1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2019 The DPDK contributors
4 .. include:: <isonum.txt>
9 .. **Read this first.**
11 The text in the sections below explains how to update the release notes.
13 Use proper spelling, capitalization and punctuation in all sections.
15 Variable and config names should be quoted as fixed width text:
18 Build the docs and view the output file to ensure the changes are correct::
22 xdg-open build/doc/html/guides/rel_notes/release_19_11.html
28 .. This section should contain new features added in this release.
31 * **Add a title in the past tense with a full stop.**
33 Add a short 1-2 sentence description in the past tense.
34 The description should be enough to allow someone scanning
35 the release notes to understand the new feature.
37 If the feature adds a lot of sub-features you can use a bullet list
40 * Added feature foo to do something.
41 * Enhanced feature bar to do something else.
43 Refer to the previous release notes for examples.
45 Suggested order in release notes items:
46 * Core libs (EAL, mempool, ring, mbuf, buses)
47 * Device abstraction libs and PMDs
49 - cryptodev (lib, PMDs)
50 - eventdev (lib, PMDs)
53 * Apps, Examples, Tools (if significant)
55 This section is a comment. Do not overwrite or remove it.
56 Also, make sure to start the actual text at the margin.
57 =========================================================
59 * **FreeBSD now supports `--base-virtaddr` EAL option.**
61 FreeBSD version now also supports setting base virtual address for mapping
62 pages and resources into its address space.
64 * **Added Lock-free Stack for aarch64.**
66 The lock-free stack implementation is enabled for aarch64 platforms.
68 * **Extended pktmbuf mempool private structure.**
70 rte_pktmbuf_pool_private structure was extended to include flags field
71 for future compatibility.
72 As per 19.11 release this field is reserved and should be set to 0
75 * **Changed mempool allocation behaviour.**
77 Objects are no longer across pages by default.
78 It may consume more memory when using small memory pages.
80 * **Added support of dynamic fields and flags in mbuf.**
82 This new feature adds the ability to dynamically register some room
83 for a field or a flag in the mbuf structure. This is typically used
84 for specific offload features, where adding a static field or flag
85 in the mbuf is not justified.
87 * **Added hairpin queue.**
89 On supported NICs, we can now setup haipin queue which will offload packets
90 from the wire, backto the wire.
92 * **Added flow tag in rte_flow.**
94 SET_TAG action and TAG item have been added to support transient flow tag.
96 * **Extended metadata support in rte_flow.**
98 Flow metadata is extended to both Rx and Tx.
100 * Tx metadata can also be set by SET_META action of rte_flow.
101 * Rx metadata is delivered to host via a dynamic field of ``rte_mbuf`` with
102 PKT_RX_DYNF_METADATA.
104 * **Added ethdev API to set supported packet types**
106 * Added new API ``rte_eth_dev_set_ptypes`` that allows an application to
107 inform PMD about reduced range of packet types to handle.
108 * This scheme will allow PMDs to avoid lookup to internal ptype table on Rx
109 and thereby improve Rx performance if application wishes do so.
111 * **Added Rx offload flag to enable or disable RSS update**
113 * Added new Rx offload flag `DEV_RX_OFFLOAD_RSS_HASH` which can be used to
114 enable/disable PMDs write to `rte_mbuf::hash::rss`.
115 * PMDs notify the validity of `rte_mbuf::hash:rss` to the application
116 by enabling `PKT_RX_RSS_HASH ` flag in `rte_mbuf::ol_flags`.
118 * **Updated the enic driver.**
120 * Added support for Geneve with options offload.
121 * Added flow API implementation based on VIC Flow Manager API.
123 * **Added Hisilicon hns3 PMD.**
125 Added the new ``hns3`` net driver for the inbuilt Hisilicon Network
126 Subsystem 3(HNS3) network engine found in the Hisilicon Kunpeng 920 SoC.
127 See the :doc:`../nics/hns3` guide for more details on this new driver.
129 * **Added NXP PFE PMD.**
131 Added the new PFE driver for the NXP LS1012A platform. See the
132 :doc:`../nics/pfe` NIC driver guide for more details on this new driver.
