1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2018 The DPDK contributors
7 .. **Read this first.**
9 The text in the sections below explains how to update the release notes.
11 Use proper spelling, capitalization and punctuation in all sections.
13 Variable and config names should be quoted as fixed width text:
16 Build the docs and view the output file to ensure the changes are correct::
20 xdg-open build/doc/html/guides/rel_notes/release_18_11.html
26 .. This section should contain new features added in this release.
29 * **Add a title in the past tense with a full stop.**
31 Add a short 1-2 sentence description in the past tense.
32 The description should be enough to allow someone scanning
33 the release notes to understand the new feature.
35 If the feature adds a lot of sub-features you can use a bullet list
38 * Added feature foo to do something.
39 * Enhanced feature bar to do something else.
41 Refer to the previous release notes for examples.
43 Suggested order in release notes items:
44 * Core libs (EAL, mempool, ring, mbuf, buses)
45 * Device abstraction libs and PMDs
47 - cryptodev (lib, PMDs)
48 - eventdev (lib, PMDs)
51 * Apps, Examples, Tools (if significative)
53 This section is a comment. Do not overwrite or remove it.
54 Also, make sure to start the actual text at the margin.
55 =========================================================
57 * **Added support for using externally allocated memory in DPDK.**
59 DPDK has gained support for creating new ``rte_malloc`` heaps referencing
60 memory that was created outside of DPDK's own page allocator, and using that
61 memory natively with any other DPDK library or data structure.
63 * **Added check for ensuring allocated memory addressable by devices.**
65 Some devices can have addressing limitations so a new function,
66 ``rte_mem_check_dma_mask``, has been added for checking allocated memory is
67 not out of the device range. Because now memory can be dynamically allocated
68 after initialization, a dma mask is kept and any new allocated memory will be
69 checked out against that dma mask and rejected if out of range. If more than
70 one device has addressing limitations, the dma mask is the more restricted one.
72 * **Updated the C11 memory model version of ring library.**
74 The latency is decreased for architectures using the C11 memory model
75 version of the ring library.
77 On Cavium ThunderX2 platform, the changes decreased latency by 27~29%
78 and 3~15% for MPMC and SPSC cases respectively (with 2 lcores). The
79 real improvements may vary with the number of contending lcores and
82 * **Added hot-unplug handle mechanism.**
84 ``rte_dev_hotplug_handle_enable`` and ``rte_dev_hotplug_handle_disable`` are
85 for enabling or disabling hotplug handle mechanism.
87 * **Support device multi-process hotplug.**
89 Hotplug and hot-unplug for devices will now be supported in multiprocessing
90 scenario. Any ethdev devices created in the primary process will be regarded
91 as shared and will be available for all DPDK processes. Synchronization
92 between processes will be done using DPDK IPC.
94 * **Added new Flow API actions to rewrite fields in packet headers.**
96 Added new Flow API actions to:
98 * Modify source and destination IP addresses in the outermost IPv4/IPv6
100 * Modify source and destination port numbers in the outermost TCP/UDP
103 * **Added new Flow API action to swap MAC addresses in Ethernet header.**
105 Added new Flow API action to swap the source and destination MAC
106 addresses in the outermost Ethernet header.
108 * **Add support to offload more flow match and actions for CXGBE PMD**
110 Flow API support has been enhanced for CXGBE Poll Mode Driver to offload:
112 * Match items: destination MAC address.
113 * Action items: push/pop/rewrite vlan header,
114 rewrite IP addresses in outermost IPv4/IPv6 header,
115 rewrite port numbers in outermost TCP/UDP header,
116 swap MAC addresses in outermost Ethernet header.
118 * **Added a devarg to use the latest supported vector path in i40e.**
119 A new devarg ``use-latest-supported-vec`` was introduced to allow users to
120 choose the latest vector path that the platform supported. For example, users
121 can use AVX2 vector path on BDW/HSW to get better performance.
123 * **Added support for SR-IOV in netvsc PMD.**
125 The ``netvsc`` poll mode driver now supports the Accelerated Networking
126 SR-IOV option in Hyper-V and Azure. This is an alternative to the previous
127 vdev_netvsc, tap, and failsafe drivers combination.
129 * **Added a new net driver for Marvell Armada 3k device.**
131 Added the new ``mvneta`` net driver for Marvell Armada 3k device. See the
132 :doc:`../nics/mvneta` NIC guide for more details on this new driver.
