bsd.own.mk: rename GCC_NO_* to CC_WNO_*

Rename compiler-warning-disable variables from
GCC_NO_warning
to
CC_WNO_warning
where warning is the full warning name as used by the compiler.

GCC_NO_IMPLICIT_FALLTHRU is CC_WNO_IMPLICIT_FALLTHROUGH

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

add support for new GCC 9 warnings that may be too much to fix
right now. new address-of-packed-member and format-overflow
warnings have new GCC_NO_ADDR_OF_PACKED_MEMBER amd
GCC_NO_FORMAT_OVERFLOW variables to remove these warnings.

apply to a bunch of the tree. mostly, these are real bugs that
should be fixed, but in many cases, only by removing the 'packed'
attribute from some structure that doesn't really need it. (i
looked at many different ones, and while perhaps 60-80% were
already properly aligned, it wasn't clear to me that the uses
were always coming from sane data vs network alignment, so it
doesn't seem safe to remove packed without careful research for
each affect struct.) clang already warned (and was not erroring)
for many of these cases, but gcc picked up dozens more.

introduce some common variables for use in GCC warning disables:

GCC_NO_FORMAT_TRUNCATION -Wno-format-truncation (GCC 7/8)
GCC_NO_STRINGOP_TRUNCATION -Wno-stringop-truncation (GCC 8)
GCC_NO_STRINGOP_OVERFLOW -Wno-stringop-overflow (GCC 8)
GCC_NO_CAST_FUNCTION_TYPE -Wno-cast-function-type (GCC 8)

use these to turn off warnings for most GCC-8 complaints. many
of these are false positives, most of the real bugs are already
commited, or are yet to come.


we plan to introduce versions of (some?) of these that use the
"-Wno-error=" form, which still displays the warnings but does
not make it an error, and all of the above will be re-considered
as either being "fix me" (warning still displayed) or "warning
is wrong."

branches: 1.39.32;
Don't depend on HAVE_GCC being always defined.

use getfsspecname()

branches: 1.37.2;
remove most of the remaining HAVE_GCC tests that are always true in
the modern world.

branches: 1.36.4;
Enable WARNS=4 by default except for:
dump dump_lfs fsck_ffs fsck_lfs fsdb mount_smbfs
newfs_ext2fs newfs_lfs resize_lfs setkey

branches: 1.35.4;
Wrap compiler-specific flags with HAVE_GCC and HAVE_PCC as necessary. Add a few flags for PCC.

Remove adding -g to CFLAGS and LDFLAGS. Ancient debug bogon?

branches: 1.33.20; 1.33.24;
sprinkle some -fno-strict-aliasing and -Wno-pointer-sign with GCC4.

Snapshot support for dump(8):

- New option `-x backup' takes the dump from a snapshot backed up by `backup'.
The snapshot will be deleted on exit.

- New option `-X' as a synonym for `-x mountpoint' where `mountpoint' is the
file system to be dumped.

Reviewed and Approved by: Manuel Bouyer <bouyer@netbsd.org>

branches: 1.31.6;
Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Use "wall -g operator" (instead of private code) to broadcast() messages to
members of the operator group.
Don't install "setgid tty", and remove now unnecessary gid/egid swapping.
Remove utmp trawling code pulled in from usr.bin/who.
The Code is now simpler, and more portable (without the utmp cruft) too.

This is derived from similar work in OpenBSD.

Use ${NETBSDSRCDIR}/some/path instead of ${.CURDIR}/../../some/path

support utmpx

document more of the debug options

nuke trailing whitespace

Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

branches: 1.22.4;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

Handle "nodump" flag on directories, by not dumping any files or directories
under it. Based on some parts of PR 6705 by Brian Grayson.
In the Makefile, add (commented out) debug options that can be turned on.

#include machine/bswap.h and remove -lutil.

Add support for non-native byte order FFS. The dump is in filesystem byte
order, restore already knows how to byteswap dumps.

CFLAGS->CPPFLAGS

enable rcmd(). garbage collect.

branches: 1.16.4;
convert to new RCS Id conventions; reduce my headache

specify man pages the new way.

Update from 4.4-Lite, with local changes.

new version from CSRG, via BSDI, with fixes

from Thomas Eberhardt <thomas@mathematik.uni-Bremen.de>:
Some reordering and modifications in the Makefiles for sh, dump,
restore to get dependicies right. The README in /usr/share/mk
states that it's a Bad Thing(tm) to add something after
.include <bsd.prog.mk>, but this seems the only way to get the
dependencies right.

oops again; fix that last

oops; fix that last

add LDSTATIC

Add RCS identifiers.

Use ${COPY}, not -c for install.

Okay, so I was confused (but only a little).

Make rdump target match normal program target.
(This should probably be in a separate directory, no?)

make install copy

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Add -D flag to allow the user to specify an alternate dumpdates file.
Closes PR #54469.

Bump date for previous.

Add -U flag to dump(8) and dump_lfs(8) to specify dumpdates entry

This address situations where dump(8) cannot figure out the device being
dumped. It also allows tracking of subvolume dumps by using virtual
device as dumpdates entry.

Document what rdump is (was once) for. PR 53442.

branches: 1.68.6; 1.68.12; 1.68.14;
dump was there from v4.
Confirmed from the TUHS hosted copies of man pages.
Bump date.

Default the read block size (-k default value) to kern.maxphys (usually
64k these days). This gives a noticable performance boost on large filesystems.

Fix typos

branches: 1.65.10;
Use more markup, fix English usage.

Sort option descriptions a bit more.

PR/45735: Bug Hunting: The dump(8) manpage, as well as the program's `usage'-line,
needs improvement

branches: 1.62.6; 1.62.8;
File system snapshots are no longer experimental.

Use a list for the list of bugs.

-i does not take an argument. Remove unnecessary Bk/Ek. New sentence, new line.

PR/42883: Greywolf: Add -i flag which brings "true incremental" capability.

Use indent-synopsis instead of iS register. The latter is a register
used in BSD derived .Nm implementation.

We don't need to dump WAPBL log files - don't dump files with SF_LOG
flag set, and document this.

branches: 1.56.20; 1.56.24;
Argument is read-blocksize, not "read blocksize". From YOMURA Masanori in private mail.

Sort SEE ALSO.

Snapshot support for dump(8):

- New option `-x backup' takes the dump from a snapshot backed up by `backup'.
The snapshot will be deleted on exit.

- New option `-X' as a synonym for `-x mountpoint' where `mountpoint' is the
file system to be dumped.

Reviewed and Approved by: Manuel Bouyer <bouyer@netbsd.org>

branches: 1.53.6;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22308, verified by myself.

* Document $RCMD_CMD in environ(7).
* Cross-reference rcmd(1), rcmd(3), and environ(7) as appropriate.

Should fix [bin/21670] from Geoff Wing.

Use "wall -g operator" (instead of private code) to broadcast() messages to
members of the operator group.
Don't install "setgid tty", and remove now unnecessary gid/egid swapping.
Remove utmp trawling code pulled in from usr.bin/who.
The Code is now simpler, and more portable (without the utmp cruft) too.

This is derived from similar work in OpenBSD.

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

"Utilize" has exactly the same meaning as "use," but it is more
difficult to read and understand. Most manuals of English style
therefore say that you should use "use".

New sentence, new line. By Robert Elz with minimal fixes.

Typo fix, from jslag@prop.walkerart.org via OpenBSD.

Fix typo in macro.

For -l: specify the timeout on the command line (in seconds) instead of
hardwiring it to 2s, as suggested on tech-userlevel.

Add -a to "auto-size" the tape, rather than relying upon other options
to define the tape size. Requires the tape driver to either return ENOSPC
at end of media, or 0 when a write is attempted (such as the "early warning"
support in st(4) enabled with "mt eew 1"). From FreeBSD.

