branches: 1.19.44;
- static/__dead
- use strpct from libutil
- fix off by one in format

Document non-literal formats

Default to WARNS=4 (except for rcp & ksh)

branches: 1.16.10;
PR/36817: Zafer Aydogan: Don't use -g to implement the same named option
in solaris to display statvfs output, use -G instead. I researched this
and all other implementations except Solaris (OS/X, FreeBSD, us before June
24, 2007) use -g to signify gigabytes. So revert to that.
PR/38154: YAMAMOTO Takashi: df -P doesn't work without -k. Fix from Anon Ymous
together with passing lint, and exiting with the right values.

branches: 1.15.24; 1.15.30;
Use ${NETBSDSRCDIR}/some/path instead of ${.CURDIR}/../../some/path (etc).

Following FreeBSD's example, remove the old pre-statfs(2) code for
examining FFS filesystems directly. This also gets rid of df(1)'s
setgid operator bit.

Fix PR#3590 - remove floating point calculation from df to avoid faulting
on machines that don't yet have complete FP support (e.g. MC68LC040)

Remove WARNS=1 from all the subdirectory Makefiles, and add it to Makefile.inc
now that all /bin has been cleaned.

Build with WARNS

Sync to 4.4BSD-Lite2; really just the SCCS ID, because we don't have vfslist.c in our mount sources.

convert to new RCS id conventions.

Deal with large file systems, and clean up a bit.

branches: 1.7.2;
don't need -lutil

Minor tweak for cross-compiling.

Incorporate new version using getbsize(), with -k flag added.

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Add -M and -q options to df(1).

-q suppresses warning output (rather than redirecting stderr, which
also buries error messages, which -q does not).

-M requires each arg given to be the name of a mount point, and
issues a warning, and ignores the arg, if it is not. This allows
scripts (etc) to have a whole list of places for which to produce df
output, while only listing the ones that are actually mounted (rather
than simply listing the filesystem containing the mount point instead -
perhaps several times). (If there are no args, -M is a no-op).

If the options aren't given, nothing alters.

Typo correction: s/givem/given/.

branches: 1.58.4;
Sort options, fix macro, add RCS Id.

Add 4 new flags:
-b (from FreeBSD) - set blocksize to blocks (512 bytes)
(overrides a contrary setting in BLOCKSIZE)
-H (from FreeBSD and Linux): -h using SI units (powers of 10). Ugh.
-N suppress the header line (except with -P which requires it).
-f show only free space (or inodes) in a minimal format (implies -N)
(that is, with one file[system] specified, print 1 number only)
With -c, show only the total.
Intended to be useful for scripting (aka, I needed it.)

While here, improve the usage message (group options where they apply,
there is no reason, for example, that -g should be shown differently
to -k -m ..., and those options aren't at all useful with -G)

Update the man page to match.

df: bump man page date

df: add grand total option

Add a grand total option, -c, similar to the du(1) -c option. Adapted from
the same option (-c) in FreeBSD df(1).

we don't need root anymore.

file system police. Fix typo. Fix macro use.

It is not just root, it is device read access (kre)

mention that -W needs root.

Print the wedge name with -W instead of mntfrom

-G cannot be specified alongside -i or -P.

Heads up by <leot>

Match SYNOPSIS with usage()

branches: 1.47.4; 1.47.6;
Remove workaround for ancient HTML generation code.

branches: 1.46.6;
Bump date
Reminded by wiz@

df was there from v1.
Confirmed from the TUHS & cat-v.org hosted copies of man pages.

\\ -> \e

sort options from anon ymous

PR/36817: Zafer Aydogan: Don't use -g to implement the same named option
in solaris to display statvfs output, use -G instead. I researched this
and all other implementations except Solaris (OS/X, FreeBSD, us before June
24, 2007) use -g to signify gigabytes. So revert to that.
PR/38154: YAMAMOTO Takashi: df -P doesn't work without -k. Fix from Anon Ymous
together with passing lint, and exiting with the right values.

branches: 1.41.4; 1.41.8;
PR/36541: Greg A. Woods: fix -P option, and add solaris -g option.

Improve wording of -a description. From FreeBSD via Zafer Aydogan.

PR/32103: Zafer Aydogan: Fix references to obsolete fstatfs and getfsstat

Mention -g in synopsis. PR#25800 from Kouichirou Hiratsuka.

