2 Copyright 2015 6WIND S.A.
4 Redistribution and use in source and binary forms, with or without
5 modification, are permitted provided that the following conditions
8 * Redistributions of source code must retain the above copyright
9 notice, this list of conditions and the following disclaimer.
10 * Redistributions in binary form must reproduce the above copyright
11 notice, this list of conditions and the following disclaimer in
12 the documentation and/or other materials provided with the
14 * Neither the name of 6WIND S.A. nor the names of its
15 contributors may be used to endorse or promote products derived
16 from this software without specific prior written permission.
18 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
21 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
23 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
24 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
25 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
26 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
28 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 The MLX5 poll mode driver library (**librte_pmd_mlx5**) provides support for
34 **Mellanox ConnectX-4** and **Mellanox ConnectX-4 Lx** families of
35 10/25/40/50/100 Gb/s adapters as well as their virtual functions (VF) in
38 Information and documentation about these adapters can be found on the
39 `Mellanox website <http://www.mellanox.com>`__. Help is also provided by the
40 `Mellanox community <http://community.mellanox.com/welcome>`__.
42 There is also a `section dedicated to this poll mode driver
43 <http://www.mellanox.com/page/products_dyn?product_family=209&mtag=pmd_for_dpdk>`__.
47 Due to external dependencies, this driver is disabled by default. It must
48 be enabled manually by setting ``CONFIG_RTE_LIBRTE_MLX5_PMD=y`` and
51 Implementation details
52 ----------------------
54 Besides its dependency on libibverbs (that implies libmlx5 and associated
55 kernel support), librte_pmd_mlx5 relies heavily on system calls for control
56 operations such as querying/updating the MTU and flow control parameters.
58 For security reasons and robustness, this driver only deals with virtual
59 memory addresses. The way resources allocations are handled by the kernel
60 combined with hardware specifications that allow it to handle virtual memory
61 addresses directly ensure that DPDK applications cannot access random
62 physical memory (or memory that does not belong to the current process).
64 This capability allows the PMD to coexist with kernel network interfaces
65 which remain functional, although they stop receiving unicast packets as
66 long as they share the same MAC address.
68 Enabling librte_pmd_mlx5 causes DPDK applications to be linked against
74 - Multiple TX and RX queues.
75 - Support for scattered TX and RX frames.
76 - IPv4, IPv6, TCPv4, TCPv6, UDPv4 and UDPv6 RSS on any number of queues.
77 - Several RSS hash keys, one for each flow type.
78 - Configurable RETA table.
79 - Support for multiple MAC addresses.
82 - Multicast promiscuous mode.
83 - Hardware checksum offloads.
84 - Flow director (RTE_FDIR_MODE_PERFECT and RTE_FDIR_MODE_PERFECT_MAC_VLAN).
89 - KVM and VMware ESX SR-IOV modes are not supported yet.
90 - Inner RSS for VXLAN frames is not supported yet.
91 - Port statistics through software counters only.
92 - Hardware checksum offloads for VXLAN inner header are not supported yet.
93 - Secondary processes are not supported yet.
101 These options can be modified in the ``.config`` file.
103 - ``CONFIG_RTE_LIBRTE_MLX5_PMD`` (default **n**)
105 Toggle compilation of librte_pmd_mlx5 itself.
107 - ``CONFIG_RTE_LIBRTE_MLX5_DEBUG`` (default **n**)
109 Toggle debugging code and stricter compilation flags. Enabling this option
110 adds additional run-time checks and debugging messages at the cost of
113 - ``CONFIG_RTE_LIBRTE_MLX5_SGE_WR_N`` (default **4**)
115 Number of scatter/gather elements (SGEs) per work request (WR). Lowering
116 this number improves performance but also limits the ability to receive
117 scattered packets (packets that do not fit a single mbuf). The default
118 value is a safe tradeoff.
120 - ``CONFIG_RTE_LIBRTE_MLX5_MAX_INLINE`` (default **0**)
122 Amount of data to be inlined during TX operations. Improves latency.
123 Can improve PPS performance when PCI backpressure is detected and may be
124 useful for scenarios involving heavy traffic on many queues.
