1 /* SPDX-License-Identifier: BSD-3-Clause
2 * Copyright(c) 2010-2017 Intel Corporation
11 * RTE Ethernet Device API
13 * The Ethernet Device API is composed of two parts:
15 * - The application-oriented Ethernet API that includes functions to setup
16 * an Ethernet device (configure it, setup its Rx and Tx queues and start it),
17 * to get its MAC address, the speed and the status of its physical link,
18 * to receive and to transmit packets, and so on.
20 * - The driver-oriented Ethernet API that exports functions allowing
21 * an Ethernet Poll Mode Driver (PMD) to allocate an Ethernet device instance,
22 * create memzone for HW rings and process registered callbacks, and so on.
23 * PMDs should include ethdev_driver.h instead of this header.
25 * By default, all the functions of the Ethernet Device API exported by a PMD
26 * are lock-free functions which assume to not be invoked in parallel on
27 * different logical cores to work on the same target object. For instance,
28 * the receive function of a PMD cannot be invoked in parallel on two logical
29 * cores to poll the same Rx queue [of the same port]. Of course, this function
30 * can be invoked in parallel by different logical cores on different Rx queues.
31 * It is the responsibility of the upper level application to enforce this rule.
33 * If needed, parallel accesses by multiple logical cores to shared queues
34 * shall be explicitly protected by dedicated inline lock-aware functions
35 * built on top of their corresponding lock-free functions of the PMD API.
37 * In all functions of the Ethernet API, the Ethernet device is
38 * designated by an integer >= 0 named the device port identifier.
40 * At the Ethernet driver level, Ethernet devices are represented by a generic
41 * data structure of type *rte_eth_dev*.
43 * Ethernet devices are dynamically registered during the PCI probing phase
44 * performed at EAL initialization time.
45 * When an Ethernet device is being probed, an *rte_eth_dev* structure and
46 * a new port identifier are allocated for that device. Then, the eth_dev_init()
47 * function supplied by the Ethernet driver matching the probed PCI
48 * device is invoked to properly initialize the device.
50 * The role of the device init function consists of resetting the hardware,
51 * checking access to Non-volatile Memory (NVM), reading the MAC address
54 * If the device init operation is successful, the correspondence between
55 * the port identifier assigned to the new device and its associated
56 * *rte_eth_dev* structure is effectively registered.
57 * Otherwise, both the *rte_eth_dev* structure and the port identifier are
60 * The functions exported by the application Ethernet API to setup a device
61 * designated by its port identifier must be invoked in the following order:
62 * - rte_eth_dev_configure()
63 * - rte_eth_tx_queue_setup()
64 * - rte_eth_rx_queue_setup()
65 * - rte_eth_dev_start()
67 * Then, the network application can invoke, in any order, the functions
68 * exported by the Ethernet API to get the MAC address of a given device, to
69 * get the speed and the status of a device physical link, to receive/transmit
70 * [burst of] packets, and so on.
72 * If the application wants to change the configuration (i.e. call
73 * rte_eth_dev_configure(), rte_eth_tx_queue_setup(), or
74 * rte_eth_rx_queue_setup()), it must call rte_eth_dev_stop() first to stop the
75 * device and then do the reconfiguration before calling rte_eth_dev_start()
76 * again. The transmit and receive functions should not be invoked when the
79 * Please note that some configuration is not stored between calls to
80 * rte_eth_dev_stop()/rte_eth_dev_start(). The following configuration will
84 * - flow control settings
85 * - receive mode configuration (promiscuous mode, all-multicast mode,
86 * hardware checksum mode, RSS/VMDq settings etc.)
87 * - VLAN filtering configuration
88 * - default MAC address
89 * - MAC addresses supplied to MAC address array
90 * - flow director filtering mode (but not filtering rules)
91 * - NIC queue statistics mappings
93 * 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_ether.h>
173 #include <rte_power_intrinsics.h>
175 #include "rte_ethdev_trace_fp.h"
176 #include "rte_dev_info.h"
178 extern int rte_eth_dev_logtype;
180 #define RTE_ETHDEV_LOG(level, ...) \
181 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__)
186 * Initializes a device iterator.
188 * This iterator allows accessing a list of devices matching some devargs.
191 * Device iterator handle initialized by the function.
192 * The fields bus_str and cls_str might be dynamically allocated,
193 * and could be freed by calling rte_eth_iterator_cleanup().
196 * Device description string.
199 * 0 on successful initialization, negative otherwise.
201 int rte_eth_iterator_init(struct rte_dev_iterator *iter, const char *devargs);
204 * Iterates on devices with devargs filter.
205 * The ownership is not checked.
207 * The next port ID is returned, and the iterator is updated.
210 * Device iterator handle initialized by rte_eth_iterator_init().
211 * Some fields bus_str and cls_str might be freed when no more port is found,
212 * by calling rte_eth_iterator_cleanup().
215 * A port ID if found, RTE_MAX_ETHPORTS otherwise.
217 uint16_t rte_eth_iterator_next(struct rte_dev_iterator *iter);
220 * Free some allocated fields of the iterator.
222 * This function is automatically called by rte_eth_iterator_next()
223 * on the last iteration (i.e. when no more matching port is found).
225 * It is safe to call this function twice; it will do nothing more.
228 * Device iterator handle initialized by rte_eth_iterator_init().
229 * The fields bus_str and cls_str are freed if needed.
231 void rte_eth_iterator_cleanup(struct rte_dev_iterator *iter);
234 * Macro to iterate over all ethdev ports matching some devargs.
236 * If a break is done before the end of the loop,
237 * the function rte_eth_iterator_cleanup() must be called.
240 * Iterated port ID of type uint16_t.
242 * Device parameters input as string of type char*.
244 * Iterator handle of type struct rte_dev_iterator, used internally.
246 #define RTE_ETH_FOREACH_MATCHING_DEV(id, devargs, iter) \
247 for (rte_eth_iterator_init(iter, devargs), \
248 id = rte_eth_iterator_next(iter); \
249 id != RTE_MAX_ETHPORTS; \
250 id = rte_eth_iterator_next(iter))
253 * A structure used to retrieve statistics for an Ethernet port.
254 * Not all statistics fields in struct rte_eth_stats are supported
255 * by any type of network interface card (NIC). If any statistics
256 * field is not supported, its value is 0.
257 * All byte-related statistics do not include Ethernet FCS regardless
258 * of whether these bytes have been delivered to the application
259 * (see RTE_ETH_RX_OFFLOAD_KEEP_CRC).
261 struct rte_eth_stats {
262 uint64_t ipackets; /**< Total number of successfully received packets. */
263 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
264 uint64_t ibytes; /**< Total number of successfully received bytes. */
265 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
267 * Total of Rx packets dropped by the HW,
268 * because there are no available buffer (i.e. Rx queues are full).
271 uint64_t ierrors; /**< Total number of erroneous received packets. */
272 uint64_t oerrors; /**< Total number of failed transmitted packets. */
273 uint64_t rx_nombuf; /**< Total number of Rx mbuf allocation failures. */
274 /* Queue stats are limited to max 256 queues */
275 /** Total number of queue Rx packets. */
276 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
277 /** Total number of queue Tx packets. */
278 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
279 /** Total number of successfully received queue bytes. */
280 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
281 /** Total number of successfully transmitted queue bytes. */
282 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
283 /** Total number of queue packets received that are dropped. */
284 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS];
287 /**@{@name Link speed capabilities
288 * Device supported speeds bitmap flags
290 #define RTE_ETH_LINK_SPEED_AUTONEG 0 /**< Autonegotiate (all speeds) */
291 #define RTE_ETH_LINK_SPEED_FIXED RTE_BIT32(0) /**< Disable autoneg (fixed speed) */
292 #define RTE_ETH_LINK_SPEED_10M_HD RTE_BIT32(1) /**< 10 Mbps half-duplex */
293 #define RTE_ETH_LINK_SPEED_10M RTE_BIT32(2) /**< 10 Mbps full-duplex */
294 #define RTE_ETH_LINK_SPEED_100M_HD RTE_BIT32(3) /**< 100 Mbps half-duplex */
295 #define RTE_ETH_LINK_SPEED_100M RTE_BIT32(4) /**< 100 Mbps full-duplex */
296 #define RTE_ETH_LINK_SPEED_1G RTE_BIT32(5) /**< 1 Gbps */
297 #define RTE_ETH_LINK_SPEED_2_5G RTE_BIT32(6) /**< 2.5 Gbps */
298 #define RTE_ETH_LINK_SPEED_5G RTE_BIT32(7) /**< 5 Gbps */
299 #define RTE_ETH_LINK_SPEED_10G RTE_BIT32(8) /**< 10 Gbps */
300 #define RTE_ETH_LINK_SPEED_20G RTE_BIT32(9) /**< 20 Gbps */
301 #define RTE_ETH_LINK_SPEED_25G RTE_BIT32(10) /**< 25 Gbps */
302 #define RTE_ETH_LINK_SPEED_40G RTE_BIT32(11) /**< 40 Gbps */
303 #define RTE_ETH_LINK_SPEED_50G RTE_BIT32(12) /**< 50 Gbps */
304 #define RTE_ETH_LINK_SPEED_56G RTE_BIT32(13) /**< 56 Gbps */
305 #define RTE_ETH_LINK_SPEED_100G RTE_BIT32(14) /**< 100 Gbps */
306 #define RTE_ETH_LINK_SPEED_200G RTE_BIT32(15) /**< 200 Gbps */
309 #define ETH_LINK_SPEED_AUTONEG RTE_DEPRECATED(ETH_LINK_SPEED_AUTONEG) RTE_ETH_LINK_SPEED_AUTONEG
310 #define ETH_LINK_SPEED_FIXED RTE_DEPRECATED(ETH_LINK_SPEED_FIXED) RTE_ETH_LINK_SPEED_FIXED
311 #define ETH_LINK_SPEED_10M_HD RTE_DEPRECATED(ETH_LINK_SPEED_10M_HD) RTE_ETH_LINK_SPEED_10M_HD
312 #define ETH_LINK_SPEED_10M RTE_DEPRECATED(ETH_LINK_SPEED_10M) RTE_ETH_LINK_SPEED_10M
313 #define ETH_LINK_SPEED_100M_HD RTE_DEPRECATED(ETH_LINK_SPEED_100M_HD) RTE_ETH_LINK_SPEED_100M_HD
314 #define ETH_LINK_SPEED_100M RTE_DEPRECATED(ETH_LINK_SPEED_100M) RTE_ETH_LINK_SPEED_100M
315 #define ETH_LINK_SPEED_1G RTE_DEPRECATED(ETH_LINK_SPEED_1G) RTE_ETH_LINK_SPEED_1G
316 #define ETH_LINK_SPEED_2_5G RTE_DEPRECATED(ETH_LINK_SPEED_2_5G) RTE_ETH_LINK_SPEED_2_5G
317 #define ETH_LINK_SPEED_5G RTE_DEPRECATED(ETH_LINK_SPEED_5G) RTE_ETH_LINK_SPEED_5G
318 #define ETH_LINK_SPEED_10G RTE_DEPRECATED(ETH_LINK_SPEED_10G) RTE_ETH_LINK_SPEED_10G
319 #define ETH_LINK_SPEED_20G RTE_DEPRECATED(ETH_LINK_SPEED_20G) RTE_ETH_LINK_SPEED_20G
320 #define ETH_LINK_SPEED_25G RTE_DEPRECATED(ETH_LINK_SPEED_25G) RTE_ETH_LINK_SPEED_25G
321 #define ETH_LINK_SPEED_40G RTE_DEPRECATED(ETH_LINK_SPEED_40G) RTE_ETH_LINK_SPEED_40G
322 #define ETH_LINK_SPEED_50G RTE_DEPRECATED(ETH_LINK_SPEED_50G) RTE_ETH_LINK_SPEED_50G
323 #define ETH_LINK_SPEED_56G RTE_DEPRECATED(ETH_LINK_SPEED_56G) RTE_ETH_LINK_SPEED_56G
324 #define ETH_LINK_SPEED_100G RTE_DEPRECATED(ETH_LINK_SPEED_100G) RTE_ETH_LINK_SPEED_100G
325 #define ETH_LINK_SPEED_200G RTE_DEPRECATED(ETH_LINK_SPEED_200G) RTE_ETH_LINK_SPEED_200G
327 /**@{@name Link speed
328 * Ethernet numeric link speeds in Mbps
330 #define RTE_ETH_SPEED_NUM_NONE 0 /**< Not defined */
331 #define RTE_ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
332 #define RTE_ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
333 #define RTE_ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
334 #define RTE_ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
335 #define RTE_ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
336 #define RTE_ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
337 #define RTE_ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
338 #define RTE_ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
339 #define RTE_ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
340 #define RTE_ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
341 #define RTE_ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
342 #define RTE_ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
343 #define RTE_ETH_SPEED_NUM_200G 200000 /**< 200 Gbps */
344 #define RTE_ETH_SPEED_NUM_UNKNOWN UINT32_MAX /**< Unknown */
347 #define ETH_SPEED_NUM_NONE RTE_DEPRECATED(ETH_SPEED_NUM_NONE) RTE_ETH_SPEED_NUM_NONE
348 #define ETH_SPEED_NUM_10M RTE_DEPRECATED(ETH_SPEED_NUM_10M) RTE_ETH_SPEED_NUM_10M
349 #define ETH_SPEED_NUM_100M RTE_DEPRECATED(ETH_SPEED_NUM_100M) RTE_ETH_SPEED_NUM_100M
350 #define ETH_SPEED_NUM_1G RTE_DEPRECATED(ETH_SPEED_NUM_1G) RTE_ETH_SPEED_NUM_1G
351 #define ETH_SPEED_NUM_2_5G RTE_DEPRECATED(ETH_SPEED_NUM_2_5G) RTE_ETH_SPEED_NUM_2_5G
352 #define ETH_SPEED_NUM_5G RTE_DEPRECATED(ETH_SPEED_NUM_5G) RTE_ETH_SPEED_NUM_5G
353 #define ETH_SPEED_NUM_10G RTE_DEPRECATED(ETH_SPEED_NUM_10G) RTE_ETH_SPEED_NUM_10G
354 #define ETH_SPEED_NUM_20G RTE_DEPRECATED(ETH_SPEED_NUM_20G) RTE_ETH_SPEED_NUM_20G
355 #define ETH_SPEED_NUM_25G RTE_DEPRECATED(ETH_SPEED_NUM_25G) RTE_ETH_SPEED_NUM_25G
356 #define ETH_SPEED_NUM_40G RTE_DEPRECATED(ETH_SPEED_NUM_40G) RTE_ETH_SPEED_NUM_40G
357 #define ETH_SPEED_NUM_50G RTE_DEPRECATED(ETH_SPEED_NUM_50G) RTE_ETH_SPEED_NUM_50G
358 #define ETH_SPEED_NUM_56G RTE_DEPRECATED(ETH_SPEED_NUM_56G) RTE_ETH_SPEED_NUM_56G
359 #define ETH_SPEED_NUM_100G RTE_DEPRECATED(ETH_SPEED_NUM_100G) RTE_ETH_SPEED_NUM_100G
360 #define ETH_SPEED_NUM_200G RTE_DEPRECATED(ETH_SPEED_NUM_200G) RTE_ETH_SPEED_NUM_200G
361 #define ETH_SPEED_NUM_UNKNOWN RTE_DEPRECATED(ETH_SPEED_NUM_UNKNOWN) RTE_ETH_SPEED_NUM_UNKNOWN
364 * A structure used to retrieve link-level information of an Ethernet port.
367 struct rte_eth_link {
368 uint32_t link_speed; /**< RTE_ETH_SPEED_NUM_ */
369 uint16_t link_duplex : 1; /**< RTE_ETH_LINK_[HALF/FULL]_DUPLEX */
370 uint16_t link_autoneg : 1; /**< RTE_ETH_LINK_[AUTONEG/FIXED] */
371 uint16_t link_status : 1; /**< RTE_ETH_LINK_[DOWN/UP] */
372 } __rte_aligned(8); /**< aligned for atomic64 read/write */
374 /**@{@name Link negotiation
375 * Constants used in link management.
377 #define RTE_ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
378 #define RTE_ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
379 #define RTE_ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
380 #define RTE_ETH_LINK_UP 1 /**< Link is up (see link_status). */
381 #define RTE_ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
382 #define RTE_ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
383 #define RTE_ETH_LINK_MAX_STR_LEN 40 /**< Max length of default link string. */
386 #define ETH_LINK_HALF_DUPLEX RTE_DEPRECATED(ETH_LINK_HALF_DUPLEX) RTE_ETH_LINK_HALF_DUPLEX
387 #define ETH_LINK_FULL_DUPLEX RTE_DEPRECATED(ETH_LINK_FULL_DUPLEX) RTE_ETH_LINK_FULL_DUPLEX
388 #define ETH_LINK_DOWN RTE_DEPRECATED(ETH_LINK_DOWN) RTE_ETH_LINK_DOWN
389 #define ETH_LINK_UP RTE_DEPRECATED(ETH_LINK_UP) RTE_ETH_LINK_UP
390 #define ETH_LINK_FIXED RTE_DEPRECATED(ETH_LINK_FIXED) RTE_ETH_LINK_FIXED
391 #define ETH_LINK_AUTONEG RTE_DEPRECATED(ETH_LINK_AUTONEG) RTE_ETH_LINK_AUTONEG
394 * A structure used to configure the ring threshold registers of an Rx/Tx
395 * queue for an Ethernet port.
397 struct rte_eth_thresh {
398 uint8_t pthresh; /**< Ring prefetch threshold. */
399 uint8_t hthresh; /**< Ring host threshold. */
400 uint8_t wthresh; /**< Ring writeback threshold. */
403 /**@{@name Multi-queue mode
404 * @see rte_eth_conf.rxmode.mq_mode.
406 #define RTE_ETH_MQ_RX_RSS_FLAG RTE_BIT32(0) /**< Enable RSS. @see rte_eth_rss_conf */
407 #define RTE_ETH_MQ_RX_DCB_FLAG RTE_BIT32(1) /**< Enable DCB. */
408 #define RTE_ETH_MQ_RX_VMDQ_FLAG RTE_BIT32(2) /**< Enable VMDq. */
411 #define ETH_MQ_RX_RSS_FLAG RTE_DEPRECATED(ETH_MQ_RX_RSS_FLAG) RTE_ETH_MQ_RX_RSS_FLAG
412 #define ETH_MQ_RX_DCB_FLAG RTE_DEPRECATED(ETH_MQ_RX_DCB_FLAG) RTE_ETH_MQ_RX_DCB_FLAG
413 #define ETH_MQ_RX_VMDQ_FLAG RTE_DEPRECATED(ETH_MQ_RX_VMDQ_FLAG) RTE_ETH_MQ_RX_VMDQ_FLAG
416 * A set of values to identify what method is to be used to route
417 * packets to multiple queues.
419 enum rte_eth_rx_mq_mode {
420 /** None of DCB, RSS or VMDq mode */
421 RTE_ETH_MQ_RX_NONE = 0,
423 /** For Rx side, only RSS is on */
424 RTE_ETH_MQ_RX_RSS = RTE_ETH_MQ_RX_RSS_FLAG,
425 /** For Rx side,only DCB is on. */
426 RTE_ETH_MQ_RX_DCB = RTE_ETH_MQ_RX_DCB_FLAG,
427 /** Both DCB and RSS enable */
428 RTE_ETH_MQ_RX_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG,
430 /** Only VMDq, no RSS nor DCB */
431 RTE_ETH_MQ_RX_VMDQ_ONLY = RTE_ETH_MQ_RX_VMDQ_FLAG,
432 /** RSS mode with VMDq */
433 RTE_ETH_MQ_RX_VMDQ_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_VMDQ_FLAG,
434 /** Use VMDq+DCB to route traffic to queues */
435 RTE_ETH_MQ_RX_VMDQ_DCB = RTE_ETH_MQ_RX_VMDQ_FLAG | RTE_ETH_MQ_RX_DCB_FLAG,
436 /** Enable both VMDq and DCB in VMDq */
437 RTE_ETH_MQ_RX_VMDQ_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG |
438 RTE_ETH_MQ_RX_VMDQ_FLAG,
441 #define ETH_MQ_RX_NONE RTE_DEPRECATED(ETH_MQ_RX_NONE) RTE_ETH_MQ_RX_NONE
442 #define ETH_MQ_RX_RSS RTE_DEPRECATED(ETH_MQ_RX_RSS) RTE_ETH_MQ_RX_RSS
443 #define ETH_MQ_RX_DCB RTE_DEPRECATED(ETH_MQ_RX_DCB) RTE_ETH_MQ_RX_DCB
444 #define ETH_MQ_RX_DCB_RSS RTE_DEPRECATED(ETH_MQ_RX_DCB_RSS) RTE_ETH_MQ_RX_DCB_RSS
445 #define ETH_MQ_RX_VMDQ_ONLY RTE_DEPRECATED(ETH_MQ_RX_VMDQ_ONLY) RTE_ETH_MQ_RX_VMDQ_ONLY
446 #define ETH_MQ_RX_VMDQ_RSS RTE_DEPRECATED(ETH_MQ_RX_VMDQ_RSS) RTE_ETH_MQ_RX_VMDQ_RSS
447 #define ETH_MQ_RX_VMDQ_DCB RTE_DEPRECATED(ETH_MQ_RX_VMDQ_DCB) RTE_ETH_MQ_RX_VMDQ_DCB
448 #define ETH_MQ_RX_VMDQ_DCB_RSS RTE_DEPRECATED(ETH_MQ_RX_VMDQ_DCB_RSS) RTE_ETH_MQ_RX_VMDQ_DCB_RSS
451 * A set of values to identify what method is to be used to transmit
452 * packets using multi-TCs.
454 enum rte_eth_tx_mq_mode {
455 RTE_ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
456 RTE_ETH_MQ_TX_DCB, /**< For Tx side,only DCB is on. */
457 RTE_ETH_MQ_TX_VMDQ_DCB, /**< For Tx side,both DCB and VT is on. */
458 RTE_ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
461 #define ETH_MQ_TX_NONE RTE_DEPRECATED(ETH_MQ_TX_NONE) RTE_ETH_MQ_TX_NONE
462 #define ETH_MQ_TX_DCB RTE_DEPRECATED(ETH_MQ_TX_DCB) RTE_ETH_MQ_TX_DCB
463 #define ETH_MQ_TX_VMDQ_DCB RTE_DEPRECATED(ETH_MQ_TX_VMDQ_DCB) RTE_ETH_MQ_TX_VMDQ_DCB
464 #define ETH_MQ_TX_VMDQ_ONLY RTE_DEPRECATED(ETH_MQ_TX_VMDQ_ONLY) RTE_ETH_MQ_TX_VMDQ_ONLY
467 * A structure used to configure the Rx features of an Ethernet port.
469 struct rte_eth_rxmode {
470 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
471 enum rte_eth_rx_mq_mode mq_mode;
472 uint32_t mtu; /**< Requested MTU. */
473 /** Maximum allowed size of LRO aggregated packet. */
474 uint32_t max_lro_pkt_size;
475 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
477 * Per-port Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags.
478 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
479 * structure are allowed to be set.
483 uint64_t reserved_64s[2]; /**< Reserved for future fields */
484 void *reserved_ptrs[2]; /**< Reserved for future fields */
488 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
489 * Note that single VLAN is treated the same as inner VLAN.
492 RTE_ETH_VLAN_TYPE_UNKNOWN = 0,
493 RTE_ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
494 RTE_ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
495 RTE_ETH_VLAN_TYPE_MAX,
498 #define ETH_VLAN_TYPE_UNKNOWN RTE_DEPRECATED(ETH_VLAN_TYPE_UNKNOWN) RTE_ETH_VLAN_TYPE_UNKNOWN
499 #define ETH_VLAN_TYPE_INNER RTE_DEPRECATED(ETH_VLAN_TYPE_INNER) RTE_ETH_VLAN_TYPE_INNER
500 #define ETH_VLAN_TYPE_OUTER RTE_DEPRECATED(ETH_VLAN_TYPE_OUTER) RTE_ETH_VLAN_TYPE_OUTER
501 #define ETH_VLAN_TYPE_MAX RTE_DEPRECATED(ETH_VLAN_TYPE_MAX) RTE_ETH_VLAN_TYPE_MAX
504 * A structure used to describe a VLAN filter.
505 * If the bit corresponding to a VID is set, such VID is on.
507 struct rte_vlan_filter_conf {
512 * A structure used to configure the Receive Side Scaling (RSS) feature
513 * of an Ethernet port.
514 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
515 * to an array holding the RSS key to use for hashing specific header
516 * fields of received packets. The length of this array should be indicated
517 * by *rss_key_len* below. Otherwise, a default random hash key is used by
520 * The *rss_key_len* field of the *rss_conf* structure indicates the length
521 * in bytes of the array pointed by *rss_key*. To be compatible, this length
522 * will be checked in i40e only. Others assume 40 bytes to be used as before.
524 * The *rss_hf* field of the *rss_conf* structure indicates the different
525 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
526 * Supplying an *rss_hf* equal to zero disables the RSS feature.
528 struct rte_eth_rss_conf {
529 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
530 uint8_t rss_key_len; /**< hash key length in bytes. */
531 uint64_t rss_hf; /**< Hash functions to apply - see below. */
535 * A packet can be identified by hardware as different flow types. Different
536 * NIC hardware may support different flow types.
537 * Basically, the NIC hardware identifies the flow type as deep protocol as
538 * possible, and exclusively. For example, if a packet is identified as
539 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types,
540 * though it is an actual IPV4 packet.
542 #define RTE_ETH_FLOW_UNKNOWN 0
543 #define RTE_ETH_FLOW_RAW 1
544 #define RTE_ETH_FLOW_IPV4 2
545 #define RTE_ETH_FLOW_FRAG_IPV4 3
546 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4
547 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5
548 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6
549 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7
550 #define RTE_ETH_FLOW_IPV6 8
551 #define RTE_ETH_FLOW_FRAG_IPV6 9
552 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10
553 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11
554 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12
555 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13
556 #define RTE_ETH_FLOW_L2_PAYLOAD 14
557 #define RTE_ETH_FLOW_IPV6_EX 15
558 #define RTE_ETH_FLOW_IPV6_TCP_EX 16
559 #define RTE_ETH_FLOW_IPV6_UDP_EX 17
560 /** Consider device port number as a flow differentiator */
561 #define RTE_ETH_FLOW_PORT 18
562 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */
563 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */
564 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */
565 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */
566 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */
567 #define RTE_ETH_FLOW_MAX 24
570 * Below macros are defined for RSS offload types, they can be used to
571 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types.
