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 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 or the queue 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 * The following configuration may be retained or not 94 * depending on the device capabilities: 95 * 96 * - flow rules 97 * - flow-related shared objects, e.g. indirect actions 98 * 99 * Any other configuration will not be stored and will need to be re-entered 100 * before a call to rte_eth_dev_start(). 101 * 102 * Finally, a network application can close an Ethernet device by invoking the 103 * rte_eth_dev_close() function. 104 * 105 * Each function of the application Ethernet API invokes a specific function 106 * of the PMD that controls the target device designated by its port 107 * identifier. 108 * For this purpose, all device-specific functions of an Ethernet driver are 109 * supplied through a set of pointers contained in a generic structure of type 110 * *eth_dev_ops*. 111 * The address of the *eth_dev_ops* structure is stored in the *rte_eth_dev* 112 * structure by the device init function of the Ethernet driver, which is 113 * invoked during the PCI probing phase, as explained earlier. 114 * 115 * In other words, each function of the Ethernet API simply retrieves the 116 * *rte_eth_dev* structure associated with the device port identifier and 117 * performs an indirect invocation of the corresponding driver function 118 * supplied in the *eth_dev_ops* structure of the *rte_eth_dev* structure. 119 * 120 * For performance reasons, the address of the burst-oriented Rx and Tx 121 * functions of the Ethernet driver are not contained in the *eth_dev_ops* 122 * structure. Instead, they are directly stored at the beginning of the 123 * *rte_eth_dev* structure to avoid an extra indirect memory access during 124 * their invocation. 125 * 126 * RTE Ethernet device drivers do not use interrupts for transmitting or 127 * receiving. Instead, Ethernet drivers export Poll-Mode receive and transmit 128 * functions to applications. 129 * Both receive and transmit functions are packet-burst oriented to minimize 130 * their cost per packet through the following optimizations: 131 * 132 * - Sharing among multiple packets the incompressible cost of the 133 * invocation of receive/transmit functions. 134 * 135 * - Enabling receive/transmit functions to take advantage of burst-oriented 136 * hardware features (L1 cache, prefetch instructions, NIC head/tail 137 * registers) to minimize the number of CPU cycles per packet, for instance, 138 * by avoiding useless read memory accesses to ring descriptors, or by 139 * systematically using arrays of pointers that exactly fit L1 cache line 140 * boundaries and sizes. 141 * 142 * The burst-oriented receive function does not provide any error notification, 143 * to avoid the corresponding overhead. As a hint, the upper-level application 144 * might check the status of the device link once being systematically returned 145 * a 0 value by the receive function of the driver for a given number of tries. 146 */ 147 148 #ifdef __cplusplus 149 extern "C" { 150 #endif 151 152 #include <stdint.h> 153 154 /* Use this macro to check if LRO API is supported */ 155 #define RTE_ETHDEV_HAS_LRO_SUPPORT 156 157 /* Alias RTE_LIBRTE_ETHDEV_DEBUG for backward compatibility. */ 158 #ifdef RTE_LIBRTE_ETHDEV_DEBUG 159 #define RTE_ETHDEV_DEBUG_RX 160 #define RTE_ETHDEV_DEBUG_TX 161 #endif 162 163 #include <rte_compat.h> 164 #include <rte_log.h> 165 #include <rte_interrupts.h> 166 #include <rte_dev.h> 167 #include <rte_devargs.h> 168 #include <rte_bitops.h> 169 #include <rte_errno.h> 170 #include <rte_common.h> 171 #include <rte_config.h> 172 #include <rte_power_intrinsics.h> 173 174 #include "rte_ethdev_trace_fp.h" 175 #include "rte_dev_info.h" 176 177 extern int rte_eth_dev_logtype; 178 179 #define RTE_ETHDEV_LOG(level, ...) \ 180 rte_log(RTE_LOG_ ## level, rte_eth_dev_logtype, "" __VA_ARGS__) 181 182 struct rte_mbuf; 183 184 /** 185 * Initializes a device iterator. 186 * 187 * This iterator allows accessing a list of devices matching some devargs. 188 * 189 * @param iter 190 * Device iterator handle initialized by the function. 191 * The fields bus_str and cls_str might be dynamically allocated, 192 * and could be freed by calling rte_eth_iterator_cleanup(). 193 * 194 * @param devargs 195 * Device description string. 196 * 197 * @return 198 * 0 on successful initialization, negative otherwise. 199 */ 200 int rte_eth_iterator_init(struct rte_dev_iterator *iter, const char *devargs); 201 202 /** 203 * Iterates on devices with devargs filter. 204 * The ownership is not checked. 205 * 206 * The next port ID is returned, and the iterator is updated. 207 * 208 * @param iter 209 * Device iterator handle initialized by rte_eth_iterator_init(). 210 * Some fields bus_str and cls_str might be freed when no more port is found, 211 * by calling rte_eth_iterator_cleanup(). 212 * 213 * @return 214 * A port ID if found, RTE_MAX_ETHPORTS otherwise. 215 */ 216 uint16_t rte_eth_iterator_next(struct rte_dev_iterator *iter); 217 218 /** 219 * Free some allocated fields of the iterator. 220 * 221 * This function is automatically called by rte_eth_iterator_next() 222 * on the last iteration (i.e. when no more matching port is found). 223 * 224 * It is safe to call this function twice; it will do nothing more. 225 * 226 * @param iter 227 * Device iterator handle initialized by rte_eth_iterator_init(). 228 * The fields bus_str and cls_str are freed if needed. 229 */ 230 void rte_eth_iterator_cleanup(struct rte_dev_iterator *iter); 231 232 /** 233 * Macro to iterate over all ethdev ports matching some devargs. 234 * 235 * If a break is done before the end of the loop, 236 * the function rte_eth_iterator_cleanup() must be called. 237 * 238 * @param id 239 * Iterated port ID of type uint16_t. 240 * @param devargs 241 * Device parameters input as string of type char*. 242 * @param iter 243 * Iterator handle of type struct rte_dev_iterator, used internally. 244 */ 245 #define RTE_ETH_FOREACH_MATCHING_DEV(id, devargs, iter) \ 246 for (rte_eth_iterator_init(iter, devargs), \ 247 id = rte_eth_iterator_next(iter); \ 248 id != RTE_MAX_ETHPORTS; \ 249 id = rte_eth_iterator_next(iter)) 250 251 /** 252 * A structure used to retrieve statistics for an Ethernet port. 253 * Not all statistics fields in struct rte_eth_stats are supported 254 * by any type of network interface card (NIC). If any statistics 255 * field is not supported, its value is 0. 256 * All byte-related statistics do not include Ethernet FCS regardless 257 * of whether these bytes have been delivered to the application 258 * (see RTE_ETH_RX_OFFLOAD_KEEP_CRC). 259 */ 260 struct rte_eth_stats { 261 uint64_t ipackets; /**< Total number of successfully received packets. */ 262 uint64_t opackets; /**< Total number of successfully transmitted packets.*/ 263 uint64_t ibytes; /**< Total number of successfully received bytes. */ 264 uint64_t obytes; /**< Total number of successfully transmitted bytes. */ 265 /** 266 * Total of Rx packets dropped by the HW, 267 * because there are no available buffer (i.e. Rx queues are full). 268 */ 269 uint64_t imissed; 270 uint64_t ierrors; /**< Total number of erroneous received packets. */ 271 uint64_t oerrors; /**< Total number of failed transmitted packets. */ 272 uint64_t rx_nombuf; /**< Total number of Rx mbuf allocation failures. */ 273 /* Queue stats are limited to max 256 queues */ 274 /** Total number of queue Rx packets. */ 275 uint64_t q_ipackets[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 276 /** Total number of queue Tx packets. */ 277 uint64_t q_opackets[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 278 /** Total number of successfully received queue bytes. */ 279 uint64_t q_ibytes[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 280 /** Total number of successfully transmitted queue bytes. */ 281 uint64_t q_obytes[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 282 /** Total number of queue packets received that are dropped. */ 283 uint64_t q_errors[RTE_ETHDEV_QUEUE_STAT_CNTRS]; 284 }; 285 286 /**@{@name Link speed capabilities 287 * Device supported speeds bitmap flags 288 */ 289 #define RTE_ETH_LINK_SPEED_AUTONEG 0 /**< Autonegotiate (all speeds) */ 290 #define RTE_ETH_LINK_SPEED_FIXED RTE_BIT32(0) /**< Disable autoneg (fixed speed) */ 291 #define RTE_ETH_LINK_SPEED_10M_HD RTE_BIT32(1) /**< 10 Mbps half-duplex */ 292 #define RTE_ETH_LINK_SPEED_10M RTE_BIT32(2) /**< 10 Mbps full-duplex */ 293 #define RTE_ETH_LINK_SPEED_100M_HD RTE_BIT32(3) /**< 100 Mbps half-duplex */ 294 #define RTE_ETH_LINK_SPEED_100M RTE_BIT32(4) /**< 100 Mbps full-duplex */ 295 #define RTE_ETH_LINK_SPEED_1G RTE_BIT32(5) /**< 1 Gbps */ 296 #define RTE_ETH_LINK_SPEED_2_5G RTE_BIT32(6) /**< 2.5 Gbps */ 297 #define RTE_ETH_LINK_SPEED_5G RTE_BIT32(7) /**< 5 Gbps */ 298 #define RTE_ETH_LINK_SPEED_10G RTE_BIT32(8) /**< 10 Gbps */ 299 #define RTE_ETH_LINK_SPEED_20G RTE_BIT32(9) /**< 20 Gbps */ 300 #define RTE_ETH_LINK_SPEED_25G RTE_BIT32(10) /**< 25 Gbps */ 301 #define RTE_ETH_LINK_SPEED_40G RTE_BIT32(11) /**< 40 Gbps */ 302 #define RTE_ETH_LINK_SPEED_50G RTE_BIT32(12) /**< 50 Gbps */ 303 #define RTE_ETH_LINK_SPEED_56G RTE_BIT32(13) /**< 56 Gbps */ 304 #define RTE_ETH_LINK_SPEED_100G RTE_BIT32(14) /**< 100 Gbps */ 305 #define RTE_ETH_LINK_SPEED_200G RTE_BIT32(15) /**< 200 Gbps */ 306 /**@}*/ 307 308 /**@{@name Link speed 309 * Ethernet numeric link speeds in Mbps 310 */ 311 #define RTE_ETH_SPEED_NUM_NONE 0 /**< Not defined */ 312 #define RTE_ETH_SPEED_NUM_10M 10 /**< 10 Mbps */ 313 #define RTE_ETH_SPEED_NUM_100M 100 /**< 100 Mbps */ 314 #define RTE_ETH_SPEED_NUM_1G 1000 /**< 1 Gbps */ 315 #define RTE_ETH_SPEED_NUM_2_5G 2500 /**< 2.5 Gbps */ 316 #define RTE_ETH_SPEED_NUM_5G 5000 /**< 5 Gbps */ 317 #define RTE_ETH_SPEED_NUM_10G 10000 /**< 10 Gbps */ 318 #define RTE_ETH_SPEED_NUM_20G 20000 /**< 20 Gbps */ 319 #define RTE_ETH_SPEED_NUM_25G 25000 /**< 25 Gbps */ 320 #define RTE_ETH_SPEED_NUM_40G 40000 /**< 40 Gbps */ 321 #define RTE_ETH_SPEED_NUM_50G 50000 /**< 50 Gbps */ 322 #define RTE_ETH_SPEED_NUM_56G 56000 /**< 56 Gbps */ 323 #define RTE_ETH_SPEED_NUM_100G 100000 /**< 100 Gbps */ 324 #define RTE_ETH_SPEED_NUM_200G 200000 /**< 200 Gbps */ 325 #define RTE_ETH_SPEED_NUM_UNKNOWN UINT32_MAX /**< Unknown */ 326 /**@}*/ 327 328 /** 329 * A structure used to retrieve link-level information of an Ethernet port. 330 */ 331 __extension__ 332 struct rte_eth_link { 333 uint32_t link_speed; /**< RTE_ETH_SPEED_NUM_ */ 334 uint16_t link_duplex : 1; /**< RTE_ETH_LINK_[HALF/FULL]_DUPLEX */ 335 uint16_t link_autoneg : 1; /**< RTE_ETH_LINK_[AUTONEG/FIXED] */ 336 uint16_t link_status : 1; /**< RTE_ETH_LINK_[DOWN/UP] */ 337 } __rte_aligned(8); /**< aligned for atomic64 read/write */ 338 339 /**@{@name Link negotiation 340 * Constants used in link management. 341 */ 342 #define RTE_ETH_LINK_HALF_DUPLEX 0 /**< Half-duplex connection (see link_duplex). */ 343 #define RTE_ETH_LINK_FULL_DUPLEX 1 /**< Full-duplex connection (see link_duplex). */ 344 #define RTE_ETH_LINK_DOWN 0 /**< Link is down (see link_status). */ 345 #define RTE_ETH_LINK_UP 1 /**< Link is up (see link_status). */ 346 #define RTE_ETH_LINK_FIXED 0 /**< No autonegotiation (see link_autoneg). */ 347 #define RTE_ETH_LINK_AUTONEG 1 /**< Autonegotiated (see link_autoneg). */ 348 #define RTE_ETH_LINK_MAX_STR_LEN 40 /**< Max length of default link string. */ 349 /**@}*/ 350 351 /** 352 * A structure used to configure the ring threshold registers of an Rx/Tx 353 * queue for an Ethernet port. 354 */ 355 struct rte_eth_thresh { 356 uint8_t pthresh; /**< Ring prefetch threshold. */ 357 uint8_t hthresh; /**< Ring host threshold. */ 358 uint8_t wthresh; /**< Ring writeback threshold. */ 359 }; 360 361 /**@{@name Multi-queue mode 362 * @see rte_eth_conf.rxmode.mq_mode. 363 */ 364 #define RTE_ETH_MQ_RX_RSS_FLAG RTE_BIT32(0) /**< Enable RSS. @see rte_eth_rss_conf */ 365 #define RTE_ETH_MQ_RX_DCB_FLAG RTE_BIT32(1) /**< Enable DCB. */ 366 #define RTE_ETH_MQ_RX_VMDQ_FLAG RTE_BIT32(2) /**< Enable VMDq. */ 367 /**@}*/ 368 369 /** 370 * A set of values to identify what method is to be used to route 371 * packets to multiple queues. 372 */ 373 enum rte_eth_rx_mq_mode { 374 /** None of DCB, RSS or VMDq mode */ 375 RTE_ETH_MQ_RX_NONE = 0, 376 377 /** For Rx side, only RSS is on */ 378 RTE_ETH_MQ_RX_RSS = RTE_ETH_MQ_RX_RSS_FLAG, 379 /** For Rx side,only DCB is on. */ 380 RTE_ETH_MQ_RX_DCB = RTE_ETH_MQ_RX_DCB_FLAG, 381 /** Both DCB and RSS enable */ 382 RTE_ETH_MQ_RX_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG, 383 384 /** Only VMDq, no RSS nor DCB */ 385 RTE_ETH_MQ_RX_VMDQ_ONLY = RTE_ETH_MQ_RX_VMDQ_FLAG, 386 /** RSS mode with VMDq */ 387 RTE_ETH_MQ_RX_VMDQ_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_VMDQ_FLAG, 388 /** Use VMDq+DCB to route traffic to queues */ 389 RTE_ETH_MQ_RX_VMDQ_DCB = RTE_ETH_MQ_RX_VMDQ_FLAG | RTE_ETH_MQ_RX_DCB_FLAG, 390 /** Enable both VMDq and DCB in VMDq */ 391 RTE_ETH_MQ_RX_VMDQ_DCB_RSS = RTE_ETH_MQ_RX_RSS_FLAG | RTE_ETH_MQ_RX_DCB_FLAG | 392 RTE_ETH_MQ_RX_VMDQ_FLAG, 393 }; 394 395 /** 396 * A set of values to identify what method is to be used to transmit 397 * packets using multi-TCs. 398 */ 399 enum rte_eth_tx_mq_mode { 400 RTE_ETH_MQ_TX_NONE = 0, /**< It is in neither DCB nor VT mode. */ 401 RTE_ETH_MQ_TX_DCB, /**< For Tx side,only DCB is on. */ 402 RTE_ETH_MQ_TX_VMDQ_DCB, /**< For Tx side,both DCB and VT is on. */ 403 RTE_ETH_MQ_TX_VMDQ_ONLY, /**< Only VT on, no DCB */ 404 }; 405 406 /** 407 * A structure used to configure the Rx features of an Ethernet port. 408 */ 409 struct rte_eth_rxmode { 410 /** The multi-queue packet distribution mode to be used, e.g. RSS. */ 411 enum rte_eth_rx_mq_mode mq_mode; 412 uint32_t mtu; /**< Requested MTU. */ 413 /** Maximum allowed size of LRO aggregated packet. */ 414 uint32_t max_lro_pkt_size; 415 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/ 416 /** 417 * Per-port Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags. 418 * Only offloads set on rx_offload_capa field on rte_eth_dev_info 419 * structure are allowed to be set. 420 */ 421 uint64_t offloads; 422 423 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 424 void *reserved_ptrs[2]; /**< Reserved for future fields */ 425 }; 426 427 /** 428 * VLAN types to indicate if it is for single VLAN, inner VLAN or outer VLAN. 429 * Note that single VLAN is treated the same as inner VLAN. 430 */ 431 enum rte_vlan_type { 432 RTE_ETH_VLAN_TYPE_UNKNOWN = 0, 433 RTE_ETH_VLAN_TYPE_INNER, /**< Inner VLAN. */ 434 RTE_ETH_VLAN_TYPE_OUTER, /**< Single VLAN, or outer VLAN. */ 435 RTE_ETH_VLAN_TYPE_MAX, 436 }; 437 438 /** 439 * A structure used to describe a VLAN filter. 440 * If the bit corresponding to a VID is set, such VID is on. 441 */ 442 struct rte_vlan_filter_conf { 443 uint64_t ids[64]; 444 }; 445 446 /** 447 * A structure used to configure the Receive Side Scaling (RSS) feature 448 * of an Ethernet port. 449 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points 450 * to an array holding the RSS key to use for hashing specific header 451 * fields of received packets. The length of this array should be indicated 452 * by *rss_key_len* below. Otherwise, a default random hash key is used by 453 * the device driver. 454 * 455 * The *rss_key_len* field of the *rss_conf* structure indicates the length 456 * in bytes of the array pointed by *rss_key*. To be compatible, this length 457 * will be checked in i40e only. Others assume 40 bytes to be used as before. 458 * 459 * The *rss_hf* field of the *rss_conf* structure indicates the different 460 * types of IPv4/IPv6 packets to which the RSS hashing must be applied. 461 * Supplying an *rss_hf* equal to zero disables the RSS feature. 462 */ 463 struct rte_eth_rss_conf { 464 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */ 465 uint8_t rss_key_len; /**< hash key length in bytes. */ 466 uint64_t rss_hf; /**< Hash functions to apply - see below. */ 467 }; 468 469 /* 470 * A packet can be identified by hardware as different flow types. Different 471 * NIC hardware may support different flow types. 472 * Basically, the NIC hardware identifies the flow type as deep protocol as 473 * possible, and exclusively. For example, if a packet is identified as 474 * 'RTE_ETH_FLOW_NONFRAG_IPV4_TCP', it will not be any of other flow types, 475 * though it is an actual IPV4 packet. 476 */ 477 #define RTE_ETH_FLOW_UNKNOWN 0 478 #define RTE_ETH_FLOW_RAW 1 479 #define RTE_ETH_FLOW_IPV4 2 480 #define RTE_ETH_FLOW_FRAG_IPV4 3 481 #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP 4 482 #define RTE_ETH_FLOW_NONFRAG_IPV4_UDP 5 483 #define RTE_ETH_FLOW_NONFRAG_IPV4_SCTP 6 484 #define RTE_ETH_FLOW_NONFRAG_IPV4_OTHER 7 485 #define RTE_ETH_FLOW_IPV6 8 486 #define RTE_ETH_FLOW_FRAG_IPV6 9 487 #define RTE_ETH_FLOW_NONFRAG_IPV6_TCP 10 488 #define RTE_ETH_FLOW_NONFRAG_IPV6_UDP 11 489 #define RTE_ETH_FLOW_NONFRAG_IPV6_SCTP 12 490 #define RTE_ETH_FLOW_NONFRAG_IPV6_OTHER 13 491 #define RTE_ETH_FLOW_L2_PAYLOAD 14 492 #define RTE_ETH_FLOW_IPV6_EX 15 493 #define RTE_ETH_FLOW_IPV6_TCP_EX 16 494 #define RTE_ETH_FLOW_IPV6_UDP_EX 17 495 /** Consider device port number as a flow differentiator */ 496 #define RTE_ETH_FLOW_PORT 18 497 #define RTE_ETH_FLOW_VXLAN 19 /**< VXLAN protocol based flow */ 498 #define RTE_ETH_FLOW_GENEVE 20 /**< GENEVE protocol based flow */ 499 #define RTE_ETH_FLOW_NVGRE 21 /**< NVGRE protocol based flow */ 500 #define RTE_ETH_FLOW_VXLAN_GPE 22 /**< VXLAN-GPE protocol based flow */ 501 #define RTE_ETH_FLOW_GTPU 23 /**< GTPU protocol based flow */ 502 #define RTE_ETH_FLOW_MAX 24 503 504 /* 505 * Below macros are defined for RSS offload types, they can be used to 506 * fill rte_eth_rss_conf.rss_hf or rte_flow_action_rss.types. 507 */ 508 #define RTE_ETH_RSS_IPV4 RTE_BIT64(2) 509 #define RTE_ETH_RSS_FRAG_IPV4 RTE_BIT64(3) 510 #define RTE_ETH_RSS_NONFRAG_IPV4_TCP RTE_BIT64(4) 511 #define RTE_ETH_RSS_NONFRAG_IPV4_UDP RTE_BIT64(5) 512 #define RTE_ETH_RSS_NONFRAG_IPV4_SCTP RTE_BIT64(6) 513 #define RTE_ETH_RSS_NONFRAG_IPV4_OTHER RTE_BIT64(7) 514 #define RTE_ETH_RSS_IPV6 RTE_BIT64(8) 515 #define RTE_ETH_RSS_FRAG_IPV6 RTE_BIT64(9) 516 #define RTE_ETH_RSS_NONFRAG_IPV6_TCP RTE_BIT64(10) 517 #define RTE_ETH_RSS_NONFRAG_IPV6_UDP RTE_BIT64(11) 518 #define RTE_ETH_RSS_NONFRAG_IPV6_SCTP RTE_BIT64(12) 519 #define RTE_ETH_RSS_NONFRAG_IPV6_OTHER RTE_BIT64(13) 520 #define RTE_ETH_RSS_L2_PAYLOAD RTE_BIT64(14) 521 #define RTE_ETH_RSS_IPV6_EX RTE_BIT64(15) 522 #define RTE_ETH_RSS_IPV6_TCP_EX RTE_BIT64(16) 523 #define RTE_ETH_RSS_IPV6_UDP_EX RTE_BIT64(17) 524 #define RTE_ETH_RSS_PORT RTE_BIT64(18) 525 #define RTE_ETH_RSS_VXLAN RTE_BIT64(19) 526 #define RTE_ETH_RSS_GENEVE RTE_BIT64(20) 527 #define RTE_ETH_RSS_NVGRE RTE_BIT64(21) 528 #define RTE_ETH_RSS_GTPU RTE_BIT64(23) 529 #define RTE_ETH_RSS_ETH RTE_BIT64(24) 530 #define RTE_ETH_RSS_S_VLAN RTE_BIT64(25) 531 #define RTE_ETH_RSS_C_VLAN RTE_BIT64(26) 532 #define RTE_ETH_RSS_ESP RTE_BIT64(27) 533 #define RTE_ETH_RSS_AH RTE_BIT64(28) 534 #define RTE_ETH_RSS_L2TPV3 RTE_BIT64(29) 535 #define RTE_ETH_RSS_PFCP RTE_BIT64(30) 536 #define RTE_ETH_RSS_PPPOE RTE_BIT64(31) 537 #define RTE_ETH_RSS_ECPRI RTE_BIT64(32) 538 #define RTE_ETH_RSS_MPLS RTE_BIT64(33) 539 #define RTE_ETH_RSS_IPV4_CHKSUM RTE_BIT64(34) 540 541 /** 542 * The RTE_ETH_RSS_L4_CHKSUM works on checksum field of any L4 header. 543 * It is similar to RTE_ETH_RSS_PORT that they don't specify the specific type of 544 * L4 header. This macro is defined to replace some specific L4 (TCP/UDP/SCTP) 545 * checksum type for constructing the use of RSS offload bits. 546 * 547 * Due to above reason, some old APIs (and configuration) don't support 548 * RTE_ETH_RSS_L4_CHKSUM. The rte_flow RSS API supports it. 549 * 550 * For the case that checksum is not used in an UDP header, 551 * it takes the reserved value 0 as input for the hash function. 552 */ 553 #define RTE_ETH_RSS_L4_CHKSUM RTE_BIT64(35) 554 555 #define RTE_ETH_RSS_L2TPV2 RTE_BIT64(36) 556 557 /* 558 * We use the following macros to combine with above RTE_ETH_RSS_* for 559 * more specific input set selection. These bits are defined starting 560 * from the high end of the 64 bits. 561 * Note: If we use above RTE_ETH_RSS_* without SRC/DST_ONLY, it represents 562 * both SRC and DST are taken into account. If SRC_ONLY and DST_ONLY of 563 * the same level are used simultaneously, it is the same case as none of 564 * them are added. 565 */ 566 #define RTE_ETH_RSS_L3_SRC_ONLY RTE_BIT64(63) 567 #define RTE_ETH_RSS_L3_DST_ONLY RTE_BIT64(62) 568 #define RTE_ETH_RSS_L4_SRC_ONLY RTE_BIT64(61) 569 #define RTE_ETH_RSS_L4_DST_ONLY RTE_BIT64(60) 570 #define RTE_ETH_RSS_L2_SRC_ONLY RTE_BIT64(59) 571 #define RTE_ETH_RSS_L2_DST_ONLY RTE_BIT64(58) 572 573 /* 574 * Only select IPV6 address prefix as RSS input set according to 575 * https://tools.ietf.org/html/rfc6052 576 * Must be combined with RTE_ETH_RSS_IPV6, RTE_ETH_RSS_NONFRAG_IPV6_UDP, 577 * RTE_ETH_RSS_NONFRAG_IPV6_TCP, RTE_ETH_RSS_NONFRAG_IPV6_SCTP. 578 */ 579 #define RTE_ETH_RSS_L3_PRE32 RTE_BIT64(57) 580 #define RTE_ETH_RSS_L3_PRE40 RTE_BIT64(56) 581 #define RTE_ETH_RSS_L3_PRE48 RTE_BIT64(55) 582 #define RTE_ETH_RSS_L3_PRE56 RTE_BIT64(54) 583 #define RTE_ETH_RSS_L3_PRE64 RTE_BIT64(53) 584 #define RTE_ETH_RSS_L3_PRE96 RTE_BIT64(52) 585 586 /* 587 * Use the following macros to combine with the above layers 588 * to choose inner and outer layers or both for RSS computation. 589 * Bits 50 and 51 are reserved for this. 590 */ 591 592 /** 593 * level 0, requests the default behavior. 594 * Depending on the packet type, it can mean outermost, innermost, 595 * anything in between or even no RSS. 596 * It basically stands for the innermost encapsulation level RSS 597 * can be performed on according to PMD and device capabilities. 598 */ 599 #define RTE_ETH_RSS_LEVEL_PMD_DEFAULT (UINT64_C(0) << 50) 600 601 /** 602 * level 1, requests RSS to be performed on the outermost packet 603 * encapsulation level. 604 */ 605 #define RTE_ETH_RSS_LEVEL_OUTERMOST (UINT64_C(1) << 50) 606 607 /** 608 * level 2, requests RSS to be performed on the specified inner packet 609 * encapsulation level, from outermost to innermost (lower to higher values). 610 */ 611 #define RTE_ETH_RSS_LEVEL_INNERMOST (UINT64_C(2) << 50) 612 #define RTE_ETH_RSS_LEVEL_MASK (UINT64_C(3) << 50) 613 614 #define RTE_ETH_RSS_LEVEL(rss_hf) ((rss_hf & RTE_ETH_RSS_LEVEL_MASK) >> 50) 615 616 /** 617 * For input set change of hash filter, if SRC_ONLY and DST_ONLY of 618 * the same level are used simultaneously, it is the same case as 619 * none of them are added. 620 * 621 * @param rss_hf 622 * RSS types with SRC/DST_ONLY. 623 * @return 624 * RSS types. 625 */ 626 static inline uint64_t 627 rte_eth_rss_hf_refine(uint64_t rss_hf) 628 { 629 if ((rss_hf & RTE_ETH_RSS_L3_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L3_DST_ONLY)) 630 rss_hf &= ~(RTE_ETH_RSS_L3_SRC_ONLY | RTE_ETH_RSS_L3_DST_ONLY); 631 632 if ((rss_hf & RTE_ETH_RSS_L4_SRC_ONLY) && (rss_hf & RTE_ETH_RSS_L4_DST_ONLY)) 633 rss_hf &= ~(RTE_ETH_RSS_L4_SRC_ONLY | RTE_ETH_RSS_L4_DST_ONLY); 634 635 return rss_hf; 636 } 637 638 #define RTE_ETH_RSS_IPV6_PRE32 ( \ 639 RTE_ETH_RSS_IPV6 | \ 640 RTE_ETH_RSS_L3_PRE32) 641 642 #define RTE_ETH_RSS_IPV6_PRE40 ( \ 643 RTE_ETH_RSS_IPV6 | \ 644 RTE_ETH_RSS_L3_PRE40) 645 646 #define RTE_ETH_RSS_IPV6_PRE48 ( \ 647 RTE_ETH_RSS_IPV6 | \ 648 RTE_ETH_RSS_L3_PRE48) 649 650 #define RTE_ETH_RSS_IPV6_PRE56 ( \ 651 RTE_ETH_RSS_IPV6 | \ 652 RTE_ETH_RSS_L3_PRE56) 653 654 #define RTE_ETH_RSS_IPV6_PRE64 ( \ 655 RTE_ETH_RSS_IPV6 | \ 656 RTE_ETH_RSS_L3_PRE64) 657 658 #define RTE_ETH_RSS_IPV6_PRE96 ( \ 659 RTE_ETH_RSS_IPV6 | \ 660 RTE_ETH_RSS_L3_PRE96) 661 662 #define RTE_ETH_RSS_IPV6_PRE32_UDP ( \ 663 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 664 RTE_ETH_RSS_L3_PRE32) 665 666 #define RTE_ETH_RSS_IPV6_PRE40_UDP ( \ 667 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 668 RTE_ETH_RSS_L3_PRE40) 669 670 #define RTE_ETH_RSS_IPV6_PRE48_UDP ( \ 671 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 672 RTE_ETH_RSS_L3_PRE48) 673 674 #define RTE_ETH_RSS_IPV6_PRE56_UDP ( \ 675 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 676 RTE_ETH_RSS_L3_PRE56) 677 678 #define RTE_ETH_RSS_IPV6_PRE64_UDP ( \ 679 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 680 RTE_ETH_RSS_L3_PRE64) 681 682 #define RTE_ETH_RSS_IPV6_PRE96_UDP ( \ 683 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 684 RTE_ETH_RSS_L3_PRE96) 685 686 #define RTE_ETH_RSS_IPV6_PRE32_TCP ( \ 687 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 688 RTE_ETH_RSS_L3_PRE32) 689 690 #define RTE_ETH_RSS_IPV6_PRE40_TCP ( \ 691 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 692 RTE_ETH_RSS_L3_PRE40) 693 694 #define RTE_ETH_RSS_IPV6_PRE48_TCP ( \ 695 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 696 RTE_ETH_RSS_L3_PRE48) 697 698 #define RTE_ETH_RSS_IPV6_PRE56_TCP ( \ 699 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 700 RTE_ETH_RSS_L3_PRE56) 701 702 #define RTE_ETH_RSS_IPV6_PRE64_TCP ( \ 703 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 704 RTE_ETH_RSS_L3_PRE64) 705 706 #define RTE_ETH_RSS_IPV6_PRE96_TCP ( \ 707 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 708 RTE_ETH_RSS_L3_PRE96) 709 710 #define RTE_ETH_RSS_IPV6_PRE32_SCTP ( \ 711 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 712 RTE_ETH_RSS_L3_PRE32) 713 714 #define RTE_ETH_RSS_IPV6_PRE40_SCTP ( \ 715 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 716 RTE_ETH_RSS_L3_PRE40) 717 718 #define RTE_ETH_RSS_IPV6_PRE48_SCTP ( \ 719 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 720 RTE_ETH_RSS_L3_PRE48) 721 722 #define RTE_ETH_RSS_IPV6_PRE56_SCTP ( \ 723 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 724 RTE_ETH_RSS_L3_PRE56) 725 726 #define RTE_ETH_RSS_IPV6_PRE64_SCTP ( \ 727 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 728 RTE_ETH_RSS_L3_PRE64) 729 730 #define RTE_ETH_RSS_IPV6_PRE96_SCTP ( \ 731 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 732 RTE_ETH_RSS_L3_PRE96) 733 734 #define RTE_ETH_RSS_IP ( \ 735 RTE_ETH_RSS_IPV4 | \ 736 RTE_ETH_RSS_FRAG_IPV4 | \ 737 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \ 738 RTE_ETH_RSS_IPV6 | \ 739 RTE_ETH_RSS_FRAG_IPV6 | \ 740 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \ 741 RTE_ETH_RSS_IPV6_EX) 742 743 #define RTE_ETH_RSS_UDP ( \ 744 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \ 745 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 746 RTE_ETH_RSS_IPV6_UDP_EX) 747 748 #define RTE_ETH_RSS_TCP ( \ 749 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \ 750 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 751 RTE_ETH_RSS_IPV6_TCP_EX) 752 753 #define RTE_ETH_RSS_SCTP ( \ 754 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \ 755 RTE_ETH_RSS_NONFRAG_IPV6_SCTP) 756 757 #define RTE_ETH_RSS_TUNNEL ( \ 758 RTE_ETH_RSS_VXLAN | \ 759 RTE_ETH_RSS_GENEVE | \ 760 RTE_ETH_RSS_NVGRE) 761 762 #define RTE_ETH_RSS_VLAN ( \ 763 RTE_ETH_RSS_S_VLAN | \ 764 RTE_ETH_RSS_C_VLAN) 765 766 /** Mask of valid RSS hash protocols */ 767 #define RTE_ETH_RSS_PROTO_MASK ( \ 768 RTE_ETH_RSS_IPV4 | \ 769 RTE_ETH_RSS_FRAG_IPV4 | \ 770 RTE_ETH_RSS_NONFRAG_IPV4_TCP | \ 771 RTE_ETH_RSS_NONFRAG_IPV4_UDP | \ 772 RTE_ETH_RSS_NONFRAG_IPV4_SCTP | \ 773 RTE_ETH_RSS_NONFRAG_IPV4_OTHER | \ 774 RTE_ETH_RSS_IPV6 | \ 775 RTE_ETH_RSS_FRAG_IPV6 | \ 776 RTE_ETH_RSS_NONFRAG_IPV6_TCP | \ 777 RTE_ETH_RSS_NONFRAG_IPV6_UDP | \ 778 RTE_ETH_RSS_NONFRAG_IPV6_SCTP | \ 779 RTE_ETH_RSS_NONFRAG_IPV6_OTHER | \ 780 RTE_ETH_RSS_L2_PAYLOAD | \ 781 RTE_ETH_RSS_IPV6_EX | \ 782 RTE_ETH_RSS_IPV6_TCP_EX | \ 783 RTE_ETH_RSS_IPV6_UDP_EX | \ 784 RTE_ETH_RSS_PORT | \ 785 RTE_ETH_RSS_VXLAN | \ 786 RTE_ETH_RSS_GENEVE | \ 787 RTE_ETH_RSS_NVGRE | \ 788 RTE_ETH_RSS_MPLS) 789 790 /* 791 * Definitions used for redirection table entry size. 792 * Some RSS RETA sizes may not be supported by some drivers, check the 793 * documentation or the description of relevant functions for more details. 794 */ 795 #define RTE_ETH_RSS_RETA_SIZE_64 64 796 #define RTE_ETH_RSS_RETA_SIZE_128 128 797 #define RTE_ETH_RSS_RETA_SIZE_256 256 798 #define RTE_ETH_RSS_RETA_SIZE_512 512 799 #define RTE_ETH_RETA_GROUP_SIZE 64 800 801 /**@{@name VMDq and DCB maximums */ 802 #define RTE_ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDq VLAN filters. */ 803 #define RTE_ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */ 804 #define RTE_ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDq DCB queues. */ 805 #define RTE_ETH_DCB_NUM_QUEUES 128 /**< Maximum nb. of DCB queues. */ 806 /**@}*/ 807 808 /**@{@name DCB capabilities */ 809 #define RTE_ETH_DCB_PG_SUPPORT RTE_BIT32(0) /**< Priority Group(ETS) support. */ 810 #define RTE_ETH_DCB_PFC_SUPPORT RTE_BIT32(1) /**< Priority Flow Control support. */ 811 /**@}*/ 812 813 /**@{@name VLAN offload bits */ 814 #define RTE_ETH_VLAN_STRIP_OFFLOAD 0x0001 /**< VLAN Strip On/Off */ 815 #define RTE_ETH_VLAN_FILTER_OFFLOAD 0x0002 /**< VLAN Filter On/Off */ 816 #define RTE_ETH_VLAN_EXTEND_OFFLOAD 0x0004 /**< VLAN Extend On/Off */ 817 #define RTE_ETH_QINQ_STRIP_OFFLOAD 0x0008 /**< QINQ Strip On/Off */ 818 819 #define RTE_ETH_VLAN_STRIP_MASK 0x0001 /**< VLAN Strip setting mask */ 820 #define RTE_ETH_VLAN_FILTER_MASK 0x0002 /**< VLAN Filter setting mask*/ 821 #define RTE_ETH_VLAN_EXTEND_MASK 0x0004 /**< VLAN Extend setting mask*/ 822 #define RTE_ETH_QINQ_STRIP_MASK 0x0008 /**< QINQ Strip setting mask */ 823 #define RTE_ETH_VLAN_ID_MAX 0x0FFF /**< VLAN ID is in lower 12 bits*/ 824 /**@}*/ 825 826 /* Definitions used for receive MAC address */ 827 #define RTE_ETH_NUM_RECEIVE_MAC_ADDR 128 /**< Maximum nb. of receive mac addr. */ 828 829 /* Definitions used for unicast hash */ 830 #define RTE_ETH_VMDQ_NUM_UC_HASH_ARRAY 128 /**< Maximum nb. of UC hash array. */ 831 832 /**@{@name VMDq Rx mode 833 * @see rte_eth_vmdq_rx_conf.rx_mode 834 */ 835 /** Accept untagged packets. */ 836 #define RTE_ETH_VMDQ_ACCEPT_UNTAG RTE_BIT32(0) 837 /** Accept packets in multicast table. */ 838 #define RTE_ETH_VMDQ_ACCEPT_HASH_MC RTE_BIT32(1) 839 /** Accept packets in unicast table. */ 840 #define RTE_ETH_VMDQ_ACCEPT_HASH_UC RTE_BIT32(2) 841 /** Accept broadcast packets. */ 842 #define RTE_ETH_VMDQ_ACCEPT_BROADCAST RTE_BIT32(3) 843 /** Multicast promiscuous. */ 844 #define RTE_ETH_VMDQ_ACCEPT_MULTICAST RTE_BIT32(4) 845 /**@}*/ 846 847 /** 848 * A structure used to configure 64 entries of Redirection Table of the 849 * Receive Side Scaling (RSS) feature of an Ethernet port. To configure 850 * more than 64 entries supported by hardware, an array of this structure 851 * is needed. 852 */ 853 struct rte_eth_rss_reta_entry64 { 854 /** Mask bits indicate which entries need to be updated/queried. */ 855 uint64_t mask; 856 /** Group of 64 redirection table entries. */ 857 uint16_t reta[RTE_ETH_RETA_GROUP_SIZE]; 858 }; 859 860 /** 861 * This enum indicates the possible number of traffic classes 862 * in DCB configurations 863 */ 864 enum rte_eth_nb_tcs { 865 RTE_ETH_4_TCS = 4, /**< 4 TCs with DCB. */ 866 RTE_ETH_8_TCS = 8 /**< 8 TCs with DCB. */ 867 }; 868 869 /** 870 * This enum indicates the possible number of queue pools 871 * in VMDq configurations. 872 */ 873 enum rte_eth_nb_pools { 874 RTE_ETH_8_POOLS = 8, /**< 8 VMDq pools. */ 875 RTE_ETH_16_POOLS = 16, /**< 16 VMDq pools. */ 876 RTE_ETH_32_POOLS = 32, /**< 32 VMDq pools. */ 877 RTE_ETH_64_POOLS = 64 /**< 64 VMDq pools. */ 878 }; 879 880 /* This structure may be extended in future. */ 881 struct rte_eth_dcb_rx_conf { 882 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs */ 883 /** Traffic class each UP mapped to. */ 884 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; 885 }; 886 887 struct rte_eth_vmdq_dcb_tx_conf { 888 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools. */ 889 /** Traffic class each UP mapped to. */ 890 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; 891 }; 892 893 struct rte_eth_dcb_tx_conf { 894 enum rte_eth_nb_tcs nb_tcs; /**< Possible DCB TCs, 4 or 8 TCs. */ 895 /** Traffic class each UP mapped to. */ 896 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; 897 }; 898 899 struct rte_eth_vmdq_tx_conf { 900 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq mode, 64 pools. */ 901 }; 902 903 /** 904 * A structure used to configure the VMDq+DCB feature 905 * of an Ethernet port. 906 * 907 * Using this feature, packets are routed to a pool of queues, based 908 * on the VLAN ID in the VLAN tag, and then to a specific queue within 909 * that pool, using the user priority VLAN tag field. 910 * 911 * A default pool may be used, if desired, to route all traffic which 912 * does not match the VLAN filter rules. 913 */ 914 struct rte_eth_vmdq_dcb_conf { 915 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */ 916 uint8_t enable_default_pool; /**< If non-zero, use a default pool */ 917 uint8_t default_pool; /**< The default pool, if applicable */ 918 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */ 919 struct { 920 uint16_t vlan_id; /**< The VLAN ID of the received frame */ 921 uint64_t pools; /**< Bitmask of pools for packet Rx */ 922 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */ 923 /** Selects a queue in a pool */ 924 uint8_t dcb_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; 925 }; 926 927 /** 928 * A structure used to configure the VMDq feature of an Ethernet port when 929 * not combined with the DCB feature. 930 * 931 * Using this feature, packets are routed to a pool of queues. By default, 932 * the pool selection is based on the MAC address, the VLAN ID in the 933 * VLAN tag as specified in the pool_map array. 934 * Passing the RTE_ETH_VMDQ_ACCEPT_UNTAG in the rx_mode field allows pool 935 * selection using only the MAC address. MAC address to pool mapping is done 936 * using the rte_eth_dev_mac_addr_add function, with the pool parameter 937 * corresponding to the pool ID. 938 * 939 * Queue selection within the selected pool will be done using RSS when 940 * it is enabled or revert to the first queue of the pool if not. 941 * 942 * A default pool may be used, if desired, to route all traffic which 943 * does not match the VLAN filter rules or any pool MAC address. 944 */ 945 struct rte_eth_vmdq_rx_conf { 946 enum rte_eth_nb_pools nb_queue_pools; /**< VMDq only mode, 8 or 64 pools */ 947 uint8_t enable_default_pool; /**< If non-zero, use a default pool */ 948 uint8_t default_pool; /**< The default pool, if applicable */ 949 uint8_t enable_loop_back; /**< Enable VT loop back */ 950 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */ 951 uint32_t rx_mode; /**< Flags from RTE_ETH_VMDQ_ACCEPT_* */ 952 struct { 953 uint16_t vlan_id; /**< The VLAN ID of the received frame */ 954 uint64_t pools; /**< Bitmask of pools for packet Rx */ 955 } pool_map[RTE_ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq VLAN pool maps. */ 956 }; 957 958 /** 959 * A structure used to configure the Tx features of an Ethernet port. 960 */ 961 struct rte_eth_txmode { 962 enum rte_eth_tx_mq_mode mq_mode; /**< Tx multi-queues mode. */ 963 /** 964 * Per-port Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags. 965 * Only offloads set on tx_offload_capa field on rte_eth_dev_info 966 * structure are allowed to be set. 967 */ 968 uint64_t offloads; 969 970 uint16_t pvid; 971 __extension__ 972 uint8_t /** If set, reject sending out tagged pkts */ 973 hw_vlan_reject_tagged : 1, 974 /** If set, reject sending out untagged pkts */ 975 hw_vlan_reject_untagged : 1, 976 /** If set, enable port based VLAN insertion */ 977 hw_vlan_insert_pvid : 1; 978 979 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 980 void *reserved_ptrs[2]; /**< Reserved for future fields */ 981 }; 982 983 /** 984 * @warning 985 * @b EXPERIMENTAL: this structure may change without prior notice. 986 * 987 * A structure used to configure an Rx packet segment to split. 988 * 989 * If RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT flag is set in offloads field, 990 * the PMD will split the received packets into multiple segments 991 * according to the specification in the description array: 992 * 993 * - The first network buffer will be allocated from the memory pool, 994 * specified in the first array element, the second buffer, from the 995 * pool in the second element, and so on. 996 * 997 * - The offsets from the segment description elements specify 998 * the data offset from the buffer beginning except the first mbuf. 999 * The first segment offset is added with RTE_PKTMBUF_HEADROOM. 1000 * 1001 * - The lengths in the elements define the maximal data amount 1002 * being received to each segment. The receiving starts with filling 1003 * up the first mbuf data buffer up to specified length. If the 1004 * there are data remaining (packet is longer than buffer in the first 1005 * mbuf) the following data will be pushed to the next segment 1006 * up to its own length, and so on. 1007 * 1008 * - If the length in the segment description element is zero 1009 * the actual buffer size will be deduced from the appropriate 1010 * memory pool properties. 1011 * 1012 * - If there is not enough elements to describe the buffer for entire 1013 * packet of maximal length the following parameters will be used 1014 * for the all remaining segments: 1015 * - pool from the last valid element 1016 * - the buffer size from this pool 1017 * - zero offset 1018 */ 1019 struct rte_eth_rxseg_split { 1020 struct rte_mempool *mp; /**< Memory pool to allocate segment from. */ 1021 uint16_t length; /**< Segment data length, configures split point. */ 1022 uint16_t offset; /**< Data offset from beginning of mbuf data buffer. */ 1023 uint32_t reserved; /**< Reserved field. */ 1024 }; 1025 1026 /** 1027 * @warning 1028 * @b EXPERIMENTAL: this structure may change without prior notice. 1029 * 1030 * A common structure used to describe Rx packet segment properties. 1031 */ 1032 union rte_eth_rxseg { 1033 /* The settings for buffer split offload. */ 1034 struct rte_eth_rxseg_split split; 1035 /* The other features settings should be added here. */ 1036 }; 1037 1038 /** 1039 * A structure used to configure an Rx ring of an Ethernet port. 1040 */ 1041 struct rte_eth_rxconf { 1042 struct rte_eth_thresh rx_thresh; /**< Rx ring threshold registers. */ 1043 uint16_t rx_free_thresh; /**< Drives the freeing of Rx descriptors. */ 1044 uint8_t rx_drop_en; /**< Drop packets if no descriptors are available. */ 1045 uint8_t rx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */ 1046 uint16_t rx_nseg; /**< Number of descriptions in rx_seg array. */ 1047 /** 1048 * Share group index in Rx domain and switch domain. 1049 * Non-zero value to enable Rx queue share, zero value disable share. 1050 * PMD is responsible for Rx queue consistency checks to avoid member 1051 * port's configuration contradict to each other. 1052 */ 1053 uint16_t share_group; 1054 uint16_t share_qid; /**< Shared Rx queue ID in group */ 1055 /** 1056 * Per-queue Rx offloads to be set using RTE_ETH_RX_OFFLOAD_* flags. 1057 * Only offloads set on rx_queue_offload_capa or rx_offload_capa 1058 * fields on rte_eth_dev_info structure are allowed to be set. 1059 */ 1060 uint64_t offloads; 1061 /** 1062 * Points to the array of segment descriptions for an entire packet. 1063 * Array elements are properties for consecutive Rx segments. 1064 * 1065 * The supported capabilities of receiving segmentation is reported 1066 * in rte_eth_dev_info.rx_seg_capa field. 1067 */ 1068 union rte_eth_rxseg *rx_seg; 1069 1070 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 1071 void *reserved_ptrs[2]; /**< Reserved for future fields */ 1072 }; 1073 1074 /** 1075 * A structure used to configure a Tx ring of an Ethernet port. 1076 */ 1077 struct rte_eth_txconf { 1078 struct rte_eth_thresh tx_thresh; /**< Tx ring threshold registers. */ 1079 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */ 1080 uint16_t tx_free_thresh; /**< Start freeing Tx buffers if there are 1081 less free descriptors than this value. */ 1082 1083 uint8_t tx_deferred_start; /**< Do not start queue with rte_eth_dev_start(). */ 1084 /** 1085 * Per-queue Tx offloads to be set using RTE_ETH_TX_OFFLOAD_* flags. 1086 * Only offloads set on tx_queue_offload_capa or tx_offload_capa 1087 * fields on rte_eth_dev_info structure are allowed to be set. 1088 */ 1089 uint64_t offloads; 1090 1091 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 1092 void *reserved_ptrs[2]; /**< Reserved for future fields */ 1093 }; 1094 1095 /** 1096 * @warning 1097 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1098 * 1099 * A structure used to return the hairpin capabilities that are supported. 1100 */ 1101 struct rte_eth_hairpin_cap { 1102 /** The max number of hairpin queues (different bindings). */ 1103 uint16_t max_nb_queues; 1104 /** Max number of Rx queues to be connected to one Tx queue. */ 1105 uint16_t max_rx_2_tx; 1106 /** Max number of Tx queues to be connected to one Rx queue. */ 1107 uint16_t max_tx_2_rx; 1108 uint16_t max_nb_desc; /**< The max num of descriptors. */ 1109 }; 1110 1111 #define RTE_ETH_MAX_HAIRPIN_PEERS 32 1112 1113 /** 1114 * @warning 1115 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1116 * 1117 * A structure used to hold hairpin peer data. 1118 */ 1119 struct rte_eth_hairpin_peer { 1120 uint16_t port; /**< Peer port. */ 1121 uint16_t queue; /**< Peer queue. */ 1122 }; 1123 1124 /** 1125 * @warning 1126 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1127 * 1128 * A structure used to configure hairpin binding. 1129 */ 1130 struct rte_eth_hairpin_conf { 1131 uint32_t peer_count:16; /**< The number of peers. */ 1132 1133 /** 1134 * Explicit Tx flow rule mode. 1135 * One hairpin pair of queues should have the same attribute. 1136 * 1137 * - When set, the user should be responsible for inserting the hairpin 1138 * Tx part flows and removing them. 1139 * - When clear, the PMD will try to handle the Tx part of the flows, 1140 * e.g., by splitting one flow into two parts. 1141 */ 1142 uint32_t tx_explicit:1; 1143 1144 /** 1145 * Manually bind hairpin queues. 1146 * One hairpin pair of queues should have the same attribute. 1147 * 1148 * - When set, to enable hairpin, the user should call the hairpin bind 1149 * function after all the queues are set up properly and the ports are 1150 * started. Also, the hairpin unbind function should be called 1151 * accordingly before stopping a port that with hairpin configured. 1152 * - When clear, the PMD will try to enable the hairpin with the queues 1153 * configured automatically during port start. 1154 */ 1155 uint32_t manual_bind:1; 1156 uint32_t reserved:14; /**< Reserved bits. */ 1157 struct rte_eth_hairpin_peer peers[RTE_ETH_MAX_HAIRPIN_PEERS]; 1158 }; 1159 1160 /** 1161 * A structure contains information about HW descriptor ring limitations. 1162 */ 1163 struct rte_eth_desc_lim { 1164 uint16_t nb_max; /**< Max allowed number of descriptors. */ 1165 uint16_t nb_min; /**< Min allowed number of descriptors. */ 1166 uint16_t nb_align; /**< Number of descriptors should be aligned to. */ 1167 1168 /** 1169 * Max allowed number of segments per whole packet. 1170 * 1171 * - For TSO packet this is the total number of data descriptors allowed 1172 * by device. 1173 * 1174 * @see nb_mtu_seg_max 1175 */ 1176 uint16_t nb_seg_max; 1177 1178 /** 1179 * Max number of segments per one MTU. 1180 * 1181 * - For non-TSO packet, this is the maximum allowed number of segments 1182 * in a single transmit packet. 1183 * 1184 * - For TSO packet each segment within the TSO may span up to this 1185 * value. 1186 * 1187 * @see nb_seg_max 1188 */ 1189 uint16_t nb_mtu_seg_max; 1190 }; 1191 1192 /** 1193 * This enum indicates the flow control mode 1194 */ 1195 enum rte_eth_fc_mode { 1196 RTE_ETH_FC_NONE = 0, /**< Disable flow control. */ 1197 RTE_ETH_FC_RX_PAUSE, /**< Rx pause frame, enable flowctrl on Tx side. */ 1198 RTE_ETH_FC_TX_PAUSE, /**< Tx pause frame, enable flowctrl on Rx side. */ 1199 RTE_ETH_FC_FULL /**< Enable flow control on both side. */ 1200 }; 1201 1202 /** 1203 * A structure used to configure Ethernet flow control parameter. 1204 * These parameters will be configured into the register of the NIC. 1205 * Please refer to the corresponding data sheet for proper value. 1206 */ 1207 struct rte_eth_fc_conf { 1208 uint32_t high_water; /**< High threshold value to trigger XOFF */ 1209 uint32_t low_water; /**< Low threshold value to trigger XON */ 1210 uint16_t pause_time; /**< Pause quota in the Pause frame */ 1211 uint16_t send_xon; /**< Is XON frame need be sent */ 1212 enum rte_eth_fc_mode mode; /**< Link flow control mode */ 1213 uint8_t mac_ctrl_frame_fwd; /**< Forward MAC control frames */ 1214 uint8_t autoneg; /**< Use Pause autoneg */ 1215 }; 1216 1217 /** 1218 * A structure used to configure Ethernet priority flow control parameter. 1219 * These parameters will be configured into the register of the NIC. 1220 * Please refer to the corresponding data sheet for proper value. 1221 */ 1222 struct rte_eth_pfc_conf { 1223 struct rte_eth_fc_conf fc; /**< General flow control parameter. */ 1224 uint8_t priority; /**< VLAN User Priority. */ 1225 }; 1226 1227 /** 1228 * @warning 1229 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1230 * 1231 * A structure used to retrieve information of queue based PFC. 1232 */ 1233 struct rte_eth_pfc_queue_info { 1234 /** 1235 * Maximum supported traffic class as per PFC (802.1Qbb) specification. 1236 */ 1237 uint8_t tc_max; 1238 /** PFC queue mode capabilities. */ 1239 enum rte_eth_fc_mode mode_capa; 1240 }; 1241 1242 /** 1243 * @warning 1244 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 1245 * 1246 * A structure used to configure Ethernet priority flow control parameters for 1247 * ethdev queues. 1248 * 1249 * rte_eth_pfc_queue_conf::rx_pause structure shall be used to configure given 1250 * tx_qid with corresponding tc. When ethdev device receives PFC frame with 1251 * rte_eth_pfc_queue_conf::rx_pause::tc, traffic will be paused on 1252 * rte_eth_pfc_queue_conf::rx_pause::tx_qid for that tc. 1253 * 1254 * rte_eth_pfc_queue_conf::tx_pause structure shall be used to configure given 1255 * rx_qid. When rx_qid is congested, PFC frames are generated with 1256 * rte_eth_pfc_queue_conf::rx_pause::tc and 1257 * rte_eth_pfc_queue_conf::rx_pause::pause_time to the peer. 1258 */ 1259 struct rte_eth_pfc_queue_conf { 1260 enum rte_eth_fc_mode mode; /**< Link flow control mode */ 1261 1262 struct { 1263 uint16_t tx_qid; /**< Tx queue ID */ 1264 /** Traffic class as per PFC (802.1Qbb) spec. The value must be 1265 * in the range [0, rte_eth_pfc_queue_info::tx_max - 1] 1266 */ 1267 uint8_t tc; 1268 } rx_pause; /* Valid when (mode == FC_RX_PAUSE || mode == FC_FULL) */ 1269 1270 struct { 1271 uint16_t pause_time; /**< Pause quota in the Pause frame */ 1272 uint16_t rx_qid; /**< Rx queue ID */ 1273 /** Traffic class as per PFC (802.1Qbb) spec. The value must be 1274 * in the range [0, rte_eth_pfc_queue_info::tx_max - 1] 1275 */ 1276 uint8_t tc; 1277 } tx_pause; /* Valid when (mode == FC_TX_PAUSE || mode == FC_FULL) */ 1278 }; 1279 1280 /** 1281 * Tunnel type for device-specific classifier configuration. 1282 * @see rte_eth_udp_tunnel 1283 */ 1284 enum rte_eth_tunnel_type { 1285 RTE_ETH_TUNNEL_TYPE_NONE = 0, 1286 RTE_ETH_TUNNEL_TYPE_VXLAN, 1287 RTE_ETH_TUNNEL_TYPE_GENEVE, 1288 RTE_ETH_TUNNEL_TYPE_TEREDO, 1289 RTE_ETH_TUNNEL_TYPE_NVGRE, 1290 RTE_ETH_TUNNEL_TYPE_IP_IN_GRE, 1291 RTE_ETH_L2_TUNNEL_TYPE_E_TAG, 1292 RTE_ETH_TUNNEL_TYPE_VXLAN_GPE, 1293 RTE_ETH_TUNNEL_TYPE_ECPRI, 1294 RTE_ETH_TUNNEL_TYPE_MAX, 1295 }; 1296 1297 /* Deprecated API file for rte_eth_dev_filter_* functions */ 1298 #include "rte_eth_ctrl.h" 1299 1300 /** 1301 * UDP tunneling configuration. 1302 * 1303 * Used to configure the classifier of a device, 1304 * associating an UDP port with a type of tunnel. 1305 * 1306 * Some NICs may need such configuration to properly parse a tunnel 1307 * with any standard or custom UDP port. 1308 */ 1309 struct rte_eth_udp_tunnel { 1310 uint16_t udp_port; /**< UDP port used for the tunnel. */ 1311 uint8_t prot_type; /**< Tunnel type. @see rte_eth_tunnel_type */ 1312 }; 1313 1314 /** 1315 * A structure used to enable/disable specific device interrupts. 1316 */ 1317 struct rte_eth_intr_conf { 1318 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */ 1319 uint32_t lsc:1; 1320 /** enable/disable rxq interrupt. 0 (default) - disable, 1 enable */ 1321 uint32_t rxq:1; 1322 /** enable/disable rmv interrupt. 0 (default) - disable, 1 enable */ 1323 uint32_t rmv:1; 1324 }; 1325 1326 #define rte_intr_conf rte_eth_intr_conf 1327 1328 /** 1329 * A structure used to configure an Ethernet port. 1330 * Depending upon the Rx multi-queue mode, extra advanced 1331 * configuration settings may be needed. 1332 */ 1333 struct rte_eth_conf { 1334 uint32_t link_speeds; /**< bitmap of RTE_ETH_LINK_SPEED_XXX of speeds to be 1335 used. RTE_ETH_LINK_SPEED_FIXED disables link 1336 autonegotiation, and a unique speed shall be 1337 set. Otherwise, the bitmap defines the set of 1338 speeds to be advertised. If the special value 1339 RTE_ETH_LINK_SPEED_AUTONEG (0) is used, all speeds 1340 supported are advertised. */ 1341 struct rte_eth_rxmode rxmode; /**< Port Rx configuration. */ 1342 struct rte_eth_txmode txmode; /**< Port Tx configuration. */ 1343 uint32_t lpbk_mode; /**< Loopback operation mode. By default the value 1344 is 0, meaning the loopback mode is disabled. 1345 Read the datasheet of given Ethernet controller 1346 for details. The possible values of this field 1347 are defined in implementation of each driver. */ 1348 struct { 1349 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */ 1350 /** Port VMDq+DCB configuration. */ 1351 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf; 1352 /** Port DCB Rx configuration. */ 1353 struct rte_eth_dcb_rx_conf dcb_rx_conf; 1354 /** Port VMDq Rx configuration. */ 1355 struct rte_eth_vmdq_rx_conf vmdq_rx_conf; 1356 } rx_adv_conf; /**< Port Rx filtering configuration. */ 1357 union { 1358 /** Port VMDq+DCB Tx configuration. */ 1359 struct rte_eth_vmdq_dcb_tx_conf vmdq_dcb_tx_conf; 1360 /** Port DCB Tx configuration. */ 1361 struct rte_eth_dcb_tx_conf dcb_tx_conf; 1362 /** Port VMDq Tx configuration. */ 1363 struct rte_eth_vmdq_tx_conf vmdq_tx_conf; 1364 } tx_adv_conf; /**< Port Tx DCB configuration (union). */ 1365 /** Currently,Priority Flow Control(PFC) are supported,if DCB with PFC 1366 is needed,and the variable must be set RTE_ETH_DCB_PFC_SUPPORT. */ 1367 uint32_t dcb_capability_en; 1368 struct rte_eth_intr_conf intr_conf; /**< Interrupt mode configuration. */ 1369 }; 1370 1371 /** 1372 * Rx offload capabilities of a device. 1373 */ 1374 #define RTE_ETH_RX_OFFLOAD_VLAN_STRIP RTE_BIT64(0) 1375 #define RTE_ETH_RX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1) 1376 #define RTE_ETH_RX_OFFLOAD_UDP_CKSUM RTE_BIT64(2) 1377 #define RTE_ETH_RX_OFFLOAD_TCP_CKSUM RTE_BIT64(3) 1378 #define RTE_ETH_RX_OFFLOAD_TCP_LRO RTE_BIT64(4) 1379 #define RTE_ETH_RX_OFFLOAD_QINQ_STRIP RTE_BIT64(5) 1380 #define RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(6) 1381 #define RTE_ETH_RX_OFFLOAD_MACSEC_STRIP RTE_BIT64(7) 1382 #define RTE_ETH_RX_OFFLOAD_HEADER_SPLIT RTE_BIT64(8) 1383 #define RTE_ETH_RX_OFFLOAD_VLAN_FILTER RTE_BIT64(9) 1384 #define RTE_ETH_RX_OFFLOAD_VLAN_EXTEND RTE_BIT64(10) 1385 #define RTE_ETH_RX_OFFLOAD_SCATTER RTE_BIT64(13) 1386 /** 1387 * Timestamp is set by the driver in RTE_MBUF_DYNFIELD_TIMESTAMP_NAME 1388 * and RTE_MBUF_DYNFLAG_RX_TIMESTAMP_NAME is set in ol_flags. 1389 * The mbuf field and flag are registered when the offload is configured. 1390 */ 1391 #define RTE_ETH_RX_OFFLOAD_TIMESTAMP RTE_BIT64(14) 1392 #define RTE_ETH_RX_OFFLOAD_SECURITY RTE_BIT64(15) 1393 #define RTE_ETH_RX_OFFLOAD_KEEP_CRC RTE_BIT64(16) 1394 #define RTE_ETH_RX_OFFLOAD_SCTP_CKSUM RTE_BIT64(17) 1395 #define RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(18) 1396 #define RTE_ETH_RX_OFFLOAD_RSS_HASH RTE_BIT64(19) 1397 #define RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT RTE_BIT64(20) 1398 1399 #define RTE_ETH_RX_OFFLOAD_CHECKSUM (RTE_ETH_RX_OFFLOAD_IPV4_CKSUM | \ 1400 RTE_ETH_RX_OFFLOAD_UDP_CKSUM | \ 1401 RTE_ETH_RX_OFFLOAD_TCP_CKSUM) 1402 #define RTE_ETH_RX_OFFLOAD_VLAN (RTE_ETH_RX_OFFLOAD_VLAN_STRIP | \ 1403 RTE_ETH_RX_OFFLOAD_VLAN_FILTER | \ 1404 RTE_ETH_RX_OFFLOAD_VLAN_EXTEND | \ 1405 RTE_ETH_RX_OFFLOAD_QINQ_STRIP) 1406 1407 /* 1408 * If new Rx offload capabilities are defined, they also must be 1409 * mentioned in rte_rx_offload_names in rte_ethdev.c file. 1410 */ 1411 1412 /** 1413 * Tx offload capabilities of a device. 1414 */ 1415 #define RTE_ETH_TX_OFFLOAD_VLAN_INSERT RTE_BIT64(0) 1416 #define RTE_ETH_TX_OFFLOAD_IPV4_CKSUM RTE_BIT64(1) 1417 #define RTE_ETH_TX_OFFLOAD_UDP_CKSUM RTE_BIT64(2) 1418 #define RTE_ETH_TX_OFFLOAD_TCP_CKSUM RTE_BIT64(3) 1419 #define RTE_ETH_TX_OFFLOAD_SCTP_CKSUM RTE_BIT64(4) 1420 #define RTE_ETH_TX_OFFLOAD_TCP_TSO RTE_BIT64(5) 1421 #define RTE_ETH_TX_OFFLOAD_UDP_TSO RTE_BIT64(6) 1422 #define RTE_ETH_TX_OFFLOAD_OUTER_IPV4_CKSUM RTE_BIT64(7) /**< Used for tunneling packet. */ 1423 #define RTE_ETH_TX_OFFLOAD_QINQ_INSERT RTE_BIT64(8) 1424 #define RTE_ETH_TX_OFFLOAD_VXLAN_TNL_TSO RTE_BIT64(9) /**< Used for tunneling packet. */ 1425 #define RTE_ETH_TX_OFFLOAD_GRE_TNL_TSO RTE_BIT64(10) /**< Used for tunneling packet. */ 1426 #define RTE_ETH_TX_OFFLOAD_IPIP_TNL_TSO RTE_BIT64(11) /**< Used for tunneling packet. */ 1427 #define RTE_ETH_TX_OFFLOAD_GENEVE_TNL_TSO RTE_BIT64(12) /**< Used for tunneling packet. */ 1428 #define RTE_ETH_TX_OFFLOAD_MACSEC_INSERT RTE_BIT64(13) 1429 /** 1430 * Multiple threads can invoke rte_eth_tx_burst() concurrently on the same 1431 * Tx queue without SW lock. 1432 */ 1433 #define RTE_ETH_TX_OFFLOAD_MT_LOCKFREE RTE_BIT64(14) 1434 /** Device supports multi segment send. */ 1435 #define RTE_ETH_TX_OFFLOAD_MULTI_SEGS RTE_BIT64(15) 1436 /** 1437 * Device supports optimization for fast release of mbufs. 1438 * When set application must guarantee that per-queue all mbufs comes from 1439 * the same mempool and has refcnt = 1. 1440 */ 1441 #define RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE RTE_BIT64(16) 1442 #define RTE_ETH_TX_OFFLOAD_SECURITY RTE_BIT64(17) 1443 /** 1444 * Device supports generic UDP tunneled packet TSO. 1445 * Application must set RTE_MBUF_F_TX_TUNNEL_UDP and other mbuf fields required 1446 * for tunnel TSO. 1447 */ 1448 #define RTE_ETH_TX_OFFLOAD_UDP_TNL_TSO RTE_BIT64(18) 1449 /** 1450 * Device supports generic IP tunneled packet TSO. 1451 * Application must set RTE_MBUF_F_TX_TUNNEL_IP and other mbuf fields required 1452 * for tunnel TSO. 1453 */ 1454 #define RTE_ETH_TX_OFFLOAD_IP_TNL_TSO RTE_BIT64(19) 1455 /** Device supports outer UDP checksum */ 1456 #define RTE_ETH_TX_OFFLOAD_OUTER_UDP_CKSUM RTE_BIT64(20) 1457 /** 1458 * Device sends on time read from RTE_MBUF_DYNFIELD_TIMESTAMP_NAME 1459 * if RTE_MBUF_DYNFLAG_TX_TIMESTAMP_NAME is set in ol_flags. 1460 * The mbuf field and flag are registered when the offload is configured. 1461 */ 1462 #define RTE_ETH_TX_OFFLOAD_SEND_ON_TIMESTAMP RTE_BIT64(21) 1463 /* 1464 * If new Tx offload capabilities are defined, they also must be 1465 * mentioned in rte_tx_offload_names in rte_ethdev.c file. 1466 */ 1467 1468 /**@{@name Device capabilities 1469 * Non-offload capabilities reported in rte_eth_dev_info.dev_capa. 1470 */ 1471 /** Device supports Rx queue setup after device started. */ 1472 #define RTE_ETH_DEV_CAPA_RUNTIME_RX_QUEUE_SETUP RTE_BIT64(0) 1473 /** Device supports Tx queue setup after device started. */ 1474 #define RTE_ETH_DEV_CAPA_RUNTIME_TX_QUEUE_SETUP RTE_BIT64(1) 1475 /** 1476 * Device supports shared Rx queue among ports within Rx domain and 1477 * switch domain. Mbufs are consumed by shared Rx queue instead of 1478 * each queue. Multiple groups are supported by share_group of Rx 1479 * queue configuration. Shared Rx queue is identified by PMD using 1480 * share_qid of Rx queue configuration. Polling any port in the group 1481 * receive packets of all member ports, source port identified by 1482 * mbuf->port field. 1483 */ 1484 #define RTE_ETH_DEV_CAPA_RXQ_SHARE RTE_BIT64(2) 1485 /** Device supports keeping flow rules across restart. */ 1486 #define RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP RTE_BIT64(3) 1487 /** Device supports keeping shared flow objects across restart. */ 1488 #define RTE_ETH_DEV_CAPA_FLOW_SHARED_OBJECT_KEEP RTE_BIT64(4) 1489 /**@}*/ 1490 1491 /* 1492 * Fallback default preferred Rx/Tx port parameters. 1493 * These are used if an application requests default parameters 1494 * but the PMD does not provide preferred values. 1495 */ 1496 #define RTE_ETH_DEV_FALLBACK_RX_RINGSIZE 512 1497 #define RTE_ETH_DEV_FALLBACK_TX_RINGSIZE 512 1498 #define RTE_ETH_DEV_FALLBACK_RX_NBQUEUES 1 1499 #define RTE_ETH_DEV_FALLBACK_TX_NBQUEUES 1 1500 1501 /** 1502 * Preferred Rx/Tx port parameters. 1503 * There are separate instances of this structure for transmission 1504 * and reception respectively. 1505 */ 1506 struct rte_eth_dev_portconf { 1507 uint16_t burst_size; /**< Device-preferred burst size */ 1508 uint16_t ring_size; /**< Device-preferred size of queue rings */ 1509 uint16_t nb_queues; /**< Device-preferred number of queues */ 1510 }; 1511 1512 /** 1513 * Default values for switch domain ID when ethdev does not support switch 1514 * domain definitions. 1515 */ 1516 #define RTE_ETH_DEV_SWITCH_DOMAIN_ID_INVALID (UINT16_MAX) 1517 1518 /** 1519 * Ethernet device associated switch information 1520 */ 1521 struct rte_eth_switch_info { 1522 const char *name; /**< switch name */ 1523 uint16_t domain_id; /**< switch domain ID */ 1524 /** 1525 * Mapping to the devices physical switch port as enumerated from the 1526 * perspective of the embedded interconnect/switch. For SR-IOV enabled 1527 * device this may correspond to the VF_ID of each virtual function, 1528 * but each driver should explicitly define the mapping of switch 1529 * port identifier to that physical interconnect/switch 1530 */ 1531 uint16_t port_id; 1532 /** 1533 * Shared Rx queue sub-domain boundary. Only ports in same Rx domain 1534 * and switch domain can share Rx queue. Valid only if device advertised 1535 * RTE_ETH_DEV_CAPA_RXQ_SHARE capability. 1536 */ 1537 uint16_t rx_domain; 1538 }; 1539 1540 /** 1541 * @warning 1542 * @b EXPERIMENTAL: this structure may change without prior notice. 1543 * 1544 * Ethernet device Rx buffer segmentation capabilities. 1545 */ 1546 struct rte_eth_rxseg_capa { 1547 __extension__ 1548 uint32_t multi_pools:1; /**< Supports receiving to multiple pools.*/ 1549 uint32_t offset_allowed:1; /**< Supports buffer offsets. */ 1550 uint32_t offset_align_log2:4; /**< Required offset alignment. */ 1551 uint16_t max_nseg; /**< Maximum amount of segments to split. */ 1552 uint16_t reserved; /**< Reserved field. */ 1553 }; 1554 1555 /** 1556 * Ethernet device information 1557 */ 1558 1559 /** 1560 * Ethernet device representor port type. 1561 */ 1562 enum rte_eth_representor_type { 1563 RTE_ETH_REPRESENTOR_NONE, /**< not a representor. */ 1564 RTE_ETH_REPRESENTOR_VF, /**< representor of Virtual Function. */ 1565 RTE_ETH_REPRESENTOR_SF, /**< representor of Sub Function. */ 1566 RTE_ETH_REPRESENTOR_PF, /**< representor of Physical Function. */ 1567 }; 1568 1569 /** 1570 * A structure used to retrieve the contextual information of 1571 * an Ethernet device, such as the controlling driver of the 1572 * device, etc... 1573 */ 1574 struct rte_eth_dev_info { 1575 struct rte_device *device; /**< Generic device information */ 1576 const char *driver_name; /**< Device Driver name. */ 1577 unsigned int if_index; /**< Index to bound host interface, or 0 if none. 1578 Use if_indextoname() to translate into an interface name. */ 1579 uint16_t min_mtu; /**< Minimum MTU allowed */ 1580 uint16_t max_mtu; /**< Maximum MTU allowed */ 1581 const uint32_t *dev_flags; /**< Device flags */ 1582 uint32_t min_rx_bufsize; /**< Minimum size of Rx buffer. */ 1583 uint32_t max_rx_pktlen; /**< Maximum configurable length of Rx pkt. */ 1584 /** Maximum configurable size of LRO aggregated packet. */ 1585 uint32_t max_lro_pkt_size; 1586 uint16_t max_rx_queues; /**< Maximum number of Rx queues. */ 1587 uint16_t max_tx_queues; /**< Maximum number of Tx queues. */ 1588 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */ 1589 /** Maximum number of hash MAC addresses for MTA and UTA. */ 1590 uint32_t max_hash_mac_addrs; 1591 uint16_t max_vfs; /**< Maximum number of VFs. */ 1592 uint16_t max_vmdq_pools; /**< Maximum number of VMDq pools. */ 1593 struct rte_eth_rxseg_capa rx_seg_capa; /**< Segmentation capability.*/ 1594 /** All Rx offload capabilities including all per-queue ones */ 1595 uint64_t rx_offload_capa; 1596 /** All Tx offload capabilities including all per-queue ones */ 1597 uint64_t tx_offload_capa; 1598 /** Device per-queue Rx offload capabilities. */ 1599 uint64_t rx_queue_offload_capa; 1600 /** Device per-queue Tx offload capabilities. */ 1601 uint64_t tx_queue_offload_capa; 1602 /** Device redirection table size, the total number of entries. */ 1603 uint16_t reta_size; 1604 uint8_t hash_key_size; /**< Hash key size in bytes */ 1605 /** Bit mask of RSS offloads, the bit offset also means flow type */ 1606 uint64_t flow_type_rss_offloads; 1607 struct rte_eth_rxconf default_rxconf; /**< Default Rx configuration */ 1608 struct rte_eth_txconf default_txconf; /**< Default Tx configuration */ 1609 uint16_t vmdq_queue_base; /**< First queue ID for VMDq pools. */ 1610 uint16_t vmdq_queue_num; /**< Queue number for VMDq pools. */ 1611 uint16_t vmdq_pool_base; /**< First ID of VMDq pools. */ 1612 struct rte_eth_desc_lim rx_desc_lim; /**< Rx descriptors limits */ 1613 struct rte_eth_desc_lim tx_desc_lim; /**< Tx descriptors limits */ 1614 uint32_t speed_capa; /**< Supported speeds bitmap (RTE_ETH_LINK_SPEED_). */ 1615 /** Configured number of Rx/Tx queues */ 1616 uint16_t nb_rx_queues; /**< Number of Rx queues. */ 1617 uint16_t nb_tx_queues; /**< Number of Tx queues. */ 1618 /** Rx parameter recommendations */ 1619 struct rte_eth_dev_portconf default_rxportconf; 1620 /** Tx parameter recommendations */ 1621 struct rte_eth_dev_portconf default_txportconf; 1622 /** Generic device capabilities (RTE_ETH_DEV_CAPA_). */ 1623 uint64_t dev_capa; 1624 /** 1625 * Switching information for ports on a device with a 1626 * embedded managed interconnect/switch. 1627 */ 1628 struct rte_eth_switch_info switch_info; 1629 1630 uint64_t reserved_64s[2]; /**< Reserved for future fields */ 1631 void *reserved_ptrs[2]; /**< Reserved for future fields */ 1632 }; 1633 1634 /**@{@name Rx/Tx queue states */ 1635 #define RTE_ETH_QUEUE_STATE_STOPPED 0 /**< Queue stopped. */ 1636 #define RTE_ETH_QUEUE_STATE_STARTED 1 /**< Queue started. */ 1637 #define RTE_ETH_QUEUE_STATE_HAIRPIN 2 /**< Queue used for hairpin. */ 1638 /**@}*/ 1639 1640 /** 1641 * Ethernet device Rx queue information structure. 1642 * Used to retrieve information about configured queue. 1643 */ 1644 struct rte_eth_rxq_info { 1645 struct rte_mempool *mp; /**< mempool used by that queue. */ 1646 struct rte_eth_rxconf conf; /**< queue config parameters. */ 1647 uint8_t scattered_rx; /**< scattered packets Rx supported. */ 1648 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */ 1649 uint16_t nb_desc; /**< configured number of RXDs. */ 1650 uint16_t rx_buf_size; /**< hardware receive buffer size. */ 1651 /** 1652 * Available Rx descriptors threshold defined as percentage 1653 * of Rx queue size. If number of available descriptors is lower, 1654 * the event RTE_ETH_EVENT_RX_AVAIL_THESH is generated. 1655 * Value 0 means that the threshold monitoring is disabled. 1656 */ 1657 uint8_t avail_thresh; 1658 } __rte_cache_min_aligned; 1659 1660 /** 1661 * Ethernet device Tx queue information structure. 1662 * Used to retrieve information about configured queue. 1663 */ 1664 struct rte_eth_txq_info { 1665 struct rte_eth_txconf conf; /**< queue config parameters. */ 1666 uint16_t nb_desc; /**< configured number of TXDs. */ 1667 uint8_t queue_state; /**< one of RTE_ETH_QUEUE_STATE_*. */ 1668 } __rte_cache_min_aligned; 1669 1670 /* Generic Burst mode flag definition, values can be ORed. */ 1671 1672 /** 1673 * If the queues have different burst mode description, this bit will be set 1674 * by PMD, then the application can iterate to retrieve burst description for 1675 * all other queues. 1676 */ 1677 #define RTE_ETH_BURST_FLAG_PER_QUEUE RTE_BIT64(0) 1678 1679 /** 1680 * Ethernet device Rx/Tx queue packet burst mode information structure. 1681 * Used to retrieve information about packet burst mode setting. 1682 */ 1683 struct rte_eth_burst_mode { 1684 uint64_t flags; /**< The ORed values of RTE_ETH_BURST_FLAG_xxx */ 1685 1686 #define RTE_ETH_BURST_MODE_INFO_SIZE 1024 /**< Maximum size for information */ 1687 char info[RTE_ETH_BURST_MODE_INFO_SIZE]; /**< burst mode information */ 1688 }; 1689 1690 /** Maximum name length for extended statistics counters */ 1691 #define RTE_ETH_XSTATS_NAME_SIZE 64 1692 1693 /** 1694 * An Ethernet device extended statistic structure 1695 * 1696 * This structure is used by rte_eth_xstats_get() to provide 1697 * statistics that are not provided in the generic *rte_eth_stats* 1698 * structure. 1699 * It maps a name ID, corresponding to an index in the array returned 1700 * by rte_eth_xstats_get_names(), to a statistic value. 1701 */ 1702 struct rte_eth_xstat { 1703 uint64_t id; /**< The index in xstats name array. */ 1704 uint64_t value; /**< The statistic counter value. */ 1705 }; 1706 1707 /** 1708 * A name element for extended statistics. 1709 * 1710 * An array of this structure is returned by rte_eth_xstats_get_names(). 1711 * It lists the names of extended statistics for a PMD. The *rte_eth_xstat* 1712 * structure references these names by their array index. 1713 * 1714 * The xstats should follow a common naming scheme. 1715 * Some names are standardized in rte_stats_strings. 1716 * Examples: 1717 * - rx_missed_errors 1718 * - tx_q3_bytes 1719 * - tx_size_128_to_255_packets 1720 */ 1721 struct rte_eth_xstat_name { 1722 char name[RTE_ETH_XSTATS_NAME_SIZE]; /**< The statistic name. */ 1723 }; 1724 1725 #define RTE_ETH_DCB_NUM_TCS 8 1726 #define RTE_ETH_MAX_VMDQ_POOL 64 1727 1728 /** 1729 * A structure used to get the information of queue and 1730 * TC mapping on both Tx and Rx paths. 1731 */ 1732 struct rte_eth_dcb_tc_queue_mapping { 1733 /** Rx queues assigned to tc per Pool */ 1734 struct { 1735 uint16_t base; 1736 uint16_t nb_queue; 1737 } tc_rxq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS]; 1738 /** Rx queues assigned to tc per Pool */ 1739 struct { 1740 uint16_t base; 1741 uint16_t nb_queue; 1742 } tc_txq[RTE_ETH_MAX_VMDQ_POOL][RTE_ETH_DCB_NUM_TCS]; 1743 }; 1744 1745 /** 1746 * A structure used to get the information of DCB. 1747 * It includes TC UP mapping and queue TC mapping. 1748 */ 1749 struct rte_eth_dcb_info { 1750 uint8_t nb_tcs; /**< number of TCs */ 1751 uint8_t prio_tc[RTE_ETH_DCB_NUM_USER_PRIORITIES]; /**< Priority to tc */ 1752 uint8_t tc_bws[RTE_ETH_DCB_NUM_TCS]; /**< Tx BW percentage for each TC */ 1753 /** Rx queues assigned to tc */ 1754 struct rte_eth_dcb_tc_queue_mapping tc_queue; 1755 }; 1756 1757 /** 1758 * This enum indicates the possible Forward Error Correction (FEC) modes 1759 * of an ethdev port. 1760 */ 1761 enum rte_eth_fec_mode { 1762 RTE_ETH_FEC_NOFEC = 0, /**< FEC is off */ 1763 RTE_ETH_FEC_AUTO, /**< FEC autonegotiation modes */ 1764 RTE_ETH_FEC_BASER, /**< FEC using common algorithm */ 1765 RTE_ETH_FEC_RS, /**< FEC using RS algorithm */ 1766 }; 1767 1768 /* Translate from FEC mode to FEC capa */ 1769 #define RTE_ETH_FEC_MODE_TO_CAPA(x) RTE_BIT32(x) 1770 1771 /* This macro indicates FEC capa mask */ 1772 #define RTE_ETH_FEC_MODE_CAPA_MASK(x) RTE_BIT32(RTE_ETH_FEC_ ## x) 1773 1774 /* A structure used to get capabilities per link speed */ 1775 struct rte_eth_fec_capa { 1776 uint32_t speed; /**< Link speed (see RTE_ETH_SPEED_NUM_*) */ 1777 uint32_t capa; /**< FEC capabilities bitmask */ 1778 }; 1779 1780 #define RTE_ETH_ALL RTE_MAX_ETHPORTS 1781 1782 /* Macros to check for valid port */ 1783 #define RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, retval) do { \ 1784 if (!rte_eth_dev_is_valid_port(port_id)) { \ 1785 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \ 1786 return retval; \ 1787 } \ 1788 } while (0) 1789 1790 #define RTE_ETH_VALID_PORTID_OR_RET(port_id) do { \ 1791 if (!rte_eth_dev_is_valid_port(port_id)) { \ 1792 RTE_ETHDEV_LOG(ERR, "Invalid port_id=%u\n", port_id); \ 1793 return; \ 1794 } \ 1795 } while (0) 1796 1797 /** 1798 * Function type used for Rx packet processing packet callbacks. 1799 * 1800 * The callback function is called on Rx with a burst of packets that have 1801 * been received on the given port and queue. 1802 * 1803 * @param port_id 1804 * The Ethernet port on which Rx is being performed. 1805 * @param queue 1806 * The queue on the Ethernet port which is being used to receive the packets. 1807 * @param pkts 1808 * The burst of packets that have just been received. 1809 * @param nb_pkts 1810 * The number of packets in the burst pointed to by "pkts". 1811 * @param max_pkts 1812 * The max number of packets that can be stored in the "pkts" array. 1813 * @param user_param 1814 * The arbitrary user parameter passed in by the application when the callback 1815 * was originally configured. 1816 * @return 1817 * The number of packets returned to the user. 1818 */ 1819 typedef uint16_t (*rte_rx_callback_fn)(uint16_t port_id, uint16_t queue, 1820 struct rte_mbuf *pkts[], uint16_t nb_pkts, uint16_t max_pkts, 1821 void *user_param); 1822 1823 /** 1824 * Function type used for Tx packet processing packet callbacks. 1825 * 1826 * The callback function is called on Tx with a burst of packets immediately 1827 * before the packets are put onto the hardware queue for transmission. 1828 * 1829 * @param port_id 1830 * The Ethernet port on which Tx is being performed. 1831 * @param queue 1832 * The queue on the Ethernet port which is being used to transmit the packets. 1833 * @param pkts 1834 * The burst of packets that are about to be transmitted. 1835 * @param nb_pkts 1836 * The number of packets in the burst pointed to by "pkts". 1837 * @param user_param 1838 * The arbitrary user parameter passed in by the application when the callback 1839 * was originally configured. 1840 * @return 1841 * The number of packets to be written to the NIC. 1842 */ 1843 typedef uint16_t (*rte_tx_callback_fn)(uint16_t port_id, uint16_t queue, 1844 struct rte_mbuf *pkts[], uint16_t nb_pkts, void *user_param); 1845 1846 /** 1847 * Possible states of an ethdev port. 1848 */ 1849 enum rte_eth_dev_state { 1850 /** Device is unused before being probed. */ 1851 RTE_ETH_DEV_UNUSED = 0, 1852 /** Device is attached when allocated in probing. */ 1853 RTE_ETH_DEV_ATTACHED, 1854 /** Device is in removed state when plug-out is detected. */ 1855 RTE_ETH_DEV_REMOVED, 1856 }; 1857 1858 struct rte_eth_dev_sriov { 1859 uint8_t active; /**< SRIOV is active with 16, 32 or 64 pools */ 1860 uint8_t nb_q_per_pool; /**< Rx queue number per pool */ 1861 uint16_t def_vmdq_idx; /**< Default pool num used for PF */ 1862 uint16_t def_pool_q_idx; /**< Default pool queue start reg index */ 1863 }; 1864 #define RTE_ETH_DEV_SRIOV(dev) ((dev)->data->sriov) 1865 1866 #define RTE_ETH_NAME_MAX_LEN RTE_DEV_NAME_MAX_LEN 1867 1868 #define RTE_ETH_DEV_NO_OWNER 0 1869 1870 #define RTE_ETH_MAX_OWNER_NAME_LEN 64 1871 1872 struct rte_eth_dev_owner { 1873 uint64_t id; /**< The owner unique identifier. */ 1874 char name[RTE_ETH_MAX_OWNER_NAME_LEN]; /**< The owner name. */ 1875 }; 1876 1877 /**@{@name Device flags 1878 * Flags internally saved in rte_eth_dev_data.dev_flags 1879 * and reported in rte_eth_dev_info.dev_flags. 1880 */ 1881 /** PMD supports thread-safe flow operations */ 1882 #define RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE RTE_BIT32(0) 1883 /** Device supports link state interrupt */ 1884 #define RTE_ETH_DEV_INTR_LSC RTE_BIT32(1) 1885 /** Device is a bonded slave */ 1886 #define RTE_ETH_DEV_BONDED_SLAVE RTE_BIT32(2) 1887 /** Device supports device removal interrupt */ 1888 #define RTE_ETH_DEV_INTR_RMV RTE_BIT32(3) 1889 /** Device is port representor */ 1890 #define RTE_ETH_DEV_REPRESENTOR RTE_BIT32(4) 1891 /** Device does not support MAC change after started */ 1892 #define RTE_ETH_DEV_NOLIVE_MAC_ADDR RTE_BIT32(5) 1893 /** 1894 * Queue xstats filled automatically by ethdev layer. 1895 * PMDs filling the queue xstats themselves should not set this flag 1896 */ 1897 #define RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS RTE_BIT32(6) 1898 /**@}*/ 1899 1900 /** 1901 * Iterates over valid ethdev ports owned by a specific owner. 1902 * 1903 * @param port_id 1904 * The ID of the next possible valid owned port. 1905 * @param owner_id 1906 * The owner identifier. 1907 * RTE_ETH_DEV_NO_OWNER means iterate over all valid ownerless ports. 1908 * @return 1909 * Next valid port ID owned by owner_id, RTE_MAX_ETHPORTS if there is none. 1910 */ 1911 uint64_t rte_eth_find_next_owned_by(uint16_t port_id, 1912 const uint64_t owner_id); 1913 1914 /** 1915 * Macro to iterate over all enabled ethdev ports owned by a specific owner. 1916 */ 1917 #define RTE_ETH_FOREACH_DEV_OWNED_BY(p, o) \ 1918 for (p = rte_eth_find_next_owned_by(0, o); \ 1919 (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \ 1920 p = rte_eth_find_next_owned_by(p + 1, o)) 1921 1922 /** 1923 * Iterates over valid ethdev ports. 1924 * 1925 * @param port_id 1926 * The ID of the next possible valid port. 1927 * @return 1928 * Next valid port ID, RTE_MAX_ETHPORTS if there is none. 1929 */ 1930 uint16_t rte_eth_find_next(uint16_t port_id); 1931 1932 /** 1933 * Macro to iterate over all enabled and ownerless ethdev ports. 1934 */ 1935 #define RTE_ETH_FOREACH_DEV(p) \ 1936 RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER) 1937 1938 /** 1939 * Iterates over ethdev ports of a specified device. 1940 * 1941 * @param port_id_start 1942 * The ID of the next possible valid port. 1943 * @param parent 1944 * The generic device behind the ports to iterate. 1945 * @return 1946 * Next port ID of the device, possibly port_id_start, 1947 * RTE_MAX_ETHPORTS if there is none. 1948 */ 1949 uint16_t 1950 rte_eth_find_next_of(uint16_t port_id_start, 1951 const struct rte_device *parent); 1952 1953 /** 1954 * Macro to iterate over all ethdev ports of a specified device. 1955 * 1956 * @param port_id 1957 * The ID of the matching port being iterated. 1958 * @param parent 1959 * The rte_device pointer matching the iterated ports. 1960 */ 1961 #define RTE_ETH_FOREACH_DEV_OF(port_id, parent) \ 1962 for (port_id = rte_eth_find_next_of(0, parent); \ 1963 port_id < RTE_MAX_ETHPORTS; \ 1964 port_id = rte_eth_find_next_of(port_id + 1, parent)) 1965 1966 /** 1967 * Iterates over sibling ethdev ports (i.e. sharing the same rte_device). 1968 * 1969 * @param port_id_start 1970 * The ID of the next possible valid sibling port. 1971 * @param ref_port_id 1972 * The ID of a reference port to compare rte_device with. 1973 * @return 1974 * Next sibling port ID, possibly port_id_start or ref_port_id itself, 1975 * RTE_MAX_ETHPORTS if there is none. 1976 */ 1977 uint16_t 1978 rte_eth_find_next_sibling(uint16_t port_id_start, uint16_t ref_port_id); 1979 1980 /** 1981 * Macro to iterate over all ethdev ports sharing the same rte_device 1982 * as the specified port. 1983 * Note: the specified reference port is part of the loop iterations. 1984 * 1985 * @param port_id 1986 * The ID of the matching port being iterated. 1987 * @param ref_port_id 1988 * The ID of the port being compared. 1989 */ 1990 #define RTE_ETH_FOREACH_DEV_SIBLING(port_id, ref_port_id) \ 1991 for (port_id = rte_eth_find_next_sibling(0, ref_port_id); \ 1992 port_id < RTE_MAX_ETHPORTS; \ 1993 port_id = rte_eth_find_next_sibling(port_id + 1, ref_port_id)) 1994 1995 /** 1996 * Get a new unique owner identifier. 1997 * An owner identifier is used to owns Ethernet devices by only one DPDK entity 1998 * to avoid multiple management of device by different entities. 1999 * 2000 * @param owner_id 2001 * Owner identifier pointer. 2002 * @return 2003 * Negative errno value on error, 0 on success. 2004 */ 2005 int rte_eth_dev_owner_new(uint64_t *owner_id); 2006 2007 /** 2008 * Set an Ethernet device owner. 2009 * 2010 * @param port_id 2011 * The identifier of the port to own. 2012 * @param owner 2013 * The owner pointer. 2014 * @return 2015 * Negative errno value on error, 0 on success. 2016 */ 2017 int rte_eth_dev_owner_set(const uint16_t port_id, 2018 const struct rte_eth_dev_owner *owner); 2019 2020 /** 2021 * Unset Ethernet device owner to make the device ownerless. 2022 * 2023 * @param port_id 2024 * The identifier of port to make ownerless. 2025 * @param owner_id 2026 * The owner identifier. 2027 * @return 2028 * 0 on success, negative errno value on error. 2029 */ 2030 int rte_eth_dev_owner_unset(const uint16_t port_id, 2031 const uint64_t owner_id); 2032 2033 /** 2034 * Remove owner from all Ethernet devices owned by a specific owner. 2035 * 2036 * @param owner_id 2037 * The owner identifier. 2038 * @return 2039 * 0 on success, negative errno value on error. 2040 */ 2041 int rte_eth_dev_owner_delete(const uint64_t owner_id); 2042 2043 /** 2044 * Get the owner of an Ethernet device. 2045 * 2046 * @param port_id 2047 * The port identifier. 2048 * @param owner 2049 * The owner structure pointer to fill. 2050 * @return 2051 * 0 on success, negative errno value on error.. 2052 */ 2053 int rte_eth_dev_owner_get(const uint16_t port_id, 2054 struct rte_eth_dev_owner *owner); 2055 2056 /** 2057 * Get the number of ports which are usable for the application. 2058 * 2059 * These devices must be iterated by using the macro 2060 * ``RTE_ETH_FOREACH_DEV`` or ``RTE_ETH_FOREACH_DEV_OWNED_BY`` 2061 * to deal with non-contiguous ranges of devices. 2062 * 2063 * @return 2064 * The count of available Ethernet devices. 2065 */ 2066 uint16_t rte_eth_dev_count_avail(void); 2067 2068 /** 2069 * Get the total number of ports which are allocated. 2070 * 2071 * Some devices may not be available for the application. 2072 * 2073 * @return 2074 * The total count of Ethernet devices. 2075 */ 2076 uint16_t rte_eth_dev_count_total(void); 2077 2078 /** 2079 * Convert a numerical speed in Mbps to a bitmap flag that can be used in 2080 * the bitmap link_speeds of the struct rte_eth_conf 2081 * 2082 * @param speed 2083 * Numerical speed value in Mbps 2084 * @param duplex 2085 * RTE_ETH_LINK_[HALF/FULL]_DUPLEX (only for 10/100M speeds) 2086 * @return 2087 * 0 if the speed cannot be mapped 2088 */ 2089 uint32_t rte_eth_speed_bitflag(uint32_t speed, int duplex); 2090 2091 /** 2092 * Get RTE_ETH_RX_OFFLOAD_* flag name. 2093 * 2094 * @param offload 2095 * Offload flag. 2096 * @return 2097 * Offload name or 'UNKNOWN' if the flag cannot be recognised. 2098 */ 2099 const char *rte_eth_dev_rx_offload_name(uint64_t offload); 2100 2101 /** 2102 * Get RTE_ETH_TX_OFFLOAD_* flag name. 2103 * 2104 * @param offload 2105 * Offload flag. 2106 * @return 2107 * Offload name or 'UNKNOWN' if the flag cannot be recognised. 2108 */ 2109 const char *rte_eth_dev_tx_offload_name(uint64_t offload); 2110 2111 /** 2112 * @warning 2113 * @b EXPERIMENTAL: this API may change without prior notice. 2114 * 2115 * Get RTE_ETH_DEV_CAPA_* flag name. 2116 * 2117 * @param capability 2118 * Capability flag. 2119 * @return 2120 * Capability name or 'UNKNOWN' if the flag cannot be recognized. 2121 */ 2122 __rte_experimental 2123 const char *rte_eth_dev_capability_name(uint64_t capability); 2124 2125 /** 2126 * Configure an Ethernet device. 2127 * This function must be invoked first before any other function in the 2128 * Ethernet API. This function can also be re-invoked when a device is in the 2129 * stopped state. 2130 * 2131 * @param port_id 2132 * The port identifier of the Ethernet device to configure. 2133 * @param nb_rx_queue 2134 * The number of receive queues to set up for the Ethernet device. 2135 * @param nb_tx_queue 2136 * The number of transmit queues to set up for the Ethernet device. 2137 * @param eth_conf 2138 * The pointer to the configuration data to be used for the Ethernet device. 2139 * The *rte_eth_conf* structure includes: 2140 * - the hardware offload features to activate, with dedicated fields for 2141 * each statically configurable offload hardware feature provided by 2142 * Ethernet devices, such as IP checksum or VLAN tag stripping for 2143 * example. 2144 * The Rx offload bitfield API is obsolete and will be deprecated. 2145 * Applications should set the ignore_bitfield_offloads bit on *rxmode* 2146 * structure and use offloads field to set per-port offloads instead. 2147 * - Any offloading set in eth_conf->[rt]xmode.offloads must be within 2148 * the [rt]x_offload_capa returned from rte_eth_dev_info_get(). 2149 * Any type of device supported offloading set in the input argument 2150 * eth_conf->[rt]xmode.offloads to rte_eth_dev_configure() is enabled 2151 * on all queues and it can't be disabled in rte_eth_[rt]x_queue_setup() 2152 * - the Receive Side Scaling (RSS) configuration when using multiple Rx 2153 * queues per port. Any RSS hash function set in eth_conf->rss_conf.rss_hf 2154 * must be within the flow_type_rss_offloads provided by drivers via 2155 * rte_eth_dev_info_get() API. 2156 * 2157 * Embedding all configuration information in a single data structure 2158 * is the more flexible method that allows the addition of new features 2159 * without changing the syntax of the API. 2160 * @return 2161 * - 0: Success, device configured. 2162 * - <0: Error code returned by the driver configuration function. 2163 */ 2164 int rte_eth_dev_configure(uint16_t port_id, uint16_t nb_rx_queue, 2165 uint16_t nb_tx_queue, const struct rte_eth_conf *eth_conf); 2166 2167 /** 2168 * Check if an Ethernet device was physically removed. 2169 * 2170 * @param port_id 2171 * The port identifier of the Ethernet device. 2172 * @return 2173 * 1 when the Ethernet device is removed, otherwise 0. 2174 */ 2175 int 2176 rte_eth_dev_is_removed(uint16_t port_id); 2177 2178 /** 2179 * Allocate and set up a receive queue for an Ethernet device. 2180 * 2181 * The function allocates a contiguous block of memory for *nb_rx_desc* 2182 * receive descriptors from a memory zone associated with *socket_id* 2183 * and initializes each receive descriptor with a network buffer allocated 2184 * from the memory pool *mb_pool*. 2185 * 2186 * @param port_id 2187 * The port identifier of the Ethernet device. 2188 * @param rx_queue_id 2189 * The index of the receive queue to set up. 2190 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2191 * to rte_eth_dev_configure(). 2192 * @param nb_rx_desc 2193 * The number of receive descriptors to allocate for the receive ring. 2194 * @param socket_id 2195 * The *socket_id* argument is the socket identifier in case of NUMA. 2196 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for 2197 * the DMA memory allocated for the receive descriptors of the ring. 2198 * @param rx_conf 2199 * The pointer to the configuration data to be used for the receive queue. 2200 * NULL value is allowed, in which case default Rx configuration 2201 * will be used. 2202 * The *rx_conf* structure contains an *rx_thresh* structure with the values 2203 * of the Prefetch, Host, and Write-Back threshold registers of the receive 2204 * ring. 2205 * In addition it contains the hardware offloads features to activate using 2206 * the RTE_ETH_RX_OFFLOAD_* flags. 2207 * If an offloading set in rx_conf->offloads 2208 * hasn't been set in the input argument eth_conf->rxmode.offloads 2209 * to rte_eth_dev_configure(), it is a new added offloading, it must be 2210 * per-queue type and it is enabled for the queue. 2211 * No need to repeat any bit in rx_conf->offloads which has already been 2212 * enabled in rte_eth_dev_configure() at port level. An offloading enabled 2213 * at port level can't be disabled at queue level. 2214 * The configuration structure also contains the pointer to the array 2215 * of the receiving buffer segment descriptions, see rx_seg and rx_nseg 2216 * fields, this extended configuration might be used by split offloads like 2217 * RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT. If mb_pool is not NULL, 2218 * the extended configuration fields must be set to NULL and zero. 2219 * @param mb_pool 2220 * The pointer to the memory pool from which to allocate *rte_mbuf* network 2221 * memory buffers to populate each descriptor of the receive ring. There are 2222 * two options to provide Rx buffer configuration: 2223 * - single pool: 2224 * mb_pool is not NULL, rx_conf.rx_nseg is 0. 2225 * - multiple segments description: 2226 * mb_pool is NULL, rx_conf.rx_seg is not NULL, rx_conf.rx_nseg is not 0. 2227 * Taken only if flag RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT is set in offloads. 2228 * 2229 * @return 2230 * - 0: Success, receive queue correctly set up. 2231 * - -EIO: if device is removed. 2232 * - -ENODEV: if *port_id* is invalid. 2233 * - -EINVAL: The memory pool pointer is null or the size of network buffers 2234 * which can be allocated from this memory pool does not fit the various 2235 * buffer sizes allowed by the device controller. 2236 * - -ENOMEM: Unable to allocate the receive ring descriptors or to 2237 * allocate network memory buffers from the memory pool when 2238 * initializing receive descriptors. 2239 */ 2240 int rte_eth_rx_queue_setup(uint16_t port_id, uint16_t rx_queue_id, 2241 uint16_t nb_rx_desc, unsigned int socket_id, 2242 const struct rte_eth_rxconf *rx_conf, 2243 struct rte_mempool *mb_pool); 2244 2245 /** 2246 * @warning 2247 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 2248 * 2249 * Allocate and set up a hairpin receive queue for an Ethernet device. 2250 * 2251 * The function set up the selected queue to be used in hairpin. 2252 * 2253 * @param port_id 2254 * The port identifier of the Ethernet device. 2255 * @param rx_queue_id 2256 * The index of the receive queue to set up. 2257 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2258 * to rte_eth_dev_configure(). 2259 * @param nb_rx_desc 2260 * The number of receive descriptors to allocate for the receive ring. 2261 * 0 means the PMD will use default value. 2262 * @param conf 2263 * The pointer to the hairpin configuration. 2264 * 2265 * @return 2266 * - (0) if successful. 2267 * - (-ENODEV) if *port_id* is invalid. 2268 * - (-ENOTSUP) if hardware doesn't support. 2269 * - (-EINVAL) if bad parameter. 2270 * - (-ENOMEM) if unable to allocate the resources. 2271 */ 2272 __rte_experimental 2273 int rte_eth_rx_hairpin_queue_setup 2274 (uint16_t port_id, uint16_t rx_queue_id, uint16_t nb_rx_desc, 2275 const struct rte_eth_hairpin_conf *conf); 2276 2277 /** 2278 * Allocate and set up a transmit queue for an Ethernet device. 2279 * 2280 * @param port_id 2281 * The port identifier of the Ethernet device. 2282 * @param tx_queue_id 2283 * The index of the transmit queue to set up. 2284 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2285 * to rte_eth_dev_configure(). 2286 * @param nb_tx_desc 2287 * The number of transmit descriptors to allocate for the transmit ring. 2288 * @param socket_id 2289 * The *socket_id* argument is the socket identifier in case of NUMA. 2290 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for 2291 * the DMA memory allocated for the transmit descriptors of the ring. 2292 * @param tx_conf 2293 * The pointer to the configuration data to be used for the transmit queue. 2294 * NULL value is allowed, in which case default Tx configuration 2295 * will be used. 2296 * The *tx_conf* structure contains the following data: 2297 * - The *tx_thresh* structure with the values of the Prefetch, Host, and 2298 * Write-Back threshold registers of the transmit ring. 2299 * When setting Write-Back threshold to the value greater then zero, 2300 * *tx_rs_thresh* value should be explicitly set to one. 2301 * - The *tx_free_thresh* value indicates the [minimum] number of network 2302 * buffers that must be pending in the transmit ring to trigger their 2303 * [implicit] freeing by the driver transmit function. 2304 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit 2305 * descriptors that must be pending in the transmit ring before setting the 2306 * RS bit on a descriptor by the driver transmit function. 2307 * The *tx_rs_thresh* value should be less or equal then 2308 * *tx_free_thresh* value, and both of them should be less then 2309 * *nb_tx_desc* - 3. 2310 * - The *offloads* member contains Tx offloads to be enabled. 2311 * If an offloading set in tx_conf->offloads 2312 * hasn't been set in the input argument eth_conf->txmode.offloads 2313 * to rte_eth_dev_configure(), it is a new added offloading, it must be 2314 * per-queue type and it is enabled for the queue. 2315 * No need to repeat any bit in tx_conf->offloads which has already been 2316 * enabled in rte_eth_dev_configure() at port level. An offloading enabled 2317 * at port level can't be disabled at queue level. 2318 * 2319 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces 2320 * the transmit function to use default values. 2321 * @return 2322 * - 0: Success, the transmit queue is correctly set up. 2323 * - -ENOMEM: Unable to allocate the transmit ring descriptors. 2324 */ 2325 int rte_eth_tx_queue_setup(uint16_t port_id, uint16_t tx_queue_id, 2326 uint16_t nb_tx_desc, unsigned int socket_id, 2327 const struct rte_eth_txconf *tx_conf); 2328 2329 /** 2330 * @warning 2331 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 2332 * 2333 * Allocate and set up a transmit hairpin queue for an Ethernet device. 2334 * 2335 * @param port_id 2336 * The port identifier of the Ethernet device. 2337 * @param tx_queue_id 2338 * The index of the transmit queue to set up. 2339 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2340 * to rte_eth_dev_configure(). 2341 * @param nb_tx_desc 2342 * The number of transmit descriptors to allocate for the transmit ring. 2343 * 0 to set default PMD value. 2344 * @param conf 2345 * The hairpin configuration. 2346 * 2347 * @return 2348 * - (0) if successful. 2349 * - (-ENODEV) if *port_id* is invalid. 2350 * - (-ENOTSUP) if hardware doesn't support. 2351 * - (-EINVAL) if bad parameter. 2352 * - (-ENOMEM) if unable to allocate the resources. 2353 */ 2354 __rte_experimental 2355 int rte_eth_tx_hairpin_queue_setup 2356 (uint16_t port_id, uint16_t tx_queue_id, uint16_t nb_tx_desc, 2357 const struct rte_eth_hairpin_conf *conf); 2358 2359 /** 2360 * @warning 2361 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 2362 * 2363 * Get all the hairpin peer Rx / Tx ports of the current port. 2364 * The caller should ensure that the array is large enough to save the ports 2365 * list. 2366 * 2367 * @param port_id 2368 * The port identifier of the Ethernet device. 2369 * @param peer_ports 2370 * Pointer to the array to store the peer ports list. 2371 * @param len 2372 * Length of the array to store the port identifiers. 2373 * @param direction 2374 * Current port to peer port direction 2375 * positive - current used as Tx to get all peer Rx ports. 2376 * zero - current used as Rx to get all peer Tx ports. 2377 * 2378 * @return 2379 * - (0 or positive) actual peer ports number. 2380 * - (-EINVAL) if bad parameter. 2381 * - (-ENODEV) if *port_id* invalid 2382 * - (-ENOTSUP) if hardware doesn't support. 2383 * - Others detailed errors from PMDs. 2384 */ 2385 __rte_experimental 2386 int rte_eth_hairpin_get_peer_ports(uint16_t port_id, uint16_t *peer_ports, 2387 size_t len, uint32_t direction); 2388 2389 /** 2390 * @warning 2391 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 2392 * 2393 * Bind all hairpin Tx queues of one port to the Rx queues of the peer port. 2394 * It is only allowed to call this function after all hairpin queues are 2395 * configured properly and the devices are in started state. 2396 * 2397 * @param tx_port 2398 * The identifier of the Tx port. 2399 * @param rx_port 2400 * The identifier of peer Rx port. 2401 * RTE_MAX_ETHPORTS is allowed for the traversal of all devices. 2402 * Rx port ID could have the same value as Tx port ID. 2403 * 2404 * @return 2405 * - (0) if successful. 2406 * - (-ENODEV) if Tx port ID is invalid. 2407 * - (-EBUSY) if device is not in started state. 2408 * - (-ENOTSUP) if hardware doesn't support. 2409 * - Others detailed errors from PMDs. 2410 */ 2411 __rte_experimental 2412 int rte_eth_hairpin_bind(uint16_t tx_port, uint16_t rx_port); 2413 2414 /** 2415 * @warning 2416 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 2417 * 2418 * Unbind all hairpin Tx queues of one port from the Rx queues of the peer port. 2419 * This should be called before closing the Tx or Rx devices, if the bind 2420 * function is called before. 2421 * After unbinding the hairpin ports pair, it is allowed to bind them again. 2422 * Changing queues configuration should be after stopping the device(s). 2423 * 2424 * @param tx_port 2425 * The identifier of the Tx port. 2426 * @param rx_port 2427 * The identifier of peer Rx port. 2428 * RTE_MAX_ETHPORTS is allowed for traversal of all devices. 2429 * Rx port ID could have the same value as Tx port ID. 2430 * 2431 * @return 2432 * - (0) if successful. 2433 * - (-ENODEV) if Tx port ID is invalid. 2434 * - (-EBUSY) if device is in stopped state. 2435 * - (-ENOTSUP) if hardware doesn't support. 2436 * - Others detailed errors from PMDs. 2437 */ 2438 __rte_experimental 2439 int rte_eth_hairpin_unbind(uint16_t tx_port, uint16_t rx_port); 2440 2441 /** 2442 * Return the NUMA socket to which an Ethernet device is connected 2443 * 2444 * @param port_id 2445 * The port identifier of the Ethernet device 2446 * @return 2447 * The NUMA socket ID to which the Ethernet device is connected or 2448 * a default of zero if the socket could not be determined. 2449 * -1 is returned is the port_id value is out of range. 2450 */ 2451 int rte_eth_dev_socket_id(uint16_t port_id); 2452 2453 /** 2454 * Check if port_id of device is attached 2455 * 2456 * @param port_id 2457 * The port identifier of the Ethernet device 2458 * @return 2459 * - 0 if port is out of range or not attached 2460 * - 1 if device is attached 2461 */ 2462 int rte_eth_dev_is_valid_port(uint16_t port_id); 2463 2464 /** 2465 * Start specified Rx queue of a port. It is used when rx_deferred_start 2466 * flag of the specified queue is true. 2467 * 2468 * @param port_id 2469 * The port identifier of the Ethernet device 2470 * @param rx_queue_id 2471 * The index of the Rx queue to update the ring. 2472 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2473 * to rte_eth_dev_configure(). 2474 * @return 2475 * - 0: Success, the receive queue is started. 2476 * - -ENODEV: if *port_id* is invalid. 2477 * - -EINVAL: The queue_id out of range or belong to hairpin. 2478 * - -EIO: if device is removed. 2479 * - -ENOTSUP: The function not supported in PMD. 2480 */ 2481 int rte_eth_dev_rx_queue_start(uint16_t port_id, uint16_t rx_queue_id); 2482 2483 /** 2484 * Stop specified Rx queue of a port 2485 * 2486 * @param port_id 2487 * The port identifier of the Ethernet device 2488 * @param rx_queue_id 2489 * The index of the Rx queue to update the ring. 2490 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 2491 * to rte_eth_dev_configure(). 2492 * @return 2493 * - 0: Success, the receive queue is stopped. 2494 * - -ENODEV: if *port_id* is invalid. 2495 * - -EINVAL: The queue_id out of range or belong to hairpin. 2496 * - -EIO: if device is removed. 2497 * - -ENOTSUP: The function not supported in PMD. 2498 */ 2499 int rte_eth_dev_rx_queue_stop(uint16_t port_id, uint16_t rx_queue_id); 2500 2501 /** 2502 * Start Tx for specified queue of a port. It is used when tx_deferred_start 2503 * flag of the specified queue is true. 2504 * 2505 * @param port_id 2506 * The port identifier of the Ethernet device 2507 * @param tx_queue_id 2508 * The index of the Tx queue to update the ring. 2509 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2510 * to rte_eth_dev_configure(). 2511 * @return 2512 * - 0: Success, the transmit queue is started. 2513 * - -ENODEV: if *port_id* is invalid. 2514 * - -EINVAL: The queue_id out of range or belong to hairpin. 2515 * - -EIO: if device is removed. 2516 * - -ENOTSUP: The function not supported in PMD. 2517 */ 2518 int rte_eth_dev_tx_queue_start(uint16_t port_id, uint16_t tx_queue_id); 2519 2520 /** 2521 * Stop specified Tx queue of a port 2522 * 2523 * @param port_id 2524 * The port identifier of the Ethernet device 2525 * @param tx_queue_id 2526 * The index of the Tx queue to update the ring. 2527 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2528 * to rte_eth_dev_configure(). 2529 * @return 2530 * - 0: Success, the transmit queue is stopped. 2531 * - -ENODEV: if *port_id* is invalid. 2532 * - -EINVAL: The queue_id out of range or belong to hairpin. 2533 * - -EIO: if device is removed. 2534 * - -ENOTSUP: The function not supported in PMD. 2535 */ 2536 int rte_eth_dev_tx_queue_stop(uint16_t port_id, uint16_t tx_queue_id); 2537 2538 /** 2539 * Start an Ethernet device. 2540 * 2541 * The device start step is the last one and consists of setting the configured 2542 * offload features and in starting the transmit and the receive units of the 2543 * device. 2544 * 2545 * Device RTE_ETH_DEV_NOLIVE_MAC_ADDR flag causes MAC address to be set before 2546 * PMD port start callback function is invoked. 2547 * 2548 * On success, all basic functions exported by the Ethernet API (link status, 2549 * receive/transmit, and so on) can be invoked. 2550 * 2551 * @param port_id 2552 * The port identifier of the Ethernet device. 2553 * @return 2554 * - 0: Success, Ethernet device started. 2555 * - <0: Error code of the driver device start function. 2556 */ 2557 int rte_eth_dev_start(uint16_t port_id); 2558 2559 /** 2560 * Stop an Ethernet device. The device can be restarted with a call to 2561 * rte_eth_dev_start() 2562 * 2563 * @param port_id 2564 * The port identifier of the Ethernet device. 2565 * @return 2566 * - 0: Success, Ethernet device stopped. 2567 * - <0: Error code of the driver device stop function. 2568 */ 2569 int rte_eth_dev_stop(uint16_t port_id); 2570 2571 /** 2572 * Link up an Ethernet device. 2573 * 2574 * Set device link up will re-enable the device Rx/Tx 2575 * functionality after it is previously set device linked down. 2576 * 2577 * @param port_id 2578 * The port identifier of the Ethernet device. 2579 * @return 2580 * - 0: Success, Ethernet device linked up. 2581 * - <0: Error code of the driver device link up function. 2582 */ 2583 int rte_eth_dev_set_link_up(uint16_t port_id); 2584 2585 /** 2586 * Link down an Ethernet device. 2587 * The device Rx/Tx functionality will be disabled if success, 2588 * and it can be re-enabled with a call to 2589 * rte_eth_dev_set_link_up() 2590 * 2591 * @param port_id 2592 * The port identifier of the Ethernet device. 2593 */ 2594 int rte_eth_dev_set_link_down(uint16_t port_id); 2595 2596 /** 2597 * Close a stopped Ethernet device. The device cannot be restarted! 2598 * The function frees all port resources. 2599 * 2600 * @param port_id 2601 * The port identifier of the Ethernet device. 2602 * @return 2603 * - Zero if the port is closed successfully. 2604 * - Negative if something went wrong. 2605 */ 2606 int rte_eth_dev_close(uint16_t port_id); 2607 2608 /** 2609 * Reset a Ethernet device and keep its port ID. 2610 * 2611 * When a port has to be reset passively, the DPDK application can invoke 2612 * this function. For example when a PF is reset, all its VFs should also 2613 * be reset. Normally a DPDK application can invoke this function when 2614 * RTE_ETH_EVENT_INTR_RESET event is detected, but can also use it to start 2615 * a port reset in other circumstances. 2616 * 2617 * When this function is called, it first stops the port and then calls the 2618 * PMD specific dev_uninit( ) and dev_init( ) to return the port to initial 2619 * state, in which no Tx and Rx queues are setup, as if the port has been 2620 * reset and not started. The port keeps the port ID it had before the 2621 * function call. 2622 * 2623 * After calling rte_eth_dev_reset( ), the application should use 2624 * rte_eth_dev_configure( ), rte_eth_rx_queue_setup( ), 2625 * rte_eth_tx_queue_setup( ), and rte_eth_dev_start( ) 2626 * to reconfigure the device as appropriate. 2627 * 2628 * Note: To avoid unexpected behavior, the application should stop calling 2629 * Tx and Rx functions before calling rte_eth_dev_reset( ). For thread 2630 * safety, all these controlling functions should be called from the same 2631 * thread. 2632 * 2633 * @param port_id 2634 * The port identifier of the Ethernet device. 2635 * 2636 * @return 2637 * - (0) if successful. 2638 * - (-ENODEV) if *port_id* is invalid. 2639 * - (-ENOTSUP) if hardware doesn't support this function. 2640 * - (-EPERM) if not ran from the primary process. 2641 * - (-EIO) if re-initialisation failed or device is removed. 2642 * - (-ENOMEM) if the reset failed due to OOM. 2643 * - (-EAGAIN) if the reset temporarily failed and should be retried later. 2644 */ 2645 int rte_eth_dev_reset(uint16_t port_id); 2646 2647 /** 2648 * Enable receipt in promiscuous mode for an Ethernet device. 2649 * 2650 * @param port_id 2651 * The port identifier of the Ethernet device. 2652 * @return 2653 * - (0) if successful. 2654 * - (-ENOTSUP) if support for promiscuous_enable() does not exist 2655 * for the device. 2656 * - (-ENODEV) if *port_id* invalid. 2657 */ 2658 int rte_eth_promiscuous_enable(uint16_t port_id); 2659 2660 /** 2661 * Disable receipt in promiscuous mode for an Ethernet device. 2662 * 2663 * @param port_id 2664 * The port identifier of the Ethernet device. 2665 * @return 2666 * - (0) if successful. 2667 * - (-ENOTSUP) if support for promiscuous_disable() does not exist 2668 * for the device. 2669 * - (-ENODEV) if *port_id* invalid. 2670 */ 2671 int rte_eth_promiscuous_disable(uint16_t port_id); 2672 2673 /** 2674 * Return the value of promiscuous mode for an Ethernet device. 2675 * 2676 * @param port_id 2677 * The port identifier of the Ethernet device. 2678 * @return 2679 * - (1) if promiscuous is enabled 2680 * - (0) if promiscuous is disabled. 2681 * - (-1) on error 2682 */ 2683 int rte_eth_promiscuous_get(uint16_t port_id); 2684 2685 /** 2686 * Enable the receipt of any multicast frame by an Ethernet device. 2687 * 2688 * @param port_id 2689 * The port identifier of the Ethernet device. 2690 * @return 2691 * - (0) if successful. 2692 * - (-ENOTSUP) if support for allmulticast_enable() does not exist 2693 * for the device. 2694 * - (-ENODEV) if *port_id* invalid. 2695 */ 2696 int rte_eth_allmulticast_enable(uint16_t port_id); 2697 2698 /** 2699 * Disable the receipt of all multicast frames by an Ethernet device. 2700 * 2701 * @param port_id 2702 * The port identifier of the Ethernet device. 2703 * @return 2704 * - (0) if successful. 2705 * - (-ENOTSUP) if support for allmulticast_disable() does not exist 2706 * for the device. 2707 * - (-ENODEV) if *port_id* invalid. 2708 */ 2709 int rte_eth_allmulticast_disable(uint16_t port_id); 2710 2711 /** 2712 * Return the value of allmulticast mode for an Ethernet device. 2713 * 2714 * @param port_id 2715 * The port identifier of the Ethernet device. 2716 * @return 2717 * - (1) if allmulticast is enabled 2718 * - (0) if allmulticast is disabled. 2719 * - (-1) on error 2720 */ 2721 int rte_eth_allmulticast_get(uint16_t port_id); 2722 2723 /** 2724 * Retrieve the link status (up/down), the duplex mode (half/full), 2725 * the negotiation (auto/fixed), and if available, the speed (Mbps). 2726 * 2727 * It might need to wait up to 9 seconds. 2728 * @see rte_eth_link_get_nowait. 2729 * 2730 * @param port_id 2731 * The port identifier of the Ethernet device. 2732 * @param link 2733 * Link information written back. 2734 * @return 2735 * - (0) if successful. 2736 * - (-ENOTSUP) if the function is not supported in PMD. 2737 * - (-ENODEV) if *port_id* invalid. 2738 * - (-EINVAL) if bad parameter. 2739 */ 2740 int rte_eth_link_get(uint16_t port_id, struct rte_eth_link *link); 2741 2742 /** 2743 * Retrieve the link status (up/down), the duplex mode (half/full), 2744 * the negotiation (auto/fixed), and if available, the speed (Mbps). 2745 * 2746 * @param port_id 2747 * The port identifier of the Ethernet device. 2748 * @param link 2749 * Link information written back. 2750 * @return 2751 * - (0) if successful. 2752 * - (-ENOTSUP) if the function is not supported in PMD. 2753 * - (-ENODEV) if *port_id* invalid. 2754 * - (-EINVAL) if bad parameter. 2755 */ 2756 int rte_eth_link_get_nowait(uint16_t port_id, struct rte_eth_link *link); 2757 2758 /** 2759 * @warning 2760 * @b EXPERIMENTAL: this API may change without prior notice. 2761 * 2762 * The function converts a link_speed to a string. It handles all special 2763 * values like unknown or none speed. 2764 * 2765 * @param link_speed 2766 * link_speed of rte_eth_link struct 2767 * @return 2768 * Link speed in textual format. It's pointer to immutable memory. 2769 * No free is required. 2770 */ 2771 __rte_experimental 2772 const char *rte_eth_link_speed_to_str(uint32_t link_speed); 2773 2774 /** 2775 * @warning 2776 * @b EXPERIMENTAL: this API may change without prior notice. 2777 * 2778 * The function converts a rte_eth_link struct representing a link status to 2779 * a string. 2780 * 2781 * @param str 2782 * A pointer to a string to be filled with textual representation of 2783 * device status. At least RTE_ETH_LINK_MAX_STR_LEN bytes should be allocated to 2784 * store default link status text. 2785 * @param len 2786 * Length of available memory at 'str' string. 2787 * @param eth_link 2788 * Link status returned by rte_eth_link_get function 2789 * @return 2790 * Number of bytes written to str array or -EINVAL if bad parameter. 2791 */ 2792 __rte_experimental 2793 int rte_eth_link_to_str(char *str, size_t len, 2794 const struct rte_eth_link *eth_link); 2795 2796 /** 2797 * Retrieve the general I/O statistics of an Ethernet device. 2798 * 2799 * @param port_id 2800 * The port identifier of the Ethernet device. 2801 * @param stats 2802 * A pointer to a structure of type *rte_eth_stats* to be filled with 2803 * the values of device counters for the following set of statistics: 2804 * - *ipackets* with the total of successfully received packets. 2805 * - *opackets* with the total of successfully transmitted packets. 2806 * - *ibytes* with the total of successfully received bytes. 2807 * - *obytes* with the total of successfully transmitted bytes. 2808 * - *ierrors* with the total of erroneous received packets. 2809 * - *oerrors* with the total of failed transmitted packets. 2810 * @return 2811 * Zero if successful. Non-zero otherwise. 2812 */ 2813 int rte_eth_stats_get(uint16_t port_id, struct rte_eth_stats *stats); 2814 2815 /** 2816 * Reset the general I/O statistics of an Ethernet device. 2817 * 2818 * @param port_id 2819 * The port identifier of the Ethernet device. 2820 * @return 2821 * - (0) if device notified to reset stats. 2822 * - (-ENOTSUP) if hardware doesn't support. 2823 * - (-ENODEV) if *port_id* invalid. 2824 * - (<0): Error code of the driver stats reset function. 2825 */ 2826 int rte_eth_stats_reset(uint16_t port_id); 2827 2828 /** 2829 * Retrieve names of extended statistics of an Ethernet device. 2830 * 2831 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched 2832 * by array index: 2833 * xstats_names[i].name => xstats[i].value 2834 * 2835 * And the array index is same with id field of 'struct rte_eth_xstat': 2836 * xstats[i].id == i 2837 * 2838 * This assumption makes key-value pair matching less flexible but simpler. 2839 * 2840 * @param port_id 2841 * The port identifier of the Ethernet device. 2842 * @param xstats_names 2843 * An rte_eth_xstat_name array of at least *size* elements to 2844 * be filled. If set to NULL, the function returns the required number 2845 * of elements. 2846 * @param size 2847 * The size of the xstats_names array (number of elements). 2848 * @return 2849 * - A positive value lower or equal to size: success. The return value 2850 * is the number of entries filled in the stats table. 2851 * - A positive value higher than size: error, the given statistics table 2852 * is too small. The return value corresponds to the size that should 2853 * be given to succeed. The entries in the table are not valid and 2854 * shall not be used by the caller. 2855 * - A negative value on error (invalid port ID). 2856 */ 2857 int rte_eth_xstats_get_names(uint16_t port_id, 2858 struct rte_eth_xstat_name *xstats_names, 2859 unsigned int size); 2860 2861 /** 2862 * Retrieve extended statistics of an Ethernet device. 2863 * 2864 * There is an assumption that 'xstat_names' and 'xstats' arrays are matched 2865 * by array index: 2866 * xstats_names[i].name => xstats[i].value 2867 * 2868 * And the array index is same with id field of 'struct rte_eth_xstat': 2869 * xstats[i].id == i 2870 * 2871 * This assumption makes key-value pair matching less flexible but simpler. 2872 * 2873 * @param port_id 2874 * The port identifier of the Ethernet device. 2875 * @param xstats 2876 * A pointer to a table of structure of type *rte_eth_xstat* 2877 * to be filled with device statistics ids and values. 2878 * This parameter can be set to NULL if and only if n is 0. 2879 * @param n 2880 * The size of the xstats array (number of elements). 2881 * If lower than the required number of elements, the function returns 2882 * the required number of elements. 2883 * If equal to zero, the xstats must be NULL, the function returns the 2884 * required number of elements. 2885 * @return 2886 * - A positive value lower or equal to n: success. The return value 2887 * is the number of entries filled in the stats table. 2888 * - A positive value higher than n: error, the given statistics table 2889 * is too small. The return value corresponds to the size that should 2890 * be given to succeed. The entries in the table are not valid and 2891 * shall not be used by the caller. 2892 * - A negative value on error (invalid port ID). 2893 */ 2894 int rte_eth_xstats_get(uint16_t port_id, struct rte_eth_xstat *xstats, 2895 unsigned int n); 2896 2897 /** 2898 * Retrieve names of extended statistics of an Ethernet device. 2899 * 2900 * @param port_id 2901 * The port identifier of the Ethernet device. 2902 * @param xstats_names 2903 * Array to be filled in with names of requested device statistics. 2904 * Must not be NULL if @p ids are specified (not NULL). 2905 * @param size 2906 * Number of elements in @p xstats_names array (if not NULL) and in 2907 * @p ids array (if not NULL). Must be 0 if both array pointers are NULL. 2908 * @param ids 2909 * IDs array given by app to retrieve specific statistics. May be NULL to 2910 * retrieve names of all available statistics or, if @p xstats_names is 2911 * NULL as well, just the number of available statistics. 2912 * @return 2913 * - A positive value lower or equal to size: success. The return value 2914 * is the number of entries filled in the stats table. 2915 * - A positive value higher than size: success. The given statistics table 2916 * is too small. The return value corresponds to the size that should 2917 * be given to succeed. The entries in the table are not valid and 2918 * shall not be used by the caller. 2919 * - A negative value on error. 2920 */ 2921 int 2922 rte_eth_xstats_get_names_by_id(uint16_t port_id, 2923 struct rte_eth_xstat_name *xstats_names, unsigned int size, 2924 uint64_t *ids); 2925 2926 /** 2927 * Retrieve extended statistics of an Ethernet device. 2928 * 2929 * @param port_id 2930 * The port identifier of the Ethernet device. 2931 * @param ids 2932 * IDs array given by app to retrieve specific statistics. May be NULL to 2933 * retrieve all available statistics or, if @p values is NULL as well, 2934 * just the number of available statistics. 2935 * @param values 2936 * Array to be filled in with requested device statistics. 2937 * Must not be NULL if ids are specified (not NULL). 2938 * @param size 2939 * Number of elements in @p values array (if not NULL) and in @p ids 2940 * array (if not NULL). Must be 0 if both array pointers are NULL. 2941 * @return 2942 * - A positive value lower or equal to size: success. The return value 2943 * is the number of entries filled in the stats table. 2944 * - A positive value higher than size: success: The given statistics table 2945 * is too small. The return value corresponds to the size that should 2946 * be given to succeed. The entries in the table are not valid and 2947 * shall not be used by the caller. 2948 * - A negative value on error. 2949 */ 2950 int rte_eth_xstats_get_by_id(uint16_t port_id, const uint64_t *ids, 2951 uint64_t *values, unsigned int size); 2952 2953 /** 2954 * Gets the ID of a statistic from its name. 2955 * 2956 * This function searches for the statistics using string compares, and 2957 * as such should not be used on the fast-path. For fast-path retrieval of 2958 * specific statistics, store the ID as provided in *id* from this function, 2959 * and pass the ID to rte_eth_xstats_get() 2960 * 2961 * @param port_id The port to look up statistics from 2962 * @param xstat_name The name of the statistic to return 2963 * @param[out] id A pointer to an app-supplied uint64_t which should be 2964 * set to the ID of the stat if the stat exists. 2965 * @return 2966 * 0 on success 2967 * -ENODEV for invalid port_id, 2968 * -EIO if device is removed, 2969 * -EINVAL if the xstat_name doesn't exist in port_id 2970 * -ENOMEM if bad parameter. 2971 */ 2972 int rte_eth_xstats_get_id_by_name(uint16_t port_id, const char *xstat_name, 2973 uint64_t *id); 2974 2975 /** 2976 * Reset extended statistics of an Ethernet device. 2977 * 2978 * @param port_id 2979 * The port identifier of the Ethernet device. 2980 * @return 2981 * - (0) if device notified to reset extended stats. 2982 * - (-ENOTSUP) if pmd doesn't support both 2983 * extended stats and basic stats reset. 2984 * - (-ENODEV) if *port_id* invalid. 2985 * - (<0): Error code of the driver xstats reset function. 2986 */ 2987 int rte_eth_xstats_reset(uint16_t port_id); 2988 2989 /** 2990 * Set a mapping for the specified transmit queue to the specified per-queue 2991 * statistics counter. 2992 * 2993 * @param port_id 2994 * The port identifier of the Ethernet device. 2995 * @param tx_queue_id 2996 * The index of the transmit queue for which a queue stats mapping is required. 2997 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 2998 * to rte_eth_dev_configure(). 2999 * @param stat_idx 3000 * The per-queue packet statistics functionality number that the transmit 3001 * queue is to be assigned. 3002 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1]. 3003 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256. 3004 * @return 3005 * Zero if successful. Non-zero otherwise. 3006 */ 3007 int rte_eth_dev_set_tx_queue_stats_mapping(uint16_t port_id, 3008 uint16_t tx_queue_id, uint8_t stat_idx); 3009 3010 /** 3011 * Set a mapping for the specified receive queue to the specified per-queue 3012 * statistics counter. 3013 * 3014 * @param port_id 3015 * The port identifier of the Ethernet device. 3016 * @param rx_queue_id 3017 * The index of the receive queue for which a queue stats mapping is required. 3018 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3019 * to rte_eth_dev_configure(). 3020 * @param stat_idx 3021 * The per-queue packet statistics functionality number that the receive 3022 * queue is to be assigned. 3023 * The value must be in the range [0, RTE_ETHDEV_QUEUE_STAT_CNTRS - 1]. 3024 * Max RTE_ETHDEV_QUEUE_STAT_CNTRS being 256. 3025 * @return 3026 * Zero if successful. Non-zero otherwise. 3027 */ 3028 int rte_eth_dev_set_rx_queue_stats_mapping(uint16_t port_id, 3029 uint16_t rx_queue_id, 3030 uint8_t stat_idx); 3031 3032 /** 3033 * Retrieve the Ethernet address of an Ethernet device. 3034 * 3035 * @param port_id 3036 * The port identifier of the Ethernet device. 3037 * @param mac_addr 3038 * A pointer to a structure of type *ether_addr* to be filled with 3039 * the Ethernet address of the Ethernet device. 3040 * @return 3041 * - (0) if successful 3042 * - (-ENODEV) if *port_id* invalid. 3043 * - (-EINVAL) if bad parameter. 3044 */ 3045 int rte_eth_macaddr_get(uint16_t port_id, struct rte_ether_addr *mac_addr); 3046 3047 /** 3048 * @warning 3049 * @b EXPERIMENTAL: this API may change without prior notice 3050 * 3051 * Retrieve the Ethernet addresses of an Ethernet device. 3052 * 3053 * @param port_id 3054 * The port identifier of the Ethernet device. 3055 * @param ma 3056 * A pointer to an array of structures of type *ether_addr* to be filled with 3057 * the Ethernet addresses of the Ethernet device. 3058 * @param num 3059 * Number of elements in the @p ma array. 3060 * Note that rte_eth_dev_info::max_mac_addrs can be used to retrieve 3061 * max number of Ethernet addresses for given port. 3062 * @return 3063 * - number of retrieved addresses if successful 3064 * - (-ENODEV) if *port_id* invalid. 3065 * - (-EINVAL) if bad parameter. 3066 */ 3067 __rte_experimental 3068 int rte_eth_macaddrs_get(uint16_t port_id, struct rte_ether_addr *ma, 3069 unsigned int num); 3070 3071 /** 3072 * Retrieve the contextual information of an Ethernet device. 3073 * 3074 * As part of this function, a number of of fields in dev_info will be 3075 * initialized as follows: 3076 * 3077 * rx_desc_lim = lim 3078 * tx_desc_lim = lim 3079 * 3080 * Where lim is defined within the rte_eth_dev_info_get as 3081 * 3082 * const struct rte_eth_desc_lim lim = { 3083 * .nb_max = UINT16_MAX, 3084 * .nb_min = 0, 3085 * .nb_align = 1, 3086 * .nb_seg_max = UINT16_MAX, 3087 * .nb_mtu_seg_max = UINT16_MAX, 3088 * }; 3089 * 3090 * device = dev->device 3091 * min_mtu = RTE_ETHER_MIN_LEN - RTE_ETHER_HDR_LEN - RTE_ETHER_CRC_LEN 3092 * max_mtu = UINT16_MAX 3093 * 3094 * The following fields will be populated if support for dev_infos_get() 3095 * exists for the device and the rte_eth_dev 'dev' has been populated 3096 * successfully with a call to it: 3097 * 3098 * driver_name = rte_driver_name(rte_dev_driver(dev->device)); 3099 * nb_rx_queues = dev->data->nb_rx_queues 3100 * nb_tx_queues = dev->data->nb_tx_queues 3101 * dev_flags = &dev->data->dev_flags 3102 * 3103 * @param port_id 3104 * The port identifier of the Ethernet device. 3105 * @param dev_info 3106 * A pointer to a structure of type *rte_eth_dev_info* to be filled with 3107 * the contextual information of the Ethernet device. 3108 * @return 3109 * - (0) if successful. 3110 * - (-ENOTSUP) if support for dev_infos_get() does not exist for the device. 3111 * - (-ENODEV) if *port_id* invalid. 3112 * - (-EINVAL) if bad parameter. 3113 */ 3114 int rte_eth_dev_info_get(uint16_t port_id, struct rte_eth_dev_info *dev_info); 3115 3116 /** 3117 * @warning 3118 * @b EXPERIMENTAL: this API may change without prior notice. 3119 * 3120 * Retrieve the configuration of an Ethernet device. 3121 * 3122 * @param port_id 3123 * The port identifier of the Ethernet device. 3124 * @param dev_conf 3125 * Location for Ethernet device configuration to be filled in. 3126 * @return 3127 * - (0) if successful. 3128 * - (-ENODEV) if *port_id* invalid. 3129 * - (-EINVAL) if bad parameter. 3130 */ 3131 __rte_experimental 3132 int rte_eth_dev_conf_get(uint16_t port_id, struct rte_eth_conf *dev_conf); 3133 3134 /** 3135 * Retrieve the firmware version of a device. 3136 * 3137 * @param port_id 3138 * The port identifier of the device. 3139 * @param fw_version 3140 * A pointer to a string array storing the firmware version of a device, 3141 * the string includes terminating null. This pointer is allocated by caller. 3142 * @param fw_size 3143 * The size of the string array pointed by fw_version, which should be 3144 * large enough to store firmware version of the device. 3145 * @return 3146 * - (0) if successful. 3147 * - (-ENOTSUP) if operation is not supported. 3148 * - (-ENODEV) if *port_id* invalid. 3149 * - (-EIO) if device is removed. 3150 * - (-EINVAL) if bad parameter. 3151 * - (>0) if *fw_size* is not enough to store firmware version, return 3152 * the size of the non truncated string. 3153 */ 3154 int rte_eth_dev_fw_version_get(uint16_t port_id, 3155 char *fw_version, size_t fw_size); 3156 3157 /** 3158 * Retrieve the supported packet types of an Ethernet device. 3159 * 3160 * When a packet type is announced as supported, it *must* be recognized by 3161 * the PMD. For instance, if RTE_PTYPE_L2_ETHER, RTE_PTYPE_L2_ETHER_VLAN 3162 * and RTE_PTYPE_L3_IPV4 are announced, the PMD must return the following 3163 * packet types for these packets: 3164 * - Ether/IPv4 -> RTE_PTYPE_L2_ETHER | RTE_PTYPE_L3_IPV4 3165 * - Ether/VLAN/IPv4 -> RTE_PTYPE_L2_ETHER_VLAN | RTE_PTYPE_L3_IPV4 3166 * - Ether/[anything else] -> RTE_PTYPE_L2_ETHER 3167 * - Ether/VLAN/[anything else] -> RTE_PTYPE_L2_ETHER_VLAN 3168 * 3169 * When a packet is received by a PMD, the most precise type must be 3170 * returned among the ones supported. However a PMD is allowed to set 3171 * packet type that is not in the supported list, at the condition that it 3172 * is more precise. Therefore, a PMD announcing no supported packet types 3173 * can still set a matching packet type in a received packet. 3174 * 3175 * @note 3176 * Better to invoke this API after the device is already started or Rx burst 3177 * function is decided, to obtain correct supported ptypes. 3178 * @note 3179 * if a given PMD does not report what ptypes it supports, then the supported 3180 * ptype count is reported as 0. 3181 * @param port_id 3182 * The port identifier of the Ethernet device. 3183 * @param ptype_mask 3184 * A hint of what kind of packet type which the caller is interested in. 3185 * @param ptypes 3186 * An array pointer to store adequate packet types, allocated by caller. 3187 * @param num 3188 * Size of the array pointed by param ptypes. 3189 * @return 3190 * - (>=0) Number of supported ptypes. If the number of types exceeds num, 3191 * only num entries will be filled into the ptypes array, but the full 3192 * count of supported ptypes will be returned. 3193 * - (-ENODEV) if *port_id* invalid. 3194 * - (-EINVAL) if bad parameter. 3195 */ 3196 int rte_eth_dev_get_supported_ptypes(uint16_t port_id, uint32_t ptype_mask, 3197 uint32_t *ptypes, int num); 3198 /** 3199 * Inform Ethernet device about reduced range of packet types to handle. 3200 * 3201 * Application can use this function to set only specific ptypes that it's 3202 * interested. This information can be used by the PMD to optimize Rx path. 3203 * 3204 * The function accepts an array `set_ptypes` allocated by the caller to 3205 * store the packet types set by the driver, the last element of the array 3206 * is set to RTE_PTYPE_UNKNOWN. The size of the `set_ptype` array should be 3207 * `rte_eth_dev_get_supported_ptypes() + 1` else it might only be filled 3208 * partially. 3209 * 3210 * @param port_id 3211 * The port identifier of the Ethernet device. 3212 * @param ptype_mask 3213 * The ptype family that application is interested in should be bitwise OR of 3214 * RTE_PTYPE_*_MASK or 0. 3215 * @param set_ptypes 3216 * An array pointer to store set packet types, allocated by caller. The 3217 * function marks the end of array with RTE_PTYPE_UNKNOWN. 3218 * @param num 3219 * Size of the array pointed by param ptypes. 3220 * Should be rte_eth_dev_get_supported_ptypes() + 1 to accommodate the 3221 * set ptypes. 3222 * @return 3223 * - (0) if Success. 3224 * - (-ENODEV) if *port_id* invalid. 3225 * - (-EINVAL) if *ptype_mask* is invalid (or) set_ptypes is NULL and 3226 * num > 0. 3227 */ 3228 int rte_eth_dev_set_ptypes(uint16_t port_id, uint32_t ptype_mask, 3229 uint32_t *set_ptypes, unsigned int num); 3230 3231 /** 3232 * Retrieve the MTU of an Ethernet device. 3233 * 3234 * @param port_id 3235 * The port identifier of the Ethernet device. 3236 * @param mtu 3237 * A pointer to a uint16_t where the retrieved MTU is to be stored. 3238 * @return 3239 * - (0) if successful. 3240 * - (-ENODEV) if *port_id* invalid. 3241 * - (-EINVAL) if bad parameter. 3242 */ 3243 int rte_eth_dev_get_mtu(uint16_t port_id, uint16_t *mtu); 3244 3245 /** 3246 * Change the MTU of an Ethernet device. 3247 * 3248 * @param port_id 3249 * The port identifier of the Ethernet device. 3250 * @param mtu 3251 * A uint16_t for the MTU to be applied. 3252 * @return 3253 * - (0) if successful. 3254 * - (-ENOTSUP) if operation is not supported. 3255 * - (-ENODEV) if *port_id* invalid. 3256 * - (-EIO) if device is removed. 3257 * - (-EINVAL) if *mtu* invalid, validation of mtu can occur within 3258 * rte_eth_dev_set_mtu if dev_infos_get is supported by the device or 3259 * when the mtu is set using dev->dev_ops->mtu_set. 3260 * - (-EBUSY) if operation is not allowed when the port is running 3261 */ 3262 int rte_eth_dev_set_mtu(uint16_t port_id, uint16_t mtu); 3263 3264 /** 3265 * Enable/Disable hardware filtering by an Ethernet device of received 3266 * VLAN packets tagged with a given VLAN Tag Identifier. 3267 * 3268 * @param port_id 3269 * The port identifier of the Ethernet device. 3270 * @param vlan_id 3271 * The VLAN Tag Identifier whose filtering must be enabled or disabled. 3272 * @param on 3273 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*. 3274 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*. 3275 * @return 3276 * - (0) if successful. 3277 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured. 3278 * - (-ENODEV) if *port_id* invalid. 3279 * - (-EIO) if device is removed. 3280 * - (-ENOSYS) if VLAN filtering on *port_id* disabled. 3281 * - (-EINVAL) if *vlan_id* > 4095. 3282 */ 3283 int rte_eth_dev_vlan_filter(uint16_t port_id, uint16_t vlan_id, int on); 3284 3285 /** 3286 * Enable/Disable hardware VLAN Strip by a Rx queue of an Ethernet device. 3287 * 3288 * @param port_id 3289 * The port identifier of the Ethernet device. 3290 * @param rx_queue_id 3291 * The index of the receive queue for which a queue stats mapping is required. 3292 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3293 * to rte_eth_dev_configure(). 3294 * @param on 3295 * If 1, Enable VLAN Stripping of the receive queue of the Ethernet port. 3296 * If 0, Disable VLAN Stripping of the receive queue of the Ethernet port. 3297 * @return 3298 * - (0) if successful. 3299 * - (-ENOTSUP) if hardware-assisted VLAN stripping not configured. 3300 * - (-ENODEV) if *port_id* invalid. 3301 * - (-EINVAL) if *rx_queue_id* invalid. 3302 */ 3303 int rte_eth_dev_set_vlan_strip_on_queue(uint16_t port_id, uint16_t rx_queue_id, 3304 int on); 3305 3306 /** 3307 * Set the Outer VLAN Ether Type by an Ethernet device, it can be inserted to 3308 * the VLAN header. 3309 * 3310 * @param port_id 3311 * The port identifier of the Ethernet device. 3312 * @param vlan_type 3313 * The VLAN type. 3314 * @param tag_type 3315 * The Tag Protocol ID 3316 * @return 3317 * - (0) if successful. 3318 * - (-ENOTSUP) if hardware-assisted VLAN TPID setup is not supported. 3319 * - (-ENODEV) if *port_id* invalid. 3320 * - (-EIO) if device is removed. 3321 */ 3322 int rte_eth_dev_set_vlan_ether_type(uint16_t port_id, 3323 enum rte_vlan_type vlan_type, 3324 uint16_t tag_type); 3325 3326 /** 3327 * Set VLAN offload configuration on an Ethernet device. 3328 * 3329 * @param port_id 3330 * The port identifier of the Ethernet device. 3331 * @param offload_mask 3332 * The VLAN Offload bit mask can be mixed use with "OR" 3333 * RTE_ETH_VLAN_STRIP_OFFLOAD 3334 * RTE_ETH_VLAN_FILTER_OFFLOAD 3335 * RTE_ETH_VLAN_EXTEND_OFFLOAD 3336 * RTE_ETH_QINQ_STRIP_OFFLOAD 3337 * @return 3338 * - (0) if successful. 3339 * - (-ENOTSUP) if hardware-assisted VLAN filtering not configured. 3340 * - (-ENODEV) if *port_id* invalid. 3341 * - (-EIO) if device is removed. 3342 */ 3343 int rte_eth_dev_set_vlan_offload(uint16_t port_id, int offload_mask); 3344 3345 /** 3346 * Read VLAN Offload configuration from an Ethernet device 3347 * 3348 * @param port_id 3349 * The port identifier of the Ethernet device. 3350 * @return 3351 * - (>0) if successful. Bit mask to indicate 3352 * RTE_ETH_VLAN_STRIP_OFFLOAD 3353 * RTE_ETH_VLAN_FILTER_OFFLOAD 3354 * RTE_ETH_VLAN_EXTEND_OFFLOAD 3355 * RTE_ETH_QINQ_STRIP_OFFLOAD 3356 * - (-ENODEV) if *port_id* invalid. 3357 */ 3358 int rte_eth_dev_get_vlan_offload(uint16_t port_id); 3359 3360 /** 3361 * Set port based Tx VLAN insertion on or off. 3362 * 3363 * @param port_id 3364 * The port identifier of the Ethernet device. 3365 * @param pvid 3366 * Port based Tx VLAN identifier together with user priority. 3367 * @param on 3368 * Turn on or off the port based Tx VLAN insertion. 3369 * 3370 * @return 3371 * - (0) if successful. 3372 * - negative if failed. 3373 */ 3374 int rte_eth_dev_set_vlan_pvid(uint16_t port_id, uint16_t pvid, int on); 3375 3376 /** 3377 * @warning 3378 * @b EXPERIMENTAL: this API may change without prior notice. 3379 * 3380 * Set Rx queue available descriptors threshold. 3381 * 3382 * @param port_id 3383 * The port identifier of the Ethernet device. 3384 * @param queue_id 3385 * The index of the receive queue. 3386 * @param avail_thresh 3387 * The available descriptors threshold is percentage of Rx queue size 3388 * which describes the availability of Rx queue for hardware. 3389 * If the Rx queue availability is below it, 3390 * the event RTE_ETH_EVENT_RX_AVAIL_THRESH is triggered. 3391 * [1-99] to set a new available descriptors threshold. 3392 * 0 to disable threshold monitoring. 3393 * 3394 * @return 3395 * - 0 if successful. 3396 * - (-ENODEV) if @p port_id is invalid. 3397 * - (-EINVAL) if bad parameter. 3398 * - (-ENOTSUP) if available Rx descriptors threshold is not supported. 3399 * - (-EIO) if device is removed. 3400 */ 3401 __rte_experimental 3402 int rte_eth_rx_avail_thresh_set(uint16_t port_id, uint16_t queue_id, 3403 uint8_t avail_thresh); 3404 3405 /** 3406 * @warning 3407 * @b EXPERIMENTAL: this API may change without prior notice. 3408 * 3409 * Find Rx queue with RTE_ETH_EVENT_RX_AVAIL_THRESH event pending. 3410 * 3411 * @param port_id 3412 * The port identifier of the Ethernet device. 3413 * @param[inout] queue_id 3414 * On input starting Rx queue index to search from. 3415 * If the queue_id is bigger than maximum queue ID of the port, 3416 * search is started from 0. So that application can keep calling 3417 * this function to handle all pending events with a simple increment 3418 * of queue_id on the next call. 3419 * On output if return value is 1, Rx queue index with the event pending. 3420 * @param[out] avail_thresh 3421 * Location for available descriptors threshold of the found Rx queue. 3422 * 3423 * @return 3424 * - 1 if an Rx queue with pending event is found. 3425 * - 0 if no Rx queue with pending event is found. 3426 * - (-ENODEV) if @p port_id is invalid. 3427 * - (-EINVAL) if bad parameter (e.g. @p queue_id is NULL). 3428 * - (-ENOTSUP) if operation is not supported. 3429 * - (-EIO) if device is removed. 3430 */ 3431 __rte_experimental 3432 int rte_eth_rx_avail_thresh_query(uint16_t port_id, uint16_t *queue_id, 3433 uint8_t *avail_thresh); 3434 3435 typedef void (*buffer_tx_error_fn)(struct rte_mbuf **unsent, uint16_t count, 3436 void *userdata); 3437 3438 /** 3439 * Structure used to buffer packets for future Tx 3440 * Used by APIs rte_eth_tx_buffer and rte_eth_tx_buffer_flush 3441 */ 3442 struct rte_eth_dev_tx_buffer { 3443 buffer_tx_error_fn error_callback; 3444 void *error_userdata; 3445 uint16_t size; /**< Size of buffer for buffered Tx */ 3446 uint16_t length; /**< Number of packets in the array */ 3447 /** Pending packets to be sent on explicit flush or when full */ 3448 struct rte_mbuf *pkts[]; 3449 }; 3450 3451 /** 3452 * Calculate the size of the Tx buffer. 3453 * 3454 * @param sz 3455 * Number of stored packets. 3456 */ 3457 #define RTE_ETH_TX_BUFFER_SIZE(sz) \ 3458 (sizeof(struct rte_eth_dev_tx_buffer) + (sz) * sizeof(struct rte_mbuf *)) 3459 3460 /** 3461 * Initialize default values for buffered transmitting 3462 * 3463 * @param buffer 3464 * Tx buffer to be initialized. 3465 * @param size 3466 * Buffer size 3467 * @return 3468 * 0 if no error 3469 */ 3470 int 3471 rte_eth_tx_buffer_init(struct rte_eth_dev_tx_buffer *buffer, uint16_t size); 3472 3473 /** 3474 * Configure a callback for buffered packets which cannot be sent 3475 * 3476 * Register a specific callback to be called when an attempt is made to send 3477 * all packets buffered on an Ethernet port, but not all packets can 3478 * successfully be sent. The callback registered here will be called only 3479 * from calls to rte_eth_tx_buffer() and rte_eth_tx_buffer_flush() APIs. 3480 * The default callback configured for each queue by default just frees the 3481 * packets back to the calling mempool. If additional behaviour is required, 3482 * for example, to count dropped packets, or to retry transmission of packets 3483 * which cannot be sent, this function should be used to register a suitable 3484 * callback function to implement the desired behaviour. 3485 * The example callback "rte_eth_count_unsent_packet_callback()" is also 3486 * provided as reference. 3487 * 3488 * @param buffer 3489 * The port identifier of the Ethernet device. 3490 * @param callback 3491 * The function to be used as the callback. 3492 * @param userdata 3493 * Arbitrary parameter to be passed to the callback function 3494 * @return 3495 * 0 on success, or -EINVAL if bad parameter 3496 */ 3497 int 3498 rte_eth_tx_buffer_set_err_callback(struct rte_eth_dev_tx_buffer *buffer, 3499 buffer_tx_error_fn callback, void *userdata); 3500 3501 /** 3502 * Callback function for silently dropping unsent buffered packets. 3503 * 3504 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to 3505 * adjust the default behavior when buffered packets cannot be sent. This 3506 * function drops any unsent packets silently and is used by Tx buffered 3507 * operations as default behavior. 3508 * 3509 * NOTE: this function should not be called directly, instead it should be used 3510 * as a callback for packet buffering. 3511 * 3512 * NOTE: when configuring this function as a callback with 3513 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter 3514 * should point to an uint64_t value. 3515 * 3516 * @param pkts 3517 * The previously buffered packets which could not be sent 3518 * @param unsent 3519 * The number of unsent packets in the pkts array 3520 * @param userdata 3521 * Not used 3522 */ 3523 void 3524 rte_eth_tx_buffer_drop_callback(struct rte_mbuf **pkts, uint16_t unsent, 3525 void *userdata); 3526 3527 /** 3528 * Callback function for tracking unsent buffered packets. 3529 * 3530 * This function can be passed to rte_eth_tx_buffer_set_err_callback() to 3531 * adjust the default behavior when buffered packets cannot be sent. This 3532 * function drops any unsent packets, but also updates a user-supplied counter 3533 * to track the overall number of packets dropped. The counter should be an 3534 * uint64_t variable. 3535 * 3536 * NOTE: this function should not be called directly, instead it should be used 3537 * as a callback for packet buffering. 3538 * 3539 * NOTE: when configuring this function as a callback with 3540 * rte_eth_tx_buffer_set_err_callback(), the final, userdata parameter 3541 * should point to an uint64_t value. 3542 * 3543 * @param pkts 3544 * The previously buffered packets which could not be sent 3545 * @param unsent 3546 * The number of unsent packets in the pkts array 3547 * @param userdata 3548 * Pointer to an uint64_t value, which will be incremented by unsent 3549 */ 3550 void 3551 rte_eth_tx_buffer_count_callback(struct rte_mbuf **pkts, uint16_t unsent, 3552 void *userdata); 3553 3554 /** 3555 * Request the driver to free mbufs currently cached by the driver. The 3556 * driver will only free the mbuf if it is no longer in use. It is the 3557 * application's responsibility to ensure rte_eth_tx_buffer_flush(..) is 3558 * called if needed. 3559 * 3560 * @param port_id 3561 * The port identifier of the Ethernet device. 3562 * @param queue_id 3563 * The index of the transmit queue through which output packets must be 3564 * sent. 3565 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 3566 * to rte_eth_dev_configure(). 3567 * @param free_cnt 3568 * Maximum number of packets to free. Use 0 to indicate all possible packets 3569 * should be freed. Note that a packet may be using multiple mbufs. 3570 * @return 3571 * Failure: < 0 3572 * -ENODEV: Invalid interface 3573 * -EIO: device is removed 3574 * -ENOTSUP: Driver does not support function 3575 * Success: >= 0 3576 * 0-n: Number of packets freed. More packets may still remain in ring that 3577 * are in use. 3578 */ 3579 int 3580 rte_eth_tx_done_cleanup(uint16_t port_id, uint16_t queue_id, uint32_t free_cnt); 3581 3582 /** 3583 * Subtypes for IPsec offload event(@ref RTE_ETH_EVENT_IPSEC) raised by 3584 * eth device. 3585 */ 3586 enum rte_eth_event_ipsec_subtype { 3587 /** Unknown event type */ 3588 RTE_ETH_EVENT_IPSEC_UNKNOWN = 0, 3589 /** Sequence number overflow */ 3590 RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW, 3591 /** Soft time expiry of SA */ 3592 RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY, 3593 /** Soft byte expiry of SA */ 3594 RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY, 3595 /** Max value of this enum */ 3596 RTE_ETH_EVENT_IPSEC_MAX 3597 }; 3598 3599 /** 3600 * Descriptor for @ref RTE_ETH_EVENT_IPSEC event. Used by eth dev to send extra 3601 * information of the IPsec offload event. 3602 */ 3603 struct rte_eth_event_ipsec_desc { 3604 /** Type of RTE_ETH_EVENT_IPSEC_* event */ 3605 enum rte_eth_event_ipsec_subtype subtype; 3606 /** 3607 * Event specific metadata. 3608 * 3609 * For the following events, *userdata* registered 3610 * with the *rte_security_session* would be returned 3611 * as metadata, 3612 * 3613 * - @ref RTE_ETH_EVENT_IPSEC_ESN_OVERFLOW 3614 * - @ref RTE_ETH_EVENT_IPSEC_SA_TIME_EXPIRY 3615 * - @ref RTE_ETH_EVENT_IPSEC_SA_BYTE_EXPIRY 3616 * 3617 * @see struct rte_security_session_conf 3618 * 3619 */ 3620 uint64_t metadata; 3621 }; 3622 3623 /** 3624 * The eth device event type for interrupt, and maybe others in the future. 3625 */ 3626 enum rte_eth_event_type { 3627 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */ 3628 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */ 3629 /** queue state event (enabled/disabled) */ 3630 RTE_ETH_EVENT_QUEUE_STATE, 3631 /** reset interrupt event, sent to VF on PF reset */ 3632 RTE_ETH_EVENT_INTR_RESET, 3633 RTE_ETH_EVENT_VF_MBOX, /**< message from the VF received by PF */ 3634 RTE_ETH_EVENT_MACSEC, /**< MACsec offload related event */ 3635 RTE_ETH_EVENT_INTR_RMV, /**< device removal event */ 3636 RTE_ETH_EVENT_NEW, /**< port is probed */ 3637 RTE_ETH_EVENT_DESTROY, /**< port is released */ 3638 RTE_ETH_EVENT_IPSEC, /**< IPsec offload related event */ 3639 RTE_ETH_EVENT_FLOW_AGED,/**< New aged-out flows is detected */ 3640 /** 3641 * Number of available Rx descriptors is smaller than the threshold. 3642 * @see rte_eth_rx_avail_thresh_set() 3643 */ 3644 RTE_ETH_EVENT_RX_AVAIL_THRESH, 3645 RTE_ETH_EVENT_MAX /**< max value of this enum */ 3646 }; 3647 3648 /** User application callback to be registered for interrupts. */ 3649 typedef int (*rte_eth_dev_cb_fn)(uint16_t port_id, 3650 enum rte_eth_event_type event, void *cb_arg, void *ret_param); 3651 3652 /** 3653 * Register a callback function for port event. 3654 * 3655 * @param port_id 3656 * Port ID. 3657 * RTE_ETH_ALL means register the event for all port ids. 3658 * @param event 3659 * Event interested. 3660 * @param cb_fn 3661 * User supplied callback function to be called. 3662 * @param cb_arg 3663 * Pointer to the parameters for the registered callback. 3664 * 3665 * @return 3666 * - On success, zero. 3667 * - On failure, a negative value. 3668 */ 3669 int rte_eth_dev_callback_register(uint16_t port_id, 3670 enum rte_eth_event_type event, 3671 rte_eth_dev_cb_fn cb_fn, void *cb_arg); 3672 3673 /** 3674 * Unregister a callback function for port event. 3675 * 3676 * @param port_id 3677 * Port ID. 3678 * RTE_ETH_ALL means unregister the event for all port ids. 3679 * @param event 3680 * Event interested. 3681 * @param cb_fn 3682 * User supplied callback function to be called. 3683 * @param cb_arg 3684 * Pointer to the parameters for the registered callback. -1 means to 3685 * remove all for the same callback address and same event. 3686 * 3687 * @return 3688 * - On success, zero. 3689 * - On failure, a negative value. 3690 */ 3691 int rte_eth_dev_callback_unregister(uint16_t port_id, 3692 enum rte_eth_event_type event, 3693 rte_eth_dev_cb_fn cb_fn, void *cb_arg); 3694 3695 /** 3696 * When there is no Rx packet coming in Rx Queue for a long time, we can 3697 * sleep lcore related to Rx Queue for power saving, and enable Rx interrupt 3698 * to be triggered when Rx packet arrives. 3699 * 3700 * The rte_eth_dev_rx_intr_enable() function enables Rx queue 3701 * interrupt on specific Rx queue of a port. 3702 * 3703 * @param port_id 3704 * The port identifier of the Ethernet device. 3705 * @param queue_id 3706 * The index of the receive queue from which to retrieve input packets. 3707 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3708 * to rte_eth_dev_configure(). 3709 * @return 3710 * - (0) if successful. 3711 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3712 * that operation. 3713 * - (-ENODEV) if *port_id* invalid. 3714 * - (-EIO) if device is removed. 3715 */ 3716 int rte_eth_dev_rx_intr_enable(uint16_t port_id, uint16_t queue_id); 3717 3718 /** 3719 * When lcore wakes up from Rx interrupt indicating packet coming, disable Rx 3720 * interrupt and returns to polling mode. 3721 * 3722 * The rte_eth_dev_rx_intr_disable() function disables Rx queue 3723 * interrupt on specific Rx queue of a port. 3724 * 3725 * @param port_id 3726 * The port identifier of the Ethernet device. 3727 * @param queue_id 3728 * The index of the receive queue from which to retrieve input packets. 3729 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3730 * to rte_eth_dev_configure(). 3731 * @return 3732 * - (0) if successful. 3733 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3734 * that operation. 3735 * - (-ENODEV) if *port_id* invalid. 3736 * - (-EIO) if device is removed. 3737 */ 3738 int rte_eth_dev_rx_intr_disable(uint16_t port_id, uint16_t queue_id); 3739 3740 /** 3741 * Rx Interrupt control per port. 3742 * 3743 * @param port_id 3744 * The port identifier of the Ethernet device. 3745 * @param epfd 3746 * Epoll instance fd which the intr vector associated to. 3747 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance. 3748 * @param op 3749 * The operation be performed for the vector. 3750 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}. 3751 * @param data 3752 * User raw data. 3753 * @return 3754 * - On success, zero. 3755 * - On failure, a negative value. 3756 */ 3757 int rte_eth_dev_rx_intr_ctl(uint16_t port_id, int epfd, int op, void *data); 3758 3759 /** 3760 * Rx Interrupt control per queue. 3761 * 3762 * @param port_id 3763 * The port identifier of the Ethernet device. 3764 * @param queue_id 3765 * The index of the receive queue from which to retrieve input packets. 3766 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3767 * to rte_eth_dev_configure(). 3768 * @param epfd 3769 * Epoll instance fd which the intr vector associated to. 3770 * Using RTE_EPOLL_PER_THREAD allows to use per thread epoll instance. 3771 * @param op 3772 * The operation be performed for the vector. 3773 * Operation type of {RTE_INTR_EVENT_ADD, RTE_INTR_EVENT_DEL}. 3774 * @param data 3775 * User raw data. 3776 * @return 3777 * - On success, zero. 3778 * - On failure, a negative value. 3779 */ 3780 int rte_eth_dev_rx_intr_ctl_q(uint16_t port_id, uint16_t queue_id, 3781 int epfd, int op, void *data); 3782 3783 /** 3784 * Get interrupt fd per Rx queue. 3785 * 3786 * @param port_id 3787 * The port identifier of the Ethernet device. 3788 * @param queue_id 3789 * The index of the receive queue from which to retrieve input packets. 3790 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 3791 * to rte_eth_dev_configure(). 3792 * @return 3793 * - (>=0) the interrupt fd associated to the requested Rx queue if 3794 * successful. 3795 * - (-1) on error. 3796 */ 3797 int 3798 rte_eth_dev_rx_intr_ctl_q_get_fd(uint16_t port_id, uint16_t queue_id); 3799 3800 /** 3801 * Turn on the LED on the Ethernet device. 3802 * This function turns on the LED on the Ethernet device. 3803 * 3804 * @param port_id 3805 * The port identifier of the Ethernet device. 3806 * @return 3807 * - (0) if successful. 3808 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3809 * that operation. 3810 * - (-ENODEV) if *port_id* invalid. 3811 * - (-EIO) if device is removed. 3812 */ 3813 int rte_eth_led_on(uint16_t port_id); 3814 3815 /** 3816 * Turn off the LED on the Ethernet device. 3817 * This function turns off the LED on the Ethernet device. 3818 * 3819 * @param port_id 3820 * The port identifier of the Ethernet device. 3821 * @return 3822 * - (0) if successful. 3823 * - (-ENOTSUP) if underlying hardware OR driver doesn't support 3824 * that operation. 3825 * - (-ENODEV) if *port_id* invalid. 3826 * - (-EIO) if device is removed. 3827 */ 3828 int rte_eth_led_off(uint16_t port_id); 3829 3830 /** 3831 * @warning 3832 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 3833 * 3834 * Get Forward Error Correction(FEC) capability. 3835 * 3836 * @param port_id 3837 * The port identifier of the Ethernet device. 3838 * @param speed_fec_capa 3839 * speed_fec_capa is out only with per-speed capabilities. 3840 * If set to NULL, the function returns the required number 3841 * of required array entries. 3842 * @param num 3843 * a number of elements in an speed_fec_capa array. 3844 * 3845 * @return 3846 * - A non-negative value lower or equal to num: success. The return value 3847 * is the number of entries filled in the fec capa array. 3848 * - A non-negative value higher than num: error, the given fec capa array 3849 * is too small. The return value corresponds to the num that should 3850 * be given to succeed. The entries in fec capa array are not valid and 3851 * shall not be used by the caller. 3852 * - (-ENOTSUP) if underlying hardware OR driver doesn't support. 3853 * that operation. 3854 * - (-EIO) if device is removed. 3855 * - (-ENODEV) if *port_id* invalid. 3856 * - (-EINVAL) if *num* or *speed_fec_capa* invalid 3857 */ 3858 __rte_experimental 3859 int rte_eth_fec_get_capability(uint16_t port_id, 3860 struct rte_eth_fec_capa *speed_fec_capa, 3861 unsigned int num); 3862 3863 /** 3864 * @warning 3865 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 3866 * 3867 * Get current Forward Error Correction(FEC) mode. 3868 * If link is down and AUTO is enabled, AUTO is returned, otherwise, 3869 * configured FEC mode is returned. 3870 * If link is up, current FEC mode is returned. 3871 * 3872 * @param port_id 3873 * The port identifier of the Ethernet device. 3874 * @param fec_capa 3875 * A bitmask of enabled FEC modes. If AUTO bit is set, other 3876 * bits specify FEC modes which may be negotiated. If AUTO 3877 * bit is clear, specify FEC modes to be used (only one valid 3878 * mode per speed may be set). 3879 * @return 3880 * - (0) if successful. 3881 * - (-ENOTSUP) if underlying hardware OR driver doesn't support. 3882 * that operation. 3883 * - (-EIO) if device is removed. 3884 * - (-ENODEV) if *port_id* invalid. 3885 */ 3886 __rte_experimental 3887 int rte_eth_fec_get(uint16_t port_id, uint32_t *fec_capa); 3888 3889 /** 3890 * @warning 3891 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 3892 * 3893 * Set Forward Error Correction(FEC) mode. 3894 * 3895 * @param port_id 3896 * The port identifier of the Ethernet device. 3897 * @param fec_capa 3898 * A bitmask of allowed FEC modes. If AUTO bit is set, other 3899 * bits specify FEC modes which may be negotiated. If AUTO 3900 * bit is clear, specify FEC modes to be used (only one valid 3901 * mode per speed may be set). 3902 * @return 3903 * - (0) if successful. 3904 * - (-EINVAL) if the FEC mode is not valid. 3905 * - (-ENOTSUP) if underlying hardware OR driver doesn't support. 3906 * - (-EIO) if device is removed. 3907 * - (-ENODEV) if *port_id* invalid. 3908 */ 3909 __rte_experimental 3910 int rte_eth_fec_set(uint16_t port_id, uint32_t fec_capa); 3911 3912 /** 3913 * Get current status of the Ethernet link flow control for Ethernet device 3914 * 3915 * @param port_id 3916 * The port identifier of the Ethernet device. 3917 * @param fc_conf 3918 * The pointer to the structure where to store the flow control parameters. 3919 * @return 3920 * - (0) if successful. 3921 * - (-ENOTSUP) if hardware doesn't support flow control. 3922 * - (-ENODEV) if *port_id* invalid. 3923 * - (-EIO) if device is removed. 3924 * - (-EINVAL) if bad parameter. 3925 */ 3926 int rte_eth_dev_flow_ctrl_get(uint16_t port_id, 3927 struct rte_eth_fc_conf *fc_conf); 3928 3929 /** 3930 * Configure the Ethernet link flow control for Ethernet device 3931 * 3932 * @param port_id 3933 * The port identifier of the Ethernet device. 3934 * @param fc_conf 3935 * The pointer to the structure of the flow control parameters. 3936 * @return 3937 * - (0) if successful. 3938 * - (-ENOTSUP) if hardware doesn't support flow control mode. 3939 * - (-ENODEV) if *port_id* invalid. 3940 * - (-EINVAL) if bad parameter 3941 * - (-EIO) if flow control setup failure or device is removed. 3942 */ 3943 int rte_eth_dev_flow_ctrl_set(uint16_t port_id, 3944 struct rte_eth_fc_conf *fc_conf); 3945 3946 /** 3947 * Configure the Ethernet priority flow control under DCB environment 3948 * for Ethernet device. 3949 * 3950 * @param port_id 3951 * The port identifier of the Ethernet device. 3952 * @param pfc_conf 3953 * The pointer to the structure of the priority flow control parameters. 3954 * @return 3955 * - (0) if successful. 3956 * - (-ENOTSUP) if hardware doesn't support priority flow control mode. 3957 * - (-ENODEV) if *port_id* invalid. 3958 * - (-EINVAL) if bad parameter 3959 * - (-EIO) if flow control setup failure or device is removed. 3960 */ 3961 int rte_eth_dev_priority_flow_ctrl_set(uint16_t port_id, 3962 struct rte_eth_pfc_conf *pfc_conf); 3963 3964 /** 3965 * Add a MAC address to the set used for filtering incoming packets. 3966 * 3967 * @param port_id 3968 * The port identifier of the Ethernet device. 3969 * @param mac_addr 3970 * The MAC address to add. 3971 * @param pool 3972 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is 3973 * not enabled, this should be set to 0. 3974 * @return 3975 * - (0) if successfully added or *mac_addr* was already added. 3976 * - (-ENOTSUP) if hardware doesn't support this feature. 3977 * - (-ENODEV) if *port* is invalid. 3978 * - (-EIO) if device is removed. 3979 * - (-ENOSPC) if no more MAC addresses can be added. 3980 * - (-EINVAL) if MAC address is invalid. 3981 */ 3982 int rte_eth_dev_mac_addr_add(uint16_t port_id, struct rte_ether_addr *mac_addr, 3983 uint32_t pool); 3984 3985 /** 3986 * @warning 3987 * @b EXPERIMENTAL: this API may change without prior notice. 3988 * 3989 * Retrieve the information for queue based PFC. 3990 * 3991 * @param port_id 3992 * The port identifier of the Ethernet device. 3993 * @param pfc_queue_info 3994 * A pointer to a structure of type *rte_eth_pfc_queue_info* to be filled with 3995 * the information about queue based PFC. 3996 * @return 3997 * - (0) if successful. 3998 * - (-ENOTSUP) if support for priority_flow_ctrl_queue_info_get does not exist. 3999 * - (-ENODEV) if *port_id* invalid. 4000 * - (-EINVAL) if bad parameter. 4001 */ 4002 __rte_experimental 4003 int rte_eth_dev_priority_flow_ctrl_queue_info_get(uint16_t port_id, 4004 struct rte_eth_pfc_queue_info *pfc_queue_info); 4005 4006 /** 4007 * @warning 4008 * @b EXPERIMENTAL: this API may change without prior notice. 4009 * 4010 * Configure the queue based priority flow control for a given queue 4011 * for Ethernet device. 4012 * 4013 * @note When an ethdev port switches to queue based PFC mode, the 4014 * unconfigured queues shall be configured by the driver with 4015 * default values such as lower priority value for TC etc. 4016 * 4017 * @param port_id 4018 * The port identifier of the Ethernet device. 4019 * @param pfc_queue_conf 4020 * The pointer to the structure of the priority flow control parameters 4021 * for the queue. 4022 * @return 4023 * - (0) if successful. 4024 * - (-ENOTSUP) if hardware doesn't support queue based PFC mode. 4025 * - (-ENODEV) if *port_id* invalid. 4026 * - (-EINVAL) if bad parameter 4027 * - (-EIO) if flow control setup queue failure 4028 */ 4029 __rte_experimental 4030 int rte_eth_dev_priority_flow_ctrl_queue_configure(uint16_t port_id, 4031 struct rte_eth_pfc_queue_conf *pfc_queue_conf); 4032 4033 /** 4034 * Remove a MAC address from the internal array of addresses. 4035 * 4036 * @param port_id 4037 * The port identifier of the Ethernet device. 4038 * @param mac_addr 4039 * MAC address to remove. 4040 * @return 4041 * - (0) if successful, or *mac_addr* didn't exist. 4042 * - (-ENOTSUP) if hardware doesn't support. 4043 * - (-ENODEV) if *port* invalid. 4044 * - (-EADDRINUSE) if attempting to remove the default MAC address. 4045 * - (-EINVAL) if MAC address is invalid. 4046 */ 4047 int rte_eth_dev_mac_addr_remove(uint16_t port_id, 4048 struct rte_ether_addr *mac_addr); 4049 4050 /** 4051 * Set the default MAC address. 4052 * 4053 * @param port_id 4054 * The port identifier of the Ethernet device. 4055 * @param mac_addr 4056 * New default MAC address. 4057 * @return 4058 * - (0) if successful, or *mac_addr* didn't exist. 4059 * - (-ENOTSUP) if hardware doesn't support. 4060 * - (-ENODEV) if *port* invalid. 4061 * - (-EINVAL) if MAC address is invalid. 4062 */ 4063 int rte_eth_dev_default_mac_addr_set(uint16_t port_id, 4064 struct rte_ether_addr *mac_addr); 4065 4066 /** 4067 * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device. 4068 * 4069 * @param port_id 4070 * The port identifier of the Ethernet device. 4071 * @param reta_conf 4072 * RETA to update. 4073 * @param reta_size 4074 * Redirection table size. The table size can be queried by 4075 * rte_eth_dev_info_get(). 4076 * @return 4077 * - (0) if successful. 4078 * - (-ENODEV) if *port_id* is invalid. 4079 * - (-ENOTSUP) if hardware doesn't support. 4080 * - (-EINVAL) if bad parameter. 4081 * - (-EIO) if device is removed. 4082 */ 4083 int rte_eth_dev_rss_reta_update(uint16_t port_id, 4084 struct rte_eth_rss_reta_entry64 *reta_conf, 4085 uint16_t reta_size); 4086 4087 /** 4088 * Query Redirection Table(RETA) of Receive Side Scaling of Ethernet device. 4089 * 4090 * @param port_id 4091 * The port identifier of the Ethernet device. 4092 * @param reta_conf 4093 * RETA to query. For each requested reta entry, corresponding bit 4094 * in mask must be set. 4095 * @param reta_size 4096 * Redirection table size. The table size can be queried by 4097 * rte_eth_dev_info_get(). 4098 * @return 4099 * - (0) if successful. 4100 * - (-ENODEV) if *port_id* is invalid. 4101 * - (-ENOTSUP) if hardware doesn't support. 4102 * - (-EINVAL) if bad parameter. 4103 * - (-EIO) if device is removed. 4104 */ 4105 int rte_eth_dev_rss_reta_query(uint16_t port_id, 4106 struct rte_eth_rss_reta_entry64 *reta_conf, 4107 uint16_t reta_size); 4108 4109 /** 4110 * Updates unicast hash table for receiving packet with the given destination 4111 * MAC address, and the packet is routed to all VFs for which the Rx mode is 4112 * accept packets that match the unicast hash table. 4113 * 4114 * @param port_id 4115 * The port identifier of the Ethernet device. 4116 * @param addr 4117 * Unicast MAC address. 4118 * @param on 4119 * 1 - Set an unicast hash bit for receiving packets with the MAC address. 4120 * 0 - Clear an unicast hash bit. 4121 * @return 4122 * - (0) if successful. 4123 * - (-ENOTSUP) if hardware doesn't support. 4124 * - (-ENODEV) if *port_id* invalid. 4125 * - (-EIO) if device is removed. 4126 * - (-EINVAL) if bad parameter. 4127 */ 4128 int rte_eth_dev_uc_hash_table_set(uint16_t port_id, struct rte_ether_addr *addr, 4129 uint8_t on); 4130 4131 /** 4132 * Updates all unicast hash bitmaps for receiving packet with any Unicast 4133 * Ethernet MAC addresses,the packet is routed to all VFs for which the Rx 4134 * mode is accept packets that match the unicast hash table. 4135 * 4136 * @param port_id 4137 * The port identifier of the Ethernet device. 4138 * @param on 4139 * 1 - Set all unicast hash bitmaps for receiving all the Ethernet 4140 * MAC addresses 4141 * 0 - Clear all unicast hash bitmaps 4142 * @return 4143 * - (0) if successful. 4144 * - (-ENOTSUP) if hardware doesn't support. 4145 * - (-ENODEV) if *port_id* invalid. 4146 * - (-EIO) if device is removed. 4147 * - (-EINVAL) if bad parameter. 4148 */ 4149 int rte_eth_dev_uc_all_hash_table_set(uint16_t port_id, uint8_t on); 4150 4151 /** 4152 * Set the rate limitation for a queue on an Ethernet device. 4153 * 4154 * @param port_id 4155 * The port identifier of the Ethernet device. 4156 * @param queue_idx 4157 * The queue ID. 4158 * @param tx_rate 4159 * The Tx rate in Mbps. Allocated from the total port link speed. 4160 * @return 4161 * - (0) if successful. 4162 * - (-ENOTSUP) if hardware doesn't support this feature. 4163 * - (-ENODEV) if *port_id* invalid. 4164 * - (-EIO) if device is removed. 4165 * - (-EINVAL) if bad parameter. 4166 */ 4167 int rte_eth_set_queue_rate_limit(uint16_t port_id, uint16_t queue_idx, 4168 uint16_t tx_rate); 4169 4170 /** 4171 * Configuration of Receive Side Scaling hash computation of Ethernet device. 4172 * 4173 * @param port_id 4174 * The port identifier of the Ethernet device. 4175 * @param rss_conf 4176 * The new configuration to use for RSS hash computation on the port. 4177 * @return 4178 * - (0) if successful. 4179 * - (-ENODEV) if port identifier is invalid. 4180 * - (-EIO) if device is removed. 4181 * - (-ENOTSUP) if hardware doesn't support. 4182 * - (-EINVAL) if bad parameter. 4183 */ 4184 int rte_eth_dev_rss_hash_update(uint16_t port_id, 4185 struct rte_eth_rss_conf *rss_conf); 4186 4187 /** 4188 * Retrieve current configuration of Receive Side Scaling hash computation 4189 * of Ethernet device. 4190 * 4191 * @param port_id 4192 * The port identifier of the Ethernet device. 4193 * @param rss_conf 4194 * Where to store the current RSS hash configuration of the Ethernet device. 4195 * @return 4196 * - (0) if successful. 4197 * - (-ENODEV) if port identifier is invalid. 4198 * - (-EIO) if device is removed. 4199 * - (-ENOTSUP) if hardware doesn't support RSS. 4200 * - (-EINVAL) if bad parameter. 4201 */ 4202 int 4203 rte_eth_dev_rss_hash_conf_get(uint16_t port_id, 4204 struct rte_eth_rss_conf *rss_conf); 4205 4206 /** 4207 * Add UDP tunneling port for a type of tunnel. 4208 * 4209 * Some NICs may require such configuration to properly parse a tunnel 4210 * with any standard or custom UDP port. 4211 * The packets with this UDP port will be parsed for this type of tunnel. 4212 * The device parser will also check the rest of the tunnel headers 4213 * before classifying the packet. 4214 * 4215 * With some devices, this API will affect packet classification, i.e.: 4216 * - mbuf.packet_type reported on Rx 4217 * - rte_flow rules with tunnel items 4218 * 4219 * @param port_id 4220 * The port identifier of the Ethernet device. 4221 * @param tunnel_udp 4222 * UDP tunneling configuration. 4223 * 4224 * @return 4225 * - (0) if successful. 4226 * - (-ENODEV) if port identifier is invalid. 4227 * - (-EIO) if device is removed. 4228 * - (-ENOTSUP) if hardware doesn't support tunnel type. 4229 */ 4230 int 4231 rte_eth_dev_udp_tunnel_port_add(uint16_t port_id, 4232 struct rte_eth_udp_tunnel *tunnel_udp); 4233 4234 /** 4235 * Delete UDP tunneling port for a type of tunnel. 4236 * 4237 * The packets with this UDP port will not be classified as this type of tunnel 4238 * anymore if the device use such mapping for tunnel packet classification. 4239 * 4240 * @see rte_eth_dev_udp_tunnel_port_add 4241 * 4242 * @param port_id 4243 * The port identifier of the Ethernet device. 4244 * @param tunnel_udp 4245 * UDP tunneling configuration. 4246 * 4247 * @return 4248 * - (0) if successful. 4249 * - (-ENODEV) if port identifier is invalid. 4250 * - (-EIO) if device is removed. 4251 * - (-ENOTSUP) if hardware doesn't support tunnel type. 4252 */ 4253 int 4254 rte_eth_dev_udp_tunnel_port_delete(uint16_t port_id, 4255 struct rte_eth_udp_tunnel *tunnel_udp); 4256 4257 /** 4258 * Get DCB information on an Ethernet device. 4259 * 4260 * @param port_id 4261 * The port identifier of the Ethernet device. 4262 * @param dcb_info 4263 * DCB information. 4264 * @return 4265 * - (0) if successful. 4266 * - (-ENODEV) if port identifier is invalid. 4267 * - (-EIO) if device is removed. 4268 * - (-ENOTSUP) if hardware doesn't support. 4269 * - (-EINVAL) if bad parameter. 4270 */ 4271 int rte_eth_dev_get_dcb_info(uint16_t port_id, 4272 struct rte_eth_dcb_info *dcb_info); 4273 4274 struct rte_eth_rxtx_callback; 4275 4276 /** 4277 * Add a callback to be called on packet Rx on a given port and queue. 4278 * 4279 * This API configures a function to be called for each burst of 4280 * packets received on a given NIC port queue. The return value is a pointer 4281 * that can be used to later remove the callback using 4282 * rte_eth_remove_rx_callback(). 4283 * 4284 * Multiple functions are called in the order that they are added. 4285 * 4286 * @param port_id 4287 * The port identifier of the Ethernet device. 4288 * @param queue_id 4289 * The queue on the Ethernet device on which the callback is to be added. 4290 * @param fn 4291 * The callback function 4292 * @param user_param 4293 * A generic pointer parameter which will be passed to each invocation of the 4294 * callback function on this port and queue. Inter-thread synchronization 4295 * of any user data changes is the responsibility of the user. 4296 * 4297 * @return 4298 * NULL on error. 4299 * On success, a pointer value which can later be used to remove the callback. 4300 */ 4301 const struct rte_eth_rxtx_callback * 4302 rte_eth_add_rx_callback(uint16_t port_id, uint16_t queue_id, 4303 rte_rx_callback_fn fn, void *user_param); 4304 4305 /** 4306 * Add a callback that must be called first on packet Rx on a given port 4307 * and queue. 4308 * 4309 * This API configures a first function to be called for each burst of 4310 * packets received on a given NIC port queue. The return value is a pointer 4311 * that can be used to later remove the callback using 4312 * rte_eth_remove_rx_callback(). 4313 * 4314 * Multiple functions are called in the order that they are added. 4315 * 4316 * @param port_id 4317 * The port identifier of the Ethernet device. 4318 * @param queue_id 4319 * The queue on the Ethernet device on which the callback is to be added. 4320 * @param fn 4321 * The callback function 4322 * @param user_param 4323 * A generic pointer parameter which will be passed to each invocation of the 4324 * callback function on this port and queue. Inter-thread synchronization 4325 * of any user data changes is the responsibility of the user. 4326 * 4327 * @return 4328 * NULL on error. 4329 * On success, a pointer value which can later be used to remove the callback. 4330 */ 4331 const struct rte_eth_rxtx_callback * 4332 rte_eth_add_first_rx_callback(uint16_t port_id, uint16_t queue_id, 4333 rte_rx_callback_fn fn, void *user_param); 4334 4335 /** 4336 * Add a callback to be called on packet Tx on a given port and queue. 4337 * 4338 * This API configures a function to be called for each burst of 4339 * packets sent on a given NIC port queue. The return value is a pointer 4340 * that can be used to later remove the callback using 4341 * rte_eth_remove_tx_callback(). 4342 * 4343 * Multiple functions are called in the order that they are added. 4344 * 4345 * @param port_id 4346 * The port identifier of the Ethernet device. 4347 * @param queue_id 4348 * The queue on the Ethernet device on which the callback is to be added. 4349 * @param fn 4350 * The callback function 4351 * @param user_param 4352 * A generic pointer parameter which will be passed to each invocation of the 4353 * callback function on this port and queue. Inter-thread synchronization 4354 * of any user data changes is the responsibility of the user. 4355 * 4356 * @return 4357 * NULL on error. 4358 * On success, a pointer value which can later be used to remove the callback. 4359 */ 4360 const struct rte_eth_rxtx_callback * 4361 rte_eth_add_tx_callback(uint16_t port_id, uint16_t queue_id, 4362 rte_tx_callback_fn fn, void *user_param); 4363 4364 /** 4365 * Remove an Rx packet callback from a given port and queue. 4366 * 4367 * This function is used to removed callbacks that were added to a NIC port 4368 * queue using rte_eth_add_rx_callback(). 4369 * 4370 * Note: the callback is removed from the callback list but it isn't freed 4371 * since the it may still be in use. The memory for the callback can be 4372 * subsequently freed back by the application by calling rte_free(): 4373 * 4374 * - Immediately - if the port is stopped, or the user knows that no 4375 * callbacks are in flight e.g. if called from the thread doing Rx/Tx 4376 * on that queue. 4377 * 4378 * - After a short delay - where the delay is sufficient to allow any 4379 * in-flight callbacks to complete. Alternately, the RCU mechanism can be 4380 * used to detect when data plane threads have ceased referencing the 4381 * callback memory. 4382 * 4383 * @param port_id 4384 * The port identifier of the Ethernet device. 4385 * @param queue_id 4386 * The queue on the Ethernet device from which the callback is to be removed. 4387 * @param user_cb 4388 * User supplied callback created via rte_eth_add_rx_callback(). 4389 * 4390 * @return 4391 * - 0: Success. Callback was removed. 4392 * - -ENODEV: If *port_id* is invalid. 4393 * - -ENOTSUP: Callback support is not available. 4394 * - -EINVAL: The queue_id is out of range, or the callback 4395 * is NULL or not found for the port/queue. 4396 */ 4397 int rte_eth_remove_rx_callback(uint16_t port_id, uint16_t queue_id, 4398 const struct rte_eth_rxtx_callback *user_cb); 4399 4400 /** 4401 * Remove a Tx packet callback from a given port and queue. 4402 * 4403 * This function is used to removed callbacks that were added to a NIC port 4404 * queue using rte_eth_add_tx_callback(). 4405 * 4406 * Note: the callback is removed from the callback list but it isn't freed 4407 * since the it may still be in use. The memory for the callback can be 4408 * subsequently freed back by the application by calling rte_free(): 4409 * 4410 * - Immediately - if the port is stopped, or the user knows that no 4411 * callbacks are in flight e.g. if called from the thread doing Rx/Tx 4412 * on that queue. 4413 * 4414 * - After a short delay - where the delay is sufficient to allow any 4415 * in-flight callbacks to complete. Alternately, the RCU mechanism can be 4416 * used to detect when data plane threads have ceased referencing the 4417 * callback memory. 4418 * 4419 * @param port_id 4420 * The port identifier of the Ethernet device. 4421 * @param queue_id 4422 * The queue on the Ethernet device from which the callback is to be removed. 4423 * @param user_cb 4424 * User supplied callback created via rte_eth_add_tx_callback(). 4425 * 4426 * @return 4427 * - 0: Success. Callback was removed. 4428 * - -ENODEV: If *port_id* is invalid. 4429 * - -ENOTSUP: Callback support is not available. 4430 * - -EINVAL: The queue_id is out of range, or the callback 4431 * is NULL or not found for the port/queue. 4432 */ 4433 int rte_eth_remove_tx_callback(uint16_t port_id, uint16_t queue_id, 4434 const struct rte_eth_rxtx_callback *user_cb); 4435 4436 /** 4437 * Retrieve information about given port's Rx queue. 4438 * 4439 * @param port_id 4440 * The port identifier of the Ethernet device. 4441 * @param queue_id 4442 * The Rx queue on the Ethernet device for which information 4443 * will be retrieved. 4444 * @param qinfo 4445 * A pointer to a structure of type *rte_eth_rxq_info_info* to be filled with 4446 * the information of the Ethernet device. 4447 * 4448 * @return 4449 * - 0: Success 4450 * - -ENODEV: If *port_id* is invalid. 4451 * - -ENOTSUP: routine is not supported by the device PMD. 4452 * - -EINVAL: The queue_id is out of range, or the queue 4453 * is hairpin queue. 4454 */ 4455 int rte_eth_rx_queue_info_get(uint16_t port_id, uint16_t queue_id, 4456 struct rte_eth_rxq_info *qinfo); 4457 4458 /** 4459 * Retrieve information about given port's Tx queue. 4460 * 4461 * @param port_id 4462 * The port identifier of the Ethernet device. 4463 * @param queue_id 4464 * The Tx queue on the Ethernet device for which information 4465 * will be retrieved. 4466 * @param qinfo 4467 * A pointer to a structure of type *rte_eth_txq_info_info* to be filled with 4468 * the information of the Ethernet device. 4469 * 4470 * @return 4471 * - 0: Success 4472 * - -ENODEV: If *port_id* is invalid. 4473 * - -ENOTSUP: routine is not supported by the device PMD. 4474 * - -EINVAL: The queue_id is out of range, or the queue 4475 * is hairpin queue. 4476 */ 4477 int rte_eth_tx_queue_info_get(uint16_t port_id, uint16_t queue_id, 4478 struct rte_eth_txq_info *qinfo); 4479 4480 /** 4481 * Retrieve information about the Rx packet burst mode. 4482 * 4483 * @param port_id 4484 * The port identifier of the Ethernet device. 4485 * @param queue_id 4486 * The Rx queue on the Ethernet device for which information 4487 * will be retrieved. 4488 * @param mode 4489 * A pointer to a structure of type *rte_eth_burst_mode* to be filled 4490 * with the information of the packet burst mode. 4491 * 4492 * @return 4493 * - 0: Success 4494 * - -ENODEV: If *port_id* is invalid. 4495 * - -ENOTSUP: routine is not supported by the device PMD. 4496 * - -EINVAL: The queue_id is out of range. 4497 */ 4498 int rte_eth_rx_burst_mode_get(uint16_t port_id, uint16_t queue_id, 4499 struct rte_eth_burst_mode *mode); 4500 4501 /** 4502 * Retrieve information about the Tx packet burst mode. 4503 * 4504 * @param port_id 4505 * The port identifier of the Ethernet device. 4506 * @param queue_id 4507 * The Tx queue on the Ethernet device for which information 4508 * will be retrieved. 4509 * @param mode 4510 * A pointer to a structure of type *rte_eth_burst_mode* to be filled 4511 * with the information of the packet burst mode. 4512 * 4513 * @return 4514 * - 0: Success 4515 * - -ENODEV: If *port_id* is invalid. 4516 * - -ENOTSUP: routine is not supported by the device PMD. 4517 * - -EINVAL: The queue_id is out of range. 4518 */ 4519 int rte_eth_tx_burst_mode_get(uint16_t port_id, uint16_t queue_id, 4520 struct rte_eth_burst_mode *mode); 4521 4522 /** 4523 * @warning 4524 * @b EXPERIMENTAL: this API may change without prior notice. 4525 * 4526 * Retrieve the monitor condition for a given receive queue. 4527 * 4528 * @param port_id 4529 * The port identifier of the Ethernet device. 4530 * @param queue_id 4531 * The Rx queue on the Ethernet device for which information 4532 * will be retrieved. 4533 * @param pmc 4534 * The pointer to power-optimized monitoring condition structure. 4535 * 4536 * @return 4537 * - 0: Success. 4538 * -ENOTSUP: Operation not supported. 4539 * -EINVAL: Invalid parameters. 4540 * -ENODEV: Invalid port ID. 4541 */ 4542 __rte_experimental 4543 int rte_eth_get_monitor_addr(uint16_t port_id, uint16_t queue_id, 4544 struct rte_power_monitor_cond *pmc); 4545 4546 /** 4547 * Retrieve device registers and register attributes (number of registers and 4548 * register size) 4549 * 4550 * @param port_id 4551 * The port identifier of the Ethernet device. 4552 * @param info 4553 * Pointer to rte_dev_reg_info structure to fill in. If info->data is 4554 * NULL the function fills in the width and length fields. If non-NULL 4555 * the registers are put into the buffer pointed at by the data field. 4556 * @return 4557 * - (0) if successful. 4558 * - (-ENOTSUP) if hardware doesn't support. 4559 * - (-EINVAL) if bad parameter. 4560 * - (-ENODEV) if *port_id* invalid. 4561 * - (-EIO) if device is removed. 4562 * - others depends on the specific operations implementation. 4563 */ 4564 int rte_eth_dev_get_reg_info(uint16_t port_id, struct rte_dev_reg_info *info); 4565 4566 /** 4567 * Retrieve size of device EEPROM 4568 * 4569 * @param port_id 4570 * The port identifier of the Ethernet device. 4571 * @return 4572 * - (>=0) EEPROM size if successful. 4573 * - (-ENOTSUP) if hardware doesn't support. 4574 * - (-ENODEV) if *port_id* invalid. 4575 * - (-EIO) if device is removed. 4576 * - others depends on the specific operations implementation. 4577 */ 4578 int rte_eth_dev_get_eeprom_length(uint16_t port_id); 4579 4580 /** 4581 * Retrieve EEPROM and EEPROM attribute 4582 * 4583 * @param port_id 4584 * The port identifier of the Ethernet device. 4585 * @param info 4586 * The template includes buffer for return EEPROM data and 4587 * EEPROM attributes to be filled. 4588 * @return 4589 * - (0) if successful. 4590 * - (-ENOTSUP) if hardware doesn't support. 4591 * - (-EINVAL) if bad parameter. 4592 * - (-ENODEV) if *port_id* invalid. 4593 * - (-EIO) if device is removed. 4594 * - others depends on the specific operations implementation. 4595 */ 4596 int rte_eth_dev_get_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info); 4597 4598 /** 4599 * Program EEPROM with provided data 4600 * 4601 * @param port_id 4602 * The port identifier of the Ethernet device. 4603 * @param info 4604 * The template includes EEPROM data for programming and 4605 * EEPROM attributes to be filled 4606 * @return 4607 * - (0) if successful. 4608 * - (-ENOTSUP) if hardware doesn't support. 4609 * - (-ENODEV) if *port_id* invalid. 4610 * - (-EINVAL) if bad parameter. 4611 * - (-EIO) if device is removed. 4612 * - others depends on the specific operations implementation. 4613 */ 4614 int rte_eth_dev_set_eeprom(uint16_t port_id, struct rte_dev_eeprom_info *info); 4615 4616 /** 4617 * @warning 4618 * @b EXPERIMENTAL: this API may change without prior notice. 4619 * 4620 * Retrieve the type and size of plugin module EEPROM 4621 * 4622 * @param port_id 4623 * The port identifier of the Ethernet device. 4624 * @param modinfo 4625 * The type and size of plugin module EEPROM. 4626 * @return 4627 * - (0) if successful. 4628 * - (-ENOTSUP) if hardware doesn't support. 4629 * - (-ENODEV) if *port_id* invalid. 4630 * - (-EINVAL) if bad parameter. 4631 * - (-EIO) if device is removed. 4632 * - others depends on the specific operations implementation. 4633 */ 4634 __rte_experimental 4635 int 4636 rte_eth_dev_get_module_info(uint16_t port_id, 4637 struct rte_eth_dev_module_info *modinfo); 4638 4639 /** 4640 * @warning 4641 * @b EXPERIMENTAL: this API may change without prior notice. 4642 * 4643 * Retrieve the data of plugin module EEPROM 4644 * 4645 * @param port_id 4646 * The port identifier of the Ethernet device. 4647 * @param info 4648 * The template includes the plugin module EEPROM attributes, and the 4649 * buffer for return plugin module EEPROM data. 4650 * @return 4651 * - (0) if successful. 4652 * - (-ENOTSUP) if hardware doesn't support. 4653 * - (-EINVAL) if bad parameter. 4654 * - (-ENODEV) if *port_id* invalid. 4655 * - (-EIO) if device is removed. 4656 * - others depends on the specific operations implementation. 4657 */ 4658 __rte_experimental 4659 int 4660 rte_eth_dev_get_module_eeprom(uint16_t port_id, 4661 struct rte_dev_eeprom_info *info); 4662 4663 /** 4664 * Set the list of multicast addresses to filter on an Ethernet device. 4665 * 4666 * @param port_id 4667 * The port identifier of the Ethernet device. 4668 * @param mc_addr_set 4669 * The array of multicast addresses to set. Equal to NULL when the function 4670 * is invoked to flush the set of filtered addresses. 4671 * @param nb_mc_addr 4672 * The number of multicast addresses in the *mc_addr_set* array. Equal to 0 4673 * when the function is invoked to flush the set of filtered addresses. 4674 * @return 4675 * - (0) if successful. 4676 * - (-ENODEV) if *port_id* invalid. 4677 * - (-EIO) if device is removed. 4678 * - (-ENOTSUP) if PMD of *port_id* doesn't support multicast filtering. 4679 * - (-ENOSPC) if *port_id* has not enough multicast filtering resources. 4680 * - (-EINVAL) if bad parameter. 4681 */ 4682 int rte_eth_dev_set_mc_addr_list(uint16_t port_id, 4683 struct rte_ether_addr *mc_addr_set, 4684 uint32_t nb_mc_addr); 4685 4686 /** 4687 * Enable IEEE1588/802.1AS timestamping for an Ethernet device. 4688 * 4689 * @param port_id 4690 * The port identifier of the Ethernet device. 4691 * 4692 * @return 4693 * - 0: Success. 4694 * - -ENODEV: The port ID is invalid. 4695 * - -EIO: if device is removed. 4696 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4697 */ 4698 int rte_eth_timesync_enable(uint16_t port_id); 4699 4700 /** 4701 * Disable IEEE1588/802.1AS timestamping for an Ethernet device. 4702 * 4703 * @param port_id 4704 * The port identifier of the Ethernet device. 4705 * 4706 * @return 4707 * - 0: Success. 4708 * - -ENODEV: The port ID is invalid. 4709 * - -EIO: if device is removed. 4710 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4711 */ 4712 int rte_eth_timesync_disable(uint16_t port_id); 4713 4714 /** 4715 * Read an IEEE1588/802.1AS Rx timestamp from an Ethernet device. 4716 * 4717 * @param port_id 4718 * The port identifier of the Ethernet device. 4719 * @param timestamp 4720 * Pointer to the timestamp struct. 4721 * @param flags 4722 * Device specific flags. Used to pass the Rx timesync register index to 4723 * i40e. Unused in igb/ixgbe, pass 0 instead. 4724 * 4725 * @return 4726 * - 0: Success. 4727 * - -EINVAL: No timestamp is available. 4728 * - -ENODEV: The port ID is invalid. 4729 * - -EIO: if device is removed. 4730 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4731 */ 4732 int rte_eth_timesync_read_rx_timestamp(uint16_t port_id, 4733 struct timespec *timestamp, uint32_t flags); 4734 4735 /** 4736 * Read an IEEE1588/802.1AS Tx timestamp from an Ethernet device. 4737 * 4738 * @param port_id 4739 * The port identifier of the Ethernet device. 4740 * @param timestamp 4741 * Pointer to the timestamp struct. 4742 * 4743 * @return 4744 * - 0: Success. 4745 * - -EINVAL: No timestamp is available. 4746 * - -ENODEV: The port ID is invalid. 4747 * - -EIO: if device is removed. 4748 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4749 */ 4750 int rte_eth_timesync_read_tx_timestamp(uint16_t port_id, 4751 struct timespec *timestamp); 4752 4753 /** 4754 * Adjust the timesync clock on an Ethernet device. 4755 * 4756 * This is usually used in conjunction with other Ethdev timesync functions to 4757 * synchronize the device time using the IEEE1588/802.1AS protocol. 4758 * 4759 * @param port_id 4760 * The port identifier of the Ethernet device. 4761 * @param delta 4762 * The adjustment in nanoseconds. 4763 * 4764 * @return 4765 * - 0: Success. 4766 * - -ENODEV: The port ID is invalid. 4767 * - -EIO: if device is removed. 4768 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4769 */ 4770 int rte_eth_timesync_adjust_time(uint16_t port_id, int64_t delta); 4771 4772 /** 4773 * Read the time from the timesync clock on an Ethernet device. 4774 * 4775 * This is usually used in conjunction with other Ethdev timesync functions to 4776 * synchronize the device time using the IEEE1588/802.1AS protocol. 4777 * 4778 * @param port_id 4779 * The port identifier of the Ethernet device. 4780 * @param time 4781 * Pointer to the timespec struct that holds the time. 4782 * 4783 * @return 4784 * - 0: Success. 4785 * - -EINVAL: Bad parameter. 4786 */ 4787 int rte_eth_timesync_read_time(uint16_t port_id, struct timespec *time); 4788 4789 /** 4790 * Set the time of the timesync clock on an Ethernet device. 4791 * 4792 * This is usually used in conjunction with other Ethdev timesync functions to 4793 * synchronize the device time using the IEEE1588/802.1AS protocol. 4794 * 4795 * @param port_id 4796 * The port identifier of the Ethernet device. 4797 * @param time 4798 * Pointer to the timespec struct that holds the time. 4799 * 4800 * @return 4801 * - 0: Success. 4802 * - -EINVAL: No timestamp is available. 4803 * - -ENODEV: The port ID is invalid. 4804 * - -EIO: if device is removed. 4805 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4806 */ 4807 int rte_eth_timesync_write_time(uint16_t port_id, const struct timespec *time); 4808 4809 /** 4810 * @warning 4811 * @b EXPERIMENTAL: this API may change without prior notice. 4812 * 4813 * Read the current clock counter of an Ethernet device 4814 * 4815 * This returns the current raw clock value of an Ethernet device. It is 4816 * a raw amount of ticks, with no given time reference. 4817 * The value returned here is from the same clock than the one 4818 * filling timestamp field of Rx packets when using hardware timestamp 4819 * offload. Therefore it can be used to compute a precise conversion of 4820 * the device clock to the real time. 4821 * 4822 * E.g, a simple heuristic to derivate the frequency would be: 4823 * uint64_t start, end; 4824 * rte_eth_read_clock(port, start); 4825 * rte_delay_ms(100); 4826 * rte_eth_read_clock(port, end); 4827 * double freq = (end - start) * 10; 4828 * 4829 * Compute a common reference with: 4830 * uint64_t base_time_sec = current_time(); 4831 * uint64_t base_clock; 4832 * rte_eth_read_clock(port, base_clock); 4833 * 4834 * Then, convert the raw mbuf timestamp with: 4835 * base_time_sec + (double)(*timestamp_dynfield(mbuf) - base_clock) / freq; 4836 * 4837 * This simple example will not provide a very good accuracy. One must 4838 * at least measure multiple times the frequency and do a regression. 4839 * To avoid deviation from the system time, the common reference can 4840 * be repeated from time to time. The integer division can also be 4841 * converted by a multiplication and a shift for better performance. 4842 * 4843 * @param port_id 4844 * The port identifier of the Ethernet device. 4845 * @param clock 4846 * Pointer to the uint64_t that holds the raw clock value. 4847 * 4848 * @return 4849 * - 0: Success. 4850 * - -ENODEV: The port ID is invalid. 4851 * - -ENOTSUP: The function is not supported by the Ethernet driver. 4852 * - -EINVAL: if bad parameter. 4853 */ 4854 __rte_experimental 4855 int 4856 rte_eth_read_clock(uint16_t port_id, uint64_t *clock); 4857 4858 /** 4859 * Get the port ID from device name. The device name should be specified 4860 * as below: 4861 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:2:00.0 4862 * - SoC device name, for example- fsl-gmac0 4863 * - vdev dpdk name, for example- net_[pcap0|null0|tap0] 4864 * 4865 * @param name 4866 * pci address or name of the device 4867 * @param port_id 4868 * pointer to port identifier of the device 4869 * @return 4870 * - (0) if successful and port_id is filled. 4871 * - (-ENODEV or -EINVAL) on failure. 4872 */ 4873 int 4874 rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id); 4875 4876 /** 4877 * Get the device name from port ID. The device name is specified as below: 4878 * - PCIe address (Domain:Bus:Device.Function), for example- 0000:02:00.0 4879 * - SoC device name, for example- fsl-gmac0 4880 * - vdev dpdk name, for example- net_[pcap0|null0|tun0|tap0] 4881 * 4882 * @param port_id 4883 * Port identifier of the device. 4884 * @param name 4885 * Buffer of size RTE_ETH_NAME_MAX_LEN to store the name. 4886 * @return 4887 * - (0) if successful. 4888 * - (-ENODEV) if *port_id* is invalid. 4889 * - (-EINVAL) on failure. 4890 */ 4891 int 4892 rte_eth_dev_get_name_by_port(uint16_t port_id, char *name); 4893 4894 /** 4895 * Check that numbers of Rx and Tx descriptors satisfy descriptors limits from 4896 * the Ethernet device information, otherwise adjust them to boundaries. 4897 * 4898 * @param port_id 4899 * The port identifier of the Ethernet device. 4900 * @param nb_rx_desc 4901 * A pointer to a uint16_t where the number of receive 4902 * descriptors stored. 4903 * @param nb_tx_desc 4904 * A pointer to a uint16_t where the number of transmit 4905 * descriptors stored. 4906 * @return 4907 * - (0) if successful. 4908 * - (-ENOTSUP, -ENODEV or -EINVAL) on failure. 4909 */ 4910 int rte_eth_dev_adjust_nb_rx_tx_desc(uint16_t port_id, 4911 uint16_t *nb_rx_desc, 4912 uint16_t *nb_tx_desc); 4913 4914 /** 4915 * Test if a port supports specific mempool ops. 4916 * 4917 * @param port_id 4918 * Port identifier of the Ethernet device. 4919 * @param [in] pool 4920 * The name of the pool operations to test. 4921 * @return 4922 * - 0: best mempool ops choice for this port. 4923 * - 1: mempool ops are supported for this port. 4924 * - -ENOTSUP: mempool ops not supported for this port. 4925 * - -ENODEV: Invalid port Identifier. 4926 * - -EINVAL: Pool param is null. 4927 */ 4928 int 4929 rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool); 4930 4931 /** 4932 * Get the security context for the Ethernet device. 4933 * 4934 * @param port_id 4935 * Port identifier of the Ethernet device 4936 * @return 4937 * - NULL on error. 4938 * - pointer to security context on success. 4939 */ 4940 void * 4941 rte_eth_dev_get_sec_ctx(uint16_t port_id); 4942 4943 /** 4944 * @warning 4945 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 4946 * 4947 * Query the device hairpin capabilities. 4948 * 4949 * @param port_id 4950 * The port identifier of the Ethernet device. 4951 * @param cap 4952 * Pointer to a structure that will hold the hairpin capabilities. 4953 * @return 4954 * - (0) if successful. 4955 * - (-ENOTSUP) if hardware doesn't support. 4956 * - (-EINVAL) if bad parameter. 4957 */ 4958 __rte_experimental 4959 int rte_eth_dev_hairpin_capability_get(uint16_t port_id, 4960 struct rte_eth_hairpin_cap *cap); 4961 4962 /** 4963 * @warning 4964 * @b EXPERIMENTAL: this structure may change without prior notice. 4965 * 4966 * Ethernet device representor ID range entry 4967 */ 4968 struct rte_eth_representor_range { 4969 enum rte_eth_representor_type type; /**< Representor type */ 4970 int controller; /**< Controller index */ 4971 int pf; /**< Physical function index */ 4972 __extension__ 4973 union { 4974 int vf; /**< VF start index */ 4975 int sf; /**< SF start index */ 4976 }; 4977 uint32_t id_base; /**< Representor ID start index */ 4978 uint32_t id_end; /**< Representor ID end index */ 4979 char name[RTE_DEV_NAME_MAX_LEN]; /**< Representor name */ 4980 }; 4981 4982 /** 4983 * @warning 4984 * @b EXPERIMENTAL: this structure may change without prior notice. 4985 * 4986 * Ethernet device representor information 4987 */ 4988 struct rte_eth_representor_info { 4989 uint16_t controller; /**< Controller ID of caller device. */ 4990 uint16_t pf; /**< Physical function ID of caller device. */ 4991 uint32_t nb_ranges_alloc; /**< Size of the ranges array. */ 4992 uint32_t nb_ranges; /**< Number of initialized ranges. */ 4993 struct rte_eth_representor_range ranges[];/**< Representor ID range. */ 4994 }; 4995 4996 /** 4997 * Retrieve the representor info of the device. 4998 * 4999 * Get device representor info to be able to calculate a unique 5000 * representor ID. @see rte_eth_representor_id_get helper. 5001 * 5002 * @param port_id 5003 * The port identifier of the device. 5004 * @param info 5005 * A pointer to a representor info structure. 5006 * NULL to return number of range entries and allocate memory 5007 * for next call to store detail. 5008 * The number of ranges that were written into this structure 5009 * will be placed into its nb_ranges field. This number cannot be 5010 * larger than the nb_ranges_alloc that by the user before calling 5011 * this function. It can be smaller than the value returned by the 5012 * function, however. 5013 * @return 5014 * - (-ENOTSUP) if operation is not supported. 5015 * - (-ENODEV) if *port_id* invalid. 5016 * - (-EIO) if device is removed. 5017 * - (>=0) number of available representor range entries. 5018 */ 5019 __rte_experimental 5020 int rte_eth_representor_info_get(uint16_t port_id, 5021 struct rte_eth_representor_info *info); 5022 5023 /** The NIC is able to deliver flag (if set) with packets to the PMD. */ 5024 #define RTE_ETH_RX_METADATA_USER_FLAG RTE_BIT64(0) 5025 5026 /** The NIC is able to deliver mark ID with packets to the PMD. */ 5027 #define RTE_ETH_RX_METADATA_USER_MARK RTE_BIT64(1) 5028 5029 /** The NIC is able to deliver tunnel ID with packets to the PMD. */ 5030 #define RTE_ETH_RX_METADATA_TUNNEL_ID RTE_BIT64(2) 5031 5032 /** 5033 * Negotiate the NIC's ability to deliver specific kinds of metadata to the PMD. 5034 * 5035 * Invoke this API before the first rte_eth_dev_configure() invocation 5036 * to let the PMD make preparations that are inconvenient to do later. 5037 * 5038 * The negotiation process is as follows: 5039 * 5040 * - the application requests features intending to use at least some of them; 5041 * - the PMD responds with the guaranteed subset of the requested feature set; 5042 * - the application can retry negotiation with another set of features; 5043 * - the application can pass zero to clear the negotiation result; 5044 * - the last negotiated result takes effect upon 5045 * the ethdev configure and start. 5046 * 5047 * @note 5048 * The PMD is supposed to first consider enabling the requested feature set 5049 * in its entirety. Only if it fails to do so, does it have the right to 5050 * respond with a smaller set of the originally requested features. 5051 * 5052 * @note 5053 * Return code (-ENOTSUP) does not necessarily mean that the requested 5054 * features are unsupported. In this case, the application should just 5055 * assume that these features can be used without prior negotiations. 5056 * 5057 * @param port_id 5058 * Port (ethdev) identifier 5059 * 5060 * @param[inout] features 5061 * Feature selection buffer 5062 * 5063 * @return 5064 * - (-EBUSY) if the port can't handle this in its current state; 5065 * - (-ENOTSUP) if the method itself is not supported by the PMD; 5066 * - (-ENODEV) if *port_id* is invalid; 5067 * - (-EINVAL) if *features* is NULL; 5068 * - (-EIO) if the device is removed; 5069 * - (0) on success 5070 */ 5071 int rte_eth_rx_metadata_negotiate(uint16_t port_id, uint64_t *features); 5072 5073 /** Flag to offload IP reassembly for IPv4 packets. */ 5074 #define RTE_ETH_DEV_REASSEMBLY_F_IPV4 (RTE_BIT32(0)) 5075 /** Flag to offload IP reassembly for IPv6 packets. */ 5076 #define RTE_ETH_DEV_REASSEMBLY_F_IPV6 (RTE_BIT32(1)) 5077 5078 /** 5079 * A structure used to get/set IP reassembly configuration. It is also used 5080 * to get the maximum capability values that a PMD can support. 5081 * 5082 * If rte_eth_ip_reassembly_capability_get() returns 0, IP reassembly can be 5083 * enabled using rte_eth_ip_reassembly_conf_set() and params values lower than 5084 * capability params can be set in the PMD. 5085 */ 5086 struct rte_eth_ip_reassembly_params { 5087 /** Maximum time in ms which PMD can wait for other fragments. */ 5088 uint32_t timeout_ms; 5089 /** Maximum number of fragments that can be reassembled. */ 5090 uint16_t max_frags; 5091 /** 5092 * Flags to enable reassembly of packet types - 5093 * RTE_ETH_DEV_REASSEMBLY_F_xxx. 5094 */ 5095 uint16_t flags; 5096 }; 5097 5098 /** 5099 * @warning 5100 * @b EXPERIMENTAL: this API may change without prior notice 5101 * 5102 * Get IP reassembly capabilities supported by the PMD. This is the first API 5103 * to be called for enabling the IP reassembly offload feature. PMD will return 5104 * the maximum values of parameters that PMD can support and user can call 5105 * rte_eth_ip_reassembly_conf_set() with param values lower than capability. 5106 * 5107 * @param port_id 5108 * The port identifier of the device. 5109 * @param capa 5110 * A pointer to rte_eth_ip_reassembly_params structure. 5111 * @return 5112 * - (-ENOTSUP) if offload configuration is not supported by device. 5113 * - (-ENODEV) if *port_id* invalid. 5114 * - (-EIO) if device is removed. 5115 * - (-EINVAL) if device is not configured or *capa* passed is NULL. 5116 * - (0) on success. 5117 */ 5118 __rte_experimental 5119 int rte_eth_ip_reassembly_capability_get(uint16_t port_id, 5120 struct rte_eth_ip_reassembly_params *capa); 5121 5122 /** 5123 * @warning 5124 * @b EXPERIMENTAL: this API may change without prior notice 5125 * 5126 * Get IP reassembly configuration parameters currently set in PMD. 5127 * The API will return error if the configuration is not already 5128 * set using rte_eth_ip_reassembly_conf_set() before calling this API or if 5129 * the device is not configured. 5130 * 5131 * @param port_id 5132 * The port identifier of the device. 5133 * @param conf 5134 * A pointer to rte_eth_ip_reassembly_params structure. 5135 * @return 5136 * - (-ENOTSUP) if offload configuration is not supported by device. 5137 * - (-ENODEV) if *port_id* invalid. 5138 * - (-EIO) if device is removed. 5139 * - (-EINVAL) if device is not configured or if *conf* passed is NULL or if 5140 * configuration is not set using rte_eth_ip_reassembly_conf_set(). 5141 * - (0) on success. 5142 */ 5143 __rte_experimental 5144 int rte_eth_ip_reassembly_conf_get(uint16_t port_id, 5145 struct rte_eth_ip_reassembly_params *conf); 5146 5147 /** 5148 * @warning 5149 * @b EXPERIMENTAL: this API may change without prior notice 5150 * 5151 * Set IP reassembly configuration parameters if the PMD supports IP reassembly 5152 * offload. User should first call rte_eth_ip_reassembly_capability_get() to 5153 * check the maximum values supported by the PMD before setting the 5154 * configuration. The use of this API is mandatory to enable this feature and 5155 * should be called before rte_eth_dev_start(). 5156 * 5157 * In datapath, PMD cannot guarantee that IP reassembly is always successful. 5158 * Hence, PMD shall register mbuf dynamic field and dynamic flag using 5159 * rte_eth_ip_reassembly_dynfield_register() to denote incomplete IP reassembly. 5160 * If dynfield is not successfully registered, error will be returned and 5161 * IP reassembly offload cannot be used. 5162 * 5163 * @param port_id 5164 * The port identifier of the device. 5165 * @param conf 5166 * A pointer to rte_eth_ip_reassembly_params structure. 5167 * @return 5168 * - (-ENOTSUP) if offload configuration is not supported by device. 5169 * - (-ENODEV) if *port_id* invalid. 5170 * - (-EIO) if device is removed. 5171 * - (-EINVAL) if device is not configured or if device is already started or 5172 * if *conf* passed is NULL or if mbuf dynfield is not registered 5173 * successfully by the PMD. 5174 * - (0) on success. 5175 */ 5176 __rte_experimental 5177 int rte_eth_ip_reassembly_conf_set(uint16_t port_id, 5178 const struct rte_eth_ip_reassembly_params *conf); 5179 5180 /** 5181 * In case of IP reassembly offload failure, packet will be updated with 5182 * dynamic flag - RTE_MBUF_DYNFLAG_IP_REASSEMBLY_INCOMPLETE_NAME and packets 5183 * will be returned without alteration. 5184 * The application can retrieve the attached fragments using mbuf dynamic field 5185 * RTE_MBUF_DYNFIELD_IP_REASSEMBLY_NAME. 5186 */ 5187 typedef struct { 5188 /** 5189 * Next fragment packet. Application should fetch dynamic field of 5190 * each fragment until a NULL is received and nb_frags is 0. 5191 */ 5192 struct rte_mbuf *next_frag; 5193 /** Time spent(in ms) by HW in waiting for further fragments. */ 5194 uint16_t time_spent; 5195 /** Number of more fragments attached in mbuf dynamic fields. */ 5196 uint16_t nb_frags; 5197 } rte_eth_ip_reassembly_dynfield_t; 5198 5199 /** 5200 * @warning 5201 * @b EXPERIMENTAL: this API may change, or be removed, without prior notice 5202 * 5203 * Dump private info from device to a file. Provided data and the order depends 5204 * on the PMD. 5205 * 5206 * @param port_id 5207 * The port identifier of the Ethernet device. 5208 * @param file 5209 * A pointer to a file for output. 5210 * @return 5211 * - (0) on success. 5212 * - (-ENODEV) if *port_id* is invalid. 5213 * - (-EINVAL) if null file. 5214 * - (-ENOTSUP) if the device does not support this function. 5215 * - (-EIO) if device is removed. 5216 */ 5217 __rte_experimental 5218 int rte_eth_dev_priv_dump(uint16_t port_id, FILE *file); 5219 5220 #include <rte_ethdev_core.h> 5221 5222 /** 5223 * @internal 5224 * Helper routine for rte_eth_rx_burst(). 5225 * Should be called at exit from PMD's rte_eth_rx_bulk implementation. 5226 * Does necessary post-processing - invokes Rx callbacks if any, etc. 5227 * 5228 * @param port_id 5229 * The port identifier of the Ethernet device. 5230 * @param queue_id 5231 * The index of the receive queue from which to retrieve input packets. 5232 * @param rx_pkts 5233 * The address of an array of pointers to *rte_mbuf* structures that 5234 * have been retrieved from the device. 5235 * @param nb_rx 5236 * The number of packets that were retrieved from the device. 5237 * @param nb_pkts 5238 * The number of elements in @p rx_pkts array. 5239 * @param opaque 5240 * Opaque pointer of Rx queue callback related data. 5241 * 5242 * @return 5243 * The number of packets effectively supplied to the @p rx_pkts array. 5244 */ 5245 uint16_t rte_eth_call_rx_callbacks(uint16_t port_id, uint16_t queue_id, 5246 struct rte_mbuf **rx_pkts, uint16_t nb_rx, uint16_t nb_pkts, 5247 void *opaque); 5248 5249 /** 5250 * 5251 * Retrieve a burst of input packets from a receive queue of an Ethernet 5252 * device. The retrieved packets are stored in *rte_mbuf* structures whose 5253 * pointers are supplied in the *rx_pkts* array. 5254 * 5255 * The rte_eth_rx_burst() function loops, parsing the Rx ring of the 5256 * receive queue, up to *nb_pkts* packets, and for each completed Rx 5257 * descriptor in the ring, it performs the following operations: 5258 * 5259 * - Initialize the *rte_mbuf* data structure associated with the 5260 * Rx descriptor according to the information provided by the NIC into 5261 * that Rx descriptor. 5262 * 5263 * - Store the *rte_mbuf* data structure into the next entry of the 5264 * *rx_pkts* array. 5265 * 5266 * - Replenish the Rx descriptor with a new *rte_mbuf* buffer 5267 * allocated from the memory pool associated with the receive queue at 5268 * initialization time. 5269 * 5270 * When retrieving an input packet that was scattered by the controller 5271 * into multiple receive descriptors, the rte_eth_rx_burst() function 5272 * appends the associated *rte_mbuf* buffers to the first buffer of the 5273 * packet. 5274 * 5275 * The rte_eth_rx_burst() function returns the number of packets 5276 * actually retrieved, which is the number of *rte_mbuf* data structures 5277 * effectively supplied into the *rx_pkts* array. 5278 * A return value equal to *nb_pkts* indicates that the Rx queue contained 5279 * at least *rx_pkts* packets, and this is likely to signify that other 5280 * received packets remain in the input queue. Applications implementing 5281 * a "retrieve as much received packets as possible" policy can check this 5282 * specific case and keep invoking the rte_eth_rx_burst() function until 5283 * a value less than *nb_pkts* is returned. 5284 * 5285 * This receive method has the following advantages: 5286 * 5287 * - It allows a run-to-completion network stack engine to retrieve and 5288 * to immediately process received packets in a fast burst-oriented 5289 * approach, avoiding the overhead of unnecessary intermediate packet 5290 * queue/dequeue operations. 5291 * 5292 * - Conversely, it also allows an asynchronous-oriented processing 5293 * method to retrieve bursts of received packets and to immediately 5294 * queue them for further parallel processing by another logical core, 5295 * for instance. However, instead of having received packets being 5296 * individually queued by the driver, this approach allows the caller 5297 * of the rte_eth_rx_burst() function to queue a burst of retrieved 5298 * packets at a time and therefore dramatically reduce the cost of 5299 * enqueue/dequeue operations per packet. 5300 * 5301 * - It allows the rte_eth_rx_burst() function of the driver to take 5302 * advantage of burst-oriented hardware features (CPU cache, 5303 * prefetch instructions, and so on) to minimize the number of CPU 5304 * cycles per packet. 5305 * 5306 * To summarize, the proposed receive API enables many 5307 * burst-oriented optimizations in both synchronous and asynchronous 5308 * packet processing environments with no overhead in both cases. 5309 * 5310 * @note 5311 * Some drivers using vector instructions require that *nb_pkts* is 5312 * divisible by 4 or 8, depending on the driver implementation. 5313 * 5314 * The rte_eth_rx_burst() function does not provide any error 5315 * notification to avoid the corresponding overhead. As a hint, the 5316 * upper-level application might check the status of the device link once 5317 * being systematically returned a 0 value for a given number of tries. 5318 * 5319 * @param port_id 5320 * The port identifier of the Ethernet device. 5321 * @param queue_id 5322 * The index of the receive queue from which to retrieve input packets. 5323 * The value must be in the range [0, nb_rx_queue - 1] previously supplied 5324 * to rte_eth_dev_configure(). 5325 * @param rx_pkts 5326 * The address of an array of pointers to *rte_mbuf* structures that 5327 * must be large enough to store *nb_pkts* pointers in it. 5328 * @param nb_pkts 5329 * The maximum number of packets to retrieve. 5330 * The value must be divisible by 8 in order to work with any driver. 5331 * @return 5332 * The number of packets actually retrieved, which is the number 5333 * of pointers to *rte_mbuf* structures effectively supplied to the 5334 * *rx_pkts* array. 5335 */ 5336 static inline uint16_t 5337 rte_eth_rx_burst(uint16_t port_id, uint16_t queue_id, 5338 struct rte_mbuf **rx_pkts, const uint16_t nb_pkts) 5339 { 5340 uint16_t nb_rx; 5341 struct rte_eth_fp_ops *p; 5342 void *qd; 5343 5344 #ifdef RTE_ETHDEV_DEBUG_RX 5345 if (port_id >= RTE_MAX_ETHPORTS || 5346 queue_id >= RTE_MAX_QUEUES_PER_PORT) { 5347 RTE_ETHDEV_LOG(ERR, 5348 "Invalid port_id=%u or queue_id=%u\n", 5349 port_id, queue_id); 5350 return 0; 5351 } 5352 #endif 5353 5354 /* fetch pointer to queue data */ 5355 p = &rte_eth_fp_ops[port_id]; 5356 qd = p->rxq.data[queue_id]; 5357 5358 #ifdef RTE_ETHDEV_DEBUG_RX 5359 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0); 5360 5361 if (qd == NULL) { 5362 RTE_ETHDEV_LOG(ERR, "Invalid Rx queue_id=%u for port_id=%u\n", 5363 queue_id, port_id); 5364 return 0; 5365 } 5366 #endif 5367 5368 nb_rx = p->rx_pkt_burst(qd, rx_pkts, nb_pkts); 5369 5370 #ifdef RTE_ETHDEV_RXTX_CALLBACKS 5371 { 5372 void *cb; 5373 5374 /* __ATOMIC_RELEASE memory order was used when the 5375 * call back was inserted into the list. 5376 * Since there is a clear dependency between loading 5377 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is 5378 * not required. 5379 */ 5380 cb = __atomic_load_n((void **)&p->rxq.clbk[queue_id], 5381 __ATOMIC_RELAXED); 5382 if (unlikely(cb != NULL)) 5383 nb_rx = rte_eth_call_rx_callbacks(port_id, queue_id, 5384 rx_pkts, nb_rx, nb_pkts, cb); 5385 } 5386 #endif 5387 5388 rte_ethdev_trace_rx_burst(port_id, queue_id, (void **)rx_pkts, nb_rx); 5389 return nb_rx; 5390 } 5391 5392 /** 5393 * Get the number of used descriptors of a Rx queue 5394 * 5395 * @param port_id 5396 * The port identifier of the Ethernet device. 5397 * @param queue_id 5398 * The queue ID on the specific port. 5399 * @return 5400 * The number of used descriptors in the specific queue, or: 5401 * - (-ENODEV) if *port_id* is invalid. 5402 * (-EINVAL) if *queue_id* is invalid 5403 * (-ENOTSUP) if the device does not support this function 5404 */ 5405 static inline int 5406 rte_eth_rx_queue_count(uint16_t port_id, uint16_t queue_id) 5407 { 5408 struct rte_eth_fp_ops *p; 5409 void *qd; 5410 5411 if (port_id >= RTE_MAX_ETHPORTS || 5412 queue_id >= RTE_MAX_QUEUES_PER_PORT) { 5413 RTE_ETHDEV_LOG(ERR, 5414 "Invalid port_id=%u or queue_id=%u\n", 5415 port_id, queue_id); 5416 return -EINVAL; 5417 } 5418 5419 /* fetch pointer to queue data */ 5420 p = &rte_eth_fp_ops[port_id]; 5421 qd = p->rxq.data[queue_id]; 5422 5423 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV); 5424 if (*p->rx_queue_count == NULL) 5425 return -ENOTSUP; 5426 if (qd == NULL) 5427 return -EINVAL; 5428 5429 return (int)(*p->rx_queue_count)(qd); 5430 } 5431 5432 /**@{@name Rx hardware descriptor states 5433 * @see rte_eth_rx_descriptor_status 5434 */ 5435 #define RTE_ETH_RX_DESC_AVAIL 0 /**< Desc available for hw. */ 5436 #define RTE_ETH_RX_DESC_DONE 1 /**< Desc done, filled by hw. */ 5437 #define RTE_ETH_RX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */ 5438 /**@}*/ 5439 5440 /** 5441 * Check the status of a Rx descriptor in the queue 5442 * 5443 * It should be called in a similar context than the Rx function: 5444 * - on a dataplane core 5445 * - not concurrently on the same queue 5446 * 5447 * Since it's a dataplane function, no check is performed on port_id and 5448 * queue_id. The caller must therefore ensure that the port is enabled 5449 * and the queue is configured and running. 5450 * 5451 * Note: accessing to a random descriptor in the ring may trigger cache 5452 * misses and have a performance impact. 5453 * 5454 * @param port_id 5455 * A valid port identifier of the Ethernet device which. 5456 * @param queue_id 5457 * A valid Rx queue identifier on this port. 5458 * @param offset 5459 * The offset of the descriptor starting from tail (0 is the next 5460 * packet to be received by the driver). 5461 * 5462 * @return 5463 * - (RTE_ETH_RX_DESC_AVAIL): Descriptor is available for the hardware to 5464 * receive a packet. 5465 * - (RTE_ETH_RX_DESC_DONE): Descriptor is done, it is filled by hw, but 5466 * not yet processed by the driver (i.e. in the receive queue). 5467 * - (RTE_ETH_RX_DESC_UNAVAIL): Descriptor is unavailable, either hold by 5468 * the driver and not yet returned to hw, or reserved by the hw. 5469 * - (-EINVAL) bad descriptor offset. 5470 * - (-ENOTSUP) if the device does not support this function. 5471 * - (-ENODEV) bad port or queue (only if compiled with debug). 5472 */ 5473 static inline int 5474 rte_eth_rx_descriptor_status(uint16_t port_id, uint16_t queue_id, 5475 uint16_t offset) 5476 { 5477 struct rte_eth_fp_ops *p; 5478 void *qd; 5479 5480 #ifdef RTE_ETHDEV_DEBUG_RX 5481 if (port_id >= RTE_MAX_ETHPORTS || 5482 queue_id >= RTE_MAX_QUEUES_PER_PORT) { 5483 RTE_ETHDEV_LOG(ERR, 5484 "Invalid port_id=%u or queue_id=%u\n", 5485 port_id, queue_id); 5486 return -EINVAL; 5487 } 5488 #endif 5489 5490 /* fetch pointer to queue data */ 5491 p = &rte_eth_fp_ops[port_id]; 5492 qd = p->rxq.data[queue_id]; 5493 5494 #ifdef RTE_ETHDEV_DEBUG_RX 5495 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV); 5496 if (qd == NULL) 5497 return -ENODEV; 5498 #endif 5499 if (*p->rx_descriptor_status == NULL) 5500 return -ENOTSUP; 5501 return (*p->rx_descriptor_status)(qd, offset); 5502 } 5503 5504 /**@{@name Tx hardware descriptor states 5505 * @see rte_eth_tx_descriptor_status 5506 */ 5507 #define RTE_ETH_TX_DESC_FULL 0 /**< Desc filled for hw, waiting xmit. */ 5508 #define RTE_ETH_TX_DESC_DONE 1 /**< Desc done, packet is transmitted. */ 5509 #define RTE_ETH_TX_DESC_UNAVAIL 2 /**< Desc used by driver or hw. */ 5510 /**@}*/ 5511 5512 /** 5513 * Check the status of a Tx descriptor in the queue. 5514 * 5515 * It should be called in a similar context than the Tx function: 5516 * - on a dataplane core 5517 * - not concurrently on the same queue 5518 * 5519 * Since it's a dataplane function, no check is performed on port_id and 5520 * queue_id. The caller must therefore ensure that the port is enabled 5521 * and the queue is configured and running. 5522 * 5523 * Note: accessing to a random descriptor in the ring may trigger cache 5524 * misses and have a performance impact. 5525 * 5526 * @param port_id 5527 * A valid port identifier of the Ethernet device which. 5528 * @param queue_id 5529 * A valid Tx queue identifier on this port. 5530 * @param offset 5531 * The offset of the descriptor starting from tail (0 is the place where 5532 * the next packet will be send). 5533 * 5534 * @return 5535 * - (RTE_ETH_TX_DESC_FULL) Descriptor is being processed by the hw, i.e. 5536 * in the transmit queue. 5537 * - (RTE_ETH_TX_DESC_DONE) Hardware is done with this descriptor, it can 5538 * be reused by the driver. 5539 * - (RTE_ETH_TX_DESC_UNAVAIL): Descriptor is unavailable, reserved by the 5540 * driver or the hardware. 5541 * - (-EINVAL) bad descriptor offset. 5542 * - (-ENOTSUP) if the device does not support this function. 5543 * - (-ENODEV) bad port or queue (only if compiled with debug). 5544 */ 5545 static inline int rte_eth_tx_descriptor_status(uint16_t port_id, 5546 uint16_t queue_id, uint16_t offset) 5547 { 5548 struct rte_eth_fp_ops *p; 5549 void *qd; 5550 5551 #ifdef RTE_ETHDEV_DEBUG_TX 5552 if (port_id >= RTE_MAX_ETHPORTS || 5553 queue_id >= RTE_MAX_QUEUES_PER_PORT) { 5554 RTE_ETHDEV_LOG(ERR, 5555 "Invalid port_id=%u or queue_id=%u\n", 5556 port_id, queue_id); 5557 return -EINVAL; 5558 } 5559 #endif 5560 5561 /* fetch pointer to queue data */ 5562 p = &rte_eth_fp_ops[port_id]; 5563 qd = p->txq.data[queue_id]; 5564 5565 #ifdef RTE_ETHDEV_DEBUG_TX 5566 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV); 5567 if (qd == NULL) 5568 return -ENODEV; 5569 #endif 5570 if (*p->tx_descriptor_status == NULL) 5571 return -ENOTSUP; 5572 return (*p->tx_descriptor_status)(qd, offset); 5573 } 5574 5575 /** 5576 * @internal 5577 * Helper routine for rte_eth_tx_burst(). 5578 * Should be called before entry PMD's rte_eth_tx_bulk implementation. 5579 * Does necessary pre-processing - invokes Tx callbacks if any, etc. 5580 * 5581 * @param port_id 5582 * The port identifier of the Ethernet device. 5583 * @param queue_id 5584 * The index of the transmit queue through which output packets must be 5585 * sent. 5586 * @param tx_pkts 5587 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures 5588 * which contain the output packets. 5589 * @param nb_pkts 5590 * The maximum number of packets to transmit. 5591 * @return 5592 * The number of output packets to transmit. 5593 */ 5594 uint16_t rte_eth_call_tx_callbacks(uint16_t port_id, uint16_t queue_id, 5595 struct rte_mbuf **tx_pkts, uint16_t nb_pkts, void *opaque); 5596 5597 /** 5598 * Send a burst of output packets on a transmit queue of an Ethernet device. 5599 * 5600 * The rte_eth_tx_burst() function is invoked to transmit output packets 5601 * on the output queue *queue_id* of the Ethernet device designated by its 5602 * *port_id*. 5603 * The *nb_pkts* parameter is the number of packets to send which are 5604 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them 5605 * allocated from a pool created with rte_pktmbuf_pool_create(). 5606 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets, 5607 * up to the number of transmit descriptors available in the Tx ring of the 5608 * transmit queue. 5609 * For each packet to send, the rte_eth_tx_burst() function performs 5610 * the following operations: 5611 * 5612 * - Pick up the next available descriptor in the transmit ring. 5613 * 5614 * - Free the network buffer previously sent with that descriptor, if any. 5615 * 5616 * - Initialize the transmit descriptor with the information provided 5617 * in the *rte_mbuf data structure. 5618 * 5619 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers, 5620 * the rte_eth_tx_burst() function uses several transmit descriptors 5621 * of the ring. 5622 * 5623 * The rte_eth_tx_burst() function returns the number of packets it 5624 * actually sent. A return value equal to *nb_pkts* means that all packets 5625 * have been sent, and this is likely to signify that other output packets 5626 * could be immediately transmitted again. Applications that implement a 5627 * "send as many packets to transmit as possible" policy can check this 5628 * specific case and keep invoking the rte_eth_tx_burst() function until 5629 * a value less than *nb_pkts* is returned. 5630 * 5631 * It is the responsibility of the rte_eth_tx_burst() function to 5632 * transparently free the memory buffers of packets previously sent. 5633 * This feature is driven by the *tx_free_thresh* value supplied to the 5634 * rte_eth_dev_configure() function at device configuration time. 5635 * When the number of free Tx descriptors drops below this threshold, the 5636 * rte_eth_tx_burst() function must [attempt to] free the *rte_mbuf* buffers 5637 * of those packets whose transmission was effectively completed. 5638 * 5639 * If the PMD is RTE_ETH_TX_OFFLOAD_MT_LOCKFREE capable, multiple threads can 5640 * invoke this function concurrently on the same Tx queue without SW lock. 5641 * @see rte_eth_dev_info_get, struct rte_eth_txconf::offloads 5642 * 5643 * @see rte_eth_tx_prepare to perform some prior checks or adjustments 5644 * for offloads. 5645 * 5646 * @param port_id 5647 * The port identifier of the Ethernet device. 5648 * @param queue_id 5649 * The index of the transmit queue through which output packets must be 5650 * sent. 5651 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 5652 * to rte_eth_dev_configure(). 5653 * @param tx_pkts 5654 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures 5655 * which contain the output packets. 5656 * @param nb_pkts 5657 * The maximum number of packets to transmit. 5658 * @return 5659 * The number of output packets actually stored in transmit descriptors of 5660 * the transmit ring. The return value can be less than the value of the 5661 * *tx_pkts* parameter when the transmit ring is full or has been filled up. 5662 */ 5663 static inline uint16_t 5664 rte_eth_tx_burst(uint16_t port_id, uint16_t queue_id, 5665 struct rte_mbuf **tx_pkts, uint16_t nb_pkts) 5666 { 5667 struct rte_eth_fp_ops *p; 5668 void *qd; 5669 5670 #ifdef RTE_ETHDEV_DEBUG_TX 5671 if (port_id >= RTE_MAX_ETHPORTS || 5672 queue_id >= RTE_MAX_QUEUES_PER_PORT) { 5673 RTE_ETHDEV_LOG(ERR, 5674 "Invalid port_id=%u or queue_id=%u\n", 5675 port_id, queue_id); 5676 return 0; 5677 } 5678 #endif 5679 5680 /* fetch pointer to queue data */ 5681 p = &rte_eth_fp_ops[port_id]; 5682 qd = p->txq.data[queue_id]; 5683 5684 #ifdef RTE_ETHDEV_DEBUG_TX 5685 RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, 0); 5686 5687 if (qd == NULL) { 5688 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n", 5689 queue_id, port_id); 5690 return 0; 5691 } 5692 #endif 5693 5694 #ifdef RTE_ETHDEV_RXTX_CALLBACKS 5695 { 5696 void *cb; 5697 5698 /* __ATOMIC_RELEASE memory order was used when the 5699 * call back was inserted into the list. 5700 * Since there is a clear dependency between loading 5701 * cb and cb->fn/cb->next, __ATOMIC_ACQUIRE memory order is 5702 * not required. 5703 */ 5704 cb = __atomic_load_n((void **)&p->txq.clbk[queue_id], 5705 __ATOMIC_RELAXED); 5706 if (unlikely(cb != NULL)) 5707 nb_pkts = rte_eth_call_tx_callbacks(port_id, queue_id, 5708 tx_pkts, nb_pkts, cb); 5709 } 5710 #endif 5711 5712 nb_pkts = p->tx_pkt_burst(qd, tx_pkts, nb_pkts); 5713 5714 rte_ethdev_trace_tx_burst(port_id, queue_id, (void **)tx_pkts, nb_pkts); 5715 return nb_pkts; 5716 } 5717 5718 /** 5719 * Process a burst of output packets on a transmit queue of an Ethernet device. 5720 * 5721 * The rte_eth_tx_prepare() function is invoked to prepare output packets to be 5722 * transmitted on the output queue *queue_id* of the Ethernet device designated 5723 * by its *port_id*. 5724 * The *nb_pkts* parameter is the number of packets to be prepared which are 5725 * supplied in the *tx_pkts* array of *rte_mbuf* structures, each of them 5726 * allocated from a pool created with rte_pktmbuf_pool_create(). 5727 * For each packet to send, the rte_eth_tx_prepare() function performs 5728 * the following operations: 5729 * 5730 * - Check if packet meets devices requirements for Tx offloads. 5731 * 5732 * - Check limitations about number of segments. 5733 * 5734 * - Check additional requirements when debug is enabled. 5735 * 5736 * - Update and/or reset required checksums when Tx offload is set for packet. 5737 * 5738 * Since this function can modify packet data, provided mbufs must be safely 5739 * writable (e.g. modified data cannot be in shared segment). 5740 * 5741 * The rte_eth_tx_prepare() function returns the number of packets ready to be 5742 * sent. A return value equal to *nb_pkts* means that all packets are valid and 5743 * ready to be sent, otherwise stops processing on the first invalid packet and 5744 * leaves the rest packets untouched. 5745 * 5746 * When this functionality is not implemented in the driver, all packets are 5747 * are returned untouched. 5748 * 5749 * @param port_id 5750 * The port identifier of the Ethernet device. 5751 * The value must be a valid port ID. 5752 * @param queue_id 5753 * The index of the transmit queue through which output packets must be 5754 * sent. 5755 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 5756 * to rte_eth_dev_configure(). 5757 * @param tx_pkts 5758 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures 5759 * which contain the output packets. 5760 * @param nb_pkts 5761 * The maximum number of packets to process. 5762 * @return 5763 * The number of packets correct and ready to be sent. The return value can be 5764 * less than the value of the *tx_pkts* parameter when some packet doesn't 5765 * meet devices requirements with rte_errno set appropriately: 5766 * - EINVAL: offload flags are not correctly set 5767 * - ENOTSUP: the offload feature is not supported by the hardware 5768 * - ENODEV: if *port_id* is invalid (with debug enabled only) 5769 * 5770 */ 5771 5772 #ifndef RTE_ETHDEV_TX_PREPARE_NOOP 5773 5774 static inline uint16_t 5775 rte_eth_tx_prepare(uint16_t port_id, uint16_t queue_id, 5776 struct rte_mbuf **tx_pkts, uint16_t nb_pkts) 5777 { 5778 struct rte_eth_fp_ops *p; 5779 void *qd; 5780 5781 #ifdef RTE_ETHDEV_DEBUG_TX 5782 if (port_id >= RTE_MAX_ETHPORTS || 5783 queue_id >= RTE_MAX_QUEUES_PER_PORT) { 5784 RTE_ETHDEV_LOG(ERR, 5785 "Invalid port_id=%u or queue_id=%u\n", 5786 port_id, queue_id); 5787 rte_errno = ENODEV; 5788 return 0; 5789 } 5790 #endif 5791 5792 /* fetch pointer to queue data */ 5793 p = &rte_eth_fp_ops[port_id]; 5794 qd = p->txq.data[queue_id]; 5795 5796 #ifdef RTE_ETHDEV_DEBUG_TX 5797 if (!rte_eth_dev_is_valid_port(port_id)) { 5798 RTE_ETHDEV_LOG(ERR, "Invalid Tx port_id=%u\n", port_id); 5799 rte_errno = ENODEV; 5800 return 0; 5801 } 5802 if (qd == NULL) { 5803 RTE_ETHDEV_LOG(ERR, "Invalid Tx queue_id=%u for port_id=%u\n", 5804 queue_id, port_id); 5805 rte_errno = EINVAL; 5806 return 0; 5807 } 5808 #endif 5809 5810 if (!p->tx_pkt_prepare) 5811 return nb_pkts; 5812 5813 return p->tx_pkt_prepare(qd, tx_pkts, nb_pkts); 5814 } 5815 5816 #else 5817 5818 /* 5819 * Native NOOP operation for compilation targets which doesn't require any 5820 * preparations steps, and functional NOOP may introduce unnecessary performance 5821 * drop. 5822 * 5823 * Generally this is not a good idea to turn it on globally and didn't should 5824 * be used if behavior of tx_preparation can change. 5825 */ 5826 5827 static inline uint16_t 5828 rte_eth_tx_prepare(__rte_unused uint16_t port_id, 5829 __rte_unused uint16_t queue_id, 5830 __rte_unused struct rte_mbuf **tx_pkts, uint16_t nb_pkts) 5831 { 5832 return nb_pkts; 5833 } 5834 5835 #endif 5836 5837 /** 5838 * Send any packets queued up for transmission on a port and HW queue 5839 * 5840 * This causes an explicit flush of packets previously buffered via the 5841 * rte_eth_tx_buffer() function. It returns the number of packets successfully 5842 * sent to the NIC, and calls the error callback for any unsent packets. Unless 5843 * explicitly set up otherwise, the default callback simply frees the unsent 5844 * packets back to the owning mempool. 5845 * 5846 * @param port_id 5847 * The port identifier of the Ethernet device. 5848 * @param queue_id 5849 * The index of the transmit queue through which output packets must be 5850 * sent. 5851 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 5852 * to rte_eth_dev_configure(). 5853 * @param buffer 5854 * Buffer of packets to be transmit. 5855 * @return 5856 * The number of packets successfully sent to the Ethernet device. The error 5857 * callback is called for any packets which could not be sent. 5858 */ 5859 static inline uint16_t 5860 rte_eth_tx_buffer_flush(uint16_t port_id, uint16_t queue_id, 5861 struct rte_eth_dev_tx_buffer *buffer) 5862 { 5863 uint16_t sent; 5864 uint16_t to_send = buffer->length; 5865 5866 if (to_send == 0) 5867 return 0; 5868 5869 sent = rte_eth_tx_burst(port_id, queue_id, buffer->pkts, to_send); 5870 5871 buffer->length = 0; 5872 5873 /* All packets sent, or to be dealt with by callback below */ 5874 if (unlikely(sent != to_send)) 5875 buffer->error_callback(&buffer->pkts[sent], 5876 (uint16_t)(to_send - sent), 5877 buffer->error_userdata); 5878 5879 return sent; 5880 } 5881 5882 /** 5883 * Buffer a single packet for future transmission on a port and queue 5884 * 5885 * This function takes a single mbuf/packet and buffers it for later 5886 * transmission on the particular port and queue specified. Once the buffer is 5887 * full of packets, an attempt will be made to transmit all the buffered 5888 * packets. In case of error, where not all packets can be transmitted, a 5889 * callback is called with the unsent packets as a parameter. If no callback 5890 * is explicitly set up, the unsent packets are just freed back to the owning 5891 * mempool. The function returns the number of packets actually sent i.e. 5892 * 0 if no buffer flush occurred, otherwise the number of packets successfully 5893 * flushed 5894 * 5895 * @param port_id 5896 * The port identifier of the Ethernet device. 5897 * @param queue_id 5898 * The index of the transmit queue through which output packets must be 5899 * sent. 5900 * The value must be in the range [0, nb_tx_queue - 1] previously supplied 5901 * to rte_eth_dev_configure(). 5902 * @param buffer 5903 * Buffer used to collect packets to be sent. 5904 * @param tx_pkt 5905 * Pointer to the packet mbuf to be sent. 5906 * @return 5907 * 0 = packet has been buffered for later transmission 5908 * N > 0 = packet has been buffered, and the buffer was subsequently flushed, 5909 * causing N packets to be sent, and the error callback to be called for 5910 * the rest. 5911 */ 5912 static __rte_always_inline uint16_t 5913 rte_eth_tx_buffer(uint16_t port_id, uint16_t queue_id, 5914 struct rte_eth_dev_tx_buffer *buffer, struct rte_mbuf *tx_pkt) 5915 { 5916 buffer->pkts[buffer->length++] = tx_pkt; 5917 if (buffer->length < buffer->size) 5918 return 0; 5919 5920 return rte_eth_tx_buffer_flush(port_id, queue_id, buffer); 5921 } 5922 5923 #ifdef __cplusplus 5924 } 5925 #endif 5926 5927 #endif /* _RTE_ETHDEV_H_ */ 5928