2018-03-20 19:20:35 +00:00
.. SPDX-License-Identifier: BSD-3-Clause
2015-10-30 18:52:42 +00:00
Copyright 2015 6WIND S.A.
2018-03-20 19:20:35 +00:00
Copyright 2015 Mellanox Technologies, Ltd
2015-10-30 18:52:42 +00:00
MLX5 poll mode driver
=====================
2017-01-06 00:49:31 +00:00
The MLX5 poll mode driver library (**librte_pmd_mlx5** ) provides support
2018-05-15 06:12:50 +00:00
for **Mellanox ConnectX-4** , **Mellanox ConnectX-4 Lx** , **Mellanox
2019-05-13 11:41:02 +00:00
ConnectX-5**, ** Mellanox ConnectX-6** and ** Mellanox BlueField** families
2018-12-31 12:43:48 +00:00
of 10/25/40/50/100/200 Gb/s adapters as well as their virtual functions (VF)
in SR-IOV context.
2015-10-30 18:52:42 +00:00
Information and documentation about these adapters can be found on the
`Mellanox website <http://www.mellanox.com> `__ . Help is also provided by the
`Mellanox community <http://community.mellanox.com/welcome> `__ .
There is also a `section dedicated to this poll mode driver
<http://www.mellanox.com/page/products_dyn?product_family=209&mtag=pmd_for_dpdk> `__.
.. note ::
Due to external dependencies, this driver is disabled by default. It must
be enabled manually by setting `` CONFIG_RTE_LIBRTE_MLX5_PMD=y `` and
recompiling DPDK.
Implementation details
----------------------
Besides its dependency on libibverbs (that implies libmlx5 and associated
kernel support), librte_pmd_mlx5 relies heavily on system calls for control
operations such as querying/updating the MTU and flow control parameters.
For security reasons and robustness, this driver only deals with virtual
memory addresses. The way resources allocations are handled by the kernel
combined with hardware specifications that allow it to handle virtual memory
addresses directly ensure that DPDK applications cannot access random
physical memory (or memory that does not belong to the current process).
This capability allows the PMD to coexist with kernel network interfaces
which remain functional, although they stop receiving unicast packets as
long as they share the same MAC address.
2017-07-28 14:28:33 +00:00
This means legacy linux control tools (for example: ethtool, ifconfig and
more) can operate on the same network interfaces that owned by the DPDK
application.
2015-10-30 18:52:42 +00:00
Enabling librte_pmd_mlx5 causes DPDK applications to be linked against
libibverbs.
2015-10-30 18:55:19 +00:00
Features
--------
2018-07-12 12:01:31 +00:00
- Multi arch support: x86_64, POWER8, ARMv8, i686.
2015-10-30 18:55:19 +00:00
- Multiple TX and RX queues.
- Support for scattered TX and RX frames.
2015-12-12 19:43:24 +00:00
- IPv4, IPv6, TCPv4, TCPv6, UDPv4 and UDPv6 RSS on any number of queues.
2015-10-30 18:55:19 +00:00
- Several RSS hash keys, one for each flow type.
2018-11-04 12:10:20 +00:00
- Default RSS operation with no hash key specification.
2015-12-12 19:43:24 +00:00
- Configurable RETA table.
2019-07-30 15:06:21 +00:00
- Link flow control (pause frame).
2015-10-30 18:55:19 +00:00
- Support for multiple MAC addresses.
- VLAN filtering.
2016-03-03 14:26:44 +00:00
- RX VLAN stripping.
2016-03-17 15:38:58 +00:00
- TX VLAN insertion.
2016-03-17 15:38:56 +00:00
- RX CRC stripping configuration.
2019-05-13 11:41:02 +00:00
- Promiscuous mode on PF and VF.
- Multicast promiscuous mode on PF and VF.
2015-12-12 19:43:24 +00:00
- Hardware checksum offloads.
2017-10-09 14:45:05 +00:00
- Flow director (RTE_FDIR_MODE_PERFECT, RTE_FDIR_MODE_PERFECT_MAC_VLAN and
RTE_ETH_FDIR_REJECT).
2019-01-30 11:20:19 +00:00
- Flow API, including :ref: `flow_isolated_mode` .
2017-10-06 15:45:51 +00:00
- Multiple process.
2016-06-08 09:43:31 +00:00
- KVM and VMware ESX SR-IOV modes are supported.
2016-09-28 12:11:18 +00:00
- RSS hash result is supported.
2018-05-09 00:14:49 +00:00
- Hardware TSO for generic IP or UDP tunnel, including VXLAN and GRE.
- Hardware checksum Tx offload for generic IP or UDP tunnel, including VXLAN and GRE.
2017-07-28 14:28:33 +00:00
- RX interrupts.
- Statistics query including Basic, Extended and per queue.
2017-10-30 10:58:25 +00:00
- Rx HW timestamp.
2019-07-10 14:59:45 +00:00
- Tunnel types: VXLAN, L3 VXLAN, VXLAN-GPE, GRE, MPLSoGRE, MPLSoUDP, IP-in-IP.
2018-04-23 12:33:10 +00:00
- Tunnel HW offloads: packet type, inner/outer RSS, IP and UDP checksum verification.
2019-05-13 11:41:02 +00:00
- NIC HW offloads: encapsulation (vxlan, gre, mplsoudp, mplsogre), NAT, routing, TTL
increment/decrement, count, drop, mark. For details please see :ref: `Supported hardware offloads using rte_flow API` .
- Flow insertion rate of more then million flows per second, when using Direct Rules.
- Support for multiple rte_flow groups.
2019-07-22 14:51:59 +00:00
- Hardware LRO.
2015-10-30 18:55:19 +00:00
Limitations
-----------
2018-01-25 16:17:59 +00:00
- For secondary process:
- Forked secondary process not supported.
2019-04-01 21:17:53 +00:00
- External memory unregistered in EAL memseg list cannot be used for DMA
unless such memory has been registered by `` mlx5_mr_update_ext_mp() `` in
primary process and remapped to the same virtual address in secondary
process. If the external memory is registered by primary process but has
different virtual address in secondary process, unexpected error may happen.
2018-01-25 16:17:59 +00:00
2017-07-28 14:28:33 +00:00
- Flow pattern without any specific vlan will match for vlan packets as well:
When VLAN spec is not specified in the pattern, the matching rule will be created with VLAN as a wild card.
Meaning, the flow rule::
flow create 0 ingress pattern eth / vlan vid is 3 / ipv4 / end ...
Will only match vlan packets with vid=3. and the flow rules::
flow create 0 ingress pattern eth / ipv4 / end ...
Or::
flow create 0 ingress pattern eth / vlan / ipv4 / end ...
Will match any ipv4 packet (VLAN included).
2015-10-30 18:55:19 +00:00
2017-09-14 10:50:39 +00:00
- A multi segment packet must have less than 6 segments in case the Tx burst function
is set to multi-packet send or Enhanced multi-packet send. Otherwise it must have
less than 50 segments.
2018-05-09 11:13:50 +00:00
2017-10-31 15:51:11 +00:00
- Flows with a VXLAN Network Identifier equal (or ends to be equal)
to 0 are not supported.