Add a -l (autoload) flag. For multivolume dumps, this makes dump eject the
tape when a volume is full, and try to reopen the tape drive for 2 mn.
To be used with tape changers which load the next tape when the current one is
ejected.
While I'm there fix eject handling for remote tape.

One '\' too many.

relax the restriction on -F that the file system image argument must be a
regular file

Whitespace fixes, sort SEE ALSO, sort sections.

Add the capability for dump to print timestamps on all informational
messages.

- implement -F, which indicates that files-to-dump is a regular file
containing a file system image (instead of being a single file as
part of a subtree dump)
- use "file system" instead of "filesystem" or "file-system"

Drop arguments of .Os.

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

Add a new option "e" to "dump" which allows to eject tapes automatically
if a tape change is required.

branches: 1.34.8;
Note the default tape device is taken from _PATH_DEFTAPE in /usr/include/paths.h

Default device is nrst0 not rst0

branches: 1.32.2;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

Handle "nodump" flag on directories, by not dumping any files or directories
under it. Based on some parts of PR 6705 by Brian Grayson.
In the Makefile, add (commented out) debug options that can be turned on.

Clean up SYNOPSIS formatting.

s/1998/1999/

add support for '-L label', which file(1) and restore(1) can extract
from the header.
mostly based on [bin/6715] from Brian Grayson <bgrayson@ece.utexas.edu>

Add reference to "st(4)" as suggested by Brian Grayson in PR bin/6713.

A dump level above 0 saves new or modified files since the last dump of
a lower level, not the same level. Fixes pr 5402 by Kevin Sullivan.

PR#5697 + nroff cleanup

fix bad .Xr references

Change occurrences of "UNIX" to .Ux or .At as appropriate.

add missing .Nm sections.

branches: 1.21.2;
Fix .Nm usage.

Various changes from Brian Grayson <bgrayson@ece.utexas.edu> in [bin/4103]
* exit after providing an estimate if -S was given. the PR used -e,
but checking around indicated prior art in Solaris usin -S.
* remove superfluous 'DUMP:' prefix in two messages
* initialise blocksperfile explicitly (not necessary, but everything
else in that section gets initialised, so be consistant :)
* display the ``pretty'' name of the dumped directory, so the user
knows if it's a subset or not

resolve conflicts from lite-2 import

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

cleanup xrefs

* Add the ability to dump specific files & directories of a single
filesystem. This uses fts(3) to access the directory structure (and
not the raw device), so the standard access permissions are adhered
to (unlike dumping an entire filesystem, which just requires read
access to the raw disk device).
* Support SIGINFO status reporting.
* Remove now unused variables that previously stored the (e)uid.
* Be more informative in a couple of error messages.

use _PATH_DEFTAPE from <paths.h>
change doco references/examples from /dev/rmt* to /dev/rst*

support $TAPE. update manual to reflect reality (#1937).

Add /dev/rst* to FILES section. (pr 601) (Matches mt man page.)

convert to new RCS Id conventions; reduce my headache

Fix an old formatting glitch.

Use getopt(3), with obsolete() from restore(8) for backward compatibility.

Update from 4.4-Lite, with local changes.

More spelling errors.

new version from CSRG, via BSDI, with fixes

Use ".At v6" macro instead of spelling out "Version 6 AT&T UNIX".

Dump is now in /sbin.

Add RCS indentifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Add external attribute dumping and restoring support from FreeBSD.
Does not fully work yet, attributes are being saved and restored correctly,
but don't appear in the restored files somehow.

PR bin/55834

count blocks written in unsigned 64 bit counter
rather than signed int which overflows after 2^31-1
blocks (2TiB) after which neither the 5 minute
status updates or SIGINFO (^T) reports are issued
until the negative numbers increase past 0 and
wildly inaccurate reports would be written.

Fix depenency on common symbols in sbin.

branches: 1.57.2;
Add -U flag to dump(8) and dump_lfs(8) to specify dumpdates entry

This address situations where dump(8) cannot figure out the device being
dumped. It also allows tracking of subvolume dumps by using virtual
device as dumpdates entry.

Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

- mark Exit() __dead
- bump a buffer size to avoid possible truncation
- adding missing fallthru comment

branches: 1.54.8; 1.54.16;
- update NAME_MAX to match the kernel.
- add a comment about where the constant is used.

Hack up dinode usage to be 64 vs. 32 as needed. Part 1.

(This part changes the native lfs code; the ufs-derived code already
has 64 vs. 32 logic, but as aspects of it are unsafe, and don't
entirely interoperate cleanly with the lfs 64/32 stuff, pass 2 will be
rehashing that.)

Add a new lfs header file: lfs_accessors.h.

This contains all the accessor functions and macros out of lfs.h.
Add an include of lfs_accessors.h after all uses of lfs.h... except
for code that wants to define its own struct lfs-alike that the
accessors are supposed to play along with. For these, set STRUCT_LFS
and include lfs_accessors.h after the necessary structure has been
defined, so that lfs_accessors.h can emit functions in terms of it.

handle new lfs split

Stuff UFS_ in front of a few of ufs's symbols to reduce namespace
pollution. Specifically:
ROOTINO -> UFS_ROOTINO
WINO -> UFS_WINO
NXADDR -> UFS_NXADDR
NDADDR -> UFS_NDADDR
NIADDR -> UFS_NIADDR
MAXSYMLINKLEN -> UFS_MAXSYMLINKLEN
MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency)

Sort out ext2fs's misuse of NDADDR and NIADDR; fortunately, these have
the same values in ext2fs and ffs.

No functional change intended.

Use __printflike.

Add some const.

branches: 1.47.2;
remove ancient crap

branches: 1.46.6;
PR/42883: Greywolf: Add -i flag which brings "true incremental" capability.

Fix some inconsisent/conflicting definitions and missing parameters

branches: 1.44.8; 1.44.10;
When using a snapshot take the snapshot raw device on further open.

Fixes PR #34923 dump(8) only dumps a corefile with -X (snapshots)

Approved by: Manuel Bouyer <bouyer@netbsd.org>

Change LFCNWRAP{STOP,GO} to make them more suitable for snapshotting; in
particular, the caller can now choose whether to wait for the condition
to be met, and if the caller of LFCNWRAPSTOP dies or otherwise closes
the descriptor, the filesystem is started again. Updated the ckckp
regression test to use the new semantics.

dump_lfs(8) now uses the fcntls to implement LFS-style snapshotting through
the -X flag, addressing PR#33457 albeit not using fss(4). Fixed a couple
other problems with dump_lfs that manifested themselves during testing.

Don't use cast expressions as lvalues as newer versions of gcc warn.

Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.

sprinkle const

branches: 1.39.2; 1.39.4; 1.39.6;
Replace the statfs() family of system calls with statvfs().
Retain binary compatibility.

Adjust tstart_volume (volume start time) by the time spent in query(),
similar to the adjustment for tstart_writing made in optr.c rev 1.4.
Should fix PR bin/19711 from matthew green.

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Use "wall -g operator" (instead of private code) to broadcast() messages to
members of the operator group.
Don't install "setgid tty", and remove now unnecessary gid/egid swapping.
Remove utmp trawling code pulled in from usr.bin/who.
The Code is now simpler, and more portable (without the utmp cruft) too.

This is derived from similar work in OpenBSD.

Bump daddr_t to 64 bits. Replace it with int32_t in all places where
it was used on-disk, so that on-disk formats remain the same.
Remove ufs_daddr_t and ufs_lbn_t for the time being.

