4 * Copyright(c) 2010-2012 Intel Corporation. All rights reserved.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
11 * * Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * * Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in
15 * the documentation and/or other materials provided with the
17 * * Neither the name of Intel Corporation nor the names of its
18 * contributors may be used to endorse or promote products derived
19 * from this software without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
24 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
25 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
26 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
27 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
28 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
29 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
30 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
31 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
35 #ifndef _RTE_ETHDEV_H_
36 #define _RTE_ETHDEV_H_
41 * RTE Ethernet Device API
43 * The Ethernet Device API is composed of two parts:
45 * - The application-oriented Ethernet API that includes functions to setup
46 * an Ethernet device (configure it, setup its RX and TX queues and start it),
47 * to get its MAC address, the speed and the status of its physical link,
48 * to receive and to transmit packets, and so on.
50 * - The driver-oriented Ethernet API that exports a function allowing
51 * an Ethernet Poll Mode Driver (PMD) to simultaneously register itself as
52 * an Ethernet device driver and as a PCI driver for a set of matching PCI
53 * [Ethernet] devices classes.
55 * By default, all the functions of the Ethernet Device API exported by a PMD
56 * are lock-free functions which assume to not be invoked in parallel on
57 * different logical cores to work on the same target object. For instance,
58 * the receive function of a PMD cannot be invoked in parallel on two logical
59 * cores to poll the same RX queue [of the same port]. Of course, this function
60 * can be invoked in parallel by different logical cores on different RX queues.
61 * It is the responsibility of the upper level application to enforce this rule.
63 * If needed, parallel accesses by multiple logical cores to shared queues
64 * shall be explicitly protected by dedicated inline lock-aware functions
65 * built on top of their corresponding lock-free functions of the PMD API.
67 * In all functions of the Ethernet API, the Ethernet device is
68 * designated by an integer >= 0 named the device port identifier.
70 * At the Ethernet driver level, Ethernet devices are represented by a generic
71 * data structure of type *rte_eth_dev*.
73 * Ethernet devices are dynamically registered during the PCI probing phase
74 * performed at EAL initialization time.
75 * When an Ethernet device is being probed, an *rte_eth_dev* structure and
76 * a new port identifier are allocated for that device. Then, the eth_dev_init()
77 * function supplied by the Ethernet driver matching the probed PCI
78 * device is invoked to properly initialize the device.
80 * The role of the device init function consists of resetting the hardware,
81 * checking access to Non-volatile Memory (NVM), reading the MAC address
84 * If the device init operation is successful, the correspondence between
85 * the port identifier assigned to the new device and its associated
86 * *rte_eth_dev* structure is effectively registered.
87 * Otherwise, both the *rte_eth_dev* structure and the port identifier are
90 * The functions exported by the application Ethernet API to setup a device
91 * designated by its port identifier must be invoked in the following order:
92 * - rte_eth_dev_configure()
93 * - rte_eth_tx_queue_setup()
94 * - rte_eth_rx_queue_setup()
95 * - rte_eth_dev_start()
97 * Then, the network application can invoke, in any order, the functions
98 * exported by the Ethernet API to get the MAC address of a given device, to
99 * get the speed and the status of a device physical link, to receive/transmit
100 * [burst of] packets, and so on.
102 * If the application wants to change the configuration (i.e. call
103 * rte_eth_dev_configure(), rte_eth_tx_queue_setup(), or
104 * rte_eth_rx_queue_setup()), it must call rte_eth_dev_stop() first to stop the
105 * device and then do the reconfiguration before calling rte_eth_dev_start()
106 * again. The tramsit and receive functions should not be invoked when the
109 * Please note that some configuration is not stored between calls to
110 * rte_eth_dev_stop()/rte_eth_dev_start(). The following configuration will
113 * - flow control settings
114 * - receive mode configuration (promiscuous mode, hardware checksum mode,
115 * RSS/VMDQ settings etc.)
116 * - VLAN filtering configuration
117 * - MAC addresses supplied to MAC address array
118 * - flow director filtering mode (but not filtering rules)
120 * Any other configuration will not be stored and will need to be re-entered
121 * after a call to rte_eth_dev_start().
123 * Finally, a network application can close an Ethernet device by invoking the
124 * rte_eth_dev_close() function.
126 * Each function of the application Ethernet API invokes a specific function
127 * of the PMD that controls the target device designated by its port
129 * For this purpose, all device-specific functions of an Ethernet driver are
130 * supplied through a set of pointers contained in a generic structure of type
132 * The address of the *eth_dev_ops* structure is stored in the *rte_eth_dev*
133 * structure by the device init function of the Ethernet driver, which is
134 * invoked during the PCI probing phase, as explained earlier.
136 * In other words, each function of the Ethernet API simply retrieves the
137 * *rte_eth_dev* structure associated with the device port identifier and
138 * performs an indirect invocation of the corresponding driver function
139 * supplied in the *eth_dev_ops* structure of the *rte_eth_dev* structure.
141 * For performance reasons, the address of the burst-oriented RX and TX
142 * functions of the Ethernet driver are not contained in the *eth_dev_ops*
143 * structure. Instead, they are directly stored at the beginning of the
144 * *rte_eth_dev* structure to avoid an extra indirect memory access during
147 * RTE ethernet device drivers do not use interrupts for transmitting or
148 * receiving. Instead, Ethernet drivers export Poll-Mode receive and transmit
149 * functions to applications.
150 * Both receive and transmit functions are packet-burst oriented to minimize
151 * their cost per packet through the following optimizations:
153 * - Sharing among multiple packets the incompressible cost of the
154 * invocation of receive/transmit functions.
156 * - Enabling receive/transmit functions to take advantage of burst-oriented
157 * hardware features (L1 cache, prefetch instructions, NIC head/tail
158 * registers) to minimize the number of CPU cycles per packet, for instance,
159 * by avoiding useless read memory accesses to ring descriptors, or by
160 * systematically using arrays of pointers that exactly fit L1 cache line
161 * boundaries and sizes.
163 * The burst-oriented receive function does not provide any error notification,
164 * to avoid the corresponding overhead. As a hint, the upper-level application
165 * might check the status of the device link once being systematically returned
166 * a 0 value by the receive function of the driver for a given number of tries.
176 #include <rte_interrupts.h>
178 #include <rte_mbuf.h>
179 #include "rte_ether.h"
182 * A structure used to retrieve statistics for an Ethernet port.
184 struct rte_eth_stats {
185 uint64_t ipackets; /**< Total number of successfully received packets. */
186 uint64_t opackets; /**< Total number of successfully transmitted packets.*/
187 uint64_t ibytes; /**< Total number of successfully received bytes. */
188 uint64_t obytes; /**< Total number of successfully transmitted bytes. */
189 uint64_t ierrors; /**< Total number of erroneous received packets. */
190 uint64_t oerrors; /**< Total number of failed transmitted packets. */
191 uint64_t imcasts; /**< Total number of multicast received packets. */
192 uint64_t rx_nombuf; /**< Total number of RX mbuf allocation failures. */
193 uint64_t fdirmatch; /**< Total number of RX packets matching a filter. */
194 uint64_t fdirmiss; /**< Total number of RX packets not matching any filter. */
198 * A structure used to retrieve link-level information of an Ethernet port.
200 struct rte_eth_link {
201 uint16_t link_speed; /**< ETH_LINK_SPEED_[10, 100, 1000, 10000] */
202 uint16_t link_duplex; /**< ETH_LINK_[HALF_DUPLEX, FULL_DUPLEX] */
203 uint8_t link_status : 1; /**< 1 -> link up, 0 -> link down */
204 }__attribute__((aligned(8))); /**< aligned for atomic64 read/write */
206 #define ETH_LINK_SPEED_AUTONEG 0 /**< Auto-negotiate link speed. */
207 #define ETH_LINK_SPEED_10 10 /**< 10 megabits/second. */
208 #define ETH_LINK_SPEED_100 100 /**< 100 megabits/second. */
209 #define ETH_LINK_SPEED_1000 1000 /**< 1 gigabits/second. */
210 #define ETH_LINK_SPEED_10000 10000 /**< 10 gigabits/second. */
212 #define ETH_LINK_AUTONEG_DUPLEX 0 /**< Auto-negotiate duplex. */
213 #define ETH_LINK_HALF_DUPLEX 1 /**< Half-duplex connection. */
214 #define ETH_LINK_FULL_DUPLEX 2 /**< Full-duplex connection. */
217 * A structure used to configure the ring threshold registers of an RX/TX
218 * queue for an Ethernet port.
