1 /* SPDX-License-Identifier: BSD-3-Clause 2 * Copyright(c) 2010-2017 Intel Corporation 3 */ 4 5 #ifndef _RTE_ETHDEV_H_ 6 #define _RTE_ETHDEV_H_ 7 8 /** 9 * @file 10 * 11 * RTE Ethernet Device API 12 * 13 * The Ethernet Device API is composed of two parts: 14 * 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. 19 * 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. 24 * 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. 32 * 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. 36 * 37 * In all functions of the Ethernet API, the Ethernet device is 38 * designated by an integer >= 0 named the device port identifier. 39 * 40 * At the Ethernet driver level, Ethernet devices are represented by a generic 41 * data structure of type *rte_eth_dev*. 42 * 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. 49 * 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 52 * from NVM etc. 53 * 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 58 * freed. 59 * 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() 66 * 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. 71 * 72 * If the application wants to change the configuration (i.e. call 73 * rte_eth_dev_configure(), rte_eth_tx_queue_setup(), or 74 * rte_eth_rx_queue_setup()), it must call rte_eth_dev_stop() first to stop the 75 * device and then do the reconfiguration before calling rte_eth_dev_start() 76 * again. The transmit and receive functions should not be invoked when the 77 * device is stopped. 78 * 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 81 * be retained: 82 * 83 * - MTU 84 * - flow control settings 85 * - receive mode configuration (promiscuous mode, all-multicast mode, 86 * hardware checksum mode, RSS/VMDQ settings etc.) 87 * - VLAN filtering configuration 88 * - default MAC address 89 * - MAC addresses supplied to MAC address array 90 * - flow director filtering mode (but not filtering rules) 91 * - NIC queue statistics mappings 92 * 93 * Any other configuration will not be stored and will need to be re-entered 94 * before a call to rte_eth_dev_start(). 95 * 96 * Finally, a network application can close an Ethernet device by invoking the 97 * rte_eth_dev_close() function. 98 * 99 * Each function of the application Ethernet API invokes a specific function 100 * of the PMD that controls the target device designated by its port 101 * identifier. 102 * For this purpose, all device-specific functions of an Ethernet driver are 103 * supplied through a set of pointers contained in a generic structure of type 104 * *eth_dev_ops*. 105 * The address of the *eth_dev_ops* structure is stored in the *rte_eth_dev* 106 * structure by the device init function of the Ethernet driver, which is 107 * invoked during the PCI probing phase, as explained earlier. 108 * 109 * In other words, each function of the Ethernet API simply retrieves the 110 * *rte_eth_dev* structure associated with the device port identifier and 111 * performs an indirect invocation of the corresponding driver function 112 * supplied in the *eth_dev_ops* structure of the *rte_eth_dev* structure. 113 * 114 * For performance reasons, the address of the burst-oriented RX and TX 115 * functions of the Ethernet driver are not contained in the *eth_dev_ops* 116 * structure. Instead, they are directly stored at the beginning of the 117 * *rte_eth_dev* structure to avoid an extra indirect memory access during 118 * their invocation. 119 * 120 * RTE ethernet device drivers do not use interrupts for transmitting or 121 * receiving. Instead, Ethernet drivers export Poll-Mode receive and transmit 122 * functions to applications. 123 * Both receive and transmit functions are packet-burst oriented to minimize 124 * their cost per packet through the following optimizations: 125 * 126 * - Sharing among multiple packets the incompressible cost of the 127 * invocation of receive/transmit functions. 128 * 129 * - Enabling receive/transmit functions to take advantage of burst-oriented 130 * hardware features (L1 cache, prefetch instructions, NIC head/tail 131 * registers) to minimize the number of CPU cycles per packet, for instance, 132 * by avoiding useless read memory accesses to ring descriptors, or by 133 * systematically using arrays of pointers that exactly fit L1 cache line 134 * boundaries and sizes. 135 * 136 * The burst-oriented receive function does not provide any error notification, 137 * to avoid the corresponding overhead. As a hint, the upper-level application 138 * might check the status of the device link once being systematically returned 139 * a 0 value by the receive function of the driver for a given number of tries. 140 */ 141 142 #ifdef __cplusplus 143 extern "C" { 144 #endif 145 146 #include <stdint.h> 147 148 /* Use this macro to check if LRO API is supported */ 149 #define RTE_ETHDEV_HAS_LRO_SUPPORT 150 151 #include <rte_compat.h> 152 #include <rte_log.h> 153 #include <rte_interrupts.h> 154 #include <rte_dev.h> 155 #include <rte_devargs.h> 156 #include <rte_errno.h> 157 #include <rte_common.h> 158 #include <rte_config.h> 159 #include <rte_ether.h> 160 161 #include "rte_dev_info.h" 162 163 extern int rte_eth_dev_logtype; 164 165 #define RTE_ETHDEV_LOG(level, ...) \ 166 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__) 167 168 struct rte_mbuf; 169 170 /** 171 * Initializes a device iterator. 172 * 173 * This iterator allows accessing a list of devices matching some devargs. 174 * 175 * @param iter 176 * Device iterator handle initialized by the function. 177 * The fields bus_str and cls_str might be dynamically allocated, 178 * and could be freed by calling rte_eth_iterator_cleanup(). 179 * 180 * @param devargs 181 * Device description string. 182 * 183 * @return 184 * 0 on successful initialization, negative otherwise. 185 */ 186 int rte_eth_iterator_init(struct rte_dev_iterator *iter, const char *devargs); 187 188 /** 189 * Iterates on devices with devargs filter. 190 * The ownership is not checked. 191 * 192 * The next port id is returned, and the iterator is updated. 193 * 194 * @param iter 195 * Device iterator handle initialized by rte_eth_iterator_init(). 196 * Some fields bus_str and cls_str might be freed when no more port is found, 197 * by calling rte_eth_iterator_cleanup(). 198 * 199 * @return 200 * A port id if found, RTE_MAX_ETHPORTS otherwise. 201 */ 202 uint16_t rte_eth_iterator_next(struct rte_dev_iterator *iter); 203 204 /** 205 * Free some allocated fields of the iterator. 206 * 207 * This function is automatically called by rte_eth_iterator_next() 208 * on the last iteration (i.e. when no more matching port is found). 209 * 210 * It is safe to call this function twice; it will do nothing more. 211 * 212 * @param iter 213 * Device iterator handle initialized by rte_eth_iterator_init(). 214 * The fields bus_str and cls_str are freed if needed. 215 */ 216 void rte_eth_iterator_cleanup(struct rte_dev_iterator *iter); 217 218 /** 219 * Macro to iterate over all ethdev ports matching some devargs. 220 * 221 * If a break is done before the end of the loop, 222 * the function rte_eth_iterator_cleanup() must be called. 223 * 224 * @param id 225 * Iterated port id of type uint16_t. 226 * @param devargs 227 * Device parameters input as string of type char*. 228 * @param iter 229 * Iterator handle of type struct rte_dev_iterator, used internally. 230 */ 231 #define RTE_ETH_FOREACH_MATCHING_DEV(id, devargs, iter) \ 232 for (rte_eth_iterator_init(iter, devargs), \ 233 id = rte_eth_iterator_next(iter); \ 234 id != RTE_MAX_ETHPORTS; \ 235 id = rte_eth_iterator_next(iter)) 236 237 /** 238 * A structure used to retrieve statistics for an Ethernet port. 239 * Not all statistics fields in struct rte_eth_stats are supported 240 * by any type of network interface card (NIC). If any statistics 241 * field is not supported, its value is 0. 242 */ 243 struct rte_eth_stats { 244 uint64_t ipackets; /**< Total number of successfully received packets. */ 245 uint64_t opackets; /**< Total number of successfully transmitted packets.*/ 246 uint64_t ibytes; /**< Total number of successfully received bytes. */ 247 uint64_t obytes; /**< Total number of successfully transmitted bytes. */ 248 uint64_t imissed; 249 /**< Total of RX packets dropped by the HW, 250 * because there are no available buffer (i.e. RX queues are full). 251 */ 252 uint64_t ierrors; /**< Total number of erroneous received packets. */ 253 uint64_t oerrors; /**< Total number of failed transmitted packets. */ 254 uint64_t rx_nombuf; /**< Total number of RX mbuf allocation failures. */ 255 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 256 /**< Total number of queue RX packets. */ 257 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 258 /**< Total number of queue TX packets. */ 259 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 260 /**< Total number of successfully received queue bytes. */ 261 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 262 /**< Total number of successfully transmitted queue bytes. */ 263 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 264 /**< Total number of queue packets received that are dropped. */ 265 }; 266 267 /** 268 * Device supported speeds bitmap flags 269 */ 270 #define ETH_LINK_SPEED_AUTONEG (0 << 0) /**< Autonegotiate (all speeds) */ 271 #define ETH_LINK_SPEED_FIXED (1 << 0) /**< Disable autoneg (fixed speed) */ 272 #define ETH_LINK_SPEED_10M_HD (1 << 1) /**< 10 Mbps half-duplex */ 273 #define ETH_LINK_SPEED_10M (1 << 2) /**< 10 Mbps full-duplex */ 274 #define ETH_LINK_SPEED_100M_HD (1 << 3) /**< 100 Mbps half-duplex */ 275 #define ETH_LINK_SPEED_100M (1 << 4) /**< 100 Mbps full-duplex */ 276 #define ETH_LINK_SPEED_1G (1 << 5) /**< 1 Gbps */ 277 #define ETH_LINK_SPEED_2_5G (1 << 6) /**< 2.5 Gbps */ 278 #define ETH_LINK_SPEED_5G (1 << 7) /**< 5 Gbps */ 279 #define ETH_LINK_SPEED_10G (1 << 8) /**< 10 Gbps */ 280 #define ETH_LINK_SPEED_20G (1 << 9) /**< 20 Gbps */ 281 #define ETH_LINK_SPEED_25G (1 << 10) /**< 25 Gbps */ 282 #define ETH_LINK_SPEED_40G (1 << 11) /**< 40 Gbps */ 283 #define ETH_LINK_SPEED_50G (1 << 12) /**< 50 Gbps */ 284 #define ETH_LINK_SPEED_56G (1 << 13) /**< 56 Gbps */ 285 #define ETH_LINK_SPEED_100G (1 << 14) /**< 100 Gbps */ 286 287 /** 288 * Ethernet numeric link speeds in Mbps 289 */ 290 #define ETH_SPEED_NUM_NONE 0 /**< Not defined */ 291 #define ETH_SPEED_NUM_10M 10 /**< 10 Mbps */ 292 #define ETH_SPEED_NUM_100M 100 /**< 100 Mbps */ 293 #define ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */ 294 #define ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */ 295 #define ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */ 296 #define ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */ 297 #define ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */ 298 #define ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */ 299 #define ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */ 300 #define ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */ 301 #define ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */ 302 #define ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */ 303 304 /** 305 * A structure used to retrieve link-level information of an Ethernet port. 306 */ 307 __extension__ 308 struct rte_eth_link { 309 uint32_t link_speed; /**< ETH_SPEED_NUM_ */ 310 uint16_t link_duplex : 1; /**< ETH_LINK_[HALF/FULL]_DUPLEX */ 311 uint16_t link_autoneg : 1; /**< ETH_LINK_[AUTONEG/FIXED] */ 312 uint16_t link_status : 1; /**< ETH_LINK_[DOWN/UP] */ 313 } __attribute__((aligned(8))); /**< aligned for atomic64 read/write */ 314 315 /* Utility constants */ 316 #define ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */ 317 #define ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */ 318 #define ETH_LINK_DOWN 0 /**< Link is down (see link_status). */ 319 #define ETH_LINK_UP 1 /**< Link is up (see link_status). */ 320 #define ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */ 321 #define ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */ 322 323 /** 324 * A structure used to configure the ring threshold registers of an RX/TX 325 * queue for an Ethernet port. 326 */ 327 struct rte_eth_thresh { 328 uint8_t pthresh; /**< Ring prefetch threshold. */ 329 uint8_t hthresh; /**< Ring host threshold. */ 330 uint8_t wthresh; /**< Ring writeback threshold. */ 331 }; 332 333 /** 334 * Simple flags are used for rte_eth_conf.rxmode.mq_mode. 335 */ 336 #define ETH_MQ_RX_RSS_FLAG 0x1 337 #define ETH_MQ_RX_DCB_FLAG 0x2 338 #define ETH_MQ_RX_VMDQ_FLAG 0x4 339 340 /** 341 * A set of values to identify what method is to be used to route 342 * packets to multiple queues. 343 */ 344 enum rte_eth_rx_mq_mode { 345 /** None of DCB,RSS or VMDQ mode */ 346 ETH_MQ_RX_NONE = 0, 347 348 /** For RX side, only RSS is on */ 349 ETH_MQ_RX_RSS = ETH_MQ_RX_RSS_FLAG, 350 /** For RX side,only DCB is on. */ 351 ETH_MQ_RX_DCB = ETH_MQ_RX_DCB_FLAG, 352 /** Both DCB and RSS enable */ 353 ETH_MQ_RX_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG, 354 355 /** Only VMDQ, no RSS nor DCB */ 356 ETH_MQ_RX_VMDQ_ONLY = ETH_MQ_RX_VMDQ_FLAG, 357 /** RSS mode with VMDQ */ 358 ETH_MQ_RX_VMDQ_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_VMDQ_FLAG, 359 /** Use VMDQ+DCB to route traffic to queues */ 360 ETH_MQ_RX_VMDQ_DCB = ETH_MQ_RX_VMDQ_FLAG | ETH_MQ_RX_DCB_FLAG, 361 /** Enable both VMDQ and DCB in VMDq */ 362 ETH_MQ_RX_VMDQ_DCB_RSS = ETH_MQ_RX_RSS_FLAG | ETH_MQ_RX_DCB_FLAG | 363 ETH_MQ_RX_VMDQ_FLAG, 364 }; 365 366 /** 367 * for rx mq mode backward compatible 368 */ 369 #define ETH_RSS ETH_MQ_RX_RSS 370 #define VMDQ_DCB ETH_MQ_RX_VMDQ_DCB 371 #define ETH_DCB_RX ETH_MQ_RX_DCB 372 373 /** 374 * A set of values to identify what method is to be used to transmit 375 * packets using multi-TCs. 376 */ 377 enum rte_eth_tx_mq_mode { 378 ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */ 379 ETH_MQ_TX_DCB, /**< For TX side,only DCB is on. */ 380 ETH_MQ_TX_VMDQ_DCB, /**< For TX side,both DCB and VT is on. */ 381 ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */ 382 }; 383 384 /** 385 * for tx mq mode backward compatible 386 */ 387 #define ETH_DCB_NONE ETH_MQ_TX_NONE 388 #define ETH_VMDQ_DCB_TX ETH_MQ_TX_VMDQ_DCB 389 #define ETH_DCB_TX ETH_MQ_TX_DCB 390 391 /** 392 * A structure used to configure the RX features of an Ethernet port. 393 */ 394 struct rte_eth_rxmode { 395 /** The multi-queue packet distribution mode to be used, e.g. RSS. */ 396 enum rte_eth_rx_mq_mode mq_mode; 397 uint32_t max_rx_pkt_len; /**< Only used if JUMBO_FRAME enabled. */ 398 /** Maximum allowed size of LRO aggregated packet. */ 399 uint32_t max_lro_pkt_size; 400 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/ 401 /** 402 * Per-port Rx offloads to be set using DEV_RX_OFFLOAD_* flags. 403 * Only offloads set on rx_offload_capa field on rte_eth_dev_info 404 * structure are allowed to be set. 405 */ 406 uint64_t offloads; 407 408 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 409 void *reserved_ptrs[2]; /**< Reserved for future fields */ 410 }; 411 412 /** 413 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN. 414 * Note that single VLAN is treated the same as inner VLAN. 415 */ 416 enum rte_vlan_type { 417 ETH_VLAN_TYPE_UNKNOWN = 0, 418 ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */ 419 ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */ 420 ETH_VLAN_TYPE_MAX, 421 }; 422 423 /** 424 * A structure used to describe a vlan filter. 425 * If the bit corresponding to a VID is set, such VID is on. 426 */ 427 struct rte_vlan_filter_conf { 428 uint64_t ids[64]; 429 }; 430 431 /** 432 * A structure used to configure the Receive Side Scaling (RSS) feature 433 * of an Ethernet port. 434 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points 435 * to an array holding the RSS key to use for hashing specific header 436 * fields of received packets. The length of this array should be indicated 437 * by *rss_key_len* below. Otherwise, a default random hash key is used by 438 * the device driver. 439 * 440 * The *rss_key_len* field of the *rss_conf* structure indicates the length 441 * in bytes of the array pointed by *rss_key*. To be compatible, this length 442 * will be checked in i40e only. Others assume 40 bytes to be used as before. 443 * 444 * The *rss_hf* field of the *rss_conf* structure indicates the different 445 * types of IPv4/IPv6 packets to which the RSS hashing must be applied. 446 * Supplying an *rss_hf* equal to zero disables the RSS feature. 447 */ 448 struct rte_eth_rss_conf { 449 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */ 450 uint8_t rss_key_len; /**< hash key length in bytes. */ 451 uint64_t rss_hf; /**< Hash functions to apply - see below. */ 452 }; 453 454 /* 455 * A packet can be identified by hardware as different flow types. Different 456 * NIC hardware may support different flow types. 457 * Basically, the NIC hardware identifies the flow type as deep protocol as 458 * possible, and exclusively. For example, if a packet is identified as 459 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types, 460 * though it is an actual IPV4 packet. 461 */ 462 #define RTE_ETH_FLOW_UNKNOWN 0 463 #define RTE_ETH_FLOW_RAW 1 464 #define RTE_ETH_FLOW_IPV4 2 465 #define RTE_ETH_FLOW_FRAG_IPV4 3 466 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4 467 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5 468 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6 469 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7 470 #define RTE_ETH_FLOW_IPV6 8 471 #define RTE_ETH_FLOW_FRAG_IPV6 9 472 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10 473 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11 474 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12 475 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13 476 #define RTE_ETH_FLOW_L2_PAYLOAD 14 477 #define RTE_ETH_FLOW_IPV6_EX 15 478 #define RTE_ETH_FLOW_IPV6_TCP_EX 16 479 #define RTE_ETH_FLOW_IPV6_UDP_EX 17 480 #define RTE_ETH_FLOW_PORT 18 481 /**< Consider device port number as a flow differentiator */ 482 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */ 483 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */ 484 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */ 485 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */ 486 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */ 487 #define RTE_ETH_FLOW_MAX 24 488 489 /* 490 * Below macros are defined for RSS offload types, they can be used to 491 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types. 492 */ 493 #define ETH_RSS_IPV4 (1ULL << 2) 494 #define ETH_RSS_FRAG_IPV4 (1ULL << 3) 495 #define ETH_RSS_NONFRAG_IPV4_TCP (1ULL << 4) 496 #define ETH_RSS_NONFRAG_IPV4_UDP (1ULL << 5) 497 #define ETH_RSS_NONFRAG_IPV4_SCTP (1ULL << 6) 498 #define ETH_RSS_NONFRAG_IPV4_OTHER (1ULL << 7) 499 #define ETH_RSS_IPV6 (1ULL << 8) 500 #define ETH_RSS_FRAG_IPV6 (1ULL << 9) 501 #define ETH_RSS_NONFRAG_IPV6_TCP (1ULL << 10) 502 #define ETH_RSS_NONFRAG_IPV6_UDP (1ULL << 11) 503 #define ETH_RSS_NONFRAG_IPV6_SCTP (1ULL << 12) 504 #define ETH_RSS_NONFRAG_IPV6_OTHER (1ULL << 13) 505 #define ETH_RSS_L2_PAYLOAD (1ULL << 14) 506 #define ETH_RSS_IPV6_EX (1ULL << 15) 507 #define ETH_RSS_IPV6_TCP_EX (1ULL << 16) 508 #define ETH_RSS_IPV6_UDP_EX (1ULL << 17) 509 #define ETH_RSS_PORT (1ULL << 18) 510 #define ETH_RSS_VXLAN (1ULL << 19) 511 #define ETH_RSS_GENEVE (1ULL << 20) 512 #define ETH_RSS_NVGRE (1ULL << 21) 513 #define ETH_RSS_GTPU (1ULL << 23) 514 515 /* 516 * We use the following macros to combine with above ETH_RSS_* for 517 * more specific input set selection. These bits are defined starting 518 * from the high end of the 64 bits. 519 * Note: If we use above ETH_RSS_* without SRC/DST_ONLY, it represents 520 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of 521 * the same level are used simultaneously, it is the same case as none of 522 * them are added. 523 */ 524 #define ETH_RSS_L3_SRC_ONLY (1ULL << 63) 525 #define ETH_RSS_L3_DST_ONLY (1ULL << 62) 526 #define ETH_RSS_L4_SRC_ONLY (1ULL << 61) 527 #define ETH_RSS_L4_DST_ONLY (1ULL << 60) 528 529 /** 530 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of 531 * the same level are used simultaneously, it is the same case as 532 * none of them are added. 533 * 534 * @param rss_hf 535 * RSS types with SRC/DST_ONLY. 536 * @return 537 * RSS types. 538 */ 539 static inline uint64_t 540 rte_eth_rss_hf_refine(uint64_t rss_hf) 541 { 542 if ((rss_hf & ETH_RSS_L3_SRC_ONLY) && (rss_hf & ETH_RSS_L3_DST_ONLY)) 543 rss_hf &= ~(ETH_RSS_L3_SRC_ONLY | ETH_RSS_L3_DST_ONLY); 544 545 if ((rss_hf & ETH_RSS_L4_SRC_ONLY) && (rss_hf & ETH_RSS_L4_DST_ONLY)) 546 rss_hf &= ~(ETH_RSS_L4_SRC_ONLY | ETH_RSS_L4_DST_ONLY); 547 548 return rss_hf; 549 } 550 551 #define ETH_RSS_IP ( \ 552 ETH_RSS_IPV4 | \ 553 ETH_RSS_FRAG_IPV4 | \ 554 ETH_RSS_NONFRAG_IPV4_OTHER | \ 555 ETH_RSS_IPV6 | \ 556 ETH_RSS_FRAG_IPV6 | \ 557 ETH_RSS_NONFRAG_IPV6_OTHER | \ 558 ETH_RSS_IPV6_EX) 559 560 #define ETH_RSS_UDP ( \ 561 ETH_RSS_NONFRAG_IPV4_UDP | \ 562 ETH_RSS_NONFRAG_IPV6_UDP | \ 563 ETH_RSS_IPV6_UDP_EX) 564 565 #define ETH_RSS_TCP ( \ 566 ETH_RSS_NONFRAG_IPV4_TCP | \ 567 ETH_RSS_NONFRAG_IPV6_TCP | \ 568 ETH_RSS_IPV6_TCP_EX) 569 570 #define ETH_RSS_SCTP ( \ 571 ETH_RSS_NONFRAG_IPV4_SCTP | \ 572 ETH_RSS_NONFRAG_IPV6_SCTP) 573 574 #define ETH_RSS_TUNNEL ( \ 575 ETH_RSS_VXLAN | \ 576 ETH_RSS_GENEVE | \ 577 ETH_RSS_NVGRE) 578 579 /**< Mask of valid RSS hash protocols */ 580 #define ETH_RSS_PROTO_MASK ( \ 581 ETH_RSS_IPV4 | \ 582 ETH_RSS_FRAG_IPV4 | \ 583 ETH_RSS_NONFRAG_IPV4_TCP | \ 584 ETH_RSS_NONFRAG_IPV4_UDP | \ 585 ETH_RSS_NONFRAG_IPV4_SCTP | \ 586 ETH_RSS_NONFRAG_IPV4_OTHER | \ 587 ETH_RSS_IPV6 | \ 588 ETH_RSS_FRAG_IPV6 | \ 589 ETH_RSS_NONFRAG_IPV6_TCP | \ 590 ETH_RSS_NONFRAG_IPV6_UDP | \ 591 ETH_RSS_NONFRAG_IPV6_SCTP | \ 592 ETH_RSS_NONFRAG_IPV6_OTHER | \ 593 ETH_RSS_L2_PAYLOAD | \ 594 ETH_RSS_IPV6_EX | \ 595 ETH_RSS_IPV6_TCP_EX | \ 596 ETH_RSS_IPV6_UDP_EX | \ 597 ETH_RSS_PORT | \ 598 ETH_RSS_VXLAN | \ 599 ETH_RSS_GENEVE | \ 600 ETH_RSS_NVGRE) 601 602 /* 603 * Definitions used for redirection table entry size. 604 * Some RSS RETA sizes may not be supported by some drivers, check the 605 * documentation or the description of relevant functions for more details. 