Commit Graph

134 Commits

Author SHA1 Message Date
John Baldwin
e876f6d052 Cap the percent CPU of individual threads at 100% to fix some of the
more obvious imprecision in the previous top changes.

Specifically, top uses a delta of clock_gettime() calls right after
invoking the kern.proc sysctl to fetch the process/thread list to
compute the time delta between the fetches.  However, the kern.proc
sysctl handler does not run in constant time.  It can spin on locks,
be preempted by an interrupt handler, etc.  As a result, the time
between the gathering of stats for individual processes or threads
between subsequent kern.proc handlers can vary.  If a "slow" kern.proc
run is followed by a "fast" kern.proc run, then the threads/processes
at the start of the "slow" run will have a longer time delta than the
threads/processes at the end.  If the clock_gettime() time delta is
not itself skewed by preemption, then the delta may be too short for
a given thread/process resulting in a higher percent CPU than actual.
However, there is no good way to calculate the exact amount of overage,
nor to know which threads to subtract the overage from.  Instead, just
punt and fix the definitely-wrong case of an individual thread having
more than 100% CPU.

Discussed with:	zonk
2014-06-20 19:54:23 +00:00
John Baldwin
a90539544c Rework the notion of CPU used in top. In particular, for subsequent
displays after a pause, use the difference in runtime divided by the
length of the pause as the percentage of CPU used instead of the value
calculated by the kernel.  In addition, when determing if a process or
thread is idle or not, treat any process or thread that has used any
runtime or performed any context switches during the interval as busy.

Note that the percent CPU is calculated as a double and stored in an
array to avoid recalculating the value multiple times in the comparison
method used to sort processes in the CPU display.

Tested by:	Jamie Landeg-Jones <jamie@dyslexicfish.net>
Reviewed by:	emaste (earlier version)
MFC after:	1 week
2014-05-30 21:18:53 +00:00
John Baldwin
5648de3b31 Fix a couple of size_t != int warnings. 2014-05-27 20:22:36 +00:00
Bryan Drewery
069f49cfe8 Fix width/alignment of JID column. Make it support up to the maximum
7-wide JIDs. On a system using jails for common tasks the JID can
quickly increase.

MFC after:	2 weeks
2014-05-03 15:03:47 +00:00
Bryan Drewery
1b60f97272 Add -J command/flag to filter by jail name/jid. This will automatically
display the JID as well (the -j command/flag).

  0 displays host.
  + displays all.

MFC after:	2 weeks
Relnotes:	yes
2014-05-02 23:30:39 +00:00
John Baldwin
49e3bde165 Bump the CPU/WCPU column width by one so that it fits values from 100% up
to 999.99% CPU.  It still won't be aligned if you have a multithreaded
process using more than 1000% CPU (e.g. idle process on an idle 12-way
system), but 100% is a common case.

Submitted by:	Jeremy Chadwick (partial)
MFC after:	1 week
2013-06-11 19:05:29 +00:00
John Baldwin
ee66c8b94e Fix the 'C' field for a running thread to match the behavior described
in the manpage by having it display the current CPU (ki_oncpu) rather
than the previously used CPU (ki_lastcpu).  ki_lastcpu is still used for
all other thread states.

Reported by:	Chris Ross <cross+freebsd@distal.com>
MFC after:	1 week
2013-03-11 16:33:05 +00:00
Rui Paulo
78210af90c Use the correct size when allocating the cmdbuf string.
cmdlengthdelta is the size of the header and we were using it to
allocate a buffer to store the command line. This would mean that
the cmdbuf could be too short. In practice this was never noticed unless
you usually run top -a. On a stock FreeBSD system you can see the
problem by running sendmail and then running top -a on a big terminal
window. In practice this doubles to size available to cmdbuf since the
header is around 65-68 bytes.

