1GETRUSAGE(2) Linux Programmer's Manual GETRUSAGE(2)
2
6 getrusage - get resource usage
7
9 #include <sys/resource.h>
10 int getrusage(int who, struct rusage *usage);
12
14 getrusage() returns resource usage measures for who, which can be one
15 of the following:
16 RUSAGE_SELF
18 Return resource usage statistics for the calling process, which
19 is the sum of resources used by all threads in the process.
20 RUSAGE_CHILDREN
22 Return resource usage statistics for all children of the calling
23 process that have terminated and been waited for. These statis‐
24 tics will include the resources used by grandchildren, and fur‐
25 ther removed descendants, if all of the intervening descendants
26 waited on their terminated children.
27 RUSAGE_THREAD (since Linux 2.6.26)
29 Return resource usage statistics for the calling thread. The
30 _GNU_SOURCE feature test macro must be defined (before including
31 any header file) in order to obtain the definition of this con‐
32 stant from <sys/resource.h>.
33 The resource usages are returned in the structure pointed to by usage,
35 which has the following form:
36 struct rusage {
38 struct timeval ru_utime; /* user CPU time used */
39 struct timeval ru_stime; /* system CPU time used */
40 long ru_maxrss; /* maximum resident set size */
41 long ru_ixrss; /* integral shared memory size */
42 long ru_idrss; /* integral unshared data size */
43 long ru_isrss; /* integral unshared stack size */
44 long ru_minflt; /* page reclaims (soft page faults) */
45 long ru_majflt; /* page faults (hard page faults) */
46 long ru_nswap; /* swaps */
47 long ru_inblock; /* block input operations */
48 long ru_oublock; /* block output operations */
49 long ru_msgsnd; /* IPC messages sent */
50 long ru_msgrcv; /* IPC messages received */
51 long ru_nsignals; /* signals received */
52 long ru_nvcsw; /* voluntary context switches */
53 long ru_nivcsw; /* involuntary context switches */
54 };
55 Not all fields are completed; unmaintained fields are set to zero by
57 the kernel. (The unmaintained fields are provided for compatibility
58 with other systems, and because they may one day be supported on
59 Linux.) The fields are interpreted as follows:
60 ru_utime
62 This is the total amount of time spent executing in user mode,
63 expressed in a timeval structure (seconds plus microseconds).
64 ru_stime
66 This is the total amount of time spent executing in kernel mode,
67 expressed in a timeval structure (seconds plus microseconds).
68 ru_maxrss (since Linux 2.6.32)
70 This is the maximum resident set size used (in kilobytes). For
71 RUSAGE_CHILDREN, this is the resident set size of the largest
72 child, not the maximum resident set size of the process tree.
73 ru_ixrss (unmaintained)
75 This field is currently unused on Linux.
76 ru_idrss (unmaintained)
78 This field is currently unused on Linux.
79 ru_isrss (unmaintained)
81 This field is currently unused on Linux.
82 ru_minflt
84 The number of page faults serviced without any I/O activity;
85 here I/O activity is avoided by “reclaiming” a page frame from
86 the list of pages awaiting reallocation.
87 ru_majflt
89 The number of page faults serviced that required I/O activity.
90 ru_nswap (unmaintained)
92 This field is currently unused on Linux.
93 ru_inblock (since Linux 2.6.22)
95 The number of times the filesystem had to perform input.
96 ru_oublock (since Linux 2.6.22)
98 The number of times the filesystem had to perform output.
99 ru_msgsnd (unmaintained)
101 This field is currently unused on Linux.
102 ru_msgrcv (unmaintained)
104 This field is currently unused on Linux.
105 ru_nsignals (unmaintained)
107 This field is currently unused on Linux.
108 ru_nvcsw (since Linux 2.6)
110 The number of times a context switch resulted due to a process
111 voluntarily giving up the processor before its time slice was
112 completed (usually to await availability of a resource).
113 ru_nivcsw (since Linux 2.6)
115 The number of times a context switch resulted due to a higher
116 priority process becoming runnable or because the current
117 process exceeded its time slice.
118
120 On success, zero is returned. On error, -1 is returned, and errno is
121 set to indicate the error.
122
124 EFAULT usage points outside the accessible address space.
125 EINVAL who is invalid.
127
129 For an explanation of the terms used in this section, see at‐
130 tributes(7).
131 ┌────────────────────────────────────────────┬───────────────┬─────────┐
133 │Interface │ Attribute │ Value │
134 ├────────────────────────────────────────────┼───────────────┼─────────┤
135 │getrusage() │ Thread safety │ MT-Safe │
136 └────────────────────────────────────────────┴───────────────┴─────────┘
137
139 POSIX.1-2001, POSIX.1-2008, SVr4, 4.3BSD. POSIX.1 specifies
140 getrusage(), but specifies only the fields ru_utime and ru_stime.
141 RUSAGE_THREAD is Linux-specific.
143
145 Resource usage metrics are preserved across an execve(2).
146 In Linux kernel versions before 2.6.9, if the disposition of SIGCHLD is
148 set to SIG_IGN then the resource usages of child processes are automat‐
149 ically included in the value returned by RUSAGE_CHILDREN, although
150 POSIX.1-2001 explicitly prohibits this. This nonconformance is recti‐
151 fied in Linux 2.6.9 and later.
152 The structure definition shown at the start of this page was taken from
154 4.3BSD Reno.
155 Ancient systems provided a vtimes() function with a similar purpose to
157 getrusage(). For backward compatibility, glibc (up until version 2.32)
158 also provides vtimes(). All new applications should be written using
159 getrusage(). (Since version 2.33, glibc no longer provides an vtimes()
160 implementation.)
161 See also the description of /proc/[pid]/stat in proc(5).
163
165 clock_gettime(2), getrlimit(2), times(2), wait(2), wait4(2), clock(3)
166
168 This page is part of release 5.13 of the Linux man-pages project. A
169 description of the project, information about reporting bugs, and the
170 latest version of this page, can be found at
171 https://www.kernel.org/doc/man-pages/.
172Linux 2021-03-22 GETRUSAGE(2)