Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 1 | Linux* Base Driver for the Intel(R) PRO/1000 Family of Adapters |
| 2 | =============================================================== |
| 3 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 4 | Intel Gigabit Linux driver. |
| 5 | Copyright(c) 1999 - 2010 Intel Corporation. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 6 | |
| 7 | Contents |
| 8 | ======== |
| 9 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 10 | - Identifying Your Adapter |
| 11 | - Command Line Parameters |
| 12 | - Speed and Duplex Configuration |
| 13 | - Additional Configurations |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 14 | - Support |
| 15 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 16 | Identifying Your Adapter |
| 17 | ======================== |
| 18 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 19 | For more information on how to identify your adapter, go to the Adapter & |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 20 | Driver ID Guide at: |
| 21 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 22 | http://support.intel.com/support/go/network/adapter/idguide.htm |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 23 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 24 | For the latest Intel network drivers for Linux, refer to the following |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 25 | website. In the search field, enter your adapter name or type, or use the |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 26 | networking link on the left to search for your adapter: |
| 27 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 28 | http://support.intel.com/support/go/network/adapter/home.htm |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 29 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 30 | Command Line Parameters |
| 31 | ======================= |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 32 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 33 | The default value for each parameter is generally the recommended setting, |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 34 | unless otherwise noted. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 35 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 36 | NOTES: For more information about the AutoNeg, Duplex, and Speed |
| 37 | parameters, see the "Speed and Duplex Configuration" section in |
| 38 | this document. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 39 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 40 | For more information about the InterruptThrottleRate, |
| 41 | RxIntDelay, TxIntDelay, RxAbsIntDelay, and TxAbsIntDelay |
| 42 | parameters, see the application note at: |
| 43 | http://www.intel.com/design/network/applnots/ap450.htm |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 44 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 45 | AutoNeg |
| 46 | ------- |
| 47 | (Supported only on adapters with copper connections) |
| 48 | Valid Range: 0x01-0x0F, 0x20-0x2F |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 49 | Default Value: 0x2F |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 50 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 51 | This parameter is a bit-mask that specifies the speed and duplex settings |
| 52 | advertised by the adapter. When this parameter is used, the Speed and |
| 53 | Duplex parameters must not be specified. |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 54 | |
| 55 | NOTE: Refer to the Speed and Duplex section of this readme for more |
| 56 | information on the AutoNeg parameter. |
| 57 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 58 | Duplex |
| 59 | ------ |
| 60 | (Supported only on adapters with copper connections) |
| 61 | Valid Range: 0-2 (0=auto-negotiate, 1=half, 2=full) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 62 | Default Value: 0 |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 63 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 64 | This defines the direction in which data is allowed to flow. Can be |
| 65 | either one or two-directional. If both Duplex and the link partner are |
| 66 | set to auto-negotiate, the board auto-detects the correct duplex. If the |
| 67 | link partner is forced (either full or half), Duplex defaults to half- |
| 68 | duplex. |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 69 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 70 | FlowControl |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 71 | ----------- |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 72 | Valid Range: 0-3 (0=none, 1=Rx only, 2=Tx only, 3=Rx&Tx) |
| 73 | Default Value: Reads flow control settings from the EEPROM |
| 74 | |
| 75 | This parameter controls the automatic generation(Tx) and response(Rx) |
| 76 | to Ethernet PAUSE frames. |
| 77 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 78 | InterruptThrottleRate |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 79 | --------------------- |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 80 | (not supported on Intel(R) 82542, 82543 or 82544-based adapters) |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 81 | Valid Range: 0,1,3,4,100-100000 (0=off, 1=dynamic, 3=dynamic conservative, |
