forked from intersvyaz/ixgbe
ixgbe driver mirror
License
AlexeyManikin/ixgbe
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
ixgbe Linux* Base Driver for Intel(R) Ethernet Network Connection =============================================================== ================================================================================ October 5, 2015 ================================================================================ Contents -------- - Important Note - Overview - Building and Installation - Command Line Parameters - Additional Configurations - Known Issues/Troubleshooting - Support - License ================================================================================ Important Notes --------------- Disable LRO if enabling ip forwarding or bridging ------------------------------------------------- WARNING: The ixgbe driver supports the Large Receive Offload (LRO) feature. This option offers the lowest CPU utilization for receives but is completely incompatible with *routing/ip forwarding* and *bridging*. If enabling ip forwarding or bridging is a requirement, it is necessary to disable LRO using compile time options as noted in the LRO section later in this document. The result of not disabling LRO when combined with ip forwarding or bridging can be low throughput or even a kernel panic. Do not unload port driver if VF with active VM is bound to it ------------------------------------------------------------- Do not unload a port's driver if a Virtual Function (VF) with an active Virtual Machine (VM) is bound to it. Doing so will cause the port to appear to hang. Once the VM shuts down, or otherwise releases the VF, the command will complete. Configuring SR-IOV for improved network security ------------------------------------------------ In a virtualized environment, on Intel(R) Server Adapters that support SR-IOV, the virtual function (VF) may be subject to malicious behavior. Software- generated layer two frames, like IEEE 802.3x (link flow control), IEEE 802.1Qbb (priority based flow-control), and others of this type, are not expected and can throttle traffic between the host and the virtual switch, reducing performance. To resolve this issue, configure all SR-IOV enabled ports for VLAN tagging. This configuration allows unexpected, and potentially malicious, frames to be dropped. Overview -------- This document describes the ixgbe Linux* Base Driver for the 10 Gigabit PCI Express Family of Adapters. The Linux* base driver supports the following kernel versions: 2.6.x and newer It includes support for any Linux supported system, including Itanium(R)2 based systems, x86_64, i686, and PPC. This driver is only supported as a loadable module at this time. Intel is not supplying patches against the kernel source to allow for static linking of the drivers. A version of the driver may already be included by your distribution and/or the kernel.org kernel. For questions related to hardware requirements, refer to the documentation supplied with your Intel adapter. All hardware requirements listed apply to use with Linux. The following features are now available in supported kernels: - Native VLANs - Channel Bonding (teaming) - SNMP - Generic Receive Offload - Data Center Bridging Adapter teaming is implemented using the native Linux Channel bonding module. This is included in supported Linux kernels. Channel Bonding documentation can be found in the Linux kernel source: /documentation/networking/bonding.txt The driver information previously displayed in the /proc file system is not supported in this release. Driver information can be obtained using ethtool, lspci, and ifconfig. Instructions on updating ethtool can be found in the section Additional Configurations later in this document. Identifying Your Adapter ------------------------ The driver in this release is compatible with 82598, 82599, X540, x550, and X552-based Intel Ethernet Network Connections and Adapters. For information on how to identify your adapter, go to the Adapter & Driver ID Guide at: http://support.intel.com/support/go/network/adapter/proidguide.htm For the latest Intel network drivers, refer to the following website and select your adapter. http://www.intel.com/support SFP+ Devices with Pluggable Optics ---------------------------------- 82599-BASED ADAPTERS -------------------- NOTES: - If your 82599-based Intel(R) Network Adapter came with Intel optics or is an Intel(R) Ethernet Server Adapter X520-2, then it only supports Intel optics and/or the direct attach cables listed below. - When 82599-based SFP+ devices are connected back to back, they should be set to the same Speed setting via ethtool. Results may vary if you mix speed settings. Supplier Type Part Numbers -------- ---- ------------ SR Modules Intel DUAL RATE 1G/10G SFP+ SR (bailed) FTLX8571D3BCV-IT Intel DUAL RATE 1G/10G SFP+ SR (bailed) AFBR-703SDZ-IN2 Intel DUAL RATE 1G/10G SFP+ SR (bailed) AFBR-703SDDZ-IN1 LR Modules Intel DUAL RATE 1G/10G SFP+ LR (bailed) FTLX1471D3BCV-IT Intel DUAL RATE 1G/10G SFP+ LR (bailed) AFCT-701SDZ-IN2 Intel DUAL RATE 1G/10G SFP+ LR (bailed) AFCT-701SDDZ-IN1 The following is a list of 3rd party SFP+ modules that have received some testing. Not all modules are applicable to all devices. Supplier Type Part Numbers -------- ---- ------------ Finisar SFP+ SR bailed, 10g single rate FTLX8571D3BCL Avago SFP+ SR bailed, 10g single rate AFBR-700SDZ Finisar SFP+ LR bailed, 10g single rate FTLX1471D3BCL Finisar DUAL RATE 1G/10G SFP+ SR (No Bail) FTLX8571D3QCV-IT Avago DUAL RATE 1G/10G SFP+ SR (No Bail) AFBR-703SDZ-IN1 Finisar DUAL RATE 1G/10G SFP+ LR (No Bail) FTLX1471D3QCV-IT Avago DUAL RATE 1G/10G SFP+ LR (No Bail) AFCT-701SDZ-IN1 Finisar 1000BASE-T SFP FCLF8522P2BTL Avago 1000BASE-T ABCU-5710RZ HP 1000BASE-SX SFP 453153-001 82599-based adapters support all passive and active limiting direct attach cables that comply with SFF-8431 v4.1 and SFF-8472 v10.4 specifications. Laser turns off for SFP+ when ifconfig ethX down ------------------------------------------------ "ifconfig ethX down" turns off the laser for 82599-based SFP+ fiber adapters. "ifconfig ethX up" turns on the laser. 82599-based QSFP+ Adapters -------------------------- NOTES: - If your 82599-based Intel(R) Network Adapter came with Intel optics, it only supports Intel optics. - 82599-based QSFP+ adapters only support 4x10 Gbps connections. 