fix various typos in comments and output/log messages.

Add a prog_setsockopt (thanks kre@)

route: Enable receive error reporting on the routing socket

Avoid undefined behavior

route.c:1523:20, left shift of 1073741824 by 1 places cannot be represented
in type 'int'

route: address flags are more useful than interface flags for address msgs

route: Decode the correct flags for interface address messages

route(8): Show addres for RTM_LOSING

While here, add /* FALLTHROUGH */.

route(8): show addrs for RTM_MISS

route(8): print RTM_CHANGE messages

XXX Pullup -9

branches: 1.160.2;
Resolve every route(4) message type before printing anything more about it
other than the type and length.

This solves the issue where RTM_ONEWADDR is received and it tries to
parse route information from it.

branches: 1.159.2;
Handle the routing socket overflowing gracefully.

branches: 1.158.2;
Use the definitions from the standard header files and replace homebrew
snprintb with the libutil one.

Fix typo in flag name. We should probably just use IFFBITS string
that <net/if.h> defines.

PR bin/52815

Support -c <count> option for route monitor

route command exits if it receives <count> routing messages where
<count> is a value specified by -c.

The option is useful to get only particular message(s) in a test script.

branches: 1.155.4;
Report the PID for the process changing the address.

branches: 1.154.2; 1.154.4;
Separate nexthop caches from the routing table

By this change, nexthop caches (IP-MAC address pair) are not stored
in the routing table anymore. Instead nexthop caches are stored in
each network interface; we already have lltable/llentry data structure
for this purpose. This change also obsoletes the concept of cloning/cloned
routes. Cloned routes no longer exist while cloning routes still exist
with renamed to connected routes.

Noticeable changes are:
- Nexthop caches aren't listed in route show/netstat -r
- sysctl(NET_RT_DUMP) doesn't return them
- If RTF_LLDATA is specified, it returns nexthop caches
- Several definitions of routing flags and messages are removed
- RTF_CLONING, RTF_XRESOLVE, RTF_LLINFO, RTF_CLONED and RTM_RESOLVE
- RTF_CONNECTED is added
- It has the same value of RTF_CLONING for backward compatibility
- route's -xresolve, -[no]cloned and -llinfo options are removed
- -[no]cloning remains because it seems there are users
- -[no]connected is introduced and recommended
to be used instead of -[no]cloning
- route show/netstat -r drops some flags
- 'L' and 'c' are not seen anymore
- 'C' now indicates a connected route
- Gateway value of a route of an interface address is now not
a L2 address but "link#N" like a connected (cloning) route
- Proxy ARP: "arp -s ... pub" doesn't create a route

You can know details of behavior changes by seeing diffs under tests/.

Proposed on tech-net and tech-kern:
http://mail-index.netbsd.org/tech-net/2016/03/11/msg005701.html

no need for <sys/mbuf.h>

PR/50671: David Binderman: fix memory leak.

Add RTF_BROADCAST to mark routes used for the broadcast address when
they are created on the fly. This makes it clear what the route is for
and allows an optimisation in ip_output() by avoiding a call to
in_broadcast() because most of the time we do talk to a host.
It also avoids a needless allocation for the storage of llinfo_arp and
thus vanishes from arp(8) - it showed as incomplete anyway so this
is a nice side effect.

Guard against this and routes marked with RTF_BLACKHOLE in
ip_fastforward().
While here, guard against routes marked with RTF_BLACKHOLE in
ip6_fastforward().
RTF_BROADCAST is IPv4 only, so don't bother checking that here.

Teach route(8) about RTF_LOCAL.

MKINET6=no build fixes

http://mail-index.netbsd.org/tech-net/2014/12/13/msg004898.html

Note that RTM_CHANGE can also change the gateway

PR/47704: Takahiro HAYASHI: Add -L flag

Now that the code can support tags and verbose route printing, enable them.
Also document previously undocumented flags.

Factor out the netstat route printing code and use it here. There is no
point in having 2 different copies; fixes PR/49371

branches: 1.144.4;
use correct function

use symbolic flags

use scopeid functions

avoid pointer gymnastics

Retire OSI network stack. OK core@

Fix fallout from 1.129 that converted sou::so_foo from unions to
pointers but missed (char *)&soup->so_foo => (char *)soup->so_foo in
mask_addr(). It worked by luck - unless it didn't: due to pointer
numerology on amd64 route add -net ClassC without explicit /24 prefix
length specification would result into /16 destination instead of /24.

branches: 1.138.2;
remove useless rump headers.

PR/42179: Christoph Badura: Be a little friendlier about missing args.

branches: 1.136.2;
Off by one and clarity improvement from Christian Biere

PR/45661: Henning Petersen: Overlapping buffer in route.c.

Move RTF_ANNOUNCE flag so that it no longer conflicts with RTF_PROTO2.

RTF_ANNOUNCE was defined as RTF_PROTO2. The flag is used to indicated
that host should act as a proxy for a link level arp or ndp request.
(If RTF_PROTO2 is used as an experimental flag (as advertised),
various problems can occur.)

This commit provides a first-class definition with its own bit for
RTF_ANNOUNCE, removes the old aliasing definitions, and adds support
for the new RTF_ANNOUNCE flag to netstat(8) and route(8).,

Also, remove unused RTF_ flags that collide with RTF_PROTO1:
netinet/icmp6.h defined RTF_PROBEMTU as RTF_PROTO1
netinet/if_inarp.h defined RTF_USETRAILERS as RTF_PROTO1
(Neither of these flags are used anywhere. Both have been removed
to reduce chances of collision with RTF_PROTO1.)

Figuring this out and the diff are the work of Beverly Schwartz of
BBN.

(Passed release build, boot in VM, with no apparently related atf
failures.)

Approved for Public Release, Distribution Unlimited
This material is based upon work supported by the Defense Advanced
Research Projects Agency and Space and Naval Warfare Systems Center,
Pacific, under Contract No. N66001-09-C-2073.

branches: 1.133.2;
Simplify.

Use __dead

Fix memset usage.

don't compile tag functions if SMALL is defined. Should fix the build

Alloc dynamically sockunions in sou.
Teach route(8) about multiple tags. E.G.: -tag 100,20,33
XXX: needs documentation

Add a new AF/PF_ROUTE which is 64-bit clean which makes the routing socket
interface (and its associated sysctls) act identically for both 32 and 64 bit
programs. The old unclean one remains for backward compatibility.

branches: 1.127.2;
Convert from RUMP_ACTION to RUMPPRG.

