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
77 * device or the queue is stopped.
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 * The following configuration may be retained or not
94 * depending on the device capabilities:
97 * - flow-related shared objects, e.g. indirect actions
99 * Any other configuration will not be stored and will need to be re-entered
100 * before a call to rte_eth_dev_start().
102 * Finally, a network application can close an Ethernet device by invoking the
103 * rte_eth_dev_close() function.
105 * Each function of the application Ethernet API invokes a specific function
106 * of the PMD that controls the target device designated by its port
108 * For this purpose, all device-specific functions of an Ethernet driver are
109 * supplied through a set of pointers contained in a generic structure of type
111 * The address of the *eth_dev_ops* structure is stored in the *rte_eth_dev*
112 * structure by the device init function of the Ethernet driver, which is
113 * invoked during the PCI probing phase, as explained earlier.
115 * In other words, each function of the Ethernet API simply retrieves the
116 * *rte_eth_dev* structure associated with the device port identifier and
117 * performs an indirect invocation of the corresponding driver function
118 * supplied in the *eth_dev_ops* structure of the *rte_eth_dev* structure.
120 * For performance reasons, the address of the burst-oriented Rx and Tx
121 * functions of the Ethernet driver are not contained in the *eth_dev_ops*
122 * structure. Instead, they are directly stored at the beginning of the
123 * *rte_eth_dev* structure to avoid an extra indirect memory access during
126 * RTE Ethernet device drivers do not use interrupts for transmitting or
127 * receiving. Instead, Ethernet drivers export Poll-Mode receive and transmit
128 * functions to applications.
129 * Both receive and transmit functions are packet-burst oriented to minimize
130 * their cost per packet through the following optimizations:
132 * - Sharing among multiple packets the incompressible cost of the
133 * invocation of receive/transmit functions.
135 * - Enabling receive/transmit functions to take advantage of burst-oriented
136 * hardware features (L1 cache, prefetch instructions, NIC head/tail
137 * registers) to minimize the number of CPU cycles per packet, for instance,
138 * by avoiding useless read memory accesses to ring descriptors, or by
139 * systematically using arrays of pointers that exactly fit L1 cache line
140 * boundaries and sizes.
142 * The burst-oriented receive function does not provide any error notification,
143 * to avoid the corresponding overhead. As a hint, the upper-level application
144 * might check the status of the device link once being systematically returned
145 * a 0 value by the receive function of the driver for a given number of tries.
154 /* Use this macro to check if LRO API is supported */
155 #define RTE_ETHDEV_HAS_LRO_SUPPORT
157 /* Alias RTE_LIBRTE_ETHDEV_DEBUG for backward compatibility. */
158 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
159 #define RTE_ETHDEV_DEBUG_RX
160 #define RTE_ETHDEV_DEBUG_TX
163 #include <rte_compat.h>
165 #include <rte_interrupts.h>
167 #include <rte_devargs.h>
168 #include <rte_bitops.h>
169 #include <rte_errno.h>
170 #include <rte_common.h>
171 #include <rte_config.h>
172 #include <rte_power_intrinsics.h>
174 #include "rte_ethdev_trace_fp.h"
175 #include "rte_dev_info.h"
177 extern int rte_eth_dev_logtype;
179 #define RTE_ETHDEV_LOG(level, ...) \
180 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__)
185 * Initializes a device iterator.
187 * This iterator allows accessing a list of devices matching some devargs.
190 * Device iterator handle initialized by the function.
191 * The fields bus_str and cls_str might be dynamically allocated,
192 * and could be freed by calling rte_eth_iterator_cleanup().
195 * Device description string.
198 * 0 on successful initialization, negative otherwise.
200 int rte_eth_iterator_init(struct rte_dev_iterator *iter, const char *devargs);
203 * Iterates on devices with devargs filter.
204 * The ownership is not checked.
206 * The next port ID is returned, and the iterator is updated.
209 * Device iterator handle initialized by rte_eth_iterator_init().
210 * Some fields bus_str and cls_str might be freed when no more port is found,
211 * by calling rte_eth_iterator_cleanup().
214 * A port ID if found, RTE_MAX_ETHPORTS otherwise.
216 uint16_t rte_eth_iterator_next(struct rte_dev_iterator *iter);
219 * Free some allocated fields of the iterator.
221 * This function is automatically called by rte_eth_iterator_next()
222 * on the last iteration (i.e. when no more matching port is found).
224 * It is safe to call this function twice; it will do nothing more.
227 * Device iterator handle initialized by rte_eth_iterator_init().
228 * The fields bus_str and cls_str are freed if needed.
230 void rte_eth_iterator_cleanup(struct rte_dev_iterator *iter);
233 * Macro to iterate over all ethdev ports matching some devargs.
235 * If a break is done before the end of the loop,
236 * the function rte_eth_iterator_cleanup() must be called.
239 * Iterated port ID of type uint16_t.
241 * Device parameters input as string of type char*.
243 * Iterator handle of type struct rte_dev_iterator, used internally.
245 #define RTE_ETH_FOREACH_MATCHING_DEV(id, devargs, iter) \
246 for (rte_eth_iterator_init(iter, devargs), \
247 id = rte_eth_iterator_next(iter); \
248 id != RTE_MAX_ETHPORTS; \
249 id = rte_eth_iterator_next(iter))
252 * A structure used to retrieve statistics for an Ethernet port.
253 * Not all statistics fields in struct rte_eth_stats are supported
254 * by any type of network interface card (NIC). If any statistics
255 * field is not supported, its value is 0.
256 * All byte-related statistics do not include Ethernet FCS regardless
257 * of whether these bytes have been delivered to the application
258 * (see RTE_ETH_RX_OFFLOAD_KEEP_CRC).
260 struct rte_eth_stats {
261 uint64_t ipackets; /**< Total number of successfully received packets. */
262 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
263 uint64_t ibytes; /**< Total number of successfully received bytes. */
264 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
266 * Total of Rx packets dropped by the HW,
267 * because there are no available buffer (i.e. Rx queues are full).
270 uint64_t ierrors; /**< Total number of erroneous received packets. */
271 uint64_t oerrors; /**< Total number of failed transmitted packets. */
272 uint64_t rx_nombuf; /**< Total number of Rx mbuf allocation failures. */
273 /* Queue stats are limited to max 256 queues */
274 /** Total number of queue Rx packets. */
275 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
276 /** Total number of queue Tx packets. */
277 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
278 /** Total number of successfully received queue bytes. */
279 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
280 /** Total number of successfully transmitted queue bytes. */
281 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
282 /** Total number of queue packets received that are dropped. */
283 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS];
286 /**@{@name Link speed capabilities
287 * Device supported speeds bitmap flags
289 #define RTE_ETH_LINK_SPEED_AUTONEG 0 /**< Autonegotiate (all speeds) */
290 #define RTE_ETH_LINK_SPEED_FIXED RTE_BIT32(0) /**< Disable autoneg (fixed speed) */
291 #define RTE_ETH_LINK_SPEED_10M_HD RTE_BIT32(1) /**< 10 Mbps half-duplex */
292 #define RTE_ETH_LINK_SPEED_10M RTE_BIT32(2) /**< 10 Mbps full-duplex */
293 #define RTE_ETH_LINK_SPEED_100M_HD RTE_BIT32(3) /**< 100 Mbps half-duplex */
294 #define RTE_ETH_LINK_SPEED_100M RTE_BIT32(4) /**< 100 Mbps full-duplex */
295 #define RTE_ETH_LINK_SPEED_1G RTE_BIT32(5) /**< 1 Gbps */
296 #define RTE_ETH_LINK_SPEED_2_5G RTE_BIT32(6) /**< 2.5 Gbps */
297 #define RTE_ETH_LINK_SPEED_5G RTE_BIT32(7) /**< 5 Gbps */
298 #define RTE_ETH_LINK_SPEED_10G RTE_BIT32(8) /**< 10 Gbps */
299 #define RTE_ETH_LINK_SPEED_20G RTE_BIT32(9) /**< 20 Gbps */
300 #define RTE_ETH_LINK_SPEED_25G RTE_BIT32(10) /**< 25 Gbps */
301 #define RTE_ETH_LINK_SPEED_40G RTE_BIT32(11) /**< 40 Gbps */
302 #define RTE_ETH_LINK_SPEED_50G RTE_BIT32(12) /**< 50 Gbps */
303 #define RTE_ETH_LINK_SPEED_56G RTE_BIT32(13) /**< 56 Gbps */
304 #define RTE_ETH_LINK_SPEED_100G RTE_BIT32(14) /**< 100 Gbps */
305 #define RTE_ETH_LINK_SPEED_200G RTE_BIT32(15) /**< 200 Gbps */
308 #define ETH_LINK_SPEED_AUTONEG RTE_DEPRECATED(ETH_LINK_SPEED_AUTONEG) RTE_ETH_LINK_SPEED_AUTONEG
309 #define ETH_LINK_SPEED_FIXED RTE_DEPRECATED(ETH_LINK_SPEED_FIXED) RTE_ETH_LINK_SPEED_FIXED
310 #define ETH_LINK_SPEED_10M_HD RTE_DEPRECATED(ETH_LINK_SPEED_10M_HD) RTE_ETH_LINK_SPEED_10M_HD
311 #define ETH_LINK_SPEED_10M RTE_DEPRECATED(ETH_LINK_SPEED_10M) RTE_ETH_LINK_SPEED_10M
312 #define ETH_LINK_SPEED_100M_HD RTE_DEPRECATED(ETH_LINK_SPEED_100M_HD) RTE_ETH_LINK_SPEED_100M_HD
313 #define ETH_LINK_SPEED_100M RTE_DEPRECATED(ETH_LINK_SPEED_100M) RTE_ETH_LINK_SPEED_100M
314 #define ETH_LINK_SPEED_1G RTE_DEPRECATED(ETH_LINK_SPEED_1G) RTE_ETH_LINK_SPEED_1G
315 #define ETH_LINK_SPEED_2_5G RTE_DEPRECATED(ETH_LINK_SPEED_2_5G) RTE_ETH_LINK_SPEED_2_5G
316 #define ETH_LINK_SPEED_5G RTE_DEPRECATED(ETH_LINK_SPEED_5G) RTE_ETH_LINK_SPEED_5G
317 #define ETH_LINK_SPEED_10G RTE_DEPRECATED(ETH_LINK_SPEED_10G) RTE_ETH_LINK_SPEED_10G
318 #define ETH_LINK_SPEED_20G RTE_DEPRECATED(ETH_LINK_SPEED_20G) RTE_ETH_LINK_SPEED_20G
319 #define ETH_LINK_SPEED_25G RTE_DEPRECATED(ETH_LINK_SPEED_25G) RTE_ETH_LINK_SPEED_25G
320 #define ETH_LINK_SPEED_40G RTE_DEPRECATED(ETH_LINK_SPEED_40G) RTE_ETH_LINK_SPEED_40G
321 #define ETH_LINK_SPEED_50G RTE_DEPRECATED(ETH_LINK_SPEED_50G) RTE_ETH_LINK_SPEED_50G
322 #define ETH_LINK_SPEED_56G RTE_DEPRECATED(ETH_LINK_SPEED_56G) RTE_ETH_LINK_SPEED_56G
323 #define ETH_LINK_SPEED_100G RTE_DEPRECATED(ETH_LINK_SPEED_100G) RTE_ETH_LINK_SPEED_100G
324 #define ETH_LINK_SPEED_200G RTE_DEPRECATED(ETH_LINK_SPEED_200G) RTE_ETH_LINK_SPEED_200G
326 /**@{@name Link speed
327 * Ethernet numeric link speeds in Mbps
329 #define RTE_ETH_SPEED_NUM_NONE 0 /**< Not defined */
330 #define RTE_ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
331 #define RTE_ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
332 #define RTE_ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
333 #define RTE_ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
334 #define RTE_ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
335 #define RTE_ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
336 #define RTE_ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
337 #define RTE_ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
338 #define RTE_ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
339 #define RTE_ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
340 #define RTE_ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
341 #define RTE_ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
342 #define RTE_ETH_SPEED_NUM_200G 200000 /**< 200 Gbps */
343 #define RTE_ETH_SPEED_NUM_UNKNOWN UINT32_MAX /**< Unknown */
346 #define ETH_SPEED_NUM_NONE RTE_DEPRECATED(ETH_SPEED_NUM_NONE) RTE_ETH_SPEED_NUM_NONE
347 #define ETH_SPEED_NUM_10M RTE_DEPRECATED(ETH_SPEED_NUM_10M) RTE_ETH_SPEED_NUM_10M
348 #define ETH_SPEED_NUM_100M RTE_DEPRECATED(ETH_SPEED_NUM_100M) RTE_ETH_SPEED_NUM_100M
349 #define ETH_SPEED_NUM_1G RTE_DEPRECATED(ETH_SPEED_NUM_1G) RTE_ETH_SPEED_NUM_1G
350 #define ETH_SPEED_NUM_2_5G RTE_DEPRECATED(ETH_SPEED_NUM_2_5G) RTE_ETH_SPEED_NUM_2_5G
351 #define ETH_SPEED_NUM_5G RTE_DEPRECATED(ETH_SPEED_NUM_5G) RTE_ETH_SPEED_NUM_5G
352 #define ETH_SPEED_NUM_10G RTE_DEPRECATED(ETH_SPEED_NUM_10G) RTE_ETH_SPEED_NUM_10G
353 #define ETH_SPEED_NUM_20G RTE_DEPRECATED(ETH_SPEED_NUM_20G) RTE_ETH_SPEED_NUM_20G
354 #define ETH_SPEED_NUM_25G RTE_DEPRECATED(ETH_SPEED_NUM_25G) RTE_ETH_SPEED_NUM_25G
355 #define ETH_SPEED_NUM_40G RTE_DEPRECATED(ETH_SPEED_NUM_40G) RTE_ETH_SPEED_NUM_40G
356 #define ETH_SPEED_NUM_50G RTE_DEPRECATED(ETH_SPEED_NUM_50G) RTE_ETH_SPEED_NUM_50G
357 #define ETH_SPEED_NUM_56G RTE_DEPRECATED(ETH_SPEED_NUM_56G) RTE_ETH_SPEED_NUM_56G
358 #define ETH_SPEED_NUM_100G RTE_DEPRECATED(ETH_SPEED_NUM_100G) RTE_ETH_SPEED_NUM_100G
359 #define ETH_SPEED_NUM_200G RTE_DEPRECATED(ETH_SPEED_NUM_200G) RTE_ETH_SPEED_NUM_200G
360 #define ETH_SPEED_NUM_UNKNOWN RTE_DEPRECATED(ETH_SPEED_NUM_UNKNOWN) RTE_ETH_SPEED_NUM_UNKNOWN
363 * A structure used to retrieve link-level information of an Ethernet port.
366 struct rte_eth_link {
367 uint32_t link_speed; /**< RTE_ETH_SPEED_NUM_ */
368 uint16_t link_duplex : 1; /**< RTE_ETH_LINK_[HALF/FULL]_DUPLEX */
369 uint16_t link_autoneg : 1; /**< RTE_ETH_LINK_[AUTONEG/FIXED] */
370 uint16_t link_status : 1; /**< RTE_ETH_LINK_[DOWN/UP] */
371 } __rte_aligned(8); /**< aligned for atomic64 read/write */
373 /**@{@name Link negotiation
374 * Constants used in link management.
376 #define RTE_ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
377 #define RTE_ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
378 #define RTE_ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
379 #define RTE_ETH_LINK_UP 1 /**< Link is up (see link_status). */
380 #define RTE_ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
381 #define RTE_ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
382 #define RTE_ETH_LINK_MAX_STR_LEN 40 /**< Max length of default link string. */
385 #define ETH_LINK_HALF_DUPLEX RTE_DEPRECATED(ETH_LINK_HALF_DUPLEX) RTE_ETH_LINK_HALF_DUPLEX
386 #define ETH_LINK_FULL_DUPLEX RTE_DEPRECATED(ETH_LINK_FULL_DUPLEX) RTE_ETH_LINK_FULL_DUPLEX
387 #define ETH_LINK_DOWN RTE_DEPRECATED(ETH_LINK_DOWN) RTE_ETH_LINK_DOWN
388 #define ETH_LINK_UP RTE_DEPRECATED(ETH_LINK_UP) RTE_ETH_LINK_UP
389 #define ETH_LINK_FIXED RTE_DEPRECATED(ETH_LINK_FIXED) RTE_ETH_LINK_FIXED
390 #define ETH_LINK_AUTONEG RTE_DEPRECATED(ETH_LINK_AUTONEG) RTE_ETH_LINK_AUTONEG
393 * A structure used to configure the ring threshold registers of an Rx/Tx
394 * queue for an Ethernet port.
396 struct rte_eth_thresh {
397 uint8_t pthresh; /**< Ring prefetch threshold. */
398 uint8_t hthresh; /**< Ring host threshold. */
399 uint8_t wthresh; /**< Ring writeback threshold. */
402 /**@{@name Multi-queue mode
403 * @see rte_eth_conf.rxmode.mq_mode.
405 #define RTE_ETH_MQ_RX_RSS_FLAG RTE_BIT32(0) /**< Enable RSS. @see rte_eth_rss_conf */
406 #define RTE_ETH_MQ_RX_DCB_FLAG RTE_BIT32(1) /**< Enable DCB. */
407 #define RTE_ETH_MQ_RX_VMDQ_FLAG RTE_BIT32(2) /**< Enable VMDq. */
410 #define ETH_MQ_RX_RSS_FLAG RTE_DEPRECATED(ETH_MQ_RX_RSS_FLAG) RTE_ETH_MQ_RX_RSS_FLAG
411 #define ETH_MQ_RX_DCB_FLAG RTE_DEPRECATED(ETH_MQ_RX_DCB_FLAG) RTE_ETH_MQ_RX_DCB_FLAG
412 #define ETH_MQ_RX_VMDQ_FLAG RTE_DEPRECATED(ETH_MQ_RX_VMDQ_FLAG) RTE_ETH_MQ_RX_VMDQ_FLAG
415 * A set of values to identify what method is to be used to route
416 * packets to multiple queues.
418 enum rte_eth_rx_mq_mode {
419 /** None of DCB, RSS or VMDq mode */
420 RTE_ETH_MQ_RX_NONE = 0,
422 /** For Rx side, only RSS is on */
423 RTE_ETH_MQ_RX_RSS = RTE_ETH_MQ_RX_RSS_FLAG,
424 /** For Rx side,only DCB is on. */
425 RTE_ETH_MQ_RX_DCB = RTE_ETH_MQ_RX_DCB_FLAG,
426 /** Both DCB and RSS enable */
427 RTE_ETH_MQ_RX_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG,
429 /** Only VMDq, no RSS nor DCB */
430 RTE_ETH_MQ_RX_VMDQ_ONLY = RTE_ETH_MQ_RX_VMDQ_FLAG,
431 /** RSS mode with VMDq */
432 RTE_ETH_MQ_RX_VMDQ_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_VMDQ_FLAG,
433 /** Use VMDq+DCB to route traffic to queues */
434 RTE_ETH_MQ_RX_VMDQ_DCB = RTE_ETH_MQ_RX_VMDQ_FLAG | RTE_ETH_MQ_RX_DCB_FLAG,
435 /** Enable both VMDq and DCB in VMDq */
436 RTE_ETH_MQ_RX_VMDQ_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG |
437 RTE_ETH_MQ_RX_VMDQ_FLAG,
440 #define ETH_MQ_RX_NONE RTE_DEPRECATED(ETH_MQ_RX_NONE) RTE_ETH_MQ_RX_NONE
441 #define ETH_MQ_RX_RSS RTE_DEPRECATED(ETH_MQ_RX_RSS) RTE_ETH_MQ_RX_RSS
442 #define ETH_MQ_RX_DCB RTE_DEPRECATED(ETH_MQ_RX_DCB) RTE_ETH_MQ_RX_DCB
443 #define ETH_MQ_RX_DCB_RSS RTE_DEPRECATED(ETH_MQ_RX_DCB_RSS) RTE_ETH_MQ_RX_DCB_RSS
444 #define ETH_MQ_RX_VMDQ_ONLY RTE_DEPRECATED(ETH_MQ_RX_VMDQ_ONLY) RTE_ETH_MQ_RX_VMDQ_ONLY
445 #define ETH_MQ_RX_VMDQ_RSS RTE_DEPRECATED(ETH_MQ_RX_VMDQ_RSS) RTE_ETH_MQ_RX_VMDQ_RSS
446 #define ETH_MQ_RX_VMDQ_DCB RTE_DEPRECATED(ETH_MQ_RX_VMDQ_DCB) RTE_ETH_MQ_RX_VMDQ_DCB
447 #define ETH_MQ_RX_VMDQ_DCB_RSS RTE_DEPRECATED(ETH_MQ_RX_VMDQ_DCB_RSS) RTE_ETH_MQ_RX_VMDQ_DCB_RSS
450 * A set of values to identify what method is to be used to transmit
451 * packets using multi-TCs.
453 enum rte_eth_tx_mq_mode {
454 RTE_ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
455 RTE_ETH_MQ_TX_DCB, /**< For Tx side,only DCB is on. */
456 RTE_ETH_MQ_TX_VMDQ_DCB, /**< For Tx side,both DCB and VT is on. */
457 RTE_ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
460 #define ETH_MQ_TX_NONE RTE_DEPRECATED(ETH_MQ_TX_NONE) RTE_ETH_MQ_TX_NONE
461 #define ETH_MQ_TX_DCB RTE_DEPRECATED(ETH_MQ_TX_DCB) RTE_ETH_MQ_TX_DCB
462 #define ETH_MQ_TX_VMDQ_DCB RTE_DEPRECATED(ETH_MQ_TX_VMDQ_DCB) RTE_ETH_MQ_TX_VMDQ_DCB
463 #define ETH_MQ_TX_VMDQ_ONLY RTE_DEPRECATED(ETH_MQ_TX_VMDQ_ONLY) RTE_ETH_MQ_TX_VMDQ_ONLY
466 * A structure used to configure the Rx features of an Ethernet port.
468 struct rte_eth_rxmode {
469 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
470 enum rte_eth_rx_mq_mode mq_mode;
471 uint32_t mtu; /**< Requested MTU. */
472 /** Maximum allowed size of LRO aggregated packet. */
473 uint32_t max_lro_pkt_size;
474 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
476 * Per-port Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags.
477 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
478 * structure are allowed to be set.
482 uint64_t reserved_64s[2]; /**< Reserved for future fields */
483 void *reserved_ptrs[2]; /**< Reserved for future fields */
487 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
488 * Note that single VLAN is treated the same as inner VLAN.
491 RTE_ETH_VLAN_TYPE_UNKNOWN = 0,
492 RTE_ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
493 RTE_ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
494 RTE_ETH_VLAN_TYPE_MAX,
497 #define ETH_VLAN_TYPE_UNKNOWN RTE_DEPRECATED(ETH_VLAN_TYPE_UNKNOWN) RTE_ETH_VLAN_TYPE_UNKNOWN
498 #define ETH_VLAN_TYPE_INNER RTE_DEPRECATED(ETH_VLAN_TYPE_INNER) RTE_ETH_VLAN_TYPE_INNER
499 #define ETH_VLAN_TYPE_OUTER RTE_DEPRECATED(ETH_VLAN_TYPE_OUTER) RTE_ETH_VLAN_TYPE_OUTER
500 #define ETH_VLAN_TYPE_MAX RTE_DEPRECATED(ETH_VLAN_TYPE_MAX) RTE_ETH_VLAN_TYPE_MAX
503 * A structure used to describe a VLAN filter.
504 * If the bit corresponding to a VID is set, such VID is on.
506 struct rte_vlan_filter_conf {
511 * A structure used to configure the Receive Side Scaling (RSS) feature
512 * of an Ethernet port.
513 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
514 * to an array holding the RSS key to use for hashing specific header
515 * fields of received packets. The length of this array should be indicated
516 * by *rss_key_len* below. Otherwise, a default random hash key is used by
519 * The *rss_key_len* field of the *rss_conf* structure indicates the length
520 * in bytes of the array pointed by *rss_key*. To be compatible, this length
521 * will be checked in i40e only. Others assume 40 bytes to be used as before.
523 * The *rss_hf* field of the *rss_conf* structure indicates the different
524 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
525 * Supplying an *rss_hf* equal to zero disables the RSS feature.
527 struct rte_eth_rss_conf {
528 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
529 uint8_t rss_key_len; /**< hash key length in bytes. */
530 uint64_t rss_hf; /**< Hash functions to apply - see below. */
534 * A packet can be identified by hardware as different flow types. Different
535 * NIC hardware may support different flow types.
536 * Basically, the NIC hardware identifies the flow type as deep protocol as
537 * possible, and exclusively. For example, if a packet is identified as
538 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types,
539 * though it is an actual IPV4 packet.
541 #define RTE_ETH_FLOW_UNKNOWN 0
542 #define RTE_ETH_FLOW_RAW 1
543 #define RTE_ETH_FLOW_IPV4 2
544 #define RTE_ETH_FLOW_FRAG_IPV4 3
545 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4
546 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5
547 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6
548 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7
549 #define RTE_ETH_FLOW_IPV6 8
550 #define RTE_ETH_FLOW_FRAG_IPV6 9
551 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10
552 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11
553 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12
554 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13
555 #define RTE_ETH_FLOW_L2_PAYLOAD 14
556 #define RTE_ETH_FLOW_IPV6_EX 15
557 #define RTE_ETH_FLOW_IPV6_TCP_EX 16
558 #define RTE_ETH_FLOW_IPV6_UDP_EX 17
559 /** Consider device port number as a flow differentiator */
560 #define RTE_ETH_FLOW_PORT 18
561 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */
562 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */
563 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */
564 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */
565 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */
566 #define RTE_ETH_FLOW_MAX 24
569 * Below macros are defined for RSS offload types, they can be used to
570 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types.