606 */ 607 #define ETH_RSS_RETA_SIZE_64 64 608 #define ETH_RSS_RETA_SIZE_128 128 609 #define ETH_RSS_RETA_SIZE_256 256 610 #define ETH_RSS_RETA_SIZE_512 512 611 #define RTE_RETA_GROUP_SIZE 64 612 613 /* Definitions used for VMDQ and DCB functionality */ 614 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */ 615 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */ 616 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */ 617 #define ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */ 618 619 /* DCB capability defines */ 620 #define ETH_DCB_PG_SUPPORT 0x00000001 /**< Priority Group(ETS) support. */ 621 #define ETH_DCB_PFC_SUPPORT 0x00000002 /**< Priority Flow Control support. */ 622 623 /* Definitions used for VLAN Offload functionality */ 624 #define ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */ 625 #define ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */ 626 #define ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */ 627 #define ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */ 628 629 /* Definitions used for mask VLAN setting */ 630 #define ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */ 631 #define ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/ 632 #define ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/ 633 #define ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */ 634 #define ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/ 635 636 /* Definitions used for receive MAC address */ 637 #define ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */ 638 639 /* Definitions used for unicast hash */ 640 #define ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */ 641 642 /* Definitions used for VMDQ pool rx mode setting */ 643 #define ETH_VMDQ_ACCEPT_UNTAG 0x0001 /**< accept untagged packets. */ 644 #define ETH_VMDQ_ACCEPT_HASH_MC 0x0002 /**< accept packets in multicast table . */ 645 #define ETH_VMDQ_ACCEPT_HASH_UC 0x0004 /**< accept packets in unicast table. */ 646 #define ETH_VMDQ_ACCEPT_BROADCAST 0x0008 /**< accept broadcast packets. */ 647 #define ETH_VMDQ_ACCEPT_MULTICAST 0x0010 /**< multicast promiscuous. */ 648 649 /** Maximum nb. of vlan per mirror rule */ 650 #define ETH_MIRROR_MAX_VLANS 64 651 652 #define ETH_MIRROR_VIRTUAL_POOL_UP 0x01 /**< Virtual Pool uplink Mirroring. */ 653 #define ETH_MIRROR_UPLINK_PORT 0x02 /**< Uplink Port Mirroring. */ 654 #define ETH_MIRROR_DOWNLINK_PORT 0x04 /**< Downlink Port Mirroring. */ 655 #define ETH_MIRROR_VLAN 0x08 /**< VLAN Mirroring. */ 656 #define ETH_MIRROR_VIRTUAL_POOL_DOWN 0x10 /**< Virtual Pool downlink Mirroring. */ 657 658 /** 659 * A structure used to configure VLAN traffic mirror of an Ethernet port. 660 */ 661 struct rte_eth_vlan_mirror { 662 uint64_t vlan_mask; /**< mask for valid VLAN ID. */ 663 /** VLAN ID list for vlan mirroring. */ 664 uint16_t vlan_id[ETH_MIRROR_MAX_VLANS]; 665 }; 666 667 /** 668 * A structure used to configure traffic mirror of an Ethernet port. 669 */ 670 struct rte_eth_mirror_conf { 671 uint8_t rule_type; /**< Mirroring rule type */ 672 uint8_t dst_pool; /**< Destination pool for this mirror rule. */ 673 uint64_t pool_mask; /**< Bitmap of pool for pool mirroring */ 674 /** VLAN ID setting for VLAN mirroring. */ 675 struct rte_eth_vlan_mirror vlan; 676 }; 677 678 /** 679 * A structure used to configure 64 entries of Redirection Table of the 680 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure 681 * more than 64 entries supported by hardware, an array of this structure 682 * is needed. 683 */ 684 struct rte_eth_rss_reta_entry64 { 685 uint64_t mask; 686 /**< Mask bits indicate which entries need to be updated/queried. */ 687 uint16_t reta[RTE_RETA_GROUP_SIZE]; 688 /**< Group of 64 redirection table entries. */ 689 }; 690 691 /** 692 * This enum indicates the possible number of traffic classes 693 * in DCB configurations 694 */ 695 enum rte_eth_nb_tcs { 696 ETH_4_TCS = 4, /**< 4 TCs with DCB. */ 697 ETH_8_TCS = 8 /**< 8 TCs with DCB. */ 698 }; 699 700 /** 701 * This enum indicates the possible number of queue pools 702 * in VMDQ configurations. 703 */ 704 enum rte_eth_nb_pools { 705 ETH_8_POOLS = 8, /**< 8 VMDq pools. */ 706 ETH_16_POOLS = 16, /**< 16 VMDq pools. */ 707 ETH_32_POOLS = 32, /**< 32 VMDq pools. */ 708 ETH_64_POOLS = 64 /**< 64 VMDq pools. */ 709 }; 710 711 /* This structure may be extended in future. */ 712 struct rte_eth_dcb_rx_conf { 713 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */ 714 /** Traffic class each UP mapped to. */ 715 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES]; 716 }; 717 718 struct rte_eth_vmdq_dcb_tx_conf { 719 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */ 720 /** Traffic class each UP mapped to. */ 721 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES]; 722 }; 723 724 struct rte_eth_dcb_tx_conf { 725 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */ 726 /** Traffic class each UP mapped to. */ 727 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES]; 728 }; 729 730 struct rte_eth_vmdq_tx_conf { 731 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */ 732 }; 733 734 /** 735 * A structure used to configure the VMDQ+DCB feature 736 * of an Ethernet port. 737 * 738 * Using this feature, packets are routed to a pool of queues, based 739 * on the vlan id in the vlan tag, and then to a specific queue within 740 * that pool, using the user priority vlan tag field. 741 * 742 * A default pool may be used, if desired, to route all traffic which 743 * does not match the vlan filter rules. 744 */ 745 struct rte_eth_vmdq_dcb_conf { 746 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */ 747 uint8_t enable_default_pool; /**< If non-zero, use a default pool */ 748 uint8_t default_pool; /**< The default pool, if applicable */ 749 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */ 750 struct { 751 uint16_t vlan_id; /**< The vlan id of the received frame */ 752 uint64_t pools; /**< Bitmask of pools for packet rx */ 753 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */ 754 uint8_t dcb_tc[ETH_DCB_NUM_USER_PRIORITIES]; 755 /**< Selects a queue in a pool */ 756 }; 757 758 /** 759 * A structure used to configure the VMDQ feature of an Ethernet port when 760 * not combined with the DCB feature. 761 * 762 * Using this feature, packets are routed to a pool of queues. By default, 763 * the pool selection is based on the MAC address, the vlan id in the 764 * vlan tag as specified in the pool_map array. 765 * Passing the ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool 766 * selection using only the MAC address. MAC address to pool mapping is done 767 * using the rte_eth_dev_mac_addr_add function, with the pool parameter 768 * corresponding to the pool id. 769 * 770 * Queue selection within the selected pool will be done using RSS when 771 * it is enabled or revert to the first queue of the pool if not. 772 * 773 * A default pool may be used, if desired, to route all traffic which 774 * does not match the vlan filter rules or any pool MAC address. 775 */ 776 struct rte_eth_vmdq_rx_conf { 777 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */ 778 uint8_t enable_default_pool; /**< If non-zero, use a default pool */ 779 uint8_t default_pool; /**< The default pool, if applicable */ 780 uint8_t enable_loop_back; /**< Enable VT loop back */ 781 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */ 782 uint32_t rx_mode; /**< Flags from ETH_VMDQ_ACCEPT_* */ 783 struct { 784 uint16_t vlan_id; /**< The vlan id of the received frame */ 785 uint64_t pools; /**< Bitmask of pools for packet rx */ 786 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */ 787 }; 788 789 /** 790 * A structure used to configure the TX features of an Ethernet port. 791 */ 792 struct rte_eth_txmode { 793 enum rte_eth_tx_mq_mode mq_mode; /**< TX multi-queues mode. */ 794 /** 795 * Per-port Tx offloads to be set using DEV_TX_OFFLOAD_* flags. 796 * Only offloads set on tx_offload_capa field on rte_eth_dev_info 797 * structure are allowed to be set. 798 */ 799 uint64_t offloads; 800 801 /* For i40e specifically */ 802 uint16_t pvid; 803 __extension__ 804 uint8_t hw_vlan_reject_tagged : 1, 805 /**< If set, reject sending out tagged pkts */ 806 hw_vlan_reject_untagged : 1, 807 /**< If set, reject sending out untagged pkts */ 808 hw_vlan_insert_pvid : 1; 809 /**< If set, enable port based VLAN insertion */ 810 811 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 812 void *reserved_ptrs[2]; /**< Reserved for future fields */ 813 }; 814 815 /** 816 * A structure used to configure an RX ring of an Ethernet port. 817 */ 818 struct rte_eth_rxconf { 819 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */ 820 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */ 821 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */ 822 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */ 823 /** 824 * Per-queue Rx offloads to be set using DEV_RX_OFFLOAD_* flags. 825 * Only offloads set on rx_queue_offload_capa or rx_offload_capa 826 * fields on rte_eth_dev_info structure are allowed to be set. 827 */ 828 uint64_t offloads; 829 830 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 831 void *reserved_ptrs[2]; /**< Reserved for future fields */ 832 }; 833 834 /** 835 * A structure used to configure a TX ring of an Ethernet port. 836 */ 837 struct rte_eth_txconf { 838 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */ 839 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */ 840 uint16_t tx_free_thresh; /**< Start freeing TX buffers if there are 841 less free descriptors than this value. */ 842 843 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */ 844 /** 845 * Per-queue Tx offloads to be set using DEV_TX_OFFLOAD_* flags. 846 * Only offloads set on tx_queue_offload_capa or tx_offload_capa 847 * fields on rte_eth_dev_info structure are allowed to be set. 848 */ 849 uint64_t offloads; 850 851 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 852 void *reserved_ptrs[2]; /**< Reserved for future fields */ 853 }; 854 855 /** 856 * @warning 857 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 858 * 859 * A structure used to return the hairpin capabilities that are supported. 860 */ 861 struct rte_eth_hairpin_cap { 862 /** The max number of hairpin queues (different bindings). */ 863 uint16_t max_nb_queues; 864 /** Max number of Rx queues to be connected to one Tx queue. */ 865 uint16_t max_rx_2_tx; 866 /** Max number of Tx queues to be connected to one Rx queue. */ 867 uint16_t max_tx_2_rx; 868 uint16_t max_nb_desc; /**< The max num of descriptors. */ 869 }; 870 871 #define RTE_ETH_MAX_HAIRPIN_PEERS 32 872 873 /** 874 * @warning 875 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 876 * 877 * A structure used to hold hairpin peer data. 878 */ 879 struct rte_eth_hairpin_peer { 880 uint16_t port; /**< Peer port. */ 881 uint16_t queue; /**< Peer queue. */ 882 }; 883 884 /** 885 * @warning 886 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 887 * 888 * A structure used to configure hairpin binding. 889 */ 890 struct rte_eth_hairpin_conf { 891 uint16_t peer_count; /**< The number of peers. */ 892 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS]; 893 }; 894 895 /** 896 * A structure contains information about HW descriptor ring limitations. 897 */ 898 struct rte_eth_desc_lim { 899 uint16_t nb_max; /**< Max allowed number of descriptors. */ 900 uint16_t nb_min; /**< Min allowed number of descriptors. */ 901 uint16_t nb_align; /**< Number of descriptors should be aligned to. */ 902 903 /** 904 * Max allowed number of segments per whole packet. 905 * 906 * - For TSO packet this is the total number of data descriptors allowed 907 * by device. 908 * 909 * @see nb_mtu_seg_max 910 */ 911 uint16_t nb_seg_max; 912 913 /** 914 * Max number of segments per one MTU. 915 * 916 * - For non-TSO packet, this is the maximum allowed number of segments 917 * in a single transmit packet. 918 * 919 * - For TSO packet each segment within the TSO may span up to this 920 * value. 921 * 922 * @see nb_seg_max 923 */ 924 uint16_t nb_mtu_seg_max; 925 }; 926 927 /** 928 * This enum indicates the flow control mode 929 */ 930 enum rte_eth_fc_mode { 931 RTE_FC_NONE = 0, /**< Disable flow control. */ 932 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */ 933 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */ 934 RTE_FC_FULL /**< Enable flow control on both side. */ 935 }; 936 937 /** 938 * A structure used to configure Ethernet flow control parameter. 939 * These parameters will be configured into the register of the NIC. 940 * Please refer to the corresponding data sheet for proper value. 941 */ 942 struct rte_eth_fc_conf { 943 uint32_t high_water; /**< High threshold value to trigger XOFF */ 944 uint32_t low_water; /**< Low threshold value to trigger XON */ 945 uint16_t pause_time; /**< Pause quota in the Pause frame */ 946 uint16_t send_xon; /**< Is XON frame need be sent */ 947 enum rte_eth_fc_mode mode; /**< Link flow control mode */ 948 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */ 949 uint8_t autoneg; /**< Use Pause autoneg */ 950 }; 951 952 /** 953 * A structure used to configure Ethernet priority flow control parameter. 954 * These parameters will be configured into the register of the NIC. 955 * Please refer to the corresponding data sheet for proper value. 956 */ 957 struct rte_eth_pfc_conf { 958 struct rte_eth_fc_conf fc; /**< General flow control parameter. */ 959 uint8_t priority; /**< VLAN User Priority. */ 960 }; 961 962 /** 963 * Tunneled type. 964 */ 965 enum rte_eth_tunnel_type { 966 RTE_TUNNEL_TYPE_NONE = 0, 967 RTE_TUNNEL_TYPE_VXLAN, 968 RTE_TUNNEL_TYPE_GENEVE, 969 RTE_TUNNEL_TYPE_TEREDO, 970 RTE_TUNNEL_TYPE_NVGRE, 971 RTE_TUNNEL_TYPE_IP_IN_GRE, 972 RTE_L2_TUNNEL_TYPE_E_TAG, 973 RTE_TUNNEL_TYPE_VXLAN_GPE, 974 RTE_TUNNEL_TYPE_MAX, 975 }; 976 977 /* Deprecated API file for rte_eth_dev_filter_* functions */ 978 #include "rte_eth_ctrl.h" 979 980 /** 981 * Memory space that can be configured to store Flow Director filters 982 * in the board memory. 983 */ 984 enum rte_fdir_pballoc_type { 985 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */ 986 RTE_FDIR_PBALLOC_128K, /**< 128k. */ 987 RTE_FDIR_PBALLOC_256K, /**< 256k. */ 988 }; 989 990 /** 991 * Select report mode of FDIR hash information in RX descriptors. 992 */ 993 enum rte_fdir_status_mode { 994 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */ 995 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */ 996 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */ 997 }; 998 999 /** 1000 * A structure used to configure the Flow Director (FDIR) feature 1001 * of an Ethernet port. 1002 * 1003 * If mode is RTE_FDIR_MODE_NONE, the pballoc value is ignored. 1004 */ 1005 struct rte_fdir_conf { 1006 enum rte_fdir_mode mode; /**< Flow Director mode. */ 1007 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */ 1008 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */ 1009 /** RX queue of packets matching a "drop" filter in perfect mode. */ 1010 uint8_t drop_queue; 1011 struct rte_eth_fdir_masks mask; 1012 struct rte_eth_fdir_flex_conf flex_conf; 1013 /**< Flex payload configuration. */ 1014 }; 1015 1016 /** 1017 * UDP tunneling configuration. 1018 * Used to config the UDP port for a type of tunnel. 1019 * NICs need the UDP port to identify the tunnel type. 1020 * Normally a type of tunnel has a default UDP port, this structure can be used 1021 * in case if the users want to change or support more UDP port. 1022 */ 1023 struct rte_eth_udp_tunnel { 1024 uint16_t udp_port; /**< UDP port used for the tunnel. */ 1025 uint8_t prot_type; /**< Tunnel type. Defined in rte_eth_tunnel_type. */ 1026 }; 1027 1028 /** 1029 * A structure used to enable/disable specific device interrupts. 1030 */ 1031 struct rte_intr_conf { 1032 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */ 1033 uint32_t lsc:1; 1034 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */ 1035 uint32_t rxq:1; 1036 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */ 1037 uint32_t rmv:1; 1038 }; 1039 1040 /** 1041 * A structure used to configure an Ethernet port. 1042 * Depending upon the RX multi-queue mode, extra advanced 1043 * configuration settings may be needed. 1044 */ 1045 struct rte_eth_conf { 1046 uint32_t link_speeds; /**< bitmap of ETH_LINK_SPEED_XXX of speeds to be 1047 used. ETH_LINK_SPEED_FIXED disables link 1048 autonegotiation, and a unique speed shall be 1049 set. Otherwise, the bitmap defines the set of 1050 speeds to be advertised. If the special value 1051 ETH_LINK_SPEED_AUTONEG (0) is used, all speeds 1052 supported are advertised. */ 1053 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */ 1054 struct rte_eth_txmode txmode; /**< Port TX configuration. */ 1055 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value 1056 is 0, meaning the loopback mode is disabled. 1057 Read the datasheet of given ethernet controller 1058 for details. The possible values of this field 1059 are defined in implementation of each driver. */ 1060 struct { 1061 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */ 1062 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf; 1063 /**< Port vmdq+dcb configuration. */ 1064 struct rte_eth_dcb_rx_conf dcb_rx_conf; 1065 /**< Port dcb RX configuration. */ 1066 struct rte_eth_vmdq_rx_conf vmdq_rx_conf; 1067 /**< Port vmdq RX configuration. */ 1068 } rx_adv_conf; /**< Port RX filtering configuration. */ 1069 union { 1070 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf; 1071 /**< Port vmdq+dcb TX configuration. */ 1072 struct rte_eth_dcb_tx_conf dcb_tx_conf; 1073 /**< Port dcb TX configuration. */ 1074 struct rte_eth_vmdq_tx_conf vmdq_tx_conf; 1075 /**< Port vmdq TX configuration. */ 1076 } tx_adv_conf; /**< Port TX DCB configuration (union). */ 1077 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC 1078 is needed,and the variable must be set ETH_DCB_PFC_SUPPORT. */ 1079 uint32_t dcb_capability_en; 1080 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. DEPRECATED */ 1081 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */ 1082 }; 1083 1084 /** 1085 * RX offload capabilities of a device. 1086 */ 1087 #define DEV_RX_OFFLOAD_VLAN_STRIP 0x00000001 1088 #define DEV_RX_OFFLOAD_IPV4_CKSUM 0x00000002 1089 #define DEV_RX_OFFLOAD_UDP_CKSUM 0x00000004 1090 #define DEV_RX_OFFLOAD_TCP_CKSUM 0x00000008 1091 #define DEV_RX_OFFLOAD_TCP_LRO 0x00000010 1092 #define DEV_RX_OFFLOAD_QINQ_STRIP 0x00000020 1093 #define DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000040 1094 #define DEV_RX_OFFLOAD_MACSEC_STRIP 0x00000080 1095 #define DEV_RX_OFFLOAD_HEADER_SPLIT 0x00000100 1096 #define DEV_RX_OFFLOAD_VLAN_FILTER 0x00000200 1097 #define DEV_RX_OFFLOAD_VLAN_EXTEND 0x00000400 1098 #define DEV_RX_OFFLOAD_JUMBO_FRAME 0x00000800 1099 #define DEV_RX_OFFLOAD_SCATTER 0x00002000 1100 #define DEV_RX_OFFLOAD_TIMESTAMP 0x00004000 1101 #define DEV_RX_OFFLOAD_SECURITY 0x00008000 1102 #define DEV_RX_OFFLOAD_KEEP_CRC 0x00010000 1103 #define DEV_RX_OFFLOAD_SCTP_CKSUM 0x00020000 1104 #define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM 0x00040000 1105 #define DEV_RX_OFFLOAD_RSS_HASH 0x00080000 1106 1107 #define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \ 1108 DEV_RX_OFFLOAD_UDP_CKSUM | \ 1109 DEV_RX_OFFLOAD_TCP_CKSUM) 1110 #define DEV_RX_OFFLOAD_VLAN (DEV_RX_OFFLOAD_VLAN_STRIP | \ 1111 DEV_RX_OFFLOAD_VLAN_FILTER | \ 1112 DEV_RX_OFFLOAD_VLAN_EXTEND | \ 1113 DEV_RX_OFFLOAD_QINQ_STRIP) 1114 1115 /* 1116 * If new Rx offload capabilities are defined, they also must be 1117 * mentioned in rte_rx_offload_names in rte_ethdev.c file. 1118 */ 1119 1120 /** 1121 * TX offload capabilities of a device. 1122 */ 1123 #define DEV_TX_OFFLOAD_VLAN_INSERT 0x00000001 1124 #define DEV_TX_OFFLOAD_IPV4_CKSUM 0x00000002 1125 #define DEV_TX_OFFLOAD_UDP_CKSUM 0x00000004 1126 #define DEV_TX_OFFLOAD_TCP_CKSUM 0x00000008 1127 #define DEV_TX_OFFLOAD_SCTP_CKSUM 0x00000010 1128 #define DEV_TX_OFFLOAD_TCP_TSO 0x00000020 1129 #define DEV_TX_OFFLOAD_UDP_TSO 0x00000040 1130 #define DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM 0x00000080 /**< Used for tunneling packet. */ 1131 #define DEV_TX_OFFLOAD_QINQ_INSERT 0x00000100 1132 #define DEV_TX_OFFLOAD_VXLAN_TNL_TSO 0x00000200 /**< Used for tunneling packet. */ 1133 #define DEV_TX_OFFLOAD_GRE_TNL_TSO 0x00000400 /**< Used for tunneling packet. */ 1134 #define DEV_TX_OFFLOAD_IPIP_TNL_TSO 0x00000800 /**< Used for tunneling packet. */ 1135 #define DEV_TX_OFFLOAD_GENEVE_TNL_TSO 0x00001000 /**< Used for tunneling packet. */ 1136 #define DEV_TX_OFFLOAD_MACSEC_INSERT 0x00002000 1137 #define DEV_TX_OFFLOAD_MT_LOCKFREE 0x00004000 1138 /**< Multiple threads can invoke rte_eth_tx_burst() concurrently on the same 1139 * tx queue without SW lock. 1140 */ 1141 #define DEV_TX_OFFLOAD_MULTI_SEGS 0x00008000 1142 /**< Device supports multi segment send. */ 1143 #define DEV_TX_OFFLOAD_MBUF_FAST_FREE 0x00010000 1144 /**< Device supports optimization for fast release of mbufs. 1145 * When set application must guarantee that per-queue all mbufs comes from 1146 * the same mempool and has refcnt = 1. 1147 */ 1148 #define DEV_TX_OFFLOAD_SECURITY 0x00020000 1149 /** 1150 * Device supports generic UDP tunneled packet TSO. 1151 * Application must set PKT_TX_TUNNEL_UDP and other mbuf fields required 1152 * for tunnel TSO. 1153 */ 1154 #define DEV_TX_OFFLOAD_UDP_TNL_TSO 0x00040000 1155 /** 1156 * Device supports generic IP tunneled packet TSO. 1157 * Application must set PKT_TX_TUNNEL_IP and other mbuf fields required 1158 * for tunnel TSO. 1159 */ 1160 #define DEV_TX_OFFLOAD_IP_TNL_TSO 0x00080000 1161 /** Device supports outer UDP checksum */ 1162 #define DEV_TX_OFFLOAD_OUTER_UDP_CKSUM 0x00100000 1163 /* 1164 * If new Tx offload capabilities are defined, they also must be 1165 * mentioned in rte_tx_offload_names in rte_ethdev.c file. 1166 */ 1167 1168 /**@{@name Device capabilities 1169 * Non-offload capabilities reported in rte_eth_dev_info.dev_capa. 1170 */ 1171 /** Device supports Rx queue setup after device started. */ 1172 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP 0x00000001 1173 /** Device supports Tx queue setup after device started. */ 1174 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP 0x00000002 1175 /**@}*/ 1176 1177 /* 1178 * Fallback default preferred Rx/Tx port parameters. 1179 * These are used if an application requests default parameters 1180 * but the PMD does not provide preferred values. 1181 */ 1182 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512 1183 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512 1184 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1 1185 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1 1186 1187 /** 1188 * Preferred Rx/Tx port parameters. 1189 * There are separate instances of this structure for transmission 1190 * and reception respectively. 1191 */ 1192 struct rte_eth_dev_portconf { 1193 uint16_t burst_size; /**< Device-preferred burst size */ 1194 uint16_t ring_size; /**< Device-preferred size of queue rings */ 1195 uint16_t nb_queues; /**< Device-preferred number of queues */ 1196 }; 1197 1198 /** 1199 * Default values for switch domain id when ethdev does not support switch 1200 * domain definitions. 