1 .\" Copyright (c) 2001-2003, Intel Corporation
2 .\" All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions are met:
7 .\" 1. Redistributions of source code must retain the above copyright notice,
8 .\" this list of conditions and the following disclaimer.
10 .\" 2. Redistributions in binary form must reproduce the above copyright
11 .\" notice, this list of conditions and the following disclaimer in the
12 .\" documentation and/or other materials provided with the distribution.
14 .\" 3. Neither the name of the Intel Corporation nor the names of its
15 .\" contributors may be used to endorse or promote products derived from
16 .\" this software without specific prior written permission.
18 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
19 .\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
22 .\" LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
23 .\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
24 .\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
25 .\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
26 .\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
27 .\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28 .\" POSSIBILITY OF SUCH DAMAGE.
30 .\" * Other names and brands may be claimed as the property of others.
32 .\" $FreeBSD: src/share/man/man4/igb.4,v 1.2 2010/05/14 20:11:30 marius Exp $
39 .Nd "Intel(R) PRO/1000 PCI Express Gigabit Ethernet adapter driver"
41 To compile this driver into the kernel,
42 place the following line in your
43 kernel configuration file:
44 .Bd -ragged -offset indent
49 Alternatively, to load the driver as a
50 module at boot time, place the following line in
52 .Bd -literal -offset indent
59 driver provides support for PCI Express Gigabit Ethernet adapters based on
67 Ethernet controller chips.
72 .Bl -item -offset indent -compact
74 Transmit/Receive checksum offload for IP/UDP/TCP.
79 TCP segmentation offload (TSO)
81 Receive side scaling (RSS)
83 Multiple tranmission queues
87 VLAN tag stripping and inserting
96 driver will try enabling as many reception queues and transmission queues
97 as are allowed by the number of CPUs in the system.
99 If multiple transmission queues are used,
100 the round-robin arbitration is performed among the transmission queues.
101 It should be noted that
102 if both TSO and multiple transmission queues are used,
103 for 82575, the round-robin arbitration between transmission queues is done
104 at the TSO packet boundary;
105 for the reset of the hardware, the round-robin arbitration
106 between transmission queues is done at the TCP segment boundary after
107 the hardware segmentation is performed.
109 82575 supports 4 reception queues and 4 transmission queues.
110 MSI-X is not enabled due to hardware errata.
111 Under MSI or legacy interrupt mode,
112 2 reception queues are enabled for hardware RSS hash
113 and only 1 transmission queue is enable.
115 82576 supports 16 reception queues and 16 transmission queues.
116 MSI-X is enable by default.
118 due to the number of MSI-X vectors is 10,
119 at most 8 reception queues and 8 transmission queues will be enabled
123 is enabled on the devices,
124 at most 16 reception queues and 16 transmission queues will be enabled.
127 I350 and I354 support 8 reception queues and 8 transmission queues.
128 MSI-X is enabled by default.
130 I210 supports 4 reception queues and 4 transmission queues.
131 MSI-X is enabled by default.
133 I211 supports 2 reception queues and 2 transmission queues.
134 MSI-X is enabled by default.
136 .\"For questions related to hardware requirements, refer to the
137 .\"documentation supplied with your Intel PRO/1000 adapter.
138 .\"All hardware requirements listed apply to use with
141 .\"Support for Jumbo Frames is provided via the interface MTU setting.
142 .\"Selecting an MTU larger than 1500 bytes with the
144 .\"utility configures the adapter to receive and transmit Jumbo Frames.
145 .\"The maximum MTU size for Jumbo Frames is 9216.
149 driver supports the following media types:
150 .Bl -tag -width ".Cm 10baseT/UTP"
152 Enables auto-negotiation for speed and duplex.
154 Sets 10Mbps operation.
161 Sets 100Mbps operation.
168 Sets 1000Mbps operation.
171 mode is supported at this speed.
173 Sets 1000Mbps operation.
176 mode is supported at this speed.
181 driver supports the following media options:
182 .Bl -tag -width ".Cm full-duplex"
184 Forces full-duplex operation
186 Forces half-duplex operation.
195 is not specified, the driver defaults to
198 For more information on configuring this device, see
207 driver supports Gigabit Ethernet adapters based on the Intel
218 Intel 82575EB Gigabit Ethernet Controller
220 Intel 82576 Gigabit Ethernet Controller
222 Intel 82580EB Gigabit Ethernet Controller
224 Intel Ethernet Controller I210 Series
226 Intel Ethernet Controller I211 Series
228 Intel Ethernet Controller I350
230 Intel Ethernet Controller I354
232 Intel Ethernet Server Adapter I210-T1
234 Intel Ethernet Server Adapter I340-F4
236 Intel Ethernet Server Adapter I340-T4
238 Intel Ethernet Server Adapter I350-F2
240 Intel Ethernet Server Adapter I350-F4
242 Intel Ethernet Server Adapter I350-T2
244 Intel Ethernet Server Adapter I350-T4
246 Intel Gigabit EF Dual Port Server Adapter
248 Intel Gigabit ET Dual Port Server Adapter
250 Intel Gigabit ET Quad Port Server Adapter
252 Intel Gigabit ET2 Quad Port Server Adapter
254 Intel Gigabit VT Quad Port Server Adapter
257 Tunables can be set at the
259 prompt before booting the kernel or stored in
262 is the device unit number.
263 .Bl -tag -width ".Va hw.igbX.irq.unshared"
264 .It Va hw.igb.rxd Va hw.igbX.rxd
265 Number of receive descriptors allocated by the driver.
266 The default value is 512.
268 and the maximum is 4096.