Refer statvfs instead of removed statfs.

branches: 1.36.2;
Replace hard to understand wording with a different
one from further down in the same page. Suggested by jmc@openbsd.

comma-separated with a "-" in between. From Jared Yanovich via jmc@openbsd.

mention that -h also overrides BLOCKSIZE, and sort around there

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

Patches provided by Joel Baker in PR 22249, verified by myself.

add -g to ENVIRONMENT.

change -G arg to -g.

-G was chosen as SuSv2 specified -g was already used in ls(1), but ls
has no blocksize options like df/du, so this was redundant.

bump date in man page.

add -G argument for output in gigabytes.

ok'd by perry@, mrg@

grammar fix.

matche->match, from Igor Sobrado in PR 19670

human-readable output via -h flag. output matches Linux and other BSDs.
partly from kstailey@openbsd.org. approved itojun and perry.

New policy: New sentences start on a new line.
Patches by Robert Elz <kre at munnari oz au>, with minimal changes by me.

Generate <>& symbolically. I'm avoiding .../dist/... directories for now.

Use standard headers, sort SEE ALSO.

put -l explanation before -m, so that the options are in alphabet order
Fixes misc/14471 by Hiramatsu Yoshifumi.

precede, not preceed.

Xref getbsize(3).

Add -m option, similar to -k, but display output in 1024*1024 byte units.

Patch submitted by Chris Pinnock <C.J.E.Pinnock@qmw.ac.uk> in PR 11849,
adjusted for NetBSD-current by me.

Support the 1003.2-92 -P flag for a `portable output format'; based on
Ben Harris' patch in standards/11208 but extended about output alignment
and documentation.

Add -a flag now that we have support for MNT_IGNORE

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.

tunefs lives in section 8, not 1.

Close PR 6919, submitted by Jaromir Dolecek <jdolecek@per4mance.cz>.

branches: 1.14.2;
Fix .Nm usage.

Make df(1) more liberal (and intelligent) about what it accepts
as arguments (for -t and -l flags, specifically).
From Hubert Feyrer <feyrer@rfhs8002.fh-regensburg.de>, PR #2869.

ENVIRONMENTAL -> ENVIRONMENT

Use "utility" instead of "command". Modern definitions of these terms
are distinct (See POSIX.2 glossary).

A utility is a executable, script or shell builtin; while a command
can be any of those things plus lists, pipelines, compound commands
(if, for, while) and shell function definitions.

convert to new RCS id conventions.

Document the `-t' option as in umount(8).

Add `-t' option from Matthew Green, hacked a bit by me.

Deal with large file systems, and clean up a bit.

branches: 1.6.2;
Add `-l' option.

Incorporate new version using getbsize(), with -k flag added.

Add RCS indentifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

PR lib/59811 allow any rounding in strpct(3).

Add str[s]pct_r (same as str[s]pct with an addition "how to round" param) and
strpct_round to set the rounding used by the older str[s]pct functions.
The default remains rounding down (toward zero) for compatibility.

See the PR (and the updated man page) for the details.

For df(1) use "round away from zero" mode, as POSIX requires.

XXX pullup -11

branches: 1.104.2;
PR/58718: Christof Meerwald: "total files" and "free files" values swapped
in df -G output

Add -M and -q options to df(1).

-q suppresses warning output (rather than redirecting stderr, which
also buries error messages, which -q does not).

-M requires each arg given to be the name of a mount point, and
issues a warning, and ignores the arg, if it is not. This allows
scripts (etc) to have a whole list of places for which to produce df
output, while only listing the ones that are actually mounted (rather
than simply listing the filesystem containing the mount point instead -
perhaps several times). (If there are no args, -M is a no-op).

If the options aren't given, nothing alters.

branches: 1.102.2;
Correctly compute the "Filesystem" column with when using -W
(that added "NAME=" occupies space!)

XXX pullup -10

branches: 1.101.2;
Sync usage with man page.

Add 4 new flags:
-b (from FreeBSD) - set blocksize to blocks (512 bytes)
(overrides a contrary setting in BLOCKSIZE)
-H (from FreeBSD and Linux): -h using SI units (powers of 10). Ugh.
-N suppress the header line (except with -P which requires it).
-f show only free space (or inodes) in a minimal format (implies -N)
(that is, with one file[system] specified, print 1 number only)
With -c, show only the total.
Intended to be useful for scripting (aka, I needed it.)

