October 30, 2013
-
Important Note
-
Overview
-
Building and Installation
-
Command Line Parameters
-
Additional Configurations
-
Known Issues/Troubleshooting
-
Support
-
License
Important Note
WARNING: The ixgbe driver compiles by default with the Large Receive Offload (LRO) feature enabled. 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. Overview -------- This document describes the ixgbe Linux* Base Driver for the 10 Gigabit PCI Express Family of Adapters. The Linux* base driver supports the 2.6.x and 3.x kernels and includes support for any Linux supported system, including Itanium(R)2, x86_64, i686, and PPC. These drivers are 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 driver. 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 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. Alternatively, you can use ethtool (version 1.6 or later), lspci, and ifconfig to obtain the same information. 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 and X540-based Intel Ethernet Network Connections. For more 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 for Linux, refer to the following website. Select the link for your adapter. http://support.intel.com/support/go/network/adapter/home.htm SFP+ Devices with Pluggable Optics ---------------------------------- 82599-BASED ADAPTERS -------------------- NOTES: - If your 82599-based Intel® Network Adapter came with Intel optics or is an Intel® 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 SFP 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® Ethernet Network Adapters that support removable optical modules only support their original module type (for example, the Intel® 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.
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.
-
Move the base driver tar file to the directory of your choice. For example, use /home/username/ixgbe or /usr/local/src/ixgbe.
-
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
-
Change to the driver src directory, where <x.x.x> is the version number for the driver tar: cd ixgbe-<x.x.x>/src/
-
Compile the driver module: # make install
The binary will be installed as: /lib/modules/<KERNEL VERSION>/kernel/drivers/net/ixgbe/ixgbe.[k]o
The install location listed above is the default location. This may differ for various Linux distributions.
-
Load the module. Use the modprobe command: modprobe ixgbe <parameter>=<value>
Ensure that older ixgbe drivers are removed from the kernel, before loading the new module: rmmod ixgbe; modprobe ixgbe
-
Assign an IP address to the interface by entering the following, where x is the interface number: ifconfig eth<x> <IP_address> netmask <netmask>
-
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>
-
If your kernel supports DCA, the driver will build by default with DCA enabled.
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 InterruptThrottleRate parameter, 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.
Receive Side Scaling (RSS) or multiple queues for receives
----------------------------------------------------------
Valid Range: 0-16
Default Value: 1
0 = As# 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.
RSS affects 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 kernels 2.6.23 to 2.6.26. Other
options enable multiqueue in 2.6.27 and newer kernels.
Multi Queue (MQ)
----------------
Valid Range: 0, 1
Default Value: 1
0 = Disables MQ support
1 = Enables MQ support, which is a prerequisite for RSS
Direct Cache Access (DCA)
-------------------------
Valid Range: 0, 1
Default Value: 1 (when IXGBE_DCA is enabled)
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 fo
the feature.
IntMode
-------
Valid Range: 0-2 (0 = Legacy Int, 1 = MSI and 2 = MSI-X)
Default Value: 2
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: 956-488281 (0=off, 1=dynamic)
Default Value: 1
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-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.
Low Latency Interrupts (LLI)
----------------------------
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. MSI or MSI-X
interrupt modes must be used to successfully use LLI. (To verify MSI or MSI-X
is being used, see cat /proc/interrupts.)
LLIPort
-------
Valid Range: 0-65535
Default Value: 0 (disabled)
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 hardware 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.
LLIPush
-------
Valid Range: 0-1
Default Value: 0 (disabled)
LLIPush can be set to 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.
LLISize
-------
Valid Range: 0-1500
Default Value: 0 (disabled)
LLISize causes an immediate interrupt if the hardware receives a packet
smaller than the specified size.
LLIEType
--------
Valid Range: 0-0x8FFF
Default Value: 0 (disabled)
This parameter specifies the LLI Ethernet protocol type.
LLIVLANP
--------
Valid Range: 0-7
Default Value: 0 (disabled)
This parameter specifies the LLI on VLAN priority threshold.
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. If you want to disable a flow control capable link partner, use
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.
Intel® Ethernet Flow Director
-----------------------------
NOTE: Flow director parameters are only supported on kernel versions 2.6.30 or
later.
This parameter supports advanced filters that direct receive packets by their
flows to different queues. It enables tight control on routing a flow in the
platform. It matches flows and CPU cores for flow affinity. It also supports
multiple parameters for flexible flow classification and load balancing.
Flow director is enabled only if the kernel is multiple TX queue capable.
An included script (set_irq_affinity) automates setting the IRQ to CPU
affinity.
You can verify that the driver is using Flow Director by looking at the
counter in ethtool: fdir_miss and fdir_match.
Other ethtool Commands:
- To enable Flow Director: ethtool -K ethX ntuple on
- To add a filter use -U switch: ethtool -U ethX flow-type tcp4 src-ip
192.168.0.100 action 1
- To see the list of filters currently present: ethtool -u 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 hasing 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
Default Value: 1 (64k)
This specifies the Flow Director allocated packet buffer size.
