2 Copyright 2015 6WIND S.A.
3 Copyright 2015 Mellanox
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11 * Redistributions in binary form must reproduce the above copyright
12 notice, this list of conditions and the following disclaimer in
13 the documentation and/or other materials provided with the
15 * Neither the name of 6WIND S.A. nor the names of its
16 contributors may be used to endorse or promote products derived
17 from this software without specific prior written permission.
19 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
34 The MLX5 poll mode driver library (**librte_pmd_mlx5**) provides support
35 for **Mellanox ConnectX-4**, **Mellanox ConnectX-4 Lx** and **Mellanox
36 ConnectX-5** families of 10/25/40/50/100 Gb/s adapters as well as their
37 virtual functions (VF) in SR-IOV context.
39 Information and documentation about these adapters can be found on the
40 `Mellanox website <http://www.mellanox.com>`__. Help is also provided by the
41 `Mellanox community <http://community.mellanox.com/welcome>`__.
43 There is also a `section dedicated to this poll mode driver
44 <http://www.mellanox.com/page/products_dyn?product_family=209&mtag=pmd_for_dpdk>`__.
48 Due to external dependencies, this driver is disabled by default. It must
49 be enabled manually by setting ``CONFIG_RTE_LIBRTE_MLX5_PMD=y`` and
52 Implementation details
53 ----------------------
55 Besides its dependency on libibverbs (that implies libmlx5 and associated
56 kernel support), librte_pmd_mlx5 relies heavily on system calls for control
57 operations such as querying/updating the MTU and flow control parameters.
59 For security reasons and robustness, this driver only deals with virtual
60 memory addresses. The way resources allocations are handled by the kernel
61 combined with hardware specifications that allow it to handle virtual memory
62 addresses directly ensure that DPDK applications cannot access random
63 physical memory (or memory that does not belong to the current process).
65 This capability allows the PMD to coexist with kernel network interfaces
66 which remain functional, although they stop receiving unicast packets as
67 long as they share the same MAC address.
68 This means legacy linux control tools (for example: ethtool, ifconfig and
69 more) can operate on the same network interfaces that owned by the DPDK
72 Enabling librte_pmd_mlx5 causes DPDK applications to be linked against
78 - Multi arch support: x86_64, POWER8, ARMv8.
79 - Multiple TX and RX queues.
80 - Support for scattered TX and RX frames.
81 - IPv4, IPv6, TCPv4, TCPv6, UDPv4 and UDPv6 RSS on any number of queues.
82 - Several RSS hash keys, one for each flow type.
83 - Configurable RETA table.
84 - Support for multiple MAC addresses.
88 - RX CRC stripping configuration.
90 - Multicast promiscuous mode.
91 - Hardware checksum offloads.
92 - Flow director (RTE_FDIR_MODE_PERFECT, RTE_FDIR_MODE_PERFECT_MAC_VLAN and
96 - KVM and VMware ESX SR-IOV modes are supported.
97 - RSS hash result is supported.
99 - Hardware checksum TX offload for VXLAN and GRE.
101 - Statistics query including Basic, Extended and per queue.
107 - Inner RSS for VXLAN frames is not supported yet.
108 - Hardware checksum RX offloads for VXLAN inner header are not supported yet.
109 - For secondary process:
111 - Forked secondary process not supported.
112 - All mempools must be initialized before rte_eth_dev_start().
114 - Flow pattern without any specific vlan will match for vlan packets as well:
116 When VLAN spec is not specified in the pattern, the matching rule will be created with VLAN as a wild card.
117 Meaning, the flow rule::
119 flow create 0 ingress pattern eth / vlan vid is 3 / ipv4 / end ...
121 Will only match vlan packets with vid=3. and the flow rules::
123 flow create 0 ingress pattern eth / ipv4 / end ...
127 flow create 0 ingress pattern eth / vlan / ipv4 / end ...
129 Will match any ipv4 packet (VLAN included).
131 - A multi segment packet must have less than 6 segments in case the Tx burst function
132 is set to multi-packet send or Enhanced multi-packet send. Otherwise it must have
133 less than 50 segments.
134 - Count action for RTE flow is **only supported in Mellanox OFED**.
135 - Flows with a VXLAN Network Identifier equal (or ends to be equal)
136 to 0 are not supported.
