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 a function allowing
21 * an Ethernet Poll Mode Driver (PMD) to simultaneously register itself as
22 * an Ethernet device driver and as a PCI driver for a set of matching PCI
23 * [Ethernet] devices classes.
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
83 * - flow control settings
84 * - receive mode configuration (promiscuous mode, hardware checksum mode,
85 * RSS/VMDQ settings etc.)
86 * - VLAN filtering configuration
87 * - MAC addresses supplied to MAC address array
88 * - flow director filtering mode (but not filtering rules)
89 * - NIC queue statistics mappings
91 * Any other configuration will not be stored and will need to be re-entered
92 * before a call to rte_eth_dev_start().
94 * Finally, a network application can close an Ethernet device by invoking the
95 * rte_eth_dev_close() function.
97 * Each function of the application Ethernet API invokes a specific function
98 * of the PMD that controls the target device designated by its port
100 * For this purpose, all device-specific functions of an Ethernet driver are
101 * supplied through a set of pointers contained in a generic structure of type
103 * The address of the *eth_dev_ops* structure is stored in the *rte_eth_dev*
104 * structure by the device init function of the Ethernet driver, which is
105 * invoked during the PCI probing phase, as explained earlier.
107 * In other words, each function of the Ethernet API simply retrieves the
108 * *rte_eth_dev* structure associated with the device port identifier and
109 * performs an indirect invocation of the corresponding driver function
110 * supplied in the *eth_dev_ops* structure of the *rte_eth_dev* structure.
112 * For performance reasons, the address of the burst-oriented RX and TX
113 * functions of the Ethernet driver are not contained in the *eth_dev_ops*
114 * structure. Instead, they are directly stored at the beginning of the
115 * *rte_eth_dev* structure to avoid an extra indirect memory access during
118 * RTE ethernet device drivers do not use interrupts for transmitting or
119 * receiving. Instead, Ethernet drivers export Poll-Mode receive and transmit
120 * functions to applications.
121 * Both receive and transmit functions are packet-burst oriented to minimize
122 * their cost per packet through the following optimizations:
124 * - Sharing among multiple packets the incompressible cost of the
125 * invocation of receive/transmit functions.
127 * - Enabling receive/transmit functions to take advantage of burst-oriented
128 * hardware features (L1 cache, prefetch instructions, NIC head/tail
129 * registers) to minimize the number of CPU cycles per packet, for instance,
130 * by avoiding useless read memory accesses to ring descriptors, or by
131 * systematically using arrays of pointers that exactly fit L1 cache line
132 * boundaries and sizes.
134 * The burst-oriented receive function does not provide any error notification,
135 * to avoid the corresponding overhead. As a hint, the upper-level application
136 * might check the status of the device link once being systematically returned
137 * a 0 value by the receive function of the driver for a given number of tries.
146 /* Use this macro to check if LRO API is supported */
147 #define RTE_ETHDEV_HAS_LRO_SUPPORT
150 #include <rte_interrupts.h>
152 #include <rte_devargs.h>
153 #include <rte_errno.h>
154 #include <rte_common.h>
156 #include "rte_ether.h"
157 #include "rte_eth_ctrl.h"
158 #include "rte_dev_info.h"
163 * A structure used to retrieve statistics for an Ethernet port.
164 * Not all statistics fields in struct rte_eth_stats are supported
165 * by any type of network interface card (NIC). If any statistics
166 * field is not supported, its value is 0.
168 struct rte_eth_stats {
169 uint64_t ipackets; /**< Total number of successfully received packets. */
170 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
171 uint64_t ibytes; /**< Total number of successfully received bytes. */
172 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
174 /**< Total of RX packets dropped by the HW,
175 * because there are no available buffer (i.e. RX queues are full).
177 uint64_t ierrors; /**< Total number of erroneous received packets. */
178 uint64_t oerrors; /**< Total number of failed transmitted packets. */
179 uint64_t rx_nombuf; /**< Total number of RX mbuf allocation failures. */
180 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
181 /**< Total number of queue RX packets. */
182 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
183 /**< Total number of queue TX packets. */
184 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
185 /**< Total number of successfully received queue bytes. */
186 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
187 /**< Total number of successfully transmitted queue bytes. */
188 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS];
189 /**< Total number of queue packets received that are dropped. */
193 * Device supported speeds bitmap flags
195 #define ETH_LINK_SPEED_AUTONEG (0 << 0) /**< Autonegotiate (all speeds) */
196 #define ETH_LINK_SPEED_FIXED (1 << 0) /**< Disable autoneg (fixed speed) */
197 #define ETH_LINK_SPEED_10M_HD (1 << 1) /**< 10 Mbps half-duplex */
198 #define ETH_LINK_SPEED_10M (1 << 2) /**< 10 Mbps full-duplex */
199 #define ETH_LINK_SPEED_100M_HD (1 << 3) /**< 100 Mbps half-duplex */
200 #define ETH_LINK_SPEED_100M (1 << 4) /**< 100 Mbps full-duplex */
201 #define ETH_LINK_SPEED_1G (1 << 5) /**< 1 Gbps */
202 #define ETH_LINK_SPEED_2_5G (1 << 6) /**< 2.5 Gbps */
203 #define ETH_LINK_SPEED_5G (1 << 7) /**< 5 Gbps */
204 #define ETH_LINK_SPEED_10G (1 << 8) /**< 10 Gbps */
205 #define ETH_LINK_SPEED_20G (1 << 9) /**< 20 Gbps */
206 #define ETH_LINK_SPEED_25G (1 << 10) /**< 25 Gbps */
207 #define ETH_LINK_SPEED_40G (1 << 11) /**< 40 Gbps */
208 #define ETH_LINK_SPEED_50G (1 << 12) /**< 50 Gbps */
209 #define ETH_LINK_SPEED_56G (1 << 13) /**< 56 Gbps */
210 #define ETH_LINK_SPEED_100G (1 << 14) /**< 100 Gbps */
213 * Ethernet numeric link speeds in Mbps
215 #define ETH_SPEED_NUM_NONE 0 /**< Not defined */
216 #define ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
217 #define ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
218 #define ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
219 #define ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
220 #define ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
221 #define ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
222 #define ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
223 #define ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
224 #define ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
225 #define ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
226 #define ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
227 #define ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
230 * A structure used to retrieve link-level information of an Ethernet port.
233 struct rte_eth_link {
234 uint32_t link_speed; /**< ETH_SPEED_NUM_ */
235 uint16_t link_duplex : 1; /**< ETH_LINK_[HALF/FULL]_DUPLEX */
236 uint16_t link_autoneg : 1; /**< ETH_LINK_[AUTONEG/FIXED] */
237 uint16_t link_status : 1; /**< ETH_LINK_[DOWN/UP] */
238 } __attribute__((aligned(8))); /**< aligned for atomic64 read/write */
240 /* Utility constants */
241 #define ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
242 #define ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
243 #define ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
244 #define ETH_LINK_UP 1 /**< Link is up (see link_status). */
245 #define ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
246 #define ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
249 * A structure used to configure the ring threshold registers of an RX/TX
250 * queue for an Ethernet port.
252 struct rte_eth_thresh {
253 uint8_t pthresh; /**< Ring prefetch threshold. */
254 uint8_t hthresh; /**< Ring host threshold. */
255 uint8_t wthresh; /**< Ring writeback threshold. */
259 * Simple flags are used for rte_eth_conf.rxmode.mq_mode.
261 #define ETH_MQ_RX_RSS_FLAG 0x1
262 #define ETH_MQ_RX_DCB_FLAG 0x2
263 #define ETH_MQ_RX_VMDQ_FLAG 0x4
266 * A set of values to identify what method is to be used to route
267 * packets to multiple queues.
269 enum rte_eth_rx_mq_mode {
270 /** None of DCB,RSS or VMDQ mode */
273 /** For RX side, only RSS is on */
274 ETH_MQ_RX_RSS = ETH_MQ_RX_RSS_FLAG,
275 /** For RX side,only DCB is on. */
276 ETH_MQ_RX_DCB = ETH_MQ_RX_DCB_FLAG,
277 /** Both DCB and RSS enable */
278 ETH_MQ_RX_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG,
280 /** Only VMDQ, no RSS nor DCB */
281 ETH_MQ_RX_VMDQ_ONLY = ETH_MQ_RX_VMDQ_FLAG,
282 /** RSS mode with VMDQ */
283 ETH_MQ_RX_VMDQ_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_VMDQ_FLAG,
284 /** Use VMDQ+DCB to route traffic to queues */
285 ETH_MQ_RX_VMDQ_DCB = ETH_MQ_RX_VMDQ_FLAG | ETH_MQ_RX_DCB_FLAG,
286 /** Enable both VMDQ and DCB in VMDq */
287 ETH_MQ_RX_VMDQ_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG |
292 * for rx mq mode backward compatible
294 #define ETH_RSS ETH_MQ_RX_RSS
295 #define VMDQ_DCB ETH_MQ_RX_VMDQ_DCB
296 #define ETH_DCB_RX ETH_MQ_RX_DCB
299 * A set of values to identify what method is to be used to transmit
300 * packets using multi-TCs.
302 enum rte_eth_tx_mq_mode {
303 ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
304 ETH_MQ_TX_DCB, /**< For TX side,only DCB is on. */
305 ETH_MQ_TX_VMDQ_DCB, /**< For TX side,both DCB and VT is on. */
306 ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
310 * for tx mq mode backward compatible
312 #define ETH_DCB_NONE ETH_MQ_TX_NONE
313 #define ETH_VMDQ_DCB_TX ETH_MQ_TX_VMDQ_DCB
314 #define ETH_DCB_TX ETH_MQ_TX_DCB
317 * A structure used to configure the RX features of an Ethernet port.
319 struct rte_eth_rxmode {
320 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
321 enum rte_eth_rx_mq_mode mq_mode;
322 uint32_t max_rx_pkt_len; /**< Only used if jumbo_frame enabled. */
323 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
325 * Per-port Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
326 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
327 * structure are allowed to be set.
332 * Below bitfield API is obsolete. Application should
333 * enable per-port offloads using the offload field
336 uint16_t header_split : 1, /**< Header Split enable. */
337 hw_ip_checksum : 1, /**< IP/UDP/TCP checksum offload enable. */
338 hw_vlan_filter : 1, /**< VLAN filter enable. */
339 hw_vlan_strip : 1, /**< VLAN strip enable. */
340 hw_vlan_extend : 1, /**< Extended VLAN enable. */
341 jumbo_frame : 1, /**< Jumbo Frame Receipt enable. */
342 hw_strip_crc : 1, /**< Enable CRC stripping by hardware. */
343 enable_scatter : 1, /**< Enable scatter packets rx handler */
344 enable_lro : 1, /**< Enable LRO */
345 hw_timestamp : 1, /**< Enable HW timestamp */
346 security : 1, /**< Enable rte_security offloads */
348 * When set the offload bitfield should be ignored.
349 * Instead per-port Rx offloads should be set on offloads
351 * Per-queue offloads shuold be set on rte_eth_rxq_conf
353 * This bit is temporary till rxmode bitfield offloads API will
356 ignore_offload_bitfield : 1;
360 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
361 * Note that single VLAN is treated the same as inner VLAN.
364 ETH_VLAN_TYPE_UNKNOWN = 0,
365 ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
366 ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
371 * A structure used to describe a vlan filter.
372 * If the bit corresponding to a VID is set, such VID is on.
374 struct rte_vlan_filter_conf {
379 * A structure used to configure the Receive Side Scaling (RSS) feature
380 * of an Ethernet port.
381 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
382 * to an array holding the RSS key to use for hashing specific header
383 * fields of received packets. The length of this array should be indicated
384 * by *rss_key_len* below. Otherwise, a default random hash key is used by
387 * The *rss_key_len* field of the *rss_conf* structure indicates the length
388 * in bytes of the array pointed by *rss_key*. To be compatible, this length
389 * will be checked in i40e only. Others assume 40 bytes to be used as before.
391 * The *rss_hf* field of the *rss_conf* structure indicates the different
392 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
393 * Supplying an *rss_hf* equal to zero disables the RSS feature.
395 struct rte_eth_rss_conf {
396 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
397 uint8_t rss_key_len; /**< hash key length in bytes. */
398 uint64_t rss_hf; /**< Hash functions to apply - see below. */
402 * The RSS offload types are defined based on flow types which are defined
403 * in rte_eth_ctrl.h. Different NIC hardwares may support different RSS offload
404 * types. The supported flow types or RSS offload types can be queried by
405 * rte_eth_dev_info_get().
407 #define ETH_RSS_IPV4 (1ULL << RTE_ETH_FLOW_IPV4)
408 #define ETH_RSS_FRAG_IPV4 (1ULL << RTE_ETH_FLOW_FRAG_IPV4)
409 #define ETH_RSS_NONFRAG_IPV4_TCP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_TCP)
410 #define ETH_RSS_NONFRAG_IPV4_UDP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_UDP)
411 #define ETH_RSS_NONFRAG_IPV4_SCTP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_SCTP)
412 #define ETH_RSS_NONFRAG_IPV4_OTHER (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_OTHER)
413 #define ETH_RSS_IPV6 (1ULL << RTE_ETH_FLOW_IPV6)
414 #define ETH_RSS_FRAG_IPV6 (1ULL << RTE_ETH_FLOW_FRAG_IPV6)
415 #define ETH_RSS_NONFRAG_IPV6_TCP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_TCP)
416 #define ETH_RSS_NONFRAG_IPV6_UDP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_UDP)
417 #define ETH_RSS_NONFRAG_IPV6_SCTP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_SCTP)
418 #define ETH_RSS_NONFRAG_IPV6_OTHER (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_OTHER)
419 #define ETH_RSS_L2_PAYLOAD (1ULL << RTE_ETH_FLOW_L2_PAYLOAD)
420 #define ETH_RSS_IPV6_EX (1ULL << RTE_ETH_FLOW_IPV6_EX)
421 #define ETH_RSS_IPV6_TCP_EX (1ULL << RTE_ETH_FLOW_IPV6_TCP_EX)
422 #define ETH_RSS_IPV6_UDP_EX (1ULL << RTE_ETH_FLOW_IPV6_UDP_EX)
423 #define ETH_RSS_PORT (1ULL << RTE_ETH_FLOW_PORT)
424 #define ETH_RSS_VXLAN (1ULL << RTE_ETH_FLOW_VXLAN)
425 #define ETH_RSS_GENEVE (1ULL << RTE_ETH_FLOW_GENEVE)
426 #define ETH_RSS_NVGRE (1ULL << RTE_ETH_FLOW_NVGRE)
428 #define ETH_RSS_IP ( \
430 ETH_RSS_FRAG_IPV4 | \
431 ETH_RSS_NONFRAG_IPV4_OTHER | \
433 ETH_RSS_FRAG_IPV6 | \
434 ETH_RSS_NONFRAG_IPV6_OTHER | \
437 #define ETH_RSS_UDP ( \
438 ETH_RSS_NONFRAG_IPV4_UDP | \
439 ETH_RSS_NONFRAG_IPV6_UDP | \
442 #define ETH_RSS_TCP ( \
443 ETH_RSS_NONFRAG_IPV4_TCP | \
444 ETH_RSS_NONFRAG_IPV6_TCP | \
447 #define ETH_RSS_SCTP ( \
448 ETH_RSS_NONFRAG_IPV4_SCTP | \
449 ETH_RSS_NONFRAG_IPV6_SCTP)
451 #define ETH_RSS_TUNNEL ( \
457 /**< Mask of valid RSS hash protocols */
458 #define ETH_RSS_PROTO_MASK ( \
460 ETH_RSS_FRAG_IPV4 | \
461 ETH_RSS_NONFRAG_IPV4_TCP | \
462 ETH_RSS_NONFRAG_IPV4_UDP | \
463 ETH_RSS_NONFRAG_IPV4_SCTP | \
464 ETH_RSS_NONFRAG_IPV4_OTHER | \
466 ETH_RSS_FRAG_IPV6 | \
467 ETH_RSS_NONFRAG_IPV6_TCP | \
468 ETH_RSS_NONFRAG_IPV6_UDP | \
469 ETH_RSS_NONFRAG_IPV6_SCTP | \
470 ETH_RSS_NONFRAG_IPV6_OTHER | \
471 ETH_RSS_L2_PAYLOAD | \
473 ETH_RSS_IPV6_TCP_EX | \
474 ETH_RSS_IPV6_UDP_EX | \
481 * Definitions used for redirection table entry size.
482 * Some RSS RETA sizes may not be supported by some drivers, check the
483 * documentation or the description of relevant functions for more details.
485 #define ETH_RSS_RETA_SIZE_64 64
486 #define ETH_RSS_RETA_SIZE_128 128
487 #define ETH_RSS_RETA_SIZE_256 256
488 #define ETH_RSS_RETA_SIZE_512 512
489 #define RTE_RETA_GROUP_SIZE 64
491 /* Definitions used for VMDQ and DCB functionality */
492 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */
493 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
494 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */
495 #define ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
497 /* DCB capability defines */
498 #define ETH_DCB_PG_SUPPORT 0x00000001 /**< Priority Group(ETS) support. */
499 #define ETH_DCB_PFC_SUPPORT 0x00000002 /**< Priority Flow Control support. */
501 /* Definitions used for VLAN Offload functionality */
502 #define ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
503 #define ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
504 #define ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
506 /* Definitions used for mask VLAN setting */
507 #define ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
508 #define ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
509 #define ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
510 #define ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
512 /* Definitions used for receive MAC address */
513 #define ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
515 /* Definitions used for unicast hash */
516 #define ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
518 /* Definitions used for VMDQ pool rx mode setting */
519 #define ETH_VMDQ_ACCEPT_UNTAG 0x0001 /**< accept untagged packets. */
520 #define ETH_VMDQ_ACCEPT_HASH_MC 0x0002 /**< accept packets in multicast table . */
521 #define ETH_VMDQ_ACCEPT_HASH_UC 0x0004 /**< accept packets in unicast table. */
522 #define ETH_VMDQ_ACCEPT_BROADCAST 0x0008 /**< accept broadcast packets. */
523 #define ETH_VMDQ_ACCEPT_MULTICAST 0x0010 /**< multicast promiscuous. */
525 /** Maximum nb. of vlan per mirror rule */
526 #define ETH_MIRROR_MAX_VLANS 64
528 #define ETH_MIRROR_VIRTUAL_POOL_UP 0x01 /**< Virtual Pool uplink Mirroring. */
529 #define ETH_MIRROR_UPLINK_PORT 0x02 /**< Uplink Port Mirroring. */
530 #define ETH_MIRROR_DOWNLINK_PORT 0x04 /**< Downlink Port Mirroring. */
531 #define ETH_MIRROR_VLAN 0x08 /**< VLAN Mirroring. */
532 #define ETH_MIRROR_VIRTUAL_POOL_DOWN 0x10 /**< Virtual Pool downlink Mirroring. */
535 * A structure used to configure VLAN traffic mirror of an Ethernet port.
