1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2019 The DPDK contributors
4 .. include:: <isonum.txt>
12 * **EAL will now pick IOVA as VA mode as the default in most cases.**
14 Previously, the preferred default IOVA mode was selected to be IOVA as PA. The
15 behavior has now been changed to handle IOVA mode detection in a more complex
16 manner, and will default to IOVA as VA in most cases.
20 MCS lock provides scalability by spinning on a CPU/thread local variable
21 which avoids expensive cache bouncing.
22 It provides fairness by maintaining a list of acquirers and passing
23 the lock to each CPU/thread in the order they acquired the lock.
25 * **Updated the EAL Pseudo-random Number Generator.**
27 The ``lrand48()`` based ``rte_rand()`` function is replaced with a
28 DPDK-native combined Linear Feedback Shift Register (LFSR)
29 pseudo-random number generator (PRNG).
31 This new PRNG implementation is multi-thread safe, provides
32 higher-quality pseudo-random numbers (including full 64 bit
33 support) and improved performance.
35 In addition, ``<rte_random.h>`` is extended with a new function
36 ``rte_rand_max()`` which supplies unbiased, bounded pseudo-random
39 * **Updated the Broadcom bnxt PMD.**
41 Updated the Broadcom bnxt PMD. The major enhancements include:
43 * Performance optimizations in non-vector Tx path.
44 * Added support for SSE vector mode.
45 * Updated HWRM API to version 1.10.0.91.
47 * **Added support for Broadcom NetXtreme-E BCM57500 Ethernet controllers.**
49 Added support to the Broadcom bnxt PMD for the BCM57500 (a.k.a. "Thor") family
50 of Ethernet controllers. These controllers support link speeds up to
51 200Gbps, 50G PAM-4, and PCIe 4.0.
53 * **Added Huawei hinic PMD.**
55 Added the new ``hinic`` net driver for Huawei Intelligent PCIE Network
56 Adapters based on the Huawei Ethernet Controller Hi1822.
57 See the :doc:`../nics/hinic` guide for more details on this new driver.
59 * **Updated the Intel ice driver.**
61 Updated the Intel ice driver with new features and improvements, including:
63 * Enabled Tx outer/inner L3/L4 checksum offload.
64 * Enabled generic filter framework and supported switch filter.
65 * Supported UDP tunnel port add.
67 * **Updated the Intel i40e driver.**
69 Updated tje Intel i40e driver with new features and improvements, including:
71 * Added support for MARK + RSS action in rte_flow (non-vector RX path only)
73 * **Updated Mellanox mlx5 driver.**
75 Updated Mellanox mlx5 driver with new features and improvements, including:
77 * Updated the packet header modification feature. Added support of TCP header
78 sequence number and acknowledgment number modification.
79 * Added support for match on ICMP/ICMP6 code and type.
80 * Added support for matching on GRE's key and C,K,S present bits.
81 * Added support for IP-in-IP tunnel.
82 * Accelerated flows with count action creation and destroy.
83 * Accelerated flows counter query.
84 * Improved Tx datapath performance with enabled HW offloads.
85 * Added support for LRO.
87 * **Updated Solarflare network PMD.**
89 Updated the Solarflare ``sfc_efx`` driver with changes including:
91 * Added support for Rx interrupts.
93 * **Added memif PMD.**
95 Added a new Shared Memory Packet Interface (``memif``) PMD.
96 See the :doc:`../nics/memif` guide for more details on this new driver.
98 * **Updated the AF_XDP PMD.**
100 Updated the AF_XDP PMD. The new features include:
102 * Enabled zero copy through mbuf's external memory mechanism to achieve
104 * Added multi-queue support to allow one af_xdp vdev with multiple netdev
106 * Enabled "need_wakeup" feature which can provide efficient support for the
107 use case where the application and driver executing on the same core.
109 * **Enabled infinite Rx in the PCAP PMD.**
111 Added an infinite Rx feature to the PCAP PMD which allows packets in the Rx
112 PCAP to be received repeatedly at a high rate. This can be useful for quick
113 performance testing of DPDK apps.
115 * **Enabled receiving no packet in the PCAP PMD.**
117 Added function to allow users to run the PCAP PMD without receiving any
118 packets on PCAP Rx. When the function is called, a dummy queue is created
119 for each Tx queue argument passed.
