1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright 2021 The DPDK contributors 3 4.. include:: <isonum.txt> 5 6DPDK Release 21.11 7================== 8 9New Features 10------------ 11 12* **Enabled new devargs parser.** 13 14 * Enabled devargs syntax: 15 ``bus=X,paramX=x/class=Y,paramY=y/driver=Z,paramZ=z``. 16 * Added bus-level parsing of the devargs syntax. 17 * Kept compatibility with the legacy syntax as parsing fallback. 18 19* **Updated EAL hugetlbfs mount handling for Linux.** 20 21 * Modified EAL to allow ``--huge-dir`` option to specify a sub-directory 22 within a hugetlbfs mountpoint. 23 24* **Added dmadev library.** 25 26 * Added a DMA device framework for management and provision of 27 hardware and software DMA devices. 28 * Added generic API which support a number of different DMA 29 operations. 30 * Added multi-process support. 31 32* **Updated default KNI behavior on net devices control callbacks.** 33 34 Updated KNI net devices control callbacks to run with ``rtnl`` kernel lock 35 held by default. A newly added ``enable_bifurcated`` KNI kernel module 36 parameter can be used to run callbacks with ``rtnl`` lock released. 37 38* **Added HiSilicon DMA driver.** 39 40 The HiSilicon DMA driver provides device drivers for the Kunpeng's DMA devices. 41 This device driver can be used through the generic dmadev API. 42 43* **Added IDXD dmadev driver implementation.** 44 45 The IDXD dmadev driver provides device drivers for the Intel DSA devices. 46 This device driver can be used through the generic dmadev API. 47 48* **Added IOAT dmadev driver implementation.** 49 50 The Intel I/O Acceleration Technology (IOAT) dmadev driver provides a device 51 driver for Intel IOAT devices such as Crystal Beach DMA (CBDMA) on Ice Lake, 52 Skylake and Broadwell. This device driver can be used through the generic dmadev API. 53 54* **Added Marvell CNXK DMA driver.** 55 56 Added dmadev driver for the DPI DMA hardware accelerator 57 of Marvell OCTEONTX2 and OCTEONTX3 family of SoCs. 58 59* **Added NXP DPAA DMA driver.** 60 61 Added a new dmadev driver for the NXP DPAA platform. 62 63* **Added support to get all MAC addresses of a device.** 64 65 Added ``rte_eth_macaddrs_get`` to allow a user to retrieve all Ethernet 66 addresses assigned to a given Ethernet port. 67 68* **Introduced GPU device class.** 69 70 Introduced the GPU device class with initial features: 71 72 * Device information. 73 * Memory management. 74 * Communication flag and list. 75 76* **Added NVIDIA GPU driver implemented with CUDA library.** 77 78 Added NVIDIA GPU driver implemented with CUDA library under the new 79 GPU device interface. 80 81* **Added new RSS offload types for IPv4/L4 checksum in RSS flow.** 82 83 Added macros ``ETH_RSS_IPV4_CHKSUM`` and ``ETH_RSS_L4_CHKSUM``. The IPv4 and 84 TCP/UDP/SCTP header checksum field can now be used as input set for RSS. 85 86* **Added L2TPv2 and PPP protocol support in flow API.** 87 88 Added flow pattern items and header formats for the L2TPv2 and PPP protocols. 89 90* **Added flow flex item.** 91 92 The configurable flow flex item provides the capability to introduce 93 an arbitrary user-specified network protocol header, 94 configure the hardware accordingly, and perform match on this header 95 with desired patterns and masks. 96 97* **Added ethdev support to control delivery of Rx metadata from the HW to the PMD.** 98 99 A new API, ``rte_eth_rx_metadata_negotiate()``, was added. 100 The following parts of Rx metadata were defined: 101 102 * ``RTE_ETH_RX_METADATA_USER_FLAG`` 103 * ``RTE_ETH_RX_METADATA_USER_MARK`` 104 * ``RTE_ETH_RX_METADATA_TUNNEL_ID`` 105 106* **Added an API to get a proxy port to manage "transfer" flows.** 107 108 A new API, ``rte_flow_pick_transfer_proxy()``, was added. 109 110* **Added ethdev shared Rx queue support.** 111 112 * Added new device capability flag and Rx domain field to switch info. 113 * Added share group and share queue ID to Rx queue configuration. 114 * Added testpmd support and dedicated forwarding engine. 115 116* **Updated af_packet ethdev driver.** 117 118 * The default VLAN strip behavior has changed. The VLAN tag won't be stripped 119 unless ``DEV_RX_OFFLOAD_VLAN_STRIP`` offload is enabled. 120 121* **Added API to get device configuration in ethdev.** 122 123 Added an ethdev API which can help users get device configuration. 124 125* **Updated AF_XDP PMD.** 126 127 * Disabled secondary process support due to insufficient state shared 128 between processes which causes a crash. This will be fixed/re-enabled 129 in the next release. 130 131* **Updated Amazon ENA PMD.** 132 133 Updated the Amazon ENA PMD. The new driver version (v2.5.0) introduced 134 bug fixes and improvements, including: 135 136 * Support for the ``tx_free_thresh`` and ``rx_free_thresh`` configuration parameters. 137 * NUMA aware allocations for the queue helper structures. 138 * A Watchdog feature which is checking for missing Tx completions. 139 140* **Updated Broadcom bnxt PMD.** 141 142 * Added flow offload support for Thor. 143 * Added TruFlow and AFM SRAM partitioning support. 144 * Implemented support for tunnel offload. 145 * Updated HWRM API to version 1.10.2.68. 146 * Added NAT support for destination IP and port combination. 147 * Added support for socket redirection. 148 * Added wildcard match support for ingress flows. 149 * Added support for inner IP header for GRE tunnel flows. 150 * Updated support for RSS action in flow rules. 151 * Removed devargs option for stats accumulation. 152 153* **Updated Cisco enic driver.** 154 155 * Added rte_flow support for matching GTP, GTP-C and GTP-U headers. 156 157* **Updated Intel e1000 emulated driver.** 158 159 * Added Intel e1000 support on Windows. 160 161* **Updated Intel iavf driver.** 162 163 * Added Intel iavf support on Windows. 164 * Added IPv4 and L4 (TCP/UDP/SCTP) checksum hash support in RSS flow. 165 * Added PPPoL2TPv2oUDP RSS hash based on inner IP address and TCP/UDP port. 166 * Added Intel iavf inline crypto support. 167 168* **Updated Intel ice driver.** 169 170 * Added protocol agnostic flow offloading support in Flow Director. 171 * Added protocol agnostic flow offloading support in RSS hash. 172 * Added 1PPS out support via devargs. 173 * Added IPv4 and L4 (TCP/UDP/SCTP) checksum hash support in RSS flow. 174 * Added ``DEV_RX_OFFLOAD_TIMESTAMP`` support. 175 * Added timesync API support under scalar path. 176 * Added DCF reset API support. 177 178* **Updated Intel ixgbe driver.** 179 180 * Added Intel ixgbe support on Windows. 181 182* **Updated Marvell cnxk ethdev driver.** 183 184 * Added rte_flow support for dual VLAN insert and strip actions. 185 * Added rte_tm support. 186 * Added support for Inline IPsec for CN9K event mode and CN10K 187 poll mode and event mode. 188 * Added support for ingress meter for CN10K platform. 189 190* **Updated Mellanox mlx5 driver.** 191 192 Updated the Mellanox mlx5 driver with new features and improvements, including: 193 194 * Added implicit mempool registration to avoid data path hiccups (opt-out). 195 * Added delay drop support for Rx queues. 196 * Added NIC offloads for the PMD on Windows (TSO, VLAN strip, CRC keep). 197 * Added socket direct mode bonding support. 198 199* **Added NXP ENETFEC PMD [EXPERIMENTAL].** 200 201 Added the new ENETFEC driver for the NXP IMX8MMEVK platform. See the 202 :doc:`../nics/enetfec` NIC driver guide for more details on this new driver. 203 204* **Updated Solarflare network PMD.** 205 206 Updated the Solarflare ``sfc_efx`` driver with changes including: 207 208 * Added port representors support on SN1000 SmartNICs 209 * Added flow API transfer proxy support 210 * Added SN1000 virtual functions (VF) support 211 * Added support for flow counters without service cores 212 * Added support for regioned DMA mapping required on SN1022 SoC 213 214* **Added power monitor API in vhost library.** 215 216 Added an API to support power monitor in vhost library. 217 218* **Updated vhost PMD.** 219 220 Add power monitor support in vhost PMD. 221 222* **Updated virtio PMD.** 223 224 * Initial support for RSS receive mode has been added to the Virtio PMD, 225 with the capability for the application to configure the hash key, 226 the RETA and the hash types. 227 Virtio hash reporting is yet to be added. 228 * Added power monitor support in virtio PMD. 229 230* **Updated Wangxun ngbe driver.** 231 232 * Added offloads and packet type on RxTx. 233 * Added VLAN and MAC filters. 