10. DPDK Release 18.11
10.1. New Features
Added support for using externally allocated memory in DPDK.
DPDK has added support for creating new
rte_malloc
heaps referencing memory that was created outside of DPDK’s own page allocator, and using that memory natively with any other DPDK library or data structure.Added check for ensuring allocated memory is addressable by devices.
Some devices can have addressing limitations so a new function,
rte_mem_check_dma_mask()
, has been added for checking that allocated memory is not out of the device range. Since memory can now be allocated dynamically after initialization, a DMA mask is stored and any new allocated memory will be checked against it and rejected if it is out of range. If more than one device has addressing limitations, the DMA mask is the more restrictive one.Updated the C11 memory model version of the ring library.
Added changes to decrease latency for architectures using the C11 memory model version of the ring library.
On Cavium ThunderX2 platform, the changes decreased latency by 27-29% and 3-15% for MPMC and SPSC cases respectively (with 2 lcores). The real improvements may vary with the number of contending lcores and the size of the ring.
Added hot-unplug handle mechanism.
Added
rte_dev_hotplug_handle_enable()
andrte_dev_hotplug_handle_disable()
for enabling or disabling the hotplug handle mechanism.Added support for device multi-process hotplug.
Added support for hotplug and hot-unplug in a multiprocessing scenario. Any ethdev devices created in the primary process will be regarded as shared and will be available for all DPDK processes. Synchronization between processes will be done using DPDK IPC.
Added new Flow API actions to rewrite fields in packet headers.
Added new Flow API actions to:
Modify source and destination IP addresses in the outermost IPv4/IPv6 headers.
Modify source and destination port numbers in the outermost TCP/UDP headers.
Added new Flow API action to swap MAC addresses in Ethernet header.
Added new Flow API action to swap the source and destination MAC addresses in the outermost Ethernet header.
Add support to offload more flow match and actions for CXGBE PMD.
The Flow API support has been enhanced for the CXGBE Poll Mode Driver to offload:
Match items: destination MAC address.
Action items: push/pop/rewrite vlan header, rewrite IP addresses in outermost IPv4/IPv6 header, rewrite port numbers in outermost TCP/UDP header, swap MAC addresses in outermost Ethernet header.
Added a devarg to use the latest supported vector path in i40e.
A new devarg
use-latest-supported-vec
was introduced to allow users to choose the latest vector path that the platform supported. For example, users can use AVX2 vector path on BDW/HSW to get better performance.Added support for SR-IOV in netvsc PMD.
The
netvsc
poll mode driver now supports the Accelerated Networking SR-IOV option in Hyper-V and Azure. This is an alternative to the previous vdev_netvsc, tap, and failsafe drivers combination.Added a new net driver for Marvell Armada 3k device.
Added the new
mvneta
net driver for Marvell Armada 3k device. See the MVNETA Poll Mode Driver NIC guide for more details on this new driver.Added NXP ENETC PMD.
Added the new enetc driver for the NXP enetc platform. See the ENETC Poll Mode Driver NIC driver guide for more details on this new driver.
Added Ethernet poll mode driver for Aquantia aQtion family of 10G devices.
Added the new
atlantic
ethernet poll mode driver for Aquantia XGBE devices. See the Aquantia Atlantic DPDK Driver NIC driver guide for more details on this driver.Updated mlx5 driver.
Updated the mlx5 driver including the following changes:
Improved security of PMD to prevent the NIC from getting stuck when the application misbehaves.
Reworked flow engine to supported e-switch flow rules (transfer attribute).
Added support for header re-write(L2-L4), VXLAN encap/decap, count, match on TCP flags and multiple flow groups with e-switch flow rules.
Added support for match on metadata, VXLAN and MPLS encap/decap with flow rules.
Added support for
RTE_ETH_DEV_CLOSE_REMOVE
flag to provide better support for representors.Added support for meson build.
Fixed build issue with PPC.
Added support for BlueField VF.
Added support for externally allocated static memory for DMA.
Updated Solarflare network PMD.
Updated the sfc_efx driver including the following changes:
Added support for Rx scatter in EF10 datapath implementation.
Added support for Rx descriptor status API in EF10 datapath implementation.
Added support for TSO in EF10 datapath implementation.
Added support for Tx descriptor status API in EF10 (ef10 and ef10_simple) datapaths implementation.
Updated the enic driver.
Added AVX2-based vectorized Rx handler.
Added VLAN and checksum offloads to the simple Tx handler.
Added the “count” flow action.
Enabled the virtual address IOVA mode.
Updated the failsafe driver.
