man-pages/man2/getitimer.2

.\" Copyright 7/93 by Darren Senn <sinster@scintilla.santa-clara.ca.us>
.\" and Copyright (C) 2016, Michael Kerrisk <mtk.manpages@gmail.com>
.\" Based on a similar page Copyright 1992 by Rick Faith
.\"
.\" %%%LICENSE_START(FREELY_REDISTRIBUTABLE)
.\" May be freely distributed and modified
.\" %%%LICENSE_END
.\"
.\" Modified Tue Oct 22 00:22:35 EDT 1996 by Eric S. Raymond <esr@thyrsus.com>
.\" 2005-04-06 mtk, Matthias Lang <matthias@corelatus.se>
.\" 	Noted MAX_SEC_IN_JIFFIES ceiling
.\"
.TH GETITIMER 2 2021-03-22 "Linux" "Linux Programmer's Manual"
.SH NAME
getitimer, setitimer \- get or set value of an interval timer
.SH SYNOPSIS
.nf
.B #include <sys/time.h>
.PP
.BI "int getitimer(int " which ", struct itimerval *" curr_value );
.BI "int setitimer(int " which ", const struct itimerval *restrict " new_value ,
.BI "              struct itimerval *restrict " old_value );
.fi
.SH DESCRIPTION
These system calls provide access to interval timers, that is,
timers that initially expire at some point in the future,
and (optionally) at regular intervals after that.
When a timer expires, a signal is generated for the calling process,
and the timer is reset to the specified interval
(if the interval is nonzero).
.PP
Three types of timers\(emspecified via the
.IR which
argument\(emare provided,
each of which counts against a different clock and
generates a different signal on timer expiration:
.TP
.B ITIMER_REAL
This timer counts down in real (i.e., wall clock) time.
At each expiration, a
.B SIGALRM
signal is generated.
.TP
.B ITIMER_VIRTUAL
This timer counts down against the user-mode CPU time consumed by the process.
(The measurement includes CPU time consumed by all threads in the process.)
At each expiration, a
.B SIGVTALRM
signal is generated.
.TP
.B ITIMER_PROF
This timer counts down against the total (i.e., both user and system)
CPU time consumed by the process.
(The measurement includes CPU time consumed by all threads in the process.)
At each expiration, a
.B SIGPROF
signal is generated.
.IP
In conjunction with
.BR ITIMER_VIRTUAL ,
this timer can be used to profile user and system CPU time
consumed by the process.
.PP
A process has only one of each of the three types of timers.
.PP
Timer values are defined by the following structures:
.PP
.in +4n
.EX
struct itimerval {
    struct timeval it_interval; /* Interval for periodic timer */
    struct timeval it_value;    /* Time until next expiration */
};

struct timeval {
    time_t      tv_sec;         /* seconds */
    suseconds_t tv_usec;        /* microseconds */
};
.EE
.in
.\"
.SS getitimer()
The function
.BR getitimer ()
places the current value of the timer specified by
.IR which
in the buffer pointed to by
.IR curr_value .
.PP
The
.IR it_value
substructure is populated with the amount of time remaining until
the next expiration of the specified timer.
This value changes as the timer counts down, and will be reset to
.IR it_interval
when the timer expires.
If both fields of
.IR it_value
are zero, then this timer is currently disarmed (inactive).
.PP
The
.IR it_interval
substructure is populated with the timer interval.
If both fields of
.IR it_interval
are zero, then this is a single-shot timer (i.e., it expires just once).
.SS setitimer()
The function
.BR setitimer ()
arms or disarms the timer specified by
.IR which ,
by setting the timer to the value specified by
.IR new_value .
If
.I old_value
is non-NULL,
the buffer it points to is used to return the previous value of the timer
(i.e., the same information that is returned by
.BR getitimer ()).
.PP
If either field in
.IR new_value.it_value
is nonzero,
then the timer is armed to initially expire at the specified time.
If both fields in
.IR new_value.it_value
are zero, then the timer is disarmed.
.PP
The
.IR new_value.it_interval
field specifies the new interval for the timer;
if both of its subfields are zero, the timer is single-shot.
.SH RETURN VALUE
On success, zero is returned.
On error, \-1 is returned, and
.I errno
is set to indicate the error.
.SH ERRORS
.TP
.B EFAULT
.IR new_value ,
.IR old_value ,
or
.I curr_value
is not valid a pointer.
.TP
.B EINVAL
.I which
is not one of
.BR ITIMER_REAL ,
.BR ITIMER_VIRTUAL ,
or
.BR ITIMER_PROF ;
or (since Linux 2.6.22) one of the
.I tv_usec
fields in the structure pointed to by
.I new_value
contains a value outside the range 0 to 999999.
.SH CONFORMING TO
POSIX.1-2001, SVr4, 4.4BSD (this call first appeared in 4.2BSD).
POSIX.1-2008 marks
.BR getitimer ()
and
.BR setitimer ()
obsolete, recommending the use of the POSIX timers API
.RB ( timer_gettime (2),
.BR timer_settime (2),
etc.) instead.
.SH NOTES
Timers will never expire before the requested time,
but may expire some (short) time afterward, which depends
on the system timer resolution and on the system load; see
.BR time (7).
(But see BUGS below.)
If the timer expires while the process is active (always true for
.BR ITIMER_VIRTUAL ),
the signal will be delivered immediately when generated.
.PP
A child created via
.BR fork (2)
does not inherit its parent's interval timers.
Interval timers are preserved across an
.BR execve (2).
.PP
POSIX.1 leaves the
interaction between
.BR setitimer ()
and the three interfaces
.BR alarm (2),
.BR sleep (3),
and
.BR usleep (3)
unspecified.
.PP
The standards are silent on the meaning of the call:
.PP
    setitimer(which, NULL, &old_value);
.PP
Many systems (Solaris, the BSDs, and perhaps others)
treat this as equivalent to:
.PP
    getitimer(which, &old_value);
.PP
In Linux, this is treated as being equivalent to a call in which the
.I new_value
fields are zero; that is, the timer is disabled.
.IR "Don't use this Linux misfeature" :
it is nonportable and unnecessary.
.SH BUGS
The generation and delivery of a signal are distinct, and
only one instance of each of the signals listed above may be pending
for a process.
Under very heavy loading, an
.B ITIMER_REAL
timer may expire before the signal from a previous expiration
has been delivered.
The second signal in such an event will be lost.
.PP
On Linux kernels before 2.6.16, timer values are represented in jiffies.
If a request is made set a timer with a value whose jiffies
representation exceeds
.B MAX_SEC_IN_JIFFIES
(defined in
.IR include/linux/jiffies.h ),
then the timer is silently truncated to this ceiling value.
On Linux/i386 (where, since Linux 2.6.13,
the default jiffy is 0.004 seconds),
this means that the ceiling value for a timer is
approximately 99.42 days.
Since Linux 2.6.16,
the kernel uses a different internal representation for times,
and this ceiling is removed.
.PP
On certain systems (including i386),
Linux kernels before version 2.6.12 have a bug which will produce
premature timer expirations of up to one jiffy under some circumstances.
This bug is fixed in kernel 2.6.12.
.\" 4 Jul 2005: It looks like this bug may remain in 2.4.x.
.\"	http://lkml.org/lkml/2005/7/1/165
.PP
POSIX.1-2001 says that
.BR setitimer ()
should fail if a
.I tv_usec
value is specified that is outside of the range 0 to 999999.
However, in kernels up to and including 2.6.21,
Linux does not give an error, but instead silently
adjusts the corresponding seconds value for the timer.
From kernel 2.6.22 onward,
this nonconformance has been repaired:
an improper
.I tv_usec
value results in an
.B EINVAL
error.
.\" Bugzilla report 25 Apr 2006:
.\" http://bugzilla.kernel.org/show_bug.cgi?id=6443
.\" "setitimer() should reject noncanonical arguments"
.SH SEE ALSO
.BR gettimeofday (2),
.BR sigaction (2),
.BR signal (2),
.BR timer_create (2),
.BR timerfd_create (2),
.BR time (7)