1x40 Gbps connections are not supported. QSFP+ link partners must be configured for 4x10 Gbps. - 82599-based QSFP+ adapters do not support automatic link speed detection. The link speed must be configured to either 10 Gbps or 1 Gbps to match the link partners speed capabilities. Incorrect speed configurations will result in failure to link. - Intel(R) Ethernet Converged Network Adapter X520-Q1 only supports the optics and direct attach cables listed below. Supplier Type Part Numbers -------- ---- ------------ Intel DUAL RATE 1G/10G QSFP+ SRL (bailed) E10GQSFPSR 82599-based QSFP+ adapters support all passive and active limiting QSFP+ direct attach cables that comply with SFF-8436 v4.1 specifications. 82598-BASED ADAPTERS -------------------- NOTES: - Intel(r) Ethernet Network Adapters that support removable optical modules only support their original module type (for example, the Intel(R) 10 Gigabit SR Dual Port Express Module only supports SR optical modules). If you plug in a different type of module, the driver will not load. - Hot Swapping/hot plugging optical modules is not supported. - Only single speed, 10 gigabit modules are supported. - LAN on Motherboard (LOMs) may support DA, SR, or LR modules. Other module types are not supported. Please see your system documentation for details. The following is a list of SFP+ modules and direct attach cables that have received some testing. Not all modules are applicable to all devices. Supplier Type Part Numbers -------- ---- ------------ Finisar SFP+ SR bailed, 10g single rate FTLX8571D3BCL Avago SFP+ SR bailed, 10g single rate AFBR-700SDZ Finisar SFP+ LR bailed, 10g single rate FTLX1471D3BCL 82598-based adapters support all passive direct attach cables that comply with SFF-8431 v4.1 and SFF-8472 v10.4 specifications. Active direct attach cables are not supported. Third party optic modules and cables referred to above are listed only for the purpose of highlighting third party specifications and potential compatibility, and are not recommendations or endorsements or sponsorship of any third party's product by Intel. Intel is not endorsing or promoting products made by any third party and the third party reference is provided only to share information regarding certain optic modules and cables with the above specifications. There may be other manufacturers or suppliers, producing or supplying optic modules and cables with similar or matching descriptions. Customers must use their own discretion and diligence to purchase optic modules and cables from any third party of their choice. Customers are solely responsible for assessing the suitability of the product and/or devices and for the selection of the vendor for purchasing any product. THE OPTIC MODULES AND CABLES REFERRED TO ABOVE ARE NOT WARRANTED OR SUPPORTED BY INTEL. INTEL ASSUMES NO LIABILITY WHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO SALE AND/OR USE OF SUCH THIRD PARTY PRODUCTS OR SELECTION OF VENDOR BY CUSTOMERS. ================================================================================ Building and Installation ------------------------- To build a binary RPM* package of this driver, run 'rpmbuild -tb ixgbe-<x.x.x>.tar.gz', where <x.x.x> is the version number for the driver tar file. NOTES: - For the build to work properly, the currently running kernel MUST match the version and configuration of the installed kernel sources. If you have just recompiled the kernel reboot the system before building. - RPM functionality has only been tested in Red Hat distributions. 1. Move the base driver tar file to the directory of your choice. For example, use '/home/username/ixgbe' or '/usr/local/src/ixgbe'. 2. Untar/unzip the archive, where <x.x.x> is the version number for the driver tar file: tar zxf ixgbe-<x.x.x>.tar.gz 3. Change to the driver src directory, where <x.x.x> is the version number for the driver tar: cd ixgbe-<x.x.x>/src/ 4. Compile the driver module: # make install The binary will be installed as: /lib/modules/<KERNEL VERSION>/updates/drivers/net/ethernet/intel/ixgbe/ixgbe.ko The install location listed above is the default location. This may differ for various Linux distributions. 5. Load the module using the modprobe command: modprobe <ixgbe> [parameter=port1_value,port2_value] Make sure that any older ixgbe drivers are removed from the kernel before loading the new module: rmmod ixgbe; modprobe ixgbe 6. Assign an IP address to the interface by entering the following, where ethX is the interface name that was shown in dmesg after modprobe: ip address add <IP_address>/<netmask bits> dev ethX 7. Verify that the interface works. Enter the following, where IP_address is the IP address for another machine on the same subnet as the interface that is being tested: ping <IP_address> NOTE: For certain distributions like (but not limited to) RedHat Enterprise Linux 7 and Ubuntu, once the driver is installed the initrd/initramfs file may need to be updated to prevent the OS loading old versions of the ixgbe driver. The dracut utility may be used on RedHat distributions: # dracut --force For Ubuntu: # update-initramfs -u ================================================================================ Command Line Parameters ----------------------- If the driver is built as a module, the following optional parameters are used by entering them on the command line with the modprobe command using this syntax: modprobe ixgbe [<option>=<VAL1>,<VAL2>,...] There needs to be a <VAL#> for each network port in the system supported by this driver. The values will be applied to each instance, in function order. For example: modprobe ixgbe InterruptThrottleRate=16000,16000 In this case, there are two network ports supported by ixgbe in the system. The default value for each parameter is generally the recommended setting, unless otherwise noted. NOTES: - For more information about the command line parameters, see the application note at: http://www.intel.com/design/network/applnots/ap450.htm. - A descriptor describes a data buffer and attributes related to the data buffer. This information is accessed by the hardware. RSS --- Valid Range: 0-16 0 = Assign up to the lesser value of the number of CPUs or the number of queues X = Assign X queues, where X is less than or equal to the maximum number of queues (16 queues). RSS also effects the number of transmit queues allocated