Updated the failsafe driver including the following changes:
Added support for Rx and Tx queues start and stop.
Added support for Rx and Tx queues deferred start.
Added support for runtime Rx and Tx queues setup.
Added support multicast MAC address set.
Added a devarg to use a PCAP interface physical MAC address.
A new devarg
phy_mac
was introduced to allow users to use the physical MAC address of the selected PCAP interface.Added TAP Rx/Tx queues sharing with a secondary process.
Added support to allow a secondary process to attach a TAP device created in the primary process, probe the queues, and process Rx/Tx in a secondary process.
Added classification and metering support to SoftNIC PMD.
Added support for flow classification (rte_flow API), and metering and policing (rte_mtr API) to the SoftNIC PMD.
Added Crypto support to the Softnic PMD.
The Softnic is now capable of processing symmetric crypto workloads such as cipher, cipher-authentication chaining, and AEAD encryption and decryption. This is achieved by calling DPDK Cryptodev APIs.
Added cryptodev port to port library.
Cryptodev port is a shim layer in the port library that interacts with DPDK Cryptodev PMDs including burst enqueuing and dequeuing crypto operations.
Added symmetric cryptographic actions to the pipeline library.
In the pipeline library support was added for symmetric crypto action parsing and an action handler was implemented. The action allows automatic preparation of the crypto operation with the rules specified such as algorithm, key, and IV, etc. for the cryptodev port to process.
Updated the AESNI MB PMD.
The AESNI MB PMD has been updated with additional support for the AES-GCM algorithm.
Added NXP CAAM JR PMD.
Added the new caam job ring driver for NXP platforms. See the NXP CAAM JOB RING (caam_jr) guide for more details on this new driver.
Added support for GEN3 devices to Intel QAT driver.
Added support for the third generation of Intel QuickAssist devices.
Updated the QAT PMD.
The QAT PMD was updated with additional support for:
The AES-CMAC algorithm.
Added support for Dynamic Huffman Encoding to Intel QAT comp PMD.
The Intel QuickAssist (QAT) compression PMD has been updated with support for Dynamic Huffman Encoding for the Deflate algorithm.
Added Event Ethernet Tx Adapter.
Added event ethernet Tx adapter library that provides configuration and data path APIs for the ethernet transmit stage of an event driven packet processing application. These APIs abstract the implementation of the transmit stage and allow the application to use eventdev PMD support or a common implementation.
Added Distributed Software Eventdev PMD.
Added the new Distributed Software Event Device (DSW), which is a pure-software eventdev driver distributing the work of scheduling among all eventdev ports and the lcores using them. DSW, compared to the SW eventdev PMD, sacrifices load balancing performance to gain better event scheduling throughput and scalability.
Added extendable bucket feature to hash library (rte_hash).
This new “extendable bucket” feature provides 100% insertion guarantee to the capacity specified by the user by extending hash table with extra buckets when needed to accommodate the unlikely event of intensive hash collisions. In addition, the internal hashing algorithm was changed to use partial-key hashing to improve memory efficiency and lookup performance.
Added lock free reader/writer concurrency to hash library (rte_hash).
Lock free reader/writer concurrency prevents the readers from getting blocked due to a preempted writer thread. This allows the hash library to be used in scenarios where the writer thread runs on the control plane.
Added Traffic Pattern Aware Power Control Library.
Added an experimental library that extends the Power Library and provides empty_poll APIs. This feature measures how many times empty_polls are executed per core and uses the number of empty polls as a hint for system power management.
See the Power Management section of the DPDK Programmers Guide document for more information.
Added JSON power policy interface for containers.
Extended the Power Library and vm_power_manager sample app to allow power policies to be submitted via a FIFO using JSON formatted strings. Previously limited to Virtual Machines, this feature extends power policy functionality to containers and host applications that need to have their cores frequency controlled based on the rules contained in the policy.
Added Telemetry API.
Added a new telemetry API which allows applications to transparently expose their telemetry in JSON via a UNIX socket. The JSON can be consumed by any Service Assurance agent, such as CollectD.
Updated KNI kernel module, rte_kni library, and KNI sample application.
Updated the KNI kernel module with a new kernel module parameter,
carrier=[on|off]
to allow the user to control the default carrier state of the KNI kernel network interfaces. The default carrier state is now set tooff
, so the interfaces cannot be used until the carrier state is set toon
viarte_kni_update_link
or by writing1
to/sys/devices/virtual/net/<iface>/carrier
. In previous versions the default carrier state was left undefined. See Kernel NIC Interface for more information.Also added the new API function
rte_kni_update_link()
to allow the user to set the carrier state of the KNI kernel network interface.Also added a new command line flag
-m
to the KNI sample application to monitor and automatically reflect the physical NIC carrier state to the KNI kernel network interface with the newrte_kni_update_link()
API. See Kernel NIC Interface Sample Application for more information.Added ability to switch queue deferred start flag on testpmd app.
Added a console command to testpmd app, giving ability to switch
rx_deferred_start
ortx_deferred_start
flag of the specified queue of the specified port. The port must be stopped before the command call in order to reconfigure queues.Add a new sample application for vDPA.
The vdpa sample application creates vhost-user sockets by using the vDPA backend. vDPA stands for vhost Data Path Acceleration which utilizes virtio ring compatible devices to serve virtio driver directly to enable datapath acceleration. As vDPA driver can help to set up vhost datapath, this application doesn’t need to launch dedicated worker threads for vhost enqueue/dequeue operations.
Added cryptodev FIPS validation example application.
Added an example application to parse and perform symmetric cryptography computation to the NIST Cryptographic Algorithm Validation Program (CAVP) test vectors.
Allow unit test binary to take parameters from the environment.
The unit test “test”, or “dpdk-test”, binary is often called from scripts, which can make passing additional parameters, such as a coremask, difficult. Support has been added to the application to allow it to take additional command-line parameter values from the
DPDK_TEST_PARAMS
environment variable to make this application easier to use.
10.2. API Changes
eal:
rte_memseg_list
structure now has an additional flag indicating whether the memseg list is externally allocated. This will have implications for any users of memseg-walk-related functions, as they will now have to skip externally allocated segments in most cases if the intent is to only iterate over internal DPDK memory.In addition the
socket_id
parameter across the entire DPDK has gained additional meaning, as some socket ID’s will now be representing externally allocated memory. No changes will be required for existing code as backwards compatibility will be kept, and those who do not use this feature will not see these extra socket ID’s. Any new API’s must not check socket ID parameters themselves, and must instead leave it to the memory subsystem to decide whether socket ID is a valid one.eal: The following devargs functions, which were deprecated in 18.05, were removed in 18.11:
rte_eal_parse_devargs_str()
,rte_eal_devargs_add()
,rte_eal_devargs_type_count()
, andrte_eal_devargs_dump()
.eal: The parameters of the function
rte_devargs_remove()
have changed from bus and device names tostruct rte_devargs
.eal: The deprecated functions attach/detach were removed in 18.11.
rte_eal_dev_attach
can be replaced byrte_dev_probe
orrte_eal_hotplug_add
.rte_eal_dev_detach
can be replaced byrte_dev_remove
orrte_eal_hotplug_remove
.eal: The scope of
rte_eal_hotplug_add()
/rte_dev_probe()
andrte_eal_hotplug_remove()
/rte_dev_remove()
has been extended. In the multi-process model, they will guarantee that the device is attached or detached on all processes.mbuf: The
__rte_mbuf_raw_free()
and__rte_pktmbuf_prefree_seg()
functions were deprecated since 17.05 and are replaced byrte_mbuf_raw_free()
andrte_pktmbuf_prefree_seg()
.ethdev: The deprecated functions attach/detach were removed in 18.11.
rte_eth_dev_attach()
can be replaced byRTE_ETH_FOREACH_MATCHING_DEV
andrte_dev_probe()
orrte_eal_hotplug_add()
.rte_eth_dev_detach()
can be replaced byrte_dev_remove()
orrte_eal_hotplug_remove()
.ethdev: A call to
rte_eth_dev_release_port()
has been added inrte_eth_dev_close()
. As a consequence, a closed port is freed and seen as invalid because of its stateRTE_ETH_DEV_UNUSED
. This new behavior is enabled per driver for a migration period.A new device flag,
RTE_ETH_DEV_NOLIVE_MAC_ADDR
, changes the order of actions insiderte_eth_dev_start()
regarding MAC set. Some NICs do not support MAC changes once the port has started and with this new device flag the MAC can be properly configured in any case. This is particularly important for bonding.The default behavior of CRC strip offload has changed in this release. Without any specific Rx offload flag, default behavior by a PMD is now to strip CRC.