Support the new RTM_IFINFO and RTM_CHGADDR messages.

adapt to new rumpclient_init() signature

#ifdef variable correctly

be consistent in what to leave outside of SMALL

Support RUMP_ACTION to compile a rump kernel client.

add MPLS clue - mpls and tag keywords

PR/40455: Mihai Chelaru: Add noreject noblackhole options

parse unsigned long with strtoul not atoi

branches: 1.118.6;
Get the socket addresses for dst/gateway/netmask/interface/ifaddr
out of the global namespace.

Don't cast a sockunion * to sockaddr *. Instead, reference the
sockaddr inside of the sockunion.

Delete superfluous casts to void *.

Constify.

Constify.

Constify.

For clarity's sake, do not assign a new value to inet_makenetandmask()'s
argument, `inet'.

Extract subroutine extract_addrs() for reuse, later.

Replace the anonymous constant, 0, with AF_UNSPEC.

Constify.

Shorten a staircase. Remove superfluous parentheses from return
statements. Constify.

Use a union and sizeof() to avoid a cast and an anonymous constant
buffer length, respectively.

Use bool.

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

branches: 1.106.12; 1.106.16;
TSEL() is const char* now

Fix a bug that caused 0.0.0.0/0 to be treated like 0.0.0.0/32.

foo << 32 is undefined per section 6.5.7 of the C99 standard
if foo is (or is promoted to) a 32-bit type.

Let us qualify 'route flush' and 'route show' commands with -llinfo
and -host flags, which is useful for displaying/flushing ARP entries:

# route -n show -inet -llinfo -host
Routing table

Internet:
Destination Gateway Flags
169.254.1.119 link#1 UH
169.254.230.110 00:02:6f:21:e6:6e UH
# route -n flush -inet -llinfo -host
169.254.1.119 done
169.254.230.110 0.2.6f.21.e6.6e done
169.254.237.70 done
# route -n show -inet -llinfo -host
Routing table
#

This will help me address bin/11079, "dhclient may require arp and sed".

Extract common code from flushroutes() and show(), creating
parse_show_opts().

While I'm here, make small cosmetic changes to flushroutes().

Move the #ifndef SMALL code together.

Use the symbol AF_UNSPEC instead of 0 as appropriate.

Remove dead code.

Remove XNS and CCITT/X25 bits.

KNF err(3), errx(3) usage, indentation, braces.

Add route keywords -nocloned, -nocloning.

Use default option order (AaBbCc...).

Print the INET netmask if it is non-contiguous.

PR/32632: Yves-Emmanuel JUTARD: Improvment suggestion in 'route' display.
Add a new -S flag that prints a space for missing flags.

PR/32629: Yves-Emmanuel JUTARD: Type errors in sbin/route/route.c, another
u_char/char conflict.

Add full IPv6 syntax support when route is built with -DSMALL and -DINET6.
Fixes PR/23937. Adapted from supplied patch.

Clean up some minor nits:
+ style (KNF).
. don't cast NULL
. white space clean up
+ use consistent format for error messages.
+ plug memory leaks and avoid malloc(0).
+ verbose message should display even when routing table is empty
(flushroutes).

Make sure that a netmask socket address value prints appropriately when the
family is unset. Either the destination or interface address socket address
family value is used. This change prevents the following sort of output:

RTM_NEWADDR: address being added to iface: len 60, metric 0, flags:<UP>
sockaddrs: <NETMASK,IFA,BRD>
(0) 00.00.ff.00.00 127.0.0.1 127.0.0.1

With this change the last line becomes:

255.0.0.0 127.0.0.1 127.0.0.1

Also make sure that when using a generated inet netmask the address family
is set. With this change a generated netmask will print appropriately
when using the verbose (-v) option.

Make sure the get command always exits non-zero on error. This makes things
like the following work as expected.

#!/bin/sh

if ! route -sn get default >/dev/null 2>&1; then
echo default gateway not set
exit 1
fi
echo default gateway set

Handle routing socket write(2) errors when they occur. This produces
better diagnostics by allowing for handling of the special route errno
values ESRCH, EBUSY, and ENOBUFS even with the quiet (-q) option or
when doing a get command.

Trim the local domain name in routename() for INET6 lookups; similar to
INET lookups.

Make this build for __GNUC__ <= 2 (i.e. for vax). The cast to
caddr_t removes a const qualifier. Instead, cast the first element
of the subtraction to "struct sockaddr *" to work around the problem.

Accept protocol flags, i.e. -inet, -inet6, with the show command.
[From OpenBSD via DragonFly BSD]

Fix the default case in sodump(). Printing the address of the "which"
string isn't really that useful. Instead print the string and the value
of sa_family.

More style nits (KNF); casts and sizeof's are not followed by a space.

switch is not a function, so add a space; sort switch elements
alphabetically (KNF).

* Exit non-zero error code when flushroute() fails operating on the
routing socket, e.g. running route flush as non-root.

* When quiet (-q) don't append error string to a non-existent
diagnostic output (stdout). The error diagnostic is still written
to stderr. [from FreeBSD]

* Use warn(3) in place of perror(3).

const poisoning.

Teach route(8) to print the messages generated by net80211.

Convert prefixlen() to ANSI function declaration. Missed by earlier
conversion.

Due to a limitation in inet_addr(3) it was not possible to specify a
netmask value of 0xffffffff (INADDR_NONE). Fix by using inet_aton(3).
Note it was possible to specify the same mask as the dotted quad
255.255.255.255 because of a later call to gethostbyname(3).
[From FreeBSD]

Fix botched prior commit. Additional debugging error messages mistakenly
committed.

* Honor -net when using CIDR (net/mask) notation. This differs
from the default CIDR (host/mask) interpretation. As discussed on
tech-net.

* Use errx() in place of fprintf() and exit(). Make the error message
formats consistent.

Fix macro NEXTADDR so that the names, like so_dst, are printed
rather than u. The preprocessor doesn't do replacement within
strings, so the # string creation operator must be used.

Kill __P(), use ANSI function declarations.

NI_WITHSCOPEID was not picked up by IETF standardization process.

bump buffer size for any_ntoa(), for huge sa_len

2-byte read overrun for unknown sockaddr. from openbsd
XXX out[64], is it safe? what if sa_len > 23?

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

print prefixlen on -show

Remove a comment that referred to a non-committed part of the patch I
submitted for bin/17613. Ok'd by mrg@.

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

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

This is an updated submitted patch originally written by Jim Rees
and sent in by Greg Hudson as seen in PR misc/3227. Basically what it
does is adds a flushall option which deletes all but localhost routes.
This is done by andoring in a flag called doall (1 means do all routes
including gateway, 0 means do a regular flush). I have seen some
platforms that do this. I tested it out on ipv4 only, it works as
advertised. Commit was approved by christos@.

