 * - (-ENOTSUP) if the device does not support this function.
5435 * - (-ENODEV) bad port or queue (only if compiled with debug).
5438 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
5441 struct rte_eth_fp_ops *p;
5444 #ifdef RTE_ETHDEV_DEBUG_RX
5445 if (port_id >= RTE_MAX_ETHPORTS ||
5446 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5448 "Invalid port_id=%u or queue_id=%u\n",
5454 /* fetch pointer to queue data */
5455 p = &rte_eth_fp_ops[port_id];
5456 qd = p->rxq.data[queue_id];
5458 #ifdef RTE_ETHDEV_DEBUG_RX
5459 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5463 RTE_FUNC_PTR_OR_ERR_RET(*p->rx_descriptor_status, -ENOTSUP);
5464 return (*p->rx_descriptor_status)(qd, offset);
5467 /**@{@name Tx hardware descriptor states
5468 * @see rte_eth_tx_descriptor_status
5470 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
5471 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
5472 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
5476 * Check the status of a Tx descriptor in the queue.
5478 * It should be called in a similar context than the Tx function:
5479 * - on a dataplane core
5480 * - not concurrently on the same queue
5482 * Since it's a dataplane function, no check is performed on port_id and
5483 * queue_id. The caller must therefore ensure that the port is enabled
5484 * and the queue is configured and running.
5486 * Note: accessing to a random descriptor in the ring may trigger cache
5487 * misses and have a performance impact.
5490 * A valid port identifier of the Ethernet device which.
5492 * A valid Tx queue identifier on this port.
5494 * The offset of the descriptor starting from tail (0 is the place where
5495 * the next packet will be send).
5498 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
5499 * in the transmit queue.
5500 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
5501 * be reused by the driver.
5502 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
5503 * driver or the hardware.
5504 * - (-EINVAL) bad descriptor offset.
5505 * - (-ENOTSUP) if the device does not support this function.
5506 * - (-ENODEV) bad port or queue (only if compiled with debug).
5508 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
5509 uint16_t queue_id, uint16_t offset)
5511 struct rte_eth_fp_ops *p;
5514 #ifdef RTE_ETHDEV_DEBUG_TX
5515 if (port_id >= RTE_MAX_ETHPORTS ||
5516 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5518 "Invalid port_id=%u or queue_id=%u\n",
5524 /* fetch pointer to queue data */
5525 p = &rte_eth_fp_ops[port_id];
5526 qd = p->txq.data[queue_id];
5528 #ifdef RTE_ETHDEV_DEBUG_TX
5529 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5533 RTE_FUNC_PTR_OR_ERR_RET(*p->tx_descriptor_status, -ENOTSUP);
5534 return (*p->tx_descriptor_status)(qd, offset);
5539 * Helper routine for rte_eth_tx_burst().
5540 * Should be called before entry PMD's rte_eth_tx_bulk implementation.
5541 * Does necessary pre-processing - invokes Tx callbacks if any, etc.
5544 * The port identifier of the Ethernet device.
5546 * The index of the transmit queue through which output packets must be
5549 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5550 * which contain the output packets.
5552 * The maximum number of packets to transmit.
5554 * The number of output packets to transmit.
5556 uint16_t rte_eth_call_tx_callbacks(uint16_t port_id, uint16_t queue_id,
5557 struct rte_mbuf **tx_pkts, uint16_t nb_pkts, void *opaque);
5560 * Send a burst of output packets on a transmit queue of an Ethernet device.
5562 * The rte_eth_tx_burst() function is invoked to transmit output packets
5563 * on the output queue *queue_id* of the Ethernet device designated by its
5565 * The *nb_pkts* parameter is the number of packets to send which are
5566 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
5567 * allocated from a pool created with rte_pktmbuf_pool_create().
5568 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
5569 * up to the number of transmit descriptors available in the Tx ring of the
5571 * For each packet to send, the rte_eth_tx_burst() function performs
5572 * the following operations:
5574 * - Pick up the next available descriptor in the transmit ring.
5576 * - Free the network buffer previously sent with that descriptor, if any.
