4 .. **Read this first.**
6 The text below explains how to update the release notes.
8 Use proper spelling, capitalization and punctuation in all sections.
10 Variable and config names should be quoted as fixed width text: ``LIKE_THIS``.
12 Build the docs and view the output file to ensure the changes are correct::
16 firefox build/doc/html/guides/rel_notes/release_16_07.html
22 .. This section should contain new features added in this release. Sample format:
24 * **Add a title in the past tense with a full stop.**
26 Add a short 1-2 sentence description in the past tense. The description
27 should be enough to allow someone scanning the release notes to understand
30 If the feature adds a lot of sub-features you can use a bullet list like this.
32 * Added feature foo to do something.
33 * Enhanced feature bar to do something else.
35 Refer to the previous release notes for examples.
37 * **Removed mempool cache if not needed.**
39 The size of the mempool structure is reduced if the per-lcore cache is disabled.
41 * **Added mempool external cache for non-EAL thread.**
43 Added new functions to create, free or flush a user-owned mempool
44 cache for non-EAL threads. Previously the cache was always disabled
47 * **Changed the memory allocation in mempool library.**
49 * Added ability to allocate a large mempool in virtually fragmented memory.
50 * Added new APIs to populate a mempool with memory.
51 * Added an API to free a mempool.
52 * Modified the API of the ``rte_mempool_obj_iter()`` function.
53 * Dropped specific Xen Dom0 code.
54 * Dropped specific anonymous mempool code in testpmd.
56 * **Added new driver for Broadcom NetXtreme-C devices.**
58 Added the new bnxt driver for Broadcom NetXtreme-C devices. See the
59 "Network Interface Controller Drivers" document for more details on this
62 * **Added new driver for ThunderX nicvf device.**
64 * **Added mailbox interrupt support for ixgbe and igb VFs.**
66 When the physical NIC link comes up or down, the PF driver will send a
67 mailbox message to notify each VF. To handle this link up/down event,
68 support have been added for a mailbox interrupt to receive the message and
69 allow the application to register a callback for it.
71 * **Updated the ixgbe base driver.**
73 The ixgbe base driver was updated with changes including the
76 * Added sgmii link for X550.
77 * Added MAC link setup for X550a SFP and SFP+.
78 * Added KR support for X550em_a.
79 * Added new PHY definitions for M88E1500.
80 * Added support for the VLVF to be bypassed when adding/removing a VFTA entry.
81 * Added X550a flow control auto negotiation support.
83 * **Updated the i40e base driver.**
85 Updated the i40e base driver including support for new devices IDs.
87 * **Updated the enic driver.**
89 The enic driver was updated with changes including the following:
91 * Optimized the Tx function.
92 * Added Scattered Rx capability.
93 * Improved packet type identification.
94 * Added MTU update in non Scattered Rx mode and enabled MTU of up to 9208
95 with UCS Software release 2.2 on 1300 series VICs.
97 * **Added support for virtio on IBM POWER8.**
99 The ioports are mapped in memory when using Linux UIO.
101 * **Added support for Virtio in containers.**
103 Add a new virtual device, named virtio_user, to support virtio for containers.
107 * Control queue and multi-queue are not supported yet.
108 * Doesn't work with ``--huge-unlink``.
109 * Doesn't work with ``--no-huge``.
110 * Doesn't work when there are more than ``VHOST_MEMORY_MAX_NREGIONS(8)`` hugepages.
111 * Root privilege is required for sorting hugepages by physical address.
112 * Can only be used with the vhost user backend.
114 * **Added vhost-user client mode.**
116 DPDK vhost-user now supports client mode as well as server mode. Client mode
117 is enabled when the ``RTE_VHOST_USER_CLIENT`` flag is set while calling
118 ``rte_vhost_driver_register``.
120 When DPDK vhost-user restarts from an normal or abnormal exit (such as a
121 crash), the client mode allows DPDK to establish the connection again. Note
122 that QEMU version v2.7 or above is required for this feature.
124 DPDK vhost-user will also try to reconnect by default when:
126 * The first connect fails (when QEMU is not started yet).
127 * The connection is broken (when QEMU restarts).
129 It can be turned off by setting the ``RTE_VHOST_USER_NO_RECONNECT`` flag.
131 * **Added NSH packet recognition in i40e.**
133 * **Added AES-CTR support to AESNI MB PMD.**
135 Now AESNI MB PMD supports 128/192/256-bit counter mode AES encryption and
138 * **Added support for AES counter mode with Intel QuickAssist devices.**
140 Enabled support for the AES CTR algorithm for Intel QuickAssist devices.
141 Provided support for algorithm-chaining operations.
143 * **Added KASUMI SW PMD.**
145 A new Crypto PMD has been added, which provides KASUMI F8 (UEA1) ciphering
146 and KASUMI F9 (UIA1) hashing.
148 * **Added multi-writer support for RTE Hash with Intel TSX.**
150 The following features/modifications have been added to rte_hash library:
152 * Enabled application developers to use an extra flag for ``rte_hash``
153 creation to specify default behavior (multi-thread safe/unsafe) with the
154 ``rte_hash_add_key`` function.
155 * Changed the Cuckoo Hash Search algorithm to breadth first search for
156 multi-writer routines and split Cuckoo Hash Search and Move operations in
157 order to reduce transactional code region and improve TSX performance.
158 * Added a hash multi-writer test case to the test app.
160 * **Improved IP Pipeline Application.**
162 The following features have been added to the ip_pipeline application:
164 * Configure the MAC address in the routing pipeline and automatic route
165 updates with change in link state.
166 * Enable RSS per network interface through the configuration file.
167 * Streamline the CLI code.
169 * **Added keepalive enhancements.**
171 Added support for reporting of core states other than dead to
172 monitoring applications, enabling the support of broader liveness
173 reporting to external processes.
175 * **Added packet capture framework.**
177 * A new library ``librte_pdump`` is added to provide a packet capture API.
178 * A new ``app/pdump`` tool is added to demonstrate capture packets in DPDK.
181 * **Added floating VEB support for i40e PF driver.**
183 A "floating VEB" is a special Virtual Ethernet Bridge (VEB) which does not
184 have an upload port, but instead is used for switching traffic between
185 virtual functions (VFs) on a port.
187 For information on this feature, please see the "I40E Poll Mode Driver"
188 section of the "Network Interface Controller Drivers" document.
190 * **Added support for live migration of a VM with SRIOV VF.**
192 Live migration of a VM with Virtio and VF PMD's using the bonding PMD.
198 .. This section should contain bug fixes added to the relevant sections. Sample format:
200 * **code/section Fixed issue in the past tense with a full stop.**
202 Add a short 1-2 sentence description of the resolved issue in the past tense.
203 The title should contain the code/lib section like a commit message.
204 Add the entries in alphabetic order in the relevant sections below.
210 * **igb_uio: Fixed possible mmap failure for Linux >= 4.5.**
212 The mmaping of the iomem range of the PCI device fails for kernels that
213 enabled the ``CONFIG_IO_STRICT_DEVMEM`` option. The error seen by the
214 user is as similar to the following::
216 EAL: pci_map_resource():
218 cannot mmap(39, 0x7f1c51800000, 0x100000, 0x0):
219 Invalid argument (0xffffffffffffffff)
221 The ``CONFIG_IO_STRICT_DEVMEM`` kernel option was introduced in Linux v4.5.
223 The issues was resolve by updating ``igb_uio`` to stop reserving PCI memory
224 resources. From the kernel point of view the iomem region looks like idle
225 and mmap works again. This matches the ``uio_pci_generic`` usage.