While here, improve the usage message (group options where they apply,
there is no reason, for example, that -g should be shown differently
to -k -m ..., and those options aren't at all useful with -G)

Update the man page to match.

Set totals.f_frsize to DEV_BSIZE instead of totals.f_bsize so that
addstat() uses an initialised value for total size calculations.
Fixes core dump for "df -c".

df: add grand total option

Add a grand total option, -c, similar to the du(1) -c option. Adapted from
the same option (-c) in FreeBSD df(1).

- Bump Size/Used/Avail from 10 to 12, and iUsed/iAvail from 8 to 10 for double-digit terabyte storage.
- set the field width to the optimal, if "-h" is specified.
- if blocksize is greater than M or G, reduce the size field width.

field width of blocksize and inode can be changed by #define

Add a new member to struct vfsstat and grow the unused members
The new member is caled f_mntfromlabel and it is the dkw_wname
of the corresponding wedge. This is now used by df -W to display
the mountpoint name as NAME=

Print the wedge name with -W instead of mntfrom

-G cannot be specified alongside -i or -P.

Heads up by <leot>

branches: 1.92.8; 1.92.14; 1.92.16;
Correct display of df(1) with the -i parameter

Now the "Mounted on" column should be aligned with /entries.

Patch by Michal Mazurek.

PR/50422: Robert Elz: df -G prints the wrong value for fragsize (+FIX)
For df -G, print the block and fragment size instead of the iosize
and the blocksize. If we need to print the iosize, it should be done
in a different field. Nevertheless printing the blocksize in the fragment
size field is just wrong.
XXX: pullup-6, pullup-7

use strspct.

branches: 1.89.2;
Use the same column spacing for the -h output as with the normal output,
as the same header line format is used in both cases and it can't line
up correctly with both of them otherwise.

No need to print internal state once debug is done.

- static/__dead
- use strpct from libutil
- fix off by one in format

block numbers are measured in f_frsize units. Make -P option
use this instead of f_bsize. Also account for reserved blocks
like normal non-P output.
Fixes PR bin/41541.

sign-compare fix for amd64

fix -Wsign-compare issues

branches: 1.83.4; 1.83.8;
Remove the \n and tabs from the __COPYRIGHT() strings.

PR/38755: murray armfield: /bin posix programs missing setprogname and/or
setlocale

branches: 1.81.2; 1.81.4;
do the -g -G swap for real.

PR/36817: Zafer Aydogan: Don't use -g to implement the same named option
in solaris to display statvfs output, use -G instead. I researched this
and all other implementations except Solaris (OS/X, FreeBSD, us before June
24, 2007) use -g to signify gigabytes. So revert to that.
PR/38154: YAMAMOTO Takashi: df -P doesn't work without -k. Fix from Anon Ymous
together with passing lint, and exiting with the right values.

restore whitespace between 'Filesystem' and 'blocks'.

branches: 1.78.2;
prtstat: PRIdMAX -> jd as suggested by Joerg Sonnenberger.

don't cast 64bit values to u_long.

branches: 1.76.4;
kill MFSNAMELEN

Fix '-Ph' segfault; from Greg Woods

spacing fixes from Geoff Wing

PR/36541: Greg A. Woods: fix -P option, and add solaris -g option.

fix formatting of df -h; from Zafer.

-h and -[mgk] are mutually exclusive. Only use the last specified. [from zafer]
Use one variable instead of 3 flags to hold the scale, so that we don't have
to reset each flag.

branches: 1.70.2;
Handle malloc returning NULL.

PR/23688: don't say we're printing in 1k-blocks if -h was also specified.

sprinkle a little const, and now everything compiles with WARNS=3

Use the same width (9) for both the Used and Avail columns.

Drop LSBs when denominator is big.

Remove now unnecesary/incorrect cast.

- Remove unnecessary empty line.
- Wrap long line.

Available block count may be negative - so make the calculation signed.
This avoids output like:
Filesystem 1K-blocks Used Avail Capacity Mounted on
/dev/wd1a 11799 11276 9223372036854775741 100% /boot

Replace the statfs() family of system calls with statvfs().
Retain binary compatibility.

Combine two printf()s back to one. There was a reason to split in my
old patch but no reason to do so in the final version.

Prevent sign extension for positive value, and f_bavail if it is considered
as positive value.

Fix an output of df -h when the value is negative; we need to provide
space for minus sign.

No need to reference through a pointer; just refer the array of
character directly.

Don't put space before close paren.

Use signed 64bit integer type instead of unsigned long long
since f_bavail may be negative.

Backout recent undocumented change to keep an output same as before.

fix compilation on LP64 systems.

cast to unsigned long long before computing to avoid overflow

use unsigned long long to print block count.

Homogenize usage messages: make the 'usage' word all lowercase, as this seems
to be the most common practice in our tree.

Some file systems, like FAT, have no inodes in the ffs sense and don't
run out of inodes. df -i was saying, however, that such file systems
had 100% of their inodes in use, which would do things like trigger
alarms in scripts looking for file systems that have run out.

Instead, say 0% are in use, which although not strictly true is at
least less wrong, fixes scripts and is less worrying in nightly reports.

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

Patches provided by Joel Baker in PR 22249, verified by myself.

Bump ifree/iused field width.

change -G arg to -g.

-G was chosen as SuSv2 specified -g was already used in ls(1), but ls
has no blocksize options like df/du, so this was redundant.

bump date in man page.

add -G argument for output in gigabytes.

ok'd by perry@, mrg@

sync usage args with reality.

human-readable output via -h flag. output matches Linux and other BSDs.
partly from kstailey@openbsd.org. approved itojun and perry.

Following FreeBSD's example, remove the old pre-statfs(2) code for
examining FFS filesystems directly. This also gets rid of df(1)'s
setgid operator bit.

branches: 1.42.2;
if -a is not given, we need to walk the filesystem list to check the
flags in order to get the correct count (from enami)

KNF.

use getprogname()

Adjust output column in printf().

eliminate nested extern

use more standard %ll_ in favour of %q_

Add -m option, similar to -k, but display output in 1024*1024 byte units.

Patch submitted by Chris Pinnock <C.J.E.Pinnock@qmw.ac.uk> in PR 11849,
adjusted for NetBSD-current by me.

Support the 1003.2-92 -P flag for a `portable output format'; based on
Ben Harris' patch in standards/11208 but extended about output alignment
and documentation.

Add -a flag now that we have support for MNT_IGNORE

Fix overflow in df on big partitions' raw device.
Supplied by Greg Oster, problem noted in OpenBSD PR#924.

branches: 1.32.2;
Fix PR3474: df -t/df -l confused by stacked local mounts.
(Ross's patch in the PR appears to Just Work).

branches: 1.31.2;
Cast args to strpct() u_long, as in the prototype, not ulong.

Delint.

Delint.

Use mkdtemp(3).

minor KNF.

Fix PR#3590 - remove floating point calculation from df to avoid faulting
on machines that don't yet have complete FP support (e.g. MC68LC040)

Merge with Lite2 + local changes

- Use __COPYRIGHT() and __RCSID().
- Fix compiler warnings.

Make df(1) more liberal (and intelligent) about what it accepts
as arguments (for -t and -l flags, specifically).
From Hubert Feyrer <feyrer@rfhs8002.fh-regensburg.de>, PR #2869.

merge in changes from 1.1 release branch

branches: 1.21.2;
lose unused variable; PR #1297

don't assume f_fstypename is larger than MFSNAMELEN or is nul-terminated

convert to new RCS id conventions.

Use S_IS*().

Various:
* Redo `-t' handling like mount(8) and umount(8).
* Don't force a stat of every fs unless we need it.
* When using an explicit list, stat everything before printing anything, and
minimize the column width.

Add `-t' option from Matthew Green, hacked a bit by me.

Deal with large file systems, and clean up a bit.

fix typo

branches: 1.13.2;
fixup for new fs code.

copyright foo

queue/struct changes

fs type names will soon be strings

check return code. noticed while discovering a already fixed bug

when you move a fn. call to another function, you shouldn't depend
on the locals' values it used to set...

Add `-l' option.

new definition of getbsize

Incorporate new version using getbsize(), with -k flag added.

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision