man/man4/ix.4

.\" Copyright (c) 2001-2008, Intel Corporation
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions are met:
.\"
.\" 1. Redistributions of source code must retain the above copyright notice,
.\"    this list of conditions and the following disclaimer.
.\"
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\"
.\" 3. Neither the name of the Intel Corporation nor the names of its
.\"    contributors may be used to endorse or promote products derived from
.\"    this software without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
.\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
.\" LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
.\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
.\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
.\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
.\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
.\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
.\" POSSIBILITY OF SUCH DAMAGE.
.\"
.\" * Other names and brands may be claimed as the property of others.
.\"
.\" $FreeBSD: src/share/man/man4/ixgbe.4,v 1.2 2008/06/17 21:14:02 brueffer Exp $
.\"
.Dd May 11, 2014
.Dt IX 4
.Os
.Sh NAME
.Nm ix
.Nd "Intel(R) 10Gb Ethernet driver"
.Sh SYNOPSIS
To compile this driver into the kernel,
place the following line in your
kernel configuration file:
.Bd -ragged -offset indent
.Cd "device ix"
.Ed
.Pp
Alternatively, to load the driver as a
module at boot time, place the following line in
.Xr loader.conf 5 :
.Bd -literal -offset indent
if_ix_load="YES"
.Ed
.Sh DESCRIPTION
The
.Nm
driver provides support for PCI Express 10Gb Ethernet adapters based on
the Intel
82598,
82599,
and
X540
Ethernet controller chips.
The
.Nm
driver supports:
.Pp
.Bl -item -offset indent -compact
.It
Transmit/Receive checksum offload for IP/UDP/TCP.
.\"Jumbo Frames.
.It
Interrupt moderation
.It
TCP segmentation offload (TSO)
.It
Receive side scaling (RSS)
.It
Multiple tranmission queues
.It
Multiple vector MSI-X
.It
VLAN tag stripping and inserting
.El
.Pp
If
.Xr polling 4
or MSI-X is used,
by default,
the
.Nm
driver will try enabling as many reception queues and transmission queues
as are allowed by the number of CPUs in the system.
.Pp
If multiple transmission queues are used,
the round-robin arbitration is performed among the transmission queues.
And the round-robin arbitration between transmission queues is done at the
TCP segment boundary after the hardware segmentation is performed.
.Pp
82598 supports 16 reception queues and 32 transmission queues.
MSI-X is not enabled due to hardware errata.
Under MSI or legacy interrupt mode,
2 reception queues are enabled for hardware RSS hash
and only 1 transmission queue is enable.
.Pp
82599 and X540 supports 16 reception queues and 64 transmission queues.
MSI-X is enable by default.
However,
due to the number of MSI-X vectors is 64,
at most 16 reception queues and 32 transmission queues will be enabled
under MSI-X mode.
.Pp
For more information on configuring this device, see
.Xr ifconfig 8 .
The
.Nm
driver supports
.Xr polling 4 .
.Sh HARDWARE
The
.Nm
driver supports Gigabit Ethernet adapters based on the Intel
82598,
82599,
and
X540
controller chips:
.Pp
.Bl -bullet -compact
.It
Intel 10 Gigabit AF DA Dual Port Server Adapter
.It
Intel 10 Gigabit AT Server Adapter
.It
Intel 10 Gigabit AT2 Server Adapter
.It
Intel 10 Gigabit CX4 Dual Port Server Adapter
.It
Intel 10 Gigabit XF LR Server Adapter
.It
Intel 10 Gigabit XF SR Dual Port Server Adapter
.It
Intel 10 Gigabit XF SR Server Adapter
.It
Intel 82598 10 Gigabit Ethernet Controller
.It
Intel 82599 10 Gigabit Ethernet Controller
.It
Intel Ethernet Controller X540-AT2
.It
Intel Ethernet Converged Network Adapter X520 Series
.It
Intel Ethernet Converged Network Adapter X540-T1
.It
Intel Ethernet Converged Network Adapter X540-T2
.It
Intel Ethernet Server Adapter X520 Series
.It
Intel Ethernet Server Adapter X520-DA2
.It
Intel Ethernet Server Adapter X520-LR1
.It
Intel Ethernet Server Adapter X520-SR1
.It
Intel Ethernet Server Adapter X520-SR2
.It
Intel Ethernet Server Adapter X520-T2
.El
.Sh TUNABLES
Tunables can be set at the
.Xr loader 8
prompt before booting the kernel or stored in
.Xr loader.conf 5 .
