2 Copyright (c) 2017, Cisco Systems, Inc.
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
9 1. Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
12 2. Redistributions in binary form must reproduce the above copyright
13 notice, this list of conditions and the following disclaimer in
14 the documentation and/or other materials provided with the
17 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
20 FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
21 COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
22 INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
23 BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
24 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
25 CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
27 ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28 POSSIBILITY OF SUCH DAMAGE.
33 ENIC PMD is the DPDK poll-mode driver for the Cisco System Inc. VIC Ethernet
34 NICs. These adapters are also referred to as vNICs below. If you are running
35 or would like to run DPDK software applications on Cisco UCS servers using
36 Cisco VIC adapters the following documentation is relevant.
38 How to obtain ENIC PMD integrated DPDK
39 --------------------------------------
41 ENIC PMD support is integrated into the DPDK suite. dpdk-<version>.tar.gz
42 should be downloaded from http://dpdk.org
45 Configuration information
46 -------------------------
48 - **DPDK Configuration Parameters**
50 The following configuration options are available for the ENIC PMD:
52 - **CONFIG_RTE_LIBRTE_ENIC_PMD** (default y): Enables or disables inclusion
53 of the ENIC PMD driver in the DPDK compilation.
55 - **CONFIG_RTE_LIBRTE_ENIC_DEBUG** (default n): Enables or disables debug
56 logging within the ENIC PMD driver.
58 - **CONFIG_RTE_LIBRTE_ENIC_DEBUG_FLOW** (default n): Enables or disables flow
59 API related debug logging within the ENIC PMD driver.
61 - **vNIC Configuration Parameters**
63 - **Number of Queues**
65 The maximum number of receive queues (RQs), work queues (WQs) and
66 completion queues (CQs) are configurable on a per vNIC basis
67 through the Cisco UCS Manager (CIMC or UCSM).
69 These values should be configured as follows:
71 - The number of WQs should be greater or equal to the value of the
72 expected nb_tx_q parameter in the call to
73 rte_eth_dev_configure()
75 - The number of RQs configured in the vNIC should be greater or
76 equal to *twice* the value of the expected nb_rx_q parameter in
77 the call to rte_eth_dev_configure(). With the addition of Rx
78 scatter, a pair of RQs on the vnic is needed for each receive
79 queue used by DPDK, even if Rx scatter is not being used.
80 Having a vNIC with only 1 RQ is not a valid configuration, and
81 will fail with an error message.
83 - The number of CQs should set so that there is one CQ for each
84 WQ, and one CQ for each pair of RQs.
86 For example: If the application requires 3 Rx queues, and 3 Tx
87 queues, the vNIC should be configured to have at least 3 WQs, 6
88 RQs (3 pairs), and 6 CQs (3 for use by WQs + 3 for use by the 3
93 Likewise, the number of receive and transmit descriptors are configurable on
94 a per-vNIC basis via the UCS Manager and should be greater than or equal to
95 the nb_rx_desc and nb_tx_desc parameters expected to be used in the calls
96 to rte_eth_rx_queue_setup() and rte_eth_tx_queue_setup() respectively.
97 An application requesting more than the set size will be limited to that
100 Unless there is a lack of resources due to creating many vNICs, it
101 is recommended that the WQ and RQ sizes be set to the maximum. This
102 gives the application the greatest amount of flexibility in its
105 - *Note*: Since the introduction of Rx scatter, for performance
106 reasons, this PMD uses two RQs on the vNIC per receive queue in
107 DPDK. One RQ holds descriptors for the start of a packet, and the
108 second RQ holds the descriptors for the rest of the fragments of
109 a packet. This means that the nb_rx_desc parameter to
110 rte_eth_rx_queue_setup() can be a greater than 4096. The exact
111 amount will depend on the size of the mbufs being used for
112 receives, and the MTU size.
114 For example: If the mbuf size is 2048, and the MTU is 9000, then
115 receiving a full size packet will take 5 descriptors, 1 from the
116 start-of-packet queue, and 4 from the second queue. Assuming
117 that the RQ size was set to the maximum of 4096, then the
118 application can specify up to 1024 + 4096 as the nb_rx_desc
119 parameter to rte_eth_rx_queue_setup().
123 Only one interrupt per vNIC interface should be configured in the UCS
124 manager regardless of the number receive/transmit queues. The ENIC PMD
125 uses this interrupt to get information about link status and errors
128 .. _enic-flow-director:
130 Flow director support
131 ---------------------
133 Advanced filtering support was added to 1300 series VIC firmware starting
134 with version 2.0.13 for C-series UCS servers and version 3.1.2 for UCSM
135 managed blade servers. In order to enable advanced filtering the 'Advanced
136 filter' radio button should be enabled via CIMC or UCSM followed by a reboot
139 With advanced filters, perfect matching of all fields of IPv4, IPv6 headers
140 as well as TCP, UDP and SCTP L4 headers is available through flow director.
141 Masking of these fields for partial match is also supported.
143 Without advanced filter support, the flow director is limited to IPv4
144 perfect filtering of the 5-tuple with no masking of fields supported.
146 SR-IOV mode utilization
147 -----------------------
149 UCS blade servers configured with dynamic vNIC connection policies in UCS
150 manager are capable of supporting assigned devices on virtual machines (VMs)
151 through a KVM hypervisor. Assigned devices, also known as 'passthrough'
152 devices, are SR-IOV virtual functions (VFs) on the host which are exposed
155 The Cisco Virtual Machine Fabric Extender (VM-FEX) gives the VM a dedicated
156 interface on the Fabric Interconnect (FI). Layer 2 switching is done at
157 the FI. This may eliminate the requirement for software switching on the
158 host to route intra-host VM traffic.
160 Please refer to `Creating a Dynamic vNIC Connection Policy
161 <http://www.cisco.com/c/en/us/td/docs/unified_computing/ucs/sw/vm_fex/vmware/gui/config_guide/b_GUI_VMware_VM-FEX_UCSM_Configuration_Guide/b_GUI_VMware_VM-FEX_UCSM_Configuration_Guide_chapter_010.html#task_433E01651F69464783A68E66DA8A47A5>`_
162 for information on configuring SR-IOV adapter policies using UCS manager.
164 Once the policies are in place and the host OS is rebooted, VFs should be
165 visible on the host, E.g.:
167 .. code-block:: console
169 # lspci | grep Cisco | grep Ethernet
170 0d:00.0 Ethernet controller: Cisco Systems Inc VIC Ethernet NIC (rev a2)
171 0d:00.1 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
172 0d:00.2 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
173 0d:00.3 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
174 0d:00.4 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
175 0d:00.5 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
176 0d:00.6 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
177 0d:00.7 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
179 Enable Intel IOMMU on the host and install KVM and libvirt. A VM instance should
180 be created with an assigned device. When using libvirt, this configuration can
181 be done within the domain (i.e. VM) config file. For example this entry maps
182 host VF 0d:00:01 into the VM.
184 .. code-block:: console
186 <interface type='hostdev' managed='yes'>
187 <mac address='52:54:00:ac:ff:b6'/>
189 <address type='pci' domain='0x0000' bus='0x0d' slot='0x00' function='0x1'/>
192 Alternatively, the configuration can be done in a separate file using the
193 ``network`` keyword. These methods are described in the libvirt documentation for
194 `Network XML format <https://libvirt.org/formatnetwork.html>`_.
196 When the VM instance is started, the ENIC KVM driver will bind the host VF to
197 vfio, complete provisioning on the FI and bring up the link.
201 It is not possible to use a VF directly from the host because it is not
202 fully provisioned until the hypervisor brings up the VM that it is assigned
205 In the VM instance, the VF will now be visible. E.g., here the VF 00:04.0 is
206 seen on the VM instance and should be available for binding to a DPDK.