Add -a to "auto-size" the tape, rather than relying upon other options
to define the tape size. Requires the tape driver to either return ENOSPC
at end of media, or 0 when a write is attempted (such as the "early warning"
support in st(4) enabled with "mt eew 1"). From FreeBSD.

Fixes from FreeBSD:
- Fix error if first tape was write protected. Fix NetBSD PRs 4754 and 6098.
- Make dump exit codes confirm to manual page.
- Use \a instead of \7 to make noise.
- Fix estimated number of tapes for huge dumps to cartridges.
- Use <sys/queue.h> SLIST_* instead of home-rolled lists.
- Do not exit if unable to read or create /etc/dumpdates.
- Support output (tape) device returning ENOSPC for end-of-media on a write.

Fixes by me:
- Remove unused ddates_in.
- Don't dump core if SIGINFO is received before 1 second has elapsed.
- Only process SIGINFO in current "active" child.
- Don't dump core in -w if dumpdates wasn't readable and ddatev == NULL
- Minor KNF; wrap some lines

- use correct type for minTime in findlru() (size_t instead of int)
- clean up whitespace

Add a -l (autoload) flag. For multivolume dumps, this makes dump eject the
tape when a volume is full, and try to reopen the tape drive for 2 mn.
To be used with tape changers which load the next tape when the current one is
ejected.
While I'm there fix eject handling for remote tape.

Add the capability for dump to print timestamps on all informational
messages.

s/filesystem/file system/

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

knf to ansi

Fix bug in new "e" option: only eject tape if it is full, not if the dump
was finished.

Add a new option "e" to "dump" which allows to eject tapes automatically
if a tape change is required.

restore needs the rcmd operations.

fix redundant decls

`dev_bsize' needs to be declared extern.

Format string cleanups by Bill Sommerfeld.

branches: 1.19.6;
Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

branches: 1.16.2; 1.16.4;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

#include machine/bswap.h and remove -lutil.

Add support for non-native byte order FFS. The dump is in filesystem byte
order, restore already knows how to byteswap dumps.

resolve conflicts from lite-2 import

remove __dead

* Add the ability to dump specific files & directories of a single
filesystem. This uses fts(3) to access the directory structure (and
not the raw device), so the standard access permissions are adhered
to (unlike dumping an entire filesystem, which just requires read
access to the raw disk device).
* Support SIGINFO status reporting.
* Remove now unused variables that previously stored the (e)uid.
* Be more informative in a couple of error messages.

Implement changes from [bin/1915]
* 'w' and 'W' display the year
* verbosity additions displaying total & per-volume transfer times and rates

convert to new RCS Id conventions; reduce my headache

Eliminate uses of some obsolete functions.

Update from 4.4-Lite, with local changes.

new version from CSRG, via BSDI, with fixes

Avoid core dumping if /etc/dumpdates is not present. Note that only file
systems listed in dumpdates are display by `dump w' and `dump W'; this is
probably a bug.
Original bug reported by John Brezak <brezak@ch.hp.com>.

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Clean up deleted files.

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Clean up deleted files.

Change the four inode loops to the form:
for (ino = 1; ino <= maxino; ino++)
Based on a bug fix from Havard Eidnes <Havard.Eidnes@runit.sintef.no> and
suggestion by Chris Torek <torek@horse.ee.lbl.gov>.

Add RCS identifiers.

Don't overrun end of argv.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Clean up deleted files.

Add RCS identifiers.

Don't dereference NULL pointer.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

handle new lfs split

Use <ctype.h> functions correctly.

Add some const.

branches: 1.34.2; 1.34.8;
Use __dead

fix sign-compare issue

branches: 1.32.22;
Convert a couple of "(void)&" constructs to volatile.
Flag various unused parameters in restore and dump so that these will
compile with -Wextra. (Note: restore uses some stuff from dump.)

sprinkle const

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Fixes from FreeBSD:
- Fix error if first tape was write protected. Fix NetBSD PRs 4754 and 6098.
- Make dump exit codes confirm to manual page.
- Use \a instead of \7 to make noise.
- Fix estimated number of tapes for huge dumps to cartridges.
- Use <sys/queue.h> SLIST_* instead of home-rolled lists.
- Do not exit if unable to read or create /etc/dumpdates.
- Support output (tape) device returning ENOSPC for end-of-media on a write.

Fixes by me:
- Remove unused ddates_in.
- Don't dump core if SIGINFO is received before 1 second has elapsed.
- Only process SIGINFO in current "active" child.
- Don't dump core in -w if dumpdates wasn't readable and ddatev == NULL
- Minor KNF; wrap some lines

unifdef sunos

Add a -l (autoload) flag. For multivolume dumps, this makes dump eject the
tape when a volume is full, and try to reopen the tape drive for 2 mn.
To be used with tape changers which load the next tape when the current one is
ejected.
While I'm there fix eject handling for remote tape.

fix -Wshadow warnings

unifdef !__STDC__ (missed in previous)

knf to ansi

use strdup appropriately

fix redundant decls

branches: 1.21.10;
Turn on TCP_NODELAY on the remote tape socket, from FreeBSD. Here is the
commit message from Bill Fenner:

Turn on TCP_NODELAY on the remote socket, to turn off sender silly window
syndrome avoidance. The combination of SWS avoidance and ack-every-other
causes low throughput if the block size divided by the MSS is odd (which
is true with the default block size and MSS).
Turning on TCP_NODELAY disables the Nagle algorithm and sender SWS avoidance.
The rdump request/response protocol can not invoke Nagle and cannot cause
SWS, so this has no negative effects.

Also, put back the code that sets the TOS to "throughput", which seems to
have been erroneously removed during the Lite-2 merge.

use <errno.h>

resolve conflicts from lite-2 import

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

fix PR#3710, reported by Tatoku Ogaito <tacha@trap.fukui-med.ac.jp>.
- don't pass pw->pw_name into functions; make a copy.

* Add the ability to dump specific files & directories of a single
filesystem. This uses fts(3) to access the directory structure (and
not the raw device), so the standard access permissions are adhered
to (unlike dumping an entire filesystem, which just requires read
access to the raw disk device).
* Support SIGINFO status reporting.
* Remove now unused variables that previously stored the (e)uid.
* Be more informative in a couple of error messages.

enable rcmd(). garbage collect.

use orcmd() for now.

be safe with buffers.

remove use of "register".

- If rcmd() fails, don't do any of the setsockopt() calls.
- Since the TCP_MAXSEG setsockopt can only descrease the MSS, and never
increase it, do an IP_TOS:IPTOS_THROUGHPUT setsockopt instead, since
dump is a bulk transfer.
Submitted by Bill Fenner <fenner@parc.xerox.com>.

Be more paranoid with our effective uid.

convert to new RCS Id conventions; reduce my headache

Eliminate uses of some obsolete functions.

err(3)/warn(3) cleanup

Update from 4.4-Lite, with local changes.

new version from CSRG, via BSDI, with fixes

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Clean up deleted files.

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Clean up deleted files.

Change the four inode loops to the form:
for (ino = 1; ino <= maxino; ino++)
Based on a bug fix from Havard Eidnes <Havard.Eidnes@runit.sintef.no> and
suggestion by Chris Torek <torek@horse.ee.lbl.gov>.

Add RCS identifiers.

Add support for fast symlinks.

from patch 110, by bostic:
dump not mapping last inode in fs

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Restore backward compatibility of UFS2 with previous NetBSD releases by
disabling support in UFS2 for extended attributes (including ACLs).
Add a new variant of UFS2 called "UFS2ea" that does support extended attributes.
Add new fsck_ffs operations "-c ea" and "-c no-ea" to convert file systems
from UFS2 to UFS2ea and vice-versa (both of which delete all existing extended
attributes in the process).

Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

branches: 1.22.20; 1.22.28;
fsbtodb() -> FFS_FSBTODB(), EXT2_FSBTODB(), or MFS_FSBTODB()
dbtofsb() -> FFS_DBTOFSB() or EXT2_DBTOFSB()

(Christos already did the lfs ones a few days back)

Rename ambiguous macros:
MAXDIRSIZE -> UFS_MAXDIRSIZE or LFS_MAXDIRSIZE
NINDIR -> FFS_NINDIR, EXT2_NINDIR, LFS_NINDIR, or MFS_NINDIR
INOPB -> FFS_INOPB, LFS_INOPB
INOPF -> FFS_INOPF, LFS_INOPF
blksize -> ffs_blksize, ext2_blksize, or lfs_blksize
sblksize -> ffs_blksize

These are not the only ambiguously defined filesystem macros, of
course, there's a pile more. I may not have found all the ambiguous
definitions of blksize(), too, as there are a lot of other things
called 'blksize' in the system.

Stuff UFS_ in front of a few of ufs's symbols to reduce namespace
pollution. Specifically:
ROOTINO -> UFS_ROOTINO
WINO -> UFS_WINO
NXADDR -> UFS_NXADDR
NDADDR -> UFS_NDADDR
NIADDR -> UFS_NIADDR
MAXSYMLINKLEN -> UFS_MAXSYMLINKLEN
MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency)

Sort out ext2fs's misuse of NDADDR and NIADDR; fortunately, these have
the same values in ext2fs and ffs.

No functional change intended.

Use static; remove unused global var this uncovered.

branches: 1.18.18; 1.18.24;
Remove the \n and tabs from the __COPYRIGHT() strings.
(Tweak some to use a consistent format.)

branches: 1.17.12; 1.17.16;
Convert a couple of "(void)&" constructs to volatile.
Flag various unused parameters in restore and dump so that these will
compile with -Wextra. (Note: restore uses some stuff from dump.)

Don't require ffsv2 fs have FS_FLAGS_UPDATED set

After determining whether the filesystem is byte swapped,
we better return that fact to the caller.

When searching for the superblock, don't pick an ffsv1 superblock from the
location where we expect to find an ffsv2 superblock.
It could be the first alternate for a ffsv1 filesystem with 64k blocks.
Fixes part of PR kern/24809

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Bump daddr_t to 64 bits. Replace it with int32_t in all places where
it was used on-disk, so that on-disk formats remain the same.
Remove ufs_daddr_t and ufs_lbn_t for the time being.

unifdef sunos

- use correct type for minTime in findlru() (size_t instead of int)
- clean up whitespace

fix -Wshadow warnings

remove third argument (`int ns') from ffs_sb_swap(), and let ffs_sb_swap()
determine the endianness of the `struct fs *o' superblock from o->fs_magic
and set needswap as necessary, rather than trusting the caller to get
it right. invariably, almost every caller of ffs_sb_swap() was calling it
with ns set to the wrong value for ns anyway!
ansi KNF ffs_bswap.c declarations whilst here.

this fixes all sorts of problems when trying to use other-endian file systems,
notably the kernel trying to access memory *way* off, possibly corrupting or
panicing, and userland programs SEGVing and/or corrupting things (e.g,
"fsck_ffs -B" to swap a file system endianness).

whilst the previous rev of ffs_bswap.c (1.10, 2000/12/23) made this problem
worse, i suspect that the problem was always there and previous versions
just happened not to trash things at the wrong time.

FFS_EI should now be a lot more stable.

s/filesystem/file system/dump.h

unifdef !__STDC__ (missed in previous)

knf to ansi

branches: 1.3.4; 1.3.8;
Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

Add -U flag to dump(8) and dump_lfs(8) to specify dumpdates entry

This address situations where dump(8) cannot figure out the device being
dumped. It also allows tracking of subvolume dumps by using virtual
device as dumpdates entry.

Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

branches: 1.20.20; 1.20.28;
handle new lfs split

Add some const.

branches: 1.18.6; 1.18.12;
PR/42952: The Grey Wolf: Fix for "true incremental". Don't overload level
9, use a new level "i" which is logically 10.

PR/42883: Greywolf: Add -i flag which brings "true incremental" capability.

Coverity CID 3275: Plug memory leak.

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

Patches provided by Joel Baker in PR 22308, verified by myself.

use strlcpy. use sizeof() instead of xxLEN to avoid de-synchronization

Fixes from FreeBSD:
- Fix error if first tape was write protected. Fix NetBSD PRs 4754 and 6098.
- Make dump exit codes confirm to manual page.
- Use \a instead of \7 to make noise.
- Fix estimated number of tapes for huge dumps to cartridges.
- Use <sys/queue.h> SLIST_* instead of home-rolled lists.
- Do not exit if unable to read or create /etc/dumpdates.
- Support output (tape) device returning ENOSPC for end-of-media on a write.

Fixes by me:
- Remove unused ddates_in.
- Don't dump core if SIGINFO is received before 1 second has elapsed.
- Only process SIGINFO in current "active" child.
- Don't dump core in -w if dumpdates wasn't readable and ddatev == NULL
- Minor KNF; wrap some lines

unifdef sunos

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

unifdef !__STDC__ (missed in previous)

knf to ansi

branches: 1.8.10;
PR bin/5346 from haszlaki@UAccess.NET:
After parsing the entries from /etc/dumpdates, remember ou entrie, in order
to print the date after /etc/dumpdates has been rewritten.

Need <time.h> for ctime() prototype.

Add support for non-native byte order FFS. The dump is in filesystem byte
order, restore already knows how to byteswap dumps.

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

remove use of "register".

convert to new RCS Id conventions; reduce my headache

Update from 4.4-Lite, with local changes.

branches: 1.1.1;
new version from CSRG, via BSDI, with fixes

dump: suppress lint warning about "effectively discards 'const'"

branches: 1.78.8;

PR bin/55834

count blocks written in unsigned 64 bit counter
rather than signed int which overflows after 2^31-1
blocks (2TiB) after which neither the 5 minute
status updates or SIGINFO (^T) reports are issued
until the negative numbers increase past 0 and
wildly inaccurate reports would be written.

Fix depenency on common symbols in sbin.

Add -D flag to allow the user to specify an alternate dumpdates file.
Closes PR #54469.

branches: 1.75.2;
Add -U flag to dump(8) and dump_lfs(8) to specify dumpdates entry

This address situations where dump(8) cannot figure out the device being
dumped. It also allows tracking of subvolume dumps by using virtual
device as dumpdates entry.

Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

branches: 1.73.8; 1.73.16;
Consistently use iswap64() and 64bit intermediate variable, as well as
64bit formats for c_tapea. Fixes tape useage report for large filesystems.

Default the read block size (-k default value) to kern.maxphys (usually
64k these days). This gives a noticable performance boost on large filesystems.

branches: 1.71.4;
Fix handling of NAME aliases in /etc/fstab.

handle new lfs split

branches: 1.69.2;
factor out rawname() from dump, fsck, savecore.

use getfsspecname()

PR/45735: Bug Hunting: The dump(8) manpage, as well as the program's `usage'-line,
needs improvement

branches: 1.66.6; 1.66.8;
PR/42952: The Grey Wolf: Fix for "true incremental". Don't overload level
9, use a new level "i" which is logically 10.

PR/42883: Greywolf: Add -i flag which brings "true incremental" capability.

Remove the \n and tabs from the __COPYRIGHT() strings.
(Tweak some to use a consistent format.)

branches: 1.63.18; 1.63.22;
When using a snapshot take the snapshot raw device on further open.