on 2.6.23 and newer kernels with CONFIG_NETDEVICES_MULTIQUEUE set in the kernel .config file. CONFIG_NETDEVICES_MULTIQUEUE only exists from 2.6.23 to 2.6.26. Other options enable multiqueue in 2.6.27 and newer kernels. Multiqueue ---------- Valid Range: 0, 1 0 = Disables Multiple Queue support 1 = Enabled Multiple Queue support (a prerequisite for RSS) Direct Cache Access (DCA) ------------------------- Valid Range: 0, 1 0 = Disables DCA support in the driver 1 = Enables DCA support in the driver If the driver is enabled for DCA, this parameter allows load-time control of the feature. Note: DCA is not supported on X550-based adapters. IntMode ------- Valid Range: 0-2 (0 = Legacy Int, 1 = MSI and 2 = MSI-X) IntMode controls allow load time control over the type of interrupt registered for by the driver. MSI-X is required for multiple queue support, and some kernels and combinations of kernel .config options will force a lower level of interrupt support. 'cat /proc/interrupts' will show different values for each type of interrupt. InterruptThrottleRate --------------------- Valid Range: 0=off 1=dynamic <min_ITR>-<max_ITR> Interrupt Throttle Rate controls the number of interrupts each interrupt vector can generate per second. Increasing ITR lowers latency at the cost of increased CPU utilization, though it may help throughput in some circumstances. 0 = Setting InterruptThrottleRate to 0 turns off any interrupt moderation and may improve small packet latency. However, this is generally not suitable for bulk throughput traffic due to the increased CPU utilization of the higher interrupt rate. NOTES: - On 82599, and X540, and X550-based adapters, disabling InterruptThrottleRate will also result in the driver disabling HW RSC. - On 82598-based adapters, disabling InterruptThrottleRate will also result in disabling LRO (Large Receive Offloads). 1 = Setting InterruptThrottleRate to Dynamic mode attempts to moderate interrupts per vector while maintaining very low latency. This can sometimes cause extra CPU utilization. If planning on deploying ixgbe in a latency sensitive environment, this parameter should be considered. <min_ITR>-<max_ITR> = 956-488281 Setting InterruptThrottleRate to a value greater or equal to <min_ITR> will program the adapter to send at most that many interrupts per second, even if more packets have come in. This reduces interrupt load on the system and can lower CPU utilization under heavy load, but will increase latency as packets are not processed as quickly. LLI (Low Latency Interrupts) ---------------------------- LLI allows for immediate generation of an interrupt upon processing receive packets that match certain criteria as set by the parameters described below. LLI parameters are not enabled when Legacy interrupts are used. You must be using MSI or MSI-X (see cat /proc/interrupts) to successfully use LLI. Note: LLI is not supported on X550-based adapters. LLIPort ------- Valid Range: 0-65535 LLI is configured with the LLIPort command-line parameter, which specifies which TCP port should generate Low Latency Interrupts. For example, using LLIPort=80 would cause the board to generate an immediate interrupt upon receipt of any packet sent to TCP port 80 on the local machine. WARNING: Enabling LLI can result in an excessive number of interrupts/second that may cause problems with the system and in some cases may cause a kernel panic. Note: LLI is not supported on X550-based adapters. LLIPush ------- Valid Range: 0-1 LLIPush can be set to be enabled or disabled (default). It is most effective in an environment with many small transactions. NOTE: Enabling LLIPush may allow a denial of service attack. Note: LLI is not supported on X550-based adapters. LLISize ------- Valid Range: 0-1500 LLISize causes an immediate interrupt if the board receives a packet smaller than the specified size. Note: LLI is not supported on X550-based adapters. LLIEType -------- Valid Range: 0-0x8FFF This parameter specifies the Low Latency Interrupt (LLI) Ethernet protocol type. Note: LLI is not supported on X550-based adapters. LLIVLANP -------- Valid Range: 0-7 This parameter specifies the LLI on VLAN priority threshold. Note: LLI is not supported on X550-based adapters. Flow Control ------------ Ethernet Flow Control (IEEE 802.3x) can be configured with ethtool to enable receiving and transmitting pause frames for ixgbe. When transmit is enabled, pause frames are generated when the receive packet buffer crosses a predefined threshold. When receive is enabled, the transmit unit will halt for the time delay specified when a pause frame is received. Flow Control is enabled by default. Use ethtool to change the flow control settings. ethtool: ethtool -A eth? autoneg off rx off tx off NOTE: For 82598 backplane cards entering 1 gigabit mode, flow control default behavior is changed to off. Flow control in 1 gigabit mode on these devices can lead to transmit hangs. NOTE: You must have a flow control capable link partner. Intel(R) Ethernet Flow Director ------------------------------- NOTE: Flow director parameters are only supported on kernel versions 2.6.30 or newer. The Flow Director performs the following tasks: - Directs receive packets according to their flows to different queues. - Enables tight control on routing a flow in the platform. - Matches flows and CPU cores for flow affinity. - Supports multiple parameters for flexible flow classification and load balancing. NOTES: - The Flow Director is enabled only if the kernel supports multiple transmit queues. - An included script (set_irq_affinity) automates setting the IRQ to CPU affinity. - Flow director masking works in the opposite manner from subnet masking. In the following command: #ethtool -N eth11 flow-type ip4 src-ip 172.4.1.2 m 255.0.0.0 dst-ip \ 172.21.1.1 m 255.128.0.0 action 31 The src-ip value that is written to the filter will be 0.4.1.2, not 172.0.0.0 as might be expected. Similarly, the dst-ip value written to the filter will be 0.21.1.1, not 172.0.0.0. ethtool commands: - To enable or disable the Flow Director # ethtool -K ethX ntuple <on|off> When disabling ntuple filters all the user programed filters are flushed from the driver cache and hardware. Filters must be re-added if they are needed when ntuple is re-enabled. - To add a filter that directs packet to queue 2, use -U or -N switch # ethtool -N ethX flow-type