Previously, when there is an even number of offsets, we did the average
of the two middle offets but would set the REFID from one of them.
Instead, we simply select the middle offset with the lowest delay.
diff from Mike Miller <mmiller mgm51 com> (many thanks!)
OK phessler@, henning@

is undefined.  Add a warning to that effect.  OK deraadt@ jmc@

modern standards, with people having usb cameras, music players,
smartcard readers, UPSs, wifi scanners, rocket launchers and so
on...
OK sthen@, djm@

ok dcoppa ajacoutot

per SoC.  It's heavily based on the SGI port.
ok florian@

ok rapha@

ok dcoppa@ sthen@

ok matthieu@

Our stack is able to tell if the address is local or not.
ok todd@, krw@

spotted by Nathanael Rensen, thanks!

No regression has been reported since libusb became the prefered
solution to work with USB scanners.
req. by mpi@
ok ian@ mpi@ miod@

* _afs: arla/nnpfs got disabled then removed long ago
* _kdc, _kadmin: these were added 10 years ago but never used for
anything afaict
ok dcoppa@ beck@

when parsing them.
Add a couple [ug]id_t --> u_int casts for consistency with rest of code.
Based on a diff from Nathanael Rensen (nathanael (at) polymorpheus.com)

Provide a declaration for userspace
Fix the kernel's sanity check on the buflen argument
lack of prototype pointed out by sthen@ and landry@
ok kettenis@ otto@

routable. See RFC6666. Diff by Gleydson Soares.

From: Jan Stary

From: Jan Stary

members to 64bit types.  Assign new syscall numbers for (almost
all) the syscalls that involve the affected types, including anything
with time_t, timeval, itimerval, timespec, rusage, dirent, stat,
or kevent arguments.  Add a d_off member to struct dirent and replace
getdirentries() with getdents(), thus immensely simplifying and
accelerating telldir/seekdir.  Build perl with -DBIG_TIME.
Bump the major on every single base library: the compat bits included
here are only good enough to make the transition; the T32 compat
option will be burned as soon as we've reached the new world are
are happy with the snapshots for all architectures.
DANGER: ABI incompatibility.  Updating to this kernel requires extra
work or you won't be able to login: install a snapshot instead.
Much assistance in fixing userland issues from deraadt@ and tedu@
and build assistance from todd@ and otto@

headers from any subdirectory of dev/pci/drm.  If we ever add more drm(4)
drivers any headers needed by userland should end up in dev/pci/drm.
ok mpi@, miod@

Move identd_flags away from the inetd-capable daemons and fix usage.
ok deraadt@

ok deraadt@

ok sthen@ deraadt@

ok dlg@

elsewhere for running with multiple cards.

ok sthen@ deraadt@

unsigned, just like we do the existing records.  OK deraadt@

Change the logic depending upon COMPILER_VERSION everywhere, to assume gcc4
is the norm and to explicitely test for gcc3 when a different behaviour
is required.
No functional change intended. Be sure to `make install' in share/mk before
attempting to do anything.

running gcc 3, for gcc will output a reserved operand for these.
ok martynas@ matthew@ espie@ guenther@

ok deraadt@

code isn't used yet, thankfully.
first observed by djm running regress. ok deraadt djm

facts checked by sobrado@
style tweaks and ok jmc@

first sort by manual section (1, 8, 6, 2, 3, ...),
then by manual tree (share, X11R6, local)
and only for ties prefer cat over man.
For example, this avoids that /usr/local/man/cat7/roff.0
overrides /usr/share/man/man7/roff.7.
Triggered by a question from Jan Stary <hans at stare dot cz>,
ok dcoppa@ bentley@ jca@ pascal@.

no objection from deraadt@

Remove useless README -- the info(1) and man(1) pages contains all the
required information.
ok dcoppa@ robert@ beck@