copy libc's swab.c into dd as dd_swab(), and remove the restrict.

our implementation was fine, but the restrict marker is problematic
as gcc 8 is now more strict about checking for restrict issues.

this is the only actual consumer of swab(3) in our tree, though,
besides the test for it. oh well.

Drop -D_INCOMPLETE_XOPEN_C063 from dd(1)

branches: 1.18.16;
Add iflag and oflag operands to dd(1)

Like GNU dd(1) similar operands, iflag and oflag allow specifying the
O_* flags given to open(2) for the input and the output file. The values
are comma-sepratated, lower-case, O_ prefix-stripped constants documented
in open(2).

Since iflag and oflag override default values, specifying oflag means
O_CREATE is not set by default and must be specified explicitely.

Some values do not make sense (e.g.: iflag=directory) but are still used
and will raise a warning. For oflag, values rdonly, rdwr and wronly are
filtered out with a warning (dd(1) attempts open(2) with O_RDWR and
then O_WRONLY on failure).

Specifying oflag=trunc along with (seek, oseek or conv=notrunc) is
contradictory and will raise an error.

iflag and oflag are disabled if building with -DMALLPROG

branches: 1.17.12;
let the standard rules deal with librumpclient

Add a new command to dd(1): msgfmt. The command modifies the
output of the information summary returned by dd(1). This can be used
to specify messages in a more usable (or parseable) format like
human-readable values.

My intent is to re-use this for building image files and quick I/O
benchmarking.

Reviewed by tsutsui@ on tech-userlevel. See also
http://mail-index.netbsd.org/tech-userlevel/2010/12/03/msg004179.html

Some examples:

$ dd if=/dev/zero of=/dev/null bs=1m count=1 msgfmt=human
1+0 records in
1+0 records out
1048576 bytes (1,0 MB) transferred in 0.001 secs (1048576000 bytes/sec - 1,0 GB/sec)

$ dd if=/dev/zero of=/dev/null count=1 msgfmt='
> <speed>%E</speed>
> <time>%s</time>
> <bytes>%b</bytes>
> '
<speed>500 KB/sec</speed>
<time>0.001</time>
<bytes>512</bytes>

branches: 1.15.4;
Remove the rif/rof options and add rump.dd. This makes usage
consistent with other rump clients. Copying between kernels is
done using the host pipe, e.g.:

dd if=foo rof=bar skip=1 seek=1 => dd if=foo skip=1 | rump.dd of=bar seek=1

Also, the pipe idiom extends to copying between different rump
kernels, e.g.:

env RUMP_SERVER=unix://srv1 rump.dd if=thefile \
| env RUMP_SERVER=unix://srv2 rump.dd of=thefile

Pipe approach suggested by yamt (thanks!)

branches: 1.14.2;
extra crunchgen hoops

Add two new operands: "rif" and "rof". They operate exactly like
"if" and "of" with the exception that the communicate with a rump
kernel instead of the host kernel.

For example, to write stdout to /tmp/file.txt in a rump kernel namespace:
dd rof=/tmp/file.txt

copy /file1 to /file2 inside a rump kernel:
dd rif=/file1 rof=/file2

copy a snippet from /dev/rmd0d on the rump kernel to the host fs:
dd rif=/dev/rmd0d of=save seek=1000 count=3

Eat that, usermode OS.

(I'll document the operands one I have some manpage to refer to
for rump client use).

Convert to using raise_default_signal(3).

branches: 1.11.10;
sprinkle a little const, and now everything compiles with WARNS=3

convert to using strsuftoll(3) from libc

.ifdef SMALLPROG, compile with -DNO_CONV and don't compile in conv_tab.o

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

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

Fix compiler warnings.
Add WARNS=1

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

Don't use the signal handler to terminate when nothing is to do (count=0, files=0).

The signal handler tries to raise the signal again, so that a parent can retrieve
the signal from the exit code. Calling the terminate handler with signal code 0
doesn't raise a signal and dd continues with exit(127) making this case an error.

The "ibm" and "oldibm" tables are identical, because POSIX just
standardised the table from V7. Nobody, including the original authors,
seems to have noticed this. Merge them and update the documentation.
Also fix the odd, inconsistent, spelling of "pre-4.3BSD-Reno").
(From nabijaczleweli)

dd: fix lint warnings about effectively discarding 'const'

args.c(153): warning: call to 'bsearch' effectively discards 'const'
from argument [346]
args.c(383): dito
args.c(475): dito

No functional change.

adjust the open flags available for dd to match actual reality
of what matters. remove "search" for now, since O_SEARCH has
no backend. document them all.

branches: 1.39.16;
Add iflag and oflag operands to dd(1)

Like GNU dd(1) similar operands, iflag and oflag allow specifying the
O_* flags given to open(2) for the input and the output file. The values
are comma-sepratated, lower-case, O_ prefix-stripped constants documented
in open(2).

Since iflag and oflag override default values, specifying oflag means
O_CREATE is not set by default and must be specified explicitely.

