fdisk: remove unreachable options 'T' and 't', fix lint warnings

branches: 1.44.2;
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.

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: 1.42.32;
Don't use "#if defined(__i386__) || defined(__x86_64__)" conditionals
in C source file to define option features.
Define proper options in each Makefile per ${MACHINE_ARCH} variable instead.

Previously if a host is x86 and it has /usr/mdec/mbr file in its system,
tools fdisk implicitly installs it as mbr bootcode even for !x86 targets.

branches: 1.41.6; 1.41.8;
Remove fdisk/mbr for macppc which was blindly pulled from OpenBSD.
It just fakes MBR partition map which contains 1MB FAT16B partition
and ~1GB OpenBSD partition, and we can always create necessary
MBR partitions for OpenFirmware by the fdisk(8) command itself.

Delete GPT partitions if we overwrite with MBR to avoid confusion with disks
that have both MBR and a secondary left over GPT partition signature. From
Mike M. Volokhov

remove mispelt WARNS. (it will be set in ../Makefile.inc "soon")

branches: 1.38.2; 1.38.16;
Make this compile without -Wno-pointer-sign
(Needs updated bootblock.h)

branches: 1.37.4;
WARNS= 4

don't do the -Wno-pointer-sign dance for host programs.

sprinkle some -fno-strict-aliasing and -Wno-pointer-sign with GCC4.

If HOSTPROG is empty, do not pull in libutil, since we do not need it in that
case, and several host platforms do not have libutil. Fixes the build of
these programs as host tools on platforms without libutil.

Turn fdisk(8) into a host tool. Still todo: commit the host-tool
build infrastructure to src/tools/fdisk/.

Make fdisk(8) available on all platforms, since MS-DOS partitioning
is the de-facto standard for many removable media types.

WARNS=2 is the default defined in sbin/Makefile.inc. (thanks wiz)

Remove main() prototype; WARNS=2

Initial import of iyonix port.

The Iyonix is a desktop machine from Castle Technology, based on a 600MHz
XScale[tm] 80321 processor.

* Uses the bootloader from NetBSD/acorn32, which is now 32-bit compatible.
* Currently boots multiuser with a serial console.
* Device support is not yet complete.

With help from abs.

Don't descend into mbr_bootsel.
i386/and64 mbr code is built in arch i386/stand/mbr

x86_64 -> amd64

Compile mbr and mbr_bootsel on x86_64

Added fdisk on hpcarm

fdisk is now useful for macppc, build it

add hpcsh

Build fdisk(8) on netwinder.

fdisk(8) is needed on mvmeppc since PPCBUG uses PeeCee-style MBR
partition info when booting from disk.

playstation2 support

also install on prep.

Also install on cobalt.

Build fdisk also on arc

Oops, needs an operator.

Build fdisk also on hpcmips.

Don't decend into the subdirectory mbr_bootsel except i386,
since it currently doesn't contain MI thing and not compile
on ofppc box (thanks Wolfgang) and probably on bebox box.

branches: 1.13.2;
Build fdisk on bebox and ofppc(replaced powerpc).

Since fdisk/mbr contains a man page, we need to descend into it regardless
of which ${MACHINE} we are.

Add and build a bootselecting MBR (i386).

branches: 1.10.2;
Separate master bootcode from fdisk.

switch to using opendisk(3) from libutil

define WARNS?=1 for all of sbin/*

* support ``short'' names of disks (e.g, ``sd0'') instead of requiring
the full name (e.g, ``/dev/rsd0d'') a la disklabel(8) (where the code
was effectively lifted from ;-)
* prefix hexadecimal numbers in output with '0x'

-- add -b flag for setting fdisk's idea of BIOS geometry from command line
-- cleanup so WARNS=1 works
-- man page updated to reality.

move man pages into share/man.

convert to new RCS Id conventions; reduce my headache

specify man pages the new way.

Add RCS identifiers.

added a 386bsd fdisk program, by julian elischer. (patch 111)

fdisk: remove unreachable options 'T' and 't', fix lint warnings

branches: 1.92.8;
Use Nx, fix formatting nit.

Add -g flag to preserve GPT headers when updating MBR.

Fix xrefs.