2018-05-09 11:13:50 +00:00
2017-10-30 10:58:25 +00:00
- VXLAN TSO and checksum offloads are not supported on VM.
2018-05-09 11:13:50 +00:00
2018-05-15 11:07:14 +00:00
- L3 VXLAN and VXLAN-GPE tunnels cannot be supported together with MPLSoGRE and MPLSoUDP.
2018-04-05 15:07:19 +00:00
- VF: flow rules created on VF devices can only match traffic targeted at the
configured MAC addresses (see `` rte_eth_dev_mac_addr_add() `` ).
.. note ::
MAC addresses not already present in the bridge table of the associated
kernel network device will be added and cleaned up by the PMD when closing
the device. In case of ungraceful program termination, some entries may
remain present and should be removed manually by other means.
2017-09-14 10:50:39 +00:00
2018-05-09 11:13:50 +00:00
- When Multi-Packet Rx queue is configured (`` mprq_en `` ), a Rx packet can be
externally attached to a user-provided mbuf with having EXT_ATTACHED_MBUF in
ol_flags. As the mempool for the external buffer is managed by PMD, all the
Rx mbufs must be freed before the device is closed. Otherwise, the mempool of
the external buffers will be freed by PMD and the application which still
holds the external buffers may be corrupted.
2018-06-26 12:39:24 +00:00
- If Multi-Packet Rx queue is configured (`` mprq_en `` ) and Rx CQE compression is
enabled (`` rxq_cqe_comp_en `` ) at the same time, RSS hash result is not fully
supported. Some Rx packets may not have PKT_RX_RSS_HASH.
2018-11-12 11:31:06 +00:00
- IPv6 Multicast messages are not supported on VM, while promiscuous mode
and allmulticast mode are both set to off.
To receive IPv6 Multicast messages on VM, explicitly set the relevant
MAC address using rte_eth_dev_mac_addr_add() API.
2019-05-13 11:41:02 +00:00
- E-Switch decapsulation Flow:
2018-11-22 13:49:16 +00:00
2019-03-31 10:12:09 +00:00
- can be applied to PF port only.
2018-11-22 13:49:16 +00:00
- must specify VF port action (packet redirection from PF to VF).
- optionally may specify tunnel inner source and destination MAC addresses.
2019-05-13 11:41:02 +00:00
- E-Switch encapsulation Flow:
2018-11-22 13:49:16 +00:00
- can be applied to VF ports only.
- must specify PF port action (packet redirection from VF to PF).
2019-07-03 07:22:49 +00:00
- ICMP/ICMP6 code/type matching cannot be supported togeter with IP-in-IP tunnel.
2019-07-22 14:52:03 +00:00
- LRO:
2019-07-29 11:53:27 +00:00
- KEEP_CRC offload cannot be supported with LRO.
- The first mbuf length, without head-room, must be big enough to include the
TCP header (122B).
2019-07-22 14:52:03 +00:00
2017-11-29 11:29:07 +00:00
Statistics
----------
MLX5 supports various of methods to report statistics:
2018-11-23 08:03:37 +00:00
Port statistics can be queried using `` rte_eth_stats_get() `` . The received and sent statistics are through SW only and counts the number of packets received or sent successfully by the PMD. The imissed counter is the amount of packets that could not be delivered to SW because a queue was full. Packets not received due to congestion in the bus or on the NIC can be queried via the rx_discards_phy xstats counter.
2017-11-29 11:29:07 +00:00
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.
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.
2015-10-30 18:52:42 +00:00
Configuration
-------------
Compilation options
~~~~~~~~~~~~~~~~~~~
These options can be modified in the `` .config `` file.
- `` CONFIG_RTE_LIBRTE_MLX5_PMD `` (default **n** )
Toggle compilation of librte_pmd_mlx5 itself.
2019-01-09 14:23:18 +00:00
- `` CONFIG_RTE_IBVERBS_LINK_DLOPEN `` (default **n** )
2018-01-30 15:34:58 +00:00
Build PMD with additional code to make it loadable without hard
dependencies on **libibverbs** nor **libmlx5** , which may not be installed
on the target system.
In this mode, their presence is still required for it to run properly,
however their absence won't prevent a DPDK application from starting (with
`` CONFIG_RTE_BUILD_SHARED_LIB `` disabled) and they won't show up as
missing with `` ldd(1) `` .
2018-02-02 16:46:18 +00:00
It works by moving these dependencies to a purpose-built rdma-core "glue"
2018-03-02 14:15:17 +00:00
plug-in which must either be installed in a directory whose name is based
on `` CONFIG_RTE_EAL_PMD_PATH `` suffixed with `` -glue `` if set, or in a
standard location for the dynamic linker (e.g. `` /lib `` ) if left to the
default empty string (`` "" `` ).
2018-02-02 16:46:18 +00:00
2018-01-30 15:34:58 +00:00
This option has no performance impact.
2019-01-09 14:23:19 +00:00
- `` CONFIG_RTE_IBVERBS_LINK_STATIC `` (default **n** )
2019-04-26 15:14:21 +00:00
Embed static flavor of the dependencies **libibverbs** and **libmlx5**
2019-01-09 14:23:19 +00:00
in the PMD shared library or the executable static binary.
2015-10-30 18:52:42 +00:00
- `` CONFIG_RTE_LIBRTE_MLX5_DEBUG `` (default **n** )
Toggle debugging code and stricter compilation flags. Enabling this option
adds additional run-time checks and debugging messages at the cost of
lower performance.
2019-01-30 07:53:19 +00:00
.. note ::
2019-05-13 11:41:02 +00:00
For BlueField, target should be set to `` arm64-bluefield-linux-gcc `` . This
2019-01-30 07:53:19 +00:00
will enable `` CONFIG_RTE_LIBRTE_MLX5_PMD `` and set `` RTE_CACHE_LINE_SIZE `` to
64. Default armv8a configuration of make build and meson build set it to 128
then brings performance degradation.
2015-11-23 14:44:49 +00:00
Environment variables
~~~~~~~~~~~~~~~~~~~~~
2018-02-02 16:46:18 +00:00
- `` MLX5_GLUE_PATH ``
A list of directories in which to search for the rdma-core "glue" plug-in,
separated by colons or semi-colons.
2019-01-09 14:23:18 +00:00
Only matters when compiled with `` CONFIG_RTE_IBVERBS_LINK_DLOPEN ``
2018-02-02 16:46:18 +00:00
enabled and most useful when `` CONFIG_RTE_EAL_PMD_PATH `` is also set,
since `` LD_LIBRARY_PATH `` has no effect in this case.
2017-10-25 00:27:25 +00:00
- `` MLX5_SHUT_UP_BF ``
Configures HW Tx doorbell register as IO-mapped.
By default, the HW Tx doorbell is configured as a write-combining register.
The register would be flushed to HW usually when the write-combining buffer
becomes full, but it depends on CPU design.
Except for vectorized Tx burst routines, a write memory barrier is enforced
after updating the register so that the update can be immediately visible to
HW.
When vectorized Tx burst is called, the barrier is set only if the burst size
is not aligned to MLX5_VPMD_TX_MAX_BURST. However, setting this environmental
variable will bring better latency even though the maximum throughput can
slightly decline.