Jeff Kirsher | 96c3c03 | 2010-12-09 23:42:00 -0800 | [diff] [blame] | 82 | 4=simplified balancing) |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 83 | Default Value: 3 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 84 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 85 | The driver can limit the amount of interrupts per second that the adapter |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 86 | will generate for incoming packets. It does this by writing a value to the |
| 87 | adapter that is based on the maximum amount of interrupts that the adapter |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 88 | will generate per second. |
| 89 | |
| 90 | Setting InterruptThrottleRate to a value greater or equal to 100 |
| 91 | will program the adapter to send out a maximum of that many interrupts |
| 92 | per second, even if more packets have come in. This reduces interrupt |
| 93 | load on the system and can lower CPU utilization under heavy load, |
| 94 | but will increase latency as packets are not processed as quickly. |
| 95 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 96 | The default behaviour of the driver previously assumed a static |
| 97 | InterruptThrottleRate value of 8000, providing a good fallback value for |
| 98 | all traffic types,but lacking in small packet performance and latency. |
| 99 | The hardware can handle many more small packets per second however, and |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 100 | for this reason an adaptive interrupt moderation algorithm was implemented. |
| 101 | |
| 102 | Since 7.3.x, the driver has two adaptive modes (setting 1 or 3) in which |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 103 | it dynamically adjusts the InterruptThrottleRate value based on the traffic |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 104 | that it receives. After determining the type of incoming traffic in the last |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 105 | timeframe, it will adjust the InterruptThrottleRate to an appropriate value |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 106 | for that traffic. |
| 107 | |
| 108 | The algorithm classifies the incoming traffic every interval into |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 109 | classes. Once the class is determined, the InterruptThrottleRate value is |
| 110 | adjusted to suit that traffic type the best. There are three classes defined: |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 111 | "Bulk traffic", for large amounts of packets of normal size; "Low latency", |
| 112 | for small amounts of traffic and/or a significant percentage of small |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 113 | packets; and "Lowest latency", for almost completely small packets or |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 114 | minimal traffic. |
| 115 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 116 | In dynamic conservative mode, the InterruptThrottleRate value is set to 4000 |
| 117 | for traffic that falls in class "Bulk traffic". If traffic falls in the "Low |
| 118 | latency" or "Lowest latency" class, the InterruptThrottleRate is increased |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 119 | stepwise to 20000. This default mode is suitable for most applications. |
| 120 | |
| 121 | For situations where low latency is vital such as cluster or |
| 122 | grid computing, the algorithm can reduce latency even more when |
| 123 | InterruptThrottleRate is set to mode 1. In this mode, which operates |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 124 | the same as mode 3, the InterruptThrottleRate will be increased stepwise to |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 125 | 70000 for traffic in class "Lowest latency". |
| 126 | |
Jeff Kirsher | 96c3c03 | 2010-12-09 23:42:00 -0800 | [diff] [blame] | 127 | In simplified mode the interrupt rate is based on the ratio of TX and |
| 128 | RX traffic. If the bytes per second rate is approximately equal, the |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 129 | interrupt rate will drop as low as 2000 interrupts per second. If the |
| 130 | traffic is mostly transmit or mostly receive, the interrupt rate could |
| 131 | be as high as 8000. |
| 132 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 133 | Setting InterruptThrottleRate to 0 turns off any interrupt moderation |
| 134 | and may improve small packet latency, but is generally not suitable |
| 135 | for bulk throughput traffic. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 136 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 137 | NOTE: InterruptThrottleRate takes precedence over the TxAbsIntDelay and |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 138 | RxAbsIntDelay parameters. In other words, minimizing the receive |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 139 | and/or transmit absolute delays does not force the controller to |
| 140 | generate more interrupts than what the Interrupt Throttle Rate |
| 141 | allows. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 142 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 143 | CAUTION: If you are using the Intel(R) PRO/1000 CT Network Connection |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 144 | (controller 82547), setting InterruptThrottleRate to a value |
| 145 | greater than 75,000, may hang (stop transmitting) adapters |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 146 | under certain network conditions. If this occurs a NETDEV |