PR/54756: germain: Avoid confusion caused by overuse of the word "file" both
as a noun and an argument name.

mention that -B is used to set the timeout for mbr_bootsel

branches: 1.87.12;
Fix spelling of simultaneous.

branches: 1.86.2;
- Document the HISTORY of fdisk based on the original post to comp.unix.bsd by
Julian Elischer [1] and the Mach 2.5 Installation notes [2].
I was unable to pin point the exact version of Mach the fdisk utility appeared
as I didn't find documentation older than version 2.5 & no source code or repo
history. fdisk utility appears as a separate utility[3] in v2.5. Due to this,
I've avoided stating the exact version fdisk first appeared in Mach.
- Make correction pointed by textproc/igor
- Bump date

[1] https://groups.google.com/forum/#!searchin/comp.unix.bsd/14-Dec-89$20Robert$20Baron$20(rvb)$20at$20Carnegie-Mellon$20University%7Csort:relevance/comp.unix.bsd/Hhi45vAHxDg

[2] ftp://ftp.mcs.vuw.ac.nz/doc/misc/mach-i386-doc/i386_install.ps
[3] ftp://ftp.mcs.vuw.ac.nz/doc/misc/mach-i386-doc/i386_manpages.ps

branches: 1.85.8;
-s:
If the optional arguments are not provided, they stay as before or
use the same defaults as the interactive mode, if new.

(feel free to improve)

Fix a bug in previous and update detailed -s description.

Update SYNOPSIS.
Christos, please check.

Bring back the playstation2 port - now that a toolchain for it is
available again.

Suggest using -b 1023/255/63 for modern disks where the BIOS is
configured to use LBA-Assisted translation.

Revert my previous commit, I misunderstood the intention.
Noted by njoly.

Previous slightly more consistently.

Fix indented blocks rendering with mandoc. Switch from .in unsupported
macro to .Bd/.Ed ones. While here kill some unneeded .Pp/.br macros.

branches: 1.77.2;
Remove unnecessary Bk/Ek pairs from SYNOPSIS.
No effective change except where I used the opportunity to sort options
and/or option descriptions.

support sector sizes > 512.

Remove trailing whitespace.

According to most other OSes and documentation sources, an MBR has
one "extended" partition and the partitions inside that are "logical"
partitions. Make fdisk(8) man page follow suit.

Patch from Julian Fagir in PR#45695.

Remove trailing whitespace.

Add a flag to support writing overlapping partitions and explain why.

branches: 1.71.4;
Use Nx.

Minor punctuation fixes, improve and sort SYNOPSIS, sort flag descriptions.
From Bug Hunting in PR 43130.

Add info for -A ptn_alignment[/ptn_0_offset]

Use implicit -F for regular files, much like disklabel(8).

Remove support for NetBSD/playstation2.

Add a bunch of useful typical examples. Also use appropriate markup there.

Update -u usage to follow changes in fdisk.c rev 1.125:
>> Allow MB, GB and CYL (not just M, G and C) and lower case.
>> Don't output a splurious 'd' before "cyl".
>> Fixes PR/37414.

XXX "NNcy" is also allowed?

Clarify description of -f and -u options.
(-u means updating partition data, not interactive mode)

Do not use .Xo / .Xc without a good reason.

an example.

Small grammar tweak.

Delete GPT partitions if we overwrite with MBR to avoid confusion with disks
that have both MBR and a secondary left over GPT partition signature. From
Mike M. Volokhov

Note the existance of mbr_com0 and mbr_com0_9600.

Remove note about the first track being reserved for the BIOS (it's
not), and replace it with a note about the first partition traditionally
beginning in the second track of the first cylinder.

Report the first active partition.

branches: 1.56.4; 1.56.6;
Plural fix.

Bump date for previous.

Document the new checks for whether a PBR is bootable.
Also add more detail in the introduction, and a few
more references.

branches: 1.53.2;
Use more macros. New sentence, new line.

PR# 28660: how to remove bootmenu text from some partition

Allow partition sizes to be input in GB

s/Nm/Nx/, as it's NetBSD that has ports, not fdisk.

Bump date for previous; new sentence, new line;
sort list of ports where fdisk is available.

Add note top man page explaining that fdisk is not used on all platforms
and list the ones that is is.

Partial fix to PR#25943

Use .Pp instead of .br as appropriate.
(This needs more work)

Nm needs no argument; also, correct wrong Nm usage.

Bump date for last. Use Sq for single quotes. Drop trailing space.
New sentence, new line. Sort xrefs. Fix a spelling error.

Add flag -F, options -t disktab and -T disktype to fdisk. Flag -F
indicates that the target of the fdisk operation is a file, not a
disk. With -t and -T, the user tells fdisk the geometry for the
file If the default geometry that fdisk will "fake up" for a file
are not satisfactory, the user may tell the geometry to fdisk using
-t disktab -T disktype.

be consistent about "MS-DOS"

add missing . before Fl

MBR, not mbr.

Change a couple of .Em to .Ar

Replace some raw roff with mdoc.

Improve formatting for [/bootmenu], requested by dsl; add some commas.

a mbr -> an mbr

Drop trailing space.

Rototilled fdisk that supports:
- editing of the extended partition list
- user friendly default start/size for new partitions
- partition start/size input in sectors, cylinders or MB
- bootsel menu names configured with paritition bounds
- partition update loop asks used which partition to change
(instead of asking about each partition in turn).
- detection (and erroring) of overlapped partitions.
- automatic installation of correct mbr_xxx code (after prompting user)

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

Drop trailing space (in last).

Slightly clarify/improve the documentation for the -B (boot selector)
option

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

PR/16073: Jachym _Freza_ Holecek: Add the ability to list known partition types.
While I am here, KNF a bit.

Drop duplicate .Pp.

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

Cross-link fdisk(8), mbr(8), and mbrlabel(8) in the SEE ALSO sections,
as suggested in PR 14288. Also add a little white space to mbr(8) to
make it a little more readable.

Whitespace nits

Drop arguments of .Os.

list the options in a specific options list (as per the `traditional'
man page). i find this much easier to read than option descriptions
scattered amongst paragraphs of text.

xref (disklabel.8, fdisk.8, mbrlabel.8) as appropriate

Note that sysid 0 marks a partition as unused.

branches: 1.21.4;
Add note that "fdisk -i" is the same as "fdisk /mbr" on MSDOS.

branches: 1.20.2;
Fix tyop and clean up usage of commas a little.

Remove confusing entry about cylinder boundaries.

Make the bootselector configurable by fdisk. Add a -B flag to this
effect (i386 only of course). Also change one branch in the
bootselector code to an explicit 16 bit one, and check both
the boot menu and partition tables to see if the partition
requested by the user should be booted. This check just in case,
should the bootselector menu and partition table get out of sync
somehow. (mmm, bytesqueezing)

branches: 1.17.2;
Clean up SYNOPSIS formatting.

Separate master bootcode from fdisk.

Automagically initialize boot code and partition table if not yet valid.
Don't touch the partition table on -i, if previously valid.

update for new partition ID

macroize BSD, NetBSD, FreeBSD and misc cleanup

Typos, English.

branches: 1.11.2;
correct use of various mdoc macros to cleanup resultant man page

-- add -b flag for setting fdisk's idea of BIOS geometry from command line
-- cleanup so WARNS=1 works
-- man page updated to reality.

Upgrade for non-interactive use. Interactive use is nearly unchanged.
-i no longer implies -u, -i now does not use the entire first track
because some BIOSes refuse to boot from a partition in the first track.
-s renamed to -S, new -s for non-interactive setting of a partition's
id, offset and size. Updated documentation to match.

convert to new RCS Id conventions; reduce my headache

Clean up a bit more.

Clean this up a bit.

More spelling errors.

Fix spelling errors

386bsd -> NetBSD

Add RCS identifiers, remove some completely useless RCS logs and patchkit
headers, and a few other insignificant changes.

added a 386bsd fdisk program, by julian elischer. (patch 111)

s/sightly/slightly/ and s/sligtly/slightly/ in documentation and comment.

fdisk: remove unreachable options 'T' and 't', fix lint warnings

branches: 1.161.4;
Avoid crashes with invalid or tiny sector sizes.

fdisk(8): Convert realloc(x * y) to reallocarr. Eliminate temp variables.

Add -g to usage.

Add -g flag to preserve GPT headers when updating MBR.

print the name of the disk causing the error.

In validate_bootsel, the code assumed that arr[j][i] can be accessed
with something like arr[0][j*ARRAYWIDTH+i]. gcc no longer allows such
hacks and discards the code silently because of undefined behaviour.