137 - VXLAN TSO and checksum offloads are not supported on VM.
142 MLX5 supports various of methods to report statistics:
144 Port statistics can be queried using ``rte_eth_stats_get()``. The port statistics are through SW only and counts the number of packets received or sent successfully by the PMD.
146 Extended statistics can be queried using ``rte_eth_xstats_get()``. The extended statistics expose a wider set of counters counted by the device. The extended port statistics counts the number of packets received or sent successfully by the port. As Mellanox NICs are using the :ref:`Bifurcated Linux Driver <linux_gsg_linux_drivers>` those counters counts also packet received or sent by the Linux kernel. The counters with ``_phy`` suffix counts the total events on the physical port, therefore not valid for VF.
148 Finally per-flow statistics can by queried using ``rte_flow_query`` when attaching a count action for specific flow. The flow counter counts the number of packets received successfully by the port and match the specific flow.
156 These options can be modified in the ``.config`` file.
158 - ``CONFIG_RTE_LIBRTE_MLX5_PMD`` (default **n**)
160 Toggle compilation of librte_pmd_mlx5 itself.
162 - ``CONFIG_RTE_LIBRTE_MLX5_DLOPEN_DEPS`` (default **n**)
164 Build PMD with additional code to make it loadable without hard
165 dependencies on **libibverbs** nor **libmlx5**, which may not be installed
166 on the target system.
168 In this mode, their presence is still required for it to run properly,
169 however their absence won't prevent a DPDK application from starting (with
170 ``CONFIG_RTE_BUILD_SHARED_LIB`` disabled) and they won't show up as
171 missing with ``ldd(1)``.
173 It works by moving these dependencies to a purpose-built rdma-core "glue"
174 plug-in which must either be installed in a directory whose name is based
175 on ``CONFIG_RTE_EAL_PMD_PATH`` suffixed with ``-glue`` if set, or in a
176 standard location for the dynamic linker (e.g. ``/lib``) if left to the
177 default empty string (``""``).
179 This option has no performance impact.
181 - ``CONFIG_RTE_LIBRTE_MLX5_DEBUG`` (default **n**)
183 Toggle debugging code and stricter compilation flags. Enabling this option
184 adds additional run-time checks and debugging messages at the cost of
187 - ``CONFIG_RTE_LIBRTE_MLX5_TX_MP_CACHE`` (default **8**)
189 Maximum number of cached memory pools (MPs) per TX queue. Each MP from
190 which buffers are to be transmitted must be associated to memory regions
191 (MRs). This is a slow operation that must be cached.
193 This value is always 1 for RX queues since they use a single MP.
195 Environment variables
196 ~~~~~~~~~~~~~~~~~~~~~
200 A list of directories in which to search for the rdma-core "glue" plug-in,
201 separated by colons or semi-colons.
203 Only matters when compiled with ``CONFIG_RTE_LIBRTE_MLX5_DLOPEN_DEPS``
204 enabled and most useful when ``CONFIG_RTE_EAL_PMD_PATH`` is also set,
205 since ``LD_LIBRARY_PATH`` has no effect in this case.
207 - ``MLX5_PMD_ENABLE_PADDING``
209 Enables HW packet padding in PCI bus transactions.
211 When packet size is cache aligned and CRC stripping is enabled, 4 fewer
212 bytes are written to the PCI bus. Enabling padding makes such packets
215 In cases where PCI bandwidth is the bottleneck, padding can improve
218 This is disabled by default since this can also decrease performance for
219 unaligned packet sizes.
221 - ``MLX5_SHUT_UP_BF``
223 Configures HW Tx doorbell register as IO-mapped.
225 By default, the HW Tx doorbell is configured as a write-combining register.
226 The register would be flushed to HW usually when the write-combining buffer
227 becomes full, but it depends on CPU design.
229 Except for vectorized Tx burst routines, a write memory barrier is enforced
230 after updating the register so that the update can be immediately visible to
233 When vectorized Tx burst is called, the barrier is set only if the burst size
234 is not aligned to MLX5_VPMD_TX_MAX_BURST. However, setting this environmental
235 variable will bring better latency even though the maximum throughput can
238 Run-time configuration
239 ~~~~~~~~~~~~~~~~~~~~~~
241 - librte_pmd_mlx5 brings kernel network interfaces up during initialization
242 because it is affected by their state. Forcing them down prevents packets
245 - **ethtool** operations on related kernel interfaces also affect the PMD.