.Em X
is the device unit number.
.Bl -tag -width ".Va hw.ixX.unsupported_sfp"
.It Va hw.ix.rxd Va hw.ixX.rxd
Number of receive descriptors allocated by the driver.
The default value is 2048.
The minimum is 64,
and the maximum is 4096.
.It Va hw.ix.txd Va hw.ixX.txd
Number of transmit descriptors allocated by the driver.
The default value is 2048.
The minimum is 64,
and the maximum is 4096.
.It Va hw.ix.rxr Va hw.ixX.rxr
This tunable specifies the number of reception queues could be enabled.
Maximum allowed value for these tunables is device specific
and it must be power of 2 aligned.
Setting these tunables to 0 allows the driver to make
as many reception queues ready-for-use as allowed by the number of CPUs.
.It Va hw.ix.txr Va hw.ixX.txr
This tunable specifies the number of transmission queues could be enabled.
Maximum allowed value for these tunables is device specific
and it must be power of 2 aligned.
Setting these tunables to 0 allows the driver to make
as many transmission queues ready-for-use as allowed by the number of CPUs.
.It Va hw.ix.msix.enable Va hw.ixX.msix.enable
By default,
the driver will use MSI-X if it is supported.
This behaviour can be turned off by setting this tunable to 0.
.It Va hw.ix.msix.agg_rxtx Va hw.ixX.msix.agg_rxtx
If MSI-X is used,
the driver aggregates transmission queue and reception queue processing
by default.
This behaviour could be turned off by setting this tunable to 0.
If the number of MSI-X vectors is not enough to
put transmission queue processing and reception queue processing
onto independent MSI-X vector,
then transmission queue and reception queue processing are always
aggregated.
.It Va hw.ixX.msix.off
If MSI-X is used,
and transmission queue and reception queue processing are aggregated,
this tunable specifies the leading target CPU for
transmission and reception queues processing.
The value specificed must be aligned to the maximum of
the number of reception queues
and the number of transmission queues enabled,
and must be less than the power of 2 number of CPUs.
.It Va hw.ixX.msix.rxoff
If MSI-X is used,
and transmission queue and reception queue processing are not aggregated,
this tunable specifies the leading target CPU for reception queues processing.
The value specificed must be aligned to the number of reception queues enabled
and must be less than the power of 2 number of CPUs.
.It Va hw.ixX.msix.txoff
If MSI-X is used,
and transmission queue and reception queue processing are not aggregated,
this tunable specifies the leading target CPU
for transmission queues processing.
The value specificed must be aligned to
the number of transmission queues enabled
and must be less than the power of 2 number of CPUs.
.It Va hw.ix.msi.enable Va hw.ixX.msi.enable
If MSI-X is disabled and MSI is supported,
the driver will use MSI.
This behavior can be turned off by setting this tunable to 0.
.It Va hw.ixX.msi.cpu
If MSI is used,
it specifies the MSI's target CPU.
.It Va hw.ixX.npoll.txoff
This tunable specifies the leading target CPU for
transmission queue
.Xr polling 4
processing.
The value specificed must be aligned to the number of transmission queues
enabled and must be less than the power of 2 number of CPUs.
.It Va hw.ixX.npoll.rxoff
This tunable specifies the leading target CPU for
reception queue
.Xr polling 4
processing.
The value specificed must be aligned to the number of reception queues
enabled and must be less than the power of 2 number of CPUs.
.It Va hw.ix.unsupported_sfp
By default,
this driver does not allow "unsupported" SFP modules.
