1 .\" This is _*_ nroff _*_ source. Emacs, gimme all those colors :)
3 .\" Copyright (c) International Business Machines orp., 2006
5 .\" This program is free software; you can redistribute it and/or
6 .\" modify it under the terms of the GNU General Public License as
7 .\" published by the Free Software Foundation; either version 2 of
8 .\" the License, or (at your option) any later version.
10 .\" This program is distributed in the hope that it will be useful,
11 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
12 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
13 .\" the GNU General Public License for more details.
15 .\" You should have received a copy of the GNU General Public License
16 .\" along with this program; if not, write to the Free Software
17 .\" Foundation, Inc., 59 Temple Place, Suite 330, Boston,
21 .\" 2006-04-27, created by Eduardo M. Fleury <efleury@br.ibm.com>
22 .\" with various additions by Michael Kerrisk <mtk.manpages@gmail.com>
25 .TH IOPRIO_SET 2 2012-07-13 "Linux" "Linux Programmer's Manual"
27 ioprio_get, ioprio_set \- get/set I/O scheduling class and priority
30 .BI "int ioprio_get(int " which ", int " who );
31 .BI "int ioprio_set(int " which ", int " who ", int " ioprio );
35 There are no glibc wrappers for these system calls; see NOTES.
41 system calls respectively get and set the I/O scheduling class and
42 priority of one or more processes.
48 arguments identify the process(es) on which the system
52 argument determines how
54 is interpreted, and has one of the following values:
58 is a process ID identifying a single process.
62 is a process group ID identifying all the members of a process group.
66 is a user ID identifying all of the processes that
67 have a matching real UID.
77 and more than one process matches
79 then the returned priority will be the highest one found among
80 all of the matching processes.
81 One priority is said to be
82 higher than another one if it belongs to a higher priority
85 is the highest priority class;
88 or if it belongs to the same priority class as the other process but
89 has a higher priority level (a lower priority number means a
90 higher priority level).
96 is a bit mask that specifies both the scheduling class and the
97 priority to be assigned to the target process(es).
98 The following macros are used for assembling and dissecting
102 .BI IOPRIO_PRIO_VALUE( class ", " data )
107 this macro combines the two values to produce an
109 value, which is returned as the result of the macro.
111 .BI IOPRIO_PRIO_CLASS( mask )
116 value), this macro returns its I/O class component, that is,
118 .BR IOPRIO_CLASS_RT ,
119 .BR IOPRIO_CLASS_BE ,
121 .BR IOPRIO_CLASS_IDLE .
123 .BI IOPRIO_PRIO_DATA( mask )
128 value), this macro returns its priority
132 See the NOTES section for more
133 information on scheduling classes and priorities.
135 I/O priorities are supported for reads and for synchronous
139 I/O priorities are not supported for asynchronous
140 writes because they are issued outside the context of the program
141 dirtying the memory, and thus program-specific priorities do not apply.
147 value of the process with highest I/O priority of any of the processes
148 that match the criteria specified in
152 On error, \-1 is returned, and
154 is set to indicate the error.
159 On error, \-1 is returned, and
161 is set to indicate the error.
169 Refer to the NOTES section for available scheduler
170 classes and priority levels for
174 The calling process does not have the privilege needed to assign this
176 to the specified process(es).
177 See the NOTES section for more information on required
182 No process(es) could be found that matched the specification in
187 These system calls have been available on Linux since
190 These system calls are Linux-specific.
192 Glibc does not provide wrapper for these system calls; call them using
195 These system calls only have an effect when used
196 in conjunction with an I/O scheduler that supports I/O priorities.
197 As at kernel 2.6.17 the only such scheduler is the Completely Fair Queuing
199 .SS "Selecting an I/O Scheduler"
200 I/O Schedulers are selected on a per-device basis via the special
202 .IR /sys/block/<device>/queue/scheduler .
204 One can view the current I/O scheduler via the
207 For example, the following command
208 displays a list of all schedulers currently loaded in the kernel:
212 .RB "$" " cat /sys/block/hda/queue/scheduler"
213 noop anticipatory deadline [cfq]
217 The scheduler surrounded by brackets is the one actually
218 in use for the device
221 Setting another scheduler is done by writing the name of the
222 new scheduler to this file.
223 For example, the following command will set the
233 .RB "#" " echo cfq > /sys/block/hda/queue/scheduler"
236 .SS "The Completely Fair Queuing (CFQ) I/O Scheduler"
237 Since v3 (aka CFQ Time Sliced) CFQ implements
238 I/O nice levels similar to those
240 These nice levels are grouped in three scheduling classes
241 each one containing one or more priority levels:
243 .BR IOPRIO_CLASS_RT " (1)"
244 This is the real-time I/O class.
245 This scheduling class is given
246 higher priority than any other class:
247 processes from this class are
248 given first access to the disk every time.
249 Thus this I/O class needs to be used with some
250 care: one I/O real-time process can starve the entire system.
251 Within the real-time class,
252 there are 8 levels of class data (priority) that determine exactly
253 how much time this process needs the disk for on each service.
254 The highest real-time priority level is 0; the lowest is 7.
255 In the future this might change to be more directly mappable to
256 performance, by passing in a desired data rate instead.
258 .BR IOPRIO_CLASS_BE " (2)"
259 This is the best-effort scheduling class,
260 which is the default for any process
261 that hasn't set a specific I/O priority.
262 The class data (priority) determines how much
263 I/O bandwidth the process will get.
264 Best-effort priority levels are analogous to CPU nice values
266 .BR getpriority (2)).
267 The priority level determines a priority relative
268 to other processes in the best-effort scheduling class.
269 Priority levels range from 0 (highest) to 7 (lowest).
271 .BR IOPRIO_CLASS_IDLE " (3)"
272 This is the idle scheduling class.
273 Processes running at this level only get I/O
274 time when no-one else needs the disk.
275 The idle class has no class data.
276 Attention is required when assigning this priority class to a process,
277 since it may become starved if higher priority processes are
278 constantly accessing the disk.
281 .I Documentation/block/ioprio.txt
282 for more information on the CFQ I/O Scheduler and an example program.
283 .SS "Required permissions to set I/O priorities"
284 Permission to change a process's priority is granted or denied based
287 .B "Process ownership"
288 An unprivileged process may only set the I/O priority of a process
290 matches the real or effective UID of the calling process.
291 A process which has the
293 capability can change the priority of any process.
295 .B "What is the desired priority"
296 Attempts to set very high priorities
297 .RB ( IOPRIO_CLASS_RT )
301 Kernel versions up to 2.6.24 also required
303 to set a very low priority
304 .RB ( IOPRIO_CLASS_IDLE ),
305 but since Linux 2.6.25, this is no longer required.
309 must follow both rules, or the call will fail with the error
312 .\" 6 May 07: Bug report raised:
313 .\" http://sources.redhat.com/bugzilla/show_bug.cgi?id=4464
314 .\" Ulrich Drepper replied that he wasn't going to add these
316 Glibc does not yet provide a suitable header file defining
317 the function prototypes and macros described on this page.
318 Suitable definitions can be found in
325 Documentation/block/ioprio.txt in the kernel source tree.