Copy stable/9 to releng/9.3 as part of the 9.3-RELEASE cycle.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

Copy head to stable/9 as part of 9.0-RELEASE release cycle.

Approved by: re (implicit)

Bump shared libraries version numbers in preparation for 9.0.
This time, only libraries which ABI has been changed compared to
stable/8, are bumped.

ABI analysis done by: Gleb Kurtsou
Approved by: re (kensmith)

Replace ERROR() macro with inline function. In-tree gcc cannot tolerate
the construct like printf("%\s", NULL) resulting from macroexpand of
ERROR(u, NULL), making it impossible to use LIBUFS_DEBUGGING.

With inline function, compiler cannot detect the NULL argument to
known function and does not try to convert it into puts().

In collaboration with: pho

Search beyond the first 1/8th of inodes.

Submitted by: jeff

- Merge soft-updates journaling from projects/suj/head into head. This
brings in support for an optional intent log which eliminates the need
for background fsck on unclean shutdown.

Sponsored by: iXsystems, Yahoo!, and Juniper.
With help from: McKusick and Peter Holm

i doesn't need to be signed here, make it unsigned.

Back to WARNS=3. The breakage wasn't what I thought it was :(

Keep Tinderbox happy.

Increased warnings weren't tested on ARM. Bump warnings back down to
0 until it can be properly tested by those raising the warnings.
Remember: make universe is required when changing the WARNS level.

Build lib/ with WARNS=6 by default.

Similar to libexec/, do the same with lib/. Make WARNS=6 the norm and
lower it when needed.

I'm setting WARNS?=0 for secure/. It seems secure/ includes the
Makefile.inc provided by lib/. I'm not going to touch that directory.
Most of the code there is contributed anyway.

Allow libufs(3) functions to operate on a regular file. This makes it possible to
use almost anything that uses libufs(3) against a file as an unprivileged user, e.g.
tunefs(8) and dumpfs(8) against a makefs(8)-created image.

Prodded by: kensmith

Bail out when memory allocation is failed, rather than referencing
a NULL pointer.

PR: kern/94480
Submitted by: Michiel Pelt <m.pelt xs4all nl>

Add a berase() function which uses ioctl(DIOCGDELETE) to erase a slab
of the disk.

The ufs_disk_fillout(3) can take special device name (with or without /dev/
prefix) as an argument and mount point path. At the end it has to find
device name file system is stored on, which means when mount point path is
given, it tries to look into /etc/fstab and find special device
corresponding to the given mount point. This is not perfect, because it
doesn't handle the case when file system is mounted by hand and mount point
is given as an argument.

I found this problem while trying to use snapinfo(8), which passes mount
points to the ufs_disk_fillout(3) function, but I had file system mounted
manually, so snapinfo(8) was exiting with the error below:

ufs_disk_fillout: No such file or directory

I modified libufs(3) to handle those arguments (the order is important):

1. special device with /dev/ prefix
2. special device without /dev/ prefix
3. mount point listed in /etc/fstab, directory exists
4. mount point listed in /etc/fstab, directory doesn't exist
5. mount point of a file system mounted by hand

Implement cgwrite1(3) function which stored a given cylinder group on disk.

Sponsored by: home.pl

minor style.Makefile(5) fixes:
- WARNS before CFLAGS
- CFLAGS -DXXX before -IXXX

Approved by: ru

Remove getino(3) manpage. It doesn't document what is here, what is here should
not get documented, and what it does document isn't going to come to CVS any
time in the immediate future.

Patience of a saint: trhodes

Disconnect getino.3 and remove MLINK I added.

Discussed with: jmallett

Hook getino.3 up to the build and link it to putino.3.

PR: 83820

Fixed xrefs.

Markup fixes.

Sort SEE ALSO references (in dictionary order, ignoring case).

Removed trailing whitespace.

Fixed spelling of the document date.

Remove unneccessary include of sys/types.h in the SYNOPSIS
header.

Sys/param.h includes sys/types.h internally unless LOCORE is
defined.

Approved by: des (mentor)

Add XXX'ed temporary bounce-buffering.

style.Makefile(5)

Stage 3 of dynamic root support. Make all the libraries needed to run
binaries in /bin and /sbin installed in /lib. Only the versioned files
reside in /lib, the .so symlink continues to live /usr/lib so the
toolchain doesn't need to be modified.

Correct an apparent typo.

PR: bin/53515
Submitted by: Lukas Ertl <l.ertl@univie.ac.at>
Approved by: jmallett

Various cleanups of careless mistakes/omissions.

PR: 53149
Submitted by: Lukas Ertl <l.ertl@univie.ac.at>

Left out the critical part of my "public domain" template, a notice saying
that this file is (these files are) in the public domain.

PR: 53149

Commit rudimentary libufs manual pages, except for that for
getino(3)/putino(3), inode.c has been reworked in Perforce to the point
where a manual page may not be accurate. Certainly putino(3) has not
even been merged back yet.

These will need a lot of improvement for most applications, but they
document the API enough to get someone on their feet, most likely. The
best documentation still exists in the form of libufs(3) consumers in the
base system.

Remove ufs_disk_ctor and ufs_disk_dtor, they never came to fruition. I do
not know of any software using them, and there is no "published API" for
libufs, as it were.

Reduce diffs with code in Perforce:
Parenthesise return values.

MFp4: Fix copy&paste English error.

Close the disk file descriptor that is RO before trying to open the
new one, and do not fall back to the RO fd. There was a bug here
in that the RO fd was never closed, if the RDRW open succeeded, but
this code is bogus anyway, and it breaks newfs of floppies, at least
for me, due to "Device busy." Anything that wants to fall back is
doing something significantly odd that it should have some more complex
code on its end.

Clean up error reporting in block.c, so that it gives honest error strings
for the sorts of errors we run into[1]. This also gives us room to put in a
vaguely appropriate casts to silence warnings since our compiler doesn't like
when we compare ssize_t to size_t[2]. Add a cast in sblock.c[3] to silence
a warning because of signed vs. size_t hell (again). Clean up nearby
excessive parenthemutilation[4].

Reviewed by: bde [2] [3]
Suggested by: bde, many [1]
Submitted by: bde [4]

An aside about [4], bde notes that we do not check for a negative value for
the fs bsize. I'm nto going to do that in every situation we use it, one must
expect a reasonable program to pass down reasonable values. Some foot shooting
protection I will tolerate, some I will not. Also he suggests some possible
conditional improvements there, which I may take to heart.

PS: For me at least, this is now WARNS=5 clean...

WARNS ?= 2, so idiocy like 1.12 of type.c doesn't have to happen again.

Missing "return 0"

Big pointy hat to: jmallett
Spotted by: peter

API for opening (and tracking) writable file descriptors per disk.

If we don't know where the sblock is (e.g. filling out a blank disk), then
get it from the fs structure. Really libufs should have interfaces to generate
both what we export, and what we import, based on eachother, and this should
be full of redundant code to make sure everything is right... But really, we
don't even deal with checksums, so plenty of room to improve.

bwrite, not sbwrite, needs to open for writing and write.

API to fillout a blank disk. For e.g. newfs.

Store not only the current cylinder group in the series (i.e. next that needs
to be read in) but also the last cylinder group in the series (i.e. what is
stored in the structure).

Don't crash when utilities are dumb and try to read less than the disk block
size (dumpfs may try to read the cylinder size (or is is sector size?) by way
of bread). Prevents a bounds error.

Add facility to read one, or a string of, cylinger groups.

Hunt for a disk to operate on, if we're passed a partition mountpoint, etc.

Concept reviewed by: phk

Fix typo.

Nuke dumb error reporting code, people can just use disk::d_error. Unify the
DEBUG and d_error initialisation into an ERROR macro, which can both trace and
set the d_error field. Much a more meaningful thing, I should say.

Create a new 32-bit fs_flags word in the superblock. Add code to move
the old 8-bit fs_old_flags to the new location the first time that the
filesystem is mounted by a new kernel. One of the unused flags in
fs_old_flags is used to indicate that the flags have been moved.
Leave the fs_old_flags word intact so that it will work properly if
used on an old kernel.

Change the fs_sblockloc superblock location field to be in units
of bytes instead of in units of filesystem fragments. The old units
did not work properly when the fragment size exceeeded the superblock
size (8192). Update old fs_sblockloc values at the same time that
the flags are moved.

Suggested by: BOUWSMA Barry <freebsd-misuser@netscum.dyndns.dk>
Sponsored by: DARPA & NAI Labs.

Use an error message closer to old dumpfs(8) in the case of truncated/no
superblock.

Submitted by: kkenn

Can't use it verbatim, at least I hate to, as the ", skipped" bit doesn't
make much sense in a library, to me.

Add the concept of a per-disk error string, and a function which prints it
along with the errno, if one is set.

Wrap the header to prevent multiple inclusion, and mark the DECLS section.

Reminded by: Rachel Hestilow <hestilow@ximian.com>

Initialise disk->d_ufs so that in sblock.c it's always initialised
(unless someone tries to use libufs support functions without using
_fillout or _ctor to construct a uufsd.)

Obtained from: jmallett_libufs Perforce branch.

Fill out (zero) and fill in (when doing getino()) the minimum and maximum
inodes in our inoblock (disk->d_ino{min,max}) appropriately.

No need to explicitly set NOMAN here.

Reviewed by: jmallett

DEBUG is a knob that means something else in FreeBSD, use LIBUFS_DEBUG to
turn on tracing.

In getino, have our DEBUG message in the unhandled case mention that it
does not know what sort of UFS filesystem this is.

Add some DEBUG(NULL)'s to function entry points.

Add libufs, a library for dealing with UFS filesystems from userland to
the build. It is here to compartmentalise functionality currently duplicated
in many notable programs in the base system. It currently handles block
reads and writes, as well as reading and writing of the filesystem superblock,
and the reading/lookup of inode data. It supports both UFS and UFS2. I
will be maintaining it, and porting programs to use it, however for now, it
is simply being built as part of world.