Fixes PR #34923 dump(8) only dumps a corefile with -X (snapshots)

Approved by: Manuel Bouyer <bouyer@netbsd.org>

Change LFCNWRAP{STOP,GO} to make them more suitable for snapshotting; in
particular, the caller can now choose whether to wait for the condition
to be met, and if the caller of LFCNWRAPSTOP dies or otherwise closes
the descriptor, the filesystem is started again. Updated the ckckp
regression test to use the new semantics.

dump_lfs(8) now uses the fcntls to implement LFS-style snapshotting through
the -X flag, addressing PR#33457 albeit not using fss(4). Fixed a couple
other problems with dump_lfs that manifested themselves during testing.

Emit "a snapshot of" message for snap_internal case too.

Coverity CID 2297: Fix memory leak.

sprinkle const

Snapshot support for dump(8):

- New option `-x backup' takes the dump from a snapshot backed up by `backup'.
The snapshot will be deleted on exit.

- New option `-X' as a synonym for `-x mountpoint' where `mountpoint' is the
file system to be dumped.

Reviewed and Approved by: Manuel Bouyer <bouyer@netbsd.org>

branches: 1.57.2;
Replace the statfs() family of system calls with statvfs().
Retain binary compatibility.

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Use "wall -g operator" (instead of private code) to broadcast() messages to
members of the operator group.
Don't install "setgid tty", and remove now unnecessary gid/egid swapping.
Remove utmp trawling code pulled in from usr.bin/who.
The Code is now simpler, and more portable (without the utmp cruft) too.

This is derived from similar work in OpenBSD.

should be safer to cast to u_long than to cast down to int

Add a cast to sizeof in printf() arg since _BSD_SIZE_T is unsigned long
on some ports.

use strlcpy. use sizeof() instead of xxLEN to avoid de-synchronization

For -l: specify the timeout on the command line (in seconds) instead of
hardwiring it to 2s, as suggested on tech-userlevel.

Add -a to "auto-size" the tape, rather than relying upon other options
to define the tape size. Requires the tape driver to either return ENOSPC
at end of media, or 0 when a write is attempted (such as the "early warning"
support in st(4) enabled with "mt eew 1"). From FreeBSD.

Fixes from FreeBSD:
- Fix error if first tape was write protected. Fix NetBSD PRs 4754 and 6098.
- Make dump exit codes confirm to manual page.
- Use \a instead of \7 to make noise.
- Fix estimated number of tapes for huge dumps to cartridges.
- Use <sys/queue.h> SLIST_* instead of home-rolled lists.
- Do not exit if unable to read or create /etc/dumpdates.
- Support output (tape) device returning ENOSPC for end-of-media on a write.

Fixes by me:
- Remove unused ddates_in.
- Don't dump core if SIGINFO is received before 1 second has elapsed.
- Only process SIGINFO in current "active" child.
- Don't dump core in -w if dumpdates wasn't readable and ddatev == NULL
- Minor KNF; wrap some lines

clean up from time_t -> int32_t change in dumprestore.h

unifdef sunos

nuke trailing whitespace

- use correct type for minTime in findlru() (size_t instead of int)
- clean up whitespace

Another little mistake, spotted by wiz.

Sync getopt() / man page with actual getopt options.

Add a -l (autoload) flag. For multivolume dumps, this makes dump eject the
tape when a volume is full, and try to reopen the tape drive for 2 mn.
To be used with tape changers which load the next tape when the current one is
ejected.
While I'm there fix eject handling for remote tape.

relax the restriction on -F that the file system image argument must be a
regular file

Add the capability for dump to print timestamps on all informational
messages.

- implement -F, which indicates that files-to-dump is a regular file
containing a file system image (instead of being a single file as
part of a subtree dump)
- use "file system" instead of "filesystem" or "file-system"

Patch to drop setgid tty privs until forking for operator notification

reset uid & gid around calling rmthost().

display what signal, not "unknown"

Assign a saved copy of dt->fs_file to mountpoint. dt->fs_file will be freed
otherwise resulting in a wrong spcl.c_filesys .

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

more KNF

knf to ansi

Add "e" option to usage message. Problem pointed out by Takahiro Kambe
in private e-mail.

Fix bug in new "e" option: only eject tape if it is full, not if the dump
was finished.

Add a new option "e" to "dump" which allows to eject tapes automatically
if a tape change is required.

Global variables don't need to be initialised to zero.

Fix format warnings

branches: 1.25.6;
Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

branches: 1.22.2; 1.22.4;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

add support for '-L label', which file(1) and restore(1) can extract from the
header. mostly based on [bin/6715] from Brian Grayson <bgrayson@ece.utexas.edu>

old filesystems don't have fs_qbmask and fs_qfmask set in the superblock;
fudge them as per code in /sys/ufs/ffs/ffs_vfsops.c::ffs_oldfscompat().
bug and suggested fix from Robert Elz <kre@munnari.OZ.AU> in [bin/6610]
XXX: there may be other variables required as well...

from Erik Bertelsen <erik@mediator.uni-c.dk>
{ put } { in } { lots } { of } { these } { to } { shut } { up } { egcs }

- use an array MAXHOSTNAMELEN+1 size to hold hostnames
- ensure hostname from gethostname() is nul-terminated in all cases
- minor KNF
- use MAXHOSTNAMELEN over various other values/defines
- be safe will buffers that hold hostnames

Add support for non-native byte order FFS. The dump is in filesystem byte
order, restore already knows how to byteswap dumps.

Various changes from Brian Grayson <bgrayson@ece.utexas.edu> in [bin/4103]
* exit after providing an estimate if -S was given. the PR used -e,
but checking around indicated prior art in Solaris usin -S.
* remove superfluous 'DUMP:' prefix in two messages
* initialise blocksperfile explicitly (not necessary, but everything
else in that section gets initialised, so be consistant :)
* display the ``pretty'' name of the dumped directory, so the user
knows if it's a subset or not

resolve conflicts from lite-2 import

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

* Add the ability to dump specific files & directories of a single
filesystem. This uses fts(3) to access the directory structure (and
not the raw device), so the standard access permissions are adhered
to (unlike dumping an entire filesystem, which just requires read
access to the raw disk device).
* Support SIGINFO status reporting.
* Remove now unused variables that previously stored the (e)uid.
* Be more informative in a couple of error messages.

enable rcmd(). garbage collect.

be safe with buffers.

remove use of "register".

Implement changes from [bin/1915]
* 'w' and 'W' display the year
* verbosity additions displaying total & per-volume transfer times and rates

set bflag so that 'dump bd M N' works properly. fix from Juergen
Hannken-Illjes in PR bin/3254.

Be more paranoid with our effective uid.

support $TAPE. update manual to reflect reality (#1937).

convert to new RCS Id conventions; reduce my headache

Use getopt(3), with obsolete() from restore(8) for backward compatibility.

Eliminate uses of some obsolete functions.

Update from 4.4-Lite, with local changes.

don't segv if called with no options!

branches: 1.1.1;
new version from CSRG, via BSDI, with fixes

Silently ignore fstab entries with NAME= entries that cannot be resolved.
Fixes PR 56249.

branches: 1.43.2;
Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

branches: 1.42.18; 1.42.26;
Fix handling of NAME aliases in /etc/fstab.

handle new lfs split

Use more static.

Add some const.

branches: 1.38.2;
factor out rawname() from dump, fsck, savecore.

use getfsspecname()

branches: 1.36.36;
Convert a couple of "(void)&" constructs to volatile.
Flag various unused parameters in restore and dump so that these will
compile with -Wextra. (Note: restore uses some stuff from dump.)