1201 */ 1202 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX) 1203 1204 /** 1205 * Ethernet device associated switch information 1206 */ 1207 struct rte_eth_switch_info { 1208 const char *name; /**< switch name */ 1209 uint16_t domain_id; /**< switch domain id */ 1210 uint16_t port_id; 1211 /**< 1212 * mapping to the devices physical switch port as enumerated from the 1213 * perspective of the embedded interconnect/switch. For SR-IOV enabled 1214 * device this may correspond to the VF_ID of each virtual function, 1215 * but each driver should explicitly define the mapping of switch 1216 * port identifier to that physical interconnect/switch 1217 */ 1218 }; 1219 1220 /** 1221 * Ethernet device information 1222 */ 1223 1224 /** 1225 * A structure used to retrieve the contextual information of 1226 * an Ethernet device, such as the controlling driver of the 1227 * device, etc... 1228 */ 1229 struct rte_eth_dev_info { 1230 struct rte_device *device; /** Generic device information */ 1231 const char *driver_name; /**< Device Driver name. */ 1232 unsigned int if_index; /**< Index to bound host interface, or 0 if none. 1233 Use if_indextoname() to translate into an interface name. */ 1234 uint16_t min_mtu; /**< Minimum MTU allowed */ 1235 uint16_t max_mtu; /**< Maximum MTU allowed */ 1236 const uint32_t *dev_flags; /**< Device flags */ 1237 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */ 1238 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */ 1239 /** Maximum configurable size of LRO aggregated packet. */ 1240 uint32_t max_lro_pkt_size; 1241 uint16_t max_rx_queues; /**< Maximum number of RX queues. */ 1242 uint16_t max_tx_queues; /**< Maximum number of TX queues. */ 1243 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */ 1244 uint32_t max_hash_mac_addrs; 1245 /** Maximum number of hash MAC addresses for MTA and UTA. */ 1246 uint16_t max_vfs; /**< Maximum number of VFs. */ 1247 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */ 1248 uint64_t rx_offload_capa; 1249 /**< All RX offload capabilities including all per-queue ones */ 1250 uint64_t tx_offload_capa; 1251 /**< All TX offload capabilities including all per-queue ones */ 1252 uint64_t rx_queue_offload_capa; 1253 /**< Device per-queue RX offload capabilities. */ 1254 uint64_t tx_queue_offload_capa; 1255 /**< Device per-queue TX offload capabilities. */ 1256 uint16_t reta_size; 1257 /**< Device redirection table size, the total number of entries. */ 1258 uint8_t hash_key_size; /**< Hash key size in bytes */ 1259 /** Bit mask of RSS offloads, the bit offset also means flow type */ 1260 uint64_t flow_type_rss_offloads; 1261 struct rte_eth_rxconf default_rxconf; /**< Default RX configuration */ 1262 struct rte_eth_txconf default_txconf; /**< Default TX configuration */ 1263 uint16_t vmdq_queue_base; /**< First queue ID for VMDQ pools. */ 1264 uint16_t vmdq_queue_num; /**< Queue number for VMDQ pools. */ 1265 uint16_t vmdq_pool_base; /**< First ID of VMDQ pools. */ 1266 struct rte_eth_desc_lim rx_desc_lim; /**< RX descriptors limits */ 1267 struct rte_eth_desc_lim tx_desc_lim; /**< TX descriptors limits */ 1268 uint32_t speed_capa; /**< Supported speeds bitmap (ETH_LINK_SPEED_). */ 1269 /** Configured number of rx/tx queues */ 1270 uint16_t nb_rx_queues; /**< Number of RX queues. */ 1271 uint16_t nb_tx_queues; /**< Number of TX queues. */ 1272 /** Rx parameter recommendations */ 1273 struct rte_eth_dev_portconf default_rxportconf; 1274 /** Tx parameter recommendations */ 1275 struct rte_eth_dev_portconf default_txportconf; 1276 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */ 1277 uint64_t dev_capa; 1278 /** 1279 * Switching information for ports on a device with a 1280 * embedded managed interconnect/switch. 1281 */ 1282 struct rte_eth_switch_info switch_info; 1283 1284 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 1285 void *reserved_ptrs[2]; /**< Reserved for future fields */ 1286 }; 1287 1288 /** 1289 * Ethernet device RX queue information structure. 1290 * Used to retrieve information about configured queue. 1291 */ 1292 struct rte_eth_rxq_info { 1293 struct rte_mempool *mp; /**< mempool used by that queue. */ 1294 struct rte_eth_rxconf conf; /**< queue config parameters. */ 1295 uint8_t scattered_rx; /**< scattered packets RX supported. */ 1296 uint16_t nb_desc; /**< configured number of RXDs. */ 1297 } __rte_cache_min_aligned; 1298 1299 /** 1300 * Ethernet device TX queue information structure. 1301 * Used to retrieve information about configured queue. 1302 */ 1303 struct rte_eth_txq_info { 1304 struct rte_eth_txconf conf; /**< queue config parameters. */ 1305 uint16_t nb_desc; /**< configured number of TXDs. */ 1306 } __rte_cache_min_aligned; 1307 1308 /* Generic Burst mode flag definition, values can be ORed. */ 1309 1310 /** 1311 * If the queues have different burst mode description, this bit will be set 1312 * by PMD, then the application can iterate to retrieve burst description for 1313 * all other queues. 1314 */ 1315 #define RTE_ETH_BURST_FLAG_PER_QUEUE (1ULL << 0) 1316 1317 /** 1318 * Ethernet device RX/TX queue packet burst mode information structure. 1319 * Used to retrieve information about packet burst mode setting. 1320 */ 1321 struct rte_eth_burst_mode { 1322 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */ 1323 1324 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */ 1325 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */ 1326 }; 1327 1328 /** Maximum name length for extended statistics counters */ 1329 #define RTE_ETH_XSTATS_NAME_SIZE 64 1330 1331 /** 1332 * An Ethernet device extended statistic structure 1333 * 1334 * This structure is used by rte_eth_xstats_get() to provide 1335 * statistics that are not provided in the generic *rte_eth_stats* 1336 * structure. 1337 * It maps a name id, corresponding to an index in the array returned 1338 * by rte_eth_xstats_get_names(), to a statistic value. 1339 */ 1340 struct rte_eth_xstat { 1341 uint64_t id; /**< The index in xstats name array. */ 1342 uint64_t value; /**< The statistic counter value. */ 1343 }; 1344 1345 /** 1346 * A name element for extended statistics. 1347 * 1348 * An array of this structure is returned by rte_eth_xstats_get_names(). 1349 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat* 1350 * structure references these names by their array index. 1351 */ 1352 struct rte_eth_xstat_name { 1353 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */ 1354 }; 1355 1356 #define ETH_DCB_NUM_TCS 8 1357 #define ETH_MAX_VMDQ_POOL 64 1358 1359 /** 1360 * A structure used to get the information of queue and 1361 * TC mapping on both TX and RX paths. 1362 */ 1363 struct rte_eth_dcb_tc_queue_mapping { 1364 /** rx queues assigned to tc per Pool */ 1365 struct { 1366 uint8_t base; 1367 uint8_t nb_queue; 1368 } tc_rxq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS]; 1369 /** rx queues assigned to tc per Pool */ 1370 struct { 1371 uint8_t base; 1372 uint8_t nb_queue; 1373 } tc_txq[ETH_MAX_VMDQ_POOL][ETH_DCB_NUM_TCS]; 1374 }; 1375 1376 /** 1377 * A structure used to get the information of DCB. 1378 * It includes TC UP mapping and queue TC mapping. 1379 */ 1380 struct rte_eth_dcb_info { 1381 uint8_t nb_tcs; /**< number of TCs */ 1382 uint8_t prio_tc[ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */ 1383 uint8_t tc_bws[ETH_DCB_NUM_TCS]; /**< TX BW percentage for each TC */ 1384 /** rx queues assigned to tc */ 1385 struct rte_eth_dcb_tc_queue_mapping tc_queue; 1386 }; 1387 1388 #define RTE_ETH_ALL RTE_MAX_ETHPORTS 1389 1390 /* Macros to check for valid port */ 1391 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \ 1392 if (!rte_eth_dev_is_valid_port(port_id)) { \ 1393 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \ 1394 return retval; \ 1395 } \ 1396 } while (0) 1397 1398 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \ 1399 if (!rte_eth_dev_is_valid_port(port_id)) { \ 1400 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \ 1401 return; \ 1402 } \ 1403 } while (0) 1404 1405 /** 1406 * l2 tunnel configuration. 1407 */ 1408 1409 /**< l2 tunnel enable mask */ 1410 #define ETH_L2_TUNNEL_ENABLE_MASK 0x00000001 1411 /**< l2 tunnel insertion mask */ 1412 #define ETH_L2_TUNNEL_INSERTION_MASK 0x00000002 1413 /**< l2 tunnel stripping mask */ 1414 #define ETH_L2_TUNNEL_STRIPPING_MASK 0x00000004 1415 /**< l2 tunnel forwarding mask */ 1416 #define ETH_L2_TUNNEL_FORWARDING_MASK 0x00000008 1417 1418 /** 1419 * Function type used for RX packet processing packet callbacks. 1420 * 1421 * The callback function is called on RX with a burst of packets that have 1422 * been received on the given port and queue. 1423 * 1424 * @param port_id 1425 * The Ethernet port on which RX is being performed. 1426 * @param queue 1427 * The queue on the Ethernet port which is being used to receive the packets. 1428 * @param pkts 1429 * The burst of packets that have just been received. 1430 * @param nb_pkts 1431 * The number of packets in the burst pointed to by "pkts". 1432 * @param max_pkts 1433 * The max number of packets that can be stored in the "pkts" array. 1434 * @param user_param 1435 * The arbitrary user parameter passed in by the application when the callback 1436 * was originally configured. 1437 * @return 1438 * The number of packets returned to the user. 1439 */ 1440 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue, 1441 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts, 1442 void *user_param); 1443 1444 /** 1445 * Function type used for TX packet processing packet callbacks. 1446 * 1447 * The callback function is called on TX with a burst of packets immediately 1448 * before the packets are put onto the hardware queue for transmission. 1449 * 1450 * @param port_id 1451 * The Ethernet port on which TX is being performed. 1452 * @param queue 1453 * The queue on the Ethernet port which is being used to transmit the packets. 1454 * @param pkts 1455 * The burst of packets that are about to be transmitted. 1456 * @param nb_pkts 1457 * The number of packets in the burst pointed to by "pkts". 1458 * @param user_param 1459 * The arbitrary user parameter passed in by the application when the callback 1460 * was originally configured. 1461 * @return 1462 * The number of packets to be written to the NIC. 1463 */ 1464 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue, 1465 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param); 1466 1467 /** 1468 * Possible states of an ethdev port. 1469 */ 1470 enum rte_eth_dev_state { 1471 /** Device is unused before being probed. */ 1472 RTE_ETH_DEV_UNUSED = 0, 1473 /** Device is attached when allocated in probing. */ 1474 RTE_ETH_DEV_ATTACHED, 1475 /** Device is in removed state when plug-out is detected. */ 1476 RTE_ETH_DEV_REMOVED, 1477 }; 1478 1479 struct rte_eth_dev_sriov { 1480 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */ 1481 uint8_t nb_q_per_pool; /**< rx queue number per pool */ 1482 uint16_t def_vmdq_idx; /**< Default pool num used for PF */ 1483 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */ 1484 }; 1485 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov) 1486 1487 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN 1488 1489 #define RTE_ETH_DEV_NO_OWNER 0 1490 1491 #define RTE_ETH_MAX_OWNER_NAME_LEN 64 1492 1493 struct rte_eth_dev_owner { 1494 uint64_t id; /**< The owner unique identifier. */ 1495 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */ 1496 }; 1497 1498 /** 1499 * Port is released (i.e. totally freed and data erased) on close. 1500 * Temporary flag for PMD migration to new rte_eth_dev_close() behaviour. 1501 */ 1502 #define RTE_ETH_DEV_CLOSE_REMOVE 0x0001 1503 /** Device supports link state interrupt */ 1504 #define RTE_ETH_DEV_INTR_LSC 0x0002 1505 /** Device is a bonded slave */ 1506 #define RTE_ETH_DEV_BONDED_SLAVE 0x0004 1507 /** Device supports device removal interrupt */ 1508 #define RTE_ETH_DEV_INTR_RMV 0x0008 1509 /** Device is port representor */ 1510 #define RTE_ETH_DEV_REPRESENTOR 0x0010 1511 /** Device does not support MAC change after started */ 1512 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR 0x0020 1513 1514 /** 1515 * Iterates over valid ethdev ports owned by a specific owner. 1516 * 1517 * @param port_id 1518 * The id of the next possible valid owned port. 1519 * @param owner_id 1520 * The owner identifier. 1521 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports. 1522 * @return 1523 * Next valid port id owned by owner_id, RTE_MAX_ETHPORTS if there is none. 1524 */ 1525 uint64_t rte_eth_find_next_owned_by(uint16_t port_id, 1526 const uint64_t owner_id); 1527 1528 /** 1529 * Macro to iterate over all enabled ethdev ports owned by a specific owner. 1530 */ 1531 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \ 1532 for (p = rte_eth_find_next_owned_by(0, o); \ 1533 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \ 1534 p = rte_eth_find_next_owned_by(p + 1, o)) 1535 1536 /** 1537 * Iterates over valid ethdev ports. 1538 * 1539 * @param port_id 1540 * The id of the next possible valid port. 1541 * @return 1542 * Next valid port id, RTE_MAX_ETHPORTS if there is none. 1543 */ 1544 uint16_t rte_eth_find_next(uint16_t port_id); 1545 1546 /** 1547 * Macro to iterate over all enabled and ownerless ethdev ports. 1548 */ 1549 #define RTE_ETH_FOREACH_DEV(p) \ 1550 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER) 1551 1552 /** 1553 * @warning 1554 * @b EXPERIMENTAL: this API may change without prior notice. 1555 * 1556 * Iterates over ethdev ports of a specified device. 1557 * 1558 * @param port_id_start 1559 * The id of the next possible valid port. 1560 * @param parent 1561 * The generic device behind the ports to iterate. 1562 * @return 1563 * Next port id of the device, possibly port_id_start, 1564 * RTE_MAX_ETHPORTS if there is none. 1565 */ 1566 __rte_experimental 1567 uint16_t 1568 rte_eth_find_next_of(uint16_t port_id_start, 1569 const struct rte_device *parent); 1570 1571 /** 1572 * Macro to iterate over all ethdev ports of a specified device. 1573 * 1574 * @param port_id 1575 * The id of the matching port being iterated. 1576 * @param parent 1577 * The rte_device pointer matching the iterated ports. 1578 */ 1579 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \ 1580 for (port_id = rte_eth_find_next_of(0, parent); \ 1581 port_id < RTE_MAX_ETHPORTS; \ 1582 port_id = rte_eth_find_next_of(port_id + 1, parent)) 1583 1584 /** 1585 * @warning 1586 * @b EXPERIMENTAL: this API may change without prior notice. 1587 * 1588 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device). 1589 * 1590 * @param port_id_start 1591 * The id of the next possible valid sibling port. 1592 * @param ref_port_id 1593 * The id of a reference port to compare rte_device with. 1594 * @return 1595 * Next sibling port id, possibly port_id_start or ref_port_id itself, 1596 * RTE_MAX_ETHPORTS if there is none. 1597 */ 1598 __rte_experimental 1599 uint16_t 1600 rte_eth_find_next_sibling(uint16_t port_id_start, 1601 uint16_t ref_port_id); 1602 1603 /** 1604 * Macro to iterate over all ethdev ports sharing the same rte_device 1605 * as the specified port. 1606 * Note: the specified reference port is part of the loop iterations. 1607 * 1608 * @param port_id 1609 * The id of the matching port being iterated. 1610 * @param ref_port_id 1611 * The id of the port being compared. 1612 */ 1613 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \ 1614 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \ 1615 port_id < RTE_MAX_ETHPORTS; \ 1616 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id)) 1617 1618 /** 1619 * @warning 1620 * @b EXPERIMENTAL: this API may change without prior notice. 1621 * 1622 * Get a new unique owner identifier. 1623 * An owner identifier is used to owns Ethernet devices by only one DPDK entity 1624 * to avoid multiple management of device by different entities. 1625 * 1626 * @param owner_id 1627 * Owner identifier pointer. 1628 * @return 1629 * Negative errno value on error, 0 on success. 1630 */ 1631 __rte_experimental 1632 int rte_eth_dev_owner_new(uint64_t *owner_id); 1633 1634 /** 1635 * @warning 1636 * @b EXPERIMENTAL: this API may change without prior notice. 1637 * 1638 * Set an Ethernet device owner. 1639 * 1640 * @param port_id 1641 * The identifier of the port to own. 1642 * @param owner 1643 * The owner pointer. 1644 * @return 1645 * Negative errno value on error, 0 on success. 1646 */ 1647 __rte_experimental 1648 int rte_eth_dev_owner_set(const uint16_t port_id, 1649 const struct rte_eth_dev_owner *owner); 1650 1651 /** 1652 * @warning 1653 * @b EXPERIMENTAL: this API may change without prior notice. 1654 * 1655 * Unset Ethernet device owner to make the device ownerless. 1656 * 1657 * @param port_id 1658 * The identifier of port to make ownerless. 1659 * @param owner_id 1660 * The owner identifier. 1661 * @return 1662 * 0 on success, negative errno value on error. 1663 */ 1664 __rte_experimental 1665 int rte_eth_dev_owner_unset(const uint16_t port_id, 1666 const uint64_t owner_id); 1667 1668 /** 1669 * @warning 1670 * @b EXPERIMENTAL: this API may change without prior notice. 1671 * 1672 * Remove owner from all Ethernet devices owned by a specific owner. 1673 * 1674 * @param owner_id 1675 * The owner identifier. 1676 * @return 1677 * 0 on success, negative errno value on error. 1678 */ 1679 __rte_experimental 1680 int rte_eth_dev_owner_delete(const uint64_t owner_id); 1681 1682 /** 1683 * @warning 1684 * @b EXPERIMENTAL: this API may change without prior notice. 1685 * 1686 * Get the owner of an Ethernet device. 1687 * 1688 * @param port_id 1689 * The port identifier. 1690 * @param owner 1691 * The owner structure pointer to fill. 1692 * @return 1693 * 0 on success, negative errno value on error.. 1694 */ 1695 __rte_experimental 1696 int rte_eth_dev_owner_get(const uint16_t port_id, 1697 struct rte_eth_dev_owner *owner); 1698 1699 /** 1700 * Get the number of ports which are usable for the application. 1701 * 1702 * These devices must be iterated by using the macro 1703 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY`` 1704 * to deal with non-contiguous ranges of devices. 1705 * 1706 * @return 1707 * The count of available Ethernet devices. 1708 */ 1709 uint16_t rte_eth_dev_count_avail(void); 1710 1711 /** 1712 * Get the total number of ports which are allocated. 1713 * 1714 * Some devices may not be available for the application. 1715 * 1716 * @return 1717 * The total count of Ethernet devices. 1718 */ 1719 uint16_t rte_eth_dev_count_total(void); 1720 1721 /** 1722 * Convert a numerical speed in Mbps to a bitmap flag that can be used in 1723 * the bitmap link_speeds of the struct rte_eth_conf 1724 * 1725 * @param speed 1726 * Numerical speed value in Mbps 1727 * @param duplex 1728 * ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds) 1729 * @return 1730 * 0 if the speed cannot be mapped 1731 */ 1732 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex); 1733 1734 /** 1735 * Get DEV_RX_OFFLOAD_* flag name. 1736 * 1737 * @param offload 1738 * Offload flag. 1739 * @return 1740 * Offload name or 'UNKNOWN' if the flag cannot be recognised. 1741 */ 1742 const char *rte_eth_dev_rx_offload_name(uint64_t offload); 1743 1744 /** 1745 * Get DEV_TX_OFFLOAD_* flag name. 1746 * 1747 * @param offload 1748 * Offload flag. 1749 * @return 1750 * Offload name or 'UNKNOWN' if the flag cannot be recognised. 1751 */ 1752 const char *rte_eth_dev_tx_offload_name(uint64_t offload); 1753 1754 /** 1755 * Configure an Ethernet device. 1756 * This function must be invoked first before any other function in the 1757 * Ethernet API. This function can also be re-invoked when a device is in the 1758 * stopped state. 1759 * 1760 * @param port_id 1761 * The port identifier of the Ethernet device to configure. 1762 * @param nb_rx_queue 1763 * The number of receive queues to set up for the Ethernet device. 1764 * @param nb_tx_queue 1765 * The number of transmit queues to set up for the Ethernet device. 1766 * @param eth_conf 1767 * The pointer to the configuration data to be used for the Ethernet device. 1768 * The *rte_eth_conf* structure includes: 1769 * - the hardware offload features to activate, with dedicated fields for 1770 * each statically configurable offload hardware feature provided by 1771 * Ethernet devices, such as IP checksum or VLAN tag stripping for 1772 * example. 1773 * The Rx offload bitfield API is obsolete and will be deprecated. 1774 * Applications should set the ignore_bitfield_offloads bit on *rxmode* 1775 * structure and use offloads field to set per-port offloads instead. 1776 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within 1777 * the [rt]x_offload_capa returned from rte_eth_dev_info_get(). 1778 * Any type of device supported offloading set in the input argument 1779 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled 1780 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup() 1781 * - the Receive Side Scaling (RSS) configuration when using multiple RX 1782 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf 1783 * must be within the flow_type_rss_offloads provided by drivers via 1784 * rte_eth_dev_info_get() API. 1785 * 1786 * Embedding all configuration information in a single data structure 1787 * is the more flexible method that allows the addition of new features 1788 * without changing the syntax of the API. 1789 * @return 1790 * - 0: Success, device configured. 1791 * - <0: Error code returned by the driver configuration function. 1792 */ 1793 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue, 1794 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf); 1795 1796 /** 1797 * @warning 1798 * @b EXPERIMENTAL: this API may change without prior notice. 1799 * 1800 * Check if an Ethernet device was physically removed. 1801 * 1802 * @param port_id 1803 * The port identifier of the Ethernet device. 1804 * @return 1805 * 1 when the Ethernet device is removed, otherwise 0. 1806 */ 1807 __rte_experimental 1808 int 1809 rte_eth_dev_is_removed(uint16_t port_id); 1810 1811 /** 1812 * Allocate and set up a receive queue for an Ethernet device. 1813 * 1814 * The function allocates a contiguous block of memory for *nb_rx_desc* 1815 * receive descriptors from a memory zone associated with *socket_id* 1816 * and initializes each receive descriptor with a network buffer allocated 1817 * from the memory pool *mb_pool*. 1818 * 1819 * @param port_id 1820 * The port identifier of the Ethernet device. 1821 * @param rx_queue_id 1822 * The index of the receive queue to set up. 1823 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 1824 * to rte_eth_dev_configure(). 1825 * @param nb_rx_desc 1826 * The number of receive descriptors to allocate for the receive ring. 1827 * @param socket_id 1828 * The *socket_id* argument is the socket identifier in case of NUMA. 1829 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for 1830 * the DMA memory allocated for the receive descriptors of the ring. 1831 * @param rx_conf 1832 * The pointer to the configuration data to be used for the receive queue. 