2015-10-30 18:52:42 +00:00
Run-time configuration
~~~~~~~~~~~~~~~~~~~~~~
- librte_pmd_mlx5 brings kernel network interfaces up during initialization
because it is affected by their state. Forcing them down prevents packets
reception.
- **ethtool** operations on related kernel interfaces also affect the PMD.
2016-06-24 13:17:54 +00:00
- `` rxq_cqe_comp_en `` parameter [int]
A nonzero value enables the compression of CQE on RX side. This feature
2017-07-28 14:28:33 +00:00
allows to save PCI bandwidth and improve performance. Enabled by default.
2016-06-24 13:17:54 +00:00
2016-07-27 09:27:26 +00:00
Supported on:
2019-05-13 11:41:02 +00:00
- x86_64 with ConnectX-4, ConnectX-4 LX, ConnectX-5, ConnectX-6 and BlueField.
- POWER9 and ARMv8 with ConnectX-4 LX, ConnectX-5, ConnectX-6 and BlueField.
2016-07-27 09:27:26 +00:00
2018-10-25 06:24:00 +00:00
- `` rxq_cqe_pad_en `` parameter [int]
A nonzero value enables 128B padding of CQE on RX side. The size of CQE
is aligned with the size of a cacheline of the core. If cacheline size is
128B, the CQE size is configured to be 128B even though the device writes
only 64B data on the cacheline. This is to avoid unnecessary cache
invalidation by device's two consecutive writes on to one cacheline.
However in some architecture, it is more beneficial to update entire
cacheline with padding the rest 64B rather than striding because
read-modify-write could drop performance a lot. On the other hand,
writing extra data will consume more PCIe bandwidth and could also drop
the maximum throughput. It is recommended to empirically set this
parameter. Disabled by default.
Supported on:
2019-05-13 11:41:02 +00:00
- CPU having 128B cacheline with ConnectX-5 and BlueField.
2018-10-25 06:24:00 +00:00
2019-01-15 17:38:58 +00:00
- `` rxq_pkt_pad_en `` parameter [int]
A nonzero value enables padding Rx packet to the size of cacheline on PCI
transaction. This feature would waste PCI bandwidth but could improve
performance by avoiding partial cacheline write which may cause costly
read-modify-copy in memory transaction on some architectures. Disabled by
default.
Supported on:
2019-05-13 11:41:02 +00:00
- x86_64 with ConnectX-4, ConnectX-4 LX, ConnectX-5, ConnectX-6 and BlueField.
- POWER8 and ARMv8 with ConnectX-4 LX, ConnectX-5, ConnectX-6 and BlueField.
2019-01-15 17:38:58 +00:00
2018-05-09 11:13:50 +00:00
- `` mprq_en `` parameter [int]
A nonzero value enables configuring Multi-Packet Rx queues. Rx queue is
configured as Multi-Packet RQ if the total number of Rx queues is
`` rxqs_min_mprq `` or more and Rx scatter isn't configured. Disabled by
default.
Multi-Packet Rx Queue (MPRQ a.k.a Striding RQ) can further save PCIe bandwidth
by posting a single large buffer for multiple packets. Instead of posting a
buffers per a packet, one large buffer is posted in order to receive multiple
packets on the buffer. A MPRQ buffer consists of multiple fixed-size strides
and each stride receives one packet. MPRQ can improve throughput for
2019-03-31 10:12:09 +00:00
small-packet traffic.
2018-05-09 11:13:50 +00:00
When MPRQ is enabled, max_rx_pkt_len can be larger than the size of
user-provided mbuf even if DEV_RX_OFFLOAD_SCATTER isn't enabled. PMD will
configure large stride size enough to accommodate max_rx_pkt_len as long as
device allows. Note that this can waste system memory compared to enabling Rx
scatter and multi-segment packet.
- `` mprq_log_stride_num `` parameter [int]
Log 2 of the number of strides for Multi-Packet Rx queue. Configuring more
2019-03-31 10:12:09 +00:00
strides can reduce PCIe traffic further. If configured value is not in the
2018-05-09 11:13:50 +00:00
range of device capability, the default value will be set with a warning
message. The default value is 4 which is 16 strides per a buffer, valid only
if `` mprq_en `` is set.
The size of Rx queue should be bigger than the number of strides.
- `` mprq_max_memcpy_len `` parameter [int]
The maximum length of packet to memcpy in case of Multi-Packet Rx queue. Rx
packet is mem-copied to a user-provided mbuf if the size of Rx packet is less
than or equal to this parameter. Otherwise, PMD will attach the Rx packet to
the mbuf by external buffer attachment - `` rte_pktmbuf_attach_extbuf() `` .
A mempool for external buffers will be allocated and managed by PMD. If Rx
packet is externally attached, ol_flags field of the mbuf will have
EXT_ATTACHED_MBUF and this flag must be preserved. `` RTE_MBUF_HAS_EXTBUF() ``
checks the flag. The default value is 128, valid only if `` mprq_en `` is set.
- `` rxqs_min_mprq `` parameter [int]
Configure Rx queues as Multi-Packet RQ if the total number of Rx queues is
greater or equal to this value. The default value is 12, valid only if
`` mprq_en `` is set.
2016-06-24 13:17:56 +00:00
- `` txq_inline `` parameter [int]
2019-07-21 14:24:53 +00:00
Amount of data to be inlined during TX operations. This parameter is
2019-07-21 14:24:54 +00:00
deprecated and converted to the new parameter `` txq_inline_max `` providing
partial compatibility.
2016-06-24 13:17:56 +00:00
- `` txqs_min_inline `` parameter [int]
2019-07-21 14:24:54 +00:00
Enable inline data send only when the number of TX queues is greater or equal
2016-06-24 13:17:56 +00:00
to this value.
2019-07-21 14:24:54 +00:00
This option should be used in combination with `` txq_inline_max `` and
`` txq_inline_mpw `` below and does not affect `` txq_inline_min `` settings above.
If this option is not specified the default value 16 is used for BlueField
and 8 for other platforms
The data inlining consumes the CPU cycles, so this option is intended to
auto enable inline data if we have enough Tx queues, which means we have
enough CPU cores and PCI bandwidth is getting more critical and CPU
is not supposed to be bottleneck anymore.
The copying data into WQE improves latency and can improve PPS performance
when PCI back pressure is detected and may be useful for scenarios involving
heavy traffic on many queues.
Because additional software logic is necessary to handle this mode, this
option should be used with care, as it may lower performance when back
pressure is not expected.
- `` txq_inline_min `` parameter [int]
Minimal amount of data to be inlined into WQE during Tx operations. NICs
may require this minimal data amount to operate correctly. The exact value
may depend on NIC operation mode, requested offloads, etc.
If `` txq_inline_min `` key is present the specified value (may be aligned
by the driver in order not to exceed the limits and provide better descriptor
space utilization) will be used by the driver and it is guaranteed the
requested data bytes are inlined into the WQE beside other inline settings.
This keys also may update `` txq_inline_max `` value (default of specified
explicitly in devargs) to reserve the space for inline data.