537 struct rte_eth_vlan_mirror {
538 uint64_t vlan_mask; /**< mask for valid VLAN ID. */
539 /** VLAN ID list for vlan mirroring. */
540 uint16_t vlan_id[ETH_MIRROR_MAX_VLANS];
544 * A structure used to configure traffic mirror of an Ethernet port.
546 struct rte_eth_mirror_conf {
547 uint8_t rule_type; /**< Mirroring rule type */
548 uint8_t dst_pool; /**< Destination pool for this mirror rule. */
549 uint64_t pool_mask; /**< Bitmap of pool for pool mirroring */
550 /** VLAN ID setting for VLAN mirroring. */
551 struct rte_eth_vlan_mirror vlan;
555 * A structure used to configure 64 entries of Redirection Table of the
556 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
557 * more than 64 entries supported by hardware, an array of this structure
560 struct rte_eth_rss_reta_entry64 {
562 /**< Mask bits indicate which entries need to be updated/queried. */
563 uint16_t reta[RTE_RETA_GROUP_SIZE];
564 /**< Group of 64 redirection table entries. */
568 * This enum indicates the possible number of traffic classes
569 * in DCB configurations
571 enum rte_eth_nb_tcs {
572 ETH_4_TCS = 4, /**< 4 TCs with DCB. */
573 ETH_8_TCS = 8 /**< 8 TCs with DCB. */
577 * This enum indicates the possible number of queue pools
578 * in VMDQ configurations.
580 enum rte_eth_nb_pools {
581 ETH_8_POOLS = 8, /**< 8 VMDq pools. */
582 ETH_16_POOLS = 16, /**< 16 VMDq pools. */
583 ETH_32_POOLS = 32, /**< 32 VMDq pools. */
584 ETH_64_POOLS = 64 /**< 64 VMDq pools. */
587 /* This structure may be extended in future. */
588 struct rte_eth_dcb_rx_conf {
589 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
590 /** Traffic class each UP mapped to. */
591 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
594 struct rte_eth_vmdq_dcb_tx_conf {
595 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
596 /** Traffic class each UP mapped to. */
597 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
600 struct rte_eth_dcb_tx_conf {
601 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
602 /** Traffic class each UP mapped to. */
603 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
606 struct rte_eth_vmdq_tx_conf {
607 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
611 * A structure used to configure the VMDQ+DCB feature
612 * of an Ethernet port.
614 * Using this feature, packets are routed to a pool of queues, based
615 * on the vlan id in the vlan tag, and then to a specific queue within
616 * that pool, using the user priority vlan tag field.
618 * A default pool may be used, if desired, to route all traffic which
619 * does not match the vlan filter rules.
621 struct rte_eth_vmdq_dcb_conf {
622 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
623 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
624 uint8_t default_pool; /**< The default pool, if applicable */
625 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
627 uint16_t vlan_id; /**< The vlan id of the received frame */
628 uint64_t pools; /**< Bitmask of pools for packet rx */
629 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
630 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
631 /**< Selects a queue in a pool */
635 * A structure used to configure the VMDQ feature of an Ethernet port when
636 * not combined with the DCB feature.
638 * Using this feature, packets are routed to a pool of queues. By default,
639 * the pool selection is based on the MAC address, the vlan id in the
640 * vlan tag as specified in the pool_map array.
641 * Passing the ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
642 * selection using only the MAC address. MAC address to pool mapping is done
643 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
644 * corresponding to the pool id.
646 * Queue selection within the selected pool will be done using RSS when
647 * it is enabled or revert to the first queue of the pool if not.
649 * A default pool may be used, if desired, to route all traffic which
650 * does not match the vlan filter rules or any pool MAC address.
652 struct rte_eth_vmdq_rx_conf {
653 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
654 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
655 uint8_t default_pool; /**< The default pool, if applicable */
656 uint8_t enable_loop_back; /**< Enable VT loop back */
657 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
658 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
660 uint16_t vlan_id; /**< The vlan id of the received frame */
661 uint64_t pools; /**< Bitmask of pools for packet rx */
662 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
666 * A structure used to configure the TX features of an Ethernet port.
668 struct rte_eth_txmode {
669 enum rte_eth_tx_mq_mode mq_mode; /**< TX multi-queues mode. */
671 * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
672 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
673 * structure are allowed to be set.
677 /* For i40e specifically */
680 uint8_t hw_vlan_reject_tagged : 1,
681 /**< If set, reject sending out tagged pkts */
682 hw_vlan_reject_untagged : 1,
683 /**< If set, reject sending out untagged pkts */
684 hw_vlan_insert_pvid : 1;
685 /**< If set, enable port based VLAN insertion */
689 * A structure used to configure an RX ring of an Ethernet port.
691 struct rte_eth_rxconf {
692 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */
693 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */
694 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
695 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
697 * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
698 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
699 * fields on rte_eth_dev_info structure are allowed to be set.
704 #define ETH_TXQ_FLAGS_NOMULTSEGS 0x0001 /**< nb_segs=1 for all mbufs */
705 #define ETH_TXQ_FLAGS_NOREFCOUNT 0x0002 /**< refcnt can be ignored */
706 #define ETH_TXQ_FLAGS_NOMULTMEMP 0x0004 /**< all bufs come from same mempool */
707 #define ETH_TXQ_FLAGS_NOVLANOFFL 0x0100 /**< disable VLAN offload */
708 #define ETH_TXQ_FLAGS_NOXSUMSCTP 0x0200 /**< disable SCTP checksum offload */
709 #define ETH_TXQ_FLAGS_NOXSUMUDP 0x0400 /**< disable UDP checksum offload */
710 #define ETH_TXQ_FLAGS_NOXSUMTCP 0x0800 /**< disable TCP checksum offload */
711 #define ETH_TXQ_FLAGS_NOOFFLOADS \
712 (ETH_TXQ_FLAGS_NOVLANOFFL | ETH_TXQ_FLAGS_NOXSUMSCTP | \
713 ETH_TXQ_FLAGS_NOXSUMUDP | ETH_TXQ_FLAGS_NOXSUMTCP)
714 #define ETH_TXQ_FLAGS_NOXSUMS \
715 (ETH_TXQ_FLAGS_NOXSUMSCTP | ETH_TXQ_FLAGS_NOXSUMUDP | \
716 ETH_TXQ_FLAGS_NOXSUMTCP)
718 * When set the txq_flags should be ignored,
719 * instead per-queue Tx offloads will be set on offloads field
720 * located on rte_eth_txq_conf struct.
721 * This flag is temporary till the rte_eth_txq_conf.txq_flags
722 * API will be deprecated.
724 #define ETH_TXQ_FLAGS_IGNORE 0x8000
727 * A structure used to configure a TX ring of an Ethernet port.
729 struct rte_eth_txconf {
730 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */
731 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
732 uint16_t tx_free_thresh; /**< Start freeing TX buffers if there are
733 less free descriptors than this value. */
735 uint32_t txq_flags; /**< Set flags for the Tx queue */
736 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
738 * Per-queue Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
739 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
740 * fields on rte_eth_dev_info structure are allowed to be set.
746 * A structure contains information about HW descriptor ring limitations.
748 struct rte_eth_desc_lim {
749 uint16_t nb_max; /**< Max allowed number of descriptors. */
750 uint16_t nb_min; /**< Min allowed number of descriptors. */
751 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
754 * Max allowed number of segments per whole packet.
756 * - For TSO packet this is the total number of data descriptors allowed
759 * @see nb_mtu_seg_max
764 * Max number of segments per one MTU.
766 * - For non-TSO packet, this is the maximum allowed number of segments
767 * in a single transmit packet.
769 * - For TSO packet each segment within the TSO may span up to this
774 uint16_t nb_mtu_seg_max;
778 * This enum indicates the flow control mode
780 enum rte_eth_fc_mode {
781 RTE_FC_NONE = 0, /**< Disable flow control. */
782 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */
783 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */
784 RTE_FC_FULL /**< Enable flow control on both side. */
788 * A structure used to configure Ethernet flow control parameter.
789 * These parameters will be configured into the register of the NIC.
790 * Please refer to the corresponding data sheet for proper value.
792 struct rte_eth_fc_conf {
793 uint32_t high_water; /**< High threshold value to trigger XOFF */
794 uint32_t low_water; /**< Low threshold value to trigger XON */
795 uint16_t pause_time; /**< Pause quota in the Pause frame */
796 uint16_t send_xon; /**< Is XON frame need be sent */
797 enum rte_eth_fc_mode mode; /**< Link flow control mode */
798 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
799 uint8_t autoneg; /**< Use Pause autoneg */
803 * A structure used to configure Ethernet priority flow control parameter.
804 * These parameters will be configured into the register of the NIC.
805 * Please refer to the corresponding data sheet for proper value.
807 struct rte_eth_pfc_conf {
808 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
809 uint8_t priority; /**< VLAN User Priority. */
813 * Memory space that can be configured to store Flow Director filters
814 * in the board memory.
816 enum rte_fdir_pballoc_type {
817 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */
818 RTE_FDIR_PBALLOC_128K, /**< 128k. */
819 RTE_FDIR_PBALLOC_256K, /**< 256k. */
823 * Select report mode of FDIR hash information in RX descriptors.
825 enum rte_fdir_status_mode {
826 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
827 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
828 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
832 * A structure used to configure the Flow Director (FDIR) feature
833 * of an Ethernet port.
835 * If mode is RTE_FDIR_DISABLE, the pballoc value is ignored.
837 struct rte_fdir_conf {
838 enum rte_fdir_mode mode; /**< Flow Director mode. */
839 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
840 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
841 /** RX queue of packets matching a "drop" filter in perfect mode. */
843 struct rte_eth_fdir_masks mask;
844 struct rte_eth_fdir_flex_conf flex_conf;
845 /**< Flex payload configuration. */
849 * UDP tunneling configuration.
850 * Used to config the UDP port for a type of tunnel.
851 * NICs need the UDP port to identify the tunnel type.
852 * Normally a type of tunnel has a default UDP port, this structure can be used
853 * in case if the users want to change or support more UDP port.
855 struct rte_eth_udp_tunnel {
856 uint16_t udp_port; /**< UDP port used for the tunnel. */
857 uint8_t prot_type; /**< Tunnel type. Defined in rte_eth_tunnel_type. */
861 * A structure used to enable/disable specific device interrupts.
863 struct rte_intr_conf {
864 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
866 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
868 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
873 * A structure used to configure an Ethernet port.
874 * Depending upon the RX multi-queue mode, extra advanced
875 * configuration settings may be needed.
877 struct rte_eth_conf {
878 uint32_t link_speeds; /**< bitmap of ETH_LINK_SPEED_XXX of speeds to be
879 used. ETH_LINK_SPEED_FIXED disables link
880 autonegotiation, and a unique speed shall be
881 set. Otherwise, the bitmap defines the set of
882 speeds to be advertised. If the special value
883 ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
884 supported are advertised. */
885 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */
886 struct rte_eth_txmode txmode; /**< Port TX configuration. */
887 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
888 is 0, meaning the loopback mode is disabled.
889 Read the datasheet of given ethernet controller
890 for details. The possible values of this field
891 are defined in implementation of each driver. */
893 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
894 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
895 /**< Port vmdq+dcb configuration. */
896 struct rte_eth_dcb_rx_conf dcb_rx_conf;
897 /**< Port dcb RX configuration. */
898 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
899 /**< Port vmdq RX configuration. */
900 } rx_adv_conf; /**< Port RX filtering configuration. */
902 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
903 /**< Port vmdq+dcb TX configuration. */
904 struct rte_eth_dcb_tx_conf dcb_tx_conf;
905 /**< Port dcb TX configuration. */
906 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
907 /**< Port vmdq TX configuration. */
908 } tx_adv_conf; /**< Port TX DCB configuration (union). */
909 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
910 is needed,and the variable must be set ETH_DCB_PFC_SUPPORT. */
911 uint32_t dcb_capability_en;
912 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. */
913 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */
917 * A structure used to retrieve the contextual information of
918 * an Ethernet device, such as the controlling driver of the device,
919 * its PCI context, etc...
923 * RX offload capabilities of a device.
925 #define DEV_RX_OFFLOAD_VLAN_STRIP 0x00000001
926 #define DEV_RX_OFFLOAD_IPV4_CKSUM 0x00000002
927 #define DEV_RX_OFFLOAD_UDP_CKSUM 0x00000004
928 #define DEV_RX_OFFLOAD_TCP_CKSUM 0x00000008
929 #define DEV_RX_OFFLOAD_TCP_LRO 0x00000010
930 #define DEV_RX_OFFLOAD_QINQ_STRIP 0x00000020
931 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000040
932 #define DEV_RX_OFFLOAD_MACSEC_STRIP 0x00000080
933 #define DEV_RX_OFFLOAD_HEADER_SPLIT 0x00000100
934 #define DEV_RX_OFFLOAD_VLAN_FILTER 0x00000200
935 #define DEV_RX_OFFLOAD_VLAN_EXTEND 0x00000400
936 #define DEV_RX_OFFLOAD_JUMBO_FRAME 0x00000800
937 #define DEV_RX_OFFLOAD_CRC_STRIP 0x00001000
938 #define DEV_RX_OFFLOAD_SCATTER 0x00002000
939 #define DEV_RX_OFFLOAD_TIMESTAMP 0x00004000
940 #define DEV_RX_OFFLOAD_SECURITY 0x00008000
941 #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \
942 DEV_RX_OFFLOAD_UDP_CKSUM | \
943 DEV_RX_OFFLOAD_TCP_CKSUM)
944 #define DEV_RX_OFFLOAD_VLAN (DEV_RX_OFFLOAD_VLAN_STRIP | \
945 DEV_RX_OFFLOAD_VLAN_FILTER | \
946 DEV_RX_OFFLOAD_VLAN_EXTEND)
949 * TX offload capabilities of a device.
951 #define DEV_TX_OFFLOAD_VLAN_INSERT 0x00000001
952 #define DEV_TX_OFFLOAD_IPV4_CKSUM 0x00000002
953 #define DEV_TX_OFFLOAD_UDP_CKSUM 0x00000004
954 #define DEV_TX_OFFLOAD_TCP_CKSUM 0x00000008
955 #define DEV_TX_OFFLOAD_SCTP_CKSUM 0x00000010
956 #define DEV_TX_OFFLOAD_TCP_TSO 0x00000020
957 #define DEV_TX_OFFLOAD_UDP_TSO 0x00000040
958 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000080 /**< Used for tunneling packet. */
959 #define DEV_TX_OFFLOAD_QINQ_INSERT 0x00000100
960 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO 0x00000200 /**< Used for tunneling packet. */
961 #define DEV_TX_OFFLOAD_GRE_TNL_TSO 0x00000400 /**< Used for tunneling packet. */
962 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO 0x00000800 /**< Used for tunneling packet. */
963 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO 0x00001000 /**< Used for tunneling packet. */
964 #define DEV_TX_OFFLOAD_MACSEC_INSERT 0x00002000
965 #define DEV_TX_OFFLOAD_MT_LOCKFREE 0x00004000
966 /**< Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
967 * tx queue without SW lock.
969 #define DEV_TX_OFFLOAD_MULTI_SEGS 0x00008000
970 /**< Device supports multi segment send. */
971 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE 0x00010000
972 /**< Device supports optimization for fast release of mbufs.
973 * When set application must guarantee that per-queue all mbufs comes from
974 * the same mempool and has refcnt = 1.
976 #define DEV_TX_OFFLOAD_SECURITY 0x00020000
978 struct rte_pci_device;
981 * Ethernet device information
983 struct rte_eth_dev_info {
984 struct rte_pci_device *pci_dev; /**< Device PCI information. */
985 const char *driver_name; /**< Device Driver name. */
986 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
987 Use if_indextoname() to translate into an interface name. */
988 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */
989 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */
990 uint16_t max_rx_queues; /**< Maximum number of RX queues. */
991 uint16_t max_tx_queues; /**< Maximum number of TX queues. */
992 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
993 uint32_t max_hash_mac_addrs;
994 /** Maximum number of hash MAC addresses for MTA and UTA. */
995 uint16_t max_vfs; /**< Maximum number of VFs. */
996 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
997 uint64_t rx_offload_capa;
998 /**< Device per port RX offload capabilities. */
999 uint64_t tx_offload_capa;
1000 /**< Device per port TX offload capabilities. */
1001 uint64_t rx_queue_offload_capa;
1002 /**< Device per queue RX offload capabilities. */
1003 uint64_t tx_queue_offload_capa;
1004 /**< Device per queue TX offload capabilities. */
1006 /**< Device redirection table size, the total number of entries. */
1007 uint8_t hash_key_size; /**< Hash key size in bytes */
1008 /** Bit mask of RSS offloads, the bit offset also means flow type */
1009 uint64_t flow_type_rss_offloads;
1010 struct rte_eth_rxconf default_rxconf; /**< Default RX configuration */
1011 struct rte_eth_txconf default_txconf; /**< Default TX configuration */
1012 uint16_t vmdq_queue_base; /**< First queue ID for VMDQ pools. */
1013 uint16_t vmdq_queue_num; /**< Queue number for VMDQ pools. */
1014 uint16_t vmdq_pool_base; /**< First ID of VMDQ pools. */
1015 struct rte_eth_desc_lim rx_desc_lim; /**< RX descriptors limits */
1016 struct rte_eth_desc_lim tx_desc_lim; /**< TX descriptors limits */
1017 uint32_t speed_capa; /**< Supported speeds bitmap (ETH_LINK_SPEED_). */
1018 /** Configured number of rx/tx queues */
1019 uint16_t nb_rx_queues; /**< Number of RX queues. */
1020 uint16_t nb_tx_queues; /**< Number of TX queues. */
1024 * Ethernet device RX queue information structure.
1025 * Used to retieve information about configured queue.