Reviewed by:	adrian
2012-11-19 08:03:40 +00:00
Rui Paulo
8b800d3af2 Add the PID column to the list of sort keys. 2012-11-11 08:16:33 +00:00
John Baldwin
b49af68ca0 Correct the order of the MFU and MRU labels. I had reversed them.
Submitted by:	Nikolay Denev  ndenev gmail
Pointy hat to:	jhb
MFC after:	3 days
2012-10-19 12:28:26 +00:00
John Baldwin
54900c9b1f Fix two layout bugs in the previous change:
- Properly increase y_mem when per-CPU stats are enabled.
- Update y_arc for per-CPU stats being enabled/disabled.

MFC after:	3 days
2012-07-02 20:08:11 +00:00
John Baldwin
2e52fb92ff Add a new line to top that provides a brief summary of the ZFS ARC memory
usage on hosts using ZFS.  The new line displays the total amount of RAM
used by the ARC along with the size of MFU, MRU, anonymous (in flight),
headers, and other (miscellaneous) sub-categories.  The line is not
displayed on systems that are not using ZFS.

Reviewed by:	avg, fs@
MFC after:	3 days
2012-06-27 18:08:48 +00:00
John Baldwin
17cf6fc527 Clarify that the cached file data pages included in the "Wired" count
in top are the BIO-level cached data (i.e. "Buf"), since the previous
phrase was a bit ambiguous with the "Cache" count.

MFC after:	3 days
2012-06-27 12:30:56 +00:00
Konstantin Belousov
bf313207fc Fix string buffer overflow when preparing the line of output.
PR:	bin/161739
Submitted by:	Jeremy Chadwick <freebsd jdc parodius com>
MFC after:	1 week
2012-04-18 10:23:42 +00:00
Eitan Adler
50d675f7a9 Remove trailing whitespace per mdoc lint warning
Disussed with:	gavin
No objection from:	doc
Approved by:	joel
MFC after:	3 days
2012-03-29 05:02:12 +00:00
Ulrich Spörlein
487ac9ac21 Spelling fixes for usr.bin/ 2011-12-30 11:02:40 +00:00
John Baldwin
fcc3d62fa5 Rework the dynamic per-CPU stats code a bit. Always set 'statics->ncpus'
to the maximum number of CPUs to ensure that lcpustates[] array is always
allocated to the maximum size.  Previously, if top was started without
per-CPU stats it would allocate a smaller lcpustates[] array.  When
per-CPU stats were then enabled, it would overflow the array and trash
the cpustates_columns[] array causing the CPU stats to be printed in the
wrong locations.

Approved by:	re (kib)
MFC after:	1 week
2011-07-18 21:15:47 +00:00
Bjoern A. Zeeb
c618a5222e Constantly print the command name and if set include the thread
name in per-thread mode.

Discussed with:	jhb
Obtained from:	Ed Maste at Sandvine Incorporated
Sponsored by:	Sandvine Incorporated
MFC after:	1 week
Approved by:	re (kib)
2011-07-18 20:57:43 +00:00
Bjoern A. Zeeb
925af54487 Rename ki_ocomm to ki_tdname and OCOMMLEN to TDNAMLEN.
Provide backward compatibility defines under BURN_BRIDGES.

Suggested by:	jhb
Reviewed by:	emaste
Sponsored by:	Sandvine Incorporated
Approved by:	re (kib)
2011-07-18 20:06:15 +00:00
John Baldwin
2417d97ebb - Export each thread's individual resource usage in in struct kinfo_proc's
ki_rusage member when KERN_PROC_INC_THREAD is passed to one of the
  process sysctls.
- Correctly account for the current thread's cputime in the thread when
  doing the runtime fixup in calcru().
- Use TIDs as the key to lookup the previous thread to compute IO stat
  deltas in IO mode in top when thread display is enabled.

Reviewed by:	kib
Approved by:	re (kib)
2011-07-18 17:33:08 +00:00
John Baldwin
b225f36e26 Revert 130163 and let top use KERN_PROC_PROC when individual threads are
not displayed.  The KERN_PROC_PROC sysctl was fixed in 188764.
2011-07-15 17:23:45 +00:00
Bjoern A. Zeeb
f4dad666e3 Fix indentation.
Obtained from:	Ed Maste at Sandvine Incorporated
Sponsored by:	Sandvine Incorporated
MFC after:	3 days
2011-07-14 13:47:42 +00:00
John Baldwin
51f202e9ed Allow per-CPU statistics to be toggled at runtime via the 'P' key.
While here, make -P a toggle similar to other options such as -I.