234 * Added device basic statistics and extended stats. 235 * Added multi-queue and RSS. 236 * Added SRIOV. 237 * Added flow control. 238 * Added IEEE 1588. 239 240* **Added new vDPA PMD based on Xilinx devices.** 241 242 Added a new Xilinx vDPA (``sfc_vdpa``) PMD. 243 See the :doc:`../vdpadevs/sfc` guide for more details on this driver. 244 245* **Added telemetry callbacks to the cryptodev library.** 246 247 Added telemetry callback functions which allow a list of crypto devices, 248 stats for a crypto device, and other device information to be queried. 249 Also added callback to get cryptodev capabilities. 250 251* **Added telemetry to security library.** 252 253 Added telemetry callback functions to query security capabilities of 254 crypto device. 255 256* **Updated Marvell cnxk crypto PMD.** 257 258 * Added AES-CBC SHA1-HMAC support in lookaside protocol (IPsec) for CN10K. 259 * Added Transport mode support in lookaside protocol (IPsec) for CN10K. 260 * Added UDP encapsulation support in lookaside protocol (IPsec) for CN10K. 261 * Added support for lookaside protocol (IPsec) offload for CN9K. 262 * Added support for ZUC algorithm with 256-bit key length for CN10K. 263 * Added support for CN98xx dual block. 264 * Added inner checksum support in lookaside protocol (IPsec) for CN10K. 265 * Added AES-CBC NULL auth support in lookaside protocol (IPsec) for CN10K. 266 * Added ESN and anti-replay support in lookaside protocol (IPsec) for CN9K. 267 268* **Added support for event crypto adapter on Marvell CN10K and CN9K.** 269 270 * Added event crypto adapter ``OP_FORWARD`` mode support. 271 272* **Updated Mellanox mlx5 crypto driver.** 273 274 * Added Windows support. 275 * Added support for BlueField-2 and ConnectX-6 Dx. 276 277* **Updated NXP dpaa_sec crypto PMD.** 278 279 * Added DES-CBC, AES-XCBC-MAC, AES-CMAC and non-HMAC algorithm support. 280 * Added PDCP short MAC-I support. 281 * Added raw vector datapath API support. 282 283* **Updated NXP dpaa2_sec crypto PMD.** 284 285 * Added PDCP short MAC-I support. 286 * Added raw vector datapath API support. 287 288* **Added framework for consolidation of IPsec_MB dependent SW Crypto PMDs.** 289 290 * The IPsec_MB framework was added to share common code between Intel 291 SW Crypto PMDs that depend on the intel-ipsec-mb library. 292 * Multiprocess support was added for the consolidated PMDs 293 which requires v1.1 of the intel-ipsec-mb library. 294 * The following PMDs were moved into a single source folder 295 while their usage and EAL options remain unchanged. 296 * AESNI_MB PMD. 297 * AESNI_GCM PMD. 298 * KASUMI PMD. 299 * SNOW3G PMD. 300 * ZUC PMD. 301 * CHACHA20_POLY1305 - a new PMD. 302 303* **Updated the aesni_mb crypto PMD.** 304 305 * Added support for ZUC-EEA3-256 and ZUC-EIA3-256. 306 307* **Added digest appended ops support for Snow3G PMD.** 308 309 * Added support for out-of-place auth-cipher operations that encrypt 310 the digest along with the rest of the raw data. 311 * Added support for partially encrypted digest when using auth-cipher 312 operations. 313 314* **Updated the ACC100 bbdev PMD.** 315 316 Added support for more comprehensive CRC options. 317 318* **Updated the turbo_sw bbdev PMD.** 319 320 Added support for more comprehensive CRC options. 321 322* **Added NXP LA12xx baseband PMD.** 323 324 * Added a new baseband PMD for NXP LA12xx Software defined radio. 325 * See the :doc:`../bbdevs/la12xx` for more details. 326 327* **Updated Mellanox compress driver.** 328 329 * Added devargs option to allow manual setting of Huffman block size. 330 331* **Updated Mellanox regex driver.** 332 333 * Added support for new ROF file format. 334 335* **Updated IPsec library.** 336 337 * Added support for more AEAD algorithms AES_CCM, CHACHA20_POLY1305 338 and AES_GMAC. 339 * Added support for NAT-T / UDP encapsulated ESP. 340 * Added support for SA telemetry. 341 * Added support for setting a non default starting ESN value. 342 * Added support for TSO in inline crypto mode. 343 344* **Added optimized Toeplitz hash implementation.** 345 346 Added optimized Toeplitz hash implementation using Galois Fields New Instructions. 347 348* **Added multi-process support for testpmd.