Change LFCNWRAP{STOP,GO} to make them more suitable for snapshotting; in
particular, the caller can now choose whether to wait for the condition
to be met, and if the caller of LFCNWRAPSTOP dies or otherwise closes
the descriptor, the filesystem is started again. Updated the ckckp
regression test to use the new semantics.

dump_lfs(8) now uses the fcntls to implement LFS-style snapshotting through
the -X flag, addressing PR#33457 albeit not using fss(4). Fixed a couple
other problems with dump_lfs that manifested themselves during testing.

sprinkle const

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

Adjust tstart_volume (volume start time) by the time spent in query(),
similar to the adjustment for tstart_writing made in optr.c rev 1.4.
Should fix PR bin/19711 from matthew green.

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Use "wall -g operator" (instead of private code) to broadcast() messages to
members of the operator group.
Don't install "setgid tty", and remove now unnecessary gid/egid swapping.
Remove utmp trawling code pulled in from usr.bin/who.
The Code is now simpler, and more portable (without the utmp cruft) too.

This is derived from similar work in OpenBSD.

save/restore errno in signal handlers.

defend against malicious line in ut_line, which could cause unwanted
writes to anything under /dev. revoke setuid/gid privs earlier.
From: xs@kittenz.org

support utmpx

Fixes from FreeBSD:
- Fix error if first tape was write protected. Fix NetBSD PRs 4754 and 6098.
- Make dump exit codes confirm to manual page.
- Use \a instead of \7 to make noise.
- Fix estimated number of tapes for huge dumps to cartridges.
- Use <sys/queue.h> SLIST_* instead of home-rolled lists.
- Do not exit if unable to read or create /etc/dumpdates.
- Support output (tape) device returning ENOSPC for end-of-media on a write.

Fixes by me:
- Remove unused ddates_in.
- Don't dump core if SIGINFO is received before 1 second has elapsed.
- Only process SIGINFO in current "active" child.
- Don't dump core in -w if dumpdates wasn't readable and ddatev == NULL
- Minor KNF; wrap some lines

For dump_lfs(8), in getfstab(), only "lfs" (instead of "ufs" or "ffs").
Fix from Minoura Makoto in [bin/10525].

nuke trailing whitespace

- use correct type for minTime in findlru() (size_t instead of int)
- clean up whitespace

fix -Wshadow warnings

in msg(), don't call va_list using functions twice in a row without calling
va_start() in between; reorder code so that we call vsnprintf and then
just fputs that buffer. crank the size of lastmsg whilst we're here
problem noted by Hideo Saito in [bin/14348].

Add the capability for dump to print timestamps on all informational
messages.

Patch to drop setgid tty privs until forking for operator notification

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

unifdef !__STDC__ (missed in previous)

knf to ansi

Format string cleanups by Bill Sommerfeld.

branches: 1.13.2; 1.13.10;
Need <time.h> for time(), localtime() and ctime() prototypes.

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

fix off by one bug.

Don't need to account for space for NUL, as it's included in the
sizeof(_PATH_DEV), noted by Chris Demetriou <cgd@netbsd.org>

Add missing `-', noted by Erik Fair <fair@cesium.clock.org> on
current-users.

be safe with buffers.

remove use of "register".

Implement changes from [bin/1915]
* 'w' and 'W' display the year
* verbosity additions displaying total & per-volume transfer times and rates

fix pr 2977 -- only try to dump ufs/ffs partitions

Adjust the base for time estimates to ignore time we spent waiting
for operator input.
enhancements from PR #900

convert to new RCS Id conventions; reduce my headache

Update from 4.4-Lite, with local changes.

branches: 1.1.1;
new version from CSRG, via BSDI, with fixes

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Use "wall -g operator" (instead of private code) to broadcast() messages to
members of the operator group.
Don't install "setgid tty", and remove now unnecessary gid/egid swapping.
Remove utmp trawling code pulled in from usr.bin/who.
The Code is now simpler, and more portable (without the utmp cruft) too.

This is derived from similar work in OpenBSD.

use _PATH_DEFTAPE from <paths.h>
change doco references/examples from /dev/rmt* to /dev/rst*

convert to new RCS Id conventions; reduce my headache

Use /etc/rmt explicitly.

Update from 4.4-Lite, with local changes.

new version from CSRG, via BSDI, with fixes

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Default the read block size (-k default value) to kern.maxphys (usually
64k these days). This gives a noticable performance boost on large filesystems.

branches: 1.24.6;
handle new lfs split

branches: 1.23.6; 1.23.8; 1.23.12;
range-check what we assign to int cachebufs from calculations with
uint64_t usermem. This only becomes relevant if you have several TB of RAM.
Promoting cachebufs to uint64_t is not necessary as it gets limited to
(currently) 512 anyway.

fixes the last issue of PR: 19852

Remove clause 3 and 4 from TNF licenses

branches: 1.21.32; 1.21.34;
Use HW_USERMEM64 to fetch the amount of memory available.

The sysctl returning the amount of memory in the system returns "int",
which is silly. Luckily, it really does return the correct result if
interpreted as an unsigned int. This change lets dump work on 32-bit
systems that have more than 2GB of RAM.

No need to print same error message twice in the same line.

Check lseek error correctly (i.e., use == -1 rather than < 0).

Fix typo I've introduced while cleaning up for commit.

- Fix daddr_t print format inside ifdef DIAGNOSTIC.
- Start scan cache entry just filled, rather than starting from top.

Convert `nblksread' to in terms of device block size. This restores
originally intended behaviour (see tech-userlevel archive around 1999 March).

Cosmetic changes.

Merge "struct cheader" and "struct cdesc" into an union.
No more alignment problems if the alignment of these two differs.

Approved by: Manuel Bouyer <bouyer@netbsd.org>

Bump daddr_t to 64 bits. Replace it with int32_t in all places where
it was used on-disk, so that on-disk formats remain the same.
Remove ufs_daddr_t and ufs_lbn_t for the time being.

branches: 1.11.2;
rcsid

nuke trailing whitespace

When using the read cache in bread(), ensure that the current (sub)block
isn't past the end of the file system.

Should fix the "spins forever but doing nothing" bug that dump
would occasionally have.

- use correct type for minTime in findlru() (size_t instead of int)
- clean up whitespace

In rawread(), add back a check to ensure that dump doesn't try to read
past EOM and fail with "short read error". This check was part of
traverse.c::bread(), and was removed when that function was migrated
to rcache.c::rawread() as part of the "read cache" functionality.

This should fix the problem with dump barfing on "short read" when "-r 0",
but I'm still debugging the problem where dump gets to a point and
"does nothing" when the read cache is being used.

knf to ansi

Fix format warnings

branches: 1.4.6;
Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

branches: 1.1.2; 1.1.4;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

Clean up deleted files.

Add RCS indentifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

- mark Exit() __dead
- bump a buffer size to avoid possible truncation
- adding missing fallthru comment

branches: 1.6.44;
fss(4): Allow FSSIOCSET to set the initial flags. Add a new flag
"FSS_UNLINK_ON_CREATE" to unlink the backing store before
the snapshot gets created.

With this change dump(8) no longer dumps the zero-sized, but named
snapshot it is working on. Same applies to fsck_ffs(8).

branches: 1.5.2;
Add -x option which allows to run `fsck_msdos -n' on a snapshot of
a live file system.

While here modify snap_open() to accept a character device as its
first arg and remove now unneeded get_snap_device().

Reviewed by: Manuel Bouyer <bouyer@netbsd.org>

branches: 1.4.4;
Remove clause 3 and 4 from TNF licenses

branches: 1.3.18; 1.3.20;
When using a snapshot take the snapshot raw device on further open.

Fixes PR #34923 dump(8) only dumps a corefile with -X (snapshots)

Approved by: Manuel Bouyer <bouyer@netbsd.org>

branches: 1.2.2;
Fix copyright date for previous.