573 #define RTE_ETH_RSS_IPV4 RTE_BIT64(2)
574 #define RTE_ETH_RSS_FRAG_IPV4 RTE_BIT64(3)
575 #define RTE_ETH_RSS_NONFRAG_IPV4_TCP RTE_BIT64(4)
576 #define RTE_ETH_RSS_NONFRAG_IPV4_UDP RTE_BIT64(5)
577 #define RTE_ETH_RSS_NONFRAG_IPV4_SCTP RTE_BIT64(6)
578 #define RTE_ETH_RSS_NONFRAG_IPV4_OTHER RTE_BIT64(7)
579 #define RTE_ETH_RSS_IPV6 RTE_BIT64(8)
580 #define RTE_ETH_RSS_FRAG_IPV6 RTE_BIT64(9)
581 #define RTE_ETH_RSS_NONFRAG_IPV6_TCP RTE_BIT64(10)
582 #define RTE_ETH_RSS_NONFRAG_IPV6_UDP RTE_BIT64(11)
583 #define RTE_ETH_RSS_NONFRAG_IPV6_SCTP RTE_BIT64(12)
584 #define RTE_ETH_RSS_NONFRAG_IPV6_OTHER RTE_BIT64(13)
585 #define RTE_ETH_RSS_L2_PAYLOAD RTE_BIT64(14)
586 #define RTE_ETH_RSS_IPV6_EX RTE_BIT64(15)
587 #define RTE_ETH_RSS_IPV6_TCP_EX RTE_BIT64(16)
588 #define RTE_ETH_RSS_IPV6_UDP_EX RTE_BIT64(17)
589 #define RTE_ETH_RSS_PORT RTE_BIT64(18)
590 #define RTE_ETH_RSS_VXLAN RTE_BIT64(19)
591 #define RTE_ETH_RSS_GENEVE RTE_BIT64(20)
592 #define RTE_ETH_RSS_NVGRE RTE_BIT64(21)
593 #define RTE_ETH_RSS_GTPU RTE_BIT64(23)
594 #define RTE_ETH_RSS_ETH RTE_BIT64(24)
595 #define RTE_ETH_RSS_S_VLAN RTE_BIT64(25)
596 #define RTE_ETH_RSS_C_VLAN RTE_BIT64(26)
597 #define RTE_ETH_RSS_ESP RTE_BIT64(27)
598 #define RTE_ETH_RSS_AH RTE_BIT64(28)
599 #define RTE_ETH_RSS_L2TPV3 RTE_BIT64(29)
600 #define RTE_ETH_RSS_PFCP RTE_BIT64(30)
601 #define RTE_ETH_RSS_PPPOE RTE_BIT64(31)
602 #define RTE_ETH_RSS_ECPRI RTE_BIT64(32)
603 #define RTE_ETH_RSS_MPLS RTE_BIT64(33)
604 #define RTE_ETH_RSS_IPV4_CHKSUM RTE_BIT64(34)
606 #define ETH_RSS_IPV4 RTE_DEPRECATED(ETH_RSS_IPV4) RTE_ETH_RSS_IPV4
607 #define ETH_RSS_FRAG_IPV4 RTE_DEPRECATED(ETH_RSS_FRAG_IPV4) RTE_ETH_RSS_FRAG_IPV4
608 #define ETH_RSS_NONFRAG_IPV4_TCP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_TCP) RTE_ETH_RSS_NONFRAG_IPV4_TCP
609 #define ETH_RSS_NONFRAG_IPV4_UDP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_UDP) RTE_ETH_RSS_NONFRAG_IPV4_UDP
610 #define ETH_RSS_NONFRAG_IPV4_SCTP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_SCTP) RTE_ETH_RSS_NONFRAG_IPV4_SCTP
611 #define ETH_RSS_NONFRAG_IPV4_OTHER RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV4_OTHER) RTE_ETH_RSS_NONFRAG_IPV4_OTHER
612 #define ETH_RSS_IPV6 RTE_DEPRECATED(ETH_RSS_IPV6) RTE_ETH_RSS_IPV6
613 #define ETH_RSS_FRAG_IPV6 RTE_DEPRECATED(ETH_RSS_FRAG_IPV6) RTE_ETH_RSS_FRAG_IPV6
614 #define ETH_RSS_NONFRAG_IPV6_TCP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_TCP) RTE_ETH_RSS_NONFRAG_IPV6_TCP
615 #define ETH_RSS_NONFRAG_IPV6_UDP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_UDP) RTE_ETH_RSS_NONFRAG_IPV6_UDP
616 #define ETH_RSS_NONFRAG_IPV6_SCTP RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_SCTP) RTE_ETH_RSS_NONFRAG_IPV6_SCTP
617 #define ETH_RSS_NONFRAG_IPV6_OTHER RTE_DEPRECATED(ETH_RSS_NONFRAG_IPV6_OTHER) RTE_ETH_RSS_NONFRAG_IPV6_OTHER
618 #define ETH_RSS_L2_PAYLOAD RTE_DEPRECATED(ETH_RSS_L2_PAYLOAD) RTE_ETH_RSS_L2_PAYLOAD
619 #define ETH_RSS_IPV6_EX RTE_DEPRECATED(ETH_RSS_IPV6_EX) RTE_ETH_RSS_IPV6_EX
620 #define ETH_RSS_IPV6_TCP_EX RTE_DEPRECATED(ETH_RSS_IPV6_TCP_EX) RTE_ETH_RSS_IPV6_TCP_EX
621 #define ETH_RSS_IPV6_UDP_EX RTE_DEPRECATED(ETH_RSS_IPV6_UDP_EX) RTE_ETH_RSS_IPV6_UDP_EX
622 #define ETH_RSS_PORT RTE_DEPRECATED(ETH_RSS_PORT) RTE_ETH_RSS_PORT
623 #define ETH_RSS_VXLAN RTE_DEPRECATED(ETH_RSS_VXLAN) RTE_ETH_RSS_VXLAN
624 #define ETH_RSS_GENEVE RTE_DEPRECATED(ETH_RSS_GENEVE) RTE_ETH_RSS_GENEVE
625 #define ETH_RSS_NVGRE RTE_DEPRECATED(ETH_RSS_NVGRE) RTE_ETH_RSS_NVGRE
626 #define ETH_RSS_GTPU RTE_DEPRECATED(ETH_RSS_GTPU) RTE_ETH_RSS_GTPU
627 #define ETH_RSS_ETH RTE_DEPRECATED(ETH_RSS_ETH) RTE_ETH_RSS_ETH
628 #define ETH_RSS_S_VLAN RTE_DEPRECATED(ETH_RSS_S_VLAN) RTE_ETH_RSS_S_VLAN
629 #define ETH_RSS_C_VLAN RTE_DEPRECATED(ETH_RSS_C_VLAN) RTE_ETH_RSS_C_VLAN
630 #define ETH_RSS_ESP RTE_DEPRECATED(ETH_RSS_ESP) RTE_ETH_RSS_ESP
631 #define ETH_RSS_AH RTE_DEPRECATED(ETH_RSS_AH) RTE_ETH_RSS_AH
632 #define ETH_RSS_L2TPV3 RTE_DEPRECATED(ETH_RSS_L2TPV3) RTE_ETH_RSS_L2TPV3
633 #define ETH_RSS_PFCP RTE_DEPRECATED(ETH_RSS_PFCP) RTE_ETH_RSS_PFCP
634 #define ETH_RSS_PPPOE RTE_DEPRECATED(ETH_RSS_PPPOE) RTE_ETH_RSS_PPPOE
635 #define ETH_RSS_ECPRI RTE_DEPRECATED(ETH_RSS_ECPRI) RTE_ETH_RSS_ECPRI
636 #define ETH_RSS_MPLS RTE_DEPRECATED(ETH_RSS_MPLS) RTE_ETH_RSS_MPLS
637 #define ETH_RSS_IPV4_CHKSUM RTE_DEPRECATED(ETH_RSS_IPV4_CHKSUM) RTE_ETH_RSS_IPV4_CHKSUM
640 * The ETH_RSS_L4_CHKSUM works on checksum field of any L4 header.
641 * It is similar to ETH_RSS_PORT that they don't specify the specific type of
642 * L4 header. This macro is defined to replace some specific L4 (TCP/UDP/SCTP)
643 * checksum type for constructing the use of RSS offload bits.
645 * Due to above reason, some old APIs (and configuration) don't support
646 * RTE_ETH_RSS_L4_CHKSUM. The rte_flow RSS API supports it.
648 * For the case that checksum is not used in an UDP header,
649 * it takes the reserved value 0 as input for the hash function.
651 #define RTE_ETH_RSS_L4_CHKSUM RTE_BIT64(35)
652 #define ETH_RSS_L4_CHKSUM RTE_DEPRECATED(ETH_RSS_L4_CHKSUM) RTE_ETH_RSS_L4_CHKSUM
655 * We use the following macros to combine with above RTE_ETH_RSS_* for
656 * more specific input set selection. These bits are defined starting
657 * from the high end of the 64 bits.
658 * Note: If we use above RTE_ETH_RSS_* without SRC/DST_ONLY, it represents
659 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of
660 * the same level are used simultaneously, it is the same case as none of
663 #define RTE_ETH_RSS_L3_SRC_ONLY RTE_BIT64(63)
664 #define RTE_ETH_RSS_L3_DST_ONLY RTE_BIT64(62)
665 #define RTE_ETH_RSS_L4_SRC_ONLY RTE_BIT64(61)
666 #define RTE_ETH_RSS_L4_DST_ONLY RTE_BIT64(60)
667 #define RTE_ETH_RSS_L2_SRC_ONLY RTE_BIT64(59)
668 #define RTE_ETH_RSS_L2_DST_ONLY RTE_BIT64(58)
670 #define ETH_RSS_L3_SRC_ONLY RTE_DEPRECATED(ETH_RSS_L3_SRC_ONLY) RTE_ETH_RSS_L3_SRC_ONLY
671 #define ETH_RSS_L3_DST_ONLY RTE_DEPRECATED(ETH_RSS_L3_DST_ONLY) RTE_ETH_RSS_L3_DST_ONLY
672 #define ETH_RSS_L4_SRC_ONLY RTE_DEPRECATED(ETH_RSS_L4_SRC_ONLY) RTE_ETH_RSS_L4_SRC_ONLY
673 #define ETH_RSS_L4_DST_ONLY RTE_DEPRECATED(ETH_RSS_L4_DST_ONLY) RTE_ETH_RSS_L4_DST_ONLY
674 #define ETH_RSS_L2_SRC_ONLY RTE_DEPRECATED(ETH_RSS_L2_SRC_ONLY) RTE_ETH_RSS_L2_SRC_ONLY
675 #define ETH_RSS_L2_DST_ONLY RTE_DEPRECATED(ETH_RSS_L2_DST_ONLY) RTE_ETH_RSS_L2_DST_ONLY
678 * Only select IPV6 address prefix as RSS input set according to
679 * https://tools.ietf.org/html/rfc6052
680 * Must be combined with RTE_ETH_RSS_IPV6, RTE_ETH_RSS_NONFRAG_IPV6_UDP,
681 * RTE_ETH_RSS_NONFRAG_IPV6_TCP, RTE_ETH_RSS_NONFRAG_IPV6_SCTP.
683 #define RTE_ETH_RSS_L3_PRE32 RTE_BIT64(57)
684 #define RTE_ETH_RSS_L3_PRE40 RTE_BIT64(56)
685 #define RTE_ETH_RSS_L3_PRE48 RTE_BIT64(55)
686 #define RTE_ETH_RSS_L3_PRE56 RTE_BIT64(54)
687 #define RTE_ETH_RSS_L3_PRE64 RTE_BIT64(53)
688 #define RTE_ETH_RSS_L3_PRE96 RTE_BIT64(52)
691 * Use the following macros to combine with the above layers
692 * to choose inner and outer layers or both for RSS computation.
693 * Bits 50 and 51 are reserved for this.
697 * level 0, requests the default behavior.
698 * Depending on the packet type, it can mean outermost, innermost,
699 * anything in between or even no RSS.
700 * It basically stands for the innermost encapsulation level RSS
701 * can be performed on according to PMD and device capabilities.
703 #define RTE_ETH_RSS_LEVEL_PMD_DEFAULT (UINT64_C(0) << 50)
704 #define ETH_RSS_LEVEL_PMD_DEFAULT RTE_DEPRECATED(ETH_RSS_LEVEL_PMD_DEFAULT) RTE_ETH_RSS_LEVEL_PMD_DEFAULT
707 * level 1, requests RSS to be performed on the outermost packet
708 * encapsulation level.
710 #define RTE_ETH_RSS_LEVEL_OUTERMOST (UINT64_C(1) << 50)
711 #define ETH_RSS_LEVEL_OUTERMOST RTE_DEPRECATED(ETH_RSS_LEVEL_OUTERMOST) RTE_ETH_RSS_LEVEL_OUTERMOST
714 * level 2, requests RSS to be performed on the specified inner packet
715 * encapsulation level, from outermost to innermost (lower to higher values).
717 #define RTE_ETH_RSS_LEVEL_INNERMOST (UINT64_C(2) << 50)
718 #define RTE_ETH_RSS_LEVEL_MASK (UINT64_C(3) << 50)
720 #define ETH_RSS_LEVEL_INNERMOST RTE_DEPRECATED(ETH_RSS_LEVEL_INNERMOST) RTE_ETH_RSS_LEVEL_INNERMOST
721 #define ETH_RSS_LEVEL_MASK RTE_DEPRECATED(ETH_RSS_LEVEL_MASK) RTE_ETH_RSS_LEVEL_MASK
723 #define RTE_ETH_RSS_LEVEL(rss_hf) ((rss_hf & RTE_ETH_RSS_LEVEL_MASK) >> 50)
724 #define ETH_RSS_LEVEL(rss_hf) RTE_DEPRECATED(ETH_RSS_LEVEL(rss_hf)) RTE_ETH_RSS_LEVEL(rss_hf)
727 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of
728 * the same level are used simultaneously, it is the same case as
729 * none of them are added.
732 * RSS types with SRC/DST_ONLY.
736 static inline uint64_t
737 rte_eth_rss_hf_refine(uint64_t rss_hf)
739 if ((rss_hf & RTE_ETH_RSS_L3_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L3_DST_ONLY))
740 rss_hf &= ~(RTE_ETH_RSS_L3_SRC_ONLY | RTE_ETH_RSS_L3_DST_ONLY);
742 if ((rss_hf & RTE_ETH_RSS_L4_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L4_DST_ONLY))
743 rss_hf &= ~(RTE_ETH_RSS_L4_SRC_ONLY | RTE_ETH_RSS_L4_DST_ONLY);
748 #define RTE_ETH_RSS_IPV6_PRE32 ( \
750 RTE_ETH_RSS_L3_PRE32)
751 #define ETH_RSS_IPV6_PRE32 RTE_DEPRECATED(ETH_RSS_IPV6_PRE32) RTE_ETH_RSS_IPV6_PRE32
753 #define RTE_ETH_RSS_IPV6_PRE40 ( \
755 RTE_ETH_RSS_L3_PRE40)
756 #define ETH_RSS_IPV6_PRE40 RTE_DEPRECATED(ETH_RSS_IPV6_PRE40) RTE_ETH_RSS_IPV6_PRE40
758 #define RTE_ETH_RSS_IPV6_PRE48 ( \
760 RTE_ETH_RSS_L3_PRE48)
761 #define ETH_RSS_IPV6_PRE48 RTE_DEPRECATED(ETH_RSS_IPV6_PRE48) RTE_ETH_RSS_IPV6_PRE48
763 #define RTE_ETH_RSS_IPV6_PRE56 ( \
765 RTE_ETH_RSS_L3_PRE56)
766 #define ETH_RSS_IPV6_PRE56 RTE_DEPRECATED(ETH_RSS_IPV6_PRE56) RTE_ETH_RSS_IPV6_PRE56
768 #define RTE_ETH_RSS_IPV6_PRE64 ( \
770 RTE_ETH_RSS_L3_PRE64)
771 #define ETH_RSS_IPV6_PRE64 RTE_DEPRECATED(ETH_RSS_IPV6_PRE64) RTE_ETH_RSS_IPV6_PRE64
773 #define RTE_ETH_RSS_IPV6_PRE96 ( \
775 RTE_ETH_RSS_L3_PRE96)
776 #define ETH_RSS_IPV6_PRE96 RTE_DEPRECATED(ETH_RSS_IPV6_PRE96) RTE_ETH_RSS_IPV6_PRE96
778 #define RTE_ETH_RSS_IPV6_PRE32_UDP ( \
779 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
780 RTE_ETH_RSS_L3_PRE32)
781 #define ETH_RSS_IPV6_PRE32_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE32_UDP) RTE_ETH_RSS_IPV6_PRE32_UDP
783 #define RTE_ETH_RSS_IPV6_PRE40_UDP ( \
784 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
785 RTE_ETH_RSS_L3_PRE40)
786 #define ETH_RSS_IPV6_PRE40_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE40_UDP) RTE_ETH_RSS_IPV6_PRE40_UDP
788 #define RTE_ETH_RSS_IPV6_PRE48_UDP ( \
789 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
790 RTE_ETH_RSS_L3_PRE48)
791 #define ETH_RSS_IPV6_PRE48_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE48_UDP) RTE_ETH_RSS_IPV6_PRE48_UDP
793 #define RTE_ETH_RSS_IPV6_PRE56_UDP ( \
794 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
795 RTE_ETH_RSS_L3_PRE56)
796 #define ETH_RSS_IPV6_PRE56_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE56_UDP) RTE_ETH_RSS_IPV6_PRE56_UDP
798 #define RTE_ETH_RSS_IPV6_PRE64_UDP ( \
799 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
800 RTE_ETH_RSS_L3_PRE64)
801 #define ETH_RSS_IPV6_PRE64_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE64_UDP) RTE_ETH_RSS_IPV6_PRE64_UDP
803 #define RTE_ETH_RSS_IPV6_PRE96_UDP ( \
804 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
805 RTE_ETH_RSS_L3_PRE96)
806 #define ETH_RSS_IPV6_PRE96_UDP RTE_DEPRECATED(ETH_RSS_IPV6_PRE96_UDP) RTE_ETH_RSS_IPV6_PRE96_UDP
808 #define RTE_ETH_RSS_IPV6_PRE32_TCP ( \
809 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
810 RTE_ETH_RSS_L3_PRE32)
811 #define ETH_RSS_IPV6_PRE32_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE32_TCP) RTE_ETH_RSS_IPV6_PRE32_TCP
813 #define RTE_ETH_RSS_IPV6_PRE40_TCP ( \
814 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
815 RTE_ETH_RSS_L3_PRE40)
816 #define ETH_RSS_IPV6_PRE40_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE40_TCP) RTE_ETH_RSS_IPV6_PRE40_TCP
818 #define RTE_ETH_RSS_IPV6_PRE48_TCP ( \
819 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
820 RTE_ETH_RSS_L3_PRE48)
821 #define ETH_RSS_IPV6_PRE48_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE48_TCP) RTE_ETH_RSS_IPV6_PRE48_TCP
823 #define RTE_ETH_RSS_IPV6_PRE56_TCP ( \
824 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
825 RTE_ETH_RSS_L3_PRE56)
826 #define ETH_RSS_IPV6_PRE56_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE56_TCP) RTE_ETH_RSS_IPV6_PRE56_TCP
828 #define RTE_ETH_RSS_IPV6_PRE64_TCP ( \
829 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
830 RTE_ETH_RSS_L3_PRE64)
831 #define ETH_RSS_IPV6_PRE64_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE64_TCP) RTE_ETH_RSS_IPV6_PRE64_TCP
833 #define RTE_ETH_RSS_IPV6_PRE96_TCP ( \
834 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
835 RTE_ETH_RSS_L3_PRE96)
836 #define ETH_RSS_IPV6_PRE96_TCP RTE_DEPRECATED(ETH_RSS_IPV6_PRE96_TCP) RTE_ETH_RSS_IPV6_PRE96_TCP
838 #define RTE_ETH_RSS_IPV6_PRE32_SCTP ( \
839 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
840 RTE_ETH_RSS_L3_PRE32)
841 #define ETH_RSS_IPV6_PRE32_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE32_SCTP) RTE_ETH_RSS_IPV6_PRE32_SCTP
843 #define RTE_ETH_RSS_IPV6_PRE40_SCTP ( \
844 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
845 RTE_ETH_RSS_L3_PRE40)
846 #define ETH_RSS_IPV6_PRE40_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE40_SCTP) RTE_ETH_RSS_IPV6_PRE40_SCTP
848 #define RTE_ETH_RSS_IPV6_PRE48_SCTP ( \
849 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
850 RTE_ETH_RSS_L3_PRE48)
851 #define ETH_RSS_IPV6_PRE48_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE48_SCTP) RTE_ETH_RSS_IPV6_PRE48_SCTP
853 #define RTE_ETH_RSS_IPV6_PRE56_SCTP ( \
854 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
855 RTE_ETH_RSS_L3_PRE56)
856 #define ETH_RSS_IPV6_PRE56_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE56_SCTP) RTE_ETH_RSS_IPV6_PRE56_SCTP
858 #define RTE_ETH_RSS_IPV6_PRE64_SCTP ( \
859 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
860 RTE_ETH_RSS_L3_PRE64)
861 #define ETH_RSS_IPV6_PRE64_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE64_SCTP) RTE_ETH_RSS_IPV6_PRE64_SCTP
863 #define RTE_ETH_RSS_IPV6_PRE96_SCTP ( \
864 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
865 RTE_ETH_RSS_L3_PRE96)
866 #define ETH_RSS_IPV6_PRE96_SCTP RTE_DEPRECATED(ETH_RSS_IPV6_PRE96_SCTP) RTE_ETH_RSS_IPV6_PRE96_SCTP
868 #define RTE_ETH_RSS_IP ( \
870 RTE_ETH_RSS_FRAG_IPV4 | \
871 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \
873 RTE_ETH_RSS_FRAG_IPV6 | \
874 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \
876 #define ETH_RSS_IP RTE_DEPRECATED(ETH_RSS_IP) RTE_ETH_RSS_IP
878 #define RTE_ETH_RSS_UDP ( \
879 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \
880 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
881 RTE_ETH_RSS_IPV6_UDP_EX)
882 #define ETH_RSS_UDP RTE_DEPRECATED(ETH_RSS_UDP) RTE_ETH_RSS_UDP
884 #define RTE_ETH_RSS_TCP ( \
885 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \
886 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
887 RTE_ETH_RSS_IPV6_TCP_EX)
888 #define ETH_RSS_TCP RTE_DEPRECATED(ETH_RSS_TCP) RTE_ETH_RSS_TCP
890 #define RTE_ETH_RSS_SCTP ( \
891 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \
892 RTE_ETH_RSS_NONFRAG_IPV6_SCTP)
893 #define ETH_RSS_SCTP RTE_DEPRECATED(ETH_RSS_SCTP) RTE_ETH_RSS_SCTP
895 #define RTE_ETH_RSS_TUNNEL ( \
896 RTE_ETH_RSS_VXLAN | \
897 RTE_ETH_RSS_GENEVE | \
899 #define ETH_RSS_TUNNEL RTE_DEPRECATED(ETH_RSS_TUNNEL) RTE_ETH_RSS_TUNNEL
901 #define RTE_ETH_RSS_VLAN ( \
902 RTE_ETH_RSS_S_VLAN | \
904 #define ETH_RSS_VLAN RTE_DEPRECATED(ETH_RSS_VLAN) RTE_ETH_RSS_VLAN
906 /** Mask of valid RSS hash protocols */
907 #define RTE_ETH_RSS_PROTO_MASK ( \
909 RTE_ETH_RSS_FRAG_IPV4 | \
910 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \
911 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \
912 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \
913 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \
915 RTE_ETH_RSS_FRAG_IPV6 | \
916 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \
917 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \
918 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \
919 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \
920 RTE_ETH_RSS_L2_PAYLOAD | \
921 RTE_ETH_RSS_IPV6_EX | \
922 RTE_ETH_RSS_IPV6_TCP_EX | \
923 RTE_ETH_RSS_IPV6_UDP_EX | \
925 RTE_ETH_RSS_VXLAN | \
926 RTE_ETH_RSS_GENEVE | \
927 RTE_ETH_RSS_NVGRE | \
929 #define ETH_RSS_PROTO_MASK RTE_DEPRECATED(ETH_RSS_PROTO_MASK) RTE_ETH_RSS_PROTO_MASK
932 * Definitions used for redirection table entry size.
933 * Some RSS RETA sizes may not be supported by some drivers, check the
934 * documentation or the description of relevant functions for more details.
