CVS log for kame/kame/sys/net/if_gif.c

ANSIfy, to reduce diffs with *BSDs.
based on diff from dunceor at gmail.com.
XXX not tested yet, sorry, we need to ressurect buildlab configs.

random thoughts:
- IGMPv3 support needs a serious cleanup.  argument passing is horrible, style
  does not meet style(9) at all.  i'm afraid there could be a lot of bugs.
- DCCP/SCTP/NAT-PT need someone to maintain them, otherwise we should drop
  them at once.
- we should pick a platform (one of *BSD) and try to get KAME tree in sync
  with more recent version of *BSD.  otherwise, KAME tree gets more and more
  out-of-date.

moved nemo specific ioctl macro to if_nemo.h

perform sa6_recoverscope() for SIOCGIFPSRCADDR_IN6 and SIOCGIFPDSTADDR_IN6

revised scope-related interfaces, clarifying the relationship between
the kernel-internal form and the public sockaddr_in6 form.

remove freebsd[234]-specific code, since they are no longer supported in KAME-SNAP

New Mobile IPv6 code which also supports NEMO.

makes it compilable on other BSDs than freebsd5

sync with freebsd53 and makes it compilable
(their behavior is not confirmed yet...)

removed a "malloc() failure" warning message at start-up time on freebsd5.

(M_WAIT(mbuf flag) is equal to M_WAITOK(malloc() flag) in every BSD
 except freebsd5, so this problem occurs only in freebsd5)

sync w/ 3.5.
altq, pf, and carp are disabled for now.
use common altq (kame/sys/altq) instead of openbsd/sys/altq. >kjc

more #ifdef cleanups

goodbye bsdi

make it compilable on freebsd521+kame

ToDo: reflect changes in freebsd521, test, ...

snprintf

as a safety measure, clear m_flags before bpf_mtap hack

stf and gif works fine on freebsd5

KAME-origined logical-interfaces (gif, dummy, and stf for the time being)
are now recognized on freebsd5.

makes it compilable on freebsd5

makes it compilable on freebsd5, too

loop detection using packet tag (from openbsd-current)

fixed a bug that gif interface cannot be manipulated under OpenBSD+KAME due to a lack of AF_LINK address

make USE_ENCAPCHECK (in netinet*/*gif.c) to global option, GIF_ENCAPCHECK.
#ifdef out unneeded code when possible.
From: Krister Walfridsson <cato@df.lth.se>

remove freebsd4 cloning interface support.

typo if_dlt (sorry)

initialize if_dlt and if_sadl on netbsd 1.6

sync w/ 1.6.  GENERIC.KAME/i386 compiles, not tested.
- NATPT and IP6FW support were dropped (you can use ipfilter).
- altq, ecn, RTF_CACHE, SCOPEDROUTING need checking

drop too short IPv6 frame

openbsd too uses splsoftnet

altq cleanup.
reduce differences from netbsd-current and openbsd-3.0:
 - use ALTQ_DECL() for altq only variables.
 - make IFQ_ENQUEUE() take 4 args even for the non-altq case
 - the return type of altq_etherclassify() is changed from
   int to void.
   the function prototype is moved to in if_altq.h.

added a header file for a function declaration.

in gif_set_tunnel() check validity of the scope zone ID of the addresses.

remove unused extern decls

-Wunused.  noted by buildlab

reduce diff with netbsd 1.5Y.  backout variable/member renames done by freebsd
integration.  interface cloning code was originated in netbsd (1.5[A-X]) so
it would be easier to follow netbsd.

to reduce the amount of #ifdef use "gifs" tailq link on all OSes.

never break paren match with #ifdef.  it will be pain in the ass when we need
to hunt these bugs down.