Reviewed by:	arundel
MFC after:	1 week
2011-07-11 16:48:52 +00:00
John Baldwin
034b7b80b2 Always skip the kernel idle process if requested, it is not specific to
the 'CPU' mode.

PR:		bin/158677
Reported by:	arundel
MFC after:	3 days
2011-07-07 13:37:46 +00:00
John Baldwin
35b23f5d86 - Document the -H option and 'H' key alongside other options and keys
rather than at the bottom of the manpage.
- Remove an obsolete comment about SWAIT being a stale state.  It was
  resurrected for a different purpose in FreeBSD 5 to mark idle ithreads.
- Add a comment documenting that the SLEEP and LOCK states typically
  display the name of the event being waited on with lock names being
  prefixed with an asterisk and sleep event names not having a prefix.

MFC after:	1 week
2011-05-31 15:41:10 +00:00
John Baldwin
c9385548e7 Add a new option to toggle the display of the system idle process (per-CPU
idle threads).  The process is displayed by default (subject to whether or
not system processes are displayed) to preserve existing behavior.  The
system idle process can be hidden via the '-z' command line argument or the
'z' key while top is running.  When it is hidden, top more closely matches
the behavior of FreeBSD <= 4.x where idle time was not accounted to any
process.

MFC after:	2 weeks
2011-05-31 15:11:23 +00:00
Sergey Kandaurov
d9b2153b4e Add missing header file.
MFC after:	1 week
2011-05-20 17:03:23 +00:00
Giorgos Keramidas
001dcb9133 Touch up the sample memory usage numbers a bit, to avoid wrapping
on terminal boundary.  While here add definition for 'G' and fix
the indentation of 'K' units.

Submitted by:	plunket
PR:		docs/153614
MFC after:	3 days
2011-01-23 11:08:39 +00:00
Rebecca Cran
4784a27247 Memory stats are reported in bytes, not pages.
PR:	docs/151283
Submitted by:	Bas Smeelen <b.smeelen at ose.nl>
MFC after: 3 days
2010-11-13 11:43:20 +00:00
Ed Schouten
b7946da96b Build usr.bin/ with WARNS=6 by default.
Also add some missing $FreeBSD$ to keep svn happy.
2010-01-02 10:27:05 +00:00
John Baldwin
a0b9e7e332 Update top and systat for vfs.bufcache now being a long rather than an int. 2009-03-10 15:25:19 +00:00
Rui Paulo
491c34ec7e Right align the CPU column header. 2008-12-26 11:11:30 +00:00
John Baldwin
b2ce972537 Make the CPU column in top always be 2 characters and print in decimal
rather than hex.

Requested by:	rwatson
2008-10-01 22:01:07 +00:00
Stephane E. Potvin
c975a48dee Display the sum of the runtime of all the threads in a process when it's
multithreaded instead of picking the time of the first thread found.

Reviewed by:	jhb
Approved by:	kan (mentor)
MFC after:	1 month
2008-09-12 17:54:50 +00:00
Ruslan Ermilov
d3bf3b9a7a system_info.cpustates isn't sparse, so a bitmask of available CPU states
is redundant (I think it's a leftover from an older implementation).
2008-04-11 11:39:26 +00:00
Ruslan Ermilov
3aaa083285 Allocate enough memory for pcpu_cp_time[] to stop sysctl() from
writing outside of array bounds.  This fully fixes -P display on
i386, where kern.cp_times prints zeroes for non-existing CPUs.
2008-04-11 11:34:09 +00:00
Ruslan Ermilov
f89db4357e Fix "top -P" (&' mistyped as &&' and a botched logic).
The bug was unnoticed on non-i386 because mp_maxid is
initialized differently, kern.cp_times doesn't print
zeroes for non-existing CPUs, so no "writing outside of
array bounds" happens.