PR/21856: Robert Elz: Changes to route(8) to allow /n instead of -prefixlen
for INET6

permit foo/bar notation, like "route add 10.0.0.0/8 127.0.0.1".
originally by provos

strlcpy [a few mistakes found]

This addresses PR 17613 when certain methods of adding routes would
muck up the routing tables. The patch was submitted by Quentin Garnier
<netbsd@quatriemek.com> and tweaked a little after it was reviewed
by Christos Zoulas <christos@netbsd.org>. Final change/commit approved
by Christos.

Made a note in the code about the -d and -d options brought up in
PR 18674. Thanks to Chris Jones for pointing it out.

correct prefixlen handling for IPv4.

Make prefixlen check to make sure its argument is present before deref
from openbsd

sweep of errx/warnx, remove unnecessary trailing \n

permit DNS names to address portion. they will be considered as RTF_HOST
if specified as destination (as there's no getnetbyname in IPv6).

if we're unsure about prefixlen, don't say "128".

more correct ion to RTF_HOST handling.

do not have -prefixlen 128 on RTF_HOST.
From: Dawid Szymanski <dawszy@arhea.net>

branches: 1.54.2;
raise RTF_HOST if destination is a host (/128). from ume

PR/15672: Love: route monitor missing info

Make route exit with a non-zero error code when operations to the routing
socket fails. eg, running route add 1.2.3.4 5.6.7.8 as non root.

fix -Wshadow warnings

Provide a short output format for the get command that only prints the
address corresponding to RTA_GATEWAY, or nothing if it doesn't exist.
Modify the exit value of route depending on this, so that one can do
stuff like:

#!/bin/sh
gw=`route -sn get default 2>/dev/null`
if [ -z "$gw" ]; then
echo no default route
exit 1
fi
ping -w1 -c1 $gw >/dev/null 2>&1
if ! route -sn get $gw >/dev/null; then
echo default gateway not responding
exit 1
fi
echo default gateway is at $gw

Print a R on reject routes to indicate that they're not normal routes,
and check netmask (or prefix) length as well as the destination
address when determining if a route is a "default" or not. This means
that the output from 'route show' will no longer say:

Internet6:
Destination Gateway Flags
default localhost UG
default localhost UG

but instead

Internet6:
Destination Gateway Flags
::/104 localhost UGR
::/96 localhost UGR

which makes much more sense.

off-by-one error in -prefixlen arg validation (do not permit 129 on IPv6)

make prefixlen work on IPv4 too

The way pmsg_addrs() is used requires that it always prints a newline and
flushes stdout before returning.

Previously the decoded output for a message was stuck in stdio buffers until
the next arrived.

convert to use getprogname()

remove nested externs.

support -cloned command line flag.
i don't think anyone ever going to use it.

mark cloned routes with RTF_CLONED. present it with netstat -r by "c".

let static routes overwrite cloned routes, as cloned routes can come back again
if necessary. behavior same as freebsd/bsdi, code partially from bsdi42.
(NRL rt->rt_parent was not added)
should fix PR 11916 and maybe some other PRs with ARP behavior.

recompilation of usr.sbin/route6d is suggested.

support RTF_BLACKHOLE.

XXX route show is total duplicate of netstat -r, we need to either remove
route show, or share the same source code, otherwise maintenance cost
bites (and is biting) us

remove use of uninitialized route (struct ortentry).
it seems that the bug was born during 4.3reno -> 4.4Lite1 transition.

on "route monitor", print RTM_IFANNOUNCE, RTM_IFINFO (new #), and others.

branches: 1.38.4;
Allow INET6 to be used with SMALL without undefined symbols

use scoped notation for link-local multicast too.

support extended scoped address notation (fe80::1%de0).
it hides implementation detail (embedded ifid) from the users.

don't make "-prefixlen 64" a default in -inet6 case. the past behavior
was not compatible with past practice ("route get" did not work right).

sync with recent KAME.

Display local address associated with route (if any) in "route get".

branches: 1.33.2;
fix alignment issue in routing socket data, on 64bit arch (i.e. alpha).
sockaddrs are packed with ROUNDUP() in sys/net/rtsock.c so userland
programs needs to follow that.

PR: 8179

Incorrect integration. ctime takes a time_t value, not a long.

IPv6-ready route command (-inet6 and -prefixlen are added)

branches: 1.30.2;
* implement -f; flush all routes before executing command
* reset af to 0 in flushroutes() and newroute()
* cleanup -DSMALL:
- `flush' and AF_APPLETALK is supported (the latter had partial support)
- AF_NS, AF_ISO, and AF_CCITT are not supported at all (rather than being
- partially unsupported).

const poisoning.

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

64-bit sweep, "route -v show" was spinning forever.

Use {warn,err}{,x}(3) instead of printf's

Don't declare optind.

resolve conflicts from lite-2 import

* cleanup for WARNS=1
* use .Nm correctly
* deprecate register
* getopt returns -1 not EOF

Fix uninitialized pointer botch when attempting to flush an already
empty routing table. From Matthias Scheler <tron@lyssa.owl.de>, in
PR #4074.

Don't SEGV if a qualifier is passed without an argument. From
Geoff Wing <mason@primenet.com.au>, fixes PR #2700.

KNF. be safe with buffers.

- Add netatalk bits
- Add prototypes
- Fix printf formats

Remove uid checks since the kernel now handles this and route(1) is no
longer setuid.
Document the "Permission denied" msg in the man page since this is
_slightly_ less clear than route(1)'s old error msg.

As discussed some time ago, add a "show" sub-command.
Also allow compilation with -DSMALL to ommit support for
non-essential protocols (i.e. when built into a ramdisk).
Use a less hack-ish way to generate keywords.[ch] and just
check in the result (helps ../../distrib/utils/x_route).

check returns of inet_addr() and inet_network() against INADDR_NONE,
not -1, per the manual page. on 64-bit systems like the alpha, -1 will
never equal (unsigned long)INADDR_NONE.

Produce less output on -q as promised by the man page (John Hawkinson; PR#998).

Interface-specific routes can be static.

Implement the document `-llinfo' flag.

Remove extra `break;'.

convert to new RCS Id conventions; reduce my headache

kill some extraneous equal signs.

Remove some more uses of obsolete functions.

Eliminate uses of some obsolete functions.

Clean up import.

Add RCS identifiers.

Compile X.25, OSI, and NS code.

Cleanup for GCC 2.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

Add a prog_setsockopt (thanks kre@)

route: Enable receive error reporting on the routing socket

Avoid undefined behavior

route.c:1523:20, left shift of 1073741824 by 1 places cannot be represented
in type 'int'

route: address flags are more useful than interface flags for address msgs

route: Decode the correct flags for interface address messages

route(8): Show addres for RTM_LOSING

While here, add /* FALLTHROUGH */.

route(8): show addrs for RTM_MISS

route(8): print RTM_CHANGE messages

XXX Pullup -9

branches: 1.160.2;
Resolve every route(4) message type before printing anything more about it
other than the type and length.