5578 * - Initialize the transmit descriptor with the information provided
5579 * in the *rte_mbuf data structure.
5581 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
5582 * the rte_eth_tx_burst() function uses several transmit descriptors
5585 * The rte_eth_tx_burst() function returns the number of packets it
5586 * actually sent. A return value equal to *nb_pkts* means that all packets
5587 * have been sent, and this is likely to signify that other output packets
5588 * could be immediately transmitted again. Applications that implement a
5589 * "send as many packets to transmit as possible" policy can check this
5590 * specific case and keep invoking the rte_eth_tx_burst() function until
5591 * a value less than *nb_pkts* is returned.
5593 * It is the responsibility of the rte_eth_tx_burst() function to
5594 * transparently free the memory buffers of packets previously sent.
5595 * This feature is driven by the *tx_free_thresh* value supplied to the
5596 * rte_eth_dev_configure() function at device configuration time.
5597 * When the number of free Tx descriptors drops below this threshold, the
5598 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
5599 * of those packets whose transmission was effectively completed.
5601 * If the PMD is RTE_ETH_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
5602 * invoke this function concurrently on the same Tx queue without SW lock.
5603 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads
5605 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
5609 * The port identifier of the Ethernet device.
5611 * The index of the transmit queue through which output packets must be
5613 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5614 * to rte_eth_dev_configure().
5616 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5617 * which contain the output packets.
5619 * The maximum number of packets to transmit.
5621 * The number of output packets actually stored in transmit descriptors of
5622 * the transmit ring. The return value can be less than the value of the
5623 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
5625 static inline uint16_t
5626 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
5627 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5629 struct rte_eth_fp_ops *p;
5632 #ifdef RTE_ETHDEV_DEBUG_TX
5633 if (port_id >= RTE_MAX_ETHPORTS ||
5634 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5636 "Invalid port_id=%u or queue_id=%u\n",
5642 /* fetch pointer to queue data */
5643 p = &rte_eth_fp_ops[port_id];
5644 qd = p->txq.data[queue_id];
5646 #ifdef RTE_ETHDEV_DEBUG_TX
5647 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
5650 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n",
5656 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
5660 /* __ATOMIC_RELEASE memory order was used when the
5661 * call back was inserted into the list.
5662 * Since there is a clear dependency between loading
5663 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
5666 cb = __atomic_load_n((void **)&p->txq.clbk[queue_id],
5668 if (unlikely(cb != NULL))
5669 nb_pkts = rte_eth_call_tx_callbacks(port_id, queue_id,
5670 tx_pkts, nb_pkts, cb);
5674 nb_pkts = p->tx_pkt_burst(qd, tx_pkts, nb_pkts);
5676 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts, nb_pkts);
5681 * Process a burst of output packets on a transmit queue of an Ethernet device.
5683 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
5684 * transmitted on the output queue *queue_id* of the Ethernet device designated
5686 * The *nb_pkts* parameter is the number of packets to be prepared which are
5687 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
5688 * allocated from a pool created with rte_pktmbuf_pool_create().
5689 * For each packet to send, the rte_eth_tx_prepare() function performs
5690 * the following operations:
5692 * - Check if packet meets devices requirements for Tx offloads.
5694 * - Check limitations about number of segments.
5696 * - Check additional requirements when debug is enabled.
5698 * - Update and/or reset required checksums when Tx offload is set for packet.
5700 * Since this function can modify packet data, provided mbufs must be safely
5701 * writable (e.g. modified data cannot be in shared segment).
5703 * The rte_eth_tx_prepare() function returns the number of packets ready to be
5704 * sent. A return value equal to *nb_pkts* means that all packets are valid and
5705 * ready to be sent, otherwise stops processing on the first invalid packet and
5706 * leaves the rest packets untouched.
5708 * When this functionality is not implemented in the driver, all packets are
5709 * are returned untouched.
5712 * The port identifier of the Ethernet device.
5713 * The value must be a valid port ID.
5715 * The index of the transmit queue through which output packets must be
5717 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5718 * to rte_eth_dev_configure().
5720 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5721 * which contain the output packets.