** 349 350 Added command-line options to specify total number of processes and 351 current process ID. Each process owns a subset of Rx and Tx queues. 352 353* **Updated test-crypto-perf application with new cases.** 354 355 * Added support for asymmetric crypto throughput performance measurement. 356 Only modex is supported for now. 357 * Added support for lookaside IPsec protocol offload throughput measurement. 358 359* **Added lookaside protocol (IPsec) tests in dpdk-test.** 360 361 * Added known vector tests (AES-GCM 128, 192, 256). 362 * Added tests to verify error reporting with ICV corruption. 363 * Added tests to verify IV generation. 364 * Added tests to verify UDP encapsulation. 365 * Added tests to verify UDP encapsulation ports. 366 * Added tests to validate packets soft expiry. 367 * Added tests to validate packets hard expiry. 368 * Added tests to verify tunnel header verification in IPsec inbound. 369 * Added tests to verify inner checksum. 370 * Added tests for CHACHA20_POLY1305 PMD, including a new testcase for SGL OOP. 371 372* **Updated l3fwd sample application.** 373 374 * Increased number of routes to 16 for all lookup modes (LPM, EM and FIB). 375 This helps in validating SoC with many Ethernet devices. 376 * Updated EM mode to use RFC2544 reserved IP address space with RFC863 377 UDP discard protocol. 378 379* **Updated IPsec Security Gateway sample application with new features.** 380 381 * Added support for TSO (only for inline crypto TCP packets). 382 * Added support for telemetry. 383 * Added support for more AEAD algorithms: AES-GMAC, AES_CTR, AES_XCBC_MAC, 384 AES_CCM, CHACHA20_POLY1305 385 * Added support for event vectors for inline protocol offload mode. 386 387* **Revised packet capture framework.** 388 389 * New dpdk-dumpcap program that has most of the features 390 of the wireshark dumpcap utility including: 391 capture of multiple interfaces, filtering, 392 and stopping after number of bytes, packets. 393 * New library for writing pcapng packet capture files. 394 * Enhancements to the pdump library to support: 395 * Packet filter with BPF. 396 * Pcapng format with timestamps and meta-data. 397 * Fixes packet capture with stripped VLAN tags. 398 399* **Added ASan support.** 400 401 Added ASan/AddressSanitizer support. `AddressSanitizer 402 <https://github.com/google/sanitizers/wiki/AddressSanitizer>`_ 403 is a widely-used debugging tool to detect memory access errors. 404 It helps to detect issues like use-after-free, various kinds of buffer 405 overruns in C/C++ programs, and other similar errors, as well as 406 printing out detailed debug information whenever an error is detected. 407 408 409Removed Items 410------------- 411 412* eal: Removed the deprecated function ``rte_get_master_lcore()`` 413 and the iterator macro ``RTE_LCORE_FOREACH_SLAVE``. 414 415* eal: The old API arguments that were deprecated for 416 blacklist/whitelist are removed. Users must use the new 417 block/allow list arguments. 418 419* mbuf: Removed offload flag ``PKT_RX_EIP_CKSUM_BAD``. 420 The ``PKT_RX_OUTER_IP_CKSUM_BAD`` flag should be used as a replacement. 421 422* ethdev: Removed the port mirroring API. A more fine-grain flow API 423 action ``RTE_FLOW_ACTION_TYPE_SAMPLE`` should be used instead. 424 The structures ``rte_eth_mirror_conf`` and ``rte_eth_vlan_mirror`` and 425 the functions ``rte_eth_mirror_rule_set`` and 426 ``rte_eth_mirror_rule_reset`` along with the associated macros 427 ``ETH_MIRROR_*`` are removed. 428 429* ethdev: Removed the ``rte_eth_rx_descriptor_done()`` API function and its 430 driver callback. It is replaced by the more complete function 431 ``rte_eth_rx_descriptor_status()``. 432 433* ethdev: Removed deprecated ``shared`` attribute of the 434 ``struct rte_flow_action_count``. Shared counters should be managed 435 using indirect actions API (``rte_flow_action_handle_create`` etc). 436 437* i40e: Removed i40evf driver. 438 iavf already became the default VF driver for i40e devices, 439 so there is no need to maintain i40evf. 440 441 442API Changes 443----------- 444 445* eal: The lcore state ``FINISHED`` is removed from 446 the ``enum rte_lcore_state_t``. 447 The lcore state ``WAIT`` is enough to represent the same state. 448 449* eal: Made ``rte_intr_handle`` structure definition hidden. 