1 .\" Copyright (c) 2002 Luigi Rizzo
2 .\" All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
9 .\" 2. Redistributions in binary form must reproduce the above copyright
10 .\" notice, this list of conditions and the following disclaimer in the
11 .\" documentation and/or other materials provided with the distribution.
13 .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
14 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
16 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
17 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
18 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
19 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
20 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
21 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
22 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25 .\" $FreeBSD: src/share/man/man4/polling.4,v 1.27 2007/04/06 14:25:14 brueffer Exp $
26 .\" $DragonFly: src/share/man/man4/polling.4,v 1.13 2007/11/03 07:35:52 swildner Exp $
33 .Nd network device driver polling support
35 .Cd "options IFPOLL_ENABLE"
37 Network device polling
39 for brevity) refers to a technique that
40 lets the operating system periodically poll network devices, instead of
41 relying on the network devices to generate interrupts when they need attention.
42 This might seem inefficient and counterintuitive, but when done
45 gives more control to the operating system on
46 when and how to handle network devices, with a number of advantages in terms
47 of system responsiveness and performance.
51 reduces the overhead for context
52 switches which is incurred when servicing interrupts, and
53 gives more control on the scheduling of a CPU between various
54 tasks (user processes, software interrupts, device handling)
55 which ultimately reduces the chances of livelock in the system.
56 .Ss Principles of Operation
57 In the normal, interrupt-based mode, network devices generate an interrupt
58 whenever they need attention.
60 context switch and the execution of an interrupt handler
61 which performs whatever processing is needed by the network device.
62 The duration of the interrupt handler is potentially unbounded
63 unless the network device driver has been programmed with real-time
64 concerns in mind (which is generally not the case for
67 Furthermore, under heavy traffic load, the system might be
68 persistently processing interrupts without being able to
69 complete other work, either in the kernel or in userland.
71 Network device polling disables interrupts by polling network devices on
73 This way, the context switch overhead is removed.
75 the operating system can control accurately how much work to spend
76 in handling network device events, and thus prevent livelock by reserving
77 some amount of CPU to other tasks.
81 also changes the way software network interrupts
82 are scheduled, so there is never the risk of livelock because
83 packets are not processed to completion.
85 It is turned on and off with help of
88 An interface does not have to be
90 in order to turn on its
94 The following tunables can be set from
98 .Bl -tag -width indent -compact
99 .It Va net.ifpoll.burst_max
101 .Va net.ifpoll.X.rx.burst_max
104 .It Va net.ifpoll.each_burst
106 .Va net.ifpoll.X.rx.each_burst
109 .It Va net.ifpoll.user_frac
111 .Va net.ifpoll.X.rx.user_frac
114 .It Va net.ifpoll.pollhz
116 .Va net.ifpoll.X.pollhz
119 .It Va net.ifpoll.status_frac
121 .Va net.ifpoll.0.status_frac
124 .It Va net.ifpoll.tx_frac
126 .Va net.ifpoll.X.tx_frac
132 is controlled by the following per CPU
138 .Bl -tag -width indent -compact
139 .It Va net.ifpoll.X.pollhz
140 The polling frequency, whose range is 1 to 30000.
143 .It Va net.ifpoll.X.rx.user_frac
146 is enabled, and provided that there is some work to do,
147 up to this percent of the CPU cycles is reserved to userland tasks,
148 the remaining fraction being available for
153 .It Va net.ifpoll.X.rx.burst
154 Maximum number of packets grabbed from each network interface in
156 This number is dynamically adjusted by the kernel,
157 according to the programmed
158 .Va user_frac , burst_max ,
159 CPU speed, and system load.
161 .It Va net.ifpoll.X.rx.each_burst
162 The burst above is split into smaller chunks of this number of
163 packets, going round-robin among all interfaces registered for
165 This prevents the case that a large burst from a single interface
166 can saturate the IP interrupt queue.
169 .It Va net.ifpoll.X.rx.burst_max
171 .Va net.ifpoll.X.rx.burst .
174 is enabled, each interface can receive at most
175 .Pq Va pollhz No * Va burst_max
176 packets per second unless there are spare CPU cycles available for
179 This number should be tuned to match the expected load.
180 Default is 250 which is adequate for 1000Mbit network and pollhz=6000.
182 .It Va net.ifpoll.X.rx.handlers
183 How many active network devices have registered for packet reception
186 .It Va net.ifpoll.X.tx_frac
187 Controls how often (every
188 .Va tx_frac No / Va pollhz
189 seconds) the tranmission queue is checked for packet transmission
191 Increasing this value reduces the time spent on checking packets
192 transmission done events thus reduces bus load,
193 but it also increases chance
194 that the transmission queue getting saturated.
197 .It Va net.ifpoll.X.tx.handlers
198 How many active network devices have registered for packet transmission
201 .It Va net.ifpoll.0.status_frac
202 Controls how often (every
203 .Va status_frac No / Va pollhz
204 seconds) the status registers of the network device are checked for error
205 conditions and the like.
206 Increasing this value reduces the load on the bus,
207 but also delays the error detection.
210 .It Va net.ifpoll.0.status.handlers
211 How many active network devices have registered for status
214 .It Va net.ifpoll.X.rx.short_ticks
215 .It Va net.ifpoll.X.rx.lost_polls
216 .It Va net.ifpoll.X.rx.pending_polls
217 .It Va net.ifpoll.X.rx.residual_burst
218 .It Va net.ifpoll.X.rx.phase
219 .It Va net.ifpoll.X.rx.suspect
220 .It Va net.ifpoll.X.rx.stalled
221 .It Va net.ifpoll.X.tx.short_ticks
222 .It Va net.ifpoll.X.tx.lost_polls
223 .It Va net.ifpoll.X.tx.pending_polls
224 .It Va net.ifpoll.X.tx.residual_burst
225 .It Va net.ifpoll.X.tx.phase
226 .It Va net.ifpoll.X.tx.suspect
227 .It Va net.ifpoll.X.tx.stalled
230 .Sh SUPPORTED DEVICES
231 Network device polling requires explicit modifications to
232 the network device drivers.
233 As of this writing, the
254 devices are supported,
255 with others in the works.
263 support multiple reception queues based
272 support multiple transmission queues based
274 The modifications are rather straightforward, consisting in
275 the extraction of the inner part of the interrupt service routine
276 and writing a callback function,
279 to probe the network device for events and process them.
281 conditionally compiled sections of the network devices mentioned above
284 In order to reduce the latency in processing packets,
285 it is advisable to set the
288 .Va net.ifpoll.X.pollhz
291 Network device polling first appeared in
297 The network device polling code was rewritten by
299 based on the original code by
300 .An Luigi Rizzo Aq luigi@iet.unipi.it .
302 made the polling frequency settable at runtime,
303 added per CPU polling
304 and added multiple reception and tranmission queue polling support.