2 Copyright 2012 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.
31 MLX4 poll mode driver library
32 =============================
34 The MLX4 poll mode driver library (**librte_pmd_mlx4**) implements support
35 for **Mellanox ConnectX-3** and **Mellanox ConnectX-3 Pro** 10/40 Gbps adapters
36 as well as their virtual functions (VF) in SR-IOV context.
38 Information and documentation about this family of adapters can be found on
39 the `Mellanox website <http://www.mellanox.com>`_. Help is also provided by
40 the `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_MLX4_PMD=y`` and
51 Implementation details
52 ----------------------
54 Most Mellanox ConnectX-3 devices provide two ports but expose a single PCI
55 bus address, thus unlike most drivers, librte_pmd_mlx4 registers itself as a
56 PCI driver that allocates one Ethernet device per detected port.
58 For this reason, one cannot white/blacklist a single port without also
59 white/blacklisting the others on the same device.
61 Besides its dependency on libibverbs (that implies libmlx4 and associated
62 kernel support), librte_pmd_mlx4 relies heavily on system calls for control
63 operations such as querying/updating the MTU and flow control parameters.
65 For security reasons and robustness, this driver only deals with virtual
66 memory addresses. The way resources allocations are handled by the kernel
67 combined with hardware specifications that allow it to handle virtual memory
68 addresses directly ensure that DPDK applications cannot access random
69 physical memory (or memory that does not belong to the current process).
71 This capability allows the PMD to coexist with kernel network interfaces
72 which remain functional, although they stop receiving unicast packets as
73 long as they share the same MAC address.
75 Compiling librte_pmd_mlx4 causes DPDK to be linked against libibverbs.
80 - Multi arch support: x86_64 and POWER8.
81 - RSS, also known as RCA, is supported. In this mode the number of
82 configured RX queues must be a power of two.
83 - VLAN filtering is supported.
84 - Link state information is provided.
85 - Promiscuous mode is supported.
86 - All multicast mode is supported.
87 - Multiple MAC addresses (unicast, multicast) can be configured.
88 - Scattered packets are supported for TX and RX.
89 - Inner L3/L4 (IP, TCP and UDP) TX/RX checksum offloading and validation.
90 - Outer L3 (IP) TX/RX checksum offloading and validation for VXLAN frames.
96 - RSS hash key cannot be modified.
97 - RSS RETA cannot be configured
98 - RSS always includes L3 (IPv4/IPv6) and L4 (UDP/TCP). They cannot be
100 - Hardware counters are not implemented (they are software counters).
108 These options can be modified in the ``.config`` file.
110 - ``CONFIG_RTE_LIBRTE_MLX4_PMD`` (default **n**)
112 Toggle compilation of librte_pmd_mlx4 itself.
114 - ``CONFIG_RTE_LIBRTE_MLX4_DEBUG`` (default **n**)
116 Toggle debugging code and stricter compilation flags. Enabling this option
117 adds additional run-time checks and debugging messages at the cost of
120 - ``CONFIG_RTE_LIBRTE_MLX4_DEBUG_BROKEN_VERBS`` (default **n**)
122 Mellanox OFED versions earlier than 4.2 may return false errors from
123 Verbs object destruction APIs after the device is plugged out.
124 Enabling this option replaces assertion checks that cause the program
125 to abort with harmless debugging messages as a workaround.
126 Relevant only when CONFIG_RTE_LIBRTE_MLX4_DEBUG is enabled.
128 - ``CONFIG_RTE_LIBRTE_MLX4_SGE_WR_N`` (default **4**)
130 Number of scatter/gather elements (SGEs) per work request (WR). Lowering
131 this number improves performance but also limits the ability to receive
132 scattered packets (packets that do not fit a single mbuf). The default
133 value is a safe tradeoff.
