1 /* SPDX-License-Identifier: BSD-3-Clause
2 * Copyright(c) 2010-2017 Intel Corporation
11 * RTE Ethernet Device API
13 * The Ethernet Device API is composed of two parts:
15 * - The application-oriented Ethernet API that includes functions to setup
16 * an Ethernet device (configure it, setup its RX and TX queues and start it),
17 * to get its MAC address, the speed and the status of its physical link,
18 * to receive and to transmit packets, and so on.
20 * - The driver-oriented Ethernet API that exports functions allowing
21 * an Ethernet Poll Mode Driver (PMD) to allocate an Ethernet device instance,
22 * create memzone for HW rings and process registered callbacks, and so on.
23 * PMDs should include rte_ethdev_driver.h instead of this header.
25 * By default, all the functions of the Ethernet Device API exported by a PMD
26 * are lock-free functions which assume to not be invoked in parallel on
27 * different logical cores to work on the same target object. For instance,
28 * the receive function of a PMD cannot be invoked in parallel on two logical
29 * cores to poll the same RX queue [of the same port]. Of course, this function
30 * can be invoked in parallel by different logical cores on different RX queues.
31 * It is the responsibility of the upper level application to enforce this rule.
33 * If needed, parallel accesses by multiple logical cores to shared queues
34 * shall be explicitly protected by dedicated inline lock-aware functions
35 * built on top of their corresponding lock-free functions of the PMD API.
37 * In all functions of the Ethernet API, the Ethernet device is
38 * designated by an integer >= 0 named the device port identifier.
40 * At the Ethernet driver level, Ethernet devices are represented by a generic
41 * data structure of type *rte_eth_dev*.
43 * Ethernet devices are dynamically registered during the PCI probing phase
44 * performed at EAL initialization time.
45 * When an Ethernet device is being probed, an *rte_eth_dev* structure and
46 * a new port identifier are allocated for that device. Then, the eth_dev_init()
47 * function supplied by the Ethernet driver matching the probed PCI
48 * device is invoked to properly initialize the device.
50 * The role of the device init function consists of resetting the hardware,
51 * checking access to Non-volatile Memory (NVM), reading the MAC address
54 * If the device init operation is successful, the correspondence between
55 * the port identifier assigned to the new device and its associated
56 * *rte_eth_dev* structure is effectively registered.
57 * Otherwise, both the *rte_eth_dev* structure and the port identifier are
60 * The functions exported by the application Ethernet API to setup a device
61 * designated by its port identifier must be invoked in the following order:
62 * - rte_eth_dev_configure()
63 * - rte_eth_tx_queue_setup()
64 * - rte_eth_rx_queue_setup()
65 * - rte_eth_dev_start()
67 * Then, the network application can invoke, in any order, the functions
68 * exported by the Ethernet API to get the MAC address of a given device, to
69 * get the speed and the status of a device physical link, to receive/transmit
70 * [burst of] packets, and so on.
72 * If the application wants to change the configuration (i.e. call
73 * rte_eth_dev_configure(), rte_eth_tx_queue_setup(), or
74 * rte_eth_rx_queue_setup()), it must call rte_eth_dev_stop() first to stop the
75 * device and then do the reconfiguration before calling rte_eth_dev_start()
76 * again. The transmit and receive functions should not be invoked when the
79 * Please note that some configuration is not stored between calls to
80 * rte_eth_dev_stop()/rte_eth_dev_start(). The following configuration will
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
149 #include <rte_compat.h>
151 #include <rte_interrupts.h>
153 #include <rte_devargs.h>
154 #include <rte_errno.h>
155 #include <rte_common.h>
156 #include <rte_config.h>
158 #include "rte_ether.h"
159 #include "rte_eth_ctrl.h"
160 #include "rte_dev_info.h"
162 extern int rte_eth_dev_logtype;
164 #define RTE_ETHDEV_LOG(level, ...) \
165 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__)
170 * A structure used to retrieve statistics for an Ethernet port.
171 * Not all statistics fields in struct rte_eth_stats are supported
172 * by any type of network interface card (NIC). If any statistics
173 * field is not supported, its value is 0.
175 struct rte_eth_stats {
176 uint64_t ipackets; /**< Total number of successfully received packets. */
177 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
178 uint64_t ibytes; /**< Total number of successfully received bytes. */
179 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
181 /**< Total of RX packets dropped by the HW,
182 * because there are no available buffer (i.e. RX queues are full).
184 uint64_t ierrors; /**< Total number of erroneous received packets. */
185 uint64_t oerrors; /**< Total number of failed transmitted packets. */
186 uint64_t rx_nombuf; /**< Total number of RX mbuf allocation failures. */
187 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
188 /**< Total number of queue RX packets. */
189 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS];
190 /**< Total number of queue TX packets. */
191 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
192 /**< Total number of successfully received queue bytes. */
193 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS];
194 /**< Total number of successfully transmitted queue bytes. */
195 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS];
196 /**< Total number of queue packets received that are dropped. */
200 * Device supported speeds bitmap flags
202 #define ETH_LINK_SPEED_AUTONEG (0 << 0) /**< Autonegotiate (all speeds) */
203 #define ETH_LINK_SPEED_FIXED (1 << 0) /**< Disable autoneg (fixed speed) */
204 #define ETH_LINK_SPEED_10M_HD (1 << 1) /**< 10 Mbps half-duplex */
205 #define ETH_LINK_SPEED_10M (1 << 2) /**< 10 Mbps full-duplex */
206 #define ETH_LINK_SPEED_100M_HD (1 << 3) /**< 100 Mbps half-duplex */
207 #define ETH_LINK_SPEED_100M (1 << 4) /**< 100 Mbps full-duplex */
208 #define ETH_LINK_SPEED_1G (1 << 5) /**< 1 Gbps */
209 #define ETH_LINK_SPEED_2_5G (1 << 6) /**< 2.5 Gbps */
210 #define ETH_LINK_SPEED_5G (1 << 7) /**< 5 Gbps */
211 #define ETH_LINK_SPEED_10G (1 << 8) /**< 10 Gbps */
212 #define ETH_LINK_SPEED_20G (1 << 9) /**< 20 Gbps */
213 #define ETH_LINK_SPEED_25G (1 << 10) /**< 25 Gbps */
214 #define ETH_LINK_SPEED_40G (1 << 11) /**< 40 Gbps */
215 #define ETH_LINK_SPEED_50G (1 << 12) /**< 50 Gbps */
216 #define ETH_LINK_SPEED_56G (1 << 13) /**< 56 Gbps */
217 #define ETH_LINK_SPEED_100G (1 << 14) /**< 100 Gbps */
220 * Ethernet numeric link speeds in Mbps
222 #define ETH_SPEED_NUM_NONE 0 /**< Not defined */
223 #define ETH_SPEED_NUM_10M 10 /**< 10 Mbps */
224 #define ETH_SPEED_NUM_100M 100 /**< 100 Mbps */
225 #define ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */
226 #define ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */
227 #define ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */
228 #define ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */
229 #define ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */
230 #define ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */
231 #define ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */
232 #define ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */
233 #define ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */
234 #define ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */
237 * A structure used to retrieve link-level information of an Ethernet port.
240 struct rte_eth_link {
241 uint32_t link_speed; /**< ETH_SPEED_NUM_ */
242 uint16_t link_duplex : 1; /**< ETH_LINK_[HALF/FULL]_DUPLEX */
243 uint16_t link_autoneg : 1; /**< ETH_LINK_[AUTONEG/FIXED] */
244 uint16_t link_status : 1; /**< ETH_LINK_[DOWN/UP] */
245 } __attribute__((aligned(8))); /**< aligned for atomic64 read/write */
247 /* Utility constants */
248 #define ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */
249 #define ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */
250 #define ETH_LINK_DOWN 0 /**< Link is down (see link_status). */
251 #define ETH_LINK_UP 1 /**< Link is up (see link_status). */
252 #define ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */
253 #define ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */
256 * A structure used to configure the ring threshold registers of an RX/TX
257 * queue for an Ethernet port.
259 struct rte_eth_thresh {
260 uint8_t pthresh; /**< Ring prefetch threshold. */
261 uint8_t hthresh; /**< Ring host threshold. */
262 uint8_t wthresh; /**< Ring writeback threshold. */
266 * Simple flags are used for rte_eth_conf.rxmode.mq_mode.
268 #define ETH_MQ_RX_RSS_FLAG 0x1
269 #define ETH_MQ_RX_DCB_FLAG 0x2
270 #define ETH_MQ_RX_VMDQ_FLAG 0x4
273 * A set of values to identify what method is to be used to route
274 * packets to multiple queues.
276 enum rte_eth_rx_mq_mode {
277 /** None of DCB,RSS or VMDQ mode */
280 /** For RX side, only RSS is on */
281 ETH_MQ_RX_RSS = ETH_MQ_RX_RSS_FLAG,
282 /** For RX side,only DCB is on. */
283 ETH_MQ_RX_DCB = ETH_MQ_RX_DCB_FLAG,
284 /** Both DCB and RSS enable */
285 ETH_MQ_RX_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG,
287 /** Only VMDQ, no RSS nor DCB */
288 ETH_MQ_RX_VMDQ_ONLY = ETH_MQ_RX_VMDQ_FLAG,
289 /** RSS mode with VMDQ */
290 ETH_MQ_RX_VMDQ_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_VMDQ_FLAG,
291 /** Use VMDQ+DCB to route traffic to queues */
292 ETH_MQ_RX_VMDQ_DCB = ETH_MQ_RX_VMDQ_FLAG | ETH_MQ_RX_DCB_FLAG,
293 /** Enable both VMDQ and DCB in VMDq */
294 ETH_MQ_RX_VMDQ_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG |
299 * for rx mq mode backward compatible
301 #define ETH_RSS ETH_MQ_RX_RSS
302 #define VMDQ_DCB ETH_MQ_RX_VMDQ_DCB
303 #define ETH_DCB_RX ETH_MQ_RX_DCB
306 * A set of values to identify what method is to be used to transmit
307 * packets using multi-TCs.
309 enum rte_eth_tx_mq_mode {
310 ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */
311 ETH_MQ_TX_DCB, /**< For TX side,only DCB is on. */
312 ETH_MQ_TX_VMDQ_DCB, /**< For TX side,both DCB and VT is on. */
313 ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */
317 * for tx mq mode backward compatible
319 #define ETH_DCB_NONE ETH_MQ_TX_NONE
320 #define ETH_VMDQ_DCB_TX ETH_MQ_TX_VMDQ_DCB
321 #define ETH_DCB_TX ETH_MQ_TX_DCB
324 * A structure used to configure the RX features of an Ethernet port.
326 struct rte_eth_rxmode {
327 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
328 enum rte_eth_rx_mq_mode mq_mode;
329 uint32_t max_rx_pkt_len; /**< Only used if jumbo_frame enabled. */
330 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
332 * Per-port Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
333 * Only offloads set on rx_offload_capa field on rte_eth_dev_info
334 * structure are allowed to be set.
339 * Below bitfield API is obsolete. Application should
340 * enable per-port offloads using the offload field
343 uint16_t header_split : 1, /**< Header Split enable. */
344 hw_ip_checksum : 1, /**< IP/UDP/TCP checksum offload enable. */
345 hw_vlan_filter : 1, /**< VLAN filter enable. */
346 hw_vlan_strip : 1, /**< VLAN strip enable. */
347 hw_vlan_extend : 1, /**< Extended VLAN enable. */
348 jumbo_frame : 1, /**< Jumbo Frame Receipt enable. */
349 hw_strip_crc : 1, /**< Enable CRC stripping by hardware. */
350 enable_scatter : 1, /**< Enable scatter packets rx handler */
351 enable_lro : 1, /**< Enable LRO */
352 hw_timestamp : 1, /**< Enable HW timestamp */
353 security : 1, /**< Enable rte_security offloads */
355 * When set the offload bitfield should be ignored.
356 * Instead per-port Rx offloads should be set on offloads
358 * Per-queue offloads shuold be set on rte_eth_rxq_conf
360 * This bit is temporary till rxmode bitfield offloads API will
363 ignore_offload_bitfield : 1;
367 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN.
368 * Note that single VLAN is treated the same as inner VLAN.
371 ETH_VLAN_TYPE_UNKNOWN = 0,
372 ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */
373 ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */
378 * A structure used to describe a vlan filter.
379 * If the bit corresponding to a VID is set, such VID is on.
381 struct rte_vlan_filter_conf {
386 * A structure used to configure the Receive Side Scaling (RSS) feature
387 * of an Ethernet port.
388 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
389 * to an array holding the RSS key to use for hashing specific header
390 * fields of received packets. The length of this array should be indicated
391 * by *rss_key_len* below. Otherwise, a default random hash key is used by
394 * The *rss_key_len* field of the *rss_conf* structure indicates the length
395 * in bytes of the array pointed by *rss_key*. To be compatible, this length
396 * will be checked in i40e only. Others assume 40 bytes to be used as before.
398 * The *rss_hf* field of the *rss_conf* structure indicates the different
399 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
400 * Supplying an *rss_hf* equal to zero disables the RSS feature.
402 struct rte_eth_rss_conf {
403 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
404 uint8_t rss_key_len; /**< hash key length in bytes. */
405 uint64_t rss_hf; /**< Hash functions to apply - see below. */
409 * The RSS offload types are defined based on flow types which are defined
410 * in rte_eth_ctrl.h. Different NIC hardwares may support different RSS offload
411 * types. The supported flow types or RSS offload types can be queried by
412 * rte_eth_dev_info_get().
414 #define ETH_RSS_IPV4 (1ULL << RTE_ETH_FLOW_IPV4)
415 #define ETH_RSS_FRAG_IPV4 (1ULL << RTE_ETH_FLOW_FRAG_IPV4)
416 #define ETH_RSS_NONFRAG_IPV4_TCP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_TCP)
417 #define ETH_RSS_NONFRAG_IPV4_UDP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_UDP)
418 #define ETH_RSS_NONFRAG_IPV4_SCTP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_SCTP)
419 #define ETH_RSS_NONFRAG_IPV4_OTHER (1ULL << RTE_ETH_FLOW_NONFRAG_IPV4_OTHER)
420 #define ETH_RSS_IPV6 (1ULL << RTE_ETH_FLOW_IPV6)
421 #define ETH_RSS_FRAG_IPV6 (1ULL << RTE_ETH_FLOW_FRAG_IPV6)
422 #define ETH_RSS_NONFRAG_IPV6_TCP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_TCP)
423 #define ETH_RSS_NONFRAG_IPV6_UDP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_UDP)
424 #define ETH_RSS_NONFRAG_IPV6_SCTP (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_SCTP)
425 #define ETH_RSS_NONFRAG_IPV6_OTHER (1ULL << RTE_ETH_FLOW_NONFRAG_IPV6_OTHER)
426 #define ETH_RSS_L2_PAYLOAD (1ULL << RTE_ETH_FLOW_L2_PAYLOAD)
427 #define ETH_RSS_IPV6_EX (1ULL << RTE_ETH_FLOW_IPV6_EX)
428 #define ETH_RSS_IPV6_TCP_EX (1ULL << RTE_ETH_FLOW_IPV6_TCP_EX)
429 #define ETH_RSS_IPV6_UDP_EX (1ULL << RTE_ETH_FLOW_IPV6_UDP_EX)
430 #define ETH_RSS_PORT (1ULL << RTE_ETH_FLOW_PORT)
431 #define ETH_RSS_VXLAN (1ULL << RTE_ETH_FLOW_VXLAN)
432 #define ETH_RSS_GENEVE (1ULL << RTE_ETH_FLOW_GENEVE)
433 #define ETH_RSS_NVGRE (1ULL << RTE_ETH_FLOW_NVGRE)
435 #define ETH_RSS_IP ( \
437 ETH_RSS_FRAG_IPV4 | \
438 ETH_RSS_NONFRAG_IPV4_OTHER | \
440 ETH_RSS_FRAG_IPV6 | \
441 ETH_RSS_NONFRAG_IPV6_OTHER | \
444 #define ETH_RSS_UDP ( \
445 ETH_RSS_NONFRAG_IPV4_UDP | \
446 ETH_RSS_NONFRAG_IPV6_UDP | \
449 #define ETH_RSS_TCP ( \
450 ETH_RSS_NONFRAG_IPV4_TCP | \
451 ETH_RSS_NONFRAG_IPV6_TCP | \
454 #define ETH_RSS_SCTP ( \
455 ETH_RSS_NONFRAG_IPV4_SCTP | \
456 ETH_RSS_NONFRAG_IPV6_SCTP)
458 #define ETH_RSS_TUNNEL ( \
463 /**< Mask of valid RSS hash protocols */
464 #define ETH_RSS_PROTO_MASK ( \
466 ETH_RSS_FRAG_IPV4 | \
467 ETH_RSS_NONFRAG_IPV4_TCP | \
468 ETH_RSS_NONFRAG_IPV4_UDP | \
469 ETH_RSS_NONFRAG_IPV4_SCTP | \
470 ETH_RSS_NONFRAG_IPV4_OTHER | \
472 ETH_RSS_FRAG_IPV6 | \
473 ETH_RSS_NONFRAG_IPV6_TCP | \
474 ETH_RSS_NONFRAG_IPV6_UDP | \
475 ETH_RSS_NONFRAG_IPV6_SCTP | \
476 ETH_RSS_NONFRAG_IPV6_OTHER | \
477 ETH_RSS_L2_PAYLOAD | \
479 ETH_RSS_IPV6_TCP_EX | \
480 ETH_RSS_IPV6_UDP_EX | \
487 * Definitions used for redirection table entry size.