121 * **Added a FPGA_LTE_FEC bbdev PMD.**
123 Added a new ``fpga_lte_fec`` bbdev driver for the Intel\ |reg| FPGA PAC
124 (Programmable Acceleration Card) N3000. See the
125 :doc:`../bbdevs/fpga_lte_fec` BBDEV guide for more details on this new driver.
127 * **Updated the TURBO_SW bbdev PMD.**
129 Updated the ``turbo_sw`` bbdev driver with changes including:
131 * Added option to build the driver with or without dependency of external
133 * Added support for 5GNR encode/decode operations.
135 * **Updated the Intel QuickAssist Technology (QAT) symmetric crypto PMD.**
137 Added support for digest-encrypted cases where digest is appended
140 * **Added the Intel QuickData Technology PMD.**
142 Added a PMD for the Intel\ |reg| QuickData Technology, part of
143 Intel\ |reg| I/O Acceleration Technology `(Intel I/OAT)
144 <https://www.intel.com/content/www/us/en/wireless-network/accel-technology.html>`_,
145 which allows data copies to be done by hardware instead
146 of via software, reducing cycles spent copying large blocks of data in
149 * **Added Marvell OCTEON TX2 drivers.**
151 Added the new ``ethdev``, ``eventdev``, ``mempool``, ``eventdev Rx adapter``,
152 ``eventdev Tx adapter``, ``eventdev Timer adapter`` and ``rawdev DMA``
153 drivers for various HW co-processors available in ``OCTEON TX2`` SoC.
155 See ``platform/octeontx2`` and driver information:
158 * ``mempool/octeontx2``
159 * ``eventdevs/octeontx2``
160 * ``rawdevs/octeontx2_dma``
162 * **Introduced the Intel NTB PMD.**
164 Added a PMD for Intel NTB (Non-transparent Bridge). This PMD implements
165 a handshake between two separate hosts and can share local memory for peer
166 host to directly access.
168 * **Updated the IPSec library and IPsec Security Gateway application.**
170 Added the following features to ``librte_ipsec``. Corresponding changes are
171 also added in the ``ipsec-secgw`` sample application.
173 * ECN and DSCP field header reconstruction as per RFC4301.
174 * Transport mode with IPv6 extension headers.
175 * Support packets with multiple segments.
177 * **Updated telemetry library for global metrics support.**
179 Updated ``librte_telemetry`` to fetch the global metrics from the
180 ``librte_metrics`` library.
182 * **Added new telemetry mode for l3fwd-power application.**
184 Added a telemetry mode to the ``l3fwd-power`` application to report
185 application level busyness, empty and full polls of ``rte_eth_rx_burst()``.
187 * **Updated the pdump application.**
189 Add support for pdump to exit with primary process.
191 * **Updated test-compress-perf tool application.**
193 Added a multiple cores feature to the compression perf tool application.
199 * Removed KNI ethtool, ``CONFIG_RTE_KNI_KMOD_ETHTOOL``, support.
201 * build: armv8 crypto extension is disabled.
207 * The ``rte_mem_config`` structure has been made private. New accessor
208 ``rte_mcfg_*`` functions were introduced to provide replacement for direct
209 access to the shared mem config.
211 * The network structures, definitions and functions have
212 been prefixed by ``rte_`` to resolve conflicts with libc headers.
214 * malloc: The function ``rte_malloc_set_limit()`` was never implemented.
215 It is deprecated and will be removed in a future release.
217 * cryptodev: the ``uint8_t *data`` member of the ``key`` structure in the xforms
218 structure (``rte_crypto_cipher_xform``, ``rte_crypto_auth_xform``, and
219 ``rte_crypto_aead_xform``) have been changed to ``const uint8_t *data``.
221 * eventdev: No longer marked as experimental.
223 The eventdev functions are no longer marked as experimental, and have
224 become part of the normal DPDK API and ABI. Any future ABI changes will be
225 announced at least one release before the ABI change is made. There are no
226 ABI breaking changes planned.
228 * ip_frag: The IP fragmentation library converts input mbuf into fragments
229 using input MTU size via the ``rte_ipv4_fragment_packet()`` interface.
230 Once fragmentation is done, each ``mbuf->ol_flags`` are set to enable IP
231 checksum H/W offload irrespective of the platform capability.