134 * **Updated iavf PMD.**
136 Enable AVX2 data path for iavf PMD.
138 * **Updated the Intel e1000 driver.**
140 Added support for the ``RTE_ETH_DEV_CLOSE_REMOVE`` flag.
142 * **Updated the Intel ixgbe driver.**
144 Added support for the ``RTE_ETH_DEV_CLOSE_REMOVE`` flag.
146 * **Updated the Intel i40e driver.**
148 Added support for the ``RTE_ETH_DEV_CLOSE_REMOVE`` flag.
150 * **Updated the Intel fm10k driver.**
152 Added support for the ``RTE_ETH_DEV_CLOSE_REMOVE`` flag.
154 * **Added RX/TX packet burst mode get API.**
156 Added two new functions ``rte_eth_rx_burst_mode_get`` and
157 ``rte_eth_tx_burst_mode_get`` that allow an application
158 to retrieve the mode information about RX/TX packet burst
159 such as Scalar or Vector, and Vector technology like AVX2.
161 * **Updated the Intel ice driver.**
163 Updated the Intel ice driver with new features and improvements, including:
165 * Added support for device-specific DDP package loading.
166 * Added support for handling Receive Flex Descriptor.
167 * Added support for protocol extraction on per Rx queue.
168 * Added support for Flow Director filter based on generic filter framework.
169 * Added support for the ``RTE_ETH_DEV_CLOSE_REMOVE`` flag.
170 * Generic filter enhancement
171 - Supported pipeline mode.
172 - Supported new packet type like PPPoE for switch filter.
173 * Supported input set change and symmetric hash by rte_flow RSS action.
174 * Added support for GTP Tx checksum offload.
175 * Added new device IDs to support E810_XXV devices.
177 * **Added cryptodev asymmetric session-less operation.**
179 Added session-less option to cryptodev asymmetric structure. It works the same
180 way as symmetric crypto, corresponding xform is used directly by the crypto op.
182 * **Updated the Huawei hinic driver.**
184 Updated the Huawei hinic driver with new features and improvements, including:
186 * Enabled SR-IOV - Partially supported at this point, VFIO only.
187 * Supported VLAN filter and VLAN offload.
188 * Supported Unicast MAC filter and Multicast MAC filter.
189 * Supported Flow API for LACP, VRRP, BGP and so on.
190 * Supported FW version get.
192 * **Updated Mellanox mlx5 driver.**
194 Updated Mellanox mlx5 driver with new features and improvements, including:
196 * Added support for VLAN pop flow offload command.
197 * Added support for VLAN push flow offload command.
198 * Added support for VLAN set PCP offload command.
199 * Added support for VLAN set VID offload command.
200 * Added support for matching on packets withe Geneve tunnel header.
201 * Added hairpin support.
202 * Added ConnectX6-DX support.
204 * **Updated the AF_XDP PMD.**
206 Updated the AF_XDP PMD. The new features include:
208 * Enabled zero copy between application mempools and UMEM by enabling the
209 XDP_UMEM_UNALIGNED_CHUNKS UMEM flag.
211 * **Added Marvell NITROX symmetric crypto PMD.**
213 Added a symmetric crypto PMD for Marvell NITROX V security processor.
214 See the :doc:`../cryptodevs/nitrox` guide for more details on this new
216 * **Added asymmetric support to Marvell OCTEON TX crypto PMD.**
218 Added support for asymmetric operations in Marvell OCTEON TX cypto PMD.
219 Supports RSA and modexp operations.
221 * **Added Marvell OCTEON TX2 crypto PMD**
223 Added a new PMD driver for h/w crypto offload block on ``OCTEON TX2`` SoC.
225 See :doc:`../cryptodevs/octeontx2` for more details
227 * **Updated NXP crypto PMDs for PDCP support.**
229 PDCP support is added to DPAA_SEC and DPAA2_SEC PMDs using rte_security APIs.
230 Support is added for all sequence number sizes for control and user plane.
231 Test and test-crypto-perf applications are updated for unit testing.
233 * **Updated the AESNI-MB PMD.**
235 * Added support for intel-ipsec-mb version 0.53.
