ceph/src/spdk/dpdk/doc/guides/rel_notes/release_16_07.rst

   1 ..  SPDX-License-Identifier: BSD-3-Clause
   2     Copyright 2016 The DPDK contributors
   3
   4 DPDK Release 16.07
   5 ==================
   6
   7 .. **Read this first.**
   8
   9    The text below explains how to update the release notes.
  10
  11    Use proper spelling, capitalization and punctuation in all sections.
  12
  13    Variable and config names should be quoted as fixed width text: ``LIKE_THIS``.
  14
  15    Build the docs and view the output file to ensure the changes are correct::
  16
  17       make doc-guides-html
  18
  19       firefox build/doc/html/guides/rel_notes/release_16_07.html
  20
  21
  22 New Features
  23 ------------
  24
  25 .. This section should contain new features added in this release. Sample format:
  26
  27    * **Add a title in the past tense with a full stop.**
  28
  29      Add a short 1-2 sentence description in the past tense. The description
  30      should be enough to allow someone scanning the release notes to understand
  31      the new feature.
  32
  33      If the feature adds a lot of sub-features you can use a bullet list like this.
  34
  35      * Added feature foo to do something.
  36      * Enhanced feature bar to do something else.
  37
  38      Refer to the previous release notes for examples.
  39
  40 * **Removed the mempool cache memory if caching is not being used.**
  41
  42   The size of the mempool structure is reduced if the per-lcore cache is disabled.
  43
  44 * **Added mempool external cache for non-EAL thread.**
  45
  46   Added new functions to create, free or flush a user-owned mempool
  47   cache for non-EAL threads. Previously the caching was always disabled
  48   on these threads.
  49
  50 * **Changed the memory allocation scheme in the mempool library.**
  51
  52   * Added the ability to allocate a large mempool in fragmented virtual memory.
  53   * Added new APIs to populate a mempool with memory.
  54   * Added an API to free a mempool.
  55   * Modified the API of the ``rte_mempool_obj_iter()`` function.
  56   * Dropped the specific Xen Dom0 code.
  57   * Dropped the specific anonymous mempool code in testpmd.
  58
  59 * **Added a new driver for Broadcom NetXtreme-C devices.**
  60
  61   Added the new bnxt driver for Broadcom NetXtreme-C devices. See the
  62   "Network Interface Controller Drivers" document for more details on this
  63   new driver.
  64
  65 * **Added a new driver for ThunderX nicvf devices.**
  66
  67   Added the new thunderx net driver for ThunderX nicvf devices. See the
  68   "Network Interface Controller Drivers" document for more details on this new
  69   driver.
  70
  71 * **Added mailbox interrupt support for ixgbe and igb VFs.**
  72
  73   When the physical NIC link comes up or down, the PF driver will send a
  74   mailbox message to notify each VF. To handle this link up/down event,
  75   support have been added for a mailbox interrupt to receive the message and
  76   allow the application to register a callback for it.
  77
  78 * **Updated the ixgbe base driver.**
  79
  80   The ixgbe base driver was updated with changes including the
  81   following:
  82
  83   * Added sgmii link for X550.
  84   * Added MAC link setup for X550a SFP and SFP+.
  85   * Added KR support for X550em_a.
  86   * Added new PHY definitions for M88E1500.
  87   * Added support for the VLVF to be bypassed when adding/removing a VFTA entry.
  88   * Added X550a flow control auto negotiation support.
  89
  90 * **Updated the i40e base driver.**
  91
  92   Updated the i40e base driver including support for new devices IDs.
  93
  94 * **Updated the enic driver.**
  95
  96   The enic driver was updated with changes including the following:
  97
  98   * Optimized the Tx function.
  99   * Added Scattered Rx capability.
 100   * Improved packet type identification.
 101   * Added MTU update in non Scattered Rx mode and enabled MTU of up to 9208
 102     with UCS Software release 2.2 on 1300 series VICs.
 103
 104 * **Updated the mlx5 driver.**
 105
 106   The mlx5 driver was updated with changes including the following:
 107
 108   * Data path was refactored to bypass Verbs to improve RX and TX performance.
 109   * Removed compilation parameters for inline send, ``MLX5_MAX_INLINE``, and
 110     added command line parameter instead, ``txq_inline``.
 111   * Improved TX scatter gather support:
 112     Removed compilation parameter ``MLX5_PMD_SGE_WR_N``.
 113     Scatter-gather elements is set to the maximum value the NIC supports.
 114     Removed linearization logic, this decreases the memory consumption of the PMD.
 115   * Improved jumbo frames support, by dynamically setting RX scatter gather elements
 116     according to the MTU and mbuf size,
 117     no need for compilation parameter ``MLX5_PMD_SGE_WR_N``
 118
 119 * **Added support for virtio on IBM POWER8.**
 120
 121   The ioports are mapped in memory when using Linux UIO.
 122
 123 * **Added support for Virtio in containers.**
 124
 125   Add a new virtual device, named virtio_user, to support virtio for containers.
 126
 127   Known limitations:
 128
 129   * Control queue and multi-queue are not supported yet.
 130   * Doesn't work with ``--huge-unlink``.
 131   * Doesn't work with ``--no-huge``.
 132   * Doesn't work when there are more than ``VHOST_MEMORY_MAX_NREGIONS(8)`` hugepages.
 133   * Root privilege is required for sorting hugepages by physical address.
 134   * Can only be used with the vhost user backend.
 135
 136 * **Added vhost-user client mode.**
 137
 138   DPDK vhost-user now supports client mode as well as server mode. Client mode
 139   is enabled when the ``RTE_VHOST_USER_CLIENT`` flag is set while calling
 140   ``rte_vhost_driver_register``.
 141
 142   When DPDK vhost-user restarts from an normal or abnormal exit (such as a
 143   crash), the client mode allows DPDK to establish the connection again. Note
 144   that QEMU version v2.7 or above is required for this feature.
 145
 146   DPDK vhost-user will also try to reconnect by default when:
 147
 148   * The first connect fails (for example when QEMU is not started yet).
 149   * The connection is broken (for example when QEMU restarts).
 150
 151   It can be turned off by setting the ``RTE_VHOST_USER_NO_RECONNECT`` flag.
 152
 153 * **Added NSH packet recognition in i40e.**
 154
 155 * **Added AES-CTR support to AESNI MB PMD.**
 156
 157   Now AESNI MB PMD supports 128/192/256-bit counter mode AES encryption and
 158   decryption.
 159
 160 * **Added AES counter mode support for Intel QuickAssist devices.**
 161
 162   Enabled support for the AES CTR algorithm for Intel QuickAssist devices.
 163   Provided support for algorithm-chaining operations.
 164
 165 * **Added KASUMI SW PMD.**
 166
 167   A new Crypto PMD has been added, which provides KASUMI F8 (UEA1) ciphering
 168   and KASUMI F9 (UIA1) hashing.
 169
 170 * **Added multi-writer support for RTE Hash with Intel TSX.**
 171
 172   The following features/modifications have been added to rte_hash library:
 173
 174   * Enabled application developers to use an extra flag for ``rte_hash``
 175     creation to specify default behavior (multi-thread safe/unsafe) with the
 176     ``rte_hash_add_key`` function.
 177   * Changed the Cuckoo Hash Search algorithm to breadth first search for
 178     multi-writer routines and split Cuckoo Hash Search and Move operations in
 179     order to reduce transactional code region and improve TSX performance.
 180   * Added a hash multi-writer test case to the test app.
 181
 182 * **Improved IP Pipeline Application.**
 183
 184   The following features have been added to the ip_pipeline application:
 185
 186   * Configure the MAC address in the routing pipeline and automatic route
 187     updates with change in link state.
 188   * Enable RSS per network interface through the configuration file.
 189   * Streamline the CLI code.
 190
 191 * **Added keepalive enhancements.**
 192
 193   Added support for reporting of core states other than "dead" to
 194   monitoring applications, enabling the support of broader liveness
 195   reporting to external processes.
 196
 197 * **Added packet capture framework.**
 198
 199   * A new library ``librte_pdump`` is added to provide a packet capture API.
 200   * A new ``app/pdump`` tool is added to demonstrate capture packets in DPDK.
 201
 202
 203 * **Added floating VEB support for i40e PF driver.**
 204
 205   A "floating VEB" is a special Virtual Ethernet Bridge (VEB) which does not
 206   have an upload port, but instead is used for switching traffic between
 207   virtual functions (VFs) on a port.
 208
 209   For information on this feature,  please see the "I40E Poll Mode Driver"
 210   section of the "Network Interface Controller Drivers" document.
