1 .. SPDX-License-Identifier: GPL-2.0+
3 Linux* Driver for Intel(R) Ethernet Network Connection
4 ======================================================
6 Intel Gigabit Linux driver.
7 Copyright(c) 2008-2018 Intel Corporation.
12 - Identifying Your Adapter
13 - Command Line Parameters
14 - Additional Configurations
18 Identifying Your Adapter
19 ========================
20 For information on how to identify your adapter, and for the latest Intel
21 network drivers, refer to the Intel Support website:
22 https://www.intel.com/support
25 Command Line Parameters
26 =======================
27 If the driver is built as a module, the following optional parameters are used
28 by entering them on the command line with the modprobe command using this
31 modprobe e1000e [<option>=<VAL1>,<VAL2>,...]
33 There needs to be a <VAL#> for each network port in the system supported by
34 this driver. The values will be applied to each instance, in function order.
37 modprobe e1000e InterruptThrottleRate=16000,16000
39 In this case, there are two network ports supported by e1000e in the system.
40 The default value for each parameter is generally the recommended setting,
41 unless otherwise noted.
43 NOTE: A descriptor describes a data buffer and attributes related to the data
44 buffer. This information is accessed by the hardware.
48 :Valid Range: 0,1,3,4,100-100000
51 Interrupt Throttle Rate controls the number of interrupts each interrupt
52 vector can generate per second. Increasing ITR lowers latency at the cost of
53 increased CPU utilization, though it may help throughput in some circumstances.
55 Setting InterruptThrottleRate to a value greater or equal to 100
56 will program the adapter to send out a maximum of that many interrupts
57 per second, even if more packets have come in. This reduces interrupt
58 load on the system and can lower CPU utilization under heavy load,
59 but will increase latency as packets are not processed as quickly.
61 The default behaviour of the driver previously assumed a static
62 InterruptThrottleRate value of 8000, providing a good fallback value for
63 all traffic types, but lacking in small packet performance and latency.
64 The hardware can handle many more small packets per second however, and
65 for this reason an adaptive interrupt moderation algorithm was implemented.
67 The driver has two adaptive modes (setting 1 or 3) in which
68 it dynamically adjusts the InterruptThrottleRate value based on the traffic
69 that it receives. After determining the type of incoming traffic in the last
70 timeframe, it will adjust the InterruptThrottleRate to an appropriate value
73 The algorithm classifies the incoming traffic every interval into
74 classes. Once the class is determined, the InterruptThrottleRate value is
75 adjusted to suit that traffic type the best. There are three classes defined:
76 "Bulk traffic", for large amounts of packets of normal size; "Low latency",
77 for small amounts of traffic and/or a significant percentage of small
78 packets; and "Lowest latency", for almost completely small packets or
82 Turns off any interrupt moderation and may improve small packet latency.
83 However, this is generally not suitable for bulk throughput traffic due
84 to the increased CPU utilization of the higher interrupt rate.
86 This mode attempts to moderate interrupts per vector while maintaining
87 very low latency. This can sometimes cause extra CPU utilization. If
88 planning on deploying e1000e in a latency sensitive environment, this
89 parameter should be considered.
90 - 3: Dynamic Conservative mode (default)
91 In dynamic conservative mode, the InterruptThrottleRate value is set to
92 4000 for traffic that falls in class "Bulk traffic". If traffic falls in
93 the "Low latency" or "Lowest latency" class, the InterruptThrottleRate is
94 increased stepwise to 20000. This default mode is suitable for most
96 - 4: Simplified Balancing mode
97 In simplified mode the interrupt rate is based on the ratio of TX and
98 RX traffic. If the bytes per second rate is approximately equal, the
99 interrupt rate will drop as low as 2000 interrupts per second. If the
100 traffic is mostly transmit or mostly receive, the interrupt rate could
103 Setting InterruptThrottleRate to a value greater or equal to 100
104 will program the adapter to send at most that many interrupts per second,
105 even if more packets have come in. This reduces interrupt load on the
106 system and can lower CPU utilization under heavy load, but will increase
107 latency as packets are not processed as quickly.
109 NOTE: InterruptThrottleRate takes precedence over the TxAbsIntDelay and
110 RxAbsIntDelay parameters. In other words, minimizing the receive and/or
111 transmit absolute delays does not force the controller to generate more
112 interrupts than what the Interrupt Throttle Rate allows.