If `` txq_inline_min `` key is not present, the value may be queried by the
driver from the NIC via DevX if this feature is available. If there is no DevX
enabled/supported the value 18 (supposing L2 header including VLAN) is set
for ConnectX-4, value 58 (supposing L2-L4 headers, required by configurations
over E-Switch) is set for ConnectX-4 Lx, and 0 is set by default for ConnectX-5
and newer NICs. If packet is shorter the `` txq_inline_min `` value, the entire
packet is inlined.
For the ConnectX-4 and ConnectX-4 Lx NICs driver does not allow to set
this value below 18 (minimal L2 header, including VLAN).
Please, note, this minimal data inlining disengages eMPW feature (Enhanced
Multi-Packet Write), because last one does not support partial packet inlining.
This is not very critical due to minimal data inlining is mostly required
by ConnectX-4 and ConnectX-4 Lx, these NICs do not support eMPW feature.
- `` txq_inline_max `` parameter [int]
Specifies the maximal packet length to be completely inlined into WQE
Ethernet Segment for ordinary SEND method. If packet is larger than specified
value, the packet data won't be copied by the driver at all, data buffer
is addressed with a pointer. If packet length is less or equal all packet
data will be copied into WQE. This may improve PCI bandwidth utilization for
short packets significantly but requires the extra CPU cycles.
The data inline feature is controlled by number of Tx queues, if number of Tx
queues is larger than `` txqs_min_inline `` key parameter, the inline feature
is engaged, if there are not enough Tx queues (which means not enough CPU cores
and CPU resources are scarce), data inline is not performed by the driver.
Assigning `` txqs_min_inline `` with zero always enables the data inline.
The default `` txq_inline_max `` value is 290. The specified value may be adjusted
by the driver in order not to exceed the limit (930 bytes) and to provide better
WQE space filling without gaps, the adjustment is reflected in the debug log.
- `` txq_inline_mpw `` parameter [int]
Specifies the maximal packet length to be completely inlined into WQE for
Enhanced MPW method. If packet is large the specified value, the packet data
won't be copied, and data buffer is addressed with pointer. If packet length
is less or equal, all packet data will be copied into WQE. This may improve PCI
bandwidth utilization for short packets significantly but requires the extra
CPU cycles.
The data inline feature is controlled by number of TX queues, if number of Tx
queues is larger than `` txqs_min_inline `` key parameter, the inline feature
is engaged, if there are not enough Tx queues (which means not enough CPU cores
and CPU resources are scarce), data inline is not performed by the driver.
Assigning `` txqs_min_inline `` with zero always enables the data inline.
2019-08-05 13:03:53 +00:00
The default `` txq_inline_mpw `` value is 268. The specified value may be adjusted
2019-07-21 14:24:54 +00:00
by the driver in order not to exceed the limit (930 bytes) and to provide better
WQE space filling without gaps, the adjustment is reflected in the debug log.
Due to multiple packets may be included to the same WQE with Enhanced Multi
Packet Write Method and overall WQE size is limited it is not recommended to
specify large values for the `` txq_inline_mpw `` .
2017-07-28 14:28:33 +00:00
2018-11-01 17:20:32 +00:00
- `` txqs_max_vec `` parameter [int]
Enable vectorized Tx only when the number of TX queues is less than or
2019-07-21 14:24:53 +00:00
equal to this value. This parameter is deprecated and ignored, kept
for compatibility issue to not prevent driver from probing.
2018-11-01 17:20:32 +00:00
2017-03-15 23:55:44 +00:00
- `` txq_mpw_hdr_dseg_en `` parameter [int]
A nonzero value enables including two pointers in the first block of TX
2019-07-21 14:24:53 +00:00
descriptor. The parameter is deprecated and ignored, kept for compatibility
issue.
2017-03-15 23:55:44 +00:00
- `` txq_max_inline_len `` parameter [int]
Maximum size of packet to be inlined. This limits the size of packet to
be inlined. If the size of a packet is larger than configured value, the
packet isn't inlined even though there's enough space remained in the
2019-07-21 14:24:53 +00:00
descriptor. Instead, the packet is included with pointer. This parameter
2019-07-21 14:24:54 +00:00
is deprecated and converted directly to `` txq_inline_mpw `` providing full
compatibility. Valid only if eMPW feature is engaged.
- `` txq_mpw_en `` parameter [int]
A nonzero value enables Enhanced Multi-Packet Write (eMPW) for ConnectX-5,
ConnectX-6 and BlueField. eMPW allows the TX burst function to pack up multiple
packets in a single descriptor session in order to save PCI bandwidth and improve
performance at the cost of a slightly higher CPU usage. When `` txq_inline_mpw ``
is set along with `` txq_mpw_en `` , TX burst function copies entire packet
data on to TX descriptor instead of including pointer of packet.
The Enhanced Multi-Packet Write feature is enabled by default if NIC supports
it, can be disabled by explicit specifying 0 value for `` txq_mpw_en `` option.
Also, if minimal data inlining is requested by non-zero `` txq_inline_min ``
option or reported by the NIC, the eMPW feature is disengaged.
2017-03-15 23:55:44 +00:00
2017-08-02 15:32:56 +00:00
- `` tx_vec_en `` parameter [int]
2019-07-21 14:24:53 +00:00
A nonzero value enables Tx vector on ConnectX-5, ConnectX-6 and BlueField
NICs if the number of global Tx queues on the port is less than
`` txqs_max_vec `` . The parameter is deprecated and ignored.
2017-08-02 15:32:56 +00:00
- `` rx_vec_en `` parameter [int]
A nonzero value enables Rx vector if the port is not configured in
multi-segment otherwise this parameter is ignored.
Enabled by default.
2018-04-05 15:07:21 +00:00
- `` vf_nl_en `` parameter [int]
A nonzero value enables Netlink requests from the VF to add/remove MAC
addresses or/and enable/disable promiscuous/all multicast on the Netdevice.
Otherwise the relevant configuration must be run with Linux iproute2 tools.
This is a prerequisite to receive this kind of traffic.
Enabled by default, valid only on VF devices ignored otherwise.
2018-04-23 12:33:02 +00:00
- `` l3_vxlan_en `` parameter [int]
2018-04-23 12:33:08 +00:00
A nonzero value allows L3 VXLAN and VXLAN-GPE flow creation. To enable
L3 VXLAN or VXLAN-GPE, users has to configure firmware and enable this
parameter. This is a prerequisite to receive this kind of traffic.
2018-04-23 12:33:02 +00:00
Disabled by default.
2018-09-24 23:17:54 +00:00
- `` dv_flow_en `` parameter [int]
A nonzero value enables the DV flow steering assuming it is supported
by the driver.
Disabled by default.
2019-05-13 11:41:02 +00:00
- `` dv_esw_en `` parameter [int]
A nonzero value enables E-Switch using Direct Rules.
Enabled by default if supported.
2019-04-01 21:17:54 +00:00
- `` mr_ext_memseg_en `` parameter [int]
A nonzero value enables extending memseg when registering DMA memory. If
enabled, the number of entries in MR (Memory Region) lookup table on datapath
is minimized and it benefits performance. On the other hand, it worsens memory
utilization because registered memory is pinned by kernel driver. Even if a
page in the extended chunk is freed, that doesn't become reusable until the
entire memory is freed.
