use pthread_create instead of fork.

Add new Makefile knob, USE_FORT, which extends USE_SSP by turning on the
FORTIFY_SOURCE feature of libssp, thus checking the size of arguments to
various string and memory copy and set functions (as well as a few system
calls and other miscellany) where known at function entry. RedHat has
evidently built all "core system packages" with this option for some time.

This option should be used at the top of Makefiles (or Makefile.inc where
this is used for subdirectories) but after any setting of LIB.

This is only useful for userland code, and cannot be used in libc or in
any code which includes the libc internals, because it overrides certain
libc functions with macros. Some effort has been made to make USE_FORT=yes
work correctly for a full-system build by having the bsd.sys.mk logic
disable the feature where it should not be used (libc, libssp iteself,
the kernel) but no attempt has been made to build the entire system with
USE_FORT and doing so will doubtless expose numerous bugs and misfeatures.

Adjust the system build so that all programs and libraries that are setuid,
directly handle network data (including serial comm data), perform
authentication, or appear likely to have (or have a history of having)
data-driven bugs (e.g. file(1)) are built with USE_FORT=yes by default,
with the exception of libc, which cannot use USE_FORT and thus uses
only USE_SSP by default. Tested on i386 with no ill results; USE_FORT=no
per-directory or in a system build will disable if desired.

convert to new RCS Id conventions; reduce my headache

specify man pages the new way.

Update from 4.4-Lite, with local changes.

don't need -lutil

no more librpc

Add RCS identifiers.

Add some library dependencies.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Expose debugging

Sort descriptions. Fix macro. Move example to new EXAMPLES section. Add EXIT STATUS section

Let nfsd behave like all other programs: tries to use both inet4 and inet6
by default and both udp and tcp: -4 uses only inet4, -6 uses only inet6,
-t uses only tcp, -u uses only udp. For compatibility, we detect old option
usage, we warn, and DTRT.

branches: 1.21.6;
update.

branches: 1.20.10; 1.20.16;
update MAXNFSDCNT and bump date.

Consistently use 'RFC 1234' instead of 'RFC1234' or 'RFC-1234'.
From jmc@openbsd.

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

Patches provided by Joel Baker in PR 22366, verified by myself.

exeeds->exceeds, from PR misc/20741 (Igor Sobrado)

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

Whitespace nits.

Substitute the cross reference to portmap(8) with rpcbind(8).

branches: 1.13.2;
Update to use ti-rpc interface. Add -6 flag for IPv6 usage. This flag
will not cause an error if IPv6 is not available, so that things work
on v4 only kernels.

branches: 1.12.2;
Mention that the max number of daemons is 20.

Order arguments in sample usage.

Do not rely on implicit '-n' in sample usage.

branches: 1.9.2;
macroize BSD, NetBSD, FreeBSD and misc cleanup

WARNSify

Bring in NFSv3 code by Rick Macklem (from Lite2).

convert to new RCS Id conventions; reduce my headache

Update from 4.4-Lite, with local changes.

Add RCS indentifiers.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

usr.sbin: remove unnecessary CONSTCOND, lint no longer needs it

Since 2021-01-31, lint no longer requires a CONSTCOND comment in a
do-while-0 statement since this is a common code pattern, especially in
statement-like macros.

sed -i -E 's,} while \(/\* ?CONSTCOND ?\*/ ?0\),} while (0),' */*.[ch]

Revert previous, don't write to the parent too early because it can exit
before the child is ready to serve. The child will write in daemon2_detach()
when it is ready.
While here:
- return EXIT_{SUCCESS,FAILURE)
- check syscall error against -1

Fix indentation

Remove trailing whitespace

Remove superfluous empty line

Include explicitly <rump/rump_syscallshotgun.h> for previous indirect users
via <rump/rump.h>.

Don't exit if we have udp only sockets, instead wait for all the workers
to terminate.

PR/51437: Hauke Fath: write a byte to the socket so that nfsd -u does not die.

Expose debugging

Rump-ify.

Typo

Memset the whole cfg array, not piecemeal because otherwise you'll end
up with random data depending on the options set.
XXX: pullup 7

branches: 1.62.4;
When daemonizing, delay the parent exit until the daemon is
ready to provide service. Fixes one of the race conditions
of PR misc/48282.

branches: 1.61.2;
Fix memset usage.

use freeaddrinfo()

Let nfsd behave like all other programs: tries to use both inet4 and inet6
by default and both udp and tcp: -4 uses only inet4, -6 uses only inet6,
-t uses only tcp, -u uses only udp. For compatibility, we detect old option
usage, we warn, and DTRT.

branches: 1.58.2;
static + ANSIfy + __dead

PR/42486: Ed Ravin nfsd sometimes exits prematurely during port-scan fix
from freebsd bin 61084 applied (preserving errno properly). The accept
code should be merged.

Silly-rename child to worker (no pun intended .. maybe).
nfsd is threading these days.

use varargs macro to make -DDDEBUG compile

branches: 1.54.4;
Remove the \n and tabs from the __COPYRIGHT() strings.
Tweak to use a consistent format.

branches: 1.53.4;
remove MAXNFSDCNT.

The ISO stuff has never worked, remove it.

branches: 1.51.4; 1.51.8;
after switching to threads:
- g/c SIGCHLD handler.
- pthread_setname_np rather than setproctitle.

the patch provided by Andrew Doran.

use pthread_create instead of fork.

branches: 1.49.2;
PR/18906: roskens at elfin dot net: misc. select() to poll() updates.

Adapted to -current by myself, thanks for the patch!

In two TCP error cases, say that it's TCP, not UDP.
Pointed out by pee- on IRCnet #NetBSD.

branches: 1.47.2;
change (mostly) int to socklen_t. GCC 4 doesn't like that int and
socklen_t are different signness.

Since KerberosIV has gone away, so should NFSKERB.

appease gcc -Wuninitialized

Increase userspace (/usr/sbin/nfsd) limit on nfsd count from 20 to 1024.

Jason Thorpe suggests removing the limit altogether; anyone who needs
more than 1024 nfsds can still do "nfsd -n 1024" multiple times, and
this provides _some_ saftey-net against nfsd (issued by root) from
eating all the avaiable process slots.

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

Patches provided by Joel Baker in PR 22366, verified by myself.

include des.h, not kerberosIV/des.h

Fix PR 18391, from Geoff Wing.

null commit

poll.h, not sys/poll.h

Set fd of unused pollfd entries to -1.

select() -> poll()

s/IPV6_BINDV6ONLY/IPV6_V6ONLY/ to follow changes made yesterday.

comment or delete text after CPP directives.

minor knf

remove unnecessary ":" in openlog ident

Update to use ti-rpc interface. Add -6 flag for IPv6 usage. This flag
will not cause an error if IPv6 is not available, so that things work
on v4 only kernels.

branches: 1.31.6;
Per the documented behaviour in the man page, default to
serving UDP requests if none of the -t or -u options are
given.

Previously, nfsd wouldn't actually do anything at all
if it was not told to do UDP and/or TCP, which could be
confusing.

Note that at least Solaris and IRIX default to both TCP
and UDP when nfsd is run without options. This might be
useful for NetBSD as well.

Fixes PR bin/7358.

Re-indent some of the NFSKERB code, to make the entire loop its in easier
to read.

Add a missing close bracket, in some currently-unused code.

branches: 1.28.2;
fix comments

Do not block SIGTERM, so that nfsd exits gracefully at shutdown.

minor KNF.

change includes to fix compiler warning

use memcmp/memset/memmove instead of bcmp/bzero/bcopy

WARNSify

getopt(3) usage police: EOF is deprecated, check against -1.

Add missing argument to 2 warnx() calls. From <msaitoh@spa.is.uec.ac.jp>,
PR 4086.

Clear the sin_zero fields explicitly.

branches: 1.19.4;
Reorder #includes.

Bring in NFSv3 code by Rick Macklem (from Lite2).

<sys/syslog.h> -> <syslog.h>

convert to new RCS Id conventions; reduce my headache

Better patch for the previous.

Don't wedge in an infinite loop if NFSSERVER is not configured.

Change process titles to be less redundant.

Eliminate uses of some obsolete functions.

branches: 1.11.2;
Update from 4.4-Lite, with local changes.

use setproctitle

should bzero automatic sockaddr's before use

use syslog to report errors Vernon Schryver (vjs@calcite.rhyolite.com)

Add RCS identifiers.

update to use PS_STRINGS to set proc title

now you too can partake in this offense against any higher being there
may be. basically to find out whether nfsserver support is in the kernel
earlier and thus avoid the loop problem, call nfssvc() with a bogus fd and
see whether you get a -1 return value or a SIGSYS...

doesn't core when nfsserver support is not in kernel

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision