xref: /illumos-gate/usr/src/man/man4d/i40e.4d (revision 37e2cd25d56b334a2403f2540a0b0a1e6a40bcd1)
1.\"
2.\" This file and its contents are supplied under the terms of the
3.\" Common Development and Distribution License ("CDDL"), version 1.0.
4.\" You may only use this file in accordance with the terms of version
5.\" 1.0 of the CDDL.
6.\"
7.\" A full copy of the text of the CDDL should have accompanied this
8.\" source.  A copy of the CDDL is also available via the Internet at
9.\" http://www.illumos.org/license/CDDL.
10.\"
11.\"
12.\" Copyright (c) 2018 Joyent, Inc.
13.\" Copyright 2020 Ryan Zezeski
14.\" Copyright 2021 Oxide Computer Company
15.\"
16.Dd January 18, 2021
17.Dt I40E 4D
18.Os
19.Sh NAME
20.Nm i40e
21.Nd Intel 710/722 Ethernet Device Driver
22.Sh SYNOPSIS
23.Pa /dev/net/i40e*
24.Sh DESCRIPTION
25The
26.Nm
27driver is a GLDv3, multi-threaded, clonable, loadable device driver that
28supports the Data Link Provider Interface,
29.Xr dlpi 4P .
30The
31.Nm
32driver supports the Intel 710 and 722 Ethernet Controller families of
33networking interface cards which support speeds of 1 GbE, 2.5 GbE, 5 GbE, 10
34GbE, 25 GbE, and 40 GbE.
35.Pp
36In addition to basic device initialization and the sending and receiving
37of frames, it supports the following features:
38.Bl -dash -offset indent
39.It
40Jumbo frames up to 9710 bytes.
41.It
42Promiscuous access via
43.Xr snoop 8 and
44.Xr dlpi 4P
45.It
46IPv4 Checksum Offload
47.It
48TCP, UDP, and SCTP checksum offload
49.It
50IPv4 and IPv6 TCP Segmentation offload
51.El
52.Pp
53At this time, the
54.Nm
55driver does not enable the use of energy efficient Ethernet (EEE) or
56support the use of flow control through hardware pause frames.
57.Sh APPLICATION PROGRAMMING INTERFACE
58For each device supported by the
59.Nm
60installed in the system, a character-special file will be created.
61This file supports the Data Link Provider Interface (DLPI) which is documented
62in
63.Xr dlpi 4P .
64For most consumers, the use of
65.Xr libdlpi 3LIB ,
66is recommended.
67.Pp
68Each instance is assigned a unique ascending integer identifier.
69A device which has multiple ports may appear to the system as separate
70instances.
71The system does not provide a guarantee on how these will be presented.
72Using this instance identifier, one can determine the exact character-special
73file to open.
74For example, the first instance enumerated in the system, with id 0, would be
75named
76.Sy i40e0 .
77It exists in the file system at
78.Pa /dev/net/i40e0 .
79.Sh CONFIGURATION
80The
81.Nm i40e
82driver always performs auto-negotiation and depending on the model may
83negotiate to 40 Gbps, 25 Gbps, 10 Gbps, or 1 Gbps.
84At this time, the driver requires the use of auto-negotiation.
85.Pp
86The
87.Nm
88driver is managed by the
89.Xr dladm 8
90utility.
91.Xr dladm 8
92is the preferred interface for setting all properties.
93While
94.Xr driver.conf 5
95based configuration is possible,
96.Xr dladm 8
97is recommended.
98The
99.Nm
100driver may be joined into an aggregation based on the link aggregation
101control protocol (LACP) through
102.Xr dladm 8 .
103.Sh PROPERTIES
104The device supports the following properties which may be tuned through
105its driver.conf file,
106.Pa /kernel/drv/i40e.conf .
107Most of these properties cannot be changed after the device has been started.
108The device is started in response to a DLPI consumer opening the device and
109binding to it.
110This happens when an IP interfaces is plumbed or another
111.Xr dlpi 4P
112consumer such as
113.Xr snoop 8
114or an LLDP daemon is started.
115.Pp
116Some properties may be tuned at runtime with the
117.Xr dladm 8
118utility.
119Properties that can be will have the name of the dladm property called out
120explicitly.
121.Pp
122These properties are not considered stable at this time.
123They may change and should not be relied on.
124They are considered
125.Sy Volatile .
126It is not expected that administrators of the system will have to tune
127these values.
128.Bl -hang -width Ds
129.It Sy default_mtu
130.Bd -filled -compact
131Minimum:
132.Sy 1500 |
133Maximum:
134.Sy 9710 |
135Runtime Property:
136.Sy mtu
137.Ed
138.Bd -filled
139The
140.Sy default_mtu
141property determines the starting MTU of the various device instances.
142Note that the device's MTU also determines the upper bound of the MTU of
143all VNICs created over the device.
144The default MTU is
145.Sy 1500 .
146.Ed
147.It Sy mr_enable
148.Bd -filled -compact
149Minimum:
150.Sy 0 |
151Maximum:
152.Sy 1
153.Ed
154.Bd -filled
155The
156.Sy mr_enable
157property determines whether or not support for multiple rings is enabled
158for the device.
159The default is always to enable them.
160It is not recommended to to disable them.
161.Ed
162.It Sy rx_num_groups
163.Bd -filled -compact
164Minimum:
165.Sy 1 |
166Maximum:
167.Sy 32
168.Ed
169.Bd -filled
170The
171.Sy rx_num_groups
172property determines the number of receive mac groups provided by the driver.
173Each group can handle all unicast traffic for a single MAC address, more groups
174means more unicast traffic that can be steered by hardware.
175However, more groups also means more demand for kernel memory.
176If you are not making heavy use of VNICs, or do not need the efficiency gains
177of hardware steering, then reducing this number can reduce kernel memory
178taken by
179.Nm i40e.
180.Ed
181.It Sy rx_ring_size
182.Bd -filled -compact
183Minimum:
184.Sy 64 |
185Maximum:
186.Sy 4096
187.Ed
188.Bd -filled
189The
190.Sy rx_ring_size
191property determines the number of descriptors that will be used in each
192receive ring on the card.
193Administrators should not normally need to tune this value.
194Hardware requires that the ring size be a multiple of 32.
195The system will round up the set value to the nearest multiple of 32.
196.Ed
197.It Sy tx_ring_size
198.Bd -filled -compact
199Minimum:
200.Sy 64 |
201Maximum:
202.Sy 4096
203.Ed
204.Bd -filled
205The
206.Sy tx_ring_size
207property determines the number of descriptors that will be used in each
208transmit ring on the card.
209Administrators should not normally need to tune this value.
210Hardware requires that the ring size be a multiple of 32.
211The system will round up the set value to the nearest multiple of 32.
212.Ed
213.It Sy tx_resched_threshold
214.Bd -filled -compact
215Minimum:
216.Sy 8 |
217Maximum:
218.Sy Variable
219.Ed
220.Bd -filled
221The
222.Sy tx_resched_threshold
223property determines the number of descriptors that must be available for
224a frame to be transmitted.
225The maximum is variable.
226It is dependent on the value of the
227.Sy tx_ring_size
228property.
229At least eight descriptors must be available for the device to function
230correctly.
231.Ed
232.It Sy rx_limit_per_intr
233.Bd -filled -compact
234Minimum:
235.Sy 16 |
236Maximum:
237.Sy 4096
238.Ed
239.Bd -filled
240The
241.Sy rx_limit_per_intr
242property determines the maximum number of packets that will be processed
243on a given ring during a single interrupt.
244This is done to try and guarantee some amount of liveness in the system.
245It is not expected that administrators will have to tune this value.
246.Ed
247.It Sy tx_hcksum_enable
248.Bd -filled -compact
249Minimum:
250.Sy 0 |
251Maximum:
252.Sy 1
253.Ed
254.Bd -filled
255The
256.Sy tx_hcksum_enable
257property controls whether or not the device enables support for hardware
258checksumming of outgoing packets.
259The default is to always enable support for this.
260Turning it off will increase latency and decrease throughput when transmitting
261packets, but should be done if a hardware bug is suspected.
262.Ed
263.It Sy rx_hcksum_enable
264.Bd -filled -compact
265Minimum:
266.Sy 0 |
267Maximum:
268.Sy 1
269.Ed
270.Bd -filled
271The
272.Sy rx_hcksum_enable
273property controls whether or not the device enables support for hardware
274checksumming of incoming packets.
275The default is to always enable support for this.
276Turning it off will increase latency and decrease throughput when receiving
277packets, but should be done if a hardware bug is suspected.
278.Ed
279.It Sy rx_dma_threshold
280.Bd -filled -compact
281Minimum:
282.Sy 0 |
283Maximum:
284.Sy INT32_MAX |
285Runtime Property:
286.Sy _rx_dma_threshold
287.Ed
288.Bd -filled
289The
290.Sy rx_dma_threshold
291indicates the size in bytes of a received frame, including all of its
292headers, at which the driver should not copy the frame but instead bind
293DMA memory.
294By setting this property to its minimum, all frames will be processed with DMA
295binding.
296By setting this property to its maximum, all frames will be processed by copying
297the frame.
298.Ed
299.It Sy tx_lso_enable
300.Bd -filled -compact
301Minimum:
302.Sy 0 |
303Maximum:
304.Sy 1
305.Ed
306.Bd -filled
307The
308.Sy tx_lso_enable
309property controls whether or not the device enables support for Large Segment
310Offloand (LSO) when transmitting packets.
311The default is to always enable support for this.
312Turning it off will decrease throughput when transmitting packets, but should
313be done if a hardware bug is suspected.
314.Ed
315.El
316.Sh ARCHITECTURE
317The
318.Nm
319driver is only supported on
320.Sy x86
321systems at this time.
322.Sh FILES
323.Bl -tag -width Pa
324.It Pa /dev/net/i40e*
325Per-instance character device.
326.It Pa /kernel/drv/amd64/i40e
327Device driver (x86)
328.It Pa /kernel/drv/i40e.conf
329Driver configuration file
330.El
331.Sh SEE ALSO
332.Xr dlpi 4P ,
333.Xr driver.conf 5 ,
334.Xr dladm 8 ,
335.Xr snoop 8
336