xref: /dpdk/doc/guides/nics/ice.rst (revision 68a03efeed657e6e05f281479b33b51102797e15)
1..  SPDX-License-Identifier: BSD-3-Clause
2    Copyright(c) 2018 Intel Corporation.
3
4ICE Poll Mode Driver
5======================
6
7The ice PMD (**librte_net_ice**) provides poll mode driver support for
810/25/50/100 Gbps Intel® Ethernet 800 Series Network Adapters based on
9the Intel Ethernet Controller E810 and Intel Ethernet Connection E822/E823.
10
11Prerequisites
12-------------
13
14- Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup the basic DPDK environment.
15
16- To get better performance on Intel platforms, please follow the "How to get best performance with NICs on Intel platforms"
17  section of the :ref:`Getting Started Guide for Linux <linux_gsg>`.
18
19- Please follow the matching list to download specific kernel driver, firmware and DDP package from
20  `https://www.intel.com/content/www/us/en/search.html?ws=text#q=e810&t=Downloads&layout=table`.
21
22- To understand what is DDP package and how it works, please review `Intel® Ethernet Controller E810 Dynamic
23  Device Personalization (DDP) for Telecommunications Technology Guide <https://cdrdv2.intel.com/v1/dl/getContent/617015>`_.
24
25- To understand DDP for COMMs usage with DPDK, please review `Intel® Ethernet 800 Series Telecommunication (Comms)
26  Dynamic Device Personalization (DDP) Package <https://cdrdv2.intel.com/v1/dl/getContent/618651>`_.
27
28
29Recommended Matching List
30-------------------------
31
32It is highly recommended to upgrade the ice kernel driver, firmware and DDP package
33to avoid the compatibility issues with ice PMD.
34Here is the suggested matching list which has been tested and verified.
35The detailed information can refer to chapter Tested Platforms/Tested NICs in release notes.
36
37   +-----------+---------------+-----------------+-----------+-----------+
38   |    DPDK   | Kernel Driver | OS Default DDP  | COMMS DDP | Firmware  |
39   +===========+===============+=================+===========+===========+
40   |    20.11  |     1.3.0     |      1.3.20     |  1.3.24   |    2.3    |
41   +-----------+---------------+-----------------+-----------+-----------+
42
43Pre-Installation Configuration
44------------------------------
45
46
47Runtime Config Options
48~~~~~~~~~~~~~~~~~~~~~~
49
50- ``Safe Mode Support`` (default ``0``)
51
52  If driver failed to load OS package, by default driver's initialization failed.
53  But if user intend to use the device without OS package, user can take ``devargs``
54  parameter ``safe-mode-support``, for example::
55
56    -a 80:00.0,safe-mode-support=1
57
58  Then the driver will be initialized successfully and the device will enter Safe Mode.
59  NOTE: In Safe mode, only very limited features are available, features like RSS,
60  checksum, fdir, tunneling ... are all disabled.
61
62- ``Generic Flow Pipeline Mode Support`` (default ``0``)
63
64  In pipeline mode, a flow can be set at one specific stage by setting parameter
65  ``priority``. Currently, we support two stages: priority = 0 or !0. Flows with
66  priority 0 located at the first pipeline stage which typically be used as a firewall
67  to drop the packet on a blocklist(we called it permission stage). At this stage,
68  flow rules are created for the device's exact match engine: switch. Flows with priority
69  !0 located at the second stage, typically packets are classified here and be steered to
70  specific queue or queue group (we called it distribution stage), At this stage, flow
71  rules are created for device's flow director engine.
72  For none-pipeline mode, ``priority`` is ignored, a flow rule can be created as a flow director
73  rule or a switch rule depends on its pattern/action and the resource allocation situation,
74  all flows are virtually at the same pipeline stage.
75  By default, generic flow API is enabled in none-pipeline mode, user can choose to
76  use pipeline mode by setting ``devargs`` parameter ``pipeline-mode-support``,
77  for example::
78
79    -a 80:00.0,pipeline-mode-support=1
80
81- ``Protocol extraction for per queue``
82
83  Configure the RX queues to do protocol extraction into mbuf for protocol
84  handling acceleration, like checking the TCP SYN packets quickly.
85
86  The argument format is::
87
88      -a 18:00.0,proto_xtr=<queues:protocol>[<queues:protocol>...]
89      -a 18:00.0,proto_xtr=<protocol>
90
91  Queues are grouped by ``(`` and ``)`` within the group. The ``-`` character
92  is used as a range separator and ``,`` is used as a single number separator.