Some values do not make sense (e.g.: iflag=directory) but are still used
and will raise a warning. For oflag, values rdonly, rdwr and wronly are
filtered out with a warning (dd(1) attempts open(2) with O_RDWR and
then O_WRONLY on failure).

Specifying oflag=trunc along with (seek, oseek or conv=notrunc) is
contradictory and will raise an error.

iflag and oflag are disabled if building with -DMALLPROG

branches: 1.38.6;
PR/48057: psi: skip and msgfmt cannot be used together. Looks like a pasto;
fixed.

branches: 1.37.6;
As suggested by Enami, rework the msgfmt logic a bit:
- error out when an unknown specifier is used. Do this in f_msgfmt(),
before dd(1) starts operation.
- allow buffer_write() to flush the internal buffer even when NULL is
passed as parameter.

Some whitespace fixes too.

Add a new command to dd(1): msgfmt. The command modifies the
output of the information summary returned by dd(1). This can be used
to specify messages in a more usable (or parseable) format like
human-readable values.

My intent is to re-use this for building image files and quick I/O
benchmarking.

Reviewed by tsutsui@ on tech-userlevel. See also
http://mail-index.netbsd.org/tech-userlevel/2010/12/03/msg004179.html

Some examples:

$ dd if=/dev/zero of=/dev/null bs=1m count=1 msgfmt=human
1+0 records in
1+0 records out
1048576 bytes (1,0 MB) transferred in 0.001 secs (1048576000 bytes/sec - 1,0 GB/sec)

$ dd if=/dev/zero of=/dev/null count=1 msgfmt='
> <speed>%E</speed>
> <time>%s</time>
> <bytes>%b</bytes>
> '
<speed>500 KB/sec</speed>
<time>0.001</time>
<bytes>512</bytes>

branches: 1.35.2;
Use __dead for the small build too

Remove the rif/rof options and add rump.dd. This makes usage
consistent with other rump clients. Copying between kernels is
done using the host pipe, e.g.:

dd if=foo rof=bar skip=1 seek=1 => dd if=foo skip=1 | rump.dd of=bar seek=1

Also, the pipe idiom extends to copying between different rump
kernels, e.g.:

env RUMP_SERVER=unix://srv1 rump.dd if=thefile \
| env RUMP_SERVER=unix://srv2 rump.dd of=thefile

Pipe approach suggested by yamt (thanks!)

branches: 1.33.2;
No need to cast with bsearch(), it returns a void *. Use __arraycount().

Add iseek and oseek option as aliases for skip and seek respectively.
These options exist in dd of solaris or svr3. From FreeBSD.

Copy argument before modifying it so that ps shows entire argument.
From OpenBSD via FreeBSD.

Make compile on non-NetBSD.

Wrap long line.

Hmm, for 'rof', I believe that C_ROF|C_OF is invalid rather than
C_ROF|C_ROF.

Hi pooka :)

Add two new operands: "rif" and "rof". They operate exactly like
"if" and "of" with the exception that the communicate with a rump
kernel instead of the host kernel.

For example, to write stdout to /tmp/file.txt in a rump kernel namespace:
dd rof=/tmp/file.txt

copy /file1 to /file2 inside a rump kernel:
dd rif=/file1 rof=/file2

copy a snippet from /dev/rmd0d on the rump kernel to the host fs:
dd rif=/dev/rmd0d of=save seek=1000 count=3

Eat that, usermode OS.

(I'll document the operands one I have some manpage to refer to
for rump client use).

In "dd progress=N", let the value of N control how often
to print a "." to stderr. Previously, any non-zero
value behaved like "progress=1".

PR 24300

Approved by christos

add support for conv=sparse
inspired by freebsd, although this patch attempts
to avoid some potential bugs in their implementation.

Add C_ASCII and C_EBCDIC to list of options which unset C_BS semantics.
Without this fix, the bs option would cause any of the ascii<->ebcdic
conv options to be ignored.

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

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

As discusses a little while back on tech-userlevel:

If stdout is a tty, use vis(3) to print any filenames to prevent garbage
from being printed if the filename contains control- or other non-printable
characters.

While here, sprinkle some EXIT_FAILURE and NOTREACHED where appropriate.

convert to using strsuftoll(3) from libc

a bit more cosmetic changes.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

fix WARNS=2 for -DNO_CONV

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

seperate -> separate

In the words of Our Fearless Leader, I am "utterly sick" of dd(1)
lying to me about transfer rates. Do some fixed point frobbing to get
stuff like: 2486864 bytes transferred in 0.272 secs (9142882 bytes/sec)

Add "progress=1" switch to show a sign of life by printing a '.' for
every block written to the output file.

Delint.

"static" should be located at the beginning of declaration.

Lineup `case' label and `default' label with `switch'.

Put space after the keyword `switch'.

Do multiply in lseek calls as off_t arithmetic. This allows for seeks
of more than INT_MAX. Removed Check for INT_MAX as maximum seeks.
Fixes PR 2101.

Fix compiler warnings.
Add WARNS=1

add const qualifier to args and clist tables

Re-do change to use const qualifer with conversion tables.