572 #define RTE_ETH_RSS_IPV4 RTE_BIT64(2)
573 #define RTE_ETH_RSS_FRAG_IPV4 RTE_BIT64(3)
574 #define RTE_ETH_RSS_NONFRAG_IPV4_TCP RTE_BIT64(4)
575 #define RTE_ETH_RSS_NONFRAG_IPV4_UDP RTE_BIT64(5)
576 #define RTE_ETH_RSS_NONFRAG_IPV4_SCTP RTE_BIT64(6)
577 #define RTE_ETH_RSS_NONFRAG_IPV4_OTHER RTE_BIT64(7)
578 #define RTE_ETH_RSS_IPV6 RTE_BIT64(8)
579 #define RTE_ETH_RSS_FRAG_IPV6 RTE_BIT64(9)
580 #define RTE_ETH_RSS_NONFRAG_IPV6_TCP RTE_BIT64(10)
581 #define RTE_ETH_RSS_NONFRAG_IPV6_UDP RTE_BIT64(11)
582 #define RTE_ETH_RSS_NONFRAG_IPV6_SCTP RTE_BIT64(12)
583 #define RTE_ETH_RSS_NONFRAG_IPV6_OTHER RTE_BIT64(13)
584 #define RTE_ETH_RSS_L2_PAYLOAD RTE_BIT64(14)
585 #define RTE_ETH_RSS_IPV6_EX RTE_BIT64(15)
586 #define RTE_ETH_RSS_IPV6_TCP_EX RTE_BIT64(16)
587 #define RTE_ETH_RSS_IPV6_UDP_EX RTE_BIT64(17)
588 #define RTE_ETH_RSS_PORT RTE_BIT64(18)
589 #define RTE_ETH_RSS_VXLAN RTE_BIT64(19)
590 #define RTE_ETH_RSS_GENEVE RTE_BIT64(20)
591 #define RTE_ETH_RSS_NVGRE RTE_BIT64(21)
592 #define RTE_ETH_RSS_GTPU RTE_BIT64(23)
593 #define RTE_ETH_RSS_ETH RTE_BIT64(24)
594 #define RTE_ETH_RSS_S_VLAN RTE_BIT64(25)
595 #define RTE_ETH_RSS_C_VLAN RTE_BIT64(26)
596 #define RTE_ETH_RSS_ESP RTE_BIT64(27)
597 #define RTE_ETH_RSS_AH RTE_BIT64(28)
598 #define RTE_ETH_RSS_L2TPV3 RTE_BIT64(29)
599 #define RTE_ETH_RSS_PFCP RTE_BIT64(30)
600 #define RTE_ETH_RSS_PPPOE RTE_BIT64(31)
601 #define RTE_ETH_RSS_ECPRI RTE_BIT64(32)
602 #define RTE_ETH_RSS_MPLS RTE_BIT64(33)
603 #define RTE_ETH_RSS_IPV4_CHKSUM RTE_BIT64(34)
605 #define ETH_RSS_IPV4 RTE_DEPRECATED(ETH_RSS_IPV4) RTE_ETH_RSS_IPV4
606 #define ETH_RSS_FRAG_IPV4 RTE_DEPRECATED(ETH_RSS_FRAG_IPV4) RTE_ETH_RSS_FRAG_IPV4
607 #define ETH_RSS_NONFRAG_IPV4_TCP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_TCP) RTE_ETH_RSS_NONFRAG_IPV4_TCP
608 #define ETH_RSS_NONFRAG_IPV4_UDP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_UDP) RTE_ETH_RSS_NONFRAG_IPV4_UDP
609 #define ETH_RSS_NONFRAG_IPV4_SCTP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_SCTP) RTE_ETH_RSS_NONFRAG_IPV4_SCTP
610 #define ETH_RSS_NONFRAG_IPV4_OTHER RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_OTHER) RTE_ETH_RSS_NONFRAG_IPV4_OTHER
611 #define ETH_RSS_IPV6 RTE_DEPRECATED(ETH_RSS_IPV6) RTE_ETH_RSS_IPV6
612 #define ETH_RSS_FRAG_IPV6 RTE_DEPRECATED(ETH_RSS_FRAG_IPV6) RTE_ETH_RSS_FRAG_IPV6
613 #define ETH_RSS_NONFRAG_IPV6_TCP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_TCP) RTE_ETH_RSS_NONFRAG_IPV6_TCP
614 #define ETH_RSS_NONFRAG_IPV6_UDP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_UDP) RTE_ETH_RSS_NONFRAG_IPV6_UDP
615 #define ETH_RSS_NONFRAG_IPV6_SCTP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_SCTP) RTE_ETH_RSS_NONFRAG_IPV6_SCTP
616 #define ETH_RSS_NONFRAG_IPV6_OTHER RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_OTHER) RTE_ETH_RSS_NONFRAG_IPV6_OTHER
617 #define ETH_RSS_L2_PAYLOAD RTE_DEPRECATED(ETH_RSS_L2_PAYLOAD) RTE_ETH_RSS_L2_PAYLOAD
618 #define ETH_RSS_IPV6_EX RTE_DEPRECATED(ETH_RSS_IPV6_EX) RTE_ETH_RSS_IPV6_EX
619 #define ETH_RSS_IPV6_TCP_EX RTE_DEPRECATED(ETH_RSS_IPV6_TCP_EX) RTE_ETH_RSS_IPV6_TCP_EX
620 #define ETH_RSS_IPV6_UDP_EX RTE_DEPRECATED(ETH_RSS_IPV6_UDP_EX) RTE_ETH_RSS_IPV6_UDP_EX
621 #define ETH_RSS_PORT RTE_DEPRECATED(ETH_RSS_PORT) RTE_ETH_RSS_PORT
622 #define ETH_RSS_VXLAN RTE_DEPRECATED(ETH_RSS_VXLAN) RTE_ETH_RSS_VXLAN
623 #define ETH_RSS_GENEVE RTE_DEPRECATED(ETH_RSS_GENEVE) RTE_ETH_RSS_GENEVE
624 #define ETH_RSS_NVGRE RTE_DEPRECATED(ETH_RSS_NVGRE) RTE_ETH_RSS_NVGRE
625 #define ETH_RSS_GTPU RTE_DEPRECATED(ETH_RSS_GTPU) RTE_ETH_RSS_GTPU
626 #define ETH_RSS_ETH RTE_DEPRECATED(ETH_RSS_ETH) RTE_ETH_RSS_ETH
627 #define ETH_RSS_S_VLAN RTE_DEPRECATED(ETH_RSS_S_VLAN) RTE_ETH_RSS_S_VLAN
628 #define ETH_RSS_C_VLAN RTE_DEPRECATED(ETH_RSS_C_VLAN) RTE_ETH_RSS_C_VLAN
629 #define ETH_RSS_ESP RTE_DEPRECATED(ETH_RSS_ESP) RTE_ETH_RSS_ESP
630 #define ETH_RSS_AH RTE_DEPRECATED(ETH_RSS_AH) RTE_ETH_RSS_AH
631 #define ETH_RSS_L2TPV3 RTE_DEPRECATED(ETH_RSS_L2TPV3) RTE_ETH_RSS_L2TPV3
632 #define ETH_RSS_PFCP RTE_DEPRECATED(ETH_RSS_PFCP) RTE_ETH_RSS_PFCP
633 #define ETH_RSS_PPPOE RTE_DEPRECATED(ETH_RSS_PPPOE) RTE_ETH_RSS_PPPOE
634 #define ETH_RSS_ECPRI RTE_DEPRECATED(ETH_RSS_ECPRI) RTE_ETH_RSS_ECPRI
635 #define ETH_RSS_MPLS RTE_DEPRECATED(ETH_RSS_MPLS) RTE_ETH_RSS_MPLS
636 #define ETH_RSS_IPV4_CHKSUM RTE_DEPRECATED(ETH_RSS_IPV4_CHKSUM) RTE_ETH_RSS_IPV4_CHKSUM
639 * The ETH_RSS_L4_CHKSUM works on checksum field of any L4 header.
640 * It is similar to ETH_RSS_PORT that they don't specify the specific type of
641 * L4 header. This macro is defined to replace some specific L4 (TCP/UDP/SCTP)
642 * checksum type for constructing the use of RSS offload bits.
644 * Due to above reason, some old APIs (and configuration) don't support
645 * RTE_ETH_RSS_L4_CHKSUM. The rte_flow RSS API supports it.
647 * For the case that checksum is not used in an UDP header,
648 * it takes the reserved value 0 as input for the hash function.
650 #define RTE_ETH_RSS_L4_CHKSUM RTE_BIT64(35)
651 #define ETH_RSS_L4_CHKSUM RTE_DEPRECATED(ETH_RSS_L4_CHKSUM) RTE_ETH_RSS_L4_CHKSUM
653 #define RTE_ETH_RSS_L2TPV2 RTE_BIT64(36)
656 * We use the following macros to combine with above RTE_ETH_RSS_* for
657 * more specific input set selection. These bits are defined starting
658 * from the high end of the 64 bits.
659 * Note: If we use above RTE_ETH_RSS_* without SRC/DST_ONLY, it represents
660 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of
661 * the same level are used simultaneously, it is the same case as none of
664 #define RTE_ETH_RSS_L3_SRC_ONLY RTE_BIT64(63)
665 #define RTE_ETH_RSS_L3_DST_ONLY RTE_BIT64(62)
666 #define RTE_ETH_RSS_L4_SRC_ONLY RTE_BIT64(61)
667 #define RTE_ETH_RSS_L4_DST_ONLY RTE_BIT64(60)
668 #define RTE_ETH_RSS_L2_SRC_ONLY RTE_BIT64(59)
669 #define RTE_ETH_RSS_L2_DST_ONLY RTE_BIT64(58)
671 #define ETH_RSS_L3_SRC_ONLY RTE_DEPRECATED(ETH_RSS_L3_SRC_ONLY) RTE_ETH_RSS_L3_SRC_ONLY
672 #define ETH_RSS_L3_DST_ONLY RTE_DEPRECATED(ETH_RSS_L3_DST_ONLY) RTE_ETH_RSS_L3_DST_ONLY
673 #define ETH_RSS_L4_SRC_ONLY RTE_DEPRECATED(ETH_RSS_L4_SRC_ONLY) RTE_ETH_RSS_L4_SRC_ONLY
674 #define ETH_RSS_L4_DST_ONLY RTE_DEPRECATED(ETH_RSS_L4_DST_ONLY) RTE_ETH_RSS_L4_DST_ONLY
675 #define ETH_RSS_L2_SRC_ONLY RTE_DEPRECATED(ETH_RSS_L2_SRC_ONLY) RTE_ETH_RSS_L2_SRC_ONLY
676 #define ETH_RSS_L2_DST_ONLY RTE_DEPRECATED(ETH_RSS_L2_DST_ONLY) RTE_ETH_RSS_L2_DST_ONLY
679 * Only select IPV6 address prefix as RSS input set according to
680 * https://tools.ietf.org/html/rfc6052
681 * Must be combined with RTE_ETH_RSS_IPV6, RTE_ETH_RSS_NONFRAG_IPV6_UDP,
682 * RTE_ETH_RSS_NONFRAG_IPV6_TCP, RTE_ETH_RSS_NONFRAG_IPV6_SCTP.
684 #define RTE_ETH_RSS_L3_PRE32 RTE_BIT64(57)
685 #define RTE_ETH_RSS_L3_PRE40 RTE_BIT64(56)
686 #define RTE_ETH_RSS_L3_PRE48 RTE_BIT64(55)
687 #define RTE_ETH_RSS_L3_PRE56 RTE_BIT64(54)
688 #define RTE_ETH_RSS_L3_PRE64 RTE_BIT64(53)
689 #define RTE_ETH_RSS_L3_PRE96 RTE_BIT64(52)
692 * Use the following macros to combine with the above layers
693 * to choose inner and outer layers or both for RSS computation.
694 * Bits 50 and 51 are reserved for this.
698 * level 0, requests the default behavior.
699 * Depending on the packet type, it can mean outermost, innermost,
700 * anything in between or even no RSS.
701 * It basically stands for the innermost encapsulation level RSS
702 * can be performed on according to PMD and device capabilities.
704 #define RTE_ETH_RSS_LEVEL_PMD_DEFAULT (UINT64_C(0) << 50)
705 #define ETH_RSS_LEVEL_PMD_DEFAULT RTE_DEPRECATED(ETH_RSS_LEVEL_PMD_DEFAULT) RTE_ETH_RSS_LEVEL_PMD_DEFAULT
708 * level 1, requests RSS to be performed on the outermost packet
709 * encapsulation level.
711 #define RTE_ETH_RSS_LEVEL_OUTERMOST (UINT64_C(1) << 50)
712 #define ETH_RSS_LEVEL_OUTERMOST RTE_DEPRECATED(ETH_RSS_LEVEL_OUTERMOST) RTE_ETH_RSS_LEVEL_OUTERMOST
715 * level 2, requests RSS to be performed on the specified inner packet
716 * encapsulation level, from outermost to innermost (lower to higher values).
718 #define RTE_ETH_RSS_LEVEL_INNERMOST (UINT64_C(2) << 50)
719 #define RTE_ETH_RSS_LEVEL_MASK (UINT64_C(3) << 50)
721 #define ETH_RSS_LEVEL_INNERMOST RTE_DEPRECATED(ETH_RSS_LEVEL_INNERMOST) RTE_ETH_RSS_LEVEL_INNERMOST
722 #define ETH_RSS_LEVEL_MASK RTE_DEPRECATED(ETH_RSS_LEVEL_MASK) RTE_ETH_RSS_LEVEL_MASK
724 #define RTE_ETH_RSS_LEVEL(rss_hf) ((rss_hf & RTE_ETH_RSS_LEVEL_MASK) >> 50)
725 #define ETH_RSS_LEVEL(rss_hf) RTE_DEPRECATED(ETH_RSS_LEVEL(rss_hf)) RTE_ETH_RSS_LEVEL(rss_hf)
728 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of
729 * the same level are used simultaneously, it is the same case as
730 * none of them are added.
733 * RSS types with SRC/DST_ONLY.
737 static inline uint64_t
738 rte_eth_rss_hf_refine(uint64_t rss_hf)
740 if ((rss_hf & RTE_ETH_RSS_L3_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L3_DST_ONLY))
741 rss_hf &= ~(RTE_ETH_RSS_L3_SRC_ONLY | RTE_ETH_RSS_L3_DST_ONLY);
743 if ((rss_hf & RTE_ETH_RSS_L4_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L4_DST_ONLY))
744 rss_hf &= ~(RTE_ETH_RSS_L4_SRC_ONLY | RTE_ETH_RSS_L4_DST_ONLY);
749 #define RTE_ETH_RSS_IPV6_PRE32 ( \
751 RTE_ETH_RSS_L3_PRE32)
752 #define ETH_RSS_IPV6_PRE32 RTE_DEPRECATED(ETH_RSS_IPV6_PRE32) RTE_ETH_RSS_IPV6_PRE32
754 #define RTE_ETH_RSS_IPV6_PRE40 ( \
756 RTE_ETH_RSS_L3_PRE40)
757 #define ETH_RSS_IPV6_PRE40 RTE_DEPRECATED(ETH_RSS_IPV6_PRE40) RTE_ETH_RSS_IPV6_PRE40
759 #define RTE_ETH_RSS_IPV6_PRE48 ( \
761 RTE_ETH_RSS_L3_PRE48)
762 #define ETH_RSS_IPV6_PRE48 RTE_DEPRECATED(ETH_RSS_IPV6_PRE48) RTE_ETH_RSS_IPV6_PRE48
764 #define RTE_ETH_RSS_IPV6_PRE56 ( \
766 RTE_ETH_RSS_L3_PRE56)
767 #define ETH_RSS_IPV6_PRE56 RTE_DEPRECATED(ETH_RSS_IPV6_PRE56) RTE_ETH_RSS_IPV6_PRE56
769 #define RTE_ETH_RSS_IPV6_PRE64 ( \
771 RTE_ETH_RSS_L3_PRE64)
772 #define ETH_RSS_IPV6_PRE64 RTE_DEPRECATED(ETH_RSS_IPV6_PRE64) RTE_ETH_RSS_IPV6_PRE64
774 #define RTE_ETH_RSS_IPV6_PRE96 ( \
776 RTE_ETH_RSS_L3_PRE96)
777 #define ETH_RSS_IPV6_PRE96 RTE_DEPRECATED(ETH_RSS_IPV6_PRE96) RTE_ETH_RSS_IPV6_PRE96
779 #define RTE_ETH_RSS_IPV6_PRE32_UDP ( \
780 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
781 RTE_ETH_RSS_L3_PRE32)
782 #define ETH_RSS_IPV6_PRE32_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE32_UDP) RTE_ETH_RSS_IPV6_PRE32_UDP
784 #define RTE_ETH_RSS_IPV6_PRE40_UDP ( \
785 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
786 RTE_ETH_RSS_L3_PRE40)
787 #define ETH_RSS_IPV6_PRE40_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE40_UDP) RTE_ETH_RSS_IPV6_PRE40_UDP
789 #define RTE_ETH_RSS_IPV6_PRE48_UDP ( \
790 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
791 RTE_ETH_RSS_L3_PRE48)
792 #define ETH_RSS_IPV6_PRE48_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE48_UDP) RTE_ETH_RSS_IPV6_PRE48_UDP
794 #define RTE_ETH_RSS_IPV6_PRE56_UDP ( \
795 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
796 RTE_ETH_RSS_L3_PRE56)
797 #define ETH_RSS_IPV6_PRE56_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE56_UDP) RTE_ETH_RSS_IPV6_PRE56_UDP
799 #define RTE_ETH_RSS_IPV6_PRE64_UDP ( \
800 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
801 RTE_ETH_RSS_L3_PRE64)
802 #define ETH_RSS_IPV6_PRE64_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE64_UDP) RTE_ETH_RSS_IPV6_PRE64_UDP
804 #define RTE_ETH_RSS_IPV6_PRE96_UDP ( \
805 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
806 RTE_ETH_RSS_L3_PRE96)
807 #define ETH_RSS_IPV6_PRE96_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE96_UDP) RTE_ETH_RSS_IPV6_PRE96_UDP
809 #define RTE_ETH_RSS_IPV6_PRE32_TCP ( \
810 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
811 RTE_ETH_RSS_L3_PRE32)
812 #define ETH_RSS_IPV6_PRE32_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE32_TCP) RTE_ETH_RSS_IPV6_PRE32_TCP
814 #define RTE_ETH_RSS_IPV6_PRE40_TCP ( \
815 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
816 RTE_ETH_RSS_L3_PRE40)
817 #define ETH_RSS_IPV6_PRE40_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE40_TCP) RTE_ETH_RSS_IPV6_PRE40_TCP
819 #define RTE_ETH_RSS_IPV6_PRE48_TCP ( \
820 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
821 RTE_ETH_RSS_L3_PRE48)
822 #define ETH_RSS_IPV6_PRE48_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE48_TCP) RTE_ETH_RSS_IPV6_PRE48_TCP
824 #define RTE_ETH_RSS_IPV6_PRE56_TCP ( \
825 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
826 RTE_ETH_RSS_L3_PRE56)
827 #define ETH_RSS_IPV6_PRE56_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE56_TCP) RTE_ETH_RSS_IPV6_PRE56_TCP
829 #define RTE_ETH_RSS_IPV6_PRE64_TCP ( \
830 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
831 RTE_ETH_RSS_L3_PRE64)
832 #define ETH_RSS_IPV6_PRE64_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE64_TCP) RTE_ETH_RSS_IPV6_PRE64_TCP
834 #define RTE_ETH_RSS_IPV6_PRE96_TCP ( \
835 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
836 RTE_ETH_RSS_L3_PRE96)
837 #define ETH_RSS_IPV6_PRE96_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE96_TCP) RTE_ETH_RSS_IPV6_PRE96_TCP
839 #define RTE_ETH_RSS_IPV6_PRE32_SCTP ( \
840 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
841 RTE_ETH_RSS_L3_PRE32)
842 #define ETH_RSS_IPV6_PRE32_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE32_SCTP) RTE_ETH_RSS_IPV6_PRE32_SCTP
844 #define RTE_ETH_RSS_IPV6_PRE40_SCTP ( \
845 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
846 RTE_ETH_RSS_L3_PRE40)
847 #define ETH_RSS_IPV6_PRE40_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE40_SCTP) RTE_ETH_RSS_IPV6_PRE40_SCTP
849 #define RTE_ETH_RSS_IPV6_PRE48_SCTP ( \
850 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
851 RTE_ETH_RSS_L3_PRE48)
852 #define ETH_RSS_IPV6_PRE48_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE48_SCTP) RTE_ETH_RSS_IPV6_PRE48_SCTP
854 #define RTE_ETH_RSS_IPV6_PRE56_SCTP ( \
855 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
856 RTE_ETH_RSS_L3_PRE56)
857 #define ETH_RSS_IPV6_PRE56_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE56_SCTP) RTE_ETH_RSS_IPV6_PRE56_SCTP
859 #define RTE_ETH_RSS_IPV6_PRE64_SCTP ( \
860 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
861 RTE_ETH_RSS_L3_PRE64)
862 #define ETH_RSS_IPV6_PRE64_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE64_SCTP) RTE_ETH_RSS_IPV6_PRE64_SCTP
864 #define RTE_ETH_RSS_IPV6_PRE96_SCTP ( \
865 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
866 RTE_ETH_RSS_L3_PRE96)
867 #define ETH_RSS_IPV6_PRE96_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE96_SCTP) RTE_ETH_RSS_IPV6_PRE96_SCTP
869 #define RTE_ETH_RSS_IP ( \
871 RTE_ETH_RSS_FRAG_IPV4 | \
872 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \
874 RTE_ETH_RSS_FRAG_IPV6 | \
875 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \
877 #define ETH_RSS_IP RTE_DEPRECATED(ETH_RSS_IP) RTE_ETH_RSS_IP
879 #define RTE_ETH_RSS_UDP ( \
880 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \
881 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
882 RTE_ETH_RSS_IPV6_UDP_EX)
883 #define ETH_RSS_UDP RTE_DEPRECATED(ETH_RSS_UDP) RTE_ETH_RSS_UDP
885 #define RTE_ETH_RSS_TCP ( \
886 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \
887 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
888 RTE_ETH_RSS_IPV6_TCP_EX)
889 #define ETH_RSS_TCP RTE_DEPRECATED(ETH_RSS_TCP) RTE_ETH_RSS_TCP
891 #define RTE_ETH_RSS_SCTP ( \
892 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \
893 RTE_ETH_RSS_NONFRAG_IPV6_SCTP)
894 #define ETH_RSS_SCTP RTE_DEPRECATED(ETH_RSS_SCTP) RTE_ETH_RSS_SCTP
896 #define RTE_ETH_RSS_TUNNEL ( \
897 RTE_ETH_RSS_VXLAN | \
898 RTE_ETH_RSS_GENEVE | \
900 #define ETH_RSS_TUNNEL RTE_DEPRECATED(ETH_RSS_TUNNEL) RTE_ETH_RSS_TUNNEL
902 #define RTE_ETH_RSS_VLAN ( \
903 RTE_ETH_RSS_S_VLAN | \
905 #define ETH_RSS_VLAN RTE_DEPRECATED(ETH_RSS_VLAN) RTE_ETH_RSS_VLAN
907 /** Mask of valid RSS hash protocols */
908 #define RTE_ETH_RSS_PROTO_MASK ( \
910 RTE_ETH_RSS_FRAG_IPV4 | \
911 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \
912 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \
913 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \
914 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \
916 RTE_ETH_RSS_FRAG_IPV6 | \
917 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
918 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
919 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
920 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \
921 RTE_ETH_RSS_L2_PAYLOAD | \
922 RTE_ETH_RSS_IPV6_EX | \
923 RTE_ETH_RSS_IPV6_TCP_EX | \
924 RTE_ETH_RSS_IPV6_UDP_EX | \
926 RTE_ETH_RSS_VXLAN | \
927 RTE_ETH_RSS_GENEVE | \
928 RTE_ETH_RSS_NVGRE | \
930 #define ETH_RSS_PROTO_MASK RTE_DEPRECATED(ETH_RSS_PROTO_MASK) RTE_ETH_RSS_PROTO_MASK
933 * Definitions used for redirection table entry size.
934 * Some RSS RETA sizes may not be supported by some drivers, check the
935 * documentation or the description of relevant functions for more details.
937 #define RTE_ETH_RSS_RETA_SIZE_64 64
938 #define RTE_ETH_RSS_RETA_SIZE_128 128
939 #define RTE_ETH_RSS_RETA_SIZE_256 256
940 #define RTE_ETH_RSS_RETA_SIZE_512 512
941 #define RTE_ETH_RETA_GROUP_SIZE 64
943 #define ETH_RSS_RETA_SIZE_64 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_64) RTE_ETH_RSS_RETA_SIZE_64
944 #define ETH_RSS_RETA_SIZE_128 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_128) RTE_ETH_RSS_RETA_SIZE_128
945 #define ETH_RSS_RETA_SIZE_256 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_256) RTE_ETH_RSS_RETA_SIZE_256
946 #define ETH_RSS_RETA_SIZE_512 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_512) RTE_ETH_RSS_RETA_SIZE_512
947 #define RTE_RETA_GROUP_SIZE RTE_DEPRECATED(RTE_RETA_GROUP_SIZE) RTE_ETH_RETA_GROUP_SIZE
949 /**@{@name VMDq and DCB maximums */
950 #define RTE_ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDq VLAN filters. */
951 #define RTE_ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
952 #define RTE_ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDq DCB queues. */
953 #define RTE_ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
956 #define ETH_VMDQ_MAX_VLAN_FILTERS RTE_DEPRECATED(ETH_VMDQ_MAX_VLAN_FILTERS) RTE_ETH_VMDQ_MAX_VLAN_FILTERS
957 #define ETH_DCB_NUM_USER_PRIORITIES RTE_DEPRECATED(ETH_DCB_NUM_USER_PRIORITIES) RTE_ETH_DCB_NUM_USER_PRIORITIES
958 #define ETH_VMDQ_DCB_NUM_QUEUES RTE_DEPRECATED(ETH_VMDQ_DCB_NUM_QUEUES) RTE_ETH_VMDQ_DCB_NUM_QUEUES
959 #define ETH_DCB_NUM_QUEUES RTE_DEPRECATED(ETH_DCB_NUM_QUEUES) RTE_ETH_DCB_NUM_QUEUES
961 /**@{@name DCB capabilities */
962 #define RTE_ETH_DCB_PG_SUPPORT RTE_BIT32(0) /**< Priority Group(ETS) support. */
963 #define RTE_ETH_DCB_PFC_SUPPORT RTE_BIT32(1) /**< Priority Flow Control support. */
966 #define ETH_DCB_PG_SUPPORT RTE_DEPRECATED(ETH_DCB_PG_SUPPORT) RTE_ETH_DCB_PG_SUPPORT
967 #define ETH_DCB_PFC_SUPPORT RTE_DEPRECATED(ETH_DCB_PFC_SUPPORT) RTE_ETH_DCB_PFC_SUPPORT
969 /**@{@name VLAN offload bits */
970 #define RTE_ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
971 #define RTE_ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
972 #define RTE_ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
973 #define RTE_ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */
975 #define ETH_VLAN_STRIP_OFFLOAD RTE_DEPRECATED(ETH_VLAN_STRIP_OFFLOAD) RTE_ETH_VLAN_STRIP_OFFLOAD
976 #define ETH_VLAN_FILTER_OFFLOAD RTE_DEPRECATED(ETH_VLAN_FILTER_OFFLOAD) RTE_ETH_VLAN_FILTER_OFFLOAD
977 #define ETH_VLAN_EXTEND_OFFLOAD RTE_DEPRECATED(ETH_VLAN_EXTEND_OFFLOAD) RTE_ETH_VLAN_EXTEND_OFFLOAD
978 #define ETH_QINQ_STRIP_OFFLOAD RTE_DEPRECATED(ETH_QINQ_STRIP_OFFLOAD) RTE_ETH_QINQ_STRIP_OFFLOAD
980 #define RTE_ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
981 #define RTE_ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
982 #define RTE_ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
983 #define RTE_ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */
984 #define RTE_ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
987 #define ETH_VLAN_STRIP_MASK RTE_DEPRECATED(ETH_VLAN_STRIP_MASK) RTE_ETH_VLAN_STRIP_MASK
988 #define ETH_VLAN_FILTER_MASK RTE_DEPRECATED(ETH_VLAN_FILTER_MASK) RTE_ETH_VLAN_FILTER_MASK
989 #define ETH_VLAN_EXTEND_MASK RTE_DEPRECATED(ETH_VLAN_EXTEND_MASK) RTE_ETH_VLAN_EXTEND_MASK
990 #define ETH_QINQ_STRIP_MASK RTE_DEPRECATED(ETH_QINQ_STRIP_MASK) RTE_ETH_QINQ_STRIP_MASK
991 #define ETH_VLAN_ID_MAX RTE_DEPRECATED(ETH_VLAN_ID_MAX) RTE_ETH_VLAN_ID_MAX
993 /* Definitions used for receive MAC address */
994 #define RTE_ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
995 #define ETH_NUM_RECEIVE_MAC_ADDR RTE_DEPRECATED(ETH_NUM_RECEIVE_MAC_ADDR) RTE_ETH_NUM_RECEIVE_MAC_ADDR
997 /* Definitions used for unicast hash */
998 #define RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
999 #define ETH_VMDQ_NUM_UC_HASH_ARRAY RTE_DEPRECATED(ETH_VMDQ_NUM_UC_HASH_ARRAY) RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY
1001 /**@{@name VMDq Rx mode
1002 * @see rte_eth_vmdq_rx_conf.rx_mode
1004 /** Accept untagged packets. */
1005 #define RTE_ETH_VMDQ_ACCEPT_UNTAG RTE_BIT32(0)
1006 /** Accept packets in multicast table. */
1007 #define RTE_ETH_VMDQ_ACCEPT_HASH_MC RTE_BIT32(1)
1008 /** Accept packets in unicast table. */
1009 #define RTE_ETH_VMDQ_ACCEPT_HASH_UC RTE_BIT32(2)
1010 /** Accept broadcast packets. */
1011 #define RTE_ETH_VMDQ_ACCEPT_BROADCAST RTE_BIT32(3)
1012 /** Multicast promiscuous. */
1013 #define RTE_ETH_VMDQ_ACCEPT_MULTICAST RTE_BIT32(4)
1016 #define ETH_VMDQ_ACCEPT_UNTAG RTE_DEPRECATED(ETH_VMDQ_ACCEPT_UNTAG) RTE_ETH_VMDQ_ACCEPT_UNTAG
1017 #define ETH_VMDQ_ACCEPT_HASH_MC RTE_DEPRECATED(ETH_VMDQ_ACCEPT_HASH_MC) RTE_ETH_VMDQ_ACCEPT_HASH_MC
1018 #define ETH_VMDQ_ACCEPT_HASH_UC RTE_DEPRECATED(ETH_VMDQ_ACCEPT_HASH_UC) RTE_ETH_VMDQ_ACCEPT_HASH_UC
1019 #define ETH_VMDQ_ACCEPT_BROADCAST RTE_DEPRECATED(ETH_VMDQ_ACCEPT_BROADCAST) RTE_ETH_VMDQ_ACCEPT_BROADCAST
1020 #define ETH_VMDQ_ACCEPT_MULTICAST RTE_DEPRECATED(ETH_VMDQ_ACCEPT_MULTICAST) RTE_ETH_VMDQ_ACCEPT_MULTICAST
1023 * A structure used to configure 64 entries of Redirection Table of the
1024 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
1025 * more than 64 entries supported by hardware, an array of this structure
1028 struct rte_eth_rss_reta_entry64 {
1029 /** Mask bits indicate which entries need to be updated/queried. */
1031 /** Group of 64 redirection table entries. */
1032 uint16_t reta[RTE_ETH_RETA_GROUP_SIZE];
1036 * This enum indicates the possible number of traffic classes
1037 * in DCB configurations
1039 enum rte_eth_nb_tcs {
1040 RTE_ETH_4_TCS = 4, /**< 4 TCs with DCB. */
1041 RTE_ETH_8_TCS = 8 /**< 8 TCs with DCB. */
1043 #define ETH_4_TCS RTE_DEPRECATED(ETH_4_TCS) RTE_ETH_4_TCS
1044 #define ETH_8_TCS RTE_DEPRECATED(ETH_8_TCS) RTE_ETH_8_TCS
1047 * This enum indicates the possible number of queue pools
1048 * in VMDq configurations.
1050 enum rte_eth_nb_pools {
1051 RTE_ETH_8_POOLS = 8, /**< 8 VMDq pools. */
1052 RTE_ETH_16_POOLS = 16, /**< 16 VMDq pools. */
1053 RTE_ETH_32_POOLS = 32, /**< 32 VMDq pools. */
1054 RTE_ETH_64_POOLS = 64 /**< 64 VMDq pools. */
1056 #define ETH_8_POOLS RTE_DEPRECATED(ETH_8_POOLS) RTE_ETH_8_POOLS
1057 #define ETH_16_POOLS RTE_DEPRECATED(ETH_16_POOLS) RTE_ETH_16_POOLS
1058 #define ETH_32_POOLS RTE_DEPRECATED(ETH_32_POOLS) RTE_ETH_32_POOLS
1059 #define ETH_64_POOLS RTE_DEPRECATED(ETH_64_POOLS) RTE_ETH_64_POOLS
1061 /* This structure may be extended in future. */
1062 struct rte_eth_dcb_rx_conf {
1063 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
1064 /** Traffic class each UP mapped to. */
1065 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1068 struct rte_eth_vmdq_dcb_tx_conf {
1069 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
1070 /** Traffic class each UP mapped to. */
1071 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1074 struct rte_eth_dcb_tx_conf {
1075 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
1076 /** Traffic class each UP mapped to. */
1077 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1080 struct rte_eth_vmdq_tx_conf {
1081 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
1085 * A structure used to configure the VMDq+DCB feature
1086 * of an Ethernet port.
1088 * Using this feature, packets are routed to a pool of queues, based
1089 * on the VLAN ID in the VLAN tag, and then to a specific queue within
1090 * that pool, using the user priority VLAN tag field.
1092 * A default pool may be used, if desired, to route all traffic which
1093 * does not match the VLAN filter rules.
1095 struct rte_eth_vmdq_dcb_conf {
1096 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
1097 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
1098 uint8_t default_pool; /**< The default pool, if applicable */
1099 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
1101 uint16_t vlan_id; /**< The VLAN ID of the received frame */
1102 uint64_t pools; /**< Bitmask of pools for packet Rx */
1103 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */
1104 /** Selects a queue in a pool */
1105 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1109 * A structure used to configure the VMDq feature of an Ethernet port when
1110 * not combined with the DCB feature.
1112 * Using this feature, packets are routed to a pool of queues. By default,
1113 * the pool selection is based on the MAC address, the VLAN ID in the
1114 * VLAN tag as specified in the pool_map array.
1115 * Passing the RTE_ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
1116 * selection using only the MAC address. MAC address to pool mapping is done
1117 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
1118 * corresponding to the pool ID.
1120 * Queue selection within the selected pool will be done using RSS when
1121 * it is enabled or revert to the first queue of the pool if not.
1123 * A default pool may be used, if desired, to route all traffic which
1124 * does not match the VLAN filter rules or any pool MAC address.
1126 struct rte_eth_vmdq_rx_conf {
1127 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
1128 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
1129 uint8_t default_pool; /**< The default pool, if applicable */
1130 uint8_t enable_loop_back; /**< Enable VT loop back */
1131 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
1132 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
1134 uint16_t vlan_id; /**< The VLAN ID of the received frame */
1135 uint64_t pools; /**< Bitmask of pools for packet Rx */
1136 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */
1140 * A structure used to configure the Tx features of an Ethernet port.
1142 struct rte_eth_txmode {
1143 enum rte_eth_tx_mq_mode mq_mode; /**< Tx multi-queues mode. */
1145 * Per-port Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags.
1146 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
1147 * structure are allowed to be set.
1153 uint8_t /** If set, reject sending out tagged pkts */
1154 hw_vlan_reject_tagged : 1,
1155 /** If set, reject sending out untagged pkts */
1156 hw_vlan_reject_untagged : 1,
1157 /** If set, enable port based VLAN insertion */
1158 hw_vlan_insert_pvid : 1;
1160 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1161 void *reserved_ptrs[2]; /**< Reserved for future fields */
1166 * @b EXPERIMENTAL: this structure may change without prior notice.
1168 * A structure used to configure an Rx packet segment to split.
1170 * If RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT flag is set in offloads field,
1171 * the PMD will split the received packets into multiple segments
1172 * according to the specification in the description array:
1174 * - The first network buffer will be allocated from the memory pool,
1175 * specified in the first array element, the second buffer, from the
1176 * pool in the second element, and so on.
1178 * - The offsets from the segment description elements specify
1179 * the data offset from the buffer beginning except the first mbuf.
1180 * The first segment offset is added with RTE_PKTMBUF_HEADROOM.
1182 * - The lengths in the elements define the maximal data amount
1183 * being received to each segment. The receiving starts with filling
1184 * up the first mbuf data buffer up to specified length. If the
1185 * there are data remaining (packet is longer than buffer in the first
1186 * mbuf) the following data will be pushed to the next segment
1187 * up to its own length, and so on.
1189 * - If the length in the segment description element is zero
1190 * the actual buffer size will be deduced from the appropriate
1191 * memory pool properties.
1193 * - If there is not enough elements to describe the buffer for entire
1194 * packet of maximal length the following parameters will be used
1195 * for the all remaining segments:
1196 * - pool from the last valid element
1197 * - the buffer size from this pool
1200 struct rte_eth_rxseg_split {
1201 struct rte_mempool *mp; /**< Memory pool to allocate segment from. */
1202 uint16_t length; /**< Segment data length, configures split point. */
1203 uint16_t offset; /**< Data offset from beginning of mbuf data buffer. */
1204 uint32_t reserved; /**< Reserved field. */
1209 * @b EXPERIMENTAL: this structure may change without prior notice.
1211 * A common structure used to describe Rx packet segment properties.
1213 union rte_eth_rxseg {
1214 /* The settings for buffer split offload. */
1215 struct rte_eth_rxseg_split split;
1216 /* The other features settings should be added here. */
1220 * A structure used to configure an Rx ring of an Ethernet port.
1222 struct rte_eth_rxconf {
1223 struct rte_eth_thresh rx_thresh; /**< Rx ring threshold registers. */
1224 uint16_t rx_free_thresh; /**< Drives the freeing of Rx descriptors. */
1225 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
1226 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1227 uint16_t rx_nseg; /**< Number of descriptions in rx_seg array. */
1229 * Share group index in Rx domain and switch domain.
1230 * Non-zero value to enable Rx queue share, zero value disable share.
1231 * PMD is responsible for Rx queue consistency checks to avoid member
1232 * port's configuration contradict to each other.
1234 uint16_t share_group;
1235 uint16_t share_qid; /**< Shared Rx queue ID in group */
1237 * Per-queue Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags.
1238 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
1239 * fields on rte_eth_dev_info structure are allowed to be set.
1243 * Points to the array of segment descriptions for an entire packet.
1244 * Array elements are properties for consecutive Rx segments.
1246 * The supported capabilities of receiving segmentation is reported
1247 * in rte_eth_dev_info.rx_seg_capa field.
1249 union rte_eth_rxseg *rx_seg;
1251 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1252 void *reserved_ptrs[2]; /**< Reserved for future fields */
1256 * A structure used to configure a Tx ring of an Ethernet port.
1258 struct rte_eth_txconf {
1259 struct rte_eth_thresh tx_thresh; /**< Tx ring threshold registers. */
1260 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
1261 uint16_t tx_free_thresh; /**< Start freeing Tx buffers if there are
1262 less free descriptors than this value. */
1264 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1266 * Per-queue Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags.
1267 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
1268 * fields on rte_eth_dev_info structure are allowed to be set.
1272 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1273 void *reserved_ptrs[2]; /**< Reserved for future fields */
1278 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1280 * A structure used to return the hairpin capabilities that are supported.
1282 struct rte_eth_hairpin_cap {
1283 /** The max number of hairpin queues (different bindings). */
1284 uint16_t max_nb_queues;
1285 /** Max number of Rx queues to be connected to one Tx queue. */
1286 uint16_t max_rx_2_tx;
1287 /** Max number of Tx queues to be connected to one Rx queue. */
1288 uint16_t max_tx_2_rx;
1289 uint16_t max_nb_desc; /**< The max num of descriptors. */
1292 #define RTE_ETH_MAX_HAIRPIN_PEERS 32
1296 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1298 * A structure used to hold hairpin peer data.
1300 struct rte_eth_hairpin_peer {
1301 uint16_t port; /**< Peer port. */
1302 uint16_t queue; /**< Peer queue. */
1307 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1309 * A structure used to configure hairpin binding.
1311 struct rte_eth_hairpin_conf {
1312 uint32_t peer_count:16; /**< The number of peers. */
1315 * Explicit Tx flow rule mode.
1316 * One hairpin pair of queues should have the same attribute.
1318 * - When set, the user should be responsible for inserting the hairpin
1319 * Tx part flows and removing them.
1320 * - When clear, the PMD will try to handle the Tx part of the flows,
1321 * e.g., by splitting one flow into two parts.
1323 uint32_t tx_explicit:1;
1326 * Manually bind hairpin queues.
1327 * One hairpin pair of queues should have the same attribute.
1329 * - When set, to enable hairpin, the user should call the hairpin bind
1330 * function after all the queues are set up properly and the ports are
1331 * started. Also, the hairpin unbind function should be called
1332 * accordingly before stopping a port that with hairpin configured.
1333 * - When clear, the PMD will try to enable the hairpin with the queues
1334 * configured automatically during port start.
1336 uint32_t manual_bind:1;
1337 uint32_t reserved:14; /**< Reserved bits. */
1338 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
1342 * A structure contains information about HW descriptor ring limitations.
1344 struct rte_eth_desc_lim {
1345 uint16_t nb_max; /**< Max allowed number of descriptors. */
1346 uint16_t nb_min; /**< Min allowed number of descriptors. */
1347 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
1350 * Max allowed number of segments per whole packet.
1352 * - For TSO packet this is the total number of data descriptors allowed
1355 * @see nb_mtu_seg_max
1357 uint16_t nb_seg_max;
1360 * Max number of segments per one MTU.
1362 * - For non-TSO packet, this is the maximum allowed number of segments
1363 * in a single transmit packet.
1365 * - For TSO packet each segment within the TSO may span up to this
1370 uint16_t nb_mtu_seg_max;
1374 * This enum indicates the flow control mode
1376 enum rte_eth_fc_mode {
1377 RTE_ETH_FC_NONE = 0, /**< Disable flow control. */
1378 RTE_ETH_FC_RX_PAUSE, /**< Rx pause frame, enable flowctrl on Tx side. */
1379 RTE_ETH_FC_TX_PAUSE, /**< Tx pause frame, enable flowctrl on Rx side. */
1380 RTE_ETH_FC_FULL /**< Enable flow control on both side. */
1382 #define RTE_FC_NONE RTE_DEPRECATED(RTE_FC_NONE) RTE_ETH_FC_NONE
1383 #define RTE_FC_RX_PAUSE RTE_DEPRECATED(RTE_FC_RX_PAUSE) RTE_ETH_FC_RX_PAUSE
1384 #define RTE_FC_TX_PAUSE RTE_DEPRECATED(RTE_FC_TX_PAUSE) RTE_ETH_FC_TX_PAUSE
1385 #define RTE_FC_FULL RTE_DEPRECATED(RTE_FC_FULL) RTE_ETH_FC_FULL
1388 * A structure used to configure Ethernet flow control parameter.
1389 * These parameters will be configured into the register of the NIC.
1390 * Please refer to the corresponding data sheet for proper value.
1392 struct rte_eth_fc_conf {
1393 uint32_t high_water; /**< High threshold value to trigger XOFF */
1394 uint32_t low_water; /**< Low threshold value to trigger XON */
1395 uint16_t pause_time; /**< Pause quota in the Pause frame */
1396 uint16_t send_xon; /**< Is XON frame need be sent */
1397 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1398 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
1399 uint8_t autoneg; /**< Use Pause autoneg */
1403 * A structure used to configure Ethernet priority flow control parameter.
1404 * These parameters will be configured into the register of the NIC.
1405 * Please refer to the corresponding data sheet for proper value.
1407 struct rte_eth_pfc_conf {
1408 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
1409 uint8_t priority; /**< VLAN User Priority. */
1414 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1416 * A structure used to retrieve information of queue based PFC.
1418 struct rte_eth_pfc_queue_info {
1420 * Maximum supported traffic class as per PFC (802.1Qbb) specification.
1423 /** PFC queue mode capabilities. */
1424 enum rte_eth_fc_mode mode_capa;
1429 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1431 * A structure used to configure Ethernet priority flow control parameters for
1434 * rte_eth_pfc_queue_conf::rx_pause structure shall be used to configure given
1435 * tx_qid with corresponding tc. When ethdev device receives PFC frame with
1436 * rte_eth_pfc_queue_conf::rx_pause::tc, traffic will be paused on
1437 * rte_eth_pfc_queue_conf::rx_pause::tx_qid for that tc.
1439 * rte_eth_pfc_queue_conf::tx_pause structure shall be used to configure given
1440 * rx_qid. When rx_qid is congested, PFC frames are generated with
1441 * rte_eth_pfc_queue_conf::rx_pause::tc and
1442 * rte_eth_pfc_queue_conf::rx_pause::pause_time to the peer.
1444 struct rte_eth_pfc_queue_conf {
1445 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1448 uint16_t tx_qid; /**< Tx queue ID */
1449 /** Traffic class as per PFC (802.1Qbb) spec. The value must be
1450 * in the range [0, rte_eth_pfc_queue_info::tx_max - 1]
1453 } rx_pause; /* Valid when (mode == FC_RX_PAUSE || mode == FC_FULL) */
1456 uint16_t pause_time; /**< Pause quota in the Pause frame */
1457 uint16_t rx_qid; /**< Rx queue ID */
1458 /** Traffic class as per PFC (802.1Qbb) spec. The value must be
1459 * in the range [0, rte_eth_pfc_queue_info::tx_max - 1]
1462 } tx_pause; /* Valid when (mode == FC_TX_PAUSE || mode == FC_FULL) */
1466 * Tunnel type for device-specific classifier configuration.
1467 * @see rte_eth_udp_tunnel
1469 enum rte_eth_tunnel_type {
1470 RTE_ETH_TUNNEL_TYPE_NONE = 0,
1471 RTE_ETH_TUNNEL_TYPE_VXLAN,
1472 RTE_ETH_TUNNEL_TYPE_GENEVE,
1473 RTE_ETH_TUNNEL_TYPE_TEREDO,
1474 RTE_ETH_TUNNEL_TYPE_NVGRE,
1475 RTE_ETH_TUNNEL_TYPE_IP_IN_GRE,
1476 RTE_ETH_L2_TUNNEL_TYPE_E_TAG,
1477 RTE_ETH_TUNNEL_TYPE_VXLAN_GPE,
1478 RTE_ETH_TUNNEL_TYPE_ECPRI,
1479 RTE_ETH_TUNNEL_TYPE_MAX,
1481 #define RTE_TUNNEL_TYPE_NONE RTE_DEPRECATED(RTE_TUNNEL_TYPE_NONE) RTE_ETH_TUNNEL_TYPE_NONE
1482 #define RTE_TUNNEL_TYPE_VXLAN RTE_DEPRECATED(RTE_TUNNEL_TYPE_VXLAN) RTE_ETH_TUNNEL_TYPE_VXLAN
1483 #define RTE_TUNNEL_TYPE_GENEVE RTE_DEPRECATED(RTE_TUNNEL_TYPE_GENEVE) RTE_ETH_TUNNEL_TYPE_GENEVE
1484 #define RTE_TUNNEL_TYPE_TEREDO RTE_DEPRECATED(RTE_TUNNEL_TYPE_TEREDO) RTE_ETH_TUNNEL_TYPE_TEREDO
1485 #define RTE_TUNNEL_TYPE_NVGRE RTE_DEPRECATED(RTE_TUNNEL_TYPE_NVGRE) RTE_ETH_TUNNEL_TYPE_NVGRE
1486 #define RTE_TUNNEL_TYPE_IP_IN_GRE RTE_DEPRECATED(RTE_TUNNEL_TYPE_IP_IN_GRE) RTE_ETH_TUNNEL_TYPE_IP_IN_GRE
1487 #define RTE_L2_TUNNEL_TYPE_E_TAG RTE_DEPRECATED(RTE_L2_TUNNEL_TYPE_E_TAG) RTE_ETH_L2_TUNNEL_TYPE_E_TAG
1488 #define RTE_TUNNEL_TYPE_VXLAN_GPE RTE_DEPRECATED(RTE_TUNNEL_TYPE_VXLAN_GPE) RTE_ETH_TUNNEL_TYPE_VXLAN_GPE
1489 #define RTE_TUNNEL_TYPE_ECPRI RTE_DEPRECATED(RTE_TUNNEL_TYPE_ECPRI) RTE_ETH_TUNNEL_TYPE_ECPRI
1490 #define RTE_TUNNEL_TYPE_MAX RTE_DEPRECATED(RTE_TUNNEL_TYPE_MAX) RTE_ETH_TUNNEL_TYPE_MAX
1492 /* Deprecated API file for rte_eth_dev_filter_* functions */
1493 #include "rte_eth_ctrl.h"
1496 * Memory space that can be configured to store Flow Director filters
1497 * in the board memory.
1499 enum rte_eth_fdir_pballoc_type {
1500 RTE_ETH_FDIR_PBALLOC_64K = 0, /**< 64k. */
1501 RTE_ETH_FDIR_PBALLOC_128K, /**< 128k. */
1502 RTE_ETH_FDIR_PBALLOC_256K, /**< 256k. */
1504 #define rte_fdir_pballoc_type rte_eth_fdir_pballoc_type
1506 #define RTE_FDIR_PBALLOC_64K RTE_DEPRECATED(RTE_FDIR_PBALLOC_64K) RTE_ETH_FDIR_PBALLOC_64K
1507 #define RTE_FDIR_PBALLOC_128K RTE_DEPRECATED(RTE_FDIR_PBALLOC_128K) RTE_ETH_FDIR_PBALLOC_128K
1508 #define RTE_FDIR_PBALLOC_256K RTE_DEPRECATED(RTE_FDIR_PBALLOC_256K) RTE_ETH_FDIR_PBALLOC_256K
1511 * Select report mode of FDIR hash information in Rx descriptors.
1513 enum rte_fdir_status_mode {
1514 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
1515 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
1516 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
1520 * A structure used to configure the Flow Director (FDIR) feature
1521 * of an Ethernet port.
1523 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored.
1525 struct rte_eth_fdir_conf {
1526 enum rte_fdir_mode mode; /**< Flow Director mode. */
1527 enum rte_eth_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
1528 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
1529 /** Rx queue of packets matching a "drop" filter in perfect mode. */
1531 struct rte_eth_fdir_masks mask;
1532 /** Flex payload configuration. */
1533 struct rte_eth_fdir_flex_conf flex_conf;
1535 #define rte_fdir_conf rte_eth_fdir_conf
1538 * UDP tunneling configuration.
1540 * Used to configure the classifier of a device,
1541 * associating an UDP port with a type of tunnel.
1543 * Some NICs may need such configuration to properly parse a tunnel
1544 * with any standard or custom UDP port.
1546 struct rte_eth_udp_tunnel {
1547 uint16_t udp_port; /**< UDP port used for the tunnel. */
1548 uint8_t prot_type; /**< Tunnel type. @see rte_eth_tunnel_type */
1552 * A structure used to enable/disable specific device interrupts.
1554 struct rte_eth_intr_conf {
1555 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
1557 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
1559 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
1563 #define rte_intr_conf rte_eth_intr_conf
1566 * A structure used to configure an Ethernet port.
1567 * Depending upon the Rx multi-queue mode, extra advanced
1568 * configuration settings may be needed.
1570 struct rte_eth_conf {
1571 uint32_t link_speeds; /**< bitmap of RTE_ETH_LINK_SPEED_XXX of speeds to be
1572 used. RTE_ETH_LINK_SPEED_FIXED disables link
1573 autonegotiation, and a unique speed shall be
1574 set. Otherwise, the bitmap defines the set of
1575 speeds to be advertised. If the special value
1576 RTE_ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
1577 supported are advertised. */
1578 struct rte_eth_rxmode rxmode; /**< Port Rx configuration. */
1579 struct rte_eth_txmode txmode; /**< Port Tx configuration. */
1580 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
1581 is 0, meaning the loopback mode is disabled.
1582 Read the datasheet of given Ethernet controller
1583 for details. The possible values of this field
1584 are defined in implementation of each driver. */
1586 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
1587 /** Port VMDq+DCB configuration. */
1588 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
1589 /** Port DCB Rx configuration. */
1590 struct rte_eth_dcb_rx_conf dcb_rx_conf;
1591 /** Port VMDq Rx configuration. */
1592 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
1593 } rx_adv_conf; /**< Port Rx filtering configuration. */
1595 /** Port VMDq+DCB Tx configuration. */
1596 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
1597 /** Port DCB Tx configuration. */
1598 struct rte_eth_dcb_tx_conf dcb_tx_conf;
1599 /** Port VMDq Tx configuration. */
1600 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
1601 } tx_adv_conf; /**< Port Tx DCB configuration (union). */
1602 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
1603 is needed,and the variable must be set RTE_ETH_DCB_PFC_SUPPORT. */
1604 uint32_t dcb_capability_en;
1605 struct rte_eth_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */
1606 struct rte_eth_intr_conf intr_conf; /**< Interrupt mode configuration. */
1610 * Rx offload capabilities of a device.
1612 #define RTE_ETH_RX_OFFLOAD_VLAN_STRIP RTE_BIT64(0)
1613 #define RTE_ETH_RX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1)
1614 #define RTE_ETH_RX_OFFLOAD_UDP_CKSUM RTE_BIT64(2)
1615 #define RTE_ETH_RX_OFFLOAD_TCP_CKSUM RTE_BIT64(3)
1616 #define RTE_ETH_RX_OFFLOAD_TCP_LRO RTE_BIT64(4)
1617 #define RTE_ETH_RX_OFFLOAD_QINQ_STRIP RTE_BIT64(5)
1618 #define RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(6)
1619 #define RTE_ETH_RX_OFFLOAD_MACSEC_STRIP RTE_BIT64(7)
1620 #define RTE_ETH_RX_OFFLOAD_HEADER_SPLIT RTE_BIT64(8)
1621 #define RTE_ETH_RX_OFFLOAD_VLAN_FILTER RTE_BIT64(9)
1622 #define RTE_ETH_RX_OFFLOAD_VLAN_EXTEND RTE_BIT64(10)
1623 #define RTE_ETH_RX_OFFLOAD_SCATTER RTE_BIT64(13)
1625 * Timestamp is set by the driver in RTE_MBUF_DYNFIELD_TIMESTAMP_NAME
1626 * and RTE_MBUF_DYNFLAG_RX_TIMESTAMP_NAME is set in ol_flags.
1627 * The mbuf field and flag are registered when the offload is configured.