| 147 | WATCHDOG message is logged in the system event log. In |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 148 | addition, the controller is automatically reset, restoring |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 149 | the network connection. To eliminate the potential for the |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 150 | hang, ensure that InterruptThrottleRate is set no greater |
| 151 | than 75,000 and is not set to 0. |
| 152 | |
| 153 | NOTE: When e1000 is loaded with default settings and multiple adapters |
| 154 | are in use simultaneously, the CPU utilization may increase non- |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 155 | linearly. In order to limit the CPU utilization without impacting |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 156 | the overall throughput, we recommend that you load the driver as |
| 157 | follows: |
| 158 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 159 | modprobe e1000 InterruptThrottleRate=3000,3000,3000 |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 160 | |
| 161 | This sets the InterruptThrottleRate to 3000 interrupts/sec for |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 162 | the first, second, and third instances of the driver. The range |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 163 | of 2000 to 3000 interrupts per second works on a majority of |
| 164 | systems and is a good starting point, but the optimal value will |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 165 | be platform-specific. If CPU utilization is not a concern, use |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 166 | RX_POLLING (NAPI) and default driver settings. |
| 167 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 168 | RxDescriptors |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 169 | ------------- |
| 170 | Valid Range: 80-256 for 82542 and 82543-based adapters |
| 171 | 80-4096 for all other supported adapters |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 172 | Default Value: 256 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 173 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 174 | This value specifies the number of receive buffer descriptors allocated |
| 175 | by the driver. Increasing this value allows the driver to buffer more |
| 176 | incoming packets, at the expense of increased system memory utilization. |
| 177 | |
| 178 | Each descriptor is 16 bytes. A receive buffer is also allocated for each |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 179 | descriptor and can be either 2048, 4096, 8192, or 16384 bytes, depending |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 180 | on the MTU setting. The maximum MTU size is 16110. |
| 181 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 182 | NOTE: MTU designates the frame size. It only needs to be set for Jumbo |
| 183 | Frames. Depending on the available system resources, the request |
| 184 | for a higher number of receive descriptors may be denied. In this |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 185 | case, use a lower number. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 186 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 187 | RxIntDelay |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 188 | ---------- |
| 189 | Valid Range: 0-65535 (0=off) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 190 | Default Value: 0 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 191 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 192 | This value delays the generation of receive interrupts in units of 1.024 |
| 193 | microseconds. Receive interrupt reduction can improve CPU efficiency if |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 194 | properly tuned for specific network traffic. Increasing this value adds |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 195 | extra latency to frame reception and can end up decreasing the throughput |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 196 | of TCP traffic. If the system is reporting dropped receives, this value |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 197 | may be set too high, causing the driver to run out of available receive |
| 198 | descriptors. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 199 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 200 | CAUTION: When setting RxIntDelay to a value other than 0, adapters may |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 201 | hang (stop transmitting) under certain network conditions. If |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 202 | this occurs a NETDEV WATCHDOG message is logged in the system |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 203 | event log. In addition, the controller is automatically reset, |
| 204 | restoring the network connection. To eliminate the potential |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 205 | for the hang ensure that RxIntDelay is set to 0. |
| 206 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 207 | RxAbsIntDelay |
| 208 | ------------- |
| 209 | (This parameter is supported only on 82540, 82545 and later adapters.) |
| 210 | Valid Range: 0-65535 (0=off) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 211 | Default Value: 128 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 212 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 213 | This value, in units of 1.024 microseconds, limits the delay in which a |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 214 | receive interrupt is generated. Useful only if RxIntDelay is non-zero, |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 215 | this value ensures that an interrupt is generated after the initial |