93  The grouping ``()`` can be omitted for single element group. If no queues are
94  specified, PMD will use this protocol extraction type for all queues.
95
96  Protocol is : ``vlan, ipv4, ipv6, ipv6_flow, tcp, ip_offset``.
97
98  .. code-block:: console
99
100    dpdk-testpmd -a 18:00.0,proto_xtr='[(1,2-3,8-9):tcp,10-13:vlan]'
101
102  This setting means queues 1, 2-3, 8-9 are TCP extraction, queues 10-13 are
103  VLAN extraction, other queues run with no protocol extraction.
104
105  .. code-block:: console
106
107    dpdk-testpmd -a 18:00.0,proto_xtr=vlan,proto_xtr='[(1,2-3,8-9):tcp,10-23:ipv6]'
108
109  This setting means queues 1, 2-3, 8-9 are TCP extraction, queues 10-23 are
110  IPv6 extraction, other queues use the default VLAN extraction.
111
112  The extraction metadata is copied into the registered dynamic mbuf field, and
113  the related dynamic mbuf flags is set.
114
115  .. table:: Protocol extraction : ``vlan``
116
117   +----------------------------+----------------------------+
118   |           VLAN2            |           VLAN1            |
119   +======+===+=================+======+===+=================+
120   |  PCP | D |       VID       |  PCP | D |       VID       |
121   +------+---+-----------------+------+---+-----------------+
122
123  VLAN1 - single or EVLAN (first for QinQ).
124
125  VLAN2 - C-VLAN (second for QinQ).
126
127  .. table:: Protocol extraction : ``ipv4``
128
129   +----------------------------+----------------------------+
130   |           IPHDR2           |           IPHDR1           |
131   +======+=======+=============+==============+=============+
132   |  Ver |Hdr Len|    ToS      |      TTL     |  Protocol   |
133   +------+-------+-------------+--------------+-------------+
134
135  IPHDR1 - IPv4 header word 4, "TTL" and "Protocol" fields.
136
137  IPHDR2 - IPv4 header word 0, "Ver", "Hdr Len" and "Type of Service" fields.
138
139  .. table:: Protocol extraction : ``ipv6``
140
141   +----------------------------+----------------------------+
142   |           IPHDR2           |           IPHDR1           |
143   +=====+=============+========+=============+==============+
144   | Ver |Traffic class|  Flow  | Next Header |   Hop Limit  |
145   +-----+-------------+--------+-------------+--------------+
146
147  IPHDR1 - IPv6 header word 3, "Next Header" and "Hop Limit" fields.
148
149  IPHDR2 - IPv6 header word 0, "Ver", "Traffic class" and high 4 bits of
150  "Flow Label" fields.
151
152  .. table:: Protocol extraction : ``ipv6_flow``
153
154   +----------------------------+----------------------------+
155   |           IPHDR2           |           IPHDR1           |
156   +=====+=============+========+============================+
157   | Ver |Traffic class|            Flow Label               |
158   +-----+-------------+-------------------------------------+
159
160  IPHDR1 - IPv6 header word 1, 16 low bits of the "Flow Label" field.
161
162  IPHDR2 - IPv6 header word 0, "Ver", "Traffic class" and high 4 bits of
163  "Flow Label" fields.
164
165  .. table:: Protocol extraction : ``tcp``
166
167   +----------------------------+----------------------------+
168   |           TCPHDR2          |           TCPHDR1          |
169   +============================+======+======+==============+
170   |          Reserved          |Offset|  RSV |     Flags    |
171   +----------------------------+------+------+--------------+
172
173  TCPHDR1 - TCP header word 6, "Data Offset" and "Flags" fields.
174
175  TCPHDR2 - Reserved
176
177  .. table:: Protocol extraction : ``ip_offset``
178
179   +----------------------------+----------------------------+
180   |           IPHDR2           |           IPHDR1           |
181   +============================+============================+
182   |       IPv6 HDR Offset      |       IPv4 HDR Offset      |
183   +----------------------------+----------------------------+
184
185  IPHDR1 - Outer/Single IPv4 Header offset.
186
187  IPHDR2 - Outer/Single IPv6 Header offset.
188
189  Use ``rte_net_ice_dynf_proto_xtr_metadata_get`` to access the protocol
190  extraction metadata, and use ``RTE_PKT_RX_DYNF_PROTO_XTR_*`` to get the
191  metadata type of ``struct rte_mbuf::ol_flags``.
192
193  The ``rte_net_ice_dump_proto_xtr_metadata`` routine shows how to
194  access the protocol extraction result in ``struct rte_mbuf``.
195
196Driver compilation and testing
197------------------------------
198
199Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
200for details.
201
202Features
203--------
204
205Vector PMD
206~~~~~~~~~~
207
208Vector PMD for RX and TX path are selected automatically. The paths
209are chosen based on 2 conditions.
210
211- ``CPU``
212  On the X86 platform, the driver checks if the CPU supports AVX2.
213  If it's supported, AVX2 paths will be chosen. If not, SSE is chosen.
214  If the CPU supports AVX512 and EAL argument ``--force-max-simd-bitwidth``
215  is set to 512, AVX512 paths will be chosen.
216
217- ``Offload features``
218  The supported HW offload features are described in the document ice.ini,
219  A value "P" means the offload feature is not supported by vector path.
220  If any not supported features are used, ICE vector PMD is disabled and the
221  normal paths are chosen.
222
223Malicious driver detection (MDD)
224~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
225
226It's not appropriate to send a packet, if this packet's destination MAC address
227is just this port's MAC address. If SW tries to send such packets, HW will
228report a MDD event and drop the packets.
229
230The APPs based on DPDK should avoid providing such packets.
231
232Device Config Function (DCF)
233~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
234
235This section demonstrates ICE DCF PMD, which shares the core module with ICE
236PMD and iAVF PMD.
237
238A DCF (Device Config Function) PMD bounds to the device's trusted VF with ID 0,
239it can act as a sole controlling entity to exercise advance functionality (such
240as switch, ACL) for the rest VFs.
241
242The DCF PMD needs to advertise and acquire DCF capability which allows DCF to
243send AdminQ commands that it would like to execute over to the PF and receive
244responses for the same from PF.
245
246.. _figure_ice_dcf:
247
248.. figure:: img/ice_dcf.*
249
250   DCF Communication flow.
251
252#. Create the VFs::
253
254      echo 4 > /sys/bus/pci/devices/0000\:18\:00.0/sriov_numvfs
255
256#. Enable the VF0 trust on::
257
258      ip link set dev enp24s0f0 vf 0 trust on
259
260#. Bind the VF0,  and run testpmd with 'cap=dcf' devarg::
261
262      dpdk-testpmd -l 22-25 -n 4 -a 18:01.0,cap=dcf -- -i
263
264#. Monitor the VF2 interface network traffic::
265
266      tcpdump -e -nn -i enp24s1f2
267
268#. Create one flow to redirect the traffic to VF2 by DCF::
269
270      flow create 0 priority 0 ingress pattern eth / ipv4 src is 192.168.0.2 \
271      dst is 192.168.0.3 / end actions vf id 2 / end
272
273#. Send the packet, and it should be displayed on tcpdump::
274
275      sendp(Ether(src='3c:fd:fe:aa:bb:78', dst='00:00:00:01:02:03')/IP(src=' \
276      192.168.0.2', dst="192.168.0.3")/TCP(flags='S')/Raw(load='XXXXXXXXXX'), \
277      iface="enp24s0f0", count=10)
278
279Sample Application Notes
280------------------------
281
282Vlan filter
283~~~~~~~~~~~
284
285Vlan filter only works when Promiscuous mode is off.
286
287To start ``testpmd``, and add vlan 10 to port 0:
288
289.. code-block:: console
290
291    ./app/dpdk-testpmd -l 0-15 -n 4 -- -i
292    ...
293
294    testpmd> rx_vlan add 10 0
295
296Limitations or Known issues
297---------------------------
298
299The Intel E810 requires a programmable pipeline package be downloaded
300by the driver to support normal operations. The E810 has a limited
301functionality built in to allow PXE boot and other use cases, but the
302driver must download a package file during the driver initialization
303stage.
304
305The default DDP package file name is ice.pkg. For a specific NIC, the
306DDP package supposed to be loaded can have a filename: ice-xxxxxx.pkg,
307where 'xxxxxx' is the 64-bit PCIe Device Serial Number of the NIC. For
308example, if the NIC's device serial number is 00-CC-BB-FF-FF-AA-05-68,
309the device-specific DDP package filename is ice-00ccbbffffaa0568.pkg
310(in hex and all low case). During initialization, the driver searches
311in the following paths in order: /lib/firmware/updates/intel/ice/ddp
312and /lib/firmware/intel/ice/ddp. The corresponding device-specific DDP
313package will be downloaded first if the file exists. If not, then the
314driver tries to load the default package. The type of loaded package
315is stored in ``ice_adapter->active_pkg_type``.
316
317A symbolic link to the DDP package file is also ok. The same package
318file is used by both the kernel driver and the DPDK PMD.
319