5723 * The maximum number of packets to process.
5725 * The number of packets correct and ready to be sent. The return value can be
5726 * less than the value of the *tx_pkts* parameter when some packet doesn't
5727 * meet devices requirements with rte_errno set appropriately:
5728 * - EINVAL: offload flags are not correctly set
5729 * - ENOTSUP: the offload feature is not supported by the hardware
5730 * - ENODEV: if *port_id* is invalid (with debug enabled only)
5734 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
5736 static inline uint16_t
5737 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
5738 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5740 struct rte_eth_fp_ops *p;
5743 #ifdef RTE_ETHDEV_DEBUG_TX
5744 if (port_id >= RTE_MAX_ETHPORTS ||
5745 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5747 "Invalid port_id=%u or queue_id=%u\n",
5754 /* fetch pointer to queue data */
5755 p = &rte_eth_fp_ops[port_id];
5756 qd = p->txq.data[queue_id];
5758 #ifdef RTE_ETHDEV_DEBUG_TX
5759 if (!rte_eth_dev_is_valid_port(port_id)) {
5760 RTE_ETHDEV_LOG(ERR, "Invalid Tx port_id=%u\n", port_id);
5765 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n",
5772 if (!p->tx_pkt_prepare)
5775 return p->tx_pkt_prepare(qd, tx_pkts, nb_pkts);
5781 * Native NOOP operation for compilation targets which doesn't require any
5782 * preparations steps, and functional NOOP may introduce unnecessary performance
5785 * Generally this is not a good idea to turn it on globally and didn't should
5786 * be used if behavior of tx_preparation can change.
5789 static inline uint16_t
5790 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
5791 __rte_unused uint16_t queue_id,
5792 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5800 * Send any packets queued up for transmission on a port and HW queue
5802 * This causes an explicit flush of packets previously buffered via the
5803 * rte_eth_tx_buffer() function. It returns the number of packets successfully
5804 * sent to the NIC, and calls the error callback for any unsent packets. Unless
5805 * explicitly set up otherwise, the default callback simply frees the unsent
5806 * packets back to the owning mempool.
5809 * The port identifier of the Ethernet device.
5811 * The index of the transmit queue through which output packets must be
5813 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5814 * to rte_eth_dev_configure().
5816 * Buffer of packets to be transmit.
5818 * The number of packets successfully sent to the Ethernet device. The error
5819 * callback is called for any packets which could not be sent.
5821 static inline uint16_t
5822 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
5823 struct rte_eth_dev_tx_buffer *buffer)
5826 uint16_t to_send = buffer->length;
5831 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
5835 /* All packets sent, or to be dealt with by callback below */
5836 if (unlikely(sent != to_send))
5837 buffer->error_callback(&buffer->pkts[sent],
5838 (uint16_t)(to_send - sent),
5839 buffer->error_userdata);
5845 * Buffer a single packet for future transmission on a port and queue
5847 * This function takes a single mbuf/packet and buffers it for later
5848 * transmission on the particular port and queue specified. Once the buffer is
5849 * full of packets, an attempt will be made to transmit all the buffered
5850 * packets. In case of error, where not all packets can be transmitted, a
5851 * callback is called with the unsent packets as a parameter. If no callback
5852 * is explicitly set up, the unsent packets are just freed back to the owning
5853 * mempool. The function returns the number of packets actually sent i.e.
5854 * 0 if no buffer flush occurred, otherwise the number of packets successfully
5858 * The port identifier of the Ethernet device.
5860 * The index of the transmit queue through which output packets must be
5862 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5863 * to rte_eth_dev_configure().
5865 * Buffer used to collect packets to be sent.
5867 * Pointer to the packet mbuf to be sent.
5869 * 0 = packet has been buffered for later transmission
5870 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
5871 * causing N packets to be sent, and the error callback to be called for
5874 static __rte_always_inline uint16_t
5875 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
5876 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
5878 buffer->pkts[buffer->length++] = tx_pkt;
5879 if (buffer->length < buffer->size)
5882 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
5889 #endif /* _RTE_ETHDEV_H_ */