1027 struct rte_eth_rxq_info {
1028 struct rte_mempool *mp; /**< mempool used by that queue. */
1029 struct rte_eth_rxconf conf; /**< queue config parameters. */
1030 uint8_t scattered_rx; /**< scattered packets RX supported. */
1031 uint16_t nb_desc; /**< configured number of RXDs. */
1032 } __rte_cache_min_aligned;
1035 * Ethernet device TX queue information structure.
1036 * Used to retrieve information about configured queue.
1038 struct rte_eth_txq_info {
1039 struct rte_eth_txconf conf; /**< queue config parameters. */
1040 uint16_t nb_desc; /**< configured number of TXDs. */
1041 } __rte_cache_min_aligned;
1043 /** Maximum name length for extended statistics counters */
1044 #define RTE_ETH_XSTATS_NAME_SIZE 64
1047 * An Ethernet device extended statistic structure
1049 * This structure is used by rte_eth_xstats_get() to provide
1050 * statistics that are not provided in the generic *rte_eth_stats*
1052 * It maps a name id, corresponding to an index in the array returned
1053 * by rte_eth_xstats_get_names(), to a statistic value.
1055 struct rte_eth_xstat {
1056 uint64_t id; /**< The index in xstats name array. */
1057 uint64_t value; /**< The statistic counter value. */
1061 * A name element for extended statistics.
1063 * An array of this structure is returned by rte_eth_xstats_get_names().
1064 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1065 * structure references these names by their array index.
1067 struct rte_eth_xstat_name {
1068 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1071 #define ETH_DCB_NUM_TCS 8
1072 #define ETH_MAX_VMDQ_POOL 64
1075 * A structure used to get the information of queue and
1076 * TC mapping on both TX and RX paths.
1078 struct rte_eth_dcb_tc_queue_mapping {
1079 /** rx queues assigned to tc per Pool */
1083 } tc_rxq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1084 /** rx queues assigned to tc per Pool */
1088 } tc_txq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1092 * A structure used to get the information of DCB.
1093 * It includes TC UP mapping and queue TC mapping.
1095 struct rte_eth_dcb_info {
1096 uint8_t nb_tcs; /**< number of TCs */
1097 uint8_t prio_tc[ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
1098 uint8_t tc_bws[ETH_DCB_NUM_TCS]; /**< TX BW percentage for each TC */
1099 /** rx queues assigned to tc */
1100 struct rte_eth_dcb_tc_queue_mapping tc_queue;
1104 * RX/TX queue states
1106 #define RTE_ETH_QUEUE_STATE_STOPPED 0
1107 #define RTE_ETH_QUEUE_STATE_STARTED 1
1111 struct rte_eth_dev_callback;
1112 /** @internal Structure to keep track of registered callbacks */
1113 TAILQ_HEAD(rte_eth_dev_cb_list, rte_eth_dev_callback);
1115 /* Macros to check for valid port */
1116 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
1117 if (!rte_eth_dev_is_valid_port(port_id)) { \
1118 RTE_PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id); \
1123 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
1124 if (!rte_eth_dev_is_valid_port(port_id)) { \
1125 RTE_PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id); \
1131 * l2 tunnel configuration.
1134 /**< l2 tunnel enable mask */
1135 #define ETH_L2_TUNNEL_ENABLE_MASK 0x00000001
1136 /**< l2 tunnel insertion mask */
1137 #define ETH_L2_TUNNEL_INSERTION_MASK 0x00000002
1138 /**< l2 tunnel stripping mask */
1139 #define ETH_L2_TUNNEL_STRIPPING_MASK 0x00000004
1140 /**< l2 tunnel forwarding mask */
1141 #define ETH_L2_TUNNEL_FORWARDING_MASK 0x00000008
1144 * Definitions of all functions exported by an Ethernet driver through the
1145 * the generic structure of type *eth_dev_ops* supplied in the *rte_eth_dev*
1146 * structure associated with an Ethernet device.
1149 typedef int (*eth_dev_configure_t)(struct rte_eth_dev *dev);
1150 /**< @internal Ethernet device configuration. */
1152 typedef int (*eth_dev_start_t)(struct rte_eth_dev *dev);
1153 /**< @internal Function used to start a configured Ethernet device. */
1155 typedef void (*eth_dev_stop_t)(struct rte_eth_dev *dev);
1156 /**< @internal Function used to stop a configured Ethernet device. */
1158 typedef int (*eth_dev_set_link_up_t)(struct rte_eth_dev *dev);
1159 /**< @internal Function used to link up a configured Ethernet device. */
1161 typedef int (*eth_dev_set_link_down_t)(struct rte_eth_dev *dev);
1162 /**< @internal Function used to link down a configured Ethernet device. */
1164 typedef void (*eth_dev_close_t)(struct rte_eth_dev *dev);
1165 /**< @internal Function used to close a configured Ethernet device. */
1167 typedef int (*eth_dev_reset_t)(struct rte_eth_dev *dev);
1168 /** <@internal Function used to reset a configured Ethernet device. */
1170 typedef void (*eth_promiscuous_enable_t)(struct rte_eth_dev *dev);
1171 /**< @internal Function used to enable the RX promiscuous mode of an Ethernet device. */
1173 typedef void (*eth_promiscuous_disable_t)(struct rte_eth_dev *dev);
1174 /**< @internal Function used to disable the RX promiscuous mode of an Ethernet device. */
1176 typedef void (*eth_allmulticast_enable_t)(struct rte_eth_dev *dev);
1177 /**< @internal Enable the receipt of all multicast packets by an Ethernet device. */
1179 typedef void (*eth_allmulticast_disable_t)(struct rte_eth_dev *dev);
1180 /**< @internal Disable the receipt of all multicast packets by an Ethernet device. */
1182 typedef int (*eth_link_update_t)(struct rte_eth_dev *dev,
1183 int wait_to_complete);
1184 /**< @internal Get link speed, duplex mode and state (up/down) of an Ethernet device. */
1186 typedef int (*eth_stats_get_t)(struct rte_eth_dev *dev,
1187 struct rte_eth_stats *igb_stats);
1188 /**< @internal Get global I/O statistics of an Ethernet device. */
1190 typedef void (*eth_stats_reset_t)(struct rte_eth_dev *dev);
1191 /**< @internal Reset global I/O statistics of an Ethernet device to 0. */
1193 typedef int (*eth_xstats_get_t)(struct rte_eth_dev *dev,
1194 struct rte_eth_xstat *stats, unsigned n);
1195 /**< @internal Get extended stats of an Ethernet device. */
1197 typedef int (*eth_xstats_get_by_id_t)(struct rte_eth_dev *dev,
1198 const uint64_t *ids,
1201 /**< @internal Get extended stats of an Ethernet device. */
1203 typedef void (*eth_xstats_reset_t)(struct rte_eth_dev *dev);
1204 /**< @internal Reset extended stats of an Ethernet device. */
1206 typedef int (*eth_xstats_get_names_t)(struct rte_eth_dev *dev,
1207 struct rte_eth_xstat_name *xstats_names, unsigned size);
1208 /**< @internal Get names of extended stats of an Ethernet device. */
1210 typedef int (*eth_xstats_get_names_by_id_t)(struct rte_eth_dev *dev,
1211 struct rte_eth_xstat_name *xstats_names, const uint64_t *ids,
1213 /**< @internal Get names of extended stats of an Ethernet device. */
1215 typedef int (*eth_queue_stats_mapping_set_t)(struct rte_eth_dev *dev,
1219 /**< @internal Set a queue statistics mapping for a tx/rx queue of an Ethernet device. */
1221 typedef void (*eth_dev_infos_get_t)(struct rte_eth_dev *dev,
1222 struct rte_eth_dev_info *dev_info);
1223 /**< @internal Get specific informations of an Ethernet device. */
1225 typedef const uint32_t *(*eth_dev_supported_ptypes_get_t)(struct rte_eth_dev *dev);
1226 /**< @internal Get supported ptypes of an Ethernet device. */
1228 typedef int (*eth_queue_start_t)(struct rte_eth_dev *dev,
1230 /**< @internal Start rx and tx of a queue of an Ethernet device. */
1232 typedef int (*eth_queue_stop_t)(struct rte_eth_dev *dev,
1234 /**< @internal Stop rx and tx of a queue of an Ethernet device. */
1236 typedef int (*eth_rx_queue_setup_t)(struct rte_eth_dev *dev,
1237 uint16_t rx_queue_id,
1238 uint16_t nb_rx_desc,
1239 unsigned int socket_id,
1240 const struct rte_eth_rxconf *rx_conf,
1241 struct rte_mempool *mb_pool);
1242 /**< @internal Set up a receive queue of an Ethernet device. */
1244 typedef int (*eth_tx_queue_setup_t)(struct rte_eth_dev *dev,
1245 uint16_t tx_queue_id,
1246 uint16_t nb_tx_desc,
1247 unsigned int socket_id,
1248 const struct rte_eth_txconf *tx_conf);
1249 /**< @internal Setup a transmit queue of an Ethernet device. */
1251 typedef int (*eth_rx_enable_intr_t)(struct rte_eth_dev *dev,
1252 uint16_t rx_queue_id);
1253 /**< @internal Enable interrupt of a receive queue of an Ethernet device. */
1255 typedef int (*eth_rx_disable_intr_t)(struct rte_eth_dev *dev,
1256 uint16_t rx_queue_id);
1257 /**< @internal Disable interrupt of a receive queue of an Ethernet device. */
1259 typedef void (*eth_queue_release_t)(void *queue);
1260 /**< @internal Release memory resources allocated by given RX/TX queue. */
1262 typedef uint32_t (*eth_rx_queue_count_t)(struct rte_eth_dev *dev,
1263 uint16_t rx_queue_id);
1264 /**< @internal Get number of used descriptors on a receive queue. */
1266 typedef int (*eth_rx_descriptor_done_t)(void *rxq, uint16_t offset);
1267 /**< @internal Check DD bit of specific RX descriptor */
1269 typedef int (*eth_rx_descriptor_status_t)(void *rxq, uint16_t offset);
1270 /**< @internal Check the status of a Rx descriptor */
1272 typedef int (*eth_tx_descriptor_status_t)(void *txq, uint16_t offset);
1273 /**< @internal Check the status of a Tx descriptor */
1275 typedef int (*eth_fw_version_get_t)(struct rte_eth_dev *dev,
1276 char *fw_version, size_t fw_size);
1277 /**< @internal Get firmware information of an Ethernet device. */
1279 typedef int (*eth_tx_done_cleanup_t)(void *txq, uint32_t free_cnt);
1280 /**< @internal Force mbufs to be from TX ring. */
1282 typedef void (*eth_rxq_info_get_t)(struct rte_eth_dev *dev,
1283 uint16_t rx_queue_id, struct rte_eth_rxq_info *qinfo);
1285 typedef void (*eth_txq_info_get_t)(struct rte_eth_dev *dev,
1286 uint16_t tx_queue_id, struct rte_eth_txq_info *qinfo);
1288 typedef int (*mtu_set_t)(struct rte_eth_dev *dev, uint16_t mtu);
1289 /**< @internal Set MTU. */
1291 typedef int (*vlan_filter_set_t)(struct rte_eth_dev *dev,
1294 /**< @internal filtering of a VLAN Tag Identifier by an Ethernet device. */
1296 typedef int (*vlan_tpid_set_t)(struct rte_eth_dev *dev,
1297 enum rte_vlan_type type, uint16_t tpid);
1298 /**< @internal set the outer/inner VLAN-TPID by an Ethernet device. */
1300 typedef int (*vlan_offload_set_t)(struct rte_eth_dev *dev, int mask);
1301 /**< @internal set VLAN offload function by an Ethernet device. */
1303 typedef int (*vlan_pvid_set_t)(struct rte_eth_dev *dev,
1306 /**< @internal set port based TX VLAN insertion by an Ethernet device. */
1308 typedef void (*vlan_strip_queue_set_t)(struct rte_eth_dev *dev,
1309 uint16_t rx_queue_id,
1311 /**< @internal VLAN stripping enable/disable by an queue of Ethernet device. */
1313 typedef uint16_t (*eth_rx_burst_t)(void *rxq,
1314 struct rte_mbuf **rx_pkts,
1316 /**< @internal Retrieve input packets from a receive queue of an Ethernet device. */
1318 typedef uint16_t (*eth_tx_burst_t)(void *txq,
1319 struct rte_mbuf **tx_pkts,
1321 /**< @internal Send output packets on a transmit queue of an Ethernet device. */
1323 typedef uint16_t (*eth_tx_prep_t)(void *txq,
1324 struct rte_mbuf **tx_pkts,
1326 /**< @internal Prepare output packets on a transmit queue of an Ethernet device. */
1328 typedef int (*flow_ctrl_get_t)(struct rte_eth_dev *dev,
1329 struct rte_eth_fc_conf *fc_conf);
1330 /**< @internal Get current flow control parameter on an Ethernet device */
1332 typedef int (*flow_ctrl_set_t)(struct rte_eth_dev *dev,
1333 struct rte_eth_fc_conf *fc_conf);
1334 /**< @internal Setup flow control parameter on an Ethernet device */
1336 typedef int (*priority_flow_ctrl_set_t)(struct rte_eth_dev *dev,
1337 struct rte_eth_pfc_conf *pfc_conf);
1338 /**< @internal Setup priority flow control parameter on an Ethernet device */
1340 typedef int (*reta_update_t)(struct rte_eth_dev *dev,
1341 struct rte_eth_rss_reta_entry64 *reta_conf,
1342 uint16_t reta_size);
1343 /**< @internal Update RSS redirection table on an Ethernet device */
1345 typedef int (*reta_query_t)(struct rte_eth_dev *dev,
1346 struct rte_eth_rss_reta_entry64 *reta_conf,
1347 uint16_t reta_size);
1348 /**< @internal Query RSS redirection table on an Ethernet device */
1350 typedef int (*rss_hash_update_t)(struct rte_eth_dev *dev,
1351 struct rte_eth_rss_conf *rss_conf);
1352 /**< @internal Update RSS hash configuration of an Ethernet device */
1354 typedef int (*rss_hash_conf_get_t)(struct rte_eth_dev *dev,
1355 struct rte_eth_rss_conf *rss_conf);
1356 /**< @internal Get current RSS hash configuration of an Ethernet device */
1358 typedef int (*eth_dev_led_on_t)(struct rte_eth_dev *dev);
1359 /**< @internal Turn on SW controllable LED on an Ethernet device */
1361 typedef int (*eth_dev_led_off_t)(struct rte_eth_dev *dev);
1362 /**< @internal Turn off SW controllable LED on an Ethernet device */
1364 typedef void (*eth_mac_addr_remove_t)(struct rte_eth_dev *dev, uint32_t index);
1365 /**< @internal Remove MAC address from receive address register */
1367 typedef int (*eth_mac_addr_add_t)(struct rte_eth_dev *dev,
1368 struct ether_addr *mac_addr,
1371 /**< @internal Set a MAC address into Receive Address Address Register */
1373 typedef void (*eth_mac_addr_set_t)(struct rte_eth_dev *dev,
1374 struct ether_addr *mac_addr);
1375 /**< @internal Set a MAC address into Receive Address Address Register */
1377 typedef int (*eth_uc_hash_table_set_t)(struct rte_eth_dev *dev,
1378 struct ether_addr *mac_addr,
1380 /**< @internal Set a Unicast Hash bitmap */
1382 typedef int (*eth_uc_all_hash_table_set_t)(struct rte_eth_dev *dev,
1384 /**< @internal Set all Unicast Hash bitmap */
1386 typedef int (*eth_set_queue_rate_limit_t)(struct rte_eth_dev *dev,
1389 /**< @internal Set queue TX rate */
1391 typedef int (*eth_mirror_rule_set_t)(struct rte_eth_dev *dev,
1392 struct rte_eth_mirror_conf *mirror_conf,
1395 /**< @internal Add a traffic mirroring rule on an Ethernet device */
1397 typedef int (*eth_mirror_rule_reset_t)(struct rte_eth_dev *dev,
1399 /**< @internal Remove a traffic mirroring rule on an Ethernet device */
1401 typedef int (*eth_udp_tunnel_port_add_t)(struct rte_eth_dev *dev,
1402 struct rte_eth_udp_tunnel *tunnel_udp);
1403 /**< @internal Add tunneling UDP port */
1405 typedef int (*eth_udp_tunnel_port_del_t)(struct rte_eth_dev *dev,
1406 struct rte_eth_udp_tunnel *tunnel_udp);
1407 /**< @internal Delete tunneling UDP port */
1409 typedef int (*eth_set_mc_addr_list_t)(struct rte_eth_dev *dev,
1410 struct ether_addr *mc_addr_set,
1411 uint32_t nb_mc_addr);
1412 /**< @internal set the list of multicast addresses on an Ethernet device */
1414 typedef int (*eth_timesync_enable_t)(struct rte_eth_dev *dev);
1415 /**< @internal Function used to enable IEEE1588/802.1AS timestamping. */
1417 typedef int (*eth_timesync_disable_t)(struct rte_eth_dev *dev);
1418 /**< @internal Function used to disable IEEE1588/802.1AS timestamping. */
1420 typedef int (*eth_timesync_read_rx_timestamp_t)(struct rte_eth_dev *dev,
1421 struct timespec *timestamp,
1423 /**< @internal Function used to read an RX IEEE1588/802.1AS timestamp. */
1425 typedef int (*eth_timesync_read_tx_timestamp_t)(struct rte_eth_dev *dev,
1426 struct timespec *timestamp);
1427 /**< @internal Function used to read a TX IEEE1588/802.1AS timestamp. */
1429 typedef int (*eth_timesync_adjust_time)(struct rte_eth_dev *dev, int64_t);
1430 /**< @internal Function used to adjust the device clock */
1432 typedef int (*eth_timesync_read_time)(struct rte_eth_dev *dev,
1433 struct timespec *timestamp);
1434 /**< @internal Function used to get time from the device clock. */
1436 typedef int (*eth_timesync_write_time)(struct rte_eth_dev *dev,
1437 const struct timespec *timestamp);
1438 /**< @internal Function used to get time from the device clock */
1440 typedef int (*eth_get_reg_t)(struct rte_eth_dev *dev,
1441 struct rte_dev_reg_info *info);
1442 /**< @internal Retrieve registers */
1444 typedef int (*eth_get_eeprom_length_t)(struct rte_eth_dev *dev);
1445 /**< @internal Retrieve eeprom size */
1447 typedef int (*eth_get_eeprom_t)(struct rte_eth_dev *dev,
1448 struct rte_dev_eeprom_info *info);
1449 /**< @internal Retrieve eeprom data */
1451 typedef int (*eth_set_eeprom_t)(struct rte_eth_dev *dev,
1452 struct rte_dev_eeprom_info *info);
1453 /**< @internal Program eeprom data */
1455 typedef int (*eth_l2_tunnel_eth_type_conf_t)
1456 (struct rte_eth_dev *dev, struct rte_eth_l2_tunnel_conf *l2_tunnel);
1457 /**< @internal config l2 tunnel ether type */
1459 typedef int (*eth_l2_tunnel_offload_set_t)
1460 (struct rte_eth_dev *dev,
1461 struct rte_eth_l2_tunnel_conf *l2_tunnel,
1464 /**< @internal enable/disable the l2 tunnel offload functions */
1467 typedef int (*eth_filter_ctrl_t)(struct rte_eth_dev *dev,
1468 enum rte_filter_type filter_type,
1469 enum rte_filter_op filter_op,
1471 /**< @internal Take operations to assigned filter type on an Ethernet device */
1473 typedef int (*eth_tm_ops_get_t)(struct rte_eth_dev *dev, void *ops);
1474 /**< @internal Get Traffic Management (TM) operations on an Ethernet device */
1476 typedef int (*eth_mtr_ops_get_t)(struct rte_eth_dev *dev, void *ops);
1477 /**< @internal Get Trafffic Metering and Policing (MTR) operations */
1479 typedef int (*eth_get_dcb_info)(struct rte_eth_dev *dev,
1480 struct rte_eth_dcb_info *dcb_info);
1481 /**< @internal Get dcb information on an Ethernet device */
1483 typedef int (*eth_pool_ops_supported_t)(struct rte_eth_dev *dev,
1485 /**< @internal Test if a port supports specific mempool ops */
1488 * @internal A structure containing the functions exported by an Ethernet driver.