134 * **Added NXP ENETC PMD.**
136 Added the new enetc driver for NXP enetc platform. See the
137 "ENETC Poll Mode Driver" document for more details on this new driver.
139 * **Added Ethernet poll mode driver for Aquantia aQtion family of 10G devices.**
141 Added the new ``atlantic`` ethernet poll mode driver for Aquantia XGBE devices.
142 See the :doc:`../nics/atlantic` nic driver guide for more details on this
145 * **Updated mlx5 driver.**
147 Updated the mlx5 driver including the following changes:
149 * Improved security of PMD to prevent the NIC from getting stuck when
150 application misbehaves.
151 * reworked flow engine to supported e-switch flow rules (transfer attribute).
152 * Supported header re-write(L2-L4), VXLAN encap/decap, count ,match
153 on TCP flags and multiple flow groups with e-switch flow rules.
154 * Supported match on medatada, VXLAN and MPLS encap/decap with flow rules.
155 * Supported RTE_ETH_DEV_CLOSE_REMOVE flag to provide better support for
157 * Supported meson build.
158 * Fixed build issue with PPC.
159 * Supported BlueField VF.
160 * Supported externally allocated static memory for DMA.
162 * **Updated Solarflare network PMD.**
164 Updated the sfc_efx driver including the following changes:
166 * Added support for Rx scatter in EF10 datapath implementation.
167 * Added support for Rx descriptor status API in EF10 datapath implementation.
168 * Added support for TSO in EF10 datapath implementation.
169 * Added support for Tx descriptor status API in EF10 (ef10 and ef10_simple)
170 datapaths implementation.
172 * **Updated the enic driver.**
174 * Added AVX2-based vectorized Rx handler.
175 * Added VLAN and checksum offloads to the simple Tx handler.
176 * Added the count flow action.
177 * Enabled the virtual address IOVA mode.
179 * **Updated failsafe driver.**
181 Updated the failsafe driver including the following changes:
183 * Support for Rx and Tx queues start and stop.
184 * Support for Rx and Tx queues deferred start.
185 * Support for runtime Rx and Tx queues setup.
186 * Support multicast MAC address set.
188 * **Added a devarg to use PCAP interface physical MAC address.**
189 A new devarg ``phy_mac`` was introduced to allow users to use physical
190 MAC address of the selected PCAP interface.
192 * **Added TAP Rx/Tx queues sharing with a secondary process.**
194 A secondary process can attach a TAP device created in the primary process,
195 probe the queues, and process Rx/Tx in a secondary process.
197 * **Added classification and metering support to SoftNIC PMD.**
199 Added support for flow classification (rte_flow API), and metering and
200 policing (rte_mtr API) to the SoftNIC PMD.
202 * **Added Crypto support to Softnic PMD.**
204 The Softnic is now capable of processing symmetric crypto workloads such
205 as cipher, cipher-authentication chaining, and aead encryption and
206 decryption. This is achieved by calling DPDK Cryptodev APIs.
208 * **Added cryptodev port to port library.**
210 Cryptodev port is a shim layer in the port library that interacts with DPDK
211 Cryptodev PMDs including burst enqueuing and dequeuing crypto operations.
213 * **Added symmetric cryptographic actions to the pipeline library.**
215 In the pipeline library an added symmetric crypto action parsing and action
216 handler are implemented. The action allows automatically preparing the crypto
217 operation with the rules specified such as algorithm, key, and IV, etc for
218 the cryptodev port to process.
220 * **Added support for GEN3 devices to Intel QAT driver .**
222 Added support for the third generation of Intel QuickAssist devices.
224 * **Updated the QAT PMD.**
226 The QAT PMD was updated with additional support for:
228 * AES-CMAC algorithm.
230 * **Updated the AESNI MB PMD.**
232 The AESNI MB PMD has been updated with additional support for AES-GCM
235 * **Added NXP CAAM JR PMD.**
237 Added the new caam job ring driver for NXP platforms. See the
238 "NXP CAAM JOB RING (caam_jr)" document for more details on this new driver.
240 * **Added support for Dynamic Huffman Encoding to Intel QAT comp PMD.**
242 The Intel QuickAssist (QAT) compression PMD has been updated with support
243 for Dynamic Huffman Encoding for the Deflate algorithm.
245 * **Added Event Ethernet Tx Adapter.**
247 Added event ethernet Tx adapter library that provides configuration and
248 data path APIs for the ethernet transmit stage of an event driven packet
249 processing application. These APIs abstract the implementation of the
250 transmit stage and allow the application to use eventdev PMD support or
251 a common implementation.
253 * **Added Distributed Software Eventdev PMD.**
255 Added the new Distributed Software Event Device (DSW), which is a
256 pure-software eventdev driver distributing the work of scheduling
257 among all eventdev ports and the lcores using them. DSW, compared to
258 the SW eventdev PMD, sacrifices load balancing performance to
259 gain better event scheduling throughput and scalability.
261 * **Added extendable bucket feature to hash library (rte_hash).**
263 This new “extendable bucket” feature provides 100% insertion guarantee to
264 the capacity specified by the user by extending hash table with extra
265 buckets when needed to accommodate the unlikely event of intensive hash
266 collisions. In addition, the internal hashing algorithm was changed to use
267 partial-key hashing to improve memory efficiency and lookup performance.
269 * **Added lock free reader/writer concurrency to hash library (rte_hash).**
271 Lock free reader/writer concurrency prevents the readers from getting
272 blocked due to a pre-empted writer thread. This allows the hash library
273 to be used in scenarios where the writer thread runs on control plane.
275 * **Added Traffic Pattern Aware Power Control Library**
277 Added an experimental library. This extend Power Library and provide
278 empty_poll APIs. This feature measure how many times empty_poll are
279 executed per core, use the number of empty polls as a hint for system
282 See the :doc:`../prog_guide/power_man` section of the DPDK Programmers
283 Guide document for more information.
285 * **Added JSON power policy interface for containers.**
287 Extended the Power Library and vm_power_manager sample app to allow power
288 policies to be submitted via a FIFO using JSON formatted strings. Previously
289 limited to Virtual Machines, this feature extends power policy functionality
290 to containers and host applications that need to have their cores frequency
291 controlled based on the rules contained in the policy.
293 * **Added Telemetry API.**
295 Added the telemetry API which allows applications to transparently expose
296 their telemetry via a UNIX socket in JSON. The JSON can be consumed by any
297 Service Assurance agent, such as CollectD.
299 * **Updated KNI kernel module, rte_kni library, and KNI sample application.**
301 Updated the KNI kernel module with a new kernel module parameter,
302 ``carrier=[on|off]`` to allow the user to control the default carrier
303 state of KNI kernel network interfaces. The default carrier state
304 is now set to ``off``, so the interfaces cannot be used until the
305 carrier state is set to ``on`` via ``rte_kni_update_link`` or
306 by writing ``1`` to ``/sys/devices/virtual/net/<iface>/carrier``.
307 In previous versions the default carrier state was left undefined.
308 See :doc:`../prog_guide/kernel_nic_interface` for more information.
310 Added the new API function ``rte_kni_update_link`` to allow the user
311 to set the carrier state of the KNI kernel network interface.
313 Added a new command line flag ``-m`` to the KNI sample application to
314 monitor and automatically reflect the physical NIC carrier state to the
315 KNI kernel network interface with the new ``rte_kni_update_link`` API.
316 See :doc:`../sample_app_ug/kernel_nic_interface` for more information.
318 * **Added ability to switch queue deferred start flag on testpmd app.**
320 Added a console command to testpmd app, giving ability to switch
321 ``rx_deferred_start`` or ``tx_deferred_start`` flag of the specified queue of
322 the specified port. The port must be stopped before the command call in order
323 to reconfigure queues.
325 * **Add a new sample for vDPA**
327 The vdpa sample application creates vhost-user sockets by using the
328 vDPA backend. vDPA stands for vhost Data Path Acceleration which utilizes
329 virtio ring compatible devices to serve virtio driver directly to enable
330 datapath acceleration. As vDPA driver can help to set up vhost datapath,
331 this application doesn't need to launch dedicated worker threads for vhost
332 enqueue/dequeue operations.
334 * **Added cryptodev FIPS validation example application.**
336 Added an example application to parse and perform symmetric cryptography
337 computation to the NIST Cryptographic Algorithm Validation Program (CAVP)
340 * **Allow unit test binary to take parameters from the environment**
342 The unit test "test", or "dpdk-test", binary is often called from scripts,
343 which can make passing additional parameters, such as a coremask, to it more
344 awkward. Support has been added to the application to allow it to take
345 additional command-line parameter values from the "DPDK_TEST_PARAMS"
346 environment variable to make this application easier to use.