126 Since the additional software logic necessary to handle this mode can
127 lower performance when there is no backpressure, it is not enabled by
130 - ``CONFIG_RTE_LIBRTE_MLX5_TX_MP_CACHE`` (default **8**)
132 Maximum number of cached memory pools (MPs) per TX queue. Each MP from
133 which buffers are to be transmitted must be associated to memory regions
134 (MRs). This is a slow operation that must be cached.
136 This value is always 1 for RX queues since they use a single MP.
138 Environment variables
139 ~~~~~~~~~~~~~~~~~~~~~
141 - ``MLX5_ENABLE_CQE_COMPRESSION``
143 A nonzero value lets ConnectX-4 return smaller completion entries to
144 improve performance when PCI backpressure is detected. It is most useful
145 for scenarios involving heavy traffic on many queues.
147 Since the additional software logic necessary to handle this mode can
148 lower performance when there is no backpressure, it is not enabled by
151 Run-time configuration
152 ~~~~~~~~~~~~~~~~~~~~~~
154 - librte_pmd_mlx5 brings kernel network interfaces up during initialization
155 because it is affected by their state. Forcing them down prevents packets
158 - **ethtool** operations on related kernel interfaces also affect the PMD.
163 This driver relies on external libraries and kernel drivers for resources
164 allocations and initialization. The following dependencies are not part of
165 DPDK and must be installed separately:
169 User space Verbs framework used by librte_pmd_mlx5. This library provides
170 a generic interface between the kernel and low-level user space drivers
173 It allows slow and privileged operations (context initialization, hardware
174 resources allocations) to be managed by the kernel and fast operations to
175 never leave user space.
179 Low-level user space driver library for Mellanox ConnectX-4 devices,
180 it is automatically loaded by libibverbs.
182 This library basically implements send/receive calls to the hardware
185 - **Kernel modules** (mlnx-ofed-kernel)
187 They provide the kernel-side Verbs API and low level device drivers that
188 manage actual hardware initialization and resources sharing with user
191 Unlike most other PMDs, these modules must remain loaded and bound to
194 - mlx5_core: hardware driver managing Mellanox ConnectX-4 devices and
195 related Ethernet kernel network devices.
196 - mlx5_ib: InifiniBand device driver.
197 - ib_uverbs: user space driver for Verbs (entry point for libibverbs).
199 - **Firmware update**
201 Mellanox OFED releases include firmware updates for ConnectX-4 adapters.
203 Because each release provides new features, these updates must be applied to
204 match the kernel modules and libraries they come with.
208 Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
211 Currently supported by DPDK:
213 - Mellanox OFED **3.1-1.0.3**, **3.1-1.5.7.1** or **3.2-2.0.0.0** depending
216 The following features are supported with version **3.1-1.5.7.1** and
219 - IPv6, UPDv6, TCPv6 RSS.
220 - RX checksum offloads.
223 The following features are supported with version **3.2-2.0.0.0** and
228 - Minimum firmware version:
230 With MLNX_OFED **3.1-1.0.3**:
232 - ConnectX-4: **12.12.1240**
233 - ConnectX-4 Lx: **14.12.1100**
235 With MLNX_OFED **3.1-1.5.7.1**:
237 - ConnectX-4: **12.13.0144**
238 - ConnectX-4 Lx: **14.13.0144**
240 With MLNX_OFED **3.2-2.0.0.0**:
242 - ConnectX-4: **12.14.2036**
243 - ConnectX-4 Lx: **14.14.2036**
245 Getting Mellanox OFED
246 ~~~~~~~~~~~~~~~~~~~~~
248 While these libraries and kernel modules are available on OpenFabrics
249 Alliance's `website <https://www.openfabrics.org/>`__ and provided by package
250 managers on most distributions, this PMD requires Ethernet extensions that
251 may not be supported at the moment (this is a work in progress).
254 <http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux>`__
255 includes the necessary support and should be used in the meantime. For DPDK,
256 only libibverbs, libmlx5, mlnx-ofed-kernel packages and firmware updates are
257 required from that distribution.
261 Several versions of Mellanox OFED are available. Installing the version
262 this DPDK release was developed and tested against is strongly
263 recommended. Please check the `prerequisites`_.
268 Compared to librte_pmd_mlx4 that implements a single RSS configuration per
269 port, librte_pmd_mlx5 supports per-protocol RSS configuration.