1833 * NULL value is allowed, in which case default RX configuration 1834 * will be used. 1835 * The *rx_conf* structure contains an *rx_thresh* structure with the values 1836 * of the Prefetch, Host, and Write-Back threshold registers of the receive 1837 * ring. 1838 * In addition it contains the hardware offloads features to activate using 1839 * the DEV_RX_OFFLOAD_* flags. 1840 * If an offloading set in rx_conf->offloads 1841 * hasn't been set in the input argument eth_conf->rxmode.offloads 1842 * to rte_eth_dev_configure(), it is a new added offloading, it must be 1843 * per-queue type and it is enabled for the queue. 1844 * No need to repeat any bit in rx_conf->offloads which has already been 1845 * enabled in rte_eth_dev_configure() at port level. An offloading enabled 1846 * at port level can't be disabled at queue level. 1847 * @param mb_pool 1848 * The pointer to the memory pool from which to allocate *rte_mbuf* network 1849 * memory buffers to populate each descriptor of the receive ring. 1850 * @return 1851 * - 0: Success, receive queue correctly set up. 1852 * - -EIO: if device is removed. 1853 * - -EINVAL: The memory pool pointer is null or the size of network buffers 1854 * which can be allocated from this memory pool does not fit the various 1855 * buffer sizes allowed by the device controller. 1856 * - -ENOMEM: Unable to allocate the receive ring descriptors or to 1857 * allocate network memory buffers from the memory pool when 1858 * initializing receive descriptors. 1859 */ 1860 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id, 1861 uint16_t nb_rx_desc, unsigned int socket_id, 1862 const struct rte_eth_rxconf *rx_conf, 1863 struct rte_mempool *mb_pool); 1864 1865 /** 1866 * @warning 1867 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1868 * 1869 * Allocate and set up a hairpin receive queue for an Ethernet device. 1870 * 1871 * The function set up the selected queue to be used in hairpin. 1872 * 1873 * @param port_id 1874 * The port identifier of the Ethernet device. 1875 * @param rx_queue_id 1876 * The index of the receive queue to set up. 1877 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 1878 * to rte_eth_dev_configure(). 1879 * @param nb_rx_desc 1880 * The number of receive descriptors to allocate for the receive ring. 1881 * 0 means the PMD will use default value. 1882 * @param conf 1883 * The pointer to the hairpin configuration. 1884 * 1885 * @return 1886 * - (0) if successful. 1887 * - (-ENOTSUP) if hardware doesn't support. 1888 * - (-EINVAL) if bad parameter. 1889 * - (-ENOMEM) if unable to allocate the resources. 1890 */ 1891 __rte_experimental 1892 int rte_eth_rx_hairpin_queue_setup 1893 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc, 1894 const struct rte_eth_hairpin_conf *conf); 1895 1896 /** 1897 * Allocate and set up a transmit queue for an Ethernet device. 1898 * 1899 * @param port_id 1900 * The port identifier of the Ethernet device. 1901 * @param tx_queue_id 1902 * The index of the transmit queue to set up. 1903 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 1904 * to rte_eth_dev_configure(). 1905 * @param nb_tx_desc 1906 * The number of transmit descriptors to allocate for the transmit ring. 1907 * @param socket_id 1908 * The *socket_id* argument is the socket identifier in case of NUMA. 1909 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for 1910 * the DMA memory allocated for the transmit descriptors of the ring. 1911 * @param tx_conf 1912 * The pointer to the configuration data to be used for the transmit queue. 1913 * NULL value is allowed, in which case default TX configuration 1914 * will be used. 1915 * The *tx_conf* structure contains the following data: 1916 * - The *tx_thresh* structure with the values of the Prefetch, Host, and 1917 * Write-Back threshold registers of the transmit ring. 1918 * When setting Write-Back threshold to the value greater then zero, 1919 * *tx_rs_thresh* value should be explicitly set to one. 1920 * - The *tx_free_thresh* value indicates the [minimum] number of network 1921 * buffers that must be pending in the transmit ring to trigger their 1922 * [implicit] freeing by the driver transmit function. 1923 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit 1924 * descriptors that must be pending in the transmit ring before setting the 1925 * RS bit on a descriptor by the driver transmit function. 1926 * The *tx_rs_thresh* value should be less or equal then 1927 * *tx_free_thresh* value, and both of them should be less then 1928 * *nb_tx_desc* - 3. 1929 * - The *offloads* member contains Tx offloads to be enabled. 1930 * If an offloading set in tx_conf->offloads 1931 * hasn't been set in the input argument eth_conf->txmode.offloads 1932 * to rte_eth_dev_configure(), it is a new added offloading, it must be 1933 * per-queue type and it is enabled for the queue. 1934 * No need to repeat any bit in tx_conf->offloads which has already been 1935 * enabled in rte_eth_dev_configure() at port level. An offloading enabled 1936 * at port level can't be disabled at queue level. 1937 * 1938 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces 1939 * the transmit function to use default values. 1940 * @return 1941 * - 0: Success, the transmit queue is correctly set up. 1942 * - -ENOMEM: Unable to allocate the transmit ring descriptors. 1943 */ 1944 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id, 1945 uint16_t nb_tx_desc, unsigned int socket_id, 1946 const struct rte_eth_txconf *tx_conf); 1947 1948 /** 1949 * @warning 1950 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1951 * 1952 * Allocate and set up a transmit hairpin queue for an Ethernet device. 1953 * 1954 * @param port_id 1955 * The port identifier of the Ethernet device. 1956 * @param tx_queue_id 1957 * The index of the transmit queue to set up. 1958 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 1959 * to rte_eth_dev_configure(). 1960 * @param nb_tx_desc 1961 * The number of transmit descriptors to allocate for the transmit ring. 1962 * 0 to set default PMD value. 1963 * @param conf 1964 * The hairpin configuration. 1965 * 1966 * @return 1967 * - (0) if successful. 1968 * - (-ENOTSUP) if hardware doesn't support. 1969 * - (-EINVAL) if bad parameter. 1970 * - (-ENOMEM) if unable to allocate the resources. 1971 */ 1972 __rte_experimental 1973 int rte_eth_tx_hairpin_queue_setup 1974 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc, 1975 const struct rte_eth_hairpin_conf *conf); 1976 1977 /** 1978 * Return the NUMA socket to which an Ethernet device is connected 1979 * 1980 * @param port_id 1981 * The port identifier of the Ethernet device 1982 * @return 1983 * The NUMA socket id to which the Ethernet device is connected or 1984 * a default of zero if the socket could not be determined. 1985 * -1 is returned is the port_id value is out of range. 1986 */ 1987 int rte_eth_dev_socket_id(uint16_t port_id); 1988 1989 /** 1990 * Check if port_id of device is attached 1991 * 1992 * @param port_id 1993 * The port identifier of the Ethernet device 1994 * @return 1995 * - 0 if port is out of range or not attached 1996 * - 1 if device is attached 1997 */ 1998 int rte_eth_dev_is_valid_port(uint16_t port_id); 1999 2000 /** 2001 * Start specified RX queue of a port. It is used when rx_deferred_start 2002 * flag of the specified queue is true. 2003 * 2004 * @param port_id 2005 * The port identifier of the Ethernet device 2006 * @param rx_queue_id 2007 * The index of the rx queue to update the ring. 2008 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2009 * to rte_eth_dev_configure(). 2010 * @return 2011 * - 0: Success, the receive queue is started. 2012 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin. 2013 * - -EIO: if device is removed. 2014 * - -ENOTSUP: The function not supported in PMD driver. 2015 */ 2016 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id); 2017 2018 /** 2019 * Stop specified RX queue of a port 2020 * 2021 * @param port_id 2022 * The port identifier of the Ethernet device 2023 * @param rx_queue_id 2024 * The index of the rx queue to update the ring. 2025 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2026 * to rte_eth_dev_configure(). 2027 * @return 2028 * - 0: Success, the receive queue is stopped. 2029 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin. 2030 * - -EIO: if device is removed. 2031 * - -ENOTSUP: The function not supported in PMD driver. 2032 */ 2033 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id); 2034 2035 /** 2036 * Start TX for specified queue of a port. It is used when tx_deferred_start 2037 * flag of the specified queue is true. 2038 * 2039 * @param port_id 2040 * The port identifier of the Ethernet device 2041 * @param tx_queue_id 2042 * The index of the tx queue to update the ring. 2043 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2044 * to rte_eth_dev_configure(). 2045 * @return 2046 * - 0: Success, the transmit queue is started. 2047 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin. 2048 * - -EIO: if device is removed. 2049 * - -ENOTSUP: The function not supported in PMD driver. 2050 */ 2051 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id); 2052 2053 /** 2054 * Stop specified TX queue of a port 2055 * 2056 * @param port_id 2057 * The port identifier of the Ethernet device 2058 * @param tx_queue_id 2059 * The index of the tx queue to update the ring. 2060 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2061 * to rte_eth_dev_configure(). 2062 * @return 2063 * - 0: Success, the transmit queue is stopped. 2064 * - -EINVAL: The port_id or the queue_id out of range or belong to hairpin. 2065 * - -EIO: if device is removed. 2066 * - -ENOTSUP: The function not supported in PMD driver. 2067 */ 2068 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id); 2069 2070 /** 2071 * Start an Ethernet device. 2072 * 2073 * The device start step is the last one and consists of setting the configured 2074 * offload features and in starting the transmit and the receive units of the 2075 * device. 2076 * 2077 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before 2078 * PMD port start callback function is invoked. 2079 * 2080 * On success, all basic functions exported by the Ethernet API (link status, 2081 * receive/transmit, and so on) can be invoked. 2082 * 2083 * @param port_id 2084 * The port identifier of the Ethernet device. 2085 * @return 2086 * - 0: Success, Ethernet device started. 2087 * - <0: Error code of the driver device start function. 2088 */ 2089 int rte_eth_dev_start(uint16_t port_id); 2090 2091 /** 2092 * Stop an Ethernet device. The device can be restarted with a call to 2093 * rte_eth_dev_start() 2094 * 2095 * @param port_id 2096 * The port identifier of the Ethernet device. 2097 */ 2098 void rte_eth_dev_stop(uint16_t port_id); 2099 2100 /** 2101 * Link up an Ethernet device. 2102 * 2103 * Set device link up will re-enable the device rx/tx 2104 * functionality after it is previously set device linked down. 2105 * 2106 * @param port_id 2107 * The port identifier of the Ethernet device. 2108 * @return 2109 * - 0: Success, Ethernet device linked up. 2110 * - <0: Error code of the driver device link up function. 2111 */ 2112 int rte_eth_dev_set_link_up(uint16_t port_id); 2113 2114 /** 2115 * Link down an Ethernet device. 2116 * The device rx/tx functionality will be disabled if success, 2117 * and it can be re-enabled with a call to 2118 * rte_eth_dev_set_link_up() 2119 * 2120 * @param port_id 2121 * The port identifier of the Ethernet device. 2122 */ 2123 int rte_eth_dev_set_link_down(uint16_t port_id); 2124 2125 /** 2126 * Close a stopped Ethernet device. The device cannot be restarted! 2127 * The function frees all port resources if the driver supports 2128 * the flag RTE_ETH_DEV_CLOSE_REMOVE. 2129 * 2130 * @param port_id 2131 * The port identifier of the Ethernet device. 2132 */ 2133 void rte_eth_dev_close(uint16_t port_id); 2134 2135 /** 2136 * Reset a Ethernet device and keep its port id. 2137 * 2138 * When a port has to be reset passively, the DPDK application can invoke 2139 * this function. For example when a PF is reset, all its VFs should also 2140 * be reset. Normally a DPDK application can invoke this function when 2141 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start 2142 * a port reset in other circumstances. 2143 * 2144 * When this function is called, it first stops the port and then calls the 2145 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial 2146 * state, in which no Tx and Rx queues are setup, as if the port has been 2147 * reset and not started. The port keeps the port id it had before the 2148 * function call. 2149 * 2150 * After calling rte_eth_dev_reset( ), the application should use 2151 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ), 2152 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( ) 2153 * to reconfigure the device as appropriate. 2154 * 2155 * Note: To avoid unexpected behavior, the application should stop calling 2156 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread 2157 * safety, all these controlling functions should be called from the same 2158 * thread. 2159 * 2160 * @param port_id 2161 * The port identifier of the Ethernet device. 2162 * 2163 * @return 2164 * - (0) if successful. 2165 * - (-EINVAL) if port identifier is invalid. 2166 * - (-ENOTSUP) if hardware doesn't support this function. 2167 * - (-EPERM) if not ran from the primary process. 2168 * - (-EIO) if re-initialisation failed or device is removed. 2169 * - (-ENOMEM) if the reset failed due to OOM. 2170 * - (-EAGAIN) if the reset temporarily failed and should be retried later. 2171 */ 2172 int rte_eth_dev_reset(uint16_t port_id); 2173 2174 /** 2175 * Enable receipt in promiscuous mode for an Ethernet device. 2176 * 2177 * @param port_id 2178 * The port identifier of the Ethernet device. 2179 * @return 2180 * - (0) if successful. 2181 * - (-ENOTSUP) if support for promiscuous_enable() does not exist 2182 * for the device. 2183 * - (-ENODEV) if *port_id* invalid. 2184 */ 2185 int rte_eth_promiscuous_enable(uint16_t port_id); 2186 2187 /** 2188 * Disable receipt in promiscuous mode for an Ethernet device. 2189 * 2190 * @param port_id 2191 * The port identifier of the Ethernet device. 2192 * @return 2193 * - (0) if successful. 2194 * - (-ENOTSUP) if support for promiscuous_disable() does not exist 2195 * for the device. 2196 * - (-ENODEV) if *port_id* invalid. 2197 */ 2198 int rte_eth_promiscuous_disable(uint16_t port_id); 2199 2200 /** 2201 * Return the value of promiscuous mode for an Ethernet device. 2202 * 2203 * @param port_id 2204 * The port identifier of the Ethernet device. 2205 * @return 2206 * - (1) if promiscuous is enabled 2207 * - (0) if promiscuous is disabled. 2208 * - (-1) on error 2209 */ 2210 int rte_eth_promiscuous_get(uint16_t port_id); 2211 2212 /** 2213 * Enable the receipt of any multicast frame by an Ethernet device. 2214 * 2215 * @param port_id 2216 * The port identifier of the Ethernet device. 2217 * @return 2218 * - (0) if successful. 2219 * - (-ENOTSUP) if support for allmulticast_enable() does not exist 2220 * for the device. 2221 * - (-ENODEV) if *port_id* invalid. 2222 */ 2223 int rte_eth_allmulticast_enable(uint16_t port_id); 2224 2225 /** 2226 * Disable the receipt of all multicast frames by an Ethernet device. 2227 * 2228 * @param port_id 2229 * The port identifier of the Ethernet device. 2230 * @return 2231 * - (0) if successful. 2232 * - (-ENOTSUP) if support for allmulticast_disable() does not exist 2233 * for the device. 2234 * - (-ENODEV) if *port_id* invalid. 2235 */ 2236 int rte_eth_allmulticast_disable(uint16_t port_id); 2237 2238 /** 2239 * Return the value of allmulticast mode for an Ethernet device. 2240 * 2241 * @param port_id 2242 * The port identifier of the Ethernet device. 2243 * @return 2244 * - (1) if allmulticast is enabled 2245 * - (0) if allmulticast is disabled. 2246 * - (-1) on error 2247 */ 2248 int rte_eth_allmulticast_get(uint16_t port_id); 2249 2250 /** 2251 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX 2252 * or FULL-DUPLEX) of the physical link of an Ethernet device. It might need 2253 * to wait up to 9 seconds in it. 2254 * 2255 * @param port_id 2256 * The port identifier of the Ethernet device. 2257 * @param link 2258 * A pointer to an *rte_eth_link* structure to be filled with 2259 * the status, the speed and the mode of the Ethernet device link. 2260 * @return 2261 * - (0) if successful. 2262 * - (-ENOTSUP) if the function is not supported in PMD driver. 2263 * - (-ENODEV) if *port_id* invalid. 2264 */ 2265 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link); 2266 2267 /** 2268 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX 2269 * or FULL-DUPLEX) of the physical link of an Ethernet device. It is a no-wait 2270 * version of rte_eth_link_get(). 2271 * 2272 * @param port_id 2273 * The port identifier of the Ethernet device. 2274 * @param link 2275 * A pointer to an *rte_eth_link* structure to be filled with 2276 * the status, the speed and the mode of the Ethernet device link. 2277 * @return 2278 * - (0) if successful. 2279 * - (-ENOTSUP) if the function is not supported in PMD driver. 2280 * - (-ENODEV) if *port_id* invalid. 2281 */ 2282 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link); 2283 2284 /** 2285 * Retrieve the general I/O statistics of an Ethernet device. 2286 * 2287 * @param port_id 2288 * The port identifier of the Ethernet device. 2289 * @param stats 2290 * A pointer to a structure of type *rte_eth_stats* to be filled with 2291 * the values of device counters for the following set of statistics: 2292 * - *ipackets* with the total of successfully received packets. 2293 * - *opackets* with the total of successfully transmitted packets. 2294 * - *ibytes* with the total of successfully received bytes. 2295 * - *obytes* with the total of successfully transmitted bytes. 2296 * - *ierrors* with the total of erroneous received packets. 2297 * - *oerrors* with the total of failed transmitted packets. 2298 * @return 2299 * Zero if successful. Non-zero otherwise. 2300 */ 2301 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats); 2302 2303 /** 2304 * Reset the general I/O statistics of an Ethernet device. 2305 * 2306 * @param port_id 2307 * The port identifier of the Ethernet device. 2308 * @return 2309 * - (0) if device notified to reset stats. 2310 * - (-ENOTSUP) if hardware doesn't support. 2311 * - (-ENODEV) if *port_id* invalid. 2312 * - (<0): Error code of the driver stats reset function. 2313 */ 2314 int rte_eth_stats_reset(uint16_t port_id); 2315 2316 /** 2317 * Retrieve names of extended statistics of an Ethernet device. 2318 * 2319 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched 2320 * by array index: 2321 * xstats_names[i].name => xstats[i].value 2322 * 2323 * And the array index is same with id field of 'struct rte_eth_xstat': 2324 * xstats[i].id == i 2325 * 2326 * This assumption makes key-value pair matching less flexible but simpler. 2327 * 2328 * @param port_id 2329 * The port identifier of the Ethernet device. 2330 * @param xstats_names 2331 * An rte_eth_xstat_name array of at least *size* elements to 2332 * be filled. If set to NULL, the function returns the required number 2333 * of elements. 2334 * @param size 2335 * The size of the xstats_names array (number of elements). 2336 * @return 2337 * - A positive value lower or equal to size: success. The return value 2338 * is the number of entries filled in the stats table. 2339 * - A positive value higher than size: error, the given statistics table 2340 * is too small. The return value corresponds to the size that should 2341 * be given to succeed. The entries in the table are not valid and 2342 * shall not be used by the caller. 2343 * - A negative value on error (invalid port id). 2344 */ 2345 int rte_eth_xstats_get_names(uint16_t port_id, 2346 struct rte_eth_xstat_name *xstats_names, 2347 unsigned int size); 2348 2349 /** 2350 * Retrieve extended statistics of an Ethernet device. 2351 * 2352 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched 2353 * by array index: 2354 * xstats_names[i].name => xstats[i].value 2355 * 2356 * And the array index is same with id field of 'struct rte_eth_xstat': 2357 * xstats[i].id == i 2358 * 2359 * This assumption makes key-value pair matching less flexible but simpler. 2360 * 2361 * @param port_id 2362 * The port identifier of the Ethernet device. 2363 * @param xstats 2364 * A pointer to a table of structure of type *rte_eth_xstat* 2365 * to be filled with device statistics ids and values. 2366 * This parameter can be set to NULL if n is 0. 2367 * @param n 2368 * The size of the xstats array (number of elements). 2369 * @return 2370 * - A positive value lower or equal to n: success. The return value 2371 * is the number of entries filled in the stats table. 2372 * - A positive value higher than n: error, the given statistics table 2373 * is too small. The return value corresponds to the size that should 2374 * be given to succeed. The entries in the table are not valid and 2375 * shall not be used by the caller. 2376 * - A negative value on error (invalid port id). 2377 */ 2378 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats, 2379 unsigned int n); 2380 2381 /** 2382 * Retrieve names of extended statistics of an Ethernet device. 2383 * 2384 * @param port_id 2385 * The port identifier of the Ethernet device. 2386 * @param xstats_names 2387 * An rte_eth_xstat_name array of at least *size* elements to 2388 * be filled. If set to NULL, the function returns the required number 2389 * of elements. 2390 * @param ids 2391 * IDs array given by app to retrieve specific statistics 2392 * @param size 2393 * The size of the xstats_names array (number of elements). 2394 * @return 2395 * - A positive value lower or equal to size: success. The return value 2396 * is the number of entries filled in the stats table. 2397 * - A positive value higher than size: error, the given statistics table 2398 * is too small. The return value corresponds to the size that should 2399 * be given to succeed. The entries in the table are not valid and 2400 * shall not be used by the caller. 2401 * - A negative value on error (invalid port id). 2402 */ 2403 int 2404 rte_eth_xstats_get_names_by_id(uint16_t port_id, 2405 struct rte_eth_xstat_name *xstats_names, unsigned int size, 2406 uint64_t *ids); 2407 2408 /** 2409 * Retrieve extended statistics of an Ethernet device. 2410 * 2411 * @param port_id 2412 * The port identifier of the Ethernet device. 2413 * @param ids 2414 * A pointer to an ids array passed by application. This tells which 2415 * statistics values function should retrieve. This parameter 2416 * can be set to NULL if size is 0. In this case function will retrieve 2417 * all available statistics. 