116 :Valid Range: 0-65535 (0=off)
119 This value delays the generation of receive interrupts in units of 1.024
120 microseconds. Receive interrupt reduction can improve CPU efficiency if
121 properly tuned for specific network traffic. Increasing this value adds extra
122 latency to frame reception and can end up decreasing the throughput of TCP
123 traffic. If the system is reporting dropped receives, this value may be set
124 too high, causing the driver to run out of available receive descriptors.
126 CAUTION: When setting RxIntDelay to a value other than 0, adapters may hang
127 (stop transmitting) under certain network conditions. If this occurs a NETDEV
128 WATCHDOG message is logged in the system event log. In addition, the
129 controller is automatically reset, restoring the network connection. To
130 eliminate the potential for the hang ensure that RxIntDelay is set to 0.
134 :Valid Range: 0-65535 (0=off)
137 This value, in units of 1.024 microseconds, limits the delay in which a
138 receive interrupt is generated. This value ensures that an interrupt is
139 generated after the initial packet is received within the set amount of time,
140 which is useful only if RxIntDelay is non-zero. Proper tuning, along with
141 RxIntDelay, may improve traffic throughput in specific network conditions.
145 :Valid Range: 0-65535 (0=off)
148 This value delays the generation of transmit interrupts in units of 1.024
149 microseconds. Transmit interrupt reduction can improve CPU efficiency if
150 properly tuned for specific network traffic. If the system is reporting
151 dropped transmits, this value may be set too high causing the driver to run
152 out of available transmit descriptors.
156 :Valid Range: 0-65535 (0=off)
159 This value, in units of 1.024 microseconds, limits the delay in which a
160 transmit interrupt is generated. It is useful only if TxIntDelay is non-zero.
161 It ensures that an interrupt is generated after the initial Packet is sent on
162 the wire within the set amount of time. Proper tuning, along with TxIntDelay,
163 may improve traffic throughput in specific network conditions.
167 :Valid Range: 0-xxxxxxx (0=off)
170 The driver copies all packets below or equaling this size to a fresh receive
171 buffer before handing it up the stack.
172 This parameter differs from other parameters because it is a single (not 1,1,1
173 etc.) parameter applied to all driver instances and it is also available
174 during runtime at /sys/module/e1000e/parameters/copybreak.
176 To use copybreak, type::
178 modprobe e1000e.ko copybreak=128
183 :Default Value: 0 (disabled)
185 Allows the PHY to turn off in lower power states. The user can turn off this
186 parameter in supported chipsets.
191 :Default Value: 1 (enabled)
193 This workaround skips resetting the PHY at shutdown for the initial silicon
194 releases of ICH8 systems.
201 +-------+----------------+
202 | Value | Interrupt Mode |
203 +=======+================+
205 +-------+----------------+
207 +-------+----------------+
209 +-------+----------------+
211 IntMode allows load time control over the type of interrupt registered for by
212 the driver. MSI-X is required for multiple queue support, and some kernels and
213 combinations of kernel .config options will force a lower level of interrupt
216 This command will show different values for each type of interrupt::
223 :Default Value: 1 (enabled)
225 Strip the CRC from received packets before sending up the network stack. If
226 you have a machine with a BMC enabled but cannot receive IPMI traffic after
227 loading or enabling the driver, try disabling this feature.
232 :Default Value: 1 (enabled)
234 If set to 1, configure the hardware to ignore all write/erase cycles to the
235 GbE region in the ICHx NVM (in order to prevent accidental corruption of the
236 NVM). This feature can be disabled by setting the parameter to 0 during initial
239 NOTE: The machine must be power cycled (full off/on) when enabling NVM writes
240 via setting the parameter to zero. Once the NVM has been locked (via the
241 parameter at 1 when the driver loads) it cannot be unlocked except via power
246 :Valid Range: 0-16 (0=none,...,16=all)
249 This parameter adjusts the level of debug messages displayed in the system logs.
252 Additional Features and Configurations
253 ======================================
257 Jumbo Frames support is enabled by changing the Maximum Transmission Unit (MTU)
258 to a value larger than the default value of 1500.
260 Use the ifconfig command to increase the MTU size. For example, enter the
261 following where <x> is the interface number::
263 ifconfig eth<x> mtu 9000 up
265 Alternatively, you can use the ip command as follows::
267 ip link set mtu 9000 dev eth<x>
268 ip link set up dev eth<x>
270 This setting is not saved across reboots. The setting change can be made
271 permanent by adding 'MTU=9000' to the file:
273 - For RHEL: /etc/sysconfig/network-scripts/ifcfg-eth<x>
274 - For SLES: /etc/sysconfig/network/<config_file>
276 NOTE: The maximum MTU setting for Jumbo Frames is 8996. This value coincides
277 with the maximum Jumbo Frames size of 9018 bytes.