488 * Some RSS RETA sizes may not be supported by some drivers, check the
489 * documentation or the description of relevant functions for more details.
491 #define ETH_RSS_RETA_SIZE_64 64
492 #define ETH_RSS_RETA_SIZE_128 128
493 #define ETH_RSS_RETA_SIZE_256 256
494 #define ETH_RSS_RETA_SIZE_512 512
495 #define RTE_RETA_GROUP_SIZE 64
497 /* Definitions used for VMDQ and DCB functionality */
498 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */
499 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
500 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */
501 #define ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */
503 /* DCB capability defines */
504 #define ETH_DCB_PG_SUPPORT 0x00000001 /**< Priority Group(ETS) support. */
505 #define ETH_DCB_PFC_SUPPORT 0x00000002 /**< Priority Flow Control support. */
507 /* Definitions used for VLAN Offload functionality */
508 #define ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */
509 #define ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */
510 #define ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */
512 /* Definitions used for mask VLAN setting */
513 #define ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */
514 #define ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/
515 #define ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/
516 #define ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/
518 /* Definitions used for receive MAC address */
519 #define ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */
521 /* Definitions used for unicast hash */
522 #define ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */
524 /* Definitions used for VMDQ pool rx mode setting */
525 #define ETH_VMDQ_ACCEPT_UNTAG 0x0001 /**< accept untagged packets. */
526 #define ETH_VMDQ_ACCEPT_HASH_MC 0x0002 /**< accept packets in multicast table . */
527 #define ETH_VMDQ_ACCEPT_HASH_UC 0x0004 /**< accept packets in unicast table. */
528 #define ETH_VMDQ_ACCEPT_BROADCAST 0x0008 /**< accept broadcast packets. */
529 #define ETH_VMDQ_ACCEPT_MULTICAST 0x0010 /**< multicast promiscuous. */
531 /** Maximum nb. of vlan per mirror rule */
532 #define ETH_MIRROR_MAX_VLANS 64
534 #define ETH_MIRROR_VIRTUAL_POOL_UP 0x01 /**< Virtual Pool uplink Mirroring. */
535 #define ETH_MIRROR_UPLINK_PORT 0x02 /**< Uplink Port Mirroring. */
536 #define ETH_MIRROR_DOWNLINK_PORT 0x04 /**< Downlink Port Mirroring. */
537 #define ETH_MIRROR_VLAN 0x08 /**< VLAN Mirroring. */
538 #define ETH_MIRROR_VIRTUAL_POOL_DOWN 0x10 /**< Virtual Pool downlink Mirroring. */
541 * A structure used to configure VLAN traffic mirror of an Ethernet port.
543 struct rte_eth_vlan_mirror {
544 uint64_t vlan_mask; /**< mask for valid VLAN ID. */
545 /** VLAN ID list for vlan mirroring. */
546 uint16_t vlan_id[ETH_MIRROR_MAX_VLANS];
550 * A structure used to configure traffic mirror of an Ethernet port.
552 struct rte_eth_mirror_conf {
553 uint8_t rule_type; /**< Mirroring rule type */
554 uint8_t dst_pool; /**< Destination pool for this mirror rule. */
555 uint64_t pool_mask; /**< Bitmap of pool for pool mirroring */
556 /** VLAN ID setting for VLAN mirroring. */
557 struct rte_eth_vlan_mirror vlan;
561 * A structure used to configure 64 entries of Redirection Table of the
562 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure
563 * more than 64 entries supported by hardware, an array of this structure
566 struct rte_eth_rss_reta_entry64 {
568 /**< Mask bits indicate which entries need to be updated/queried. */
569 uint16_t reta[RTE_RETA_GROUP_SIZE];
570 /**< Group of 64 redirection table entries. */
574 * This enum indicates the possible number of traffic classes
575 * in DCB configurations
577 enum rte_eth_nb_tcs {
578 ETH_4_TCS = 4, /**< 4 TCs with DCB. */
579 ETH_8_TCS = 8 /**< 8 TCs with DCB. */
583 * This enum indicates the possible number of queue pools
584 * in VMDQ configurations.
586 enum rte_eth_nb_pools {
587 ETH_8_POOLS = 8, /**< 8 VMDq pools. */
588 ETH_16_POOLS = 16, /**< 16 VMDq pools. */
589 ETH_32_POOLS = 32, /**< 32 VMDq pools. */
590 ETH_64_POOLS = 64 /**< 64 VMDq pools. */
593 /* This structure may be extended in future. */
594 struct rte_eth_dcb_rx_conf {
595 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */
596 /** Traffic class each UP mapped to. */
597 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
600 struct rte_eth_vmdq_dcb_tx_conf {
601 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */
602 /** Traffic class each UP mapped to. */
603 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
606 struct rte_eth_dcb_tx_conf {
607 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */
608 /** Traffic class each UP mapped to. */
609 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
612 struct rte_eth_vmdq_tx_conf {
613 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */
617 * A structure used to configure the VMDQ+DCB feature
618 * of an Ethernet port.
620 * Using this feature, packets are routed to a pool of queues, based
621 * on the vlan id in the vlan tag, and then to a specific queue within
622 * that pool, using the user priority vlan tag field.
624 * A default pool may be used, if desired, to route all traffic which
625 * does not match the vlan filter rules.
627 struct rte_eth_vmdq_dcb_conf {
628 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
629 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
630 uint8_t default_pool; /**< The default pool, if applicable */
631 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
633 uint16_t vlan_id; /**< The vlan id of the received frame */
634 uint64_t pools; /**< Bitmask of pools for packet rx */
635 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
636 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES];
637 /**< Selects a queue in a pool */
641 * A structure used to configure the VMDQ feature of an Ethernet port when
642 * not combined with the DCB feature.
644 * Using this feature, packets are routed to a pool of queues. By default,
645 * the pool selection is based on the MAC address, the vlan id in the
646 * vlan tag as specified in the pool_map array.
647 * Passing the ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool
648 * selection using only the MAC address. MAC address to pool mapping is done
649 * using the rte_eth_dev_mac_addr_add function, with the pool parameter
650 * corresponding to the pool id.
652 * Queue selection within the selected pool will be done using RSS when
653 * it is enabled or revert to the first queue of the pool if not.
655 * A default pool may be used, if desired, to route all traffic which
656 * does not match the vlan filter rules or any pool MAC address.
658 struct rte_eth_vmdq_rx_conf {
659 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */
660 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
661 uint8_t default_pool; /**< The default pool, if applicable */
662 uint8_t enable_loop_back; /**< Enable VT loop back */
663 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
664 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */
666 uint16_t vlan_id; /**< The vlan id of the received frame */
667 uint64_t pools; /**< Bitmask of pools for packet rx */
668 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
672 * A structure used to configure the TX features of an Ethernet port.
674 struct rte_eth_txmode {
675 enum rte_eth_tx_mq_mode mq_mode; /**< TX multi-queues mode. */
677 * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
678 * Only offloads set on tx_offload_capa field on rte_eth_dev_info
679 * structure are allowed to be set.
683 /* For i40e specifically */
686 uint8_t hw_vlan_reject_tagged : 1,
687 /**< If set, reject sending out tagged pkts */
688 hw_vlan_reject_untagged : 1,
689 /**< If set, reject sending out untagged pkts */
690 hw_vlan_insert_pvid : 1;
691 /**< If set, enable port based VLAN insertion */
695 * A structure used to configure an RX ring of an Ethernet port.
697 struct rte_eth_rxconf {
698 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */
699 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */
700 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */
701 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
703 * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags.
704 * Only offloads set on rx_queue_offload_capa or rx_offload_capa
705 * fields on rte_eth_dev_info structure are allowed to be set.
710 #define ETH_TXQ_FLAGS_NOMULTSEGS 0x0001 /**< nb_segs=1 for all mbufs */
711 #define ETH_TXQ_FLAGS_NOREFCOUNT 0x0002 /**< refcnt can be ignored */
712 #define ETH_TXQ_FLAGS_NOMULTMEMP 0x0004 /**< all bufs come from same mempool */
713 #define ETH_TXQ_FLAGS_NOVLANOFFL 0x0100 /**< disable VLAN offload */
714 #define ETH_TXQ_FLAGS_NOXSUMSCTP 0x0200 /**< disable SCTP checksum offload */
715 #define ETH_TXQ_FLAGS_NOXSUMUDP 0x0400 /**< disable UDP checksum offload */
716 #define ETH_TXQ_FLAGS_NOXSUMTCP 0x0800 /**< disable TCP checksum offload */
717 #define ETH_TXQ_FLAGS_NOOFFLOADS \
718 (ETH_TXQ_FLAGS_NOVLANOFFL | ETH_TXQ_FLAGS_NOXSUMSCTP | \
719 ETH_TXQ_FLAGS_NOXSUMUDP | ETH_TXQ_FLAGS_NOXSUMTCP)
720 #define ETH_TXQ_FLAGS_NOXSUMS \
721 (ETH_TXQ_FLAGS_NOXSUMSCTP | ETH_TXQ_FLAGS_NOXSUMUDP | \
722 ETH_TXQ_FLAGS_NOXSUMTCP)
724 * When set the txq_flags should be ignored,
725 * instead per-queue Tx offloads will be set on offloads field
726 * located on rte_eth_txq_conf struct.
727 * This flag is temporary till the rte_eth_txq_conf.txq_flags
728 * API will be deprecated.
730 #define ETH_TXQ_FLAGS_IGNORE 0x8000
733 * A structure used to configure a TX ring of an Ethernet port.
735 struct rte_eth_txconf {
736 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */
737 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
738 uint16_t tx_free_thresh; /**< Start freeing TX buffers if there are
739 less free descriptors than this value. */
741 uint32_t txq_flags; /**< Set flags for the Tx queue */
742 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */
744 * Per-queue Tx offloads to be set using DEV_TX_OFFLOAD_* flags.
745 * Only offloads set on tx_queue_offload_capa or tx_offload_capa
746 * fields on rte_eth_dev_info structure are allowed to be set.
752 * A structure contains information about HW descriptor ring limitations.
754 struct rte_eth_desc_lim {
755 uint16_t nb_max; /**< Max allowed number of descriptors. */
756 uint16_t nb_min; /**< Min allowed number of descriptors. */
757 uint16_t nb_align; /**< Number of descriptors should be aligned to. */
760 * Max allowed number of segments per whole packet.
762 * - For TSO packet this is the total number of data descriptors allowed
765 * @see nb_mtu_seg_max
770 * Max number of segments per one MTU.
772 * - For non-TSO packet, this is the maximum allowed number of segments
773 * in a single transmit packet.
775 * - For TSO packet each segment within the TSO may span up to this
780 uint16_t nb_mtu_seg_max;
784 * This enum indicates the flow control mode
786 enum rte_eth_fc_mode {
787 RTE_FC_NONE = 0, /**< Disable flow control. */
788 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */
789 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */
790 RTE_FC_FULL /**< Enable flow control on both side. */
794 * A structure used to configure Ethernet flow control parameter.
795 * These parameters will be configured into the register of the NIC.
796 * Please refer to the corresponding data sheet for proper value.
798 struct rte_eth_fc_conf {
799 uint32_t high_water; /**< High threshold value to trigger XOFF */
800 uint32_t low_water; /**< Low threshold value to trigger XON */
801 uint16_t pause_time; /**< Pause quota in the Pause frame */
802 uint16_t send_xon; /**< Is XON frame need be sent */
803 enum rte_eth_fc_mode mode; /**< Link flow control mode */
804 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */
805 uint8_t autoneg; /**< Use Pause autoneg */
809 * A structure used to configure Ethernet priority flow control parameter.
810 * These parameters will be configured into the register of the NIC.
811 * Please refer to the corresponding data sheet for proper value.
813 struct rte_eth_pfc_conf {
814 struct rte_eth_fc_conf fc; /**< General flow control parameter. */
815 uint8_t priority; /**< VLAN User Priority. */
819 * Memory space that can be configured to store Flow Director filters
820 * in the board memory.
822 enum rte_fdir_pballoc_type {
823 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */
824 RTE_FDIR_PBALLOC_128K, /**< 128k. */
825 RTE_FDIR_PBALLOC_256K, /**< 256k. */
829 * Select report mode of FDIR hash information in RX descriptors.
831 enum rte_fdir_status_mode {
832 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
833 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
834 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
838 * A structure used to configure the Flow Director (FDIR) feature
839 * of an Ethernet port.
841 * If mode is RTE_FDIR_DISABLE, the pballoc value is ignored.
843 struct rte_fdir_conf {
844 enum rte_fdir_mode mode; /**< Flow Director mode. */
845 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
846 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
847 /** RX queue of packets matching a "drop" filter in perfect mode. */
849 struct rte_eth_fdir_masks mask;
850 struct rte_eth_fdir_flex_conf flex_conf;
851 /**< Flex payload configuration. */
855 * UDP tunneling configuration.
856 * Used to config the UDP port for a type of tunnel.
857 * NICs need the UDP port to identify the tunnel type.
858 * Normally a type of tunnel has a default UDP port, this structure can be used
859 * in case if the users want to change or support more UDP port.
861 struct rte_eth_udp_tunnel {
862 uint16_t udp_port; /**< UDP port used for the tunnel. */
863 uint8_t prot_type; /**< Tunnel type. Defined in rte_eth_tunnel_type. */
867 * A structure used to enable/disable specific device interrupts.
869 struct rte_intr_conf {
870 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
872 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */
874 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */
879 * A structure used to configure an Ethernet port.
880 * Depending upon the RX multi-queue mode, extra advanced
881 * configuration settings may be needed.
883 struct rte_eth_conf {
884 uint32_t link_speeds; /**< bitmap of ETH_LINK_SPEED_XXX of speeds to be
885 used. ETH_LINK_SPEED_FIXED disables link
886 autonegotiation, and a unique speed shall be
887 set. Otherwise, the bitmap defines the set of
888 speeds to be advertised. If the special value
889 ETH_LINK_SPEED_AUTONEG (0) is used, all speeds
890 supported are advertised. */
891 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */
892 struct rte_eth_txmode txmode; /**< Port TX configuration. */
893 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value
894 is 0, meaning the loopback mode is disabled.
895 Read the datasheet of given ethernet controller
896 for details. The possible values of this field
897 are defined in implementation of each driver. */
899 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
900 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
901 /**< Port vmdq+dcb configuration. */
902 struct rte_eth_dcb_rx_conf dcb_rx_conf;
903 /**< Port dcb RX configuration. */
904 struct rte_eth_vmdq_rx_conf vmdq_rx_conf;
905 /**< Port vmdq RX configuration. */
906 } rx_adv_conf; /**< Port RX filtering configuration. */
908 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf;
909 /**< Port vmdq+dcb TX configuration. */
910 struct rte_eth_dcb_tx_conf dcb_tx_conf;
911 /**< Port dcb TX configuration. */
912 struct rte_eth_vmdq_tx_conf vmdq_tx_conf;
913 /**< Port vmdq TX configuration. */
914 } tx_adv_conf; /**< Port TX DCB configuration (union). */
915 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC
916 is needed,and the variable must be set ETH_DCB_PFC_SUPPORT. */
917 uint32_t dcb_capability_en;
918 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. */
919 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */
923 * A structure used to retrieve the contextual information of
924 * an Ethernet device, such as the controlling driver of the device,
925 * its PCI context, etc...
929 * RX offload capabilities of a device.