tcp4 src-ip 192.168.10.1 dst-ip \ 192.168.10.2 src-port 2000 dst-port 2001 action 2 [loc 1] - To see the list of filters currently present # ethtool <-u|-n> ethX Perfect Filter -------------- Perfect filter is an interface to load the filter table that funnels all flow into queue_0 unless an alternative queue is specified using "action". In that case, any flow that matches the filter criteria will be directed to the appropriate queue. Support for Virtual Function (VF) is through the user data field. ethtool must be updated to the version built for the 2.6.40 kernel. Perfect Filter is supported on all kernels 2.6.30 and later. Rules may be deleted from the table itself. This is done using "ethtool -U ethX delete N", where N is the rule number to be deleted. NOTE: Flow Director Perfect Filters can run in single queue mode when SR-IOV is enabled or when DCB is enabled. If the queue is defined as -1, the filter will drop matching packets. To account for filter matches and misses, there are two stats in ethtool: fdir_match and fdir_miss. In addition, rx_queue_N_packets shows the number of packets processed by the Nth queue. NOTES: - Receive Packet Steering (RPS) and Receive Flow Steering (RFS) are not compatible with Flow Director. If Flow Director is enabled, these will be disabled - For VLAN Masks only four masks are supported. - Once a rule is defined, you must supply the same fields and masks (if masks are specified). Support for UDP RSS ------------------- This feature adds an ON/OFF switch for hashing over certain flow types. Only UDP can be turned on. The default setting is disabled. Only support for enabling/disabling hashing on ports for UDP over IPv4 (UDP4) or IPv6 (UDP6) is supported. NOTE: Fragmented packets may arrive out of order when RSS UDP support is configured. Supported Ethtool Commands and Options: -n --show-nfc Retrieves the receive network flow classification configurations. rx-flow-hash tcp4|udp4|ah4|esp4|sctp4|tcp6|udp6|ah6|esp6|sctp6 Retrieves the hash options for the specified network traffic type. -N --config-nfc Configures the receive network flow classification. rx-flow-hash tcp4|udp4|ah4|esp4|sctp4|tcp6|udp6|ah6|esp6|sctp6 m|v|t|s|d|f|n|r... Configures the hash options for the specified network traffic type. udp4 UDP over IPv4 udp6 UDP over IPv6 f Hash on bytes 0 and 1 of the Layer 4 header of the rx packet. n Hash on bytes 2 and 3 of the Layer 4 header of the rx packet. The following is an example using udp4 (UDP over IPv4): - To include UDP port numbers in RSS hashing run: ethtool -N ethX rx-flow-hash udp4 sdfn - To exclude UDP port numbers from RSS hashing run: ethtool -N ethX rx-flow-hash udp4 sd - To display UDP hashing current configuration run: ethtool -n ethX rx-flow-hash udp4 The results of running that call will be the following, if UDP hashing is enabled. UDP over IPV4 flows use these fields for computing Hash flow key: IP SA IP DA L4 bytes 0 & 1 [TCP/UDP src port] L4 bytes 2 & 3 [TCP/UDP dst port] The results if UDP hashing is disabled are shown below. UDP over IPV4 flows use these fields for computing Hash flow key: IP SA IP DA Parameters FdirPballoc and AtrSampleRate impact Flow Director. FdirPballoc ----------- Valid Range: 1-3 Specifies the Flow Director allocated packet buffer size. 1 = 64k 2 = 128k 3 = 256k AtrSampleRate ------------- Valid Range: 0-255 This parameter is used with the Flow Director and is the software ATR transmit packet sample rate. For example, when AtrSampleRate is set to 20, every 20th packet looks to see if the packet will create a new flow. A value of 0 indicates that ATR should be disabled and no samples will be taken. max_vfs ------- Valid Range: 1-63 If the value is greater than 0 it will also force the VMDq parameter to be 1 or more. NOTE: This parameter is only used on kernel 3.7.x and below. On kernel 3.8.x and above, use sysfs to enable VFs. For example: #echo $num_vf_enabled > /sys/class/net/$dev/device/sriov_numvfs //enable VFs #echo 0 > /sys/class/net/$dev/device/sriov_numvfs //disable VFs The parameters for the driver are referenced by position. Thus, if you have a dual port adapter, or more than one adapter in your system, and want N virtual functions per port, you must specify a number for each port with each parameter separated by a comma. For example: modprobe ixgbe max_vfs=4,1 NOTE: Caution must be used in loading the driver with these parameters. Depending on your system configuration, number of slots, etc., it is impossible to predict in all cases where the positions would be on the command line. This parameter adds support for SR-IOV. It causes the driver to spawn up to max_vfs worth of virtual functions. NOTE: When either SR-IOV mode or VMDq mode is enabled, hardware VLAN filtering and VLAN tag stripping/insertion will remain enabled. Please remove the old VLAN filter before the new VLAN filter is added. For example, ip link set eth0 vf 0 vlan 100 // set vlan 100 for VF 0 ip link set eth0 vf 0 vlan 0 // Delete vlan 100 ip link set eth0 vf 0 vlan 200 // set a new vlan 200 for VF 0 With kernel 3.6, the driver supports the simultaneous usage of max_vfs and DCB features, subject to the constraints described below. Prior to kernel 3.6, the driver did not support the simultaneous operation of max_vfs greater than 0 and the DCB features (multiple traffic classes utilizing Priority Flow Control and Extended Transmission Selection). When DCB is enabled, network traffic is transmitted and received through multiple traffic classes (packet buffers in the NIC). The traffic is associated with a specific class based on priority, which has a value of 0 through 7 used in the VLAN tag. When SR-IOV is not enabled, each traffic class is associated with a set of receive/transmit descriptor queue pairs. The number of queue pairs for a given traffic class depends on the hardware configuration. When SR-IOV is enabled, the descriptor queue pairs are grouped into pools. The Physical Function (PF) and each Virtual Function (VF) is allocated a pool of receive/transmit descriptor queue pairs. When multiple traffic classes are configured (for example, DCB is enabled), each pool contains a queue pair from each traffic class. When a single traffic class is configured in the hardware, the pools contain multiple queue pairs from the single traffic class. The number of VFs that can be allocated depends on the number of traffic