135 - ``CONFIG_RTE_LIBRTE_MLX4_MAX_INLINE`` (default **0**)
137 Amount of data to be inlined during TX operations. Improves latency but
140 - ``CONFIG_RTE_LIBRTE_MLX4_TX_MP_CACHE`` (default **8**)
142 Maximum number of cached memory pools (MPs) per TX queue. Each MP from
143 which buffers are to be transmitted must be associated to memory regions
144 (MRs). This is a slow operation that must be cached.
146 This value is always 1 for RX queues since they use a single MP.
148 - ``CONFIG_RTE_LIBRTE_MLX4_SOFT_COUNTERS`` (default **1**)
150 Toggle software counters. No counters are available if this option is
151 disabled since hardware counters are not supported.
153 Environment variables
154 ~~~~~~~~~~~~~~~~~~~~~
156 - ``MLX4_INLINE_RECV_SIZE``
158 A nonzero value enables inline receive for packets up to that size. May
159 significantly improve performance in some cases but lower it in
160 others. Requires careful testing.
162 Run-time configuration
163 ~~~~~~~~~~~~~~~~~~~~~~
165 - The only constraint when RSS mode is requested is to make sure the number
166 of RX queues is a power of two. This is a hardware requirement.
168 - librte_pmd_mlx4 brings kernel network interfaces up during initialization
169 because it is affected by their state. Forcing them down prevents packets
172 - **ethtool** operations on related kernel interfaces also affect the PMD.
174 - ``port`` parameter [int]
176 This parameter provides a physical port to probe and can be specified multiple
177 times for additional ports. All ports are probed by default if left
180 Kernel module parameters
181 ~~~~~~~~~~~~~~~~~~~~~~~~
183 The **mlx4_core** kernel module has several parameters that affect the
184 behavior and/or the performance of librte_pmd_mlx4. Some of them are described
187 - **num_vfs** (integer or triplet, optionally prefixed by device address
190 Create the given number of VFs on the specified devices.
192 - **log_num_mgm_entry_size** (integer)
194 Device-managed flow steering (DMFS) is required by DPDK applications. It is
195 enabled by using a negative value, the last four bits of which have a
198 - **-1**: force device-managed flow steering (DMFS).
199 - **-7**: configure optimized steering mode to improve performance with the
200 following limitation: VLAN filtering is not supported with this mode.
201 This is the recommended mode in case VLAN filter is not needed.
206 This driver relies on external libraries and kernel drivers for resources
207 allocations and initialization. The following dependencies are not part of
208 DPDK and must be installed separately:
212 User space verbs framework used by librte_pmd_mlx4. This library provides
213 a generic interface between the kernel and low-level user space drivers
216 It allows slow and privileged operations (context initialization, hardware
217 resources allocations) to be managed by the kernel and fast operations to
218 never leave user space.
222 Low-level user space driver library for Mellanox ConnectX-3 devices,
223 it is automatically loaded by libibverbs.
225 This library basically implements send/receive calls to the hardware
228 - **Kernel modules** (mlnx-ofed-kernel)
230 They provide the kernel-side verbs API and low level device drivers that
231 manage actual hardware initialization and resources sharing with user
234 Unlike most other PMDs, these modules must remain loaded and bound to
237 - mlx4_core: hardware driver managing Mellanox ConnectX-3 devices.
238 - mlx4_en: Ethernet device driver that provides kernel network interfaces.
239 - mlx4_ib: InifiniBand device driver.
240 - ib_uverbs: user space driver for verbs (entry point for libibverbs).
242 - **Firmware update**
244 Mellanox OFED releases include firmware updates for ConnectX-3 adapters.
246 Because each release provides new features, these updates must be applied to
247 match the kernel modules and libraries they come with.
251 Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
254 Currently supported by DPDK:
256 - Mellanox OFED **4.1**.
257 - Firmware version **2.36.5000** and above.
259 Getting Mellanox OFED
260 ~~~~~~~~~~~~~~~~~~~~~
262 While these libraries and kernel modules are available on OpenFabrics
263 Alliance's `website <https://www.openfabrics.org/>`_ and provided by package
264 managers on most distributions, this PMD requires Ethernet extensions that
265 may not be supported at the moment (this is a work in progress).