232 Cleared IP checksum H/W offload flag from the library. The application must
233 set this flag if it is supported by the platform and application wishes to
236 * ip_frag: IP reassembly library converts the list of fragments into a
237 reassembled packet via ``rte_ipv4_frag_reassemble_packet()`` interface.
238 Once reassembly is done, ``mbuf->ol_flags`` are set to enable IP checksum H/W
239 offload irrespective of the platform capability. Cleared IP checksum H/W
240 offload flag from the library. The application must set this flag if it is
241 supported by the platform and application wishes to use it.
243 * sched: Macros ``RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS`` and
244 ``RTE_SCHED_PIPE_PROFILES_PER_PORT`` are removed for flexible configuration
245 of pipe traffic classes and their queues size, and for runtime configuration
246 of the maximum number of pipe profiles, respectively. In addition, the
247 ``wrr_weights`` field of struct ``rte_sched_pipe_params`` is modified to be
248 used only for best-effort tc, and the ``qsize`` field of struct
249 ``rte_sched_port_params`` is changed to allow different sizes for each
256 * eventdev: Event based Rx adapter callback
258 The mbuf pointer array in the event eth Rx adapter callback
259 has been replaced with an event array. Using
260 an event array allows the application to change attributes
261 of the events enqueued by the SW adapter.
263 The callback can drop packets and populate
264 a callback argument with the number of dropped packets.
265 Add a Rx adapter stats field to keep track of the total
266 number of dropped packets.
268 * cryptodev: New member in ``rte_cryptodev_config`` to allow applications to
269 disable features supported by the crypto device. Only the following features
270 would be allowed to be disabled this way,
272 - ``RTE_CRYPTODEV_FF_SYMMETRIC_CRYPTO``.
273 - ``RTE_CRYPTODEV_FF_ASYMMETRIC_CRYPTO``.
274 - ``RTE_CRYPTODEV_FF_SECURITY``.
276 Disabling unused features would facilitate efficient usage of HW/SW offload.
278 * bbdev: New operations and parameters have been added to support new 5GNR
279 operations. The bbdev ABI is still kept experimental.
281 * rawdev: The driver names have been changed to ``librte_rawdev_*``.
282 Now they all have the same prefix, and same name with make and meson builds.
285 Shared Library Versions
286 -----------------------
288 The libraries prepended with a plus sign were incremented in this version.
294 librte_bitratestats.so.2
297 librte_bus_fslmc.so.2
298 librte_bus_ifpga.so.2
301 librte_bus_vmbus.so.2
304 librte_compressdev.so.1
305 + librte_cryptodev.so.8
306 librte_distributor.so.1
310 + librte_eventdev.so.7
311 librte_flow_classify.so.1
320 librte_latencystats.so.1
334 librte_pmd_ixgbe.so.2
335 librte_pmd_dpaa2_qdma.so.1
337 librte_pmd_softnic.so.1
338 librte_pmd_vhost.so.2
356 * **Unsuitable IOVA mode may be picked as the default.**
358 Not all kernel drivers and not all devices support all IOVA modes. EAL will
359 attempt to pick a reasonable default based on a number of factors, but
360 there may be cases where the default is unsuitable.
362 It is recommended to use the `--iova-mode` command-line parameter if the
363 default is not suitable.