Fixed bug where pre-computed upper/lower case conversion tables were
used. This won't work if user selected a different locale.

Add -DNO_CONV to make this smaller in my ramdisk.

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

branches: 1.17.98;
Move UCB-licensed code from 4-clause to 3-clause licence.

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

Include stdlib.h so EXIT_FAILURE is defined.

As discusses a little while back on tech-userlevel:

If stdout is a tty, use vis(3) to print any filenames to prevent garbage
from being printed if the filename contains control- or other non-printable
characters.

While here, sprinkle some EXIT_FAILURE and NOTREACHED where appropriate.

Replace {u_,}longlong_t usage with {u,}int64_t, as the former has
"issues" and really shouldn't be used outside of the Sun XDR stuff.
Thanks again to Chuck Silvers for reminding me.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

In the words of Our Fearless Leader, I am "utterly sick" of dd(1)
lying to me about transfer rates. Do some fixed point frobbing to get
stuff like: 2486864 bytes transferred in 0.272 secs (9142882 bytes/sec)

avoid C sequence point issues warned about by development version of gcc.

errx?/warnx? audit. do not pass variable alone, use %s. idea from openbsd

branches: 1.8.10;
Delint.

Fix compiler warnings.
Add WARNS=1

Re-do change to use const qualifer with conversion tables.

Fixed bug where pre-computed upper/lower case conversion tables were
used. This won't work if user selected a different locale.

Add -DNO_CONV to make this smaller in my ramdisk.

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

The "ibm" and "oldibm" tables are identical, because POSIX just
standardised the table from V7. Nobody, including the original authors,
seems to have noticed this. Merge them and update the documentation.
Also fix the odd, inconsistent, spelling of "pre-4.3BSD-Reno").
(From nabijaczleweli)

branches: 1.9.98;
Move UCB-licensed code from 4-clause to 3-clause licence.

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

Fix compiler warnings.
Add WARNS=1

Re-do change to use const qualifer with conversion tables.

Fixed bug where pre-computed upper/lower case conversion tables were
used. This won't work if user selected a different locale.

Previous change (Use const qualifier with conversion tables) is broken.
(1) It won't compile, and (2) there's code in dd.c which actually
does modify (some, all, used?) conversion tables, so they can't _all_ be
const (if any can).

Use const qualifier with conversion tables

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

The "ibm" and "oldibm" tables are identical, because POSIX just
standardised the table from V7. Nobody, including the original authors,
seems to have noticed this. Merge them and update the documentation.
Also fix the odd, inconsistent, spelling of "pre-4.3BSD-Reno").
(From nabijaczleweli)

dd.1: add an example of writing a NetBSD image

branches: 1.36.6;
Fix typos; use American spelling consistently. Remove an unnecessary macro.

adjust the open flags available for dd to match actual reality
of what matters. remove "search" for now, since O_SEARCH has
no backend. document them all.

branches: 1.34.6;
Use Ex.

Remove workaround for ancient HTML generation code.

bump date

dd first appeared in V5, not V1
Heads up by Ingo Schwarze
http://www.tuhs.org/cgi-bin/utree.pl?file=V5/usr/source/s1/dd.c

Document the version dd first appeared.
Bump date.

Pp not required before Bd, picked up by mandoc lint

The macro to cross-reference is Xr, not Xt.
Fix typo error picked up by mandoc lint.

New sentence, new line. Fix typos. Bump date for previous.

Add iflag and oflag operands to dd(1)

Like GNU dd(1) similar operands, iflag and oflag allow specifying the
O_* flags given to open(2) for the input and the output file. The values
are comma-sepratated, lower-case, O_ prefix-stripped constants documented
in open(2).

Since iflag and oflag override default values, specifying oflag means
O_CREATE is not set by default and must be specified explicitely.

Some values do not make sense (e.g.: iflag=directory) but are still used
and will raise a warning. For oflag, values rdonly, rdwr and wronly are
filtered out with a warning (dd(1) attempts open(2) with O_RDWR and
then O_WRONLY on failure).

Specifying oflag=trunc along with (seek, oseek or conv=notrunc) is
contradictory and will raise an error.

iflag and oflag are disabled if building with -DMALLPROG

branches: 1.25.12;
- sync `SYNOPSIS' with dd.c's `usage' line;
- correct spelling (prepositions).

From Bug Hunting.

Add a new command to dd(1): msgfmt. The command modifies the
output of the information summary returned by dd(1). This can be used
to specify messages in a more usable (or parseable) format like
human-readable values.

My intent is to re-use this for building image files and quick I/O
benchmarking.