220 struct rte_eth_thresh {
221 uint8_t pthresh; /**< Ring prefetch threshold. */
222 uint8_t hthresh; /**< Ring host threshold. */
223 uint8_t wthresh; /**< Ring writeback threshold. */
227 * A set of values to identify what method is to be used to route
228 * packets to multiple queues.
230 enum rte_eth_rx_mq_mode {
231 ETH_RSS = 0, /**< Default to RSS mode */
232 ETH_VMDQ_DCB /**< Use VMDQ+DCB to route traffic to queues */
236 * A structure used to configure the RX features of an Ethernet port.
238 struct rte_eth_rxmode {
239 /** The multi-queue packet distribution mode to be used, e.g. RSS. */
240 enum rte_eth_rx_mq_mode mq_mode;
241 uint32_t max_rx_pkt_len; /**< Only used if jumbo_frame enabled. */
242 uint16_t split_hdr_size; /**< hdr buf size (header_split enabled).*/
243 uint8_t header_split : 1, /**< Header Split enable. */
244 hw_ip_checksum : 1, /**< IP/UDP/TCP checksum offload enable. */
245 hw_vlan_filter : 1, /**< VLAN filter enable. */
246 jumbo_frame : 1, /**< Jumbo Frame Receipt enable. */
247 hw_strip_crc : 1; /**< Enable CRC stripping by hardware. */
251 * A structure used to configure the Receive Side Scaling (RSS) feature
252 * of an Ethernet port.
253 * If not NULL, the *rss_key* pointer of the *rss_conf* structure points
254 * to an array of 40 bytes holding the RSS key to use for hashing specific
255 * header fields of received packets.
256 * Otherwise, a default random hash key is used by the device driver.
258 * The *rss_hf* field of the *rss_conf* structure indicates the different
259 * types of IPv4/IPv6 packets to which the RSS hashing must be applied.
260 * Supplying an *rss_hf* equal to zero disables the RSS feature.
262 struct rte_eth_rss_conf {
263 uint8_t *rss_key; /**< If not NULL, 40-byte hash key. */
264 uint16_t rss_hf; /**< Hash functions to apply - see below. */
267 #define ETH_RSS_IPV4 0x0001 /**< IPv4 packet. */
268 #define ETH_RSS_IPV4_TCP 0x0002 /**< IPv4/TCP packet. */
269 #define ETH_RSS_IPV6 0x0004 /**< IPv6 packet. */
270 #define ETH_RSS_IPV6_EX 0x0008 /**< IPv6 packet with extension headers.*/
271 #define ETH_RSS_IPV6_TCP 0x0010 /**< IPv6/TCP packet. */
272 #define ETH_RSS_IPV6_TCP_EX 0x0020 /**< IPv6/TCP with extension headers. */
273 /* Intel RSS extensions to UDP packets */
274 #define ETH_RSS_IPV4_UDP 0x0040 /**< IPv4/UDP packet. */
275 #define ETH_RSS_IPV6_UDP 0x0080 /**< IPv6/UDP packet. */
276 #define ETH_RSS_IPV6_UDP_EX 0x0100 /**< IPv6/UDP with extension headers. */
278 /* Definitions used for VMDQ and DCB functionality */
279 #define ETH_VMDQ_MAX_VLAN_FILTERS 64 /**< Maximum nb. of VMDQ vlan filters. */
280 #define ETH_DCB_NUM_USER_PRIORITIES 8 /**< Maximum nb. of DCB priorities. */
281 #define ETH_VMDQ_DCB_NUM_QUEUES 128 /**< Maximum nb. of VMDQ DCB queues. */
284 * This enum indicates the possible number of queue pools
285 * in VMDQ+DCB configurations.
287 enum rte_eth_nb_pools {
288 ETH_16_POOLS = 16, /**< 16 pools with DCB. */
289 ETH_32_POOLS = 32 /**< 32 pools with DCB. */
293 * A structure used to configure the VMDQ+DCB feature
294 * of an Ethernet port.
296 * Using this feature, packets are routed to a pool of queues, based
297 * on the vlan id in the vlan tag, and then to a specific queue within
298 * that pool, using the user priority vlan tag field.
300 * A default pool may be used, if desired, to route all traffic which
301 * does not match the vlan filter rules.
303 struct rte_eth_vmdq_dcb_conf {
304 enum rte_eth_nb_pools nb_queue_pools; /**< With DCB, 16 or 32 pools */
305 uint8_t enable_default_pool; /**< If non-zero, use a default pool */
306 uint8_t default_pool; /**< The default pool, if applicable */
307 uint8_t nb_pool_maps; /**< We can have up to 64 filters/mappings */
309 uint16_t vlan_id; /**< The vlan id of the received frame */
310 uint64_t pools; /**< Bitmask of pools for packet rx */
311 } pool_map[ETH_VMDQ_MAX_VLAN_FILTERS]; /**< VMDq vlan pool maps. */
312 uint8_t dcb_queue[ETH_DCB_NUM_USER_PRIORITIES];
313 /**< Selects a queue in a pool */
317 * A structure used to configure the TX features of an Ethernet port.
318 * For future extensions.
320 struct rte_eth_txmode {
324 * A structure used to configure an RX ring of an Ethernet port.
326 struct rte_eth_rxconf {
327 struct rte_eth_thresh rx_thresh; /**< RX ring threshold registers. */
328 uint16_t rx_free_thresh; /**< Drives the freeing of RX descriptors. */
332 * A structure used to configure a TX ring of an Ethernet port.
334 struct rte_eth_txconf {
335 struct rte_eth_thresh tx_thresh; /**< TX ring threshold registers. */
336 uint16_t tx_rs_thresh; /**< Drives the setting of RS bit on TXDs. */
337 uint16_t tx_free_thresh; /**< Drives the freeing of TX buffers. */
341 * This enum indicates the flow control mode
343 enum rte_eth_fc_mode {
344 RTE_FC_NONE = 0, /**< Disable flow control. */
345 RTE_FC_RX_PAUSE, /**< RX pause frame, enable flowctrl on TX side. */
346 RTE_FC_TX_PAUSE, /**< TX pause frame, enable flowctrl on RX side. */
347 RTE_FC_FULL /**< Enable flow control on both side. */
351 * A structure used to configure Ethernet flow control parameter.
352 * These parameters will be configured into the register of the NIC.
353 * Please refer to the corresponding data sheet for proper value.
355 struct rte_eth_fc_conf {
356 uint32_t high_water; /**< High threshold value to trigger XOFF */
357 uint32_t low_water; /**< Low threshold value to trigger XON */
358 uint16_t pause_time; /**< Pause quota in the Pause frame */
359 uint16_t send_xon; /**< Is XON frame need be sent */
360 enum rte_eth_fc_mode mode; /**< Link flow control mode */
364 * Flow Director setting modes: none (default), signature or perfect.
367 RTE_FDIR_MODE_NONE = 0, /**< Disable FDIR support. */
368 RTE_FDIR_MODE_SIGNATURE, /**< Enable FDIR signature filter mode. */
369 RTE_FDIR_MODE_PERFECT, /**< Enable FDIR perfect filter mode. */
373 * Memory space that can be configured to store Flow Director filters
374 * in the board memory.
376 enum rte_fdir_pballoc_type {
377 RTE_FDIR_PBALLOC_64K = 0, /**< 64k. */
378 RTE_FDIR_PBALLOC_128K, /**< 128k. */
379 RTE_FDIR_PBALLOC_256K, /**< 256k. */
383 * Select report mode of FDIR hash information in RX descriptors.
385 enum rte_fdir_status_mode {
386 RTE_FDIR_NO_REPORT_STATUS = 0, /**< Never report FDIR hash. */
387 RTE_FDIR_REPORT_STATUS, /**< Only report FDIR hash for matching pkts. */
388 RTE_FDIR_REPORT_STATUS_ALWAYS, /**< Always report FDIR hash. */
392 * A structure used to configure the Flow Director (FDIR) feature
393 * of an Ethernet port.
395 * If mode is RTE_FDIR_DISABLE, the pballoc value is ignored.
397 struct rte_fdir_conf {
398 enum rte_fdir_mode mode; /**< Flow Director mode. */
399 enum rte_fdir_pballoc_type pballoc; /**< Space for FDIR filters. */
400 enum rte_fdir_status_mode status; /**< How to report FDIR hash. */
401 /* Offset of flexbytes field in RX packets (in 16-bit word units). */
402 uint8_t flexbytes_offset;
403 /* RX queue of packets matching a "drop" filter in perfect mode. */
408 * Possible l4type of FDIR filters.
411 RTE_FDIR_L4TYPE_NONE = 0, /**< None. */
412 RTE_FDIR_L4TYPE_UDP, /**< UDP. */
413 RTE_FDIR_L4TYPE_TCP, /**< TCP. */
414 RTE_FDIR_L4TYPE_SCTP, /**< SCTP. */
418 * Select IPv4 or IPv6 FDIR filters.
421 RTE_FDIR_IPTYPE_IPV4 = 0, /**< IPv4. */
422 RTE_FDIR_IPTYPE_IPV6 , /**< IPv6. */
426 * A structure used to define a FDIR packet filter.
428 struct rte_fdir_filter {
429 uint16_t flex_bytes; /**< Flex bytes value to match. */
430 uint16_t vlan_id; /**< VLAN ID value to match, 0 otherwise. */
431 uint16_t port_src; /**< Source port to match, 0 otherwise. */
432 uint16_t port_dst; /**< Destination port to match, 0 otherwise. */
434 uint32_t ipv4_addr; /**< IPv4 source address to match. */
435 uint32_t ipv6_addr[4]; /**< IPv6 source address to match. */
436 } ip_src; /**< IPv4/IPv6 source address to match (union of above). */
438 uint32_t ipv4_addr; /**< IPv4 destination address to match. */
439 uint32_t ipv6_addr[4]; /**< IPv6 destination address to match */
440 } ip_dst; /**< IPv4/IPv6 destination address to match (union of above). */
441 enum rte_l4type l4type; /**< l4type to match: NONE/UDP/TCP/SCTP. */
442 enum rte_iptype iptype; /**< IP packet type to match: IPv4 or IPv6. */
446 * A structure used to configure FDIR masks that are used by the device
447 * to match the various fields of RX packet headers.
448 * @note The only_ip_flow field has the opposite meaning compared to other
451 struct rte_fdir_masks {
452 /** When set to 1, packet l4type is \b NOT relevant in filters, and
453 source and destination port masks must be set to zero. */
454 uint8_t only_ip_flow;
455 /** If set to 1, vlan_id is relevant in filters. */
457 /** If set to 1, vlan_prio is relevant in filters. */
459 /** If set to 1, flexbytes is relevant in filters. */
461 /** Mask of Destination IPv4 Address. All bits set to 1 define the
462 relevant bits to use in the destination address of an IPv4 packet
463 when matching it against FDIR filters. */
464 uint32_t dst_ipv4_mask;
465 /** Mask of Source IPv4 Address. All bits set to 1 define
466 the relevant bits to use in the source address of an IPv4 packet
467 when matching it against FDIR filters. */
468 uint32_t src_ipv4_mask;
469 /** Mask of Source IPv6 Address. All bits set to 1 define the
470 relevant BYTES to use in the source address of an IPv6 packet
471 when matching it against FDIR filters. */
472 uint16_t src_ipv6_mask;
473 /** Mask of Source Port. All bits set to 1 define the relevant
474 bits to use in the source port of an IP packets when matching it
475 against FDIR filters. */
476 uint16_t src_port_mask;
477 /** Mask of Destination Port. All bits set to 1 define the relevant
478 bits to use in the destination port of an IP packet when matching it
479 against FDIR filters. */
480 uint16_t dst_port_mask;
484 * A structure used to report the status of the flow director filters in use.
486 struct rte_eth_fdir {
487 /** Number of filters with collision indication. */
489 /** Number of free (non programmed) filters. */
491 /** The Lookup hash value of the added filter that updated the value
492 of the MAXLEN field */
494 /** Longest linked list of filters in the table. */
496 /** Number of added filters. */
498 /** Number of removed filters. */
500 /** Number of failed added filters (no more space in device). */
502 /** Number of failed removed filters. */
507 * A structure used to enable/disable specific device interrupts.
509 struct rte_intr_conf {
510 /** enable/disable lsc interrupt. 0 (default) - disable, 1 enable */
515 * A structure used to configure an Ethernet port.
516 * Depending upon the RX multi-queue mode, extra advanced
517 * configuration settings may be needed.
519 struct rte_eth_conf {
521 /**< ETH_LINK_SPEED_10[0|00|000], or 0 for autonegotation */
522 uint16_t link_duplex;
523 /**< ETH_LINK_[HALF_DUPLEX|FULL_DUPLEX], or 0 for autonegotation */
524 struct rte_eth_rxmode rxmode; /**< Port RX configuration. */
525 struct rte_eth_txmode txmode; /**< Port TX configuration. */
527 struct rte_eth_rss_conf rss_conf; /**< Port RSS configuration */
528 struct rte_eth_vmdq_dcb_conf vmdq_dcb_conf;
529 /**< Port vmdq+dcb configuration. */
530 } rx_adv_conf; /**< Port RX filtering configuration (union). */
531 struct rte_fdir_conf fdir_conf; /**< FDIR configuration. */
532 struct rte_intr_conf intr_conf; /**< Interrupt mode configuration. */
536 * A structure used to retrieve the contextual information of
537 * an Ethernet device, such as the controlling driver of the device,
538 * its PCI context, etc...
540 struct rte_eth_dev_info {
541 struct rte_pci_device *pci_dev; /**< Device PCI information. */
542 const char *driver_name; /**< Device Driver name. */
543 uint32_t min_rx_bufsize; /**< Minimum size of RX buffer. */
544 uint32_t max_rx_pktlen; /**< Maximum configurable length of RX pkt. */
545 uint16_t max_rx_queues; /**< Maximum number of RX queues. */
546 uint16_t max_tx_queues; /**< Maximum number of TX queues. */
547 uint32_t max_mac_addrs; /**< Maximum number of MAC addresses. */
554 struct rte_eth_dev_callback;
555 /** @internal Structure to keep track of registered callbacks */
556 TAILQ_HEAD(rte_eth_dev_cb_list, rte_eth_dev_callback);
559 * Definitions of all functions exported by an Ethernet driver through the
560 * the generic structure of type *eth_dev_ops* supplied in the *rte_eth_dev*
561 * structure associated with an Ethernet device.
564 typedef int (*eth_dev_configure_t)(struct rte_eth_dev *dev, uint16_t nb_rx_q,
566 /**< @internal Ethernet device configuration. */
568 typedef int (*eth_dev_start_t)(struct rte_eth_dev *dev);
569 /**< @internal Function used to start a configured Ethernet device. */
571 typedef void (*eth_dev_stop_t)(struct rte_eth_dev *dev);
572 /**< @internal Function used to stop a configured Ethernet device. */
574 typedef void (*eth_dev_close_t)(struct rte_eth_dev *dev);
575 /**< @internal Function used to close a configured Ethernet device. */
577 typedef void (*eth_promiscuous_enable_t)(struct rte_eth_dev *dev);
578 /**< @internal Function used to enable the RX promiscuous mode of an Ethernet device. */
580 typedef void (*eth_promiscuous_disable_t)(struct rte_eth_dev *dev);
581 /**< @internal Function used to disable the RX promiscuous mode of an Ethernet device. */
583 typedef void (*eth_allmulticast_enable_t)(struct rte_eth_dev *dev);
584 /**< @internal Enable the receipt of all multicast packets by an Ethernet device. */
586 typedef void (*eth_allmulticast_disable_t)(struct rte_eth_dev *dev);
587 /**< @internal Disable the receipt of all multicast packets by an Ethernet device. */
589 typedef int (*eth_link_update_t)(struct rte_eth_dev *dev,
590 int wait_to_complete);
591 /**< @internal Get link speed, duplex mode and state (up/down) of an Ethernet device. */
593 typedef void (*eth_stats_get_t)(struct rte_eth_dev *dev,
594 struct rte_eth_stats *igb_stats);
595 /**< @internal Get global I/O statistics of an Ethernet device. */
597 typedef void (*eth_stats_reset_t)(struct rte_eth_dev *dev);
598 /**< @internal Reset global I/O statistics of an Ethernet device to 0. */
600 typedef void (*eth_dev_infos_get_t)(struct rte_eth_dev *dev,
601 struct rte_eth_dev_info *dev_info);
602 /**< @internal Get specific informations of an Ethernet device. */
604 typedef int (*eth_rx_queue_setup_t)(struct rte_eth_dev *dev,
605 uint16_t rx_queue_id,
607 unsigned int socket_id,
608 const struct rte_eth_rxconf *rx_conf,
609 struct rte_mempool *mb_pool);
610 /**< @internal Set up a receive queue of an Ethernet device. */
612 typedef int (*eth_tx_queue_setup_t)(struct rte_eth_dev *dev,
613 uint16_t tx_queue_id,
615 unsigned int socket_id,
616 const struct rte_eth_txconf *tx_conf);
617 /**< @internal Setup a transmit queue of an Ethernet device. */
619 typedef void (*vlan_filter_set_t)(struct rte_eth_dev *dev,
622 /**< @internal filtering of a VLAN Tag Identifier by an Ethernet device. */
624 typedef uint16_t (*eth_rx_burst_t)(struct igb_rx_queue *rxq,
625 struct rte_mbuf **rx_pkts,
627 /**< @internal Retrieve input packets from a receive queue of an Ethernet device. */
629 typedef uint16_t (*eth_tx_burst_t)(struct igb_tx_queue *txq,
630 struct rte_mbuf **tx_pkts,
632 /**< @internal Send output packets on a transmit queue of an Ethernet device. */
634 typedef int (*fdir_add_signature_filter_t)(struct rte_eth_dev *dev,
635 struct rte_fdir_filter *fdir_ftr,
637 /**< @internal Setup a new signature filter rule on an Ethernet device */
639 typedef int (*fdir_update_signature_filter_t)(struct rte_eth_dev *dev,
640 struct rte_fdir_filter *fdir_ftr,
642 /**< @internal Update a signature filter rule on an Ethernet device */
644 typedef int (*fdir_remove_signature_filter_t)(struct rte_eth_dev *dev,
645 struct rte_fdir_filter *fdir_ftr);
646 /**< @internal Remove a signature filter rule on an Ethernet device */
648 typedef void (*fdir_infos_get_t)(struct rte_eth_dev *dev,
649 struct rte_eth_fdir *fdir);
650 /**< @internal Get information about fdir status */
652 typedef int (*fdir_add_perfect_filter_t)(struct rte_eth_dev *dev,
653 struct rte_fdir_filter *fdir_ftr,
654 uint16_t soft_id, uint8_t rx_queue,
656 /**< @internal Setup a new perfect filter rule on an Ethernet device */
658 typedef int (*fdir_update_perfect_filter_t)(struct rte_eth_dev *dev,
659 struct rte_fdir_filter *fdir_ftr,
660 uint16_t soft_id, uint8_t rx_queue,
662 /**< @internal Update a perfect filter rule on an Ethernet device */
664 typedef int (*fdir_remove_perfect_filter_t)(struct rte_eth_dev *dev,
665 struct rte_fdir_filter *fdir_ftr,
667 /**< @internal Remove a perfect filter rule on an Ethernet device */
669 typedef int (*fdir_set_masks_t)(struct rte_eth_dev *dev,
670 struct rte_fdir_masks *fdir_masks);
671 /**< @internal Setup flow director masks on an Ethernet device */
673 typedef int (*flow_ctrl_set_t)(struct rte_eth_dev *dev,
674 struct rte_eth_fc_conf *fc_conf);
675 /**< @internal Setup flow control parameter on an Ethernet device */
677 typedef int (*eth_dev_led_on_t)(struct rte_eth_dev *dev);
678 /**< @internal Turn on SW controllable LED on an Ethernet device */
680 typedef int (*eth_dev_led_off_t)(struct rte_eth_dev *dev);
681 /**< @internal Turn off SW controllable LED on an Ethernet device */
683 typedef void (*eth_mac_addr_remove_t)(struct rte_eth_dev *dev, uint32_t index);
684 /**< @internal Remove MAC address from receive address register */
686 typedef void (*eth_mac_addr_add_t)(struct rte_eth_dev *dev,
687 struct ether_addr *mac_addr,
690 /**< @internal Set a MAC address into Receive Address Address Register */
693 * @internal A structure containing the functions exported by an Ethernet driver.
696 eth_dev_configure_t dev_configure; /**< Configure device. */
697 eth_dev_start_t dev_start; /**< Start device. */
698 eth_dev_stop_t dev_stop; /**< Stop device. */
699 eth_dev_close_t dev_close; /**< Close device. */
700 eth_promiscuous_enable_t promiscuous_enable; /**< Promiscuous ON. */
701 eth_promiscuous_disable_t promiscuous_disable;/**< Promiscuous OFF. */
702 eth_allmulticast_enable_t allmulticast_enable;/**< RX multicast ON. */
703 eth_allmulticast_disable_t allmulticast_disable;/**< RX multicast OF. */
704 eth_link_update_t link_update; /**< Get device link state. */
705 eth_stats_get_t stats_get; /**< Get device statistics. */
706 eth_stats_reset_t stats_reset; /**< Reset device statistics. */
707 eth_dev_infos_get_t dev_infos_get; /**< Get device info. */
708 vlan_filter_set_t vlan_filter_set; /**< Filter VLAN Setup. */
709 eth_rx_queue_setup_t rx_queue_setup;/**< Set up device RX queue.*/
710 eth_tx_queue_setup_t tx_queue_setup;/**< Set up device TX queue.*/
711 eth_dev_led_on_t dev_led_on; /**< Turn on LED. */
712 eth_dev_led_off_t dev_led_off; /**< Turn off LED. */
713 flow_ctrl_set_t flow_ctrl_set; /**< Setup flow control. */
714 eth_mac_addr_remove_t mac_addr_remove; /**< Remove MAC address */
715 eth_mac_addr_add_t mac_addr_add; /**< Add a MAC address */
717 /** Add a signature filter. */
718 fdir_add_signature_filter_t fdir_add_signature_filter;
719 /** Update a signature filter. */
720 fdir_update_signature_filter_t fdir_update_signature_filter;
721 /** Remove a signature filter. */
722 fdir_remove_signature_filter_t fdir_remove_signature_filter;
723 /** Get information about FDIR status. */
724 fdir_infos_get_t fdir_infos_get;
725 /** Add a perfect filter. */
726 fdir_add_perfect_filter_t fdir_add_perfect_filter;
727 /** Update a perfect filter. */
728 fdir_update_perfect_filter_t fdir_update_perfect_filter;
729 /** Remove a perfect filter. */
730 fdir_remove_perfect_filter_t fdir_remove_perfect_filter;
731 /** Setup masks for FDIR filtering. */
732 fdir_set_masks_t fdir_set_masks;
737 * The generic data structure associated with each ethernet device.
739 * Pointers to burst-oriented packet receive and transmit functions are
740 * located at the beginning of the structure, along with the pointer to
741 * where all the data elements for the particular device are stored in shared
742 * memory. This split allows the function pointer and driver data to be per-
743 * process, while the actual configuration data for the device is shared.
746 eth_rx_burst_t rx_pkt_burst; /**< Pointer to PMD receive function. */
747 eth_tx_burst_t tx_pkt_burst; /**< Pointer to PMD transmit function. */
748 struct rte_eth_dev_data *data; /**< Pointer to device data */
749 const struct eth_driver *driver;/**< Driver for this device */
750 struct eth_dev_ops *dev_ops; /**< Functions exported by PMD */
751 struct rte_pci_device *pci_dev; /**< PCI info. supplied by probing */
752 struct rte_eth_dev_cb_list callbacks; /**< User application callbacks */
757 * The data part, with no function pointers, associated with each ethernet device.
759 * This structure is safe to place in shared memory to be common among different
760 * processes in a multi-process configuration.
762 struct rte_eth_dev_data {
763 struct igb_rx_queue **rx_queues; /**< Array of pointers to RX queues. */
764 struct igb_tx_queue **tx_queues; /**< Array of pointers to TX queues. */
765 uint16_t nb_rx_queues; /**< Number of RX queues. */
766 uint16_t nb_tx_queues; /**< Number of TX queues. */
768 void *dev_private; /**< PMD-specific private data */
770 struct rte_eth_link dev_link;
771 /**< Link-level information & status */
773 struct rte_eth_conf dev_conf; /**< Configuration applied to device. */
774 uint16_t max_frame_size; /**< Default is ETHER_MAX_LEN (1518). */
776 uint64_t rx_mbuf_alloc_failed; /**< RX ring mbuf allocation failures. */
777 struct ether_addr* mac_addrs;/**< Device Ethernet Link address. */
778 uint8_t port_id; /**< Device [external] port identifier. */
779 uint8_t promiscuous : 1, /**< RX promiscuous mode ON(1) / OFF(0). */
780 scattered_rx : 1, /**< RX of scattered packets is ON(1) / OFF(0) */
781 all_multicast : 1, /**< RX all multicast mode ON(1) / OFF(0). */
782 dev_started : 1; /**< Device state: STARTED(1) / STOPPED(0). */
787 * The pool of *rte_eth_dev* structures. The size of the pool
788 * is configured at compile-time in the <rte_ethdev.c> file.
790 extern struct rte_eth_dev rte_eth_devices[];
793 * Get the total number of Ethernet devices that have been successfully
794 * initialized by the [matching] Ethernet driver during the PCI probing phase.
795 * All devices whose port identifier is in the range
796 * [0, rte_eth_dev_count() - 1] can be operated on by network applications.
799 * - The total number of usable Ethernet devices.
801 extern uint8_t rte_eth_dev_count(void);
806 * Initialization function of an Ethernet driver invoked for each matching
807 * Ethernet PCI device detected during the PCI probing phase.
810 * The pointer to the [matching] Ethernet driver structure supplied by
811 * the PMD when it registered itself.
813 * The *eth_dev* pointer is the address of the *rte_eth_dev* structure
814 * associated with the matching device and which have been [automatically]
815 * allocated in the *rte_eth_devices* array.
816 * The *eth_dev* structure is supplied to the driver initialization function
817 * with the following fields already initialized:
819 * - *pci_dev*: Holds the pointers to the *rte_pci_device* structure which
820 * contains the generic PCI information of the matching device.
822 * - *dev_private*: Holds a pointer to the device private data structure.
824 * - *max_frame_size*: Contains the default Ethernet maximum frame length
827 * - *port_id*: Contains the port index of the device (actually the index
828 * of the *eth_dev* structure in the *rte_eth_devices* array).
831 * - 0: Success, the device is properly initialized by the driver.
832 * In particular, the driver MUST have set up the *dev_ops* pointer
833 * of the *eth_dev* structure.
834 * - <0: Error code of the device initialization failure.
836 typedef int (*eth_dev_init_t)(struct eth_driver *eth_drv,
837 struct rte_eth_dev *eth_dev);
841 * The structure associated with a PMD Ethernet driver.
843 * Each Ethernet driver acts as a PCI driver and is represented by a generic
844 * *eth_driver* structure that holds:
846 * - An *rte_pci_driver* structure (which must be the first field).
848 * - The *eth_dev_init* function invoked for each matching PCI device.
850 * - The size of the private data to allocate for each matching device.
853 struct rte_pci_driver pci_drv; /**< The PMD is also a PCI driver. */
854 eth_dev_init_t eth_dev_init; /**< Device init function. */
855 unsigned int dev_private_size; /**< Size of device private data. */
860 * A function invoked by the initialization function of an Ethernet driver
861 * to simultaneously register itself as a PCI driver and as an Ethernet
862 * Poll Mode Driver (PMD).
865 * The pointer to the *eth_driver* structure associated with
866 * the Ethernet driver.
868 extern void rte_eth_driver_register(struct eth_driver *eth_drv);
871 * The initialization function of the driver for
872 * Intel(r) IGB Gigabit Ethernet Controller devices.
873 * This function is invoked once at EAL start time.
877 extern int rte_igb_pmd_init(void);
880 * The initialization function of the driver for
881 * Intel(r) EM Gigabit Ethernet Controller devices.
882 * This function is invoked once at EAL start time.
886 extern int rte_em_pmd_init(void);
889 * The initialization function of the driver for 1Gbps Intel IGB_VF
891 * Invoked once at EAL start time.
895 extern int rte_igbvf_pmd_init(void);
898 * The initialization function of the driver for 10Gbps Intel IXGBE
900 * Invoked once at EAL start time.
904 extern int rte_ixgbe_pmd_init(void);
907 * The initialization function of the driver for 10Gbps Intel IXGBE_VF
909 * Invoked once at EAL start time.
913 extern int rte_ixgbevf_pmd_init(void);
916 * The initialization function of *all* supported and enabled drivers.
917 * Right now, the following PMDs are supported:
923 * This function is invoked once at EAL start time.
926 * Error code of the device initialization failure,
927 * -ENODEV if there are no drivers available
928 * (e.g. if all driver config options are = n).
931 int rte_pmd_init_all(void)
935 #ifdef RTE_LIBRTE_IGB_PMD
936 if ((ret = rte_igb_pmd_init()) != 0) {
937 RTE_LOG(ERR, PMD, "Cannot init igb PMD\n");
940 if ((ret = rte_igbvf_pmd_init()) != 0) {
941 RTE_LOG(ERR, PMD, "Cannot init igbvf PMD\n");
944 #endif /* RTE_LIBRTE_IGB_PMD */
946 #ifdef RTE_LIBRTE_EM_PMD
947 if ((ret = rte_em_pmd_init()) != 0) {
948 RTE_LOG(ERR, PMD, "Cannot init em PMD\n");
951 #endif /* RTE_LIBRTE_EM_PMD */
953 #ifdef RTE_LIBRTE_IXGBE_PMD
954 if ((ret = rte_ixgbe_pmd_init()) != 0) {
955 RTE_LOG(ERR, PMD, "Cannot init ixgbe PMD\n");
958 if ((ret = rte_ixgbevf_pmd_init()) != 0) {
959 RTE_LOG(ERR, PMD, "Cannot init ixgbevf PMD\n");
962 #endif /* RTE_LIBRTE_IXGBE_PMD */
965 RTE_LOG(ERR, PMD, "No PMD(s) are configured\n");
970 * Configure an Ethernet device.
971 * This function must be invoked first before any other function in the
972 * Ethernet API. This function can also be re-invoked when a device is in the
976 * The port identifier of the Ethernet device to configure.
978 * The number of receive queues to set up for the Ethernet device.
980 * The number of transmit queues to set up for the Ethernet device.
982 * The pointer to the configuration data to be used for the Ethernet device.
983 * The *rte_eth_conf* structure includes:
984 * - the hardware offload features to activate, with dedicated fields for
985 * each statically configurable offload hardware feature provided by
986 * Ethernet devices, such as IP checksum or VLAN tag stripping for
988 * - the Receive Side Scaling (RSS) configuration when using multiple RX
991 * Embedding all configuration information in a single data structure
992 * is the more flexible method that allows the addition of new features
993 * without changing the syntax of the API.
995 * - 0: Success, device configured.
996 * - <0: Error code returned by the driver configuration function.
998 extern int rte_eth_dev_configure(uint8_t port_id,
999 uint16_t nb_rx_queue,
1000 uint16_t nb_tx_queue,
1001 const struct rte_eth_conf *eth_conf);
1004 * Allocate and set up a receive queue for an Ethernet device.
1006 * The function allocates a contiguous block of memory for *nb_rx_desc*
1007 * receive descriptors from a memory zone associated with *socket_id*
1008 * and initializes each receive descriptor with a network buffer allocated
1009 * from the memory pool *mb_pool*.
1012 * The port identifier of the Ethernet device.
1013 * @param rx_queue_id
1014 * The index of the receive queue to set up.
1015 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1016 * to rte_eth_dev_configure().
1018 * The number of receive descriptors to allocate for the receive ring.
1020 * The *socket_id* argument is the socket identifier in case of NUMA.
1021 * The value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1022 * the DMA memory allocated for the receive descriptors of the ring.
1024 * The pointer to the configuration data to be used for the receive queue.
1025 * The *rx_conf* structure contains an *rx_thresh* structure with the values
1026 * of the Prefetch, Host, and Write-Back threshold registers of the receive
1029 * The pointer to the memory pool from which to allocate *rte_mbuf* network
1030 * memory buffers to populate each descriptor of the receive ring.
1032 * - 0: Success, receive queue correctly set up.
1033 * - -EINVAL: The size of network buffers which can be allocated from the
1034 * memory pool does not fit the various buffer sizes allowed by the
1035 * device controller.
1036 * - -ENOMEM: Unable to allocate the receive ring descriptors or to
1037 * allocate network memory buffers from the memory pool when
1038 * initializing receive descriptors.
1040 extern int rte_eth_rx_queue_setup(uint8_t port_id, uint16_t rx_queue_id,
1041 uint16_t nb_rx_desc, unsigned int socket_id,
1042 const struct rte_eth_rxconf *rx_conf,
1043 struct rte_mempool *mb_pool);
1046 * Allocate and set up a transmit queue for an Ethernet device.
1049 * The port identifier of the Ethernet device.
1050 * @param tx_queue_id
1051 * The index of the transmit queue to set up.
1052 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1053 * to rte_eth_dev_configure().
1055 * The number of transmit descriptors to allocate for the transmit ring.
1057 * The *socket_id* argument is the socket identifier in case of NUMA.
1058 * Its value can be *SOCKET_ID_ANY* if there is no NUMA constraint for
1059 * the DMA memory allocated for the transmit descriptors of the ring.
1061 * The pointer to the configuration data to be used for the transmit queue.
1062 * The *tx_conf* structure contains the following data:
1063 * - The *tx_thresh* structure with the values of the Prefetch, Host, and
1064 * Write-Back threshold registers of the transmit ring.
1065 * When setting Write-Back threshold to the value greater then zero,
1066 * *tx_rs_thresh* value should be explicitly set to one.
1067 * - The *tx_free_thresh* value indicates the [minimum] number of network
1068 * buffers that must be pending in the transmit ring to trigger their
1069 * [implicit] freeing by the driver transmit function.
1070 * - The *tx_rs_thresh* value indicates the [minimum] number of transmit
1071 * descriptors that must be pending in the transmit ring before setting the
1072 * RS bit on a descriptor by the driver transmit function.
1073 * The *tx_rs_thresh* value should be less or equal then
1074 * *tx_free_thresh* value, and both of them should be less then
1077 * Note that setting *tx_free_thresh* or *tx_rs_thresh* value to 0 forces
1078 * the transmit function to use default values.
1080 * - 0: Success, the transmit queue is correctly set up.
1081 * - -ENOMEM: Unable to allocate the transmit ring descriptors.
1083 extern int rte_eth_tx_queue_setup(uint8_t port_id, uint16_t tx_queue_id,
1084 uint16_t nb_tx_desc, unsigned int socket_id,
1085 const struct rte_eth_txconf *tx_conf);
1088 * Start an Ethernet device.
1090 * The device start step is the last one and consists of setting the configured
1091 * offload features and in starting the transmit and the receive units of the
1093 * On success, all basic functions exported by the Ethernet API (link status,
1094 * receive/transmit, and so on) can be invoked.
1097 * The port identifier of the Ethernet device.
1099 * - 0: Success, Ethernet device started.
1100 * - <0: Error code of the driver device start function.
1102 extern int rte_eth_dev_start(uint8_t port_id);
1105 * Stop an Ethernet device. The device can be restarted with a call to
1106 * rte_eth_dev_start()
1109 * The port identifier of the Ethernet device.
1111 extern void rte_eth_dev_stop(uint8_t port_id);
1114 * Close an Ethernet device. The device cannot be restarted!
1117 * The port identifier of the Ethernet device.
1119 extern void rte_eth_dev_close(uint8_t port_id);
1122 * Enable receipt in promiscuous mode for an Ethernet device.
1125 * The port identifier of the Ethernet device.
1127 extern void rte_eth_promiscuous_enable(uint8_t port_id);
1130 * Disable receipt in promiscuous mode for an Ethernet device.
1133 * The port identifier of the Ethernet device.
1135 extern void rte_eth_promiscuous_disable(uint8_t port_id);
1138 * Return the value of promiscuous mode for an Ethernet device.
1141 * The port identifier of the Ethernet device.
1143 * - (1) if promiscuous is enabled
1144 * - (0) if promiscuous is disabled.
1147 extern int rte_eth_promiscuous_get(uint8_t port_id);
1150 * Enable the receipt of any multicast frame by an Ethernet device.
1153 * The port identifier of the Ethernet device.
1155 extern void rte_eth_allmulticast_enable(uint8_t port_id);
1158 * Disable the receipt of all multicast frames by an Ethernet device.
1161 * The port identifier of the Ethernet device.
1163 extern void rte_eth_allmulticast_disable(uint8_t port_id);
1166 * Return the value of allmulticast mode for an Ethernet device.
1169 * The port identifier of the Ethernet device.
1171 * - (1) if allmulticast is enabled
1172 * - (0) if allmulticast is disabled.
1175 extern int rte_eth_allmulticast_get(uint8_t port_id);
1178 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
1179 * or FULL-DUPLEX) of the physical link of an Ethernet device. It might need
1180 * to wait up to 9 seconds in it.
1183 * The port identifier of the Ethernet device.
1185 * A pointer to an *rte_eth_link* structure to be filled with
1186 * the status, the speed and the mode of the Ethernet device link.
1188 extern void rte_eth_link_get(uint8_t port_id, struct rte_eth_link *link);
1191 * Retrieve the status (ON/OFF), the speed (in Mbps) and the mode (HALF-DUPLEX
1192 * or FULL-DUPLEX) of the physical link of an Ethernet device. It is a no-wait
1193 * version of rte_eth_link_get().
1196 * The port identifier of the Ethernet device.
1198 * A pointer to an *rte_eth_link* structure to be filled with
1199 * the status, the speed and the mode of the Ethernet device link.
1201 extern void rte_eth_link_get_nowait(uint8_t port_id,
1202 struct rte_eth_link *link);
1205 * Retrieve the general I/O statistics of an Ethernet device.
1208 * The port identifier of the Ethernet device.
1210 * A pointer to a structure of type *rte_eth_stats* to be filled with
1211 * the values of device counters for the following set of statistics:
1212 * - *ipackets* with the total of successfully received packets.
1213 * - *opackets* with the total of successfully transmitted packets.
1214 * - *ibytes* with the total of successfully received bytes.
1215 * - *obytes* with the total of successfully transmitted bytes.
1216 * - *ierrors* with the total of erroneous received packets.
1217 * - *oerrors* with the total of failed transmitted packets.
1219 extern void rte_eth_stats_get(uint8_t port_id, struct rte_eth_stats *stats);
1222 * Reset the general I/O statistics of an Ethernet device.
1225 * The port identifier of the Ethernet device.
1227 extern void rte_eth_stats_reset(uint8_t port_id);
1230 * Retrieve the Ethernet address of an Ethernet device.
1233 * The port identifier of the Ethernet device.
1235 * A pointer to a structure of type *ether_addr* to be filled with
1236 * the Ethernet address of the Ethernet device.
1238 extern void rte_eth_macaddr_get(uint8_t port_id, struct ether_addr *mac_addr);
1241 * Retrieve the contextual information of an Ethernet device.
1244 * The port identifier of the Ethernet device.
1246 * A pointer to a structure of type *rte_eth_dev_info* to be filled with
1247 * the contextual information of the Ethernet device.
1249 extern void rte_eth_dev_info_get(uint8_t port_id,
1250 struct rte_eth_dev_info *dev_info);
1253 * Enable/Disable hardware filtering by an Ethernet device of received
1254 * VLAN packets tagged with a given VLAN Tag Identifier.
1257 * The port identifier of the Ethernet device.
1259 * The VLAN Tag Identifier whose filtering must be enabled or disabled.
1261 * If > 0, enable VLAN filtering of VLAN packets tagged with *vlan_id*.
1262 * Otherwise, disable VLAN filtering of VLAN packets tagged with *vlan_id*.
1264 * - (0) if successful.
1265 * - (-ENOSUP) if hardware-assisted VLAN filtering not configured.
1266 * - (-ENODEV) if *port_id* invalid.
1267 * - (-ENOSYS) if VLAN filtering on *port_id* disabled.
1268 * - (-EINVAL) if *vlan_id* > 4095.
1270 extern int rte_eth_dev_vlan_filter(uint8_t port_id, uint16_t vlan_id, int on);
1274 * Retrieve a burst of input packets from a receive queue of an Ethernet
1275 * device. The retrieved packets are stored in *rte_mbuf* structures whose
1276 * pointers are supplied in the *rx_pkts* array.
1278 * The rte_eth_rx_burst() function loops, parsing the RX ring of the
1279 * receive queue, up to *nb_pkts* packets, and for each completed RX
1280 * descriptor in the ring, it performs the following operations:
1282 * - Initialize the *rte_mbuf* data structure associated with the
1283 * RX descriptor according to the information provided by the NIC into
1284 * that RX descriptor.
1286 * - Store the *rte_mbuf* data structure into the next entry of the
1289 * - Replenish the RX descriptor with a new *rte_mbuf* buffer
1290 * allocated from the memory pool associated with the receive queue at
1291 * initialization time.
1293 * When retrieving an input packet that was scattered by the controller
1294 * into multiple receive descriptors, the rte_eth_rx_burst() function
1295 * appends the associated *rte_mbuf* buffers to the first buffer of the
1298 * The rte_eth_rx_burst() function returns the number of packets
1299 * actually retrieved, which is the number of *rte_mbuf* data structures
1300 * effectively supplied into the *rx_pkts* array.
1301 * A return value equal to *nb_pkts* indicates that the RX queue contained
1302 * at least *rx_pkts* packets, and this is likely to signify that other
1303 * received packets remain in the input queue. Applications implementing
1304 * a "retrieve as much received packets as possible" policy can check this
1305 * specific case and keep invoking the rte_eth_rx_burst() function until
1306 * a value less than *nb_pkts* is returned.
1308 * This receive method has the following advantages:
1310 * - It allows a run-to-completion network stack engine to retrieve and
1311 * to immediately process received packets in a fast burst-oriented
1312 * approach, avoiding the overhead of unnecessary intermediate packet
1313 * queue/dequeue operations.
1315 * - Conversely, it also allows an asynchronous-oriented processing
1316 * method to retrieve bursts of received packets and to immediately
1317 * queue them for further parallel processing by another logical core,
1318 * for instance. However, instead of having received packets being
1319 * individually queued by the driver, this approach allows the invoker
1320 * of the rte_eth_rx_burst() function to queue a burst of retrieved
1321 * packets at a time and therefore dramatically reduce the cost of
1322 * enqueue/dequeue operations per packet.
1324 * - It allows the rte_eth_rx_burst() function of the driver to take
1325 * advantage of burst-oriented hardware features (CPU cache,
1326 * prefetch instructions, and so on) to minimize the number of CPU
1327 * cycles per packet.
1329 * To summarize, the proposed receive API enables many
1330 * burst-oriented optimizations in both synchronous and asynchronous
1331 * packet processing environments with no overhead in both cases.
1333 * The rte_eth_rx_burst() function does not provide any error
1334 * notification to avoid the corresponding overhead. As a hint, the
1335 * upper-level application might check the status of the device link once
1336 * being systematically returned a 0 value for a given number of tries.
1339 * The port identifier of the Ethernet device.
1341 * The index of the receive queue from which to retrieve input packets.
1342 * The value must be in the range [0, nb_rx_queue - 1] previously supplied
1343 * to rte_eth_dev_configure().
1345 * The address of an array of pointers to *rte_mbuf* structures that
1346 * must be large enough to store *nb_pkts* pointers in it.
1348 * The maximum number of packets to retrieve.
1350 * The number of packets actually retrieved, which is the number
1351 * of pointers to *rte_mbuf* structures effectively supplied to the
1354 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
1355 extern uint16_t rte_eth_rx_burst(uint8_t port_id, uint16_t queue_id,
1356 struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
1358 static inline uint16_t
1359 rte_eth_rx_burst(uint8_t port_id, uint16_t queue_id,
1360 struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
1362 struct rte_eth_dev *dev;
1364 dev = &rte_eth_devices[port_id];
1365 return (*dev->rx_pkt_burst)(dev->data->rx_queues[queue_id], rx_pkts, nb_pkts);
1370 * Send a burst of output packets on a transmit queue of an Ethernet device.
1372 * The rte_eth_tx_burst() function is invoked to transmit output packets
1373 * on the output queue *queue_id* of the Ethernet device designated by its
1375 * The *nb_pkts* parameter is the number of packets to send which are
1376 * supplied in the *tx_pkts* array of *rte_mbuf* structures.
1377 * The rte_eth_tx_burst() function loops, sending *nb_pkts* packets,
1378 * up to the number of transmit descriptors available in the TX ring of the
1380 * For each packet to send, the rte_eth_tx_burst() function performs
1381 * the following operations:
1383 * - Pick up the next available descriptor in the transmit ring.
1385 * - Free the network buffer previously sent with that descriptor, if any.
1387 * - Initialize the transmit descriptor with the information provided
1388 * in the *rte_mbuf data structure.
1390 * In the case of a segmented packet composed of a list of *rte_mbuf* buffers,
1391 * the rte_eth_tx_burst() function uses several transmit descriptors
1394 * The rte_eth_tx_burst() function returns the number of packets it
1395 * actually sent. A return value equal to *nb_pkts* means that all packets
1396 * have been sent, and this is likely to signify that other output packets
1397 * could be immediately transmitted again. Applications that implement a
1398 * "send as many packets to transmit as possible" policy can check this
1399 * specific case and keep invoking the rte_eth_tx_burst() function until
1400 * a value less than *nb_pkts* is returned.
1402 * It is the responsibility of the rte_eth_tx_burst() function to
1403 * transparently free the memory buffers of packets previously sent.
1404 * This feature is driven by the *tx_free_thresh* value supplied to the
1405 * rte_eth_dev_configure() function at device configuration time.
1406 * When the number of previously sent packets reached the "minimum transmit
1407 * packets to free" threshold, the rte_eth_tx_burst() function must
1408 * [attempt to] free the *rte_mbuf* buffers of those packets whose
1409 * transmission was effectively completed.
1412 * The port identifier of the Ethernet device.
1414 * The index of the transmit queue through which output packets must be
1416 * The value must be in the range [0, nb_tx_queue - 1] previously supplied
1417 * to rte_eth_dev_configure().
1419 * The address of an array of *nb_pkts* pointers to *rte_mbuf* structures
1420 * which contain the output packets.
1422 * The maximum number of packets to transmit.
1424 * The number of output packets actually stored in transmit descriptors of
1425 * the transmit ring. The return value can be less than the value of the
1426 * *tx_pkts* parameter when the transmit ring is full or has been filled up.
1428 #ifdef RTE_LIBRTE_ETHDEV_DEBUG
1429 extern uint16_t rte_eth_tx_burst(uint8_t port_id, uint16_t queue_id,
1430 struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
1432 static inline uint16_t
1433 rte_eth_tx_burst(uint8_t port_id, uint16_t queue_id,
1434 struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
1436 struct rte_eth_dev *dev;
1438 dev = &rte_eth_devices[port_id];
1439 return (*dev->tx_pkt_burst)(dev->data->tx_queues[queue_id], tx_pkts, nb_pkts);
1444 * Setup a new signature filter rule on an Ethernet device
1447 * The port identifier of the Ethernet device.
1448 * @param fdir_filter
1449 * The pointer to the fdir filter structure describing the signature filter
1451 * The *rte_fdir_filter* structure includes the values of the different fields
1452 * to match: source and destination IP addresses, vlan id, flexbytes, source
1453 * and destination ports, and so on.
1455 * The index of the RX queue where to store RX packets matching the added
1456 * signature filter defined in fdir_filter.
1458 * - (0) if successful.
1459 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1460 * - (-ENODEV) if *port_id* invalid.
1461 * - (-ENOSYS) if the FDIR mode is not configured in signature mode
1463 * - (-EINVAL) if the fdir_filter information is not correct.
1465 int rte_eth_dev_fdir_add_signature_filter(uint8_t port_id,
1466 struct rte_fdir_filter *fdir_filter,
1470 * Update a signature filter rule on an Ethernet device.
1471 * If the rule doesn't exits, it is created.
1474 * The port identifier of the Ethernet device.
1476 * The pointer to the structure describing the signature filter rule.
1477 * The *rte_fdir_filter* structure includes the values of the different fields
1478 * to match: source and destination IP addresses, vlan id, flexbytes, source
1479 * and destination ports, and so on.
1481 * The index of the RX queue where to store RX packets matching the added
1482 * signature filter defined in fdir_ftr.
1484 * - (0) if successful.
1485 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1486 * - (-ENODEV) if *port_id* invalid.
1487 * - (-ENOSYS) if the flow director mode is not configured in signature mode
1489 * - (-EINVAL) if the fdir_filter information is not correct.
1491 int rte_eth_dev_fdir_update_signature_filter(uint8_t port_id,
1492 struct rte_fdir_filter *fdir_ftr,
1496 * Remove a signature filter rule on an Ethernet device.
1499 * The port identifier of the Ethernet device.
1501 * The pointer to the structure describing the signature filter rule.
1502 * The *rte_fdir_filter* structure includes the values of the different fields
1503 * to match: source and destination IP addresses, vlan id, flexbytes, source
1504 * and destination ports, and so on.
1506 * - (0) if successful.
1507 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1508 * - (-ENODEV) if *port_id* invalid.
1509 * - (-ENOSYS) if the flow director mode is not configured in signature mode
1511 * - (-EINVAL) if the fdir_filter information is not correct.
1513 int rte_eth_dev_fdir_remove_signature_filter(uint8_t port_id,
1514 struct rte_fdir_filter *fdir_ftr);
1517 * Retrieve the flow director information of an Ethernet device.
1520 * The port identifier of the Ethernet device.
1522 * A pointer to a structure of type *rte_eth_dev_fdir* to be filled with
1523 * the flow director information of the Ethernet device.
1525 * - (0) if successful.
1526 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1527 * - (-ENODEV) if *port_id* invalid.
1528 * - (-ENOSYS) if the flow director mode is not configured on *port_id*.
1530 int rte_eth_dev_fdir_get_infos(uint8_t port_id, struct rte_eth_fdir *fdir);
1533 * Add a new perfect filter rule on an Ethernet device.
1536 * The port identifier of the Ethernet device.
1537 * @param fdir_filter
1538 * The pointer to the structure describing the perfect filter rule.
1539 * The *rte_fdir_filter* structure includes the values of the different fields
1540 * to match: source and destination IP addresses, vlan id, flexbytes, source
1541 * and destination ports, and so on.
1542 * IPv6 are not supported.
1544 * The 16-bit value supplied in the field hash.fdir.id of mbuf for RX
1545 * packets matching the perfect filter.
1547 * The index of the RX queue where to store RX packets matching the added
1548 * perfect filter defined in fdir_filter.
1550 * If drop is set to 1, matching RX packets are stored into the RX drop
1551 * queue defined in the rte_fdir_conf.
1553 * - (0) if successful.
1554 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1555 * - (-ENODEV) if *port_id* invalid.
1556 * - (-ENOSYS) if the flow director mode is not configured in perfect mode
1558 * - (-EINVAL) if the fdir_filter information is not correct.
1560 int rte_eth_dev_fdir_add_perfect_filter(uint8_t port_id,
1561 struct rte_fdir_filter *fdir_filter,
1562 uint16_t soft_id, uint8_t rx_queue,
1566 * Update a perfect filter rule on an Ethernet device.
1567 * If the rule doesn't exits, it is created.
1570 * The port identifier of the Ethernet device.
1571 * @param fdir_filter
1572 * The pointer to the structure describing the perfect filter rule.
1573 * The *rte_fdir_filter* structure includes the values of the different fields
1574 * to match: source and destination IP addresses, vlan id, flexbytes, source
1575 * and destination ports, and so on.
1576 * IPv6 are not supported.
1578 * The 16-bit value supplied in the field hash.fdir.id of mbuf for RX
1579 * packets matching the perfect filter.
1581 * The index of the RX queue where to store RX packets matching the added
1582 * perfect filter defined in fdir_filter.
1584 * If drop is set to 1, matching RX packets are stored into the RX drop
1585 * queue defined in the rte_fdir_conf.
1587 * - (0) if successful.
1588 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1589 * - (-ENODEV) if *port_id* invalid.
1590 * - (-ENOSYS) if the flow director mode is not configured in perfect mode
1592 * - (-EINVAL) if the fdir_filter information is not correct.
1594 int rte_eth_dev_fdir_update_perfect_filter(uint8_t port_id,
1595 struct rte_fdir_filter *fdir_filter,
1596 uint16_t soft_id, uint8_t rx_queue,
1600 * Remove a perfect filter rule on an Ethernet device.
1603 * The port identifier of the Ethernet device.
1604 * @param fdir_filter
1605 * The pointer to the structure describing the perfect filter rule.
1606 * The *rte_fdir_filter* structure includes the values of the different fields
1607 * to match: source and destination IP addresses, vlan id, flexbytes, source
1608 * and destination ports, and so on.
1609 * IPv6 are not supported.
1611 * The soft_id value provided when adding/updating the removed filter.
1613 * - (0) if successful.
1614 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1615 * - (-ENODEV) if *port_id* invalid.
1616 * - (-ENOSYS) if the flow director mode is not configured in perfect mode
1618 * - (-EINVAL) if the fdir_filter information is not correct.
1620 int rte_eth_dev_fdir_remove_perfect_filter(uint8_t port_id,
1621 struct rte_fdir_filter *fdir_filter,
1624 * Configure globally the masks for flow director mode for an Ethernet device.
1625 * For example, the device can match packets with only the first 24 bits of
1626 * the IPv4 source address.
1628 * The following fields can be masked: IPv4 addresses and L4 port numbers.
1629 * The following fields can be either enabled or disabled completely for the
1630 * matching functionality: VLAN ID tag; VLAN Priority + CFI bit; Flexible 2-byte
1632 * IPv6 masks are not supported.
1634 * All filters must comply with the masks previously configured.
1635 * For example, with a mask equal to 255.255.255.0 for the source IPv4 address,
1636 * all IPv4 filters must be created with a source IPv4 address that fits the
1639 * This function flushes all filters that have been previously added in
1643 * The port identifier of the Ethernet device.
1645 * The pointer to the fdir mask structure describing relevant headers fields
1646 * and relevant bits to use when matching packets addresses and ports.
1647 * IPv6 masks are not supported.
1649 * - (0) if successful.
1650 * - (-ENOTSUP) if hardware doesn't support flow director mode.
1651 * - (-ENODEV) if *port_id* invalid.
1652 * - (-ENOSYS) if the flow director mode is not configured in perfect
1653 * mode on *port_id*.
1654 * - (-EINVAL) if the fdir_filter information is not correct
1656 int rte_eth_dev_fdir_set_masks(uint8_t port_id,
1657 struct rte_fdir_masks *fdir_mask);
1660 * The eth device event type for interrupt, and maybe others in the future.
1662 enum rte_eth_event_type {
1663 RTE_ETH_EVENT_UNKNOWN, /**< unknown event type */
1664 RTE_ETH_EVENT_INTR_LSC, /**< lsc interrupt event */
1665 RTE_ETH_EVENT_MAX /**< max value of this enum */
1668 typedef void (*rte_eth_dev_cb_fn)(uint8_t port_id, \
1669 enum rte_eth_event_type event, void *cb_arg);
1670 /**< user application callback to be registered for interrupts */
1675 * Register a callback function for specific port id.
1682 * User supplied callback function to be called.
1684 * Pointer to the parameters for the registered callback.
1687 * - On success, zero.
1688 * - On failure, a negative value.
1690 int rte_eth_dev_callback_register(uint8_t port_id,
1691 enum rte_eth_event_type event,
1692 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
1695 * Unregister a callback function for specific port id.
1702 * User supplied callback function to be called.
1704 * Pointer to the parameters for the registered callback. -1 means to
1705 * remove all for the same callback address and same event.
1708 * - On success, zero.
1709 * - On failure, a negative value.
1711 int rte_eth_dev_callback_unregister(uint8_t port_id,
1712 enum rte_eth_event_type event,
1713 rte_eth_dev_cb_fn cb_fn, void *cb_arg);
1716 * @internal Executes all the user application registered callbacks for
1717 * the specific device. It is for DPDK internal user only. User
1718 * application should not call it directly.
1721 * Pointer to struct rte_eth_dev.
1723 * Eth device interrupt event type.
1728 void _rte_eth_dev_callback_process(struct rte_eth_dev *dev,
1729 enum rte_eth_event_type event);
1732 * Turn on the LED on the Ethernet device.
1733 * This function turns on the LED on the Ethernet device.
1736 * The port identifier of the Ethernet device.
1738 * - (0) if successful.
1739 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
1741 * - (-ENODEV) if *port_id* invalid.
1743 int rte_eth_led_on(uint8_t port_id);
1746 * Turn off the LED on the Ethernet device.
1747 * This function turns off the LED on the Ethernet device.
1750 * The port identifier of the Ethernet device.
1752 * - (0) if successful.
1753 * - (-ENOTSUP) if underlying hardware OR driver doesn't support
1755 * - (-ENODEV) if *port_id* invalid.
1757 int rte_eth_led_off(uint8_t port_id);
1760 * Configure the Ethernet link flow control for Ethernet device
1763 * The port identifier of the Ethernet device.
1765 * The pointer to the structure of the flow control parameters.
1767 * - (0) if successful.
1768 * - (-ENOTSUP) if hardware doesn't support flow control mode.
1769 * - (-ENODEV) if *port_id* invalid.
1770 * - (-EINVAL) if bad parameter
1771 * - (-EIO) if flow control setup failure
1773 int rte_eth_dev_flow_ctrl_set(uint8_t port_id,
1774 struct rte_eth_fc_conf *fc_conf);
1777 * Add a MAC address to an internal array of addresses used to enable whitelist
1778 * filtering to accept packets only if the destination MAC address matches.
1781 * The port identifier of the Ethernet device.
1783 * The MAC address to add.
1785 * VMDq pool index to associate address with (if VMDq is enabled). If VMDq is
1786 * not enabled, this should be set to 0.
1788 * - (0) if successfully added or *mac_addr" was already added.
1789 * - (-ENOTSUP) if hardware doesn't support this feature.
1790 * - (-ENODEV) if *port* is invalid.
1791 * - (-ENOSPC) if no more MAC addresses can be added.
1792 * - (-EINVAL) if MAC address is invalid.
1794 int rte_eth_dev_mac_addr_add(uint8_t port, struct ether_addr *mac_addr,
1798 * Remove a MAC address from the internal array of addresses.
1801 * The port identifier of the Ethernet device.
1803 * MAC address to remove.
1805 * - (0) if successful, or *mac_addr* didn't exist.
1806 * - (-ENOTSUP) if hardware doesn't support.
1807 * - (-ENODEV) if *port* invalid.
1808 * - (-EADDRINUSE) if attempting to remove the default MAC address
1810 int rte_eth_dev_mac_addr_remove(uint8_t port, struct ether_addr *mac_addr);
1817 #endif /* _RTE_ETHDEV_H_ */