This behavior can be changed by setting this tunable to 1.
.El
.Sh MIB Variables
A number of per-interface variables are implemented in the
.Va hw.ix Ns Em X
branch of the
.Xr sysctl 3
MIB.
.Bl -tag -width "rxtx_intr_rate"
.It Va rxr
Number of reception queues could be enabled (read-only).
Use the tunable
.Va hw.ix.rxr
or
.Va hw.ixX.rxr
to configure it.
.It Va rxr_inuse
Number of reception queues being used (read-only).
.It Va txr
Number of transmission queues could be enabled (read-only).
Use the tunable
.Va hw.ix.txr
or
.Va hw.ixX.txr
to configure it.
.It Va txr_inuse
Number of transmission queues being used (read-only).
.It Va rxd
Number of descriptors per reception queue (read-only).
Use the tunable
.Va hw.ix.rxd
or
.Va hw.ixX.rxd
to configure it.
.It Va txd
Number of descriptors per transmission queue (read-only).
Use the tunable
.Va hw.ix.txd
or
.Va hw.ixX.txd
to configure it.
.It Va rxtx_intr_rate
If MSI or legacy interrupt is used,
this sysctl controls the highest possible frequency
that interrupt could be generated by the device.
If MSI-X is used,
this sysctl controls the highest possible frequency
that interrupt could be generated by the MSI-X vectors,
which aggregate transmission queue and reception queue procecssing.
It is 8000 by default (125us).
.It Va rx_intr_rate
If MSI-X is used,
this sysctl controls the highest possible frequency
that interrupt could be generated by the MSI-X vectors,
which only process reception queue.
It is 8000 by default (125us).
.It Va tx_intr_rate
If MSI-X is used,
this sysctl controls the highest possible frequency
that interrupt could be generated by the MSI-X vectors,
which only process transmission queue.
It is 6000 by default (~150us).
.It Va sts_intr_rate
If MSI-X is used,
this sysctl controls the highest possible frequency
that interrupt could be generated by the MSI-X vectors,
which only process chip status changes.
It is 8000 by default (125us).
.It Va tx_intr_nsegs
Transmission interrupt is asked to be generated upon every
.Va tx_intr_nsegs
transmission descritors having been setup.
The default value is 1/16 of the number of transmission descriptors per queue.
.It Va tx_wreg_nsegs
The number of transmission descriptors should be setup
before the hardware register is written.
Setting this value too high will have negative effect
on transmission timeliness.
Setting this value too low will hurt overall transmission performance
due to the frequent hardware register writing.
The default value is 8.
.It Va rx_wreg_nsegs
The number of reception descriptors should be setup
before the hardware register is written.
Setting this value too high will make device drop incoming packets.
Setting this value too low will hurt overall reception performance
due to the frequent hardware register writing.
The default value is 32.
.It Va npoll_rxoff
See the tunable
.Va hw.ixX.npoll.rxoff .
The set value will take effect the next time
.Xr polling 4
is enabled on the device.
.It Va npoll_txoff
See the tunable
.Va hw.ixX.npoll.txoff .
The set value will take effect the next time
.Xr polling 4
is enabled on the device.
.It Va flowctrl
Flow control setting.
Set it to 0 to turn off flow control.
Set it to 1 to enable only the reception of pause frames.
Set it to 2 to enable only the generation of pause frames.
Set it to 3 to enable both the reception of pause frames
and generation of pause frames,
i.e. full flow control.
.El
.Sh SEE ALSO
.Xr altq 4 ,
.Xr arp 4 ,
.Xr ifmedia 4 ,
.Xr netintro 4 ,
.Xr ng_ether 4 ,
.Xr polling 4 ,
.Xr vlan 4 ,
.Xr ifconfig 8
.Sh HISTORY
The
.Nm
device driver first appeared in
.Dx 3.1 .
.Sh AUTHORS
The
.Nm
driver was written by
.An Intel Corporation Aq Mt freebsdnic@mailbox.intel.com .