diff options
Diffstat (limited to 'man5/proc_timer_stats.5')
| -rw-r--r-- | man5/proc_timer_stats.5 | 117 |
1 files changed, 117 insertions, 0 deletions
diff --git a/man5/proc_timer_stats.5 b/man5/proc_timer_stats.5 new file mode 100644 index 0000000..3503f11 --- /dev/null +++ b/man5/proc_timer_stats.5 @@ -0,0 +1,117 @@ +.\" Copyright (C) 1994, 1995, Daniel Quinlan <quinlan@yggdrasil.com> +.\" Copyright (C) 2002-2008, 2017, Michael Kerrisk <mtk.manpages@gmail.com> +.\" Copyright (C) 2023, Alejandro Colomar <alx@kernel.org> +.\" +.\" SPDX-License-Identifier: GPL-3.0-or-later +.\" +.TH proc_timer_stats 5 2023-08-15 "Linux man-pages 6.7" +.SH NAME +/proc/timer_stats \- timer statistics +.SH DESCRIPTION +.TP +.IR /proc/timer_stats " (from Linux 2.6.21 until Linux 4.10)" +.\" commit 82f67cd9fca8c8762c15ba7ed0d5747588c1e221 +.\" Date: Fri Feb 16 01:28:13 2007 -0800 +.\" Text largely derived from Documentation/timers/timer_stats.txt +.\" removed in commit dfb4357da6ddbdf57d583ba64361c9d792b0e0b1 +.\" Date: Wed Feb 8 11:26:59 2017 -0800 +This is a debugging facility to make timer (ab)use in a Linux +system visible to kernel and user-space developers. +It can be used by kernel and user-space developers to verify that +their code does not make undue use of timers. +The goal is to avoid unnecessary wakeups, +thereby optimizing power consumption. +.IP +If enabled in the kernel +.RB ( CONFIG_TIMER_STATS ), +but not used, +it has almost zero run-time overhead and a relatively small +data-structure overhead. +Even if collection is enabled at run time, overhead is low: +all the locking is per-CPU and lookup is hashed. +.IP +The +.I /proc/timer_stats +file is used both to control sampling facility and to read out the +sampled information. +.IP +The +.I timer_stats +functionality is inactive on bootup. +A sampling period can be started using the following command: +.IP +.in +4n +.EX +# echo 1 > /proc/timer_stats +.EE +.in +.IP +The following command stops a sampling period: +.IP +.in +4n +.EX +# echo 0 > /proc/timer_stats +.EE +.in +.IP +The statistics can be retrieved by: +.IP +.in +4n +.EX +$ cat /proc/timer_stats +.EE +.in +.IP +While sampling is enabled, each readout from +.I /proc/timer_stats +will see +newly updated statistics. +Once sampling is disabled, the sampled information +is kept until a new sample period is started. +This allows multiple readouts. +.IP +Sample output from +.IR /proc/timer_stats : +.IP +.in +4n +.EX +.RB $ " cat /proc/timer_stats" +Timer Stats Version: v0.3 +Sample period: 1.764 s +Collection: active + 255, 0 swapper/3 hrtimer_start_range_ns (tick_sched_timer) + 71, 0 swapper/1 hrtimer_start_range_ns (tick_sched_timer) + 58, 0 swapper/0 hrtimer_start_range_ns (tick_sched_timer) + 4, 1694 gnome\-shell mod_delayed_work_on (delayed_work_timer_fn) + 17, 7 rcu_sched rcu_gp_kthread (process_timeout) +\&... + 1, 4911 kworker/u16:0 mod_delayed_work_on (delayed_work_timer_fn) + 1D, 2522 kworker/0:0 queue_delayed_work_on (delayed_work_timer_fn) +1029 total events, 583.333 events/sec +.EE +.in +.IP +The output columns are: +.RS +.IP [1] 5 +a count of the number of events, +optionally (since Linux 2.6.23) followed by the letter \[aq]D\[aq] +.\" commit c5c061b8f9726bc2c25e19dec227933a13d1e6b7 deferrable timers +if this is a deferrable timer; +.IP [2] +the PID of the process that initialized the timer; +.IP [3] +the name of the process that initialized the timer; +.IP [4] +the function where the timer was initialized; and +(in parentheses) +the callback function that is associated with the timer. +.RE +.IP +During the Linux 4.11 development cycle, +this file was removed because of security concerns, +as it exposes information across namespaces. +Furthermore, it is possible to obtain +the same information via in-kernel tracing facilities such as ftrace. +.SH SEE ALSO +.BR proc (5) |