352 .. This section should contain API changes. Sample format:
354 * Add a short 1-2 sentence description of the API change.
355 Use fixed width quotes for ``function_names`` or ``struct_names``.
358 This section is a comment. Do not overwrite or remove it.
359 Also, make sure to start the actual text at the margin.
360 =========================================================
362 * eal: ``rte_memseg_list`` structure now has an additional flag indicating
363 whether the memseg list is externally allocated. This will have implications
364 for any users of memseg-walk-related functions, as they will now have to skip
365 externally allocated segments in most cases if the intent is to only iterate
366 over internal DPDK memory.
367 ``socket_id`` parameter across the entire DPDK has gained additional meaning,
368 as some socket ID's will now be representing externally allocated memory. No
369 changes will be required for existing code as backwards compatibility will be
370 kept, and those who do not use this feature will not see these extra socket
371 ID's. Any new API's must not check socket ID parameters themselves, and must
372 instead leave it to the memory subsystem to decide whether socket ID is a
375 * eal: The following devargs functions, which were deprecated in 18.05,
376 were removed in 18.11:
377 ``rte_eal_parse_devargs_str()``, ``rte_eal_devargs_add()``,
378 ``rte_eal_devargs_type_count()``, and ``rte_eal_devargs_dump()``.
380 * eal: The parameters of the function ``rte_devargs_remove()`` have changed
381 from bus and device names to ``struct rte_devargs``.
383 * eal: The deprecated functions attach/detach were removed in 18.11.
384 ``rte_eal_dev_attach`` can be replaced by
385 ``rte_dev_probe`` or ``rte_eal_hotplug_add``.
386 ``rte_eal_dev_detach`` can be replaced by
387 ``rte_dev_remove`` or ``rte_eal_hotplug_remove``.
389 * eal: The scope of ``rte_eal_hotplug_add()``/``rte_dev_probe()``
390 and ``rte_eal_hotplug_remove()``/``rte_dev_remove()`` is extended.
391 In multi-process model, they will guarantee that the device is
392 attached or detached on all processes.
394 * mbuf: The ``__rte_mbuf_raw_free()`` and ``__rte_pktmbuf_prefree_seg()``
395 functions were deprecated since 17.05 and are replaced by
396 ``rte_mbuf_raw_free()`` and ``rte_pktmbuf_prefree_seg()``.
398 * ethdev: The deprecated functions attach/detach were removed in 18.11.
399 ``rte_eth_dev_attach`` can be replaced by ``RTE_ETH_FOREACH_MATCHING_DEV``
400 and ``rte_dev_probe`` or ``rte_eal_hotplug_add``.
401 ``rte_eth_dev_detach`` can be replaced by
402 ``rte_dev_remove`` or ``rte_eal_hotplug_remove``.
404 * ethdev: A call to ``rte_eth_dev_release_port()`` has been added in
405 ``rte_eth_dev_close()``. As a consequence, a closed port is freed
406 and seen as invalid because of its state ``RTE_ETH_DEV_UNUSED``.
407 This new behaviour is enabled per driver for a migration period.
409 * A new device flag, RTE_ETH_DEV_NOLIVE_MAC_ADDR, changes the order of
410 actions inside rte_eth_dev_start regarding MAC set. Some NICs do not
411 support MAC changes once the port has started and with this new device
412 flag the MAC can be properly configured in any case. This is particularly
413 important for bonding.
415 * The default behaviour of CRC strip offload changed. Without any specific Rx
416 offload flag, default behavior by PMD is now to strip CRC.
417 DEV_RX_OFFLOAD_CRC_STRIP offload flag has been removed.
418 To request keeping CRC, application should set ``DEV_RX_OFFLOAD_KEEP_CRC`` Rx
421 * eventdev: Type of 2nd parameter to ``rte_event_eth_rx_adapter_caps_get()``
422 has been changed from uint8_t to uint16_t.
424 * kni: By default, interface carrier status is ``off`` which means there won't
425 be any traffic. It can be set to ``on`` via ``rte_kni_update_link()`` API
426 or via ``sysfs`` interface:
427 ``echo 1 > /sys/class/net/vEth0/carrier``.
428 Note interface should be ``up`` to be able to read/write sysfs interface.
429 When KNI sample application is used, ``-m`` parameter can be used to
430 automatically update the carrier status for the interface.
432 * kni: When ethtool support enabled (``CONFIG_RTE_KNI_KMOD_ETHTOOL=y``)
433 ethtool commands ``ETHTOOL_GSET & ETHTOOL_SSET`` are no more supported for the
434 kernels that has ``ETHTOOL_GLINKSETTINGS & ETHTOOL_SLINKSETTINGS`` support.
435 This means ``ethtool "-a|--show-pause", "-s|--change"`` won't work, and
436 ``ethtool <iface>`` output will have less information.
442 .. This section should contain ABI changes. Sample format:
444 * Add a short 1-2 sentence description of the ABI change
445 that was announced in the previous releases and made in this release.
446 Use fixed width quotes for ``function_names`` or ``struct_names``.
449 This section is a comment. Do not overwrite or remove it.
450 Also, make sure to start the actual text at the margin.
451 =========================================================
453 * eal: added ``legacy_mem`` and ``single_file_segments`` values to
454 ``rte_config`` structure on account of improving DPDK usability when
455 using either ``--legacy-mem`` or ``--single-file-segments`` flags.
457 * eal: EAL library ABI version was changed due to previously announced work on
458 supporting external memory in DPDK:
459 - structure ``rte_memseg_list`` now has a new field indicating length
460 of memory addressed by the segment list
461 - structure ``rte_memseg_list`` now has a new flag indicating whether
462 the memseg list refers to external memory
463 - structure ``rte_malloc_heap`` now has a new field indicating socket
464 ID the malloc heap belongs to
465 - structure ``rte_mem_config`` has had its ``malloc_heaps`` array
466 resized from ``RTE_MAX_NUMA_NODES`` to ``RTE_MAX_HEAPS`` value
467 - structure ``rte_malloc_heap`` now has a ``heap_name`` member
468 - structure ``rte_eal_memconfig`` has been extended to contain next
469 socket ID for externally allocated segments
471 * eal: Added ``dma_maskbits`` to ``rte_mem_config`` for keeping more restricted
472 dma mask based on devices addressing limitations.
474 * eal: The structure ``rte_device`` got a new field to reference a ``rte_bus``.
475 It is changing the size of the ``struct rte_device`` and the inherited
476 device structures of all buses.
482 .. This section should contain removed items in this release. Sample format:
484 * Add a short 1-2 sentence description of the removed item
487 This section is a comment. Do not overwrite or remove it.
488 Also, make sure to start the actual text at the margin.
489 =========================================================
492 Shared Library Versions
493 -----------------------
495 .. Update any library version updated in this release
496 and prepend with a ``+`` sign, like this:
499 + librte_cfgfile.so.2
502 This section is a comment. Do not overwrite or remove it.
503 =========================================================
505 The libraries prepended with a plus sign were incremented in this version.
511 librte_bitratestats.so.2
513 + librte_bus_dpaa.so.2
514 + librte_bus_fslmc.so.2
515 + librte_bus_ifpga.so.2
516 + librte_bus_pci.so.2
517 + librte_bus_vdev.so.2
518 + librte_bus_vmbus.so.2
521 librte_compressdev.so.1
522 librte_cryptodev.so.5
523 librte_distributor.so.1
526 + librte_ethdev.so.11
527 + librte_eventdev.so.6
528 librte_flow_classify.so.1
536 librte_latencystats.so.1
550 librte_pmd_ixgbe.so.2
551 librte_pmd_dpaa2_qdma.so.1
553 librte_pmd_softnic.so.1
554 librte_pmd_vhost.so.2
570 .. This section should contain new known issues in this release. Sample format:
572 * **Add title in present tense with full stop.**
574 Add a short 1-2 sentence description of the known issue
575 in the present tense. Add information on any known workarounds.
577 This section is a comment. Do not overwrite or remove it.
578 Also, make sure to start the actual text at the margin.
579 =========================================================
581 * When using SR-IOV (VF) support with netvsc PMD and the Mellanox mlx5 bifurcated
582 driver; the Linux netvsc device must be brought up before the netvsc device is
583 unbound and passed to the DPDK.
585 * IBM Power8 is not supported by this release of DPDK. IBM Power9 is supported.
591 .. This section should contain a list of platforms that were tested
596 * <vendor> platform with <vendor> <type of devices> combinations
601 * Other relevant details...
603 This section is a comment. Do not overwrite or remove it.
604 Also, make sure to start the actual text at the margin.
605 =========================================================