1629 #define RTE_ETH_RX_OFFLOAD_TIMESTAMP RTE_BIT64(14)
1630 #define RTE_ETH_RX_OFFLOAD_SECURITY RTE_BIT64(15)
1631 #define RTE_ETH_RX_OFFLOAD_KEEP_CRC RTE_BIT64(16)
1632 #define RTE_ETH_RX_OFFLOAD_SCTP_CKSUM RTE_BIT64(17)
1633 #define RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(18)
1634 #define RTE_ETH_RX_OFFLOAD_RSS_HASH RTE_BIT64(19)
1635 #define RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT RTE_BIT64(20)
1637 #define DEV_RX_OFFLOAD_VLAN_STRIP RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN_STRIP) RTE_ETH_RX_OFFLOAD_VLAN_STRIP
1638 #define DEV_RX_OFFLOAD_IPV4_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_IPV4_CKSUM) RTE_ETH_RX_OFFLOAD_IPV4_CKSUM
1639 #define DEV_RX_OFFLOAD_UDP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_UDP_CKSUM) RTE_ETH_RX_OFFLOAD_UDP_CKSUM
1640 #define DEV_RX_OFFLOAD_TCP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_TCP_CKSUM) RTE_ETH_RX_OFFLOAD_TCP_CKSUM
1641 #define DEV_RX_OFFLOAD_TCP_LRO RTE_DEPRECATED(DEV_RX_OFFLOAD_TCP_LRO) RTE_ETH_RX_OFFLOAD_TCP_LRO
1642 #define DEV_RX_OFFLOAD_QINQ_STRIP RTE_DEPRECATED(DEV_RX_OFFLOAD_QINQ_STRIP) RTE_ETH_RX_OFFLOAD_QINQ_STRIP
1643 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM) RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM
1644 #define DEV_RX_OFFLOAD_MACSEC_STRIP RTE_DEPRECATED(DEV_RX_OFFLOAD_MACSEC_STRIP) RTE_ETH_RX_OFFLOAD_MACSEC_STRIP
1645 #define DEV_RX_OFFLOAD_HEADER_SPLIT RTE_DEPRECATED(DEV_RX_OFFLOAD_HEADER_SPLIT) RTE_ETH_RX_OFFLOAD_HEADER_SPLIT
1646 #define DEV_RX_OFFLOAD_VLAN_FILTER RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN_FILTER) RTE_ETH_RX_OFFLOAD_VLAN_FILTER
1647 #define DEV_RX_OFFLOAD_VLAN_EXTEND RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN_EXTEND) RTE_ETH_RX_OFFLOAD_VLAN_EXTEND
1648 #define DEV_RX_OFFLOAD_SCATTER RTE_DEPRECATED(DEV_RX_OFFLOAD_SCATTER) RTE_ETH_RX_OFFLOAD_SCATTER
1649 #define DEV_RX_OFFLOAD_TIMESTAMP RTE_DEPRECATED(DEV_RX_OFFLOAD_TIMESTAMP) RTE_ETH_RX_OFFLOAD_TIMESTAMP
1650 #define DEV_RX_OFFLOAD_SECURITY RTE_DEPRECATED(DEV_RX_OFFLOAD_SECURITY) RTE_ETH_RX_OFFLOAD_SECURITY
1651 #define DEV_RX_OFFLOAD_KEEP_CRC RTE_DEPRECATED(DEV_RX_OFFLOAD_KEEP_CRC) RTE_ETH_RX_OFFLOAD_KEEP_CRC
1652 #define DEV_RX_OFFLOAD_SCTP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_SCTP_CKSUM) RTE_ETH_RX_OFFLOAD_SCTP_CKSUM
1653 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_OUTER_UDP_CKSUM) RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM
1654 #define DEV_RX_OFFLOAD_RSS_HASH RTE_DEPRECATED(DEV_RX_OFFLOAD_RSS_HASH) RTE_ETH_RX_OFFLOAD_RSS_HASH
1656 #define RTE_ETH_RX_OFFLOAD_CHECKSUM (RTE_ETH_RX_OFFLOAD_IPV4_CKSUM | \
1657 RTE_ETH_RX_OFFLOAD_UDP_CKSUM | \
1658 RTE_ETH_RX_OFFLOAD_TCP_CKSUM)
1659 #define DEV_RX_OFFLOAD_CHECKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_CHECKSUM) RTE_ETH_RX_OFFLOAD_CHECKSUM
1660 #define RTE_ETH_RX_OFFLOAD_VLAN (RTE_ETH_RX_OFFLOAD_VLAN_STRIP | \
1661 RTE_ETH_RX_OFFLOAD_VLAN_FILTER | \
1662 RTE_ETH_RX_OFFLOAD_VLAN_EXTEND | \
1663 RTE_ETH_RX_OFFLOAD_QINQ_STRIP)
1664 #define DEV_RX_OFFLOAD_VLAN RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN) RTE_ETH_RX_OFFLOAD_VLAN
1667 * If new Rx offload capabilities are defined, they also must be
1668 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
1672 * Tx offload capabilities of a device.
1674 #define RTE_ETH_TX_OFFLOAD_VLAN_INSERT RTE_BIT64(0)
1675 #define RTE_ETH_TX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1)
1676 #define RTE_ETH_TX_OFFLOAD_UDP_CKSUM RTE_BIT64(2)
1677 #define RTE_ETH_TX_OFFLOAD_TCP_CKSUM RTE_BIT64(3)
1678 #define RTE_ETH_TX_OFFLOAD_SCTP_CKSUM RTE_BIT64(4)
1679 #define RTE_ETH_TX_OFFLOAD_TCP_TSO RTE_BIT64(5)
1680 #define RTE_ETH_TX_OFFLOAD_UDP_TSO RTE_BIT64(6)
1681 #define RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(7) /**< Used for tunneling packet. */
1682 #define RTE_ETH_TX_OFFLOAD_QINQ_INSERT RTE_BIT64(8)
1683 #define RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO RTE_BIT64(9) /**< Used for tunneling packet. */
1684 #define RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO RTE_BIT64(10) /**< Used for tunneling packet. */
1685 #define RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO RTE_BIT64(11) /**< Used for tunneling packet. */
1686 #define RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO RTE_BIT64(12) /**< Used for tunneling packet. */
1687 #define RTE_ETH_TX_OFFLOAD_MACSEC_INSERT RTE_BIT64(13)
1689 * Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
1690 * Tx queue without SW lock.
1692 #define RTE_ETH_TX_OFFLOAD_MT_LOCKFREE RTE_BIT64(14)
1693 /** Device supports multi segment send. */
1694 #define RTE_ETH_TX_OFFLOAD_MULTI_SEGS RTE_BIT64(15)
1696 * Device supports optimization for fast release of mbufs.
1697 * When set application must guarantee that per-queue all mbufs comes from
1698 * the same mempool and has refcnt = 1.
1700 #define RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE RTE_BIT64(16)
1701 #define RTE_ETH_TX_OFFLOAD_SECURITY RTE_BIT64(17)
1703 * Device supports generic UDP tunneled packet TSO.
1704 * Application must set RTE_MBUF_F_TX_TUNNEL_UDP and other mbuf fields required
1707 #define RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO RTE_BIT64(18)
1709 * Device supports generic IP tunneled packet TSO.
1710 * Application must set RTE_MBUF_F_TX_TUNNEL_IP and other mbuf fields required
1713 #define RTE_ETH_TX_OFFLOAD_IP_TNL_TSO RTE_BIT64(19)
1714 /** Device supports outer UDP checksum */
1715 #define RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(20)
1717 * Device sends on time read from RTE_MBUF_DYNFIELD_TIMESTAMP_NAME
1718 * if RTE_MBUF_DYNFLAG_TX_TIMESTAMP_NAME is set in ol_flags.
1719 * The mbuf field and flag are registered when the offload is configured.
1721 #define RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_BIT64(21)
1723 * If new Tx offload capabilities are defined, they also must be
1724 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1727 #define DEV_TX_OFFLOAD_VLAN_INSERT RTE_DEPRECATED(DEV_TX_OFFLOAD_VLAN_INSERT) RTE_ETH_TX_OFFLOAD_VLAN_INSERT
1728 #define DEV_TX_OFFLOAD_IPV4_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_IPV4_CKSUM) RTE_ETH_TX_OFFLOAD_IPV4_CKSUM
1729 #define DEV_TX_OFFLOAD_UDP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_UDP_CKSUM) RTE_ETH_TX_OFFLOAD_UDP_CKSUM
1730 #define DEV_TX_OFFLOAD_TCP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_TCP_CKSUM) RTE_ETH_TX_OFFLOAD_TCP_CKSUM
1731 #define DEV_TX_OFFLOAD_SCTP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_SCTP_CKSUM) RTE_ETH_TX_OFFLOAD_SCTP_CKSUM
1732 #define DEV_TX_OFFLOAD_TCP_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_TCP_TSO) RTE_ETH_TX_OFFLOAD_TCP_TSO
1733 #define DEV_TX_OFFLOAD_UDP_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_UDP_TSO) RTE_ETH_TX_OFFLOAD_UDP_TSO
1734 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM) RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
1735 #define DEV_TX_OFFLOAD_QINQ_INSERT RTE_DEPRECATED(DEV_TX_OFFLOAD_QINQ_INSERT) RTE_ETH_TX_OFFLOAD_QINQ_INSERT
1736 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_VXLAN_TNL_TSO) RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO
1737 #define DEV_TX_OFFLOAD_GRE_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_GRE_TNL_TSO) RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO
1738 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_IPIP_TNL_TSO) RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO
1739 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_GENEVE_TNL_TSO) RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO
1740 #define DEV_TX_OFFLOAD_MACSEC_INSERT RTE_DEPRECATED(DEV_TX_OFFLOAD_MACSEC_INSERT) RTE_ETH_TX_OFFLOAD_MACSEC_INSERT
1741 #define DEV_TX_OFFLOAD_MT_LOCKFREE RTE_DEPRECATED(DEV_TX_OFFLOAD_MT_LOCKFREE) RTE_ETH_TX_OFFLOAD_MT_LOCKFREE
1742 #define DEV_TX_OFFLOAD_MULTI_SEGS RTE_DEPRECATED(DEV_TX_OFFLOAD_MULTI_SEGS) RTE_ETH_TX_OFFLOAD_MULTI_SEGS
1743 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE RTE_DEPRECATED(DEV_TX_OFFLOAD_MBUF_FAST_FREE) RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE
1744 #define DEV_TX_OFFLOAD_SECURITY RTE_DEPRECATED(DEV_TX_OFFLOAD_SECURITY) RTE_ETH_TX_OFFLOAD_SECURITY
1745 #define DEV_TX_OFFLOAD_UDP_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_UDP_TNL_TSO) RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO
1746 #define DEV_TX_OFFLOAD_IP_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_IP_TNL_TSO) RTE_ETH_TX_OFFLOAD_IP_TNL_TSO
1747 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_OUTER_UDP_CKSUM) RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM
1748 #define DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_DEPRECATED(DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP) RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP
1750 /**@{@name Device capabilities
1751 * Non-offload capabilities reported in rte_eth_dev_info.dev_capa.
1753 /** Device supports Rx queue setup after device started. */
1754 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP RTE_BIT64(0)
1755 /** Device supports Tx queue setup after device started. */
1756 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP RTE_BIT64(1)
1758 * Device supports shared Rx queue among ports within Rx domain and
1759 * switch domain. Mbufs are consumed by shared Rx queue instead of
1760 * each queue. Multiple groups are supported by share_group of Rx
1761 * queue configuration. Shared Rx queue is identified by PMD using
1762 * share_qid of Rx queue configuration. Polling any port in the group
1763 * receive packets of all member ports, source port identified by
1766 #define RTE_ETH_DEV_CAPA_RXQ_SHARE RTE_BIT64(2)
1767 /** Device supports keeping flow rules across restart. */
1768 #define RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP RTE_BIT64(3)
1769 /** Device supports keeping shared flow objects across restart. */
1770 #define RTE_ETH_DEV_CAPA_FLOW_SHARED_OBJECT_KEEP RTE_BIT64(4)
1774 * Fallback default preferred Rx/Tx port parameters.
1775 * These are used if an application requests default parameters
1776 * but the PMD does not provide preferred values.
1778 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1779 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1780 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1781 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1784 * Preferred Rx/Tx port parameters.
1785 * There are separate instances of this structure for transmission
1786 * and reception respectively.
1788 struct rte_eth_dev_portconf {
1789 uint16_t burst_size; /**< Device-preferred burst size */
1790 uint16_t ring_size; /**< Device-preferred size of queue rings */
1791 uint16_t nb_queues; /**< Device-preferred number of queues */
1795 * Default values for switch domain ID when ethdev does not support switch
1796 * domain definitions.
1798 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX)
1801 * Ethernet device associated switch information
1803 struct rte_eth_switch_info {
1804 const char *name; /**< switch name */
1805 uint16_t domain_id; /**< switch domain ID */
1807 * Mapping to the devices physical switch port as enumerated from the
1808 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1809 * device this may correspond to the VF_ID of each virtual function,
1810 * but each driver should explicitly define the mapping of switch
1811 * port identifier to that physical interconnect/switch
1815 * Shared Rx queue sub-domain boundary. Only ports in same Rx domain
1816 * and switch domain can share Rx queue. Valid only if device advertised
1817 * RTE_ETH_DEV_CAPA_RXQ_SHARE capability.
1824 * @b EXPERIMENTAL: this structure may change without prior notice.
1826 * Ethernet device Rx buffer segmentation capabilities.
1828 struct rte_eth_rxseg_capa {
1830 uint32_t multi_pools:1; /**< Supports receiving to multiple pools.*/
1831 uint32_t offset_allowed:1; /**< Supports buffer offsets. */
1832 uint32_t offset_align_log2:4; /**< Required offset alignment. */
1833 uint16_t max_nseg; /**< Maximum amount of segments to split. */
1834 uint16_t reserved; /**< Reserved field. */
1838 * Ethernet device information
1842 * Ethernet device representor port type.
1844 enum rte_eth_representor_type {
1845 RTE_ETH_REPRESENTOR_NONE, /**< not a representor. */
1846 RTE_ETH_REPRESENTOR_VF, /**< representor of Virtual Function. */
1847 RTE_ETH_REPRESENTOR_SF, /**< representor of Sub Function. */
1848 RTE_ETH_REPRESENTOR_PF, /**< representor of Physical Function. */
1852 * A structure used to retrieve the contextual information of
1853 * an Ethernet device, such as the controlling driver of the
1856 struct rte_eth_dev_info {
1857 struct rte_device *device; /**< Generic device information */
1858 const char *driver_name; /**< Device Driver name. */
1859 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1860 Use if_indextoname() to translate into an interface name. */
1861 uint16_t min_mtu; /**< Minimum MTU allowed */
1862 uint16_t max_mtu; /**< Maximum MTU allowed */
1863 const uint32_t *dev_flags; /**< Device flags */
1864 uint32_t min_rx_bufsize; /**< Minimum size of Rx buffer. */
1865 uint32_t max_rx_pktlen; /**< Maximum configurable length of Rx pkt. */
1866 /** Maximum configurable size of LRO aggregated packet. */
1867 uint32_t max_lro_pkt_size;
1868 uint16_t max_rx_queues; /**< Maximum number of Rx queues. */
1869 uint16_t max_tx_queues; /**< Maximum number of Tx queues. */
1870 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1871 /** Maximum number of hash MAC addresses for MTA and UTA. */
1872 uint32_t max_hash_mac_addrs;
1873 uint16_t max_vfs; /**< Maximum number of VFs. */
1874 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1875 struct rte_eth_rxseg_capa rx_seg_capa; /**< Segmentation capability.*/
1876 /** All Rx offload capabilities including all per-queue ones */
1877 uint64_t rx_offload_capa;
1878 /** All Tx offload capabilities including all per-queue ones */
1879 uint64_t tx_offload_capa;
1880 /** Device per-queue Rx offload capabilities. */
1881 uint64_t rx_queue_offload_capa;
1882 /** Device per-queue Tx offload capabilities. */
1883 uint64_t tx_queue_offload_capa;
1884 /** Device redirection table size, the total number of entries. */
1886 uint8_t hash_key_size; /**< Hash key size in bytes */
1887 /** Bit mask of RSS offloads, the bit offset also means flow type */
1888 uint64_t flow_type_rss_offloads;
1889 struct rte_eth_rxconf default_rxconf; /**< Default Rx configuration */
1890 struct rte_eth_txconf default_txconf; /**< Default Tx configuration */
1891 uint16_t vmdq_queue_base; /**< First queue ID for VMDq pools. */
1892 uint16_t vmdq_queue_num; /**< Queue number for VMDq pools. */
1893 uint16_t vmdq_pool_base; /**< First ID of VMDq pools. */
1894 struct rte_eth_desc_lim rx_desc_lim; /**< Rx descriptors limits */
1895 struct rte_eth_desc_lim tx_desc_lim; /**< Tx descriptors limits */
1896 uint32_t speed_capa; /**< Supported speeds bitmap (RTE_ETH_LINK_SPEED_). */
1897 /** Configured number of Rx/Tx queues */
1898 uint16_t nb_rx_queues; /**< Number of Rx queues. */
1899 uint16_t nb_tx_queues; /**< Number of Tx queues. */
1900 /** Rx parameter recommendations */
1901 struct rte_eth_dev_portconf default_rxportconf;
1902 /** Tx parameter recommendations */
1903 struct rte_eth_dev_portconf default_txportconf;
1904 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1907 * Switching information for ports on a device with a
1908 * embedded managed interconnect/switch.
1910 struct rte_eth_switch_info switch_info;
1912 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1913 void *reserved_ptrs[2]; /**< Reserved for future fields */
1916 /**@{@name Rx/Tx queue states */
1917 #define RTE_ETH_QUEUE_STATE_STOPPED 0 /**< Queue stopped. */
1918 #define RTE_ETH_QUEUE_STATE_STARTED 1 /**< Queue started. */
1919 #define RTE_ETH_QUEUE_STATE_HAIRPIN 2 /**< Queue used for hairpin. */
1923 * Ethernet device Rx queue information structure.
1924 * Used to retrieve information about configured queue.
1926 struct rte_eth_rxq_info {
1927 struct rte_mempool *mp; /**< mempool used by that queue. */
1928 struct rte_eth_rxconf conf; /**< queue config parameters. */
1929 uint8_t scattered_rx; /**< scattered packets Rx supported. */
1930 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */
1931 uint16_t nb_desc; /**< configured number of RXDs. */
1932 uint16_t rx_buf_size; /**< hardware receive buffer size. */
1934 * Available Rx descriptors threshold defined as percentage
1935 * of Rx queue size. If number of available descriptors is lower,
1936 * the event RTE_ETH_EVENT_RX_AVAIL_THESH is generated.
1937 * Value 0 means that the threshold monitoring is disabled.
1939 uint8_t avail_thresh;
1940 } __rte_cache_min_aligned;
1943 * Ethernet device Tx queue information structure.
1944 * Used to retrieve information about configured queue.
1946 struct rte_eth_txq_info {
1947 struct rte_eth_txconf conf; /**< queue config parameters. */
1948 uint16_t nb_desc; /**< configured number of TXDs. */
1949 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */
1950 } __rte_cache_min_aligned;
1952 /* Generic Burst mode flag definition, values can be ORed. */
1955 * If the queues have different burst mode description, this bit will be set
1956 * by PMD, then the application can iterate to retrieve burst description for
1959 #define RTE_ETH_BURST_FLAG_PER_QUEUE RTE_BIT64(0)
1962 * Ethernet device Rx/Tx queue packet burst mode information structure.
1963 * Used to retrieve information about packet burst mode setting.
1965 struct rte_eth_burst_mode {
1966 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */
1968 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */
1969 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */
1972 /** Maximum name length for extended statistics counters */
1973 #define RTE_ETH_XSTATS_NAME_SIZE 64
1976 * An Ethernet device extended statistic structure
1978 * This structure is used by rte_eth_xstats_get() to provide
1979 * statistics that are not provided in the generic *rte_eth_stats*
1981 * It maps a name ID, corresponding to an index in the array returned
1982 * by rte_eth_xstats_get_names(), to a statistic value.
1984 struct rte_eth_xstat {
1985 uint64_t id; /**< The index in xstats name array. */
1986 uint64_t value; /**< The statistic counter value. */
1990 * A name element for extended statistics.
1992 * An array of this structure is returned by rte_eth_xstats_get_names().
1993 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1994 * structure references these names by their array index.
1996 * The xstats should follow a common naming scheme.
1997 * Some names are standardized in rte_stats_strings.
1999 * - rx_missed_errors
2001 * - tx_size_128_to_255_packets
2003 struct rte_eth_xstat_name {
2004 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
2007 #define RTE_ETH_DCB_NUM_TCS 8
2008 #define RTE_ETH_MAX_VMDQ_POOL 64
2010 #define ETH_DCB_NUM_TCS RTE_DEPRECATED(ETH_DCB_NUM_TCS) RTE_ETH_DCB_NUM_TCS
2011 #define ETH_MAX_VMDQ_POOL RTE_DEPRECATED(ETH_MAX_VMDQ_POOL) RTE_ETH_MAX_VMDQ_POOL
2014 * A structure used to get the information of queue and
2015 * TC mapping on both Tx and Rx paths.
2017 struct rte_eth_dcb_tc_queue_mapping {
2018 /** Rx queues assigned to tc per Pool */
2022 } tc_rxq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS];
2023 /** Rx queues assigned to tc per Pool */
2027 } tc_txq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS];
2031 * A structure used to get the information of DCB.
2032 * It includes TC UP mapping and queue TC mapping.
2034 struct rte_eth_dcb_info {
2035 uint8_t nb_tcs; /**< number of TCs */
2036 uint8_t prio_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
2037 uint8_t tc_bws[RTE_ETH_DCB_NUM_TCS]; /**< Tx BW percentage for each TC */
2038 /** Rx queues assigned to tc */
2039 struct rte_eth_dcb_tc_queue_mapping tc_queue;
2043 * This enum indicates the possible Forward Error Correction (FEC) modes
2044 * of an ethdev port.
2046 enum rte_eth_fec_mode {
2047 RTE_ETH_FEC_NOFEC = 0, /**< FEC is off */
2048 RTE_ETH_FEC_AUTO, /**< FEC autonegotiation modes */
2049 RTE_ETH_FEC_BASER, /**< FEC using common algorithm */
2050 RTE_ETH_FEC_RS, /**< FEC using RS algorithm */
2053 /* Translate from FEC mode to FEC capa */
2054 #define RTE_ETH_FEC_MODE_TO_CAPA(x) RTE_BIT32(x)
2056 /* This macro indicates FEC capa mask */
2057 #define RTE_ETH_FEC_MODE_CAPA_MASK(x) RTE_BIT32(RTE_ETH_FEC_ ## x)
2059 /* A structure used to get capabilities per link speed */
2060 struct rte_eth_fec_capa {
2061 uint32_t speed; /**< Link speed (see RTE_ETH_SPEED_NUM_*) */
2062 uint32_t capa; /**< FEC capabilities bitmask */
2065 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
2067 /* Macros to check for valid port */
2068 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
2069 if (!rte_eth_dev_is_valid_port(port_id)) { \
2070 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
2075 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
2076 if (!rte_eth_dev_is_valid_port(port_id)) { \
2077 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
2083 * Function type used for Rx packet processing packet callbacks.
2085 * The callback function is called on Rx with a burst of packets that have
2086 * been received on the given port and queue.
2089 * The Ethernet port on which Rx is being performed.
2091 * The queue on the Ethernet port which is being used to receive the packets.
2093 * The burst of packets that have just been received.
2095 * The number of packets in the burst pointed to by "pkts".
2097 * The max number of packets that can be stored in the "pkts" array.
2099 * The arbitrary user parameter passed in by the application when the callback
2100 * was originally configured.
2102 * The number of packets returned to the user.
2104 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
2105 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
2109 * Function type used for Tx packet processing packet callbacks.
2111 * The callback function is called on Tx with a burst of packets immediately
2112 * before the packets are put onto the hardware queue for transmission.
2115 * The Ethernet port on which Tx is being performed.
2117 * The queue on the Ethernet port which is being used to transmit the packets.
2119 * The burst of packets that are about to be transmitted.
2121 * The number of packets in the burst pointed to by "pkts".
2123 * The arbitrary user parameter passed in by the application when the callback
2124 * was originally configured.
2126 * The number of packets to be written to the NIC.
2128 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
2129 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
2132 * Possible states of an ethdev port.
2134 enum rte_eth_dev_state {
2135 /** Device is unused before being probed. */
2136 RTE_ETH_DEV_UNUSED = 0,
2137 /** Device is attached when allocated in probing. */
2138 RTE_ETH_DEV_ATTACHED,
2139 /** Device is in removed state when plug-out is detected. */
2140 RTE_ETH_DEV_REMOVED,
2143 struct rte_eth_dev_sriov {
2144 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
2145 uint8_t nb_q_per_pool; /**< Rx queue number per pool */
2146 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
2147 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
2149 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
2151 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
2153 #define RTE_ETH_DEV_NO_OWNER 0
2155 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
2157 struct rte_eth_dev_owner {
2158 uint64_t id; /**< The owner unique identifier. */
2159 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
2162 /**@{@name Device flags
2163 * Flags internally saved in rte_eth_dev_data.dev_flags
2164 * and reported in rte_eth_dev_info.dev_flags.
2166 /** PMD supports thread-safe flow operations */
2167 #define RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE RTE_BIT32(0)
2168 /** Device supports link state interrupt */
2169 #define RTE_ETH_DEV_INTR_LSC RTE_BIT32(1)
2170 /** Device is a bonded slave */
2171 #define RTE_ETH_DEV_BONDED_SLAVE RTE_BIT32(2)
2172 /** Device supports device removal interrupt */
2173 #define RTE_ETH_DEV_INTR_RMV RTE_BIT32(3)
2174 /** Device is port representor */
2175 #define RTE_ETH_DEV_REPRESENTOR RTE_BIT32(4)
2176 /** Device does not support MAC change after started */
2177 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR RTE_BIT32(5)
2179 * Queue xstats filled automatically by ethdev layer.
2180 * PMDs filling the queue xstats themselves should not set this flag
2182 #define RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS RTE_BIT32(6)
2186 * Iterates over valid ethdev ports owned by a specific owner.
2189 * The ID of the next possible valid owned port.
2191 * The owner identifier.
2192 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
2194 * Next valid port ID owned by owner_id, RTE_MAX_ETHPORTS if there is none.
2196 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
2197 const uint64_t owner_id);
2200 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
2202 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
2203 for (p = rte_eth_find_next_owned_by(0, o); \
2204 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
2205 p = rte_eth_find_next_owned_by(p + 1, o))
2208 * Iterates over valid ethdev ports.
2211 * The ID of the next possible valid port.
2213 * Next valid port ID, RTE_MAX_ETHPORTS if there is none.
2215 uint16_t rte_eth_find_next(uint16_t port_id);
2218 * Macro to iterate over all enabled and ownerless ethdev ports.
2220 #define RTE_ETH_FOREACH_DEV(p) \
2221 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
2224 * Iterates over ethdev ports of a specified device.
2226 * @param port_id_start
2227 * The ID of the next possible valid port.
2229 * The generic device behind the ports to iterate.
2231 * Next port ID of the device, possibly port_id_start,
2232 * RTE_MAX_ETHPORTS if there is none.
2235 rte_eth_find_next_of(uint16_t port_id_start,
2236 const struct rte_device *parent);
2239 * Macro to iterate over all ethdev ports of a specified device.
2242 * The ID of the matching port being iterated.
2244 * The rte_device pointer matching the iterated ports.
2246 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \
2247 for (port_id = rte_eth_find_next_of(0, parent); \
2248 port_id < RTE_MAX_ETHPORTS; \
2249 port_id = rte_eth_find_next_of(port_id + 1, parent))
2252 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device).
2254 * @param port_id_start
2255 * The ID of the next possible valid sibling port.
2256 * @param ref_port_id
2257 * The ID of a reference port to compare rte_device with.
2259 * Next sibling port ID, possibly port_id_start or ref_port_id itself,
2260 * RTE_MAX_ETHPORTS if there is none.
2263 rte_eth_find_next_sibling(uint16_t port_id_start, uint16_t ref_port_id);
2266 * Macro to iterate over all ethdev ports sharing the same rte_device
2267 * as the specified port.
2268 * Note: the specified reference port is part of the loop iterations.
2271 * The ID of the matching port being iterated.
2272 * @param ref_port_id
2273 * The ID of the port being compared.
2275 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \
2276 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \
2277 port_id < RTE_MAX_ETHPORTS; \
2278 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id))
2281 * Get a new unique owner identifier.
2282 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
2283 * to avoid multiple management of device by different entities.