936 #define RTE_ETH_RSS_RETA_SIZE_64 64
937 #define RTE_ETH_RSS_RETA_SIZE_128 128
938 #define RTE_ETH_RSS_RETA_SIZE_256 256
939 #define RTE_ETH_RSS_RETA_SIZE_512 512
940 #define RTE_ETH_RETA_GROUP_SIZE 64
942 #define ETH_RSS_RETA_SIZE_64 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_64) RTE_ETH_RSS_RETA_SIZE_64
943 #define ETH_RSS_RETA_SIZE_128 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_128) RTE_ETH_RSS_RETA_SIZE_128
944 #define ETH_RSS_RETA_SIZE_256 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_256) RTE_ETH_RSS_RETA_SIZE_256
945 #define ETH_RSS_RETA_SIZE_512 RTE_DEPRECATED(ETH_RSS_RETA_SIZE_512) RTE_ETH_RSS_RETA_SIZE_512
946 #define RTE_RETA_GROUP_SIZE RTE_DEPRECATED(RTE_RETA_GROUP_SIZE) RTE_ETH_RETA_GROUP_SIZE
948 /**@{@name VMDq and DCB maximums */
949 #define RTE_ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDq VLAN filters. */
950 #define RTE_ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
951 #define RTE_ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDq DCB queues. */
952 #define RTE_ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
955 #define ETH_VMDQ_MAX_VLAN_FILTERS RTE_DEPRECATED(ETH_VMDQ_MAX_VLAN_FILTERS) RTE_ETH_VMDQ_MAX_VLAN_FILTERS
956 #define ETH_DCB_NUM_USER_PRIORITIES RTE_DEPRECATED(ETH_DCB_NUM_USER_PRIORITIES) RTE_ETH_DCB_NUM_USER_PRIORITIES
957 #define ETH_VMDQ_DCB_NUM_QUEUES RTE_DEPRECATED(ETH_VMDQ_DCB_NUM_QUEUES) RTE_ETH_VMDQ_DCB_NUM_QUEUES
958 #define ETH_DCB_NUM_QUEUES RTE_DEPRECATED(ETH_DCB_NUM_QUEUES) RTE_ETH_DCB_NUM_QUEUES
960 /**@{@name DCB capabilities */
961 #define RTE_ETH_DCB_PG_SUPPORT RTE_BIT32(0) /**< Priority Group(ETS) support. */
962 #define RTE_ETH_DCB_PFC_SUPPORT RTE_BIT32(1) /**< Priority Flow Control support. */
965 #define ETH_DCB_PG_SUPPORT RTE_DEPRECATED(ETH_DCB_PG_SUPPORT) RTE_ETH_DCB_PG_SUPPORT
966 #define ETH_DCB_PFC_SUPPORT RTE_DEPRECATED(ETH_DCB_PFC_SUPPORT) RTE_ETH_DCB_PFC_SUPPORT
968 /**@{@name VLAN offload bits */
969 #define RTE_ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
970 #define RTE_ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
971 #define RTE_ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
972 #define RTE_ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */
974 #define ETH_VLAN_STRIP_OFFLOAD RTE_DEPRECATED(ETH_VLAN_STRIP_OFFLOAD) RTE_ETH_VLAN_STRIP_OFFLOAD
975 #define ETH_VLAN_FILTER_OFFLOAD RTE_DEPRECATED(ETH_VLAN_FILTER_OFFLOAD) RTE_ETH_VLAN_FILTER_OFFLOAD
976 #define ETH_VLAN_EXTEND_OFFLOAD RTE_DEPRECATED(ETH_VLAN_EXTEND_OFFLOAD) RTE_ETH_VLAN_EXTEND_OFFLOAD
977 #define ETH_QINQ_STRIP_OFFLOAD RTE_DEPRECATED(ETH_QINQ_STRIP_OFFLOAD) RTE_ETH_QINQ_STRIP_OFFLOAD
979 #define RTE_ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
980 #define RTE_ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
981 #define RTE_ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
982 #define RTE_ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */
983 #define RTE_ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
986 #define ETH_VLAN_STRIP_MASK RTE_DEPRECATED(ETH_VLAN_STRIP_MASK) RTE_ETH_VLAN_STRIP_MASK
987 #define ETH_VLAN_FILTER_MASK RTE_DEPRECATED(ETH_VLAN_FILTER_MASK) RTE_ETH_VLAN_FILTER_MASK
988 #define ETH_VLAN_EXTEND_MASK RTE_DEPRECATED(ETH_VLAN_EXTEND_MASK) RTE_ETH_VLAN_EXTEND_MASK
989 #define ETH_QINQ_STRIP_MASK RTE_DEPRECATED(ETH_QINQ_STRIP_MASK) RTE_ETH_QINQ_STRIP_MASK
990 #define ETH_VLAN_ID_MAX RTE_DEPRECATED(ETH_VLAN_ID_MAX) RTE_ETH_VLAN_ID_MAX
992 /* Definitions used for receive MAC address */
993 #define RTE_ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
994 #define ETH_NUM_RECEIVE_MAC_ADDR RTE_DEPRECATED(ETH_NUM_RECEIVE_MAC_ADDR) RTE_ETH_NUM_RECEIVE_MAC_ADDR
996 /* Definitions used for unicast hash */
997 #define RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
998 #define ETH_VMDQ_NUM_UC_HASH_ARRAY RTE_DEPRECATED(ETH_VMDQ_NUM_UC_HASH_ARRAY) RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY
1000 /**@{@name VMDq Rx mode
1001 * @see rte_eth_vmdq_rx_conf.rx_mode
1003 /** Accept untagged packets. */
1004 #define RTE_ETH_VMDQ_ACCEPT_UNTAG RTE_BIT32(0)
1005 /** Accept packets in multicast table. */
1006 #define RTE_ETH_VMDQ_ACCEPT_HASH_MC RTE_BIT32(1)
1007 /** Accept packets in unicast table. */
1008 #define RTE_ETH_VMDQ_ACCEPT_HASH_UC RTE_BIT32(2)
1009 /** Accept broadcast packets. */
1010 #define RTE_ETH_VMDQ_ACCEPT_BROADCAST RTE_BIT32(3)
1011 /** Multicast promiscuous. */
1012 #define RTE_ETH_VMDQ_ACCEPT_MULTICAST RTE_BIT32(4)
1015 #define ETH_VMDQ_ACCEPT_UNTAG RTE_DEPRECATED(ETH_VMDQ_ACCEPT_UNTAG) RTE_ETH_VMDQ_ACCEPT_UNTAG
1016 #define ETH_VMDQ_ACCEPT_HASH_MC RTE_DEPRECATED(ETH_VMDQ_ACCEPT_HASH_MC) RTE_ETH_VMDQ_ACCEPT_HASH_MC
1017 #define ETH_VMDQ_ACCEPT_HASH_UC RTE_DEPRECATED(ETH_VMDQ_ACCEPT_HASH_UC) RTE_ETH_VMDQ_ACCEPT_HASH_UC
1018 #define ETH_VMDQ_ACCEPT_BROADCAST RTE_DEPRECATED(ETH_VMDQ_ACCEPT_BROADCAST) RTE_ETH_VMDQ_ACCEPT_BROADCAST
1019 #define ETH_VMDQ_ACCEPT_MULTICAST RTE_DEPRECATED(ETH_VMDQ_ACCEPT_MULTICAST) RTE_ETH_VMDQ_ACCEPT_MULTICAST
1022 * A structure used to configure 64 entries of Redirection Table of the
1023 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
1024 * more than 64 entries supported by hardware, an array of this structure
1027 struct rte_eth_rss_reta_entry64 {
1028 /** Mask bits indicate which entries need to be updated/queried. */
1030 /** Group of 64 redirection table entries. */
1031 uint16_t reta[RTE_ETH_RETA_GROUP_SIZE];
1035 * This enum indicates the possible number of traffic classes
1036 * in DCB configurations
1038 enum rte_eth_nb_tcs {
1039 RTE_ETH_4_TCS = 4, /**< 4 TCs with DCB. */
1040 RTE_ETH_8_TCS = 8 /**< 8 TCs with DCB. */
1042 #define ETH_4_TCS RTE_DEPRECATED(ETH_4_TCS) RTE_ETH_4_TCS
1043 #define ETH_8_TCS RTE_DEPRECATED(ETH_8_TCS) RTE_ETH_8_TCS
1046 * This enum indicates the possible number of queue pools
1047 * in VMDq configurations.
1049 enum rte_eth_nb_pools {
1050 RTE_ETH_8_POOLS = 8, /**< 8 VMDq pools. */
1051 RTE_ETH_16_POOLS = 16, /**< 16 VMDq pools. */
1052 RTE_ETH_32_POOLS = 32, /**< 32 VMDq pools. */
1053 RTE_ETH_64_POOLS = 64 /**< 64 VMDq pools. */
1055 #define ETH_8_POOLS RTE_DEPRECATED(ETH_8_POOLS) RTE_ETH_8_POOLS
1056 #define ETH_16_POOLS RTE_DEPRECATED(ETH_16_POOLS) RTE_ETH_16_POOLS
1057 #define ETH_32_POOLS RTE_DEPRECATED(ETH_32_POOLS) RTE_ETH_32_POOLS
1058 #define ETH_64_POOLS RTE_DEPRECATED(ETH_64_POOLS) RTE_ETH_64_POOLS
1060 /* This structure may be extended in future. */
1061 struct rte_eth_dcb_rx_conf {
1062 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
1063 /** Traffic class each UP mapped to. */
1064 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1067 struct rte_eth_vmdq_dcb_tx_conf {
1068 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
1069 /** Traffic class each UP mapped to. */
1070 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1073 struct rte_eth_dcb_tx_conf {
1074 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
1075 /** Traffic class each UP mapped to. */
1076 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1079 struct rte_eth_vmdq_tx_conf {
1080 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
1084 * A structure used to configure the VMDq+DCB feature
1085 * of an Ethernet port.
1087 * Using this feature, packets are routed to a pool of queues, based
1088 * on the VLAN ID in the VLAN tag, and then to a specific queue within
1089 * that pool, using the user priority VLAN tag field.
1091 * A default pool may be used, if desired, to route all traffic which
1092 * does not match the VLAN filter rules.
1094 struct rte_eth_vmdq_dcb_conf {
1095 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
1096 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
1097 uint8_t default_pool; /**< The default pool, if applicable */
1098 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
1100 uint16_t vlan_id; /**< The VLAN ID of the received frame */
1101 uint64_t pools; /**< Bitmask of pools for packet Rx */
1102 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */
1103 /** Selects a queue in a pool */
1104 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES];
1108 * A structure used to configure the VMDq feature of an Ethernet port when
1109 * not combined with the DCB feature.
1111 * Using this feature, packets are routed to a pool of queues. By default,
1112 * the pool selection is based on the MAC address, the VLAN ID in the
1113 * VLAN tag as specified in the pool_map array.
1114 * Passing the RTE_ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
1115 * selection using only the MAC address. MAC address to pool mapping is done
1116 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
1117 * corresponding to the pool ID.
1119 * Queue selection within the selected pool will be done using RSS when
1120 * it is enabled or revert to the first queue of the pool if not.
1122 * A default pool may be used, if desired, to route all traffic which
1123 * does not match the VLAN filter rules or any pool MAC address.
1125 struct rte_eth_vmdq_rx_conf {
1126 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
1127 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
1128 uint8_t default_pool; /**< The default pool, if applicable */
1129 uint8_t enable_loop_back; /**< Enable VT loop back */
1130 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
1131 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
1133 uint16_t vlan_id; /**< The VLAN ID of the received frame */
1134 uint64_t pools; /**< Bitmask of pools for packet Rx */
1135 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */
1139 * A structure used to configure the Tx features of an Ethernet port.
1141 struct rte_eth_txmode {
1142 enum rte_eth_tx_mq_mode mq_mode; /**< Tx multi-queues mode. */
1144 * Per-port Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags.
1145 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
1146 * structure are allowed to be set.
1152 uint8_t /** If set, reject sending out tagged pkts */
1153 hw_vlan_reject_tagged : 1,
1154 /** If set, reject sending out untagged pkts */
1155 hw_vlan_reject_untagged : 1,
1156 /** If set, enable port based VLAN insertion */
1157 hw_vlan_insert_pvid : 1;
1159 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1160 void *reserved_ptrs[2]; /**< Reserved for future fields */
1165 * @b EXPERIMENTAL: this structure may change without prior notice.
1167 * A structure used to configure an Rx packet segment to split.
1169 * If RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT flag is set in offloads field,
1170 * the PMD will split the received packets into multiple segments
1171 * according to the specification in the description array:
1173 * - The first network buffer will be allocated from the memory pool,
1174 * specified in the first array element, the second buffer, from the
1175 * pool in the second element, and so on.
1177 * - The offsets from the segment description elements specify
1178 * the data offset from the buffer beginning except the first mbuf.
1179 * The first segment offset is added with RTE_PKTMBUF_HEADROOM.
1181 * - The lengths in the elements define the maximal data amount
1182 * being received to each segment. The receiving starts with filling
1183 * up the first mbuf data buffer up to specified length. If the
1184 * there are data remaining (packet is longer than buffer in the first
1185 * mbuf) the following data will be pushed to the next segment
1186 * up to its own length, and so on.
1188 * - If the length in the segment description element is zero
1189 * the actual buffer size will be deduced from the appropriate
1190 * memory pool properties.
1192 * - If there is not enough elements to describe the buffer for entire
1193 * packet of maximal length the following parameters will be used
1194 * for the all remaining segments:
1195 * - pool from the last valid element
1196 * - the buffer size from this pool
1199 struct rte_eth_rxseg_split {
1200 struct rte_mempool *mp; /**< Memory pool to allocate segment from. */
1201 uint16_t length; /**< Segment data length, configures split point. */
1202 uint16_t offset; /**< Data offset from beginning of mbuf data buffer. */
1203 uint32_t reserved; /**< Reserved field. */
1208 * @b EXPERIMENTAL: this structure may change without prior notice.
1210 * A common structure used to describe Rx packet segment properties.
1212 union rte_eth_rxseg {
1213 /* The settings for buffer split offload. */
1214 struct rte_eth_rxseg_split split;
1215 /* The other features settings should be added here. */
1219 * A structure used to configure an Rx ring of an Ethernet port.
1221 struct rte_eth_rxconf {
1222 struct rte_eth_thresh rx_thresh; /**< Rx ring threshold registers. */
1223 uint16_t rx_free_thresh; /**< Drives the freeing of Rx descriptors. */
1224 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
1225 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1226 uint16_t rx_nseg; /**< Number of descriptions in rx_seg array. */
1228 * Share group index in Rx domain and switch domain.
1229 * Non-zero value to enable Rx queue share, zero value disable share.
1230 * PMD is responsible for Rx queue consistency checks to avoid member
1231 * port's configuration contradict to each other.
1233 uint16_t share_group;
1234 uint16_t share_qid; /**< Shared Rx queue ID in group */
1236 * Per-queue Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags.
1237 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
1238 * fields on rte_eth_dev_info structure are allowed to be set.
1242 * Points to the array of segment descriptions for an entire packet.
1243 * Array elements are properties for consecutive Rx segments.
1245 * The supported capabilities of receiving segmentation is reported
1246 * in rte_eth_dev_info.rx_seg_capa field.
1248 union rte_eth_rxseg *rx_seg;
1250 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1251 void *reserved_ptrs[2]; /**< Reserved for future fields */
1255 * A structure used to configure a Tx ring of an Ethernet port.
1257 struct rte_eth_txconf {
1258 struct rte_eth_thresh tx_thresh; /**< Tx ring threshold registers. */
1259 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
1260 uint16_t tx_free_thresh; /**< Start freeing Tx buffers if there are
1261 less free descriptors than this value. */
1263 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
1265 * Per-queue Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags.
1266 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
1267 * fields on rte_eth_dev_info structure are allowed to be set.
1271 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1272 void *reserved_ptrs[2]; /**< Reserved for future fields */
1277 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1279 * A structure used to return the hairpin capabilities that are supported.
1281 struct rte_eth_hairpin_cap {
1282 /** The max number of hairpin queues (different bindings). */
1283 uint16_t max_nb_queues;
1284 /** Max number of Rx queues to be connected to one Tx queue. */
1285 uint16_t max_rx_2_tx;
1286 /** Max number of Tx queues to be connected to one Rx queue. */
1287 uint16_t max_tx_2_rx;
1288 uint16_t max_nb_desc; /**< The max num of descriptors. */
1291 #define RTE_ETH_MAX_HAIRPIN_PEERS 32
1295 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1297 * A structure used to hold hairpin peer data.
1299 struct rte_eth_hairpin_peer {
1300 uint16_t port; /**< Peer port. */
1301 uint16_t queue; /**< Peer queue. */
1306 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1308 * A structure used to configure hairpin binding.
1310 struct rte_eth_hairpin_conf {
1311 uint32_t peer_count:16; /**< The number of peers. */
1314 * Explicit Tx flow rule mode.
1315 * One hairpin pair of queues should have the same attribute.
1317 * - When set, the user should be responsible for inserting the hairpin
1318 * Tx part flows and removing them.
1319 * - When clear, the PMD will try to handle the Tx part of the flows,
1320 * e.g., by splitting one flow into two parts.
1322 uint32_t tx_explicit:1;
1325 * Manually bind hairpin queues.
1326 * One hairpin pair of queues should have the same attribute.
1328 * - When set, to enable hairpin, the user should call the hairpin bind
1329 * function after all the queues are set up properly and the ports are
1330 * started. Also, the hairpin unbind function should be called
1331 * accordingly before stopping a port that with hairpin configured.
1332 * - When clear, the PMD will try to enable the hairpin with the queues
1333 * configured automatically during port start.
1335 uint32_t manual_bind:1;
1336 uint32_t reserved:14; /**< Reserved bits. */
1337 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS];
1341 * A structure contains information about HW descriptor ring limitations.
1343 struct rte_eth_desc_lim {
1344 uint16_t nb_max; /**< Max allowed number of descriptors. */
1345 uint16_t nb_min; /**< Min allowed number of descriptors. */
1346 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
1349 * Max allowed number of segments per whole packet.
1351 * - For TSO packet this is the total number of data descriptors allowed
1354 * @see nb_mtu_seg_max
1356 uint16_t nb_seg_max;
1359 * Max number of segments per one MTU.
1361 * - For non-TSO packet, this is the maximum allowed number of segments
1362 * in a single transmit packet.
1364 * - For TSO packet each segment within the TSO may span up to this
1369 uint16_t nb_mtu_seg_max;
1373 * This enum indicates the flow control mode
1375 enum rte_eth_fc_mode {
1376 RTE_ETH_FC_NONE = 0, /**< Disable flow control. */
1377 RTE_ETH_FC_RX_PAUSE, /**< Rx pause frame, enable flowctrl on Tx side. */
1378 RTE_ETH_FC_TX_PAUSE, /**< Tx pause frame, enable flowctrl on Rx side. */
1379 RTE_ETH_FC_FULL /**< Enable flow control on both side. */
1381 #define RTE_FC_NONE RTE_DEPRECATED(RTE_FC_NONE) RTE_ETH_FC_NONE
1382 #define RTE_FC_RX_PAUSE RTE_DEPRECATED(RTE_FC_RX_PAUSE) RTE_ETH_FC_RX_PAUSE
1383 #define RTE_FC_TX_PAUSE RTE_DEPRECATED(RTE_FC_TX_PAUSE) RTE_ETH_FC_TX_PAUSE
1384 #define RTE_FC_FULL RTE_DEPRECATED(RTE_FC_FULL) RTE_ETH_FC_FULL
1387 * A structure used to configure Ethernet flow control parameter.
1388 * These parameters will be configured into the register of the NIC.
1389 * Please refer to the corresponding data sheet for proper value.
1391 struct rte_eth_fc_conf {
1392 uint32_t high_water; /**< High threshold value to trigger XOFF */
1393 uint32_t low_water; /**< Low threshold value to trigger XON */
1394 uint16_t pause_time; /**< Pause quota in the Pause frame */
1395 uint16_t send_xon; /**< Is XON frame need be sent */
1396 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1397 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
1398 uint8_t autoneg; /**< Use Pause autoneg */
1402 * A structure used to configure Ethernet priority flow control parameter.
1403 * These parameters will be configured into the register of the NIC.
1404 * Please refer to the corresponding data sheet for proper value.
1406 struct rte_eth_pfc_conf {
1407 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
1408 uint8_t priority; /**< VLAN User Priority. */
1413 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1415 * A structure used to retrieve information of queue based PFC.
1417 struct rte_eth_pfc_queue_info {
1419 * Maximum supported traffic class as per PFC (802.1Qbb) specification.
1422 /** PFC queue mode capabilities. */
1423 enum rte_eth_fc_mode mode_capa;
1428 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
1430 * A structure used to configure Ethernet priority flow control parameters for
1433 * rte_eth_pfc_queue_conf::rx_pause structure shall be used to configure given
1434 * tx_qid with corresponding tc. When ethdev device receives PFC frame with
1435 * rte_eth_pfc_queue_conf::rx_pause::tc, traffic will be paused on
1436 * rte_eth_pfc_queue_conf::rx_pause::tx_qid for that tc.
1438 * rte_eth_pfc_queue_conf::tx_pause structure shall be used to configure given
1439 * rx_qid. When rx_qid is congested, PFC frames are generated with
1440 * rte_eth_pfc_queue_conf::rx_pause::tc and
1441 * rte_eth_pfc_queue_conf::rx_pause::pause_time to the peer.
1443 struct rte_eth_pfc_queue_conf {
1444 enum rte_eth_fc_mode mode; /**< Link flow control mode */
1447 uint16_t tx_qid; /**< Tx queue ID */
1448 /** Traffic class as per PFC (802.1Qbb) spec. The value must be
1449 * in the range [0, rte_eth_pfc_queue_info::tx_max - 1]
1452 } rx_pause; /* Valid when (mode == FC_RX_PAUSE || mode == FC_FULL) */
1455 uint16_t pause_time; /**< Pause quota in the Pause frame */
1456 uint16_t rx_qid; /**< Rx queue ID */
1457 /** Traffic class as per PFC (802.1Qbb) spec. The value must be
1458 * in the range [0, rte_eth_pfc_queue_info::tx_max - 1]
1461 } tx_pause; /* Valid when (mode == FC_TX_PAUSE || mode == FC_FULL) */
1465 * Tunnel type for device-specific classifier configuration.
1466 * @see rte_eth_udp_tunnel
1468 enum rte_eth_tunnel_type {
1469 RTE_ETH_TUNNEL_TYPE_NONE = 0,
1470 RTE_ETH_TUNNEL_TYPE_VXLAN,
1471 RTE_ETH_TUNNEL_TYPE_GENEVE,
1472 RTE_ETH_TUNNEL_TYPE_TEREDO,
1473 RTE_ETH_TUNNEL_TYPE_NVGRE,
1474 RTE_ETH_TUNNEL_TYPE_IP_IN_GRE,
1475 RTE_ETH_L2_TUNNEL_TYPE_E_TAG,
1476 RTE_ETH_TUNNEL_TYPE_VXLAN_GPE,
1477 RTE_ETH_TUNNEL_TYPE_ECPRI,
1478 RTE_ETH_TUNNEL_TYPE_MAX,
1480 #define RTE_TUNNEL_TYPE_NONE RTE_DEPRECATED(RTE_TUNNEL_TYPE_NONE) RTE_ETH_TUNNEL_TYPE_NONE
1481 #define RTE_TUNNEL_TYPE_VXLAN RTE_DEPRECATED(RTE_TUNNEL_TYPE_VXLAN) RTE_ETH_TUNNEL_TYPE_VXLAN
1482 #define RTE_TUNNEL_TYPE_GENEVE RTE_DEPRECATED(RTE_TUNNEL_TYPE_GENEVE) RTE_ETH_TUNNEL_TYPE_GENEVE
1483 #define RTE_TUNNEL_TYPE_TEREDO RTE_DEPRECATED(RTE_TUNNEL_TYPE_TEREDO) RTE_ETH_TUNNEL_TYPE_TEREDO
1484 #define RTE_TUNNEL_TYPE_NVGRE RTE_DEPRECATED(RTE_TUNNEL_TYPE_NVGRE) RTE_ETH_TUNNEL_TYPE_NVGRE
1485 #define RTE_TUNNEL_TYPE_IP_IN_GRE RTE_DEPRECATED(RTE_TUNNEL_TYPE_IP_IN_GRE) RTE_ETH_TUNNEL_TYPE_IP_IN_GRE
1486 #define RTE_L2_TUNNEL_TYPE_E_TAG RTE_DEPRECATED(RTE_L2_TUNNEL_TYPE_E_TAG) RTE_ETH_L2_TUNNEL_TYPE_E_TAG
1487 #define RTE_TUNNEL_TYPE_VXLAN_GPE RTE_DEPRECATED(RTE_TUNNEL_TYPE_VXLAN_GPE) RTE_ETH_TUNNEL_TYPE_VXLAN_GPE
1488 #define RTE_TUNNEL_TYPE_ECPRI RTE_DEPRECATED(RTE_TUNNEL_TYPE_ECPRI) RTE_ETH_TUNNEL_TYPE_ECPRI
1489 #define RTE_TUNNEL_TYPE_MAX RTE_DEPRECATED(RTE_TUNNEL_TYPE_MAX) RTE_ETH_TUNNEL_TYPE_MAX
1491 /* Deprecated API file for rte_eth_dev_filter_* functions */
1492 #include "rte_eth_ctrl.h"
1495 * Memory space that can be configured to store Flow Director filters
1496 * in the board memory.
1498 enum rte_eth_fdir_pballoc_type {
1499 RTE_ETH_FDIR_PBALLOC_64K = 0, /**< 64k. */
1500 RTE_ETH_FDIR_PBALLOC_128K, /**< 128k. */
1501 RTE_ETH_FDIR_PBALLOC_256K, /**< 256k. */
1503 #define rte_fdir_pballoc_type rte_eth_fdir_pballoc_type
1505 #define RTE_FDIR_PBALLOC_64K RTE_DEPRECATED(RTE_FDIR_PBALLOC_64K) RTE_ETH_FDIR_PBALLOC_64K
1506 #define RTE_FDIR_PBALLOC_128K RTE_DEPRECATED(RTE_FDIR_PBALLOC_128K) RTE_ETH_FDIR_PBALLOC_128K
1507 #define RTE_FDIR_PBALLOC_256K RTE_DEPRECATED(RTE_FDIR_PBALLOC_256K) RTE_ETH_FDIR_PBALLOC_256K
1510 * Select report mode of FDIR hash information in Rx descriptors.
1512 enum rte_fdir_status_mode {
1513 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
1514 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
1515 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
1519 * A structure used to configure the Flow Director (FDIR) feature
1520 * of an Ethernet port.
1522 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored.
1524 struct rte_eth_fdir_conf {
1525 enum rte_fdir_mode mode; /**< Flow Director mode. */
1526 enum rte_eth_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
1527 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
1528 /** Rx queue of packets matching a "drop" filter in perfect mode. */
1530 struct rte_eth_fdir_masks mask;
1531 /** Flex payload configuration. */
1532 struct rte_eth_fdir_flex_conf flex_conf;
1534 #define rte_fdir_conf rte_eth_fdir_conf
1537 * UDP tunneling configuration.
1539 * Used to configure the classifier of a device,
1540 * associating an UDP port with a type of tunnel.
1542 * Some NICs may need such configuration to properly parse a tunnel
1543 * with any standard or custom UDP port.
1545 struct rte_eth_udp_tunnel {
1546 uint16_t udp_port; /**< UDP port used for the tunnel. */
1547 uint8_t prot_type; /**< Tunnel type. @see rte_eth_tunnel_type */
1551 * A structure used to enable/disable specific device interrupts.
1553 struct rte_eth_intr_conf {
1554 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
1556 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
1558 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
1562 #define rte_intr_conf rte_eth_intr_conf
1565 * A structure used to configure an Ethernet port.
1566 * Depending upon the Rx multi-queue mode, extra advanced
1567 * configuration settings may be needed.
1569 struct rte_eth_conf {
1570 uint32_t link_speeds; /**< bitmap of RTE_ETH_LINK_SPEED_XXX of speeds to be
1571 used. RTE_ETH_LINK_SPEED_FIXED disables link
1572 autonegotiation, and a unique speed shall be
1573 set. Otherwise, the bitmap defines the set of
1574 speeds to be advertised. If the special value
1575 RTE_ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
1576 supported are advertised. */
1577 struct rte_eth_rxmode rxmode; /**< Port Rx configuration. */
1578 struct rte_eth_txmode txmode; /**< Port Tx configuration. */
1579 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
1580 is 0, meaning the loopback mode is disabled.
1581 Read the datasheet of given Ethernet controller
1582 for details. The possible values of this field
1583 are defined in implementation of each driver. */
1585 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
1586 /** Port VMDq+DCB configuration. */
1587 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
1588 /** Port DCB Rx configuration. */
1589 struct rte_eth_dcb_rx_conf dcb_rx_conf;
1590 /** Port VMDq Rx configuration. */
1591 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
1592 } rx_adv_conf; /**< Port Rx filtering configuration. */
1594 /** Port VMDq+DCB Tx configuration. */
1595 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
1596 /** Port DCB Tx configuration. */
1597 struct rte_eth_dcb_tx_conf dcb_tx_conf;
1598 /** Port VMDq Tx configuration. */
1599 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
1600 } tx_adv_conf; /**< Port Tx DCB configuration (union). */
1601 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
1602 is needed,and the variable must be set RTE_ETH_DCB_PFC_SUPPORT. */
1603 uint32_t dcb_capability_en;
1604 struct rte_eth_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */
1605 struct rte_eth_intr_conf intr_conf; /**< Interrupt mode configuration. */
1609 * Rx offload capabilities of a device.
1611 #define RTE_ETH_RX_OFFLOAD_VLAN_STRIP RTE_BIT64(0)
1612 #define RTE_ETH_RX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1)
1613 #define RTE_ETH_RX_OFFLOAD_UDP_CKSUM RTE_BIT64(2)
1614 #define RTE_ETH_RX_OFFLOAD_TCP_CKSUM RTE_BIT64(3)
1615 #define RTE_ETH_RX_OFFLOAD_TCP_LRO RTE_BIT64(4)
1616 #define RTE_ETH_RX_OFFLOAD_QINQ_STRIP RTE_BIT64(5)
1617 #define RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(6)
1618 #define RTE_ETH_RX_OFFLOAD_MACSEC_STRIP RTE_BIT64(7)
1619 #define RTE_ETH_RX_OFFLOAD_HEADER_SPLIT RTE_BIT64(8)
1620 #define RTE_ETH_RX_OFFLOAD_VLAN_FILTER RTE_BIT64(9)
1621 #define RTE_ETH_RX_OFFLOAD_VLAN_EXTEND RTE_BIT64(10)
1622 #define RTE_ETH_RX_OFFLOAD_SCATTER RTE_BIT64(13)
1624 * Timestamp is set by the driver in RTE_MBUF_DYNFIELD_TIMESTAMP_NAME
1625 * and RTE_MBUF_DYNFLAG_RX_TIMESTAMP_NAME is set in ol_flags.
1626 * The mbuf field and flag are registered when the offload is configured.
1628 #define RTE_ETH_RX_OFFLOAD_TIMESTAMP RTE_BIT64(14)
1629 #define RTE_ETH_RX_OFFLOAD_SECURITY RTE_BIT64(15)
1630 #define RTE_ETH_RX_OFFLOAD_KEEP_CRC RTE_BIT64(16)
1631 #define RTE_ETH_RX_OFFLOAD_SCTP_CKSUM RTE_BIT64(17)
1632 #define RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(18)
1633 #define RTE_ETH_RX_OFFLOAD_RSS_HASH RTE_BIT64(19)
1634 #define RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT RTE_BIT64(20)
1636 #define DEV_RX_OFFLOAD_VLAN_STRIP RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN_STRIP) RTE_ETH_RX_OFFLOAD_VLAN_STRIP
1637 #define DEV_RX_OFFLOAD_IPV4_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_IPV4_CKSUM) RTE_ETH_RX_OFFLOAD_IPV4_CKSUM
1638 #define DEV_RX_OFFLOAD_UDP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_UDP_CKSUM) RTE_ETH_RX_OFFLOAD_UDP_CKSUM
1639 #define DEV_RX_OFFLOAD_TCP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_TCP_CKSUM) RTE_ETH_RX_OFFLOAD_TCP_CKSUM
1640 #define DEV_RX_OFFLOAD_TCP_LRO RTE_DEPRECATED(DEV_RX_OFFLOAD_TCP_LRO) RTE_ETH_RX_OFFLOAD_TCP_LRO
1641 #define DEV_RX_OFFLOAD_QINQ_STRIP RTE_DEPRECATED(DEV_RX_OFFLOAD_QINQ_STRIP) RTE_ETH_RX_OFFLOAD_QINQ_STRIP
1642 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM) RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM
1643 #define DEV_RX_OFFLOAD_MACSEC_STRIP RTE_DEPRECATED(DEV_RX_OFFLOAD_MACSEC_STRIP) RTE_ETH_RX_OFFLOAD_MACSEC_STRIP
1644 #define DEV_RX_OFFLOAD_HEADER_SPLIT RTE_DEPRECATED(DEV_RX_OFFLOAD_HEADER_SPLIT) RTE_ETH_RX_OFFLOAD_HEADER_SPLIT
1645 #define DEV_RX_OFFLOAD_VLAN_FILTER RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN_FILTER) RTE_ETH_RX_OFFLOAD_VLAN_FILTER
1646 #define DEV_RX_OFFLOAD_VLAN_EXTEND RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN_EXTEND) RTE_ETH_RX_OFFLOAD_VLAN_EXTEND
1647 #define DEV_RX_OFFLOAD_SCATTER RTE_DEPRECATED(DEV_RX_OFFLOAD_SCATTER) RTE_ETH_RX_OFFLOAD_SCATTER
1648 #define DEV_RX_OFFLOAD_TIMESTAMP RTE_DEPRECATED(DEV_RX_OFFLOAD_TIMESTAMP) RTE_ETH_RX_OFFLOAD_TIMESTAMP
1649 #define DEV_RX_OFFLOAD_SECURITY RTE_DEPRECATED(DEV_RX_OFFLOAD_SECURITY) RTE_ETH_RX_OFFLOAD_SECURITY
1650 #define DEV_RX_OFFLOAD_KEEP_CRC RTE_DEPRECATED(DEV_RX_OFFLOAD_KEEP_CRC) RTE_ETH_RX_OFFLOAD_KEEP_CRC
1651 #define DEV_RX_OFFLOAD_SCTP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_SCTP_CKSUM) RTE_ETH_RX_OFFLOAD_SCTP_CKSUM
1652 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_OUTER_UDP_CKSUM) RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM
1653 #define DEV_RX_OFFLOAD_RSS_HASH RTE_DEPRECATED(DEV_RX_OFFLOAD_RSS_HASH) RTE_ETH_RX_OFFLOAD_RSS_HASH
1655 #define RTE_ETH_RX_OFFLOAD_CHECKSUM (RTE_ETH_RX_OFFLOAD_IPV4_CKSUM | \
1656 RTE_ETH_RX_OFFLOAD_UDP_CKSUM | \
1657 RTE_ETH_RX_OFFLOAD_TCP_CKSUM)
1658 #define DEV_RX_OFFLOAD_CHECKSUM RTE_DEPRECATED(DEV_RX_OFFLOAD_CHECKSUM) RTE_ETH_RX_OFFLOAD_CHECKSUM
1659 #define RTE_ETH_RX_OFFLOAD_VLAN (RTE_ETH_RX_OFFLOAD_VLAN_STRIP | \
1660 RTE_ETH_RX_OFFLOAD_VLAN_FILTER | \
1661 RTE_ETH_RX_OFFLOAD_VLAN_EXTEND | \
1662 RTE_ETH_RX_OFFLOAD_QINQ_STRIP)
1663 #define DEV_RX_OFFLOAD_VLAN RTE_DEPRECATED(DEV_RX_OFFLOAD_VLAN) RTE_ETH_RX_OFFLOAD_VLAN
1666 * If new Rx offload capabilities are defined, they also must be
1667 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
1671 * Tx offload capabilities of a device.
1673 #define RTE_ETH_TX_OFFLOAD_VLAN_INSERT RTE_BIT64(0)
1674 #define RTE_ETH_TX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1)
1675 #define RTE_ETH_TX_OFFLOAD_UDP_CKSUM RTE_BIT64(2)
1676 #define RTE_ETH_TX_OFFLOAD_TCP_CKSUM RTE_BIT64(3)
1677 #define RTE_ETH_TX_OFFLOAD_SCTP_CKSUM RTE_BIT64(4)
1678 #define RTE_ETH_TX_OFFLOAD_TCP_TSO RTE_BIT64(5)
1679 #define RTE_ETH_TX_OFFLOAD_UDP_TSO RTE_BIT64(6)
1680 #define RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(7) /**< Used for tunneling packet. */
1681 #define RTE_ETH_TX_OFFLOAD_QINQ_INSERT RTE_BIT64(8)
1682 #define RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO RTE_BIT64(9) /**< Used for tunneling packet. */
1683 #define RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO RTE_BIT64(10) /**< Used for tunneling packet. */
1684 #define RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO RTE_BIT64(11) /**< Used for tunneling packet. */
1685 #define RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO RTE_BIT64(12) /**< Used for tunneling packet. */
1686 #define RTE_ETH_TX_OFFLOAD_MACSEC_INSERT RTE_BIT64(13)
1688 * Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
1689 * Tx queue without SW lock.
1691 #define RTE_ETH_TX_OFFLOAD_MT_LOCKFREE RTE_BIT64(14)
1692 /** Device supports multi segment send. */
1693 #define RTE_ETH_TX_OFFLOAD_MULTI_SEGS RTE_BIT64(15)
1695 * Device supports optimization for fast release of mbufs.
1696 * When set application must guarantee that per-queue all mbufs comes from
1697 * the same mempool and has refcnt = 1.
1699 #define RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE RTE_BIT64(16)
1700 #define RTE_ETH_TX_OFFLOAD_SECURITY RTE_BIT64(17)
1702 * Device supports generic UDP tunneled packet TSO.
1703 * Application must set RTE_MBUF_F_TX_TUNNEL_UDP and other mbuf fields required
1706 #define RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO RTE_BIT64(18)
1708 * Device supports generic IP tunneled packet TSO.
1709 * Application must set RTE_MBUF_F_TX_TUNNEL_IP and other mbuf fields required
1712 #define RTE_ETH_TX_OFFLOAD_IP_TNL_TSO RTE_BIT64(19)
1713 /** Device supports outer UDP checksum */
1714 #define RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(20)
1716 * Device sends on time read from RTE_MBUF_DYNFIELD_TIMESTAMP_NAME
1717 * if RTE_MBUF_DYNFLAG_TX_TIMESTAMP_NAME is set in ol_flags.
1718 * The mbuf field and flag are registered when the offload is configured.
1720 #define RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_BIT64(21)
1722 * If new Tx offload capabilities are defined, they also must be
1723 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1726 #define DEV_TX_OFFLOAD_VLAN_INSERT RTE_DEPRECATED(DEV_TX_OFFLOAD_VLAN_INSERT) RTE_ETH_TX_OFFLOAD_VLAN_INSERT
1727 #define DEV_TX_OFFLOAD_IPV4_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_IPV4_CKSUM) RTE_ETH_TX_OFFLOAD_IPV4_CKSUM
1728 #define DEV_TX_OFFLOAD_UDP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_UDP_CKSUM) RTE_ETH_TX_OFFLOAD_UDP_CKSUM
1729 #define DEV_TX_OFFLOAD_TCP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_TCP_CKSUM) RTE_ETH_TX_OFFLOAD_TCP_CKSUM
1730 #define DEV_TX_OFFLOAD_SCTP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_SCTP_CKSUM) RTE_ETH_TX_OFFLOAD_SCTP_CKSUM
1731 #define DEV_TX_OFFLOAD_TCP_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_TCP_TSO) RTE_ETH_TX_OFFLOAD_TCP_TSO
1732 #define DEV_TX_OFFLOAD_UDP_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_UDP_TSO) RTE_ETH_TX_OFFLOAD_UDP_TSO
1733 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM) RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM
1734 #define DEV_TX_OFFLOAD_QINQ_INSERT RTE_DEPRECATED(DEV_TX_OFFLOAD_QINQ_INSERT) RTE_ETH_TX_OFFLOAD_QINQ_INSERT
1735 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_VXLAN_TNL_TSO) RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO
1736 #define DEV_TX_OFFLOAD_GRE_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_GRE_TNL_TSO) RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO
1737 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_IPIP_TNL_TSO) RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO
1738 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_GENEVE_TNL_TSO) RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO
1739 #define DEV_TX_OFFLOAD_MACSEC_INSERT RTE_DEPRECATED(DEV_TX_OFFLOAD_MACSEC_INSERT) RTE_ETH_TX_OFFLOAD_MACSEC_INSERT
1740 #define DEV_TX_OFFLOAD_MT_LOCKFREE RTE_DEPRECATED(DEV_TX_OFFLOAD_MT_LOCKFREE) RTE_ETH_TX_OFFLOAD_MT_LOCKFREE
1741 #define DEV_TX_OFFLOAD_MULTI_SEGS RTE_DEPRECATED(DEV_TX_OFFLOAD_MULTI_SEGS) RTE_ETH_TX_OFFLOAD_MULTI_SEGS
1742 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE RTE_DEPRECATED(DEV_TX_OFFLOAD_MBUF_FAST_FREE) RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE
1743 #define DEV_TX_OFFLOAD_SECURITY RTE_DEPRECATED(DEV_TX_OFFLOAD_SECURITY) RTE_ETH_TX_OFFLOAD_SECURITY
1744 #define DEV_TX_OFFLOAD_UDP_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_UDP_TNL_TSO) RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO
1745 #define DEV_TX_OFFLOAD_IP_TNL_TSO RTE_DEPRECATED(DEV_TX_OFFLOAD_IP_TNL_TSO) RTE_ETH_TX_OFFLOAD_IP_TNL_TSO
1746 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_DEPRECATED(DEV_TX_OFFLOAD_OUTER_UDP_CKSUM) RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM
1747 #define DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_DEPRECATED(DEV_TX_OFFLOAD_SEND_ON_TIMESTAMP) RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP
1749 /**@{@name Device capabilities
1750 * Non-offload capabilities reported in rte_eth_dev_info.dev_capa.
1752 /** Device supports Rx queue setup after device started. */
1753 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP RTE_BIT64(0)
1754 /** Device supports Tx queue setup after device started. */
1755 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP RTE_BIT64(1)
1757 * Device supports shared Rx queue among ports within Rx domain and
1758 * switch domain. Mbufs are consumed by shared Rx queue instead of
1759 * each queue. Multiple groups are supported by share_group of Rx
1760 * queue configuration. Shared Rx queue is identified by PMD using
1761 * share_qid of Rx queue configuration. Polling any port in the group
1762 * receive packets of all member ports, source port identified by
1765 #define RTE_ETH_DEV_CAPA_RXQ_SHARE RTE_BIT64(2)
1766 /** Device supports keeping flow rules across restart. */
1767 #define RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP RTE_BIT64(3)
1768 /** Device supports keeping shared flow objects across restart. */
1769 #define RTE_ETH_DEV_CAPA_FLOW_SHARED_OBJECT_KEEP RTE_BIT64(4)
1773 * Fallback default preferred Rx/Tx port parameters.
1774 * These are used if an application requests default parameters
1775 * but the PMD does not provide preferred values.
1777 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1778 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1779 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1780 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1783 * Preferred Rx/Tx port parameters.
1784 * There are separate instances of this structure for transmission
1785 * and reception respectively.
1787 struct rte_eth_dev_portconf {
1788 uint16_t burst_size; /**< Device-preferred burst size */
1789 uint16_t ring_size; /**< Device-preferred size of queue rings */
1790 uint16_t nb_queues; /**< Device-preferred number of queues */
1794 * Default values for switch domain ID when ethdev does not support switch
1795 * domain definitions.
1797 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX)
1800 * Ethernet device associated switch information
1802 struct rte_eth_switch_info {
1803 const char *name; /**< switch name */
1804 uint16_t domain_id; /**< switch domain ID */
1806 * Mapping to the devices physical switch port as enumerated from the
1807 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1808 * device this may correspond to the VF_ID of each virtual function,
1809 * but each driver should explicitly define the mapping of switch
1810 * port identifier to that physical interconnect/switch
1814 * Shared Rx queue sub-domain boundary. Only ports in same Rx domain
1815 * and switch domain can share Rx queue. Valid only if device advertised
1816 * RTE_ETH_DEV_CAPA_RXQ_SHARE capability.
1823 * @b EXPERIMENTAL: this structure may change without prior notice.
1825 * Ethernet device Rx buffer segmentation capabilities.
1827 struct rte_eth_rxseg_capa {
1829 uint32_t multi_pools:1; /**< Supports receiving to multiple pools.*/
1830 uint32_t offset_allowed:1; /**< Supports buffer offsets. */
1831 uint32_t offset_align_log2:4; /**< Required offset alignment. */
1832 uint16_t max_nseg; /**< Maximum amount of segments to split. */
1833 uint16_t reserved; /**< Reserved field. */
1837 * Ethernet device information
1841 * Ethernet device representor port type.
1843 enum rte_eth_representor_type {
1844 RTE_ETH_REPRESENTOR_NONE, /**< not a representor. */
1845 RTE_ETH_REPRESENTOR_VF, /**< representor of Virtual Function. */
1846 RTE_ETH_REPRESENTOR_SF, /**< representor of Sub Function. */
1847 RTE_ETH_REPRESENTOR_PF, /**< representor of Physical Function. */
1851 * A structure used to retrieve the contextual information of
1852 * an Ethernet device, such as the controlling driver of the
1855 struct rte_eth_dev_info {
1856 struct rte_device *device; /** Generic device information */
1857 const char *driver_name; /**< Device Driver name. */
1858 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1859 Use if_indextoname() to translate into an interface name. */
1860 uint16_t min_mtu; /**< Minimum MTU allowed */
1861 uint16_t max_mtu; /**< Maximum MTU allowed */
1862 const uint32_t *dev_flags; /**< Device flags */
1863 uint32_t min_rx_bufsize; /**< Minimum size of Rx buffer. */
1864 uint32_t max_rx_pktlen; /**< Maximum configurable length of Rx pkt. */
1865 /** Maximum configurable size of LRO aggregated packet. */
1866 uint32_t max_lro_pkt_size;
1867 uint16_t max_rx_queues; /**< Maximum number of Rx queues. */
1868 uint16_t max_tx_queues; /**< Maximum number of Tx queues. */
1869 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1870 uint32_t max_hash_mac_addrs;
1871 /** Maximum number of hash MAC addresses for MTA and UTA. */
1872 uint16_t max_vfs; /**< Maximum number of VFs. */
1873 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1874 struct rte_eth_rxseg_capa rx_seg_capa; /**< Segmentation capability.*/
1875 /** All Rx offload capabilities including all per-queue ones */
1876 uint64_t rx_offload_capa;
1877 /** All Tx offload capabilities including all per-queue ones */
1878 uint64_t tx_offload_capa;
1879 /** Device per-queue Rx offload capabilities. */
1880 uint64_t rx_queue_offload_capa;
1881 /** Device per-queue Tx offload capabilities. */
1882 uint64_t tx_queue_offload_capa;
1883 /** Device redirection table size, the total number of entries. */
1885 uint8_t hash_key_size; /**< Hash key size in bytes */
1886 /** Bit mask of RSS offloads, the bit offset also means flow type */
1887 uint64_t flow_type_rss_offloads;
1888 struct rte_eth_rxconf default_rxconf; /**< Default Rx configuration */
1889 struct rte_eth_txconf default_txconf; /**< Default Tx configuration */
1890 uint16_t vmdq_queue_base; /**< First queue ID for VMDq pools. */
1891 uint16_t vmdq_queue_num; /**< Queue number for VMDq pools. */
1892 uint16_t vmdq_pool_base; /**< First ID of VMDq pools. */
1893 struct rte_eth_desc_lim rx_desc_lim; /**< Rx descriptors limits */
1894 struct rte_eth_desc_lim tx_desc_lim; /**< Tx descriptors limits */
1895 uint32_t speed_capa; /**< Supported speeds bitmap (RTE_ETH_LINK_SPEED_). */
1896 /** Configured number of Rx/Tx queues */
1897 uint16_t nb_rx_queues; /**< Number of Rx queues. */
1898 uint16_t nb_tx_queues; /**< Number of Tx queues. */
1899 /** Rx parameter recommendations */
1900 struct rte_eth_dev_portconf default_rxportconf;
1901 /** Tx parameter recommendations */
1902 struct rte_eth_dev_portconf default_txportconf;
1903 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1906 * Switching information for ports on a device with a
1907 * embedded managed interconnect/switch.
1909 struct rte_eth_switch_info switch_info;
1911 uint64_t reserved_64s[2]; /**< Reserved for future fields */
1912 void *reserved_ptrs[2]; /**< Reserved for future fields */
1915 /**@{@name Rx/Tx queue states */
1916 #define RTE_ETH_QUEUE_STATE_STOPPED 0 /**< Queue stopped. */
1917 #define RTE_ETH_QUEUE_STATE_STARTED 1 /**< Queue started. */
1918 #define RTE_ETH_QUEUE_STATE_HAIRPIN 2 /**< Queue used for hairpin. */
1922 * Ethernet device Rx queue information structure.
1923 * Used to retrieve information about configured queue.
1925 struct rte_eth_rxq_info {
1926 struct rte_mempool *mp; /**< mempool used by that queue. */
1927 struct rte_eth_rxconf conf; /**< queue config parameters. */
1928 uint8_t scattered_rx; /**< scattered packets Rx supported. */
1929 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */
1930 uint16_t nb_desc; /**< configured number of RXDs. */
1931 uint16_t rx_buf_size; /**< hardware receive buffer size. */
1932 } __rte_cache_min_aligned;
1935 * Ethernet device Tx queue information structure.
1936 * Used to retrieve information about configured queue.
1938 struct rte_eth_txq_info {
1939 struct rte_eth_txconf conf; /**< queue config parameters. */
1940 uint16_t nb_desc; /**< configured number of TXDs. */
1941 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */
1942 } __rte_cache_min_aligned;
1944 /* Generic Burst mode flag definition, values can be ORed. */
1947 * If the queues have different burst mode description, this bit will be set
1948 * by PMD, then the application can iterate to retrieve burst description for
1951 #define RTE_ETH_BURST_FLAG_PER_QUEUE RTE_BIT64(0)
1954 * Ethernet device Rx/Tx queue packet burst mode information structure.
1955 * Used to retrieve information about packet burst mode setting.
1957 struct rte_eth_burst_mode {
1958 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */
1960 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */
1961 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */
1964 /** Maximum name length for extended statistics counters */
1965 #define RTE_ETH_XSTATS_NAME_SIZE 64
1968 * An Ethernet device extended statistic structure
1970 * This structure is used by rte_eth_xstats_get() to provide
1971 * statistics that are not provided in the generic *rte_eth_stats*
1973 * It maps a name ID, corresponding to an index in the array returned
1974 * by rte_eth_xstats_get_names(), to a statistic value.
1976 struct rte_eth_xstat {
1977 uint64_t id; /**< The index in xstats name array. */
1978 uint64_t value; /**< The statistic counter value. */
1982 * A name element for extended statistics.
1984 * An array of this structure is returned by rte_eth_xstats_get_names().
1985 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1986 * structure references these names by their array index.
1988 * The xstats should follow a common naming scheme.
1989 * Some names are standardized in rte_stats_strings.
1991 * - rx_missed_errors
1993 * - tx_size_128_to_255_packets
1995 struct rte_eth_xstat_name {
1996 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1999 #define RTE_ETH_DCB_NUM_TCS 8
2000 #define RTE_ETH_MAX_VMDQ_POOL 64
2002 #define ETH_DCB_NUM_TCS RTE_DEPRECATED(ETH_DCB_NUM_TCS) RTE_ETH_DCB_NUM_TCS
2003 #define ETH_MAX_VMDQ_POOL RTE_DEPRECATED(ETH_MAX_VMDQ_POOL) RTE_ETH_MAX_VMDQ_POOL
2006 * A structure used to get the information of queue and
2007 * TC mapping on both Tx and Rx paths.
2009 struct rte_eth_dcb_tc_queue_mapping {
2010 /** Rx queues assigned to tc per Pool */
2014 } tc_rxq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS];
2015 /** Rx queues assigned to tc per Pool */
2019 } tc_txq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS];
2023 * A structure used to get the information of DCB.
2024 * It includes TC UP mapping and queue TC mapping.
2026 struct rte_eth_dcb_info {
2027 uint8_t nb_tcs; /**< number of TCs */
2028 uint8_t prio_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
2029 uint8_t tc_bws[RTE_ETH_DCB_NUM_TCS]; /**< Tx BW percentage for each TC */
2030 /** Rx queues assigned to tc */
2031 struct rte_eth_dcb_tc_queue_mapping tc_queue;
2035 * This enum indicates the possible Forward Error Correction (FEC) modes
2036 * of an ethdev port.
2038 enum rte_eth_fec_mode {
2039 RTE_ETH_FEC_NOFEC = 0, /**< FEC is off */
2040 RTE_ETH_FEC_AUTO, /**< FEC autonegotiation modes */
2041 RTE_ETH_FEC_BASER, /**< FEC using common algorithm */
2042 RTE_ETH_FEC_RS, /**< FEC using RS algorithm */
2045 /* Translate from FEC mode to FEC capa */
2046 #define RTE_ETH_FEC_MODE_TO_CAPA(x) RTE_BIT32(x)
2048 /* This macro indicates FEC capa mask */
2049 #define RTE_ETH_FEC_MODE_CAPA_MASK(x) RTE_BIT32(RTE_ETH_FEC_ ## x)
2051 /* A structure used to get capabilities per link speed */
2052 struct rte_eth_fec_capa {
2053 uint32_t speed; /**< Link speed (see RTE_ETH_SPEED_NUM_*) */
2054 uint32_t capa; /**< FEC capabilities bitmask */
2057 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
2059 /* Macros to check for valid port */
2060 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
2061 if (!rte_eth_dev_is_valid_port(port_id)) { \
2062 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
2067 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
2068 if (!rte_eth_dev_is_valid_port(port_id)) { \
2069 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
2075 * Function type used for Rx packet processing packet callbacks.
2077 * The callback function is called on Rx with a burst of packets that have
2078 * been received on the given port and queue.
2081 * The Ethernet port on which Rx is being performed.
2083 * The queue on the Ethernet port which is being used to receive the packets.
2085 * The burst of packets that have just been received.
2087 * The number of packets in the burst pointed to by "pkts".
2089 * The max number of packets that can be stored in the "pkts" array.
2091 * The arbitrary user parameter passed in by the application when the callback
2092 * was originally configured.
2094 * The number of packets returned to the user.
2096 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
2097 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
2101 * Function type used for Tx packet processing packet callbacks.
2103 * The callback function is called on Tx with a burst of packets immediately
2104 * before the packets are put onto the hardware queue for transmission.
2107 * The Ethernet port on which Tx is being performed.
2109 * The queue on the Ethernet port which is being used to transmit the packets.
2111 * The burst of packets that are about to be transmitted.
2113 * The number of packets in the burst pointed to by "pkts".
2115 * The arbitrary user parameter passed in by the application when the callback
2116 * was originally configured.
2118 * The number of packets to be written to the NIC.
2120 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
2121 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
2124 * Possible states of an ethdev port.
2126 enum rte_eth_dev_state {
2127 /** Device is unused before being probed. */
2128 RTE_ETH_DEV_UNUSED = 0,
2129 /** Device is attached when allocated in probing. */
2130 RTE_ETH_DEV_ATTACHED,
2131 /** Device is in removed state when plug-out is detected. */
2132 RTE_ETH_DEV_REMOVED,
2135 struct rte_eth_dev_sriov {
2136 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
2137 uint8_t nb_q_per_pool; /**< Rx queue number per pool */
2138 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
2139 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
2141 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
2143 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
2145 #define RTE_ETH_DEV_NO_OWNER 0
2147 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
2149 struct rte_eth_dev_owner {
2150 uint64_t id; /**< The owner unique identifier. */
2151 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
2154 /**@{@name Device flags
2155 * Flags internally saved in rte_eth_dev_data.dev_flags
2156 * and reported in rte_eth_dev_info.dev_flags.
2158 /** PMD supports thread-safe flow operations */
2159 #define RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE RTE_BIT32(0)
2160 /** Device supports link state interrupt */
2161 #define RTE_ETH_DEV_INTR_LSC RTE_BIT32(1)
2162 /** Device is a bonded slave */
2163 #define RTE_ETH_DEV_BONDED_SLAVE RTE_BIT32(2)
2164 /** Device supports device removal interrupt */
2165 #define RTE_ETH_DEV_INTR_RMV RTE_BIT32(3)
2166 /** Device is port representor */
2167 #define RTE_ETH_DEV_REPRESENTOR RTE_BIT32(4)
2168 /** Device does not support MAC change after started */
2169 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR RTE_BIT32(5)
2171 * Queue xstats filled automatically by ethdev layer.
2172 * PMDs filling the queue xstats themselves should not set this flag
2174 #define RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS RTE_BIT32(6)
2178 * Iterates over valid ethdev ports owned by a specific owner.
2181 * The ID of the next possible valid owned port.
2183 * The owner identifier.
2184 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
2186 * Next valid port ID owned by owner_id, RTE_MAX_ETHPORTS if there is none.
2188 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
2189 const uint64_t owner_id);
2192 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
2194 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
2195 for (p = rte_eth_find_next_owned_by(0, o); \
2196 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
2197 p = rte_eth_find_next_owned_by(p + 1, o))
2200 * Iterates over valid ethdev ports.
2203 * The ID of the next possible valid port.
2205 * Next valid port ID, RTE_MAX_ETHPORTS if there is none.
2207 uint16_t rte_eth_find_next(uint16_t port_id);
2210 * Macro to iterate over all enabled and ownerless ethdev ports.
2212 #define RTE_ETH_FOREACH_DEV(p) \
2213 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
2216 * Iterates over ethdev ports of a specified device.
2218 * @param port_id_start
2219 * The ID of the next possible valid port.
2221 * The generic device behind the ports to iterate.
2223 * Next port ID of the device, possibly port_id_start,
2224 * RTE_MAX_ETHPORTS if there is none.
2227 rte_eth_find_next_of(uint16_t port_id_start,
2228 const struct rte_device *parent);
2231 * Macro to iterate over all ethdev ports of a specified device.
2234 * The ID of the matching port being iterated.
2236 * The rte_device pointer matching the iterated ports.
2238 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \
2239 for (port_id = rte_eth_find_next_of(0, parent); \
2240 port_id < RTE_MAX_ETHPORTS; \
2241 port_id = rte_eth_find_next_of(port_id + 1, parent))
2244 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device).
2246 * @param port_id_start
2247 * The ID of the next possible valid sibling port.
2248 * @param ref_port_id
2249 * The ID of a reference port to compare rte_device with.
2251 * Next sibling port ID, possibly port_id_start or ref_port_id itself,
2252 * RTE_MAX_ETHPORTS if there is none.
2255 rte_eth_find_next_sibling(uint16_t port_id_start, uint16_t ref_port_id);
2258 * Macro to iterate over all ethdev ports sharing the same rte_device
2259 * as the specified port.
2260 * Note: the specified reference port is part of the loop iterations.
2263 * The ID of the matching port being iterated.
2264 * @param ref_port_id
2265 * The ID of the port being compared.
2267 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \
2268 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \
2269 port_id < RTE_MAX_ETHPORTS; \
2270 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id))
2274 * @b EXPERIMENTAL: this API may change without prior notice.
2276 * Get a new unique owner identifier.
2277 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
2278 * to avoid multiple management of device by different entities.
2281 * Owner identifier pointer.
2283 * Negative errno value on error, 0 on success.
2286 int rte_eth_dev_owner_new(uint64_t *owner_id);
2290 * @b EXPERIMENTAL: this API may change without prior notice.
2292 * Set an Ethernet device owner.
2295 * The identifier of the port to own.
2297 * The owner pointer.
2299 * 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);
2307 * @b EXPERIMENTAL: this API may change without prior notice.
2309 * Unset Ethernet device owner to make the device ownerless.
2312 * The identifier of port to make ownerless.
2314 * The owner identifier.
2316 * 0 on success, negative errno value on error.
2319 int rte_eth_dev_owner_unset(const uint16_t port_id,
2320 const uint64_t owner_id);
2324 * @b EXPERIMENTAL: this API may change without prior notice.
2326 * Remove owner from all Ethernet devices owned by a specific owner.
2329 * The owner identifier.
2331 * 0 on success, negative errno value on error.
2334 int rte_eth_dev_owner_delete(const uint64_t owner_id);
2338 * @b EXPERIMENTAL: this API may change without prior notice.
2340 * Get the owner of an Ethernet device.
2343 * The port identifier.
2345 * The owner structure pointer to fill.
2347 * 0 on success, negative errno value on error..
2350 int rte_eth_dev_owner_get(const uint16_t port_id,
2351 struct rte_eth_dev_owner *owner);
2354 * Get the number of ports which are usable for the application.
2356 * These devices must be iterated by using the macro
2357 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
2358 * to deal with non-contiguous ranges of devices.
2361 * The count of available Ethernet devices.
2363 uint16_t rte_eth_dev_count_avail(void);
2366 * Get the total number of ports which are allocated.
2368 * Some devices may not be available for the application.
2371 * The total count of Ethernet devices.
2373 uint16_t rte_eth_dev_count_total(void);
2376 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
2377 * the bitmap link_speeds of the struct rte_eth_conf
2380 * Numerical speed value in Mbps
2382 * RTE_ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
2384 * 0 if the speed cannot be mapped
2386 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
2389 * Get RTE_ETH_RX_OFFLOAD_* flag name.
2394 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
2396 const char *rte_eth_dev_rx_offload_name(uint64_t offload);
2399 * Get RTE_ETH_TX_OFFLOAD_* flag name.
2404 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
2406 const char *rte_eth_dev_tx_offload_name(uint64_t offload);
2410 * @b EXPERIMENTAL: this API may change without prior notice.
2412 * Get RTE_ETH_DEV_CAPA_* flag name.
2417 * Capability name or 'UNKNOWN' if the flag cannot be recognized.
2420 const char *rte_eth_dev_capability_name(uint64_t capability);
2423 * Configure an Ethernet device.
2424 * This function must be invoked first before any other function in the
2425 * Ethernet API. This function can also be re-invoked when a device is in the
2429 * The port identifier of the Ethernet device to configure.
2430 * @param nb_rx_queue
2431 * The number of receive queues to set up for the Ethernet device.
2432 * @param nb_tx_queue
2433 * The number of transmit queues to set up for the Ethernet device.
2435 * The pointer to the configuration data to be used for the Ethernet device.
2436 * The *rte_eth_conf* structure includes:
2437 * - the hardware offload features to activate, with dedicated fields for
2438 * each statically configurable offload hardware feature provided by
2439 * Ethernet devices, such as IP checksum or VLAN tag stripping for
2441 * The Rx offload bitfield API is obsolete and will be deprecated.
2442 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
2443 * structure and use offloads field to set per-port offloads instead.
2444 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
2445 * the [rt]x_offload_capa returned from rte_eth_dev_info_get().
2446 * Any type of device supported offloading set in the input argument
2447 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
2448 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup()
2449 * - the Receive Side Scaling (RSS) configuration when using multiple Rx
2450 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf
2451 * must be within the flow_type_rss_offloads provided by drivers via
2452 * rte_eth_dev_info_get() API.
2454 * Embedding all configuration information in a single data structure
2455 * is the more flexible method that allows the addition of new features
2456 * without changing the syntax of the API.
2458 * - 0: Success, device configured.
2459 * - <0: Error code returned by the driver configuration function.
2461 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
2462 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
2465 * Check if an Ethernet device was physically removed.
2468 * The port identifier of the Ethernet device.
2470 * 1 when the Ethernet device is removed, otherwise 0.
2473 rte_eth_dev_is_removed(uint16_t port_id);
2476 * Allocate and set up a receive queue for an Ethernet device.
2478 * The function allocates a contiguous block of memory for *nb_rx_desc*
2479 * receive descriptors from a memory zone associated with *socket_id*
2480 * and initializes each receive descriptor with a network buffer allocated
2481 * from the memory pool *mb_pool*.
2484 * The port identifier of the Ethernet device.
2485 * @param rx_queue_id
2486 * The index of the receive queue to set up.
2487 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2488 * to rte_eth_dev_configure().
2490 * The number of receive descriptors to allocate for the receive ring.
2492 * The *socket_id* argument is the socket identifier in case of NUMA.
2493 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2494 * the DMA memory allocated for the receive descriptors of the ring.
2496 * The pointer to the configuration data to be used for the receive queue.
2497 * NULL value is allowed, in which case default Rx configuration
2499 * The *rx_conf* structure contains an *rx_thresh* structure with the values
2500 * of the Prefetch, Host, and Write-Back threshold registers of the receive
2502 * In addition it contains the hardware offloads features to activate using
2503 * the RTE_ETH_RX_OFFLOAD_* flags.
2504 * If an offloading set in rx_conf->offloads
2505 * hasn't been set in the input argument eth_conf->rxmode.offloads
2506 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2507 * per-queue type and it is enabled for the queue.
2508 * No need to repeat any bit in rx_conf->offloads which has already been
2509 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2510 * at port level can't be disabled at queue level.
2511 * The configuration structure also contains the pointer to the array
2512 * of the receiving buffer segment descriptions, see rx_seg and rx_nseg
2513 * fields, this extended configuration might be used by split offloads like
2514 * RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT. If mb_pool is not NULL,
2515 * the extended configuration fields must be set to NULL and zero.
2517 * The pointer to the memory pool from which to allocate *rte_mbuf* network
2518 * memory buffers to populate each descriptor of the receive ring. There are
2519 * two options to provide Rx buffer configuration:
2521 * mb_pool is not NULL, rx_conf.rx_nseg is 0.
2522 * - multiple segments description:
2523 * mb_pool is NULL, rx_conf.rx_seg is not NULL, rx_conf.rx_nseg is not 0.
2524 * Taken only if flag RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT is set in offloads.
2527 * - 0: Success, receive queue correctly set up.
2528 * - -EIO: if device is removed.
2529 * - -ENODEV: if *port_id* is invalid.
2530 * - -EINVAL: The memory pool pointer is null or the size of network buffers
2531 * which can be allocated from this memory pool does not fit the various
2532 * buffer sizes allowed by the device controller.
2533 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
2534 * allocate network memory buffers from the memory pool when
2535 * initializing receive descriptors.
2537 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
2538 uint16_t nb_rx_desc, unsigned int socket_id,
2539 const struct rte_eth_rxconf *rx_conf,
2540 struct rte_mempool *mb_pool);
2544 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2546 * Allocate and set up a hairpin receive queue for an Ethernet device.
2548 * The function set up the selected queue to be used in hairpin.
2551 * The port identifier of the Ethernet device.
2552 * @param rx_queue_id
2553 * The index of the receive queue to set up.
2554 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2555 * to rte_eth_dev_configure().
2557 * The number of receive descriptors to allocate for the receive ring.
2558 * 0 means the PMD will use default value.
2560 * The pointer to the hairpin configuration.
2563 * - (0) if successful.
2564 * - (-ENODEV) if *port_id* is invalid.
2565 * - (-ENOTSUP) if hardware doesn't support.
2566 * - (-EINVAL) if bad parameter.
2567 * - (-ENOMEM) if unable to allocate the resources.
2570 int rte_eth_rx_hairpin_queue_setup
2571 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc,
2572 const struct rte_eth_hairpin_conf *conf);
2575 * Allocate and set up a transmit queue for an Ethernet device.
2578 * The port identifier of the Ethernet device.
2579 * @param tx_queue_id
2580 * The index of the transmit queue to set up.
2581 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2582 * to rte_eth_dev_configure().
2584 * The number of transmit descriptors to allocate for the transmit ring.
2586 * The *socket_id* argument is the socket identifier in case of NUMA.
2587 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2588 * the DMA memory allocated for the transmit descriptors of the ring.
2590 * The pointer to the configuration data to be used for the transmit queue.
2591 * NULL value is allowed, in which case default Tx configuration
2593 * The *tx_conf* structure contains the following data:
2594 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
2595 * Write-Back threshold registers of the transmit ring.
2596 * When setting Write-Back threshold to the value greater then zero,
2597 * *tx_rs_thresh* value should be explicitly set to one.
2598 * - The *tx_free_thresh* value indicates the [minimum] number of network
2599 * buffers that must be pending in the transmit ring to trigger their
2600 * [implicit] freeing by the driver transmit function.
2601 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
2602 * descriptors that must be pending in the transmit ring before setting the
2603 * RS bit on a descriptor by the driver transmit function.
2604 * The *tx_rs_thresh* value should be less or equal then
2605 * *tx_free_thresh* value, and both of them should be less then
2607 * - The *offloads* member contains Tx offloads to be enabled.
2608 * If an offloading set in tx_conf->offloads
2609 * hasn't been set in the input argument eth_conf->txmode.offloads
2610 * to rte_eth_dev_configure(), it is a new added offloading, it must be
2611 * per-queue type and it is enabled for the queue.
2612 * No need to repeat any bit in tx_conf->offloads which has already been
2613 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
2614 * at port level can't be disabled at queue level.
2616 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
2617 * the transmit function to use default values.
2619 * - 0: Success, the transmit queue is correctly set up.
2620 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
2622 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
2623 uint16_t nb_tx_desc, unsigned int socket_id,
2624 const struct rte_eth_txconf *tx_conf);
2628 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2630 * Allocate and set up a transmit hairpin queue for an Ethernet device.
2633 * The port identifier of the Ethernet device.
2634 * @param tx_queue_id
2635 * The index of the transmit queue to set up.
2636 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2637 * to rte_eth_dev_configure().
2639 * The number of transmit descriptors to allocate for the transmit ring.
2640 * 0 to set default PMD value.
2642 * The hairpin configuration.
2645 * - (0) if successful.
2646 * - (-ENODEV) if *port_id* is invalid.
2647 * - (-ENOTSUP) if hardware doesn't support.
2648 * - (-EINVAL) if bad parameter.
2649 * - (-ENOMEM) if unable to allocate the resources.
2652 int rte_eth_tx_hairpin_queue_setup
2653 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc,
2654 const struct rte_eth_hairpin_conf *conf);
2658 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2660 * Get all the hairpin peer Rx / Tx ports of the current port.
2661 * The caller should ensure that the array is large enough to save the ports
2665 * The port identifier of the Ethernet device.
2667 * Pointer to the array to store the peer ports list.
2669 * Length of the array to store the port identifiers.
2671 * Current port to peer port direction
2672 * positive - current used as Tx to get all peer Rx ports.
2673 * zero - current used as Rx to get all peer Tx ports.
2676 * - (0 or positive) actual peer ports number.
2677 * - (-EINVAL) if bad parameter.
2678 * - (-ENODEV) if *port_id* invalid
2679 * - (-ENOTSUP) if hardware doesn't support.
2680 * - Others detailed errors from PMDs.
2683 int rte_eth_hairpin_get_peer_ports(uint16_t port_id, uint16_t *peer_ports,
2684 size_t len, uint32_t direction);
2688 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2690 * Bind all hairpin Tx queues of one port to the Rx queues of the peer port.
2691 * It is only allowed to call this function after all hairpin queues are
2692 * configured properly and the devices are in started state.
2695 * The identifier of the Tx port.
2697 * The identifier of peer Rx port.
2698 * RTE_MAX_ETHPORTS is allowed for the traversal of all devices.
2699 * Rx port ID could have the same value as Tx port ID.
2702 * - (0) if successful.
2703 * - (-ENODEV) if Tx port ID is invalid.
2704 * - (-EBUSY) if device is not in started state.
2705 * - (-ENOTSUP) if hardware doesn't support.
2706 * - Others detailed errors from PMDs.
2709 int rte_eth_hairpin_bind(uint16_t tx_port, uint16_t rx_port);
2713 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
2715 * Unbind all hairpin Tx queues of one port from the Rx queues of the peer port.
2716 * This should be called before closing the Tx or Rx devices, if the bind
2717 * function is called before.
2718 * After unbinding the hairpin ports pair, it is allowed to bind them again.
2719 * Changing queues configuration should be after stopping the device(s).
2722 * The identifier of the Tx port.
2724 * The identifier of peer Rx port.
2725 * RTE_MAX_ETHPORTS is allowed for traversal of all devices.
2726 * Rx port ID could have the same value as Tx port ID.
2729 * - (0) if successful.
2730 * - (-ENODEV) if Tx port ID is invalid.
2731 * - (-EBUSY) if device is in stopped state.
2732 * - (-ENOTSUP) if hardware doesn't support.
2733 * - Others detailed errors from PMDs.
2736 int rte_eth_hairpin_unbind(uint16_t tx_port, uint16_t rx_port);
2739 * Return the NUMA socket to which an Ethernet device is connected
2742 * The port identifier of the Ethernet device
2744 * The NUMA socket ID to which the Ethernet device is connected or
2745 * a default of zero if the socket could not be determined.
2746 * -1 is returned is the port_id value is out of range.
2748 int rte_eth_dev_socket_id(uint16_t port_id);
2751 * Check if port_id of device is attached
2754 * The port identifier of the Ethernet device
2756 * - 0 if port is out of range or not attached
2757 * - 1 if device is attached
2759 int rte_eth_dev_is_valid_port(uint16_t port_id);
2762 * Start specified Rx queue of a port. It is used when rx_deferred_start
2763 * flag of the specified queue is true.
2766 * The port identifier of the Ethernet device
2767 * @param rx_queue_id
2768 * The index of the Rx queue to update the ring.
2769 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2770 * to rte_eth_dev_configure().
2772 * - 0: Success, the receive queue is started.
2773 * - -ENODEV: if *port_id* is invalid.
2774 * - -EINVAL: The queue_id out of range or belong to hairpin.
2775 * - -EIO: if device is removed.
2776 * - -ENOTSUP: The function not supported in PMD.
2778 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2781 * Stop specified Rx queue of a port
2784 * The port identifier of the Ethernet device
2785 * @param rx_queue_id
2786 * The index of the Rx queue to update the ring.
2787 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2788 * to rte_eth_dev_configure().
2790 * - 0: Success, the receive queue is stopped.
2791 * - -ENODEV: if *port_id* is invalid.
2792 * - -EINVAL: The queue_id out of range or belong to hairpin.
2793 * - -EIO: if device is removed.
2794 * - -ENOTSUP: The function not supported in PMD.
2796 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2799 * Start Tx for specified queue of a port. It is used when tx_deferred_start
2800 * flag of the specified queue is true.
2803 * The port identifier of the Ethernet device
2804 * @param tx_queue_id
2805 * The index of the Tx queue to update the ring.
2806 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2807 * to rte_eth_dev_configure().
2809 * - 0: Success, the transmit queue is started.
2810 * - -ENODEV: if *port_id* is invalid.
2811 * - -EINVAL: The queue_id out of range or belong to hairpin.
2812 * - -EIO: if device is removed.
2813 * - -ENOTSUP: The function not supported in PMD.
2815 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2818 * Stop specified Tx queue of a port
2821 * The port identifier of the Ethernet device
2822 * @param tx_queue_id
2823 * The index of the Tx queue to update the ring.
2824 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2825 * to rte_eth_dev_configure().
2827 * - 0: Success, the transmit queue is stopped.
2828 * - -ENODEV: if *port_id* is invalid.
2829 * - -EINVAL: The queue_id out of range or belong to hairpin.
2830 * - -EIO: if device is removed.
2831 * - -ENOTSUP: The function not supported in PMD.
2833 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2836 * Start an Ethernet device.
2838 * The device start step is the last one and consists of setting the configured
2839 * offload features and in starting the transmit and the receive units of the
2842 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before
2843 * PMD port start callback function is invoked.
2845 * On success, all basic functions exported by the Ethernet API (link status,
2846 * receive/transmit, and so on) can be invoked.
2849 * The port identifier of the Ethernet device.
2851 * - 0: Success, Ethernet device started.
2852 * - <0: Error code of the driver device start function.
2854 int rte_eth_dev_start(uint16_t port_id);
2857 * Stop an Ethernet device. The device can be restarted with a call to
2858 * rte_eth_dev_start()
2861 * The port identifier of the Ethernet device.
2863 * - 0: Success, Ethernet device stopped.
2864 * - <0: Error code of the driver device stop function.
2866 int rte_eth_dev_stop(uint16_t port_id);
2869 * Link up an Ethernet device.
2871 * Set device link up will re-enable the device Rx/Tx
2872 * functionality after it is previously set device linked down.
2875 * The port identifier of the Ethernet device.
2877 * - 0: Success, Ethernet device linked up.
2878 * - <0: Error code of the driver device link up function.
2880 int rte_eth_dev_set_link_up(uint16_t port_id);
2883 * Link down an Ethernet device.
2884 * The device Rx/Tx functionality will be disabled if success,
2885 * and it can be re-enabled with a call to
2886 * rte_eth_dev_set_link_up()
2889 * The port identifier of the Ethernet device.
2891 int rte_eth_dev_set_link_down(uint16_t port_id);
2894 * Close a stopped Ethernet device. The device cannot be restarted!
2895 * The function frees all port resources.
2898 * The port identifier of the Ethernet device.
2900 * - Zero if the port is closed successfully.
2901 * - Negative if something went wrong.
2903 int rte_eth_dev_close(uint16_t port_id);
2906 * Reset a Ethernet device and keep its port ID.
2908 * When a port has to be reset passively, the DPDK application can invoke
2909 * this function. For example when a PF is reset, all its VFs should also
2910 * be reset. Normally a DPDK application can invoke this function when
2911 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2912 * a port reset in other circumstances.
2914 * When this function is called, it first stops the port and then calls the
2915 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2916 * state, in which no Tx and Rx queues are setup, as if the port has been
2917 * reset and not started. The port keeps the port ID it had before the
2920 * After calling rte_eth_dev_reset( ), the application should use
2921 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2922 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2923 * to reconfigure the device as appropriate.
2925 * Note: To avoid unexpected behavior, the application should stop calling
2926 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2927 * safety, all these controlling functions should be called from the same
2931 * The port identifier of the Ethernet device.
2934 * - (0) if successful.
2935 * - (-ENODEV) if *port_id* is invalid.
2936 * - (-ENOTSUP) if hardware doesn't support this function.
2937 * - (-EPERM) if not ran from the primary process.
2938 * - (-EIO) if re-initialisation failed or device is removed.
2939 * - (-ENOMEM) if the reset failed due to OOM.
2940 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2942 int rte_eth_dev_reset(uint16_t port_id);
2945 * Enable receipt in promiscuous mode for an Ethernet device.
2948 * The port identifier of the Ethernet device.
2950 * - (0) if successful.
2951 * - (-ENOTSUP) if support for promiscuous_enable() does not exist
2953 * - (-ENODEV) if *port_id* invalid.
2955 int rte_eth_promiscuous_enable(uint16_t port_id);
2958 * Disable receipt in promiscuous mode for an Ethernet device.
2961 * The port identifier of the Ethernet device.
2963 * - (0) if successful.
2964 * - (-ENOTSUP) if support for promiscuous_disable() does not exist
2966 * - (-ENODEV) if *port_id* invalid.
2968 int rte_eth_promiscuous_disable(uint16_t port_id);
2971 * Return the value of promiscuous mode for an Ethernet device.
2974 * The port identifier of the Ethernet device.
2976 * - (1) if promiscuous is enabled
2977 * - (0) if promiscuous is disabled.
2980 int rte_eth_promiscuous_get(uint16_t port_id);
2983 * Enable the receipt of any multicast frame by an Ethernet device.
2986 * The port identifier of the Ethernet device.
2988 * - (0) if successful.
2989 * - (-ENOTSUP) if support for allmulticast_enable() does not exist
2991 * - (-ENODEV) if *port_id* invalid.
2993 int rte_eth_allmulticast_enable(uint16_t port_id);
2996 * Disable the receipt of all multicast frames by an Ethernet device.
2999 * The port identifier of the Ethernet device.
3001 * - (0) if successful.
3002 * - (-ENOTSUP) if support for allmulticast_disable() does not exist
3004 * - (-ENODEV) if *port_id* invalid.
3006 int rte_eth_allmulticast_disable(uint16_t port_id);
3009 * Return the value of allmulticast mode for an Ethernet device.
3012 * The port identifier of the Ethernet device.
3014 * - (1) if allmulticast is enabled
3015 * - (0) if allmulticast is disabled.
3018 int rte_eth_allmulticast_get(uint16_t port_id);
3021 * Retrieve the link status (up/down), the duplex mode (half/full),
3022 * the negotiation (auto/fixed), and if available, the speed (Mbps).
3024 * It might need to wait up to 9 seconds.
3025 * @see rte_eth_link_get_nowait.
3028 * The port identifier of the Ethernet device.
3030 * Link information written back.
3032 * - (0) if successful.
3033 * - (-ENOTSUP) if the function is not supported in PMD.
3034 * - (-ENODEV) if *port_id* invalid.
3035 * - (-EINVAL) if bad parameter.
3037 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
3040 * Retrieve the link status (up/down), the duplex mode (half/full),
3041 * the negotiation (auto/fixed), and if available, the speed (Mbps).
3044 * The port identifier of the Ethernet device.
3046 * Link information written back.
3048 * - (0) if successful.
3049 * - (-ENOTSUP) if the function is not supported in PMD.
3050 * - (-ENODEV) if *port_id* invalid.
3051 * - (-EINVAL) if bad parameter.
3053 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
3057 * @b EXPERIMENTAL: this API may change without prior notice.
3059 * The function converts a link_speed to a string. It handles all special
3060 * values like unknown or none speed.
3063 * link_speed of rte_eth_link struct
3065 * Link speed in textual format. It's pointer to immutable memory.
3066 * No free is required.
3069 const char *rte_eth_link_speed_to_str(uint32_t link_speed);
3073 * @b EXPERIMENTAL: this API may change without prior notice.
3075 * The function converts a rte_eth_link struct representing a link status to
3079 * A pointer to a string to be filled with textual representation of
3080 * device status. At least RTE_ETH_LINK_MAX_STR_LEN bytes should be allocated to
3081 * store default link status text.
3083 * Length of available memory at 'str' string.
3085 * Link status returned by rte_eth_link_get function
3087 * Number of bytes written to str array or -EINVAL if bad parameter.
3090 int rte_eth_link_to_str(char *str, size_t len,
3091 const struct rte_eth_link *eth_link);
3094 * Retrieve the general I/O statistics of an Ethernet device.
3097 * The port identifier of the Ethernet device.
3099 * A pointer to a structure of type *rte_eth_stats* to be filled with
3100 * the values of device counters for the following set of statistics:
3101 * - *ipackets* with the total of successfully received packets.
3102 * - *opackets* with the total of successfully transmitted packets.
3103 * - *ibytes* with the total of successfully received bytes.
3104 * - *obytes* with the total of successfully transmitted bytes.
3105 * - *ierrors* with the total of erroneous received packets.
3106 * - *oerrors* with the total of failed transmitted packets.
3108 * Zero if successful. Non-zero otherwise.
3110 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
3113 * Reset the general I/O statistics of an Ethernet device.
3116 * The port identifier of the Ethernet device.
3118 * - (0) if device notified to reset stats.
3119 * - (-ENOTSUP) if hardware doesn't support.
3120 * - (-ENODEV) if *port_id* invalid.
3121 * - (<0): Error code of the driver stats reset function.
3123 int rte_eth_stats_reset(uint16_t port_id);
3126 * Retrieve names of extended statistics of an Ethernet device.
3128 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
3130 * xstats_names[i].name => xstats[i].value
3132 * And the array index is same with id field of 'struct rte_eth_xstat':
3135 * This assumption makes key-value pair matching less flexible but simpler.
3138 * The port identifier of the Ethernet device.
3139 * @param xstats_names
3140 * An rte_eth_xstat_name array of at least *size* elements to
3141 * be filled. If set to NULL, the function returns the required number
3144 * The size of the xstats_names array (number of elements).
3146 * - A positive value lower or equal to size: success. The return value
3147 * is the number of entries filled in the stats table.
3148 * - A positive value higher than size: error, the given statistics table
3149 * is too small. The return value corresponds to the size that should
3150 * be given to succeed. The entries in the table are not valid and
3151 * shall not be used by the caller.
3152 * - A negative value on error (invalid port ID).
3154 int rte_eth_xstats_get_names(uint16_t port_id,
3155 struct rte_eth_xstat_name *xstats_names,
3159 * Retrieve extended statistics of an Ethernet device.
3161 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched
3163 * xstats_names[i].name => xstats[i].value
3165 * And the array index is same with id field of 'struct rte_eth_xstat':
3168 * This assumption makes key-value pair matching less flexible but simpler.
3171 * The port identifier of the Ethernet device.
3173 * A pointer to a table of structure of type *rte_eth_xstat*
3174 * to be filled with device statistics ids and values.
3175 * This parameter can be set to NULL if n is 0.
3177 * The size of the xstats array (number of elements).
3179 * - A positive value lower or equal to n: success. The return value
3180 * is the number of entries filled in the stats table.
3181 * - A positive value higher than n: error, the given statistics table
3182 * is too small. The return value corresponds to the size that should
3183 * be given to succeed. The entries in the table are not valid and
3184 * shall not be used by the caller.
3185 * - A negative value on error (invalid port ID).
3187 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
3191 * Retrieve names of extended statistics of an Ethernet device.
3194 * The port identifier of the Ethernet device.
3195 * @param xstats_names
3196 * Array to be filled in with names of requested device statistics.
3197 * Must not be NULL if @p ids are specified (not NULL).
3199 * Number of elements in @p xstats_names array (if not NULL) and in
3200 * @p ids array (if not NULL). Must be 0 if both array pointers are NULL.
3202 * IDs array given by app to retrieve specific statistics. May be NULL to
3203 * retrieve names of all available statistics or, if @p xstats_names is
3204 * NULL as well, just the number of available statistics.
3206 * - A positive value lower or equal to size: success. The return value
3207 * is the number of entries filled in the stats table.
3208 * - A positive value higher than size: success. The given statistics table
3209 * is too small. The return value corresponds to the size that should
3210 * be given to succeed. The entries in the table are not valid and
3211 * shall not be used by the caller.
3212 * - A negative value on error.
3215 rte_eth_xstats_get_names_by_id(uint16_t port_id,
3216 struct rte_eth_xstat_name *xstats_names, unsigned int size,
3220 * Retrieve extended statistics of an Ethernet device.
3223 * The port identifier of the Ethernet device.
3225 * IDs array given by app to retrieve specific statistics. May be NULL to
3226 * retrieve all available statistics or, if @p values is NULL as well,
3227 * just the number of available statistics.
3229 * Array to be filled in with requested device statistics.
3230 * Must not be NULL if ids are specified (not NULL).
3232 * Number of elements in @p values array (if not NULL) and in @p ids
3233 * array (if not NULL). Must be 0 if both array pointers are NULL.
3235 * - A positive value lower or equal to size: success. The return value
3236 * is the number of entries filled in the stats table.
3237 * - A positive value higher than size: success: The given statistics table
3238 * is too small. The return value corresponds to the size that should
3239 * be given to succeed. The entries in the table are not valid and
3240 * shall not be used by the caller.
3241 * - A negative value on error.
3243 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
3244 uint64_t *values, unsigned int size);
3247 * Gets the ID of a statistic from its name.
3249 * This function searches for the statistics using string compares, and
3250 * as such should not be used on the fast-path. For fast-path retrieval of
3251 * specific statistics, store the ID as provided in *id* from this function,
3252 * and pass the ID to rte_eth_xstats_get()
3254 * @param port_id The port to look up statistics from
3255 * @param xstat_name The name of the statistic to return
3256 * @param[out] id A pointer to an app-supplied uint64_t which should be
3257 * set to the ID of the stat if the stat exists.
3260 * -ENODEV for invalid port_id,
3261 * -EIO if device is removed,
3262 * -EINVAL if the xstat_name doesn't exist in port_id
3263 * -ENOMEM if bad parameter.
3265 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
3269 * Reset extended statistics of an Ethernet device.
3272 * The port identifier of the Ethernet device.
3274 * - (0) if device notified to reset extended stats.
3275 * - (-ENOTSUP) if pmd doesn't support both
3276 * extended stats and basic stats reset.
3277 * - (-ENODEV) if *port_id* invalid.
3278 * - (<0): Error code of the driver xstats reset function.
3280 int rte_eth_xstats_reset(uint16_t port_id);
3283 * Set a mapping for the specified transmit queue to the specified per-queue
3284 * statistics counter.
3287 * The port identifier of the Ethernet device.
3288 * @param tx_queue_id
3289 * The index of the transmit queue for which a queue stats mapping is required.
3290 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3291 * to rte_eth_dev_configure().
3293 * The per-queue packet statistics functionality number that the transmit
3294 * queue is to be assigned.
3295 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
3296 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256.
3298 * Zero if successful. Non-zero otherwise.
3300 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
3301 uint16_t tx_queue_id, uint8_t stat_idx);
3304 * Set a mapping for the specified receive queue to the specified per-queue
3305 * statistics counter.
3308 * The port identifier of the Ethernet device.
3309 * @param rx_queue_id
3310 * The index of the receive queue for which a queue stats mapping is required.
3311 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3312 * to rte_eth_dev_configure().
3314 * The per-queue packet statistics functionality number that the receive
3315 * queue is to be assigned.
3316 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1].
3317 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256.
3319 * Zero if successful. Non-zero otherwise.
3321 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
3322 uint16_t rx_queue_id,
3326 * Retrieve the Ethernet address of an Ethernet device.
3329 * The port identifier of the Ethernet device.
3331 * A pointer to a structure of type *ether_addr* to be filled with
3332 * the Ethernet address of the Ethernet device.
3334 * - (0) if successful
3335 * - (-ENODEV) if *port_id* invalid.
3336 * - (-EINVAL) if bad parameter.
3338 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr);
3342 * @b EXPERIMENTAL: this API may change without prior notice
3344 * Retrieve the Ethernet addresses of an Ethernet device.
3347 * The port identifier of the Ethernet device.
3349 * A pointer to an array of structures of type *ether_addr* to be filled with
3350 * the Ethernet addresses of the Ethernet device.
3352 * Number of elements in the @p ma array.
3353 * Note that rte_eth_dev_info::max_mac_addrs can be used to retrieve
3354 * max number of Ethernet addresses for given port.
3356 * - number of retrieved addresses if successful
3357 * - (-ENODEV) if *port_id* invalid.
3358 * - (-EINVAL) if bad parameter.
3361 int rte_eth_macaddrs_get(uint16_t port_id, struct rte_ether_addr *ma,
3365 * Retrieve the contextual information of an Ethernet device.
3367 * As part of this function, a number of of fields in dev_info will be
3368 * initialized as follows:
3373 * Where lim is defined within the rte_eth_dev_info_get as
3375 * const struct rte_eth_desc_lim lim = {
3376 * .nb_max = UINT16_MAX,
3379 * .nb_seg_max = UINT16_MAX,
3380 * .nb_mtu_seg_max = UINT16_MAX,
3383 * device = dev->device
3384 * min_mtu = RTE_ETHER_MIN_LEN - RTE_ETHER_HDR_LEN - RTE_ETHER_CRC_LEN
3385 * max_mtu = UINT16_MAX
3387 * The following fields will be populated if support for dev_infos_get()
3388 * exists for the device and the rte_eth_dev 'dev' has been populated
3389 * successfully with a call to it:
3391 * driver_name = dev->device->driver->name
3392 * nb_rx_queues = dev->data->nb_rx_queues
3393 * nb_tx_queues = dev->data->nb_tx_queues
3394 * dev_flags = &dev->data->dev_flags
3397 * The port identifier of the Ethernet device.
3399 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
3400 * the contextual information of the Ethernet device.
3402 * - (0) if successful.
3403 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device.
3404 * - (-ENODEV) if *port_id* invalid.
3405 * - (-EINVAL) if bad parameter.
3407 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
3411 * @b EXPERIMENTAL: this API may change without prior notice.
3413 * Retrieve the configuration of an Ethernet device.
3416 * The port identifier of the Ethernet device.
3418 * Location for Ethernet device configuration to be filled in.
3420 * - (0) if successful.
3421 * - (-ENODEV) if *port_id* invalid.
3422 * - (-EINVAL) if bad parameter.
3425 int rte_eth_dev_conf_get(uint16_t port_id, struct rte_eth_conf *dev_conf);
3428 * Retrieve the firmware version of a device.
3431 * The port identifier of the device.
3433 * A pointer to a string array storing the firmware version of a device,
3434 * the string includes terminating null. This pointer is allocated by caller.
3436 * The size of the string array pointed by fw_version, which should be
3437 * large enough to store firmware version of the device.
3439 * - (0) if successful.
3440 * - (-ENOTSUP) if operation is not supported.
3441 * - (-ENODEV) if *port_id* invalid.
3442 * - (-EIO) if device is removed.
3443 * - (-EINVAL) if bad parameter.
3444 * - (>0) if *fw_size* is not enough to store firmware version, return
3445 * the size of the non truncated string.
3447 int rte_eth_dev_fw_version_get(uint16_t port_id,
3448 char *fw_version, size_t fw_size);
3451 * Retrieve the supported packet types of an Ethernet device.
3453 * When a packet type is announced as supported, it *must* be recognized by
3454 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
3455 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
3456 * packet types for these packets:
3457 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
3458 * - Ether/VLAN/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
3459 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
3460 * - Ether/VLAN/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
3462 * When a packet is received by a PMD, the most precise type must be
3463 * returned among the ones supported. However a PMD is allowed to set
3464 * packet type that is not in the supported list, at the condition that it
3465 * is more precise. Therefore, a PMD announcing no supported packet types
3466 * can still set a matching packet type in a received packet.
3469 * Better to invoke this API after the device is already started or Rx burst
3470 * function is decided, to obtain correct supported ptypes.
3472 * if a given PMD does not report what ptypes it supports, then the supported
3473 * ptype count is reported as 0.
3475 * The port identifier of the Ethernet device.
3477 * A hint of what kind of packet type which the caller is interested in.
3479 * An array pointer to store adequate packet types, allocated by caller.
3481 * Size of the array pointed by param ptypes.
3483 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
3484 * only num entries will be filled into the ptypes array, but the full
3485 * count of supported ptypes will be returned.
3486 * - (-ENODEV) if *port_id* invalid.
3487 * - (-EINVAL) if bad parameter.
3489 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
3490 uint32_t *ptypes, int num);
3492 * Inform Ethernet device about reduced range of packet types to handle.
3494 * Application can use this function to set only specific ptypes that it's
3495 * interested. This information can be used by the PMD to optimize Rx path.
3497 * The function accepts an array `set_ptypes` allocated by the caller to
3498 * store the packet types set by the driver, the last element of the array
3499 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be
3500 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled
3504 * The port identifier of the Ethernet device.
3506 * The ptype family that application is interested in should be bitwise OR of
3507 * RTE_PTYPE_*_MASK or 0.
3509 * An array pointer to store set packet types, allocated by caller. The
3510 * function marks the end of array with RTE_PTYPE_UNKNOWN.
3512 * Size of the array pointed by param ptypes.
3513 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the
3517 * - (-ENODEV) if *port_id* invalid.
3518 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and
3521 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask,
3522 uint32_t *set_ptypes, unsigned int num);
3525 * Retrieve the MTU of an Ethernet device.
3528 * The port identifier of the Ethernet device.
3530 * A pointer to a uint16_t where the retrieved MTU is to be stored.
3532 * - (0) if successful.
3533 * - (-ENODEV) if *port_id* invalid.
3534 * - (-EINVAL) if bad parameter.
3536 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
3539 * Change the MTU of an Ethernet device.
3542 * The port identifier of the Ethernet device.
3544 * A uint16_t for the MTU to be applied.
3546 * - (0) if successful.
3547 * - (-ENOTSUP) if operation is not supported.
3548 * - (-ENODEV) if *port_id* invalid.
3549 * - (-EIO) if device is removed.
3550 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within
3551 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or
3552 * when the mtu is set using dev->dev_ops->mtu_set.
3553 * - (-EBUSY) if operation is not allowed when the port is running
3555 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
3558 * Enable/Disable hardware filtering by an Ethernet device of received
3559 * VLAN packets tagged with a given VLAN Tag Identifier.
3562 * The port identifier of the Ethernet device.
3564 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
3566 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
3567 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
3569 * - (0) if successful.
3570 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
3571 * - (-ENODEV) if *port_id* invalid.
3572 * - (-EIO) if device is removed.
3573 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
3574 * - (-EINVAL) if *vlan_id* > 4095.
3576 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
3579 * Enable/Disable hardware VLAN Strip by a Rx queue of an Ethernet device.
3582 * The port identifier of the Ethernet device.
3583 * @param rx_queue_id
3584 * The index of the receive queue for which a queue stats mapping is required.
3585 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3586 * to rte_eth_dev_configure().
3588 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
3589 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
3591 * - (0) if successful.
3592 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured.
3593 * - (-ENODEV) if *port_id* invalid.
3594 * - (-EINVAL) if *rx_queue_id* invalid.
3596 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
3600 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
3604 * The port identifier of the Ethernet device.
3608 * The Tag Protocol ID
3610 * - (0) if successful.
3611 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported.
3612 * - (-ENODEV) if *port_id* invalid.
3613 * - (-EIO) if device is removed.
3615 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
3616 enum rte_vlan_type vlan_type,
3620 * Set VLAN offload configuration on an Ethernet device.
3623 * The port identifier of the Ethernet device.
3624 * @param offload_mask
3625 * The VLAN Offload bit mask can be mixed use with "OR"
3626 * RTE_ETH_VLAN_STRIP_OFFLOAD
3627 * RTE_ETH_VLAN_FILTER_OFFLOAD
3628 * RTE_ETH_VLAN_EXTEND_OFFLOAD
3629 * RTE_ETH_QINQ_STRIP_OFFLOAD
3631 * - (0) if successful.
3632 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured.
3633 * - (-ENODEV) if *port_id* invalid.
3634 * - (-EIO) if device is removed.
3636 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
3639 * Read VLAN Offload configuration from an Ethernet device
3642 * The port identifier of the Ethernet device.
3644 * - (>0) if successful. Bit mask to indicate
3645 * RTE_ETH_VLAN_STRIP_OFFLOAD
3646 * RTE_ETH_VLAN_FILTER_OFFLOAD
3647 * RTE_ETH_VLAN_EXTEND_OFFLOAD
3648 * RTE_ETH_QINQ_STRIP_OFFLOAD
3649 * - (-ENODEV) if *port_id* invalid.
3651 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
3654 * Set port based Tx VLAN insertion on or off.
3657 * The port identifier of the Ethernet device.
3659 * Port based Tx VLAN identifier together with user priority.
3661 * Turn on or off the port based Tx VLAN insertion.
3664 * - (0) if successful.
3665 * - negative if failed.
3667 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
3669 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
3673 * Structure used to buffer packets for future Tx
3674 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
3676 struct rte_eth_dev_tx_buffer {
3677 buffer_tx_error_fn error_callback;
3678 void *error_userdata;
3679 uint16_t size; /**< Size of buffer for buffered Tx */
3680 uint16_t length; /**< Number of packets in the array */
3681 /** Pending packets to be sent on explicit flush or when full */
3682 struct rte_mbuf *pkts[];
3686 * Calculate the size of the Tx buffer.
3689 * Number of stored packets.
3691 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
3692 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
3695 * Initialize default values for buffered transmitting
3698 * Tx buffer to be initialized.
3705 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
3708 * Configure a callback for buffered packets which cannot be sent
3710 * Register a specific callback to be called when an attempt is made to send
3711 * all packets buffered on an Ethernet port, but not all packets can
3712 * successfully be sent. The callback registered here will be called only
3713 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
3714 * The default callback configured for each queue by default just frees the
3715 * packets back to the calling mempool. If additional behaviour is required,
3716 * for example, to count dropped packets, or to retry transmission of packets
3717 * which cannot be sent, this function should be used to register a suitable
3718 * callback function to implement the desired behaviour.
3719 * The example callback "rte_eth_count_unsent_packet_callback()" is also
3720 * provided as reference.
3723 * The port identifier of the Ethernet device.
3725 * The function to be used as the callback.
3727 * Arbitrary parameter to be passed to the callback function
3729 * 0 on success, or -EINVAL if bad parameter
3732 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
3733 buffer_tx_error_fn callback, void *userdata);
3736 * Callback function for silently dropping unsent buffered packets.
3738 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3739 * adjust the default behavior when buffered packets cannot be sent. This
3740 * function drops any unsent packets silently and is used by Tx buffered
3741 * operations as default behavior.
3743 * NOTE: this function should not be called directly, instead it should be used
3744 * as a callback for packet buffering.
3746 * NOTE: when configuring this function as a callback with
3747 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3748 * should point to an uint64_t value.
3751 * The previously buffered packets which could not be sent
3753 * The number of unsent packets in the pkts array
3758 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
3762 * Callback function for tracking unsent buffered packets.
3764 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3765 * adjust the default behavior when buffered packets cannot be sent. This
3766 * function drops any unsent packets, but also updates a user-supplied counter
3767 * to track the overall number of packets dropped. The counter should be an
3768 * uint64_t variable.
3770 * NOTE: this function should not be called directly, instead it should be used
3771 * as a callback for packet buffering.
3773 * NOTE: when configuring this function as a callback with
3774 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3775 * should point to an uint64_t value.
3778 * The previously buffered packets which could not be sent
3780 * The number of unsent packets in the pkts array
3782 * Pointer to an uint64_t value, which will be incremented by unsent
3785 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
3789 * Request the driver to free mbufs currently cached by the driver. The
3790 * driver will only free the mbuf if it is no longer in use. It is the
3791 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is
3795 * The port identifier of the Ethernet device.
3797 * The index of the transmit queue through which output packets must be
3799 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3800 * to rte_eth_dev_configure().
3802 * Maximum number of packets to free. Use 0 to indicate all possible packets
3803 * should be freed. Note that a packet may be using multiple mbufs.
3806 * -ENODEV: Invalid interface
3807 * -EIO: device is removed
3808 * -ENOTSUP: Driver does not support function
3810 * 0-n: Number of packets freed. More packets may still remain in ring that
3814 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
3817 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
3820 enum rte_eth_event_ipsec_subtype {
3821 /** Unknown event type */
3822 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
3823 /** Sequence number overflow */
3824 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
3825 /** Soft time expiry of SA */
3826 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
3827 /** Soft byte expiry of SA */
3828 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
3829 /** Max value of this enum */
3830 RTE_ETH_EVENT_IPSEC_MAX
3834 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
3835 * information of the IPsec offload event.
3837 struct rte_eth_event_ipsec_desc {
3838 /** Type of RTE_ETH_EVENT_IPSEC_* event */
3839 enum rte_eth_event_ipsec_subtype subtype;
3841 * Event specific metadata.
3843 * For the following events, *userdata* registered
3844 * with the *rte_security_session* would be returned
3847 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
3848 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
3849 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
3851 * @see struct rte_security_session_conf
3858 * The eth device event type for interrupt, and maybe others in the future.
3860 enum rte_eth_event_type {
3861 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3862 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3863 /** queue state event (enabled/disabled) */
3864 RTE_ETH_EVENT_QUEUE_STATE,
3865 /** reset interrupt event, sent to VF on PF reset */
3866 RTE_ETH_EVENT_INTR_RESET,
3867 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3868 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3869 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3870 RTE_ETH_EVENT_NEW, /**< port is probed */
3871 RTE_ETH_EVENT_DESTROY, /**< port is released */
3872 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
3873 RTE_ETH_EVENT_FLOW_AGED,/**< New aged-out flows is detected */
3874 RTE_ETH_EVENT_MAX /**< max value of this enum */
3877 /** User application callback to be registered for interrupts. */
3878 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3879 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3882 * Register a callback function for port event.
3886 * RTE_ETH_ALL means register the event for all port ids.
3890 * User supplied callback function to be called.
3892 * Pointer to the parameters for the registered callback.
3895 * - On success, zero.
3896 * - On failure, a negative value.
3898 int rte_eth_dev_callback_register(uint16_t port_id,
3899 enum rte_eth_event_type event,
3900 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3903 * Unregister a callback function for port event.
3907 * RTE_ETH_ALL means unregister the event for all port ids.
3911 * User supplied callback function to be called.
3913 * Pointer to the parameters for the registered callback. -1 means to
3914 * remove all for the same callback address and same event.
3917 * - On success, zero.
3918 * - On failure, a negative value.
3920 int rte_eth_dev_callback_unregister(uint16_t port_id,
3921 enum rte_eth_event_type event,
3922 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3925 * When there is no Rx packet coming in Rx Queue for a long time, we can
3926 * sleep lcore related to Rx Queue for power saving, and enable Rx interrupt
3927 * to be triggered when Rx packet arrives.
3929 * The rte_eth_dev_rx_intr_enable() function enables Rx queue
3930 * interrupt on specific Rx queue of a port.
3933 * The port identifier of the Ethernet device.
3935 * The index of the receive queue from which to retrieve input packets.
3936 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3937 * to rte_eth_dev_configure().
3939 * - (0) if successful.
3940 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3942 * - (-ENODEV) if *port_id* invalid.
3943 * - (-EIO) if device is removed.
3945 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
3948 * When lcore wakes up from Rx interrupt indicating packet coming, disable Rx
3949 * interrupt and returns to polling mode.
3951 * The rte_eth_dev_rx_intr_disable() function disables Rx queue
3952 * interrupt on specific Rx queue of a port.
3955 * The port identifier of the Ethernet device.
3957 * The index of the receive queue from which to retrieve input packets.
3958 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3959 * to rte_eth_dev_configure().
3961 * - (0) if successful.
3962 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3964 * - (-ENODEV) if *port_id* invalid.
3965 * - (-EIO) if device is removed.
3967 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
3970 * Rx Interrupt control per port.
3973 * The port identifier of the Ethernet device.
3975 * Epoll instance fd which the intr vector associated to.
3976 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3978 * The operation be performed for the vector.
3979 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3983 * - On success, zero.
3984 * - On failure, a negative value.
3986 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
3989 * Rx Interrupt control per queue.
3992 * The port identifier of the Ethernet device.
3994 * The index of the receive queue from which to retrieve input packets.
3995 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3996 * to rte_eth_dev_configure().
3998 * Epoll instance fd which the intr vector associated to.
3999 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
4001 * The operation be performed for the vector.
4002 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
4006 * - On success, zero.
4007 * - On failure, a negative value.
4009 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
4010 int epfd, int op, void *data);
4013 * Get interrupt fd per Rx queue.
4016 * The port identifier of the Ethernet device.
4018 * The index of the receive queue from which to retrieve input packets.
4019 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
4020 * to rte_eth_dev_configure().
4022 * - (>=0) the interrupt fd associated to the requested Rx queue if
4027 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id);
4030 * Turn on the LED on the Ethernet device.
4031 * This function turns on the LED on the Ethernet device.
4034 * The port identifier of the Ethernet device.
4036 * - (0) if successful.
4037 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
4039 * - (-ENODEV) if *port_id* invalid.
4040 * - (-EIO) if device is removed.
4042 int rte_eth_led_on(uint16_t port_id);
4045 * Turn off the LED on the Ethernet device.
4046 * This function turns off the LED on the Ethernet device.
4049 * The port identifier of the Ethernet device.
4051 * - (0) if successful.
4052 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
4054 * - (-ENODEV) if *port_id* invalid.
4055 * - (-EIO) if device is removed.
4057 int rte_eth_led_off(uint16_t port_id);
4061 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4063 * Get Forward Error Correction(FEC) capability.
4066 * The port identifier of the Ethernet device.
4067 * @param speed_fec_capa
4068 * speed_fec_capa is out only with per-speed capabilities.
4069 * If set to NULL, the function returns the required number
4070 * of required array entries.
4072 * a number of elements in an speed_fec_capa array.
4075 * - A non-negative value lower or equal to num: success. The return value
4076 * is the number of entries filled in the fec capa array.
4077 * - A non-negative value higher than num: error, the given fec capa array
4078 * is too small. The return value corresponds to the num that should
4079 * be given to succeed. The entries in fec capa array are not valid and
4080 * shall not be used by the caller.
4081 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4083 * - (-EIO) if device is removed.
4084 * - (-ENODEV) if *port_id* invalid.