231 * **i40e: Fixed vlan stripping from inner header.**
233 Previously, for tunnel packets, such as VXLAN/NVGRE, the vlan
234 tags of the inner header will be stripped without putting vlan
236 Now this issue is fixed by disabling vlan stripping from inner header.
238 * **i40e: Fixed the type issue of a single VLAN type.**
240 Currently, if a single VLAN header is added in a packet, it's treated
241 as inner VLAN. But generally, a single VLAN header is treated as the
243 This issue is fixed by changing corresponding register for single VLAN.
245 * **enic: Fixed several issues when stopping then restarting ports and queues.**
247 Fixed several crashes related to stopping then restarting ports and queues.
248 Fixed possible crash when re-configuring the number of Rx queue descriptors.
250 * **enic: Fixed Rx data mis-alignment if mbuf data offset modified.**
252 Fixed possible Rx corruption when mbufs were returned to a pool with data
253 offset other than RTE_PKTMBUF_HEADROOM.
255 * **enic: Fixed Tx IP/UDP/TCP checksum offload and VLAN insertion.**
257 * **enic: Fixed Rx error and missed counters.**
263 * **mbuf: Fixed refcnt update when detaching.**
265 Fix the ``rte_pktmbuf_detach()`` function to decrement the direct mbuf's
266 reference counter. The previous behavior was not to affect the reference
267 counter. This lead to a memory leak of the direct mbuf.
281 .. This section should contain new known issues in this release. Sample format:
283 * **Add title in present tense with full stop.**
285 Add a short 1-2 sentence description of the known issue in the present
286 tense. Add information on any known workarounds.
292 .. This section should contain API changes. Sample format:
294 * Add a short 1-2 sentence description of the API change. Use fixed width
295 quotes for ``rte_function_names`` or ``rte_struct_names``. Use the past tense.
297 * The following counters are removed from the ``rte_eth_stats`` structure:
309 * The extended statistics are fetched by ids with ``rte_eth_xstats_get``
310 after a lookup by name ``rte_eth_xstats_get_names``.
312 * The function ``rte_eth_dev_info_get`` fill the new fields ``nb_rx_queues``
313 and ``nb_tx_queues`` in the structure ``rte_eth_dev_info``.
315 * The vhost function ``rte_vring_available_entries`` is renamed to
316 ``rte_vhost_avail_entries``.
318 * All existing vhost APIs and callbacks with ``virtio_net`` struct pointer
319 as the parameter have been changed due to the ABI refactoring described
320 below. It is replaced by ``int vid``.
322 * The function ``rte_vhost_enqueue_burst`` no longer supports concurrent enqueuing
323 packets to the same queue.
325 * The function ``rte_eth_dev_set_mtu`` adds a new return value ``-EBUSY``, which
326 indicates the operation is forbidden because the port is running.
328 * The script ``dpdk_nic_bind.py`` is renamed to ``dpdk-devbind.py``.
329 And the script ``setup.sh`` is renamed to ``dpdk-setup.sh``.
335 .. * Add a short 1-2 sentence description of the ABI change that was announced in
336 the previous releases and made in this release. Use fixed width quotes for
337 ``rte_function_names`` or ``rte_struct_names``. Use the past tense.
339 * The ``rte_port_source_params`` structure has new fields to support PCAP files.
340 It was already in release 16.04 with ``RTE_NEXT_ABI`` flag.
342 * The ``rte_eth_dev_info`` structure has new fields ``nb_rx_queues`` and ``nb_tx_queues``
343 to support the number of queues configured by software.
345 * A Vhost ABI refactoring has been made: the ``virtio_net`` structure is no
346 longer exported directly to the application. Instead, a handle, ``vid``, has
347 been used to represent this structure internally.
350 Shared Library Versions
351 -----------------------
353 .. Update any library version updated in this release and prepend with a ``+`` sign.
355 The libraries prepended with a plus sign were incremented in this version.
363 librte_cryptodev.so.1
364 librte_distributor.so.1
374 + librte_mempool.so.2
393 .. This section should contain a list of platforms that were tested with this
407 .. This section should contain a list of NICs that were tested with this release.