237 * **Updated the AESNI-GCM PMD.**
239 * Added support for intel-ipsec-mb version 0.53.
240 * Supported in-place chained mbufs on AES-GCM algorithm.
242 * **Enabled Single Pass GCM acceleration on QAT GEN3.**
244 Added support for Single Pass GCM, available on QAT GEN3 only (Intel
245 QuickAssist Technology C4xxx). It is automatically chosen instead of the
246 classic 2-pass mode when running on QAT GEN3, significantly improving
247 the performance of AES GCM operations.
249 * **Updated the Intel QuickAssist Technology (QAT) asymmetric crypto PMD.**
251 * Added support for asymmetric session-less operations.
252 * Added support for RSA algorithm with pair (n, d) private key representation.
253 * Added support for RSA algorithm with quintuple private key representation.
255 * **Updated the Intel QuickAssist Technology (QAT) compression PMD.**
257 Added stateful decompression support in the Intel QuickAssist Technology PMD.
258 Please note that stateful compression is not supported.
260 * **Added external buffers support for dpdk-test-compress-perf tool.**
262 Added a command line option to dpdk-test-compress-perf tool to allocate
263 and use memory zones as external buffers instead of keeping the data directly
266 * **Updated the IPSec library.**
268 * Added SA Database API to ``librte_ipsec``. A new test-sad application is also
269 introduced to evaluate and perform custom functional and performance tests
270 for IPsec SAD implementation.
272 * Support fragmented packets in inline crypto processing mode with fallback
273 ``lookaside-none`` session. Corresponding changes are also added in IPsec
274 Security Gateway application.
276 * **Introduced FIFO for NTB PMD.**
278 Introduced FIFO for NTB (Non-transparent Bridge) PMD to support
279 packet based processing.
281 * **Added eBPF JIT support for arm64.**
283 Added eBPF JIT support for arm64 architecture to improve the eBPF program
286 * **Added RIB and FIB (Routing/Forwarding Information Base) libraries.**
288 RIB and FIB can replace the LPM (Longest Prefix Match) library
289 with better control plane (RIB) performance.
290 The data plane (FIB) can be extended with new algorithms.
292 * **Updated testpmd.**
294 * Added a console command to testpmd app, ``show port (port_id) ptypes`` which
295 gives ability to print port supported ptypes in different protocol layers.
296 * Packet type detection disabled by default for the supported PMDs.
298 * **Added new example l2fwd-event application.**
300 Added an example application `l2fwd-event` that adds event device support to
301 traditional l2fwd example. It demonstrates usage of poll and event mode IO
302 mechanism under a single application.
304 * **Added build support for Link Time Optimization.**
306 LTO is an optimization technique used by the compiler to perform whole
307 program analysis and optimization at link time. In order to do that
308 compilers store their internal representation of the source code that
309 the linker uses at the final stage of compilation process.
311 See :doc:`../prog_guide/lto` for more information:
313 * **Added IOVA as VA support for KNI.**
315 * Added IOVA = VA support for KNI, KNI can operate in IOVA = VA mode when
316 `iova-mode=va` EAL option is passed to the application or when bus IOVA
317 scheme is selected as RTE_IOVA_VA. This mode only works on Linux Kernel
318 versions above 4.9.0.
320 * Due to IOVA to KVA address translations, based on the KNI use case there
321 can be a performance impact. For mitigation, forcing IOVA to PA via EAL
322 "--iova-mode=pa" option can be used, IOVA_DC bus iommu scheme can also
323 result in IOVA as PA.
329 .. This section should contain removed items in this release. Sample format:
331 * Add a short 1-2 sentence description of the removed item
334 This section is a comment. Do not overwrite or remove it.
335 Also, make sure to start the actual text at the margin.
336 =========================================================
338 * Removed duplicated set of commands for Rx offload configuration from testpmd::
340 port config all crc-strip|scatter|rx-cksum|rx-timestamp|
341 hw-vlan|hw-vlan-filter|hw-vlan-strip|hw-vlan-extend on|off
343 The testpmd commands set that can be used instead
344 in order to enable or disable Rx offloading on all Rx queues of a port is::
346 port config <port_id> rx_offload crc_strip|scatter|
347 ipv4_cksum|udp_cksum|tcp_cksum|timestamp|
348 vlan_strip|vlan_filter|vlan_extend on|off
350 * Removed AF_XDP pmd_zero copy vdev argument. Support is now auto-detected.