4085 * - (-EINVAL) if *num* or *speed_fec_capa* invalid
4088 int rte_eth_fec_get_capability(uint16_t port_id,
4089 struct rte_eth_fec_capa *speed_fec_capa,
4094 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4096 * Get current Forward Error Correction(FEC) mode.
4097 * If link is down and AUTO is enabled, AUTO is returned, otherwise,
4098 * configured FEC mode is returned.
4099 * If link is up, current FEC mode is returned.
4102 * The port identifier of the Ethernet device.
4104 * A bitmask of enabled FEC modes. If AUTO bit is set, other
4105 * bits specify FEC modes which may be negotiated. If AUTO
4106 * bit is clear, specify FEC modes to be used (only one valid
4107 * mode per speed may be set).
4109 * - (0) if successful.
4110 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4112 * - (-EIO) if device is removed.
4113 * - (-ENODEV) if *port_id* invalid.
4116 int rte_eth_fec_get(uint16_t port_id, uint32_t *fec_capa);
4120 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
4122 * Set Forward Error Correction(FEC) mode.
4125 * The port identifier of the Ethernet device.
4127 * A bitmask of allowed FEC modes. If AUTO bit is set, other
4128 * bits specify FEC modes which may be negotiated. If AUTO
4129 * bit is clear, specify FEC modes to be used (only one valid
4130 * mode per speed may be set).
4132 * - (0) if successful.
4133 * - (-EINVAL) if the FEC mode is not valid.
4134 * - (-ENOTSUP) if underlying hardware OR driver doesn't support.
4135 * - (-EIO) if device is removed.
4136 * - (-ENODEV) if *port_id* invalid.
4139 int rte_eth_fec_set(uint16_t port_id, uint32_t fec_capa);
4142 * Get current status of the Ethernet link flow control for Ethernet device
4145 * The port identifier of the Ethernet device.
4147 * The pointer to the structure where to store the flow control parameters.
4149 * - (0) if successful.
4150 * - (-ENOTSUP) if hardware doesn't support flow control.
4151 * - (-ENODEV) if *port_id* invalid.
4152 * - (-EIO) if device is removed.
4153 * - (-EINVAL) if bad parameter.
4155 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
4156 struct rte_eth_fc_conf *fc_conf);
4159 * Configure the Ethernet link flow control for Ethernet device
4162 * The port identifier of the Ethernet device.
4164 * The pointer to the structure of the flow control parameters.
4166 * - (0) if successful.
4167 * - (-ENOTSUP) if hardware doesn't support flow control mode.
4168 * - (-ENODEV) if *port_id* invalid.
4169 * - (-EINVAL) if bad parameter
4170 * - (-EIO) if flow control setup failure or device is removed.
4172 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
4173 struct rte_eth_fc_conf *fc_conf);
4176 * Configure the Ethernet priority flow control under DCB environment
4177 * for Ethernet device.
4180 * The port identifier of the Ethernet device.
4182 * The pointer to the structure of the priority flow control parameters.
4184 * - (0) if successful.
4185 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
4186 * - (-ENODEV) if *port_id* invalid.
4187 * - (-EINVAL) if bad parameter
4188 * - (-EIO) if flow control setup failure or device is removed.
4190 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
4191 struct rte_eth_pfc_conf *pfc_conf);
4194 * Add a MAC address to the set used for filtering incoming packets.
4197 * The port identifier of the Ethernet device.
4199 * The MAC address to add.
4201 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
4202 * not enabled, this should be set to 0.
4204 * - (0) if successfully added or *mac_addr* was already added.
4205 * - (-ENOTSUP) if hardware doesn't support this feature.
4206 * - (-ENODEV) if *port* is invalid.
4207 * - (-EIO) if device is removed.
4208 * - (-ENOSPC) if no more MAC addresses can be added.
4209 * - (-EINVAL) if MAC address is invalid.
4211 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr,
4216 * @b EXPERIMENTAL: this API may change without prior notice.
4218 * Retrieve the information for queue based PFC.
4221 * The port identifier of the Ethernet device.
4222 * @param pfc_queue_info
4223 * A pointer to a structure of type *rte_eth_pfc_queue_info* to be filled with
4224 * the information about queue based PFC.
4226 * - (0) if successful.
4227 * - (-ENOTSUP) if support for priority_flow_ctrl_queue_info_get does not exist.
4228 * - (-ENODEV) if *port_id* invalid.
4229 * - (-EINVAL) if bad parameter.
4232 int rte_eth_dev_priority_flow_ctrl_queue_info_get(uint16_t port_id,
4233 struct rte_eth_pfc_queue_info *pfc_queue_info);
4237 * @b EXPERIMENTAL: this API may change without prior notice.
4239 * Configure the queue based priority flow control for a given queue
4240 * for Ethernet device.
4242 * @note When an ethdev port switches to queue based PFC mode, the
4243 * unconfigured queues shall be configured by the driver with
4244 * default values such as lower priority value for TC etc.
4247 * The port identifier of the Ethernet device.
4248 * @param pfc_queue_conf
4249 * The pointer to the structure of the priority flow control parameters
4252 * - (0) if successful.
4253 * - (-ENOTSUP) if hardware doesn't support queue based PFC mode.
4254 * - (-ENODEV) if *port_id* invalid.
4255 * - (-EINVAL) if bad parameter
4256 * - (-EIO) if flow control setup queue failure
4259 int rte_eth_dev_priority_flow_ctrl_queue_configure(uint16_t port_id,
4260 struct rte_eth_pfc_queue_conf *pfc_queue_conf);
4263 * Remove a MAC address from the internal array of addresses.
4266 * The port identifier of the Ethernet device.
4268 * MAC address to remove.
4270 * - (0) if successful, or *mac_addr* didn't exist.
4271 * - (-ENOTSUP) if hardware doesn't support.
4272 * - (-ENODEV) if *port* invalid.
4273 * - (-EADDRINUSE) if attempting to remove the default MAC address.
4274 * - (-EINVAL) if MAC address is invalid.
4276 int rte_eth_dev_mac_addr_remove(uint16_t port_id,
4277 struct rte_ether_addr *mac_addr);
4280 * Set the default MAC address.
4283 * The port identifier of the Ethernet device.
4285 * New default MAC address.
4287 * - (0) if successful, or *mac_addr* didn't exist.
4288 * - (-ENOTSUP) if hardware doesn't support.
4289 * - (-ENODEV) if *port* invalid.
4290 * - (-EINVAL) if MAC address is invalid.
4292 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
4293 struct rte_ether_addr *mac_addr);
4296 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
4299 * The port identifier of the Ethernet device.
4303 * Redirection table size. The table size can be queried by
4304 * rte_eth_dev_info_get().
4306 * - (0) if successful.
4307 * - (-ENODEV) if *port_id* is invalid.
4308 * - (-ENOTSUP) if hardware doesn't support.
4309 * - (-EINVAL) if bad parameter.
4310 * - (-EIO) if device is removed.
4312 int rte_eth_dev_rss_reta_update(uint16_t port_id,
4313 struct rte_eth_rss_reta_entry64 *reta_conf,
4314 uint16_t reta_size);
4317 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
4320 * The port identifier of the Ethernet device.
4322 * RETA to query. For each requested reta entry, corresponding bit
4323 * in mask must be set.
4325 * Redirection table size. The table size can be queried by
4326 * rte_eth_dev_info_get().
4328 * - (0) if successful.
4329 * - (-ENODEV) if *port_id* is invalid.
4330 * - (-ENOTSUP) if hardware doesn't support.
4331 * - (-EINVAL) if bad parameter.
4332 * - (-EIO) if device is removed.
4334 int rte_eth_dev_rss_reta_query(uint16_t port_id,
4335 struct rte_eth_rss_reta_entry64 *reta_conf,
4336 uint16_t reta_size);
4339 * Updates unicast hash table for receiving packet with the given destination
4340 * MAC address, and the packet is routed to all VFs for which the Rx mode is
4341 * accept packets that match the unicast hash table.
4344 * The port identifier of the Ethernet device.
4346 * Unicast MAC address.
4348 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
4349 * 0 - Clear an unicast hash bit.
4351 * - (0) if successful.
4352 * - (-ENOTSUP) if hardware doesn't support.
4353 * - (-ENODEV) if *port_id* invalid.
4354 * - (-EIO) if device is removed.
4355 * - (-EINVAL) if bad parameter.
4357 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr,
4361 * Updates all unicast hash bitmaps for receiving packet with any Unicast
4362 * Ethernet MAC addresses,the packet is routed to all VFs for which the Rx
4363 * mode is accept packets that match the unicast hash table.
4366 * The port identifier of the Ethernet device.
4368 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
4370 * 0 - Clear all unicast hash bitmaps
4372 * - (0) if successful.
4373 * - (-ENOTSUP) if hardware doesn't support.
4374 * - (-ENODEV) if *port_id* invalid.
4375 * - (-EIO) if device is removed.
4376 * - (-EINVAL) if bad parameter.
4378 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
4381 * Set the rate limitation for a queue on an Ethernet device.
4384 * The port identifier of the Ethernet device.
4388 * The Tx rate in Mbps. Allocated from the total port link speed.
4390 * - (0) if successful.
4391 * - (-ENOTSUP) if hardware doesn't support this feature.
4392 * - (-ENODEV) if *port_id* invalid.
4393 * - (-EIO) if device is removed.
4394 * - (-EINVAL) if bad parameter.
4396 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
4400 * Configuration of Receive Side Scaling hash computation of Ethernet device.
4403 * The port identifier of the Ethernet device.
4405 * The new configuration to use for RSS hash computation on the port.
4407 * - (0) if successful.
4408 * - (-ENODEV) if port identifier is invalid.
4409 * - (-EIO) if device is removed.
4410 * - (-ENOTSUP) if hardware doesn't support.
4411 * - (-EINVAL) if bad parameter.
4413 int rte_eth_dev_rss_hash_update(uint16_t port_id,
4414 struct rte_eth_rss_conf *rss_conf);
4417 * Retrieve current configuration of Receive Side Scaling hash computation
4418 * of Ethernet device.
4421 * The port identifier of the Ethernet device.
4423 * Where to store the current RSS hash configuration of the Ethernet device.
4425 * - (0) if successful.
4426 * - (-ENODEV) if port identifier is invalid.
4427 * - (-EIO) if device is removed.
4428 * - (-ENOTSUP) if hardware doesn't support RSS.
4429 * - (-EINVAL) if bad parameter.
4432 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
4433 struct rte_eth_rss_conf *rss_conf);
4436 * Add UDP tunneling port for a type of tunnel.
4438 * Some NICs may require such configuration to properly parse a tunnel
4439 * with any standard or custom UDP port.
4440 * The packets with this UDP port will be parsed for this type of tunnel.
4441 * The device parser will also check the rest of the tunnel headers
4442 * before classifying the packet.
4444 * With some devices, this API will affect packet classification, i.e.:
4445 * - mbuf.packet_type reported on Rx
4446 * - rte_flow rules with tunnel items
4449 * The port identifier of the Ethernet device.
4451 * UDP tunneling configuration.
4454 * - (0) if successful.
4455 * - (-ENODEV) if port identifier is invalid.
4456 * - (-EIO) if device is removed.
4457 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4460 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
4461 struct rte_eth_udp_tunnel *tunnel_udp);
4464 * Delete UDP tunneling port for a type of tunnel.
4466 * The packets with this UDP port will not be classified as this type of tunnel
4467 * anymore if the device use such mapping for tunnel packet classification.
4469 * @see rte_eth_dev_udp_tunnel_port_add
4472 * The port identifier of the Ethernet device.
4474 * UDP tunneling configuration.
4477 * - (0) if successful.
4478 * - (-ENODEV) if port identifier is invalid.
4479 * - (-EIO) if device is removed.
4480 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4483 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
4484 struct rte_eth_udp_tunnel *tunnel_udp);
4487 * Get DCB information on an Ethernet device.
4490 * The port identifier of the Ethernet device.
4494 * - (0) if successful.
4495 * - (-ENODEV) if port identifier is invalid.
4496 * - (-EIO) if device is removed.
4497 * - (-ENOTSUP) if hardware doesn't support.
4498 * - (-EINVAL) if bad parameter.
4500 int rte_eth_dev_get_dcb_info(uint16_t port_id,
4501 struct rte_eth_dcb_info *dcb_info);
4503 struct rte_eth_rxtx_callback;
4506 * Add a callback to be called on packet Rx on a given port and queue.
4508 * This API configures a function to be called for each burst of
4509 * packets received on a given NIC port queue. The return value is a pointer
4510 * that can be used to later remove the callback using
4511 * rte_eth_remove_rx_callback().
4513 * Multiple functions are called in the order that they are added.
4516 * The port identifier of the Ethernet device.
4518 * The queue on the Ethernet device on which the callback is to be added.
4520 * The callback function
4522 * A generic pointer parameter which will be passed to each invocation of the
4523 * callback function on this port and queue. Inter-thread synchronization
4524 * of any user data changes is the responsibility of the user.
4528 * On success, a pointer value which can later be used to remove the callback.
4530 const struct rte_eth_rxtx_callback *
4531 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
4532 rte_rx_callback_fn fn, void *user_param);
4535 * Add a callback that must be called first on packet Rx on a given port
4538 * This API configures a first function to be called for each burst of
4539 * packets received on a given NIC port queue. The return value is a pointer
4540 * that can be used to later remove the callback using
4541 * rte_eth_remove_rx_callback().
4543 * Multiple functions are called in the order that they are added.
4546 * The port identifier of the Ethernet device.
4548 * The queue on the Ethernet device on which the callback is to be added.
4550 * The callback function
4552 * A generic pointer parameter which will be passed to each invocation of the
4553 * callback function on this port and queue. Inter-thread synchronization
4554 * of any user data changes is the responsibility of the user.
4558 * On success, a pointer value which can later be used to remove the callback.
4560 const struct rte_eth_rxtx_callback *
4561 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
4562 rte_rx_callback_fn fn, void *user_param);
4565 * Add a callback to be called on packet Tx on a given port and queue.
4567 * This API configures a function to be called for each burst of
4568 * packets sent on a given NIC port queue. The return value is a pointer
4569 * that can be used to later remove the callback using
4570 * rte_eth_remove_tx_callback().
4572 * Multiple functions are called in the order that they are added.
4575 * The port identifier of the Ethernet device.
4577 * The queue on the Ethernet device on which the callback is to be added.
4579 * The callback function
4581 * A generic pointer parameter which will be passed to each invocation of the
4582 * callback function on this port and queue. Inter-thread synchronization
4583 * of any user data changes is the responsibility of the user.
4587 * On success, a pointer value which can later be used to remove the callback.
4589 const struct rte_eth_rxtx_callback *
4590 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
4591 rte_tx_callback_fn fn, void *user_param);
4594 * Remove an Rx packet callback from a given port and queue.
4596 * This function is used to removed callbacks that were added to a NIC port
4597 * queue using rte_eth_add_rx_callback().
4599 * Note: the callback is removed from the callback list but it isn't freed
4600 * since the it may still be in use. The memory for the callback can be
4601 * subsequently freed back by the application by calling rte_free():
4603 * - Immediately - if the port is stopped, or the user knows that no
4604 * callbacks are in flight e.g. if called from the thread doing Rx/Tx
4607 * - After a short delay - where the delay is sufficient to allow any
4608 * in-flight callbacks to complete. Alternately, the RCU mechanism can be
4609 * used to detect when data plane threads have ceased referencing the
4613 * The port identifier of the Ethernet device.
4615 * The queue on the Ethernet device from which the callback is to be removed.
4617 * User supplied callback created via rte_eth_add_rx_callback().
4620 * - 0: Success. Callback was removed.
4621 * - -ENODEV: If *port_id* is invalid.
4622 * - -ENOTSUP: Callback support is not available.
4623 * - -EINVAL: The queue_id is out of range, or the callback
4624 * is NULL or not found for the port/queue.
4626 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
4627 const struct rte_eth_rxtx_callback *user_cb);
4630 * Remove a Tx packet callback from a given port and queue.
4632 * This function is used to removed callbacks that were added to a NIC port
4633 * queue using rte_eth_add_tx_callback().
4635 * Note: the callback is removed from the callback list but it isn't freed
4636 * since the it may still be in use. The memory for the callback can be
4637 * subsequently freed back by the application by calling rte_free():
4639 * - Immediately - if the port is stopped, or the user knows that no
4640 * callbacks are in flight e.g. if called from the thread doing Rx/Tx
4643 * - After a short delay - where the delay is sufficient to allow any
4644 * in-flight callbacks to complete. Alternately, the RCU mechanism can be
4645 * used to detect when data plane threads have ceased referencing the
4649 * The port identifier of the Ethernet device.
4651 * The queue on the Ethernet device from which the callback is to be removed.
4653 * User supplied callback created via rte_eth_add_tx_callback().
4656 * - 0: Success. Callback was removed.
4657 * - -ENODEV: If *port_id* is invalid.
4658 * - -ENOTSUP: Callback support is not available.
4659 * - -EINVAL: The queue_id is out of range, or the callback
4660 * is NULL or not found for the port/queue.
4662 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
4663 const struct rte_eth_rxtx_callback *user_cb);
4666 * Retrieve information about given port's Rx queue.
4669 * The port identifier of the Ethernet device.
4671 * The Rx queue on the Ethernet device for which information
4672 * will be retrieved.
4674 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
4675 * the information of the Ethernet device.
4679 * - -ENODEV: If *port_id* is invalid.
4680 * - -ENOTSUP: routine is not supported by the device PMD.
4681 * - -EINVAL: The queue_id is out of range, or the queue
4684 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4685 struct rte_eth_rxq_info *qinfo);
4688 * Retrieve information about given port's Tx queue.
4691 * The port identifier of the Ethernet device.
4693 * The Tx queue on the Ethernet device for which information
4694 * will be retrieved.
4696 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
4697 * the information of the Ethernet device.
4701 * - -ENODEV: If *port_id* is invalid.
4702 * - -ENOTSUP: routine is not supported by the device PMD.
4703 * - -EINVAL: The queue_id is out of range, or the queue
4706 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4707 struct rte_eth_txq_info *qinfo);
4710 * Retrieve information about the Rx packet burst mode.
4713 * The port identifier of the Ethernet device.
4715 * The Rx queue on the Ethernet device for which information
4716 * will be retrieved.
4718 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4719 * with the information of the packet burst mode.
4723 * - -ENODEV: If *port_id* is invalid.
4724 * - -ENOTSUP: routine is not supported by the device PMD.
4725 * - -EINVAL: The queue_id is out of range.
4727 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4728 struct rte_eth_burst_mode *mode);
4731 * Retrieve information about the Tx packet burst mode.
4734 * The port identifier of the Ethernet device.
4736 * The Tx queue on the Ethernet device for which information
4737 * will be retrieved.
4739 * A pointer to a structure of type *rte_eth_burst_mode* to be filled
4740 * with the information of the packet burst mode.
4744 * - -ENODEV: If *port_id* is invalid.
4745 * - -ENOTSUP: routine is not supported by the device PMD.
4746 * - -EINVAL: The queue_id is out of range.
4748 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id,
4749 struct rte_eth_burst_mode *mode);
4753 * @b EXPERIMENTAL: this API may change without prior notice.
4755 * Retrieve the monitor condition for a given receive queue.
4758 * The port identifier of the Ethernet device.
4760 * The Rx queue on the Ethernet device for which information
4761 * will be retrieved.
4763 * The pointer to power-optimized monitoring condition structure.
4767 * -ENOTSUP: Operation not supported.
4768 * -EINVAL: Invalid parameters.
4769 * -ENODEV: Invalid port ID.
4772 int rte_eth_get_monitor_addr(uint16_t port_id, uint16_t queue_id,
4773 struct rte_power_monitor_cond *pmc);
4776 * Retrieve device registers and register attributes (number of registers and
4780 * The port identifier of the Ethernet device.
4782 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
4783 * NULL the function fills in the width and length fields. If non-NULL
4784 * the registers are put into the buffer pointed at by the data field.
4786 * - (0) if successful.
4787 * - (-ENOTSUP) if hardware doesn't support.
4788 * - (-EINVAL) if bad parameter.
4789 * - (-ENODEV) if *port_id* invalid.
4790 * - (-EIO) if device is removed.
4791 * - others depends on the specific operations implementation.
4793 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
4796 * Retrieve size of device EEPROM
4799 * The port identifier of the Ethernet device.
4801 * - (>=0) EEPROM size if successful.
4802 * - (-ENOTSUP) if hardware doesn't support.
4803 * - (-ENODEV) if *port_id* invalid.
4804 * - (-EIO) if device is removed.
4805 * - others depends on the specific operations implementation.
4807 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
4810 * Retrieve EEPROM and EEPROM attribute
4813 * The port identifier of the Ethernet device.
4815 * The template includes buffer for return EEPROM data and
4816 * EEPROM attributes to be filled.
4818 * - (0) if successful.
4819 * - (-ENOTSUP) if hardware doesn't support.
4820 * - (-EINVAL) if bad parameter.
4821 * - (-ENODEV) if *port_id* invalid.
4822 * - (-EIO) if device is removed.
4823 * - others depends on the specific operations implementation.
4825 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4828 * Program EEPROM with provided data
4831 * The port identifier of the Ethernet device.
4833 * The template includes EEPROM data for programming and
4834 * EEPROM attributes to be filled
4836 * - (0) if successful.
4837 * - (-ENOTSUP) if hardware doesn't support.
4838 * - (-ENODEV) if *port_id* invalid.
4839 * - (-EINVAL) if bad parameter.
4840 * - (-EIO) if device is removed.
4841 * - others depends on the specific operations implementation.
4843 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4847 * @b EXPERIMENTAL: this API may change without prior notice.
4849 * Retrieve the type and size of plugin module EEPROM
4852 * The port identifier of the Ethernet device.
4854 * The type and size of plugin module EEPROM.
4856 * - (0) if successful.
4857 * - (-ENOTSUP) if hardware doesn't support.
4858 * - (-ENODEV) if *port_id* invalid.
4859 * - (-EINVAL) if bad parameter.
4860 * - (-EIO) if device is removed.
4861 * - others depends on the specific operations implementation.
4865 rte_eth_dev_get_module_info(uint16_t port_id,
4866 struct rte_eth_dev_module_info *modinfo);
4870 * @b EXPERIMENTAL: this API may change without prior notice.
4872 * Retrieve the data of plugin module EEPROM
4875 * The port identifier of the Ethernet device.
4877 * The template includes the plugin module EEPROM attributes, and the
4878 * buffer for return plugin module EEPROM data.
4880 * - (0) if successful.
4881 * - (-ENOTSUP) if hardware doesn't support.
4882 * - (-EINVAL) if bad parameter.
4883 * - (-ENODEV) if *port_id* invalid.
4884 * - (-EIO) if device is removed.
4885 * - others depends on the specific operations implementation.
4889 rte_eth_dev_get_module_eeprom(uint16_t port_id,
4890 struct rte_dev_eeprom_info *info);
4893 * Set the list of multicast addresses to filter on an Ethernet device.
4896 * The port identifier of the Ethernet device.
4897 * @param mc_addr_set
4898 * The array of multicast addresses to set. Equal to NULL when the function
4899 * is invoked to flush the set of filtered addresses.
4901 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
4902 * when the function is invoked to flush the set of filtered addresses.
4904 * - (0) if successful.
4905 * - (-ENODEV) if *port_id* invalid.
4906 * - (-EIO) if device is removed.
4907 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
4908 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
4909 * - (-EINVAL) if bad parameter.
4911 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
4912 struct rte_ether_addr *mc_addr_set,
4913 uint32_t nb_mc_addr);
4916 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
4919 * The port identifier of the Ethernet device.
4923 * - -ENODEV: The port ID is invalid.
4924 * - -EIO: if device is removed.
4925 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4927 int rte_eth_timesync_enable(uint16_t port_id);
4930 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
4933 * The port identifier of the Ethernet device.
4937 * - -ENODEV: The port ID is invalid.
4938 * - -EIO: if device is removed.
4939 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4941 int rte_eth_timesync_disable(uint16_t port_id);
4944 * Read an IEEE1588/802.1AS Rx timestamp from an Ethernet device.
4947 * The port identifier of the Ethernet device.
4949 * Pointer to the timestamp struct.
4951 * Device specific flags. Used to pass the Rx timesync register index to
4952 * i40e. Unused in igb/ixgbe, pass 0 instead.
4956 * - -EINVAL: No timestamp is available.
4957 * - -ENODEV: The port ID is invalid.
4958 * - -EIO: if device is removed.
4959 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4961 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
4962 struct timespec *timestamp, uint32_t flags);
4965 * Read an IEEE1588/802.1AS Tx timestamp from an Ethernet device.
4968 * The port identifier of the Ethernet device.
4970 * Pointer to the timestamp struct.
4974 * - -EINVAL: No timestamp is available.
4975 * - -ENODEV: The port ID is invalid.
4976 * - -EIO: if device is removed.
4977 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4979 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
4980 struct timespec *timestamp);
4983 * Adjust the timesync clock on an Ethernet device.
4985 * This is usually used in conjunction with other Ethdev timesync functions to
4986 * synchronize the device time using the IEEE1588/802.1AS protocol.
4989 * The port identifier of the Ethernet device.
4991 * The adjustment in nanoseconds.
4995 * - -ENODEV: The port ID is invalid.
4996 * - -EIO: if device is removed.
4997 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4999 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
5002 * Read the time from the timesync clock on an Ethernet device.
5004 * This is usually used in conjunction with other Ethdev timesync functions to
5005 * synchronize the device time using the IEEE1588/802.1AS protocol.
5008 * The port identifier of the Ethernet device.
5010 * Pointer to the timespec struct that holds the time.
5014 * - -EINVAL: Bad parameter.
5016 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
5019 * Set the time of the timesync clock on an Ethernet device.
5021 * This is usually used in conjunction with other Ethdev timesync functions to
5022 * synchronize the device time using the IEEE1588/802.1AS protocol.
5025 * The port identifier of the Ethernet device.
5027 * Pointer to the timespec struct that holds the time.
5031 * - -EINVAL: No timestamp is available.
5032 * - -ENODEV: The port ID is invalid.
5033 * - -EIO: if device is removed.
5034 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5036 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
5040 * @b EXPERIMENTAL: this API may change without prior notice.
5042 * Read the current clock counter of an Ethernet device
5044 * This returns the current raw clock value of an Ethernet device. It is
5045 * a raw amount of ticks, with no given time reference.
5046 * The value returned here is from the same clock than the one
5047 * filling timestamp field of Rx packets when using hardware timestamp
5048 * offload. Therefore it can be used to compute a precise conversion of
5049 * the device clock to the real time.
