7 * **Introduce ARMv7 and ARMv8 architectures**
9 * It is now possible to build DPDK for the ARMv7 and ARMv8 platforms.
10 * ARMv7 can be tested with virtual PMD drivers.
11 * ARMv8 can be tested with virtual and physical PMD drivers.
13 * **Enabled freeing of ring.**
15 New function rte_ring_free() allows the user to free a ring
16 if it was created with rte_ring_create().
18 * **Added keepalive support to EAL and example application.**
20 * **Added experimental cryptodev API**
22 The cryptographic processing of packet is provided as a preview
25 * Intel QuickAssist devices
26 * Intel AES-NI multi-buffer library
28 Due to its experimental state, the API may change without prior notice.
30 * **Added ethdev API to support IEEE1588.**
32 Added functions to read, write and adjust system time in the NIC.
34 * **Extended Statistics**
36 Define extended statistics naming scheme to store metadata in the name
37 string name of each statistic, refer to the Extended Statistics section
38 of the programmers guide. Implemented the extended stats API for these
48 * **Added API in ethdev to retrieve RX/TX queue information.**
50 * Add the ability for the upper layer to query RX/TX queue information.
51 * Add into rte_eth_dev_info new fields to represent information about
52 RX/TX descriptors min/max/align numbers per queue for the device.
54 * **Added RSS dynamic configuration to bonding.**
56 * **Added e1000 Rx interrupt support.**
58 * **Added igb TSO support for both PF and VF.**
60 * **RSS enhancement on Intel x550 NIC**
62 * Support 512 entries RSS redirection table.
63 * Support per VF RSS redirection table.
65 * **Flow director enhancement on Intel x550 NIC**
67 * Add 2 new flow director modes on x550.
68 * One is MAC VLAN mode, the other is tunnel mode.
70 * **Added i40e vector RX/TX.**
72 * **Added i40e Rx interrupt support.**
74 * **Added i40e flow control support.**
76 * **Added DCB support to i40e PF driver.**
78 * **Added RSS/FD input set granularity on Intel X710/XL710.**
80 * **Added different GRE key length for input set on Intel X710/XL710.**
82 * **Added flow director support in i40e VF.**
84 * **Added fm10k vector RX/TX.**
86 * **Added fm10k TSO support for both PF and VF.**
88 * **Added fm10k VMDQ support.**
90 * **New NIC Boulder Rapid support.**
92 Boulder Rapid is a new NIC of Intel's fm10k family.
94 * **Enhanced support for the Chelsio CXGBE driver.**
96 * Added support for Jumbo Frames.
97 * Optimize forwarding performance for Chelsio T5 40GbE cards.
99 * **Improved enic Tx packet rate.**
101 Reduced frequency of Tx tail pointer updates to the NIC.
103 * **Added support for Mellanox ConnectX-4 adapters (mlx5).**
105 The mlx5 poll-mode driver implements support for Mellanox ConnectX-4 EN
106 and Mellanox ConnectX-4 Lx EN families of 10/25/40/50/100 Gb/s adapters.
108 Like mlx4, this PMD is only available for Linux and is disabled by default
109 due to external dependencies (libibverbs and libmlx5).
111 * **Added driver for Netronome nfp-6xxx card.**
113 Support for using Netronome nfp-6xxx with PCI VFs.
115 * **Added virtual szedata2 driver for COMBO cards.**
117 Added virtual PMD for COMBO-100G and COMBO-80G cards.
118 PMD is disabled in default configuration.
120 * **Enhanced support for virtio driver.**
122 * Virtio ring layout optimization (fixed avail ring)
126 * **Added vhost-user multiple queue support.**
128 * **Added port hotplug support to vmxnet3.**
130 * **Added port hotplug support to xenvirt.**
132 * **Added ethtool shim and sample application.**
134 * **Added experimental performance thread example application.**
136 The application demonstrates L3 fowarding with different threading
137 models: pthreads, cgroups, or lighweight threads. The example inludes
138 a simple cooperative scheduler.
140 Due to its experimental state this application may change without notice.
141 The application is supported only for Linux x86_64.
150 * **eal/linux: Fixed epoll timeout.**
152 Fixed issue where the ``rte_epoll_wait()`` function didn't return when the
153 underlying call to ``epoll_wait()`` timed out.
159 * **igb: Fixed IEEE1588 frame identification in I210.**
161 Fixed issue where the flag PKT_RX_IEEE1588_PTP was not being set
162 in Intel I210 NIC, as EtherType in RX descriptor is in bits 8:10 of
163 Packet Type and not in the default bits 0:2.
165 * **igb: Fixed VF start with PF stopped.**
167 VF needs the PF interrupt support initialized even if not started.
169 * **igb: Fixed VF MAC address when using with DPDK PF.**
171 Assign a random MAC address in VF when not assigned by PF.
173 * **ixgbe: Fixed issue with X550 DCB.**
175 Fixed a DCB issue with x550 where for 8 TCs (Traffic Classes), if a packet
176 with user priority 6 or 7 was injected to the NIC, then the NIC would only
177 put 3 packets into the queue. There was also a similar issue for 4 TCs.
179 * **ixgbe: Removed burst size restriction of vector RX.**
181 Fixed issue where a burst size less than 32 didn't receive anything.
183 * **ixgbe: Fixed VF start with PF stopped.**
185 VF needs the PF interrupt support initialized even if not started.
187 * **ixgbe: Fixed TX hang when RS distance exceeds HW limit.**
189 Fixed an issue when TX queue can hang when a lot of highly fragmented
190 packets have to be sent.
191 As part of that fix, tx_rs_thresh for ixgbe PMD is not allowed to be greater
192 then to 32 to comply with HW restrictions.
194 * **i40e: Fixed base driver allocation when not using first numa node.**
196 Fixed i40e issue that occurred when a DPDK application didn't initialize
197 ports if memory wasn't available on socket 0.
199 * **i40e: Fixed maximum of 64 queues per port.**
201 Fixed the issue in i40e of cannot supporting more than 64 queues per port,
202 though hardware actually supports that. The real number of queues may vary,
203 as long as the total number of queues used in PF, VFs, VMDq and FD does not
204 exceeds the hardware maximum.
206 * **i40e: Fixed statistics of packets.**
208 Added discarding packets on VSI to the stats and rectify the old statistics.
210 * **i40e: Fixed issue of not freeing memzone.**
212 Fixed the issue of not freeing memzone in the call to free the memory for
215 * **mlx: Fixed driver loading.**
217 The mlx drivers were unable to load when built as a shared library,
218 due to a missing symbol in mempool library.
220 * **vhost: Fixed Qemu shutdown.**
222 Fixed issue with libvirt ``virsh destroy`` not killing the VM.
224 * **virtio: Fixed crash after changing link state.**
226 Fixed io permission in the interrupt handler.
228 * **virtio: Fixed crash when releasing queue.**
230 Fixed issue when releasing null control queue.
236 * **hash: Fixed memory allocation of Cuckoo Hash key table.**
238 Fixed issue where an incorrect Cuckoo Hash key table size could be
239 calculated limiting the size to 4GB.
241 * **hash: Fixed incorrect lookup if key is all zero.**
243 Fixed issue in hash library that occurred if an all zero
244 key was not added in the table and the key was looked up,
245 resulting in an incorrect hit.
247 * **hash: Fixed thread scaling by reducing contention.**
249 Fixed issue in hash library where, using multiple cores with
250 hardware transactional memory support, thread scaling did not work,
251 due to the global ring that is shared by all cores.
257 * **l3fwd: Fixed crash with IPv6.**
259 * **vhost_xen: Fixed compile error.**
265 * This release drops compatibility with Linux kernel 2.6.33. The minimum
266 kernel requirement is now 2.6.34.
272 * Some drivers do not fill the packet type when receiving.
273 As the l3fwd example application requires this info, the i40e vector
274 driver must be disabled to benefit of the packet type with i40e.
280 * The deprecated flow director API is removed.
281 It was replaced by rte_eth_dev_filter_ctrl().
283 * The dcb_queue is renamed to dcb_tc in following dcb configuration
284 structures: rte_eth_dcb_rx_conf, rte_eth_dcb_tx_conf,
285 rte_eth_vmdq_dcb_conf, rte_eth_vmdq_dcb_tx_conf.
287 * The rte_eth_rx_queue_count() function now returns "int" instead of "uint32_t"
288 to allow the use of negative values as error codes on return.
290 * The function rte_eal_pci_close_one() is removed.
291 It was replaced by rte_eal_pci_detach().
293 * The deprecated ACL API ipv4vlan is removed.
295 * The deprecated hash function rte_jhash2() is removed.
296 It was replaced by rte_jhash_32b().
298 * The deprecated KNI functions are removed:
299 rte_kni_create(), rte_kni_get_port_id() and rte_kni_info_get().
301 * The deprecated ring PMD functions are removed:
302 rte_eth_ring_pair_create() and rte_eth_ring_pair_attach().
304 * The devargs union field virtual is renamed to virt for C++ compatibility.
309 * The EAL and ethdev structures rte_intr_handle and rte_eth_conf were changed
310 to support Rx interrupt. It was already done in 2.1 for CONFIG_RTE_NEXT_ABI.
312 * The ethdev flow director entries for SCTP were changed.
313 It was already done in 2.1 for CONFIG_RTE_NEXT_ABI.
315 * The ethdev flow director structure rte_eth_fdir_flow_ext structure is changed.
316 New fields are added to support flow director filtering in VF.
318 * The size of the ethdev structure rte_eth_hash_filter_info is changed
319 by adding a new element rte_eth_input_set_conf in an union.
321 * The new fields rx_desc_lim and tx_desc_lim are added into rte_eth_dev_info
324 * For debug builds, the functions rte_eth_rx_burst(), rte_eth_tx_burst()
325 rte_eth_rx_descriptor_done() and rte_eth_rx_queue_count() will
326 no longer be separate functions in the DPDK libraries. Instead, they will
327 only be present in the rte_ethdev.h header file.
329 * The maximum number of queues per port CONFIG_RTE_MAX_QUEUES_PER_PORT is
332 * The mbuf structure was changed to support unified packet type.
333 It was already done in 2.1 for CONFIG_RTE_NEXT_ABI.
335 * The dummy malloc library is removed. The content was moved into EAL in 2.1.
337 * The LPM structure is changed. The deprecated field mem_location is removed.
339 * librte_table LPM: A new parameter to hold the table name will be added to
340 the LPM table parameter structure.
342 * librte_table hash: The key mask parameter is added to the hash table
343 parameter structure for 8-byte key and 16-byte key extendible bucket
346 * librte_port: Macros to access the packet meta-data stored within the packet
347 buffer has been adjusted to cover the packet mbuf structure.
349 * librte_cfgfile: Allow longer names and values by increasing the constants
350 CFG_NAME_LEN and CFG_VALUE_LEN to 64 and 256 respectively.
353 Shared Library Versions
354 -----------------------
356 The libraries prepended with a plus sign were incremented in this version.
362 + librte_cfgfile.so.2
364 librte_distributor.so.1
376 + librte_pipeline.so.2
378 + librte_pmd_ring.so.2