352 * The following sample applications have been removed in this release:
355 * L3 Forwarding in a Virtualization Environment
357 * Netmap Compatibility
358 * Quota and Watermark
361 * Removed arm64-dpaa2-* build config. arm64-dpaa-* can now build for both
362 dpaa and dpaa2 platforms.
368 .. This section should contain API changes. Sample format:
370 * sample: Add a short 1-2 sentence description of the API change
371 which was announced in the previous releases and made in this release.
372 Start with a scope label like "ethdev:".
373 Use fixed width quotes for ``function_names`` or ``struct_names``.
376 This section is a comment. Do not overwrite or remove it.
377 Also, make sure to start the actual text at the margin.
378 =========================================================
380 * eal: made the ``lcore_config`` struct and global symbol private.
382 * eal: removed the ``rte_cpu_check_supported`` function, replaced by
383 ``rte_cpu_is_supported`` since dpdk v17.08.
385 * eal: removed the ``rte_malloc_virt2phy`` function, replaced by
386 ``rte_malloc_virt2iova`` since v17.11.
388 * eal: made the ``rte_config`` struct and ``rte_eal_get_configuration``
391 * mem: hid the internal ``malloc_heap`` structure and the
392 ``rte_malloc_heap.h`` header.
394 * vfio: removed ``rte_vfio_dma_map`` and ``rte_vfio_dma_unmap`` that have
395 been marked as deprecated in release 19.05.
396 ``rte_vfio_container_dma_map`` and ``rte_vfio_container_dma_unmap`` can
397 be used as substitutes.
399 * pci: removed the following functions deprecated since dpdk v17.11:
401 - ``eal_parse_pci_BDF`` replaced by ``rte_pci_addr_parse``
402 - ``eal_parse_pci_DomBDF`` replaced by ``rte_pci_addr_parse``
403 - ``rte_eal_compare_pci_addr`` replaced by ``rte_pci_addr_cmp``
405 * The network structure ``esp_tail`` has been prefixed by ``rte_``.
407 * The network definitions of PPPoE ethertypes have been prefixed by ``RTE_``.
409 * The network structure for MPLS has been prefixed by ``rte_``.
411 * ethdev: changed ``rte_eth_dev_infos_get`` return value from ``void`` to
412 ``int`` to provide a way to report various error conditions.
414 * ethdev: changed ``rte_eth_promiscuous_enable`` and
415 ``rte_eth_promiscuous_disable`` return value from ``void`` to ``int`` to
416 provide a way to report various error conditions.
418 * ethdev: changed ``rte_eth_allmulticast_enable`` and
419 ``rte_eth_allmulticast_disable`` return value from ``void`` to ``int`` to
420 provide a way to report various error conditions.
422 * ethdev: changed ``rte_eth_dev_xstats_reset`` return value from ``void`` to
423 ``int`` to provide a way to report various error conditions.
425 * ethdev: changed ``rte_eth_link_get`` and ``rte_eth_link_get_nowait``
426 return value from ``void`` to ``int`` to provide a way to report various
429 * ethdev: changed ``rte_eth_macaddr_get`` return value from ``void`` to
430 ``int`` to provide a way to report various error conditions.
432 * ethdev: changed ``rte_eth_dev_owner_delete`` return value from ``void`` to
433 ``int`` to provide a way to report various error conditions.
435 * ethdev: The deprecated function ``rte_eth_dev_count`` was removed.
436 The function ``rte_eth_dev_count_avail`` is a drop-in replacement.
437 If the intent is to iterate over ports, ``RTE_ETH_FOREACH_*`` macros
438 are better port iterators.
440 * ethdev: RTE_FLOW_ITEM_TYPE_META data endianness altered to host one.