1490 struct eth_dev_ops {
1491 eth_dev_configure_t dev_configure; /**< Configure device. */
1492 eth_dev_start_t dev_start; /**< Start device. */
1493 eth_dev_stop_t dev_stop; /**< Stop device. */
1494 eth_dev_set_link_up_t dev_set_link_up; /**< Device link up. */
1495 eth_dev_set_link_down_t dev_set_link_down; /**< Device link down. */
1496 eth_dev_close_t dev_close; /**< Close device. */
1497 eth_dev_reset_t dev_reset; /**< Reset device. */
1498 eth_link_update_t link_update; /**< Get device link state. */
1500 eth_promiscuous_enable_t promiscuous_enable; /**< Promiscuous ON. */
1501 eth_promiscuous_disable_t promiscuous_disable;/**< Promiscuous OFF. */
1502 eth_allmulticast_enable_t allmulticast_enable;/**< RX multicast ON. */
1503 eth_allmulticast_disable_t allmulticast_disable;/**< RX multicast OFF. */
1504 eth_mac_addr_remove_t mac_addr_remove; /**< Remove MAC address. */
1505 eth_mac_addr_add_t mac_addr_add; /**< Add a MAC address. */
1506 eth_mac_addr_set_t mac_addr_set; /**< Set a MAC address. */
1507 eth_set_mc_addr_list_t set_mc_addr_list; /**< set list of mcast addrs. */
1508 mtu_set_t mtu_set; /**< Set MTU. */
1510 eth_stats_get_t stats_get; /**< Get generic device statistics. */
1511 eth_stats_reset_t stats_reset; /**< Reset generic device statistics. */
1512 eth_xstats_get_t xstats_get; /**< Get extended device statistics. */
1513 eth_xstats_reset_t xstats_reset; /**< Reset extended device statistics. */
1514 eth_xstats_get_names_t xstats_get_names;
1515 /**< Get names of extended statistics. */
1516 eth_queue_stats_mapping_set_t queue_stats_mapping_set;
1517 /**< Configure per queue stat counter mapping. */
1519 eth_dev_infos_get_t dev_infos_get; /**< Get device info. */
1520 eth_rxq_info_get_t rxq_info_get; /**< retrieve RX queue information. */
1521 eth_txq_info_get_t txq_info_get; /**< retrieve TX queue information. */
1522 eth_fw_version_get_t fw_version_get; /**< Get firmware version. */
1523 eth_dev_supported_ptypes_get_t dev_supported_ptypes_get;
1524 /**< Get packet types supported and identified by device. */
1526 vlan_filter_set_t vlan_filter_set; /**< Filter VLAN Setup. */
1527 vlan_tpid_set_t vlan_tpid_set; /**< Outer/Inner VLAN TPID Setup. */
1528 vlan_strip_queue_set_t vlan_strip_queue_set; /**< VLAN Stripping on queue. */
1529 vlan_offload_set_t vlan_offload_set; /**< Set VLAN Offload. */
1530 vlan_pvid_set_t vlan_pvid_set; /**< Set port based TX VLAN insertion. */
1532 eth_queue_start_t rx_queue_start;/**< Start RX for a queue. */
1533 eth_queue_stop_t rx_queue_stop; /**< Stop RX for a queue. */
1534 eth_queue_start_t tx_queue_start;/**< Start TX for a queue. */
1535 eth_queue_stop_t tx_queue_stop; /**< Stop TX for a queue. */
1536 eth_rx_queue_setup_t rx_queue_setup;/**< Set up device RX queue. */
1537 eth_queue_release_t rx_queue_release; /**< Release RX queue. */
1538 eth_rx_queue_count_t rx_queue_count;
1539 /**< Get the number of used RX descriptors. */
1540 eth_rx_descriptor_done_t rx_descriptor_done; /**< Check rxd DD bit. */
1541 eth_rx_descriptor_status_t rx_descriptor_status;
1542 /**< Check the status of a Rx descriptor. */
1543 eth_tx_descriptor_status_t tx_descriptor_status;
1544 /**< Check the status of a Tx descriptor. */
1545 eth_rx_enable_intr_t rx_queue_intr_enable; /**< Enable Rx queue interrupt. */
1546 eth_rx_disable_intr_t rx_queue_intr_disable; /**< Disable Rx queue interrupt. */
1547 eth_tx_queue_setup_t tx_queue_setup;/**< Set up device TX queue. */
1548 eth_queue_release_t tx_queue_release; /**< Release TX queue. */
1549 eth_tx_done_cleanup_t tx_done_cleanup;/**< Free tx ring mbufs */
1551 eth_dev_led_on_t dev_led_on; /**< Turn on LED. */
1552 eth_dev_led_off_t dev_led_off; /**< Turn off LED. */
1554 flow_ctrl_get_t flow_ctrl_get; /**< Get flow control. */
1555 flow_ctrl_set_t flow_ctrl_set; /**< Setup flow control. */
1556 priority_flow_ctrl_set_t priority_flow_ctrl_set; /**< Setup priority flow control. */
1558 eth_uc_hash_table_set_t uc_hash_table_set; /**< Set Unicast Table Array. */
1559 eth_uc_all_hash_table_set_t uc_all_hash_table_set; /**< Set Unicast hash bitmap. */
1561 eth_mirror_rule_set_t mirror_rule_set; /**< Add a traffic mirror rule. */
1562 eth_mirror_rule_reset_t mirror_rule_reset; /**< reset a traffic mirror rule. */
1564 eth_udp_tunnel_port_add_t udp_tunnel_port_add; /** Add UDP tunnel port. */
1565 eth_udp_tunnel_port_del_t udp_tunnel_port_del; /** Del UDP tunnel port. */
1566 eth_l2_tunnel_eth_type_conf_t l2_tunnel_eth_type_conf;
1567 /** Config ether type of l2 tunnel. */
1568 eth_l2_tunnel_offload_set_t l2_tunnel_offload_set;
1569 /** Enable/disable l2 tunnel offload functions. */
1571 eth_set_queue_rate_limit_t set_queue_rate_limit; /**< Set queue rate limit. */
1573 rss_hash_update_t rss_hash_update; /** Configure RSS hash protocols. */
1574 rss_hash_conf_get_t rss_hash_conf_get; /** Get current RSS hash configuration. */
1575 reta_update_t reta_update; /** Update redirection table. */
1576 reta_query_t reta_query; /** Query redirection table. */
1578 eth_get_reg_t get_reg; /**< Get registers. */
1579 eth_get_eeprom_length_t get_eeprom_length; /**< Get eeprom length. */
1580 eth_get_eeprom_t get_eeprom; /**< Get eeprom data. */
1581 eth_set_eeprom_t set_eeprom; /**< Set eeprom. */
1584 eth_filter_ctrl_t filter_ctrl; /**< common filter control. */
1586 eth_get_dcb_info get_dcb_info; /** Get DCB information. */
1588 eth_timesync_enable_t timesync_enable;
1589 /** Turn IEEE1588/802.1AS timestamping on. */
1590 eth_timesync_disable_t timesync_disable;
1591 /** Turn IEEE1588/802.1AS timestamping off. */
1592 eth_timesync_read_rx_timestamp_t timesync_read_rx_timestamp;
1593 /** Read the IEEE1588/802.1AS RX timestamp. */
1594 eth_timesync_read_tx_timestamp_t timesync_read_tx_timestamp;
1595 /** Read the IEEE1588/802.1AS TX timestamp. */
1596 eth_timesync_adjust_time timesync_adjust_time; /** Adjust the device clock. */
1597 eth_timesync_read_time timesync_read_time; /** Get the device clock time. */
1598 eth_timesync_write_time timesync_write_time; /** Set the device clock time. */
1600 eth_xstats_get_by_id_t xstats_get_by_id;
1601 /**< Get extended device statistic values by ID. */
1602 eth_xstats_get_names_by_id_t xstats_get_names_by_id;
1603 /**< Get name of extended device statistics by ID. */
1605 eth_tm_ops_get_t tm_ops_get;
1606 /**< Get Traffic Management (TM) operations. */
1608 eth_mtr_ops_get_t mtr_ops_get;
1609 /**< Get Traffic Metering and Policing (MTR) operations. */
1611 eth_pool_ops_supported_t pool_ops_supported;
1612 /**< Test if a port supports specific mempool ops */
1616 * Function type used for RX packet processing packet callbacks.
1618 * The callback function is called on RX with a burst of packets that have
1619 * been received on the given port and queue.
1622 * The Ethernet port on which RX is being performed.
1624 * The queue on the Ethernet port which is being used to receive the packets.
1626 * The burst of packets that have just been received.
1628 * The number of packets in the burst pointed to by "pkts".
1630 * The max number of packets that can be stored in the "pkts" array.
1632 * The arbitrary user parameter passed in by the application when the callback
1633 * was originally configured.
1635 * The number of packets returned to the user.
1637 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port, uint16_t queue,
1638 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
1642 * Function type used for TX packet processing packet callbacks.
1644 * The callback function is called on TX with a burst of packets immediately
1645 * before the packets are put onto the hardware queue for transmission.
1648 * The Ethernet port on which TX is being performed.
1650 * The queue on the Ethernet port which is being used to transmit the packets.
1652 * The burst of packets that are about to be transmitted.
1654 * The number of packets in the burst pointed to by "pkts".
1656 * The arbitrary user parameter passed in by the application when the callback
1657 * was originally configured.
1659 * The number of packets to be written to the NIC.
1661 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port, uint16_t queue,
1662 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
1666 * Structure used to hold information about the callbacks to be called for a
1667 * queue on RX and TX.
1669 struct rte_eth_rxtx_callback {
1670 struct rte_eth_rxtx_callback *next;
1672 rte_rx_callback_fn rx;
1673 rte_tx_callback_fn tx;
1679 * A set of values to describe the possible states of an eth device.
1681 enum rte_eth_dev_state {
1682 RTE_ETH_DEV_UNUSED = 0,
1683 RTE_ETH_DEV_ATTACHED,
1684 RTE_ETH_DEV_DEFERRED,
1689 * The generic data structure associated with each ethernet device.
1691 * Pointers to burst-oriented packet receive and transmit functions are
1692 * located at the beginning of the structure, along with the pointer to
1693 * where all the data elements for the particular device are stored in shared
1694 * memory. This split allows the function pointer and driver data to be per-
1695 * process, while the actual configuration data for the device is shared.
1697 struct rte_eth_dev {
1698 eth_rx_burst_t rx_pkt_burst; /**< Pointer to PMD receive function. */
1699 eth_tx_burst_t tx_pkt_burst; /**< Pointer to PMD transmit function. */
1700 eth_tx_prep_t tx_pkt_prepare; /**< Pointer to PMD transmit prepare function. */
1701 struct rte_eth_dev_data *data; /**< Pointer to device data */
1702 const struct eth_dev_ops *dev_ops; /**< Functions exported by PMD */
1703 struct rte_device *device; /**< Backing device */
1704 struct rte_intr_handle *intr_handle; /**< Device interrupt handle */
1705 /** User application callbacks for NIC interrupts */
1706 struct rte_eth_dev_cb_list link_intr_cbs;
1708 * User-supplied functions called from rx_burst to post-process
1709 * received packets before passing them to the user
1711 struct rte_eth_rxtx_callback *post_rx_burst_cbs[RTE_MAX_QUEUES_PER_PORT];
1713 * User-supplied functions called from tx_burst to pre-process
1714 * received packets before passing them to the driver for transmission.
1716 struct rte_eth_rxtx_callback *pre_tx_burst_cbs[RTE_MAX_QUEUES_PER_PORT];
1717 enum rte_eth_dev_state state; /**< Flag indicating the port state */
1718 void *security_ctx; /**< Context for security ops */
1719 } __rte_cache_aligned;
1722 rte_eth_dev_get_sec_ctx(uint8_t port_id);
1724 struct rte_eth_dev_sriov {
1725 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
1726 uint8_t nb_q_per_pool; /**< rx queue number per pool */
1727 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
1728 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
1730 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
1732 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
1736 * The data part, with no function pointers, associated with each ethernet device.
1738 * This structure is safe to place in shared memory to be common among different
1739 * processes in a multi-process configuration.
1741 struct rte_eth_dev_data {
1742 char name[RTE_ETH_NAME_MAX_LEN]; /**< Unique identifier name */
1744 void **rx_queues; /**< Array of pointers to RX queues. */
1745 void **tx_queues; /**< Array of pointers to TX queues. */
1746 uint16_t nb_rx_queues; /**< Number of RX queues. */
1747 uint16_t nb_tx_queues; /**< Number of TX queues. */
1749 struct rte_eth_dev_sriov sriov; /**< SRIOV data */
1751 void *dev_private; /**< PMD-specific private data */
1753 struct rte_eth_link dev_link;
1754 /**< Link-level information & status */
1756 struct rte_eth_conf dev_conf; /**< Configuration applied to device. */
1757 uint16_t mtu; /**< Maximum Transmission Unit. */
1759 uint32_t min_rx_buf_size;
1760 /**< Common rx buffer size handled by all queues */
1762 uint64_t rx_mbuf_alloc_failed; /**< RX ring mbuf allocation failures. */
1763 struct ether_addr* mac_addrs;/**< Device Ethernet Link address. */
1764 uint64_t mac_pool_sel[ETH_NUM_RECEIVE_MAC_ADDR];
1765 /** bitmap array of associating Ethernet MAC addresses to pools */
1766 struct ether_addr* hash_mac_addrs;
1767 /** Device Ethernet MAC addresses of hash filtering. */
1768 uint16_t port_id; /**< Device [external] port identifier. */
1770 uint8_t promiscuous : 1, /**< RX promiscuous mode ON(1) / OFF(0). */
1771 scattered_rx : 1, /**< RX of scattered packets is ON(1) / OFF(0) */
1772 all_multicast : 1, /**< RX all multicast mode ON(1) / OFF(0). */
1773 dev_started : 1, /**< Device state: STARTED(1) / STOPPED(0). */
1774 lro : 1; /**< RX LRO is ON(1) / OFF(0) */
1775 uint8_t rx_queue_state[RTE_MAX_QUEUES_PER_PORT];
1776 /** Queues state: STARTED(1) / STOPPED(0) */
1777 uint8_t tx_queue_state[RTE_MAX_QUEUES_PER_PORT];
1778 /** Queues state: STARTED(1) / STOPPED(0) */
1779 uint32_t dev_flags; /**< Capabilities */
1780 enum rte_kernel_driver kdrv; /**< Kernel driver passthrough */
1781 int numa_node; /**< NUMA node connection */
1782 struct rte_vlan_filter_conf vlan_filter_conf;
1783 /**< VLAN filter configuration. */
1786 /** Device supports link state interrupt */
1787 #define RTE_ETH_DEV_INTR_LSC 0x0002
1788 /** Device is a bonded slave */
1789 #define RTE_ETH_DEV_BONDED_SLAVE 0x0004
1790 /** Device supports device removal interrupt */
1791 #define RTE_ETH_DEV_INTR_RMV 0x0008
1795 * The pool of *rte_eth_dev* structures. The size of the pool
1796 * is configured at compile-time in the <rte_ethdev.c> file.
1798 extern struct rte_eth_dev rte_eth_devices[];
1801 * Iterates over valid ethdev ports.
1804 * The id of the next possible valid port.
1806 * Next valid port id, RTE_MAX_ETHPORTS if there is none.
1808 uint16_t rte_eth_find_next(uint16_t port_id);
1811 * Macro to iterate over all enabled ethdev ports.
1813 #define RTE_ETH_FOREACH_DEV(p) \
1814 for (p = rte_eth_find_next(0); \
1815 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
1816 p = rte_eth_find_next(p + 1))
1820 * Get the total number of Ethernet devices that have been successfully
1821 * initialized by the matching Ethernet driver during the PCI probing phase
1822 * and that are available for applications to use. These devices must be
1823 * accessed by using the ``RTE_ETH_FOREACH_DEV()`` macro to deal with
1824 * non-contiguous ranges of devices.
1825 * These non-contiguous ranges can be created by calls to hotplug functions or
1829 * - The total number of usable Ethernet devices.
1831 uint16_t rte_eth_dev_count(void);
1835 * Returns a ethdev slot specified by the unique identifier name.
1838 * The pointer to the Unique identifier name for each Ethernet device
1840 * - The pointer to the ethdev slot, on success. NULL on error
1842 struct rte_eth_dev *rte_eth_dev_allocated(const char *name);
1846 * Allocates a new ethdev slot for an ethernet device and returns the pointer
1847 * to that slot for the driver to use.
1849 * @param name Unique identifier name for each Ethernet device
1850 * @param type Device type of this Ethernet device
1852 * - Slot in the rte_dev_devices array for a new device;
1854 struct rte_eth_dev *rte_eth_dev_allocate(const char *name);
1858 * Attach to the ethdev already initialized by the primary
1861 * @param name Ethernet device's name.