269 .It Va hw.igb.txd Va hw.igbX.txd
270 Number of transmit descriptors allocated by the driver.
271 The default value is 1024.
273 and the maximum is 4096.
274 .It Va hw.igb.rxr Va hw.igbX.rxr
275 This tunable specifies the number of reception queues could be enabled.
276 Maximum allowed value for these tunables is device specific
277 and it must be power of 2 aligned.
278 Setting these tunables to 0 allows the driver to make
279 as many reception queues ready-for-use as allowed by the number of CPUs.
280 .It Va hw.igb.txr Va hw.igbX.txr
281 This tunable specifies the number of transmission queues could be enabled.
282 Maximum allowed value for these tunables is device specific
283 and it must be power of 2 aligned.
284 Setting these tunables to 0 allows the driver to make
285 as many transmission queues ready-for-use as allowed by the number of CPUs.
286 .It Va hw.igb.msix.enable Va hw.igbX.msix.enable
288 the driver will use MSI-X if it is supported.
289 This behaviour can be turned off by setting this tunable to 0.
290 .It Va hw.igbX.msix.off
292 and the number of MSI-X vectors is not enough to
293 put transmission queue processing and reception queue processing
294 onto independent MSI-X vector,
295 this tunable specifies the leading target CPU for
296 transmission and reception queues processing.
297 The value specificed must be aligned to the maximum of
298 the number of reception queues
299 and the number of transmission queues enabled,
300 and must be less than the power of 2 number of CPUs.
301 .It Va hw.igbX.msix.rxoff
303 and the number of MSI-X vectors is enough to
304 put transmission queue processing and reception queue processing
305 onto independent MSI-X vector,
306 this tunable specifies the leading target CPU for reception queues processing.
307 The value specificed must be aligned to the number of reception queues enabled
308 and must be less than the power of 2 number of CPUs.
309 .It Va hw.igbX.msix.txoff
311 and the number of MSI-X vectors is enough to
312 put transmission queue processing and reception queue processing
313 onto independent MSI-X vector,
314 this tunable specifies the leading target CPU
315 for transmission queues processing.
316 The value specificed must be aligned to
317 the number of transmission queues enabled
318 and must be less than the power of 2 number of CPUs.
319 .It Va hw.igb.msi.enable Va hw.igbX.msi.enable
320 If MSI-X is disabled and MSI is supported,
321 the driver will use MSI.
322 This behavior can be turned off by setting this tunable to 0.
323 .It Va hw.igbX.msi.cpu
325 it specifies the MSI's target CPU.
326 .It Va hw.igbX.irq.unshared
327 If legacy interrupt is used,
329 the driver assumes the interrupt could be shared.
330 Setting this tunable to 1 allows the driver to perform certain
331 optimization based on the knowledge that the interrupt is not shared.
332 .It Va hw.igbX.npoll.txoff
333 This tunable specifies the leading target CPU for
337 The value specificed must be aligned to the number of transmission queues
338 enabled and must be less than the power of 2 number of CPUs.
339 .It Va hw.igbX.npoll.rxoff
340 This tunable specifies the leading target CPU for
344 The value specificed must be aligned to the number of reception queues
345 enabled and must be less than the power of 2 number of CPUs.
348 A number of per-interface variables are implemented in the
353 .Bl -tag -width "tx_intr_nsegs"
355 Number of reception queues could be enabled (read-only).
362 Number of reception queues being used (read-only).
364 Number of transmission queues could be enabled (read-only).
371 Number of transmission queues being used (read-only).
373 Number of descriptors per reception queue (read-only).
380 Number of descriptors per transmission queue (read-only).
387 If MSI or legacy interrupt is used,
388 this sysctl controls the highest possible frequency
389 that interrupt could be generated by the device.
390 It is 6000 by default (~150us).
393 this sysctl controls the highest possible frequency
394 that interrupt could be generated by the
397 For reception processing only MSI-X vector,
398 the default value is 6000 (~150us).
399 For transmission processing only MSI-X vector,
400 the default value is 4000 (250us).
401 For MSI-X vector which handles both reception and transmission,
402 the default value is 6000 (~150us).
404 Transmission interrupt is asked to be generated upon every
406 transmission descritors having been setup.
407 The default value is 1/16 of the number of transmission descriptors per queue.
409 The number of transmission descriptors should be setup
410 before the hardware register is written.
411 Setting this value too high will have negative effect
412 on transmission timeliness.
413 Setting this value too low will hurt overall transmission performance
414 due to the frequent hardware register writing.
415 The default value is 8.
417 The number of reception descriptors should be setup
418 before the hardware register is written.
419 Setting this value too high will make device drop incoming packets.
420 Setting this value too low will hurt overall reception performance
421 due to the frequent hardware register writing.
422 The default value is 32.
425 .Va hw.igbX.npoll.rxoff .
426 The set value will take effect the next time
428 is enabled on the device.
431 .Va hw.igbX.npoll.txoff .
432 The set value will take effect the next time
434 is enabled on the device.
437 .\"For general information and support,
438 .\"go to the Intel support website at:
439 .\".Pa http://support.intel.com .
441 .\"If an issue is identified with the released source code on the supported kernel
442 .\"with a supported adapter, email the specific information related to the
444 .\".Aq Mt freebsdnic@mailbox.intel.com .
457 device driver first appeared in
462 driver was written by
463 .An Intel Corporation Aq Mt freebsdnic@mailbox.intel.com .