DEV_RX_OFFLOAD_CRC_STRIP
offload flag has been removed. To request keeping CRC, application should setDEV_RX_OFFLOAD_KEEP_CRC
Rx offload.eventdev: The type of the second parameter to
rte_event_eth_rx_adapter_caps_get()
has been changed from uint8_t to uint16_t.kni: By default, interface carrier status is
off
which means there won’t be any traffic. It can be set toon
viarte_kni_update_link()
API or viasysfs
interface:echo 1 > /sys/class/net/vEth0/carrier
.Note interface should be
up
to be able to read/write sysfs interface. When KNI sample application is used,-m
parameter can be used to automatically update the carrier status for the interface.kni: When ethtool support is enabled (
CONFIG_RTE_KNI_KMOD_ETHTOOL=y
) ethtool commandsETHTOOL_GSET & ETHTOOL_SSET
are no longer supported for kernels that haveETHTOOL_GLINKSETTINGS & ETHTOOL_SLINKSETTINGS
support. This meansethtool "-a|--show-pause", "-s|--change"
won’t work, andethtool <iface>
output will have less information.
10.3. ABI Changes
eal: added
legacy_mem
andsingle_file_segments
values torte_config
structure on account of improving DPDK usability when using either--legacy-mem
or--single-file-segments
flags.eal: EAL library ABI version was changed due to previously announced work on supporting external memory in DPDK:
Structure
rte_memseg_list
now has a new field indicating length of memory addressed by the segment listStructure
rte_memseg_list
now has a new flag indicating whether the memseg list refers to external memoryStructure
rte_malloc_heap
now has a new field indicating socket ID the malloc heap belongs toStructure
rte_mem_config
has had itsmalloc_heaps
array resized fromRTE_MAX_NUMA_NODES
toRTE_MAX_HEAPS
valueStructure
rte_malloc_heap
now has aheap_name
memberStructure
rte_eal_memconfig
has been extended to contain next socket ID for externally allocated segments
eal: Added
dma_maskbits
torte_mem_config
for keeping the most restrictive DMA mask based on the devices addressing limitations.eal: The structure
rte_device
has a new field to reference arte_bus
. It thus changes the size of thestruct rte_device
and the inherited device structures of all buses.
10.5. Known Issues
When using SR-IOV (VF) support with netvsc PMD and the Mellanox mlx5 bifurcated driver the Linux netvsc device must be brought up before the netvsc device is unbound and passed to the DPDK.
IBM Power8 is not supported in this release of DPDK. IBM Power9 is supported.
AVX-512
support has been disabled forGCC
builds [1] because of a crash [2]. This can affectnative
machine type build targets on the platforms that supportAVX512F
likeIntel Skylake
processors, and can cause a possible performance drop. The immediate workaround is to useclang
compiler on these platforms. The issue has been identified as a GCC defect and reported to the GCC community [3]. Further actions will be taken based on the GCC defect result.[1]: Commit 8d07c82b239f (“mk: disable gcc AVX512F support”)
10.6. Tested Platforms
Intel(R) platforms with Intel(R) NICs combinations
CPU
Intel(R) Atom(TM) CPU C3758 @ 2.20GHz
Intel(R) Xeon(R) CPU D-1541 @ 2.10GHz
Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz
Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz
Intel(R) Xeon(R) Platinum 8180 CPU @ 2.50GHz
OS:
CentOS 7.5
Fedora 25
Fedora 28
FreeBSD 11.2
Red Hat Enterprise Linux Server release 7.5
Open SUSE 15
Wind River Linux 8
Ubuntu 14.04
Ubuntu 16.04
Ubuntu 16.10
Ubuntu 17.10
Ubuntu 18.04
NICs:
Intel(R) 82599ES 10 Gigabit Ethernet Controller
Firmware version: 0x61bf0001
Device id (pf/vf): 8086:10fb / 8086:10ed
Driver version: 5.2.3 (ixgbe)
Intel(R) Corporation Ethernet Connection X552/X557-AT 10GBASE-T
Firmware version: 0x800003e7
Device id (pf/vf): 8086:15ad / 8086:15a8
Driver version: 4.4.6 (ixgbe)
Intel(R) Ethernet Converged Network Adapter X710-DA4 (4x10G)
Firmware version: 6.01 0x80003221
Device id (pf/vf): 8086:1572 / 8086:154c
Driver version: 2.4.6 (i40e)
Intel(R) Corporation Ethernet Connection X722 for 10GbE SFP+ (4x10G)
Firmware version: 3.33 0x80000fd5 0.0.0