369 * Intel(R) platforms with Intel(R) NICs combinations
373 * Intel(R) Atom(TM) CPU C3758 @ 2.20GHz
374 * Intel(R) Xeon(R) CPU D-1541 @ 2.10GHz
375 * Intel(R) Xeon(R) CPU D-1553N @ 2.30GHz
376 * Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
377 * Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz
378 * Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz
379 * Intel(R) Xeon(R) Gold 6139 CPU @ 2.30GHz
380 * Intel(R) Xeon(R) Platinum 8180 CPU @ 2.50GHz
381 * Intel(R) Xeon(R) Platinum 8280M CPU @ 2.70GHz
388 * Red Hat Enterprise Linux Server release 8.0
389 * Red Hat Enterprise Linux Server release 7.6
398 * Intel(R) 82599ES 10 Gigabit Ethernet Controller
400 * Firmware version: 0x61bf0001
401 * Device id (pf/vf): 8086:10fb / 8086:10ed
402 * Driver version: 5.6.1 (ixgbe)
404 * Intel(R) Corporation Ethernet Connection X552/X557-AT 10GBASE-T
406 * Firmware version: 0x800003e7
407 * Device id (pf/vf): 8086:15ad / 8086:15a8
408 * Driver version: 5.1.0 (ixgbe)
410 * Intel Corporation Ethernet Controller 10G X550T
412 * Firmware version: 0x80000482
413 * Device id (pf): 8086:1563
414 * Driver version: 5.6.1 (ixgbe)
416 * Intel(R) Ethernet Converged Network Adapter X710-DA4 (4x10G)
418 * Firmware version: 7.00 0x80004cdb
419 * Device id (pf/vf): 8086:1572 / 8086:154c
420 * Driver version: 2.9.21 (i40e)
422 * Intel(R) Corporation Ethernet Connection X722 for 10GbE SFP+ (4x10G)
424 * Firmware version: 4.10 0x80001a3c
425 * Device id (pf/vf): 8086:37d0 / 8086:37cd
426 * Driver version: 2.9.21 (i40e)
428 * Intel(R) Ethernet Converged Network Adapter XXV710-DA2 (2x25G)
430 * Firmware version: 7.00 0x80004cf8
431 * Device id (pf/vf): 8086:158b / 8086:154c
432 * Driver version: 2.9.21 (i40e)
434 * Intel(R) Ethernet Converged Network Adapter XL710-QDA2 (2X40G)
436 * Firmware version: 7.00 0x80004c97
437 * Device id (pf/vf): 8086:1583 / 8086:154c
438 * Driver version: 2.9.21 (i40e)
440 * Intel(R) Corporation I350 Gigabit Network Connection
442 * Firmware version: 1.63, 0x80000cbc
443 * Device id (pf/vf): 8086:1521 / 8086:1520
444 * Driver version: 5.4.0-k (igb)
446 * Intel Corporation I210 Gigabit Network Connection
448 * Firmware version: 3.25, 0x800006eb
449 * Device id (pf): 8086:1533
450 * Driver version: 5.4.0-k(igb)
452 * Intel(R) platforms with Mellanox(R) NICs combinations
456 * Intel(R) Xeon(R) Gold 6154 CPU @ 3.00GHz
457 * Intel(R) Xeon(R) CPU E5-2697A v4 @ 2.60GHz
458 * Intel(R) Xeon(R) CPU E5-2697 v3 @ 2.60GHz
459 * Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
460 * Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz
461 * Intel(R) Xeon(R) CPU E5-2640 @ 2.50GHz
462 * Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz
466 * Red Hat Enterprise Linux Server release 7.6 (Maipo)
467 * Red Hat Enterprise Linux Server release 7.5 (Maipo)
468 * Red Hat Enterprise Linux Server release 7.4 (Maipo)
469 * Red Hat Enterprise Linux Server release 7.3 (Maipo)
470 * Red Hat Enterprise Linux Server release 7.2 (Maipo)
475 * SUSE Linux Enterprise Server 15
479 * MLNX_OFED 4.6-1.0.1.1
480 * MLNX_OFED 4.6-4.1.2.0
484 * Mellanox(R) ConnectX(R)-3 Pro 40G MCX354A-FCC_Ax (2x40G)
486 * Host interface: PCI Express 3.0 x8
487 * Device ID: 15b3:1007
488 * Firmware version: 2.42.5000
490 * Mellanox(R) ConnectX(R)-4 10G MCX4111A-XCAT (1x10G)
492 * Host interface: PCI Express 3.0 x8
493 * Device ID: 15b3:1013
