Remove KVM related options.

branches: 1.24.12; 1.24.22;
Maintain 5-character width of MB/s column by dynamically adjusting the
decimal precision.

This should work until disk transfer rates exceed 99999 MB/s.

update a comment.

Update to use the new kernel io statistics.

Add tape statistics.

use NETBSDSRCDIR as appropriate

Remove check for sparc64 which adds -msoft-quad-float as that is now the
default for the compiler on sparc64.

Drop the setgid'ness now that ../../usr.bin/vmstat/dkstats.c uses
sysctls now.

use -msoft-quad-float on the sparc64.

use CPPFLAGS instead of CFLAGS

branches: 1.15.2;
Back out the .PATH.c changes. The .depend problem (and others)
will be fixed using the new .NOPATH make feature instead.

Use .PATH.c: ...

New userland code to read the statistics kept by the NetBSD generic
disk framework, from John M. Vinopal <banshee@gabriella.resort.com>

Update Makefiles for usr.bin/systat and usr.sbin/iostat to be
consistent with usr.bin/vmstat/Makefile wrt CLFAGS and -I, since
all three compile vmstat/names.c. (Now compiles on a pmax.)

New-style RCS ids.

specify man pages the new way.

do not need -I/sys

add -I to make this build properly for hp300

add machine define if m68k arch (for vmstat/names.c)

Restore "${DESTDIR}" in CFLAGS.

clean up import. that was weird.

needs -lkvm, not -lutil

-I/sys --> -I${DESTDIR}/sys, to support cross-compilation.

Add RCS identifiers.

branches: 1.1.1;
added accton, edquota, flcopy, gettable, htable, iostat, quotaon,
repquota, trpt, and trsp, per (currently alpha) patch 125.

ostat: use standard option order (AaBb...)

Fix macro usage while here

add iostat -X option, limited alternative statistics.

this is like the "-D" option, except it skips the xfers/s, and
reports MB/s instead of KB/s. allows for far more devices to
be displayed per line by default (about 50% more.)

Significant changes to output layout methods - except for -x and -y
which have not (yet) been touched (-xD needs *serious* improvements).

While this still has no run-time configurability, it is now easy to
adjust the column widths in the source and recompile. Dynamic (auto)
column width sizing is probably out of the question (requires predicting
the future) but options to allow the widths to be set isn't out of the
question.

The columns are now (mostly) considerably wider than they were before,
hence wider windows are needed to view the same info. In an 80
column window the default display (with tty & cpu included) displays
just 2 drives. 160 columns will fit 7 (but with -I, just 4).

One new option added (-z) suppresses output which is true 0 (but still
prints 0 for values rounded down to 0) for everything except tty stats.
For drive output, the drive must have done nothing during the interval
to get its output data blanked.

Also options -H h -W w to set the output size (page height & width), the
former used to decide when to print headers, and the latter to calculate
the number of drives to print when no drive names were given. Env vars
LINES and COLUMNS are used if the options are not given, with fallback
to the terminal size (if output is to a terminal, and its sizes are
known), and if all else fails, 20 lines, 80 columns. Specifying 0
means unlimited (infinite). So "iostat -W 0" will show all of the
drives, across one (often very) long line. Wedges count as drives.

When drives are specified, the output will now appear in the order they were
given on the command line, rather than the order the system discovered them
during auto-configuration. If specified as an fnmatch(3) pattern, drives that
match will appear in auto-conf order, but that's generally what is wanted.
When none are specified, you still get the first N (however many fits based
upon the options selected) in auto-conf order (usually useless, more so now
given that less will fit).

Lastly, for those who looked at the patch I sent to current-users@
and were horrified at how kludgey it was, rest assured, that was just
a quick hack to demonstrate what the output format changes would look
like. This version (I hope) is not nearly so disgusting.

Fix mandoc warnings.

Document "in" field in CPU section
Fix ps(1) Xr raised by mandoc -Tlint

branches: 1.25.10; 1.25.12;
Enhance disk metrics by calculating a weighted sum that is incremented
by the number of concurrent I/O requests. Also introduce a new disk_wait()
function to measure requests waiting in a bufq.
iostat -y now reports data about waiting and active requests.

So far only drivers using dksubr and dk, ccd, wd and xbd collect data about
waiting requests.

branches: 1.24.2; 1.24.4;
support fnmatch(3) patterns for disknames. eg, "iostat wd*" works now.
update the default disk number selection to note that it uses the current
terminal size to fit disks, not hard coded 4.

branches: 1.23.2;
Remove KVM related options.

branches: 1.22.60; 1.22.72;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22366, verified by myself.

Increase width now we mention flag arguments.

In the options list, show which ones require an argument.
New sentence, new line.

Bump date for recent change.

Sync with the behaviour change.

iostat(8) and systat(8) support for separate read/write disk statistics.
"iostat -x" now shows these (ala linux/solaris), but this is only splitting
the read/write bytes/transactions, not adding any new metrics. "systat
iostat" now has two new commands to switch between combined/separate mode
for both it's numbers & bar modes.

Uppercase cpu.

Change xref for vmstat(8) to vmstat(1).
Whitespace, punctuation and spelling fixes while I'm here.