Device id (pf/vf): 8086:37d0 / 8086:37cd
Driver version: 2.4.6 (i40e)
Intel(R) Ethernet Converged Network Adapter XXV710-DA2 (2x25G)
Firmware version: 6.01 0x80003221
Device id (pf/vf): 8086:158b / 8086:154c
Driver version: 2.4.6 (i40e)
Intel(R) Ethernet Converged Network Adapter XL710-QDA2 (2X40G)
Firmware version: 6.01 0x8000321c
Device id (pf/vf): 8086:1583 / 8086:154c
Driver version: 2.4.6 (i40e)
Intel(R) Corporation I350 Gigabit Network Connection
Firmware version: 1.63, 0x80000dda
Device id (pf/vf): 8086:1521 / 8086:1520
Driver version: 5.4.0-k (igb)
Intel(R) platforms with Mellanox(R) NICs combinations
CPU:
Intel(R) Xeon(R) Gold 6154 CPU @ 3.00GHz
Intel(R) Xeon(R) CPU E5-2697A v4 @ 2.60GHz
Intel(R) Xeon(R) CPU E5-2697 v3 @ 2.60GHz
Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz
Intel(R) Xeon(R) CPU E5-2640 @ 2.50GHz
Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz
OS:
Red Hat Enterprise Linux Server release 7.6 (Maipo)
Red Hat Enterprise Linux Server release 7.5 (Maipo)
Red Hat Enterprise Linux Server release 7.4 (Maipo)
Red Hat Enterprise Linux Server release 7.3 (Maipo)
Red Hat Enterprise Linux Server release 7.2 (Maipo)
Ubuntu 18.10
Ubuntu 18.04
Ubuntu 17.10
Ubuntu 16.04
SUSE Linux Enterprise Server 15
MLNX_OFED: 4.4-2.0.1.0
MLNX_OFED: 4.5-0.3.1.0
NICs:
Mellanox(R) ConnectX(R)-3 Pro 40G MCX354A-FCC_Ax (2x40G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1007
Firmware version: 2.42.5000
Mellanox(R) ConnectX(R)-4 10G MCX4111A-XCAT (1x10G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 10G MCX4121A-XCAT (2x10G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 25G MCX4111A-ACAT (1x25G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 25G MCX4121A-ACAT (2x25G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 40G MCX4131A-BCAT/MCX413A-BCAT (1x40G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 40G MCX415A-BCAT (1x40G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 50G MCX4131A-GCAT/MCX413A-GCAT (1x50G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 50G MCX414A-BCAT (2x50G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 50G MCX415A-GCAT/MCX416A-BCAT/MCX416A-GCAT (2x50G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 50G MCX415A-CCAT (1x100G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 100G MCX416A-CCAT (2x100G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:1013
Firmware version: 12.23.8022 and above
Mellanox(R) ConnectX(R)-4 Lx 10G MCX4121A-XCAT (2x10G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1015
Firmware version: 14.23.8022 and above
Mellanox(R) ConnectX(R)-4 Lx 25G MCX4121A-ACAT (2x25G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1015
Firmware version: 14.23.8022 and above
Mellanox(R) ConnectX(R)-5 100G MCX556A-ECAT (2x100G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:1017
Firmware version: 16.23.8022 and above
Mellanox(R) ConnectX(R)-5 Ex EN 100G MCX516A-CDAT (2x100G)
Host interface: PCI Express 4.0 x16
Device ID: 15b3:1019
Firmware version: 16.23.8022 and above
ARM platforms with Mellanox(R) NICs combinations
CPU:
Qualcomm ARM 1.1 2500MHz
OS:
Red Hat Enterprise Linux Server release 7.5 (Maipo)
NICs:
Mellanox(R) ConnectX(R)-4 Lx 25G MCX4121A-ACAT (2x25G)
Host interface: PCI Express 3.0 x8
Device ID: 15b3:1015
Firmware version: 14.24.0220
Mellanox(R) ConnectX(R)-5 100G MCX556A-ECAT (2x100G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:1017
Firmware version: 16.24.0220
Mellanox(R) BlueField SmartNIC
Mellanox(R) BlueField SmartNIC MT416842 (2x25G)
Host interface: PCI Express 3.0 x16
Device ID: 15b3:a2d2
Firmware version: 18.24.0246
SoC ARM cores running OS:
CentOS Linux release 7.4.1708 (AltArch)
MLNX_OFED 4.4-2.5.3.0
DPDK application running on ARM cores inside SmartNIC
ARM SoC combinations from NXP (with integrated NICs)
SoC:
NXP/Freescale QorIQ LS1046A with ARM Cortex A72
NXP/Freescale QorIQ LS2088A with ARM Cortex A72
OS:
Ubuntu 18.04.1 LTS with NXP QorIQ LSDK 1809 support packages
Ubuntu 16.04.3 LTS with NXP QorIQ LSDK 1803 support packages