247 - ``rxq_cqe_comp_en`` parameter [int]
249 A nonzero value enables the compression of CQE on RX side. This feature
250 allows to save PCI bandwidth and improve performance. Enabled by default.
254 - x86_64 with ConnectX-4, ConnectX-4 LX and ConnectX-5.
255 - POWER8 and ARMv8 with ConnectX-4 LX and ConnectX-5.
257 - ``txq_inline`` parameter [int]
259 Amount of data to be inlined during TX operations. Improves latency.
260 Can improve PPS performance when PCI back pressure is detected and may be
261 useful for scenarios involving heavy traffic on many queues.
263 Because additional software logic is necessary to handle this mode, this
264 option should be used with care, as it can lower performance when back
265 pressure is not expected.
267 - ``txqs_min_inline`` parameter [int]
269 Enable inline send only when the number of TX queues is greater or equal
272 This option should be used in combination with ``txq_inline`` above.
274 On ConnectX-4, ConnectX-4 LX and ConnectX-5 without Enhanced MPW:
276 - Disabled by default.
277 - In case ``txq_inline`` is set recommendation is 4.
279 On ConnectX-5 with Enhanced MPW:
281 - Set to 8 by default.
283 - ``txq_mpw_en`` parameter [int]
285 A nonzero value enables multi-packet send (MPS) for ConnectX-4 Lx and
286 enhanced multi-packet send (Enhanced MPS) for ConnectX-5. MPS allows the
287 TX burst function to pack up multiple packets in a single descriptor
288 session in order to save PCI bandwidth and improve performance at the
289 cost of a slightly higher CPU usage. When ``txq_inline`` is set along
290 with ``txq_mpw_en``, TX burst function tries to copy entire packet data
291 on to TX descriptor instead of including pointer of packet only if there
292 is enough room remained in the descriptor. ``txq_inline`` sets
293 per-descriptor space for either pointers or inlined packets. In addition,
294 Enhanced MPS supports hybrid mode - mixing inlined packets and pointers
295 in the same descriptor.
297 This option cannot be used with certain offloads such as ``DEV_TX_OFFLOAD_TCP_TSO,
298 DEV_TX_OFFLOAD_VXLAN_TNL_TSO, DEV_TX_OFFLOAD_GRE_TNL_TSO, DEV_TX_OFFLOAD_VLAN_INSERT``.
299 When those offloads are requested the MPS send function will not be used.
301 It is currently only supported on the ConnectX-4 Lx and ConnectX-5
302 families of adapters. Enabled by default.
304 - ``txq_mpw_hdr_dseg_en`` parameter [int]
306 A nonzero value enables including two pointers in the first block of TX
307 descriptor. This can be used to lessen CPU load for memory copy.
309 Effective only when Enhanced MPS is supported. Disabled by default.
311 - ``txq_max_inline_len`` parameter [int]
313 Maximum size of packet to be inlined. This limits the size of packet to
314 be inlined. If the size of a packet is larger than configured value, the
315 packet isn't inlined even though there's enough space remained in the
316 descriptor. Instead, the packet is included with pointer.
318 Effective only when Enhanced MPS is supported. The default value is 256.
320 - ``tx_vec_en`` parameter [int]
322 A nonzero value enables Tx vector on ConnectX-5 only NIC if the number of
323 global Tx queues on the port is lesser than MLX5_VPMD_MIN_TXQS.
325 This option cannot be used with certain offloads such as ``DEV_TX_OFFLOAD_TCP_TSO,
326 DEV_TX_OFFLOAD_VXLAN_TNL_TSO, DEV_TX_OFFLOAD_GRE_TNL_TSO, DEV_TX_OFFLOAD_VLAN_INSERT``.
327 When those offloads are requested the MPS send function will not be used.
329 Enabled by default on ConnectX-5.
331 - ``rx_vec_en`` parameter [int]
333 A nonzero value enables Rx vector if the port is not configured in
334 multi-segment otherwise this parameter is ignored.
341 This driver relies on external libraries and kernel drivers for resources
342 allocations and initialization. The following dependencies are not part of
343 DPDK and must be installed separately:
347 User space Verbs framework used by librte_pmd_mlx5. This library provides
348 a generic interface between the kernel and low-level user space drivers
351 It allows slow and privileged operations (context initialization, hardware
352 resources allocations) to be managed by the kernel and fast operations to
353 never leave user space.
357 Low-level user space driver library for Mellanox ConnectX-4/ConnectX-5
358 devices, it is automatically loaded by libibverbs.
360 This library basically implements send/receive calls to the hardware
365 They provide the kernel-side Verbs API and low level device drivers that
366 manage actual hardware initialization and resources sharing with user
369 Unlike most other PMDs, these modules must remain loaded and bound to
372 - mlx5_core: hardware driver managing Mellanox ConnectX-4/ConnectX-5
373 devices and related Ethernet kernel network devices.
374 - mlx5_ib: InifiniBand device driver.
375 - ib_uverbs: user space driver for Verbs (entry point for libibverbs).
377 - **Firmware update**
379 Mellanox OFED releases include firmware updates for ConnectX-4/ConnectX-5
382 Because each release provides new features, these updates must be applied to
383 match the kernel modules and libraries they come with.
387 Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
393 Either RDMA Core library with a recent enough Linux kernel release
394 (recommended) or Mellanox OFED, which provides compatibility with older
397 RMDA Core with Linux Kernel
398 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
400 - Minimal kernel version : v4.14 or the most recent 4.14-rc (see `Linux installation documentation`_)
401 - Minimal rdma-core version: v15+ commit 0c5f5765213a ("Merge pull request #227 from yishaih/tm")
402 (see `RDMA Core installation documentation`_)
404 .. _`Linux installation documentation`: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git/plain/Documentation/admin-guide/README.rst
405 .. _`RDMA Core installation documentation`: https://raw.githubusercontent.com/linux-rdma/rdma-core/master/README.md
410 - Mellanox OFED version: **4.2, 4.3**.
413 - ConnectX-4: **12.21.1000** and above.
414 - ConnectX-4 Lx: **14.21.1000** and above.
415 - ConnectX-5: **16.21.1000** and above.
416 - ConnectX-5 Ex: **16.21.1000** and above.
418 While these libraries and kernel modules are available on OpenFabrics
419 Alliance's `website <https://www.openfabrics.org/>`__ and provided by package
420 managers on most distributions, this PMD requires Ethernet extensions that
421 may not be supported at the moment (this is a work in progress).
424 <http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux>`__
425 includes the necessary support and should be used in the meantime. For DPDK,
426 only libibverbs, libmlx5, mlnx-ofed-kernel packages and firmware updates are
427 required from that distribution.
431 Several versions of Mellanox OFED are available. Installing the version
432 this DPDK release was developed and tested against is strongly
433 recommended. Please check the `prerequisites`_.
438 * Mellanox(R) ConnectX(R)-4 10G MCX4111A-XCAT (1x10G)
439 * Mellanox(R) ConnectX(R)-4 10G MCX4121A-XCAT (2x10G)
440 * Mellanox(R) ConnectX(R)-4 25G MCX4111A-ACAT (1x25G)
441 * Mellanox(R) ConnectX(R)-4 25G MCX4121A-ACAT (2x25G)
442 * Mellanox(R) ConnectX(R)-4 40G MCX4131A-BCAT (1x40G)
443 * Mellanox(R) ConnectX(R)-4 40G MCX413A-BCAT (1x40G)
444 * Mellanox(R) ConnectX(R)-4 40G MCX415A-BCAT (1x40G)
445 * Mellanox(R) ConnectX(R)-4 50G MCX4131A-GCAT (1x50G)
446 * Mellanox(R) ConnectX(R)-4 50G MCX413A-GCAT (1x50G)
447 * Mellanox(R) ConnectX(R)-4 50G MCX414A-BCAT (2x50G)
448 * Mellanox(R) ConnectX(R)-4 50G MCX415A-GCAT (2x50G)
449 * Mellanox(R) ConnectX(R)-4 50G MCX416A-BCAT (2x50G)
450 * Mellanox(R) ConnectX(R)-4 50G MCX416A-GCAT (2x50G)
451 * Mellanox(R) ConnectX(R)-4 50G MCX415A-CCAT (1x100G)
452 * Mellanox(R) ConnectX(R)-4 100G MCX416A-CCAT (2x100G)
453 * Mellanox(R) ConnectX(R)-4 Lx 10G MCX4121A-XCAT (2x10G)
454 * Mellanox(R) ConnectX(R)-4 Lx 25G MCX4121A-ACAT (2x25G)
455 * Mellanox(R) ConnectX(R)-5 100G MCX556A-ECAT (2x100G)
456 * Mellanox(R) ConnectX(R)-5 Ex EN 100G MCX516A-CDAT (2x100G)
458 Quick Start Guide on OFED
459 -------------------------
461 1. Download latest Mellanox OFED. For more info check the `prerequisites`_.
