1 Linux* Base Driver for the Intel(R) PRO/1000 Family of Adapters
2 ===============================================================
11 - Identifying Your Adapter
12 - Command Line Parameters
13 - Speed and Duplex Configuration
14 - Additional Configurations
22 This file describes the Linux* Base Driver for the Intel(R) PRO/1000 Family
23 of Adapters, version 5.x.x. This driver includes support for Itanium(TM)2
26 For questions related to hardware requirements, refer to the documentation
27 supplied with your Intel PRO/1000 adapter. All hardware requirements listed
28 apply to use with Linux.
30 Native VLANs are now available with supported kernels.
32 Identifying Your Adapter
33 ========================
35 For more information on how to identify your adapter, go to the Adapter &
38 http://support.intel.com/support/network/adapter/pro100/21397.htm
40 For the latest Intel network drivers for Linux, refer to the following
41 website. In the search field, enter your adapter name or type, or use the
42 networking link on the left to search for your adapter:
44 http://downloadfinder.intel.com/scripts-df/support_intel.asp
46 Command Line Parameters
47 =======================
49 If the driver is built as a module, the following optional parameters are
50 used by entering them on the command line with the modprobe or insmod command
53 modprobe e1000 [<option>=<VAL1>,<VAL2>,...]
55 insmod e1000 [<option>=<VAL1>,<VAL2>,...]
57 For example, with two PRO/1000 PCI adapters, entering:
59 insmod e1000 TxDescriptors=80,128
61 loads the e1000 driver with 80 TX descriptors for the first adapter and 128 TX
62 descriptors for the second adapter.
64 The default value for each parameter is generally the recommended setting,
65 unless otherwise noted. Also, if the driver is statically built into the
66 kernel, the driver is loaded with the default values for all the parameters.
67 Ethtool can be used to change some of the parameters at runtime.
69 NOTES: For more information about the AutoNeg, Duplex, and Speed
70 parameters, see the "Speed and Duplex Configuration" section in
73 For more information about the InterruptThrottleRate, RxIntDelay,
74 TxIntDelay, RxAbsIntDelay, and TxAbsIntDelay parameters, see the
76 http://www.intel.com/design/network/applnots/ap450.htm
78 A descriptor describes a data buffer and attributes related to the
79 data buffer. This information is accessed by the hardware.
81 AutoNeg (adapters using copper connections only)
82 Valid Range: 0x01-0x0F, 0x20-0x2F
84 This parameter is a bit mask that specifies which speed and duplex
85 settings the board advertises. When this parameter is used, the Speed and
86 Duplex parameters must not be specified.
87 NOTE: Refer to the Speed and Duplex section of this readme for more
88 information on the AutoNeg parameter.
90 Duplex (adapters using copper connections only)
91 Valid Range: 0-2 (0=auto-negotiate, 1=half, 2=full)
93 Defines the direction in which data is allowed to flow. Can be either one
94 or two-directional. If both Duplex and the link partner are set to auto-
95 negotiate, the board auto-detects the correct duplex. If the link partner
96 is forced (either full or half), Duplex defaults to half-duplex.
99 Valid Range: 0-3 (0=none, 1=Rx only, 2=Tx only, 3=Rx&Tx)
100 Default: Read flow control settings from the EEPROM
101 This parameter controls the automatic generation(Tx) and response(Rx) to
102 Ethernet PAUSE frames.
104 InterruptThrottleRate
105 Valid Range: 100-100000 (0=off, 1=dynamic)
107 This value represents the maximum number of interrupts per second the
108 controller generates. InterruptThrottleRate is another setting used in
109 interrupt moderation. Dynamic mode uses a heuristic algorithm to adjust
110 InterruptThrottleRate based on the current traffic load.
111 Un-supported Adapters: InterruptThrottleRate is NOT supported by 82542, 82543
112 or 82544-based adapters.
114 NOTE: InterruptThrottleRate takes precedence over the TxAbsIntDelay and
115 RxAbsIntDelay parameters. In other words, minimizing the receive
116 and/or transmit absolute delays does not force the controller to
117 generate more interrupts than what the Interrupt Throttle Rate
119 CAUTION: If you are using the Intel PRO/1000 CT Network Connection
120 (controller 82547), setting InterruptThrottleRate to a value
121 greater than 75,000, may hang (stop transmitting) adapters under
122 certain network conditions. If this occurs a NETDEV WATCHDOG
123 message is logged in the system event log. In addition, the
124 controller is automatically reset, restoring the network
125 connection. To eliminate the potential for the hang, ensure
126 that InterruptThrottleRate is set no greater than 75,000 and is
128 NOTE: When e1000 is loaded with default settings and multiple adapters are
129 in use simultaneously, the CPU utilization may increase non-linearly.
130 In order to limit the CPU utilization without impacting the overall
131 throughput, we recommend that you load the driver as follows:
133 insmod e1000.o InterruptThrottleRate=3000,3000,3000
135 This sets the InterruptThrottleRate to 3000 interrupts/sec for the
136 first, second, and third instances of the driver. The range of 2000 to
137 3000 interrupts per second works on a majority of systems and is a
138 good starting point, but the optimal value will be platform-specific.