1863 * - Success: Slot in the rte_dev_devices array for attached
1865 * - Error: Null pointer.
1867 struct rte_eth_dev *rte_eth_dev_attach_secondary(const char *name);
1871 * Release the specified ethdev port.
1874 * The *eth_dev* pointer is the address of the *rte_eth_dev* structure.
1876 * - 0 on success, negative on error
1878 int rte_eth_dev_release_port(struct rte_eth_dev *eth_dev);
1881 * Attach a new Ethernet device specified by arguments.
1884 * A pointer to a strings array describing the new device
1885 * to be attached. The strings should be a pci address like
1886 * '0000:01:00.0' or virtual device name like 'net_pcap0'.
1888 * A pointer to a port identifier actually attached.
1890 * 0 on success and port_id is filled, negative on error
1892 int rte_eth_dev_attach(const char *devargs, uint16_t *port_id);
1895 * Detach a Ethernet device specified by port identifier.
1896 * This function must be called when the device is in the
1900 * The port identifier of the device to detach.
1902 * A pointer to a buffer that will be filled with the device name.
1903 * This buffer must be at least RTE_DEV_NAME_MAX_LEN long.
1905 * 0 on success and devname is filled, negative on error
1907 int rte_eth_dev_detach(uint16_t port_id, char *devname);
1910 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
1911 * the bitmap link_speeds of the struct rte_eth_conf
1914 * Numerical speed value in Mbps
1916 * ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
1918 * 0 if the speed cannot be mapped
1920 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
1923 * Configure an Ethernet device.
1924 * This function must be invoked first before any other function in the
1925 * Ethernet API. This function can also be re-invoked when a device is in the
1929 * The port identifier of the Ethernet device to configure.
1930 * @param nb_rx_queue
1931 * The number of receive queues to set up for the Ethernet device.
1932 * @param nb_tx_queue
1933 * The number of transmit queues to set up for the Ethernet device.
1935 * The pointer to the configuration data to be used for the Ethernet device.
1936 * The *rte_eth_conf* structure includes:
1937 * - the hardware offload features to activate, with dedicated fields for
1938 * each statically configurable offload hardware feature provided by
1939 * Ethernet devices, such as IP checksum or VLAN tag stripping for
1941 * The Rx offload bitfield API is obsolete and will be deprecated.
1942 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
1943 * structure and use offloads field to set per-port offloads instead.
1944 * - the Receive Side Scaling (RSS) configuration when using multiple RX
1947 * Embedding all configuration information in a single data structure
1948 * is the more flexible method that allows the addition of new features
1949 * without changing the syntax of the API.
1951 * - 0: Success, device configured.
1952 * - <0: Error code returned by the driver configuration function.
1954 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
1955 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
1959 * Release device queues and clear its configuration to force the user
1960 * application to reconfigure it. It is for internal use only.
1963 * Pointer to struct rte_eth_dev.
1968 void _rte_eth_dev_reset(struct rte_eth_dev *dev);
1971 * Allocate and set up a receive queue for an Ethernet device.
1973 * The function allocates a contiguous block of memory for *nb_rx_desc*
1974 * receive descriptors from a memory zone associated with *socket_id*
1975 * and initializes each receive descriptor with a network buffer allocated
1976 * from the memory pool *mb_pool*.
1979 * The port identifier of the Ethernet device.
1980 * @param rx_queue_id
1981 * The index of the receive queue to set up.
1982 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1983 * to rte_eth_dev_configure().
1985 * The number of receive descriptors to allocate for the receive ring.
1987 * The *socket_id* argument is the socket identifier in case of NUMA.
1988 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1989 * the DMA memory allocated for the receive descriptors of the ring.
1991 * The pointer to the configuration data to be used for the receive queue.
1992 * NULL value is allowed, in which case default RX configuration
1994 * The *rx_conf* structure contains an *rx_thresh* structure with the values
1995 * of the Prefetch, Host, and Write-Back threshold registers of the receive
1997 * In addition it contains the hardware offloads features to activate using
1998 * the DEV_RX_OFFLOAD_* flags.
2000 * The pointer to the memory pool from which to allocate *rte_mbuf* network
2001 * memory buffers to populate each descriptor of the receive ring.
2003 * - 0: Success, receive queue correctly set up.
2004 * - -EINVAL: The size of network buffers which can be allocated from the
2005 * memory pool does not fit the various buffer sizes allowed by the
2006 * device controller.
2007 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
2008 * allocate network memory buffers from the memory pool when
2009 * initializing receive descriptors.
2011 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
2012 uint16_t nb_rx_desc, unsigned int socket_id,
2013 const struct rte_eth_rxconf *rx_conf,
2014 struct rte_mempool *mb_pool);
2017 * Allocate and set up a transmit queue for an Ethernet device.
2020 * The port identifier of the Ethernet device.
2021 * @param tx_queue_id
2022 * The index of the transmit queue to set up.
2023 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2024 * to rte_eth_dev_configure().
2026 * The number of transmit descriptors to allocate for the transmit ring.
2028 * The *socket_id* argument is the socket identifier in case of NUMA.
2029 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
2030 * the DMA memory allocated for the transmit descriptors of the ring.
2032 * The pointer to the configuration data to be used for the transmit queue.
2033 * NULL value is allowed, in which case default TX configuration
2035 * The *tx_conf* structure contains the following data:
2036 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
2037 * Write-Back threshold registers of the transmit ring.
2038 * When setting Write-Back threshold to the value greater then zero,
2039 * *tx_rs_thresh* value should be explicitly set to one.
2040 * - The *tx_free_thresh* value indicates the [minimum] number of network
2041 * buffers that must be pending in the transmit ring to trigger their
2042 * [implicit] freeing by the driver transmit function.
2043 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
2044 * descriptors that must be pending in the transmit ring before setting the
2045 * RS bit on a descriptor by the driver transmit function.
2046 * The *tx_rs_thresh* value should be less or equal then
2047 * *tx_free_thresh* value, and both of them should be less then
2049 * - The *txq_flags* member contains flags to pass to the TX queue setup
2050 * function to configure the behavior of the TX queue. This should be set
2051 * to 0 if no special configuration is required.
2052 * This API is obsolete and will be deprecated. Applications
2053 * should set it to ETH_TXQ_FLAGS_IGNORE and use
2054 * the offloads field below.
2055 * - The *offloads* member contains Tx offloads to be enabled.
2056 * Offloads which are not set cannot be used on the datapath.
2058 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
2059 * the transmit function to use default values.
2061 * - 0: Success, the transmit queue is correctly set up.
2062 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
2064 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
2065 uint16_t nb_tx_desc, unsigned int socket_id,
2066 const struct rte_eth_txconf *tx_conf);
2069 * Return the NUMA socket to which an Ethernet device is connected
2072 * The port identifier of the Ethernet device
2074 * The NUMA socket id to which the Ethernet device is connected or
2075 * a default of zero if the socket could not be determined.
2076 * -1 is returned is the port_id value is out of range.
2078 int rte_eth_dev_socket_id(uint16_t port_id);
2081 * Check if port_id of device is attached
2084 * The port identifier of the Ethernet device
2086 * - 0 if port is out of range or not attached
2087 * - 1 if device is attached
2089 int rte_eth_dev_is_valid_port(uint16_t port_id);
2092 * Start specified RX queue of a port. It is used when rx_deferred_start
2093 * flag of the specified queue is true.
2096 * The port identifier of the Ethernet device
2097 * @param rx_queue_id
2098 * The index of the rx queue to update the ring.
2099 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2100 * to rte_eth_dev_configure().
2102 * - 0: Success, the receive queue is started.
2103 * - -EINVAL: The port_id or the queue_id out of range.
2104 * - -ENOTSUP: The function not supported in PMD driver.
2106 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
2109 * Stop specified RX queue of a port
2112 * The port identifier of the Ethernet device
2113 * @param rx_queue_id
2114 * The index of the rx queue to update the ring.
2115 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2116 * to rte_eth_dev_configure().
2118 * - 0: Success, the receive queue is stopped.
2119 * - -EINVAL: The port_id or the queue_id out of range.
2120 * - -ENOTSUP: The function not supported in PMD driver.
2122 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
2125 * Start TX for specified queue of a port. It is used when tx_deferred_start
2126 * flag of the specified queue is true.
2129 * The port identifier of the Ethernet device
2130 * @param tx_queue_id
2131 * The index of the tx queue to update the ring.
2132 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2133 * to rte_eth_dev_configure().
2135 * - 0: Success, the transmit queue is started.
2136 * - -EINVAL: The port_id or the queue_id out of range.
2137 * - -ENOTSUP: The function not supported in PMD driver.
2139 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
2142 * Stop specified TX queue of a port
2145 * The port identifier of the Ethernet device
2146 * @param tx_queue_id
2147 * The index of the tx queue to update the ring.
2148 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2149 * to rte_eth_dev_configure().
2151 * - 0: Success, the transmit queue is stopped.
2152 * - -EINVAL: The port_id or the queue_id out of range.
2153 * - -ENOTSUP: The function not supported in PMD driver.
2155 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
2160 * Start an Ethernet device.
2162 * The device start step is the last one and consists of setting the configured
2163 * offload features and in starting the transmit and the receive units of the
2165 * On success, all basic functions exported by the Ethernet API (link status,
2166 * receive/transmit, and so on) can be invoked.
2169 * The port identifier of the Ethernet device.
2171 * - 0: Success, Ethernet device started.
2172 * - <0: Error code of the driver device start function.
2174 int rte_eth_dev_start(uint16_t port_id);
2177 * Stop an Ethernet device. The device can be restarted with a call to
2178 * rte_eth_dev_start()
2181 * The port identifier of the Ethernet device.
2183 void rte_eth_dev_stop(uint16_t port_id);
2187 * Link up an Ethernet device.
2189 * Set device link up will re-enable the device rx/tx
2190 * functionality after it is previously set device linked down.
2193 * The port identifier of the Ethernet device.
2195 * - 0: Success, Ethernet device linked up.
2196 * - <0: Error code of the driver device link up function.
2198 int rte_eth_dev_set_link_up(uint16_t port_id);
2201 * Link down an Ethernet device.
2202 * The device rx/tx functionality will be disabled if success,
2203 * and it can be re-enabled with a call to
2204 * rte_eth_dev_set_link_up()
2207 * The port identifier of the Ethernet device.
2209 int rte_eth_dev_set_link_down(uint16_t port_id);
2212 * Close a stopped Ethernet device. The device cannot be restarted!
2213 * The function frees all resources except for needed by the
2214 * closed state. To free these resources, call rte_eth_dev_detach().
2217 * The port identifier of the Ethernet device.
2219 void rte_eth_dev_close(uint16_t port_id);
2222 * Reset a Ethernet device and keep its port id.
2224 * When a port has to be reset passively, the DPDK application can invoke
2225 * this function. For example when a PF is reset, all its VFs should also
2226 * be reset. Normally a DPDK application can invoke this function when
2227 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
2228 * a port reset in other circumstances.
2230 * When this function is called, it first stops the port and then calls the
2231 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
2232 * state, in which no Tx and Rx queues are setup, as if the port has been
2233 * reset and not started. The port keeps the port id it had before the
2236 * After calling rte_eth_dev_reset( ), the application should use
2237 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
2238 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
2239 * to reconfigure the device as appropriate.
2241 * Note: To avoid unexpected behavior, the application should stop calling
2242 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
2243 * safety, all these controlling functions should be called from the same
2247 * The port identifier of the Ethernet device.
2250 * - (0) if successful.
2251 * - (-EINVAL) if port identifier is invalid.
2252 * - (-ENOTSUP) if hardware doesn't support this function.
2253 * - (-EPERM) if not ran from the primary process.
2254 * - (-EIO) if re-initialisation failed.
2255 * - (-ENOMEM) if the reset failed due to OOM.
2256 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
2258 int rte_eth_dev_reset(uint16_t port_id);
2261 * Enable receipt in promiscuous mode for an Ethernet device.
2264 * The port identifier of the Ethernet device.
2266 void rte_eth_promiscuous_enable(uint16_t port_id);
2269 * Disable receipt in promiscuous mode for an Ethernet device.
2272 * The port identifier of the Ethernet device.
2274 void rte_eth_promiscuous_disable(uint16_t port_id);
2277 * Return the value of promiscuous mode for an Ethernet device.
2280 * The port identifier of the Ethernet device.
2282 * - (1) if promiscuous is enabled
2283 * - (0) if promiscuous is disabled.
2286 int rte_eth_promiscuous_get(uint16_t port_id);
2289 * Enable the receipt of any multicast frame by an Ethernet device.
2292 * The port identifier of the Ethernet device.
2294 void rte_eth_allmulticast_enable(uint16_t port_id);
2297 * Disable the receipt of all multicast frames by an Ethernet device.
2300 * The port identifier of the Ethernet device.
2302 void rte_eth_allmulticast_disable(uint16_t port_id);
2305 * Return the value of allmulticast mode for an Ethernet device.
2308 * The port identifier of the Ethernet device.
2310 * - (1) if allmulticast is enabled
2311 * - (0) if allmulticast is disabled.
2314 int rte_eth_allmulticast_get(uint16_t port_id);
2317 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
2318 * or FULL-DUPLEX) of the physical link of an Ethernet device. It might need
2319 * to wait up to 9 seconds in it.
2322 * The port identifier of the Ethernet device.
2324 * A pointer to an *rte_eth_link* structure to be filled with
2325 * the status, the speed and the mode of the Ethernet device link.
2327 void rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
2330 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
2331 * or FULL-DUPLEX) of the physical link of an Ethernet device. It is a no-wait
2332 * version of rte_eth_link_get().
2335 * The port identifier of the Ethernet device.
2337 * A pointer to an *rte_eth_link* structure to be filled with
2338 * the status, the speed and the mode of the Ethernet device link.
2340 void rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
2343 * Retrieve the general I/O statistics of an Ethernet device.
2346 * The port identifier of the Ethernet device.
2348 * A pointer to a structure of type *rte_eth_stats* to be filled with
2349 * the values of device counters for the following set of statistics:
2350 * - *ipackets* with the total of successfully received packets.
2351 * - *opackets* with the total of successfully transmitted packets.
2352 * - *ibytes* with the total of successfully received bytes.
2353 * - *obytes* with the total of successfully transmitted bytes.
2354 * - *ierrors* with the total of erroneous received packets.
2355 * - *oerrors* with the total of failed transmitted packets.
2357 * Zero if successful. Non-zero otherwise.
2359 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
2362 * Reset the general I/O statistics of an Ethernet device.
2365 * The port identifier of the Ethernet device.
2367 * - (0) if device notified to reset stats.
2368 * - (-ENOTSUP) if hardware doesn't support.
2369 * - (-ENODEV) if *port_id* invalid.
2371 int rte_eth_stats_reset(uint16_t port_id);
2374 * Retrieve names of extended statistics of an Ethernet device.
2377 * The port identifier of the Ethernet device.
2378 * @param xstats_names
2379 * An rte_eth_xstat_name array of at least *size* elements to
2380 * be filled. If set to NULL, the function returns the required number
2383 * The size of the xstats_names array (number of elements).
2385 * - A positive value lower or equal to size: success. The return value
2386 * is the number of entries filled in the stats table.
2387 * - A positive value higher than size: error, the given statistics table
2388 * is too small. The return value corresponds to the size that should
2389 * be given to succeed. The entries in the table are not valid and
2390 * shall not be used by the caller.
2391 * - A negative value on error (invalid port id).
2393 int rte_eth_xstats_get_names(uint16_t port_id,
2394 struct rte_eth_xstat_name *xstats_names,
2398 * Retrieve extended statistics of an Ethernet device.
2401 * The port identifier of the Ethernet device.
2403 * A pointer to a table of structure of type *rte_eth_xstat*
2404 * to be filled with device statistics ids and values: id is the
2405 * index of the name string in xstats_names (see rte_eth_xstats_get_names()),
2406 * and value is the statistic counter.
2407 * This parameter can be set to NULL if n is 0.
2409 * The size of the xstats array (number of elements).
2411 * - A positive value lower or equal to n: success. The return value
2412 * is the number of entries filled in the stats table.
2413 * - A positive value higher than n: error, the given statistics table
2414 * is too small. The return value corresponds to the size that should
2415 * be given to succeed. The entries in the table are not valid and
2416 * shall not be used by the caller.
2417 * - A negative value on error (invalid port id).
2419 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
2423 * Retrieve names of extended statistics of an Ethernet device.
2426 * The port identifier of the Ethernet device.
2427 * @param xstats_names
2428 * An rte_eth_xstat_name array of at least *size* elements to
2429 * be filled. If set to NULL, the function returns the required number
2432 * IDs array given by app to retrieve specific statistics
2434 * The size of the xstats_names array (number of elements).
2436 * - A positive value lower or equal to size: success. The return value
2437 * is the number of entries filled in the stats table.
2438 * - A positive value higher than size: error, the given statistics table
2439 * is too small. The return value corresponds to the size that should
2440 * be given to succeed. The entries in the table are not valid and
2441 * shall not be used by the caller.
2442 * - A negative value on error (invalid port id).
2445 rte_eth_xstats_get_names_by_id(uint16_t port_id,
2446 struct rte_eth_xstat_name *xstats_names, unsigned int size,
2450 * Retrieve extended statistics of an Ethernet device.
2453 * The port identifier of the Ethernet device.
2455 * A pointer to an ids array passed by application. This tells which
2456 * statistics values function should retrieve. This parameter
2457 * can be set to NULL if size is 0. In this case function will retrieve
2458 * all avalible statistics.
2460 * A pointer to a table to be filled with device statistics values.
2462 * The size of the ids array (number of elements).
2464 * - A positive value lower or equal to size: success. The return value
2465 * is the number of entries filled in the stats table.
2466 * - A positive value higher than size: error, the given statistics table
2467 * is too small. The return value corresponds to the size that should
2468 * be given to succeed. The entries in the table are not valid and
2469 * shall not be used by the caller.
2470 * - A negative value on error (invalid port id).
2472 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
2473 uint64_t *values, unsigned int size);
2476 * Gets the ID of a statistic from its name.
2478 * This function searches for the statistics using string compares, and
2479 * as such should not be used on the fast-path. For fast-path retrieval of
2480 * specific statistics, store the ID as provided in *id* from this function,
2481 * and pass the ID to rte_eth_xstats_get()
2483 * @param port_id The port to look up statistics from
2484 * @param xstat_name The name of the statistic to return
2485 * @param[out] id A pointer to an app-supplied uint64_t which should be
2486 * set to the ID of the stat if the stat exists.