450 451* kvargs: The experimental function ``rte_kvargs_strcmp()`` has been 452 removed. Its usages have been replaced by a new function 453 ``rte_kvargs_get_with_value()``. 454 455* cmdline: ``cmdline_stdin_exit()`` now frees the ``cmdline`` structure. 456 Calls to ``cmdline_free()`` after it need to be deleted from applications. 457 458* cmdline: Made ``cmdline`` structure definition hidden on Linux and FreeBSD. 459 460* cmdline: Made ``rdline`` structure definition hidden. Functions are added 461 to dynamically allocate and free it, and to access user data in callbacks. 462 463* mempool: Added ``RTE_MEMPOOL_F_NON_IO`` flag to give a hint to DPDK components 464 that objects from this pool will not be used for device IO (e.g. DMA). 465 466* mempool: The mempool flags ``MEMPOOL_F_*`` will be deprecated in the future. 467 Newly added flags with ``RTE_MEMPOOL_F_`` prefix should be used instead. 468 469* mempool: Helper macro ``MEMPOOL_HEADER_SIZE()`` is deprecated. 470 The replacement macro ``RTE_MEMPOOL_HEADER_SIZE()`` is internal only. 471 472* mempool: Macro to register mempool driver ``MEMPOOL_REGISTER_OPS()`` is 473 deprecated. Use replacement ``RTE_MEMPOOL_REGISTER_OPS()``. 474 475* mempool: The mempool API macros ``MEMPOOL_PG_*`` are deprecated and 476 will be removed in DPDK 22.11. 477 478* mbuf: The mbuf offload flags ``PKT_*`` are renamed as ``RTE_MBUF_F_*``. A 479 compatibility layer will be kept until DPDK 22.11. 480* net: Renamed ``s_addr`` and ``d_addr`` fields of ``rte_ether_hdr`` structure 481 to ``src_addr`` and ``dst_addr``, respectively. 482 483* net: Added ``version`` and ``ihl`` bit-fields to ``struct rte_ipv4_hdr``. 484 Existing ``version_ihl`` field is kept for backward compatibility. 485 486* ethdev: Added items and actions ``PORT_REPRESENTOR``, ``REPRESENTED_PORT`` to 487 flow API. 488 489* ethdev: Deprecated items and actions ``PF``, ``VF``, ``PHY_PORT``, ``PORT_ID``. 490 Suggested items and actions ``PORT_REPRESENTOR``, ``REPRESENTED_PORT`` instead. 491 492* ethdev: Deprecated the use of attributes ``ingress`` / ``egress`` combined 493 with ``transfer``. See items ``PORT_REPRESENTOR``, ``REPRESENTED_PORT``. 494 495* ethdev: ``rte_flow_action_modify_data`` structure updated, immediate data 496 array is extended, data pointer field is explicitly added to union, the 497 action behavior is defined in a more strict fashion and documentation updated. 498 The immediate value behavior has been changed, the entire immediate field 499 should be provided, and offset for immediate source bitfield is assigned 500 from the destination one. 501 502* vhost: ``rte_vdpa_register_device``, ``rte_vdpa_unregister_device``, 503 ``rte_vhost_host_notifier_ctrl`` and ``rte_vdpa_relay_vring_used`` vDPA 504 driver interface are marked as internal. 505 506* cryptodev: The API ``rte_cryptodev_pmd_is_valid_dev()`` is modified to 507 ``rte_cryptodev_is_valid_dev()`` as it can be used by the application as 508 well as the PMD to check whether the device is valid or not. 509 510* cryptodev: The ``rte_cryptodev_pmd.*`` files are renamed to ``cryptodev_pmd.*`` 511 since they are for drivers only and should be private to DPDK, and not 512 installed for app use. 513 514* cryptodev: A ``reserved`` byte from structure ``rte_crypto_op`` was 515 renamed to ``aux_flags`` to indicate warnings and other information from 516 the crypto/security operation. This field will be used to communicate 517 events such as soft expiry with IPsec in lookaside mode. 518 519* cryptodev: The field ``dataunit_len`` of the ``struct rte_crypto_cipher_xform`` 520 moved to the end of the structure and extended to ``uint32_t``. 521 522* cryptodev: The structure ``rte_crypto_vec`` was updated to add ``tot_len`` 523 field to support total buffer length to facilitate protocol offload case. 524 525* cryptodev: The structure ``rte_crypto_sym_vec`` was updated to add 526 ``dest_sgl`` to support out of place processing. 527 528* bbdev: Added device info related to data byte endianness processing. 529 530* eventdev: Moved memory used by timer adapters to hugepage. This will prevent 531 TLB misses if any and aligns to memory structure of other subsystems. 