494 * Firmware version: 12.25.6406 and above
496 * Mellanox(R) ConnectX(R)-4 10G MCX4121A-XCAT (2x10G)
498 * Host interface: PCI Express 3.0 x8
499 * Device ID: 15b3:1013
500 * Firmware version: 12.25.6406 and above
502 * Mellanox(R) ConnectX(R)-4 25G MCX4111A-ACAT (1x25G)
504 * Host interface: PCI Express 3.0 x8
505 * Device ID: 15b3:1013
506 * Firmware version: 12.25.6406 and above
508 * Mellanox(R) ConnectX(R)-4 25G MCX4121A-ACAT (2x25G)
510 * Host interface: PCI Express 3.0 x8
511 * Device ID: 15b3:1013
512 * Firmware version: 12.25.6406 and above
514 * Mellanox(R) ConnectX(R)-4 40G MCX4131A-BCAT/MCX413A-BCAT (1x40G)
516 * Host interface: PCI Express 3.0 x8
517 * Device ID: 15b3:1013
518 * Firmware version: 12.25.6406 and above
520 * Mellanox(R) ConnectX(R)-4 40G MCX415A-BCAT (1x40G)
522 * Host interface: PCI Express 3.0 x16
523 * Device ID: 15b3:1013
524 * Firmware version: 12.25.6406 and above
526 * Mellanox(R) ConnectX(R)-4 50G MCX4131A-GCAT/MCX413A-GCAT (1x50G)
528 * Host interface: PCI Express 3.0 x8
529 * Device ID: 15b3:1013
530 * Firmware version: 12.25.6406 and above
532 * Mellanox(R) ConnectX(R)-4 50G MCX414A-BCAT (2x50G)
534 * Host interface: PCI Express 3.0 x8
535 * Device ID: 15b3:1013
536 * Firmware version: 12.25.6406 and above
538 * Mellanox(R) ConnectX(R)-4 50G MCX415A-GCAT/MCX416A-BCAT/MCX416A-GCAT (2x50G)
540 * Host interface: PCI Express 3.0 x16
541 * Device ID: 15b3:1013
542 * Firmware version: 12.25.6406 and above
543 * Firmware version: 12.25.6406 and above
545 * Mellanox(R) ConnectX(R)-4 50G MCX415A-CCAT (1x100G)
547 * Host interface: PCI Express 3.0 x16
548 * Device ID: 15b3:1013
549 * Firmware version: 12.25.6406 and above
551 * Mellanox(R) ConnectX(R)-4 100G MCX416A-CCAT (2x100G)
553 * Host interface: PCI Express 3.0 x16
554 * Device ID: 15b3:1013
555 * Firmware version: 12.25.6406 and above
557 * Mellanox(R) ConnectX(R)-4 Lx 10G MCX4121A-XCAT (2x10G)
559 * Host interface: PCI Express 3.0 x8
560 * Device ID: 15b3:1015
561 * Firmware version: 14.25.6406 and above
563 * Mellanox(R) ConnectX(R)-4 Lx 25G MCX4121A-ACAT (2x25G)
565 * Host interface: PCI Express 3.0 x8
566 * Device ID: 15b3:1015
567 * Firmware version: 14.25.6406 and above
569 * Mellanox(R) ConnectX(R)-5 100G MCX556A-ECAT (2x100G)
571 * Host interface: PCI Express 3.0 x16
572 * Device ID: 15b3:1017
573 * Firmware version: 16.25.6406 and above
575 * Mellanox(R) ConnectX(R)-5 Ex EN 100G MCX516A-CDAT (2x100G)
577 * Host interface: PCI Express 4.0 x16
578 * Device ID: 15b3:1019
579 * Firmware version: 16.25.6406 and above
581 * Mellanox(R) BlueField SmartNIC
583 * Mellanox(R) BlueField SmartNIC MT416842 (2x25G)
585 * Host interface: PCI Express 3.0 x16
586 * Device ID: 15b3:a2d2
587 * Firmware version: 18.25.6600
589 * SoC Arm cores running OS:
591 * CentOS Linux release 7.5.1804 (AltArch)
592 * MLNX_OFED 4.6-3.5.8.0
594 * DPDK application running on Arm cores inside SmartNIC
596 * IBM Power 9 platforms with Mellanox(R) NICs combinations
600 * POWER9 2.2 (pvr 004e 1202) 2300MHz
604 * Ubuntu 18.04.1 LTS (Bionic Beaver)
608 * Mellanox(R) ConnectX(R)-5 100G MCX556A-ECAT (2x100G)
610 * Host interface: PCI Express 3.0 x16
611 * Device ID: 15b3:1017
612 * Firmware version: 16.25.6406
616 * MLNX_OFED 4.6-4.1.2.0
618 * ARMv8 SoC combinations from Marvell (with integrated NICs)
622 * CN83xx, CN96xx, CNF95xx, CN93xx
624 * OS (Based on Marvell OCTEON TX SDK 10.0):