2489 * -ENODEV for invalid port_id,
2490 * -EINVAL if the xstat_name doesn't exist in port_id
2492 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
2496 * Reset extended statistics of an Ethernet device.
2499 * The port identifier of the Ethernet device.
2501 void rte_eth_xstats_reset(uint16_t port_id);
2504 * Set a mapping for the specified transmit queue to the specified per-queue
2505 * statistics counter.
2508 * The port identifier of the Ethernet device.
2509 * @param tx_queue_id
2510 * The index of the transmit queue for which a queue stats mapping is required.
2511 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2512 * to rte_eth_dev_configure().
2514 * The per-queue packet statistics functionality number that the transmit
2515 * queue is to be assigned.
2516 * The value must be in the range [0, RTE_MAX_ETHPORT_QUEUE_STATS_MAPS - 1].
2518 * Zero if successful. Non-zero otherwise.
2520 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
2521 uint16_t tx_queue_id, uint8_t stat_idx);
2524 * Set a mapping for the specified receive queue to the specified per-queue
2525 * statistics counter.
2528 * The port identifier of the Ethernet device.
2529 * @param rx_queue_id
2530 * The index of the receive queue for which a queue stats mapping is required.
2531 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2532 * to rte_eth_dev_configure().
2534 * The per-queue packet statistics functionality number that the receive
2535 * queue is to be assigned.
2536 * The value must be in the range [0, RTE_MAX_ETHPORT_QUEUE_STATS_MAPS - 1].
2538 * Zero if successful. Non-zero otherwise.
2540 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
2541 uint16_t rx_queue_id,
2545 * Retrieve the Ethernet address of an Ethernet device.
2548 * The port identifier of the Ethernet device.
2550 * A pointer to a structure of type *ether_addr* to be filled with
2551 * the Ethernet address of the Ethernet device.
2553 void rte_eth_macaddr_get(uint16_t port_id, struct ether_addr *mac_addr);
2556 * Retrieve the contextual information of an Ethernet device.
2559 * The port identifier of the Ethernet device.
2561 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
2562 * the contextual information of the Ethernet device.
2564 void rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
2567 * Retrieve the firmware version of a device.
2570 * The port identifier of the device.
2572 * A pointer to a string array storing the firmware version of a device,
2573 * the string includes terminating null. This pointer is allocated by caller.
2575 * The size of the string array pointed by fw_version, which should be
2576 * large enough to store firmware version of the device.
2578 * - (0) if successful.
2579 * - (-ENOTSUP) if operation is not supported.
2580 * - (-ENODEV) if *port_id* invalid.
2581 * - (>0) if *fw_size* is not enough to store firmware version, return
2582 * the size of the non truncated string.
2584 int rte_eth_dev_fw_version_get(uint16_t port_id,
2585 char *fw_version, size_t fw_size);
2588 * Retrieve the supported packet types of an Ethernet device.
2590 * When a packet type is announced as supported, it *must* be recognized by
2591 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
2592 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
2593 * packet types for these packets:
2594 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
2595 * - Ether/Vlan/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
2596 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
2597 * - Ether/Vlan/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
2599 * When a packet is received by a PMD, the most precise type must be
2600 * returned among the ones supported. However a PMD is allowed to set
2601 * packet type that is not in the supported list, at the condition that it
2602 * is more precise. Therefore, a PMD announcing no supported packet types
2603 * can still set a matching packet type in a received packet.
2606 * Better to invoke this API after the device is already started or rx burst
2607 * function is decided, to obtain correct supported ptypes.
2609 * if a given PMD does not report what ptypes it supports, then the supported
2610 * ptype count is reported as 0.
2612 * The port identifier of the Ethernet device.
2614 * A hint of what kind of packet type which the caller is interested in.
2616 * An array pointer to store adequate packet types, allocated by caller.
2618 * Size of the array pointed by param ptypes.
2620 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
2621 * only num entries will be filled into the ptypes array, but the full
2622 * count of supported ptypes will be returned.
2623 * - (-ENODEV) if *port_id* invalid.
2625 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
2626 uint32_t *ptypes, int num);
2629 * Retrieve the MTU of an Ethernet device.
2632 * The port identifier of the Ethernet device.
2634 * A pointer to a uint16_t where the retrieved MTU is to be stored.
2636 * - (0) if successful.
2637 * - (-ENODEV) if *port_id* invalid.
2639 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
2642 * Change the MTU of an Ethernet device.
2645 * The port identifier of the Ethernet device.
2647 * A uint16_t for the MTU to be applied.
2649 * - (0) if successful.
2650 * - (-ENOTSUP) if operation is not supported.
2651 * - (-ENODEV) if *port_id* invalid.
2652 * - (-EINVAL) if *mtu* invalid.
2653 * - (-EBUSY) if operation is not allowed when the port is running
2655 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
2658 * Enable/Disable hardware filtering by an Ethernet device of received
2659 * VLAN packets tagged with a given VLAN Tag Identifier.
2662 * The port identifier of the Ethernet device.
2664 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
2666 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
2667 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
2669 * - (0) if successful.
2670 * - (-ENOSUP) if hardware-assisted VLAN filtering not configured.
2671 * - (-ENODEV) if *port_id* invalid.
2672 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
2673 * - (-EINVAL) if *vlan_id* > 4095.
2675 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
2678 * Enable/Disable hardware VLAN Strip by a rx queue of an Ethernet device.
2679 * 82599/X540/X550 can support VLAN stripping at the rx queue level
2682 * The port identifier of the Ethernet device.
2683 * @param rx_queue_id
2684 * The index of the receive queue for which a queue stats mapping is required.
2685 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2686 * to rte_eth_dev_configure().
2688 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
2689 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
2691 * - (0) if successful.
2692 * - (-ENOSUP) if hardware-assisted VLAN stripping not configured.
2693 * - (-ENODEV) if *port_id* invalid.
2694 * - (-EINVAL) if *rx_queue_id* invalid.
2696 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
2700 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
2701 * the VLAN Header. This is a register setup available on some Intel NIC, not
2702 * but all, please check the data sheet for availability.
2705 * The port identifier of the Ethernet device.
2709 * The Tag Protocol ID
2711 * - (0) if successful.
2712 * - (-ENOSUP) if hardware-assisted VLAN TPID setup is not supported.
2713 * - (-ENODEV) if *port_id* invalid.
2715 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
2716 enum rte_vlan_type vlan_type,
2720 * Set VLAN offload configuration on an Ethernet device
2721 * Enable/Disable Extended VLAN by an Ethernet device, This is a register setup
2722 * available on some Intel NIC, not but all, please check the data sheet for
2724 * Enable/Disable VLAN Strip can be done on rx queue for certain NIC, but here
2725 * the configuration is applied on the port level.
2728 * The port identifier of the Ethernet device.
2729 * @param offload_mask
2730 * The VLAN Offload bit mask can be mixed use with "OR"
2731 * ETH_VLAN_STRIP_OFFLOAD
2732 * ETH_VLAN_FILTER_OFFLOAD
2733 * ETH_VLAN_EXTEND_OFFLOAD
2735 * - (0) if successful.
2736 * - (-ENOSUP) if hardware-assisted VLAN filtering not configured.
2737 * - (-ENODEV) if *port_id* invalid.
2739 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
2742 * Read VLAN Offload configuration from an Ethernet device
2745 * The port identifier of the Ethernet device.
2747 * - (>0) if successful. Bit mask to indicate
2748 * ETH_VLAN_STRIP_OFFLOAD
2749 * ETH_VLAN_FILTER_OFFLOAD
2750 * ETH_VLAN_EXTEND_OFFLOAD
2751 * - (-ENODEV) if *port_id* invalid.
2753 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
2756 * Set port based TX VLAN insertion on or off.
2759 * The port identifier of the Ethernet device.
2761 * Port based TX VLAN identifier together with user priority.
2763 * Turn on or off the port based TX VLAN insertion.
2766 * - (0) if successful.
2767 * - negative if failed.
2769 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
2773 * Retrieve a burst of input packets from a receive queue of an Ethernet
2774 * device. The retrieved packets are stored in *rte_mbuf* structures whose
2775 * pointers are supplied in the *rx_pkts* array.
2777 * The rte_eth_rx_burst() function loops, parsing the RX ring of the
2778 * receive queue, up to *nb_pkts* packets, and for each completed RX
2779 * descriptor in the ring, it performs the following operations:
2781 * - Initialize the *rte_mbuf* data structure associated with the
2782 * RX descriptor according to the information provided by the NIC into
2783 * that RX descriptor.
2785 * - Store the *rte_mbuf* data structure into the next entry of the
2788 * - Replenish the RX descriptor with a new *rte_mbuf* buffer
2789 * allocated from the memory pool associated with the receive queue at
2790 * initialization time.
2792 * When retrieving an input packet that was scattered by the controller
2793 * into multiple receive descriptors, the rte_eth_rx_burst() function
2794 * appends the associated *rte_mbuf* buffers to the first buffer of the
2797 * The rte_eth_rx_burst() function returns the number of packets
2798 * actually retrieved, which is the number of *rte_mbuf* data structures
2799 * effectively supplied into the *rx_pkts* array.
2800 * A return value equal to *nb_pkts* indicates that the RX queue contained
2801 * at least *rx_pkts* packets, and this is likely to signify that other
2802 * received packets remain in the input queue. Applications implementing
2803 * a "retrieve as much received packets as possible" policy can check this
2804 * specific case and keep invoking the rte_eth_rx_burst() function until
2805 * a value less than *nb_pkts* is returned.
2807 * This receive method has the following advantages:
2809 * - It allows a run-to-completion network stack engine to retrieve and
2810 * to immediately process received packets in a fast burst-oriented
2811 * approach, avoiding the overhead of unnecessary intermediate packet
2812 * queue/dequeue operations.
2814 * - Conversely, it also allows an asynchronous-oriented processing
2815 * method to retrieve bursts of received packets and to immediately
2816 * queue them for further parallel processing by another logical core,
2817 * for instance. However, instead of having received packets being
2818 * individually queued by the driver, this approach allows the caller
2819 * of the rte_eth_rx_burst() function to queue a burst of retrieved
2820 * packets at a time and therefore dramatically reduce the cost of
2821 * enqueue/dequeue operations per packet.
2823 * - It allows the rte_eth_rx_burst() function of the driver to take
2824 * advantage of burst-oriented hardware features (CPU cache,
2825 * prefetch instructions, and so on) to minimize the number of CPU
2826 * cycles per packet.
2828 * To summarize, the proposed receive API enables many
2829 * burst-oriented optimizations in both synchronous and asynchronous
2830 * packet processing environments with no overhead in both cases.
2832 * The rte_eth_rx_burst() function does not provide any error
2833 * notification to avoid the corresponding overhead. As a hint, the
2834 * upper-level application might check the status of the device link once
2835 * being systematically returned a 0 value for a given number of tries.
2838 * The port identifier of the Ethernet device.
2840 * The index of the receive queue from which to retrieve input packets.
2841 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2842 * to rte_eth_dev_configure().
2844 * The address of an array of pointers to *rte_mbuf* structures that
2845 * must be large enough to store *nb_pkts* pointers in it.
2847 * The maximum number of packets to retrieve.
2849 * The number of packets actually retrieved, which is the number
2850 * of pointers to *rte_mbuf* structures effectively supplied to the
2853 static inline uint16_t
2854 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
2855 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
2857 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
2859 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
2860 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
2861 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_pkt_burst, 0);
2863 if (queue_id >= dev->data->nb_rx_queues) {
2864 RTE_PMD_DEBUG_TRACE("Invalid RX queue_id=%d\n", queue_id);
2868 int16_t nb_rx = (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id],
2871 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
2872 struct rte_eth_rxtx_callback *cb = dev->post_rx_burst_cbs[queue_id];
2874 if (unlikely(cb != NULL)) {
2876 nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx,
2877 nb_pkts, cb->param);
2879 } while (cb != NULL);
2887 * Get the number of used descriptors of a rx queue
2890 * The port identifier of the Ethernet device.
2892 * The queue id on the specific port.
2894 * The number of used descriptors in the specific queue, or:
2895 * (-EINVAL) if *port_id* or *queue_id* is invalid
2896 * (-ENOTSUP) if the device does not support this function
2899 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
2901 struct rte_eth_dev *dev;
2903 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
2904 dev = &rte_eth_devices[port_id];
2905 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_queue_count, -ENOTSUP);
2906 if (queue_id >= dev->data->nb_rx_queues)
2909 return (*dev->dev_ops->rx_queue_count)(dev, queue_id);
2913 * Check if the DD bit of the specific RX descriptor in the queue has been set
2916 * The port identifier of the Ethernet device.
2918 * The queue id on the specific port.
2920 * The offset of the descriptor ID from tail.
2922 * - (1) if the specific DD bit is set.
2923 * - (0) if the specific DD bit is not set.
2924 * - (-ENODEV) if *port_id* invalid.
2925 * - (-ENOTSUP) if the device does not support this function
2928 rte_eth_rx_descriptor_done(uint16_t port_id, uint16_t queue_id, uint16_t offset)
2930 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
2931 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
2932 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_done, -ENOTSUP);
2933 return (*dev->dev_ops->rx_descriptor_done)( \
2934 dev->data->rx_queues[queue_id], offset);
2937 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
2938 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
2939 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
2942 * Check the status of a Rx descriptor in the queue
2944 * It should be called in a similar context than the Rx function:
2945 * - on a dataplane core
2946 * - not concurrently on the same queue
2948 * Since it's a dataplane function, no check is performed on port_id and
2949 * queue_id. The caller must therefore ensure that the port is enabled
2950 * and the queue is configured and running.
2952 * Note: accessing to a random descriptor in the ring may trigger cache
2953 * misses and have a performance impact.
2956 * A valid port identifier of the Ethernet device which.
2958 * A valid Rx queue identifier on this port.
2960 * The offset of the descriptor starting from tail (0 is the next
2961 * packet to be received by the driver).
2964 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
2966 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
2967 * not yet processed by the driver (i.e. in the receive queue).
2968 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
2969 * the driver and not yet returned to hw, or reserved by the hw.
2970 * - (-EINVAL) bad descriptor offset.
2971 * - (-ENOTSUP) if the device does not support this function.
2972 * - (-ENODEV) bad port or queue (only if compiled with debug).
2975 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
2978 struct rte_eth_dev *dev;
2981 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
2982 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
2984 dev = &rte_eth_devices[port_id];
2985 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
2986 if (queue_id >= dev->data->nb_rx_queues)
2989 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_status, -ENOTSUP);
2990 rxq = dev->data->rx_queues[queue_id];
2992 return (*dev->dev_ops->rx_descriptor_status)(rxq, offset);
2995 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
2996 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
2997 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
3000 * Check the status of a Tx descriptor in the queue.
3002 * It should be called in a similar context than the Tx function:
3003 * - on a dataplane core
3004 * - not concurrently on the same queue
3006 * Since it's a dataplane function, no check is performed on port_id and
3007 * queue_id. The caller must therefore ensure that the port is enabled
3008 * and the queue is configured and running.
3010 * Note: accessing to a random descriptor in the ring may trigger cache
3011 * misses and have a performance impact.
3014 * A valid port identifier of the Ethernet device which.
3016 * A valid Tx queue identifier on this port.
3018 * The offset of the descriptor starting from tail (0 is the place where
3019 * the next packet will be send).
3022 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
3023 * in the transmit queue.
3024 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
3025 * be reused by the driver.
3026 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
3027 * driver or the hardware.
3028 * - (-EINVAL) bad descriptor offset.
3029 * - (-ENOTSUP) if the device does not support this function.
3030 * - (-ENODEV) bad port or queue (only if compiled with debug).
3032 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
3033 uint16_t queue_id, uint16_t offset)
3035 struct rte_eth_dev *dev;
3038 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3039 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
3041 dev = &rte_eth_devices[port_id];
3042 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3043 if (queue_id >= dev->data->nb_tx_queues)
3046 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->tx_descriptor_status, -ENOTSUP);
3047 txq = dev->data->tx_queues[queue_id];
3049 return (*dev->dev_ops->tx_descriptor_status)(txq, offset);
3053 * Send a burst of output packets on a transmit queue of an Ethernet device.
3055 * The rte_eth_tx_burst() function is invoked to transmit output packets
3056 * on the output queue *queue_id* of the Ethernet device designated by its
3058 * The *nb_pkts* parameter is the number of packets to send which are
3059 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
3060 * allocated from a pool created with rte_pktmbuf_pool_create().
3061 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
3062 * up to the number of transmit descriptors available in the TX ring of the
3064 * For each packet to send, the rte_eth_tx_burst() function performs
3065 * the following operations:
3067 * - Pick up the next available descriptor in the transmit ring.
3069 * - Free the network buffer previously sent with that descriptor, if any.
3071 * - Initialize the transmit descriptor with the information provided
3072 * in the *rte_mbuf data structure.
3074 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
3075 * the rte_eth_tx_burst() function uses several transmit descriptors
3078 * The rte_eth_tx_burst() function returns the number of packets it
3079 * actually sent. A return value equal to *nb_pkts* means that all packets
3080 * have been sent, and this is likely to signify that other output packets
3081 * could be immediately transmitted again. Applications that implement a
3082 * "send as many packets to transmit as possible" policy can check this
3083 * specific case and keep invoking the rte_eth_tx_burst() function until
3084 * a value less than *nb_pkts* is returned.
3086 * It is the responsibility of the rte_eth_tx_burst() function to
3087 * transparently free the memory buffers of packets previously sent.
3088 * This feature is driven by the *tx_free_thresh* value supplied to the
3089 * rte_eth_dev_configure() function at device configuration time.
3090 * When the number of free TX descriptors drops below this threshold, the
3091 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
3092 * of those packets whose transmission was effectively completed.
3094 * If the PMD is DEV_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
3095 * invoke this function concurrently on the same tx queue without SW lock.
3096 * @see rte_eth_dev_info_get, struct rte_eth_txconf::txq_flags
3099 * The port identifier of the Ethernet device.
3101 * The index of the transmit queue through which output packets must be
3103 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3104 * to rte_eth_dev_configure().
3106 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
3107 * which contain the output packets.
3109 * The maximum number of packets to transmit.