139 If CPU utilization is not a concern, use RX_POLLING (NAPI) and default
143 Valid Range: 80-256 for 82542 and 82543-based adapters
144 80-4096 for all other supported adapters
146 This value is the number of receive descriptors allocated by the driver.
147 Increasing this value allows the driver to buffer more incoming packets.
148 Each descriptor is 16 bytes. A receive buffer is also allocated for each
149 descriptor and can be either 2048, 4096, 8192, or 16384 bytes, depending
150 on the MTU setting. The maximum MTU size is 16110.
152 NOTE: MTU designates the frame size. It only needs to be set for Jumbo
154 NOTE: Depending on the available system resources, the request for a
155 higher number of receive descriptors may be denied. In this case,
159 Valid Range: 0-65535 (0=off)
161 This value delays the generation of receive interrupts in units of 1.024
162 microseconds. Receive interrupt reduction can improve CPU efficiency if
163 properly tuned for specific network traffic. Increasing this value adds
164 extra latency to frame reception and can end up decreasing the throughput
165 of TCP traffic. If the system is reporting dropped receives, this value
166 may be set too high, causing the driver to run out of available receive
169 CAUTION: When setting RxIntDelay to a value other than 0, adapters may
170 hang (stop transmitting) under certain network conditions. If
171 this occurs a NETDEV WATCHDOG message is logged in the system
172 event log. In addition, the controller is automatically reset,
173 restoring the network connection. To eliminate the potential for
174 the hang ensure that RxIntDelay is set to 0.
176 RxAbsIntDelay (82540, 82545 and later adapters only)
177 Valid Range: 0-65535 (0=off)
179 This value, in units of 1.024 microseconds, limits the delay in which a
180 receive interrupt is generated. Useful only if RxIntDelay is non-zero,
181 this value ensures that an interrupt is generated after the initial
182 packet is received within the set amount of time. Proper tuning,
183 along with RxIntDelay, may improve traffic throughput in specific network
186 Speed (adapters using copper connections only)
187 Valid Settings: 0, 10, 100, 1000
188 Default Value: 0 (auto-negotiate at all supported speeds)
189 Speed forces the line speed to the specified value in megabits per second
190 (Mbps). If this parameter is not specified or is set to 0 and the link
191 partner is set to auto-negotiate, the board will auto-detect the correct
192 speed. Duplex should also be set when Speed is set to either 10 or 100.
195 Valid Range: 80-256 for 82542 and 82543-based adapters
196 80-4096 for all other supported adapters
198 This value is the number of transmit descriptors allocated by the driver.
199 Increasing this value allows the driver to queue more transmits. Each
200 descriptor is 16 bytes.
202 NOTE: Depending on the available system resources, the request for a
203 higher number of transmit descriptors may be denied. In this case,
207 Valid Range: 0-65535 (0=off)
209 This value delays the generation of transmit interrupts in units of
210 1.024 microseconds. Transmit interrupt reduction can improve CPU
211 efficiency if properly tuned for specific network traffic. If the
212 system is reporting dropped transmits, this value may be set too high
213 causing the driver to run out of available transmit descriptors.
215 TxAbsIntDelay (82540, 82545 and later adapters only)
216 Valid Range: 0-65535 (0=off)
218 This value, in units of 1.024 microseconds, limits the delay in which a
219 transmit interrupt is generated. Useful only if TxIntDelay is non-zero,
220 this value ensures that an interrupt is generated after the initial
221 packet is sent on the wire within the set amount of time. Proper tuning,
222 along with TxIntDelay, may improve traffic throughput in specific
225 XsumRX (not available on the 82542-based adapter)
228 A value of '1' indicates that the driver should enable IP checksum
229 offload for received packets (both UDP and TCP) to the adapter hardware.
231 Speed and Duplex Configuration
232 ==============================
234 Three keywords are used to control the speed and duplex configuration. These
235 keywords are Speed, Duplex, and AutoNeg.
237 If the board uses a fiber interface, these keywords are ignored, and the
238 fiber interface board only links at 1000 Mbps full-duplex.
240 For copper-based boards, the keywords interact as follows:
242 The default operation is auto-negotiate. The board advertises all supported
243 speed and duplex combinations, and it links at the highest common speed and
244 duplex mode IF the link partner is set to auto-negotiate.
246 If Speed = 1000, limited auto-negotiation is enabled and only 1000 Mbps is
247 advertised (The 1000BaseT spec requires auto-negotiation.)
249 If Speed = 10 or 100, then both Speed and Duplex should be set. Auto-
250 negotiation is disabled, and the AutoNeg parameter is ignored. Partner SHOULD
253 The AutoNeg parameter is used when more control is required over the auto-
254 negotiation process. When this parameter is used, Speed and Duplex must not
255 be specified. This parameter is a bitmap that specifies which speed and
256 duplex settings are advertised to the link partner.
259 Speed (Mbps) N/A N/A 1000 N/A 100 100 10 10
260 Duplex Full Full Half Full Half
262 For example to limit the negotiated speed/duplex on the interface to 10 Mbps
263 Half or Full duplex, set AutoNeg to 0x02:
264 insmod e1000 AutoNeg=0x02
266 Note that setting AutoNeg does not guarantee that the board will link at the
267 highest specified speed or duplex mode, but the board will link at the
268 highest possible speed/duplex of the link partner IF the link partner is also
269 set to auto-negotiate. If the link partner is forced speed/duplex, the
270 adapter MUST be forced to the same speed/duplex.
273 Additional Configurations
274 =========================
276 Configuring the Driver on Different Distributions
277 -------------------------------------------------
279 Configuring a network driver to load properly when the system is started is
280 distribution dependent. Typically, the configuration process involves adding
281 an alias line to /etc/modules.conf as well as editing other system startup
282 scripts and/or configuration files. Many popular Linux distributions ship
283 with tools to make these changes for you. To learn the proper way to
284 configure a network device for your system, refer to your distribution
285 documentation. If during this process you are asked for the driver or module
286 name, the name for the Linux Base Driver for the Intel PRO/1000 Family of
289 As an example, if you install the e1000 driver for two PRO/1000 adapters
290 (eth0 and eth1) and set the speed and duplex to 10full and 100half, add the
291 following to modules.conf:
295 options e1000 Speed=10,100 Duplex=2,1
297 Viewing Link Messages
298 ---------------------
300 Link messages will not be displayed to the console if the distribution is
301 restricting system messages. In order to see network driver link messages on
302 your console, set dmesg to eight by entering the following:
306 NOTE: This setting is not saved across reboots.
311 The driver supports Jumbo Frames for all adapters except 82542-based
312 adapters. Jumbo Frames support is enabled by changing the MTU to a value
313 larger than the default of 1500. Use the ifconfig command to increase the
314 MTU size. For example:
316 ifconfig ethx mtu 9000 up
318 The maximum MTU setting for Jumbo Frames is 16110. This value coincides
319 with the maximum Jumbo Frames size of 16128.
321 NOTE: Jumbo Frames are supported at 1000 Mbps only. Using Jumbo Frames at
322 10 or 100 Mbps may result in poor performance or loss of link.
325 NOTE: MTU designates the frame size. To enable Jumbo Frames, increase the
326 MTU size on the interface beyond 1500.
331 The driver utilizes the ethtool interface for driver configuration and
332 diagnostics, as well as displaying statistical information. Ethtool
333 version 1.6 or later is required for this functionality.
335 The latest release of ethtool can be found from
336 http://sf.net/projects/gkernel. After ethtool is installed,
337 ethtool-copy.h must be copied and renamed to ethtool.h in your kernel
338 source tree at <linux_kernel_src>/include/linux. Backup the original
339 ethtool.h as needed before copying. The driver then must be recompiled
340 in order to take advantage of the latest ethtool features.
342 NOTE: Ethtool 1.6 only supports a limited set of ethtool options. Support
343 for a more complete ethtool feature set can be enabled by upgrading
344 ethtool to ethtool-1.8.1.
346 Enabling Wake on LAN* (WoL)
347 ---------------------------
349 WoL is configured through the Ethtool* utility. Ethtool is included with
350 all versions of Red Hat after Red Hat 7.2. For other Linux distributions,
351 download and install Ethtool from the following website:
352 http://sourceforge.net/projects/gkernel.
354 For instructions on enabling WoL with Ethtool, refer to the website listed
357 WoL will be enabled on the system during the next shut down or reboot.
358 For this driver version, in order to enable WoL, the e1000 driver must be
359 loaded when shutting down or rebooting the system.
364 NAPI (Rx polling mode) is supported in the e1000 driver. NAPI is enabled
365 or disabled based on the configuration of the kernel.
367 See www.cyberus.ca/~hadi/usenix-paper.tgz for more information on NAPI.
373 Jumbo Frames System Requirement
374 -------------------------------
376 Memory allocation failures have been observed on Linux systems with 64 MB
377 of RAM or less that are running Jumbo Frames. If you are using Jumbo Frames,
378 your system may require more than the advertised minimum requirement of 64 MB
385 For general information, go to the Intel support website at:
387 http://support.intel.com
389 If an issue is identified with the released source code on the supported
390 kernel with a supported adapter, email the specific information related to
391 the issue to linux.nics@intel.com.
397 This software program is released under the terms of a license agreement
398 between you ('Licensee') and Intel. Do not use or load this software or any
399 associated materials (collectively, the 'Software') until you have carefully
400 read the full terms and conditions of the LICENSE located in this software
401 package. By loading or using the Software, you agree to the terms of this
402 Agreement. If you do not agree with the terms of this Agreement, do not
403 install or use the Software.
405 * Other names and brands may be claimed as the property of others.