This solves the issue where RTM_ONEWADDR is received and it tries to
parse route information from it.

branches: 1.159.2;
Handle the routing socket overflowing gracefully.

branches: 1.158.2;
Use the definitions from the standard header files and replace homebrew
snprintb with the libutil one.

Fix typo in flag name. We should probably just use IFFBITS string
that <net/if.h> defines.

PR bin/52815

Support -c <count> option for route monitor

route command exits if it receives <count> routing messages where
<count> is a value specified by -c.

The option is useful to get only particular message(s) in a test script.

branches: 1.155.4;
Report the PID for the process changing the address.

branches: 1.154.2; 1.154.4;
Separate nexthop caches from the routing table

By this change, nexthop caches (IP-MAC address pair) are not stored
in the routing table anymore. Instead nexthop caches are stored in
each network interface; we already have lltable/llentry data structure
for this purpose. This change also obsoletes the concept of cloning/cloned
routes. Cloned routes no longer exist while cloning routes still exist
with renamed to connected routes.

Noticeable changes are:
- Nexthop caches aren't listed in route show/netstat -r
- sysctl(NET_RT_DUMP) doesn't return them
- If RTF_LLDATA is specified, it returns nexthop caches
- Several definitions of routing flags and messages are removed
- RTF_CLONING, RTF_XRESOLVE, RTF_LLINFO, RTF_CLONED and RTM_RESOLVE
- RTF_CONNECTED is added
- It has the same value of RTF_CLONING for backward compatibility
- route's -xresolve, -[no]cloned and -llinfo options are removed
- -[no]cloning remains because it seems there are users
- -[no]connected is introduced and recommended
to be used instead of -[no]cloning
- route show/netstat -r drops some flags
- 'L' and 'c' are not seen anymore
- 'C' now indicates a connected route
- Gateway value of a route of an interface address is now not
a L2 address but "link#N" like a connected (cloning) route
- Proxy ARP: "arp -s ... pub" doesn't create a route

You can know details of behavior changes by seeing diffs under tests/.

Proposed on tech-net and tech-kern:
http://mail-index.netbsd.org/tech-net/2016/03/11/msg005701.html

no need for <sys/mbuf.h>

PR/50671: David Binderman: fix memory leak.

Add RTF_BROADCAST to mark routes used for the broadcast address when
they are created on the fly. This makes it clear what the route is for
and allows an optimisation in ip_output() by avoiding a call to
in_broadcast() because most of the time we do talk to a host.
It also avoids a needless allocation for the storage of llinfo_arp and
thus vanishes from arp(8) - it showed as incomplete anyway so this
is a nice side effect.

Guard against this and routes marked with RTF_BLACKHOLE in
ip_fastforward().
While here, guard against routes marked with RTF_BLACKHOLE in
ip6_fastforward().
RTF_BROADCAST is IPv4 only, so don't bother checking that here.

Teach route(8) about RTF_LOCAL.

MKINET6=no build fixes

http://mail-index.netbsd.org/tech-net/2014/12/13/msg004898.html

Note that RTM_CHANGE can also change the gateway

PR/47704: Takahiro HAYASHI: Add -L flag

Now that the code can support tags and verbose route printing, enable them.
Also document previously undocumented flags.

Factor out the netstat route printing code and use it here. There is no
point in having 2 different copies; fixes PR/49371

branches: 1.144.4;
use correct function

use symbolic flags

use scopeid functions

avoid pointer gymnastics

Retire OSI network stack. OK core@

Fix fallout from 1.129 that converted sou::so_foo from unions to
pointers but missed (char *)&soup->so_foo => (char *)soup->so_foo in
mask_addr(). It worked by luck - unless it didn't: due to pointer
numerology on amd64 route add -net ClassC without explicit /24 prefix
length specification would result into /16 destination instead of /24.

branches: 1.138.2;
remove useless rump headers.

PR/42179: Christoph Badura: Be a little friendlier about missing args.

branches: 1.136.2;
Off by one and clarity improvement from Christian Biere

PR/45661: Henning Petersen: Overlapping buffer in route.c.

Move RTF_ANNOUNCE flag so that it no longer conflicts with RTF_PROTO2.

RTF_ANNOUNCE was defined as RTF_PROTO2. The flag is used to indicated
that host should act as a proxy for a link level arp or ndp request.
(If RTF_PROTO2 is used as an experimental flag (as advertised),
various problems can occur.)

This commit provides a first-class definition with its own bit for
RTF_ANNOUNCE, removes the old aliasing definitions, and adds support
for the new RTF_ANNOUNCE flag to netstat(8) and route(8).,

Also, remove unused RTF_ flags that collide with RTF_PROTO1:
netinet/icmp6.h defined RTF_PROBEMTU as RTF_PROTO1
netinet/if_inarp.h defined RTF_USETRAILERS as RTF_PROTO1
(Neither of these flags are used anywhere. Both have been removed
to reduce chances of collision with RTF_PROTO1.)

Figuring this out and the diff are the work of Beverly Schwartz of
BBN.

(Passed release build, boot in VM, with no apparently related atf
failures.)

Approved for Public Release, Distribution Unlimited
This material is based upon work supported by the Defense Advanced
Research Projects Agency and Space and Naval Warfare Systems Center,
Pacific, under Contract No. N66001-09-C-2073.

branches: 1.133.2;
Simplify.

Use __dead

Fix memset usage.

don't compile tag functions if SMALL is defined. Should fix the build

Alloc dynamically sockunions in sou.
Teach route(8) about multiple tags. E.G.: -tag 100,20,33
XXX: needs documentation

Add a new AF/PF_ROUTE which is 64-bit clean which makes the routing socket
interface (and its associated sysctls) act identically for both 32 and 64 bit
programs. The old unclean one remains for backward compatibility.

branches: 1.127.2;
Convert from RUMP_ACTION to RUMPPRG.

Support the new RTM_IFINFO and RTM_CHGADDR messages.

adapt to new rumpclient_init() signature

#ifdef variable correctly

be consistent in what to leave outside of SMALL

Support RUMP_ACTION to compile a rump kernel client.

add MPLS clue - mpls and tag keywords

PR/40455: Mihai Chelaru: Add noreject noblackhole options

parse unsigned long with strtoul not atoi

branches: 1.118.6;
Get the socket addresses for dst/gateway/netmask/interface/ifaddr
out of the global namespace.