| 216 | packet is received within the set amount of time. Proper tuning, |
| 217 | along with RxIntDelay, may improve traffic throughput in specific network |
| 218 | conditions. |
| 219 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 220 | Speed |
| 221 | ----- |
| 222 | (This parameter is supported only on adapters with copper connections.) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 223 | Valid Settings: 0, 10, 100, 1000 |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 224 | Default Value: 0 (auto-negotiate at all supported speeds) |
| 225 | |
| 226 | Speed forces the line speed to the specified value in megabits per second |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 227 | (Mbps). If this parameter is not specified or is set to 0 and the link |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 228 | partner is set to auto-negotiate, the board will auto-detect the correct |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 229 | speed. Duplex should also be set when Speed is set to either 10 or 100. |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 230 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 231 | TxDescriptors |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 232 | ------------- |
| 233 | Valid Range: 80-256 for 82542 and 82543-based adapters |
| 234 | 80-4096 for all other supported adapters |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 235 | Default Value: 256 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 236 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 237 | This value is the number of transmit descriptors allocated by the driver. |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 238 | Increasing this value allows the driver to queue more transmits. Each |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 239 | descriptor is 16 bytes. |
| 240 | |
| 241 | NOTE: Depending on the available system resources, the request for a |
| 242 | higher number of transmit descriptors may be denied. In this case, |
| 243 | use a lower number. |
| 244 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 245 | TxDescriptorStep |
| 246 | ---------------- |
| 247 | Valid Range: 1 (use every Tx Descriptor) |
Jeff Kirsher | 96c3c03 | 2010-12-09 23:42:00 -0800 | [diff] [blame] | 248 | 4 (use every 4th Tx Descriptor) |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 249 | |
| 250 | Default Value: 1 (use every Tx Descriptor) |
| 251 | |
| 252 | On certain non-Intel architectures, it has been observed that intense TX |
| 253 | traffic bursts of short packets may result in an improper descriptor |
| 254 | writeback. If this occurs, the driver will report a "TX Timeout" and reset |
| 255 | the adapter, after which the transmit flow will restart, though data may |
| 256 | have stalled for as much as 10 seconds before it resumes. |
| 257 | |
| 258 | The improper writeback does not occur on the first descriptor in a system |
| 259 | memory cache-line, which is typically 32 bytes, or 4 descriptors long. |
| 260 | |
| 261 | Setting TxDescriptorStep to a value of 4 will ensure that all TX descriptors |
| 262 | are aligned to the start of a system memory cache line, and so this problem |
| 263 | will not occur. |
| 264 | |
| 265 | NOTES: Setting TxDescriptorStep to 4 effectively reduces the number of |
| 266 | TxDescriptors available for transmits to 1/4 of the normal allocation. |
| 267 | This has a possible negative performance impact, which may be |
| 268 | compensated for by allocating more descriptors using the TxDescriptors |
| 269 | module parameter. |
| 270 | |
| 271 | There are other conditions which may result in "TX Timeout", which will |
| 272 | not be resolved by the use of the TxDescriptorStep parameter. As the |
| 273 | issue addressed by this parameter has never been observed on Intel |
| 274 | Architecture platforms, it should not be used on Intel platforms. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 275 | |
| 276 | TxIntDelay |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 277 | ---------- |
| 278 | Valid Range: 0-65535 (0=off) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 279 | Default Value: 64 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 280 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 281 | This value delays the generation of transmit interrupts in units of |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 282 | 1.024 microseconds. Transmit interrupt reduction can improve CPU |
| 283 | efficiency if properly tuned for specific network traffic. If the |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 284 | system is reporting dropped transmits, this value may be set too high |
| 285 | causing the driver to run out of available transmit descriptors. |
| 286 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 287 | TxAbsIntDelay |
| 288 | ------------- |
| 289 | (This parameter is supported only on 82540, 82545 and later adapters.) |
| 290 | Valid Range: 0-65535 (0=off) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 291 | Default Value: 64 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 292 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 293 | This value, in units of 1.024 microseconds, limits the delay in which a |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 294 | transmit interrupt is generated. Useful only if TxIntDelay is non-zero, |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 295 | this value ensures that an interrupt is generated after the initial |