 211
 212 * **Added support for live migration of a VM with SRIOV VF.**
 213
 214   Live migration of a VM with Virtio and VF PMD's using the bonding PMD.
 215
 216
 217 Resolved Issues
 218 ---------------
 219
 220 .. This section should contain bug fixes added to the relevant sections. Sample format:
 221
 222    * **code/section Fixed issue in the past tense with a full stop.**
 223
 224      Add a short 1-2 sentence description of the resolved issue in the past tense.
 225      The title should contain the code/lib section like a commit message.
 226      Add the entries in alphabetic order in the relevant sections below.
 227
 228
 229 EAL
 230 ~~~
 231
 232 * **igb_uio: Fixed possible mmap failure for Linux >= 4.5.**
 233
 234   The mmaping of the iomem range of the PCI device fails for kernels that
 235   enabled the ``CONFIG_IO_STRICT_DEVMEM`` option. The error seen by the
 236   user is as similar to the following::
 237
 238       EAL: pci_map_resource():
 239
 240           cannot mmap(39, 0x7f1c51800000, 0x100000, 0x0):
 241           Invalid argument (0xffffffffffffffff)
 242
 243   The ``CONFIG_IO_STRICT_DEVMEM`` kernel option was introduced in Linux v4.5.
 244
 245   The issues was resolve by updating ``igb_uio`` to stop reserving PCI memory
 246   resources. From the kernel point of view the iomem region looks like idle
 247   and mmap works again. This matches the ``uio_pci_generic`` usage.
 248
 249
 250 Drivers
 251 ~~~~~~~
 252
 253 * **i40e: Fixed vlan stripping from inner header.**
 254
 255   Previously, for tunnel packets, such as VXLAN/NVGRE, the vlan
 256   tags of the inner header will be stripped without putting vlan
 257   info to descriptor.
 258   Now this issue is fixed by disabling vlan stripping from inner header.
 259
 260 * **i40e: Fixed the type issue of a single VLAN type.**
 261
 262   Currently, if a single VLAN header is added in a packet, it's treated
 263   as inner VLAN. But generally, a single VLAN header is treated as the
 264   outer VLAN header.
 265   This issue is fixed by changing corresponding register for single VLAN.
 266
 267 * **enic: Fixed several issues when stopping then restarting ports and queues.**
 268
 269   Fixed several crashes related to stopping then restarting ports and queues.
 270   Fixed possible crash when re-configuring the number of Rx queue descriptors.
 271
 272 * **enic: Fixed Rx data mis-alignment if mbuf data offset modified.**
 273
 274   Fixed possible Rx corruption when mbufs were returned to a pool with data
 275   offset other than RTE_PKTMBUF_HEADROOM.
 276
 277 * **enic: Fixed Tx IP/UDP/TCP checksum offload and VLAN insertion.**
 278
 279 * **enic: Fixed Rx error and missed counters.**
 280
 281
 282 Libraries
 283 ~~~~~~~~~
 284
 285 * **mbuf: Fixed refcnt update when detaching.**
 286
 287   Fix the ``rte_pktmbuf_detach()`` function to decrement the direct mbuf's
 288   reference counter. The previous behavior was not to affect the reference
 289   counter. This lead to a memory leak of the direct mbuf.
 290
 291
 292 Examples
 293 ~~~~~~~~
 294
 295
 296 Other
 297 ~~~~~
 298
 299
 300 Known Issues
 301 ------------
 302
 303 .. This section should contain new known issues in this release. Sample format:
 304
 305    * **Add title in present tense with full stop.**
 306
 307      Add a short 1-2 sentence description of the known issue in the present
 308      tense. Add information on any known workarounds.
 309
 310
 311 API Changes
 312 -----------
 313
 314 .. This section should contain API changes. Sample format:
 315
 316    * Add a short 1-2 sentence description of the API change. Use fixed width
 317      quotes for ``rte_function_names`` or ``rte_struct_names``. Use the past tense.
 318
 319 * The following counters are removed from the ``rte_eth_stats`` structure:
 320
 321   * ``ibadcrc``
 322   * ``ibadlen``
 323   * ``imcasts``
 324   * ``fdirmatch``
 325   * ``fdirmiss``
 326   * ``tx_pause_xon``
 327   * ``rx_pause_xon``
 328   * ``tx_pause_xoff``
 329   * ``rx_pause_xoff``
 330
 331 * The extended statistics are fetched by ids with ``rte_eth_xstats_get``
 332   after a lookup by name ``rte_eth_xstats_get_names``.