Snapshot support for dump(8):

- New option `-x backup' takes the dump from a snapshot backed up by `backup'.
The snapshot will be deleted on exit.

- New option `-X' as a synonym for `-x mountpoint' where `mountpoint' is the
file system to be dumped.

Reviewed and Approved by: Manuel Bouyer <bouyer@netbsd.org>

Remove clause 3 and 4 from TNF licenses

branches: 1.3.18; 1.3.20;
When using a snapshot take the snapshot raw device on further open.

Fixes PR #34923 dump(8) only dumps a corefile with -X (snapshots)

Approved by: Manuel Bouyer <bouyer@netbsd.org>

branches: 1.2.2;
Fix copyright date for previous.

Snapshot support for dump(8):

- New option `-x backup' takes the dump from a snapshot backed up by `backup'.
The snapshot will be deleted on exit.

- New option `-X' as a synonym for `-x mountpoint' where `mountpoint' is the
file system to be dumped.

Reviewed and Approved by: Manuel Bouyer <bouyer@netbsd.org>

Add external attribute dumping and restoring support from FreeBSD.
Does not fully work yet, attributes are being saved and restored correctly,
but don't appear in the restored files somehow.

Bitmaps (TS_BITS and TS_CLRI) dont use the "c_addr" array as they
cannot have holes. As bitmaps are written without TS_ADDR records
"c_count" may be larger than the "c_addr" size resulting in a
segmentation violation reading "c_addr" beyond its end.

Compute "blks" for TS_INODE and TS_ADDR only -- its used for multi
volume dumps and the bitmaps must both be on the first volume.

branches: 1.55.2;
Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

branches: 1.54.8; 1.54.16;
Consistently use iswap64() and 64bit intermediate variable, as well as
64bit formats for c_tapea. Fixes tape useage report for large filesystems.

branches: 1.53.6;
handle new lfs split

Use void * for I/O functions and remove no-longer-needed casts.

Add some const.

branches: 1.50.2; 1.50.4; 1.50.8;
NULL does not need a cast, here

Fix some inconsisent/conflicting definitions and missing parameters

branches: 1.48.4; 1.48.6;
Convert a couple of "(void)&" constructs to volatile.
Flag various unused parameters in restore and dump so that these will
compile with -Wextra. (Note: restore uses some stuff from dump.)

When using a snapshot take the snapshot raw device on further open.

Fixes PR #34923 dump(8) only dumps a corefile with -X (snapshots)

Approved by: Manuel Bouyer <bouyer@netbsd.org>

Change LFCNWRAP{STOP,GO} to make them more suitable for snapshotting; in
particular, the caller can now choose whether to wait for the condition
to be met, and if the caller of LFCNWRAPSTOP dies or otherwise closes
the descriptor, the filesystem is started again. Updated the ckckp
regression test to use the new semantics.

dump_lfs(8) now uses the fcntls to implement LFS-style snapshotting through
the -X flag, addressing PR#33457 albeit not using fss(4). Fixed a couple
other problems with dump_lfs that manifested themselves during testing.

appease gcc -Wuninitialized

branches: 1.44.2; 1.44.4; 1.44.6;
Signal handlers should manipulate 'volatile sig_atomic_t' instead of 'int'.

Adjust tstart_volume (volume start time) by the time spent in query(),
similar to the adjustment for tstart_writing made in optr.c rev 1.4.
Should fix PR bin/19711 from matthew green.

Avoid signal race condition. If a slave gets the SIGUSR2 signal between
the "setjmp" and the "ready = 1" statements the slave will pause forever
and the complet dump hangs.

Fixes PR #24453

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Rename `sigset' locals to avoid symbol shadowing warning.

save/restore errno in signal handlers.

For -l: specify the timeout on the command line (in seconds) instead of
hardwiring it to 2s, as suggested on tech-userlevel.

Add -a to "auto-size" the tape, rather than relying upon other options
to define the tape size. Requires the tape driver to either return ENOSPC
at end of media, or 0 when a write is attempted (such as the "early warning"
support in st(4) enabled with "mt eew 1"). From FreeBSD.

Fixes from FreeBSD:
- Fix error if first tape was write protected. Fix NetBSD PRs 4754 and 6098.
- Make dump exit codes confirm to manual page.
- Use \a instead of \7 to make noise.
- Fix estimated number of tapes for huge dumps to cartridges.
- Use <sys/queue.h> SLIST_* instead of home-rolled lists.
- Do not exit if unable to read or create /etc/dumpdates.
- Support output (tape) device returning ENOSPC for end-of-media on a write.

Fixes by me:
- Remove unused ddates_in.
- Don't dump core if SIGINFO is received before 1 second has elapsed.
- Only process SIGINFO in current "active" child.
- Don't dump core in -w if dumpdates wasn't readable and ddatev == NULL
- Minor KNF; wrap some lines

unifdef sunos

nuke trailing whitespace

- use correct type for minTime in findlru() (size_t instead of int)
- clean up whitespace

Add a -l (autoload) flag. For multivolume dumps, this makes dump eject the
tape when a volume is full, and try to reopen the tape drive for 2 mn.
To be used with tape changers which load the next tape when the current one is
ejected.
While I'm there fix eject handling for remote tape.

fix -Wshadow warnings

s/filesystem/file system/dump.h

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

unifdef !__STDC__ (missed in previous)

knf to ansi

Fix bug in new "e" option: only eject tape if it is full, not if the dump
was finished.

Add a new option "e" to "dump" which allows to eject tapes automatically
if a tape change is required.

Better fix for format warnings.
The type of time_t varies between ports, so we need to cast
before printing. It appears to be sufficient and safe to use (int).

Fix format warnings

branches: 1.21.6;
Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

branches: 1.18.2; 1.18.4;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

use AF_LOCAL instead of AF_UNIX

Add support for non-native byte order FFS. The dump is in filesystem byte
order, restore already knows how to byteswap dumps.

make these compile on the alpha after WARNS=1.

resolve conflicts from lite-2 import

remove __dead

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

* Add the ability to dump specific files & directories of a single
filesystem. This uses fts(3) to access the directory structure (and
not the raw device), so the standard access permissions are adhered
to (unlike dumping an entire filesystem, which just requires read
access to the raw disk device).
* Support SIGINFO status reporting.
* Remove now unused variables that previously stored the (e)uid.
* Be more informative in a couple of error messages.

use _PATH_DEFTAPE from <paths.h>
change doco references/examples from /dev/rmt* to /dev/rst*

remove use of "register".

Implement changes from [bin/1915]
* 'w' and 'W' display the year
* verbosity additions displaying total & per-volume transfer times and rates

Use POSIX signals.

convert to new RCS Id conventions; reduce my headache

correct return values.

Eliminate uses of some obsolete functions.

Update from 4.4-Lite, with local changes.

light clean, and don't forget to close remote tape when aborting.
Inspired by Andreas Schulz.

branches: 1.1.1;
new version from CSRG, via BSDI, with fixes

fix always true conditional by removing the address-of operator.

found by GCC 12.

PR/56643: Paul Goyette: Disable the last block adjustment for now. It seems
to break restore.

PR/56270: Matthias Scheler: dump fails on ffsv1: Zero c_extsize since ffsv1
does not support extended attributes.

Add external attribute dumping and restoring support from FreeBSD.
Does not fully work yet, attributes are being saved and restored correctly,
but don't appear in the restored files somehow.

Use getfsspecname() to fill the filesystem argument in dumpdates.
While here, make sure that the error strings terminate with newline
consistently, and add a function that adds the system error string.