931 #define DEV_RX_OFFLOAD_VLAN_STRIP 0x00000001
932 #define DEV_RX_OFFLOAD_IPV4_CKSUM 0x00000002
933 #define DEV_RX_OFFLOAD_UDP_CKSUM 0x00000004
934 #define DEV_RX_OFFLOAD_TCP_CKSUM 0x00000008
935 #define DEV_RX_OFFLOAD_TCP_LRO 0x00000010
936 #define DEV_RX_OFFLOAD_QINQ_STRIP 0x00000020
937 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000040
938 #define DEV_RX_OFFLOAD_MACSEC_STRIP 0x00000080
939 #define DEV_RX_OFFLOAD_HEADER_SPLIT 0x00000100
940 #define DEV_RX_OFFLOAD_VLAN_FILTER 0x00000200
941 #define DEV_RX_OFFLOAD_VLAN_EXTEND 0x00000400
942 #define DEV_RX_OFFLOAD_JUMBO_FRAME 0x00000800
943 #define DEV_RX_OFFLOAD_CRC_STRIP 0x00001000
944 #define DEV_RX_OFFLOAD_SCATTER 0x00002000
945 #define DEV_RX_OFFLOAD_TIMESTAMP 0x00004000
946 #define DEV_RX_OFFLOAD_SECURITY 0x00008000
947 #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \
948 DEV_RX_OFFLOAD_UDP_CKSUM | \
949 DEV_RX_OFFLOAD_TCP_CKSUM)
950 #define DEV_RX_OFFLOAD_VLAN (DEV_RX_OFFLOAD_VLAN_STRIP | \
951 DEV_RX_OFFLOAD_VLAN_FILTER | \
952 DEV_RX_OFFLOAD_VLAN_EXTEND)
955 * If new Rx offload capabilities are defined, they also must be
956 * mentioned in rte_rx_offload_names in rte_ethdev.c file.
960 * TX offload capabilities of a device.
962 #define DEV_TX_OFFLOAD_VLAN_INSERT 0x00000001
963 #define DEV_TX_OFFLOAD_IPV4_CKSUM 0x00000002
964 #define DEV_TX_OFFLOAD_UDP_CKSUM 0x00000004
965 #define DEV_TX_OFFLOAD_TCP_CKSUM 0x00000008
966 #define DEV_TX_OFFLOAD_SCTP_CKSUM 0x00000010
967 #define DEV_TX_OFFLOAD_TCP_TSO 0x00000020
968 #define DEV_TX_OFFLOAD_UDP_TSO 0x00000040
969 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000080 /**< Used for tunneling packet. */
970 #define DEV_TX_OFFLOAD_QINQ_INSERT 0x00000100
971 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO 0x00000200 /**< Used for tunneling packet. */
972 #define DEV_TX_OFFLOAD_GRE_TNL_TSO 0x00000400 /**< Used for tunneling packet. */
973 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO 0x00000800 /**< Used for tunneling packet. */
974 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO 0x00001000 /**< Used for tunneling packet. */
975 #define DEV_TX_OFFLOAD_MACSEC_INSERT 0x00002000
976 #define DEV_TX_OFFLOAD_MT_LOCKFREE 0x00004000
977 /**< Multiple threads can invoke rte_eth_tx_burst() concurrently on the same
978 * tx queue without SW lock.
980 #define DEV_TX_OFFLOAD_MULTI_SEGS 0x00008000
981 /**< Device supports multi segment send. */
982 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE 0x00010000
983 /**< Device supports optimization for fast release of mbufs.
984 * When set application must guarantee that per-queue all mbufs comes from
985 * the same mempool and has refcnt = 1.
987 #define DEV_TX_OFFLOAD_SECURITY 0x00020000
989 * Device supports generic UDP tunneled packet TSO.
990 * Application must set PKT_TX_TUNNEL_UDP and other mbuf fields required
993 #define DEV_TX_OFFLOAD_UDP_TNL_TSO 0x00040000
995 * Device supports generic IP tunneled packet TSO.
996 * Application must set PKT_TX_TUNNEL_IP and other mbuf fields required
999 #define DEV_TX_OFFLOAD_IP_TNL_TSO 0x00080000
1001 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP 0x00000001
1002 /**< Device supports Rx queue setup after device started*/
1003 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP 0x00000002
1004 /**< Device supports Tx queue setup after device started*/
1007 * If new Tx offload capabilities are defined, they also must be
1008 * mentioned in rte_tx_offload_names in rte_ethdev.c file.
1011 struct rte_pci_device;
1014 * Fallback default preferred Rx/Tx port parameters.
1015 * These are used if an application requests default parameters
1016 * but the PMD does not provide preferred values.
1018 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512
1019 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512
1020 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1
1021 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1
1024 * Preferred Rx/Tx port parameters.
1025 * There are separate instances of this structure for transmission
1026 * and reception respectively.
1028 struct rte_eth_dev_portconf {
1029 uint16_t burst_size; /**< Device-preferred burst size */
1030 uint16_t ring_size; /**< Device-preferred size of queue rings */
1031 uint16_t nb_queues; /**< Device-preferred number of queues */
1035 * Default values for switch domain id when ethdev does not support switch
1036 * domain definitions.
1038 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (0)
1041 * Ethernet device associated switch information
1043 struct rte_eth_switch_info {
1044 const char *name; /**< switch name */
1045 uint16_t domain_id; /**< switch domain id */
1048 * mapping to the devices physical switch port as enumerated from the
1049 * perspective of the embedded interconnect/switch. For SR-IOV enabled
1050 * device this may correspond to the VF_ID of each virtual function,
1051 * but each driver should explicitly define the mapping of switch
1052 * port identifier to that physical interconnect/switch
1057 * Ethernet device information
1059 struct rte_eth_dev_info {
1060 struct rte_device *device; /** Generic device information */
1061 const char *driver_name; /**< Device Driver name. */
1062 unsigned int if_index; /**< Index to bound host interface, or 0 if none.
1063 Use if_indextoname() to translate into an interface name. */
1064 const uint32_t *dev_flags; /**< Device flags */
1065 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */
1066 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */
1067 uint16_t max_rx_queues; /**< Maximum number of RX queues. */
1068 uint16_t max_tx_queues; /**< Maximum number of TX queues. */
1069 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
1070 uint32_t max_hash_mac_addrs;
1071 /** Maximum number of hash MAC addresses for MTA and UTA. */
1072 uint16_t max_vfs; /**< Maximum number of VFs. */
1073 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */
1074 uint64_t rx_offload_capa;
1075 /**< All RX offload capabilities including all per-queue ones */
1076 uint64_t tx_offload_capa;
1077 /**< All TX offload capabilities including all per-queue ones */
1078 uint64_t rx_queue_offload_capa;
1079 /**< Device per-queue RX offload capabilities. */
1080 uint64_t tx_queue_offload_capa;
1081 /**< Device per-queue TX offload capabilities. */
1083 /**< Device redirection table size, the total number of entries. */
1084 uint8_t hash_key_size; /**< Hash key size in bytes */
1085 /** Bit mask of RSS offloads, the bit offset also means flow type */
1086 uint64_t flow_type_rss_offloads;
1087 struct rte_eth_rxconf default_rxconf; /**< Default RX configuration */
1088 struct rte_eth_txconf default_txconf; /**< Default TX configuration */
1089 uint16_t vmdq_queue_base; /**< First queue ID for VMDQ pools. */
1090 uint16_t vmdq_queue_num; /**< Queue number for VMDQ pools. */
1091 uint16_t vmdq_pool_base; /**< First ID of VMDQ pools. */
1092 struct rte_eth_desc_lim rx_desc_lim; /**< RX descriptors limits */
1093 struct rte_eth_desc_lim tx_desc_lim; /**< TX descriptors limits */
1094 uint32_t speed_capa; /**< Supported speeds bitmap (ETH_LINK_SPEED_). */
1095 /** Configured number of rx/tx queues */
1096 uint16_t nb_rx_queues; /**< Number of RX queues. */
1097 uint16_t nb_tx_queues; /**< Number of TX queues. */
1098 /** Rx parameter recommendations */
1099 struct rte_eth_dev_portconf default_rxportconf;
1100 /** Tx parameter recommendations */
1101 struct rte_eth_dev_portconf default_txportconf;
1102 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */
1105 * Switching information for ports on a device with a
1106 * embedded managed interconnect/switch.
1108 struct rte_eth_switch_info switch_info;
1112 * Ethernet device RX queue information structure.
1113 * Used to retieve information about configured queue.
1115 struct rte_eth_rxq_info {
1116 struct rte_mempool *mp; /**< mempool used by that queue. */
1117 struct rte_eth_rxconf conf; /**< queue config parameters. */
1118 uint8_t scattered_rx; /**< scattered packets RX supported. */
1119 uint16_t nb_desc; /**< configured number of RXDs. */
1120 } __rte_cache_min_aligned;
1123 * Ethernet device TX queue information structure.
1124 * Used to retrieve information about configured queue.
1126 struct rte_eth_txq_info {
1127 struct rte_eth_txconf conf; /**< queue config parameters. */
1128 uint16_t nb_desc; /**< configured number of TXDs. */
1129 } __rte_cache_min_aligned;
1131 /** Maximum name length for extended statistics counters */
1132 #define RTE_ETH_XSTATS_NAME_SIZE 64
1135 * An Ethernet device extended statistic structure
1137 * This structure is used by rte_eth_xstats_get() to provide
1138 * statistics that are not provided in the generic *rte_eth_stats*
1140 * It maps a name id, corresponding to an index in the array returned
1141 * by rte_eth_xstats_get_names(), to a statistic value.
1143 struct rte_eth_xstat {
1144 uint64_t id; /**< The index in xstats name array. */
1145 uint64_t value; /**< The statistic counter value. */
1149 * A name element for extended statistics.
1151 * An array of this structure is returned by rte_eth_xstats_get_names().
1152 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat*
1153 * structure references these names by their array index.
1155 struct rte_eth_xstat_name {
1156 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */
1159 #define ETH_DCB_NUM_TCS 8
1160 #define ETH_MAX_VMDQ_POOL 64
1163 * A structure used to get the information of queue and
1164 * TC mapping on both TX and RX paths.
1166 struct rte_eth_dcb_tc_queue_mapping {
1167 /** rx queues assigned to tc per Pool */
1171 } tc_rxq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1172 /** rx queues assigned to tc per Pool */
1176 } tc_txq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS];
1180 * A structure used to get the information of DCB.
1181 * It includes TC UP mapping and queue TC mapping.
1183 struct rte_eth_dcb_info {
1184 uint8_t nb_tcs; /**< number of TCs */
1185 uint8_t prio_tc[ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */
1186 uint8_t tc_bws[ETH_DCB_NUM_TCS]; /**< TX BW percentage for each TC */
1187 /** rx queues assigned to tc */
1188 struct rte_eth_dcb_tc_queue_mapping tc_queue;
1192 * RX/TX queue states
1194 #define RTE_ETH_QUEUE_STATE_STOPPED 0
1195 #define RTE_ETH_QUEUE_STATE_STARTED 1
1197 #define RTE_ETH_ALL RTE_MAX_ETHPORTS
1199 /* Macros to check for valid port */
1200 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \
1201 if (!rte_eth_dev_is_valid_port(port_id)) { \
1202 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1207 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \
1208 if (!rte_eth_dev_is_valid_port(port_id)) { \
1209 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \
1215 * l2 tunnel configuration.
1218 /**< l2 tunnel enable mask */
1219 #define ETH_L2_TUNNEL_ENABLE_MASK 0x00000001
1220 /**< l2 tunnel insertion mask */
1221 #define ETH_L2_TUNNEL_INSERTION_MASK 0x00000002
1222 /**< l2 tunnel stripping mask */
1223 #define ETH_L2_TUNNEL_STRIPPING_MASK 0x00000004
1224 /**< l2 tunnel forwarding mask */
1225 #define ETH_L2_TUNNEL_FORWARDING_MASK 0x00000008
1228 * Function type used for RX packet processing packet callbacks.
1230 * The callback function is called on RX with a burst of packets that have
1231 * been received on the given port and queue.
1234 * The Ethernet port on which RX is being performed.
1236 * The queue on the Ethernet port which is being used to receive the packets.
1238 * The burst of packets that have just been received.
1240 * The number of packets in the burst pointed to by "pkts".
1242 * The max number of packets that can be stored in the "pkts" array.
1244 * The arbitrary user parameter passed in by the application when the callback
1245 * was originally configured.
1247 * The number of packets returned to the user.
1249 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue,
1250 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts,
1254 * Function type used for TX packet processing packet callbacks.
1256 * The callback function is called on TX with a burst of packets immediately
1257 * before the packets are put onto the hardware queue for transmission.
1260 * The Ethernet port on which TX is being performed.
1262 * The queue on the Ethernet port which is being used to transmit the packets.
1264 * The burst of packets that are about to be transmitted.
1266 * The number of packets in the burst pointed to by "pkts".
1268 * The arbitrary user parameter passed in by the application when the callback
1269 * was originally configured.
1271 * The number of packets to be written to the NIC.
1273 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue,
1274 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param);
1277 * Possible states of an ethdev port.
1279 enum rte_eth_dev_state {
1280 /** Device is unused before being probed. */
1281 RTE_ETH_DEV_UNUSED = 0,
1282 /** Device is attached when allocated in probing. */
1283 RTE_ETH_DEV_ATTACHED,
1284 /** The deferred state is useless and replaced by ownership. */
1285 RTE_ETH_DEV_DEFERRED,
1286 /** Device is in removed state when plug-out is detected. */
1287 RTE_ETH_DEV_REMOVED,
1290 struct rte_eth_dev_sriov {
1291 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */
1292 uint8_t nb_q_per_pool; /**< rx queue number per pool */
1293 uint16_t def_vmdq_idx; /**< Default pool num used for PF */
1294 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */
1296 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov)
1298 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN
1300 #define RTE_ETH_DEV_NO_OWNER 0
1302 #define RTE_ETH_MAX_OWNER_NAME_LEN 64
1304 struct rte_eth_dev_owner {
1305 uint64_t id; /**< The owner unique identifier. */
1306 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */
1309 /** Device supports link state interrupt */
1310 #define RTE_ETH_DEV_INTR_LSC 0x0002
1311 /** Device is a bonded slave */
1312 #define RTE_ETH_DEV_BONDED_SLAVE 0x0004
1313 /** Device supports device removal interrupt */
1314 #define RTE_ETH_DEV_INTR_RMV 0x0008
1315 /** Device is port representor */
1316 #define RTE_ETH_DEV_REPRESENTOR 0x0010
1319 * Iterates over valid ethdev ports owned by a specific owner.
1322 * The id of the next possible valid owned port.
1324 * The owner identifier.
1325 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports.
1327 * Next valid port id owned by owner_id, RTE_MAX_ETHPORTS if there is none.
1329 uint64_t rte_eth_find_next_owned_by(uint16_t port_id,
1330 const uint64_t owner_id);
1333 * Macro to iterate over all enabled ethdev ports owned by a specific owner.
1335 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \
1336 for (p = rte_eth_find_next_owned_by(0, o); \
1337 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
1338 p = rte_eth_find_next_owned_by(p + 1, o))
1341 * Iterates over valid ethdev ports.
1344 * The id of the next possible valid port.
1346 * Next valid port id, RTE_MAX_ETHPORTS if there is none.
1348 uint16_t rte_eth_find_next(uint16_t port_id);
1351 * Macro to iterate over all enabled and ownerless ethdev ports.
1353 #define RTE_ETH_FOREACH_DEV(p) \
1354 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
1359 * @b EXPERIMENTAL: this API may change without prior notice.
1361 * Get a new unique owner identifier.
1362 * An owner identifier is used to owns Ethernet devices by only one DPDK entity
1363 * to avoid multiple management of device by different entities.
1366 * Owner identifier pointer.
1368 * Negative errno value on error, 0 on success.
1370 int __rte_experimental rte_eth_dev_owner_new(uint64_t *owner_id);
1374 * @b EXPERIMENTAL: this API may change without prior notice.
1376 * Set an Ethernet device owner.
1379 * The identifier of the port to own.
1381 * The owner pointer.
1383 * Negative errno value on error, 0 on success.
1385 int __rte_experimental rte_eth_dev_owner_set(const uint16_t port_id,
1386 const struct rte_eth_dev_owner *owner);
1390 * @b EXPERIMENTAL: this API may change without prior notice.
1392 * Unset Ethernet device owner to make the device ownerless.