2418 * @param values 2419 * A pointer to a table to be filled with device statistics values. 2420 * @param size 2421 * The size of the ids array (number of elements). 2422 * @return 2423 * - A positive value lower or equal to size: success. The return value 2424 * is the number of entries filled in the stats table. 2425 * - A positive value higher than size: error, the given statistics table 2426 * is too small. The return value corresponds to the size that should 2427 * be given to succeed. The entries in the table are not valid and 2428 * shall not be used by the caller. 2429 * - A negative value on error (invalid port id). 2430 */ 2431 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids, 2432 uint64_t *values, unsigned int size); 2433 2434 /** 2435 * Gets the ID of a statistic from its name. 2436 * 2437 * This function searches for the statistics using string compares, and 2438 * as such should not be used on the fast-path. For fast-path retrieval of 2439 * specific statistics, store the ID as provided in *id* from this function, 2440 * and pass the ID to rte_eth_xstats_get() 2441 * 2442 * @param port_id The port to look up statistics from 2443 * @param xstat_name The name of the statistic to return 2444 * @param[out] id A pointer to an app-supplied uint64_t which should be 2445 * set to the ID of the stat if the stat exists. 2446 * @return 2447 * 0 on success 2448 * -ENODEV for invalid port_id, 2449 * -EIO if device is removed, 2450 * -EINVAL if the xstat_name doesn't exist in port_id 2451 */ 2452 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name, 2453 uint64_t *id); 2454 2455 /** 2456 * Reset extended statistics of an Ethernet device. 2457 * 2458 * @param port_id 2459 * The port identifier of the Ethernet device. 2460 * @return 2461 * - (0) if device notified to reset extended stats. 2462 * - (-ENOTSUP) if pmd doesn't support both 2463 * extended stats and basic stats reset. 2464 * - (-ENODEV) if *port_id* invalid. 2465 * - (<0): Error code of the driver xstats reset function. 2466 */ 2467 int rte_eth_xstats_reset(uint16_t port_id); 2468 2469 /** 2470 * Set a mapping for the specified transmit queue to the specified per-queue 2471 * statistics counter. 2472 * 2473 * @param port_id 2474 * The port identifier of the Ethernet device. 2475 * @param tx_queue_id 2476 * The index of the transmit queue for which a queue stats mapping is required. 2477 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2478 * to rte_eth_dev_configure(). 2479 * @param stat_idx 2480 * The per-queue packet statistics functionality number that the transmit 2481 * queue is to be assigned. 2482 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1]. 2483 * @return 2484 * Zero if successful. Non-zero otherwise. 2485 */ 2486 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id, 2487 uint16_t tx_queue_id, uint8_t stat_idx); 2488 2489 /** 2490 * Set a mapping for the specified receive queue to the specified per-queue 2491 * statistics counter. 2492 * 2493 * @param port_id 2494 * The port identifier of the Ethernet device. 2495 * @param rx_queue_id 2496 * The index of the receive queue for which a queue stats mapping is required. 2497 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2498 * to rte_eth_dev_configure(). 2499 * @param stat_idx 2500 * The per-queue packet statistics functionality number that the receive 2501 * queue is to be assigned. 2502 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1]. 2503 * @return 2504 * Zero if successful. Non-zero otherwise. 2505 */ 2506 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id, 2507 uint16_t rx_queue_id, 2508 uint8_t stat_idx); 2509 2510 /** 2511 * Retrieve the Ethernet address of an Ethernet device. 2512 * 2513 * @param port_id 2514 * The port identifier of the Ethernet device. 2515 * @param mac_addr 2516 * A pointer to a structure of type *ether_addr* to be filled with 2517 * the Ethernet address of the Ethernet device. 2518 * @return 2519 * - (0) if successful 2520 * - (-ENODEV) if *port_id* invalid. 2521 */ 2522 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr); 2523 2524 /** 2525 * Retrieve the contextual information of an Ethernet device. 2526 * 2527 * As part of this function, a number of of fields in dev_info will be 2528 * initialized as follows: 2529 * 2530 * rx_desc_lim = lim 2531 * tx_desc_lim = lim 2532 * 2533 * Where lim is defined within the rte_eth_dev_info_get as 2534 * 2535 * const struct rte_eth_desc_lim lim = { 2536 * .nb_max = UINT16_MAX, 2537 * .nb_min = 0, 2538 * .nb_align = 1, 2539 * .nb_seg_max = UINT16_MAX, 2540 * .nb_mtu_seg_max = UINT16_MAX, 2541 * }; 2542 * 2543 * device = dev->device 2544 * min_mtu = RTE_ETHER_MIN_MTU 2545 * max_mtu = UINT16_MAX 2546 * 2547 * The following fields will be populated if support for dev_infos_get() 2548 * exists for the device and the rte_eth_dev 'dev' has been populated 2549 * successfully with a call to it: 2550 * 2551 * driver_name = dev->device->driver->name 2552 * nb_rx_queues = dev->data->nb_rx_queues 2553 * nb_tx_queues = dev->data->nb_tx_queues 2554 * dev_flags = &dev->data->dev_flags 2555 * 2556 * @param port_id 2557 * The port identifier of the Ethernet device. 2558 * @param dev_info 2559 * A pointer to a structure of type *rte_eth_dev_info* to be filled with 2560 * the contextual information of the Ethernet device. 2561 * @return 2562 * - (0) if successful. 2563 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device. 2564 * - (-ENODEV) if *port_id* invalid. 2565 */ 2566 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info); 2567 2568 /** 2569 * Retrieve the firmware version of a device. 2570 * 2571 * @param port_id 2572 * The port identifier of the device. 2573 * @param fw_version 2574 * A pointer to a string array storing the firmware version of a device, 2575 * the string includes terminating null. This pointer is allocated by caller. 2576 * @param fw_size 2577 * The size of the string array pointed by fw_version, which should be 2578 * large enough to store firmware version of the device. 2579 * @return 2580 * - (0) if successful. 2581 * - (-ENOTSUP) if operation is not supported. 2582 * - (-ENODEV) if *port_id* invalid. 2583 * - (-EIO) if device is removed. 2584 * - (>0) if *fw_size* is not enough to store firmware version, return 2585 * the size of the non truncated string. 2586 */ 2587 int rte_eth_dev_fw_version_get(uint16_t port_id, 2588 char *fw_version, size_t fw_size); 2589 2590 /** 2591 * Retrieve the supported packet types of an Ethernet device. 2592 * 2593 * When a packet type is announced as supported, it *must* be recognized by 2594 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN 2595 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following 2596 * packet types for these packets: 2597 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4 2598 * - Ether/Vlan/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4 2599 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER 2600 * - Ether/Vlan/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN 2601 * 2602 * When a packet is received by a PMD, the most precise type must be 2603 * returned among the ones supported. However a PMD is allowed to set 2604 * packet type that is not in the supported list, at the condition that it 2605 * is more precise. Therefore, a PMD announcing no supported packet types 2606 * can still set a matching packet type in a received packet. 2607 * 2608 * @note 2609 * Better to invoke this API after the device is already started or rx burst 2610 * function is decided, to obtain correct supported ptypes. 2611 * @note 2612 * if a given PMD does not report what ptypes it supports, then the supported 2613 * ptype count is reported as 0. 2614 * @param port_id 2615 * The port identifier of the Ethernet device. 2616 * @param ptype_mask 2617 * A hint of what kind of packet type which the caller is interested in. 2618 * @param ptypes 2619 * An array pointer to store adequate packet types, allocated by caller. 2620 * @param num 2621 * Size of the array pointed by param ptypes. 2622 * @return 2623 * - (>=0) Number of supported ptypes. If the number of types exceeds num, 2624 * only num entries will be filled into the ptypes array, but the full 2625 * count of supported ptypes will be returned. 2626 * - (-ENODEV) if *port_id* invalid. 2627 */ 2628 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask, 2629 uint32_t *ptypes, int num); 2630 /** 2631 * @warning 2632 * @b EXPERIMENTAL: this API may change without prior notice. 2633 * 2634 * Inform Ethernet device about reduced range of packet types to handle. 2635 * 2636 * Application can use this function to set only specific ptypes that it's 2637 * interested. This information can be used by the PMD to optimize Rx path. 2638 * 2639 * The function accepts an array `set_ptypes` allocated by the caller to 2640 * store the packet types set by the driver, the last element of the array 2641 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be 2642 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled 2643 * partially. 2644 * 2645 * @param port_id 2646 * The port identifier of the Ethernet device. 2647 * @param ptype_mask 2648 * The ptype family that application is interested in should be bitwise OR of 2649 * RTE_PTYPE_*_MASK or 0. 2650 * @param set_ptypes 2651 * An array pointer to store set packet types, allocated by caller. The 2652 * function marks the end of array with RTE_PTYPE_UNKNOWN. 2653 * @param num 2654 * Size of the array pointed by param ptypes. 2655 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the 2656 * set ptypes. 2657 * @return 2658 * - (0) if Success. 2659 * - (-ENODEV) if *port_id* invalid. 2660 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and 2661 * num > 0. 2662 */ 2663 __rte_experimental 2664 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask, 2665 uint32_t *set_ptypes, unsigned int num); 2666 2667 /** 2668 * Retrieve the MTU of an Ethernet device. 2669 * 2670 * @param port_id 2671 * The port identifier of the Ethernet device. 2672 * @param mtu 2673 * A pointer to a uint16_t where the retrieved MTU is to be stored. 2674 * @return 2675 * - (0) if successful. 2676 * - (-ENODEV) if *port_id* invalid. 2677 */ 2678 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu); 2679 2680 /** 2681 * Change the MTU of an Ethernet device. 2682 * 2683 * @param port_id 2684 * The port identifier of the Ethernet device. 2685 * @param mtu 2686 * A uint16_t for the MTU to be applied. 2687 * @return 2688 * - (0) if successful. 2689 * - (-ENOTSUP) if operation is not supported. 2690 * - (-ENODEV) if *port_id* invalid. 2691 * - (-EIO) if device is removed. 2692 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within 2693 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or 2694 * when the mtu is set using dev->dev_ops->mtu_set. 2695 * - (-EBUSY) if operation is not allowed when the port is running 2696 */ 2697 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu); 2698 2699 /** 2700 * Enable/Disable hardware filtering by an Ethernet device of received 2701 * VLAN packets tagged with a given VLAN Tag Identifier. 2702 * 2703 * @param port_id 2704 * The port identifier of the Ethernet device. 2705 * @param vlan_id 2706 * The VLAN Tag Identifier whose filtering must be enabled or disabled. 2707 * @param on 2708 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*. 2709 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*. 2710 * @return 2711 * - (0) if successful. 2712 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured. 2713 * - (-ENODEV) if *port_id* invalid. 2714 * - (-EIO) if device is removed. 2715 * - (-ENOSYS) if VLAN filtering on *port_id* disabled. 2716 * - (-EINVAL) if *vlan_id* > 4095. 2717 */ 2718 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on); 2719 2720 /** 2721 * Enable/Disable hardware VLAN Strip by a rx queue of an Ethernet device. 2722 * 82599/X540/X550 can support VLAN stripping at the rx queue level 2723 * 2724 * @param port_id 2725 * The port identifier of the Ethernet device. 2726 * @param rx_queue_id 2727 * The index of the receive queue for which a queue stats mapping is required. 2728 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2729 * to rte_eth_dev_configure(). 2730 * @param on 2731 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port. 2732 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port. 2733 * @return 2734 * - (0) if successful. 2735 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured. 2736 * - (-ENODEV) if *port_id* invalid. 2737 * - (-EINVAL) if *rx_queue_id* invalid. 2738 */ 2739 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id, 2740 int on); 2741 2742 /** 2743 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to 2744 * the VLAN Header. This is a register setup available on some Intel NIC, not 2745 * but all, please check the data sheet for availability. 2746 * 2747 * @param port_id 2748 * The port identifier of the Ethernet device. 2749 * @param vlan_type 2750 * The vlan type. 2751 * @param tag_type 2752 * The Tag Protocol ID 2753 * @return 2754 * - (0) if successful. 2755 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported. 2756 * - (-ENODEV) if *port_id* invalid. 2757 * - (-EIO) if device is removed. 2758 */ 2759 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id, 2760 enum rte_vlan_type vlan_type, 2761 uint16_t tag_type); 2762 2763 /** 2764 * Set VLAN offload configuration on an Ethernet device 2765 * Enable/Disable Extended VLAN by an Ethernet device, This is a register setup 2766 * available on some Intel NIC, not but all, please check the data sheet for 2767 * availability. 2768 * Enable/Disable VLAN Strip can be done on rx queue for certain NIC, but here 2769 * the configuration is applied on the port level. 2770 * 2771 * @param port_id 2772 * The port identifier of the Ethernet device. 2773 * @param offload_mask 2774 * The VLAN Offload bit mask can be mixed use with "OR" 2775 * ETH_VLAN_STRIP_OFFLOAD 2776 * ETH_VLAN_FILTER_OFFLOAD 2777 * ETH_VLAN_EXTEND_OFFLOAD 2778 * ETH_QINQ_STRIP_OFFLOAD 2779 * @return 2780 * - (0) if successful. 2781 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured. 2782 * - (-ENODEV) if *port_id* invalid. 2783 * - (-EIO) if device is removed. 2784 */ 2785 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask); 2786 2787 /** 2788 * Read VLAN Offload configuration from an Ethernet device 2789 * 2790 * @param port_id 2791 * The port identifier of the Ethernet device. 2792 * @return 2793 * - (>0) if successful. Bit mask to indicate 2794 * ETH_VLAN_STRIP_OFFLOAD 2795 * ETH_VLAN_FILTER_OFFLOAD 2796 * ETH_VLAN_EXTEND_OFFLOAD 2797 * ETH_QINQ_STRIP_OFFLOAD 2798 * - (-ENODEV) if *port_id* invalid. 2799 */ 2800 int rte_eth_dev_get_vlan_offload(uint16_t port_id); 2801 2802 /** 2803 * Set port based TX VLAN insertion on or off. 2804 * 2805 * @param port_id 2806 * The port identifier of the Ethernet device. 2807 * @param pvid 2808 * Port based TX VLAN identifier together with user priority. 2809 * @param on 2810 * Turn on or off the port based TX VLAN insertion. 2811 * 2812 * @return 2813 * - (0) if successful. 2814 * - negative if failed. 2815 */ 2816 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on); 2817 2818 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count, 2819 void *userdata); 2820 2821 /** 2822 * Structure used to buffer packets for future TX 2823 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush 2824 */ 2825 struct rte_eth_dev_tx_buffer { 2826 buffer_tx_error_fn error_callback; 2827 void *error_userdata; 2828 uint16_t size; /**< Size of buffer for buffered tx */ 2829 uint16_t length; /**< Number of packets in the array */ 2830 struct rte_mbuf *pkts[]; 2831 /**< Pending packets to be sent on explicit flush or when full */ 2832 }; 2833 2834 /** 2835 * Calculate the size of the tx buffer. 2836 * 2837 * @param sz 2838 * Number of stored packets. 2839 */ 2840 #define RTE_ETH_TX_BUFFER_SIZE(sz) \ 2841 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *)) 2842 2843 /** 2844 * Initialize default values for buffered transmitting 2845 * 2846 * @param buffer 2847 * Tx buffer to be initialized. 2848 * @param size 2849 * Buffer size 2850 * @return 2851 * 0 if no error 2852 */ 2853 int 2854 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size); 2855 2856 /** 2857 * Configure a callback for buffered packets which cannot be sent 2858 * 2859 * Register a specific callback to be called when an attempt is made to send 2860 * all packets buffered on an ethernet port, but not all packets can 2861 * successfully be sent. The callback registered here will be called only 2862 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs. 2863 * The default callback configured for each queue by default just frees the 2864 * packets back to the calling mempool. If additional behaviour is required, 2865 * for example, to count dropped packets, or to retry transmission of packets 2866 * which cannot be sent, this function should be used to register a suitable 2867 * callback function to implement the desired behaviour. 2868 * The example callback "rte_eth_count_unsent_packet_callback()" is also 2869 * provided as reference. 2870 * 2871 * @param buffer 2872 * The port identifier of the Ethernet device. 2873 * @param callback 2874 * The function to be used as the callback. 2875 * @param userdata 2876 * Arbitrary parameter to be passed to the callback function 2877 * @return 2878 * 0 on success, or -1 on error with rte_errno set appropriately 2879 */ 2880 int 2881 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer, 2882 buffer_tx_error_fn callback, void *userdata); 2883 2884 /** 2885 * Callback function for silently dropping unsent buffered packets. 2886 * 2887 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to 2888 * adjust the default behavior when buffered packets cannot be sent. This 2889 * function drops any unsent packets silently and is used by tx buffered 2890 * operations as default behavior. 2891 * 2892 * NOTE: this function should not be called directly, instead it should be used 2893 * as a callback for packet buffering. 2894 * 2895 * NOTE: when configuring this function as a callback with 2896 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter 2897 * should point to an uint64_t value. 2898 * 2899 * @param pkts 2900 * The previously buffered packets which could not be sent 2901 * @param unsent 2902 * The number of unsent packets in the pkts array 2903 * @param userdata 2904 * Not used 2905 */ 2906 void 2907 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent, 2908 void *userdata); 2909 2910 /** 2911 * Callback function for tracking unsent buffered packets. 2912 * 2913 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to 2914 * adjust the default behavior when buffered packets cannot be sent. This 2915 * function drops any unsent packets, but also updates a user-supplied counter 2916 * to track the overall number of packets dropped. The counter should be an 2917 * uint64_t variable. 2918 * 2919 * NOTE: this function should not be called directly, instead it should be used 2920 * as a callback for packet buffering. 2921 * 2922 * NOTE: when configuring this function as a callback with 2923 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter 2924 * should point to an uint64_t value. 2925 * 2926 * @param pkts 2927 * The previously buffered packets which could not be sent 2928 * @param unsent 2929 * The number of unsent packets in the pkts array 2930 * @param userdata 2931 * Pointer to an uint64_t value, which will be incremented by unsent 2932 */ 2933 void 2934 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent, 2935 void *userdata); 2936 2937 /** 2938 * Request the driver to free mbufs currently cached by the driver. The 2939 * driver will only free the mbuf if it is no longer in use. It is the 2940 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is 2941 * called if needed. 2942 * 2943 * @param port_id 2944 * The port identifier of the Ethernet device. 2945 * @param queue_id 2946 * The index of the transmit queue through which output packets must be 2947 * sent. 2948 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2949 * to rte_eth_dev_configure(). 2950 * @param free_cnt 2951 * Maximum number of packets to free. Use 0 to indicate all possible packets 2952 * should be freed. Note that a packet may be using multiple mbufs. 2953 * @return 2954 * Failure: < 0 2955 * -ENODEV: Invalid interface 2956 * -EIO: device is removed 2957 * -ENOTSUP: Driver does not support function 2958 * Success: >= 0 2959 * 0-n: Number of packets freed. More packets may still remain in ring that 2960 * are in use. 2961 */ 2962 int 2963 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt); 2964 2965 /** 2966 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by 2967 * eth device. 2968 */ 2969 enum rte_eth_event_ipsec_subtype { 2970 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0, 2971 /**< Unknown event type */ 2972 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW, 2973 /**< Sequence number overflow */ 2974 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY, 2975 /**< Soft time expiry of SA */ 2976 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY, 2977 /**< Soft byte expiry of SA */ 2978 RTE_ETH_EVENT_IPSEC_MAX 2979 /**< Max value of this enum */ 2980 }; 2981 2982 /** 2983 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra 2984 * information of the IPsec offload event. 2985 */ 2986 struct rte_eth_event_ipsec_desc { 2987 enum rte_eth_event_ipsec_subtype subtype; 2988 /**< Type of RTE_ETH_EVENT_IPSEC_* event */ 2989 uint64_t metadata; 2990 /**< Event specific metadata 2991 * 2992 * For the following events, *userdata* registered 2993 * with the *rte_security_session* would be returned 2994 * as metadata, 2995 * 2996 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW 2997 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY 2998 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY 2999 * 3000 * @see struct rte_security_session_conf 3001 * 3002 */ 3003 }; 3004 3005 /** 3006 * The eth device event type for interrupt, and maybe others in the future. 3007 */ 3008 enum rte_eth_event_type { 3009 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */ 3010 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */ 3011 RTE_ETH_EVENT_QUEUE_STATE, 3012 /**< queue state event (enabled/disabled) */ 3013 RTE_ETH_EVENT_INTR_RESET, 3014 /**< reset interrupt event, sent to VF on PF reset */ 3015 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */ 3016 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */ 3017 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */ 3018 RTE_ETH_EVENT_NEW, /**< port is probed */ 3019 RTE_ETH_EVENT_DESTROY, /**< port is released */ 3020 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */ 3021 RTE_ETH_EVENT_MAX /**< max value of this enum */ 3022 }; 3023 3024 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id, 3025 enum rte_eth_event_type event, void *cb_arg, void *ret_param); 3026 /**< user application callback to be registered for interrupts */ 3027 3028 /** 3029 * Register a callback function for port event. 