.\" Copyright (C) 2008, Michael Kerrisk <mtk.manpages@gmail.com>
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.TH ACCT 5 2008-06-15 "Linux" "Linux Programmer's Manual"
.SH NAME
acct \- process accounting file
.SH SYNOPSIS
.B #include <sys/acct.h>
.SH DESCRIPTION
If the kernel is built with the process accounting option enabled
.RB ( CONFIG_BSD_PROCESS_ACCT ),
then calling
.BR acct (2)
starts process accounting, for example:

.in +4n
acct("/var/log/pacct");
.in

When process accounting is enabled, the kernel writes a record
to the accounting file as each process on the system terminates.
This record contains information about the terminated process,
and is defined in
.I <sys/acct.h>
as follows:

.in +4n
.nf
#define ACCT_COMM 16

typedef u_int16_t comp_t;

struct acct {
    char ac_flag;           /* Accounting flags */
    u_int16_t ac_uid;       /* Accounting user ID */
    u_int16_t ac_gid;       /* Accounting group ID */
    u_int16_t ac_tty;       /* Controlling terminal */
    u_int32_t ac_btime;     /* Process creation time
                               (seconds since the Epoch) */
    comp_t    ac_utime;     /* User CPU time */
    comp_t    ac_stime;     /* System CPU time */
    comp_t    ac_etime;     /* Elapsed time */
    comp_t    ac_mem;       /* Average memory usage (kB) */
    comp_t    ac_io;        /* Characters transferred (unused) */
    comp_t    ac_rw;        /* Blocks read or written (unused) */
    comp_t    ac_minflt;    /* Minor page faults */
    comp_t    ac_majflt;    /* Major page faults */
    comp_t    ac_swaps;     /* Number of swaps (unused) */
    u_int32_t ac_exitcode;  /* Process termination status
                               (see wait(2)) */
    char      ac_comm[ACCT_COMM+1];
                            /* Command name (basename of last
                               executed command; null-terminated) */
    char      ac_pad[\fIX\fP];    /* padding bytes */
};

enum {          /* Bits that may be set in ac_flag field */
    AFORK = 0x01,           /* Has executed fork, but no exec */
    ASU   = 0x02,           /* Used superuser privileges */
    ACORE = 0x08,           /* Dumped core */
    AXSIG = 0x10            /* Killed by a signal */
};
.fi
.in
.PP
The
.I comp_t
data type is a floating-point value consisting of a 3-bit, base-8 exponent,
and a 13-bit mantissa.
A value,
.IR c ,
of this type can be converted to a (long) integer as follows:
.nf

    v = (c & 0x1fff) << (((c >> 13) & 0x7) * 3);
.fi
.PP
The
.IR ac_utime ,
.IR ac_stime ,
and
.I ac_etime
fields measure time in "clock ticks"; divide these values by
.I sysconf(_SC_CLK_TCK)
to convert them to seconds.
.SS Version 3 accounting file format
Since kernel 2.6.8,
an optional alternative version of the accounting file can be produced
if the
.B CONFIG_BSD_PROCESS_ACCT_V3
option is set when building the kernel.
With this option is set,
the records written to the accounting file contain additional fields,
and the width of
.I c_uid
and
.I ac_gid
fields is widened from 16 to 32 bits
(in line with the increased size of UID and GIDs in Linux 2.4 and later).
The records are defined as follows:

.in +4n
.nf
struct acct_v3 {
    char      ac_flag;      /* Flags */
    char      ac_version;   /* Always set to ACCT_VERSION (3) */
    u_int16_t ac_tty;       /* Controlling terminal */
    u_int32_t ac_exitcode;  /* Process termination status */
    u_int32_t ac_uid;       /* Real user ID */
    u_int32_t ac_gid;       /* Real group ID */
    u_int32_t ac_pid;       /* Process ID */
    u_int32_t ac_ppid;      /* Parent process ID */
    u_int32_t ac_btime;     /* Process creation time */
    float     ac_etime;     /* Elapsed time */
    comp_t    ac_utime;     /* User CPU time */
    comp_t    ac_stime;     /* System time */
    comp_t    ac_mem;       /* Average memory usage (kB) */
    comp_t    ac_io;        /* Characters transferred (unused) */
    comp_t    ac_rw;        /* Blocks read or written
                               (unused) */
    comp_t    ac_minflt;    /* Minor page faults */
    comp_t    ac_majflt;    /* Major page faults */
    comp_t    ac_swaps;     /* Number of swaps (unused) */
    char      ac_comm[ACCT_COMM]; /* Command name */
};
.fi
.in
.SH VERSIONS
The
.I acct_v3
structure is defined in glibc since version 2.6.
.SH CONFORMING TO
Process accounting originated on BSD.
Although it is present on most systems, it is not standardized,
and the details vary somewhat between systems.
.SH NOTES
Records in the accounting file are ordered by termination time of
the process.

In kernels up to and including 2.6.9,
a separate accounting record is written for each thread created using
the NPTL threading library;
since Linux 2.6.10,
a single accounting record is written for the entire process
on termination of the last thread in the process.

The
.I proc/sys/kernel/acct
file, described in
.BR proc (5),
defines settings that control the behavior of process accounting
when disk space runs low.
.SH SEE ALSO
.BR lastcomm (1),
.BR acct (2),
.BR accton (8),
.BR sa (8)