Reviewed by tsutsui@ on tech-userlevel. See also
http://mail-index.netbsd.org/tech-userlevel/2010/12/03/msg004179.html

Some examples:

$ dd if=/dev/zero of=/dev/null bs=1m count=1 msgfmt=human
1+0 records in
1+0 records out
1048576 bytes (1,0 MB) transferred in 0.001 secs (1048576000 bytes/sec - 1,0 GB/sec)

$ dd if=/dev/zero of=/dev/null count=1 msgfmt='
> <speed>%E</speed>
> <time>%s</time>
> <bytes>%b</bytes>
> '
<speed>500 KB/sec</speed>
<time>0.001</time>
<bytes>512</bytes>

branches: 1.23.6;
Add iseek and oseek option as aliases for skip and seek respectively.
These options exist in dd of solaris or svr3. From FreeBSD.

Don't workaround macro argument limit in ancient troff.

branches: 1.21.20;
Use IEC 60027-2 prefixes for 2^n based prefixes.

Bump date for progress=n description change.

In "dd progress=N", let the value of N control how often
to print a "." to stderr. Previously, any non-zero
value behaved like "progress=1".

PR 24300

Approved by christos

Spell the plural of suffix "suffixes", not "suffices".
Inspired by PR 24400 by Todd Vierling.

add support for conv=sparse
inspired by freebsd, although this patch attempts
to avoid some potential bugs in their implementation.

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

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

.Nm does not need a dummy argument ("") before punctuation or
for correct formatting of the SYNOPSIS any longer.

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

Drop a space and an unnecessary .Pp.

fix conv= description

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

For commands and utilities, use EXIT STATUS rather than RETURN VALUES as
appropriate (and documented in mdoc(7)).

Add 'RETURN VALUE' section header.

branches: 1.8.8;
Add "progress=1" switch to show a sign of life by printing a '.' for
every block written to the output file.

macroize BSD, NetBSD, FreeBSD and misc cleanup

Fix .Nm usage.

convert to new RCS id conventions.

ECBDIC->EBCDIC

Merge 4.4-Lite version.

Add RCS indentifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

copy libc's swab.c into dd as dd_swab(), and remove the restrict.

our implementation was fine, but the restrict marker is problematic
as gcc 8 is now more strict about checking for restrict issues.

this is the only actual consumer of swab(3) in our tree, though,
besides the test for it. oh well.

When buffers are at least page sized, explicitely request page alignment.

branches: 1.51.14;
Drop main() prototype.

Add iflag and oflag operands to dd(1)

Like GNU dd(1) similar operands, iflag and oflag allow specifying the
O_* flags given to open(2) for the input and the output file. The values
are comma-sepratated, lower-case, O_ prefix-stripped constants documented
in open(2).

Since iflag and oflag override default values, specifying oflag means
O_CREATE is not set by default and must be specified explicitely.

Some values do not make sense (e.g.: iflag=directory) but are still used
and will raise a warning. For oflag, values rdonly, rdwr and wronly are
filtered out with a warning (dd(1) attempts open(2) with O_RDWR and
then O_WRONLY on failure).

Specifying oflag=trunc along with (seek, oseek or conv=notrunc) is
contradictory and will raise an error.

iflag and oflag are disabled if building with -DMALLPROG

branches: 1.49.12;
Fix a very old bug. When allocating the buffer and doing just a bs= transfer,
hen we only need a single buffer equal to that blocksize in length.

Add a new command to dd(1): msgfmt. The command modifies the
output of the information summary returned by dd(1). This can be used
to specify messages in a more usable (or parseable) format like
human-readable values.

My intent is to re-use this for building image files and quick I/O
benchmarking.

Reviewed by tsutsui@ on tech-userlevel. See also
http://mail-index.netbsd.org/tech-userlevel/2010/12/03/msg004179.html

Some examples:

$ dd if=/dev/zero of=/dev/null bs=1m count=1 msgfmt=human
1+0 records in
1+0 records out
1048576 bytes (1,0 MB) transferred in 0.001 secs (1048576000 bytes/sec - 1,0 GB/sec)

$ dd if=/dev/zero of=/dev/null count=1 msgfmt='
> <speed>%E</speed>
> <time>%s</time>
> <bytes>%b</bytes>
> '
<speed>500 KB/sec</speed>
<time>0.001</time>
<bytes>512</bytes>

branches: 1.47.4;
Remove the rif/rof options and add rump.dd. This makes usage
consistent with other rump clients. Copying between kernels is
done using the host pipe, e.g.:

dd if=foo rof=bar skip=1 seek=1 => dd if=foo skip=1 | rump.dd of=bar seek=1

Also, the pipe idiom extends to copying between different rump
kernels, e.g.:

env RUMP_SERVER=unix://srv1 rump.dd if=thefile \
| env RUMP_SERVER=unix://srv2 rump.dd of=thefile

Pipe approach suggested by yamt (thanks!)

branches: 1.46.2;
Fix speling ("deferred," not "defered") as reported in #netbsd-code on
freenode IRC. While I'm here, clean up the wording later in the comment.

remove some leftover development garbage

Add two new operands: "rif" and "rof". They operate exactly like
"if" and "of" with the exception that the communicate with a rump
kernel instead of the host kernel.