| 296 | packet is sent on the wire within the set amount of time. Proper tuning, |
| 297 | along with TxIntDelay, may improve traffic throughput in specific |
| 298 | network conditions. |
| 299 | |
| 300 | XsumRX |
| 301 | ------ |
| 302 | (This parameter is NOT supported on the 82542-based adapter.) |
| 303 | Valid Range: 0-1 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 304 | Default Value: 1 |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 305 | |
| 306 | A value of '1' indicates that the driver should enable IP checksum |
| 307 | offload for received packets (both UDP and TCP) to the adapter hardware. |
| 308 | |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 309 | Copybreak |
| 310 | --------- |
| 311 | Valid Range: 0-xxxxxxx (0=off) |
| 312 | Default Value: 256 |
| 313 | Usage: insmod e1000.ko copybreak=128 |
| 314 | |
Jeff Kirsher | 96c3c03 | 2010-12-09 23:42:00 -0800 | [diff] [blame] | 315 | Driver copies all packets below or equaling this size to a fresh RX |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 316 | buffer before handing it up the stack. |
| 317 | |
| 318 | This parameter is different than other parameters, in that it is a |
| 319 | single (not 1,1,1 etc.) parameter applied to all driver instances and |
| 320 | it is also available during runtime at |
| 321 | /sys/module/e1000/parameters/copybreak |
| 322 | |
| 323 | SmartPowerDownEnable |
| 324 | -------------------- |
| 325 | Valid Range: 0-1 |
| 326 | Default Value: 0 (disabled) |
| 327 | |
| 328 | Allows PHY to turn off in lower power states. The user can turn off |
| 329 | this parameter in supported chipsets. |
| 330 | |
| 331 | KumeranLockLoss |
| 332 | --------------- |
| 333 | Valid Range: 0-1 |
| 334 | Default Value: 1 (enabled) |
| 335 | |
| 336 | This workaround skips resetting the PHY at shutdown for the initial |
| 337 | silicon releases of ICH8 systems. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 338 | |
| 339 | Speed and Duplex Configuration |
| 340 | ============================== |
| 341 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 342 | Three keywords are used to control the speed and duplex configuration. |
| 343 | These keywords are Speed, Duplex, and AutoNeg. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 344 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 345 | If the board uses a fiber interface, these keywords are ignored, and the |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 346 | fiber interface board only links at 1000 Mbps full-duplex. |
| 347 | |
| 348 | For copper-based boards, the keywords interact as follows: |
| 349 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 350 | The default operation is auto-negotiate. The board advertises all |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 351 | supported speed and duplex combinations, and it links at the highest |
| 352 | common speed and duplex mode IF the link partner is set to auto-negotiate. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 353 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 354 | If Speed = 1000, limited auto-negotiation is enabled and only 1000 Mbps |
| 355 | is advertised (The 1000BaseT spec requires auto-negotiation.) |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 356 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 357 | If Speed = 10 or 100, then both Speed and Duplex should be set. Auto- |
| 358 | negotiation is disabled, and the AutoNeg parameter is ignored. Partner |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 359 | SHOULD also be forced. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 360 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 361 | The AutoNeg parameter is used when more control is required over the |
| 362 | auto-negotiation process. It should be used when you wish to control which |
| 363 | speed and duplex combinations are advertised during the auto-negotiation |
| 364 | process. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 365 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 366 | The parameter may be specified as either a decimal or hexadecimal value as |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 367 | determined by the bitmap below. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 368 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 369 | Bit position 7 6 5 4 3 2 1 0 |
| 370 | Decimal Value 128 64 32 16 8 4 2 1 |
| 371 | Hex value 80 40 20 10 8 4 2 1 |
| 372 | Speed (Mbps) N/A N/A 1000 N/A 100 100 10 10 |
| 373 | Duplex Full Full Half Full Half |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 374 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 375 | Some examples of using AutoNeg: |
| 376 | |
| 377 | modprobe e1000 AutoNeg=0x01 (Restricts autonegotiation to 10 Half) |
| 378 | modprobe e1000 AutoNeg=1 (Same as above) |
| 379 | modprobe e1000 AutoNeg=0x02 (Restricts autonegotiation to 10 Full) |
| 380 | modprobe e1000 AutoNeg=0x03 (Restricts autonegotiation to 10 Half or 10 Full) |
| 381 | modprobe e1000 AutoNeg=0x04 (Restricts autonegotiation to 100 Half) |
| 382 | modprobe e1000 AutoNeg=0x05 (Restricts autonegotiation to 10 Half or 100 |
| 383 | Half) |