3111 * The number of output packets actually stored in transmit descriptors of
3112 * the transmit ring. The return value can be less than the value of the
3113 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
3115 static inline uint16_t
3116 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
3117 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
3119 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
3121 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3122 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
3123 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_pkt_burst, 0);
3125 if (queue_id >= dev->data->nb_tx_queues) {
3126 RTE_PMD_DEBUG_TRACE("Invalid TX queue_id=%d\n", queue_id);
3131 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
3132 struct rte_eth_rxtx_callback *cb = dev->pre_tx_burst_cbs[queue_id];
3134 if (unlikely(cb != NULL)) {
3136 nb_pkts = cb->fn.tx(port_id, queue_id, tx_pkts, nb_pkts,
3139 } while (cb != NULL);
3143 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts);
3148 * @b EXPERIMENTAL: this API may change without prior notice
3150 * Process a burst of output packets on a transmit queue of an Ethernet device.
3152 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
3153 * transmitted on the output queue *queue_id* of the Ethernet device designated
3155 * The *nb_pkts* parameter is the number of packets to be prepared which are
3156 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
3157 * allocated from a pool created with rte_pktmbuf_pool_create().
3158 * For each packet to send, the rte_eth_tx_prepare() function performs
3159 * the following operations:
3161 * - Check if packet meets devices requirements for tx offloads.
3163 * - Check limitations about number of segments.
3165 * - Check additional requirements when debug is enabled.
3167 * - Update and/or reset required checksums when tx offload is set for packet.
3169 * Since this function can modify packet data, provided mbufs must be safely
3170 * writable (e.g. modified data cannot be in shared segment).
3172 * The rte_eth_tx_prepare() function returns the number of packets ready to be
3173 * sent. A return value equal to *nb_pkts* means that all packets are valid and
3174 * ready to be sent, otherwise stops processing on the first invalid packet and
3175 * leaves the rest packets untouched.
3177 * When this functionality is not implemented in the driver, all packets are
3178 * are returned untouched.
3181 * The port identifier of the Ethernet device.
3182 * The value must be a valid port id.
3184 * The index of the transmit queue through which output packets must be
3186 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3187 * to rte_eth_dev_configure().
3189 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
3190 * which contain the output packets.
3192 * The maximum number of packets to process.
3194 * The number of packets correct and ready to be sent. The return value can be
3195 * less than the value of the *tx_pkts* parameter when some packet doesn't
3196 * meet devices requirements with rte_errno set appropriately:
3197 * - -EINVAL: offload flags are not correctly set
3198 * - -ENOTSUP: the offload feature is not supported by the hardware
3202 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
3204 static inline uint16_t
3205 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
3206 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
3208 struct rte_eth_dev *dev;
3210 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3211 if (!rte_eth_dev_is_valid_port(port_id)) {
3212 RTE_PMD_DEBUG_TRACE("Invalid TX port_id=%d\n", port_id);
3213 rte_errno = -EINVAL;
3218 dev = &rte_eth_devices[port_id];
3220 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3221 if (queue_id >= dev->data->nb_tx_queues) {
3222 RTE_PMD_DEBUG_TRACE("Invalid TX queue_id=%d\n", queue_id);
3223 rte_errno = -EINVAL;
3228 if (!dev->tx_pkt_prepare)
3231 return (*dev->tx_pkt_prepare)(dev->data->tx_queues[queue_id],
3238 * Native NOOP operation for compilation targets which doesn't require any
3239 * preparations steps, and functional NOOP may introduce unnecessary performance
3242 * Generally this is not a good idea to turn it on globally and didn't should
3243 * be used if behavior of tx_preparation can change.
3246 static inline uint16_t
3247 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
3248 __rte_unused uint16_t queue_id,
3249 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
3256 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
3260 * Structure used to buffer packets for future TX
3261 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
3263 struct rte_eth_dev_tx_buffer {
3264 buffer_tx_error_fn error_callback;
3265 void *error_userdata;
3266 uint16_t size; /**< Size of buffer for buffered tx */
3267 uint16_t length; /**< Number of packets in the array */
3268 struct rte_mbuf *pkts[];
3269 /**< Pending packets to be sent on explicit flush or when full */
3273 * Calculate the size of the tx buffer.
3276 * Number of stored packets.
3278 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
3279 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
3282 * Initialize default values for buffered transmitting
3285 * Tx buffer to be initialized.
3292 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
3295 * Send any packets queued up for transmission on a port and HW queue
3297 * This causes an explicit flush of packets previously buffered via the
3298 * rte_eth_tx_buffer() function. It returns the number of packets successfully
3299 * sent to the NIC, and calls the error callback for any unsent packets. Unless
3300 * explicitly set up otherwise, the default callback simply frees the unsent
3301 * packets back to the owning mempool.
3304 * The port identifier of the Ethernet device.
3306 * The index of the transmit queue through which output packets must be
3308 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3309 * to rte_eth_dev_configure().
3311 * Buffer of packets to be transmit.
3313 * The number of packets successfully sent to the Ethernet device. The error
3314 * callback is called for any packets which could not be sent.
3316 static inline uint16_t
3317 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
3318 struct rte_eth_dev_tx_buffer *buffer)
3321 uint16_t to_send = buffer->length;
3326 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
3330 /* All packets sent, or to be dealt with by callback below */
3331 if (unlikely(sent != to_send))
3332 buffer->error_callback(&buffer->pkts[sent], to_send - sent,
3333 buffer->error_userdata);
3339 * Buffer a single packet for future transmission on a port and queue
3341 * This function takes a single mbuf/packet and buffers it for later
3342 * transmission on the particular port and queue specified. Once the buffer is
3343 * full of packets, an attempt will be made to transmit all the buffered
3344 * packets. In case of error, where not all packets can be transmitted, a
3345 * callback is called with the unsent packets as a parameter. If no callback
3346 * is explicitly set up, the unsent packets are just freed back to the owning
3347 * mempool. The function returns the number of packets actually sent i.e.
3348 * 0 if no buffer flush occurred, otherwise the number of packets successfully
3352 * The port identifier of the Ethernet device.
3354 * The index of the transmit queue through which output packets must be
3356 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3357 * to rte_eth_dev_configure().
3359 * Buffer used to collect packets to be sent.
3361 * Pointer to the packet mbuf to be sent.
3363 * 0 = packet has been buffered for later transmission
3364 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
3365 * causing N packets to be sent, and the error callback to be called for
3368 static __rte_always_inline uint16_t
3369 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
3370 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
3372 buffer->pkts[buffer->length++] = tx_pkt;
3373 if (buffer->length < buffer->size)
3376 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
3380 * Configure a callback for buffered packets which cannot be sent
3382 * Register a specific callback to be called when an attempt is made to send
3383 * all packets buffered on an ethernet port, but not all packets can
3384 * successfully be sent. The callback registered here will be called only
3385 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
3386 * The default callback configured for each queue by default just frees the
3387 * packets back to the calling mempool. If additional behaviour is required,
3388 * for example, to count dropped packets, or to retry transmission of packets
3389 * which cannot be sent, this function should be used to register a suitable
3390 * callback function to implement the desired behaviour.
3391 * The example callback "rte_eth_count_unsent_packet_callback()" is also
3392 * provided as reference.
3395 * The port identifier of the Ethernet device.
3397 * The function to be used as the callback.
3399 * Arbitrary parameter to be passed to the callback function
3401 * 0 on success, or -1 on error with rte_errno set appropriately
3404 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
3405 buffer_tx_error_fn callback, void *userdata);
3408 * Callback function for silently dropping unsent buffered packets.
3410 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3411 * adjust the default behavior when buffered packets cannot be sent. This
3412 * function drops any unsent packets silently and is used by tx buffered
3413 * operations as default behavior.
3415 * NOTE: this function should not be called directly, instead it should be used
3416 * as a callback for packet buffering.
3418 * NOTE: when configuring this function as a callback with
3419 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3420 * should point to an uint64_t value.
3423 * The previously buffered packets which could not be sent
3425 * The number of unsent packets in the pkts array
3430 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
3434 * Callback function for tracking unsent buffered packets.
3436 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
3437 * adjust the default behavior when buffered packets cannot be sent. This
3438 * function drops any unsent packets, but also updates a user-supplied counter
3439 * to track the overall number of packets dropped. The counter should be an
3440 * uint64_t variable.
3442 * NOTE: this function should not be called directly, instead it should be used
3443 * as a callback for packet buffering.
3445 * NOTE: when configuring this function as a callback with
3446 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
3447 * should point to an uint64_t value.
3450 * The previously buffered packets which could not be sent
3452 * The number of unsent packets in the pkts array
3454 * Pointer to an uint64_t value, which will be incremented by unsent
3457 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
3461 * Request the driver to free mbufs currently cached by the driver. The
3462 * driver will only free the mbuf if it is no longer in use. It is the
3463 * application's responsibity to ensure rte_eth_tx_buffer_flush(..) is
3467 * The port identifier of the Ethernet device.
3469 * The index of the transmit queue through which output packets must be
3471 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
3472 * to rte_eth_dev_configure().
3474 * Maximum number of packets to free. Use 0 to indicate all possible packets
3475 * should be freed. Note that a packet may be using multiple mbufs.
3478 * -ENODEV: Invalid interface
3479 * -ENOTSUP: Driver does not support function
3481 * 0-n: Number of packets freed. More packets may still remain in ring that
3485 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
3488 * The eth device event type for interrupt, and maybe others in the future.
3490 enum rte_eth_event_type {
3491 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
3492 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
3493 RTE_ETH_EVENT_QUEUE_STATE,
3494 /**< queue state event (enabled/disabled) */
3495 RTE_ETH_EVENT_INTR_RESET,
3496 /**< reset interrupt event, sent to VF on PF reset */
3497 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
3498 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
3499 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
3500 RTE_ETH_EVENT_MAX /**< max value of this enum */
3503 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
3504 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
3505 /**< user application callback to be registered for interrupts */
3510 * Register a callback function for specific port id.
3517 * User supplied callback function to be called.
3519 * Pointer to the parameters for the registered callback.
3522 * - On success, zero.
3523 * - On failure, a negative value.
3525 int rte_eth_dev_callback_register(uint16_t port_id,
3526 enum rte_eth_event_type event,
3527 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3530 * Unregister a callback function for specific port id.
3537 * User supplied callback function to be called.
3539 * Pointer to the parameters for the registered callback. -1 means to
3540 * remove all for the same callback address and same event.
3543 * - On success, zero.
3544 * - On failure, a negative value.
3546 int rte_eth_dev_callback_unregister(uint16_t port_id,
3547 enum rte_eth_event_type event,
3548 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
3551 * @internal Executes all the user application registered callbacks for
3552 * the specific device. It is for DPDK internal user only. User
3553 * application should not call it directly.
3556 * Pointer to struct rte_eth_dev.
3558 * Eth device interrupt event type.
3560 * To pass data back to user application.
3561 * This allows the user application to decide if a particular function
3562 * is permitted or not.
3567 int _rte_eth_dev_callback_process(struct rte_eth_dev *dev,
3568 enum rte_eth_event_type event, void *ret_param);
3571 * When there is no rx packet coming in Rx Queue for a long time, we can
3572 * sleep lcore related to RX Queue for power saving, and enable rx interrupt
3573 * to be triggered when Rx packet arrives.
3575 * The rte_eth_dev_rx_intr_enable() function enables rx queue
3576 * interrupt on specific rx queue of a port.
3579 * The port identifier of the Ethernet device.
3581 * The index of the receive queue from which to retrieve input packets.
3582 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3583 * to rte_eth_dev_configure().
3585 * - (0) if successful.
3586 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3588 * - (-ENODEV) if *port_id* invalid.
3590 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
3593 * When lcore wakes up from rx interrupt indicating packet coming, disable rx
3594 * interrupt and returns to polling mode.
3596 * The rte_eth_dev_rx_intr_disable() function disables rx queue
3597 * interrupt on specific rx queue of a port.
3600 * The port identifier of the Ethernet device.
3602 * The index of the receive queue from which to retrieve input packets.
3603 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3604 * to rte_eth_dev_configure().
3606 * - (0) if successful.
3607 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3609 * - (-ENODEV) if *port_id* invalid.
3611 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
3614 * RX Interrupt control per port.
3617 * The port identifier of the Ethernet device.
3619 * Epoll instance fd which the intr vector associated to.
3620 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3622 * The operation be performed for the vector.
3623 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3627 * - On success, zero.
3628 * - On failure, a negative value.
3630 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
3633 * RX Interrupt control per queue.
3636 * The port identifier of the Ethernet device.
3638 * The index of the receive queue from which to retrieve input packets.
3639 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3640 * to rte_eth_dev_configure().
3642 * Epoll instance fd which the intr vector associated to.
3643 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
3645 * The operation be performed for the vector.
3646 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
3650 * - On success, zero.
3651 * - On failure, a negative value.
3653 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
3654 int epfd, int op, void *data);
3657 * Turn on the LED on the Ethernet device.
3658 * This function turns on the LED on the Ethernet device.
3661 * The port identifier of the Ethernet device.
3663 * - (0) if successful.
3664 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3666 * - (-ENODEV) if *port_id* invalid.
3668 int rte_eth_led_on(uint16_t port_id);
3671 * Turn off the LED on the Ethernet device.
3672 * This function turns off the LED on the Ethernet device.
3675 * The port identifier of the Ethernet device.
3677 * - (0) if successful.
3678 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
3680 * - (-ENODEV) if *port_id* invalid.
3682 int rte_eth_led_off(uint16_t port_id);
3685 * Get current status of the Ethernet link flow control for Ethernet device
3688 * The port identifier of the Ethernet device.
3690 * The pointer to the structure where to store the flow control parameters.
3692 * - (0) if successful.
3693 * - (-ENOTSUP) if hardware doesn't support flow control.
3694 * - (-ENODEV) if *port_id* invalid.
3696 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
3697 struct rte_eth_fc_conf *fc_conf);
3700 * Configure the Ethernet link flow control for Ethernet device
3703 * The port identifier of the Ethernet device.
3705 * The pointer to the structure of the flow control parameters.
3707 * - (0) if successful.
3708 * - (-ENOTSUP) if hardware doesn't support flow control mode.
3709 * - (-ENODEV) if *port_id* invalid.
3710 * - (-EINVAL) if bad parameter
3711 * - (-EIO) if flow control setup failure
3713 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
3714 struct rte_eth_fc_conf *fc_conf);
3717 * Configure the Ethernet priority flow control under DCB environment
3718 * for Ethernet device.
3721 * The port identifier of the Ethernet device.
3723 * The pointer to the structure of the priority flow control parameters.
3725 * - (0) if successful.
3726 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
3727 * - (-ENODEV) if *port_id* invalid.
3728 * - (-EINVAL) if bad parameter
3729 * - (-EIO) if flow control setup failure
3731 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
3732 struct rte_eth_pfc_conf *pfc_conf);
3735 * Add a MAC address to an internal array of addresses used to enable whitelist
3736 * filtering to accept packets only if the destination MAC address matches.
3739 * The port identifier of the Ethernet device.
3741 * The MAC address to add.
3743 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
3744 * not enabled, this should be set to 0.
3746 * - (0) if successfully added or *mac_addr" was already added.
3747 * - (-ENOTSUP) if hardware doesn't support this feature.
3748 * - (-ENODEV) if *port* is invalid.
3749 * - (-ENOSPC) if no more MAC addresses can be added.
3750 * - (-EINVAL) if MAC address is invalid.
3752 int rte_eth_dev_mac_addr_add(uint16_t port, struct ether_addr *mac_addr,
3756 * Remove a MAC address from the internal array of addresses.
3759 * The port identifier of the Ethernet device.
3761 * MAC address to remove.
3763 * - (0) if successful, or *mac_addr* didn't exist.
3764 * - (-ENOTSUP) if hardware doesn't support.
3765 * - (-ENODEV) if *port* invalid.
3766 * - (-EADDRINUSE) if attempting to remove the default MAC address
3768 int rte_eth_dev_mac_addr_remove(uint16_t port, struct ether_addr *mac_addr);
3771 * Set the default MAC address.
3774 * The port identifier of the Ethernet device.
3776 * New default MAC address.
3778 * - (0) if successful, or *mac_addr* didn't exist.
3779 * - (-ENOTSUP) if hardware doesn't support.
3780 * - (-ENODEV) if *port* invalid.
3781 * - (-EINVAL) if MAC address is invalid.
3783 int rte_eth_dev_default_mac_addr_set(uint16_t port,
3784 struct ether_addr *mac_addr);
3787 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3790 * The port identifier of the Ethernet device.
3794 * Redirection table size. The table size can be queried by
3795 * rte_eth_dev_info_get().
3797 * - (0) if successful.
3798 * - (-ENOTSUP) if hardware doesn't support.
3799 * - (-EINVAL) if bad parameter.
3801 int rte_eth_dev_rss_reta_update(uint16_t port,
3802 struct rte_eth_rss_reta_entry64 *reta_conf,
3803 uint16_t reta_size);
3806 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
3809 * The port identifier of the Ethernet device.
3813 * Redirection table size. The table size can be queried by
3814 * rte_eth_dev_info_get().
3816 * - (0) if successful.
3817 * - (-ENOTSUP) if hardware doesn't support.
3818 * - (-EINVAL) if bad parameter.
3820 int rte_eth_dev_rss_reta_query(uint16_t port,
3821 struct rte_eth_rss_reta_entry64 *reta_conf,
3822 uint16_t reta_size);
3825 * Updates unicast hash table for receiving packet with the given destination
3826 * MAC address, and the packet is routed to all VFs for which the RX mode is
3827 * accept packets that match the unicast hash table.
3830 * The port identifier of the Ethernet device.
3832 * Unicast MAC address.
3834 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
3835 * 0 - Clear an unicast hash bit.
3837 * - (0) if successful.
3838 * - (-ENOTSUP) if hardware doesn't support.
3839 * - (-ENODEV) if *port_id* invalid.
3840 * - (-EINVAL) if bad parameter.
3842 int rte_eth_dev_uc_hash_table_set(uint16_t port, struct ether_addr *addr,
3846 * Updates all unicast hash bitmaps for receiving packet with any Unicast
3847 * Ethernet MAC addresses,the packet is routed to all VFs for which the RX
3848 * mode is accept packets that match the unicast hash table.
3851 * The port identifier of the Ethernet device.
3853 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
3855 * 0 - Clear all unicast hash bitmaps
3857 * - (0) if successful.
3858 * - (-ENOTSUP) if hardware doesn't support.
3859 * - (-ENODEV) if *port_id* invalid.
3860 * - (-EINVAL) if bad parameter.