1395 * The identifier of port to make ownerless.
1397 * The owner identifier.
1399 * 0 on success, negative errno value on error.
1401 int __rte_experimental rte_eth_dev_owner_unset(const uint16_t port_id,
1402 const uint64_t owner_id);
1406 * @b EXPERIMENTAL: this API may change without prior notice.
1408 * Remove owner from all Ethernet devices owned by a specific owner.
1411 * The owner identifier.
1413 void __rte_experimental rte_eth_dev_owner_delete(const uint64_t owner_id);
1417 * @b EXPERIMENTAL: this API may change without prior notice.
1419 * Get the owner of an Ethernet device.
1422 * The port identifier.
1424 * The owner structure pointer to fill.
1426 * 0 on success, negative errno value on error..
1428 int __rte_experimental rte_eth_dev_owner_get(const uint16_t port_id,
1429 struct rte_eth_dev_owner *owner);
1432 * Get the total number of Ethernet devices that have been successfully
1433 * initialized by the matching Ethernet driver during the PCI probing phase
1434 * and that are available for applications to use. These devices must be
1435 * accessed by using the ``RTE_ETH_FOREACH_DEV()`` macro to deal with
1436 * non-contiguous ranges of devices.
1437 * These non-contiguous ranges can be created by calls to hotplug functions or
1441 * - The total number of usable Ethernet devices.
1444 uint16_t rte_eth_dev_count(void);
1447 * Get the number of ports which are usable for the application.
1449 * These devices must be iterated by using the macro
1450 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY``
1451 * to deal with non-contiguous ranges of devices.
1454 * The count of available Ethernet devices.
1456 uint16_t rte_eth_dev_count_avail(void);
1459 * Get the total number of ports which are allocated.
1461 * Some devices may not be available for the application.
1464 * The total count of Ethernet devices.
1466 uint16_t __rte_experimental rte_eth_dev_count_total(void);
1469 * Attach a new Ethernet device specified by arguments.
1472 * A pointer to a strings array describing the new device
1473 * to be attached. The strings should be a pci address like
1474 * '0000:01:00.0' or virtual device name like 'net_pcap0'.
1476 * A pointer to a port identifier actually attached.
1478 * 0 on success and port_id is filled, negative on error
1480 int rte_eth_dev_attach(const char *devargs, uint16_t *port_id);
1483 * Detach a Ethernet device specified by port identifier.
1484 * This function must be called when the device is in the
1488 * The port identifier of the device to detach.
1490 * A pointer to a buffer that will be filled with the device name.
1491 * This buffer must be at least RTE_DEV_NAME_MAX_LEN long.
1493 * 0 on success and devname is filled, negative on error
1495 int rte_eth_dev_detach(uint16_t port_id, char *devname);
1498 * Convert a numerical speed in Mbps to a bitmap flag that can be used in
1499 * the bitmap link_speeds of the struct rte_eth_conf
1502 * Numerical speed value in Mbps
1504 * ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds)
1506 * 0 if the speed cannot be mapped
1508 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex);
1512 * @b EXPERIMENTAL: this API may change without prior notice
1514 * Get DEV_RX_OFFLOAD_* flag name.
1519 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1521 const char * __rte_experimental rte_eth_dev_rx_offload_name(uint64_t offload);
1525 * @b EXPERIMENTAL: this API may change without prior notice
1527 * Get DEV_TX_OFFLOAD_* flag name.
1532 * Offload name or 'UNKNOWN' if the flag cannot be recognised.
1534 const char * __rte_experimental rte_eth_dev_tx_offload_name(uint64_t offload);
1537 * Configure an Ethernet device.
1538 * This function must be invoked first before any other function in the
1539 * Ethernet API. This function can also be re-invoked when a device is in the
1543 * The port identifier of the Ethernet device to configure.
1544 * @param nb_rx_queue
1545 * The number of receive queues to set up for the Ethernet device.
1546 * @param nb_tx_queue
1547 * The number of transmit queues to set up for the Ethernet device.
1549 * The pointer to the configuration data to be used for the Ethernet device.
1550 * The *rte_eth_conf* structure includes:
1551 * - the hardware offload features to activate, with dedicated fields for
1552 * each statically configurable offload hardware feature provided by
1553 * Ethernet devices, such as IP checksum or VLAN tag stripping for
1555 * The Rx offload bitfield API is obsolete and will be deprecated.
1556 * Applications should set the ignore_bitfield_offloads bit on *rxmode*
1557 * structure and use offloads field to set per-port offloads instead.
1558 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within
1559 * the [rt]x_offload_capa returned from rte_eth_dev_infos_get().
1560 * Any type of device supported offloading set in the input argument
1561 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled
1562 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup().
1563 * - the Receive Side Scaling (RSS) configuration when using multiple RX
1566 * Embedding all configuration information in a single data structure
1567 * is the more flexible method that allows the addition of new features
1568 * without changing the syntax of the API.
1570 * - 0: Success, device configured.
1571 * - <0: Error code returned by the driver configuration function.
1573 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue,
1574 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf);
1578 * @b EXPERIMENTAL: this API may change without prior notice.
1580 * Check if an Ethernet device was physically removed.
1583 * The port identifier of the Ethernet device.
1585 * 1 when the Ethernet device is removed, otherwise 0.
1587 int __rte_experimental
1588 rte_eth_dev_is_removed(uint16_t port_id);
1591 * Allocate and set up a receive queue for an Ethernet device.
1593 * The function allocates a contiguous block of memory for *nb_rx_desc*
1594 * receive descriptors from a memory zone associated with *socket_id*
1595 * and initializes each receive descriptor with a network buffer allocated
1596 * from the memory pool *mb_pool*.
1599 * The port identifier of the Ethernet device.
1600 * @param rx_queue_id
1601 * The index of the receive queue to set up.
1602 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1603 * to rte_eth_dev_configure().
1605 * The number of receive descriptors to allocate for the receive ring.
1607 * The *socket_id* argument is the socket identifier in case of NUMA.
1608 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1609 * the DMA memory allocated for the receive descriptors of the ring.
1611 * The pointer to the configuration data to be used for the receive queue.
1612 * NULL value is allowed, in which case default RX configuration
1614 * The *rx_conf* structure contains an *rx_thresh* structure with the values
1615 * of the Prefetch, Host, and Write-Back threshold registers of the receive
1617 * In addition it contains the hardware offloads features to activate using
1618 * the DEV_RX_OFFLOAD_* flags.
1619 * If an offloading set in rx_conf->offloads
1620 * hasn't been set in the input argument eth_conf->rxmode.offloads
1621 * to rte_eth_dev_configure(), it is a new added offloading, it must be
1622 * per-queue type and it is enabled for the queue.
1623 * No need to repeat any bit in rx_conf->offloads which has already been
1624 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
1625 * at port level can't be disabled at queue level.
1627 * The pointer to the memory pool from which to allocate *rte_mbuf* network
1628 * memory buffers to populate each descriptor of the receive ring.
1630 * - 0: Success, receive queue correctly set up.
1631 * - -EIO: if device is removed.
1632 * - -EINVAL: The size of network buffers which can be allocated from the
1633 * memory pool does not fit the various buffer sizes allowed by the
1634 * device controller.
1635 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
1636 * allocate network memory buffers from the memory pool when
1637 * initializing receive descriptors.
1639 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id,
1640 uint16_t nb_rx_desc, unsigned int socket_id,
1641 const struct rte_eth_rxconf *rx_conf,
1642 struct rte_mempool *mb_pool);
1645 * Allocate and set up a transmit queue for an Ethernet device.
1648 * The port identifier of the Ethernet device.
1649 * @param tx_queue_id
1650 * The index of the transmit queue to set up.
1651 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1652 * to rte_eth_dev_configure().
1654 * The number of transmit descriptors to allocate for the transmit ring.
1656 * The *socket_id* argument is the socket identifier in case of NUMA.
1657 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1658 * the DMA memory allocated for the transmit descriptors of the ring.
1660 * The pointer to the configuration data to be used for the transmit queue.
1661 * NULL value is allowed, in which case default TX configuration
1663 * The *tx_conf* structure contains the following data:
1664 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
1665 * Write-Back threshold registers of the transmit ring.
1666 * When setting Write-Back threshold to the value greater then zero,
1667 * *tx_rs_thresh* value should be explicitly set to one.
1668 * - The *tx_free_thresh* value indicates the [minimum] number of network
1669 * buffers that must be pending in the transmit ring to trigger their
1670 * [implicit] freeing by the driver transmit function.
1671 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
1672 * descriptors that must be pending in the transmit ring before setting the
1673 * RS bit on a descriptor by the driver transmit function.
1674 * The *tx_rs_thresh* value should be less or equal then
1675 * *tx_free_thresh* value, and both of them should be less then
1677 * - The *txq_flags* member contains flags to pass to the TX queue setup
1678 * function to configure the behavior of the TX queue. This should be set
1679 * to 0 if no special configuration is required.
1680 * This API is obsolete and will be deprecated. Applications
1681 * should set it to ETH_TXQ_FLAGS_IGNORE and use
1682 * the offloads field below.
1683 * - The *offloads* member contains Tx offloads to be enabled.
1684 * If an offloading set in tx_conf->offloads
1685 * hasn't been set in the input argument eth_conf->txmode.offloads
1686 * to rte_eth_dev_configure(), it is a new added offloading, it must be
1687 * per-queue type and it is enabled for the queue.
1688 * No need to repeat any bit in tx_conf->offloads which has already been
1689 * enabled in rte_eth_dev_configure() at port level. An offloading enabled
1690 * at port level can't be disabled at queue level.
1692 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
1693 * the transmit function to use default values.
1695 * - 0: Success, the transmit queue is correctly set up.
1696 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
1698 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id,
1699 uint16_t nb_tx_desc, unsigned int socket_id,
1700 const struct rte_eth_txconf *tx_conf);
1703 * Return the NUMA socket to which an Ethernet device is connected
1706 * The port identifier of the Ethernet device
1708 * The NUMA socket id to which the Ethernet device is connected or
1709 * a default of zero if the socket could not be determined.
1710 * -1 is returned is the port_id value is out of range.
1712 int rte_eth_dev_socket_id(uint16_t port_id);
1715 * Check if port_id of device is attached
1718 * The port identifier of the Ethernet device
1720 * - 0 if port is out of range or not attached
1721 * - 1 if device is attached
1723 int rte_eth_dev_is_valid_port(uint16_t port_id);
1726 * Start specified RX queue of a port. It is used when rx_deferred_start
1727 * flag of the specified queue is true.
1730 * The port identifier of the Ethernet device
1731 * @param rx_queue_id
1732 * The index of the rx queue to update the ring.
1733 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1734 * to rte_eth_dev_configure().
1736 * - 0: Success, the receive queue is started.
1737 * - -EINVAL: The port_id or the queue_id out of range.
1738 * - -EIO: if device is removed.
1739 * - -ENOTSUP: The function not supported in PMD driver.
1741 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id);
1744 * Stop specified RX queue of a port
1747 * The port identifier of the Ethernet device
1748 * @param rx_queue_id
1749 * The index of the rx queue to update the ring.
1750 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1751 * to rte_eth_dev_configure().
1753 * - 0: Success, the receive queue is stopped.
1754 * - -EINVAL: The port_id or the queue_id out of range.
1755 * - -EIO: if device is removed.
1756 * - -ENOTSUP: The function not supported in PMD driver.
1758 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id);
1761 * Start TX for specified queue of a port. It is used when tx_deferred_start
1762 * flag of the specified queue is true.
1765 * The port identifier of the Ethernet device
1766 * @param tx_queue_id
1767 * The index of the tx queue to update the ring.
1768 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1769 * to rte_eth_dev_configure().
1771 * - 0: Success, the transmit queue is started.
1772 * - -EINVAL: The port_id or the queue_id out of range.
1773 * - -EIO: if device is removed.
1774 * - -ENOTSUP: The function not supported in PMD driver.
1776 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id);
1779 * Stop specified TX queue of a port
1782 * The port identifier of the Ethernet device
1783 * @param tx_queue_id
1784 * The index of the tx queue to update the ring.
1785 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1786 * to rte_eth_dev_configure().
1788 * - 0: Success, the transmit queue is stopped.
1789 * - -EINVAL: The port_id or the queue_id out of range.
1790 * - -EIO: if device is removed.
1791 * - -ENOTSUP: The function not supported in PMD driver.
1793 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id);
1796 * Start an Ethernet device.
1798 * The device start step is the last one and consists of setting the configured
1799 * offload features and in starting the transmit and the receive units of the
1801 * On success, all basic functions exported by the Ethernet API (link status,
1802 * receive/transmit, and so on) can be invoked.
1805 * The port identifier of the Ethernet device.
1807 * - 0: Success, Ethernet device started.
1808 * - <0: Error code of the driver device start function.
1810 int rte_eth_dev_start(uint16_t port_id);
1813 * Stop an Ethernet device. The device can be restarted with a call to
1814 * rte_eth_dev_start()
1817 * The port identifier of the Ethernet device.
1819 void rte_eth_dev_stop(uint16_t port_id);
1822 * Link up an Ethernet device.
1824 * Set device link up will re-enable the device rx/tx
1825 * functionality after it is previously set device linked down.
1828 * The port identifier of the Ethernet device.
1830 * - 0: Success, Ethernet device linked up.
1831 * - <0: Error code of the driver device link up function.
1833 int rte_eth_dev_set_link_up(uint16_t port_id);
1836 * Link down an Ethernet device.
1837 * The device rx/tx functionality will be disabled if success,
1838 * and it can be re-enabled with a call to
1839 * rte_eth_dev_set_link_up()
1842 * The port identifier of the Ethernet device.
1844 int rte_eth_dev_set_link_down(uint16_t port_id);
1847 * Close a stopped Ethernet device. The device cannot be restarted!
1848 * The function frees all resources except for needed by the
1849 * closed state. To free these resources, call rte_eth_dev_detach().
1852 * The port identifier of the Ethernet device.
1854 void rte_eth_dev_close(uint16_t port_id);
1857 * Reset a Ethernet device and keep its port id.
1859 * When a port has to be reset passively, the DPDK application can invoke
1860 * this function. For example when a PF is reset, all its VFs should also
1861 * be reset. Normally a DPDK application can invoke this function when
1862 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start
1863 * a port reset in other circumstances.
1865 * When this function is called, it first stops the port and then calls the
1866 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial
1867 * state, in which no Tx and Rx queues are setup, as if the port has been
1868 * reset and not started. The port keeps the port id it had before the
1871 * After calling rte_eth_dev_reset( ), the application should use
1872 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ),
1873 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( )
1874 * to reconfigure the device as appropriate.
1876 * Note: To avoid unexpected behavior, the application should stop calling
1877 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread
1878 * safety, all these controlling functions should be called from the same
1882 * The port identifier of the Ethernet device.
1885 * - (0) if successful.
1886 * - (-EINVAL) if port identifier is invalid.
1887 * - (-ENOTSUP) if hardware doesn't support this function.
1888 * - (-EPERM) if not ran from the primary process.
1889 * - (-EIO) if re-initialisation failed or device is removed.
1890 * - (-ENOMEM) if the reset failed due to OOM.
1891 * - (-EAGAIN) if the reset temporarily failed and should be retried later.
1893 int rte_eth_dev_reset(uint16_t port_id);
1896 * Enable receipt in promiscuous mode for an Ethernet device.
1899 * The port identifier of the Ethernet device.
1901 void rte_eth_promiscuous_enable(uint16_t port_id);
1904 * Disable receipt in promiscuous mode for an Ethernet device.
1907 * The port identifier of the Ethernet device.
1909 void rte_eth_promiscuous_disable(uint16_t port_id);
1912 * Return the value of promiscuous mode for an Ethernet device.
1915 * The port identifier of the Ethernet device.
1917 * - (1) if promiscuous is enabled
1918 * - (0) if promiscuous is disabled.
1921 int rte_eth_promiscuous_get(uint16_t port_id);
1924 * Enable the receipt of any multicast frame by an Ethernet device.
1927 * The port identifier of the Ethernet device.
1929 void rte_eth_allmulticast_enable(uint16_t port_id);
1932 * Disable the receipt of all multicast frames by an Ethernet device.
1935 * The port identifier of the Ethernet device.
1937 void rte_eth_allmulticast_disable(uint16_t port_id);
1940 * Return the value of allmulticast mode for an Ethernet device.
1943 * The port identifier of the Ethernet device.
1945 * - (1) if allmulticast is enabled
1946 * - (0) if allmulticast is disabled.