Enabled by default.
2018-07-10 16:04:58 +00:00
- `` representor `` parameter [list]
This parameter can be used to instantiate DPDK Ethernet devices from
existing port (or VF) representors configured on the device.
It is a standard parameter whose format is described in
:ref: `ethernet_device_standard_device_arguments` .
For instance, to probe port representors 0 through 2::
representor=[0-2]
2019-05-30 10:20:32 +00:00
- `` max_dump_files_num `` parameter [int]
The maximum number of files per PMD entity that may be created for debug information.
The files will be created in /var/log directory or in current directory.
set to 128 by default.
2019-07-22 14:51:59 +00:00
- `` lro_timeout_usec `` parameter [int]
The maximum allowed duration of an LRO session, in micro-seconds.
PMD will set the nearest value supported by HW, which is not bigger than
the input `` lro_timeout_usec `` value.
If this parameter is not specified, by default PMD will set
the smallest value supported by HW.
2018-04-23 12:33:02 +00:00
Firmware configuration
~~~~~~~~~~~~~~~~~~~~~~
2019-08-05 15:32:18 +00:00
Firmware features can be configured as key/value pairs.
2018-04-23 12:33:02 +00:00
2019-08-05 15:32:18 +00:00
The command to set a value is::
2018-04-23 12:33:02 +00:00
2019-08-05 15:32:18 +00:00
mlxconfig -d <device> set <key>=<value>
2018-04-23 12:33:02 +00:00
2019-08-05 15:32:18 +00:00
The command to query a value is::
2018-04-23 12:33:02 +00:00
2019-08-05 15:32:18 +00:00
mlxconfig -d <device> query | grep <key>
2018-04-23 12:33:02 +00:00
2019-08-05 15:32:18 +00:00
The device name for the command `` mlxconfig `` can be either the PCI address,
or the mst device name found with::
2018-04-23 12:33:02 +00:00
2019-08-05 15:32:18 +00:00
mst status
2019-07-03 07:22:49 +00:00
2019-08-05 15:32:18 +00:00
Below are some firmware configurations listed.
2019-07-03 07:22:49 +00:00
2019-08-05 15:32:18 +00:00
- link type::
2019-07-03 07:22:49 +00:00
2019-08-05 15:32:18 +00:00
LINK_TYPE_P1
LINK_TYPE_P2
value: 1=Infiniband 2=Ethernet 3=VPI(auto-sense)
2019-07-03 07:22:49 +00:00
2019-08-05 15:32:18 +00:00
- enable SR-IOV::
2019-07-03 07:22:49 +00:00
2019-08-05 15:32:18 +00:00
SRIOV_EN=1
2019-07-03 07:22:49 +00:00
2019-08-05 15:32:18 +00:00
- maximum number of SR-IOV virtual functions::
2019-07-10 14:59:45 +00:00
2019-08-05 15:32:18 +00:00
NUM_OF_VFS=<max>
2019-07-10 14:59:45 +00:00
2019-08-05 15:32:18 +00:00
- aggressive CQE zipping::
2019-07-10 14:59:45 +00:00
2019-08-05 15:32:18 +00:00
CQE_COMPRESSION=1
2019-07-10 14:59:45 +00:00
2019-08-05 15:32:18 +00:00
- L3 VXLAN and VXLAN-GPE destination UDP port::
2019-07-10 14:59:45 +00:00
2019-08-05 15:32:18 +00:00
IP_OVER_VXLAN_EN=1
IP_OVER_VXLAN_PORT=<udp dport>
- enable IP-in-IP tunnel flow matching::
FLEX_PARSER_PROFILE_ENABLE=0
- enable ICMP/ICMP6 code/type fields matching::
FLEX_PARSER_PROFILE_ENABLE=2
2019-07-10 14:59:45 +00:00
2015-10-30 18:52:42 +00:00
Prerequisites
-------------
This driver relies on external libraries and kernel drivers for resources
allocations and initialization. The following dependencies are not part of
DPDK and must be installed separately:
- **libibverbs**
User space Verbs framework used by librte_pmd_mlx5. This library provides
a generic interface between the kernel and low-level user space drivers
such as libmlx5.
It allows slow and privileged operations (context initialization, hardware
resources allocations) to be managed by the kernel and fast operations to
never leave user space.
- **libmlx5**
2018-05-15 06:12:50 +00:00
Low-level user space driver library for Mellanox
2019-05-13 11:41:02 +00:00
ConnectX-4/ConnectX-5/ConnectX-6/BlueField devices, it is automatically loaded
2018-05-15 06:12:50 +00:00
by libibverbs.
2015-10-30 18:52:42 +00:00
This library basically implements send/receive calls to the hardware
queues.
2017-09-26 15:38:24 +00:00
- **Kernel modules**
2015-10-30 18:52:42 +00:00
They provide the kernel-side Verbs API and low level device drivers that
manage actual hardware initialization and resources sharing with user
space processes.
Unlike most other PMDs, these modules must remain loaded and bound to
their devices:
2018-05-15 06:12:50 +00:00
- mlx5_core: hardware driver managing Mellanox
2019-05-13 11:41:02 +00:00
ConnectX-4/ConnectX-5/ConnectX-6/BlueField devices and related Ethernet kernel
2018-05-15 06:12:50 +00:00
network devices.
2015-10-30 18:52:42 +00:00
- mlx5_ib: InifiniBand device driver.
- ib_uverbs: user space driver for Verbs (entry point for libibverbs).
- **Firmware update**
2019-01-30 07:53:20 +00:00
Mellanox OFED/EN releases include firmware updates for
2019-05-13 11:41:02 +00:00
ConnectX-4/ConnectX-5/ConnectX-6/BlueField adapters.
2015-10-30 18:52:42 +00:00
Because each release provides new features, these updates must be applied to
match the kernel modules and libraries they come with.
.. note ::
Both libraries are BSD and GPL licensed. Linux kernel modules are GPL
licensed.
2017-09-26 15:38:24 +00:00
Installation
~~~~~~~~~~~~
2015-10-30 18:55:19 +00:00
2017-09-26 15:38:24 +00:00
Either RDMA Core library with a recent enough Linux kernel release
2019-01-30 07:53:20 +00:00
(recommended) or Mellanox OFED/EN, which provides compatibility with older
2017-09-26 15:38:24 +00:00
releases.
2019-05-08 09:24:09 +00:00
RDMA Core with Linux Kernel
2017-09-26 15:38:24 +00:00
^^^^^^^^^^^^^^^^^^^^^^^^^^^
2017-10-30 10:58:25 +00:00
- Minimal kernel version : v4.14 or the most recent 4.14-rc (see `Linux installation documentation`_ )
- Minimal rdma-core version: v15+ commit 0c5f5765213a ("Merge pull request #227 from yishaih/tm")
(see `RDMA Core installation documentation`_ )
2018-07-12 12:01:31 +00:00
- When building for i686 use:
- rdma-core version 18.0 or above built with 32bit support.
- Kernel version 4.14.41 or above.