3862 int rte_eth_dev_uc_all_hash_table_set(uint16_t port, uint8_t on);
3865 * Set a traffic mirroring rule on an Ethernet device
3868 * The port identifier of the Ethernet device.
3869 * @param mirror_conf
3870 * The pointer to the traffic mirroring structure describing the mirroring rule.
3871 * The *rte_eth_vm_mirror_conf* structure includes the type of mirroring rule,
3872 * destination pool and the value of rule if enable vlan or pool mirroring.
3875 * The index of traffic mirroring rule, we support four separated rules.
3877 * 1 - Enable a mirroring rule.
3878 * 0 - Disable a mirroring rule.
3880 * - (0) if successful.
3881 * - (-ENOTSUP) if hardware doesn't support this feature.
3882 * - (-ENODEV) if *port_id* invalid.
3883 * - (-EINVAL) if the mr_conf information is not correct.
3885 int rte_eth_mirror_rule_set(uint16_t port_id,
3886 struct rte_eth_mirror_conf *mirror_conf,
3891 * Reset a traffic mirroring rule on an Ethernet device.
3894 * The port identifier of the Ethernet device.
3896 * The index of traffic mirroring rule, we support four separated rules.
3898 * - (0) if successful.
3899 * - (-ENOTSUP) if hardware doesn't support this feature.
3900 * - (-ENODEV) if *port_id* invalid.
3901 * - (-EINVAL) if bad parameter.
3903 int rte_eth_mirror_rule_reset(uint16_t port_id,
3907 * Set the rate limitation for a queue on an Ethernet device.
3910 * The port identifier of the Ethernet device.
3914 * The tx rate in Mbps. Allocated from the total port link speed.
3916 * - (0) if successful.
3917 * - (-ENOTSUP) if hardware doesn't support this feature.
3918 * - (-ENODEV) if *port_id* invalid.
3919 * - (-EINVAL) if bad parameter.
3921 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
3925 * Configuration of Receive Side Scaling hash computation of Ethernet device.
3928 * The port identifier of the Ethernet device.
3930 * The new configuration to use for RSS hash computation on the port.
3932 * - (0) if successful.
3933 * - (-ENODEV) if port identifier is invalid.
3934 * - (-ENOTSUP) if hardware doesn't support.
3935 * - (-EINVAL) if bad parameter.
3937 int rte_eth_dev_rss_hash_update(uint16_t port_id,
3938 struct rte_eth_rss_conf *rss_conf);
3941 * Retrieve current configuration of Receive Side Scaling hash computation
3942 * of Ethernet device.
3945 * The port identifier of the Ethernet device.
3947 * Where to store the current RSS hash configuration of the Ethernet device.
3949 * - (0) if successful.
3950 * - (-ENODEV) if port identifier is invalid.
3951 * - (-ENOTSUP) if hardware doesn't support RSS.
3954 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
3955 struct rte_eth_rss_conf *rss_conf);
3958 * Add UDP tunneling port for a specific type of tunnel.
3959 * The packets with this UDP port will be identified as this type of tunnel.
3960 * Before enabling any offloading function for a tunnel, users can call this API
3961 * to change or add more UDP port for the tunnel. So the offloading function
3962 * can take effect on the packets with the specific UDP port.
3965 * The port identifier of the Ethernet device.
3967 * UDP tunneling configuration.
3970 * - (0) if successful.
3971 * - (-ENODEV) if port identifier is invalid.
3972 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3975 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
3976 struct rte_eth_udp_tunnel *tunnel_udp);
3979 * Delete UDP tunneling port a specific type of tunnel.
3980 * The packets with this UDP port will not be identified as this type of tunnel
3982 * Before enabling any offloading function for a tunnel, users can call this API
3983 * to delete a UDP port for the tunnel. So the offloading function will not take
3984 * effect on the packets with the specific UDP port.
3987 * The port identifier of the Ethernet device.
3989 * UDP tunneling configuration.
3992 * - (0) if successful.
3993 * - (-ENODEV) if port identifier is invalid.
3994 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3997 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
3998 struct rte_eth_udp_tunnel *tunnel_udp);
4001 * Check whether the filter type is supported on an Ethernet device.
4002 * All the supported filter types are defined in 'rte_eth_ctrl.h'.
4005 * The port identifier of the Ethernet device.
4006 * @param filter_type
4009 * - (0) if successful.
4010 * - (-ENOTSUP) if hardware doesn't support this filter type.
4011 * - (-ENODEV) if *port_id* invalid.
4013 int rte_eth_dev_filter_supported(uint16_t port_id,
4014 enum rte_filter_type filter_type);
4017 * Take operations to assigned filter type on an Ethernet device.
4018 * All the supported operations and filter types are defined in 'rte_eth_ctrl.h'.
4021 * The port identifier of the Ethernet device.
4022 * @param filter_type
4025 * Type of operation.
4027 * A pointer to arguments defined specifically for the operation.
4029 * - (0) if successful.
4030 * - (-ENOTSUP) if hardware doesn't support.
4031 * - (-ENODEV) if *port_id* invalid.
4032 * - others depends on the specific operations implementation.
4034 int rte_eth_dev_filter_ctrl(uint16_t port_id, enum rte_filter_type filter_type,
4035 enum rte_filter_op filter_op, void *arg);
4038 * Get DCB information on an Ethernet device.
4041 * The port identifier of the Ethernet device.
4045 * - (0) if successful.
4046 * - (-ENODEV) if port identifier is invalid.
4047 * - (-ENOTSUP) if hardware doesn't support.
4049 int rte_eth_dev_get_dcb_info(uint16_t port_id,
4050 struct rte_eth_dcb_info *dcb_info);
4053 * Add a callback to be called on packet RX on a given port and queue.
4055 * This API configures a function to be called for each burst of
4056 * packets received on a given NIC port queue. The return value is a pointer
4057 * that can be used to later remove the callback using
4058 * rte_eth_remove_rx_callback().
4060 * Multiple functions are called in the order that they are added.
4063 * The port identifier of the Ethernet device.
4065 * The queue on the Ethernet device on which the callback is to be added.
4067 * The callback function
4069 * A generic pointer parameter which will be passed to each invocation of the
4070 * callback function on this port and queue.
4074 * On success, a pointer value which can later be used to remove the callback.
4076 void *rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
4077 rte_rx_callback_fn fn, void *user_param);
4080 * Add a callback that must be called first on packet RX on a given port
4083 * This API configures a first function to be called for each burst of
4084 * packets received on a given NIC port queue. The return value is a pointer
4085 * that can be used to later remove the callback using
4086 * rte_eth_remove_rx_callback().
4088 * Multiple functions are called in the order that they are added.
4091 * The port identifier of the Ethernet device.
4093 * The queue on the Ethernet device on which the callback is to be added.
4095 * The callback function
4097 * A generic pointer parameter which will be passed to each invocation of the
4098 * callback function on this port and queue.
4102 * On success, a pointer value which can later be used to remove the callback.
4104 void *rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
4105 rte_rx_callback_fn fn, void *user_param);
4108 * Add a callback to be called on packet TX on a given port and queue.
4110 * This API configures a function to be called for each burst of
4111 * packets sent on a given NIC port queue. The return value is a pointer
4112 * that can be used to later remove the callback using
4113 * rte_eth_remove_tx_callback().
4115 * Multiple functions are called in the order that they are added.
4118 * The port identifier of the Ethernet device.
4120 * The queue on the Ethernet device on which the callback is to be added.
4122 * The callback function
4124 * A generic pointer parameter which will be passed to each invocation of the
4125 * callback function on this port and queue.
4129 * On success, a pointer value which can later be used to remove the callback.
4131 void *rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
4132 rte_tx_callback_fn fn, void *user_param);
4135 * Remove an RX packet callback from a given port and queue.
4137 * This function is used to removed callbacks that were added to a NIC port
4138 * queue using rte_eth_add_rx_callback().
4140 * Note: the callback is removed from the callback list but it isn't freed
4141 * since the it may still be in use. The memory for the callback can be
4142 * subsequently freed back by the application by calling rte_free():
4144 * - Immediately - if the port is stopped, or the user knows that no
4145 * callbacks are in flight e.g. if called from the thread doing RX/TX
4148 * - After a short delay - where the delay is sufficient to allow any
4149 * in-flight callbacks to complete.
4152 * The port identifier of the Ethernet device.
4154 * The queue on the Ethernet device from which the callback is to be removed.
4156 * User supplied callback created via rte_eth_add_rx_callback().
4159 * - 0: Success. Callback was removed.
4160 * - -ENOTSUP: Callback support is not available.
4161 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
4162 * is NULL or not found for the port/queue.
4164 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
4165 struct rte_eth_rxtx_callback *user_cb);
4168 * Remove a TX packet callback from a given port and queue.
4170 * This function is used to removed callbacks that were added to a NIC port
4171 * queue using rte_eth_add_tx_callback().
4173 * Note: the callback is removed from the callback list but it isn't freed
4174 * since the it may still be in use. The memory for the callback can be
4175 * subsequently freed back by the application by calling rte_free():
4177 * - Immediately - if the port is stopped, or the user knows that no
4178 * callbacks are in flight e.g. if called from the thread doing RX/TX
4181 * - After a short delay - where the delay is sufficient to allow any
4182 * in-flight callbacks to complete.
4185 * The port identifier of the Ethernet device.
4187 * The queue on the Ethernet device from which the callback is to be removed.
4189 * User supplied callback created via rte_eth_add_tx_callback().
4192 * - 0: Success. Callback was removed.
4193 * - -ENOTSUP: Callback support is not available.
4194 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
4195 * is NULL or not found for the port/queue.
4197 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
4198 struct rte_eth_rxtx_callback *user_cb);
4201 * Retrieve information about given port's RX queue.
4204 * The port identifier of the Ethernet device.
4206 * The RX queue on the Ethernet device for which information
4207 * will be retrieved.
4209 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
4210 * the information of the Ethernet device.
4214 * - -ENOTSUP: routine is not supported by the device PMD.
4215 * - -EINVAL: The port_id or the queue_id is out of range.
4217 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4218 struct rte_eth_rxq_info *qinfo);
4221 * Retrieve information about given port's TX queue.
4224 * The port identifier of the Ethernet device.
4226 * The TX queue on the Ethernet device for which information
4227 * will be retrieved.
4229 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
4230 * the information of the Ethernet device.
4234 * - -ENOTSUP: routine is not supported by the device PMD.
4235 * - -EINVAL: The port_id or the queue_id is out of range.
4237 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
4238 struct rte_eth_txq_info *qinfo);
4241 * Retrieve device registers and register attributes (number of registers and
4245 * The port identifier of the Ethernet device.
4247 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
4248 * NULL the function fills in the width and length fields. If non-NULL
4249 * the registers are put into the buffer pointed at by the data field.
4251 * - (0) if successful.
4252 * - (-ENOTSUP) if hardware doesn't support.
4253 * - (-ENODEV) if *port_id* invalid.
4254 * - others depends on the specific operations implementation.
4256 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
4259 * Retrieve size of device EEPROM
4262 * The port identifier of the Ethernet device.
4264 * - (>=0) EEPROM size if successful.
4265 * - (-ENOTSUP) if hardware doesn't support.
4266 * - (-ENODEV) if *port_id* invalid.
4267 * - others depends on the specific operations implementation.
4269 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
4272 * Retrieve EEPROM and EEPROM attribute
4275 * The port identifier of the Ethernet device.
4277 * The template includes buffer for return EEPROM data and
4278 * EEPROM attributes to be filled.
4280 * - (0) if successful.
4281 * - (-ENOTSUP) if hardware doesn't support.
4282 * - (-ENODEV) if *port_id* invalid.
4283 * - others depends on the specific operations implementation.
4285 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4288 * Program EEPROM with provided data
4291 * The port identifier of the Ethernet device.
4293 * The template includes EEPROM data for programming and
4294 * EEPROM attributes to be filled
4296 * - (0) if successful.
4297 * - (-ENOTSUP) if hardware doesn't support.
4298 * - (-ENODEV) if *port_id* invalid.
4299 * - others depends on the specific operations implementation.
4301 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
4304 * Set the list of multicast addresses to filter on an Ethernet device.
4307 * The port identifier of the Ethernet device.
4308 * @param mc_addr_set
4309 * The array of multicast addresses to set. Equal to NULL when the function
4310 * is invoked to flush the set of filtered addresses.
4312 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
4313 * when the function is invoked to flush the set of filtered addresses.
4315 * - (0) if successful.
4316 * - (-ENODEV) if *port_id* invalid.
4317 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
4318 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
4320 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
4321 struct ether_addr *mc_addr_set,
4322 uint32_t nb_mc_addr);
4325 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
4328 * The port identifier of the Ethernet device.
4332 * - -ENODEV: The port ID is invalid.
4333 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4335 int rte_eth_timesync_enable(uint16_t port_id);
4338 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
4341 * The port identifier of the Ethernet device.
4345 * - -ENODEV: The port ID is invalid.
4346 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4348 int rte_eth_timesync_disable(uint16_t port_id);
4351 * Read an IEEE1588/802.1AS RX timestamp from an Ethernet device.
4354 * The port identifier of the Ethernet device.
4356 * Pointer to the timestamp struct.
4358 * Device specific flags. Used to pass the RX timesync register index to
4359 * i40e. Unused in igb/ixgbe, pass 0 instead.
4363 * - -EINVAL: No timestamp is available.
4364 * - -ENODEV: The port ID is invalid.
4365 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4367 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
4368 struct timespec *timestamp, uint32_t flags);
4371 * Read an IEEE1588/802.1AS TX timestamp from an Ethernet device.
4374 * The port identifier of the Ethernet device.
4376 * Pointer to the timestamp struct.
4380 * - -EINVAL: No timestamp is available.
4381 * - -ENODEV: The port ID is invalid.
4382 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4384 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
4385 struct timespec *timestamp);
4388 * Adjust the timesync clock on an Ethernet device.
4390 * This is usually used in conjunction with other Ethdev timesync functions to
4391 * synchronize the device time using the IEEE1588/802.1AS protocol.
4394 * The port identifier of the Ethernet device.
4396 * The adjustment in nanoseconds.
4400 * - -ENODEV: The port ID is invalid.
4401 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4403 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
4406 * Read the time from the timesync clock on an Ethernet device.
4408 * This is usually used in conjunction with other Ethdev timesync functions to
4409 * synchronize the device time using the IEEE1588/802.1AS protocol.
4412 * The port identifier of the Ethernet device.
4414 * Pointer to the timespec struct that holds the time.
4419 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
4422 * Set the time of the timesync clock on an Ethernet device.
4424 * This is usually used in conjunction with other Ethdev timesync functions to
4425 * synchronize the device time using the IEEE1588/802.1AS protocol.
4428 * The port identifier of the Ethernet device.
4430 * Pointer to the timespec struct that holds the time.
4434 * - -EINVAL: No timestamp is available.
4435 * - -ENODEV: The port ID is invalid.
4436 * - -ENOTSUP: The function is not supported by the Ethernet driver.
4438 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
4441 * Create memzone for HW rings.
4442 * malloc can't be used as the physical address is needed.
4443 * If the memzone is already created, then this function returns a ptr
4447 * The *eth_dev* pointer is the address of the *rte_eth_dev* structure
4449 * The name of the memory zone
4451 * The index of the queue to add to name
4453 * The sizeof of the memory area
4455 * Alignment for resulting memzone. Must be a power of 2.
4457 * The *socket_id* argument is the socket identifier in case of NUMA.
4459 const struct rte_memzone *
4460 rte_eth_dma_zone_reserve(const struct rte_eth_dev *eth_dev, const char *name,
4461 uint16_t queue_id, size_t size,
4462 unsigned align, int socket_id);
4465 * Config l2 tunnel ether type of an Ethernet device for filtering specific
4466 * tunnel packets by ether type.
4469 * The port identifier of the Ethernet device.
4471 * l2 tunnel configuration.
4474 * - (0) if successful.
4475 * - (-ENODEV) if port identifier is invalid.
4476 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4479 rte_eth_dev_l2_tunnel_eth_type_conf(uint16_t port_id,
4480 struct rte_eth_l2_tunnel_conf *l2_tunnel);
4483 * Enable/disable l2 tunnel offload functions. Include,
4484 * 1, The ability of parsing a type of l2 tunnel of an Ethernet device.
4485 * Filtering, forwarding and offloading this type of tunnel packets depend on
4487 * 2, Stripping the l2 tunnel tag.
4488 * 3, Insertion of the l2 tunnel tag.
4489 * 4, Forwarding the packets based on the l2 tunnel tag.
4492 * The port identifier of the Ethernet device.
4494 * l2 tunnel parameters.
4496 * Indicate the offload function.
4498 * Enable or disable this function.
4501 * - (0) if successful.
4502 * - (-ENODEV) if port identifier is invalid.
4503 * - (-ENOTSUP) if hardware doesn't support tunnel type.
4506 rte_eth_dev_l2_tunnel_offload_set(uint16_t port_id,
4507 struct rte_eth_l2_tunnel_conf *l2_tunnel,
4512 * Get the port id from pci address or device name
4513 * Ex: 0000:2:00.0 or vdev name net_pcap0
4516 * pci address or name of the device
4518 * pointer to port identifier of the device
4520 * - (0) if successful and port_id is filled.
4521 * - (-ENODEV or -EINVAL) on failure.
4524 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
4527 * Get the device name from port id
4530 * pointer to port identifier of the device
4532 * pci address or name of the device
4534 * - (0) if successful.
4535 * - (-EINVAL) on failure.
4538 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
4541 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
4542 * the ethernet device information, otherwise adjust them to boundaries.
4545 * The port identifier of the Ethernet device.
4547 * A pointer to a uint16_t where the number of receive
4548 * descriptors stored.
4550 * A pointer to a uint16_t where the number of transmit
4551 * descriptors stored.
4553 * - (0) if successful.
4554 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
4556 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
4557 uint16_t *nb_rx_desc,
4558 uint16_t *nb_tx_desc);
4562 * Test if a port supports specific mempool ops.
4565 * Port identifier of the Ethernet device.
4567 * The name of the pool operations to test.
4569 * - 0: best mempool ops choice for this port.
4570 * - 1: mempool ops are supported for this port.
4571 * - -ENOTSUP: mempool ops not supported for this port.
4572 * - -ENODEV: Invalid port Identifier.
4573 * - -EINVAL: Pool param is null.
4576 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
4582 #endif /* _RTE_ETHDEV_H_ */