1949 int rte_eth_allmulticast_get(uint16_t port_id);
1952 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
1953 * or FULL-DUPLEX) of the physical link of an Ethernet device. It might need
1954 * to wait up to 9 seconds in it.
1957 * The port identifier of the Ethernet device.
1959 * A pointer to an *rte_eth_link* structure to be filled with
1960 * the status, the speed and the mode of the Ethernet device link.
1962 void rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link);
1965 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
1966 * or FULL-DUPLEX) of the physical link of an Ethernet device. It is a no-wait
1967 * version of rte_eth_link_get().
1970 * The port identifier of the Ethernet device.
1972 * A pointer to an *rte_eth_link* structure to be filled with
1973 * the status, the speed and the mode of the Ethernet device link.
1975 void rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link);
1978 * Retrieve the general I/O statistics of an Ethernet device.
1981 * The port identifier of the Ethernet device.
1983 * A pointer to a structure of type *rte_eth_stats* to be filled with
1984 * the values of device counters for the following set of statistics:
1985 * - *ipackets* with the total of successfully received packets.
1986 * - *opackets* with the total of successfully transmitted packets.
1987 * - *ibytes* with the total of successfully received bytes.
1988 * - *obytes* with the total of successfully transmitted bytes.
1989 * - *ierrors* with the total of erroneous received packets.
1990 * - *oerrors* with the total of failed transmitted packets.
1992 * Zero if successful. Non-zero otherwise.
1994 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats);
1997 * Reset the general I/O statistics of an Ethernet device.
2000 * The port identifier of the Ethernet device.
2002 * - (0) if device notified to reset stats.
2003 * - (-ENOTSUP) if hardware doesn't support.
2004 * - (-ENODEV) if *port_id* invalid.
2006 int rte_eth_stats_reset(uint16_t port_id);
2009 * Retrieve names of extended statistics of an Ethernet device.
2012 * The port identifier of the Ethernet device.
2013 * @param xstats_names
2014 * An rte_eth_xstat_name array of at least *size* elements to
2015 * be filled. If set to NULL, the function returns the required number
2018 * The size of the xstats_names array (number of elements).
2020 * - A positive value lower or equal to size: success. The return value
2021 * is the number of entries filled in the stats table.
2022 * - A positive value higher than size: error, the given statistics table
2023 * is too small. The return value corresponds to the size that should
2024 * be given to succeed. The entries in the table are not valid and
2025 * shall not be used by the caller.
2026 * - A negative value on error (invalid port id).
2028 int rte_eth_xstats_get_names(uint16_t port_id,
2029 struct rte_eth_xstat_name *xstats_names,
2033 * Retrieve extended statistics of an Ethernet device.
2036 * The port identifier of the Ethernet device.
2038 * A pointer to a table of structure of type *rte_eth_xstat*
2039 * to be filled with device statistics ids and values: id is the
2040 * index of the name string in xstats_names (see rte_eth_xstats_get_names()),
2041 * and value is the statistic counter.
2042 * This parameter can be set to NULL if n is 0.
2044 * The size of the xstats array (number of elements).
2046 * - A positive value lower or equal to n: success. The return value
2047 * is the number of entries filled in the stats table.
2048 * - A positive value higher than n: error, the given statistics table
2049 * is too small. The return value corresponds to the size that should
2050 * be given to succeed. The entries in the table are not valid and
2051 * shall not be used by the caller.
2052 * - A negative value on error (invalid port id).
2054 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats,
2058 * Retrieve names of extended statistics of an Ethernet device.
2061 * The port identifier of the Ethernet device.
2062 * @param xstats_names
2063 * An rte_eth_xstat_name array of at least *size* elements to
2064 * be filled. If set to NULL, the function returns the required number
2067 * IDs array given by app to retrieve specific statistics
2069 * The size of the xstats_names array (number of elements).
2071 * - A positive value lower or equal to size: success. The return value
2072 * is the number of entries filled in the stats table.
2073 * - A positive value higher than size: error, the given statistics table
2074 * is too small. The return value corresponds to the size that should
2075 * be given to succeed. The entries in the table are not valid and
2076 * shall not be used by the caller.
2077 * - A negative value on error (invalid port id).
2080 rte_eth_xstats_get_names_by_id(uint16_t port_id,
2081 struct rte_eth_xstat_name *xstats_names, unsigned int size,
2085 * Retrieve extended statistics of an Ethernet device.
2088 * The port identifier of the Ethernet device.
2090 * A pointer to an ids array passed by application. This tells which
2091 * statistics values function should retrieve. This parameter
2092 * can be set to NULL if size is 0. In this case function will retrieve
2093 * all avalible statistics.
2095 * A pointer to a table to be filled with device statistics values.
2097 * The size of the ids array (number of elements).
2099 * - A positive value lower or equal to size: success. The return value
2100 * is the number of entries filled in the stats table.
2101 * - A positive value higher than size: error, the given statistics table
2102 * is too small. The return value corresponds to the size that should
2103 * be given to succeed. The entries in the table are not valid and
2104 * shall not be used by the caller.
2105 * - A negative value on error (invalid port id).
2107 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids,
2108 uint64_t *values, unsigned int size);
2111 * Gets the ID of a statistic from its name.
2113 * This function searches for the statistics using string compares, and
2114 * as such should not be used on the fast-path. For fast-path retrieval of
2115 * specific statistics, store the ID as provided in *id* from this function,
2116 * and pass the ID to rte_eth_xstats_get()
2118 * @param port_id The port to look up statistics from
2119 * @param xstat_name The name of the statistic to return
2120 * @param[out] id A pointer to an app-supplied uint64_t which should be
2121 * set to the ID of the stat if the stat exists.
2124 * -ENODEV for invalid port_id,
2125 * -EIO if device is removed,
2126 * -EINVAL if the xstat_name doesn't exist in port_id
2128 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name,
2132 * Reset extended statistics of an Ethernet device.
2135 * The port identifier of the Ethernet device.
2137 void rte_eth_xstats_reset(uint16_t port_id);
2140 * Set a mapping for the specified transmit queue to the specified per-queue
2141 * statistics counter.
2144 * The port identifier of the Ethernet device.
2145 * @param tx_queue_id
2146 * The index of the transmit queue for which a queue stats mapping is required.
2147 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2148 * to rte_eth_dev_configure().
2150 * The per-queue packet statistics functionality number that the transmit
2151 * queue is to be assigned.
2152 * The value must be in the range [0, RTE_MAX_ETHPORT_QUEUE_STATS_MAPS - 1].
2154 * Zero if successful. Non-zero otherwise.
2156 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id,
2157 uint16_t tx_queue_id, uint8_t stat_idx);
2160 * Set a mapping for the specified receive queue to the specified per-queue
2161 * statistics counter.
2164 * The port identifier of the Ethernet device.
2165 * @param rx_queue_id
2166 * The index of the receive queue for which a queue stats mapping is required.
2167 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2168 * to rte_eth_dev_configure().
2170 * The per-queue packet statistics functionality number that the receive
2171 * queue is to be assigned.
2172 * The value must be in the range [0, RTE_MAX_ETHPORT_QUEUE_STATS_MAPS - 1].
2174 * Zero if successful. Non-zero otherwise.
2176 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id,
2177 uint16_t rx_queue_id,
2181 * Retrieve the Ethernet address of an Ethernet device.
2184 * The port identifier of the Ethernet device.
2186 * A pointer to a structure of type *ether_addr* to be filled with
2187 * the Ethernet address of the Ethernet device.
2189 void rte_eth_macaddr_get(uint16_t port_id, struct ether_addr *mac_addr);
2192 * Retrieve the contextual information of an Ethernet device.
2195 * The port identifier of the Ethernet device.
2197 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
2198 * the contextual information of the Ethernet device.
2200 void rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info);
2203 * Retrieve the firmware version of a device.
2206 * The port identifier of the device.
2208 * A pointer to a string array storing the firmware version of a device,
2209 * the string includes terminating null. This pointer is allocated by caller.
2211 * The size of the string array pointed by fw_version, which should be
2212 * large enough to store firmware version of the device.
2214 * - (0) if successful.
2215 * - (-ENOTSUP) if operation is not supported.
2216 * - (-ENODEV) if *port_id* invalid.
2217 * - (-EIO) if device is removed.
2218 * - (>0) if *fw_size* is not enough to store firmware version, return
2219 * the size of the non truncated string.
2221 int rte_eth_dev_fw_version_get(uint16_t port_id,
2222 char *fw_version, size_t fw_size);
2225 * Retrieve the supported packet types of an Ethernet device.
2227 * When a packet type is announced as supported, it *must* be recognized by
2228 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN
2229 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following
2230 * packet types for these packets:
2231 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4
2232 * - Ether/Vlan/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4
2233 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER
2234 * - Ether/Vlan/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN
2236 * When a packet is received by a PMD, the most precise type must be
2237 * returned among the ones supported. However a PMD is allowed to set
2238 * packet type that is not in the supported list, at the condition that it
2239 * is more precise. Therefore, a PMD announcing no supported packet types
2240 * can still set a matching packet type in a received packet.
2243 * Better to invoke this API after the device is already started or rx burst
2244 * function is decided, to obtain correct supported ptypes.
2246 * if a given PMD does not report what ptypes it supports, then the supported
2247 * ptype count is reported as 0.
2249 * The port identifier of the Ethernet device.
2251 * A hint of what kind of packet type which the caller is interested in.
2253 * An array pointer to store adequate packet types, allocated by caller.
2255 * Size of the array pointed by param ptypes.
2257 * - (>=0) Number of supported ptypes. If the number of types exceeds num,
2258 * only num entries will be filled into the ptypes array, but the full
2259 * count of supported ptypes will be returned.
2260 * - (-ENODEV) if *port_id* invalid.
2262 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask,
2263 uint32_t *ptypes, int num);
2266 * Retrieve the MTU of an Ethernet device.
2269 * The port identifier of the Ethernet device.
2271 * A pointer to a uint16_t where the retrieved MTU is to be stored.
2273 * - (0) if successful.
2274 * - (-ENODEV) if *port_id* invalid.
2276 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu);
2279 * Change the MTU of an Ethernet device.
2282 * The port identifier of the Ethernet device.
2284 * A uint16_t for the MTU to be applied.
2286 * - (0) if successful.
2287 * - (-ENOTSUP) if operation is not supported.
2288 * - (-ENODEV) if *port_id* invalid.
2289 * - (-EIO) if device is removed.
2290 * - (-EINVAL) if *mtu* invalid.
2291 * - (-EBUSY) if operation is not allowed when the port is running
2293 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu);
2296 * Enable/Disable hardware filtering by an Ethernet device of received
2297 * VLAN packets tagged with a given VLAN Tag Identifier.
2300 * The port identifier of the Ethernet device.
2302 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
2304 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
2305 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
2307 * - (0) if successful.
2308 * - (-ENOSUP) if hardware-assisted VLAN filtering not configured.
2309 * - (-ENODEV) if *port_id* invalid.
2310 * - (-EIO) if device is removed.
2311 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
2312 * - (-EINVAL) if *vlan_id* > 4095.
2314 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on);
2317 * Enable/Disable hardware VLAN Strip by a rx queue of an Ethernet device.
2318 * 82599/X540/X550 can support VLAN stripping at the rx queue level
2321 * The port identifier of the Ethernet device.
2322 * @param rx_queue_id
2323 * The index of the receive queue for which a queue stats mapping is required.
2324 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2325 * to rte_eth_dev_configure().
2327 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port.
2328 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port.
2330 * - (0) if successful.
2331 * - (-ENOSUP) if hardware-assisted VLAN stripping not configured.
2332 * - (-ENODEV) if *port_id* invalid.
2333 * - (-EINVAL) if *rx_queue_id* invalid.
2335 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id,
2339 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to
2340 * the VLAN Header. This is a register setup available on some Intel NIC, not
2341 * but all, please check the data sheet for availability.
2344 * The port identifier of the Ethernet device.
2348 * The Tag Protocol ID
2350 * - (0) if successful.
2351 * - (-ENOSUP) if hardware-assisted VLAN TPID setup is not supported.
2352 * - (-ENODEV) if *port_id* invalid.
2353 * - (-EIO) if device is removed.
2355 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id,
2356 enum rte_vlan_type vlan_type,
2360 * Set VLAN offload configuration on an Ethernet device
2361 * Enable/Disable Extended VLAN by an Ethernet device, This is a register setup
2362 * available on some Intel NIC, not but all, please check the data sheet for
2364 * Enable/Disable VLAN Strip can be done on rx queue for certain NIC, but here
2365 * the configuration is applied on the port level.
2368 * The port identifier of the Ethernet device.
2369 * @param offload_mask
2370 * The VLAN Offload bit mask can be mixed use with "OR"
2371 * ETH_VLAN_STRIP_OFFLOAD
2372 * ETH_VLAN_FILTER_OFFLOAD
2373 * ETH_VLAN_EXTEND_OFFLOAD
2375 * - (0) if successful.
2376 * - (-ENOSUP) if hardware-assisted VLAN filtering not configured.
2377 * - (-ENODEV) if *port_id* invalid.
2378 * - (-EIO) if device is removed.
2380 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask);
2383 * Read VLAN Offload configuration from an Ethernet device
2386 * The port identifier of the Ethernet device.
2388 * - (>0) if successful. Bit mask to indicate
2389 * ETH_VLAN_STRIP_OFFLOAD
2390 * ETH_VLAN_FILTER_OFFLOAD
2391 * ETH_VLAN_EXTEND_OFFLOAD
2392 * - (-ENODEV) if *port_id* invalid.
2394 int rte_eth_dev_get_vlan_offload(uint16_t port_id);
2397 * Set port based TX VLAN insertion on or off.
2400 * The port identifier of the Ethernet device.
2402 * Port based TX VLAN identifier together with user priority.
2404 * Turn on or off the port based TX VLAN insertion.
2407 * - (0) if successful.
2408 * - negative if failed.
2410 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on);
2412 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count,
2416 * Structure used to buffer packets for future TX
2417 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush
2419 struct rte_eth_dev_tx_buffer {
2420 buffer_tx_error_fn error_callback;
2421 void *error_userdata;
2422 uint16_t size; /**< Size of buffer for buffered tx */
2423 uint16_t length; /**< Number of packets in the array */
2424 struct rte_mbuf *pkts[];
2425 /**< Pending packets to be sent on explicit flush or when full */
2429 * Calculate the size of the tx buffer.
2432 * Number of stored packets.
2434 #define RTE_ETH_TX_BUFFER_SIZE(sz) \
2435 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *))
2438 * Initialize default values for buffered transmitting
2441 * Tx buffer to be initialized.
2448 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size);
2451 * Configure a callback for buffered packets which cannot be sent
2453 * Register a specific callback to be called when an attempt is made to send
2454 * all packets buffered on an ethernet port, but not all packets can
2455 * successfully be sent. The callback registered here will be called only
2456 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs.
2457 * The default callback configured for each queue by default just frees the
2458 * packets back to the calling mempool. If additional behaviour is required,
2459 * for example, to count dropped packets, or to retry transmission of packets
2460 * which cannot be sent, this function should be used to register a suitable
2461 * callback function to implement the desired behaviour.
2462 * The example callback "rte_eth_count_unsent_packet_callback()" is also
2463 * provided as reference.
2466 * The port identifier of the Ethernet device.
2468 * The function to be used as the callback.
2470 * Arbitrary parameter to be passed to the callback function
2472 * 0 on success, or -1 on error with rte_errno set appropriately
2475 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer,
2476 buffer_tx_error_fn callback, void *userdata);
2479 * Callback function for silently dropping unsent buffered packets.
2481 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
2482 * adjust the default behavior when buffered packets cannot be sent. This
2483 * function drops any unsent packets silently and is used by tx buffered
2484 * operations as default behavior.
2486 * NOTE: this function should not be called directly, instead it should be used
2487 * as a callback for packet buffering.
2489 * NOTE: when configuring this function as a callback with
2490 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
2491 * should point to an uint64_t value.
2494 * The previously buffered packets which could not be sent
2496 * The number of unsent packets in the pkts array
2501 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent,
2505 * Callback function for tracking unsent buffered packets.
2507 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to
2508 * adjust the default behavior when buffered packets cannot be sent. This
2509 * function drops any unsent packets, but also updates a user-supplied counter
2510 * to track the overall number of packets dropped. The counter should be an
2511 * uint64_t variable.
2513 * NOTE: this function should not be called directly, instead it should be used
2514 * as a callback for packet buffering.
2516 * NOTE: when configuring this function as a callback with
2517 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter
2518 * should point to an uint64_t value.
2521 * The previously buffered packets which could not be sent
2523 * The number of unsent packets in the pkts array
2525 * Pointer to an uint64_t value, which will be incremented by unsent
2528 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent,
2532 * Request the driver to free mbufs currently cached by the driver. The
2533 * driver will only free the mbuf if it is no longer in use. It is the
2534 * application's responsibity to ensure rte_eth_tx_buffer_flush(..) is
2538 * The port identifier of the Ethernet device.
2540 * The index of the transmit queue through which output packets must be
2542 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
2543 * to rte_eth_dev_configure().
2545 * Maximum number of packets to free. Use 0 to indicate all possible packets
2546 * should be freed. Note that a packet may be using multiple mbufs.
2549 * -ENODEV: Invalid interface
2550 * -EIO: device is removed
2551 * -ENOTSUP: Driver does not support function
2553 * 0-n: Number of packets freed. More packets may still remain in ring that
2557 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt);
2560 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by
2563 enum rte_eth_event_ipsec_subtype {
2564 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0,
2565 /**< Unknown event type */
2566 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW,
2567 /**< Sequence number overflow */
2568 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY,
2569 /**< Soft time expiry of SA */
2570 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY,
2571 /**< Soft byte expiry of SA */
2572 RTE_ETH_EVENT_IPSEC_MAX
2573 /**< Max value of this enum */
2577 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra
2578 * information of the IPsec offload event.
2580 struct rte_eth_event_ipsec_desc {
2581 enum rte_eth_event_ipsec_subtype subtype;
2582 /**< Type of RTE_ETH_EVENT_IPSEC_* event */
2584 /**< Event specific metadata
2586 * For the following events, *userdata* registered
2587 * with the *rte_security_session* would be returned
2590 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW
2591 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY
2592 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY
2594 * @see struct rte_security_session_conf
2600 * The eth device event type for interrupt, and maybe others in the future.
2602 enum rte_eth_event_type {
2603 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
2604 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
2605 RTE_ETH_EVENT_QUEUE_STATE,
2606 /**< queue state event (enabled/disabled) */
2607 RTE_ETH_EVENT_INTR_RESET,
2608 /**< reset interrupt event, sent to VF on PF reset */
2609 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */
2610 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */
2611 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */
2612 RTE_ETH_EVENT_NEW, /**< port is probed */
2613 RTE_ETH_EVENT_DESTROY, /**< port is released */
2614 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */
2615 RTE_ETH_EVENT_MAX /**< max value of this enum */
2618 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id,
2619 enum rte_eth_event_type event, void *cb_arg, void *ret_param);
2620 /**< user application callback to be registered for interrupts */
2623 * Register a callback function for port event.
2627 * RTE_ETH_ALL means register the event for all port ids.
2631 * User supplied callback function to be called.
2633 * Pointer to the parameters for the registered callback.
2636 * - On success, zero.
2637 * - On failure, a negative value.
2639 int rte_eth_dev_callback_register(uint16_t port_id,
2640 enum rte_eth_event_type event,
2641 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
2644 * Unregister a callback function for port event.
2648 * RTE_ETH_ALL means unregister the event for all port ids.
2652 * User supplied callback function to be called.
2654 * Pointer to the parameters for the registered callback. -1 means to
2655 * remove all for the same callback address and same event.
2658 * - On success, zero.
2659 * - On failure, a negative value.
2661 int rte_eth_dev_callback_unregister(uint16_t port_id,
2662 enum rte_eth_event_type event,
2663 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
2666 * When there is no rx packet coming in Rx Queue for a long time, we can
2667 * sleep lcore related to RX Queue for power saving, and enable rx interrupt
2668 * to be triggered when Rx packet arrives.
2670 * The rte_eth_dev_rx_intr_enable() function enables rx queue
2671 * interrupt on specific rx queue of a port.
2674 * The port identifier of the Ethernet device.
2676 * The index of the receive queue from which to retrieve input packets.
2677 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2678 * to rte_eth_dev_configure().
2680 * - (0) if successful.
2681 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
2683 * - (-ENODEV) if *port_id* invalid.
2684 * - (-EIO) if device is removed.
2686 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id);
2689 * When lcore wakes up from rx interrupt indicating packet coming, disable rx
2690 * interrupt and returns to polling mode.
2692 * The rte_eth_dev_rx_intr_disable() function disables rx queue
2693 * interrupt on specific rx queue of a port.
2696 * The port identifier of the Ethernet device.
2698 * The index of the receive queue from which to retrieve input packets.
2699 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2700 * to rte_eth_dev_configure().
2702 * - (0) if successful.
2703 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
2705 * - (-ENODEV) if *port_id* invalid.
2706 * - (-EIO) if device is removed.
2708 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id);
2711 * RX Interrupt control per port.
2714 * The port identifier of the Ethernet device.
2716 * Epoll instance fd which the intr vector associated to.
2717 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
2719 * The operation be performed for the vector.
2720 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
2724 * - On success, zero.
2725 * - On failure, a negative value.
2727 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data);
2730 * RX Interrupt control per queue.
2733 * The port identifier of the Ethernet device.
2735 * The index of the receive queue from which to retrieve input packets.
2736 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
2737 * to rte_eth_dev_configure().
2739 * Epoll instance fd which the intr vector associated to.
2740 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance.
2742 * The operation be performed for the vector.
2743 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}.
2747 * - On success, zero.
2748 * - On failure, a negative value.
2750 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id,
2751 int epfd, int op, void *data);
2754 * Turn on the LED on the Ethernet device.
2755 * This function turns on the LED on the Ethernet device.
2758 * The port identifier of the Ethernet device.
2760 * - (0) if successful.
2761 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
2763 * - (-ENODEV) if *port_id* invalid.
2764 * - (-EIO) if device is removed.
2766 int rte_eth_led_on(uint16_t port_id);
2769 * Turn off the LED on the Ethernet device.
2770 * This function turns off the LED on the Ethernet device.
2773 * The port identifier of the Ethernet device.
2775 * - (0) if successful.
2776 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
2778 * - (-ENODEV) if *port_id* invalid.
2779 * - (-EIO) if device is removed.
2781 int rte_eth_led_off(uint16_t port_id);
2784 * Get current status of the Ethernet link flow control for Ethernet device
2787 * The port identifier of the Ethernet device.
2789 * The pointer to the structure where to store the flow control parameters.
2791 * - (0) if successful.
2792 * - (-ENOTSUP) if hardware doesn't support flow control.
2793 * - (-ENODEV) if *port_id* invalid.
2794 * - (-EIO) if device is removed.
2796 int rte_eth_dev_flow_ctrl_get(uint16_t port_id,
2797 struct rte_eth_fc_conf *fc_conf);
2800 * Configure the Ethernet link flow control for Ethernet device
2803 * The port identifier of the Ethernet device.
2805 * The pointer to the structure of the flow control parameters.
2807 * - (0) if successful.
2808 * - (-ENOTSUP) if hardware doesn't support flow control mode.
2809 * - (-ENODEV) if *port_id* invalid.
2810 * - (-EINVAL) if bad parameter
2811 * - (-EIO) if flow control setup failure or device is removed.
2813 int rte_eth_dev_flow_ctrl_set(uint16_t port_id,
2814 struct rte_eth_fc_conf *fc_conf);
2817 * Configure the Ethernet priority flow control under DCB environment
2818 * for Ethernet device.
2821 * The port identifier of the Ethernet device.
2823 * The pointer to the structure of the priority flow control parameters.
2825 * - (0) if successful.
2826 * - (-ENOTSUP) if hardware doesn't support priority flow control mode.
2827 * - (-ENODEV) if *port_id* invalid.
2828 * - (-EINVAL) if bad parameter
2829 * - (-EIO) if flow control setup failure or device is removed.
2831 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id,
2832 struct rte_eth_pfc_conf *pfc_conf);
2835 * Add a MAC address to an internal array of addresses used to enable whitelist
2836 * filtering to accept packets only if the destination MAC address matches.
2839 * The port identifier of the Ethernet device.
2841 * The MAC address to add.
2843 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
2844 * not enabled, this should be set to 0.
2846 * - (0) if successfully added or *mac_addr* was already added.
2847 * - (-ENOTSUP) if hardware doesn't support this feature.
2848 * - (-ENODEV) if *port* is invalid.
2849 * - (-EIO) if device is removed.
2850 * - (-ENOSPC) if no more MAC addresses can be added.
2851 * - (-EINVAL) if MAC address is invalid.
2853 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct ether_addr *mac_addr,
2857 * Remove a MAC address from the internal array of addresses.
2860 * The port identifier of the Ethernet device.
2862 * MAC address to remove.
2864 * - (0) if successful, or *mac_addr* didn't exist.
2865 * - (-ENOTSUP) if hardware doesn't support.
2866 * - (-ENODEV) if *port* invalid.
2867 * - (-EADDRINUSE) if attempting to remove the default MAC address
2869 int rte_eth_dev_mac_addr_remove(uint16_t port_id, struct ether_addr *mac_addr);
2872 * Set the default MAC address.
2875 * The port identifier of the Ethernet device.
2877 * New default MAC address.
2879 * - (0) if successful, or *mac_addr* didn't exist.
2880 * - (-ENOTSUP) if hardware doesn't support.
2881 * - (-ENODEV) if *port* invalid.
2882 * - (-EINVAL) if MAC address is invalid.
2884 int rte_eth_dev_default_mac_addr_set(uint16_t port_id,
2885 struct ether_addr *mac_addr);
2888 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
2891 * The port identifier of the Ethernet device.
2895 * Redirection table size. The table size can be queried by
2896 * rte_eth_dev_info_get().
2898 * - (0) if successful.
2899 * - (-ENOTSUP) if hardware doesn't support.
2900 * - (-EINVAL) if bad parameter.
2901 * - (-EIO) if device is removed.
2903 int rte_eth_dev_rss_reta_update(uint16_t port_id,
2904 struct rte_eth_rss_reta_entry64 *reta_conf,
2905 uint16_t reta_size);
2908 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
2911 * The port identifier of the Ethernet device.
2915 * Redirection table size. The table size can be queried by
2916 * rte_eth_dev_info_get().
2918 * - (0) if successful.
2919 * - (-ENOTSUP) if hardware doesn't support.
2920 * - (-EINVAL) if bad parameter.
2921 * - (-EIO) if device is removed.
2923 int rte_eth_dev_rss_reta_query(uint16_t port_id,
2924 struct rte_eth_rss_reta_entry64 *reta_conf,
2925 uint16_t reta_size);
2928 * Updates unicast hash table for receiving packet with the given destination
2929 * MAC address, and the packet is routed to all VFs for which the RX mode is
2930 * accept packets that match the unicast hash table.
2933 * The port identifier of the Ethernet device.
2935 * Unicast MAC address.
2937 * 1 - Set an unicast hash bit for receiving packets with the MAC address.
2938 * 0 - Clear an unicast hash bit.
2940 * - (0) if successful.
2941 * - (-ENOTSUP) if hardware doesn't support.
2942 * - (-ENODEV) if *port_id* invalid.
2943 * - (-EIO) if device is removed.
2944 * - (-EINVAL) if bad parameter.
2946 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct ether_addr *addr,
2950 * Updates all unicast hash bitmaps for receiving packet with any Unicast
2951 * Ethernet MAC addresses,the packet is routed to all VFs for which the RX
2952 * mode is accept packets that match the unicast hash table.
2955 * The port identifier of the Ethernet device.
2957 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet
2959 * 0 - Clear all unicast hash bitmaps
2961 * - (0) if successful.
2962 * - (-ENOTSUP) if hardware doesn't support.
2963 * - (-ENODEV) if *port_id* invalid.
2964 * - (-EIO) if device is removed.
2965 * - (-EINVAL) if bad parameter.
2967 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on);
2970 * Set a traffic mirroring rule on an Ethernet device
2973 * The port identifier of the Ethernet device.
2974 * @param mirror_conf
2975 * The pointer to the traffic mirroring structure describing the mirroring rule.
2976 * The *rte_eth_vm_mirror_conf* structure includes the type of mirroring rule,
2977 * destination pool and the value of rule if enable vlan or pool mirroring.
2980 * The index of traffic mirroring rule, we support four separated rules.
2982 * 1 - Enable a mirroring rule.
2983 * 0 - Disable a mirroring rule.
2985 * - (0) if successful.
2986 * - (-ENOTSUP) if hardware doesn't support this feature.
2987 * - (-ENODEV) if *port_id* invalid.
2988 * - (-EIO) if device is removed.
2989 * - (-EINVAL) if the mr_conf information is not correct.
2991 int rte_eth_mirror_rule_set(uint16_t port_id,
2992 struct rte_eth_mirror_conf *mirror_conf,
2997 * Reset a traffic mirroring rule on an Ethernet device.
3000 * The port identifier of the Ethernet device.
3002 * The index of traffic mirroring rule, we support four separated rules.
3004 * - (0) if successful.
3005 * - (-ENOTSUP) if hardware doesn't support this feature.
3006 * - (-ENODEV) if *port_id* invalid.
3007 * - (-EIO) if device is removed.
3008 * - (-EINVAL) if bad parameter.
3010 int rte_eth_mirror_rule_reset(uint16_t port_id,
3014 * Set the rate limitation for a queue on an Ethernet device.
3017 * The port identifier of the Ethernet device.
3021 * The tx rate in Mbps. Allocated from the total port link speed.
3023 * - (0) if successful.
3024 * - (-ENOTSUP) if hardware doesn't support this feature.
3025 * - (-ENODEV) if *port_id* invalid.
3026 * - (-EIO) if device is removed.
3027 * - (-EINVAL) if bad parameter.
3029 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx,
3033 * Configuration of Receive Side Scaling hash computation of Ethernet device.
3036 * The port identifier of the Ethernet device.
3038 * The new configuration to use for RSS hash computation on the port.
3040 * - (0) if successful.
3041 * - (-ENODEV) if port identifier is invalid.
3042 * - (-EIO) if device is removed.
3043 * - (-ENOTSUP) if hardware doesn't support.
3044 * - (-EINVAL) if bad parameter.
3046 int rte_eth_dev_rss_hash_update(uint16_t port_id,
3047 struct rte_eth_rss_conf *rss_conf);
3050 * Retrieve current configuration of Receive Side Scaling hash computation
3051 * of Ethernet device.
3054 * The port identifier of the Ethernet device.
3056 * Where to store the current RSS hash configuration of the Ethernet device.
3058 * - (0) if successful.
3059 * - (-ENODEV) if port identifier is invalid.
3060 * - (-EIO) if device is removed.
3061 * - (-ENOTSUP) if hardware doesn't support RSS.
3064 rte_eth_dev_rss_hash_conf_get(uint16_t port_id,
3065 struct rte_eth_rss_conf *rss_conf);
3068 * Add UDP tunneling port for a specific type of tunnel.
3069 * The packets with this UDP port will be identified as this type of tunnel.
3070 * Before enabling any offloading function for a tunnel, users can call this API
3071 * to change or add more UDP port for the tunnel. So the offloading function
3072 * can take effect on the packets with the specific UDP port.
3075 * The port identifier of the Ethernet device.
3077 * UDP tunneling configuration.
3080 * - (0) if successful.
3081 * - (-ENODEV) if port identifier is invalid.
3082 * - (-EIO) if device is removed.
3083 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3086 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id,
3087 struct rte_eth_udp_tunnel *tunnel_udp);
3090 * Delete UDP tunneling port a specific type of tunnel.
3091 * The packets with this UDP port will not be identified as this type of tunnel
3093 * Before enabling any offloading function for a tunnel, users can call this API
3094 * to delete a UDP port for the tunnel. So the offloading function will not take
3095 * effect on the packets with the specific UDP port.
3098 * The port identifier of the Ethernet device.
3100 * UDP tunneling configuration.
3103 * - (0) if successful.
3104 * - (-ENODEV) if port identifier is invalid.
3105 * - (-EIO) if device is removed.
3106 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3109 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id,
3110 struct rte_eth_udp_tunnel *tunnel_udp);
3113 * Check whether the filter type is supported on an Ethernet device.
3114 * All the supported filter types are defined in 'rte_eth_ctrl.h'.
3117 * The port identifier of the Ethernet device.
3118 * @param filter_type
3121 * - (0) if successful.
3122 * - (-ENOTSUP) if hardware doesn't support this filter type.
3123 * - (-ENODEV) if *port_id* invalid.
3124 * - (-EIO) if device is removed.