2286 * Owner identifier pointer.
2288 * Negative errno value on error, 0 on success.
2290 int rte_eth_dev_owner_new(uint64_t *owner_id);
2293 * Set an Ethernet device owner.
2296 * The identifier of the port to own.
2298 * The owner pointer.
2300 * Negative errno value on error, 0 on success.
2302 int rte_eth_dev_owner_set(const uint16_t port_id,
2303 const struct rte_eth_dev_owner *owner);
2306 * Unset Ethernet device owner to make the device ownerless.
2309 * The identifier of port to make ownerless.
2311 * The owner identifier.
2313 * 0 on success, negative errno value on error.
2315 int rte_eth_dev_owner_unset(const uint16_t port_id,
2316 const uint64_t owner_id);
2319 * Remove owner from all Ethernet devices owned by a specific owner.
2322 * The owner identifier.
2324 * 0 on success, negative errno value on error.
2326 int rte_eth_dev_owner_delete(const uint64_t owner_id);
2329 * Get the owner of an Ethernet device.
2332 * The port identifier.
2334 * The owner structure pointer to fill.
2336 * 0 on success, negative errno value on error..
2338 int rte_eth_dev_owner_get(const uint16_t port_id,
2339 struct rte_eth_dev_owner *owner);
2342 * Get the number of ports which are usable for the application.
2344 * These devices must be iterated by using the macro
2345 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
2346 * to deal with non-contiguous ranges of devices.
2349 * The count of available Ethernet devices.
2351 uint16_t rte_eth_dev_count_avail(void);
2354 * Get the total number of ports which are allocated.
2356 * Some devices may not be available for the application.
2359 * The total count of Ethernet devices.
2361 uint16_t rte_eth_dev_count_total(void);
2364 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
2365 * the bitmap link_speeds of the struct rte_eth_conf
2368 * Numerical speed value in Mbps
2370 * RTE_ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
2372 * 0 if the speed cannot be mapped
2374 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
2377 * Get RTE_ETH_RX_OFFLOAD_* flag name.
2382 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
2384 const char *rte_eth_dev_rx_offload_name(uint64_t offload);
2387 * Get RTE_ETH_TX_OFFLOAD_* flag name.
2392 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
2394 const char *rte_eth_dev_tx_offload_name(uint64_t offload);
2398 * @b EXPERIMENTAL: this API may change without prior notice.
2400 * Get RTE_ETH_DEV_CAPA_* flag name.
2405 * Capability name or 'UNKNOWN' if the flag cannot be recognized.
2408 const char *rte_eth_dev_capability_name(uint64_t capability);
2411 * Configure an Ethernet device.
2412 * This function must be invoked first before any other function in the
2413 * Ethernet API. This function can also be re-invoked when a device is in the
2417 * The port identifier of the Ethernet device to configure.
2418 * @param nb_rx_queue
2419 * The number of receive queues to set up for the Ethernet device.
2420 * @param nb_tx_queue
2421 * The number of transmit queues to set up for the Ethernet device.
2423 * The pointer to the configuration data to be used for the Ethernet device.
2424 * The *rte_eth_conf* structure includes:
2425 * - the hardware offload features to activate, with dedicated fields for
2426 * each statically configurable offload hardware feature provided by
2427 * Ethernet devices, such as IP checksum or VLAN tag stripping for
2429 * The Rx offload bitfield API is obsolete and will be deprecated.
2430 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
2431 * structure and use offloads field to set per-port offloads instead.
2432 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
2433 * the [rt]x_offload_capa returned from rte_eth_dev_info_get().
2434 * Any type of device supported offloading set in the input argument
2435 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
2436 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup()
2437 * - the Receive Side Scaling (RSS) configuration when using multiple Rx
2438 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf
2439 * must be within the flow_type_rss_offloads provided by drivers via
2440 * rte_eth_dev_info_get() API.
2442 * Embedding all configuration information in a single data structure
2443 * is the more flexible method that allows the addition of new features
2444 * without changing the syntax of the API.
2446 * - 0: Success, device configured.
2447 * - <0: Error code returned by the driver configuration function.
2449 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
2450 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
2453 * Check if an Ethernet device was physically removed.
2456 * The port identifier of the Ethernet device.
2458 * 1 when the Ethernet device is removed, otherwise 0.
2461 rte_eth_dev_is_removed(uint16_t port_id);
2464 * Allocate and set up a receive queue for an Ethernet device.
2466 * The function allocates a contiguous block of memory for *nb_rx_desc*
2467 * receive descriptors from a memory zone associated with *socket_id*
2468 * and initializes each receive descriptor with a network buffer allocated
2469 * from the memory pool *mb_pool*.
2472 * The port identifier of the Ethernet device.
2473 * @param rx_queue_id
2474 * The index of the receive queue to set up.
2475 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2476 * to rte_eth_dev_configure().
2478 * The number of receive descriptors to allocate for the receive ring.
2480 * The *socket_id* argument is the socket identifier in case of NUMA.
2481 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2482 * the DMA memory allocated for the receive descriptors of the ring.
2484 * The pointer to the configuration data to be used for the receive queue.
2485 * NULL value is allowed, in which case default Rx configuration
2487 * The *rx_conf* structure contains an *rx_thresh* structure with the values
2488 * of the Prefetch, Host, and Write-Back threshold registers of the receive
2490 * In addition it contains the hardware offloads features to activate using
2491 * the RTE_ETH_RX_OFFLOAD_* flags.
2492 * If an offloading set in rx_conf->offloads
2493 * hasn't been set in the input argument eth_conf->rxmode.offloads
2494 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2495 * per-queue type and it is enabled for the queue.
2496 * No need to repeat any bit in rx_conf->offloads which has already been
2497 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2498 * at port level can't be disabled at queue level.
2499 * The configuration structure also contains the pointer to the array
2500 * of the receiving buffer segment descriptions, see rx_seg and rx_nseg
2501 * fields, this extended configuration might be used by split offloads like
2502 * RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT. If mb_pool is not NULL,
2503 * the extended configuration fields must be set to NULL and zero.
2505 * The pointer to the memory pool from which to allocate *rte_mbuf* network
2506 * memory buffers to populate each descriptor of the receive ring. There are
2507 * two options to provide Rx buffer configuration:
2509 * mb_pool is not NULL, rx_conf.rx_nseg is 0.
2510 * - multiple segments description:
2511 * mb_pool is NULL, rx_conf.rx_seg is not NULL, rx_conf.rx_nseg is not 0.
2512 * Taken only if flag RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT is set in offloads.
2515 * - 0: Success, receive queue correctly set up.
2516 * - -EIO: if device is removed.
2517 * - -ENODEV: if *port_id* is invalid.
2518 * - -EINVAL: The memory pool pointer is null or the size of network buffers
2519 * which can be allocated from this memory pool does not fit the various
2520 * buffer sizes allowed by the device controller.
2521 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
2522 * allocate network memory buffers from the memory pool when
2523 * initializing receive descriptors.
2525 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
2526 uint16_t nb_rx_desc, unsigned int socket_id,
2527 const struct rte_eth_rxconf *rx_conf,
2528 struct rte_mempool *mb_pool);
2532 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2534 * Allocate and set up a hairpin receive queue for an Ethernet device.
2536 * The function set up the selected queue to be used in hairpin.
2539 * The port identifier of the Ethernet device.
2540 * @param rx_queue_id
2541 * The index of the receive queue to set up.
2542 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2543 * to rte_eth_dev_configure().
2545 * The number of receive descriptors to allocate for the receive ring.
2546 * 0 means the PMD will use default value.
2548 * The pointer to the hairpin configuration.
2551 * - (0) if successful.
2552 * - (-ENODEV) if *port_id* is invalid.
2553 * - (-ENOTSUP) if hardware doesn't support.
2554 * - (-EINVAL) if bad parameter.
2555 * - (-ENOMEM) if unable to allocate the resources.
2558 int rte_eth_rx_hairpin_queue_setup
2559 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc,
2560 const struct rte_eth_hairpin_conf *conf);
2563 * Allocate and set up a transmit queue for an Ethernet device.
2566 * The port identifier of the Ethernet device.
2567 * @param tx_queue_id
2568 * The index of the transmit queue to set up.
2569 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2570 * to rte_eth_dev_configure().
2572 * The number of transmit descriptors to allocate for the transmit ring.
2574 * The *socket_id* argument is the socket identifier in case of NUMA.
2575 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2576 * the DMA memory allocated for the transmit descriptors of the ring.
2578 * The pointer to the configuration data to be used for the transmit queue.
2579 * NULL value is allowed, in which case default Tx configuration
2581 * The *tx_conf* structure contains the following data:
2582 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
2583 * Write-Back threshold registers of the transmit ring.
2584 * When setting Write-Back threshold to the value greater then zero,
2585 * *tx_rs_thresh* value should be explicitly set to one.
2586 * - The *tx_free_thresh* value indicates the [minimum] number of network
2587 * buffers that must be pending in the transmit ring to trigger their
2588 * [implicit] freeing by the driver transmit function.
2589 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
2590 * descriptors that must be pending in the transmit ring before setting the
2591 * RS bit on a descriptor by the driver transmit function.
2592 * The *tx_rs_thresh* value should be less or equal then
2593 * *tx_free_thresh* value, and both of them should be less then
2595 * - The *offloads* member contains Tx offloads to be enabled.
2596 * If an offloading set in tx_conf->offloads
2597 * hasn't been set in the input argument eth_conf->txmode.offloads
2598 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2599 * per-queue type and it is enabled for the queue.
2600 * No need to repeat any bit in tx_conf->offloads which has already been
2601 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2602 * at port level can't be disabled at queue level.
2604 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
2605 * the transmit function to use default values.
2607 * - 0: Success, the transmit queue is correctly set up.
2608 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
2610 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
2611 uint16_t nb_tx_desc, unsigned int socket_id,
2612 const struct rte_eth_txconf *tx_conf);
2616 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2618 * Allocate and set up a transmit hairpin queue for an Ethernet device.
2621 * The port identifier of the Ethernet device.
2622 * @param tx_queue_id
2623 * The index of the transmit queue to set up.
2624 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2625 * to rte_eth_dev_configure().
2627 * The number of transmit descriptors to allocate for the transmit ring.
2628 * 0 to set default PMD value.
2630 * The hairpin configuration.
2633 * - (0) if successful.
2634 * - (-ENODEV) if *port_id* is invalid.
2635 * - (-ENOTSUP) if hardware doesn't support.
2636 * - (-EINVAL) if bad parameter.
2637 * - (-ENOMEM) if unable to allocate the resources.
2640 int rte_eth_tx_hairpin_queue_setup
2641 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc,
2642 const struct rte_eth_hairpin_conf *conf);
2646 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2648 * Get all the hairpin peer Rx / Tx ports of the current port.
2649 * The caller should ensure that the array is large enough to save the ports
2653 * The port identifier of the Ethernet device.
2655 * Pointer to the array to store the peer ports list.
2657 * Length of the array to store the port identifiers.
2659 * Current port to peer port direction
2660 * positive - current used as Tx to get all peer Rx ports.
2661 * zero - current used as Rx to get all peer Tx ports.
2664 * - (0 or positive) actual peer ports number.
2665 * - (-EINVAL) if bad parameter.
2666 * - (-ENODEV) if *port_id* invalid
2667 * - (-ENOTSUP) if hardware doesn't support.
2668 * - Others detailed errors from PMDs.
2671 int rte_eth_hairpin_get_peer_ports(uint16_t port_id, uint16_t *peer_ports,
2672 size_t len, uint32_t direction);
2676 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2678 * Bind all hairpin Tx queues of one port to the Rx queues of the peer port.
2679 * It is only allowed to call this function after all hairpin queues are
2680 * configured properly and the devices are in started state.
2683 * The identifier of the Tx port.
2685 * The identifier of peer Rx port.
2686 * RTE_MAX_ETHPORTS is allowed for the traversal of all devices.
2687 * Rx port ID could have the same value as Tx port ID.
2690 * - (0) if successful.
2691 * - (-ENODEV) if Tx port ID is invalid.
2692 * - (-EBUSY) if device is not in started state.
2693 * - (-ENOTSUP) if hardware doesn't support.
2694 * - Others detailed errors from PMDs.
2697 int rte_eth_hairpin_bind(uint16_t tx_port, uint16_t rx_port);
2701 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2703 * Unbind all hairpin Tx queues of one port from the Rx queues of the peer port.
2704 * This should be called before closing the Tx or Rx devices, if the bind
2705 * function is called before.
2706 * After unbinding the hairpin ports pair, it is allowed to bind them again.
2707 * Changing queues configuration should be after stopping the device(s).
2710 * The identifier of the Tx port.
2712 * The identifier of peer Rx port.
2713 * RTE_MAX_ETHPORTS is allowed for traversal of all devices.
2714 * Rx port ID could have the same value as Tx port ID.
2717 * - (0) if successful.
2718 * - (-ENODEV) if Tx port ID is invalid.
2719 * - (-EBUSY) if device is in stopped state.
2720 * - (-ENOTSUP) if hardware doesn't support.
2721 * - Others detailed errors from PMDs.
2724 int rte_eth_hairpin_unbind(uint16_t tx_port, uint16_t rx_port);
2727 * Return the NUMA socket to which an Ethernet device is connected
2730 * The port identifier of the Ethernet device
2732 * The NUMA socket ID to which the Ethernet device is connected or
2733 * a default of zero if the socket could not be determined.
2734 * -1 is returned is the port_id value is out of range.
2736 int rte_eth_dev_socket_id(uint16_t port_id);
2739 * Check if port_id of device is attached
2742 * The port identifier of the Ethernet device
2744 * - 0 if port is out of range or not attached
2745 * - 1 if device is attached
2747 int rte_eth_dev_is_valid_port(uint16_t port_id);
2750 * Start specified Rx queue of a port. It is used when rx_deferred_start
2751 * flag of the specified queue is true.
2754 * The port identifier of the Ethernet device
2755 * @param rx_queue_id
2756 * The index of the Rx queue to update the ring.
2757 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2758 * to rte_eth_dev_configure().
2760 * - 0: Success, the receive queue is started.
2761 * - -ENODEV: if *port_id* is invalid.
2762 * - -EINVAL: The queue_id out of range or belong to hairpin.
2763 * - -EIO: if device is removed.
2764 * - -ENOTSUP: The function not supported in PMD.
2766 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2769 * Stop specified Rx queue of a port
2772 * The port identifier of the Ethernet device
2773 * @param rx_queue_id
2774 * The index of the Rx queue to update the ring.
2775 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2776 * to rte_eth_dev_configure().
2778 * - 0: Success, the receive queue is stopped.
2779 * - -ENODEV: if *port_id* is invalid.
2780 * - -EINVAL: The queue_id out of range or belong to hairpin.
2781 * - -EIO: if device is removed.
2782 * - -ENOTSUP: The function not supported in PMD.
2784 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2787 * Start Tx for specified queue of a port. It is used when tx_deferred_start
2788 * flag of the specified queue is true.
2791 * The port identifier of the Ethernet device
2792 * @param tx_queue_id
2793 * The index of the Tx queue to update the ring.
2794 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2795 * to rte_eth_dev_configure().
2797 * - 0: Success, the transmit queue is started.
2798 * - -ENODEV: if *port_id* is invalid.
2799 * - -EINVAL: The queue_id out of range or belong to hairpin.
2800 * - -EIO: if device is removed.
2801 * - -ENOTSUP: The function not supported in PMD.
2803 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2806 * Stop specified Tx queue of a port
2809 * The port identifier of the Ethernet device
2810 * @param tx_queue_id
2811 * The index of the Tx queue to update the ring.
2812 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2813 * to rte_eth_dev_configure().
2815 * - 0: Success, the transmit queue is stopped.
2816 * - -ENODEV: if *port_id* is invalid.
2817 * - -EINVAL: The queue_id out of range or belong to hairpin.
2818 * - -EIO: if device is removed.
2819 * - -ENOTSUP: The function not supported in PMD.
2821 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2824 * Start an Ethernet device.
2826 * The device start step is the last one and consists of setting the configured
2827 * offload features and in starting the transmit and the receive units of the
2830 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before
2831 * PMD port start callback function is invoked.
2833 * On success, all basic functions exported by the Ethernet API (link status,
2834 * receive/transmit, and so on) can be invoked.
2837 * The port identifier of the Ethernet device.
2839 * - 0: Success, Ethernet device started.
2840 * - <0: Error code of the driver device start function.
2842 int rte_eth_dev_start(uint16_t port_id);
2845 * Stop an Ethernet device. The device can be restarted with a call to
2846 * rte_eth_dev_start()
2849 * The port identifier of the Ethernet device.
2851 * - 0: Success, Ethernet device stopped.
2852 * - <0: Error code of the driver device stop function.
2854 int rte_eth_dev_stop(uint16_t port_id);
2857 * Link up an Ethernet device.
2859 * Set device link up will re-enable the device Rx/Tx
2860 * functionality after it is previously set device linked down.
2863 * The port identifier of the Ethernet device.
2865 * - 0: Success, Ethernet device linked up.
2866 * - <0: Error code of the driver device link up function.
2868 int rte_eth_dev_set_link_up(uint16_t port_id);
2871 * Link down an Ethernet device.
2872 * The device Rx/Tx functionality will be disabled if success,
2873 * and it can be re-enabled with a call to
2874 * rte_eth_dev_set_link_up()
2877 * The port identifier of the Ethernet device.
2879 int rte_eth_dev_set_link_down(uint16_t port_id);
2882 * Close a stopped Ethernet device. The device cannot be restarted!
2883 * The function frees all port resources.
2886 * The port identifier of the Ethernet device.
2888 * - Zero if the port is closed successfully.
2889 * - Negative if something went wrong.
2891 int rte_eth_dev_close(uint16_t port_id);
2894 * Reset a Ethernet device and keep its port ID.
2896 * When a port has to be reset passively, the DPDK application can invoke
2897 * this function. For example when a PF is reset, all its VFs should also
2898 * be reset. Normally a DPDK application can invoke this function when
2899 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2900 * a port reset in other circumstances.
2902 * When this function is called, it first stops the port and then calls the
2903 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2904 * state, in which no Tx and Rx queues are setup, as if the port has been
2905 * reset and not started. The port keeps the port ID it had before the
2908 * After calling rte_eth_dev_reset( ), the application should use
2909 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2910 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2911 * to reconfigure the device as appropriate.
2913 * Note: To avoid unexpected behavior, the application should stop calling
2914 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2915 * safety, all these controlling functions should be called from the same
2919 * The port identifier of the Ethernet device.
2922 * - (0) if successful.
2923 * - (-ENODEV) if *port_id* is invalid.
2924 * - (-ENOTSUP) if hardware doesn't support this function.
2925 * - (-EPERM) if not ran from the primary process.
2926 * - (-EIO) if re-initialisation failed or device is removed.
2927 * - (-ENOMEM) if the reset failed due to OOM.
2928 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2930 int rte_eth_dev_reset(uint16_t port_id);
2933 * Enable receipt in promiscuous mode for an Ethernet device.
2936 * The port identifier of the Ethernet device.
2938 * - (0) if successful.
2939 * - (-ENOTSUP) if support for promiscuous_enable() does not exist
2941 * - (-ENODEV) if *port_id* invalid.
2943 int rte_eth_promiscuous_enable(uint16_t port_id);
2946 * Disable receipt in promiscuous mode for an Ethernet device.
2949 * The port identifier of the Ethernet device.
2951 * - (0) if successful.
2952 * - (-ENOTSUP) if support for promiscuous_disable() does not exist
2954 * - (-ENODEV) if *port_id* invalid.
2956 int rte_eth_promiscuous_disable(uint16_t port_id);
2959 * Return the value of promiscuous mode for an Ethernet device.
2962 * The port identifier of the Ethernet device.
2964 * - (1) if promiscuous is enabled
2965 * - (0) if promiscuous is disabled.
2968 int rte_eth_promiscuous_get(uint16_t port_id);
2971 * Enable the receipt of any multicast frame by an Ethernet device.
2974 * The port identifier of the Ethernet device.
2976 * - (0) if successful.
2977 * - (-ENOTSUP) if support for allmulticast_enable() does not exist
2979 * - (-ENODEV) if *port_id* invalid.
2981 int rte_eth_allmulticast_enable(uint16_t port_id);
2984 * Disable the receipt of all multicast frames by an Ethernet device.
2987 * The port identifier of the Ethernet device.
2989 * - (0) if successful.
2990 * - (-ENOTSUP) if support for allmulticast_disable() does not exist
2992 * - (-ENODEV) if *port_id* invalid.
2994 int rte_eth_allmulticast_disable(uint16_t port_id);
2997 * Return the value of allmulticast mode for an Ethernet device.
3000 * The port identifier of the Ethernet device.
3002 * - (1) if allmulticast is enabled
3003 * - (0) if allmulticast is disabled.
3006 int rte_eth_allmulticast_get(uint16_t port_id);
3009 * Retrieve the link status (up/down), the duplex mode (half/full),
3010 * the negotiation (auto/fixed), and if available, the speed (Mbps).
3012 * It might need to wait up to 9 seconds.
3013 * @see rte_eth_link_get_nowait.
3016 * The port identifier of the Ethernet device.
3018 * Link information written back.
3020 * - (0) if successful.
3021 * - (-ENOTSUP) if the function is not supported in PMD.
3022 * - (-ENODEV) if *port_id* invalid.
3023 * - (-EINVAL) if bad parameter.
3025 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
3028 * Retrieve the link status (up/down), the duplex mode (half/full),
3029 * the negotiation (auto/fixed), and if available, the speed (Mbps).
3032 * The port identifier of the Ethernet device.
3034 * Link information written back.
3036 * - (0) if successful.
3037 * - (-ENOTSUP) if the function is not supported in PMD.
3038 * - (-ENODEV) if *port_id* invalid.
3039 * - (-EINVAL) if bad parameter.
3041 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
3045 * @b EXPERIMENTAL: this API may change without prior notice.
3047 * The function converts a link_speed to a string. It handles all special
3048 * values like unknown or none speed.
3051 * link_speed of rte_eth_link struct
3053 * Link speed in textual format. It's pointer to immutable memory.
3054 * No free is required.
3057 const char *rte_eth_link_speed_to_str(uint32_t link_speed);
3061 * @b EXPERIMENTAL: this API may change without prior notice.
3063 * The function converts a rte_eth_link struct representing a link status to
3067 * A pointer to a string to be filled with textual representation of
3068 * device status. At least RTE_ETH_LINK_MAX_STR_LEN bytes should be allocated to
3069 * store default link status text.
3071 * Length of available memory at 'str' string.
3073 * Link status returned by rte_eth_link_get function
3075 * Number of bytes written to str array or -EINVAL if bad parameter.
3078 int rte_eth_link_to_str(char *str, size_t len,
3079 const struct rte_eth_link *eth_link);
3082 * Retrieve the general I/O statistics of an Ethernet device.
3085 * The port identifier of the Ethernet device.
3087 * A pointer to a structure of type *rte_eth_stats* to be filled with
3088 * the values of device counters for the following set of statistics:
3089 * - *ipackets* with the total of successfully received packets.
3090 * - *opackets* with the total of successfully transmitted packets.
3091 * - *ibytes* with the total of successfully received bytes.
3092 * - *obytes* with the total of successfully transmitted bytes.
3093 * - *ierrors* with the total of erroneous received packets.
3094 * - *oerrors* with the total of failed transmitted packets.
3096 * Zero if successful. Non-zero otherwise.
3098 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
3101 * Reset the general I/O statistics of an Ethernet device.
3104 * The port identifier of the Ethernet device.
3106 * - (0) if device notified to reset stats.
3107 * - (-ENOTSUP) if hardware doesn't support.
3108 * - (-ENODEV) if *port_id* invalid.
3109 * - (<0): Error code of the driver stats reset function.
3111 int rte_eth_stats_reset(uint16_t port_id);
3114 * Retrieve names of extended statistics of an Ethernet device.
3116 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
3118 * xstats_names[i].name => xstats[i].value
3120 * And the array index is same with id field of 'struct rte_eth_xstat':
3123 * This assumption makes key-value pair matching less flexible but simpler.
3126 * The port identifier of the Ethernet device.
3127 * @param xstats_names
3128 * An rte_eth_xstat_name array of at least *size* elements to
3129 * be filled. If set to NULL, the function returns the required number
3132 * The size of the xstats_names array (number of elements).
3134 * - A positive value lower or equal to size: success. The return value
3135 * is the number of entries filled in the stats table.
3136 * - A positive value higher than size: error, the given statistics table
3137 * is too small. The return value corresponds to the size that should
3138 * be given to succeed. The entries in the table are not valid and
3139 * shall not be used by the caller.
3140 * - A negative value on error (invalid port ID).
3142 int rte_eth_xstats_get_names(uint16_t port_id,
3143 struct rte_eth_xstat_name *xstats_names,
3147 * Retrieve extended statistics of an Ethernet device.
3149 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
3151 * xstats_names[i].name => xstats[i].value
3153 * And the array index is same with id field of 'struct rte_eth_xstat':
3156 * This assumption makes key-value pair matching less flexible but simpler.
3159 * The port identifier of the Ethernet device.
3161 * A pointer to a table of structure of type *rte_eth_xstat*
3162 * to be filled with device statistics ids and values.
3163 * This parameter can be set to NULL if and only if n is 0.
3165 * The size of the xstats array (number of elements).
3166 * If lower than the required number of elements, the function returns
3167 * the required number of elements.
3168 * If equal to zero, the xstats must be NULL, the function returns the
3169 * required number of elements.
3171 * - A positive value lower or equal to n: success. The return value
3172 * is the number of entries filled in the stats table.
3173 * - A positive value higher than n: error, the given statistics table
3174 * is too small. The return value corresponds to the size that should
3175 * be given to succeed. The entries in the table are not valid and
3176 * shall not be used by the caller.
3177 * - A negative value on error (invalid port ID).
3179 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
3183 * Retrieve names of extended statistics of an Ethernet device.
3186 * The port identifier of the Ethernet device.
3187 * @param xstats_names
3188 * Array to be filled in with names of requested device statistics.
3189 * Must not be NULL if @p ids are specified (not NULL).
3191 * Number of elements in @p xstats_names array (if not NULL) and in
3192 * @p ids array (if not NULL). Must be 0 if both array pointers are NULL.
3194 * IDs array given by app to retrieve specific statistics. May be NULL to
3195 * retrieve names of all available statistics or, if @p xstats_names is
3196 * NULL as well, just the number of available statistics.
3198 * - A positive value lower or equal to size: success. The return value
3199 * is the number of entries filled in the stats table.
3200 * - A positive value higher than size: success. The given statistics table
3201 * is too small. The return value corresponds to the size that should
3202 * be given to succeed. The entries in the table are not valid and
3203 * shall not be used by the caller.
3204 * - A negative value on error.
3207 rte_eth_xstats_get_names_by_id(uint16_t port_id,
3208 struct rte_eth_xstat_name *xstats_names, unsigned int size,
3212 * Retrieve extended statistics of an Ethernet device.
3215 * The port identifier of the Ethernet device.
3217 * IDs array given by app to retrieve specific statistics. May be NULL to
3218 * retrieve all available statistics or, if @p values is NULL as well,
3219 * just the number of available statistics.
3221 * Array to be filled in with requested device statistics.
3222 * Must not be NULL if ids are specified (not NULL).
3224 * Number of elements in @p values array (if not NULL) and in @p ids
3225 * array (if not NULL). Must be 0 if both array pointers are NULL.