268 <http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux_sw_drivers>`_
269 includes the necessary support and should be used in the meantime. For DPDK,
270 only libibverbs, libmlx4, mlnx-ofed-kernel packages and firmware updates are
271 required from that distribution.
275 Several versions of Mellanox OFED are available. Installing the version
276 this DPDK release was developed and tested against is strongly
277 recommended. Please check the `prerequisites`_.
282 * Mellanox(R) ConnectX(R)-3 Pro 40G MCX354A-FCC_Ax (2*40G)
287 1. Download latest Mellanox OFED. For more info check the `prerequisites`_.
289 2. Install the required libraries and kernel modules either by installing
290 only the required set, or by installing the entire Mellanox OFED:
294 .. code-block:: console
298 For SR-IOV hypervisors use:
300 .. code-block:: console
302 ./mlnxofedinstall --enable-sriov -hypervisor
304 For SR-IOV virtual machine use:
306 .. code-block:: console
308 ./mlnxofedinstall --guest
310 3. Verify the firmware is the correct one:
312 .. code-block:: console
316 4. Set all ports links to Ethernet, follow instructions on the screen:
318 .. code-block:: console
322 Or in the manual way:
324 .. code-block:: console
326 PCI=<NIC PCI address>
327 echo eth > "/sys/bus/pci/devices/$PCI/mlx4_port0"
328 echo eth > "/sys/bus/pci/devices/$PCI/mlx4_port1"
330 5. In case of bare metal or hypervisor, configure optimized steering mode
331 by adding the following line to ``/etc/modprobe.d/mlx4_core.conf``:
333 .. code-block:: console
335 options mlx4_core log_num_mgm_entry_size=-7
339 If VLAN filtering is used, set log_num_mgm_entry_size=-1.
340 Performance degradation can occur on this case.
342 6. Restart the driver:
344 .. code-block:: console
346 /etc/init.d/openibd restart
350 .. code-block:: console
352 service openibd restart
354 7. Compile DPDK and you are ready to go. See instructions on
355 :ref:`Development Kit Build System <Development_Kit_Build_System>`
360 1. Verify the optimized steering mode is configured:
362 .. code-block:: console
364 cat /sys/module/mlx4_core/parameters/log_num_mgm_entry_size
366 2. Use environment variable MLX4_INLINE_RECV_SIZE=64 to get maximum
367 performance for 64B messages.
369 3. Use the CPU near local NUMA node to which the PCIe adapter is connected,
370 for better performance. For VMs, verify that the right CPU
371 and NUMA node are pinned according to the above. Run:
373 .. code-block:: console
377 to identify the NUMA node to which the PCIe adapter is connected.
379 4. If more than one adapter is used, and root complex capabilities allow
380 to put both adapters on the same NUMA node without PCI bandwidth degradation,
381 it is recommended to locate both adapters on the same NUMA node.
382 This in order to forward packets from one to the other without
383 NUMA performance penalty.
385 5. Disable pause frames:
387 .. code-block:: console
389 ethtool -A <netdev> rx off tx off
391 6. Verify IO non-posted prefetch is disabled by default. This can be checked
392 via the BIOS configuration. Please contact you server provider for more
393 information about the settings.
397 On some machines, depends on the machine integrator, it is beneficial
398 to set the PCI max read request parameter to 1K. This can be
399 done in the following way:
401 To query the read request size use:
403 .. code-block:: console
405 setpci -s <NIC PCI address> 68.w
407 If the output is different than 3XXX, set it by:
409 .. code-block:: console
411 setpci -s <NIC PCI address> 68.w=3XXX
413 The XXX can be different on different systems. Make sure to configure
414 according to the setpci output.
419 This section demonstrates how to launch **testpmd** with Mellanox ConnectX-3
420 devices managed by librte_pmd_mlx4.