| 384 | modprobe e1000 AutoNeg=0x020 (Restricts autonegotiation to 1000 Full) |
| 385 | modprobe e1000 AutoNeg=32 (Same as above) |
| 386 | |
| 387 | Note that when this parameter is used, Speed and Duplex must not be specified. |
| 388 | |
| 389 | If the link partner is forced to a specific speed and duplex, then this |
| 390 | parameter should not be used. Instead, use the Speed and Duplex parameters |
| 391 | previously mentioned to force the adapter to the same speed and duplex. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 392 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 393 | Additional Configurations |
| 394 | ========================= |
| 395 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 396 | Jumbo Frames |
| 397 | ------------ |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 398 | Jumbo Frames support is enabled by changing the MTU to a value larger than |
| 399 | the default of 1500. Use the ifconfig command to increase the MTU size. |
| 400 | For example: |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 401 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 402 | ifconfig eth<x> mtu 9000 up |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 403 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 404 | This setting is not saved across reboots. It can be made permanent if |
| 405 | you add: |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 406 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 407 | MTU=9000 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 408 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 409 | to the file /etc/sysconfig/network-scripts/ifcfg-eth<x>. This example |
| 410 | applies to the Red Hat distributions; other distributions may store this |
| 411 | setting in a different location. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 412 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 413 | Notes: |
Jeff Kirsher | 2bff89c | 2010-10-05 01:17:05 +0000 | [diff] [blame] | 414 | Degradation in throughput performance may be observed in some Jumbo frames |
| 415 | environments. If this is observed, increasing the application's socket buffer |
| 416 | size and/or increasing the /proc/sys/net/ipv4/tcp_*mem entry values may help. |
| 417 | See the specific application manual and /usr/src/linux*/Documentation/ |
| 418 | networking/ip-sysctl.txt for more details. |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 419 | |
| 420 | - The maximum MTU setting for Jumbo Frames is 16110. This value coincides |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 421 | with the maximum Jumbo Frames size of 16128. |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 422 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 423 | - Using Jumbo Frames at 10 or 100 Mbps may result in poor performance or |
| 424 | loss of link. |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 425 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 426 | - Adapters based on the Intel(R) 82542 and 82573V/E controller do not |
| 427 | support Jumbo Frames. These correspond to the following product names: |
| 428 | Intel(R) PRO/1000 Gigabit Server Adapter |
| 429 | Intel(R) PRO/1000 PM Network Connection |
| 430 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 431 | Ethtool |
| 432 | ------- |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 433 | The driver utilizes the ethtool interface for driver configuration and |
Jeff Kirsher | 68f20d9 | 2010-12-17 12:14:34 +0000 | [diff] [blame] | 434 | diagnostics, as well as displaying statistical information. The ethtool |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 435 | version 1.6 or later is required for this functionality. |
| 436 | |
| 437 | The latest release of ethtool can be found from |
Jeff Kirsher | 68f20d9 | 2010-12-17 12:14:34 +0000 | [diff] [blame] | 438 | http://ftp.kernel.org/pub/software/network/ethtool/ |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 439 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 440 | Enabling Wake on LAN* (WoL) |
| 441 | --------------------------- |
Jeff Kirsher | 68f20d9 | 2010-12-17 12:14:34 +0000 | [diff] [blame] | 442 | WoL is configured through the ethtool* utility. |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 443 | |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 444 | WoL will be enabled on the system during the next shut down or reboot. |
| 445 | For this driver version, in order to enable WoL, the e1000 driver must be |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 446 | loaded when shutting down or rebooting the system. |
| 447 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 448 | Support |
| 449 | ======= |
| 450 | |
| 451 | For general information, go to the Intel support website at: |
| 452 | |
| 453 | http://support.intel.com |
| 454 | |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 455 | or the Intel Wired Networking project hosted by Sourceforge at: |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 456 | |
| 457 | http://sourceforge.net/projects/e1000 |
| 458 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 459 | If an issue is identified with the released source code on the supported |
Jesse Brandeburg | 55be062 | 2006-03-15 12:18:13 -0800 | [diff] [blame] | 460 | kernel with a supported adapter, email the specific information related |
Auke Kok | de3edab | 2006-11-01 08:48:18 -0800 | [diff] [blame] | 461 | to the issue to e1000-devel@lists.sf.net |