More and more .Os cleanups. .Os is defined in the tmac.doc-common file,
so we shouldn't override it with versions in the manpages. Many more to
come.

* implement -x; extended disk statistics. option inspired by solaris' iostat.
* fix a few typos
* a bit of KNF and egcs cleanup
* add a HISTORY (hopefully correct)

WARNSify

alternate -> alternative, per PR 2643

`Kb' means kilobits, `Mb' means megabits. Since we're talking about bytes,
not bits, use the uppercase `B' when reporting statistics, and be explicit
about it where appropriate. Noticed by J.T.

New userland code to read the statistics kept by the NetBSD generic
disk framework, from John M. Vinopal <banshee@gabriella.resort.com>

New-style RCS ids.

vmstat is in section 8; PR #1532

clean up import. that was weird.

Got bored, converted to use mandoc macros.

Add RCS indentifiers.

Fix up FILES sections.

Update file names.

branches: 1.1.1;
added accton, edquota, flcopy, gettable, htable, iostat, quotaon,
repquota, trpt, and trsp, per (currently alpha) patch 125.

fix rev 1.70: -X is 1 line output, avoids printing the header too often.

ostat: use standard option order (AaBb...)

Fix macro usage while here

add iostat -X option, limited alternative statistics.

this is like the "-D" option, except it skips the xfers/s, and
reports MB/s instead of KB/s. allows for far more devices to
be displayed per line by default (about 50% more.)

Significant changes to output layout methods - except for -x and -y
which have not (yet) been touched (-xD needs *serious* improvements).

While this still has no run-time configurability, it is now easy to
adjust the column widths in the source and recompile. Dynamic (auto)
column width sizing is probably out of the question (requires predicting
the future) but options to allow the widths to be set isn't out of the
question.

The columns are now (mostly) considerably wider than they were before,
hence wider windows are needed to view the same info. In an 80
column window the default display (with tty & cpu included) displays
just 2 drives. 160 columns will fit 7 (but with -I, just 4).

One new option added (-z) suppresses output which is true 0 (but still
prints 0 for values rounded down to 0) for everything except tty stats.
For drive output, the drive must have done nothing during the interval
to get its output data blanked.

Also options -H h -W w to set the output size (page height & width), the
former used to decide when to print headers, and the latter to calculate
the number of drives to print when no drive names were given. Env vars
LINES and COLUMNS are used if the options are not given, with fallback
to the terminal size (if output is to a terminal, and its sizes are
known), and if all else fails, 20 lines, 80 columns. Specifying 0
means unlimited (infinite). So "iostat -W 0" will show all of the
drives, across one (often very) long line. Wedges count as drives.

When drives are specified, the output will now appear in the order they were
given on the command line, rather than the order the system discovered them
during auto-configuration. If specified as an fnmatch(3) pattern, drives that
match will appear in auto-conf order, but that's generally what is wanted.
When none are specified, you still get the first N (however many fits based
upon the options selected) in auto-conf order (usually useless, more so now
given that less will fit).

Lastly, for those who looked at the patch I sent to current-users@
and were horrified at how kludgey it was, rest assured, that was just
a quick hack to demonstrate what the output format changes would look
like. This version (I hope) is not nearly so disgusting.

Fix processing of the archaic arg format (BACKWARD_COMPATIBILITY) so it
doesn't repeat the processing every iteration. Repeatedly seeing the wait
interval does no harm, but setting the iteration count (reps) over and
over again rather defeats its purpose.

branches: 1.67.4;
Re-evaluate drive selection to catch hog-plug devices.

branches: 1.66.2;
two minor fixes for -x and -y mode:
- the header is only 1 line not 2, account for this when calculating
when to display the next header.
- when ndrives > 1, don't display the header every time, but only when
the previous one disappears.

now i don't feel the need to run "iostat -x wd0 1 & iostat -c wd1 1"
to get less repeated header output on my display.

Use I/O timestamps to compute disk statistics for better precision.

Disk statistics are collected in a fixed size array, that got corrupted
when a disk was detached. Adapt by skipping entries of detached disks
and detect reused disknames at the array end.

branches: 1.64.4;
Enhance disk metrics by calculating a weighted sum that is incremented
by the number of concurrent I/O requests. Also introduce a new disk_wait()
function to measure requests waiting in a bufq.
iostat -y now reports data about waiting and active requests.

So far only drivers using dksubr and dk, ccd, wd and xbd collect data about
waiting requests.

branches: 1.63.2; 1.63.4;
extend the "tout" column to 5 characters, it's really common on a modern
machine to have eg, build output way more than 9999 chars/sec.

this doesn't affect the default disk list on an 80-char wide screen.

support fnmatch(3) patterns for disknames. eg, "iostat wd*" works now.
update the default disk number selection to note that it uses the current
terminal size to fit disks, not hard coded 4.

branches: 1.61.2;
Remove KVM related options.

branches: 1.60.8; 1.60.18;
static + __dead

A number greater than or equal to 1.0 is a positive number.

Put space after comma and around binary operator.

branches: 1.57.2;
Fix copy/paste error that resulted in a extra space after the KB/t column.

NetBSD/vax libm now has fmax, use it again.
Also, display average transfer size using dynamic precision too.

NetBSD/vax does not implement fmax(3), substitute with existing MAX() macro.

Maintain 5-character width of MB/s column by dynamically adjusting the
decimal precision.

This should work until disk transfer rates exceed 99999 MB/s.

Fix -Wsign-compare issues

branches: 1.52.6;
Remove the \n and tabs from the __COPYRIGHT() strings.
Tweak to use a consistent format.

branches: 1.51.12;
add parens.

Since the swapping of tk_nin and tk_nout was moved to its own
function (tkswap()), actually call this function.

Fixes problems where the tin/tout fields showed total tty IO since
boot instead of just what has happened in the last second.

Update to use the new kernel io statistics.

Need to read and swap cpustats separately from the dkstats.
I wonder how many more programs use the code in vmstat/dkstats.c

print a header on each outputs if and only if the output has 2 or more lines.

reset hdrcnt whenever printing a header.

Add tape statistics.

apply the patch from PR#30680: always print the header in extended
(-x) mode, as it is much easier to see what is going on.

Add (unsigned char) cast to ctype functions

for -x -I output, use just MB instead of MB/s in header

Uppercase CPU, plural is CPUs.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22366, verified by myself.

- convert header() away from a signal function (it does all sorts of stdio
and more) and use a little sig_header() helper.
- in selectdrives() make sure we don't overwrite some arrays. this makes
"iostat 1" work again on my really wide screens where defdrives (number that
can fit) was > ndrives (number of drives), rather than dump core trying to
print (char *)1...

Use the terminal width to choose the default number of drives listed.

Suggested by matt green and (indirectly) enami tsugutomo.

Pinch code from vmstat to get the number of lines of the terminal, and
show the header once per screen instead of fixed at every 22 lines.

Bump the width of the MB/s and t/s fields and show only 3 drives by
default. The default display now fits in 80 columns, and looks much
better on modern systems.

no need to #include <sys/dkstat.h>

- Account the number of lines consumed so that -xw displays header
frequently enough.
- As we are initializing dk_select to 0 (via dkinit), no need to check it.
- Re-order the code in display() to highlight the fact that -x is
exclusive with others.
- If user really specified some drives and none of them are found,
just bail out rather than falling back to default behaviour.
- Other misc. cosmetic changes while I'm here.

PR/20435: KOIE Hidetaka: iostat -x ignores drive selection.

... and fix typos which makes wrong display.

Cosmetic changes.

Print a newline between lines in -x mode.

iostat(8) and systat(8) support for separate read/write disk statistics.
"iostat -x" now shows these (ala linux/solaris), but this is only splitting
the read/write bytes/transactions, not adding any new metrics. "systat
iostat" now has two new commands to switch between combined/separate mode
for both it's numbers & bar modes.

select() -> nanosleep()

Since the KERN_CP_TIME sysctl reports total clock ticks across all
cpus and hz is measured in ticks per cpu, divide tick count by ncpu to
determine elapsed time since last sample.

Fixes I/O rate deflation observed on multiprocessors.

We don't need the code to discard setgid privileges if we aren't setgid!

branches: 1.25.2;
Drop the setgid'ness now that ../../usr.bin/vmstat/dkstats.c uses
sysctls now.

Remove extern variables and functions now declared in dkstats.h.

ANSIfy.

Need to include <sys/sched.h> to get cp_time-related constants.

Use " %2.0f" instead of "%3.0f" to print out the CPU percentages - this
way 100 in a column doesn't merge into the previous column.

branches: 1.20.6;
Don't unnecessarily check device names if "-x" is used.

Display all drives if "iostat -x <interval>" is used.
Fixes PR bin/8389 by Mattias Pantzare.

Truncate disk names in header to 7 instead of 3 chars, we actually
have "raid0".

branches: 1.17.2;
in -x, print MB/s two 2 decimal places. noted by enami@netbsd.org

add -x in usage.

* implement -x; extended disk statistics. option inspired by solaris' iostat.
* fix a few typos
* a bit of KNF and egcs cleanup
* add a HISTORY (hopefully correct)

adapt to changed dkinit() in ../../usr.bin/vmstat/dkstats.c

KNF.

WARNSify

merge lite-2

`Kb' means kilobits, `Mb' means megabits. Since we're talking about bytes,
not bits, use the uppercase `B' when reporting statistics, and be explicit
about it where appropriate. Noticed by J.T.

New userland code to read the statistics kept by the NetBSD generic
disk framework, from John M. Vinopal <banshee@gabriella.resort.com>

New-style RCS ids.

Add a column header for CP_INTR.

clean up import. that was weird.

isa device configuration changed, and the tables got killed/renamed.
fix from Jarle Greipsland <jarle@darling.idt.unit.no>.

__386BSD__ -> i386; this file might break for m68k because of ifdefs...

Add RCS identifiers.

iostat and vmstat being taught about drives..

branches: 1.1.1;
added accton, edquota, flcopy, gettable, htable, iostat, quotaon,
repquota, trpt, and trsp, per (currently alpha) patch 125.