2017-09-26 15:38:24 +00:00
2019-01-09 14:23:19 +00:00
- Starting with rdma-core v21, static libraries can be built::
cd build
CFLAGS=-fPIC cmake -DIN_PLACE=1 -DENABLE_STATIC=1 -GNinja ..
ninja
2017-09-26 15:38:24 +00:00
.. _`Linux installation documentation`: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git/plain/Documentation/admin-guide/README.rst
.. _`RDMA Core installation documentation`: https://raw.githubusercontent.com/linux-rdma/rdma-core/master/README.md
2019-01-09 14:23:19 +00:00
If rdma-core libraries are built but not installed, DPDK makefile can link them,
thanks to these environment variables:
- `` EXTRA_CFLAGS=-I/path/to/rdma-core/build/include ``
- `` EXTRA_LDFLAGS=-L/path/to/rdma-core/build/lib ``
- `` PKG_CONFIG_PATH=/path/to/rdma-core/build/lib/pkgconfig ``
2019-01-30 07:53:20 +00:00
Mellanox OFED/EN
^^^^^^^^^^^^^^^^
2017-09-26 15:38:24 +00:00
2019-05-13 11:41:02 +00:00
- Mellanox OFED version: ** 4.5, 4.6** /
Mellanox EN version: **4.5, 4.6**
2016-11-02 13:46:43 +00:00
- firmware version:
2015-12-12 19:43:24 +00:00
2017-10-30 10:58:25 +00:00
- ConnectX-4: **12.21.1000** and above.
- ConnectX-4 Lx: **14.21.1000** and above.
- ConnectX-5: **16.21.1000** and above.
- ConnectX-5 Ex: **16.21.1000** and above.
2018-12-31 12:43:48 +00:00
- ConnectX-6: **20.99.5374** and above.
2019-05-13 11:41:02 +00:00
- BlueField: **18.25.1010** and above.
2016-03-03 14:26:43 +00:00
2015-10-30 18:52:42 +00:00
While these libraries and kernel modules are available on OpenFabrics
Alliance's `website <https://www.openfabrics.org/> `__ and provided by package
managers on most distributions, this PMD requires Ethernet extensions that
may not be supported at the moment (this is a work in progress).
`Mellanox OFED
2019-01-30 07:53:20 +00:00
<http://www.mellanox.com/page/products_dyn?product_family=26&mtag=linux> `__ and
`Mellanox EN
<http://www.mellanox.com/page/products_dyn?product_family=27&mtag=linux> `__
include the necessary support and should be used in the meantime. For DPDK,
2015-10-30 18:52:42 +00:00
only libibverbs, libmlx5, mlnx-ofed-kernel packages and firmware updates are
required from that distribution.
.. note ::
2019-01-30 07:53:20 +00:00
Several versions of Mellanox OFED/EN are available. Installing the version
2015-10-30 18:52:42 +00:00
this DPDK release was developed and tested against is strongly
recommended. Please check the `prerequisites`_ .
2017-02-09 08:32:01 +00:00
Supported NICs
--------------
* Mellanox(R) ConnectX(R)-4 10G MCX4111A-XCAT (1x10G)
* Mellanox(R) ConnectX(R)-4 10G MCX4121A-XCAT (2x10G)
* Mellanox(R) ConnectX(R)-4 25G MCX4111A-ACAT (1x25G)
* Mellanox(R) ConnectX(R)-4 25G MCX4121A-ACAT (2x25G)
* Mellanox(R) ConnectX(R)-4 40G MCX4131A-BCAT (1x40G)
* Mellanox(R) ConnectX(R)-4 40G MCX413A-BCAT (1x40G)
* Mellanox(R) ConnectX(R)-4 40G MCX415A-BCAT (1x40G)
* Mellanox(R) ConnectX(R)-4 50G MCX4131A-GCAT (1x50G)
* Mellanox(R) ConnectX(R)-4 50G MCX413A-GCAT (1x50G)
* Mellanox(R) ConnectX(R)-4 50G MCX414A-BCAT (2x50G)
* Mellanox(R) ConnectX(R)-4 50G MCX415A-GCAT (2x50G)
* Mellanox(R) ConnectX(R)-4 50G MCX416A-BCAT (2x50G)
* Mellanox(R) ConnectX(R)-4 50G MCX416A-GCAT (2x50G)
* Mellanox(R) ConnectX(R)-4 50G MCX415A-CCAT (1x100G)
* Mellanox(R) ConnectX(R)-4 100G MCX416A-CCAT (2x100G)
* Mellanox(R) ConnectX(R)-4 Lx 10G MCX4121A-XCAT (2x10G)
* Mellanox(R) ConnectX(R)-4 Lx 25G MCX4121A-ACAT (2x25G)
* Mellanox(R) ConnectX(R)-5 100G MCX556A-ECAT (2x100G)
* Mellanox(R) ConnectX(R)-5 Ex EN 100G MCX516A-CDAT (2x100G)
2019-01-30 07:53:20 +00:00
Quick Start Guide on OFED/EN
----------------------------
2017-07-28 14:28:33 +00:00
2019-01-30 07:53:20 +00:00
1. Download latest Mellanox OFED/EN. For more info check the `prerequisites`_ .
2017-06-13 10:20:58 +00:00
2017-07-28 14:28:33 +00:00
2. Install the required libraries and kernel modules either by installing
2019-01-30 07:53:20 +00:00
only the required set, or by installing the entire Mellanox OFED/EN:
2017-06-13 10:20:58 +00:00
2017-07-28 14:28:33 +00:00
.. code-block :: console
2017-06-13 10:20:58 +00:00
2017-10-30 10:58:25 +00:00
./mlnxofedinstall --upstream-libs --dpdk
2017-06-13 10:20:58 +00:00
2017-07-28 14:28:33 +00:00
3. Verify the firmware is the correct one:
2017-06-13 10:20:58 +00:00
2017-07-28 14:28:33 +00:00
.. code-block :: console
2017-06-13 10:20:58 +00:00
2017-07-28 14:28:33 +00:00
ibv_devinfo
2017-06-13 10:20:58 +00:00
2017-07-28 14:28:33 +00:00
4. Verify all ports links are set to Ethernet:
.. code-block :: console
mlxconfig -d <mst device> query | grep LINK_TYPE
LINK_TYPE_P1 ETH(2)
LINK_TYPE_P2 ETH(2)
Link types may have to be configured to Ethernet:
.. code-block :: console
mlxconfig -d <mst device> set LINK_TYPE_P1/2=1/2/3
* LINK_TYPE_P1=<1|2|3> , 1=Infiniband 2=Ethernet 3=VPI(auto-sense)
For hypervisors verify SR-IOV is enabled on the NIC:
.. code-block :: console
mlxconfig -d <mst device> query | grep SRIOV_EN
SRIOV_EN True(1)
If needed, set enable the set the relevant fields:
.. code-block :: console
mlxconfig -d <mst device> set SRIOV_EN=1 NUM_OF_VFS=16
mlxfwreset -d <mst device> reset
5. Restart the driver:
.. code-block :: console
/etc/init.d/openibd restart
or:
.. code-block :: console
service openibd restart
If link type was changed, firmware must be reset as well:
.. code-block :: console
mlxfwreset -d <mst device> reset
For hypervisors, after reset write the sysfs number of virtual functions
needed for the PF.
