cb4e3454bf
multiqueue, and correct the rxdone handling. Update the polling man page to include igb as well. Thanks to Mark Johnston for these changes.
223 lines
6.9 KiB
Groff
223 lines
6.9 KiB
Groff
.\" Copyright (c) 2002 Luigi Rizzo
|
|
.\" 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.
|
|
.\"
|
|
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
|
|
.\"
|
|
.\" $FreeBSD$
|
|
.\"
|
|
.Dd April 6, 2007
|
|
.Dt POLLING 4
|
|
.Os
|
|
.Sh NAME
|
|
.Nm polling
|
|
.Nd device polling support
|
|
.Sh SYNOPSIS
|
|
.Cd "options DEVICE_POLLING"
|
|
.Sh DESCRIPTION
|
|
Device polling
|
|
.Nm (
|
|
for brevity) refers to a technique that
|
|
lets the operating system periodically poll devices, instead of
|
|
relying on the devices to generate interrupts when they need attention.
|
|
This might seem inefficient and counterintuitive, but when done
|
|
properly,
|
|
.Nm
|
|
gives more control to the operating system on
|
|
when and how to handle devices, with a number of advantages in terms
|
|
of system responsiveness and performance.
|
|
.Pp
|
|
In particular,
|
|
.Nm
|
|
reduces the overhead for context
|
|
switches which is incurred when servicing interrupts, and
|
|
gives more control on the scheduling of the CPU between various
|
|
tasks (user processes, software interrupts, device handling)
|
|
which ultimately reduces the chances of livelock in the system.
|
|
.Ss Principles of Operation
|
|
In the normal, interrupt-based mode, devices generate an interrupt
|
|
whenever they need attention.
|
|
This in turn causes a
|
|
context switch and the execution of an interrupt handler
|
|
which performs whatever processing is needed by the device.
|
|
The duration of the interrupt handler is potentially unbounded
|
|
unless the device driver has been programmed with real-time
|
|
concerns in mind (which is generally not the case for
|
|
.Fx
|
|
drivers).
|
|
Furthermore, under heavy traffic load, the system might be
|
|
persistently processing interrupts without being able to
|
|
complete other work, either in the kernel or in userland.
|
|
.Pp
|
|
Device polling disables interrupts by polling devices at appropriate
|
|
times, i.e., on clock interrupts and within the idle loop.
|
|
This way, the context switch overhead is removed.
|
|
Furthermore,
|
|
the operating system can control accurately how much work to spend
|
|
in handling device events, and thus prevent livelock by reserving
|
|
some amount of CPU to other tasks.
|
|
.Pp
|
|
Enabling
|
|
.Nm
|
|
also changes the way software network interrupts
|
|
are scheduled, so there is never the risk of livelock because
|
|
packets are not processed to completion.
|
|
.Ss Enabling polling
|
|
Currently only network interface drivers support the
|
|
.Nm
|
|
feature.
|
|
It is turned on and off with help of
|
|
.Xr ifconfig 8
|
|
command.
|
|
.Pp
|
|
The historic
|
|
.Va kern.polling.enable ,
|
|
which enabled polling for all interfaces, can be replaced with the following
|
|
code:
|
|
.Bd -literal
|
|
for i in `ifconfig -l` ;
|
|
do ifconfig $i polling; # use -polling to disable
|
|
done
|
|
.Ed
|
|
.Ss MIB Variables
|
|
The operation of
|
|
.Nm
|
|
is controlled by the following
|
|
.Xr sysctl 8
|
|
MIB variables:
|
|
.Pp
|
|
.Bl -tag -width indent -compact
|
|
.It Va kern.polling.user_frac
|
|
When
|
|
.Nm
|
|
is enabled, and provided that there is some work to do,
|
|
up to this percent of the CPU cycles is reserved to userland tasks,
|
|
the remaining fraction being available for
|
|
.Nm
|
|
processing.
|
|
Default is 50.
|
|
.Pp
|
|
.It Va kern.polling.burst
|
|
Maximum number of packets grabbed from each network interface in
|
|
each timer tick.
|
|
This number is dynamically adjusted by the kernel,
|
|
according to the programmed
|
|
.Va user_frac , burst_max ,
|
|
CPU speed, and system load.
|
|
.Pp
|
|
.It Va kern.polling.each_burst
|
|
The burst above is split into smaller chunks of this number of
|
|
packets, going round-robin among all interfaces registered for
|
|
.Nm .
|
|
This prevents the case that a large burst from a single interface
|
|
can saturate the IP interrupt queue
|
|
.Pq Va net.inet.ip.intr_queue_maxlen .
|
|
Default is 5.
|
|
.Pp
|
|
.It Va kern.polling.burst_max
|
|
Upper bound for
|
|
.Va kern.polling.burst .
|
|
Note that when
|
|
.Nm
|
|
is enabled, each interface can receive at most
|
|
.Pq Va HZ No * Va burst_max
|
|
packets per second unless there are spare CPU cycles available for
|
|
.Nm
|
|
in the idle loop.
|
|
This number should be tuned to match the expected load
|
|
(which can be quite high with GigE cards).
|
|
Default is 150 which is adequate for 100Mbit network and HZ=1000.
|
|
.Pp
|
|
.It Va kern.polling.idle_poll
|
|
Controls if
|
|
.Nm
|
|
is enabled in the idle loop.
|
|
There are no reasons (other than power saving or bugs in the scheduler's
|
|
handling of idle priority kernel threads) to disable this.
|
|
.Pp
|
|
.It Va kern.polling.reg_frac
|
|
Controls how often (every
|
|
.Va reg_frac No / Va HZ
|
|
seconds) the status registers of the device are checked for error
|
|
conditions and the like.
|
|
Increasing this value reduces the load on the bus, but also delays
|
|
the error detection.
|
|
Default is 20.
|
|
.Pp
|
|
.It Va kern.polling.handlers
|
|
How many active devices have registered for
|
|
.Nm .
|
|
.Pp
|
|
.It Va kern.polling.short_ticks
|
|
.It Va kern.polling.lost_polls
|
|
.It Va kern.polling.pending_polls
|
|
.It Va kern.polling.residual_burst
|
|
.It Va kern.polling.phase
|
|
.It Va kern.polling.suspect
|
|
.It Va kern.polling.stalled
|
|
Debugging variables.
|
|
.El
|
|
.Sh SUPPORTED DEVICES
|
|
Device polling requires explicit modifications to the device drivers.
|
|
As of this writing, the
|
|
.Xr bge 4 ,
|
|
.Xr dc 4 ,
|
|
.Xr em 4 ,
|
|
.Xr fwe 4 ,
|
|
.Xr fwip 4 ,
|
|
.Xr fxp 4 ,
|
|
.Xr igb 4 ,
|
|
.Xr ixgb 4 ,
|
|
.Xr nfe 4 ,
|
|
.Xr nge 4 ,
|
|
.Xr re 4 ,
|
|
.Xr rl 4 ,
|
|
.Xr sf 4 ,
|
|
.Xr sis 4 ,
|
|
.Xr ste 4 ,
|
|
.Xr stge 4 ,
|
|
.Xr vge 4 ,
|
|
.Xr vr 4 ,
|
|
and
|
|
.Xr xl 4
|
|
devices are supported, with others in the works.
|
|
The modifications are rather straightforward, consisting in
|
|
the extraction of the inner part of the interrupt service routine
|
|
and writing a callback function,
|
|
.Fn *_poll ,
|
|
which is invoked
|
|
to probe the device for events and process them.
|
|
(See the
|
|
conditionally compiled sections of the devices mentioned above
|
|
for more details.)
|
|
.Pp
|
|
As in the worst case the devices are only polled on clock interrupts,
|
|
in order to reduce the latency in processing packets, it is not advisable
|
|
to decrease the frequency of the clock below 1000 Hz.
|
|
.Sh HISTORY
|
|
Device polling first appeared in
|
|
.Fx 4.6
|
|
and
|
|
.Fx 5.0 .
|
|
.Sh AUTHORS
|
|
Device polling was written by
|
|
.An Luigi Rizzo Aq luigi@iet.unipi.it .
|