3126 int rte_eth_dev_filter_supported(uint16_t port_id,
3127 enum rte_filter_type filter_type);
3130 * Take operations to assigned filter type on an Ethernet device.
3131 * All the supported operations and filter types are defined in 'rte_eth_ctrl.h'.
3134 * The port identifier of the Ethernet device.
3135 * @param filter_type
3138 * Type of operation.
3140 * A pointer to arguments defined specifically for the operation.
3142 * - (0) if successful.
3143 * - (-ENOTSUP) if hardware doesn't support.
3144 * - (-ENODEV) if *port_id* invalid.
3145 * - (-EIO) if device is removed.
3146 * - others depends on the specific operations implementation.
3148 int rte_eth_dev_filter_ctrl(uint16_t port_id, enum rte_filter_type filter_type,
3149 enum rte_filter_op filter_op, void *arg);
3152 * Get DCB information on an Ethernet device.
3155 * The port identifier of the Ethernet device.
3159 * - (0) if successful.
3160 * - (-ENODEV) if port identifier is invalid.
3161 * - (-EIO) if device is removed.
3162 * - (-ENOTSUP) if hardware doesn't support.
3164 int rte_eth_dev_get_dcb_info(uint16_t port_id,
3165 struct rte_eth_dcb_info *dcb_info);
3167 struct rte_eth_rxtx_callback;
3170 * Add a callback to be called on packet RX on a given port and queue.
3172 * This API configures a function to be called for each burst of
3173 * packets received on a given NIC port queue. The return value is a pointer
3174 * that can be used to later remove the callback using
3175 * rte_eth_remove_rx_callback().
3177 * Multiple functions are called in the order that they are added.
3180 * The port identifier of the Ethernet device.
3182 * The queue on the Ethernet device on which the callback is to be added.
3184 * The callback function
3186 * A generic pointer parameter which will be passed to each invocation of the
3187 * callback function on this port and queue.
3191 * On success, a pointer value which can later be used to remove the callback.
3193 const struct rte_eth_rxtx_callback *
3194 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id,
3195 rte_rx_callback_fn fn, void *user_param);
3198 * Add a callback that must be called first on packet RX on a given port
3201 * This API configures a first function to be called for each burst of
3202 * packets received on a given NIC port queue. The return value is a pointer
3203 * that can be used to later remove the callback using
3204 * rte_eth_remove_rx_callback().
3206 * Multiple functions are called in the order that they are added.
3209 * The port identifier of the Ethernet device.
3211 * The queue on the Ethernet device on which the callback is to be added.
3213 * The callback function
3215 * A generic pointer parameter which will be passed to each invocation of the
3216 * callback function on this port and queue.
3220 * On success, a pointer value which can later be used to remove the callback.
3222 const struct rte_eth_rxtx_callback *
3223 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id,
3224 rte_rx_callback_fn fn, void *user_param);
3227 * Add a callback to be called on packet TX on a given port and queue.
3229 * This API configures a function to be called for each burst of
3230 * packets sent on a given NIC port queue. The return value is a pointer
3231 * that can be used to later remove the callback using
3232 * rte_eth_remove_tx_callback().
3234 * Multiple functions are called in the order that they are added.
3237 * The port identifier of the Ethernet device.
3239 * The queue on the Ethernet device on which the callback is to be added.
3241 * The callback function
3243 * A generic pointer parameter which will be passed to each invocation of the
3244 * callback function on this port and queue.
3248 * On success, a pointer value which can later be used to remove the callback.
3250 const struct rte_eth_rxtx_callback *
3251 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id,
3252 rte_tx_callback_fn fn, void *user_param);
3255 * Remove an RX packet callback from a given port and queue.
3257 * This function is used to removed callbacks that were added to a NIC port
3258 * queue using rte_eth_add_rx_callback().
3260 * Note: the callback is removed from the callback list but it isn't freed
3261 * since the it may still be in use. The memory for the callback can be
3262 * subsequently freed back by the application by calling rte_free():
3264 * - Immediately - if the port is stopped, or the user knows that no
3265 * callbacks are in flight e.g. if called from the thread doing RX/TX
3268 * - After a short delay - where the delay is sufficient to allow any
3269 * in-flight callbacks to complete.
3272 * The port identifier of the Ethernet device.
3274 * The queue on the Ethernet device from which the callback is to be removed.
3276 * User supplied callback created via rte_eth_add_rx_callback().
3279 * - 0: Success. Callback was removed.
3280 * - -ENOTSUP: Callback support is not available.
3281 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3282 * is NULL or not found for the port/queue.
3284 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id,
3285 const struct rte_eth_rxtx_callback *user_cb);
3288 * Remove a TX packet callback from a given port and queue.
3290 * This function is used to removed callbacks that were added to a NIC port
3291 * queue using rte_eth_add_tx_callback().
3293 * Note: the callback is removed from the callback list but it isn't freed
3294 * since the it may still be in use. The memory for the callback can be
3295 * subsequently freed back by the application by calling rte_free():
3297 * - Immediately - if the port is stopped, or the user knows that no
3298 * callbacks are in flight e.g. if called from the thread doing RX/TX
3301 * - After a short delay - where the delay is sufficient to allow any
3302 * in-flight callbacks to complete.
3305 * The port identifier of the Ethernet device.
3307 * The queue on the Ethernet device from which the callback is to be removed.
3309 * User supplied callback created via rte_eth_add_tx_callback().
3312 * - 0: Success. Callback was removed.
3313 * - -ENOTSUP: Callback support is not available.
3314 * - -EINVAL: The port_id or the queue_id is out of range, or the callback
3315 * is NULL or not found for the port/queue.
3317 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id,
3318 const struct rte_eth_rxtx_callback *user_cb);
3321 * Retrieve information about given port's RX queue.
3324 * The port identifier of the Ethernet device.
3326 * The RX queue on the Ethernet device for which information
3327 * will be retrieved.
3329 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with
3330 * the information of the Ethernet device.
3334 * - -ENOTSUP: routine is not supported by the device PMD.
3335 * - -EINVAL: The port_id or the queue_id is out of range.
3337 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3338 struct rte_eth_rxq_info *qinfo);
3341 * Retrieve information about given port's TX queue.
3344 * The port identifier of the Ethernet device.
3346 * The TX queue on the Ethernet device for which information
3347 * will be retrieved.
3349 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with
3350 * the information of the Ethernet device.
3354 * - -ENOTSUP: routine is not supported by the device PMD.
3355 * - -EINVAL: The port_id or the queue_id is out of range.
3357 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id,
3358 struct rte_eth_txq_info *qinfo);
3361 * Retrieve device registers and register attributes (number of registers and
3365 * The port identifier of the Ethernet device.
3367 * Pointer to rte_dev_reg_info structure to fill in. If info->data is
3368 * NULL the function fills in the width and length fields. If non-NULL
3369 * the registers are put into the buffer pointed at by the data field.
3371 * - (0) if successful.
3372 * - (-ENOTSUP) if hardware doesn't support.
3373 * - (-ENODEV) if *port_id* invalid.
3374 * - (-EIO) if device is removed.
3375 * - others depends on the specific operations implementation.
3377 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info);
3380 * Retrieve size of device EEPROM
3383 * The port identifier of the Ethernet device.
3385 * - (>=0) EEPROM size if successful.
3386 * - (-ENOTSUP) if hardware doesn't support.
3387 * - (-ENODEV) if *port_id* invalid.
3388 * - (-EIO) if device is removed.
3389 * - others depends on the specific operations implementation.
3391 int rte_eth_dev_get_eeprom_length(uint16_t port_id);
3394 * Retrieve EEPROM and EEPROM attribute
3397 * The port identifier of the Ethernet device.
3399 * The template includes buffer for return EEPROM data and
3400 * EEPROM attributes to be filled.
3402 * - (0) if successful.
3403 * - (-ENOTSUP) if hardware doesn't support.
3404 * - (-ENODEV) if *port_id* invalid.
3405 * - (-EIO) if device is removed.
3406 * - others depends on the specific operations implementation.
3408 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
3411 * Program EEPROM with provided data
3414 * The port identifier of the Ethernet device.
3416 * The template includes EEPROM data for programming and
3417 * EEPROM attributes to be filled
3419 * - (0) if successful.
3420 * - (-ENOTSUP) if hardware doesn't support.
3421 * - (-ENODEV) if *port_id* invalid.
3422 * - (-EIO) if device is removed.
3423 * - others depends on the specific operations implementation.
3425 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info);
3429 * @b EXPERIMENTAL: this API may change without prior notice.
3431 * Retrieve the type and size of plugin module EEPROM
3434 * The port identifier of the Ethernet device.
3436 * The type and size of plugin module EEPROM.
3438 * - (0) if successful.
3439 * - (-ENOTSUP) if hardware doesn't support.
3440 * - (-ENODEV) if *port_id* invalid.
3441 * - (-EIO) if device is removed.
3442 * - others depends on the specific operations implementation.
3444 int __rte_experimental
3445 rte_eth_dev_get_module_info(uint16_t port_id,
3446 struct rte_eth_dev_module_info *modinfo);
3450 * @b EXPERIMENTAL: this API may change without prior notice.
3452 * Retrieve the data of plugin module EEPROM
3455 * The port identifier of the Ethernet device.
3457 * The template includes the plugin module EEPROM attributes, and the
3458 * buffer for return plugin module EEPROM data.
3460 * - (0) if successful.
3461 * - (-ENOTSUP) if hardware doesn't support.
3462 * - (-ENODEV) if *port_id* invalid.
3463 * - (-EIO) if device is removed.
3464 * - others depends on the specific operations implementation.
3466 int __rte_experimental
3467 rte_eth_dev_get_module_eeprom(uint16_t port_id,
3468 struct rte_dev_eeprom_info *info);
3471 * Set the list of multicast addresses to filter on an Ethernet device.
3474 * The port identifier of the Ethernet device.
3475 * @param mc_addr_set
3476 * The array of multicast addresses to set. Equal to NULL when the function
3477 * is invoked to flush the set of filtered addresses.
3479 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0
3480 * when the function is invoked to flush the set of filtered addresses.
3482 * - (0) if successful.
3483 * - (-ENODEV) if *port_id* invalid.
3484 * - (-EIO) if device is removed.
3485 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering.
3486 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources.
3488 int rte_eth_dev_set_mc_addr_list(uint16_t port_id,
3489 struct ether_addr *mc_addr_set,
3490 uint32_t nb_mc_addr);
3493 * Enable IEEE1588/802.1AS timestamping for an Ethernet device.
3496 * The port identifier of the Ethernet device.
3500 * - -ENODEV: The port ID is invalid.
3501 * - -EIO: if device is removed.
3502 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3504 int rte_eth_timesync_enable(uint16_t port_id);
3507 * Disable IEEE1588/802.1AS timestamping for an Ethernet device.
3510 * The port identifier of the Ethernet device.
3514 * - -ENODEV: The port ID is invalid.
3515 * - -EIO: if device is removed.
3516 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3518 int rte_eth_timesync_disable(uint16_t port_id);
3521 * Read an IEEE1588/802.1AS RX timestamp from an Ethernet device.
3524 * The port identifier of the Ethernet device.
3526 * Pointer to the timestamp struct.
3528 * Device specific flags. Used to pass the RX timesync register index to
3529 * i40e. Unused in igb/ixgbe, pass 0 instead.
3533 * - -EINVAL: No timestamp is available.
3534 * - -ENODEV: The port ID is invalid.
3535 * - -EIO: if device is removed.
3536 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3538 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id,
3539 struct timespec *timestamp, uint32_t flags);
3542 * Read an IEEE1588/802.1AS TX timestamp from an Ethernet device.
3545 * The port identifier of the Ethernet device.
3547 * Pointer to the timestamp struct.
3551 * - -EINVAL: No timestamp is available.
3552 * - -ENODEV: The port ID is invalid.
3553 * - -EIO: if device is removed.
3554 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3556 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id,
3557 struct timespec *timestamp);
3560 * Adjust the timesync clock on an Ethernet device.
3562 * This is usually used in conjunction with other Ethdev timesync functions to
3563 * synchronize the device time using the IEEE1588/802.1AS protocol.
3566 * The port identifier of the Ethernet device.
3568 * The adjustment in nanoseconds.
3572 * - -ENODEV: The port ID is invalid.
3573 * - -EIO: if device is removed.
3574 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3576 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta);
3579 * Read the time from the timesync clock on an Ethernet device.
3581 * This is usually used in conjunction with other Ethdev timesync functions to
3582 * synchronize the device time using the IEEE1588/802.1AS protocol.
3585 * The port identifier of the Ethernet device.
3587 * Pointer to the timespec struct that holds the time.
3592 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time);
3595 * Set the time of the timesync clock on an Ethernet device.
3597 * This is usually used in conjunction with other Ethdev timesync functions to
3598 * synchronize the device time using the IEEE1588/802.1AS protocol.
3601 * The port identifier of the Ethernet device.
3603 * Pointer to the timespec struct that holds the time.
3607 * - -EINVAL: No timestamp is available.
3608 * - -ENODEV: The port ID is invalid.
3609 * - -EIO: if device is removed.
3610 * - -ENOTSUP: The function is not supported by the Ethernet driver.
3612 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time);
3615 * Config l2 tunnel ether type of an Ethernet device for filtering specific
3616 * tunnel packets by ether type.
3619 * The port identifier of the Ethernet device.
3621 * l2 tunnel configuration.
3624 * - (0) if successful.
3625 * - (-ENODEV) if port identifier is invalid.
3626 * - (-EIO) if device is removed.
3627 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3630 rte_eth_dev_l2_tunnel_eth_type_conf(uint16_t port_id,
3631 struct rte_eth_l2_tunnel_conf *l2_tunnel);
3634 * Enable/disable l2 tunnel offload functions. Include,
3635 * 1, The ability of parsing a type of l2 tunnel of an Ethernet device.
3636 * Filtering, forwarding and offloading this type of tunnel packets depend on
3638 * 2, Stripping the l2 tunnel tag.
3639 * 3, Insertion of the l2 tunnel tag.
3640 * 4, Forwarding the packets based on the l2 tunnel tag.
3643 * The port identifier of the Ethernet device.
3645 * l2 tunnel parameters.
3647 * Indicate the offload function.
3649 * Enable or disable this function.
3652 * - (0) if successful.
3653 * - (-ENODEV) if port identifier is invalid.
3654 * - (-EIO) if device is removed.
3655 * - (-ENOTSUP) if hardware doesn't support tunnel type.
3658 rte_eth_dev_l2_tunnel_offload_set(uint16_t port_id,
3659 struct rte_eth_l2_tunnel_conf *l2_tunnel,
3664 * Get the port id from pci address or device name
3666 * - PCIe, 0000:2:00.0
3671 * pci address or name of the device
3673 * pointer to port identifier of the device
3675 * - (0) if successful and port_id is filled.
3676 * - (-ENODEV or -EINVAL) on failure.
3679 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id);
3682 * Get the device name from port id
3684 * - PCIe Bus:Domain:Function, 0000:02:00.0
3685 * - SoC device name, fsl-gmac0
3686 * - vdev dpdk name, net_[pcap0|null0|tun0|tap0]
3689 * Port identifier of the device.
3691 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name.
3693 * - (0) if successful.
3694 * - (-EINVAL) on failure.
3697 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name);
3700 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from
3701 * the ethernet device information, otherwise adjust them to boundaries.
3704 * The port identifier of the Ethernet device.
3706 * A pointer to a uint16_t where the number of receive
3707 * descriptors stored.
3709 * A pointer to a uint16_t where the number of transmit
3710 * descriptors stored.
3712 * - (0) if successful.
3713 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure.
3715 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id,
3716 uint16_t *nb_rx_desc,
3717 uint16_t *nb_tx_desc);
3720 * Test if a port supports specific mempool ops.
3723 * Port identifier of the Ethernet device.
3725 * The name of the pool operations to test.
3727 * - 0: best mempool ops choice for this port.
3728 * - 1: mempool ops are supported for this port.
3729 * - -ENOTSUP: mempool ops not supported for this port.
3730 * - -ENODEV: Invalid port Identifier.
3731 * - -EINVAL: Pool param is null.
3734 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);
3737 * Get the security context for the Ethernet device.
3740 * Port identifier of the Ethernet device
3743 * - pointer to security context on success.
3746 rte_eth_dev_get_sec_ctx(uint16_t port_id);
3749 #include <rte_ethdev_core.h>
3753 * Retrieve a burst of input packets from a receive queue of an Ethernet
3754 * device. The retrieved packets are stored in *rte_mbuf* structures whose
3755 * pointers are supplied in the *rx_pkts* array.