3030 * 3031 * @param port_id 3032 * Port id. 3033 * RTE_ETH_ALL means register the event for all port ids. 3034 * @param event 3035 * Event interested. 3036 * @param cb_fn 3037 * User supplied callback function to be called. 3038 * @param cb_arg 3039 * Pointer to the parameters for the registered callback. 3040 * 3041 * @return 3042 * - On success, zero. 3043 * - On failure, a negative value. 3044 */ 3045 int rte_eth_dev_callback_register(uint16_t port_id, 3046 enum rte_eth_event_type event, 3047 rte_eth_dev_cb_fn cb_fn, void *cb_arg); 3048 3049 /** 3050 * Unregister a callback function for port event. 3051 * 3052 * @param port_id 3053 * Port id. 3054 * RTE_ETH_ALL means unregister the event for all port ids. 3055 * @param event 3056 * Event interested. 3057 * @param cb_fn 3058 * User supplied callback function to be called. 3059 * @param cb_arg 3060 * Pointer to the parameters for the registered callback. -1 means to 3061 * remove all for the same callback address and same event. 3062 * 3063 * @return 3064 * - On success, zero. 3065 * - On failure, a negative value. 3066 */ 3067 int rte_eth_dev_callback_unregister(uint16_t port_id, 3068 enum rte_eth_event_type event, 3069 rte_eth_dev_cb_fn cb_fn, void *cb_arg); 3070 3071 /** 3072 * When there is no rx packet coming in Rx Queue for a long time, we can 3073 * sleep lcore related to RX Queue for power saving, and enable rx interrupt 3074 * to be triggered when Rx packet arrives. 3075 * 3076 * The rte_eth_dev_rx_intr_enable() function enables rx queue 3077 * interrupt on specific rx queue of a port. 3078 * 3079 * @param port_id 3080 * The port identifier of the Ethernet device. 3081 * @param queue_id 3082 * The index of the receive queue from which to retrieve input packets. 3083 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3084 * to rte_eth_dev_configure(). 3085 * @return 3086 * - (0) if successful. 3087 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3088 * that operation. 3089 * - (-ENODEV) if *port_id* invalid. 3090 * - (-EIO) if device is removed. 3091 */ 3092 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id); 3093 3094 /** 3095 * When lcore wakes up from rx interrupt indicating packet coming, disable rx 3096 * interrupt and returns to polling mode. 3097 * 3098 * The rte_eth_dev_rx_intr_disable() function disables rx queue 3099 * interrupt on specific rx queue of a port. 3100 * 3101 * @param port_id 3102 * The port identifier of the Ethernet device. 3103 * @param queue_id 3104 * The index of the receive queue from which to retrieve input packets. 3105 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3106 * to rte_eth_dev_configure(). 3107 * @return 3108 * - (0) if successful. 3109 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3110 * that operation. 3111 * - (-ENODEV) if *port_id* invalid. 3112 * - (-EIO) if device is removed. 3113 */ 3114 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id); 3115 3116 /** 3117 * RX Interrupt control per port. 3118 * 3119 * @param port_id 3120 * The port identifier of the Ethernet device. 3121 * @param epfd 3122 * Epoll instance fd which the intr vector associated to. 3123 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance. 3124 * @param op 3125 * The operation be performed for the vector. 3126 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}. 3127 * @param data 3128 * User raw data. 3129 * @return 3130 * - On success, zero. 3131 * - On failure, a negative value. 3132 */ 3133 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data); 3134 3135 /** 3136 * RX Interrupt control per queue. 3137 * 3138 * @param port_id 3139 * The port identifier of the Ethernet device. 3140 * @param queue_id 3141 * The index of the receive queue from which to retrieve input packets. 3142 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3143 * to rte_eth_dev_configure(). 3144 * @param epfd 3145 * Epoll instance fd which the intr vector associated to. 3146 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance. 3147 * @param op 3148 * The operation be performed for the vector. 3149 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}. 3150 * @param data 3151 * User raw data. 3152 * @return 3153 * - On success, zero. 3154 * - On failure, a negative value. 3155 */ 3156 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id, 3157 int epfd, int op, void *data); 3158 3159 /** 3160 * @warning 3161 * @b EXPERIMENTAL: this API may change without prior notice. 3162 * 3163 * Get interrupt fd per Rx queue. 3164 * 3165 * @param port_id 3166 * The port identifier of the Ethernet device. 3167 * @param queue_id 3168 * The index of the receive queue from which to retrieve input packets. 3169 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3170 * to rte_eth_dev_configure(). 3171 * @return 3172 * - (>=0) the interrupt fd associated to the requested Rx queue if 3173 * successful. 3174 * - (-1) on error. 3175 */ 3176 __rte_experimental 3177 int 3178 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id); 3179 3180 /** 3181 * Turn on the LED on the Ethernet device. 3182 * This function turns on the LED on the Ethernet device. 3183 * 3184 * @param port_id 3185 * The port identifier of the Ethernet device. 3186 * @return 3187 * - (0) if successful. 3188 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3189 * that operation. 3190 * - (-ENODEV) if *port_id* invalid. 3191 * - (-EIO) if device is removed. 3192 */ 3193 int rte_eth_led_on(uint16_t port_id); 3194 3195 /** 3196 * Turn off the LED on the Ethernet device. 3197 * This function turns off the LED on the Ethernet device. 3198 * 3199 * @param port_id 3200 * The port identifier of the Ethernet device. 3201 * @return 3202 * - (0) if successful. 3203 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3204 * that operation. 3205 * - (-ENODEV) if *port_id* invalid. 3206 * - (-EIO) if device is removed. 3207 */ 3208 int rte_eth_led_off(uint16_t port_id); 3209 3210 /** 3211 * Get current status of the Ethernet link flow control for Ethernet device 3212 * 3213 * @param port_id 3214 * The port identifier of the Ethernet device. 3215 * @param fc_conf 3216 * The pointer to the structure where to store the flow control parameters. 3217 * @return 3218 * - (0) if successful. 3219 * - (-ENOTSUP) if hardware doesn't support flow control. 3220 * - (-ENODEV) if *port_id* invalid. 3221 * - (-EIO) if device is removed. 3222 */ 3223 int rte_eth_dev_flow_ctrl_get(uint16_t port_id, 3224 struct rte_eth_fc_conf *fc_conf); 3225 3226 /** 3227 * Configure the Ethernet link flow control for Ethernet device 3228 * 3229 * @param port_id 3230 * The port identifier of the Ethernet device. 3231 * @param fc_conf 3232 * The pointer to the structure of the flow control parameters. 3233 * @return 3234 * - (0) if successful. 3235 * - (-ENOTSUP) if hardware doesn't support flow control mode. 3236 * - (-ENODEV) if *port_id* invalid. 3237 * - (-EINVAL) if bad parameter 3238 * - (-EIO) if flow control setup failure or device is removed. 3239 */ 3240 int rte_eth_dev_flow_ctrl_set(uint16_t port_id, 3241 struct rte_eth_fc_conf *fc_conf); 3242 3243 /** 3244 * Configure the Ethernet priority flow control under DCB environment 3245 * for Ethernet device. 3246 * 3247 * @param port_id 3248 * The port identifier of the Ethernet device. 3249 * @param pfc_conf 3250 * The pointer to the structure of the priority flow control parameters. 3251 * @return 3252 * - (0) if successful. 3253 * - (-ENOTSUP) if hardware doesn't support priority flow control mode. 3254 * - (-ENODEV) if *port_id* invalid. 3255 * - (-EINVAL) if bad parameter 3256 * - (-EIO) if flow control setup failure or device is removed. 3257 */ 3258 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id, 3259 struct rte_eth_pfc_conf *pfc_conf); 3260 3261 /** 3262 * Add a MAC address to an internal array of addresses used to enable whitelist 3263 * filtering to accept packets only if the destination MAC address matches. 3264 * 3265 * @param port_id 3266 * The port identifier of the Ethernet device. 3267 * @param mac_addr 3268 * The MAC address to add. 3269 * @param pool 3270 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is 3271 * not enabled, this should be set to 0. 3272 * @return 3273 * - (0) if successfully added or *mac_addr* was already added. 3274 * - (-ENOTSUP) if hardware doesn't support this feature. 3275 * - (-ENODEV) if *port* is invalid. 3276 * - (-EIO) if device is removed. 3277 * - (-ENOSPC) if no more MAC addresses can be added. 3278 * - (-EINVAL) if MAC address is invalid. 3279 */ 3280 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr, 3281 uint32_t pool); 3282 3283 /** 3284 * Remove a MAC address from the internal array of addresses. 3285 * 3286 * @param port_id 3287 * The port identifier of the Ethernet device. 3288 * @param mac_addr 3289 * MAC address to remove. 3290 * @return 3291 * - (0) if successful, or *mac_addr* didn't exist. 3292 * - (-ENOTSUP) if hardware doesn't support. 3293 * - (-ENODEV) if *port* invalid. 3294 * - (-EADDRINUSE) if attempting to remove the default MAC address 3295 */ 3296 int rte_eth_dev_mac_addr_remove(uint16_t port_id, 3297 struct rte_ether_addr *mac_addr); 3298 3299 /** 3300 * Set the default MAC address. 3301 * 3302 * @param port_id 3303 * The port identifier of the Ethernet device. 3304 * @param mac_addr 3305 * New default MAC address. 3306 * @return 3307 * - (0) if successful, or *mac_addr* didn't exist. 3308 * - (-ENOTSUP) if hardware doesn't support. 3309 * - (-ENODEV) if *port* invalid. 3310 * - (-EINVAL) if MAC address is invalid. 3311 */ 3312 int rte_eth_dev_default_mac_addr_set(uint16_t port_id, 3313 struct rte_ether_addr *mac_addr); 3314 3315 /** 3316 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device. 3317 * 3318 * @param port_id 3319 * The port identifier of the Ethernet device. 3320 * @param reta_conf 3321 * RETA to update. 3322 * @param reta_size 3323 * Redirection table size. The table size can be queried by 3324 * rte_eth_dev_info_get(). 3325 * @return 3326 * - (0) if successful. 3327 * - (-ENOTSUP) if hardware doesn't support. 3328 * - (-EINVAL) if bad parameter. 3329 * - (-EIO) if device is removed. 3330 */ 3331 int rte_eth_dev_rss_reta_update(uint16_t port_id, 3332 struct rte_eth_rss_reta_entry64 *reta_conf, 3333 uint16_t reta_size); 3334 3335 /** 3336 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device. 3337 * 3338 * @param port_id 3339 * The port identifier of the Ethernet device. 3340 * @param reta_conf 3341 * RETA to query. For each requested reta entry, corresponding bit 3342 * in mask must be set. 3343 * @param reta_size 3344 * Redirection table size. The table size can be queried by 3345 * rte_eth_dev_info_get(). 3346 * @return 3347 * - (0) if successful. 3348 * - (-ENOTSUP) if hardware doesn't support. 3349 * - (-EINVAL) if bad parameter. 3350 * - (-EIO) if device is removed. 3351 */ 3352 int rte_eth_dev_rss_reta_query(uint16_t port_id, 3353 struct rte_eth_rss_reta_entry64 *reta_conf, 3354 uint16_t reta_size); 3355 3356 /** 3357 * Updates unicast hash table for receiving packet with the given destination 3358 * MAC address, and the packet is routed to all VFs for which the RX mode is 3359 * accept packets that match the unicast hash table. 3360 * 3361 * @param port_id 3362 * The port identifier of the Ethernet device. 3363 * @param addr 3364 * Unicast MAC address. 3365 * @param on 3366 * 1 - Set an unicast hash bit for receiving packets with the MAC address. 3367 * 0 - Clear an unicast hash bit. 3368 * @return 3369 * - (0) if successful. 3370 * - (-ENOTSUP) if hardware doesn't support. 3371 * - (-ENODEV) if *port_id* invalid. 3372 * - (-EIO) if device is removed. 3373 * - (-EINVAL) if bad parameter. 3374 */ 3375 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr, 3376 uint8_t on); 3377 3378 /** 3379 * Updates all unicast hash bitmaps for receiving packet with any Unicast 3380 * Ethernet MAC addresses,the packet is routed to all VFs for which the RX 3381 * mode is accept packets that match the unicast hash table. 3382 * 3383 * @param port_id 3384 * The port identifier of the Ethernet device. 3385 * @param on 3386 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet 3387 * MAC addresses 3388 * 0 - Clear all unicast hash bitmaps 3389 * @return 3390 * - (0) if successful. 3391 * - (-ENOTSUP) if hardware doesn't support. 3392 * - (-ENODEV) if *port_id* invalid. 3393 * - (-EIO) if device is removed. 3394 * - (-EINVAL) if bad parameter. 3395 */ 3396 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on); 3397 3398 /** 3399 * Set a traffic mirroring rule on an Ethernet device 3400 * 3401 * @param port_id 3402 * The port identifier of the Ethernet device. 3403 * @param mirror_conf 3404 * The pointer to the traffic mirroring structure describing the mirroring rule. 3405 * The *rte_eth_vm_mirror_conf* structure includes the type of mirroring rule, 3406 * destination pool and the value of rule if enable vlan or pool mirroring. 3407 * 3408 * @param rule_id 3409 * The index of traffic mirroring rule, we support four separated rules. 3410 * @param on 3411 * 1 - Enable a mirroring rule. 3412 * 0 - Disable a mirroring rule. 3413 * @return 3414 * - (0) if successful. 3415 * - (-ENOTSUP) if hardware doesn't support this feature. 3416 * - (-ENODEV) if *port_id* invalid. 3417 * - (-EIO) if device is removed. 3418 * - (-EINVAL) if the mr_conf information is not correct. 3419 */ 3420 int rte_eth_mirror_rule_set(uint16_t port_id, 3421 struct rte_eth_mirror_conf *mirror_conf, 3422 uint8_t rule_id, 3423 uint8_t on); 3424 3425 /** 3426 * Reset a traffic mirroring rule on an Ethernet device. 3427 * 3428 * @param port_id 3429 * The port identifier of the Ethernet device. 3430 * @param rule_id 3431 * The index of traffic mirroring rule, we support four separated rules. 3432 * @return 3433 * - (0) if successful. 3434 * - (-ENOTSUP) if hardware doesn't support this feature. 3435 * - (-ENODEV) if *port_id* invalid. 3436 * - (-EIO) if device is removed. 3437 * - (-EINVAL) if bad parameter. 3438 */ 3439 int rte_eth_mirror_rule_reset(uint16_t port_id, 3440 uint8_t rule_id); 3441 3442 /** 3443 * Set the rate limitation for a queue on an Ethernet device. 3444 * 3445 * @param port_id 3446 * The port identifier of the Ethernet device. 3447 * @param queue_idx 3448 * The queue id. 3449 * @param tx_rate 3450 * The tx rate in Mbps. Allocated from the total port link speed. 3451 * @return 3452 * - (0) if successful. 3453 * - (-ENOTSUP) if hardware doesn't support this feature. 3454 * - (-ENODEV) if *port_id* invalid. 3455 * - (-EIO) if device is removed. 3456 * - (-EINVAL) if bad parameter. 3457 */ 3458 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx, 3459 uint16_t tx_rate); 3460 3461 /** 3462 * Configuration of Receive Side Scaling hash computation of Ethernet device. 3463 * 3464 * @param port_id 3465 * The port identifier of the Ethernet device. 3466 * @param rss_conf 3467 * The new configuration to use for RSS hash computation on the port. 3468 * @return 3469 * - (0) if successful. 3470 * - (-ENODEV) if port identifier is invalid. 3471 * - (-EIO) if device is removed. 3472 * - (-ENOTSUP) if hardware doesn't support. 3473 * - (-EINVAL) if bad parameter. 3474 */ 3475 int rte_eth_dev_rss_hash_update(uint16_t port_id, 3476 struct rte_eth_rss_conf *rss_conf); 3477 3478 /** 3479 * Retrieve current configuration of Receive Side Scaling hash computation 3480 * of Ethernet device. 3481 * 3482 * @param port_id 3483 * The port identifier of the Ethernet device. 3484 * @param rss_conf 3485 * Where to store the current RSS hash configuration of the Ethernet device. 3486 * @return 3487 * - (0) if successful. 3488 * - (-ENODEV) if port identifier is invalid. 3489 * - (-EIO) if device is removed. 3490 * - (-ENOTSUP) if hardware doesn't support RSS. 3491 */ 3492 int 3493 rte_eth_dev_rss_hash_conf_get(uint16_t port_id, 3494 struct rte_eth_rss_conf *rss_conf); 3495 3496 /** 3497 * Add UDP tunneling port for a specific type of tunnel. 3498 * The packets with this UDP port will be identified as this type of tunnel. 3499 * Before enabling any offloading function for a tunnel, users can call this API 3500 * to change or add more UDP port for the tunnel. So the offloading function 3501 * can take effect on the packets with the specific UDP port. 3502 * 3503 * @param port_id 3504 * The port identifier of the Ethernet device. 3505 * @param tunnel_udp 3506 * UDP tunneling configuration. 3507 * 3508 * @return 3509 * - (0) if successful. 3510 * - (-ENODEV) if port identifier is invalid. 3511 * - (-EIO) if device is removed. 3512 * - (-ENOTSUP) if hardware doesn't support tunnel type. 3513 */ 3514 int 3515 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id, 3516 struct rte_eth_udp_tunnel *tunnel_udp); 3517 3518 /** 3519 * Delete UDP tunneling port a specific type of tunnel. 3520 * The packets with this UDP port will not be identified as this type of tunnel 3521 * any more. 3522 * Before enabling any offloading function for a tunnel, users can call this API 3523 * to delete a UDP port for the tunnel. So the offloading function will not take 3524 * effect on the packets with the specific UDP port. 3525 * 3526 * @param port_id 3527 * The port identifier of the Ethernet device. 3528 * @param tunnel_udp 3529 * UDP tunneling configuration. 3530 * 3531 * @return 3532 * - (0) if successful. 3533 * - (-ENODEV) if port identifier is invalid. 3534 * - (-EIO) if device is removed. 3535 * - (-ENOTSUP) if hardware doesn't support tunnel type. 3536 */ 3537 int 3538 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id, 3539 struct rte_eth_udp_tunnel *tunnel_udp); 3540 3541 /** 3542 * Check whether the filter type is supported on an Ethernet device. 3543 * All the supported filter types are defined in 'rte_eth_ctrl.h'. 3544 * 3545 * @param port_id 3546 * The port identifier of the Ethernet device. 3547 * @param filter_type 3548 * Filter type. 3549 * @return 3550 * - (0) if successful. 3551 * - (-ENOTSUP) if hardware doesn't support this filter type. 3552 * - (-ENODEV) if *port_id* invalid. 3553 * - (-EIO) if device is removed. 3554 */ 3555 __rte_deprecated 3556 int rte_eth_dev_filter_supported(uint16_t port_id, 3557 enum rte_filter_type filter_type); 3558 3559 /** 3560 * Take operations to assigned filter type on an Ethernet device. 3561 * All the supported operations and filter types are defined in 'rte_eth_ctrl.h'. 3562 * 3563 * @param port_id 3564 * The port identifier of the Ethernet device. 3565 * @param filter_type 3566 * Filter type. 3567 * @param filter_op 3568 * Type of operation. 3569 * @param arg 3570 * A pointer to arguments defined specifically for the operation. 3571 * @return 3572 * - (0) if successful. 3573 * - (-ENOTSUP) if hardware doesn't support. 3574 * - (-ENODEV) if *port_id* invalid. 3575 * - (-EIO) if device is removed. 3576 * - others depends on the specific operations implementation. 3577 */ 3578 __rte_deprecated 3579 int rte_eth_dev_filter_ctrl(uint16_t port_id, enum rte_filter_type filter_type, 3580 enum rte_filter_op filter_op, void *arg); 3581 3582 /** 3583 * Get DCB information on an Ethernet device. 3584 * 3585 * @param port_id 3586 * The port identifier of the Ethernet device. 3587 * @param dcb_info 3588 * dcb information. 3589 * @return 3590 * - (0) if successful. 3591 * - (-ENODEV) if port identifier is invalid. 3592 * - (-EIO) if device is removed. 3593 * - (-ENOTSUP) if hardware doesn't support. 3594 */ 3595 int rte_eth_dev_get_dcb_info(uint16_t port_id, 3596 struct rte_eth_dcb_info *dcb_info); 3597 3598 struct rte_eth_rxtx_callback; 3599 3600 /** 3601 * Add a callback to be called on packet RX on a given port and queue. 3602 * 3603 * This API configures a function to be called for each burst of 3604 * packets received on a given NIC port queue. The return value is a pointer 3605 * that can be used to later remove the callback using 3606 * rte_eth_remove_rx_callback(). 3607 * 3608 * Multiple functions are called in the order that they are added. 3609 * 3610 * @param port_id 3611 * The port identifier of the Ethernet device. 3612 * @param queue_id 3613 * The queue on the Ethernet device on which the callback is to be added. 3614 * @param fn 3615 * The callback function 3616 * @param user_param 3617 * A generic pointer parameter which will be passed to each invocation of the 3618 * callback function on this port and queue. Inter-thread synchronization 3619 * of any user data changes is the responsibility of the user. 3620 * 3621 * @return 3622 * NULL on error. 3623 * On success, a pointer value which can later be used to remove the callback. 3624 */ 3625 const struct rte_eth_rxtx_callback * 3626 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id, 3627 rte_rx_callback_fn fn, void *user_param); 3628 3629 /** 3630 * Add a callback that must be called first on packet RX on a given port 3631 * and queue. 3632 * 3633 * This API configures a first function to be called for each burst of 3634 * packets received on a given NIC port queue. The return value is a pointer 3635 * that can be used to later remove the callback using 3636 * rte_eth_remove_rx_callback(). 3637 * 3638 * Multiple functions are called in the order that they are added. 3639 * 3640 * @param port_id 3641 * The port identifier of the Ethernet device. 3642 * @param queue_id 3643 * The queue on the Ethernet device on which the callback is to be added. 3644 * @param fn 3645 * The callback function 3646 * @param user_param 3647 * A generic pointer parameter which will be passed to each invocation of the 3648 * callback function on this port and queue. Inter-thread synchronization 3649 * of any user data changes is the responsibility of the user. 3650 * 3651 * @return 3652 * NULL on error. 3653 * On success, a pointer value which can later be used to remove the callback. 3654 */ 3655 const struct rte_eth_rxtx_callback * 3656 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id, 3657 rte_rx_callback_fn fn, void *user_param); 3658 3659 /** 3660 * Add a callback to be called on packet TX on a given port and queue. 3661 * 3662 * This API configures a function to be called for each burst of 3663 * packets sent on a given NIC port queue. The return value is a pointer 3664 * that can be used to later remove the callback using 3665 * rte_eth_remove_tx_callback(). 3666 * 3667 * Multiple functions are called in the order that they are added. 