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