For example, to write stdout to /tmp/file.txt in a rump kernel namespace:
dd rof=/tmp/file.txt

copy /file1 to /file2 inside a rump kernel:
dd rif=/file1 rof=/file2

copy a snippet from /dev/rmd0d on the rump kernel to the host fs:
dd rif=/dev/rmd0d of=save seek=1000 count=3

Eat that, usermode OS.

(I'll document the operands one I have some manpage to refer to
for rump client use).

fix -Wsign-compare issues

branches: 1.42.6; 1.42.10;
Remove the \n and tabs from the __COPYRIGHT() strings.

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

branches: 1.40.10; 1.40.12;
fix typos

In "dd progress=N", let the value of N control how often
to print a "." to stderr. Previously, any non-zero
value behaved like "progress=1".

PR 24300

Approved by christos

Fixed an off-by-one error with ASCII <-> EBCDIC conversion combined with
lcase/ucase. Approved by mrg.

minor nit, set out.dbp correctly when constructing final null block
for sparse files.

add support for conv=sparse
inspired by freebsd, although this patch attempts
to avoid some potential bugs in their implementation.

Optimise previous change, using fcntl(F_DUPFD) to locate a free descriptor
for us, rather than iterating until satisfied.

Suggested by David Laight.

Avoid corrupting the dd(1) IO streams. This would happen by accidentally
outputting to the files being manipulated by opening a file in the standard IO
descriptor space. In particular, an output file unlucky enough to be sitting
on descriptor 2 (stderr) is certain to be corrupted.

Addresses PR bin/8521, and passes the recently committed regression test
"bin/dd".

Following private discussion with kleink@ and hubertf@ and public discussion
on tech-userlevel@, back out printescaped() functionality.
kleink: ``We sell rope.''

Fix memory leak noted by Hubert Feyrer in private email.
Patch by Hubert Feyrer as well, OK by kleink.
(I'm just fixing it b/c it's my fault to begin with.)

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

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

As discusses a little while back on tech-userlevel:

If stdout is a tty, use vis(3) to print any filenames to prevent garbage
from being printed if the filename contains control- or other non-printable
characters.

While here, sprinkle some EXIT_FAILURE and NOTREACHED where appropriate.

Add comment about my rev. 1.27 change.

Ignore EINVAL returned from fsync(2) because it is the expected error code
if we try to use it on a pipe.

Detect the last nfs write error.

Convert some variables that were signed ints before my update to 64 bit sizes
from uint64_t to int64_t. Fixes [bin/15187]

a bit more cosmetic changes.

Replace {u_,}longlong_t usage with {u,}int64_t, as the former has
"issues" and really shouldn't be used outside of the Sun XDR stuff.
Thanks again to Chuck Silvers for reminding me.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

fix WARNS=2 for -DNO_CONV

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

In the words of Our Fearless Leader, I am "utterly sick" of dd(1)
lying to me about transfer rates. Do some fixed point frobbing to get
stuff like: 2486864 bytes transferred in 0.272 secs (9142882 bytes/sec)

KNF nit in previous.

Make sure we accept the getopt(3) 'last option' delimiter;
fixes PR standards/11229.

Fix conv=swab. Don't swab the entire input buffer every time we read
another block, just swab the block we just read.

rename nset to infoset from jhawk.

block siginfo during writes.

branches: 1.14.4;
Since our gcc doesn't warn about NULL format strings anymore, we can
fix the incorrect err(1, "%s", "") et al.
Closes PR bin/7592 by cgd.

branches: 1.13.4;
Add "progress=1" switch to show a sign of life by printing a '.' for
every block written to the output file.

Add some braces to make egcs happy.

Delint.

Delint.

Fix a minor bug that made conv=sync pad partial blocks with garbage,
when it is supposd to pad with zero or space as with osync.

Need <time.h> fo time() prototype.

Fix compiler warnings.
Add WARNS=1

Re-do change to use const qualifer with conversion tables.

Fixed bug where pre-computed upper/lower case conversion tables were
used. This won't work if user selected a different locale.

Add -DNO_CONV to make this smaller in my ramdisk.

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

branches: 1.16.16;
Add iflag and oflag operands to dd(1)

Like GNU dd(1) similar operands, iflag and oflag allow specifying the
O_* flags given to open(2) for the input and the output file. The values
are comma-sepratated, lower-case, O_ prefix-stripped constants documented
in open(2).

Since iflag and oflag override default values, specifying oflag means
O_CREATE is not set by default and must be specified explicitely.

Some values do not make sense (e.g.: iflag=directory) but are still used
and will raise a warning. For oflag, values rdonly, rdwr and wronly are
filtered out with a warning (dd(1) attempts open(2) with O_RDWR and
then O_WRONLY on failure).

Specifying oflag=trunc along with (seek, oseek or conv=notrunc) is
contradictory and will raise an error.

iflag and oflag are disabled if building with -DMALLPROG

branches: 1.15.24;
Remove the rif/rof options and add rump.dd. This makes usage
consistent with other rump clients. Copying between kernels is
done using the host pipe, e.g.:

dd if=foo rof=bar skip=1 seek=1 => dd if=foo skip=1 | rump.dd of=bar seek=1

Also, the pipe idiom extends to copying between different rump
kernels, e.g.:

env RUMP_SERVER=unix://srv1 rump.dd if=thefile \
| env RUMP_SERVER=unix://srv2 rump.dd of=thefile

Pipe approach suggested by yamt (thanks!)

branches: 1.14.2;
Fix most inopportune typo. from gson & jmmv.

Add two new operands: "rif" and "rof". They operate exactly like
"if" and "of" with the exception that the communicate with a rump
kernel instead of the host kernel.

For example, to write stdout to /tmp/file.txt in a rump kernel namespace:
dd rof=/tmp/file.txt

copy /file1 to /file2 inside a rump kernel:
dd rif=/file1 rof=/file2

copy a snippet from /dev/rmd0d on the rump kernel to the host fs:
dd rif=/dev/rmd0d of=save seek=1000 count=3

Eat that, usermode OS.

(I'll document the operands one I have some manpage to refer to
for rump client use).

add support for conv=sparse
inspired by freebsd, although this patch attempts
to avoid some potential bugs in their implementation.

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

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

Convert some variables that were signed ints before my update to 64 bit sizes
from uint64_t to int64_t. Fixes [bin/15187]

Replace {u_,}longlong_t usage with {u,}int64_t, as the former has
"issues" and really shouldn't be used outside of the Sun XDR stuff.
Thanks again to Chuck Silvers for reminding me.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

In the words of Our Fearless Leader, I am "utterly sick" of dd(1)
lying to me about transfer rates. Do some fixed point frobbing to get
stuff like: 2486864 bytes transferred in 0.272 secs (9142882 bytes/sec)

Use u_quad_t instead of u_long to define struct STAT.bytes, which
is used to count bytes transfered, so that it won't overflow even if
transfer is more than 4Gbytes.

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

branches: 1.1.2; 1.1.48;
Remove the rif/rof options and add rump.dd. This makes usage
consistent with other rump clients. Copying between kernels is
done using the host pipe, e.g.:

dd if=foo rof=bar skip=1 seek=1 => dd if=foo skip=1 | rump.dd of=bar seek=1

Also, the pipe idiom extends to copying between different rump
kernels, e.g.:

env RUMP_SERVER=unix://srv1 rump.dd if=thefile \
| env RUMP_SERVER=unix://srv2 rump.dd of=thefile

Pipe approach suggested by yamt (thanks!)

branches: 1.1.2; 1.1.48;
Remove the rif/rof options and add rump.dd. This makes usage
consistent with other rump clients. Copying between kernels is
done using the host pipe, e.g.:

dd if=foo rof=bar skip=1 seek=1 => dd if=foo skip=1 | rump.dd of=bar seek=1

Also, the pipe idiom extends to copying between different rump
kernels, e.g.:

env RUMP_SERVER=unix://srv1 rump.dd if=thefile \
| env RUMP_SERVER=unix://srv2 rump.dd of=thefile

Pipe approach suggested by yamt (thanks!)

PR/56443: RVP: dd conv=swab doesn't always work

branches: 1.1.4;
copy libc's swab.c into dd as dd_swab(), and remove the restrict.

our implementation was fine, but the restrict marker is problematic
as gcc 8 is now more strict about checking for restrict issues.

this is the only actual consumer of swab(3) in our tree, though,
besides the test for it. oh well.

update header too.

copy libc's swab.c into dd as dd_swab(), and remove the restrict.

our implementation was fine, but the restrict marker is problematic
as gcc 8 is now more strict about checking for restrict issues.

this is the only actual consumer of swab(3) in our tree, though,
besides the test for it. oh well.

branches: 1.23.16;
Add iflag and oflag operands to dd(1)

Like GNU dd(1) similar operands, iflag and oflag allow specifying the
O_* flags given to open(2) for the input and the output file. The values
are comma-sepratated, lower-case, O_ prefix-stripped constants documented
in open(2).

Since iflag and oflag override default values, specifying oflag means
O_CREATE is not set by default and must be specified explicitely.

Some values do not make sense (e.g.: iflag=directory) but are still used
and will raise a warning. For oflag, values rdonly, rdwr and wronly are
filtered out with a warning (dd(1) attempts open(2) with O_RDWR and
then O_WRONLY on failure).

Specifying oflag=trunc along with (seek, oseek or conv=notrunc) is
contradictory and will raise an error.

iflag and oflag are disabled if building with -DMALLPROG

branches: 1.22.20;
As suggested by Enami, rework the msgfmt logic a bit:
- error out when an unknown specifier is used. Do this in f_msgfmt(),
before dd(1) starts operation.
- allow buffer_write() to flush the internal buffer even when NULL is
passed as parameter.

Some whitespace fixes too.

Add a new command to dd(1): msgfmt. The command modifies the
output of the information summary returned by dd(1). This can be used
to specify messages in a more usable (or parseable) format like
human-readable values.

My intent is to re-use this for building image files and quick I/O
benchmarking.

Reviewed by tsutsui@ on tech-userlevel. See also
http://mail-index.netbsd.org/tech-userlevel/2010/12/03/msg004179.html

Some examples:

$ dd if=/dev/zero of=/dev/null bs=1m count=1 msgfmt=human
1+0 records in
1+0 records out
1048576 bytes (1,0 MB) transferred in 0.001 secs (1048576000 bytes/sec - 1,0 GB/sec)

$ dd if=/dev/zero of=/dev/null count=1 msgfmt='
> <speed>%E</speed>
> <time>%s</time>
> <bytes>%b</bytes>
> '
<speed>500 KB/sec</speed>
<time>0.001</time>
<bytes>512</bytes>

branches: 1.20.2;
Use __dead for the small build too

Use __dead

Add two new operands: "rif" and "rof". They operate exactly like
"if" and "of" with the exception that the communicate with a rump
kernel instead of the host kernel.

For example, to write stdout to /tmp/file.txt in a rump kernel namespace:
dd rof=/tmp/file.txt

copy /file1 to /file2 inside a rump kernel:
dd rif=/file1 rof=/file2

copy a snippet from /dev/rmd0d on the rump kernel to the host fs:
dd rif=/dev/rmd0d of=save seek=1000 count=3

Eat that, usermode OS.

(I'll document the operands one I have some manpage to refer to
for rump client use).

In "dd progress=N", let the value of N control how often
to print a "." to stderr. Previously, any non-zero
value behaved like "progress=1".

PR 24300

Approved by christos

Following private discussion with kleink@ and hubertf@ and public discussion
on tech-userlevel@, back out printescaped() functionality.
kleink: ``We sell rope.''

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

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

As discusses a little while back on tech-userlevel:

If stdout is a tty, use vis(3) to print any filenames to prevent garbage
from being printed if the filename contains control- or other non-printable
characters.

While here, sprinkle some EXIT_FAILURE and NOTREACHED where appropriate.

Replace {u_,}longlong_t usage with {u,}int64_t, as the former has
"issues" and really shouldn't be used outside of the Sun XDR stuff.
Thanks again to Chuck Silvers for reminding me.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

block siginfo during writes.

branches: 1.9.8;
Add "progress=1" switch to show a sign of life by printing a '.' for
every block written to the output file.

Fix compiler warnings.
Add WARNS=1

Re-do change to use const qualifer with conversion tables.

Fixed bug where pre-computed upper/lower case conversion tables were
used. This won't work if user selected a different locale.

Previous change (Use const qualifier with conversion tables) is broken.
(1) It won't compile, and (2) there's code in dd.c which actually
does modify (some, all, used?) conversion tables, so they can't _all_ be
const (if any can).

Use const qualifier with conversion tables

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

dd: clean up function summary

No functional change. Ideally the binary would have stayed the same, but
GCC 10.3.0 inserts a nop and reorders the code.

dd: fix lint error about 'return expr' from void function

No change to the resulting binary. The 'return' statements are necessary
for GCC to generate the exact same object code, even though they can be
removed without affecting the functionality, as seen before the 'else'.

bin: remove unnecessary lint comment CONSTCOND

Since 2021-01-31, lint no longer warns about 'do ... while (0)'.

No functional change.

branches: 1.23.42;
As suggested by Enami, rework the msgfmt logic a bit:
- error out when an unknown specifier is used. Do this in f_msgfmt(),
before dd(1) starts operation.
- allow buffer_write() to flush the internal buffer even when NULL is
passed as parameter.

Some whitespace fixes too.

Add a new command to dd(1): msgfmt. The command modifies the
output of the information summary returned by dd(1). This can be used
to specify messages in a more usable (or parseable) format like
human-readable values.

My intent is to re-use this for building image files and quick I/O
benchmarking.

Reviewed by tsutsui@ on tech-userlevel. See also
http://mail-index.netbsd.org/tech-userlevel/2010/12/03/msg004179.html

Some examples:

$ dd if=/dev/zero of=/dev/null bs=1m count=1 msgfmt=human
1+0 records in
1+0 records out
1048576 bytes (1,0 MB) transferred in 0.001 secs (1048576000 bytes/sec - 1,0 GB/sec)

$ dd if=/dev/zero of=/dev/null count=1 msgfmt='
> <speed>%E</speed>
> <time>%s</time>
> <bytes>%b</bytes>
> '
<speed>500 KB/sec</speed>
<time>0.001</time>
<bytes>512</bytes>

branches: 1.21.28;
Convert to using raise_default_signal(3).

branches: 1.20.4;
In the SIGINT handler, explicitly display a summary and then raise the
default SIGINT handler to exit, rather than exit(0) and relying upon the
atexit() to display the summary.
Per SUS6.

add support for conv=sparse
inspired by freebsd, although this patch attempts
to avoid some potential bugs in their implementation.

remove unused code left from printescaped() backput as pointed out by Jeff
Ito in PR bin/23969 and PR bin/23970.

Following private discussion with kleink@ and hubertf@ and public discussion
on tech-userlevel@, back out printescaped() functionality.
kleink: ``We sell rope.''

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

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

As discusses a little while back on tech-userlevel:

If stdout is a tty, use vis(3) to print any filenames to prevent garbage
from being printed if the filename contains control- or other non-printable
characters.

While here, sprinkle some EXIT_FAILURE and NOTREACHED where appropriate.

a bit more cosmetic changes.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

In the words of Our Fearless Leader, I am "utterly sick" of dd(1)
lying to me about transfer rates. Do some fixed point frobbing to get
stuff like: 2486864 bytes transferred in 0.272 secs (9142882 bytes/sec)

use more standard %ll_ in favour of %q_

Add "progress=1" switch to show a sign of life by printing a '.' for
every block written to the output file.

Delint.

quads must be cast to long long for printing.

Use u_quad_t instead of u_long to define struct STAT.bytes, which
is used to count bytes transfered, so that it won't overflow even if
transfer is more than 4Gbytes.

Fix compiler warnings.
Add WARNS=1

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

branches: 1.18.48;
Add two new operands: "rif" and "rof". They operate exactly like
"if" and "of" with the exception that the communicate with a rump
kernel instead of the host kernel.

For example, to write stdout to /tmp/file.txt in a rump kernel namespace:
dd rof=/tmp/file.txt

copy /file1 to /file2 inside a rump kernel:
dd rif=/file1 rof=/file2

copy a snippet from /dev/rmd0d on the rump kernel to the host fs:
dd rif=/dev/rmd0d of=save seek=1000 count=3

Eat that, usermode OS.

(I'll document the operands one I have some manpage to refer to
for rump client use).

fix -Wsign-compare issues

branches: 1.16.40;
Following private discussion with kleink@ and hubertf@ and public discussion
on tech-userlevel@, back out printescaped() functionality.
kleink: ``We sell rope.''

Minor KNF nit from rev. 1.13.

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

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

As discusses a little while back on tech-userlevel:

If stdout is a tty, use vis(3) to print any filenames to prevent garbage
from being printed if the filename contains control- or other non-printable
characters.

While here, sprinkle some EXIT_FAILURE and NOTREACHED where appropriate.

a bit more cosmetic changes.

- ANSI KNF
- WARNS=2 (mainly converting some "char *" -> "const char *")

In the words of Our Fearless Leader, I am "utterly sick" of dd(1)
lying to me about transfer rates. Do some fixed point frobbing to get
stuff like: 2486864 bytes transferred in 0.272 secs (9142882 bytes/sec)

block siginfo during writes.

branches: 1.8.6;
fix comment regarding when we attempt to seek

Use lseek(2) on character devices, damn it.

branches: 1.6.4; 1.6.6;
Do multiply in lseek calls as off_t arithmetic. This allows for seeks
of more than INT_MAX. Removed Check for INT_MAX as maximum seeks.
Fixes PR 2101.

Fix compiler warnings.
Add WARNS=1

convert to new RCS id conventions.

Merge 4.4-Lite version.

Add RCS identifiers.

branches: 1.1.1;
added dd(1) from ftp.uu.net:bsd-sources; torching GNU dd

convert to using strsuftoll(3) from libc

Add LL to 1099511627776 since it won't fit in a long on ILP32 systems.

branches: 1.7.2;
Roll in fixes to permit cross-compiling from non-NetBSD hosts. This
round has been tested on Solaris/x86 and Linux hosts.

* Add host tools cap_mkdb, ctags, m4, uudecode.
* Protect __RCSID() and __COPYRIGHT() better.
* Reduce the number of places that need to include "config.h", to keep
sources closer to their "vanilla" versions.
* Add more compat #defines and autoconf-checked functions.

Blah, oversight. Revert previous; clean up to proper syntax, and add
conditional for <err.h>.

This doesn't appear to be used in src/tools any longer; remove the config.h
glue bits.

add config.h when compiled as host tool.

Deal with LP64 printf format issues.

Replace {u_,}longlong_t usage with {u,}int64_t, as the former has
"issues" and really shouldn't be used outside of the Sun XDR stuff.
Thanks again to Chuck Silvers for reminding me.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO

convert to using strsuftoll(3) from libc

Replace {u_,}longlong_t usage with {u,}int64_t, as the former has
"issues" and really shouldn't be used outside of the Sun XDR stuff.
Thanks again to Chuck Silvers for reminding me.

- Use u_longlong_t instead of u_quad_t, u_long, or int for various buffer sizes
- Add strsuftoull(), which parses a number into a u_longlong_t, with
multiplication support, and support for 'g' (GB) and 't' (TB) suffices.
If an error occurs, print to stderr and exit.
Based on get_blk() from args.c and strsufto*() (in other programs)
- Add strsuftoullx(), which acts as per strsuftoull() but returns the
error in the supplied buffer instead (if the returned buffer != "", an
error occurred)
- Replace get_bsz() use with strsuftoull()
- Remove (now) unnecessary argument validation
- Remove unused {f,p,s,t}_stats fields in struct IO