441 Due to the new dynamic metadata field in mbuf is host-endian either, there
442 is the minor compatibility issue for applications in case of 32-bit values
445 * ethdev: the tx_metadata mbuf field is moved to dymanic one.
446 PKT_TX_METADATA flag is replaced with PKT_TX_DYNF_METADATA.
447 DEV_TX_OFFLOAD_MATCH_METADATA offload flag is removed, now metadata
448 support in PMD is engaged on dynamic field registration.
450 * event: The function ``rte_event_eth_tx_adapter_enqueue`` takes an additional
451 input as ``flags``. Flag ``RTE_EVENT_ETH_TX_ADAPTER_ENQUEUE_SAME_DEST`` which
452 has been introduced in this release is used when used when all the packets
453 enqueued in the tx adapter are destined for the same Ethernet port & Tx queue.
455 * sched: The pipe nodes configuration parameters such as number of pipes,
456 pipe queue sizes, pipe profiles, etc., are moved from port level structure
457 to subport level. This allows different subports of the same port to
458 have different configuration for the pipe nodes.
464 .. This section should contain ABI changes. Sample format:
466 * sample: Add a short 1-2 sentence description of the ABI change
467 which was announced in the previous releases and made in this release.
468 Start with a scope label like "ethdev:".
469 Use fixed width quotes for ``function_names`` or ``struct_names``.
472 This section is a comment. Do not overwrite or remove it.
473 Also, make sure to start the actual text at the margin.
474 =========================================================
476 * net: The Ethernet address and other header definitions have changed
477 attributes. They have been modified to be aligned on 2-byte boundaries.
478 These changes should not impact normal usage because drivers naturally
479 align the Ethernet header on receive and all known encapsulations
480 preserve the alignment of the header.
482 * security: The field ``replay_win_sz`` has been moved from ipsec library
483 based ``rte_ipsec_sa_prm`` structure to security library based structure
484 ``rte_security_ipsec_xform``, which specify the Anti replay window size
485 to enable sequence replay attack handling.
487 * ipsec: The field ``replay_win_sz`` has been removed from the structure
488 ``rte_ipsec_sa_prm`` as it has been added to the security library.
490 * ethdev: Added 32-bit fields for maximum LRO aggregated packet size, in
491 struct ``rte_eth_dev_info`` for the port capability and in struct
492 ``rte_eth_rxmode`` for the port configuration.
493 Application should use the new field in struct ``rte_eth_rxmode`` to configure
495 PMD should use the new field in struct ``rte_eth_dev_info`` to report the
496 supported port capability.
499 Shared Library Versions
500 -----------------------
502 .. Update any library version updated in this release
503 and prepend with a ``+`` sign, like this:
509 This section is a comment. Do not overwrite or remove it.
510 =========================================================
512 The libraries prepended with a plus sign were incremented in this version.
518 librte_bitratestats.so.2
521 librte_bus_fslmc.so.2
522 librte_bus_ifpga.so.2
525 librte_bus_vmbus.so.2
528 librte_compressdev.so.1
529 librte_cryptodev.so.8
530 librte_distributor.so.1
533 + librte_ethdev.so.13
534 + librte_eventdev.so.8
536 librte_flow_classify.so.1
545 librte_latencystats.so.1
559 librte_pmd_ixgbe.so.2
560 librte_pmd_dpaa2_qdma.so.1
562 librte_pmd_softnic.so.1
563 librte_pmd_vhost.so.2
572 + librte_security.so.3
582 .. This section should contain new known issues in this release. Sample format:
584 * **Add title in present tense with full stop.**
586 Add a short 1-2 sentence description of the known issue
587 in the present tense. Add information on any known workarounds.
589 This section is a comment. Do not overwrite or remove it.
590 Also, make sure to start the actual text at the margin.
591 =========================================================
597 .. This section should contain a list of platforms that were tested
602 * <vendor> platform with <vendor> <type of devices> combinations
607 * Other relevant details...
609 This section is a comment. Do not overwrite or remove it.
610 Also, make sure to start the actual text at the margin.
611 =========================================================