208 .. code-block:: console
211 00:04.0 Ethernet controller: Cisco Systems Inc VIC SR-IOV VF (rev a2)
213 Follow the normal DPDK install procedure, binding the VF to either ``igb_uio``
214 or ``vfio`` in non-IOMMU mode.
216 Please see :ref:`Limitations <enic_limitations>` for limitations in
219 .. _enic-genic-flow-api:
221 Generic Flow API support
222 ------------------------
224 Generic Flow API is supported. The baseline support is:
226 - **1200 series VICs**
228 5-tuple exact flow support for 1200 series adapters. This allows:
230 - Attributes: ingress
231 - Items: ipv4, ipv6, udp, tcp (must exactly match src/dst IP
232 addresses and ports and all must be specified)
233 - Actions: queue and void
236 - **1300 series VICS with advanced filters disabled**
238 With advanced filters disabled, an IPv4 or IPv6 item must be specified
241 - Attributes: ingress
242 - Items: eth, ipv4, ipv6, udp, tcp, vxlan, inner eth, ipv4, ipv6, udp, tcp
243 - Actions: queue and void
244 - Selectors: 'is', 'spec' and 'mask'. 'last' is not supported
245 - In total, up to 64 bytes of mask is allowed across all headers
247 - **1300 series VICS with advanced filters enabled**
249 - Attributes: ingress
250 - Items: eth, ipv4, ipv6, udp, tcp, vxlan, inner eth, ipv4, ipv6, udp, tcp
251 - Actions: queue, mark, flag and void
252 - Selectors: 'is', 'spec' and 'mask'. 'last' is not supported
253 - In total, up to 64 bytes of mask is allowed across all headers
255 More features may be added in future firmware and new versions of the VIC.
256 Please refer to the release notes.
258 .. _enic_limitations:
263 - **VLAN 0 Priority Tagging**
265 If a vNIC is configured in TRUNK mode by the UCS manager, the adapter will
266 priority tag egress packets according to 802.1Q if they were not already
267 VLAN tagged by software. If the adapter is connected to a properly configured
268 switch, there will be no unexpected behavior.
270 In test setups where an Ethernet port of a Cisco adapter in TRUNK mode is
271 connected point-to-point to another adapter port or connected though a router
272 instead of a switch, all ingress packets will be VLAN tagged. Programs such
273 as l3fwd which do not account for VLAN tags in packets will misbehave. The
274 solution is to enable VLAN stripping on ingress. The following code fragment is
275 an example of how to accomplish this:
277 .. code-block:: console
279 vlan_offload = rte_eth_dev_get_vlan_offload(port);
280 vlan_offload |= ETH_VLAN_STRIP_OFFLOAD;
281 rte_eth_dev_set_vlan_offload(port, vlan_offload);
283 - Limited flow director support on 1200 series and 1300 series Cisco VIC
284 adapters with old firmware. Please see :ref:`enic-flow-director`.
286 - Flow director features are not supported on generation 1 Cisco VIC adapters
291 - KVM hypervisor support only. VMware has not been tested.
292 - Requires VM-FEX, and so is only available on UCS managed servers connected
293 to Fabric Interconnects. It is not on standalone C-Series servers.
294 - VF devices are not usable directly from the host. They can only be used
295 as assigned devices on VM instances.
296 - Currently, unbind of the ENIC kernel mode driver 'enic.ko' on the VM
297 instance may hang. As a workaround, enic.ko should be blacklisted or removed
298 from the boot process.
299 - pci_generic cannot be used as the uio module in the VM. igb_uio or
300 vfio in non-IOMMU mode can be used.
301 - The number of RQs in UCSM dynamic vNIC configurations must be at least 2.
302 - The number of SR-IOV devices is limited to 256. Components on target system
303 might limit this number to fewer than 256.
307 - The number of filters that can be specified with the Generic Flow API is
308 dependent on how many header fields are being masked. Use 'flow create' in
309 a loop to determine how many filters your VIC will support (not more than
310 1000 for 1300 series VICs). Filters are checked for matching in the order they
311 were added. Since there currently is no grouping or priority support,
312 'catch-all' filters should be added last.