5051 * E.g, a simple heuristic to derivate the frequency would be:
5052 * uint64_t start, end;
5053 * rte_eth_read_clock(port, start);
5054 * rte_delay_ms(100);
5055 * rte_eth_read_clock(port, end);
5056 * double freq = (end - start) * 10;
5058 * Compute a common reference with:
5059 * uint64_t base_time_sec = current_time();
5060 * uint64_t base_clock;
5061 * rte_eth_read_clock(port, base_clock);
5063 * Then, convert the raw mbuf timestamp with:
5064 * base_time_sec + (double)(*timestamp_dynfield(mbuf) - base_clock) / freq;
5066 * This simple example will not provide a very good accuracy. One must
5067 * at least measure multiple times the frequency and do a regression.
5068 * To avoid deviation from the system time, the common reference can
5069 * be repeated from time to time. The integer division can also be
5070 * converted by a multiplication and a shift for better performance.
5073 * The port identifier of the Ethernet device.
5075 * Pointer to the uint64_t that holds the raw clock value.
5079 * - -ENODEV: The port ID is invalid.
5080 * - -ENOTSUP: The function is not supported by the Ethernet driver.
5081 * - -EINVAL: if bad parameter.
5085 rte_eth_read_clock(uint16_t port_id, uint64_t *clock);
5088 * Get the port ID from device name. The device name should be specified
5090 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0
5091 * - SoC device name, for example- fsl-gmac0
5092 * - vdev dpdk name, for example- net_[pcap0|null0|tap0]
5095 * pci address or name of the device
5097 * pointer to port identifier of the device
5099 * - (0) if successful and port_id is filled.
5100 * - (-ENODEV or -EINVAL) on failure.
5103 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
5106 * Get the device name from port ID. The device name is specified as below:
5107 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0
5108 * - SoC device name, for example- fsl-gmac0
5109 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0]
5112 * Port identifier of the device.
5114 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
5116 * - (0) if successful.
5117 * - (-ENODEV) if *port_id* is invalid.
5118 * - (-EINVAL) on failure.
5121 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
5124 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
5125 * the Ethernet device information, otherwise adjust them to boundaries.
5128 * The port identifier of the Ethernet device.
5130 * A pointer to a uint16_t where the number of receive
5131 * descriptors stored.
5133 * A pointer to a uint16_t where the number of transmit
5134 * descriptors stored.
5136 * - (0) if successful.
5137 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
5139 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
5140 uint16_t *nb_rx_desc,
5141 uint16_t *nb_tx_desc);
5144 * Test if a port supports specific mempool ops.
5147 * Port identifier of the Ethernet device.
5149 * The name of the pool operations to test.
5151 * - 0: best mempool ops choice for this port.
5152 * - 1: mempool ops are supported for this port.
5153 * - -ENOTSUP: mempool ops not supported for this port.
5154 * - -ENODEV: Invalid port Identifier.
5155 * - -EINVAL: Pool param is null.
5158 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
5161 * Get the security context for the Ethernet device.
5164 * Port identifier of the Ethernet device
5167 * - pointer to security context on success.
5170 rte_eth_dev_get_sec_ctx(uint16_t port_id);
5174 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
5176 * Query the device hairpin capabilities.
5179 * The port identifier of the Ethernet device.
5181 * Pointer to a structure that will hold the hairpin capabilities.
5183 * - (0) if successful.
5184 * - (-ENOTSUP) if hardware doesn't support.
5185 * - (-EINVAL) if bad parameter.
5188 int rte_eth_dev_hairpin_capability_get(uint16_t port_id,
5189 struct rte_eth_hairpin_cap *cap);
5193 * @b EXPERIMENTAL: this structure may change without prior notice.
5195 * Ethernet device representor ID range entry
5197 struct rte_eth_representor_range {
5198 enum rte_eth_representor_type type; /**< Representor type */
5199 int controller; /**< Controller index */
5200 int pf; /**< Physical function index */
5203 int vf; /**< VF start index */
5204 int sf; /**< SF start index */
5206 uint32_t id_base; /**< Representor ID start index */
5207 uint32_t id_end; /**< Representor ID end index */
5208 char name[RTE_DEV_NAME_MAX_LEN]; /**< Representor name */
5213 * @b EXPERIMENTAL: this structure may change without prior notice.
5215 * Ethernet device representor information
5217 struct rte_eth_representor_info {
5218 uint16_t controller; /**< Controller ID of caller device. */
5219 uint16_t pf; /**< Physical function ID of caller device. */
5220 uint32_t nb_ranges_alloc; /**< Size of the ranges array. */
5221 uint32_t nb_ranges; /**< Number of initialized ranges. */
5222 struct rte_eth_representor_range ranges[];/**< Representor ID range. */
5226 * Retrieve the representor info of the device.
5228 * Get device representor info to be able to calculate a unique
5229 * representor ID. @see rte_eth_representor_id_get helper.
5232 * The port identifier of the device.
5234 * A pointer to a representor info structure.
5235 * NULL to return number of range entries and allocate memory
5236 * for next call to store detail.
5237 * The number of ranges that were written into this structure
5238 * will be placed into its nb_ranges field. This number cannot be
5239 * larger than the nb_ranges_alloc that by the user before calling
5240 * this function. It can be smaller than the value returned by the
5241 * function, however.
5243 * - (-ENOTSUP) if operation is not supported.
5244 * - (-ENODEV) if *port_id* invalid.
5245 * - (-EIO) if device is removed.
5246 * - (>=0) number of available representor range entries.
5249 int rte_eth_representor_info_get(uint16_t port_id,
5250 struct rte_eth_representor_info *info);
5252 /** The NIC is able to deliver flag (if set) with packets to the PMD. */
5253 #define RTE_ETH_RX_METADATA_USER_FLAG RTE_BIT64(0)
5255 /** The NIC is able to deliver mark ID with packets to the PMD. */
5256 #define RTE_ETH_RX_METADATA_USER_MARK RTE_BIT64(1)
5258 /** The NIC is able to deliver tunnel ID with packets to the PMD. */
5259 #define RTE_ETH_RX_METADATA_TUNNEL_ID RTE_BIT64(2)
5263 * @b EXPERIMENTAL: this API may change without prior notice
5265 * Negotiate the NIC's ability to deliver specific kinds of metadata to the PMD.
5267 * Invoke this API before the first rte_eth_dev_configure() invocation
5268 * to let the PMD make preparations that are inconvenient to do later.
5270 * The negotiation process is as follows:
5272 * - the application requests features intending to use at least some of them;
5273 * - the PMD responds with the guaranteed subset of the requested feature set;
5274 * - the application can retry negotiation with another set of features;
5275 * - the application can pass zero to clear the negotiation result;
5276 * - the last negotiated result takes effect upon
5277 * the ethdev configure and start.
5280 * The PMD is supposed to first consider enabling the requested feature set
5281 * in its entirety. Only if it fails to do so, does it have the right to
5282 * respond with a smaller set of the originally requested features.
5285 * Return code (-ENOTSUP) does not necessarily mean that the requested
5286 * features are unsupported. In this case, the application should just
5287 * assume that these features can be used without prior negotiations.
5290 * Port (ethdev) identifier
5292 * @param[inout] features
5293 * Feature selection buffer
5296 * - (-EBUSY) if the port can't handle this in its current state;
5297 * - (-ENOTSUP) if the method itself is not supported by the PMD;
5298 * - (-ENODEV) if *port_id* is invalid;
5299 * - (-EINVAL) if *features* is NULL;
5300 * - (-EIO) if the device is removed;
5304 int rte_eth_rx_metadata_negotiate(uint16_t port_id, uint64_t *features);
5306 #include <rte_ethdev_core.h>
5310 * Helper routine for rte_eth_rx_burst().
5311 * Should be called at exit from PMD's rte_eth_rx_bulk implementation.
5312 * Does necessary post-processing - invokes Rx callbacks if any, etc.
5315 * The port identifier of the Ethernet device.
5317 * The index of the receive queue from which to retrieve input packets.
5319 * The address of an array of pointers to *rte_mbuf* structures that
5320 * have been retrieved from the device.
5322 * The number of packets that were retrieved from the device.
5324 * The number of elements in @p rx_pkts array.
5326 * Opaque pointer of Rx queue callback related data.
5329 * The number of packets effectively supplied to the @p rx_pkts array.
5331 uint16_t rte_eth_call_rx_callbacks(uint16_t port_id, uint16_t queue_id,
5332 struct rte_mbuf **rx_pkts, uint16_t nb_rx, uint16_t nb_pkts,
5337 * Retrieve a burst of input packets from a receive queue of an Ethernet
5338 * device. The retrieved packets are stored in *rte_mbuf* structures whose
5339 * pointers are supplied in the *rx_pkts* array.
5341 * The rte_eth_rx_burst() function loops, parsing the Rx ring of the
5342 * receive queue, up to *nb_pkts* packets, and for each completed Rx
5343 * descriptor in the ring, it performs the following operations:
5345 * - Initialize the *rte_mbuf* data structure associated with the
5346 * Rx descriptor according to the information provided by the NIC into
5347 * that Rx descriptor.
5349 * - Store the *rte_mbuf* data structure into the next entry of the
5352 * - Replenish the Rx descriptor with a new *rte_mbuf* buffer
5353 * allocated from the memory pool associated with the receive queue at
5354 * initialization time.
5356 * When retrieving an input packet that was scattered by the controller
5357 * into multiple receive descriptors, the rte_eth_rx_burst() function
5358 * appends the associated *rte_mbuf* buffers to the first buffer of the
5361 * The rte_eth_rx_burst() function returns the number of packets
5362 * actually retrieved, which is the number of *rte_mbuf* data structures
5363 * effectively supplied into the *rx_pkts* array.
5364 * A return value equal to *nb_pkts* indicates that the Rx queue contained
5365 * at least *rx_pkts* packets, and this is likely to signify that other
5366 * received packets remain in the input queue. Applications implementing
5367 * a "retrieve as much received packets as possible" policy can check this
5368 * specific case and keep invoking the rte_eth_rx_burst() function until
5369 * a value less than *nb_pkts* is returned.
5371 * This receive method has the following advantages:
5373 * - It allows a run-to-completion network stack engine to retrieve and
5374 * to immediately process received packets in a fast burst-oriented
5375 * approach, avoiding the overhead of unnecessary intermediate packet
5376 * queue/dequeue operations.
5378 * - Conversely, it also allows an asynchronous-oriented processing
5379 * method to retrieve bursts of received packets and to immediately
5380 * queue them for further parallel processing by another logical core,
5381 * for instance. However, instead of having received packets being
5382 * individually queued by the driver, this approach allows the caller
5383 * of the rte_eth_rx_burst() function to queue a burst of retrieved
5384 * packets at a time and therefore dramatically reduce the cost of
5385 * enqueue/dequeue operations per packet.
5387 * - It allows the rte_eth_rx_burst() function of the driver to take
5388 * advantage of burst-oriented hardware features (CPU cache,
5389 * prefetch instructions, and so on) to minimize the number of CPU
5390 * cycles per packet.
5392 * To summarize, the proposed receive API enables many
5393 * burst-oriented optimizations in both synchronous and asynchronous
5394 * packet processing environments with no overhead in both cases.
5397 * Some drivers using vector instructions require that *nb_pkts* is
5398 * divisible by 4 or 8, depending on the driver implementation.
5400 * The rte_eth_rx_burst() function does not provide any error
5401 * notification to avoid the corresponding overhead. As a hint, the
5402 * upper-level application might check the status of the device link once
5403 * being systematically returned a 0 value for a given number of tries.
5406 * The port identifier of the Ethernet device.
5408 * The index of the receive queue from which to retrieve input packets.
5409 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
5410 * to rte_eth_dev_configure().
5412 * The address of an array of pointers to *rte_mbuf* structures that
5413 * must be large enough to store *nb_pkts* pointers in it.
5415 * The maximum number of packets to retrieve.
5416 * The value must be divisible by 8 in order to work with any driver.
5418 * The number of packets actually retrieved, which is the number
5419 * of pointers to *rte_mbuf* structures effectively supplied to the
5422 static inline uint16_t
5423 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
5424 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
5427 struct rte_eth_fp_ops *p;
5430 #ifdef RTE_ETHDEV_DEBUG_RX
5431 if (port_id >= RTE_MAX_ETHPORTS ||
5432 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5434 "Invalid port_id=%u or queue_id=%u\n",
5440 /* fetch pointer to queue data */
5441 p = &rte_eth_fp_ops[port_id];
5442 qd = p->rxq.data[queue_id];
5444 #ifdef RTE_ETHDEV_DEBUG_RX
5445 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
5448 RTE_ETHDEV_LOG(ERR, "Invalid Rx queue_id=%u for port_id=%u\n",
5454 nb_rx = p->rx_pkt_burst(qd, rx_pkts, nb_pkts);
5456 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
5460 /* __ATOMIC_RELEASE memory order was used when the
5461 * call back was inserted into the list.
5462 * Since there is a clear dependency between loading
5463 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
5466 cb = __atomic_load_n((void **)&p->rxq.clbk[queue_id],
5468 if (unlikely(cb != NULL))
5469 nb_rx = rte_eth_call_rx_callbacks(port_id, queue_id,
5470 rx_pkts, nb_rx, nb_pkts, cb);
5474 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx);
5479 * Get the number of used descriptors of a Rx queue
5482 * The port identifier of the Ethernet device.
5484 * The queue ID on the specific port.
5486 * The number of used descriptors in the specific queue, or:
5487 * - (-ENODEV) if *port_id* is invalid.
5488 * (-EINVAL) if *queue_id* is invalid
5489 * (-ENOTSUP) if the device does not support this function
5492 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
5494 struct rte_eth_fp_ops *p;
5497 if (port_id >= RTE_MAX_ETHPORTS ||
5498 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5500 "Invalid port_id=%u or queue_id=%u\n",
5505 /* fetch pointer to queue data */
5506 p = &rte_eth_fp_ops[port_id];
5507 qd = p->rxq.data[queue_id];
5509 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5510 RTE_FUNC_PTR_OR_ERR_RET(*p->rx_queue_count, -ENOTSUP);
5514 return (int)(*p->rx_queue_count)(qd);
5517 /**@{@name Rx hardware descriptor states
5518 * @see rte_eth_rx_descriptor_status
5520 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
5521 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
5522 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
5526 * Check the status of a Rx descriptor in the queue
5528 * It should be called in a similar context than the Rx function:
5529 * - on a dataplane core
5530 * - not concurrently on the same queue
5532 * Since it's a dataplane function, no check is performed on port_id and
5533 * queue_id. The caller must therefore ensure that the port is enabled
5534 * and the queue is configured and running.
5536 * Note: accessing to a random descriptor in the ring may trigger cache
5537 * misses and have a performance impact.
5540 * A valid port identifier of the Ethernet device which.
5542 * A valid Rx queue identifier on this port.
5544 * The offset of the descriptor starting from tail (0 is the next
5545 * packet to be received by the driver).
5548 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
5550 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
5551 * not yet processed by the driver (i.e. in the receive queue).
5552 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
5553 * the driver and not yet returned to hw, or reserved by the hw.
5554 * - (-EINVAL) bad descriptor offset.
5555 * - (-ENOTSUP) if the device does not support this function.
5556 * - (-ENODEV) bad port or queue (only if compiled with debug).
5559 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
5562 struct rte_eth_fp_ops *p;
5565 #ifdef RTE_ETHDEV_DEBUG_RX
5566 if (port_id >= RTE_MAX_ETHPORTS ||
5567 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5569 "Invalid port_id=%u or queue_id=%u\n",
5575 /* fetch pointer to queue data */
5576 p = &rte_eth_fp_ops[port_id];
5577 qd = p->rxq.data[queue_id];
5579 #ifdef RTE_ETHDEV_DEBUG_RX
5580 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5584 RTE_FUNC_PTR_OR_ERR_RET(*p->rx_descriptor_status, -ENOTSUP);
5585 return (*p->rx_descriptor_status)(qd, offset);
5588 /**@{@name Tx hardware descriptor states
5589 * @see rte_eth_tx_descriptor_status
5591 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
5592 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
5593 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
5597 * Check the status of a Tx descriptor in the queue.
5599 * It should be called in a similar context than the Tx function:
5600 * - on a dataplane core
5601 * - not concurrently on the same queue
5603 * Since it's a dataplane function, no check is performed on port_id and
5604 * queue_id. The caller must therefore ensure that the port is enabled
5605 * and the queue is configured and running.
5607 * Note: accessing to a random descriptor in the ring may trigger cache
5608 * misses and have a performance impact.
5611 * A valid port identifier of the Ethernet device which.
5613 * A valid Tx queue identifier on this port.
5615 * The offset of the descriptor starting from tail (0 is the place where
5616 * the next packet will be send).
5619 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
5620 * in the transmit queue.
5621 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
5622 * be reused by the driver.
5623 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
5624 * driver or the hardware.
5625 * - (-EINVAL) bad descriptor offset.
5626 * - (-ENOTSUP) if the device does not support this function.
5627 * - (-ENODEV) bad port or queue (only if compiled with debug).
5629 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
5630 uint16_t queue_id, uint16_t offset)
5632 struct rte_eth_fp_ops *p;
5635 #ifdef RTE_ETHDEV_DEBUG_TX
5636 if (port_id >= RTE_MAX_ETHPORTS ||
5637 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5639 "Invalid port_id=%u or queue_id=%u\n",
5645 /* fetch pointer to queue data */
5646 p = &rte_eth_fp_ops[port_id];
5647 qd = p->txq.data[queue_id];
5649 #ifdef RTE_ETHDEV_DEBUG_TX
5650 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
5654 RTE_FUNC_PTR_OR_ERR_RET(*p->tx_descriptor_status, -ENOTSUP);
5655 return (*p->tx_descriptor_status)(qd, offset);
5660 * Helper routine for rte_eth_tx_burst().
5661 * Should be called before entry PMD's rte_eth_tx_bulk implementation.
5662 * Does necessary pre-processing - invokes Tx callbacks if any, etc.
5665 * The port identifier of the Ethernet device.
5667 * The index of the transmit queue through which output packets must be
5670 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5671 * which contain the output packets.
5673 * The maximum number of packets to transmit.
5675 * The number of output packets to transmit.
5677 uint16_t rte_eth_call_tx_callbacks(uint16_t port_id, uint16_t queue_id,
5678 struct rte_mbuf **tx_pkts, uint16_t nb_pkts, void *opaque);
5681 * Send a burst of output packets on a transmit queue of an Ethernet device.
5683 * The rte_eth_tx_burst() function is invoked to transmit output packets
5684 * on the output queue *queue_id* of the Ethernet device designated by its
5686 * The *nb_pkts* parameter is the number of packets to send which are
5687 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
5688 * allocated from a pool created with rte_pktmbuf_pool_create().
5689 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
5690 * up to the number of transmit descriptors available in the Tx ring of the
5692 * For each packet to send, the rte_eth_tx_burst() function performs
5693 * the following operations:
5695 * - Pick up the next available descriptor in the transmit ring.
5697 * - Free the network buffer previously sent with that descriptor, if any.
5699 * - Initialize the transmit descriptor with the information provided
5700 * in the *rte_mbuf data structure.
5702 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
5703 * the rte_eth_tx_burst() function uses several transmit descriptors
5706 * The rte_eth_tx_burst() function returns the number of packets it
5707 * actually sent. A return value equal to *nb_pkts* means that all packets
5708 * have been sent, and this is likely to signify that other output packets
5709 * could be immediately transmitted again. Applications that implement a
5710 * "send as many packets to transmit as possible" policy can check this
5711 * specific case and keep invoking the rte_eth_tx_burst() function until
5712 * a value less than *nb_pkts* is returned.
5714 * It is the responsibility of the rte_eth_tx_burst() function to
5715 * transparently free the memory buffers of packets previously sent.
5716 * This feature is driven by the *tx_free_thresh* value supplied to the
5717 * rte_eth_dev_configure() function at device configuration time.
5718 * When the number of free Tx descriptors drops below this threshold, the
5719 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
5720 * of those packets whose transmission was effectively completed.
5722 * If the PMD is RTE_ETH_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
5723 * invoke this function concurrently on the same Tx queue without SW lock.
5724 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads
5726 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
5730 * The port identifier of the Ethernet device.
5732 * The index of the transmit queue through which output packets must be
5734 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5735 * to rte_eth_dev_configure().
5737 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5738 * which contain the output packets.
5740 * The maximum number of packets to transmit.
5742 * The number of output packets actually stored in transmit descriptors of
5743 * the transmit ring. The return value can be less than the value of the
5744 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
5746 static inline uint16_t
5747 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
5748 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5750 struct rte_eth_fp_ops *p;
5753 #ifdef RTE_ETHDEV_DEBUG_TX
5754 if (port_id >= RTE_MAX_ETHPORTS ||
5755 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5757 "Invalid port_id=%u or queue_id=%u\n",
5763 /* fetch pointer to queue data */
5764 p = &rte_eth_fp_ops[port_id];
5765 qd = p->txq.data[queue_id];
5767 #ifdef RTE_ETHDEV_DEBUG_TX
5768 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
5771 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n",
5777 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
5781 /* __ATOMIC_RELEASE memory order was used when the
5782 * call back was inserted into the list.
5783 * Since there is a clear dependency between loading
5784 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is
5787 cb = __atomic_load_n((void **)&p->txq.clbk[queue_id],
5789 if (unlikely(cb != NULL))
5790 nb_pkts = rte_eth_call_tx_callbacks(port_id, queue_id,
5791 tx_pkts, nb_pkts, cb);
5795 nb_pkts = p->tx_pkt_burst(qd, tx_pkts, nb_pkts);
5797 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts, nb_pkts);
5802 * Process a burst of output packets on a transmit queue of an Ethernet device.
5804 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
5805 * transmitted on the output queue *queue_id* of the Ethernet device designated
5807 * The *nb_pkts* parameter is the number of packets to be prepared which are
5808 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
5809 * allocated from a pool created with rte_pktmbuf_pool_create().
5810 * For each packet to send, the rte_eth_tx_prepare() function performs
5811 * the following operations:
5813 * - Check if packet meets devices requirements for Tx offloads.
5815 * - Check limitations about number of segments.
5817 * - Check additional requirements when debug is enabled.
5819 * - Update and/or reset required checksums when Tx offload is set for packet.
5821 * Since this function can modify packet data, provided mbufs must be safely
5822 * writable (e.g. modified data cannot be in shared segment).
5824 * The rte_eth_tx_prepare() function returns the number of packets ready to be
5825 * sent. A return value equal to *nb_pkts* means that all packets are valid and
5826 * ready to be sent, otherwise stops processing on the first invalid packet and
5827 * leaves the rest packets untouched.
5829 * When this functionality is not implemented in the driver, all packets are
5830 * are returned untouched.
5833 * The port identifier of the Ethernet device.
5834 * The value must be a valid port ID.
5836 * The index of the transmit queue through which output packets must be
5838 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5839 * to rte_eth_dev_configure().
5841 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
5842 * which contain the output packets.
5844 * The maximum number of packets to process.
5846 * The number of packets correct and ready to be sent. The return value can be
5847 * less than the value of the *tx_pkts* parameter when some packet doesn't
5848 * meet devices requirements with rte_errno set appropriately:
5849 * - EINVAL: offload flags are not correctly set
5850 * - ENOTSUP: the offload feature is not supported by the hardware
5851 * - ENODEV: if *port_id* is invalid (with debug enabled only)
5855 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
5857 static inline uint16_t
5858 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
5859 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5861 struct rte_eth_fp_ops *p;
5864 #ifdef RTE_ETHDEV_DEBUG_TX
5865 if (port_id >= RTE_MAX_ETHPORTS ||
5866 queue_id >= RTE_MAX_QUEUES_PER_PORT) {
5868 "Invalid port_id=%u or queue_id=%u\n",
5875 /* fetch pointer to queue data */
5876 p = &rte_eth_fp_ops[port_id];
5877 qd = p->txq.data[queue_id];
5879 #ifdef RTE_ETHDEV_DEBUG_TX
5880 if (!rte_eth_dev_is_valid_port(port_id)) {
5881 RTE_ETHDEV_LOG(ERR, "Invalid Tx port_id=%u\n", port_id);
5886 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n",
5893 if (!p->tx_pkt_prepare)
5896 return p->tx_pkt_prepare(qd, tx_pkts, nb_pkts);
5902 * Native NOOP operation for compilation targets which doesn't require any
5903 * preparations steps, and functional NOOP may introduce unnecessary performance
5906 * Generally this is not a good idea to turn it on globally and didn't should
5907 * be used if behavior of tx_preparation can change.
5910 static inline uint16_t
5911 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
5912 __rte_unused uint16_t queue_id,
5913 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
5921 * Send any packets queued up for transmission on a port and HW queue
5923 * This causes an explicit flush of packets previously buffered via the
5924 * rte_eth_tx_buffer() function. It returns the number of packets successfully
5925 * sent to the NIC, and calls the error callback for any unsent packets. Unless
5926 * explicitly set up otherwise, the default callback simply frees the unsent
5927 * packets back to the owning mempool.
5930 * The port identifier of the Ethernet device.
5932 * The index of the transmit queue through which output packets must be
5934 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5935 * to rte_eth_dev_configure().
5937 * Buffer of packets to be transmit.
5939 * The number of packets successfully sent to the Ethernet device. The error
5940 * callback is called for any packets which could not be sent.
5942 static inline uint16_t
5943 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
5944 struct rte_eth_dev_tx_buffer *buffer)
5947 uint16_t to_send = buffer->length;
5952 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
5956 /* All packets sent, or to be dealt with by callback below */
5957 if (unlikely(sent != to_send))
5958 buffer->error_callback(&buffer->pkts[sent],
5959 (uint16_t)(to_send - sent),
5960 buffer->error_userdata);
5966 * Buffer a single packet for future transmission on a port and queue
5968 * This function takes a single mbuf/packet and buffers it for later
5969 * transmission on the particular port and queue specified. Once the buffer is
5970 * full of packets, an attempt will be made to transmit all the buffered
5971 * packets. In case of error, where not all packets can be transmitted, a
5972 * callback is called with the unsent packets as a parameter. If no callback
5973 * is explicitly set up, the unsent packets are just freed back to the owning
5974 * mempool. The function returns the number of packets actually sent i.e.
5975 * 0 if no buffer flush occurred, otherwise the number of packets successfully
5979 * The port identifier of the Ethernet device.
5981 * The index of the transmit queue through which output packets must be
5983 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
5984 * to rte_eth_dev_configure().
5986 * Buffer used to collect packets to be sent.
5988 * Pointer to the packet mbuf to be sent.
5990 * 0 = packet has been buffered for later transmission
5991 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
5992 * causing N packets to be sent, and the error callback to be called for
5995 static __rte_always_inline uint16_t
5996 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
5997 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
5999 buffer->pkts[buffer->length++] = tx_pkt;
6000 if (buffer->length < buffer->size)
6003 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
6010 #endif /* _RTE_ETHDEV_H_ */