279 NOTE: Using Jumbo frames at 10 or 100 Mbps is not supported and may result in
280 poor performance or loss of link.
282 NOTE: The following adapters limit Jumbo Frames sized packets to a maximum of
285 - Intel(R) 82578DM Gigabit Network Connection
286 - Intel(R) 82577LM Gigabit Network Connection
288 The following adapters do not support Jumbo Frames:
290 - Intel(R) PRO/1000 Gigabit Server Adapter
291 - Intel(R) PRO/1000 PM Network Connection
292 - Intel(R) 82562G 10/100 Network Connection
293 - Intel(R) 82562G-2 10/100 Network Connection
294 - Intel(R) 82562GT 10/100 Network Connection
295 - Intel(R) 82562GT-2 10/100 Network Connection
296 - Intel(R) 82562V 10/100 Network Connection
297 - Intel(R) 82562V-2 10/100 Network Connection
298 - Intel(R) 82566DC Gigabit Network Connection
299 - Intel(R) 82566DC-2 Gigabit Network Connection
300 - Intel(R) 82566DM Gigabit Network Connection
301 - Intel(R) 82566MC Gigabit Network Connection
302 - Intel(R) 82566MM Gigabit Network Connection
303 - Intel(R) 82567V-3 Gigabit Network Connection
304 - Intel(R) 82577LC Gigabit Network Connection
305 - Intel(R) 82578DC Gigabit Network Connection
307 NOTE: Jumbo Frames cannot be configured on an 82579-based Network device if
308 MACSec is enabled on the system.
313 The driver utilizes the ethtool interface for driver configuration and
314 diagnostics, as well as displaying statistical information. The latest ethtool
315 version is required for this functionality. Download it at:
317 https://www.kernel.org/pub/software/network/ethtool/
319 NOTE: When validating enable/disable tests on some parts (for example, 82578),
320 it is necessary to add a few seconds between tests when working with ethtool.
323 Speed and Duplex Configuration
324 ------------------------------
325 In addressing speed and duplex configuration issues, you need to distinguish
326 between copper-based adapters and fiber-based adapters.
328 In the default mode, an Intel(R) Ethernet Network Adapter using copper
329 connections will attempt to auto-negotiate with its link partner to determine
330 the best setting. If the adapter cannot establish link with the link partner
331 using auto-negotiation, you may need to manually configure the adapter and link
332 partner to identical settings to establish link and pass packets. This should
333 only be needed when attempting to link with an older switch that does not
334 support auto-negotiation or one that has been forced to a specific speed or
335 duplex mode. Your link partner must match the setting you choose. 1 Gbps speeds
336 and higher cannot be forced. Use the autonegotiation advertising setting to
337 manually set devices for 1 Gbps and higher.
339 Speed, duplex, and autonegotiation advertising are configured through the
342 Caution: Only experienced network administrators should force speed and duplex
343 or change autonegotiation advertising manually. The settings at the switch must
344 always match the adapter settings. Adapter performance may suffer or your
345 adapter may not operate if you configure the adapter differently from your
348 An Intel(R) Ethernet Network Adapter using fiber-based connections, however,
349 will not attempt to auto-negotiate with its link partner since those adapters
350 operate only in full duplex and only at their native speed.
353 Enabling Wake on LAN* (WoL)
354 ---------------------------
355 WoL is configured through the ethtool* utility.
357 WoL will be enabled on the system during the next shut down or reboot. For
358 this driver version, in order to enable WoL, the e1000e driver must be loaded
359 prior to shutting down or suspending the system.
361 NOTE: Wake on LAN is only supported on port A for the following devices:
362 - Intel(R) PRO/1000 PT Dual Port Network Connection
363 - Intel(R) PRO/1000 PT Dual Port Server Connection
364 - Intel(R) PRO/1000 PT Dual Port Server Adapter
365 - Intel(R) PRO/1000 PF Dual Port Server Adapter
366 - Intel(R) PRO/1000 PT Quad Port Server Adapter
367 - Intel(R) Gigabit PT Quad Port Server ExpressModule
372 For general information, go to the Intel support website at:
374 https://www.intel.com/support/
376 or the Intel Wired Networking project hosted by Sourceforge at:
378 https://sourceforge.net/projects/e1000
380 If an issue is identified with the released source code on a supported kernel
381 with a supported adapter, email the specific information related to the issue
382 to e1000-devel@lists.sf.net.