To dynamically instantiate a given number of virtual functions (VFs):
.. code-block :: console
echo [num_vfs] > /sys/class/infiniband/mlx5_0/device/sriov_numvfs
6. Compile DPDK and you are ready to go. See instructions on
:ref: `Development Kit Build System <Development_Kit_Build_System>`
2019-05-13 11:41:02 +00:00
Enable switchdev mode
---------------------
Switchdev mode is a mode in E-Switch, that binds between representor and VF.
Representor is a port in DPDK that is connected to a VF in such a way
that assuming there are no offload flows, each packet that is sent from the VF
will be received by the corresponding representor. While each packet that is
sent to a representor will be received by the VF.
This is very useful in case of SRIOV mode, where the first packet that is sent
by the VF will be received by the DPDK application which will decide if this
flow should be offloaded to the E-Switch. After offloading the flow packet
that the VF that are matching the flow will not be received any more by
the DPDK application.
1. Enable SRIOV mode:
.. code-block :: console
mlxconfig -d <mst device> set SRIOV_EN=true
2. Configure the max number of VFs:
.. code-block :: console
mlxconfig -d <mst device> set NUM_OF_VFS=<num of vfs>
3. Reset the FW:
.. code-block :: console
mlxfwreset -d <mst device> reset
3. Configure the actual number of VFs:
.. code-block :: console
echo <num of vfs > /sys/class/net/<net device>/device/sriov_numvfs
4. Unbind the device (can be rebind after the switchdev mode):
.. code-block :: console
echo -n "<device pci address" > /sys/bus/pci/drivers/mlx5_core/unbind
5. Enbale switchdev mode:
.. code-block :: console
echo switchdev > /sys/class/net/<net device>/compat/devlink/mode
2017-07-28 14:28:33 +00:00
Performance tuning
------------------
1. Configure aggressive CQE Zipping for maximum performance:
.. code-block :: console
mlxconfig -d <mst device> s CQE_COMPRESSION=1
To set it back to the default CQE Zipping mode use:
.. code-block :: console
mlxconfig -d <mst device> s CQE_COMPRESSION=0
2. In case of virtualization:
- Make sure that hypervisor kernel is 3.16 or newer.
- Configure boot with `` iommu=pt `` .
- Use 1G huge pages.
- Make sure to allocate a VM on huge pages.
- Make sure to set CPU pinning.
3. Use the CPU near local NUMA node to which the PCIe adapter is connected,
for better performance. For VMs, verify that the right CPU
and NUMA node are pinned according to the above. Run:
.. code-block :: console
lstopo-no-graphics
to identify the NUMA node to which the PCIe adapter is connected.
4. If more than one adapter is used, and root complex capabilities allow
to put both adapters on the same NUMA node without PCI bandwidth degradation,
it is recommended to locate both adapters on the same NUMA node.
This in order to forward packets from one to the other without
NUMA performance penalty.
5. Disable pause frames:
.. code-block :: console
ethtool -A <netdev> rx off tx off
6. Verify IO non-posted prefetch is disabled by default. This can be checked
via the BIOS configuration. Please contact you server provider for more
information about the settings.
.. note ::
On some machines, depends on the machine integrator, it is beneficial
to set the PCI max read request parameter to 1K. This can be
done in the following way:
To query the read request size use:
.. code-block :: console
setpci -s <NIC PCI address> 68.w
If the output is different than 3XXX, set it by:
.. code-block :: console
setpci -s <NIC PCI address> 68.w=3XXX
The XXX can be different on different systems. Make sure to configure
according to the setpci output.
2017-06-13 10:20:58 +00:00
net/mlx5: add new memory region support
This is the new design of Memory Region (MR) for mlx PMD, in order to:
- Accommodate the new memory hotplug model.
- Support non-contiguous Mempool.
There are multiple layers for MR search.
L0 is to look up the last-hit entry which is pointed by mr_ctrl->mru (Most
Recently Used). If L0 misses, L1 is to look up the address in a fixed-sized
array by linear search. L0/L1 is in an inline function -
mlx5_mr_lookup_cache().
If L1 misses, the bottom-half function is called to look up the address
from the bigger local cache of the queue. This is L2 - mlx5_mr_addr2mr_bh()
and it is not an inline function. Data structure for L2 is the Binary Tree.
If L2 misses, the search falls into the slowest path which takes locks in
order to access global device cache (priv->mr.cache) which is also a B-tree
and caches the original MR list (priv->mr.mr_list) of the device. Unless
the global cache is overflowed, it is all-inclusive of the MR list. This is
L3 - mlx5_mr_lookup_dev(). The size of the L3 cache table is limited and
can't be expanded on the fly due to deadlock. Refer to the comments in the
code for the details - mr_lookup_dev(). If L3 is overflowed, the list will
have to be searched directly bypassing the cache although it is slower.
If L3 misses, a new MR for the address should be created -
mlx5_mr_create(). When it creates a new MR, it tries to register adjacent
memsegs as much as possible which are virtually contiguous around the
address. This must take two locks - memory_hotplug_lock and
priv->mr.rwlock. Due to memory_hotplug_lock, there can't be any
allocation/free of memory inside.
In the free callback of the memory hotplug event, freed space is searched
from the MR list and corresponding bits are cleared from the bitmap of MRs.
This can fragment a MR and the MR will have multiple search entries in the
caches. Once there's a change by the event, the global cache must be
rebuilt and all the per-queue caches will be flushed as well. If memory is
frequently freed in run-time, that may cause jitter on dataplane processing
in the worst case by incurring MR cache flush and rebuild. But, it would be
the least probable scenario.
To guarantee the most optimal performance, it is highly recommended to use
an EAL option - '--socket-mem'. Then, the reserved memory will be pinned
and won't be freed dynamically. And it is also recommended to configure
per-lcore cache of Mempool. Even though there're many MRs for a device or
MRs are highly fragmented, the cache of Mempool will be much helpful to
reduce misses on per-queue caches anyway.
'--legacy-mem' is also supported.
Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
2018-05-09 11:09:04 +00:00
7. To minimize overhead of searching Memory Regions:
- '--socket-mem' is recommended to pin memory by predictable amount.
- Configure per-lcore cache when creating Mempools for packet buffer.
- Refrain from dynamically allocating/freeing memory in run-time.