314 How to build the suite
315 ----------------------
317 The build instructions for the DPDK suite should be followed. By default
318 the ENIC PMD library will be built into the DPDK library.
320 Refer to the document :ref:`compiling and testing a PMD for a NIC
321 <pmd_build_and_test>` for details.
323 For configuring and using UIO and VFIO frameworks, please refer to the
324 documentation that comes with DPDK suite.
326 Supported Cisco VIC adapters
327 ----------------------------
329 ENIC PMD supports all recent generations of Cisco VIC adapters including:
343 Supported Operating Systems
344 ---------------------------
346 Any Linux distribution fulfilling the conditions described in Dependencies
347 section of DPDK documentation.
352 - Unicast, multicast and broadcast transmission and reception
353 - Receive queue polling
354 - Port Hardware Statistics
355 - Hardware VLAN acceleration
356 - IP checksum offload
357 - Receive side VLAN stripping
358 - Multiple receive and transmit queues
359 - Flow Director ADD, UPDATE, DELETE, STATS operation support IPv4 and IPv6
361 - Setting RX VLAN (supported via UCSM/CIMC only)
362 - VLAN filtering (supported via UCSM/CIMC only)
363 - Execution of application by unprivileged system users
364 - IPV4, IPV6 and TCP RSS hashing
367 - SR-IOV on UCS managed servers connected to Fabric Interconnects
370 Known bugs and unsupported features in this release
371 ---------------------------------------------------
373 - Signature or flex byte based flow direction
374 - Drop feature of flow direction
375 - VLAN based flow direction
376 - Non-IPV4 flow direction
377 - Setting of extended VLAN
379 - MTU update only works if Scattered Rx mode is disabled
384 - Prepare the system as recommended by DPDK suite. This includes environment
385 variables, hugepages configuration, tool-chains and configuration.
386 - Insert vfio-pci kernel module using the command 'modprobe vfio-pci' if the
387 user wants to use VFIO framework.
388 - Insert uio kernel module using the command 'modprobe uio' if the user wants
389 to use UIO framework.
390 - DPDK suite should be configured based on the user's decision to use VFIO or
392 - If the vNIC device(s) to be used is bound to the kernel mode Ethernet driver
393 use 'ip' to bring the interface down. The dpdk-devbind.py tool can
394 then be used to unbind the device's bus id from the ENIC kernel mode driver.
395 - Bind the intended vNIC to vfio-pci in case the user wants ENIC PMD to use
396 VFIO framework using dpdk-devbind.py.
397 - Bind the intended vNIC to igb_uio in case the user wants ENIC PMD to use
398 UIO framework using dpdk-devbind.py.
400 At this point the system should be ready to run DPDK applications. Once the
401 application runs to completion, the vNIC can be detached from vfio-pci or
402 igb_uio if necessary.
404 Root privilege is required to bind and unbind vNICs to/from VFIO/UIO.
405 VFIO framework helps an unprivileged user to run the applications.
406 For an unprivileged user to run the applications on DPDK and ENIC PMD,
407 it may be necessary to increase the maximum locked memory of the user.
408 The following command could be used to do this.
410 .. code-block:: console
412 sudo sh -c "ulimit -l <value in Kilo Bytes>"
414 The value depends on the memory configuration of the application, DPDK and
415 PMD. Typically, the limit has to be raised to higher than 2GB.
418 The compilation of any unused drivers can be disabled using the
419 configuration file in config/ directory (e.g., config/common_linuxapp).
420 This would help in bringing down the time taken for building the
421 libraries and the initialization time of the application.
426 - https://www.cisco.com/c/en/us/products/servers-unified-computing/index.html
427 - https://www.cisco.com/c/en/us/products/interfaces-modules/unified-computing-system-adapters/index.html
432 Any questions or bugs should be reported to DPDK community and to the ENIC PMD
435 - John Daley <johndale@cisco.com>
436 - Nelson Escobar <neescoba@cisco.com>