 333
 334 * The function ``rte_eth_dev_info_get`` fill the new fields ``nb_rx_queues``
 335   and ``nb_tx_queues`` in the structure ``rte_eth_dev_info``.
 336
 337 * The vhost function ``rte_vring_available_entries`` is renamed to
 338   ``rte_vhost_avail_entries``.
 339
 340 * All existing vhost APIs and callbacks with ``virtio_net`` struct pointer
 341   as the parameter have been changed due to the ABI refactoring described
 342   below. It is replaced by ``int vid``.
 343
 344 * The function ``rte_vhost_enqueue_burst`` no longer supports concurrent enqueuing
 345   packets to the same queue.
 346
 347 * The function ``rte_eth_dev_set_mtu`` adds a new return value ``-EBUSY``, which
 348   indicates the operation is forbidden because the port is running.
 349
 350 * The script ``dpdk_nic_bind.py`` is renamed to ``dpdk-devbind.py``.
 351   And the script ``setup.sh`` is renamed to ``dpdk-setup.sh``.
 352
 353
 354 ABI Changes
 355 -----------
 356
 357 .. * Add a short 1-2 sentence description of the ABI change that was announced in
 358      the previous releases and made in this release. Use fixed width quotes for
 359      ``rte_function_names`` or ``rte_struct_names``. Use the past tense.
 360
 361 * The ``rte_port_source_params`` structure has new fields to support PCAP files.
 362   It was already in release 16.04 with ``RTE_NEXT_ABI`` flag.
 363
 364 * The ``rte_eth_dev_info`` structure has new fields ``nb_rx_queues`` and ``nb_tx_queues``
 365   to support the number of queues configured by software.
 366
 367 * A Vhost ABI refactoring has been made: the ``virtio_net`` structure is no
 368   longer exported directly to the application. Instead, a handle, ``vid``, has
 369   been used to represent this structure internally.
 370
 371
 372 Shared Library Versions
 373 -----------------------
 374
 375 .. Update any library version updated in this release and prepend with a ``+`` sign.
 376
 377 The libraries prepended with a plus sign were incremented in this version.
 378
 379 .. code-block:: diff
 380
 381    + libethdev.so.4
 382      librte_acl.so.2
 383      librte_cfgfile.so.2
 384      librte_cmdline.so.2
 385      librte_cryptodev.so.1
 386      librte_distributor.so.1
 387      librte_eal.so.2
 388      librte_hash.so.2
 389      librte_ip_frag.so.1
 390      librte_ivshmem.so.1
 391      librte_jobstats.so.1
 392      librte_kni.so.2
 393      librte_kvargs.so.1
 394      librte_lpm.so.2
 395      librte_mbuf.so.2
 396    + librte_mempool.so.2
 397      librte_meter.so.1
 398      librte_pdump.so.1
 399      librte_pipeline.so.3
 400      librte_pmd_bond.so.1
 401      librte_pmd_ring.so.2
 402    + librte_port.so.3
 403      librte_power.so.1
 404      librte_reorder.so.1
 405      librte_ring.so.1
 406      librte_sched.so.1
 407      librte_table.so.2
 408      librte_timer.so.1
 409    + librte_vhost.so.3
 410
 411
 412 Tested Platforms
 413 ----------------
 414
 415 .. This section should contain a list of platforms that were tested with this
 416    release.
 417
 418    The format is:
 419
 420    #. Platform name.
 421
 422       - Platform details.
 423       - Platform details.