464 2. Install the required libraries and kernel modules either by installing
465 only the required set, or by installing the entire Mellanox OFED:
467 .. code-block:: console
469 ./mlnxofedinstall --upstream-libs --dpdk
471 3. Verify the firmware is the correct one:
473 .. code-block:: console
477 4. Verify all ports links are set to Ethernet:
479 .. code-block:: console
481 mlxconfig -d <mst device> query | grep LINK_TYPE
485 Link types may have to be configured to Ethernet:
487 .. code-block:: console
489 mlxconfig -d <mst device> set LINK_TYPE_P1/2=1/2/3
491 * LINK_TYPE_P1=<1|2|3> , 1=Infiniband 2=Ethernet 3=VPI(auto-sense)
493 For hypervisors verify SR-IOV is enabled on the NIC:
495 .. code-block:: console
497 mlxconfig -d <mst device> query | grep SRIOV_EN
500 If needed, set enable the set the relevant fields:
502 .. code-block:: console
504 mlxconfig -d <mst device> set SRIOV_EN=1 NUM_OF_VFS=16
505 mlxfwreset -d <mst device> reset
507 5. Restart the driver:
509 .. code-block:: console
511 /etc/init.d/openibd restart
515 .. code-block:: console
517 service openibd restart
519 If link type was changed, firmware must be reset as well:
521 .. code-block:: console
523 mlxfwreset -d <mst device> reset
525 For hypervisors, after reset write the sysfs number of virtual functions
528 To dynamically instantiate a given number of virtual functions (VFs):
530 .. code-block:: console
532 echo [num_vfs] > /sys/class/infiniband/mlx5_0/device/sriov_numvfs
534 6. Compile DPDK and you are ready to go. See instructions on
535 :ref:`Development Kit Build System <Development_Kit_Build_System>`
540 1. Configure aggressive CQE Zipping for maximum performance:
542 .. code-block:: console
544 mlxconfig -d <mst device> s CQE_COMPRESSION=1
546 To set it back to the default CQE Zipping mode use:
548 .. code-block:: console
550 mlxconfig -d <mst device> s CQE_COMPRESSION=0
552 2. In case of virtualization:
554 - Make sure that hypervisor kernel is 3.16 or newer.
555 - Configure boot with ``iommu=pt``.
557 - Make sure to allocate a VM on huge pages.
558 - Make sure to set CPU pinning.
560 3. Use the CPU near local NUMA node to which the PCIe adapter is connected,
561 for better performance. For VMs, verify that the right CPU
562 and NUMA node are pinned according to the above. Run:
564 .. code-block:: console
568 to identify the NUMA node to which the PCIe adapter is connected.
570 4. If more than one adapter is used, and root complex capabilities allow
571 to put both adapters on the same NUMA node without PCI bandwidth degradation,
572 it is recommended to locate both adapters on the same NUMA node.
573 This in order to forward packets from one to the other without
574 NUMA performance penalty.
576 5. Disable pause frames:
578 .. code-block:: console
580 ethtool -A <netdev> rx off tx off
582 6. Verify IO non-posted prefetch is disabled by default. This can be checked
583 via the BIOS configuration. Please contact you server provider for more
584 information about the settings.
588 On some machines, depends on the machine integrator, it is beneficial
589 to set the PCI max read request parameter to 1K. This can be
590 done in the following way:
592 To query the read request size use:
594 .. code-block:: console
596 setpci -s <NIC PCI address> 68.w
598 If the output is different than 3XXX, set it by:
600 .. code-block:: console
602 setpci -s <NIC PCI address> 68.w=3XXX
604 The XXX can be different on different systems. Make sure to configure
605 according to the setpci output.
610 Compared to librte_pmd_mlx4 that implements a single RSS configuration per
611 port, librte_pmd_mlx5 supports per-protocol RSS configuration.
613 Since ``testpmd`` defaults to IP RSS mode and there is currently no
614 command-line parameter to enable additional protocols (UDP and TCP as well
615 as IP), the following commands must be entered from its CLI to get the same
616 behavior as librte_pmd_mlx4:
618 .. code-block:: console
621 > port config all rss all
627 This section demonstrates how to launch **testpmd** with Mellanox
628 ConnectX-4/ConnectX-5 devices managed by librte_pmd_mlx5.
630 #. Load the kernel modules:
632 .. code-block:: console
634 modprobe -a ib_uverbs mlx5_core mlx5_ib
636 Alternatively if MLNX_OFED is fully installed, the following script can
639 .. code-block:: console
641 /etc/init.d/openibd restart
645 User space I/O kernel modules (uio and igb_uio) are not used and do
646 not have to be loaded.
648 #. Make sure Ethernet interfaces are in working order and linked to kernel
649 verbs. Related sysfs entries should be present:
651 .. code-block:: console
653 ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
657 .. code-block:: console
664 #. Optionally, retrieve their PCI bus addresses for whitelisting:
666 .. code-block:: console
669 for intf in eth2 eth3 eth4 eth5;
671 (cd "/sys/class/net/${intf}/device/" && pwd -P);
674 sed -n 's,.*/\(.*\),-w \1,p'
678 .. code-block:: console
685 #. Request huge pages:
687 .. code-block:: console
689 echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
691 #. Start testpmd with basic parameters:
693 .. code-block:: console
695 testpmd -l 8-15 -n 4 -w 05:00.0 -w 05:00.1 -w 06:00.0 -w 06:00.1 -- --rxq=2 --txq=2 -i
699 .. code-block:: console
702 EAL: PCI device 0000:05:00.0 on NUMA socket 0
703 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
704 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_0" (VF: false)
705 PMD: librte_pmd_mlx5: 1 port(s) detected
706 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fe
707 EAL: PCI device 0000:05:00.1 on NUMA socket 0
708 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
709 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_1" (VF: false)
710 PMD: librte_pmd_mlx5: 1 port(s) detected
711 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:ff
712 EAL: PCI device 0000:06:00.0 on NUMA socket 0
713 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
714 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_2" (VF: false)
715 PMD: librte_pmd_mlx5: 1 port(s) detected
716 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fa
717 EAL: PCI device 0000:06:00.1 on NUMA socket 0
718 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
719 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_3" (VF: false)
720 PMD: librte_pmd_mlx5: 1 port(s) detected
721 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fb
722 Interactive-mode selected
723 Configuring Port 0 (socket 0)
724 PMD: librte_pmd_mlx5: 0x8cba80: TX queues number update: 0 -> 2
725 PMD: librte_pmd_mlx5: 0x8cba80: RX queues number update: 0 -> 2
726 Port 0: E4:1D:2D:E7:0C:FE
727 Configuring Port 1 (socket 0)
728 PMD: librte_pmd_mlx5: 0x8ccac8: TX queues number update: 0 -> 2
729 PMD: librte_pmd_mlx5: 0x8ccac8: RX queues number update: 0 -> 2
730 Port 1: E4:1D:2D:E7:0C:FF
731 Configuring Port 2 (socket 0)
732 PMD: librte_pmd_mlx5: 0x8cdb10: TX queues number update: 0 -> 2
733 PMD: librte_pmd_mlx5: 0x8cdb10: RX queues number update: 0 -> 2
734 Port 2: E4:1D:2D:E7:0C:FA
735 Configuring Port 3 (socket 0)
736 PMD: librte_pmd_mlx5: 0x8ceb58: TX queues number update: 0 -> 2
737 PMD: librte_pmd_mlx5: 0x8ceb58: RX queues number update: 0 -> 2
738 Port 3: E4:1D:2D:E7:0C:FB
739 Checking link statuses...
740 Port 0 Link Up - speed 40000 Mbps - full-duplex
741 Port 1 Link Up - speed 40000 Mbps - full-duplex
742 Port 2 Link Up - speed 10000 Mbps - full-duplex
743 Port 3 Link Up - speed 10000 Mbps - full-duplex