3668 * 3669 * @param port_id 3670 * The port identifier of the Ethernet device. 3671 * @param queue_id 3672 * The queue on the Ethernet device on which the callback is to be added. 3673 * @param fn 3674 * The callback function 3675 * @param user_param 3676 * A generic pointer parameter which will be passed to each invocation of the 3677 * callback function on this port and queue. Inter-thread synchronization 3678 * of any user data changes is the responsibility of the user. 3679 * 3680 * @return 3681 * NULL on error. 3682 * On success, a pointer value which can later be used to remove the callback. 3683 */ 3684 const struct rte_eth_rxtx_callback * 3685 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id, 3686 rte_tx_callback_fn fn, void *user_param); 3687 3688 /** 3689 * Remove an RX packet callback from a given port and queue. 3690 * 3691 * This function is used to removed callbacks that were added to a NIC port 3692 * queue using rte_eth_add_rx_callback(). 3693 * 3694 * Note: the callback is removed from the callback list but it isn't freed 3695 * since the it may still be in use. The memory for the callback can be 3696 * subsequently freed back by the application by calling rte_free(): 3697 * 3698 * - Immediately - if the port is stopped, or the user knows that no 3699 * callbacks are in flight e.g. if called from the thread doing RX/TX 3700 * on that queue. 3701 * 3702 * - After a short delay - where the delay is sufficient to allow any 3703 * in-flight callbacks to complete. Alternately, the RCU mechanism can be 3704 * used to detect when data plane threads have ceased referencing the 3705 * callback memory. 3706 * 3707 * @param port_id 3708 * The port identifier of the Ethernet device. 3709 * @param queue_id 3710 * The queue on the Ethernet device from which the callback is to be removed. 3711 * @param user_cb 3712 * User supplied callback created via rte_eth_add_rx_callback(). 3713 * 3714 * @return 3715 * - 0: Success. Callback was removed. 3716 * - -ENOTSUP: Callback support is not available. 3717 * - -EINVAL: The port_id or the queue_id is out of range, or the callback 3718 * is NULL or not found for the port/queue. 3719 */ 3720 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id, 3721 const struct rte_eth_rxtx_callback *user_cb); 3722 3723 /** 3724 * Remove a TX packet callback from a given port and queue. 3725 * 3726 * This function is used to removed callbacks that were added to a NIC port 3727 * queue using rte_eth_add_tx_callback(). 3728 * 3729 * Note: the callback is removed from the callback list but it isn't freed 3730 * since the it may still be in use. The memory for the callback can be 3731 * subsequently freed back by the application by calling rte_free(): 3732 * 3733 * - Immediately - if the port is stopped, or the user knows that no 3734 * callbacks are in flight e.g. if called from the thread doing RX/TX 3735 * on that queue. 3736 * 3737 * - After a short delay - where the delay is sufficient to allow any 3738 * in-flight callbacks to complete. Alternately, the RCU mechanism can be 3739 * used to detect when data plane threads have ceased referencing the 3740 * callback memory. 3741 * 3742 * @param port_id 3743 * The port identifier of the Ethernet device. 3744 * @param queue_id 3745 * The queue on the Ethernet device from which the callback is to be removed. 3746 * @param user_cb 3747 * User supplied callback created via rte_eth_add_tx_callback(). 3748 * 3749 * @return 3750 * - 0: Success. Callback was removed. 3751 * - -ENOTSUP: Callback support is not available. 3752 * - -EINVAL: The port_id or the queue_id is out of range, or the callback 3753 * is NULL or not found for the port/queue. 3754 */ 3755 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id, 3756 const struct rte_eth_rxtx_callback *user_cb); 3757 3758 /** 3759 * Retrieve information about given port's RX queue. 3760 * 3761 * @param port_id 3762 * The port identifier of the Ethernet device. 3763 * @param queue_id 3764 * The RX queue on the Ethernet device for which information 3765 * will be retrieved. 3766 * @param qinfo 3767 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with 3768 * the information of the Ethernet device. 3769 * 3770 * @return 3771 * - 0: Success 3772 * - -ENOTSUP: routine is not supported by the device PMD. 3773 * - -EINVAL: The port_id or the queue_id is out of range, or the queue 3774 * is hairpin queue. 3775 */ 3776 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id, 3777 struct rte_eth_rxq_info *qinfo); 3778 3779 /** 3780 * Retrieve information about given port's TX queue. 3781 * 3782 * @param port_id 3783 * The port identifier of the Ethernet device. 3784 * @param queue_id 3785 * The TX queue on the Ethernet device for which information 3786 * will be retrieved. 3787 * @param qinfo 3788 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with 3789 * the information of the Ethernet device. 3790 * 3791 * @return 3792 * - 0: Success 3793 * - -ENOTSUP: routine is not supported by the device PMD. 3794 * - -EINVAL: The port_id or the queue_id is out of range, or the queue 3795 * is hairpin queue. 3796 */ 3797 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id, 3798 struct rte_eth_txq_info *qinfo); 3799 3800 /** 3801 * Retrieve information about the Rx packet burst mode. 3802 * 3803 * @param port_id 3804 * The port identifier of the Ethernet device. 3805 * @param queue_id 3806 * The Rx queue on the Ethernet device for which information 3807 * will be retrieved. 3808 * @param mode 3809 * A pointer to a structure of type *rte_eth_burst_mode* to be filled 3810 * with the information of the packet burst mode. 3811 * 3812 * @return 3813 * - 0: Success 3814 * - -ENOTSUP: routine is not supported by the device PMD. 3815 * - -EINVAL: The port_id or the queue_id is out of range. 3816 */ 3817 __rte_experimental 3818 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id, 3819 struct rte_eth_burst_mode *mode); 3820 3821 /** 3822 * Retrieve information about the Tx packet burst mode. 3823 * 3824 * @param port_id 3825 * The port identifier of the Ethernet device. 3826 * @param queue_id 3827 * The Tx queue on the Ethernet device for which information 3828 * will be retrieved. 3829 * @param mode 3830 * A pointer to a structure of type *rte_eth_burst_mode* to be filled 3831 * with the information of the packet burst mode. 3832 * 3833 * @return 3834 * - 0: Success 3835 * - -ENOTSUP: routine is not supported by the device PMD. 3836 * - -EINVAL: The port_id or the queue_id is out of range. 3837 */ 3838 __rte_experimental 3839 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id, 3840 struct rte_eth_burst_mode *mode); 3841 3842 /** 3843 * Retrieve device registers and register attributes (number of registers and 3844 * register size) 3845 * 3846 * @param port_id 3847 * The port identifier of the Ethernet device. 3848 * @param info 3849 * Pointer to rte_dev_reg_info structure to fill in. If info->data is 3850 * NULL the function fills in the width and length fields. If non-NULL 3851 * the registers are put into the buffer pointed at by the data field. 3852 * @return 3853 * - (0) if successful. 3854 * - (-ENOTSUP) if hardware doesn't support. 3855 * - (-ENODEV) if *port_id* invalid. 3856 * - (-EIO) if device is removed. 3857 * - others depends on the specific operations implementation. 3858 */ 3859 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info); 3860 3861 /** 3862 * Retrieve size of device EEPROM 3863 * 3864 * @param port_id 3865 * The port identifier of the Ethernet device. 3866 * @return 3867 * - (>=0) EEPROM size if successful. 3868 * - (-ENOTSUP) if hardware doesn't support. 3869 * - (-ENODEV) if *port_id* invalid. 3870 * - (-EIO) if device is removed. 3871 * - others depends on the specific operations implementation. 3872 */ 3873 int rte_eth_dev_get_eeprom_length(uint16_t port_id); 3874 3875 /** 3876 * Retrieve EEPROM and EEPROM attribute 3877 * 3878 * @param port_id 3879 * The port identifier of the Ethernet device. 3880 * @param info 3881 * The template includes buffer for return EEPROM data and 3882 * EEPROM attributes to be filled. 3883 * @return 3884 * - (0) if successful. 3885 * - (-ENOTSUP) if hardware doesn't support. 3886 * - (-ENODEV) if *port_id* invalid. 3887 * - (-EIO) if device is removed. 3888 * - others depends on the specific operations implementation. 3889 */ 3890 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info); 3891 3892 /** 3893 * Program EEPROM with provided data 3894 * 3895 * @param port_id 3896 * The port identifier of the Ethernet device. 3897 * @param info 3898 * The template includes EEPROM data for programming and 3899 * EEPROM attributes to be filled 3900 * @return 3901 * - (0) if successful. 3902 * - (-ENOTSUP) if hardware doesn't support. 3903 * - (-ENODEV) if *port_id* invalid. 3904 * - (-EIO) if device is removed. 3905 * - others depends on the specific operations implementation. 3906 */ 3907 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info); 3908 3909 /** 3910 * @warning 3911 * @b EXPERIMENTAL: this API may change without prior notice. 3912 * 3913 * Retrieve the type and size of plugin module EEPROM 3914 * 3915 * @param port_id 3916 * The port identifier of the Ethernet device. 3917 * @param modinfo 3918 * The type and size of plugin module EEPROM. 3919 * @return 3920 * - (0) if successful. 3921 * - (-ENOTSUP) if hardware doesn't support. 3922 * - (-ENODEV) if *port_id* invalid. 3923 * - (-EIO) if device is removed. 3924 * - others depends on the specific operations implementation. 3925 */ 3926 __rte_experimental 3927 int 3928 rte_eth_dev_get_module_info(uint16_t port_id, 3929 struct rte_eth_dev_module_info *modinfo); 3930 3931 /** 3932 * @warning 3933 * @b EXPERIMENTAL: this API may change without prior notice. 3934 * 3935 * Retrieve the data of plugin module EEPROM 3936 * 3937 * @param port_id 3938 * The port identifier of the Ethernet device. 3939 * @param info 3940 * The template includes the plugin module EEPROM attributes, and the 3941 * buffer for return plugin module EEPROM data. 3942 * @return 3943 * - (0) if successful. 3944 * - (-ENOTSUP) if hardware doesn't support. 3945 * - (-ENODEV) if *port_id* invalid. 3946 * - (-EIO) if device is removed. 3947 * - others depends on the specific operations implementation. 3948 */ 3949 __rte_experimental 3950 int 3951 rte_eth_dev_get_module_eeprom(uint16_t port_id, 3952 struct rte_dev_eeprom_info *info); 3953 3954 /** 3955 * Set the list of multicast addresses to filter on an Ethernet device. 3956 * 3957 * @param port_id 3958 * The port identifier of the Ethernet device. 3959 * @param mc_addr_set 3960 * The array of multicast addresses to set. Equal to NULL when the function 3961 * is invoked to flush the set of filtered addresses. 3962 * @param nb_mc_addr 3963 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0 3964 * when the function is invoked to flush the set of filtered addresses. 3965 * @return 3966 * - (0) if successful. 3967 * - (-ENODEV) if *port_id* invalid. 3968 * - (-EIO) if device is removed. 3969 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering. 3970 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources. 3971 */ 3972 int rte_eth_dev_set_mc_addr_list(uint16_t port_id, 3973 struct rte_ether_addr *mc_addr_set, 3974 uint32_t nb_mc_addr); 3975 3976 /** 3977 * Enable IEEE1588/802.1AS timestamping for an Ethernet device. 3978 * 3979 * @param port_id 3980 * The port identifier of the Ethernet device. 3981 * 3982 * @return 3983 * - 0: Success. 3984 * - -ENODEV: The port ID is invalid. 3985 * - -EIO: if device is removed. 3986 * - -ENOTSUP: The function is not supported by the Ethernet driver. 3987 */ 3988 int rte_eth_timesync_enable(uint16_t port_id); 3989 3990 /** 3991 * Disable IEEE1588/802.1AS timestamping for an Ethernet device. 3992 * 3993 * @param port_id 3994 * The port identifier of the Ethernet device. 3995 * 3996 * @return 3997 * - 0: Success. 3998 * - -ENODEV: The port ID is invalid. 3999 * - -EIO: if device is removed. 4000 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4001 */ 4002 int rte_eth_timesync_disable(uint16_t port_id); 4003 4004 /** 4005 * Read an IEEE1588/802.1AS RX timestamp from an Ethernet device. 4006 * 4007 * @param port_id 4008 * The port identifier of the Ethernet device. 4009 * @param timestamp 4010 * Pointer to the timestamp struct. 4011 * @param flags 4012 * Device specific flags. Used to pass the RX timesync register index to 4013 * i40e. Unused in igb/ixgbe, pass 0 instead. 4014 * 4015 * @return 4016 * - 0: Success. 4017 * - -EINVAL: No timestamp is available. 4018 * - -ENODEV: The port ID is invalid. 4019 * - -EIO: if device is removed. 4020 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4021 */ 4022 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id, 4023 struct timespec *timestamp, uint32_t flags); 4024 4025 /** 4026 * Read an IEEE1588/802.1AS TX timestamp from an Ethernet device. 4027 * 4028 * @param port_id 4029 * The port identifier of the Ethernet device. 4030 * @param timestamp 4031 * Pointer to the timestamp struct. 4032 * 4033 * @return 4034 * - 0: Success. 4035 * - -EINVAL: No timestamp is available. 4036 * - -ENODEV: The port ID is invalid. 4037 * - -EIO: if device is removed. 4038 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4039 */ 4040 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id, 4041 struct timespec *timestamp); 4042 4043 /** 4044 * Adjust the timesync clock on an Ethernet device. 4045 * 4046 * This is usually used in conjunction with other Ethdev timesync functions to 4047 * synchronize the device time using the IEEE1588/802.1AS protocol. 4048 * 4049 * @param port_id 4050 * The port identifier of the Ethernet device. 4051 * @param delta 4052 * The adjustment in nanoseconds. 4053 * 4054 * @return 4055 * - 0: Success. 4056 * - -ENODEV: The port ID is invalid. 4057 * - -EIO: if device is removed. 4058 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4059 */ 4060 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta); 4061 4062 /** 4063 * Read the time from the timesync clock on an Ethernet device. 4064 * 4065 * This is usually used in conjunction with other Ethdev timesync functions to 4066 * synchronize the device time using the IEEE1588/802.1AS protocol. 4067 * 4068 * @param port_id 4069 * The port identifier of the Ethernet device. 4070 * @param time 4071 * Pointer to the timespec struct that holds the time. 4072 * 4073 * @return 4074 * - 0: Success. 4075 */ 4076 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time); 4077 4078 /** 4079 * Set the time of the timesync clock on an Ethernet device. 4080 * 4081 * This is usually used in conjunction with other Ethdev timesync functions to 4082 * synchronize the device time using the IEEE1588/802.1AS protocol. 4083 * 4084 * @param port_id 4085 * The port identifier of the Ethernet device. 4086 * @param time 4087 * Pointer to the timespec struct that holds the time. 4088 * 4089 * @return 4090 * - 0: Success. 4091 * - -EINVAL: No timestamp is available. 4092 * - -ENODEV: The port ID is invalid. 4093 * - -EIO: if device is removed. 4094 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4095 */ 4096 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time); 4097 4098 /** 4099 * @warning 4100 * @b EXPERIMENTAL: this API may change without prior notice. 4101 * 4102 * Read the current clock counter of an Ethernet device 4103 * 4104 * This returns the current raw clock value of an Ethernet device. It is 4105 * a raw amount of ticks, with no given time reference. 4106 * The value returned here is from the same clock than the one 4107 * filling timestamp field of Rx packets when using hardware timestamp 4108 * offload. Therefore it can be used to compute a precise conversion of 4109 * the device clock to the real time. 4110 * 4111 * E.g, a simple heuristic to derivate the frequency would be: 4112 * uint64_t start, end; 4113 * rte_eth_read_clock(port, start); 4114 * rte_delay_ms(100); 4115 * rte_eth_read_clock(port, end); 4116 * double freq = (end - start) * 10; 4117 * 4118 * Compute a common reference with: 4119 * uint64_t base_time_sec = current_time(); 4120 * uint64_t base_clock; 4121 * rte_eth_read_clock(port, base_clock); 4122 * 4123 * Then, convert the raw mbuf timestamp with: 4124 * base_time_sec + (double)(mbuf->timestamp - base_clock) / freq; 4125 * 4126 * This simple example will not provide a very good accuracy. One must 4127 * at least measure multiple times the frequency and do a regression. 4128 * To avoid deviation from the system time, the common reference can 4129 * be repeated from time to time. The integer division can also be 4130 * converted by a multiplication and a shift for better performance. 4131 * 4132 * @param port_id 4133 * The port identifier of the Ethernet device. 4134 * @param clock 4135 * Pointer to the uint64_t that holds the raw clock value. 4136 * 4137 * @return 4138 * - 0: Success. 4139 * - -ENODEV: The port ID is invalid. 4140 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4141 */ 4142 __rte_experimental 4143 int 4144 rte_eth_read_clock(uint16_t port_id, uint64_t *clock); 4145 4146 /** 4147 * Config l2 tunnel ether type of an Ethernet device for filtering specific 4148 * tunnel packets by ether type. 4149 * 4150 * @param port_id 4151 * The port identifier of the Ethernet device. 4152 * @param l2_tunnel 4153 * l2 tunnel configuration. 4154 * 4155 * @return 4156 * - (0) if successful. 4157 * - (-ENODEV) if port identifier is invalid. 4158 * - (-EIO) if device is removed. 4159 * - (-ENOTSUP) if hardware doesn't support tunnel type. 4160 */ 4161 int 4162 rte_eth_dev_l2_tunnel_eth_type_conf(uint16_t port_id, 4163 struct rte_eth_l2_tunnel_conf *l2_tunnel); 4164 4165 /** 4166 * Enable/disable l2 tunnel offload functions. Include, 4167 * 1, The ability of parsing a type of l2 tunnel of an Ethernet device. 4168 * Filtering, forwarding and offloading this type of tunnel packets depend on 4169 * this ability. 4170 * 2, Stripping the l2 tunnel tag. 4171 * 3, Insertion of the l2 tunnel tag. 4172 * 4, Forwarding the packets based on the l2 tunnel tag. 4173 * 4174 * @param port_id 4175 * The port identifier of the Ethernet device. 4176 * @param l2_tunnel 4177 * l2 tunnel parameters. 4178 * @param mask 4179 * Indicate the offload function. 4180 * @param en 4181 * Enable or disable this function. 4182 * 4183 * @return 4184 * - (0) if successful. 4185 * - (-ENODEV) if port identifier is invalid. 4186 * - (-EIO) if device is removed. 4187 * - (-ENOTSUP) if hardware doesn't support tunnel type. 4188 */ 4189 int 4190 rte_eth_dev_l2_tunnel_offload_set(uint16_t port_id, 4191 struct rte_eth_l2_tunnel_conf *l2_tunnel, 4192 uint32_t mask, 4193 uint8_t en); 4194 4195 /** 4196 * Get the port id from device name. The device name should be specified 4197 * as below: 4198 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0 4199 * - SoC device name, for example- fsl-gmac0 4200 * - vdev dpdk name, for example- net_[pcap0|null0|tap0] 4201 * 4202 * @param name 4203 * pci address or name of the device 4204 * @param port_id 4205 * pointer to port identifier of the device 4206 * @return 4207 * - (0) if successful and port_id is filled. 4208 * - (-ENODEV or -EINVAL) on failure. 4209 */ 4210 int 4211 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id); 4212 4213 /** 4214 * Get the device name from port id. The device name is specified as below: 4215 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0 4216 * - SoC device name, for example- fsl-gmac0 4217 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0] 4218 * 4219 * @param port_id 4220 * Port identifier of the device. 4221 * @param name 4222 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name. 4223 * @return 4224 * - (0) if successful. 4225 * - (-EINVAL) on failure. 4226 */ 4227 int 4228 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name); 4229 4230 /** 4231 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from 4232 * the ethernet device information, otherwise adjust them to boundaries. 4233 * 4234 * @param port_id 4235 * The port identifier of the Ethernet device. 4236 * @param nb_rx_desc 4237 * A pointer to a uint16_t where the number of receive 4238 * descriptors stored. 4239 * @param nb_tx_desc 4240 * A pointer to a uint16_t where the number of transmit 4241 * descriptors stored. 4242 * @return 4243 * - (0) if successful. 4244 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure. 4245 */ 4246 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id, 4247 uint16_t *nb_rx_desc, 4248 uint16_t *nb_tx_desc); 4249 4250 /** 4251 * Test if a port supports specific mempool ops. 4252 * 4253 * @param port_id 4254 * Port identifier of the Ethernet device. 4255 * @param [in] pool 4256 * The name of the pool operations to test. 4257 * @return 4258 * - 0: best mempool ops choice for this port. 4259 * - 1: mempool ops are supported for this port. 4260 * - -ENOTSUP: mempool ops not supported for this port. 4261 * - -ENODEV: Invalid port Identifier. 4262 * - -EINVAL: Pool param is null. 4263 */ 4264 int 4265 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool); 4266 4267 /** 4268 * Get the security context for the Ethernet device. 4269 * 4270 * @param port_id 4271 * Port identifier of the Ethernet device 4272 * @return 4273 * - NULL on error. 4274 * - pointer to security context on success. 4275 */ 4276 void * 4277 rte_eth_dev_get_sec_ctx(uint16_t port_id); 4278 4279 /** 4280 * @warning 4281 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 4282 * 4283 * Query the device hairpin capabilities. 4284 * 4285 * @param port_id 4286 * The port identifier of the Ethernet device. 4287 * @param cap 4288 * Pointer to a structure that will hold the hairpin capabilities. 4289 * @return 4290 * - (0) if successful. 4291 * - (-ENOTSUP) if hardware doesn't support. 4292 */ 4293 __rte_experimental 4294 int rte_eth_dev_hairpin_capability_get(uint16_t port_id, 4295 struct rte_eth_hairpin_cap *cap); 4296 4297 #include <rte_ethdev_core.h> 4298 4299 /** 4300 * 4301 * Retrieve a burst of input packets from a receive queue of an Ethernet 4302 * device. The retrieved packets are stored in *rte_mbuf* structures whose 4303 * pointers are supplied in the *rx_pkts* array. 4304 * 4305 * The rte_eth_rx_burst() function loops, parsing the RX ring of the 4306 * receive queue, up to *nb_pkts* packets, and for each completed RX 4307 * descriptor in the ring, it performs the following operations: 4308 * 4309 * - Initialize the *rte_mbuf* data structure associated with the 4310 * RX descriptor according to the information provided by the NIC into 4311 * that RX descriptor. 4312 * 4313 * - Store the *rte_mbuf* data structure into the next entry of the 4314 * *rx_pkts* array. 