Add missing -z flag to usage()

branches: 1.154.2; 1.154.4;
Really don't include ioctl code for tools build.

branches: 1.153.8;
explain how to clear an entry (without reading the source)

a few more \n's in errors

fix error messages containing \n

remove debugging.

instead of using ugly -1's make args optional

default to something reasonable (like the interactive mode does) instead
of 0 when -1 is specified for the start or size.

Allow -1 values in the -s <sysid>/<start>/<size> indicate use the previous
values. For example:
fdisk -f -i /dev/rsd0d # initialize mbr and create an msdos partition.
fdisk -f -u -0 -a -s 169/-1/-1 /dev/rsd0d # converts the msdos partition
to a netbsd one, and makes it active.

Always print the MBR drive serial number, even if it's 0.

Improve support for logical sector sizes greater than 512.

Don't produce spurious errors when creating labels on files.

CID/979997: missing va_end()

branches: 1.142.2;
Use ${DESTDIR}/usr/mdec for DEFAULT_BOOTDIR on tools fdisk(8)
rather than unknown build host's /usr/mdec directory.

XXX: no option to leave MBR bootcode empty on x86 targets

Don't use "#if defined(__i386__) || defined(__x86_64__)" conditionals
in C source file to define option features.
Define proper options in each Makefile per ${MACHINE_ARCH} variable instead.

Previously if a host is x86 and it has /usr/mdec/mbr file in its system,
tools fdisk implicitly installs it as mbr bootcode even for !x86 targets.

support sector sizes > 512.

Add __printflike attribution to use vprintf and friends with an argument
as format string.

branches: 1.138.2;
Also ignore out of space conditions for -I so we can create overlapping
partitions on a full disk. Document and add it to the options.

Add missing -F flag and -I flag in usage. Sort the order to match
manpage.

put the error check in the conditional where it belongs.

Add a flag to support writing overlapping partitions and explain why.

branches: 1.134.2;
fix the sparc build

one static is enough, otherwise you risk zapping yourself.

staticfy. __dead.

Fix parsing of -A option so that the [/ptn_0_offset] part is optional, as
described in the man page.

Put #endif in correct place - fixes PR/42500
Update usage() for -A, and rename one of the variables for clarity.

Add support for partition being aligned on non-cylinder boundaries.
Info can be specified with -A parameter.
Default is based on how the first partition is defined.
For empty disks larger than 128GB (arbitrary figure) use 1MB alignment.

Editing partitions larger than 1TB works better if 'decimal' returns
64 bit integer values!

Add a large comment about disk geometries.
Rename (with #defines) the variables use for aligning partitions to
separate them from the bios geometry.
All in advance of allowing other partition alignments (eg 2048 sectors).

Use implicit -F for regular files, much like disklabel(8).

Allow MB, GB and CYL (not just M, G and C) and lower case.
Don't output a splurious 'd' before "cyl".
Fixes PR/37414.
Patch in PR is wrong!

Make absence of /usr/mdec/mbr only a warning.
Fixes PR/26510

Change arguments to decimal() to be int64_t.
They need to be able to hold disk sector numbers > 2^31 and also -1.
Should fix PR/34807

make this compile one non mbr machines (move #define to the general section)

Delete GPT partitions if we overwrite with MBR to avoid confusion with disks
that have both MBR and a secondary left over GPT partition signature. From
Mike M. Volokhov

also fix a comment wrt use of the csysid/cstart/csize vars

When using -s flag, read start and size as unsigned to allow for larger
disks. (Thanks to drochner@ for catching it).

fix sign-compare issues

branches: 1.117.2;
Remove local definitions of PRIxxxx macros in favour of the
centralised definitions from tools/compat/compat_defs.h, which
are included via nbtool_config.h.

branches: 1.116.6;
mbr_dsn is unsigned, print it as such.

branches: 1.115.2; 1.115.4;
Add support for mbr code with ascii keycodes

- add a sanity check to avoid possible division by zero
- adjust sector numbers _after_ "Sanity check the data against all zeroes"

Should close PR bin/37656.

Track zero is not strictly reserved for the BIOS, that's just a
convention, so don't make it an error in check_overlap(). Sector 0 is
reserved for the MBR, so make that an error in check_overlap().

Report the first active partition.

If the partition is unused, then attempting to decode the PBR would
actually decode the MBR from sector 0 of the disk. Don't do it if
v_flag < 3.

branches: 1.110.2;
Make this compile without -Wno-pointer-sign
(Needs updated bootblock.h)

Change -F to not use opendisk().
Don't change default to the boot volume until after we've found we aren't
doing -F.
For tool-fdisk enforce -F and use #defines to make code that can no
longer possibly execute compile.

Remove unwanted ">" from a printed message.

Don't print Drive serial number if -S (setting shell variables) is specified.

This program is a host tool, so when we use newfangled C features like
PRIxyz we have to ensure it still builds on hosts with Old compilers.
Noticed while building a cross-toolchain on FreeBSD 4.

branches: 1.105.4;
Preserve the mbr 'driver serial number' when copying in the netbsd
mbr code. Should leave windows vista bootable.
Print the dsn from the mbr, not each pbr.

PR/36223: Milos Negovanovic: Don't ask if we want to install mbr_bootselect
with -f.

Try to check whether each partition is bootable. If not,
always print an error message. If bootable and -v flag was
specified, print a success message. If -v flag was specified
more than once, also try to decode more information from
the PBR.

Input from dsl and bad.

branches: 1.102.2;
Report the last sector number of a partition, instead of that of the first
of the next - with appropriate discarding of the head/sector values
if the partition is cylinder aligned.
Fixes PR 26921

PR/33633: Yoshito Komatsu: /sbin/fdisk does not work
Downgrade error getting heads to a warning.

PR/33522: Peter Bex: Sanity checks to avoid zerodivide.

Appease Coverty CID 7449. Avoid close(-1) in error path.

Check for allocation failures in malloc, calloc, realloc, asprintf, and
vasprintf and try to handle them.

So that we can build a full-featured fdisk cross-tool with -DBOOTSEL,
condition the definition of DEFAULT_BOOTxxx variables and SCAN_xxx
on defined(BOOTSEL), instead of defined(__i386__) || defined(__x86_64__).

While I'm here, only build get_diskname in the same conditions as
get_geometry is built, since only the latter calls the former.

Byteswap the constant MBR_{BS_}MAGIC throughout, should genetate better code.
Except that non all architectures have that optimisation :-(

PRI* macros should not contain percent sign.
While here, conditionalise use of "lld" for PRId64 on HAVE_LONG_LONG.

Fix cross-compilation on FreeBSD 4.x (and other OSes) where PRId64 isn't
defined.

When building as host tool don't include "../../include/util.h".

opendisk() declared in that file is not used in this case anyway, and
<util.h> pulls in unwatned includes (e.g. <utmpx.h> doesn't exist on
FreeBSD 4.*).

Move WARNS=3 to the Makefile.inc, and add a little const to the remaining
programs that did not compile before.

Turn fdisk(8) into a host tool. Still todo: commit the host-tool
build infrastructure to src/tools/fdisk/.

For portability, use standard types: s/uint/unsigned int/,
s/ulong/unsigned long/, s/u_int8_t/uint8_t/.

Allow partition sizes to be input in GB

branches: 1.88.2;
Remove main() prototype; WARNS=2

Fix typo in error message, reported by Nicolas Joly in PR 28347.

Use the 1st element's size as the argument to bsearch.

Use ptype table from <sys/bootblock.h>

Add 0xBF for Solaris. From Larry Lee <lclee at west.sun.com>

Add (unsigned char) to isdigit() calls

Fix intuit_translated_geometry() calculation.
From Izumi Tsutsui in PR bin/26917

check sysctl error return on both calls.
this prevents sysctl from coredumping if the second call fails while the
first succeeds. This isn't supposed to happen, but there is another bug
in the i386 kernel implementation of sysctl machdep.diskinfo that excites this

Insert missing space so fdisk prints "primary partition table
invalid, no magic in sector 0", not "primarypartition ...".

Use the disk size from the device driver, not the bios when checking
whether partitions are valid.
Stops problems with broken BIOS, fixes PR bin/25491

Don't output the 'Installed bootfile doesn't support required options'
message when we are updating the mbr code because the user did fdisk -i ...

branches: 1.77.2;
Only examine mbr_bootsel* #ifdef BOOTSEL.
Noted on current-users@ by Markus Hennecke.