1 = 64k
2 = 128k
3 = 256k
AtrSampleRate
-------------
Valid Range: 0-255
Default Value: 20
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
Default Value: 0
If the value is greater than 0, it will also force the VMDq parameter to 1 or
more.
This parameter adds support for SR-IOV. It causes the driver to spawn up to
max_vfs worth of virtual function.
NOTE: When either SR-IOV mode or VMDq mode is enabled, hardware VLAN filtering
and VLAN tag stripping/insertion will remain enabled. An old VLAN filter
should be removed before 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
The parameters for the driver are referenced by position. Thus, if you have a
dual port 82599- or X540-based adapter 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=63,63
NOTE: If both 82598- and 82599- or X540-based adapters are installed on the
same machine, caution must be used in loading the driver with the parameters.
Depending on system configuration, number of slots, etc., it is impossible to
predict in all cases where the positions would be on the command line and any
82598 ports needs to be specified by a zero in their positions.
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.
L2LBen
------
Valid Range: 0 (disabled), 1 (enabled)
Default Value: 1 (enabled)
This parameter controls the internal switch (L2 loopback between pf and vf).
By default the switch is enabled.
Large Receive Offload (LRO)
---------------------------
Valid Range: 0 (disabled), 1 (enabled)
Default Value: 1 (on)
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
remove support for LRO from the driver. The flag is used by adding
CFLAGS_EXTRA="-DIXGBE_NO_LRO" to the make file when it is 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_flushed. This is the total number of receives using LRO.
- lro_aggregated. This counts the total number of Ethernet packets that were
combined.
NOTE: IPv6 and UDP are not supported by LRO.
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 Linux Base Driver for the 10 Gigabit Family of Adapters is ixgbe.
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. |
The driver supports Jumbo Frames for all adapters. Jumbo Frames support is enabled by changing the MTU to a value larger than the default of 1500. The maximum value for the MTU is 9710. Use the ifconfig command to increase the MTU size. For example: ifconfig ethX mtu 9000 up
The maximum MTU setting for Jumbo Frames is 9710. This value coincides with the maximum Jumbo Frames size of 9728. This driver will attempt to use multiple page sized buffers to receive each jumbo packet. This should help 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.
The driver utilizes the ethtool interface for driver configuration and diagnostics, as well as displaying statistical information. ethtool version 3 or later is required for this functionality, although we strongly recommend downloading the latest version at:
82599 and X540-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-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.
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 sends 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
|
Note that this feature can be disabled for a specific Virtual Function (VF). |
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.
Some adapters do not support Wake on LAN. To determine if your adapter supports Wake on LAN, run ethtool ethX
Precision Time Protocol (PTP) is used to synchronize clocks in a computer network and is supported in the ixgbe driver.
The IXGBE_PTP is a compile time flag. The user may enable PTP at compile time by adding CFLAGS_EXTRA="-DIXGBE_PTP" to the make file when it is being compiled. make CFLAGS_EXTRA="-DIXGBE_PTP" install
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.
|
Note
|
After installing the driver, if your Intel Ethernet Network Connection is not working, verify that you have installed the correct driver. |
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 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
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 /etc/sysctl.conf file: net.ipv4.conf.all.arp_filter = 1 |
or, install the interfaces in separate broadcast domains (either in different switches or in a switch partitioned to VLANs)
Under small packets UDP stress test with 10GbE driver, the Linux system may drop UDP packets due to the fullness of socket buffers. The driver Flow Control variables may be changed to the minimum value for controlling packet reception, or the kernel’s default buffer sizes can be increased for UDP by changing the values in
/proc/sys/net/core/rmem_default and rmem_max
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.
"order:0" errors may occur under stress with kernels 2.6.25 and above. This is caused by the way the Linux kernel reports this stressed condition.
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-based network connections support up to 64 queues.
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.
Some PCIe x8 slots are actually configured as x4 slots. These slots have insufficient bandwidth for full 10Gbe line rate with dual port and quad port 10GbE 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 x8 slot will resolve the issue.
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 panick 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.
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.
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 drivers with CFLAGS_EXTRA=-DDISABLE_PCI_MSI if the drivers need 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 or 64-bit Microsoft* Windows* Server 2008/R2 Guest OS using Intel® Intel® 82599- or X540-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 82599 or X540-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 later kernel works fine, there is a known issue with Microsoft Windows Server 2008/R2 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, but rather that KVM emulates an older CPU model for the guests. 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 82599- or X540-based controllers in SR-IOV mode with KVM and a Microsoft Windows Server 2008/R2 guest, try the following workaround. Configure the KVM to emulate a different model of CPU when using qemu to create the KVM guest: "-cpu qemu64,model=13"
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. |
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 ixgbe PF driver, shutdown all VMs to ensure VFs are no longer active and unloading the ixgbevf driver first.
SR-IOV Virtual Functions are unable to send or receive traffic between VMs using emulated connections on a Linux Software bridge and connections that use SR-IOV VFs.
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 the supported kernel with a supported adapter, email the specific information related to the issue to e1000-devel@lists.sf.net.
Intel Gigabit Linux driver. Copyright(c) 1999 - 2013 Intel Corporation.
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".
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.