MFC after:	3 days
2008-04-10 16:17:54 +00:00
Xin LI
2f487fe4e4 Use calloc() when requesting zero'ed memory allocation rather than rolling
our own.
2008-01-29 00:06:44 +00:00
Peter Wemm
031175705e Add a -P flag to display per-cpu cpu usage stats. 2008-01-18 01:43:14 +00:00
David E. O'Brien
dbb2956620 Improve -u (limit uid lookups) behavior.
Submitted by:	David Frascone <dave@frascone.com>
PR:		119490
2008-01-09 18:06:24 +00:00
Julian Elischer
7ab24ea3b9 Introduce a way to make pure kernal threads.
kthread_add() takes the same parameters as the old kthread_create()
plus a pointer to a process structure, and adds a kernel thread
to that process.

kproc_kthread_add() takes the parameters for kthread_add,
plus a process name and a pointer to a pointer to a process instead of just
a pointer, and if the proc * is NULL, it creates the process to the
specifications required, before adding the thread to it.

All other old kthread_xxx() calls return, but act on (struct thread *)
instead of (struct proc *). One reason to change the name is so that
any old kernel modules that are lying around and expect kthread_create()
to make a process will not just accidentally link.

fix top to show  kernel threads by their thread name in -SH mode
add a tdnam formatting option to ps to show thread names.

make all idle threads actual kthreads and put them into their own idled process.
make all interrupt threads kthreads and put them in an interd process
(mainly for aesthetic and accounting reasons)
rename proc 0 to be 'kernel' and it's swapper thread is now 'swapper'

man page fixes to follow.
2007-10-26 08:00:41 +00:00
Jeff Roberson
b61ce5b0e6 - Move all of the PS_ flags into either p_flag or td_flags.
- p_sflag was mostly protected by PROC_LOCK rather than the PROC_SLOCK or
   previously the sched_lock.  These bugs have existed for some time.
 - Allow swapout to try each thread in a process individually and then
   swapin the whole process if any of these fail.  This allows us to move
   most scheduler related swap flags into td_flags.
 - Keep ki_sflag for backwards compat but change all in source tools to
   use the new and more correct location of P_INMEM.

Reported by:	pho
Reviewed by:	attilio, kib
Approved by:	re (kensmith)
2007-09-17 05:31:39 +00:00
Ruslan Ermilov
06351b59c6 Inactive pages don't have to be dirty plus improve description of
the cache queue.

Submitted by:	alc
Approved by:	re (kensmith)
2007-09-02 10:29:16 +00:00
Ruslan Ermilov
d71e591b49 Fix the description of the "Cache" memory and clarify the
description of the "Inact" memory.  (They count pages in
the cache/inactive page queues, respectively.)

Approved by:	re (kensmith)
2007-09-01 05:52:39 +00:00
Bruce Evans
e19249f7a7 Third stage of unbreaking printing of pseudo-nice values (realtime
priorities, etc.) in the NICE field:

Use a combination of pri_native and pri_user instead of pri_level to
guess the original realtime priority.  Using pri_level here has been
wrong since 2001/02/12.  Using only pri_native here would be correct
if the kernel actually initialized it reasonably.  (The kernel exports
its raw td_base_priority as pri_native, but userland mostly wants a
refined base priority).  Give up on waiting pri_native to work correctly
and only use it when there is nothing better (for kthreads).

This should reduce printing of bizarre pseudo-nice values.  Bizarre
values are still printed if we observe a transient borrowed priority
for a kthread (transient borrowing is the main thing that makes the
raw td_base_priority almost useless in userland), or if there is a
kernel bug.  One current kernel bug involves the kernel idprio thread
pagezero permanently changing its priority from PRI_MAX_IDLE (255) to
PUSER (160).  Then the bizarre value "ki-6" is printed instead of
"ki31".  Here "-6" is PRI_MIN_IDLE - PUSER = -64 truncated to 2
characters.  We are observing a transient borrowed priority that has
become permanent due to a bug.

ps/print.c:priorityr() needs similar changes (including ones in stage 2
here).
2007-06-15 12:03:07 +00:00
Rong-En Fan
6000ced11b - Add a new 'j' switch and runtime option to toggle display jail id for
each process.
- While I'm here, keep help message sorted by keys

PR:		98489, 98975
Submitted by:	clsung
Approved by:	delphij (mentor)
MFC after:	2 weeks
2007-04-17 03:12:39 +00:00
Stanislav Sedov
faac60c8fc - Add new 'a' switch and runtime option that allows 'top' to display process
titles extracted from argv vector instead of the real executable names.
  This is useful when you want to watch applications that set their status
  information via setproctitle(3).

Approved by:	alfred
MFC after:	2 weeks
2007-04-14 10:16:52 +00:00
Giorgos Keramidas
d2d6d8d537 The sigconv.awk script generates a sigdesc.h header file, which
contains a sigdec[] vector of structures, but the generated output is
missing braces around the initializer of each struct, which
triggers warnings in WARNS=3:

src/usr.bin/top/sigdesc.h:10: warning: missing braces around initializer
src/usr.bin/top/sigdesc.h:10: warning: (near initialization for `sigdesc[0]')

  * Fix the sigconv.awk script to generate a header with initializers
    which look better.

  * Add rules to usr.bin/top/Makefile that rebuilds a new sigconv.h
    header which matches the correct signal set from the build-time
    version of `${DESTDIR}/usr/include/signal.h' (so sigconv.h doesn't
    get stale once changes are made to the header).

  * Remove the old sigconv.h header, now that it is autoupdated at
    build time.

  * Various Makefile style fixes (the committed Makefile was kindly
    submitted by Ruslan):

    - Reorder .PATH, PROG, SRCS and CFLAGS to match style.Makefile(5)
    - Split off the generated sources (sigdesc.h top.local.h) in an
      SRCS+= line of their own.
    - Add entries to CLEANFILES near the rules that generate the
      respective files.
    - Move the explicit rule which builds top.1 after the implicit
      rules which generate its dependencies.

Reviewed by:	ru, bde
Submitted by:	ru (Makefile)
MFC after:	2 weeks
2006-12-01 07:01:19 +00:00
Bruce Evans
8dbbff5b81 Second stage of unbreaking thr formatting of the NICE field: decode the
priority class and use this to:

- print "-" instead of a garbage value for ithreads.  Print "-" instead
  of the unused nice value for kthreads which are (mis)classified as
  PRI_TIMESHARE.  For such threads, the nice value can be set to nonzero
  by root, but it is never used (at least by the 4bsd scheduler).  For
  ithreads, we didn't even print the unused value.

- print "i<priority>" and "r<priority>" instead of a biased "<priority>"
  for idletime and realtime threads,  Here <priority> is the priority
  parameter to idprio/rtprio(1).  Just add the prefix and remove the
  bias for now.  <priority> has been stored indirectly in the kernel
  since 2001/02/12, and even the kernel cannot recover the original
  value in all cases.  Here we need to handle more cases than pri_to_rtp(),
  but actually handle fewer cases, and end up printing garbage after
  a thread changes its current priority while in the kernel.

- for idletime and realtime threads, if they are kthreads then add a prefix
  of "k" to the previous string.

- for idletime and realtime threads, if they in the FIFO scheduling class
  then add a suffix of "F" to the previous string (if it fits; the other
  parts of the string are sure to fit unless <priority> is garbage).
2006-11-07 10:03:10 +00:00
Stefan Eßer
d1b76f3f6e Fix display of idle processes, which had been broken since rev. 1.56 of
machine.c. The traditional condition was (pctcpu > 0 || SRUN), but the
negation of the condition logic (from select to skip) made this come
out as (pctcpu > 0 && SRUN), leading to a very erratic display, except
for purely CPU bound processes.

This has been discussed in the mail lists some time ago and I have used
top with this patch on my systems for more than a year without problems
(just forgot to commit it earlier, since my systems were all fixed ...).
2006-06-11 19:18:39 +00:00