422 #. Load the kernel modules:
424 .. code-block:: console
426 modprobe -a ib_uverbs mlx4_en mlx4_core mlx4_ib
428 Alternatively if MLNX_OFED is fully installed, the following script can
431 .. code-block:: console
433 /etc/init.d/openibd restart
437 User space I/O kernel modules (uio and igb_uio) are not used and do
438 not have to be loaded.
440 #. Make sure Ethernet interfaces are in working order and linked to kernel
441 verbs. Related sysfs entries should be present:
443 .. code-block:: console
445 ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
449 .. code-block:: console
456 #. Optionally, retrieve their PCI bus addresses for whitelisting:
458 .. code-block:: console
461 for intf in eth2 eth3 eth4 eth5;
463 (cd "/sys/class/net/${intf}/device/" && pwd -P);
466 sed -n 's,.*/\(.*\),-w \1,p'
470 .. code-block:: console
479 There are only two distinct PCI bus addresses because the Mellanox
480 ConnectX-3 adapters installed on this system are dual port.
482 #. Request huge pages:
484 .. code-block:: console
486 echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
488 #. Start testpmd with basic parameters:
490 .. code-block:: console
492 testpmd -l 8-15 -n 4 -w 0000:83:00.0 -w 0000:84:00.0 -- --rxq=2 --txq=2 -i
496 .. code-block:: console
499 EAL: PCI device 0000:83:00.0 on NUMA socket 1
500 EAL: probe driver: 15b3:1007 librte_pmd_mlx4
501 PMD: librte_pmd_mlx4: PCI information matches, using device "mlx4_0" (VF: false)
502 PMD: librte_pmd_mlx4: 2 port(s) detected
503 PMD: librte_pmd_mlx4: port 1 MAC address is 00:02:c9:b5:b7:50
504 PMD: librte_pmd_mlx4: port 2 MAC address is 00:02:c9:b5:b7:51
505 EAL: PCI device 0000:84:00.0 on NUMA socket 1
506 EAL: probe driver: 15b3:1007 librte_pmd_mlx4
507 PMD: librte_pmd_mlx4: PCI information matches, using device "mlx4_1" (VF: false)
508 PMD: librte_pmd_mlx4: 2 port(s) detected
509 PMD: librte_pmd_mlx4: port 1 MAC address is 00:02:c9:b5:ba:b0
510 PMD: librte_pmd_mlx4: port 2 MAC address is 00:02:c9:b5:ba:b1
511 Interactive-mode selected
512 Configuring Port 0 (socket 0)
513 PMD: librte_pmd_mlx4: 0x867d60: TX queues number update: 0 -> 2
514 PMD: librte_pmd_mlx4: 0x867d60: RX queues number update: 0 -> 2
515 Port 0: 00:02:C9:B5:B7:50
516 Configuring Port 1 (socket 0)
517 PMD: librte_pmd_mlx4: 0x867da0: TX queues number update: 0 -> 2
518 PMD: librte_pmd_mlx4: 0x867da0: RX queues number update: 0 -> 2
519 Port 1: 00:02:C9:B5:B7:51
520 Configuring Port 2 (socket 0)
521 PMD: librte_pmd_mlx4: 0x867de0: TX queues number update: 0 -> 2
522 PMD: librte_pmd_mlx4: 0x867de0: RX queues number update: 0 -> 2
523 Port 2: 00:02:C9:B5:BA:B0
524 Configuring Port 3 (socket 0)
525 PMD: librte_pmd_mlx4: 0x867e20: TX queues number update: 0 -> 2
526 PMD: librte_pmd_mlx4: 0x867e20: RX queues number update: 0 -> 2
527 Port 3: 00:02:C9:B5:BA:B1
528 Checking link statuses...
529 Port 0 Link Up - speed 10000 Mbps - full-duplex
530 Port 1 Link Up - speed 40000 Mbps - full-duplex
531 Port 2 Link Up - speed 10000 Mbps - full-duplex
532 Port 3 Link Up - speed 40000 Mbps - full-duplex