1GETPRIORITY(2) Linux Programmer's Manual GETPRIORITY(2)
2
6 getpriority, setpriority - get/set program scheduling priority
7
9 #include <sys/time.h>
10 #include <sys/resource.h>
11 int getpriority(int which, int who);
13 int setpriority(int which, int who, int prio);
14
16 The scheduling priority of the process, process group, or user, as
17 indicated by which and who is obtained with the getpriority() call and
18 set with the setpriority() call.
19 The value which is one of PRIO_PROCESS, PRIO_PGRP, or PRIO_USER, and
21 who is interpreted relative to which (a process identifier for
22 PRIO_PROCESS, process group identifier for PRIO_PGRP, and a user ID for
23 PRIO_USER). A zero value for who denotes (respectively) the calling
24 process, the process group of the calling process, or the real user ID
25 of the calling process. Prio is a value in the range -20 to 19 (but
26 see the Notes below). The default priority is 0; lower priorities
27 cause more favorable scheduling.
28 The getpriority() call returns the highest priority (lowest numerical
30 value) enjoyed by any of the specified processes. The setpriority()
31 call sets the priorities of all of the specified processes to the spec‐
32 ified value. Only the superuser may lower priorities.
33
35 Since getpriority() can legitimately return the value -1, it is neces‐
36 sary to clear the external variable errno prior to the call, then check
37 it afterward to determine if -1 is an error or a legitimate value. The
38 setpriority() call returns 0 if there is no error, or -1 if there is.
39
41 EINVAL which was not one of PRIO_PROCESS, PRIO_PGRP, or PRIO_USER.
42 ESRCH No process was located using the which and who values specified.
44 In addition to the errors indicated above, setpriority() may fail if:
46 EACCES The caller attempted to lower a process priority, but did not
48 have the required privilege (on Linux: did not have the
49 CAP_SYS_NICE capability). Since Linux 2.6.12, this error occurs
50 only if the caller attempts to set a process priority outside
51 the range of the RLIMIT_NICE soft resource limit of the target
52 process; see getrlimit(2) for details.
53 EPERM A process was located, but its effective user ID did not match
55 either the effective or the real user ID of the caller, and was
56 not privileged (on Linux: did not have the CAP_SYS_NICE capabil‐
57 ity). But see NOTES below.
58
60 SVr4, 4.4BSD (these function calls first appeared in 4.2BSD),
61 POSIX.1-2001.
62
64 A child created by fork(2) inherits its parent's nice value. The nice
65 value is preserved across execve(2).
66 The degree to which their relative nice value affects the scheduling of
68 processes varies across UNIX systems, and, on Linux, across kernel ver‐
69 sions. Starting with kernel 2.6.23, Linux adopted an algorithm that
70 causes relative differences in nice values to have a much stronger
71 effect. This causes very low nice values (+19) to truly provide little
72 CPU to a process whenever there is any other higher priority load on
73 the system, and makes high nice values (-20) deliver most of the CPU to
74 applications that require it (e.g., some audio applications).
75 The details on the condition for EPERM depend on the system. The above
77 description is what POSIX.1-2001 says, and seems to be followed on all
78 System V-like systems. Linux kernels before 2.6.12 required the real
79 or effective user ID of the caller to match the real user of the
80 process who (instead of its effective user ID). Linux 2.6.12 and later
81 require the effective user ID of the caller to match the real or effec‐
82 tive user ID of the process who. All BSD-like systems (SunOS 4.1.3,
83 Ultrix 4.2, 4.3BSD, FreeBSD 4.3, OpenBSD-2.5, ...) behave in the same
84 manner as Linux 2.6.12 and later.
85 The actual priority range varies between kernel versions. Linux before
87 1.3.36 had -infinity..15. Since kernel 1.3.43, Linux has the range
88 -20..19. Within the kernel, nice values are actually represented using
89 the corresponding range 40..1 (since negative numbers are error codes)
90 and these are the values employed by the setpriority() and getprior‐
91 ity() system calls. The glibc wrapper functions for these system calls
92 handle the translations between the user-land and kernel representa‐
93 tions of the nice value according to the formula unice = 20 - knice.
94 On some systems, the range of nice values is -20..20.
96 Including <sys/time.h> is not required these days, but increases porta‐
98 bility. (Indeed, <sys/resource.h> defines the rusage structure with
99 fields of type struct timeval defined in <sys/time.h>.)
100
102 According to POSIX, the nice value is a per-process setting. However,
103 under the current Linux/NPTL implementation of POSIX threads, the nice
104 value is a per-thread attribute: different threads in the same process
105 can have different nice values. Portable applications should avoid
106 relying on the Linux behavior, which may be made standards conformant
107 in the future.
108
110 nice(1), renice(1), fork(2), capabilities(7)
111 Documentation/scheduler/sched-nice-design.txt in the Linux kernel
113 source tree (since Linux 2.6.23)
114
116 This page is part of release 3.53 of the Linux man-pages project. A
117 description of the project, and information about reporting bugs, can
118 be found at http://www.kernel.org/doc/man-pages/.
119Linux 2013-02-12 GETPRIORITY(2)