 424
 425 #. SuperMicro 1U
 426
 427    - BIOS: 1.0c
 428    - Processor: Intel(R) Atom(TM) CPU C2758 @ 2.40GHz
 429
 430 #. SuperMicro 1U
 431
 432    - BIOS: 1.0a
 433    - Processor: Intel(R) Xeon(R) CPU D-1540 @ 2.00GHz
 434    - Onboard NIC: Intel(R) X552/X557-AT (2x10G)
 435
 436      - Firmware-version: 0x800001cf
 437      - Device ID (PF/VF): 8086:15ad /8086:15a8
 438
 439    - kernel driver version: 4.2.5 (ixgbe)
 440
 441 #. SuperMicro 2U
 442
 443    - BIOS: 1.0a
 444    - Processor: Intel(R) Xeon(R) CPU E5-4667 v3 @ 2.00GHz
 445
 446 #. Intel(R) Server board S2600GZ
 447
 448    - BIOS: SE5C600.86B.02.02.0002.122320131210
 449    - Processor: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
 450
 451 #. Intel(R) Server board W2600CR
 452
 453    - BIOS: SE5C600.86B.02.01.0002.082220131453
 454    - Processor: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
 455
 456 #. Intel(R) Server board S2600CWT
 457
 458    - BIOS: SE5C610.86B.01.01.0009.060120151350
 459    - Processor: Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz
 460
 461 #. Intel(R) Server board S2600WTT
 462
 463    - BIOS: SE5C610.86B.01.01.0005.101720141054
 464    - Processor: Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz
 465
 466 #. Intel(R) Server board S2600WTT
 467
 468    - BIOS: SE5C610.86B.11.01.0044.090120151156
 469    - Processor: Intel(R) Xeon(R) CPU E5-2695 v4 @ 2.10GHz
 470
 471
 472 Tested NICs
 473 -----------
 474
 475 .. This section should contain a list of NICs that were tested with this release.
 476
 477    The format is:
 478
 479    #. NIC name.
 480
 481       - NIC details.
 482       - NIC details.
 483
 484 #. Intel(R) Ethernet Controller X540-AT2
 485
 486    - Firmware version: 0x80000389
 487    - Device id (pf): 8086:1528
 488    - Driver version: 3.23.2 (ixgbe)
 489
 490 #. Intel(R) 82599ES 10 Gigabit Ethernet Controller
 491
 492    - Firmware version: 0x61bf0001
 493    - Device id (pf/vf): 8086:10fb / 8086:10ed
 494    - Driver version: 4.0.1-k (ixgbe)
 495
 496 #. Intel(R) Corporation Ethernet Connection X552/X557-AT 10GBASE-T
 497
 498    - Firmware version: 0x800001cf
 499    - Device id (pf/vf): 8086:15ad / 8086:15a8
 500    - Driver version: 4.2.5 (ixgbe)
 501
 502 #. Intel(R) Ethernet Converged Network Adapter X710-DA4 (4x10G)
 503
 504    - Firmware version: 5.04
 505    - Device id (pf/vf): 8086:1572 / 8086:154c
 506    - Driver version: 1.4.26 (i40e)
 507
 508 #. Intel(R) Ethernet Converged Network Adapter X710-DA2 (2x10G)
 509
 510    - Firmware version: 5.04
 511    - Device id (pf/vf): 8086:1572 / 8086:154c
 512    - Driver version: 1.4.25 (i40e)
 513
 514 #. Intel(R) Ethernet Converged Network Adapter XL710-QDA1 (1x40G)
 515
 516    - Firmware version: 5.04
 517    - Device id (pf/vf): 8086:1584 / 8086:154c
 518    - Driver version: 1.4.25 (i40e)
 519
 520 #. Intel(R) Ethernet Converged Network Adapter XL710-QDA2 (2X40G)
 521
 522    - Firmware version: 5.04
 523    - Device id (pf/vf): 8086:1583 / 8086:154c
 524    - Driver version: 1.4.25 (i40e)
 525
 526 #. Intel(R) Corporation I350 Gigabit Network Connection
 527
 528    - Firmware version: 1.48, 0x800006e7
 529    - Device id (pf/vf): 8086:1521 / 8086:1520
 530    - Driver version: 5.2.13-k (igb)
 531
 532 #. Intel(R) Ethernet Multi-host Controller FM10000
 533
 534    - Firmware version: N/A
 535    - Device id (pf/vf): 8086:15d0
 536    - Driver version: 0.17.0.9 (fm10k)
 537
 538
 539 Tested OSes
 540 -----------
 541
 542 .. This section should contain a list of OSes that were tested with this release.
 543
 544 - CentOS 7.0
 545 - Fedora 23
 546 - Fedora 24
 547 - FreeBSD 10.3
 548 - Red Hat Enterprise Linux 7.2
 549 - SUSE Enterprise Linux 12
 550 - Ubuntu 15.10
 551 - Ubuntu 16.04 LTS
 552 - Wind River Linux 8