- mark Exit() __dead
- bump a buffer size to avoid possible truncation
- adding missing fallthru comment

branches: 1.50.20; 1.50.28;
handle new lfs split

Stuff UFS_ in front of a few of ufs's symbols to reduce namespace
pollution. Specifically:
ROOTINO -> UFS_ROOTINO
WINO -> UFS_WINO
NXADDR -> UFS_NXADDR
NDADDR -> UFS_NDADDR
NIADDR -> UFS_NIADDR
MAXSYMLINKLEN -> UFS_MAXSYMLINKLEN
MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency)

Sort out ext2fs's misuse of NDADDR and NIADDR; fortunately, these have
the same values in ext2fs and ffs.

No functional change intended.

branches: 1.48.18; 1.48.24;
We don't need to dump WAPBL log files - don't dump files with SF_LOG
flag set, and document this.

branches: 1.47.20; 1.47.24;
Change LFCNWRAP{STOP,GO} to make them more suitable for snapshotting; in
particular, the caller can now choose whether to wait for the condition
to be met, and if the caller of LFCNWRAPSTOP dies or otherwise closes
the descriptor, the filesystem is started again. Updated the ckckp
regression test to use the new semantics.

dump_lfs(8) now uses the fcntls to implement LFS-style snapshotting through
the -X flag, addressing PR#33457 albeit not using fss(4). Fixed a couple
other problems with dump_lfs that manifested themselves during testing.

Don't use cast expressions as lvalues as newer versions of gcc warn.

64 bit inode changes

Add ffs internal snapshots. Written by Marshall Kirk McKusick for FreeBSD.

- Not enabled by default. Needs kernel option FFS_SNAPSHOT.
- Change parameters of ffs_blkfree.
- Let the copy-on-write functions return an error so spec_strategy
may fail if the copy-on-write fails.
- Change genfs_*lock*() to use vp->v_vnlock instead of &vp->v_lock.
- Add flag B_METAONLY to VOP_BALLOC to return indirect block buffer.
- Add a function ffs_checkfreefile needed for snapshot creation.
- Add special handling of snapshot files:
Snapshots may not be opened for writing and the attributes are read-only.
Use the mtime as the time this snapshot was taken.
Deny mtime updates for snapshot files.
- Add function transferlockers to transfer any waiting processes from
one lock to another.
- Add vfsop VFS_SNAPSHOT to take a snapshot and make it accessible through
a vnode.
- Add snapshot support to ls, fsck_ffs and dump.

Welcome to 2.0F.

Approved by: Jason R. Thorpe <thorpej@netbsd.org>

Make the computation of estimated blocks work for frag size != 1024.

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

Patches provided by Joel Baker in PR 22308, verified by myself.

Fix byteswapping issues in dirindir(); mostly from Enami Tsugutomo.

Swap the right disk block in mapdirs() for the UFS1 case. From Enami.

Since the size passed in may be different from the fs blocksize, just
allocate a buffer each time instead of using a static one.

dblk should be static. Fixes PR 21020, by Geoff Wing. Patch supplied
by him.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

Bump daddr_t to 64 bits. Replace it with int32_t in all places where
it was used on-disk, so that on-disk formats remain the same.
Remove ufs_daddr_t and ufs_lbn_t for the time being.

Add fix from FreeBSD traverse.c 1.16:
Files in subdirectories of directories that have the nodump flag set
are sometimes incorrectly being dumped.

The problem arises because the subdirectory only gets its entry
cleared from usedinomap if it is also present in dumpinomap, and it is
the absence of a directory in usedinomap that internally indicates
that the directory is under the effects of UF_NODUMP (either directly
or inherited).

FreeBSD PR: 32414
Submitted by: David C Lawrence <tale@dd.org>

unifdef sunos

fix -Wshadow warnings

s/filesystem/file system/dump.h

* improve parsing of file-systems-to-dump when a path is given:
- if it's a path to an unmounted file-system listed in /etc/fstab, use
that instead of assuming the user wanted a subtree dump of the parent
directory. this restores the behaviour of dump before the subtree
dumping code went in.
- if it's a path to a mounted file-system which is not in /etc/fstab,
use the info from getmntinfo(3). previously, dump would choke.
* implement error checked malloc(), calloc(), strdup(), and use
appropriately (some of the calloc()s weren't being checked)
* use 'file-system' instead of 'filesystem' in the man page

unifdef !__STDC__ (missed in previous)

knf to ansi

branches: 1.28.6;
Working version of aborted dump{,_lfs} join.

Back out changes made on the 29th, not at all the changes I intended to commit.
(I don't know how I did it though ... I'll figure it out tonight.)
I'll reapply this, in a working form, this evening.

Adapt dump(8) to use filestore-independent (but still ufs-specific)
replacements for NINDIR, fsbtodb, etc. Create dump_lfs by adding a few
LFS-filestore-specific routines. As described in PR#8317.

branches: 1.25.2;
Avoid byte overflow in block estimate for large files

branches: 1.24.2;
Implement a read cache, as announced on tech-userlevel. Default is 32k
read buffer size, 512 buffer or 15% of the user memory. Can be changed
with the -k and -s options.

Handle "nodump" flag on directories, by not dumping any files or directories
under it. Based on some parts of PR 6705 by Brian Grayson.
In the Makefile, add (commented out) debug options that can be turned on.

fix estimate of blocks for subsets.
fix from Brian Grayson <bgrayson@ece.utexas.edu> in [bin/6607]

from Erik Bertelsen <erik@mediator.uni-c.dk>
{ put } { in } { lots } { of } { these } { to } { shut } { up } { egcs }

Add support for non-native byte order FFS. The dump is in filesystem byte
order, restore already knows how to byteswap dumps.

resolve conflicts from lite-2 import

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

* Add the ability to dump specific files & directories of a single
filesystem. This uses fts(3) to access the directory structure (and
not the raw device), so the standard access permissions are adhered
to (unlike dumping an entire filesystem, which just requires read
access to the raw disk device).
* Support SIGINFO status reporting.
* Remove now unused variables that previously stored the (e)uid.
* Be more informative in a couple of error messages.

remove use of "register".

patches from Tom I Helbekkmo <tih@nhh.no> to deal with type-size issues,
so this works (better, at least) on 64-bit machines (e.g. alpha).

branches: 1.14.6;
update for posixified stat structure

Back out previous change.

typeof(timeval.tv_sec) != time_t

Fix another bogus cast.

Remove incorrect cast on lseek() return value.

convert to new RCS Id conventions; reduce my headache

Don't mix stat flags and inode flags.

Eliminate uses of some obsolete functions.

Correct typo; IFMT -> S_IFMT.

Fix compatibility with old fastlinks.

Update from 4.4-Lite, with local changes.

change some #ifdef's.

fix for() starting condition in mapfiles()

branches: 1.1.1;
new version from CSRG, via BSDI, with fixes

handle new lfs split

Add some const.

branches: 1.15.56; 1.15.62;
Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22308, verified by myself.

Add support for UFS2. UFS2 is an enhanced FFS, adding support for
64 bit block pointers, extended attribute storage, and a few
other things.

This commit does not yet include the code to manipulate the extended
storage (for e.g. ACLs), this will be done later.

Originally written by Kirk McKusick and Network Associates Laboratories for
FreeBSD.

knf to ansi

Use strptime() instead of a home grown parser.

* cleanup for WARNS=1
* fix use of .Nm
* comment out some unused(?) functions

remove use of "register".

convert to new RCS Id conventions; reduce my headache

Sync with CSRG.

Update from 4.4-Lite, with local changes.

Replace emitl(), a function that converts a struct tm to a time_t, with
a call to C library function mktime(), which does the exact same thing
(using basically the same algorithm too).

new version from CSRG, via BSDI, with fixes

Add RCS identifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision