| 1 |
.\" Copyright 1994-1995, Daniel Quinlan <[email protected]> |
| 2 |
.\" Copyright 2002-2017, Michael Kerrisk <[email protected]> |
| 3 |
.\" Copyright, the authors of the Linux man-pages project |
| 4 |
.\" |
| 5 |
.\" SPDX-License-Identifier: GPL-3.0-or-later |
| 6 |
.\" |
| 7 |
.TH proc_stat 5 2026-02-08 "Linux man-pages 6.18" |
| 8 |
.SH NAME |
| 9 |
/proc/stat \- kernel system statistics |
| 10 |
.SH DESCRIPTION |
| 11 |
.TP |
| 12 |
.I /proc/stat |
| 13 |
kernel/system statistics. |
| 14 |
Varies with architecture. |
| 15 |
Common |
| 16 |
entries include: |
| 17 |
.RS |
| 18 |
.TP |
| 19 |
.I cpu 10132153 290696 3084719 46828483 16683 0 25195 0 175628 0 |
| 20 |
.TQ |
| 21 |
.I cpu0 1393280 32966 572056 13343292 6130 0 17875 0 23933 0 |
| 22 |
The amount of time, measured in units of |
| 23 |
USER_HZ (1/100ths of a second on most architectures, use |
| 24 |
.I sysconf(_SC_CLK_TCK) |
| 25 |
to obtain the right value), |
| 26 |
.\" 1024 on Alpha and ia64 |
| 27 |
that the system ("cpu" line) or the specific CPU |
| 28 |
.RI (\[dq]cpu N \[dq] |
| 29 |
line) |
| 30 |
spent in various states: |
| 31 |
.RS |
| 32 |
.TP |
| 33 |
.I user |
| 34 |
(1) Time spent in user mode. |
| 35 |
.TP |
| 36 |
.I nice |
| 37 |
(2) Time spent in user mode with low priority (nice). |
| 38 |
.TP |
| 39 |
.I system |
| 40 |
(3) Time spent in system mode. |
| 41 |
.TP |
| 42 |
.I idle |
| 43 |
(4) Time spent in the idle task. |
| 44 |
.\" FIXME . Actually, the following info about the /proc/stat 'cpu' field |
| 45 |
.\" does not seem to be quite right (at least in Linux 2.6.12 or Linux 3.6): |
| 46 |
.\" the idle time in /proc/uptime does not quite match this value |
| 47 |
This value should be USER_HZ times the |
| 48 |
second entry in the |
| 49 |
.I /proc/uptime |
| 50 |
pseudo-file. |
| 51 |
.TP |
| 52 |
.IR iowait " (since Linux 2.5.41)" |
| 53 |
(5) Time waiting for I/O to complete. |
| 54 |
This value is not reliable, for the following reasons: |
| 55 |
.\" See kernel commit 9c240d757658a3ae9968dd309e674c61f07c7f48 |
| 56 |
.RS |
| 57 |
.IP \[bu] 3 |
| 58 |
The CPU will not wait for I/O to complete; |
| 59 |
iowait is the time that a task is waiting for I/O to complete. |
| 60 |
When a CPU goes into idle state for outstanding task I/O, |
| 61 |
another task will be scheduled on this CPU. |
| 62 |
.IP \[bu] |
| 63 |
On a multi-core CPU, |
| 64 |
the task waiting for I/O to complete is not running on any CPU, |
| 65 |
so the iowait of each CPU is difficult to calculate. |
| 66 |
.IP \[bu] |
| 67 |
The value in this field may |
| 68 |
.I decrease |
| 69 |
in certain conditions. |
| 70 |
.RE |
| 71 |
.TP |
| 72 |
.IR irq " (since Linux 2.6.0)" |
| 73 |
.\" Precisely: Linux 2.6.0-test4 |
| 74 |
(6) Time servicing interrupts. |
| 75 |
.TP |
| 76 |
.IR softirq " (since Linux 2.6.0)" |
| 77 |
.\" Precisely: Linux 2.6.0-test4 |
| 78 |
(7) Time servicing softirqs. |
| 79 |
.TP |
| 80 |
.IR steal " (since Linux 2.6.11)" |
| 81 |
(8) Stolen time, which is the time spent in other operating systems when |
| 82 |
running in a virtualized environment |
| 83 |
.TP |
| 84 |
.IR guest " (since Linux 2.6.24)" |
| 85 |
(9) Time spent running a virtual CPU for guest |
| 86 |
operating systems under the control of the Linux kernel. |
| 87 |
.\" See Changelog entry for 5e84cfde51cf303d368fcb48f22059f37b3872de |
| 88 |
.TP |
| 89 |
.IR guest_nice " (since Linux 2.6.33)" |
| 90 |
.\" commit ce0e7b28fb75cb003cfc8d0238613aaf1c55e797 |
| 91 |
(10) Time spent running a niced guest (virtual CPU for guest |
| 92 |
operating systems under the control of the Linux kernel). |
| 93 |
.RE |
| 94 |
.TP |
| 95 |
.I page 5741 1808 |
| 96 |
The number of pages the system paged in and the number that were paged |
| 97 |
out (from disk). |
| 98 |
.TP |
| 99 |
.I swap 1 0 |
| 100 |
The number of swap pages that have been brought in and out. |
| 101 |
.TP |
| 102 |
.\" FIXME . The following is not the full picture for the 'intr' of |
| 103 |
.\" /proc/stat on 2.6: |
| 104 |
.I intr 1462898 |
| 105 |
This line shows counts of interrupts serviced since boot time, |
| 106 |
for each of the possible system interrupts. |
| 107 |
The first column is the total of all interrupts serviced |
| 108 |
including unnumbered architecture specific interrupts; |
| 109 |
each subsequent column is the total for that particular numbered interrupt. |
| 110 |
Unnumbered interrupts are not shown, only summed into the total. |
| 111 |
.TP |
| 112 |
.IR "disk_io: (2,0):(31,30,5764,1,2) (3,0):" ... |
| 113 |
(major,disk_idx):(noinfo, read_io_ops, blks_read, write_io_ops, blks_written) |
| 114 |
.br |
| 115 |
(Linux 2.4 only) |
| 116 |
.TP |
| 117 |
.I ctxt 115315 |
| 118 |
The number of context switches that the system underwent. |
| 119 |
.TP |
| 120 |
.I btime 769041601 |
| 121 |
boot time, in seconds since the Epoch, 1970-01-01 00:00:00 +0000 (UTC). |
| 122 |
.TP |
| 123 |
.I processes 86031 |
| 124 |
Number of forks since boot. |
| 125 |
.TP |
| 126 |
.I procs_running 6 |
| 127 |
Number of processes in runnable state. |
| 128 |
(Linux 2.5.45 onward.) |
| 129 |
.TP |
| 130 |
.I procs_blocked 2 |
| 131 |
Number of processes blocked waiting for I/O to complete. |
| 132 |
(Linux 2.5.45 onward.) |
| 133 |
.TP |
| 134 |
.I softirq 229245889 94 60001584 13619 5175704 2471304 28 51212741 59130143 0 51240672 |
| 135 |
.\" commit d3d64df21d3d0de675a0d3ffa7c10514f3644b30 |
| 136 |
This line shows the number of softirq for all CPUs. |
| 137 |
The first column is the total of all softirqs and |
| 138 |
each subsequent column is the total for particular softirq. |
| 139 |
(Linux 2.6.31 onward.) |
| 140 |
.RE |
| 141 |
.SH SEE ALSO |
| 142 |
.BR proc (5) |