4315 * 4316 * - Replenish the RX descriptor with a new *rte_mbuf* buffer 4317 * allocated from the memory pool associated with the receive queue at 4318 * initialization time. 4319 * 4320 * When retrieving an input packet that was scattered by the controller 4321 * into multiple receive descriptors, the rte_eth_rx_burst() function 4322 * appends the associated *rte_mbuf* buffers to the first buffer of the 4323 * packet. 4324 * 4325 * The rte_eth_rx_burst() function returns the number of packets 4326 * actually retrieved, which is the number of *rte_mbuf* data structures 4327 * effectively supplied into the *rx_pkts* array. 4328 * A return value equal to *nb_pkts* indicates that the RX queue contained 4329 * at least *rx_pkts* packets, and this is likely to signify that other 4330 * received packets remain in the input queue. Applications implementing 4331 * a "retrieve as much received packets as possible" policy can check this 4332 * specific case and keep invoking the rte_eth_rx_burst() function until 4333 * a value less than *nb_pkts* is returned. 4334 * 4335 * This receive method has the following advantages: 4336 * 4337 * - It allows a run-to-completion network stack engine to retrieve and 4338 * to immediately process received packets in a fast burst-oriented 4339 * approach, avoiding the overhead of unnecessary intermediate packet 4340 * queue/dequeue operations. 4341 * 4342 * - Conversely, it also allows an asynchronous-oriented processing 4343 * method to retrieve bursts of received packets and to immediately 4344 * queue them for further parallel processing by another logical core, 4345 * for instance. However, instead of having received packets being 4346 * individually queued by the driver, this approach allows the caller 4347 * of the rte_eth_rx_burst() function to queue a burst of retrieved 4348 * packets at a time and therefore dramatically reduce the cost of 4349 * enqueue/dequeue operations per packet. 4350 * 4351 * - It allows the rte_eth_rx_burst() function of the driver to take 4352 * advantage of burst-oriented hardware features (CPU cache, 4353 * prefetch instructions, and so on) to minimize the number of CPU 4354 * cycles per packet. 4355 * 4356 * To summarize, the proposed receive API enables many 4357 * burst-oriented optimizations in both synchronous and asynchronous 4358 * packet processing environments with no overhead in both cases. 4359 * 4360 * The rte_eth_rx_burst() function does not provide any error 4361 * notification to avoid the corresponding overhead. As a hint, the 4362 * upper-level application might check the status of the device link once 4363 * being systematically returned a 0 value for a given number of tries. 4364 * 4365 * @param port_id 4366 * The port identifier of the Ethernet device. 4367 * @param queue_id 4368 * The index of the receive queue from which to retrieve input packets. 4369 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 4370 * to rte_eth_dev_configure(). 4371 * @param rx_pkts 4372 * The address of an array of pointers to *rte_mbuf* structures that 4373 * must be large enough to store *nb_pkts* pointers in it. 4374 * @param nb_pkts 4375 * The maximum number of packets to retrieve. 4376 * @return 4377 * The number of packets actually retrieved, which is the number 4378 * of pointers to *rte_mbuf* structures effectively supplied to the 4379 * *rx_pkts* array. 4380 */ 4381 static inline uint16_t 4382 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id, 4383 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts) 4384 { 4385 struct rte_eth_dev *dev = &rte_eth_devices[port_id]; 4386 uint16_t nb_rx; 4387 4388 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4389 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0); 4390 RTE_FUNC_PTR_OR_ERR_RET(*dev->rx_pkt_burst, 0); 4391 4392 if (queue_id >= dev->data->nb_rx_queues) { 4393 RTE_ETHDEV_LOG(ERR, "Invalid RX queue_id=%u\n", queue_id); 4394 return 0; 4395 } 4396 #endif 4397 nb_rx = (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id], 4398 rx_pkts, nb_pkts); 4399 4400 #ifdef RTE_ETHDEV_RXTX_CALLBACKS 4401 struct rte_eth_rxtx_callback *cb; 4402 4403 /* __ATOMIC_RELEASE memory order was used when the 4404 * call back was inserted into the list. 4405 * Since there is a clear dependency between loading 4406 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is 4407 * not required. 4408 */ 4409 cb = __atomic_load_n(&dev->post_rx_burst_cbs[queue_id], 4410 __ATOMIC_RELAXED); 4411 4412 if (unlikely(cb != NULL)) { 4413 do { 4414 nb_rx = cb->fn.rx(port_id, queue_id, rx_pkts, nb_rx, 4415 nb_pkts, cb->param); 4416 cb = cb->next; 4417 } while (cb != NULL); 4418 } 4419 #endif 4420 4421 return nb_rx; 4422 } 4423 4424 /** 4425 * Get the number of used descriptors of a rx queue 4426 * 4427 * @param port_id 4428 * The port identifier of the Ethernet device. 4429 * @param queue_id 4430 * The queue id on the specific port. 4431 * @return 4432 * The number of used descriptors in the specific queue, or: 4433 * (-EINVAL) if *port_id* or *queue_id* is invalid 4434 * (-ENOTSUP) if the device does not support this function 4435 */ 4436 static inline int 4437 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id) 4438 { 4439 struct rte_eth_dev *dev; 4440 4441 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -EINVAL); 4442 dev = &rte_eth_devices[port_id]; 4443 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_queue_count, -ENOTSUP); 4444 if (queue_id >= dev->data->nb_rx_queues) 4445 return -EINVAL; 4446 4447 return (int)(*dev->dev_ops->rx_queue_count)(dev, queue_id); 4448 } 4449 4450 /** 4451 * Check if the DD bit of the specific RX descriptor in the queue has been set 4452 * 4453 * @param port_id 4454 * The port identifier of the Ethernet device. 4455 * @param queue_id 4456 * The queue id on the specific port. 4457 * @param offset 4458 * The offset of the descriptor ID from tail. 4459 * @return 4460 * - (1) if the specific DD bit is set. 4461 * - (0) if the specific DD bit is not set. 4462 * - (-ENODEV) if *port_id* invalid. 4463 * - (-ENOTSUP) if the device does not support this function 4464 */ 4465 static inline int 4466 rte_eth_rx_descriptor_done(uint16_t port_id, uint16_t queue_id, uint16_t offset) 4467 { 4468 struct rte_eth_dev *dev = &rte_eth_devices[port_id]; 4469 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV); 4470 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_done, -ENOTSUP); 4471 return (*dev->dev_ops->rx_descriptor_done)( \ 4472 dev->data->rx_queues[queue_id], offset); 4473 } 4474 4475 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */ 4476 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */ 4477 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */ 4478 4479 /** 4480 * Check the status of a Rx descriptor in the queue 4481 * 4482 * It should be called in a similar context than the Rx function: 4483 * - on a dataplane core 4484 * - not concurrently on the same queue 4485 * 4486 * Since it's a dataplane function, no check is performed on port_id and 4487 * queue_id. The caller must therefore ensure that the port is enabled 4488 * and the queue is configured and running. 4489 * 4490 * Note: accessing to a random descriptor in the ring may trigger cache 4491 * misses and have a performance impact. 4492 * 4493 * @param port_id 4494 * A valid port identifier of the Ethernet device which. 4495 * @param queue_id 4496 * A valid Rx queue identifier on this port. 4497 * @param offset 4498 * The offset of the descriptor starting from tail (0 is the next 4499 * packet to be received by the driver). 4500 * 4501 * @return 4502 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to 4503 * receive a packet. 4504 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but 4505 * not yet processed by the driver (i.e. in the receive queue). 4506 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by 4507 * the driver and not yet returned to hw, or reserved by the hw. 4508 * - (-EINVAL) bad descriptor offset. 4509 * - (-ENOTSUP) if the device does not support this function. 4510 * - (-ENODEV) bad port or queue (only if compiled with debug). 4511 */ 4512 static inline int 4513 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id, 4514 uint16_t offset) 4515 { 4516 struct rte_eth_dev *dev; 4517 void *rxq; 4518 4519 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4520 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV); 4521 #endif 4522 dev = &rte_eth_devices[port_id]; 4523 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4524 if (queue_id >= dev->data->nb_rx_queues) 4525 return -ENODEV; 4526 #endif 4527 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->rx_descriptor_status, -ENOTSUP); 4528 rxq = dev->data->rx_queues[queue_id]; 4529 4530 return (*dev->dev_ops->rx_descriptor_status)(rxq, offset); 4531 } 4532 4533 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */ 4534 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */ 4535 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */ 4536 4537 /** 4538 * Check the status of a Tx descriptor in the queue. 4539 * 4540 * It should be called in a similar context than the Tx function: 4541 * - on a dataplane core 4542 * - not concurrently on the same queue 4543 * 4544 * Since it's a dataplane function, no check is performed on port_id and 4545 * queue_id. The caller must therefore ensure that the port is enabled 4546 * and the queue is configured and running. 4547 * 4548 * Note: accessing to a random descriptor in the ring may trigger cache 4549 * misses and have a performance impact. 4550 * 4551 * @param port_id 4552 * A valid port identifier of the Ethernet device which. 4553 * @param queue_id 4554 * A valid Tx queue identifier on this port. 4555 * @param offset 4556 * The offset of the descriptor starting from tail (0 is the place where 4557 * the next packet will be send). 4558 * 4559 * @return 4560 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e. 4561 * in the transmit queue. 4562 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can 4563 * be reused by the driver. 4564 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the 4565 * driver or the hardware. 4566 * - (-EINVAL) bad descriptor offset. 4567 * - (-ENOTSUP) if the device does not support this function. 4568 * - (-ENODEV) bad port or queue (only if compiled with debug). 4569 */ 4570 static inline int rte_eth_tx_descriptor_status(uint16_t port_id, 4571 uint16_t queue_id, uint16_t offset) 4572 { 4573 struct rte_eth_dev *dev; 4574 void *txq; 4575 4576 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4577 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV); 4578 #endif 4579 dev = &rte_eth_devices[port_id]; 4580 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4581 if (queue_id >= dev->data->nb_tx_queues) 4582 return -ENODEV; 4583 #endif 4584 RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->tx_descriptor_status, -ENOTSUP); 4585 txq = dev->data->tx_queues[queue_id]; 4586 4587 return (*dev->dev_ops->tx_descriptor_status)(txq, offset); 4588 } 4589 4590 /** 4591 * Send a burst of output packets on a transmit queue of an Ethernet device. 4592 * 4593 * The rte_eth_tx_burst() function is invoked to transmit output packets 4594 * on the output queue *queue_id* of the Ethernet device designated by its 4595 * *port_id*. 4596 * The *nb_pkts* parameter is the number of packets to send which are 4597 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them 4598 * allocated from a pool created with rte_pktmbuf_pool_create(). 4599 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets, 4600 * up to the number of transmit descriptors available in the TX ring of the 4601 * transmit queue. 4602 * For each packet to send, the rte_eth_tx_burst() function performs 4603 * the following operations: 4604 * 4605 * - Pick up the next available descriptor in the transmit ring. 4606 * 4607 * - Free the network buffer previously sent with that descriptor, if any. 4608 * 4609 * - Initialize the transmit descriptor with the information provided 4610 * in the *rte_mbuf data structure. 4611 * 4612 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers, 4613 * the rte_eth_tx_burst() function uses several transmit descriptors 4614 * of the ring. 4615 * 4616 * The rte_eth_tx_burst() function returns the number of packets it 4617 * actually sent. A return value equal to *nb_pkts* means that all packets 4618 * have been sent, and this is likely to signify that other output packets 4619 * could be immediately transmitted again. Applications that implement a 4620 * "send as many packets to transmit as possible" policy can check this 4621 * specific case and keep invoking the rte_eth_tx_burst() function until 4622 * a value less than *nb_pkts* is returned. 4623 * 4624 * It is the responsibility of the rte_eth_tx_burst() function to 4625 * transparently free the memory buffers of packets previously sent. 4626 * This feature is driven by the *tx_free_thresh* value supplied to the 4627 * rte_eth_dev_configure() function at device configuration time. 4628 * When the number of free TX descriptors drops below this threshold, the 4629 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers 4630 * of those packets whose transmission was effectively completed. 4631 * 4632 * If the PMD is DEV_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can 4633 * invoke this function concurrently on the same tx queue without SW lock. 4634 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads 4635 * 4636 * @see rte_eth_tx_prepare to perform some prior checks or adjustments 4637 * for offloads. 4638 * 4639 * @param port_id 4640 * The port identifier of the Ethernet device. 4641 * @param queue_id 4642 * The index of the transmit queue through which output packets must be 4643 * sent. 4644 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 4645 * to rte_eth_dev_configure(). 4646 * @param tx_pkts 4647 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures 4648 * which contain the output packets. 4649 * @param nb_pkts 4650 * The maximum number of packets to transmit. 4651 * @return 4652 * The number of output packets actually stored in transmit descriptors of 4653 * the transmit ring. The return value can be less than the value of the 4654 * *tx_pkts* parameter when the transmit ring is full or has been filled up. 4655 */ 4656 static inline uint16_t 4657 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id, 4658 struct rte_mbuf **tx_pkts, uint16_t nb_pkts) 4659 { 4660 struct rte_eth_dev *dev = &rte_eth_devices[port_id]; 4661 4662 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4663 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0); 4664 RTE_FUNC_PTR_OR_ERR_RET(*dev->tx_pkt_burst, 0); 4665 4666 if (queue_id >= dev->data->nb_tx_queues) { 4667 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id); 4668 return 0; 4669 } 4670 #endif 4671 4672 #ifdef RTE_ETHDEV_RXTX_CALLBACKS 4673 struct rte_eth_rxtx_callback *cb; 4674 4675 /* __ATOMIC_RELEASE memory order was used when the 4676 * call back was inserted into the list. 4677 * Since there is a clear dependency between loading 4678 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is 4679 * not required. 4680 */ 4681 cb = __atomic_load_n(&dev->pre_tx_burst_cbs[queue_id], 4682 __ATOMIC_RELAXED); 4683 4684 if (unlikely(cb != NULL)) { 4685 do { 4686 nb_pkts = cb->fn.tx(port_id, queue_id, tx_pkts, nb_pkts, 4687 cb->param); 4688 cb = cb->next; 4689 } while (cb != NULL); 4690 } 4691 #endif 4692 4693 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts); 4694 } 4695 4696 /** 4697 * Process a burst of output packets on a transmit queue of an Ethernet device. 4698 * 4699 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be 4700 * transmitted on the output queue *queue_id* of the Ethernet device designated 4701 * by its *port_id*. 4702 * The *nb_pkts* parameter is the number of packets to be prepared which are 4703 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them 4704 * allocated from a pool created with rte_pktmbuf_pool_create(). 4705 * For each packet to send, the rte_eth_tx_prepare() function performs 4706 * the following operations: 4707 * 4708 * - Check if packet meets devices requirements for tx offloads. 4709 * 4710 * - Check limitations about number of segments. 4711 * 4712 * - Check additional requirements when debug is enabled. 4713 * 4714 * - Update and/or reset required checksums when tx offload is set for packet. 4715 * 4716 * Since this function can modify packet data, provided mbufs must be safely 4717 * writable (e.g. modified data cannot be in shared segment). 4718 * 4719 * The rte_eth_tx_prepare() function returns the number of packets ready to be 4720 * sent. A return value equal to *nb_pkts* means that all packets are valid and 4721 * ready to be sent, otherwise stops processing on the first invalid packet and 4722 * leaves the rest packets untouched. 4723 * 4724 * When this functionality is not implemented in the driver, all packets are 4725 * are returned untouched. 4726 * 4727 * @param port_id 4728 * The port identifier of the Ethernet device. 4729 * The value must be a valid port id. 4730 * @param queue_id 4731 * The index of the transmit queue through which output packets must be 4732 * sent. 4733 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 4734 * to rte_eth_dev_configure(). 4735 * @param tx_pkts 4736 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures 4737 * which contain the output packets. 4738 * @param nb_pkts 4739 * The maximum number of packets to process. 4740 * @return 4741 * The number of packets correct and ready to be sent. The return value can be 4742 * less than the value of the *tx_pkts* parameter when some packet doesn't 4743 * meet devices requirements with rte_errno set appropriately: 4744 * - EINVAL: offload flags are not correctly set 4745 * - ENOTSUP: the offload feature is not supported by the hardware 4746 * 4747 */ 4748 4749 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP 4750 4751 static inline uint16_t 4752 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id, 4753 struct rte_mbuf **tx_pkts, uint16_t nb_pkts) 4754 { 4755 struct rte_eth_dev *dev; 4756 4757 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4758 if (!rte_eth_dev_is_valid_port(port_id)) { 4759 RTE_ETHDEV_LOG(ERR, "Invalid TX port_id=%u\n", port_id); 4760 rte_errno = EINVAL; 4761 return 0; 4762 } 4763 #endif 4764 4765 dev = &rte_eth_devices[port_id]; 4766 4767 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 4768 if (queue_id >= dev->data->nb_tx_queues) { 4769 RTE_ETHDEV_LOG(ERR, "Invalid TX queue_id=%u\n", queue_id); 4770 rte_errno = EINVAL; 4771 return 0; 4772 } 4773 #endif 4774 4775 if (!dev->tx_pkt_prepare) 4776 return nb_pkts; 4777 4778 return (*dev->tx_pkt_prepare)(dev->data->tx_queues[queue_id], 4779 tx_pkts, nb_pkts); 4780 } 4781 4782 #else 4783 4784 /* 4785 * Native NOOP operation for compilation targets which doesn't require any 4786 * preparations steps, and functional NOOP may introduce unnecessary performance 4787 * drop. 4788 * 4789 * Generally this is not a good idea to turn it on globally and didn't should 4790 * be used if behavior of tx_preparation can change. 4791 */ 4792 4793 static inline uint16_t 4794 rte_eth_tx_prepare(__rte_unused uint16_t port_id, 4795 __rte_unused uint16_t queue_id, 4796 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts) 4797 { 4798 return nb_pkts; 4799 } 4800 4801 #endif 4802 4803 /** 4804 * Send any packets queued up for transmission on a port and HW queue 4805 * 4806 * This causes an explicit flush of packets previously buffered via the 4807 * rte_eth_tx_buffer() function. It returns the number of packets successfully 4808 * sent to the NIC, and calls the error callback for any unsent packets. Unless 4809 * explicitly set up otherwise, the default callback simply frees the unsent 4810 * packets back to the owning mempool. 4811 * 4812 * @param port_id 4813 * The port identifier of the Ethernet device. 4814 * @param queue_id 4815 * The index of the transmit queue through which output packets must be 4816 * sent. 4817 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 4818 * to rte_eth_dev_configure(). 4819 * @param buffer 4820 * Buffer of packets to be transmit. 4821 * @return 4822 * The number of packets successfully sent to the Ethernet device. The error 4823 * callback is called for any packets which could not be sent. 4824 */ 4825 static inline uint16_t 4826 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id, 4827 struct rte_eth_dev_tx_buffer *buffer) 4828 { 4829 uint16_t sent; 4830 uint16_t to_send = buffer->length; 4831 4832 if (to_send == 0) 4833 return 0; 4834 4835 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send); 4836 4837 buffer->length = 0; 4838 4839 /* All packets sent, or to be dealt with by callback below */ 4840 if (unlikely(sent != to_send)) 4841 buffer->error_callback(&buffer->pkts[sent], 4842 (uint16_t)(to_send - sent), 4843 buffer->error_userdata); 4844 4845 return sent; 4846 } 4847 4848 /** 4849 * Buffer a single packet for future transmission on a port and queue 4850 * 4851 * This function takes a single mbuf/packet and buffers it for later 4852 * transmission on the particular port and queue specified. Once the buffer is 4853 * full of packets, an attempt will be made to transmit all the buffered 4854 * packets. In case of error, where not all packets can be transmitted, a 4855 * callback is called with the unsent packets as a parameter. If no callback 4856 * is explicitly set up, the unsent packets are just freed back to the owning 4857 * mempool. The function returns the number of packets actually sent i.e. 4858 * 0 if no buffer flush occurred, otherwise the number of packets successfully 4859 * flushed 4860 * 4861 * @param port_id 4862 * The port identifier of the Ethernet device. 4863 * @param queue_id 4864 * The index of the transmit queue through which output packets must be 4865 * sent. 4866 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 4867 * to rte_eth_dev_configure(). 4868 * @param buffer 4869 * Buffer used to collect packets to be sent. 4870 * @param tx_pkt 4871 * Pointer to the packet mbuf to be sent. 4872 * @return 4873 * 0 = packet has been buffered for later transmission 4874 * N > 0 = packet has been buffered, and the buffer was subsequently flushed, 4875 * causing N packets to be sent, and the error callback to be called for 4876 * the rest. 4877 */ 4878 static __rte_always_inline uint16_t 4879 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id, 4880 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt) 4881 { 4882 buffer->pkts[buffer->length++] = tx_pkt; 4883 if (buffer->length < buffer->size) 4884 return 0; 4885 4886 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer); 4887 } 4888 4889 #ifdef __cplusplus 4890 } 4891 #endif 4892 4893 #endif /* _RTE_ETHDEV_H_ */ 4894