classes that can be enabled. The configurable number of traffic classes for each enabled VF is as follows: 0 - 15 VFs = Up to 8 traffic classes, depending on device support 16 - 31 VFs = Up to 4 traffic classes 32 - 63 VFs = 1 traffic class When VFs are configured, the PF is allocated one pool as well. The PF supports the DCB features with the constraint that each traffic class will only use a single queue pair. When zero VFs are configured, the PF can support multiple queue pairs per traffic class. Multi Queue Support for Virtual Functions in SR-IOV Mode -------------------------------------------------------- Multiple queues for virtual functions (VFs) is supported in this driver. To enable this feature at compile time, use the following command line: make CFLAGS_EXTRA="-DIXGBE_ENABLE_VF_MQ" install NOTE: The ixgbevf driver also needs to support multiple queues. Configuring SR-IOV for improved network security ------------------------------------------------ In a virtualized environment, on Intel(R) Server Adapters that support SR-IOV, the virtual function (VF) may be subject to malicious behavior. Software- generated layer two frames, like IEEE 802.3x (link flow control), IEEE 802.1Qbb (priority based flow-control), and others of this type, are not expected and can throttle traffic between the host and the virtual switch, reducing performance. To resolve this issue, configure all SR-IOV enabled ports for VLAN tagging. This configuration allows unexpected, and potentially malicious, frames to be dropped. Configuring VLAN tagging on SR-IOV enabled adapter ports -------------------------------------------------------- To configure VLAN tagging for the ports on an SR-IOV enabled adapter, use the following command. The VLAN configuration should be done before the VF driver is loaded or the VM is booted. $ ip link set dev <PF netdev id> vf <id> vlan <vlan id> For example, the following instructions will configure PF eth0 and the first VF on VLAN 10. $ ip link set dev eth0 vf 0 vlan 10 . L2LBen ------ Valid Range: 0 (disabled), 1 (enabled) This parameter controls the internal switch (L2 loopback between pf and vf). By default the switch is enabled. LRO --- Valid Range: 0(off), 1(on) Large Receive Offload (LRO) is a technique for increasing inbound throughput of high-bandwidth network connections by reducing CPU overhead. It works by aggregating multiple incoming packets from a single stream into a larger buffer before they are passed higher up the networking stack, thus reducing the number of packets that have to be processed. LRO combines multiple Ethernet frames into a single receive in the stack, thereby potentially decreasing CPU utilization for receives. IXGBE_NO_LRO is a compile time flag. The user can enable it at compile time to add support for LRO from the driver. The flag is used by adding CFLAGS_EXTRA="-DIXGBE_NO_LRO" to the make file when it's being compiled. # make CFLAGS_EXTRA="-DIXGBE_NO_LRO" install You can verify that the driver is using LRO by looking at these counters in ethtool: - lro_aggregated - counts total packets that were combined - lro_flushed - counts the number of packets flushed out of LRO NOTE: IPv6 and UDP are not supported by LRO. DMAC ---- Valid Range: 0, 41-10000 This parameter enables or disables DMA Coalescing feature. Values are in microseconds and set the internal DMA Coalescing internal timer. DMAC is available on Intel(R) X552 (and later) based adapters. DMA (Direct Memory Access) allows the network device to move packet data directly to the system's memory, reducing CPU utilization. However, the frequency and random intervals at which packets arrive do not allow the system to enter a lower power state. DMA Coalescing allows the adapter to collect packets before it initiates a DMA event. This may increase network latency but also increases the chances that the system will enter a lower power state. Turning on DMA Coalescing may save energy with kernel 2.6.32 and newer. DMA Coalescing must be enabled across all active ports in order to save platform power. InterruptThrottleRate (ITR) should be set to dynamic. When ITR=0, DMA Coalescing is automatically disabled. A whitepaper containing information on how to best configure your platform is available on the Intel website. AQRate ------ Devices that support AQRate (X550 and later) will include 2.5 Gbps and 5 Gbps in the speeds that the driver advertises during auto-negotiation, even though ethtool will not display 2.5 Gbps or 5 Gbps as "Supported link modes" or "Advertised link modes." These speeds are only available through unmodified auto-negotiation. You cannot use ethtool -s advertise to force auto-negotiation to advertise 2.5 Gbps or 5 Gbps. If a 2.5 Gbps or 5 Gbps link is created, ethtool will report the correct link speed. ================================================================================ Additional Features and Configurations ------------------------------------------- Configuring the Driver on Different Distributions ------------------------------------------------- Configuring a network driver to load properly when the system is started is distribution dependent. Typically, the configuration process involves adding an alias line to /etc/modules.conf or /etc/modprobe.conf as well as editing other system startup scripts and/or configuration files. Many popular Linux distributions ship with tools to make these changes for you. To learn the proper way to configure a network device for your system, refer to your distribution documentation. If during this process you are asked for the driver or module name, the name for the Base Driver is ixgbe. For example, if you install the ixgbe driver for two adapters (eth0 and eth1) and want to set the interrupt mode to MSI-X and MSI, respectively, add the following to modules.conf or /etc/modprobe.conf: alias eth0 ixgbe alias eth1 ixgbe options ixgbe InterruptThrottleRate=3,1 Viewing Link Messages --------------------- Link messages will not be displayed to the console if the distribution is restricting system messages. In order to see network driver link messages on your console, set dmesg to eight by entering the following: dmesg -n 8 NOTE: This setting is not saved across reboots. Jumbo Frames ------------ Jumbo Frames support is enabled by changing the Maximum Transmission Unit (MTU) to a value larger than the default value of 1500. Use the ifconfig command to increase the MTU size. For example, enter the following where <x> is the interface number: ifconfig eth<x> mtu 9000 up NOTES: - The maximum MTU setting for Jumbo Frames is 9710. This value coincides with the maximum Jumbo Frames size of 9728 bytes. - This driver will attempt to use multiple page sized buffers to receive each jumbo packet. This should help to avoid buffer starvation issues when allocating receive packets. - For 82599-based network connections, if you are enabling jumbo frames in a virtual function (VF), jumbo frames must first be enabled in the physical function (PF). The VF MTU setting cannot be larger than the PF MTU. ethtool ------- The driver utilizes the ethtool interface for driver configuration and diagnostics, as well as displaying statistical information. The latest ethtool version is required for this functionality. Download it at http://ftp.kernel.org/pub/software/network/ethtool/ Speed and Duplex Configuration ------------------------------ In addressing speed and duplex configuration issues, you need to distinguish between copper-based adapters and fiber-based adapters. In the default mode, an Intel(R) Network Adapter using copper connections will attempt to auto-negotiate with its link partner to determine the best setting. If the adapter cannot establish link with the link partner using auto-negotiation, you may need to manually configure the adapter and link partner to identical settings to establish link and pass packets. This should only be needed when attempting to link with an older switch that does not support auto-negotiation or one that has been forced to a specific speed or duplex mode. Your link partner must match the setting you choose. Speed and Duplex are configured through the ethtool* utility. ethtool is included with all versions of Red Hat after Red Hat 7.2. For other Linux distributions, download and install ethtool from the following website: http://ftp.kernel.org/pub/software/network/ethtool/ Caution: Only experienced network administrators should force speed and duplex manually. The settings at the switch must always match the adapter settings. Adapter performance may suffer or your adapter may not operate if you configure the adapter differently from your switch. An Intel(R) Network Adapter using fiber-based connections, however, will not attempt to auto-negotiate with its link partner since those adapters operate only in full duplex and only at their native speed. NOTE: For the Intel(R) Ethernet Connection X552 10 GbE SFP+ you must specify the desired speed. Hardware Receive Side Coalescing (HW RSC) ----------------------------------------- 82599 and X540, and X550-based adapters support HW RSC, which can merge multiple frames from the same IPv4 TCP/IP flow into a single structure that can span one or more descriptors. It works similarly to Software Large Receive Offload technique. By default HW RSC is enabled and SW LRO cannot be used for 82599 or X540, or X550-based adapters unless HW RSC is disabled. IXGBE_NO_HW_RSC is a compile time flag. The user can enable it at compile time to remove support for HW RSC from the driver. The flag is used by adding CFLAGS_EXTRA="-DIXGBE_NO_HW_RSC" to the make file when it is being compiled. make CFLAGS_EXTRA="-DIXGBE_NO_HW_RSC" install You can verify that the driver is using HW RSC by looking at the counter in ethtool: - hw_rsc_count. This counts the total number of Ethernet packets that were being combined. MAC and VLAN anti-spoofing feature ---------------------------------- When a malicious driver attempts to send a spoofed packet, it is dropped by the hardware and not transmitted. An interrupt is sent to the PF driver notifying it of the spoof attempt. When a spoofed packet is detected, the PF driver will send the following message to the system log (displayed by the "dmesg" command): ixgbe ethX: ixgbe_spoof_check: n spoofed packets detected where "x" is the PF interface number; and "n" is number of spoofed packets. NOTE: This feature can be disabled for a specific Virtual Function (VF). ip link set <pf dev> vf <vf id> spoofchk {off|on} IPRoute2 Tool for setting MAC address, VLAN and rate limit ---------------------------------------------------------- You can set a MAC address of a Virtual Function (VF), a default VLAN and the rate limit using the IProute2 tool. Download the latest version of the iproute2 tool from Sourceforge if your version does not have all the features you require. Wake on LAN Support (WoL) ------------------------- Some adapters do not support Wake on LAN. To determine if your adapter supports Wake on LAN, run ethtool ethX The Intel(R) Ethernet Converged Network Adapter X550-T1 and Intel(R) Ethernet Converged Network Adapter X550-T2 have a manageability/AUX power connector. These devices only support WoL if AUX power is supplied via this connector. Note that this is system and adapter specific. Some with this connector do not support WoL. Some systems do not provide the correct power connection. See your system documentation for details. IEEE 1588 Precision Time Protocol (PTP) Hardware Clock (PHC) ------------------------------------------------------------ Precision Time Protocol (PTP) is used to synchronize clocks in a computer network and is supported in the ixgbe driver. IXGBE_PTP is a compile time flag. The user can enable it at compile time to add support for PTP from the driver. The flag is used by editing the make file as follows when it is being compiled: >make CFLAGS_EXTRA="-DIXGBE_PTP" install VXLAN Overlay HW Offloading --------------------------- Virtual Extensible LAN (VXLAN) allows you to extend an L2 network over an L3 network, which may be useful in a virtualized or cloud environment. Some Intel(R) Ethernet Network devices perform VXLAN processing, offloading it from the operating system. This reduces CPU utilization. VXLAN offloading is controlled by the tx and rx checksum offload options provided by ethtool. That is, if tx checksum offload is enabled, and the adapter has the capability, VXLAN offloading is also enabled. If rx checksum offload is enabled, then the VXLAN packets rx checksum will be offloaded, unless the module parameter vxlan_rx=0,0 was used to specifically disable the VXLAN rx offload. VXLAN Overlay HW Offloading is enabled by default. To view and configure VXLAN on a VXLAN-overlay offload enabled device, use the following command: # ethtool -k ethX (This command displays the offloads and their current state.) For more information on configuring your network for overlay HW offloading support, refer to the Intel Technical Brief, "Creating Overlay Networks Using Intel Ethernet Converged Network Adapters" (Intel Networking Division, August 2013): http://www.intel.com/content/dam/www/public/us/en/documents/technology-briefs/ overlay-networks-using-converged-network-adapters-brief.pdf Virtual Function (VF) TX Rate Limit ----------------------------------- Virtual Function (VF) TX rate limit is configured with an ip command from the PF interface. # ip link set eth0 vf 0 rate 1000 (This command sets TX Rate Limit of 1000Mbps for VF 0) Note that the limit is set per queue and not for the entire VF interface. ================================================================================ Known Issues/Troubleshooting ---------------------------- MAC address of Virtual Function changes unexpectedly ---------------------------------------------------- If a Virtual Function's MAC address is not assigned in the host, then the VF (virtual function) driver will use a random MAC address. This random MAC address may change each time the VF driver is reloaded. You can assign a static MAC address in the host machine. This static MAC address will survive a VF driver reload. Hardware Issues --------------- For known hardware and troubleshooting issues, either refer to the "Release Notes" in your User Guide, or for more detailed information, go to http://www.intel.com. In the search box enter your devices controller ID followed by "spec update" (i.e., 82599 spec update). The specification update file has complete information on known hardware issues. Software Issues --------------- NOTE: After installing the driver, if your Intel Ethernet Network Connection is not working, verify that you have installed the correct driver. Intel(R) Active Management Technology 2.0, 2.1, 2.5 Not Supported in Conjunction with Linux driver. MSI-X Issues with Kernels Between 2.6.19 - 2.6.21 (inclusive) ------------------------------------------------------------- Kernel panics and instability may be observed on any MSI-X hardware if you use irqbalance with kernels between 2.6.19 and 2.6.21. If such problems are encountered, you may disable the irqbalance daemon or upgrade to a newer kernel. LRO and iSCSI Incompatibility ----------------------------- LRO is incompatible with iSCSI target or initiator traffic. A panic may occur when iSCSI traffic is received through the ixgbe driver with LRO enabled. To workaround this, the driver should be built and installed with: # make CFLAGS_EXTRA=-DIXGBE_NO_LRO install Multiple Interfaces on Same Ethernet Broadcast Network ------------------------------------------------------ Due to the default ARP behavior on Linux, it is not possible to have one system on two IP networks in the same Ethernet broadcast domain (non-partitioned switch) behave as expected. All Ethernet interfaces will respond to IP traffic for any IP address assigned to the system. This results in unbalanced receive traffic. If you have multiple interfaces in a server, either turn on ARP filtering by entering: echo 1 > /proc/sys/net/ipv4/conf/all/arp_filter This only works if your kernel's version is higher than 2.4.5. NOTE: This setting is not saved across reboots. The configuration change can be made permanent by adding the following line to the file /etc/sysctl.conf: net.ipv4.conf.all.arp_filter = 1 Another alternative is to install the interfaces in separate broadcast domains (either in different switches or in a switch partitioned to VLANs). UDP Stress Test Dropped Packet Issue ------------------------------------ Under small packet UDP stress with the ixgbedriver, the system may drop UDP packets due to socket buffers being full. Setting the driver Flow Control variables to the minimum may resolve the issue. You may also try increasing the kernel's default buffer sizes by changing the values in /proc/sys/net/core/rmem_default and rmem_max Cisco Catalyst 4948-10GE port resets may cause switch to shut down ports ------------------------------------------------------------------------ 82598-based hardware can re-establish link quickly and when connected to some switches, rapid resets within the driver may cause the switch port to become isolated due to "link flap". This is typically indicated by a yellow instead of a green link light. Several operations may cause this problem, such as repeatedly running ethtool commands that cause a reset. A potential workaround is to use the Cisco IOS command "no errdisable detect cause all" from the Global Configuration prompt which enables the switch to keep the interfaces up, regardless of errors. Rx Page Allocation Errors ------------------------- 'Page allocation failure. order:0' errors may occur under stress with kernels 2.6.25 and newer. This is caused by the way the Linux kernel reports this stressed condition. DCB: Generic segmentation offload on causes bandwidth allocation issues ----------------------------------------------------------------------- In order for DCB to work correctly, Generic Segmentation Offload (GSO), also known as software TSO, must be disabled using ethtool. Since the hardware supports TSO (hardware offload of segmentation), GSO will not be running by default. The GSO state can be queried with ethtool using ethtool -k ethX. When using 82598-based network connections, ixgbe driver only supports 16 queues on a platform with more than 16 cores. Due to known hardware limitations, RSS can only filter in a maximum of 16 receive queues. 82599 and X540, and X550-based network connections support up to 64 queues. Disable GRO when routing/bridging --------------------------------- Due to a known kernel issue, GRO must be turned off when routing/bridging. GRO can be turned off via ethtool. ethtool -K ethX gro off where ethX is the ethernet interface being modified. Lower than expected performance ------------------------------- Some PCIe x8 slots are actually configured as x4 slots. These slots have insufficient bandwidth for full line rate with dual port and quad port devices. In addition, if you put a PCIe Generation 3-capable adapter into a PCIe Generation 2 slot, you cannot get full bandwidth. The driver detects this situation and writes the following message in the system log: "PCI-Express bandwidth available for this card is not sufficient for optimal performance. For optimal performance a x8 PCI-Express slot is required." If this error occurs, moving your adapter to a true PCIe Generation 3 x8 slot will resolve the issue. ethtool may incorrectly display SFP+ fiber module as direct attached cable -------------------------------------------------------------------------- Due to kernel limitations, port type can only be correctly displayed on kernel 2.6.33 or greater. Under Redhat 5.4, system may crash when closing guest OS window after loading/unloading the Physical Function (PF) driver. Do not remove the ixgbe driver from Dom0 while Virtual Functions (VFs) are assigned to guests. VFs must first use the xm "pci-detach" command to hot-plug the VF device out of the VM it is assigned to or else shut down the VM. Unloading Physical Function (PF) driver may cause kernel panic or system reboot when VM is running and VF is loaded on the VM. On pre-3.2 Linux kernels, unloading the Physical Function (PF) driver causes system reboots when the VM is running and VF is loaded on the VM. Do not unload the PF driver (ixgbe) while VFs are assigned to guests. Running ethtool -t ethX command causes break between PF and test client ----------------------------------------------------------------------- When there are active VFs, "ethtool -t" will only run the link test. The driver will also log in syslog that VFs should be shut down to run a full diagnostic test. SLES10 SP3 Random System Panic when Reloading Driver ---------------------------------------------------- This is a known SLES-10 SP3 issue. After requesting interrupts for MSI-X vectors, system may panic. Currently, the only known workaround is to build the driver with CFLAGS_EXTRA=-DDISABLE_PCI_MSI if the driver needs to be loaded/unloaded. Otherwise, the driver can be loaded once and will be safe, but unloading it will lead to the issue. Enabling SR-IOV in a 32-bit Microsoft* Windows* Server 2008 Guest OS Using Intel® 82576-based GbE or Intel® 82599-based 10GbE Controller Under KVM KVM Hypervisor/VMM supports direct assignment of a PCIe device to a VM. This includes traditional PCIe devices, as well as SR-IOV-capable devices using Intel 82576-based and 82599-based controllers. While direct assignment of a PCIe device or an SR-IOV Virtual Function (VF) to a Linux-based VM running 2.6.32 or newer kernel is functional, there is a known issue with Microsoft Windows Server 2008 VM that results in a "yellow bang" error. This problem is within the KVM VMM itself, not the Intel driver or the SR-IOV logic of the VMM. The KVM emulates an older CPU model for the guests, and this older CPU model does not support MSI-X interrupts, which is a requirement for Intel SR-IOV. If you wish to use the Intel 82576- or 82599-based controllers in SR-IOV mode with KVM and a Microsoft Windows Server 2008 guest, try the following workaround. Configure KVM to emulate a different model of CPU when using qemu to create the KVM guest: "-cpu qemu64,model=13" Unable to obtain DHCP lease on boot with RedHat ----------------------------------------------- For configurations where the auto-negotiation process takes more than 5 seconds, the boot script may fail with the following message: "ethX: failed. No link present. Check cable?" If this error appears even though the presence of a link can be confirmed using ethtool ethX, try setting "LINKDELAY=5" in /etc/sysconfig/network-scripts/ifcfg-ethX. NOTE: Link time can take up to 30 seconds. Adjust LINKDELAY value accordingly. Host May Reboot after Removing PF when VF is Active in Guest ------------------------------------------------------------ Using kernel versions earlier than 3.2, do not unload the PF driver with active VFs. Doing this will cause your VFs to stop working until you reload the PF driver and may cause a spontaneous reboot of your system. Prior to unloading the PF driver, you must first ensure that all VFs are no longer active. Do this by shutting down all VMs and unloading the VF driver. Out of memory issues on IA32 systems ------------------------------------- The driver may consume a lot of memory based on the number of CPUs and network interfaces. This leads to memory segmentation. Thus, the driver may not be able to allocate enough memory. To resolve this, reduce the number of descriptors using ethtool -G or the number of queues through the RSS parameter. VLAN tags are stripped on kernels earlier than 2.6.36 ----------------------------------------------------- In order to support DCB, kernels earlier than 2.6.36 strip VLAN tags for VLAN0. This ensures connectivity using 802.1p frames between kernels that have built-in support and kernels that do not. If the VLAN tags are necessary AND DCB is NOT used, disable vlan stripping on older kernels at build time with the following: # make CFLAGS_EXTRA=-DIXGBE_DISABLE_8021P_SUPPORT ================================================================================ Support ------- For general information, go to the Intel support website at: www.intel.com/support/ or the Intel Wired Networking project hosted by Sourceforge at: http://sourceforge.net/projects/e1000 If an issue is identified with the released source code on a supported kernel with a supported adapter, email the specific information related to the issue to e1000-devel@lists.sf.net. ================================================================================ License ------- This program is free software; you can redistribute it and/or modify it under the terms and conditions of the GNU General Public License, version 2, as published by the Free Software Foundation. This program is distributed in the hope it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. The full GNU General Public License is included in this distribution in the file called "COPYING". Intel(R) 10 Gigabit Ethernet Network Driver Intel(R) 10 Gigabit Virtual Function Network Driver Copyright(c) 1999 - 2015 Intel Corporation. ================================================================================ Trademarks ---------- Intel, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. * Other names and brands may be claimed as the property of others.
About
ixgbe driver mirror
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published
Languages
- C 98.4%
- Other 1.6%