Don't cast a sockunion * to sockaddr *. Instead, reference the
sockaddr inside of the sockunion.

Delete superfluous casts to void *.

Constify.

Constify.

Constify.

For clarity's sake, do not assign a new value to inet_makenetandmask()'s
argument, `inet'.

Extract subroutine extract_addrs() for reuse, later.

Replace the anonymous constant, 0, with AF_UNSPEC.

Constify.

Shorten a staircase. Remove superfluous parentheses from return
statements. Constify.

Use a union and sizeof() to avoid a cast and an anonymous constant
buffer length, respectively.

Use bool.

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

branches: 1.106.12; 1.106.16;
TSEL() is const char* now

Fix a bug that caused 0.0.0.0/0 to be treated like 0.0.0.0/32.

foo << 32 is undefined per section 6.5.7 of the C99 standard
if foo is (or is promoted to) a 32-bit type.

Let us qualify 'route flush' and 'route show' commands with -llinfo
and -host flags, which is useful for displaying/flushing ARP entries:

# route -n show -inet -llinfo -host
Routing table

Internet:
Destination Gateway Flags
169.254.1.119 link#1 UH
169.254.230.110 00:02:6f:21:e6:6e UH
# route -n flush -inet -llinfo -host
169.254.1.119 done
169.254.230.110 0.2.6f.21.e6.6e done
169.254.237.70 done
# route -n show -inet -llinfo -host
Routing table
#

This will help me address bin/11079, "dhclient may require arp and sed".

Extract common code from flushroutes() and show(), creating
parse_show_opts().

While I'm here, make small cosmetic changes to flushroutes().

Move the #ifndef SMALL code together.

Use the symbol AF_UNSPEC instead of 0 as appropriate.

Remove dead code.

Remove XNS and CCITT/X25 bits.

KNF err(3), errx(3) usage, indentation, braces.

Add route keywords -nocloned, -nocloning.

Use default option order (AaBbCc...).

Print the INET netmask if it is non-contiguous.

PR/32632: Yves-Emmanuel JUTARD: Improvment suggestion in 'route' display.
Add a new -S flag that prints a space for missing flags.

PR/32629: Yves-Emmanuel JUTARD: Type errors in sbin/route/route.c, another
u_char/char conflict.

Add full IPv6 syntax support when route is built with -DSMALL and -DINET6.
Fixes PR/23937. Adapted from supplied patch.

Clean up some minor nits:
+ style (KNF).
. don't cast NULL
. white space clean up
+ use consistent format for error messages.
+ plug memory leaks and avoid malloc(0).
+ verbose message should display even when routing table is empty
(flushroutes).

Make sure that a netmask socket address value prints appropriately when the
family is unset. Either the destination or interface address socket address
family value is used. This change prevents the following sort of output:

RTM_NEWADDR: address being added to iface: len 60, metric 0, flags:<UP>
sockaddrs: <NETMASK,IFA,BRD>
(0) 00.00.ff.00.00 127.0.0.1 127.0.0.1

With this change the last line becomes:

255.0.0.0 127.0.0.1 127.0.0.1

Also make sure that when using a generated inet netmask the address family
is set. With this change a generated netmask will print appropriately
when using the verbose (-v) option.

Make sure the get command always exits non-zero on error. This makes things
like the following work as expected.

#!/bin/sh

if ! route -sn get default >/dev/null 2>&1; then
echo default gateway not set
exit 1
fi
echo default gateway set

Handle routing socket write(2) errors when they occur. This produces
better diagnostics by allowing for handling of the special route errno
values ESRCH, EBUSY, and ENOBUFS even with the quiet (-q) option or
when doing a get command.

Trim the local domain name in routename() for INET6 lookups; similar to
INET lookups.

Make this build for __GNUC__ <= 2 (i.e. for vax). The cast to
caddr_t removes a const qualifier. Instead, cast the first element
of the subtraction to "struct sockaddr *" to work around the problem.

Accept protocol flags, i.e. -inet, -inet6, with the show command.
[From OpenBSD via DragonFly BSD]

Fix the default case in sodump(). Printing the address of the "which"
string isn't really that useful. Instead print the string and the value
of sa_family.

More style nits (KNF); casts and sizeof's are not followed by a space.

switch is not a function, so add a space; sort switch elements
alphabetically (KNF).

* Exit non-zero error code when flushroute() fails operating on the
routing socket, e.g. running route flush as non-root.

* When quiet (-q) don't append error string to a non-existent
diagnostic output (stdout). The error diagnostic is still written
to stderr. [from FreeBSD]

* Use warn(3) in place of perror(3).

const poisoning.

Teach route(8) to print the messages generated by net80211.

Convert prefixlen() to ANSI function declaration. Missed by earlier
conversion.

Due to a limitation in inet_addr(3) it was not possible to specify a
netmask value of 0xffffffff (INADDR_NONE). Fix by using inet_aton(3).
Note it was possible to specify the same mask as the dotted quad
255.255.255.255 because of a later call to gethostbyname(3).
[From FreeBSD]

Fix botched prior commit. Additional debugging error messages mistakenly
committed.

* Honor -net when using CIDR (net/mask) notation. This differs
from the default CIDR (host/mask) interpretation. As discussed on
tech-net.

* Use errx() in place of fprintf() and exit(). Make the error message
formats consistent.

Fix macro NEXTADDR so that the names, like so_dst, are printed
rather than u. The preprocessor doesn't do replacement within
strings, so the # string creation operator must be used.

Kill __P(), use ANSI function declarations.

NI_WITHSCOPEID was not picked up by IETF standardization process.

bump buffer size for any_ntoa(), for huge sa_len

2-byte read overrun for unknown sockaddr. from openbsd
XXX out[64], is it safe? what if sa_len > 23?

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

print prefixlen on -show

Remove a comment that referred to a non-committed part of the patch I
submitted for bin/17613. Ok'd by mrg@.

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

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

This is an updated submitted patch originally written by Jim Rees
and sent in by Greg Hudson as seen in PR misc/3227. Basically what it
does is adds a flushall option which deletes all but localhost routes.
This is done by andoring in a flag called doall (1 means do all routes
including gateway, 0 means do a regular flush). I have seen some
platforms that do this. I tested it out on ipv4 only, it works as
advertised. Commit was approved by christos@.

PR/21856: Robert Elz: Changes to route(8) to allow /n instead of -prefixlen
for INET6

permit foo/bar notation, like "route add 10.0.0.0/8 127.0.0.1".
originally by provos

strlcpy [a few mistakes found]

This addresses PR 17613 when certain methods of adding routes would
muck up the routing tables. The patch was submitted by Quentin Garnier
<netbsd@quatriemek.com> and tweaked a little after it was reviewed
by Christos Zoulas <christos@netbsd.org>. Final change/commit approved
by Christos.

Made a note in the code about the -d and -d options brought up in
PR 18674. Thanks to Chris Jones for pointing it out.

correct prefixlen handling for IPv4.

Make prefixlen check to make sure its argument is present before deref
from openbsd

sweep of errx/warnx, remove unnecessary trailing \n

permit DNS names to address portion. they will be considered as RTF_HOST
if specified as destination (as there's no getnetbyname in IPv6).

if we're unsure about prefixlen, don't say "128".

more correct ion to RTF_HOST handling.

do not have -prefixlen 128 on RTF_HOST.
From: Dawid Szymanski <dawszy@arhea.net>

branches: 1.54.2;
raise RTF_HOST if destination is a host (/128). from ume

PR/15672: Love: route monitor missing info

Make route exit with a non-zero error code when operations to the routing
socket fails. eg, running route add 1.2.3.4 5.6.7.8 as non root.

fix -Wshadow warnings

Provide a short output format for the get command that only prints the
address corresponding to RTA_GATEWAY, or nothing if it doesn't exist.
Modify the exit value of route depending on this, so that one can do
stuff like:

#!/bin/sh
gw=`route -sn get default 2>/dev/null`
if [ -z "$gw" ]; then
echo no default route
exit 1
fi
ping -w1 -c1 $gw >/dev/null 2>&1
if ! route -sn get $gw >/dev/null; then
echo default gateway not responding
exit 1
fi
echo default gateway is at $gw

Print a R on reject routes to indicate that they're not normal routes,
and check netmask (or prefix) length as well as the destination
address when determining if a route is a "default" or not. This means
that the output from 'route show' will no longer say:

Internet6:
Destination Gateway Flags
default localhost UG
default localhost UG

but instead

Internet6:
Destination Gateway Flags
::/104 localhost UGR
::/96 localhost UGR

which makes much more sense.

off-by-one error in -prefixlen arg validation (do not permit 129 on IPv6)

make prefixlen work on IPv4 too

The way pmsg_addrs() is used requires that it always prints a newline and
flushes stdout before returning.

Previously the decoded output for a message was stuck in stdio buffers until
the next arrived.

convert to use getprogname()

remove nested externs.

support -cloned command line flag.
i don't think anyone ever going to use it.

mark cloned routes with RTF_CLONED. present it with netstat -r by "c".

let static routes overwrite cloned routes, as cloned routes can come back again
if necessary. behavior same as freebsd/bsdi, code partially from bsdi42.
(NRL rt->rt_parent was not added)
should fix PR 11916 and maybe some other PRs with ARP behavior.

recompilation of usr.sbin/route6d is suggested.

support RTF_BLACKHOLE.

XXX route show is total duplicate of netstat -r, we need to either remove
route show, or share the same source code, otherwise maintenance cost
bites (and is biting) us

remove use of uninitialized route (struct ortentry).
it seems that the bug was born during 4.3reno -> 4.4Lite1 transition.

on "route monitor", print RTM_IFANNOUNCE, RTM_IFINFO (new #), and others.

branches: 1.38.4;
Allow INET6 to be used with SMALL without undefined symbols

use scoped notation for link-local multicast too.

support extended scoped address notation (fe80::1%de0).
it hides implementation detail (embedded ifid) from the users.

don't make "-prefixlen 64" a default in -inet6 case. the past behavior
was not compatible with past practice ("route get" did not work right).

sync with recent KAME.

Display local address associated with route (if any) in "route get".

branches: 1.33.2;
fix alignment issue in routing socket data, on 64bit arch (i.e. alpha).
sockaddrs are packed with ROUNDUP() in sys/net/rtsock.c so userland
programs needs to follow that.

PR: 8179

Incorrect integration. ctime takes a time_t value, not a long.

IPv6-ready route command (-inet6 and -prefixlen are added)

branches: 1.30.2;
* implement -f; flush all routes before executing command
* reset af to 0 in flushroutes() and newroute()
* cleanup -DSMALL:
- `flush' and AF_APPLETALK is supported (the latter had partial support)
- AF_NS, AF_ISO, and AF_CCITT are not supported at all (rather than being
- partially unsupported).

const poisoning.

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

64-bit sweep, "route -v show" was spinning forever.

Use {warn,err}{,x}(3) instead of printf's

Don't declare optind.

resolve conflicts from lite-2 import

* cleanup for WARNS=1
* use .Nm correctly
* deprecate register
* getopt returns -1 not EOF

Fix uninitialized pointer botch when attempting to flush an already
empty routing table. From Matthias Scheler <tron@lyssa.owl.de>, in
PR #4074.

Don't SEGV if a qualifier is passed without an argument. From
Geoff Wing <mason@primenet.com.au>, fixes PR #2700.

KNF. be safe with buffers.

- Add netatalk bits
- Add prototypes
- Fix printf formats

Remove uid checks since the kernel now handles this and route(1) is no
longer setuid.
Document the "Permission denied" msg in the man page since this is
_slightly_ less clear than route(1)'s old error msg.

As discussed some time ago, add a "show" sub-command.
Also allow compilation with -DSMALL to ommit support for
non-essential protocols (i.e. when built into a ramdisk).
Use a less hack-ish way to generate keywords.[ch] and just
check in the result (helps ../../distrib/utils/x_route).

check returns of inet_addr() and inet_network() against INADDR_NONE,
not -1, per the manual page. on 64-bit systems like the alpha, -1 will
never equal (unsigned long)INADDR_NONE.

Produce less output on -q as promised by the man page (John Hawkinson; PR#998).

Interface-specific routes can be static.

Implement the document `-llinfo' flag.

Remove extra `break;'.

convert to new RCS Id conventions; reduce my headache

kill some extraneous equal signs.

Remove some more uses of obsolete functions.

Eliminate uses of some obsolete functions.

Clean up import.

Add RCS identifiers.

Compile X.25, OSI, and NS code.

Cleanup for GCC 2.

changed "Id" to "Header" for rcsids

added rcs ids to all files

branches: 1.1.1;
Initial revision

route: Enable receive error reporting on the routing socket

Avoid undefined behavior

route.c:1523:20, left shift of 1073741824 by 1 places cannot be represented
in type 'int'

route: address flags are more useful than interface flags for address msgs

route: Decode the correct flags for interface address messages

route(8): Show addres for RTM_LOSING

While here, add /* FALLTHROUGH */.

route(8): show addrs for RTM_MISS

route(8): print RTM_CHANGE messages

XXX Pullup -9

branches: 1.160.2;
Resolve every route(4) message type before printing anything more about it
other than the type and length.

This solves the issue where RTM_ONEWADDR is received and it tries to
parse route information from it.

branches: 1.159.2;
Handle the routing socket overflowing gracefully.

branches: 1.158.2;
Use the definitions from the standard header files and replace homebrew
snprintb with the libutil one.

Fix typo in flag name. We should probably just use IFFBITS string
that <net/if.h> defines.

PR bin/52815

Support -c <count> option for route monitor

route command exits if it receives <count> routing messages where
<count> is a value specified by -c.

The option is useful to get only particular message(s) in a test script.

branches: 1.155.4;
Report the PID for the process changing the address.

branches: 1.154.2; 1.154.4;
Separate nexthop caches from the routing table

By this change, nexthop caches (IP-MAC address pair) are not stored
in the routing table anymore. Instead nexthop caches are stored in
each network interface; we already have lltable/llentry data structure
for this purpose. This change also obsoletes the concept of cloning/cloned
routes. Cloned routes no longer exist while cloning routes still exist
with renamed to connected routes.

Noticeable changes are:
- Nexthop caches aren't listed in route show/netstat -r
- sysctl(NET_RT_DUMP) doesn't return them
- If RTF_LLDATA is specified, it returns nexthop caches
- Several definitions of routing flags and messages are removed
- RTF_CLONING, RTF_XRESOLVE, RTF_LLINFO, RTF_CLONED and RTM_RESOLVE
- RTF_CONNECTED is added
- It has the same value of RTF_CLONING for backward compatibility
- route's -xresolve, -[no]cloned and -llinfo options are removed
- -[no]cloning remains because it seems there are users
- -[no]connected is introduced and recommended
to be used instead of -[no]cloning
- route show/netstat -r drops some flags
- 'L' and 'c' are not seen anymore
- 'C' now indicates a connected route
- Gateway value of a route of an interface address is now not
a L2 address but "link#N" like a connected (cloning) route
- Proxy ARP: "arp -s ... pub" doesn't create a route

You can know details of behavior changes by seeing diffs under tests/.

Proposed on tech-net and tech-kern:
http://mail-index.netbsd.org/tech-net/2016/03/11/msg005701.html

no need for <sys/mbuf.h>

PR/50671: David Binderman: fix memory leak.

Add RTF_BROADCAST to mark routes used for the broadcast address when
they are created on the fly. This makes it clear what the route is for
and allows an optimisation in ip_output() by avoiding a call to
in_broadcast() because most of the time we do talk to a host.
It also avoids a needless allocation for the storage of llinfo_arp and
thus vanishes from arp(8) - it showed as incomplete anyway so this
is a nice side effect.

Guard against this and routes marked with RTF_BLACKHOLE in
ip_fastforward().
While here, guard against routes marked with RTF_BLACKHOLE in
ip6_fastforward().
RTF_BROADCAST is IPv4 only, so don't bother checking that here.

Teach route(8) about RTF_LOCAL.

MKINET6=no build fixes

http://mail-index.netbsd.org/tech-net/2014/12/13/msg004898.html

Note that RTM_CHANGE can also change the gateway

PR/47704: Takahiro HAYASHI: Add -L flag

Now that the code can support tags and verbose route printing, enable them.
Also document previously undocumented flags.

Factor out the netstat route printing code and use it here. There is no
point in having 2 different copies; fixes PR/49371

branches: 1.144.4;
use correct function

use symbolic flags

use scopeid functions

avoid pointer gymnastics

Retire OSI network stack. OK core@

Fix fallout from 1.129 that converted sou::so_foo from unions to
pointers but missed (char *)&soup->so_foo => (char *)soup->so_foo in
mask_addr(). It worked by luck - unless it didn't: due to pointer
numerology on amd64 route add -net ClassC without explicit /24 prefix
length specification would result into /16 destination instead of /24.

branches: 1.138.2;
remove useless rump headers.

PR/42179: Christoph Badura: Be a little friendlier about missing args.

branches: 1.136.2;
Off by one and clarity improvement from Christian Biere

PR/45661: Henning Petersen: Overlapping buffer in route.c.

Move RTF_ANNOUNCE flag so that it no longer conflicts with RTF_PROTO2.

RTF_ANNOUNCE was defined as RTF_PROTO2. The flag is used to indicated
that host should act as a proxy for a link level arp or ndp request.
(If RTF_PROTO2 is used as an experimental flag (as advertised),
various problems can occur.)

This commit provides a first-class definition with its own bit for
RTF_ANNOUNCE, removes the old aliasing definitions, and adds support
for the new RTF_ANNOUNCE flag to netstat(8) and route(8).,

Also, remove unused RTF_ flags that collide with RTF_PROTO1:
netinet/icmp6.h defined RTF_PROBEMTU as RTF_PROTO1
netinet/if_inarp.h defined RTF_USETRAILERS as RTF_PROTO1
(Neither of these flags are used anywhere. Both have been removed
to reduce chances of collision with RTF_PROTO1.)

Figuring this out and the diff are the work of Beverly Schwartz of
BBN.

(Passed release build, boot in VM, with no apparently related atf
failures.)

Approved for Public Release, Distribution Unlimited
This material is based upon work supported by the Defense Advanced
Research Projects Agency and Space and Naval Warfare Systems Center,
Pacific, under Contract No. N66001-09-C-2073.

branches: 1.133.2;
Simplify.

Use __dead

Fix memset usage.

don't compile tag functions if SMALL is defined. Should fix the build

Alloc dynamically sockunions in sou.
Teach route(8) about multiple tags. E.G.: -tag 100,20,33
XXX: needs documentation

Add a new AF/PF_ROUTE which is 64-bit clean which makes the routing socket
interface (and its associated sysctls) act identically for both 32 and 64 bit
programs. The old unclean one remains for backward compatibility.

branches: 1.127.2;
Convert from RUMP_ACTION to RUMPPRG.

Support the new RTM_IFINFO and RTM_CHGADDR messages.

adapt to new rumpclient_init() signature

#ifdef variable correctly

be consistent in what to leave outside of SMALL

Support RUMP_ACTION to compile a rump kernel client.

add MPLS clue - mpls and tag keywords

PR/40455: Mihai Chelaru: Add noreject noblackhole options

parse unsigned long with strtoul not atoi

branches: 1.118.6;
Get the socket addresses for dst/gateway/netmask/interface/ifaddr
out of the global namespace.

Don't cast a sockunion * to sockaddr *. Instead, reference the
sockaddr inside of the sockunion.

Delete superfluous casts to void *.

Constify.

Constify.

Constify.

For clarity's sake, do not assign a new value to inet_makenetandmask()'s
argument, `inet'.

Extract subroutine extract_addrs() for reuse, later.

Replace the anonymous constant, 0, with AF_UNSPEC.

Constify.

Shorten a staircase. Remove superfluous parentheses from return
statements. Constify.

Use a union and sizeof() to avoid a cast and an anonymous constant
buffer length, respectively.

Use bool.

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

branches: 1.106.12; 1.106.16;
TSEL() is const char* now

Fix a bug that caused 0.0.0.0/0 to be treated like 0.0.0.0/32.

foo << 32 is undefined per section 6.5.7 of the C99 standard
if foo is (or is promoted to) a 32-bit type.

Let us qualify 'route flush' and 'route show' commands with -llinfo
and -host flags, which is useful for displaying/flushing ARP entries:

# route -n show -inet -llinfo -host
Routing table

Internet:
Destination Gateway Flags
169.254.1.119 link#1 UH
169.254.230.110 00:02:6f:21:e6:6e UH
# route -n flush -inet -llinfo -host
169.254.1.119 done
169.254.230.110 0.2.6f.21.e6.6e done
169.254.237.70 done
# route -n show -inet -llinfo -host
Routing table
#

This will help me address bin/11079, "dhclient may require arp and sed".

Extract common code from flushroutes() and show(), creating
parse_show_opts().

While I'm here, make small cosmetic changes to flushroutes().

Move the #ifndef SMALL code together.

Use the symbol AF_UNSPEC instead of 0 as appropriate.

Remove dead code.

Remove XNS and CCITT/X25 bits.

KNF err(3), errx(3) usage, indentation, braces.

Add route keywords -nocloned, -nocloning.

Use default option order (AaBbCc...).

Print the INET netmask if it is non-contiguous.

PR/32632: Yves-Emmanuel JUTARD: Improvment suggestion in 'route' display.
Add a new -S flag that prints a space for missing flags.

PR/32629: Yves-Emmanuel JUTARD: Type errors in sbin/route/route.c, another
u_char/char conflict.

Add full IPv6 syntax support when route is built with -DSMALL and -DINET6.
Fixes PR/23937. Adapted from supplied patch.

Clean up some minor nits:
+ style (KNF).
. don't cast NULL
. white space clean up
+ use consistent format for error messages.
+ plug memory leaks and avoid malloc(0).
+ verbose message should display even when routing table is empty
(flushroutes).

Make sure that a netmask socket address value prints appropriately when the
family is unset. Either the destination or interface address socket address
family value is used. This change prevents the following sort of output:

RTM_NEWADDR: address being added to iface: len 60, metric 0, flags:<UP>
sockaddrs: <NETMASK,IFA,BRD>
(0) 00.00.ff.00.00 127.0.0.1 127.0.0.1

With this change the last line becomes:

255.0.0.0 127.0.0.1 127.0.0.1

Also make sure that when using a generated inet netmask the address family
is set. With this change a generated netmask will print appropriately
when using the verbose (-v) option.

Make sure the get command always exits non-zero on error. This makes things
like the following work as expected.

#!/bin/sh

if ! route -sn get default >/dev/null 2>&1; then
echo default gateway not set
exit 1
fi
echo default gateway set

Handle routing socket write(2) errors when they occur. This produces
better diagnostics by allowing for handling of the special route errno
values ESRCH, EBUSY, and ENOBUFS even with the quiet (-q) option or
when doing a get command.

Trim the local domain name in routename() for INET6 lookups; similar to
INET lookups.

Make this build for __GNUC__ <= 2 (i.e. for vax). The cast to
caddr_t removes a const qualifier. Instead, cast the first element
of the subtraction to "struct sockaddr *" to work around the problem.

Accept protocol flags, i.e. -inet, -inet6, with the show command.
[From OpenBSD via DragonFly BSD]

Fix the default case in sodump(). Printing the address of the "which"
string isn't really that useful. Instead print the string and the value
of sa_family.

More style nits (KNF); casts and sizeof's are not followed by a space.

switch is not a function, so add a space; sort switch elements
alphabetically (KNF).

* Exit non-zero error code when flushroute() fails operating on the
routing socket, e.g. running route flush as non-root.

* When quiet (-q) don't append error string to a non-existent
diagnostic output (stdout). The error diagnostic is still written
to stderr. [from FreeBSD]

* Use warn(3) in place of perror(3).

const poisoning.

Teach route(8) to print the messages generated by net80211.

Convert prefixlen() to ANSI function declaration. Missed by earlier
conversion.

Due to a limitation in inet_addr(3) it was not possible to specify a
netmask value of 0xffffffff (INADDR_NONE). Fix by using inet_aton(3).
Note it was possible to specify the same mask as the dotted quad
255.255.255.255 because of a later call to gethostbyname(3).
[From FreeBSD]

Fix botched prior commit. Additional debugging error messages mistakenly
committed.

* Honor -net when using CIDR (net/mask) notation. This differs
from the default CIDR (host/mask) interpretation. As discussed on
tech-net.

* Use errx() in place of fprintf() and exit(). Make the error message
formats consistent.

Fix macro NEXTADDR so that the names, like so_dst, are printed
rather than u. The preprocessor doesn't do replacement within
strings, so the # string creation operator must be used.

Kill __P(), use ANSI function declarations.

NI_WITHSCOPEID was not picked up by IETF standardization process.

bump buffer size for any_ntoa(), for huge sa_len

2-byte read overrun for unknown sockaddr. from openbsd
XXX out[64], is it safe? what if sa_len > 23?

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

print prefixlen on -show

Remove a comment that referred to a non-committed part of the patch I
submitted for bin/17613. Ok'd by mrg@.

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

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

This is an updated submitted patch originally written by Jim Rees
and sent in by Greg Hudson as seen in PR misc/3227. Basically what it
does is adds a flushall option which deletes all but localhost routes.
This is done by andoring in a flag called doall (1 means do all routes
including gateway, 0 means do a regular flush). I have seen some
platforms that do this. I tested it out on ipv4 only, it works as
advertised. Commit was approved by christos@.

PR/21856: Robert Elz: Changes to route(8) to allow /n instead of -prefixlen
for INET6

permit foo/bar notation, like "route add 10.0.0.0/8 127.0.0.1".
originally by provos

strlcpy [a few mistakes found]

This addresses PR 17613 when certain methods of adding routes would
muck up the routing tables. The patch was submitted by Quentin Garnier
<netbsd@quatriemek.com> and tweaked a little after it was reviewed
by Christos Zoulas <christos@netbsd.org>. Final change/commit approved
by Christos.

Made a note in the code about the -d and -d options brought up in
PR 18674. Thanks to Chris Jones for pointing it out.

correct prefixlen handling for IPv4.

Make prefixlen check to make sure its argument is present before deref
from openbsd