532 533* fib: Added the ``rib_ext_sz`` field to ``rte_fib_conf`` and ``rte_fib6_conf`` 534 so that user can specify the size of the RIB extension inside the FIB. 535 536* ip_frag: All macros updated to have ``RTE_IP_FRAG_`` prefix. 537 Obsolete macros are kept for compatibility. 538 DPDK components updated to use new names. 539 Experimental function ``rte_frag_table_del_expired_entries()`` was renamed 540 to ``rte_ip_frag_table_del_expired_entries()`` 541 to comply with other public API naming convention. 542 543 544ABI Changes 545----------- 546 547* ethdev: All enums and macros updated to have ``RTE_ETH`` prefix and structures 548 updated to have ``rte_eth`` prefix. DPDK components updated to use new names. 549 550* ethdev: The input parameters for ``eth_rx_queue_count_t`` were changed. 551 Instead of a pointer to ``rte_eth_dev`` and queue index, it now accepts a pointer 552 to internal queue data as an input parameter. While this change is transparent 553 to the user, it still counts as an ABI change, as ``eth_rx_queue_count_t`` 554 is used by the public inline function ``rte_eth_rx_queue_count``. 555 556* ethdev: Made ``rte_eth_dev``, ``rte_eth_dev_data``, ``rte_eth_rxtx_callback`` 557 private data structures. ``rte_eth_devices[]`` can't be accessed directly 558 by user any more. While it is an ABI breakage, this change is intended 559 to be transparent for both users (no changes in user app is required) and 560 PMD developers (no changes in PMD is required). 561 562* vhost: rename ``struct vhost_device_ops`` to ``struct rte_vhost_device_ops``. 563 564* cryptodev: Made ``rte_cryptodev``, ``rte_cryptodev_data`` private 565 structures internal to DPDK. ``rte_cryptodevs`` can't be accessed directly 566 by user any more. While it is an ABI breakage, this change is intended 567 to be transparent for both users (no changes in user app is required) and 568 PMD developers (no changes in PMD is required). 569 570* security: ``rte_security_set_pkt_metadata`` and ``rte_security_get_userdata`` 571 routines used by inline outbound and inline inbound security processing were 572 made inline and enhanced to do simple 64-bit set/get for PMDs that do not 573 have much processing in PMD specific callbacks but just 64-bit set/get. 574 This avoids a per packet function pointer jump overhead for such PMDs. 575 576* security: A new option ``iv_gen_disable`` was added in structure 577 ``rte_security_ipsec_sa_options`` to disable IV generation inside PMD, 578 so that application can provide its own IV and test known test vectors. 579 580* security: A new option ``tunnel_hdr_verify`` was added in structure 581 ``rte_security_ipsec_sa_options`` to indicate whether outer header 582 verification need to be done as part of inbound IPsec processing. 583 584* security: A new option ``udp_ports_verify`` was added in structure 585 ``rte_security_ipsec_sa_options`` to indicate whether UDP ports 586 verification need to be done as part of inbound IPsec processing. 587 588* security: A new structure ``rte_security_ipsec_lifetime`` was added to 589 replace ``esn_soft_limit`` in IPsec configuration structure 590 ``rte_security_ipsec_xform`` to allow applications to configure SA soft 591 and hard expiry limits. Limits can be either in number of packets or bytes. 592 593* security: The new options ``ip_csum_enable`` and ``l4_csum_enable`` were added 594 in structure ``rte_security_ipsec_sa_options`` to indicate whether inner 595 packet IPv4 header checksum and L4 checksum need to be offloaded to 596 security device. 597 598* security: A new structure ``esn`` was added in structure 599 ``rte_security_ipsec_xform`` to set an initial ESN value. This permits 600 applications to start from an arbitrary ESN value for debug and SA lifetime 601 enforcement purposes. 602 603* security: A new structure ``udp`` was added in structure 604 ``rte_security_ipsec_xform`` to allow setting the source and destination ports 605 for UDP encapsulated IPsec traffic. 606 607* bbdev: Added capability related to more comprehensive CRC options, 608 shifting values of the ``enum rte_bbdev_op_ldpcdec_flag_bitmasks``. 609 610* eventdev: New variables ``rx_event_buf_count`` and ``rx_event_buf_size`` 611 were added in structure ``rte_event_eth_rx_adapter_stats`` to get additional 612 status. 613 614* eventdev: A new structure ``rte_event_fp_ops`` has been added which is now used 615 by the fastpath inline functions. The structures ``rte_eventdev``, 616 ``rte_eventdev_data`` have been made internal. ``rte_eventdevs[]`` can't be 617 accessed directly by user any more. This change is transparent to both 618 applications and PMDs. 619 620* eventdev: Re-arranged fields in ``rte_event_timer`` to remove holes. 621 622* ip_frag: Increased default value for config parameter 623 ``RTE_LIBRTE_IP_FRAG_MAX_FRAG`` from ``4`` to ``8``. 624 This parameter controls maximum number of fragments per packet 625 in IP reassembly table. Increasing this value from ``4`` to ``8`` 626 will allow covering the common case with jumbo packet size of ``9000B`` 627 and fragments with default frame size ``(1500B)``. 628 629 630Tested Platforms 631---------------- 632 633* Intel\ |reg| platforms with Intel\ |reg| NICs combinations 634 635 * CPU 636 637 * Intel\ |reg| Atom\ |trade| CPU C3758 @ 2.20GHz 638 * Intel\ |reg| Xeon\ |reg| CPU E5-2680 v2 @ 2.80GHz 639 * Intel\ |reg| Xeon\ |reg| CPU E5-2699 v3 @ 2.30GHz 640 * Intel\ |reg| Xeon\ |reg| CPU E5-2699 v4 @ 2.20GHz 641 * Intel\ |reg| Xeon\ |reg| Gold 6140M CPU @ 2.30GHz 642 * Intel\ |reg| Xeon\ |reg| Gold 6139 CPU @ 2.30GHz 643 * Intel\ |reg| Xeon\ |reg| Gold 6252N CPU @ 2.30GHz 644 * Intel\ |reg| Xeon\ |reg| Gold 6348 CPU @ 2.60GHz 645 * Intel\ |reg| Xeon\ |reg| Platinum 8180M CPU @ 2.50GHz 646 * Intel\ |reg| Xeon\ |reg| Platinum 8280M CPU @ 2.70GHz 647 648 * OS: 649 650 * Fedora 34 651 * OpenWRT 21.02.0 652 * FreeBSD 13.0 653 * Red Hat Enterprise Linux Server release 8.4 654 * Suse 15 SP3 655 * Ubuntu 20.04.3 656 * Ubuntu 21.10 657 658 * NICs: 659 660 * Intel\ |reg| Ethernet Controller E810-C for SFP (4x25G) 661 662 * Firmware version: 3.10 0x8000aa86 1.3100.0 663 * Device id (pf/vf): 8086:1593 / 8086:1889 664 * Driver version: 1.7.11_7_g444e5edb (ice) 665 * OS Default DDP: 1.3.27.0 666 * COMMS DDP: 1.3.31.0 667 * Wireless Edge DDP: 1.3.7.0 668 669 * Intel\ |reg| Ethernet Controller E810-C for QSFP (2x100G) 670 671 * Firmware version: 3.10 0x8000aa66 1.3100.0 672 * Device id (pf/vf): 8086:1592 / 8086:1889 673 * Driver version: 1.7.11_7_g444e5edb (ice) 674 * OS Default DDP: 1.3.27.0 675 * COMMS DDP: 1.3.31.0 676 * Wireless Edge DDP: 1.3.7.0 677 678 * Intel\ |reg| 82599ES 10 Gigabit Ethernet Controller 679 680 * Firmware version: 0x61bf0001 681 * Device id (pf/vf): 8086:10fb / 8086:10ed 682 * Driver version(in-tree): 5.1.0-k (ixgbe) 683 * Driver version(out-tree): 5.13.4 (ixgbe) 684 685 * Intel\ |reg| Ethernet Converged Network Adapter X710-DA4 (4x10G) 686 687 * PF Firmware version: 8.30 0x8000a49d 1.2926.0 688 * VF Firmware version: 8.50 0x8000b6d9 1.3082.0 689 * Device id (pf/vf): 8086:1572 / 8086:154c 690 * Driver version: 2.17.4 (i40e) 691 692 * Intel\ |reg| Corporation Ethernet Connection X722 for 10GbE SFP+ (4x10G) 693 694 * Firmware version: 5.30 0x80002a29 1.2926.0 695 * Device id (pf/vf): 8086:37d0 / 8086:37cd 696 * Driver version: 2.17.4 (i40e) 697 698 * Intel\ |reg| Corporation Ethernet Connection X722 for 10GBASE-T (2x10G) 699 700 * Firmware version: 5.40 0x80002e2f 1.2935.0 701 * Device id (pf/vf): 8086:37d2 / 8086:37cd 702 * Driver version: 2.17.4 (i40e) 703 704 * Intel\ |reg| Ethernet Converged Network Adapter XXV710-DA2 (2x25G) 705 706 * PF Firmware version: 8.30 0x8000a483 1.2926.0 707 * VF Firmware version: 8.50 0x8000b703 1.3082.0 708 * Device id (pf/vf): 8086:158b / 8086:154c 709 * Driver version: 2.17.4 (i40e) 710 711 * Intel\ |reg| Ethernet Converged Network Adapter XL710-QDA2 (2X40G) 712 713 * PF Firmware version: 8.30 0x8000a4ae 1.2926.0 714 * VF Firmware version: 8.50 0x8000b6c7 1.3082.0 715 * Device id (pf/vf): 8086:1583 / 8086:154c 716 * Driver version: 2.17.4 (i40e) 717 718 * Intel\ |reg| Ethernet Converged Network Adapter X710-T2L 719 720 * Firmware version: 8.30 0x8000a489 1.2879.0 721 * Device id (pf): 8086:15ff 722 * Driver version: 2.17.4 (i40e) 723 724* Intel\ |reg| platforms with Mellanox\ |reg| NICs combinations 725 726 * CPU: 727 728 * Intel\ |reg| Xeon\ |reg| Gold 6154 CPU @ 3.00GHz 729 * Intel\ |reg| Xeon\ |reg| CPU E5-2697A v4 @ 2.60GHz 730 * Intel\ |reg| Xeon\ |reg| CPU E5-2697 v3 @ 2.60GHz 731 * Intel\ |reg| Xeon\ |reg| CPU E5-2680 v2 @ 2.80GHz 732 * Intel\ |reg| Xeon\ |reg| CPU E5-2670 0 @ 2.60GHz 733 * Intel\ |reg| Xeon\ |reg| CPU E5-2650 v4 @ 2.20GHz 734 * Intel\ |reg| Xeon\ |reg| CPU E5-2650 v3 @ 2.30GHz 735 * Intel\ |reg| Xeon\ |reg| CPU E5-2640 @ 2.50GHz 736 * Intel\ |reg| Xeon\ |reg| CPU E5-2650 0 @ 2.00GHz 737 * Intel\ |reg| Xeon\ |reg| CPU E5-2620 v4 @ 2.10GHz 738 739 * OS: 740 741 * Red Hat Enterprise Linux release 8.2 (Ootpa) 742 * Red Hat Enterprise Linux Server release 7.8 (Maipo) 743 * Red Hat Enterprise Linux Server release 7.6 (Maipo) 744 * Red Hat Enterprise Linux Server release 7.5 (Maipo) 745 * Red Hat Enterprise Linux Server release 7.4 (Maipo) 746 * Red Hat Enterprise Linux Server release 7.3 (Maipo) 747 * Red Hat Enterprise Linux Server release 7.2 (Maipo) 748 * Ubuntu 20.04 749 * Ubuntu 18.04 750 * Ubuntu 16.04 751 * SUSE Enterprise Linux 15 SP2 752 * SUSE Enterprise Linux 12 SP4 753 754 * OFED: 755 756 * MLNX_OFED 5.5-0.5.9.0 and above 757 * MLNX_OFED 5.4-3.1.0.0 758 759 * upstream kernel: 760 761 * Linux 5.16.0-rc2 and above 762 763 * rdma-core: 764 765 * rdma-core-37.1 and above 766 767 * NICs: 768 769 * Mellanox\ |reg| ConnectX\ |reg|-3 Pro 40G MCX354A-FCC_Ax (2x40G) 770 771 * Host interface: PCI Express 3.0 x8 772 * Device ID: 15b3:1007 773 * Firmware version: 2.42.5000 774 775 * Mellanox\ |reg| ConnectX\ |reg|-3 Pro 40G MCX354A-FCCT (2x40G) 776 777 * Host interface: PCI Express 3.0 x8 778 * Device ID: 15b3:1007 779 * Firmware version: 2.42.5000 780 781 * Mellanox\ |reg| ConnectX\ |reg|-4 Lx 25G MCX4121A-ACAT (2x25G) 782 783 * Host interface: PCI Express 3.0 x8 784 * Device ID: 15b3:1015 785 * Firmware version: 14.32.0570 and above 786 787 * Mellanox\ |reg| ConnectX\ |reg|-4 Lx 50G MCX4131A-GCAT (1x50G) 788 789 * Host interface: PCI Express 3.0 x8 790 * Device ID: 15b3:1015 791 * Firmware version: 14.32.0570 and above 792 793 * Mellanox\ |reg| ConnectX\ |reg|-5 100G MCX516A-CCAT (2x100G) 794 795 * Host interface: PCI Express 3.0 x16 796 * Device ID: 15b3:1017 797 * Firmware version: 16.32.0570 and above 798 799 * Mellanox\ |reg| ConnectX\ |reg|-5 100G MCX556A-ECAT (2x100G) 800 801 * Host interface: PCI Express 3.0 x16 802 * Device ID: 15b3:1017 803 * Firmware version: 16.32.0570 and above 804 805 * Mellanox\ |reg| ConnectX\ |reg|-5 100G MCX556A-EDAT (2x100G) 806 807 * Host interface: PCI Express 3.0 x16 808 * Device ID: 15b3:1017 809 * Firmware version: 16.32.0570 and above 810 811 * Mellanox\ |reg| ConnectX\ |reg|-5 Ex EN 100G MCX516A-CDAT (2x100G) 812 813 * Host interface: PCI Express 4.0 x16 814 * Device ID: 15b3:1019 815 * Firmware version: 16.32.0570 and above 816 817 * Mellanox\ |reg| ConnectX\ |reg|-6 Dx EN 100G MCX623106AN-CDAT (2x100G) 818 819 * Host interface: PCI Express 4.0 x16 820 * Device ID: 15b3:101d 821 * Firmware version: 22.32.0570 and above 822 823 * Mellanox\ |reg| ConnectX\ |reg|-6 Lx EN 25G MCX631102AN-ADAT (2x25G) 824 825 * Host interface: PCI Express 4.0 x8 826 * Device ID: 15b3:101f 827 * Firmware version: 26.32.0570 and above 828 829* Mellanox\ |reg| BlueField\ |reg| SmartNIC 830 831 * Mellanox\ |reg| BlueField\ |reg|-2 SmartNIC MT41686 - MBF2H332A-AEEOT_A1 (2x25G) 832 833 * Host interface: PCI Express 3.0 x16 834 * Device ID: 15b3:a2d6 835 * Firmware version: 24.32.0570 and above 836 837 * Embedded software: 838 839 * Ubuntu 20.04.3 840 * MLNX_OFED 5.5-0.5.8 and above 841 * DPDK application running on Arm cores 842 843* IBM Power 9 platforms with Mellanox\ |reg| NICs combinations 844 845 * CPU: 846 847 * POWER9 2.2 (pvr 004e 1202) 2300MHz 848 849 * OS: 850 851 * Red Hat Enterprise Linux Server release 7.6 852 853 * NICs: 854 855 * Mellanox\ |reg| ConnectX\ |reg|-5 100G MCX556A-ECAT (2x100G) 856 857 * Host interface: PCI Express 4.0 x16 858 * Device ID: 15b3:1017 859 * Firmware version: 16.32.0560 860 861 * Mellanox\ |reg| ConnectX\ |reg|-6 Dx 100G MCX623106AN-CDAT (2x100G) 862 863 * Host interface: PCI Express 4.0 x16 864 * Device ID: 15b3:101d 865 * Firmware version: 22.32.0560 866 867 * OFED: 868 869 * MLNX_OFED 5.5-0.5.9.0 870 871* NXP ARM SoCs (with integrated NICs) 872 873 * SoC: 874 875 * NXP i.MX 8M Mini with ARM Cortex A53, Cortex M4 876 877 * OS (Based on NXP LF support packages): 878 879 * Kernel version: 5.10 880 * Ubuntu 18.04 881