2019-05-13 11:41:02 +00:00
Supported hardware offloads using rte_flow API
----------------------------------------------
.. _Supported hardware offloads using rte_flow API:
.. table :: Supported hardware offloads using rte_flow API
+-----------------------+-----------------+-----------------+
| Offload | E-Switch | NIC |
| | | |
+=======================+=================+=================+
| Count | | DPDK 19.05 | | DPDK 19.02 |
| | | OFED 4.6 | | OFED 4.6 |
| | | RDMA-CORE V24 | | RDMA-CORE V23 |
| | | ConnectX-5 | | ConnectX-5 |
+-----------------------+-----------------+-----------------+
| Drop / Queue / RSS | | DPDK 19.05 | | DPDK 18.11 |
| | | OFED 4.6 | | OFED 4.5 |
| | | RDMA-CORE V24 | | RDMA-CORE V23 |
| | | ConnectX-5 | | ConnectX-4 |
+-----------------------+-----------------+-----------------+
| Encapsulation | | DPDK 19.05 | | DPDK 19.02 |
| (VXLAN / NVGRE / RAW) | | OFED 4.6.2 | | OFED 4.6 |
| | | RDMA-CORE V24 | | RDMA-CORE V23 |
| | | ConnectX-5 | | ConnectX-5 |
+-----------------------+-----------------+-----------------+
| Header rewrite | | DPDK 19.05 | | DPDK 19.02 |
| (set_ipv4_src / | | OFED 4.6.2 | | OFED 4.6.2 |
| set_ipv4_dst / | | RDMA-CORE V24 | | RDMA-CORE V23 |
| set_ipv6_src / | | ConnectX-5 | | ConnectX-5 |
| set_ipv6_dst / | | |
| set_tp_src / | | |
| set_tp_dst / | | |
| dec_ttl / | | |
| set_ttl / | | |
| set_mac_src / | | |
| set_mac_dst) | | |
+-----------------------+-----------------+-----------------+
| Jump | | DPDK 19.05 | | DPDK 19.02 |
| | | OFED 4.6.2 | | OFED 4.6.2 |
| | | RDMA-CORE V24 | | N/A |
| | | ConnectX-5 | | ConnectX-5 |
+-----------------------+-----------------+-----------------+
| Mark / Flag | | DPDK 19.05 | | DPDK 18.11 |
| | | OFED 4.6 | | OFED 4.5 |
| | | RDMA-CORE V24 | | RDMA-CORE V23 |
| | | ConnectX-5 | | ConnectX-4 |
+-----------------------+-----------------+-----------------+
| Port ID | | DPDK 19.05 | | N/A |
| | | OFED 4.6 | | N/A |
| | | RDMA-CORE V24 | | N/A |
| | | ConnectX-5 | | N/A |
+-----------------------+-----------------+-----------------+
* Minimum version for each component and nic.
2015-12-12 19:43:24 +00:00
Notes for testpmd
-----------------
Compared to librte_pmd_mlx4 that implements a single RSS configuration per
port, librte_pmd_mlx5 supports per-protocol RSS configuration.
Since `` testpmd `` defaults to IP RSS mode and there is currently no
command-line parameter to enable additional protocols (UDP and TCP as well
as IP), the following commands must be entered from its CLI to get the same
behavior as librte_pmd_mlx4:
.. code-block :: console
> port stop all
> port config all rss all
> port start all
2015-10-30 18:52:42 +00:00
Usage example
-------------
2017-01-06 00:49:31 +00:00
This section demonstrates how to launch **testpmd** with Mellanox
2019-05-13 11:41:02 +00:00
ConnectX-4/ConnectX-5/ConnectX-6/BlueField devices managed by librte_pmd_mlx5.
2015-10-30 18:52:42 +00:00
#. Load the kernel modules:
.. code-block :: console
modprobe -a ib_uverbs mlx5_core mlx5_ib
2019-01-30 07:53:20 +00:00
Alternatively if MLNX_OFED/MLNX_EN is fully installed, the following script
can be run:
2015-12-12 19:43:24 +00:00
.. code-block :: console
/etc/init.d/openibd restart
2015-10-30 18:52:42 +00:00
.. note ::
User space I/O kernel modules (uio and igb_uio) are not used and do
not have to be loaded.
#. Make sure Ethernet interfaces are in working order and linked to kernel
verbs. Related sysfs entries should be present:
.. code-block :: console
ls -d /sys/class/net/*/device/infiniband_verbs/uverbs* | cut -d / -f 5
Example output:
.. code-block :: console
eth30
eth31
eth32
eth33
#. Optionally, retrieve their PCI bus addresses for whitelisting:
.. code-block :: console
{
for intf in eth2 eth3 eth4 eth5;
do
(cd "/sys/class/net/${intf}/device/" && pwd -P);
done;
} |
sed -n 's,.*/\(.* \),-w \1,p'
Example output:
.. code-block :: console
-w 0000:05:00.1
-w 0000:06:00.0
-w 0000:06:00.1
-w 0000:05:00.0
#. Request huge pages:
.. code-block :: console
echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages/nr_hugepages
#. Start testpmd with basic parameters:
.. code-block :: console
2017-02-27 19:13:40 +00:00
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
2015-10-30 18:52:42 +00:00
Example output:
.. code-block :: console
[...]
EAL: PCI device 0000:05:00.0 on NUMA socket 0
EAL: probe driver: 15b3:1013 librte_pmd_mlx5
PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_0" (VF: false)
PMD: librte_pmd_mlx5: 1 port(s) detected
PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fe
EAL: PCI device 0000:05:00.1 on NUMA socket 0
EAL: probe driver: 15b3:1013 librte_pmd_mlx5
PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_1" (VF: false)
PMD: librte_pmd_mlx5: 1 port(s) detected
PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:ff
EAL: PCI device 0000:06:00.0 on NUMA socket 0
EAL: probe driver: 15b3:1013 librte_pmd_mlx5
PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_2" (VF: false)
PMD: librte_pmd_mlx5: 1 port(s) detected
PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fa
EAL: PCI device 0000:06:00.1 on NUMA socket 0
EAL: probe driver: 15b3:1013 librte_pmd_mlx5
PMD: librte_pmd_mlx5: PCI information matches, using device "mlx5_3" (VF: false)
PMD: librte_pmd_mlx5: 1 port(s) detected
PMD: librte_pmd_mlx5: port 1 MAC address is e4:1d:2d:e7:0c:fb
Interactive-mode selected
Configuring Port 0 (socket 0)
PMD: librte_pmd_mlx5: 0x8cba80: TX queues number update: 0 -> 2
PMD: librte_pmd_mlx5: 0x8cba80: RX queues number update: 0 -> 2
Port 0: E4:1D:2D:E7:0C:FE
Configuring Port 1 (socket 0)
PMD: librte_pmd_mlx5: 0x8ccac8: TX queues number update: 0 -> 2
PMD: librte_pmd_mlx5: 0x8ccac8: RX queues number update: 0 -> 2
Port 1: E4:1D:2D:E7:0C:FF
Configuring Port 2 (socket 0)
PMD: librte_pmd_mlx5: 0x8cdb10: TX queues number update: 0 -> 2
PMD: librte_pmd_mlx5: 0x8cdb10: RX queues number update: 0 -> 2
Port 2: E4:1D:2D:E7:0C:FA
Configuring Port 3 (socket 0)
PMD: librte_pmd_mlx5: 0x8ceb58: TX queues number update: 0 -> 2
PMD: librte_pmd_mlx5: 0x8ceb58: RX queues number update: 0 -> 2
Port 3: E4:1D:2D:E7:0C:FB
Checking link statuses...
Port 0 Link Up - speed 40000 Mbps - full-duplex
Port 1 Link Up - speed 40000 Mbps - full-duplex
Port 2 Link Up - speed 10000 Mbps - full-duplex
Port 3 Link Up - speed 10000 Mbps - full-duplex
Done
testpmd>