3757 * The rte_eth_rx_burst() function loops, parsing the RX ring of the
3758 * receive queue, up to *nb_pkts* packets, and for each completed RX
3759 * descriptor in the ring, it performs the following operations:
3761 * - Initialize the *rte_mbuf* data structure associated with the
3762 * RX descriptor according to the information provided by the NIC into
3763 * that RX descriptor.
3765 * - Store the *rte_mbuf* data structure into the next entry of the
3768 * - Replenish the RX descriptor with a new *rte_mbuf* buffer
3769 * allocated from the memory pool associated with the receive queue at
3770 * initialization time.
3772 * When retrieving an input packet that was scattered by the controller
3773 * into multiple receive descriptors, the rte_eth_rx_burst() function
3774 * appends the associated *rte_mbuf* buffers to the first buffer of the
3777 * The rte_eth_rx_burst() function returns the number of packets
3778 * actually retrieved, which is the number of *rte_mbuf* data structures
3779 * effectively supplied into the *rx_pkts* array.
3780 * A return value equal to *nb_pkts* indicates that the RX queue contained
3781 * at least *rx_pkts* packets, and this is likely to signify that other
3782 * received packets remain in the input queue. Applications implementing
3783 * a "retrieve as much received packets as possible" policy can check this
3784 * specific case and keep invoking the rte_eth_rx_burst() function until
3785 * a value less than *nb_pkts* is returned.
3787 * This receive method has the following advantages:
3789 * - It allows a run-to-completion network stack engine to retrieve and
3790 * to immediately process received packets in a fast burst-oriented
3791 * approach, avoiding the overhead of unnecessary intermediate packet
3792 * queue/dequeue operations.
3794 * - Conversely, it also allows an asynchronous-oriented processing
3795 * method to retrieve bursts of received packets and to immediately
3796 * queue them for further parallel processing by another logical core,
3797 * for instance. However, instead of having received packets being
3798 * individually queued by the driver, this approach allows the caller
3799 * of the rte_eth_rx_burst() function to queue a burst of retrieved
3800 * packets at a time and therefore dramatically reduce the cost of
3801 * enqueue/dequeue operations per packet.
3803 * - It allows the rte_eth_rx_burst() function of the driver to take
3804 * advantage of burst-oriented hardware features (CPU cache,
3805 * prefetch instructions, and so on) to minimize the number of CPU
3806 * cycles per packet.
3808 * To summarize, the proposed receive API enables many
3809 * burst-oriented optimizations in both synchronous and asynchronous
3810 * packet processing environments with no overhead in both cases.
3812 * The rte_eth_rx_burst() function does not provide any error
3813 * notification to avoid the corresponding overhead. As a hint, the
3814 * upper-level application might check the status of the device link once
3815 * being systematically returned a 0 value for a given number of tries.
3818 * The port identifier of the Ethernet device.
3820 * The index of the receive queue from which to retrieve input packets.
3821 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
3822 * to rte_eth_dev_configure().
3824 * The address of an array of pointers to *rte_mbuf* structures that
3825 * must be large enough to store *nb_pkts* pointers in it.
3827 * The maximum number of packets to retrieve.
3829 * The number of packets actually retrieved, which is the number
3830 * of pointers to *rte_mbuf* structures effectively supplied to the
3833 static inline uint16_t
3834 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id,
3835 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts)
3837 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
3840 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3841 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
3842 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_pkt_burst, 0);
3844 if (queue_id >= dev->data->nb_rx_queues) {
3845 RTE_ETHDEV_LOG(ERR, "Invalid RX queue_id=%u\n", queue_id);
3849 nb_rx = (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id],
3852 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
3853 if (unlikely(dev->post_rx_burst_cbs[queue_id] != NULL)) {
3854 struct rte_eth_rxtx_callback *cb =
3855 dev->post_rx_burst_cbs[queue_id];
3858 nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx,
3859 nb_pkts, cb->param);
3861 } while (cb != NULL);
3869 * Get the number of used descriptors of a rx queue
3872 * The port identifier of the Ethernet device.
3874 * The queue id on the specific port.
3876 * The number of used descriptors in the specific queue, or:
3877 * (-EINVAL) if *port_id* or *queue_id* is invalid
3878 * (-ENOTSUP) if the device does not support this function
3881 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id)
3883 struct rte_eth_dev *dev;
3885 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL);
3886 dev = &rte_eth_devices[port_id];
3887 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_queue_count, -ENOTSUP);
3888 if (queue_id >= dev->data->nb_rx_queues)
3891 return (int)(*dev->dev_ops->rx_queue_count)(dev, queue_id);
3895 * Check if the DD bit of the specific RX descriptor in the queue has been set
3898 * The port identifier of the Ethernet device.
3900 * The queue id on the specific port.
3902 * The offset of the descriptor ID from tail.
3904 * - (1) if the specific DD bit is set.
3905 * - (0) if the specific DD bit is not set.
3906 * - (-ENODEV) if *port_id* invalid.
3907 * - (-ENOTSUP) if the device does not support this function
3910 rte_eth_rx_descriptor_done(uint16_t port_id, uint16_t queue_id, uint16_t offset)
3912 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
3913 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
3914 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_done, -ENOTSUP);
3915 return (*dev->dev_ops->rx_descriptor_done)( \
3916 dev->data->rx_queues[queue_id], offset);
3919 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */
3920 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */
3921 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
3924 * Check the status of a Rx descriptor in the queue
3926 * It should be called in a similar context than the Rx function:
3927 * - on a dataplane core
3928 * - not concurrently on the same queue
3930 * Since it's a dataplane function, no check is performed on port_id and
3931 * queue_id. The caller must therefore ensure that the port is enabled
3932 * and the queue is configured and running.
3934 * Note: accessing to a random descriptor in the ring may trigger cache
3935 * misses and have a performance impact.
3938 * A valid port identifier of the Ethernet device which.
3940 * A valid Rx queue identifier on this port.
3942 * The offset of the descriptor starting from tail (0 is the next
3943 * packet to be received by the driver).
3946 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to
3948 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but
3949 * not yet processed by the driver (i.e. in the receive queue).
3950 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by
3951 * the driver and not yet returned to hw, or reserved by the hw.
3952 * - (-EINVAL) bad descriptor offset.
3953 * - (-ENOTSUP) if the device does not support this function.
3954 * - (-ENODEV) bad port or queue (only if compiled with debug).
3957 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id,
3960 struct rte_eth_dev *dev;
3963 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3964 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
3966 dev = &rte_eth_devices[port_id];
3967 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
3968 if (queue_id >= dev->data->nb_rx_queues)
3971 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_status, -ENOTSUP);
3972 rxq = dev->data->rx_queues[queue_id];
3974 return (*dev->dev_ops->rx_descriptor_status)(rxq, offset);
3977 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */
3978 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */
3979 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */
3982 * Check the status of a Tx descriptor in the queue.
3984 * It should be called in a similar context than the Tx function:
3985 * - on a dataplane core
3986 * - not concurrently on the same queue
3988 * Since it's a dataplane function, no check is performed on port_id and
3989 * queue_id. The caller must therefore ensure that the port is enabled
3990 * and the queue is configured and running.
3992 * Note: accessing to a random descriptor in the ring may trigger cache
3993 * misses and have a performance impact.
3996 * A valid port identifier of the Ethernet device which.
3998 * A valid Tx queue identifier on this port.
4000 * The offset of the descriptor starting from tail (0 is the place where
4001 * the next packet will be send).
4004 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e.
4005 * in the transmit queue.
4006 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can
4007 * be reused by the driver.
4008 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the
4009 * driver or the hardware.
4010 * - (-EINVAL) bad descriptor offset.
4011 * - (-ENOTSUP) if the device does not support this function.
4012 * - (-ENODEV) bad port or queue (only if compiled with debug).
4014 static inline int rte_eth_tx_descriptor_status(uint16_t port_id,
4015 uint16_t queue_id, uint16_t offset)
4017 struct rte_eth_dev *dev;
4020 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4021 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
4023 dev = &rte_eth_devices[port_id];
4024 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4025 if (queue_id >= dev->data->nb_tx_queues)
4028 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->tx_descriptor_status, -ENOTSUP);
4029 txq = dev->data->tx_queues[queue_id];
4031 return (*dev->dev_ops->tx_descriptor_status)(txq, offset);
4035 * Send a burst of output packets on a transmit queue of an Ethernet device.
4037 * The rte_eth_tx_burst() function is invoked to transmit output packets
4038 * on the output queue *queue_id* of the Ethernet device designated by its
4040 * The *nb_pkts* parameter is the number of packets to send which are
4041 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4042 * allocated from a pool created with rte_pktmbuf_pool_create().
4043 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
4044 * up to the number of transmit descriptors available in the TX ring of the
4046 * For each packet to send, the rte_eth_tx_burst() function performs
4047 * the following operations:
4049 * - Pick up the next available descriptor in the transmit ring.
4051 * - Free the network buffer previously sent with that descriptor, if any.
4053 * - Initialize the transmit descriptor with the information provided
4054 * in the *rte_mbuf data structure.
4056 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
4057 * the rte_eth_tx_burst() function uses several transmit descriptors
4060 * The rte_eth_tx_burst() function returns the number of packets it
4061 * actually sent. A return value equal to *nb_pkts* means that all packets
4062 * have been sent, and this is likely to signify that other output packets
4063 * could be immediately transmitted again. Applications that implement a
4064 * "send as many packets to transmit as possible" policy can check this
4065 * specific case and keep invoking the rte_eth_tx_burst() function until
4066 * a value less than *nb_pkts* is returned.
4068 * It is the responsibility of the rte_eth_tx_burst() function to
4069 * transparently free the memory buffers of packets previously sent.
4070 * This feature is driven by the *tx_free_thresh* value supplied to the
4071 * rte_eth_dev_configure() function at device configuration time.
4072 * When the number of free TX descriptors drops below this threshold, the
4073 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers
4074 * of those packets whose transmission was effectively completed.
4076 * If the PMD is DEV_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can
4077 * invoke this function concurrently on the same tx queue without SW lock.
4078 * @see rte_eth_dev_info_get, struct rte_eth_txconf::txq_flags
4080 * @see rte_eth_tx_prepare to perform some prior checks or adjustments
4084 * The port identifier of the Ethernet device.
4086 * The index of the transmit queue through which output packets must be
4088 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4089 * to rte_eth_dev_configure().
4091 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4092 * which contain the output packets.
4094 * The maximum number of packets to transmit.
4096 * The number of output packets actually stored in transmit descriptors of
4097 * the transmit ring. The return value can be less than the value of the
4098 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
4100 static inline uint16_t
4101 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id,
4102 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4104 struct rte_eth_dev *dev = &rte_eth_devices[port_id];
4106 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4107 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0);
4108 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_pkt_burst, 0);
4110 if (queue_id >= dev->data->nb_tx_queues) {
4111 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4116 #ifdef RTE_ETHDEV_RXTX_CALLBACKS
4117 struct rte_eth_rxtx_callback *cb = dev->pre_tx_burst_cbs[queue_id];
4119 if (unlikely(cb != NULL)) {
4121 nb_pkts = cb->fn.tx(port_id, queue_id, tx_pkts, nb_pkts,
4124 } while (cb != NULL);
4128 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts);
4133 * @b EXPERIMENTAL: this API may change without prior notice
4135 * Process a burst of output packets on a transmit queue of an Ethernet device.
4137 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be
4138 * transmitted on the output queue *queue_id* of the Ethernet device designated
4140 * The *nb_pkts* parameter is the number of packets to be prepared which are
4141 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them
4142 * allocated from a pool created with rte_pktmbuf_pool_create().
4143 * For each packet to send, the rte_eth_tx_prepare() function performs
4144 * the following operations:
4146 * - Check if packet meets devices requirements for tx offloads.
4148 * - Check limitations about number of segments.
4150 * - Check additional requirements when debug is enabled.
4152 * - Update and/or reset required checksums when tx offload is set for packet.
4154 * Since this function can modify packet data, provided mbufs must be safely
4155 * writable (e.g. modified data cannot be in shared segment).
4157 * The rte_eth_tx_prepare() function returns the number of packets ready to be
4158 * sent. A return value equal to *nb_pkts* means that all packets are valid and
4159 * ready to be sent, otherwise stops processing on the first invalid packet and
4160 * leaves the rest packets untouched.
4162 * When this functionality is not implemented in the driver, all packets are
4163 * are returned untouched.
4166 * The port identifier of the Ethernet device.
4167 * The value must be a valid port id.
4169 * The index of the transmit queue through which output packets must be
4171 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4172 * to rte_eth_dev_configure().
4174 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
4175 * which contain the output packets.
4177 * The maximum number of packets to process.
4179 * The number of packets correct and ready to be sent. The return value can be
4180 * less than the value of the *tx_pkts* parameter when some packet doesn't
4181 * meet devices requirements with rte_errno set appropriately:
4182 * - -EINVAL: offload flags are not correctly set
4183 * - -ENOTSUP: the offload feature is not supported by the hardware
4187 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP
4189 static inline uint16_t
4190 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id,
4191 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4193 struct rte_eth_dev *dev;
4195 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4196 if (!rte_eth_dev_is_valid_port(port_id)) {
4197 RTE_ETHDEV_LOG(ERR, "Invalid TX port_id=%u\n", port_id);
4198 rte_errno = -EINVAL;
4203 dev = &rte_eth_devices[port_id];
4205 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
4206 if (queue_id >= dev->data->nb_tx_queues) {
4207 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id);
4208 rte_errno = -EINVAL;
4213 if (!dev->tx_pkt_prepare)
4216 return (*dev->tx_pkt_prepare)(dev->data->tx_queues[queue_id],
4223 * Native NOOP operation for compilation targets which doesn't require any
4224 * preparations steps, and functional NOOP may introduce unnecessary performance
4227 * Generally this is not a good idea to turn it on globally and didn't should
4228 * be used if behavior of tx_preparation can change.
4231 static inline uint16_t
4232 rte_eth_tx_prepare(__rte_unused uint16_t port_id,
4233 __rte_unused uint16_t queue_id,
4234 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
4242 * Send any packets queued up for transmission on a port and HW queue
4244 * This causes an explicit flush of packets previously buffered via the
4245 * rte_eth_tx_buffer() function. It returns the number of packets successfully
4246 * sent to the NIC, and calls the error callback for any unsent packets. Unless
4247 * explicitly set up otherwise, the default callback simply frees the unsent
4248 * packets back to the owning mempool.
4251 * The port identifier of the Ethernet device.
4253 * The index of the transmit queue through which output packets must be
4255 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4256 * to rte_eth_dev_configure().
4258 * Buffer of packets to be transmit.
4260 * The number of packets successfully sent to the Ethernet device. The error
4261 * callback is called for any packets which could not be sent.
4263 static inline uint16_t
4264 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id,
4265 struct rte_eth_dev_tx_buffer *buffer)
4268 uint16_t to_send = buffer->length;
4273 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send);
4277 /* All packets sent, or to be dealt with by callback below */
4278 if (unlikely(sent != to_send))
4279 buffer->error_callback(&buffer->pkts[sent],
4280 (uint16_t)(to_send - sent),
4281 buffer->error_userdata);
4287 * Buffer a single packet for future transmission on a port and queue
4289 * This function takes a single mbuf/packet and buffers it for later
4290 * transmission on the particular port and queue specified. Once the buffer is
4291 * full of packets, an attempt will be made to transmit all the buffered
4292 * packets. In case of error, where not all packets can be transmitted, a
4293 * callback is called with the unsent packets as a parameter. If no callback
4294 * is explicitly set up, the unsent packets are just freed back to the owning
4295 * mempool. The function returns the number of packets actually sent i.e.
4296 * 0 if no buffer flush occurred, otherwise the number of packets successfully
4300 * The port identifier of the Ethernet device.
4302 * The index of the transmit queue through which output packets must be
4304 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
4305 * to rte_eth_dev_configure().
4307 * Buffer used to collect packets to be sent.
4309 * Pointer to the packet mbuf to be sent.
4311 * 0 = packet has been buffered for later transmission
4312 * N > 0 = packet has been buffered, and the buffer was subsequently flushed,
4313 * causing N packets to be sent, and the error callback to be called for
4316 static __rte_always_inline uint16_t
4317 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id,
4318 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt)
4320 buffer->pkts[buffer->length++] = tx_pkt;
4321 if (buffer->length < buffer->size)
4324 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer);
4331 #endif /* _RTE_ETHDEV_H_ */