Move mbr_bootsel from offset 404 to offset 400 in struct mbr_sector to
leave 4 bytes for the Windows NT Drive Serial Number (DSN) at 440-443
(as mbr_sector.mbr_dsn).

Ensure that all the MBR & PBR code reserves space for mbr_sector.mbr_dsn.

Leave the bootsel magic number at 444-445 as mbr_sector.mbr_bootsel_magic
(instead of mbr_sector.mbr_bootsel.mbrbs_magic), but use 0xb5e1 (MBR_BS_MAGIC)
instead of 0xaa55 (MBR_MAGIC) to indicate that this change has occurred.

Rework MBR_BS_NEWMBR to mean "mbr_bootsel has moved to 400".

Modify fdisk(8) to automatically relocate the mbr_bootsel from 404 to 400
if mbr_bootsel_magic is the old value (0xaa55), and unset MBR_BS_NEWMBR
to flag that new mbr_bootsel code must be used if updating the MBR.


These changes fixes a problem where Windows 2000 or Windows XP would corrupt
the last 3 bytes + NUL of MBR partition 3's bootsel name if the bootsel name
was 5 characters long, replacing bytes 6-9 with the DSN.
Also, by explicitly reserving the space for the DSN we prevent problems in the
future if non bootsel MBR or PBR code had other information at bytes 440-443.

Add flag -F, options -t disktab and -T disktype to fdisk. Flag -F
indicates that the target of the fdisk operation is a file, not a
disk. With -t and -T, the user tells fdisk the geometry for the
file If the default geometry that fdisk will "fake up" for a file
are not satisfactory, the user may tell the geometry to fdisk using
-t disktab -T disktype.

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

Add "DISK" to -S output.

unconditionally #define DEFAULT_BOOTDIR. should fix non-x86 builds

Overhaul MBR handling (part 1):

<sys/bootblock.h>:
* Added definitions for the Master Boot Record (MBR) used by
a variety of systems (primarily i386), including the format
of the BIOS Parameter Block (BPB).
This information was cribbed from a variety of sources
including <sys/disklabel_mbr.h> which this is a superset of.

As part of this, some data structure elements and #defines
were renamed to be more "namespace friendly" and consistent
with other bootblocks and MBR documentation.
Update all uses of the old names to the new names.

<sys/disklabel_mbr.h>:
* Deprecated in favor of <sys/bootblock.h> (the latter is more
"host tool" friendly).

amd64 & i386:
* Renamed /usr/mdec/bootxx_dosfs to /usr/mdec/bootxx_msdos, to
be consistent with the naming convention of the msdosfs tools.

* Removed /usr/mdec/bootxx_ufs, as it's equivalent to bootxx_ffsv1
and it's confusing to have two functionally equivalent bootblocks,
especially given that "ufs" has multiple meanings (it could be
a synonym for "ffs", or the group of ffs/lfs/ext2fs file systems).

* Rework pbr.S (the first sector of bootxx_*):
+ Ensure that BPB (bytes 11..89) and the partition table
(bytes 446..509) do not contain code.
+ Add support for booting from FAT partitions if BOOT_FROM_FAT
is defined. (Only set for bootxx_msdos).
+ Remove "dummy" partition 3; if people want to installboot(8)
these to the start of the disk they can use fdisk(8) to
create a real MBR partition table...
+ Compile with TERSE_ERROR so it fits because of the above.
Whilst this is less user friendly, I feel it's important
to have a valid partition table and BPB in the MBR/PBR.

* Renamed /usr/mdec/biosboot to /usr/mdec/boot, to be consistent
with other platforms.

* Enable SUPPORT_DOSFS in /usr/mdec/boot (stage2), so that
we can boot off FAT partitions.

* Crank version of /usr/mdec/boot to 3.1, and fix some of the other
entries in the version file.

installboot(8) (i386):
* Read the existing MBR of the filesystem and retain the BIOS
Parameter Block (BPB) in bytes 11..89 and the MBR partition
table in bytes 446..509. (Previously installboot(8) would
trash those two sections of the MBR.)

mbrlabel(8):
* Use sys/lib/libkern/xlat_mbr_fstype.c instead of homegrown code
to map the MBR partition type to the NetBSD disklabel type.