3227 * - A positive value lower or equal to size: success. The return value
3228 * is the number of entries filled in the stats table.
3229 * - A positive value higher than size: success: The given statistics table
3230 * is too small. The return value corresponds to the size that should
3231 * be given to succeed. The entries in the table are not valid and
3232 * shall not be used by the caller.
3233 * - A negative value on error.
3235 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
3236 uint64_t *values, unsigned int size);
3239 * Gets the ID of a statistic from its name.
3241 * This function searches for the statistics using string compares, and
3242 * as such should not be used on the fast-path. For fast-path retrieval of
3243 * specific statistics, store the ID as provided in *id* from this function,
3244 * and pass the ID to rte_eth_xstats_get()
3246 * @param port_id The port to look up statistics from
3247 * @param xstat_name The name of the statistic to return
3248 * @param[out] id A pointer to an app-supplied uint64_t which should be
3249 * set to the ID of the stat if the stat exists.
3252 * -ENODEV for invalid port_id,
3253 * -EIO if device is removed,
3254 * -EINVAL if the xstat_name doesn't exist in port_id
3255 * -ENOMEM if bad parameter.
3257 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
3261 * Reset extended statistics of an Ethernet device.
3264 * The port identifier of the Ethernet device.
3266 * - (0) if device notified to reset extended stats.
3267 * - (-ENOTSUP) if pmd doesn't support both
3268 * extended stats and basic stats reset.
3269 * - (-ENODEV) if *port_id* invalid.
3270 * - (<0): Error code of the driver xstats reset function.
3272 int rte_eth_xstats_reset(uint16_t port_id);
3275 * Set a mapping for the specified transmit queue to the specified per-queue
3276 * statistics counter.
3279 * The port identifier of the Ethernet device.
3280 * @param tx_queue_id
3281 * The index of the transmit queue for which a queue stats mapping is required.
3282 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3283 * to rte_eth_dev_configure().
3285 * The per-queue packet statistics functionality number that the transmit
3286 * queue is to be assigned.
3287 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
3288 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256.
3290 * Zero if successful. Non-zero otherwise.
3292 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
3293 uint16_t tx_queue_id, uint8_t stat_idx);
3296 * Set a mapping for the specified receive queue to the specified per-queue
3297 * statistics counter.
3300 * The port identifier of the Ethernet device.
3301 * @param rx_queue_id
3302 * The index of the receive queue for which a queue stats mapping is required.
3303 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3304 * to rte_eth_dev_configure().
3306 * The per-queue packet statistics functionality number that the receive
3307 * queue is to be assigned.
3308 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
3309 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256.
3311 * Zero if successful. Non-zero otherwise.
3313 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
3314 uint16_t rx_queue_id,
3318 * Retrieve the Ethernet address of an Ethernet device.
3321 * The port identifier of the Ethernet device.
3323 * A pointer to a structure of type *ether_addr* to be filled with
3324 * the Ethernet address of the Ethernet device.
3326 * - (0) if successful
3327 * - (-ENODEV) if *port_id* invalid.
3328 * - (-EINVAL) if bad parameter.
3330 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr);
3334 * @b EXPERIMENTAL: this API may change without prior notice
3336 * Retrieve the Ethernet addresses of an Ethernet device.
3339 * The port identifier of the Ethernet device.
3341 * A pointer to an array of structures of type *ether_addr* to be filled with
3342 * the Ethernet addresses of the Ethernet device.
3344 * Number of elements in the @p ma array.
3345 * Note that rte_eth_dev_info::max_mac_addrs can be used to retrieve
3346 * max number of Ethernet addresses for given port.
3348 * - number of retrieved addresses if successful
3349 * - (-ENODEV) if *port_id* invalid.
3350 * - (-EINVAL) if bad parameter.
3353 int rte_eth_macaddrs_get(uint16_t port_id, struct rte_ether_addr *ma,
3357 * Retrieve the contextual information of an Ethernet device.
3359 * As part of this function, a number of of fields in dev_info will be
3360 * initialized as follows:
3365 * Where lim is defined within the rte_eth_dev_info_get as
3367 * const struct rte_eth_desc_lim lim = {
3368 * .nb_max = UINT16_MAX,
3371 * .nb_seg_max = UINT16_MAX,
3372 * .nb_mtu_seg_max = UINT16_MAX,
3375 * device = dev->device
3376 * min_mtu = RTE_ETHER_MIN_LEN - RTE_ETHER_HDR_LEN - RTE_ETHER_CRC_LEN
3377 * max_mtu = UINT16_MAX
3379 * The following fields will be populated if support for dev_infos_get()
3380 * exists for the device and the rte_eth_dev 'dev' has been populated
3381 * successfully with a call to it:
3383 * driver_name = dev->device->driver->name
3384 * nb_rx_queues = dev->data->nb_rx_queues
3385 * nb_tx_queues = dev->data->nb_tx_queues
3386 * dev_flags = &dev->data->dev_flags
3389 * The port identifier of the Ethernet device.
3391 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
3392 * the contextual information of the Ethernet device.
3394 * - (0) if successful.
3395 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device.
3396 * - (-ENODEV) if *port_id* invalid.
3397 * - (-EINVAL) if bad parameter.
3399 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
3403 * @b EXPERIMENTAL: this API may change without prior notice.
3405 * Retrieve the configuration of an Ethernet device.
3408 * The port identifier of the Ethernet device.
3410 * Location for Ethernet device configuration to be filled in.
3412 * - (0) if successful.
3413 * - (-ENODEV) if *port_id* invalid.
3414 * - (-EINVAL) if bad parameter.
3417 int rte_eth_dev_conf_get(uint16_t port_id, struct rte_eth_conf *dev_conf);
3420 * Retrieve the firmware version of a device.
3423 * The port identifier of the device.
3425 * A pointer to a string array storing the firmware version of a device,
3426 * the string includes terminating null. This pointer is allocated by caller.
3428 * The size of the string array pointed by fw_version, which should be
3429 * large enough to store firmware version of the device.
3431 * - (0) if successful.
3432 * - (-ENOTSUP) if operation is not supported.
3433 * - (-ENODEV) if *port_id* invalid.
3434 * - (-EIO) if device is removed.
3435 * - (-EINVAL) if bad parameter.
3436 * - (>0) if *fw_size* is not enough to store firmware version, return
3437 * the size of the non truncated string.
3439 int rte_eth_dev_fw_version_get(uint16_t port_id,
3440 char *fw_version, size_t fw_size);
3443 * Retrieve the supported packet types of an Ethernet device.
3445 * When a packet type is announced as supported, it *must* be recognized by
3446 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
3447 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
3448 * packet types for these packets:
3449 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
3450 * - Ether/VLAN/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
3451 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
3452 * - Ether/VLAN/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
3454 * When a packet is received by a PMD, the most precise type must be
3455 * returned among the ones supported. However a PMD is allowed to set
3456 * packet type that is not in the supported list, at the condition that it
3457 * is more precise. Therefore, a PMD announcing no supported packet types
3458 * can still set a matching packet type in a received packet.
3461 * Better to invoke this API after the device is already started or Rx burst
3462 * function is decided, to obtain correct supported ptypes.
3464 * if a given PMD does not report what ptypes it supports, then the supported
3465 * ptype count is reported as 0.
3467 * The port identifier of the Ethernet device.
3469 * A hint of what kind of packet type which the caller is interested in.
3471 * An array pointer to store adequate packet types, allocated by caller.
3473 * Size of the array pointed by param ptypes.
3475 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
3476 * only num entries will be filled into the ptypes array, but the full
3477 * count of supported ptypes will be returned.
3478 * - (-ENODEV) if *port_id* invalid.
3479 * - (-EINVAL) if bad parameter.
3481 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
3482 uint32_t *ptypes, int num);
3484 * Inform Ethernet device about reduced range of packet types to handle.
3486 * Application can use this function to set only specific ptypes that it's
3487 * interested. This information can be used by the PMD to optimize Rx path.
3489 * The function accepts an array `set_ptypes` allocated by the caller to
3490 * store the packet types set by the driver, the last element of the array
3491 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be
3492 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled
3496 * The port identifier of the Ethernet device.
3498 * The ptype family that application is interested in should be bitwise OR of
3499 * RTE_PTYPE_*_MASK or 0.
3501 * An array pointer to store set packet types, allocated by caller. The
3502 * function marks the end of array with RTE_PTYPE_UNKNOWN.
3504 * Size of the array pointed by param ptypes.
3505 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the
3509 * - (-ENODEV) if *port_id* invalid.
3510 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and
3513 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask,
3514 uint32_t *set_ptypes, unsigned int num);
3517 * Retrieve the MTU of an Ethernet device.
3520 * The port identifier of the Ethernet device.
3522 * A pointer to a uint16_t where the retrieved MTU is to be stored.
3524 * - (0) if successful.
3525 * - (-ENODEV) if *port_id* invalid.
3526 * - (-EINVAL) if bad parameter.
3528 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
3531 * Change the MTU of an Ethernet device.
3534 * The port identifier of the Ethernet device.
3536 * A uint16_t for the MTU to be applied.
3538 * - (0) if successful.
3539 * - (-ENOTSUP) if operation is not supported.
3540 * - (-ENODEV) if *port_id* invalid.
3541 * - (-EIO) if device is removed.
3542 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within
3543 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or
3544 * when the mtu is set using dev->dev_ops->mtu_set.
3545 * - (-EBUSY) if operation is not allowed when the port is running
3547 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
3550 * Enable/Disable hardware filtering by an Ethernet device of received
3551 * VLAN packets tagged with a given VLAN Tag Identifier.
3554 * The port identifier of the Ethernet device.
3556 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
3558 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
3559 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
3561 * - (0) if successful.
3562 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
3563 * - (-ENODEV) if *port_id* invalid.
3564 * - (-EIO) if device is removed.
3565 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
3566 * - (-EINVAL) if *vlan_id* > 4095.
3568 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
3571 * Enable/Disable hardware VLAN Strip by a Rx queue of an Ethernet device.
3574 * The port identifier of the Ethernet device.
3575 * @param rx_queue_id
3576 * The index of the receive queue for which a queue stats mapping is required.
3577 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3578 * to rte_eth_dev_configure().
3580 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
3581 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
3583 * - (0) if successful.
3584 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured.
3585 * - (-ENODEV) if *port_id* invalid.
3586 * - (-EINVAL) if *rx_queue_id* invalid.
3588 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
3592 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
3596 * The port identifier of the Ethernet device.
3600 * The Tag Protocol ID
3602 * - (0) if successful.
3603 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported.
3604 * - (-ENODEV) if *port_id* invalid.
3605 * - (-EIO) if device is removed.
3607 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
3608 enum rte_vlan_type vlan_type,
3612 * Set VLAN offload configuration on an Ethernet device.
3615 * The port identifier of the Ethernet device.
3616 * @param offload_mask
3617 * The VLAN Offload bit mask can be mixed use with "OR"
3618 * RTE_ETH_VLAN_STRIP_OFFLOAD
3619 * RTE_ETH_VLAN_FILTER_OFFLOAD
3620 * RTE_ETH_VLAN_EXTEND_OFFLOAD
3621 * RTE_ETH_QINQ_STRIP_OFFLOAD
3623 * - (0) if successful.
3624 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
3625 * - (-ENODEV) if *port_id* invalid.
3626 * - (-EIO) if device is removed.
3628 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
3631 * Read VLAN Offload configuration from an Ethernet device
3634 * The port identifier of the Ethernet device.
3636 * - (>0) if successful. Bit mask to indicate
3637 * RTE_ETH_VLAN_STRIP_OFFLOAD
3638 * RTE_ETH_VLAN_FILTER_OFFLOAD
3639 * RTE_ETH_VLAN_EXTEND_OFFLOAD
3640 * RTE_ETH_QINQ_STRIP_OFFLOAD
3641 * - (-ENODEV) if *port_id* invalid.
3643 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
3646 * Set port based Tx VLAN insertion on or off.
3649 * The port identifier of the Ethernet device.
3651 * Port based Tx VLAN identifier together with user priority.
3653 * Turn on or off the port based Tx VLAN insertion.
3656 * - (0) if successful.
3657 * - negative if failed.
3659 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
3663 * @b EXPERIMENTAL: this API may change without prior notice.
3665 * Set Rx queue available descriptors threshold.
3668 * The port identifier of the Ethernet device.
3670 * The index of the receive queue.
3671 * @param avail_thresh
3672 * The available descriptors threshold is percentage of Rx queue size
3673 * which describes the availability of Rx queue for hardware.
3674 * If the Rx queue availability is below it,
3675 * the event RTE_ETH_EVENT_RX_AVAIL_THRESH is triggered.
3676 * [1-99] to set a new available descriptors threshold.
3677 * 0 to disable threshold monitoring.
3680 * - 0 if successful.
3681 * - (-ENODEV) if @p port_id is invalid.
3682 * - (-EINVAL) if bad parameter.
3683 * - (-ENOTSUP) if available Rx descriptors threshold is not supported.
3684 * - (-EIO) if device is removed.
3687 int rte_eth_rx_avail_thresh_set(uint16_t port_id, uint16_t queue_id,
3688 uint8_t avail_thresh);
3692 * @b EXPERIMENTAL: this API may change without prior notice.
3694 * Find Rx queue with RTE_ETH_EVENT_RX_AVAIL_THRESH event pending.
3697 * The port identifier of the Ethernet device.
3698 * @param[inout] queue_id
3699 * On input starting Rx queue index to search from.
3700 * If the queue_id is bigger than maximum queue ID of the port,
3701 * search is started from 0. So that application can keep calling
3702 * this function to handle all pending events with a simple increment
3703 * of queue_id on the next call.
3704 * On output if return value is 1, Rx queue index with the event pending.
3705 * @param[out] avail_thresh
3706 * Location for available descriptors threshold of the found Rx queue.
3709 * - 1 if an Rx queue with pending event is found.
3710 * - 0 if no Rx queue with pending event is found.
3711 * - (-ENODEV) if @p port_id is invalid.
3712 * - (-EINVAL) if bad parameter (e.g. @p queue_id is NULL).
3713 * - (-ENOTSUP) if operation is not supported.
3714 * - (-EIO) if device is removed.
3717 int rte_eth_rx_avail_thresh_query(uint16_t port_id, uint16_t *queue_id,
3718 uint8_t *avail_thresh);
3720 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
3724 * Structure used to buffer packets for future Tx
3725 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
3727 struct rte_eth_dev_tx_buffer {
3728 buffer_tx_error_fn error_callback;
3729 void *error_userdata;
3730 uint16_t size; /**< Size of buffer for buffered Tx */
3731 uint16_t length; /**< Number of packets in the array */
3732 /** Pending packets to be sent on explicit flush or when full */
3733 struct rte_mbuf *pkts[];
3737 * Calculate the size of the Tx buffer.
3740 * Number of stored packets.
3742 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
3743 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
3746 * Initialize default values for buffered transmitting
3749 * Tx buffer to be initialized.
3756 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
3759 * Configure a callback for buffered packets which cannot be sent
3761 * Register a specific callback to be called when an attempt is made to send
3762 * all packets buffered on an Ethernet port, but not all packets can
3763 * successfully be sent. The callback registered here will be called only
3764 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
3765 * The default callback configured for each queue by default just frees the
3766 * packets back to the calling mempool. If additional behaviour is required,
3767 * for example, to count dropped packets, or to retry transmission of packets
3768 * which cannot be sent, this function should be used to register a suitable
3769 * callback function to implement the desired behaviour.
3770 * The example callback "rte_eth_count_unsent_packet_callback()" is also
3771 * provided as reference.
3774 * The port identifier of the Ethernet device.
3776 * The function to be used as the callback.
3778 * Arbitrary parameter to be passed to the callback function
3780 * 0 on success, or -EINVAL if bad parameter
3783 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
3784 buffer_tx_error_fn callback, void *userdata);
3787 * Callback function for silently dropping unsent buffered packets.
3789 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3790 * adjust the default behavior when buffered packets cannot be sent. This
3791 * function drops any unsent packets silently and is used by Tx buffered
3792 * operations as default behavior.
3794 * NOTE: this function should not be called directly, instead it should be used
3795 * as a callback for packet buffering.
3797 * NOTE: when configuring this function as a callback with
3798 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3799 * should point to an uint64_t value.
3802 * The previously buffered packets which could not be sent
3804 * The number of unsent packets in the pkts array
3809 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
3813 * Callback function for tracking unsent buffered packets.
3815 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3816 * adjust the default behavior when buffered packets cannot be sent. This
3817 * function drops any unsent packets, but also updates a user-supplied counter
3818 * to track the overall number of packets dropped. The counter should be an
3819 * uint64_t variable.
3821 * NOTE: this function should not be called directly, instead it should be used
3822 * as a callback for packet buffering.
3824 * NOTE: when configuring this function as a callback with
3825 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3826 * should point to an uint64_t value.
3829 * The previously buffered packets which could not be sent
3831 * The number of unsent packets in the pkts array
3833 * Pointer to an uint64_t value, which will be incremented by unsent
3836 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
3840 * Request the driver to free mbufs currently cached by the driver. The
3841 * driver will only free the mbuf if it is no longer in use. It is the
3842 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is
3846 * The port identifier of the Ethernet device.
3848 * The index of the transmit queue through which output packets must be
3850 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3851 * to rte_eth_dev_configure().
3853 * Maximum number of packets to free. Use 0 to indicate all possible packets
3854 * should be freed. Note that a packet may be using multiple mbufs.
3857 * -ENODEV: Invalid interface
3858 * -EIO: device is removed
3859 * -ENOTSUP: Driver does not support function
3861 * 0-n: Number of packets freed. More packets may still remain in ring that
3865 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
3868 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
3871 enum rte_eth_event_ipsec_subtype {
3872 /** Unknown event type */
3873 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
3874 /** Sequence number overflow */
3875 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
3876 /** Soft time expiry of SA */
3877 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
3878 /** Soft byte expiry of SA */
3879 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
3880 /** Max value of this enum */
3881 RTE_ETH_EVENT_IPSEC_MAX
3885 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
3886 * information of the IPsec offload event.
3888 struct rte_eth_event_ipsec_desc {
3889 /** Type of RTE_ETH_EVENT_IPSEC_* event */
3890 enum rte_eth_event_ipsec_subtype subtype;
3892 * Event specific metadata.
3894 * For the following events, *userdata* registered
3895 * with the *rte_security_session* would be returned
3898 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
3899 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
3900 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
3902 * @see struct rte_security_session_conf
3909 * The eth device event type for interrupt, and maybe others in the future.
3911 enum rte_eth_event_type {
3912 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3913 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3914 /** queue state event (enabled/disabled) */
3915 RTE_ETH_EVENT_QUEUE_STATE,
3916 /** reset interrupt event, sent to VF on PF reset */
3917 RTE_ETH_EVENT_INTR_RESET,
3918 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3919 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3920 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3921 RTE_ETH_EVENT_NEW, /**< port is probed */
3922 RTE_ETH_EVENT_DESTROY, /**< port is released */
3923 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
3924 RTE_ETH_EVENT_FLOW_AGED,/**< New aged-out flows is detected */
3926 * Number of available Rx descriptors is smaller than the threshold.
3927 * @see rte_eth_rx_avail_thresh_set()
3929 RTE_ETH_EVENT_RX_AVAIL_THRESH,
3930 RTE_ETH_EVENT_MAX /**< max value of this enum */
3933 /** User application callback to be registered for interrupts. */
3934 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3935 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3938 * Register a callback function for port event.
3942 * RTE_ETH_ALL means register the event for all port ids.
3946 * User supplied callback function to be called.
3948 * Pointer to the parameters for the registered callback.
3951 * - On success, zero.
3952 * - On failure, a negative value.
3954 int rte_eth_dev_callback_register(uint16_t port_id,
3955 enum rte_eth_event_type event,
3956 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3959 * Unregister a callback function for port event.
3963 * RTE_ETH_ALL means unregister the event for all port ids.
3967 * User supplied callback function to be called.
3969 * Pointer to the parameters for the registered callback. -1 means to
3970 * remove all for the same callback address and same event.
3973 * - On success, zero.
3974 * - On failure, a negative value.
3976 int rte_eth_dev_callback_unregister(uint16_t port_id,
3977 enum rte_eth_event_type event,
3978 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3981 * When there is no Rx packet coming in Rx Queue for a long time, we can
3982 * sleep lcore related to Rx Queue for power saving, and enable Rx interrupt
3983 * to be triggered when Rx packet arrives.
3985 * The rte_eth_dev_rx_intr_enable() function enables Rx queue
3986 * interrupt on specific Rx queue of a port.
3989 * The port identifier of the Ethernet device.
3991 * The index of the receive queue from which to retrieve input packets.
3992 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3993 * to rte_eth_dev_configure().
3995 * - (0) if successful.
3996 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3998 * - (-ENODEV) if *port_id* invalid.
3999 * - (-EIO) if device is removed.
4001 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
4004 * When lcore wakes up from Rx interrupt indicating packet coming, disable Rx
4005 * interrupt and returns to polling mode.
4007 * The rte_eth_dev_rx_intr_disable() function disables Rx queue
4008 * interrupt on specific Rx queue of a port.
4011 * The port identifier of the Ethernet device.
4013 * The index of the receive queue from which to retrieve input packets.
4014 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4015 * to rte_eth_dev_configure().
4017 * - (0) if successful.
4018 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
4020 * - (-ENODEV) if *port_id* invalid.
4021 * - (-EIO) if device is removed.
4023 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
4026 * Rx Interrupt control per port.
4029 * The port identifier of the Ethernet device.
4031 * Epoll instance fd which the intr vector associated to.
4032 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
4034 * The operation be performed for the vector.
4035 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
4039 * - On success, zero.
4040 * - On failure, a negative value.
4042 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
4045 * Rx Interrupt control per queue.
4048 * The port identifier of the Ethernet device.
4050 * The index of the receive queue from which to retrieve input packets.
4051 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4052 * to rte_eth_dev_configure().
4054 * Epoll instance fd which the intr vector associated to.
4055 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
4057 * The operation be performed for the vector.
4058 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
4062 * - On success, zero.
4063 * - On failure, a negative value.
4065 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
4066 int epfd, int op, void *data);
4069 * Get interrupt fd per Rx queue.
4072 * The port identifier of the Ethernet device.
4074 * The index of the receive queue from which to retrieve input packets.
4075 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4076 * to rte_eth_dev_configure().
4078 * - (>=0) the interrupt fd associated to the requested Rx queue if
4083 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id);
4086 * Turn on the LED on the Ethernet device.
4087 * This function turns on the LED on the Ethernet device.
4090 * The port identifier of the Ethernet device.
4092 * - (0) if successful.
4093 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
4095 * - (-ENODEV) if *port_id* invalid.
4096 * - (-EIO) if device is removed.
4098 int rte_eth_led_on(uint16_t port_id);
4101 * Turn off the LED on the Ethernet device.
4102 * This function turns off the LED on the Ethernet device.
4105 * The port identifier of the Ethernet device.
4107 * - (0) if successful.
4108 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
4110 * - (-ENODEV) if *port_id* invalid.
4111 * - (-EIO) if device is removed.
4113 int rte_eth_led_off(uint16_t port_id);
4117 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4119 * Get Forward Error Correction(FEC) capability.
4122 * The port identifier of the Ethernet device.
4123 * @param speed_fec_capa
4124 * speed_fec_capa is out only with per-speed capabilities.
4125 * If set to NULL, the function returns the required number
4126 * of required array entries.
4128 * a number of elements in an speed_fec_capa array.
4131 * - A non-negative value lower or equal to num: success. The return value
4132 * is the number of entries filled in the fec capa array.
4133 * - A non-negative value higher than num: error, the given fec capa array
4134 * is too small. The return value corresponds to the num that should
4135 * be given to succeed. The entries in fec capa array are not valid and
4136 * shall not be used by the caller.
4137 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4139 * - (-EIO) if device is removed.
4140 * - (-ENODEV) if *port_id* invalid.
4141 * - (-EINVAL) if *num* or *speed_fec_capa* invalid
4144 int rte_eth_fec_get_capability(uint16_t port_id,
4145 struct rte_eth_fec_capa *speed_fec_capa,
4150 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4152 * Get current Forward Error Correction(FEC) mode.
4153 * If link is down and AUTO is enabled, AUTO is returned, otherwise,
4154 * configured FEC mode is returned.
4155 * If link is up, current FEC mode is returned.
4158 * The port identifier of the Ethernet device.
4160 * A bitmask of enabled FEC modes. If AUTO bit is set, other
4161 * bits specify FEC modes which may be negotiated. If AUTO
4162 * bit is clear, specify FEC modes to be used (only one valid
4163 * mode per speed may be set).
4165 * - (0) if successful.
4166 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4168 * - (-EIO) if device is removed.
4169 * - (-ENODEV) if *port_id* invalid.
4172 int rte_eth_fec_get(uint16_t port_id, uint32_t *fec_capa);
4176 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4178 * Set Forward Error Correction(FEC) mode.
4181 * The port identifier of the Ethernet device.
4183 * A bitmask of allowed FEC modes. If AUTO bit is set, other
4184 * bits specify FEC modes which may be negotiated. If AUTO
4185 * bit is clear, specify FEC modes to be used (only one valid
4186 * mode per speed may be set).
4188 * - (0) if successful.
4189 * - (-EINVAL) if the FEC mode is not valid.
4190 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4191 * - (-EIO) if device is removed.
4192 * - (-ENODEV) if *port_id* invalid.
4195 int rte_eth_fec_set(uint16_t port_id, uint32_t fec_capa);
4198 * Get current status of the Ethernet link flow control for Ethernet device
4201 * The port identifier of the Ethernet device.
4203 * The pointer to the structure where to store the flow control parameters.
4205 * - (0) if successful.
4206 * - (-ENOTSUP) if hardware doesn't support flow control.
4207 * - (-ENODEV) if *port_id* invalid.
4208 * - (-EIO) if device is removed.
4209 * - (-EINVAL) if bad parameter.
4211 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
4212 struct rte_eth_fc_conf *fc_conf);
4215 * Configure the Ethernet link flow control for Ethernet device
4218 * The port identifier of the Ethernet device.
4220 * The pointer to the structure of the flow control parameters.
4222 * - (0) if successful.
4223 * - (-ENOTSUP) if hardware doesn't support flow control mode.
4224 * - (-ENODEV) if *port_id* invalid.
4225 * - (-EINVAL) if bad parameter
4226 * - (-EIO) if flow control setup failure or device is removed.
4228 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
4229 struct rte_eth_fc_conf *fc_conf);
4232 * Configure the Ethernet priority flow control under DCB environment
4233 * for Ethernet device.
4236 * The port identifier of the Ethernet device.
4238 * The pointer to the structure of the priority flow control parameters.
4240 * - (0) if successful.
4241 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
4242 * - (-ENODEV) if *port_id* invalid.
4243 * - (-EINVAL) if bad parameter
4244 * - (-EIO) if flow control setup failure or device is removed.
4246 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
4247 struct rte_eth_pfc_conf *pfc_conf);
4250 * Add a MAC address to the set used for filtering incoming packets.
4253 * The port identifier of the Ethernet device.
4255 * The MAC address to add.
4257 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
4258 * not enabled, this should be set to 0.
4260 * - (0) if successfully added or *mac_addr* was already added.
4261 * - (-ENOTSUP) if hardware doesn't support this feature.
4262 * - (-ENODEV) if *port* is invalid.
4263 * - (-EIO) if device is removed.
4264 * - (-ENOSPC) if no more MAC addresses can be added.
4265 * - (-EINVAL) if MAC address is invalid.
4267 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr,
4272 * @b EXPERIMENTAL: this API may change without prior notice.
4274 * Retrieve the information for queue based PFC.
4277 * The port identifier of the Ethernet device.
4278 * @param pfc_queue_info
4279 * A pointer to a structure of type *rte_eth_pfc_queue_info* to be filled with
4280 * the information about queue based PFC.
4282 * - (0) if successful.
4283 * - (-ENOTSUP) if support for priority_flow_ctrl_queue_info_get does not exist.
4284 * - (-ENODEV) if *port_id* invalid.
4285 * - (-EINVAL) if bad parameter.
4288 int rte_eth_dev_priority_flow_ctrl_queue_info_get(uint16_t port_id,
4289 struct rte_eth_pfc_queue_info *pfc_queue_info);
4293 * @b EXPERIMENTAL: this API may change without prior notice.
4295 * Configure the queue based priority flow control for a given queue
4296 * for Ethernet device.
4298 * @note When an ethdev port switches to queue based PFC mode, the
4299 * unconfigured queues shall be configured by the driver with
4300 * default values such as lower priority value for TC etc.
4303 * The port identifier of the Ethernet device.
4304 * @param pfc_queue_conf
4305 * The pointer to the structure of the priority flow control parameters
4308 * - (0) if successful.
4309 * - (-ENOTSUP) if hardware doesn't support queue based PFC mode.
4310 * - (-ENODEV) if *port_id* invalid.
4311 * - (-EINVAL) if bad parameter
4312 * - (-EIO) if flow control setup queue failure
4315 int rte_eth_dev_priority_flow_ctrl_queue_configure(uint16_t port_id,
4316 struct rte_eth_pfc_queue_conf *pfc_queue_conf);
4319 * Remove a MAC address from the internal array of addresses.
4322 * The port identifier of the Ethernet device.
4324 * MAC address to remove.
4326 * - (0) if successful, or *mac_addr* didn't exist.
4327 * - (-ENOTSUP) if hardware doesn't support.
4328 * - (-ENODEV) if *port* invalid.
4329 * - (-EADDRINUSE) if attempting to remove the default MAC address.
4330 * - (-EINVAL) if MAC address is invalid.
4332 int rte_eth_dev_mac_addr_remove(uint16_t port_id,
4333 struct rte_ether_addr *mac_addr);
4336 * Set the default MAC address.
4339 * The port identifier of the Ethernet device.
4341 * New default MAC address.
4343 * - (0) if successful, or *mac_addr* didn't exist.
4344 * - (-ENOTSUP) if hardware doesn't support.
4345 * - (-ENODEV) if *port* invalid.
4346 * - (-EINVAL) if MAC address is invalid.
4348 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
4349 struct rte_ether_addr *mac_addr);
4352 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
4355 * The port identifier of the Ethernet device.
4359 * Redirection table size. The table size can be queried by
4360 * rte_eth_dev_info_get().
4362 * - (0) if successful.
4363 * - (-ENODEV) if *port_id* is invalid.
4364 * - (-ENOTSUP) if hardware doesn't support.
4365 * - (-EINVAL) if bad parameter.
4366 * - (-EIO) if device is removed.
4368 int rte_eth_dev_rss_reta_update(uint16_t port_id,
4369 struct rte_eth_rss_reta_entry64 *reta_conf,
4370 uint16_t reta_size);
4373 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
4376 * The port identifier of the Ethernet device.
4378 * RETA to query. For each requested reta entry, corresponding bit
4379 * in mask must be set.
4381 * Redirection table size. The table size can be queried by
4382 * rte_eth_dev_info_get().
4384 * - (0) if successful.
4385 * - (-ENODEV) if *port_id* is invalid.
4386 * - (-ENOTSUP) if hardware doesn't support.
4387 * - (-EINVAL) if bad parameter.
4388 * - (-EIO) if device is removed.
4390 int rte_eth_dev_rss_reta_query(uint16_t port_id,
4391 struct rte_eth_rss_reta_entry64 *reta_conf,
4392 uint16_t reta_size);
4395 * Updates unicast hash table for receiving packet with the given destination
4396 * MAC address, and the packet is routed to all VFs for which the Rx mode is
4397 * accept packets that match the unicast hash table.
4400 * The port identifier of the Ethernet device.
4402 * Unicast MAC address.
4404 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
4405 * 0 - Clear an unicast hash bit.
4407 * - (0) if successful.
4408 * - (-ENOTSUP) if hardware doesn't support.
4409 * - (-ENODEV) if *port_id* invalid.
4410 * - (-EIO) if device is removed.
4411 * - (-EINVAL) if bad parameter.
4413 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr,
4417 * Updates all unicast hash bitmaps for receiving packet with any Unicast
4418 * Ethernet MAC addresses,the packet is routed to all VFs for which the Rx
4419 * mode is accept packets that match the unicast hash table.
4422 * The port identifier of the Ethernet device.
4424 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
4426 * 0 - Clear all unicast hash bitmaps
4428 * - (0) if successful.
4429 * - (-ENOTSUP) if hardware doesn't support.
4430 * - (-ENODEV) if *port_id* invalid.
4431 * - (-EIO) if device is removed.
4432 * - (-EINVAL) if bad parameter.
4434 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
4437 * Set the rate limitation for a queue on an Ethernet device.
4440 * The port identifier of the Ethernet device.
4444 * The Tx rate in Mbps. Allocated from the total port link speed.
4446 * - (0) if successful.
4447 * - (-ENOTSUP) if hardware doesn't support this feature.
4448 * - (-ENODEV) if *port_id* invalid.
4449 * - (-EIO) if device is removed.
4450 * - (-EINVAL) if bad parameter.
4452 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
4456 * Configuration of Receive Side Scaling hash computation of Ethernet device.
4459 * The port identifier of the Ethernet device.
4461 * The new configuration to use for RSS hash computation on the port.
4463 * - (0) if successful.
4464 * - (-ENODEV) if port identifier is invalid.
4465 * - (-EIO) if device is removed.
4466 * - (-ENOTSUP) if hardware doesn't support.
4467 * - (-EINVAL) if bad parameter.
4469 int rte_eth_dev_rss_hash_update(uint16_t port_id,
4470 struct rte_eth_rss_conf *rss_conf);
4473 * Retrieve current configuration of Receive Side Scaling hash computation
4474 * of Ethernet device.
4477 * The port identifier of the Ethernet device.
4479 * Where to store the current RSS hash configuration of the Ethernet device.
4481 * - (0) if successful.
4482 * - (-ENODEV) if port identifier is invalid.
4483 * - (-EIO) if device is removed.
4484 * - (-ENOTSUP) if hardware doesn't support RSS.
4485 * - (-EINVAL) if bad parameter.
4488 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
4489 struct rte_eth_rss_conf *rss_conf);
4492 * Add UDP tunneling port for a type of tunnel.
4494 * Some NICs may require such configuration to properly parse a tunnel
4495 * with any standard or custom UDP port.
4496 * The packets with this UDP port will be parsed for this type of tunnel.
4497 * The device parser will also check the rest of the tunnel headers
4498 * before classifying the packet.
4500 * With some devices, this API will affect packet classification, i.e.:
4501 * - mbuf.packet_type reported on Rx
4502 * - rte_flow rules with tunnel items
4505 * The port identifier of the Ethernet device.
4507 * UDP tunneling configuration.
4510 * - (0) if successful.
4511 * - (-ENODEV) if port identifier is invalid.
4512 * - (-EIO) if device is removed.
4513 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4516 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
4517 struct rte_eth_udp_tunnel *tunnel_udp);
4520 * Delete UDP tunneling port for a type of tunnel.
4522 * The packets with this UDP port will not be classified as this type of tunnel
4523 * anymore if the device use such mapping for tunnel packet classification.
4525 * @see rte_eth_dev_udp_tunnel_port_add
4528 * The port identifier of the Ethernet device.
4530 * UDP tunneling configuration.
4533 * - (0) if successful.
4534 * - (-ENODEV) if port identifier is invalid.
4535 * - (-EIO) if device is removed.
4536 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4539 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
4540 struct rte_eth_udp_tunnel *tunnel_udp);
4543 * Get DCB information on an Ethernet device.
4546 * The port identifier of the Ethernet device.
4550 * - (0) if successful.
4551 * - (-ENODEV) if port identifier is invalid.
4552 * - (-EIO) if device is removed.
4553 * - (-ENOTSUP) if hardware doesn't support.
4554 * - (-EINVAL) if bad parameter.
4556 int rte_eth_dev_get_dcb_info(uint16_t port_id,
4557 struct rte_eth_dcb_info *dcb_info);
4559 struct rte_eth_rxtx_callback;
4562 * Add a callback to be called on packet Rx on a given port and queue.
4564 * This API configures a function to be called for each burst of
4565 * packets received on a given NIC port queue. The return value is a pointer
4566 * that can be used to later remove the callback using
4567 * rte_eth_remove_rx_callback().
4569 * Multiple functions are called in the order that they are added.
4572 * The port identifier of the Ethernet device.
4574 * The queue on the Ethernet device on which the callback is to be added.
4576 * The callback function
4578 * A generic pointer parameter which will be passed to each invocation of the
4579 * callback function on this port and queue. Inter-thread synchronization
4580 * of any user data changes is the responsibility of the user.
4584 * On success, a pointer value which can later be used to remove the callback.
4586 const struct rte_eth_rxtx_callback *
4587 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
4588 rte_rx_callback_fn fn, void *user_param);
4591 * Add a callback that must be called first on packet Rx on a given port
4594 * This API configures a first function to be called for each burst of
4595 * packets received on a given NIC port queue. The return value is a pointer
4596 * that can be used to later remove the callback using
4597 * rte_eth_remove_rx_callback().
4599 * Multiple functions are called in the order that they are added.
4602 * The port identifier of the Ethernet device.
4604 * The queue on the Ethernet device on which the callback is to be added.
4606 * The callback function
4608 * A generic pointer parameter which will be passed to each invocation of the
4609 * callback function on this port and queue. Inter-thread synchronization
4610 * of any user data changes is the responsibility of the user.
4614 * On success, a pointer value which can later be used to remove the callback.
4616 const struct rte_eth_rxtx_callback *
4617 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
4618 rte_rx_callback_fn fn, void *user_param);
4621 * Add a callback to be called on packet Tx on a given port and queue.
4623 * This API configures a function to be called for each burst of
4624 * packets sent on a given NIC port queue. The return value is a pointer
4625 * that can be used to later remove the callback using
4626 * rte_eth_remove_tx_callback().
4628 * Multiple functions are called in the order that they are added.
4631 * The port identifier of the Ethernet device.
4633 * The queue on the Ethernet device on which the callback is to be added.
4635 * The callback function
4637 * A generic pointer parameter which will be passed to each invocation of the
4638 * callback function on this port and queue. Inter-thread synchronization
4639 * of any user data changes is the responsibility of the user.
4643 * On success, a pointer value which can later be used to remove the callback.
4645 const struct rte_eth_rxtx_callback *
4646 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
4647 rte_tx_callback_fn fn, void *user_param);
4650 * Remove an Rx packet callback from a given port and queue.
4652 * This function is used to removed callbacks that were added to a NIC port
4653 * queue using rte_eth_add_rx_callback().
4655 * Note: the callback is removed from the callback list but it isn't freed
4656 * since the it may still be in use. The memory for the callback can be
4657 * subsequently freed back by the application by calling rte_free():
4659 * - Immediately - if the port is stopped, or the user knows that no
4660 * callbacks are in flight e.g. if called from the thread doing Rx/Tx
4663 * - After a short delay - where the delay is sufficient to allow any
4664 * in-flight callbacks to complete. Alternately, the RCU mechanism can be
4665 * used to detect when data plane threads have ceased referencing the
4669 * The port identifier of the Ethernet device.
4671 * The queue on the Ethernet device from which the callback is to be removed.
4673 * User supplied callback created via rte_eth_add_rx_callback().
4676 * - 0: Success. Callback was removed.
4677 * - -ENODEV: If *port_id* is invalid.
4678 * - -ENOTSUP: Callback support is not available.
4679 * - -EINVAL: The queue_id is out of range, or the callback
4680 * is NULL or not found for the port/queue.
4682 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
4683 const struct rte_eth_rxtx_callback *user_cb);
4686 * Remove a Tx packet callback from a given port and queue.
4688 * This function is used to removed callbacks that were added to a NIC port
4689 * queue using rte_eth_add_tx_callback().
4691 * Note: the callback is removed from the callback list but it isn't freed
4692 * since the it may still be in use. The memory for the callback can be
4693 * subsequently freed back by the application by calling rte_free():
4695 * - Immediately - if the port is stopped, or the user knows that no
4696 * callbacks are in flight e.g. if called from the thread doing Rx/Tx
4699 * - After a short delay - where the delay is sufficient to allow any
4700 * in-flight callbacks to complete. Alternately, the RCU mechanism can be
4701 * used to detect when data plane threads have ceased referencing the
4705 * The port identifier of the Ethernet device.
4707 * The queue on the Ethernet device from which the callback is to be removed.
4709 * User supplied callback created via rte_eth_add_tx_callback().
4712 * - 0: Success. Callback was removed.
4713 * - -ENODEV: If *port_id* is invalid.
4714 * - -ENOTSUP: Callback support is not available.
4715 * - -EINVAL: The queue_id is out of range, or the callback
4716 * is NULL or not found for the port/queue.
4718 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
4719 const struct rte_eth_rxtx_callback *user_cb);
4722 * Retrieve information about given port's Rx queue.
4725 * The port identifier of the Ethernet device.
4727 * The Rx queue on the Ethernet device for which information
4728 * will be retrieved.
4730 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
4731 * the information of the Ethernet device.
4735 * - -ENODEV: If *port_id* is invalid.
4736 * - -ENOTSUP: routine is not supported by the device PMD.
4737 * - -EINVAL: The queue_id is out of range, or the queue
4740 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4741 struct rte_eth_rxq_info *qinfo);
4744 * Retrieve information about given port's Tx queue.
4747 * The port identifier of the Ethernet device.
4749 * The Tx queue on the Ethernet device for which information
4750 * will be retrieved.
4752 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
4753 * the information of the Ethernet device.
4757 * - -ENODEV: If *port_id* is invalid.
4758 * - -ENOTSUP: routine is not supported by the device PMD.
4759 * - -EINVAL: The queue_id is out of range, or the queue
4762 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4763 struct rte_eth_txq_info *qinfo);
4766 * Retrieve information about the Rx packet burst mode.
4769 * The port identifier of the Ethernet device.
4771 * The Rx queue on the Ethernet device for which information
4772 * will be retrieved.
4774 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4775 * with the information of the packet burst mode.
4779 * - -ENODEV: If *port_id* is invalid.
4780 * - -ENOTSUP: routine is not supported by the device PMD.
4781 * - -EINVAL: The queue_id is out of range.
4783 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4784 struct rte_eth_burst_mode *mode);
4787 * Retrieve information about the Tx packet burst mode.
4790 * The port identifier of the Ethernet device.
4792 * The Tx queue on the Ethernet device for which information
4793 * will be retrieved.
4795 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4796 * with the information of the packet burst mode.
4800 * - -ENODEV: If *port_id* is invalid.
4801 * - -ENOTSUP: routine is not supported by the device PMD.
4802 * - -EINVAL: The queue_id is out of range.
4804 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4805 struct rte_eth_burst_mode *mode);
4809 * @b EXPERIMENTAL: this API may change without prior notice.
4811 * Retrieve the monitor condition for a given receive queue.
4814 * The port identifier of the Ethernet device.
4816 * The Rx queue on the Ethernet device for which information
4817 * will be retrieved.
4819 * The pointer to power-optimized monitoring condition structure.
4823 * -ENOTSUP: Operation not supported.
4824 * -EINVAL: Invalid parameters.
4825 * -ENODEV: Invalid port ID.
4828 int rte_eth_get_monitor_addr(uint16_t port_id, uint16_t queue_id,
4829 struct rte_power_monitor_cond *pmc);
4832 * Retrieve device registers and register attributes (number of registers and
4836 * The port identifier of the Ethernet device.
4838 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
4839 * NULL the function fills in the width and length fields. If non-NULL
4840 * the registers are put into the buffer pointed at by the data field.
4842 * - (0) if successful.
4843 * - (-ENOTSUP) if hardware doesn't support.
4844 * - (-EINVAL) if bad parameter.
4845 * - (-ENODEV) if *port_id* invalid.
4846 * - (-EIO) if device is removed.
4847 * - others depends on the specific operations implementation.
4849 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
4852 * Retrieve size of device EEPROM
4855 * The port identifier of the Ethernet device.
4857 * - (>=0) EEPROM size if successful.
4858 * - (-ENOTSUP) if hardware doesn't support.
4859 * - (-ENODEV) if *port_id* invalid.
4860 * - (-EIO) if device is removed.
4861 * - others depends on the specific operations implementation.
4863 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
4866 * Retrieve EEPROM and EEPROM attribute
4869 * The port identifier of the Ethernet device.
4871 * The template includes buffer for return EEPROM data and
4872 * EEPROM attributes to be filled.
4874 * - (0) if successful.
4875 * - (-ENOTSUP) if hardware doesn't support.
4876 * - (-EINVAL) if bad parameter.
4877 * - (-ENODEV) if *port_id* invalid.
4878 * - (-EIO) if device is removed.
4879 * - others depends on the specific operations implementation.
4881 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4884 * Program EEPROM with provided data
4887 * The port identifier of the Ethernet device.
4889 * The template includes EEPROM data for programming and
4890 * EEPROM attributes to be filled
4892 * - (0) if successful.
4893 * - (-ENOTSUP) if hardware doesn't support.
4894 * - (-ENODEV) if *port_id* invalid.
4895 * - (-EINVAL) if bad parameter.
4896 * - (-EIO) if device is removed.
4897 * - others depends on the specific operations implementation.
4899 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4903 * @b EXPERIMENTAL: this API may change without prior notice.
4905 * Retrieve the type and size of plugin module EEPROM
4908 * The port identifier of the Ethernet device.
4910 * The type and size of plugin module EEPROM.
4912 * - (0) if successful.
4913 * - (-ENOTSUP) if hardware doesn't support.
4914 * - (-ENODEV) if *port_id* invalid.
4915 * - (-EINVAL) if bad parameter.
4916 * - (-EIO) if device is removed.
4917 * - others depends on the specific operations implementation.
4921 rte_eth_dev_get_module_info(uint16_t port_id,
4922 struct rte_eth_dev_module_info *modinfo);
4926 * @b EXPERIMENTAL: this API may change without prior notice.
4928 * Retrieve the data of plugin module EEPROM
4931 * The port identifier of the Ethernet device.
4933 * The template includes the plugin module EEPROM attributes, and the
4934 * buffer for return plugin module EEPROM data.
4936 * - (0) if successful.
4937 * - (-ENOTSUP) if hardware doesn't support.
4938 * - (-EINVAL) if bad parameter.
4939 * - (-ENODEV) if *port_id* invalid.
4940 * - (-EIO) if device is removed.
4941 * - others depends on the specific operations implementation.
4945 rte_eth_dev_get_module_eeprom(uint16_t port_id,
4946 struct rte_dev_eeprom_info *info);
4949 * Set the list of multicast addresses to filter on an Ethernet device.
4952 * The port identifier of the Ethernet device.
4953 * @param mc_addr_set
4954 * The array of multicast addresses to set. Equal to NULL when the function
4955 * is invoked to flush the set of filtered addresses.
4957 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
4958 * when the function is invoked to flush the set of filtered addresses.
4960 * - (0) if successful.
4961 * - (-ENODEV) if *port_id* invalid.
4962 * - (-EIO) if device is removed.
4963 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
4964 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
4965 * - (-EINVAL) if bad parameter.
4967 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
4968 struct rte_ether_addr *mc_addr_set,
4969 uint32_t nb_mc_addr);
4972 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
4975 * The port identifier of the Ethernet device.
4979 * - -ENODEV: The port ID is invalid.
4980 * - -EIO: if device is removed.
4981 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4983 int rte_eth_timesync_enable(uint16_t port_id);
4986 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
4989 * The port identifier of the Ethernet device.
4993 * - -ENODEV: The port ID is invalid.
4994 * - -EIO: if device is removed.
4995 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4997 int rte_eth_timesync_disable(uint16_t port_id);
5000 * Read an IEEE1588/802.1AS Rx timestamp from an Ethernet device.
5003 * The port identifier of the Ethernet device.
5005 * Pointer to the timestamp struct.
5007 * Device specific flags. Used to pass the Rx timesync register index to
5008 * i40e. Unused in igb/ixgbe, pass 0 instead.
5012 * - -EINVAL: No timestamp is available.
5013 * - -ENODEV: The port ID is invalid.
5014 * - -EIO: if device is removed.
5015 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5017 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
5018 struct timespec *timestamp, uint32_t flags);
5021 * Read an IEEE1588/802.1AS Tx timestamp from an Ethernet device.
5024 * The port identifier of the Ethernet device.
5026 * Pointer to the timestamp struct.
5030 * - -EINVAL: No timestamp is available.
5031 * - -ENODEV: The port ID is invalid.
5032 * - -EIO: if device is removed.
5033 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5035 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
5036 struct timespec *timestamp);
5039 * Adjust the timesync clock on an Ethernet device.
5041 * This is usually used in conjunction with other Ethdev timesync functions to
5042 * synchronize the device time using the IEEE1588/802.1AS protocol.
5045 * The port identifier of the Ethernet device.
5047 * The adjustment in nanoseconds.
5051 * - -ENODEV: The port ID is invalid.
5052 * - -EIO: if device is removed.
5053 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5055 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
5058 * Read the time from the timesync clock on an Ethernet device.
5060 * This is usually used in conjunction with other Ethdev timesync functions to
5061 * synchronize the device time using the IEEE1588/802.1AS protocol.
5064 * The port identifier of the Ethernet device.
5066 * Pointer to the timespec struct that holds the time.
5070 * - -EINVAL: Bad parameter.
5072 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
5075 * Set the time of the timesync clock on an Ethernet device.
5077 * This is usually used in conjunction with other Ethdev timesync functions to
5078 * synchronize the device time using the IEEE1588/802.1AS protocol.
5081 * The port identifier of the Ethernet device.
5083 * Pointer to the timespec struct that holds the time.
5087 * - -EINVAL: No timestamp is available.
5088 * - -ENODEV: The port ID is invalid.
5089 * - -EIO: if device is removed.
5090 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5092 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
5096 * @b EXPERIMENTAL: this API may change without prior notice.
5098 * Read the current clock counter of an Ethernet device
5100 * This returns the current raw clock value of an Ethernet device. It is
5101 * a raw amount of ticks, with no given time reference.
5102 * The value returned here is from the same clock than the one
5103 * filling timestamp field of Rx packets when using hardware timestamp
5104 * offload. Therefore it can be used to compute a precise conversion of
5105 * the device clock to the real time.
5107 * E.g, a simple heuristic to derivate the frequency would be:
5108 * uint64_t start, end;
5109 * rte_eth_read_clock(port, start);
5110 * rte_delay_ms(100);
5111 * rte_eth_read_clock(port, end);
5112 * double freq = (end - start) * 10;
5114 * Compute a common reference with:
5115 * uint64_t base_time_sec = current_time();
5116 * uint64_t base_clock;
5117 * rte_eth_read_clock(port, base_clock);
5119 * Then, convert the raw mbuf timestamp with:
5120 * base_time_sec + (double)(*timestamp_dynfield(mbuf) - base_clock) / freq;
5122 * This simple example will not provide a very good accuracy. One must
5123 * at least measure multiple times the frequency and do a regression.
5124 * To avoid deviation from the system time, the common reference can
5125 * be repeated from time to time. The integer division can also be
5126 * converted by a multiplication and a shift for better performance.
5129 * The port identifier of the Ethernet device.
5131 * Pointer to the uint64_t that holds the raw clock value.
5135 * - -ENODEV: The port ID is invalid.
5136 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5137 * - -EINVAL: if bad parameter.
5141 rte_eth_read_clock(uint16_t port_id, uint64_t *clock);
5144 * Get the port ID from device name. The device name should be specified
5146 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0
5147 * - SoC device name, for example- fsl-gmac0
5148 * - vdev dpdk name, for example- net_[pcap0|null0|tap0]
5151 * pci address or name of the device
5153 * pointer to port identifier of the device
5155 * - (0) if successful and port_id is filled.
5156 * - (-ENODEV or -EINVAL) on failure.
5159 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
5162 * Get the device name from port ID. The device name is specified as below:
5163 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0
5164 * - SoC device name, for example- fsl-gmac0
5165 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0]
5168 * Port identifier of the device.
5170 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
5172 * - (0) if successful.
5173 * - (-ENODEV) if *port_id* is invalid.
5174 * - (-EINVAL) on failure.
5177 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
5180 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
5181 * the Ethernet device information, otherwise adjust them to boundaries.
5184 * The port identifier of the Ethernet device.
5186 * A pointer to a uint16_t where the number of receive
5187 * descriptors stored.
5189 * A pointer to a uint16_t where the number of transmit
5190 * descriptors stored.
5192 * - (0) if successful.
5193 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
5195 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
5196 uint16_t *nb_rx_desc,
5197 uint16_t *nb_tx_desc);
5200 * Test if a port supports specific mempool ops.
5203 * Port identifier of the Ethernet device.
5205 * The name of the pool operations to test.
5207 * - 0: best mempool ops choice for this port.
5208 * - 1: mempool ops are supported for this port.
5209 * - -ENOTSUP: mempool ops not supported for this port.
5210 * - -ENODEV: Invalid port Identifier.
5211 * - -EINVAL: Pool param is null.
5214 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
5217 * Get the security context for the Ethernet device.
5220 * Port identifier of the Ethernet device
5223 * - pointer to security context on success.
5226 rte_eth_dev_get_sec_ctx(uint16_t port_id);
5230 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
5232 * Query the device hairpin capabilities.
5235 * The port identifier of the Ethernet device.
5237 * Pointer to a structure that will hold the hairpin capabilities.
5239 * - (0) if successful.
5240 * - (-ENOTSUP) if hardware doesn't support.
5241 * - (-EINVAL) if bad parameter.
5244 int rte_eth_dev_hairpin_capability_get(uint16_t port_id,
5245 struct rte_eth_hairpin_cap *cap);
5249 * @b EXPERIMENTAL: this structure may change without prior notice.
5251 * Ethernet device representor ID range entry
5253 struct rte_eth_representor_range {
5254 enum rte_eth_representor_type type; /**< Representor type */
5255 int controller; /**< Controller index */
5256 int pf; /**< Physical function index */
5259 int vf; /**< VF start index */
5260 int sf; /**< SF start index */
5262 uint32_t id_base; /**< Representor ID start index */
5263 uint32_t id_end; /**< Representor ID end index */
5264 char name[RTE_DEV_NAME_MAX_LEN]; /**< Representor name */
5269 * @b EXPERIMENTAL: this structure may change without prior notice.
5271 * Ethernet device representor information
5273 struct rte_eth_representor_info {
5274 uint16_t controller; /**< Controller ID of caller device. */
5275 uint16_t pf; /**< Physical function ID of caller device. */
5276 uint32_t nb_ranges_alloc; /**< Size of the ranges array. */
5277 uint32_t nb_ranges; /**< Number of initialized ranges. */
5278 struct rte_eth_representor_range ranges[];/**< Representor ID range. */
5282 * Retrieve the representor info of the device.
5284 * Get device representor info to be able to calculate a unique
5285 * representor ID. @see rte_eth_representor_id_get helper.
5288 * The port identifier of the device.
5290 * A pointer to a representor info structure.
5291 * NULL to return number of range entries and allocate memory
5292 * for next call to store detail.
5293 * The number of ranges that were written into this structure
5294 * will be placed into its nb_ranges field. This number cannot be
5295 * larger than the nb_ranges_alloc that by the user before calling
5296 * this function. It can be smaller than the value returned by the
5297 * function, however.
5299 * - (-ENOTSUP) if operation is not supported.
5300 * - (-ENODEV) if *port_id* invalid.
5301 * - (-EIO) if device is removed.
5302 * - (>=0) number of available representor range entries.
5305 int rte_eth_representor_info_get(uint16_t port_id,
5306 struct rte_eth_representor_info *info);
5308 /** The NIC is able to deliver flag (if set) with packets to the PMD. */
5309 #define RTE_ETH_RX_METADATA_USER_FLAG RTE_BIT64(0)
5311 /** The NIC is able to deliver mark ID with packets to the PMD. */
5312 #define RTE_ETH_RX_METADATA_USER_MARK RTE_BIT64(1)
5314 /** The NIC is able to deliver tunnel ID with packets to the PMD. */
5315 #define RTE_ETH_RX_METADATA_TUNNEL_ID RTE_BIT64(2)
5319 * @b EXPERIMENTAL: this API may change without prior notice
5321 * Negotiate the NIC's ability to deliver specific kinds of metadata to the PMD.
5323 * Invoke this API before the first rte_eth_dev_configure() invocation
5324 * to let the PMD make preparations that are inconvenient to do later.
5326 * The negotiation process is as follows:
5328 * - the application requests features intending to use at least some of them;
5329 * - the PMD responds with the guaranteed subset of the requested feature set;
5330 * - the application can retry negotiation with another set of features;
5331 * - the application can pass zero to clear the negotiation result;
5332 * - the last negotiated result takes effect upon
5333 * the ethdev configure and start.
5336 * The PMD is supposed to first consider enabling the requested feature set
5337 * in its entirety. Only if it fails to do so, does it have the right to
5338 * respond with a smaller set of the originally requested features.
5341 * Return code (-ENOTSUP) does not necessarily mean that the requested
5342 * features are unsupported. In this case, the application should just
5343 * assume that these features can be used without prior negotiations.
5346 * Port (ethdev) identifier
5348 * @param[inout] features
5349 * Feature selection buffer
5352 * - (-EBUSY) if the port can't handle this in its current state;
5353 * - (-ENOTSUP) if the method itself is not supported by the PMD;
5354 * - (-ENODEV) if *port_id* is invalid;
5355 * - (-EINVAL) if *features* is NULL;
5356 * - (-EIO) if the device is removed;
5360 int rte_eth_rx_metadata_negotiate(uint16_t port_id, uint64_t *features);
5362 /** Flag to offload IP reassembly for IPv4 packets. */
5363 #define RTE_ETH_DEV_REASSEMBLY_F_IPV4 (RTE_BIT32(0))
5364 /** Flag to offload IP reassembly for IPv6 packets. */
5365 #define RTE_ETH_DEV_REASSEMBLY_F_IPV6 (RTE_BIT32(1))
5368 * A structure used to get/set IP reassembly configuration. It is also used
5369 * to get the maximum capability values that a PMD can support.
5371 * If rte_eth_ip_reassembly_capability_get() returns 0, IP reassembly can be
5372 * enabled using rte_eth_ip_reassembly_conf_set() and params values lower than
5373 * capability params can be set in the PMD.
5375 struct rte_eth_ip_reassembly_params {
5376 /** Maximum time in ms which PMD can wait for other fragments. */
5377 uint32_t timeout_ms;
5378 /** Maximum number of fragments that can be reassembled. */
5381 * Flags to enable reassembly of packet types -
5382 * RTE_ETH_DEV_REASSEMBLY_F_xxx.
5389 * @b EXPERIMENTAL: this API may change without prior notice
5391 * Get IP reassembly capabilities supported by the PMD. This is the first API
5392 * to be called for enabling the IP reassembly offload feature. PMD will return
5393 * the maximum values of parameters that PMD can support and user can call
5394 * rte_eth_ip_reassembly_conf_set() with param values lower than capability.
5397 * The port identifier of the device.
5399 * A pointer to rte_eth_ip_reassembly_params structure.
5401 * - (-ENOTSUP) if offload configuration is not supported by device.
5402 * - (-ENODEV) if *port_id* invalid.
5403 * - (-EIO) if device is removed.
5404 * - (-EINVAL) if device is not configured or *capa* passed is NULL.
5408 int rte_eth_ip_reassembly_capability_get(uint16_t port_id,
5409 struct rte_eth_ip_reassembly_params *capa);
5413 * @b EXPERIMENTAL: this API may change without prior notice
5415 * Get IP reassembly configuration parameters currently set in PMD.
5416 * The API will return error if the configuration is not already
5417 * set using rte_eth_ip_reassembly_conf_set() before calling this API or if
5418 * the device is not configured.
5421 * The port identifier of the device.
5423 * A pointer to rte_eth_ip_reassembly_params structure.
5425 * - (-ENOTSUP) if offload configuration is not supported by device.
5426 * - (-ENODEV) if *port_id* invalid.
5427 * - (-EIO) if device is removed.
5428 * - (-EINVAL) if device is not configured or if *conf* passed is NULL or if
5429 * configuration is not set using rte_eth_ip_reassembly_conf_set().
5433 int rte_eth_ip_reassembly_conf_get(uint16_t port_id,
5434 struct rte_eth_ip_reassembly_params *conf);
5438 * @b EXPERIMENTAL: this API may change without prior notice
5440 * Set IP reassembly configuration parameters if the PMD supports IP reassembly
5441 * offload. User should first call rte_eth_ip_reassembly_capability_get() to
5442 * check the maximum values supported by the PMD before setting the
5443 * configuration. The use of this API is mandatory to enable this feature and
5444 * should be called before rte_eth_dev_start().
5446 * In datapath, PMD cannot guarantee that IP reassembly is always successful.
5447 * Hence, PMD shall register mbuf dynamic field and dynamic flag using
5448 * rte_eth_ip_reassembly_dynfield_register() to denote incomplete IP reassembly.
5449 * If dynfield is not successfully registered, error will be returned and
5450 * IP reassembly offload cannot be used.
5453 * The port identifier of the device.
5455 * A pointer to rte_eth_ip_reassembly_params structure.
5457 * - (-ENOTSUP) if offload configuration is not supported by device.
5458 * - (-ENODEV) if *port_id* invalid.
5459 * - (-EIO) if device is removed.
5460 * - (-EINVAL) if device is not configured or if device is already started or
5461 * if *conf* passed is NULL or if mbuf dynfield is not registered
5462 * successfully by the PMD.
5466 int rte_eth_ip_reassembly_conf_set(uint16_t port_id,
5467 const struct rte_eth_ip_reassembly_params *conf);
5470 * In case of IP reassembly offload failure, packet will be updated with
5471 * dynamic flag - RTE_MBUF_DYNFLAG_IP_REASSEMBLY_INCOMPLETE_NAME and packets
5472 * will be returned without alteration.
5473 * The application can retrieve the attached fragments using mbuf dynamic field
5474 * RTE_MBUF_DYNFIELD_IP_REASSEMBLY_NAME.
5478 * Next fragment packet. Application should fetch dynamic field of
5479 * each fragment until a NULL is received and nb_frags is 0.
5481 struct rte_mbuf *next_frag;
5482 /** Time spent(in ms) by HW in waiting for further fragments. */
5483 uint16_t time_spent;
5484 /** Number of more fragments attached in mbuf dynamic fields. */
5486 } rte_eth_ip_reassembly_dynfield_t;
5490 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
5492 * Dump private info from device to a file. Provided data and the order depends
5496 * The port identifier of the Ethernet device.
5498 * A pointer to a file for output.
5501 * - (-ENODEV) if *port_id* is invalid.
5502 * - (-EINVAL) if null file.
5503 * - (-ENOTSUP) if the device does not support this function.
5504 * - (-EIO) if device is removed.
5507 int rte_eth_dev_priv_dump(uint16_t port_id, FILE *file);
5509 #include <rte_ethdev_core.h>
5513 * Helper routine for rte_eth_rx_burst().
5514 * Should be called at exit from PMD's rte_eth_rx_bulk implementation.
5515 * Does necessary post-processing - invokes Rx callbacks if any, etc.
5518 * The port identifier of the Ethernet device.
5520 * The index of the receive queue from which to retrieve input packets.
5522 * The address of an array of pointers to *rte_mbuf* structures that
5523 * have been retrieved from the device.
5525 * The number of packets that were retrieved from the device.
5527 * The number of elements in @p rx_pkts array.
5529 * Opaque pointer of Rx queue callback related data.
5532 * The number of packets effectively supplied to the @p rx_pkts array.
5534 uint16_t rte_eth_call_rx_callbacks(uint16_t port_id, uint16_t queue_id,
5535 struct rte_mbuf **rx_pkts, uint16_t nb_rx, uint16_t nb_pkts,
5540 * Retrieve a burst of input packets from a receive queue of an Ethernet
5541 * device. The retrieved packets are stored in *rte_mbuf* structures whose
5542 * pointers are supplied in the *rx_pkts* array.
5544 * The rte_eth_rx_burst() function loops, parsing the Rx ring of the
5545 * receive queue, up to *nb_pkts* packets, and for each completed Rx
5546 * descriptor in the ring, it performs the following operations:
5548 * - Initialize the *rte_mbuf* data structure associated with the
5549 * Rx descriptor according to the information provided by the NIC into
5550 * that Rx descriptor.
5552 * - Store the *rte_mbuf* data structure into the next entry of the
5555 * - Replenish the Rx descriptor with a new *rte_mbuf* buffer
5556 * allocated from the memory pool associated with the receive queue at
5557 * initialization time.
5559 * When retrieving an input packet that was scattered by the controller
5560 * into multiple receive descriptors, the rte_eth_rx_burst() function
5561 * appends the associated *rte_mbuf* buffers to the first buffer of the
5564 * The rte_eth_rx_burst() function returns the number of packets
5565 * actually retrieved, which is the number of *rte_mbuf* data structures
5566 * effectively supplied into the *rx_pkts* array.
5567 * A return value equal to *nb_pkts* indicates that the Rx queue contained
5568 * at least *rx_pkts* packets, and this is likely to signify that other
5569 * received packets remain in the input queue. Applications implementing
5570 * a "retrieve as much received packets as possible" policy can check this
5571 * specific case and keep invoking the rte_eth_rx_burst() function until
5572 * a value less than *nb_pkts* is returned.
5574 * This receive method has the following advantages:
5576 * - It allows a run-to-completion network stack engine to retrieve and
5577 * to immediately process received packets in a fast burst-oriented
5578 * approach, avoiding the overhead of unnecessary intermediate packet
5579 * queue/dequeue operations.
5581 * - Conversely, it also allows an asynchronous-oriented processing
5582 * method to retrieve bursts of received packets and to immediately
5583 * queue them for further parallel processing by another logical core,
5584 * for instance. However, instead of having received packets being
5585 * individually queued by the driver, this approach allows the caller
5586 * of the rte_eth_rx_burst() function to queue a burst of retrieved
5587 * packets at a time and therefore dramatically reduce the cost of
5588 * enqueue/dequeue operations per packet.
5590 * - It allows the rte_eth_rx_burst() function of the driver to take
5591 * advantage of burst-oriented hardware features (CPU cache,
5592 * prefetch instructions, and so on) to minimize the number of CPU
5593 * cycles per packet.
5595 * To summarize, the proposed receive API enables many
5596 * burst-oriented optimizations in both synchronous and asynchronous
5597 * packet processing environments with no overhead in both cases.
5600 * Some drivers using vector instructions require that *nb_pkts* is
5601 * divisible by 4 or 8, depending on the driver implementation.
5603 * The rte_eth_rx_burst() function does not provide any error
5604 * notification to avoid the corresponding overhead. As a hint, the
5605 * upper-level application might check the status of the device link once
5606 * being systematically returned a 0 value for a given number of tries.
5609 * The port identifier of the Ethernet device.
5611 * The index of the receive queue from which to retrieve input packets.
5612 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
5613 * to rte_eth_dev_configure().
5615 * The address of an array of pointers to *rte_mbuf* structures that
5616 * must be large enough to store *nb_pkts* pointers in it.
5618 * The maximum number of packets to retrieve.
5619 * The value must be divisible by 8 in order to work with any driver.
5621 * The number of packets actually retrieved, which is the number
5622 * of pointers to *rte_mbuf* structures effectively supplied to the
5625 static inline uint16_t
5626 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
5627 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
5630 struct rte_eth_fp_ops *p;
5633 #ifdef RTE_ETHDEV_DEBUG_RX
5634 if (port_id >= RTE_MAX_ETHPORTS ||
5635 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5637 "Invalid port_id=%u or queue_id=%u\n",
5643 /* fetch pointer to queue data */
5644 p = &rte_eth_fp_ops[port_id];
5645 qd = p->rxq.data[queue_id];
5647 #ifdef RTE_ETHDEV_DEBUG_RX
5648 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
5651 RTE_ETHDEV_LOG(ERR, "Invalid Rx queue_id=%u for port_id=%u\n",
5657 nb_rx = p->rx_pkt_burst(qd, rx_pkts, nb_pkts);
5659 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
5663 /* __ATOMIC_RELEASE memory order was used when the
5664 * call back was inserted into the list.
5665 * Since there is a clear dependency between loading
5666 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
5669 cb = __atomic_load_n((void **)&p->rxq.clbk[queue_id],
5671 if (unlikely(cb != NULL))
5672 nb_rx = rte_eth_call_rx_callbacks(port_id, queue_id,
5673 rx_pkts, nb_rx, nb_pkts, cb);
5677 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx);
5682 * Get the number of used descriptors of a Rx queue
5685 * The port identifier of the Ethernet device.
5687 * The queue ID on the specific port.
5689 * The number of used descriptors in the specific queue, or:
5690 * - (-ENODEV) if *port_id* is invalid.
5691 * (-EINVAL) if *queue_id* is invalid
5692 * (-ENOTSUP) if the device does not support this function
5695 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
5697 struct rte_eth_fp_ops *p;
5700 if (port_id >= RTE_MAX_ETHPORTS ||
5701 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5703 "Invalid port_id=%u or queue_id=%u\n",
5708 /* fetch pointer to queue data */
5709 p = &rte_eth_fp_ops[port_id];
5710 qd = p->rxq.data[queue_id];
5712 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5713 RTE_FUNC_PTR_OR_ERR_RET(*p->rx_queue_count, -ENOTSUP);
5717 return (int)(*p->rx_queue_count)(qd);
5720 /**@{@name Rx hardware descriptor states
5721 * @see rte_eth_rx_descriptor_status
5723 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
5724 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
5725 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
5729 * Check the status of a Rx descriptor in the queue
5731 * It should be called in a similar context than the Rx function:
5732 * - on a dataplane core
5733 * - not concurrently on the same queue
5735 * Since it's a dataplane function, no check is performed on port_id and
5736 * queue_id. The caller must therefore ensure that the port is enabled
5737 * and the queue is configured and running.
5739 * Note: accessing to a random descriptor in the ring may trigger cache
5740 * misses and have a performance impact.
5743 * A valid port identifier of the Ethernet device which.
5745 * A valid Rx queue identifier on this port.
5747 * The offset of the descriptor starting from tail (0 is the next
5748 * packet to be received by the driver).
5751 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
5753 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
5754 * not yet processed by the driver (i.e. in the receive queue).
5755 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
5756 * the driver and not yet returned to hw, or reserved by the hw.
5757 * - (-EINVAL) bad descriptor offset.
5758 * - (-ENOTSUP) if the device does not support this function.
5759 * - (-ENODEV) bad port or queue (only if compiled with debug).
5762 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
5765 struct rte_eth_fp_ops *p;
5768 #ifdef RTE_ETHDEV_DEBUG_RX
5769 if (port_id >= RTE_MAX_ETHPORTS ||
5770 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5772 "Invalid port_id=%u or queue_id=%u\n",
5778 /* fetch pointer to queue data */
5779 p = &rte_eth_fp_ops[port_id];
5780 qd = p->rxq.data[queue_id];
5782 #ifdef RTE_ETHDEV_DEBUG_RX
5783 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5787 RTE_FUNC_PTR_OR_ERR_RET(*p->rx_descriptor_status, -ENOTSUP);
5788 return (*p->rx_descriptor_status)(qd, offset);
5791 /**@{@name Tx hardware descriptor states
5792 * @see rte_eth_tx_descriptor_status
5794 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
5795 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
5796 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
5800 * Check the status of a Tx descriptor in the queue.
5802 * It should be called in a similar context than the Tx function:
5803 * - on a dataplane core
5804 * - not concurrently on the same queue
5806 * Since it's a dataplane function, no check is performed on port_id and
5807 * queue_id. The caller must therefore ensure that the port is enabled
5808 * and the queue is configured and running.
5810 * Note: accessing to a random descriptor in the ring may trigger cache
5811 * misses and have a performance impact.
5814 * A valid port identifier of the Ethernet device which.
5816 * A valid Tx queue identifier on this port.
5818 * The offset of the descriptor starting from tail (0 is the place where
5819 * the next packet will be send).
5822 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
5823 * in the transmit queue.
5824 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
5825 * be reused by the driver.
5826 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
5827 * driver or the hardware.
5828 * - (-EINVAL) bad descriptor offset.
5829 * - (-ENOTSUP) if the device does not support this function.
5830 * - (-ENODEV) bad port or queue (only if compiled with debug).
5832 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
5833 uint16_t queue_id, uint16_t offset)
5835 struct rte_eth_fp_ops *p;
5838 #ifdef RTE_ETHDEV_DEBUG_TX
5839 if (port_id >= RTE_MAX_ETHPORTS ||
5840 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5842 "Invalid port_id=%u or queue_id=%u\n",
5848 /* fetch pointer to queue data */
5849 p = &rte_eth_fp_ops[port_id];
5850 qd = p->txq.data[queue_id];
5852 #ifdef RTE_ETHDEV_DEBUG_TX
5853 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5857 RTE_FUNC_PTR_OR_ERR_RET(*p->tx_descriptor_status, -ENOTSUP);
5858 return (*p->tx_descriptor_status)(qd, offset);
5863 * Helper routine for rte_eth_tx_burst().
5864 * Should be called before entry PMD's rte_eth_tx_bulk implementation.
5865 * Does necessary pre-processing - invokes Tx callbacks if any, etc.
5868 * The port identifier of the Ethernet device.
5870 * The index of the transmit queue through which output packets must be
5873 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5874 * which contain the output packets.
5876 * The maximum number of packets to transmit.
5878 * The number of output packets to transmit.
5880 uint16_t rte_eth_call_tx_callbacks(uint16_t port_id, uint16_t queue_id,
5881 struct rte_mbuf **tx_pkts, uint16_t nb_pkts, void *opaque);
5884 * Send a burst of output packets on a transmit queue of an Ethernet device.
5886 * The rte_eth_tx_burst() function is invoked to transmit output packets
5887 * on the output queue *queue_id* of the Ethernet device designated by its
5889 * The *nb_pkts* parameter is the number of packets to send which are
5890 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
5891 * allocated from a pool created with rte_pktmbuf_pool_create().
5892 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
5893 * up to the number of transmit descriptors available in the Tx ring of the
5895 * For each packet to send, the rte_eth_tx_burst() function performs
5896 * the following operations:
5898 * - Pick up the next available descriptor in the transmit ring.
5900 * - Free the network buffer previously sent with that descriptor, if any.
5902 * - Initialize the transmit descriptor with the information provided
5903 * in the *rte_mbuf data structure.
5905 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
5906 * the rte_eth_tx_burst() function uses several transmit descriptors
5909 * The rte_eth_tx_burst() function returns the number of packets it
5910 * actually sent. A return value equal to *nb_pkts* means that all packets
5911 * have been sent, and this is likely to signify that other output packets
5912 * could be immediately transmitted again. Applications that implement a
5913 * "send as many packets to transmit as possible" policy can check this
5914 * specific case and keep invoking the rte_eth_tx_burst() function until
5915 * a value less than *nb_pkts* is returned.
5917 * It is the responsibility of the rte_eth_tx_burst() function to
5918 * transparently free the memory buffers of packets previously sent.
5919 * This feature is driven by the *tx_free_thresh* value supplied to the
5920 * rte_eth_dev_configure() function at device configuration time.
5921 * When the number of free Tx descriptors drops below this threshold, the
5922 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
5923 * of those packets whose transmission was effectively completed.
5925 * If the PMD is RTE_ETH_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
5926 * invoke this function concurrently on the same Tx queue without SW lock.
5927 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads
5929 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
5933 * The port identifier of the Ethernet device.
5935 * The index of the transmit queue through which output packets must be
5937 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5938 * to rte_eth_dev_configure().
5940 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5941 * which contain the output packets.
5943 * The maximum number of packets to transmit.
5945 * The number of output packets actually stored in transmit descriptors of
5946 * the transmit ring. The return value can be less than the value of the
5947 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
5949 static inline uint16_t
5950 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
5951 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5953 struct rte_eth_fp_ops *p;
5956 #ifdef RTE_ETHDEV_DEBUG_TX
5957 if (port_id >= RTE_MAX_ETHPORTS ||
5958 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5960 "Invalid port_id=%u or queue_id=%u\n",
5966 /* fetch pointer to queue data */
5967 p = &rte_eth_fp_ops[port_id];
5968 qd = p->txq.data[queue_id];
5970 #ifdef RTE_ETHDEV_DEBUG_TX
5971 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
5974 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n",
5980 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
5984 /* __ATOMIC_RELEASE memory order was used when the
5985 * call back was inserted into the list.
5986 * Since there is a clear dependency between loading
5987 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
5990 cb = __atomic_load_n((void **)&p->txq.clbk[queue_id],
5992 if (unlikely(cb != NULL))
5993 nb_pkts = rte_eth_call_tx_callbacks(port_id, queue_id,
5994 tx_pkts, nb_pkts, cb);
5998 nb_pkts = p->tx_pkt_burst(qd, tx_pkts, nb_pkts);
6000 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts, nb_pkts);
6005 * Process a burst of output packets on a transmit queue of an Ethernet device.
6007 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
6008 * transmitted on the output queue *queue_id* of the Ethernet device designated
6010 * The *nb_pkts* parameter is the number of packets to be prepared which are
6011 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
6012 * allocated from a pool created with rte_pktmbuf_pool_create().
6013 * For each packet to send, the rte_eth_tx_prepare() function performs
6014 * the following operations:
6016 * - Check if packet meets devices requirements for Tx offloads.
6018 * - Check limitations about number of segments.
6020 * - Check additional requirements when debug is enabled.
6022 * - Update and/or reset required checksums when Tx offload is set for packet.
6024 * Since this function can modify packet data, provided mbufs must be safely
6025 * writable (e.g. modified data cannot be in shared segment).
6027 * The rte_eth_tx_prepare() function returns the number of packets ready to be
6028 * sent. A return value equal to *nb_pkts* means that all packets are valid and
6029 * ready to be sent, otherwise stops processing on the first invalid packet and
6030 * leaves the rest packets untouched.
6032 * When this functionality is not implemented in the driver, all packets are
6033 * are returned untouched.
6036 * The port identifier of the Ethernet device.
6037 * The value must be a valid port ID.
6039 * The index of the transmit queue through which output packets must be
6041 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
6042 * to rte_eth_dev_configure().
6044 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
6045 * which contain the output packets.
6047 * The maximum number of packets to process.
6049 * The number of packets correct and ready to be sent. The return value can be
6050 * less than the value of the *tx_pkts* parameter when some packet doesn't
6051 * meet devices requirements with rte_errno set appropriately:
6052 * - EINVAL: offload flags are not correctly set
6053 * - ENOTSUP: the offload feature is not supported by the hardware
6054 * - ENODEV: if *port_id* is invalid (with debug enabled only)
6058 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
6060 static inline uint16_t
6061 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
6062 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
6064 struct rte_eth_fp_ops *p;
6067 #ifdef RTE_ETHDEV_DEBUG_TX
6068 if (port_id >= RTE_MAX_ETHPORTS ||
6069 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
6071 "Invalid port_id=%u or queue_id=%u\n",
6078 /* fetch pointer to queue data */
6079 p = &rte_eth_fp_ops[port_id];
6080 qd = p->txq.data[queue_id];
6082 #ifdef RTE_ETHDEV_DEBUG_TX
6083 if (!rte_eth_dev_is_valid_port(port_id)) {
6084 RTE_ETHDEV_LOG(ERR, "Invalid Tx port_id=%u\n", port_id);
6089 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n",
6096 if (!p->tx_pkt_prepare)
6099 return p->tx_pkt_prepare(qd, tx_pkts, nb_pkts);
6105 * Native NOOP operation for compilation targets which doesn't require any
6106 * preparations steps, and functional NOOP may introduce unnecessary performance
6109 * Generally this is not a good idea to turn it on globally and didn't should
6110 * be used if behavior of tx_preparation can change.
6113 static inline uint16_t
6114 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
6115 __rte_unused uint16_t queue_id,
6116 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
6124 * Send any packets queued up for transmission on a port and HW queue
6126 * This causes an explicit flush of packets previously buffered via the
6127 * rte_eth_tx_buffer() function. It returns the number of packets successfully
6128 * sent to the NIC, and calls the error callback for any unsent packets. Unless
6129 * explicitly set up otherwise, the default callback simply frees the unsent
6130 * packets back to the owning mempool.
6133 * The port identifier of the Ethernet device.
6135 * The index of the transmit queue through which output packets must be
6137 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
6138 * to rte_eth_dev_configure().
6140 * Buffer of packets to be transmit.
6142 * The number of packets successfully sent to the Ethernet device. The error
6143 * callback is called for any packets which could not be sent.
6145 static inline uint16_t
6146 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
6147 struct rte_eth_dev_tx_buffer *buffer)
6150 uint16_t to_send = buffer->length;
6155 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
6159 /* All packets sent, or to be dealt with by callback below */
6160 if (unlikely(sent != to_send))
6161 buffer->error_callback(&buffer->pkts[sent],
6162 (uint16_t)(to_send - sent),
6163 buffer->error_userdata);
6169 * Buffer a single packet for future transmission on a port and queue
6171 * This function takes a single mbuf/packet and buffers it for later
6172 * transmission on the particular port and queue specified. Once the buffer is
6173 * full of packets, an attempt will be made to transmit all the buffered
6174 * packets. In case of error, where not all packets can be transmitted, a
6175 * callback is called with the unsent packets as a parameter. If no callback
6176 * is explicitly set up, the unsent packets are just freed back to the owning
6177 * mempool. The function returns the number of packets actually sent i.e.
6178 * 0 if no buffer flush occurred, otherwise the number of packets successfully
6182 * The port identifier of the Ethernet device.
6184 * The index of the transmit queue through which output packets must be
6186 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
6187 * to rte_eth_dev_configure().
6189 * Buffer used to collect packets to be sent.
6191 * Pointer to the packet mbuf to be sent.
6193 * 0 = packet has been buffered for later transmission
6194 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
6195 * causing N packets to be sent, and the error callback to be called for
6198 static __rte_always_inline uint16_t
6199 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
6200 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
6202 buffer->pkts[buffer->length++] = tx_pkt;
6203 if (buffer->length < buffer->size)
6206 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
6213 #endif /* _RTE_ETHDEV_H_ */