ixgbe Linux* Base Driver for Intel® Ethernet Network Connection
===============================================================


================================================================================

February 04, 2015

================================================================================


Contents
--------

- 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, X540, X550, 
and X552-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.


================================================================================


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>/kernel/drivers/net/ixgbe/ixgbe.[k]o

   The install location listed above is the default location. This may differ
   for various Linux distributions.

5. 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

6. Assign an IP address to the interface by entering the following, where x
   is the interface number:
   ifconfig eth<x> <IP_address> <netmask>

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>


To build ixgbe driver with DCA
------------------------------
If your kernel supports DCA, the driver will build by default with DCA enabled.

Note: DCA is not supported on X550-based adapters.


================================================================================


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 = 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.

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 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)

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, 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.


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.)

Note: LLI is not supported on X550-based adapters.


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.

Note: LLI is not supported on X550-based adapters.


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.

Note: LLI is not supported on X550-based adapters.


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.

Note: LLI is not supported on X550-based adapters.


LLIEType
--------
Valid Range:    0-0x8FFF

Default Value:  0 (disabled)

This parameter specifies the LLI Ethernet protocol type.

Note: LLI is not supported on X550-based adapters.


LLIVLANP
--------
Valid Range:    0-7

Default Value:  0 (disabled)

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. 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(R) Ethernet Flow Director
-------------------------------
NOTE: Flow director parameters are only supported on kernel versions 2.6.30 or
newer.

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 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

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, X540, or X550-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 an 82598-based adapter is installed in the same system as an 82599,
X540, or X550-based adapter, 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.


DMAC
----
Valid Range:    0, 41-10000

Default Value:  0 (disabled)	

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.


================================================================================


Additional 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 Linux Base Driver for the 10 Gigabit
Family of Adapters is ixgbe.


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
------------
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.


ethtool
-------
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:

http://ftp.kernel.org/pub/software/network/ethtool/.


Hardware Receive Side Coalescing (HW RSC)
-----------------------------------------
82599, 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,
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 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).


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.


IEEE 1588 PTP
-------------
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


VXLAN Overlay HW Offloading
---------------------------
VXLAN Overlay HW Offloading is enabled by default. The ixgbe Linux driver
features VXLAN Overlay HW Offloading support. To view and configure
VXLAN on a VXLAN-overlay offload enabled device,  use the following
commands:

  # ethtool -k ethX
    (This command displays the offloads and their current state.)
  # ethtool -K ethX tx-udp_tnl-segmentation [off|on]
    (This enables/disables VXLAN support in the driver.)

For more information on configuring your network for VXLAN overlay 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
----------------------------


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.


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 /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)


UDP Stress Test Dropped Packet Issue
------------------------------------
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


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
-------------------------
"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, 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 on dual port and quad port 10GbE devices
------------------------------------------------------------------------
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.


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 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 newer 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"


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 ixgbe PF driver, shutdown all VMs to ensure VFs are no longer
active and unloading the ixgbevf driver first.


Software bridging does not work with SR-IOV Virtual Functions
-------------------------------------------------------------
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.


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 the supported
kernel with a supported adapter, email the specific information related to the
issue to [email protected].


================================================================================


License
-------
Intel 10 Gigabit Linux driver.
Copyright(c) 1999 - 2015 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.