Test built "make release" for i386, and new bootblocks verified to work
(even off FAT!).

realloc pedant

Define and use PRIdaddr for printing disk block numbers.
Should fix bin/22638 for compilation on amd64.

Put cylinder number 1023 into the mbr for sectors beyond the CHS limit
(rather than encoding the highest CHS addressable sector).

Fix code that detects bios geometry from mbr table.

Only look at each pair of CHS values once when trying to guess geometry.

asprintf is easier

delete a diagnostic printf that was accidentally left in

Remove some unused variabled from get_params_to_use() in the !i386 && !amd64
case.

Rototilled fdisk that supports:
- editing of the extended partition list
- user friendly default start/size for new partitions
- partition start/size input in sectors, cylinders or MB
- bootsel menu names configured with paritition bounds
- partition update loop asks used which partition to change
(instead of asking about each partition in turn).
- detection (and erroring) of overlapped partitions.
- automatic installation of correct mbr_xxx code (after prompting user)

Move some definitions from fdisk.c to disklabel_mbr.h so they don't have
to be duplicated elsewhere (eg in mbr.S).
Modify fdisk.c so that it builds with the new disklabel.h

Fix range check for the default boot option case.

Print missing newline in two bootselector-related messages. Approved by wiz.

Enable bootselector code for x86_64.
XXX the ifdef-ed parts should be split off into a seperate file.

Avoid format warning for off_t printf.

Make the -l flage more useful, by actually printing the systype number
alongside the sysid string (instead of just the array index of the
struct part_type they are found in).

Now fdisk -l shows the familiar 169 for NetBSD, 165 for FreeBSD or
386BSD or old NetBSD, and other possibly-familiar (131 for Linux native,
015 for Ext. Partition - LBA) values in with their correct numbers.

add partition type 0xaf for Apple HFS.

Clean up printf->err conversions.

exit on EOF from stdin.

branches: 1.52.2;
Build fdisk(8) on netwinder.

nicer printing.

PR/16073: Jachym _Freza_ Holecek: Add the ability to list known partition types.
While I am here, KNF a bit.

add "Apple UFS" and "Apple Boot" partition ids used by Darwin

if sysctl kern.root_device is availalbe, use that the default device
(rather than /dev/rwd0d). display the disk device used when printing
the partition table.

opendisk(3) checks the device type for us

add fix by Minoura Makoto to correctly support recursive extended partitions.
the offset of an extended sub-partition is the offset of the top-level
extended partition, not the partition before it (this is annoying, and
makes `clean' recursive mbr descent difficult). fixes PRs 11829 and 12677.

use MBR_IS_EXTENDED() instead of local test (which was missing a test case)

* Recursively print out extended partition entries. An extended partition is
effectively an MBR with it's own partition table which contains another
4 `slots', each of which can be another extended partition...
This involved reworking some of the internal functions.
* Use off_t appropriately (so we can manipulate sectors past 4GB).
* Tweak to compile with WARNS=2

convert to ANSI KNF

When fetching the disk's hardware geometry, first attempt the
DIOCGDEFLABEL ioctl, then fall back onto the DIOCGDINFO ioctl
if that fails. This ensures that we will get the actual hardware
geometry info rather than any bogus info that might have been
previously written to the disk's label area.

branches: 1.41.2;
Fix QNX partition IDs, confirmed by looking at FreeBSD's fdisk.
Note: 77 == 0x4d

Truncate the CHS values correctly in case a partition boundary exceeds
their range.

branches: 1.39.2;
Build fdisk also on arc

Update usage().
Prevent buffer overflows when configuring boot selector.
Cycle through partitions when entering boot names.
Use '<UNUSED>' consistently.

Reviewed by: fvdl

Make sure the boot selector actually gets updated if you only change
the timeout and/or the default menu choice.

Make check if we need int13 extensions for the bootselector more correct.

Only use int13 extensions if the sectors can not be read through the
older CHS interface. This works around stupid BIOSs who report that
int13 extensions are present and functional, but fail when you actually
use them. Like Adaptec SCSI BIOSs.

For the bootselector, there was no space to get the CHS info from
the BIOS. Instead, use a flag that can be set by fdisk. fdisk will
set it if one of the partitions on the disk is out of CHS reach
for this disk/BIOS, so that the bootselector will use int13 extensions.
This isn't so bad, because it needs to be configured via fdisk anyway.

Change the mbr manualpage to reflect some shorter error messages.

Make the bootselector configurable by fdisk. Add a -B flag to this
effect (i386 only of course). Also change one branch in the
bootselector code to an explicit 16 bit one, and check both
the boot menu and partition tables to see if the partition
requested by the user should be booted. This check just in case,
should the bootselector menu and partition table get out of sync
somehow. (mmm, bytesqueezing)

branches: 1.33.2;
add BeOS entry, from URA Hiroshi, PR 6974

Use <sys/disklabel_mbr.h>.

Set otherwise-unused global variable `disksectors' to disklabel.d_secperunit.
Change fdisk -S to emit disksectors as DLSIZE (for benefit of sysinst.)

Separate master bootcode from fdisk.

Completely zero out a partition table entry,
if type, start and size are given as zero.

Automagically initialize boot code and partition table if not yet valid.
Don't touch the partition table on -i, if previously valid.

Dont report more than 1023 cylinders for DOS/BIOS because
the code below will truncate to a 10 bit field. This
can cause major problems later for DOS.

Use the new NetBSD partition ID for first time setup of an MBR.

Add new NetBSD partition ID.
Update information about partition types - add missing types, insert
"reserved" for IDs which are marked as "officially reserved" in other
lists.
Sources of information:
-"How it Works -- Partition Tables" by Hale Landis <hlandis@ibm.net>
-Ralph Brown's interrupt list

When estimating the number of cylinders in the BIOS geometry, use the
total number of sectors as reported by disklabel.d_secperunit rather
than multiplying out the disklabel geometry, which is less accurate.

Don't print "DOS partition initialized" if -S set; from Nathan
Williams in PR bin/4893.

Add missing cast. This is a minimum fix to avoid overflow which causes
core dumps with some disk geometry.

branches: 1.21.2;
Backout the prototype of opendisk().

Prototype opendisk().

switch to using opendisk(3) from libutil

* support ``short'' names of disks (e.g, ``sd0'') instead of requiring
the full name (e.g, ``/dev/rsd0d'') a la disklabel(8) (where the code
was effectively lifted from ;-)
* prefix hexadecimal numbers in output with '0x'

Don't print some things if -f flag given.

-- add -b flag for setting fdisk's idea of BIOS geometry from command line
-- cleanup so WARNS=1 works
-- man page updated to reality.

Upgrade for non-interactive use. Interactive use is nearly unchanged.
-i no longer implies -u, -i now does not use the entire first track
because some BIOSes refuse to boot from a partition in the first track.
-s renamed to -S, new -s for non-interactive setting of a partition's
id, offset and size. Updated documentation to match.

Add code for a -s flag where the output is a series of VAL=xyz lines
so /bin/sh can "eval `fdisk -s`" and get variables defined with the
fdisk information.

Add a whole lot of MBR partition types, derived from a couple of sources
including Linux.
Is there a master list of this somewhere?

Make this work on big-endian systems, such as the PowerPC. From
Wolfgang Solfrank <ws@tools.de>

Add code to look at the existing partition entries to deduce the
translated geometry of the disk. This code will only change the
disk geometry if it can find a geometry which is consistent with
all partition entries, so it's very conservative about choosing
a new disk geometry. Under some circumstances, it will choose a
number of cylinders which is a little lower than it needs to be,
but this code is enough so that you can create new partitions
with the correct physical addresses.

The right solution to this problem is to have the boot loader
communicate the BIOS geometry of each disk to the kernel, in a
manner similar to the symbol table, and then provide a kernel
interface to get at that information. I'm unfortunately not
really qualified to make such a change, and I want something
that works better than the status quo.

convert to new RCS Id conventions; reduce my headache

Recognize Linux file systems correctly.

more cleanups from Jim Jegers, passed over by me.

Allow `partitioning' of a regaular file.

Remove some more uses of obsolete functions.

Clean up a bit more.

Clean this up a bit.

386bsd -> NetBSD

Add RCS identifiers, remove some completely useless RCS logs and patchkit
headers, and a few other insignificant changes.

added a 386bsd fdisk program, by julian elischer. (patch 111)