271 Since ``testpmd`` defaults to IP RSS mode and there is currently no
272 command-line parameter to enable additional protocols (UDP and TCP as well
273 as IP), the following commands must be entered from its CLI to get the same
274 behavior as librte_pmd_mlx4:
276 .. code-block:: console
279 > port config all rss all
285 This section demonstrates how to launch **testpmd** with Mellanox ConnectX-4
286 devices managed by librte_pmd_mlx5.
288 #. Load the kernel modules:
290 .. code-block:: console
292 modprobe -a ib_uverbs mlx5_core mlx5_ib
294 Alternatively if MLNX_OFED is fully installed, the following script can
297 .. code-block:: console
299 /etc/init.d/openibd restart
303 User space I/O kernel modules (uio and igb_uio) are not used and do
304 not have to be loaded.
306 #. Make sure Ethernet interfaces are in working order and linked to kernel
307 verbs. Related sysfs entries should be present:
309 .. code-block:: console
311 ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
315 .. code-block:: console
322 #. Optionally, retrieve their PCI bus addresses for whitelisting:
324 .. code-block:: console
327 for intf in eth2 eth3 eth4 eth5;
329 (cd "/sys/class/net/${intf}/device/" && pwd -P);
332 sed -n 's,.*/\(.*\),-w \1,p'
336 .. code-block:: console
343 #. Request huge pages:
345 .. code-block:: console
347 echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
349 #. Start testpmd with basic parameters:
351 .. code-block:: console
353 testpmd -c 0xff00 -n 4 -w 05:00.0 -w 05:00.1 -w 06:00.0 -w 06:00.1 -- --rxq=2 --txq=2 -i
357 .. code-block:: console
360 EAL: PCI device 0000:05:00.0 on NUMA socket 0
361 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
362 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_0" (VF: false)
363 PMD: librte_pmd_mlx5: 1 port(s) detected
364 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fe
365 EAL: PCI device 0000:05:00.1 on NUMA socket 0
366 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
367 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_1" (VF: false)
368 PMD: librte_pmd_mlx5: 1 port(s) detected
369 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:ff
370 EAL: PCI device 0000:06:00.0 on NUMA socket 0
371 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
372 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_2" (VF: false)
373 PMD: librte_pmd_mlx5: 1 port(s) detected
374 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fa
375 EAL: PCI device 0000:06:00.1 on NUMA socket 0
376 EAL: probe driver: 15b3:1013 librte_pmd_mlx5
377 PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_3" (VF: false)
378 PMD: librte_pmd_mlx5: 1 port(s) detected
379 PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fb
380 Interactive-mode selected
381 Configuring Port 0 (socket 0)
382 PMD: librte_pmd_mlx5: 0x8cba80: TX queues number update: 0 -> 2
383 PMD: librte_pmd_mlx5: 0x8cba80: RX queues number update: 0 -> 2
384 Port 0: E4:1D:2D:E7:0C:FE
385 Configuring Port 1 (socket 0)
386 PMD: librte_pmd_mlx5: 0x8ccac8: TX queues number update: 0 -> 2
387 PMD: librte_pmd_mlx5: 0x8ccac8: RX queues number update: 0 -> 2
388 Port 1: E4:1D:2D:E7:0C:FF
389 Configuring Port 2 (socket 0)
390 PMD: librte_pmd_mlx5: 0x8cdb10: TX queues number update: 0 -> 2
391 PMD: librte_pmd_mlx5: 0x8cdb10: RX queues number update: 0 -> 2
392 Port 2: E4:1D:2D:E7:0C:FA
393 Configuring Port 3 (socket 0)
394 PMD: librte_pmd_mlx5: 0x8ceb58: TX queues number update: 0 -> 2
395 PMD: librte_pmd_mlx5: 0x8ceb58: RX queues number update: 0 -> 2
396 Port 3: E4:1D:2D:E7:0C:FB
397 Checking link statuses...
398 Port 0 Link Up - speed 40000 Mbps - full-duplex
399 Port 1 Link Up - speed 40000 Mbps - full-duplex
400 Port 2 Link Up - speed 10000 Mbps - full-duplex
401 Port 3 Link Up - speed 10000 Mbps - full-duplex