I had broken gif when upgrading fbsd43->44.
fix it.

make use of gifattach0() on freebsd4.  re-enable ALTQ support.

don't softintr_disestablish twice.

sync with freebsd44.

remove all ipv4 varargs functions for freebsd4.
careful test is needed.

reduce "#ifdef ALTQ" now that bsdi3/4 have ALTQ-compat queue macros.

complete ALTQ support in gif.

don't IF_ENQUE if IF_QFULL.  for non-ALTQ.  not tested.

made it compile on bsdi4:
- there's no IFQ_CLASSIFY
- IF_ENQUEUE takes only 2 args
- pktattr is defined only when ALTQ
(I'm not sure if this is really correct.  please check > itojun?)

after gif_set_tunnel(), psrc/pdst may be null.  set IFF_RUNNING accordingly.

don't softintr_disestablish on success.

netbsd-current: use generic soft interrupt on output processing.
altq classifier for outbound.
tabify and other random cleanups.

NULL != 0

s/gifp/ifp/

last arg of in6?_gif_output() is not used any more.

set IFF_UP on SIOCSIFADDR.  be consistent with others.

make EON encap/decap logic to follow 4.4BSD arg passing rules.

set if_addrlen explicitly (just in case)

multi destination mode is long gone.
expose gif_softc for sec*.

share attach logic as much as possible.
add refcnt field for if_sec.c, as we will share sec* from multiple SPDs
(we can have only one interface per an outer address pair)

missing break statement

add gif_set_tunnel(), so that we can set tunnel address from within the
kernel at ease.

have a function to do SIOCDIFPHYADDR.  sync with netbsd-current

encap_attach/detach dynamically on ioctls

move encap_attach() to dedicated function in in*_gif.c

make really sure we don't SEGV on invalid ioctl args.

comment fix

2 space after sentence

remove printf() in output path

sync interpretation of if_lastchange with RFC1573

(netbsd) change meaning of ifnet.if_lastchange, to meet with RFC1573.
sync with netbsd-current

BSD/OS and FreeBSD do not update if_lastchange when processing packets
for SNMP requirements.

do not touch "m" after IF_ENQUEUE.

sync with openbsd 2.9.  GENERIC.KAME compiles, but not tested.
ALTQ portion needs a serious checking.

raise spl level when calling if_up().  freebsd 1.8 -> 1.9

on attach, turn off ingress filter.  as discussed on core@kame.net.
the default value for merged *BSD repository needs to be picked under
consensus in *BSD community, not us.

bsd mbuf management rule in netiso processing - child function frees mbuf
on error.

minor cleanups (missing break)

no need for global *ehdr.

more comment on output dispatch

netbsd: ISO over IPv4/v6 on gif*, using EON encapsulation.  from chopps

explicitly use u_int32_t for DLT_NULL encapsulation.

clarify comment

fix bpf address family.  from chopps

set IFF_RUNNING upon assigning a pair of physical addresses
in order to perform DAD on gif.

register variable declaration, they are getting more and more
meaningless these days

correct uninitialized variable

validate parameter to SIOCSIFPHY* early.  provide SIOC[SG]LIFPHY*.

don't validate AF on SIOCGIFP*ADDR

do not require matches between gif phys addr and ioctl cmd, on
SIOCGIFP*ADDR.  gifconfig(8) relied upon the behavior.

typo

clarify bcopy size validation more

make more strict validation on SIOC[GS]IFPHY*

be more friendly with non-ipv4 build.
fix ipsecstat/ipsec6stat mixup.

sync w/ snap;
took into account differences of arguments for SIOCSIFPHYADDR and
SIOCSIFPHYADDR_IN6 to avoid invalidating commands mistakenly.

took into account differences of arguments for SIOCSIFPHYADDR and
SIOCSIFPHYADDR_IN6 to avoid invalidating commands mistakenly.

codeshare for gif.  enable stf on openbsd.

#define XBONEHACK will allow multiple gif to be configured with
the same pair of outer addresses.
reject multiple multi-destination tunnels.

always overwrite rcvif on input.  From: Greg Troxel <gdt@fnord.ir.bbn.com>

INET6 define is moved from opt_inet.h to opt_inet6.h in FreeBSD3,
I believe this helps to share codes with FreeBSD4 easily.

freebsd4 requires driver to set ifq_maxlen

NBPF -> NBPFILTER (gratuitous difference in freebsd4)

remove GIFF_INUSE (not used for years)

make gif internal variable invisible from outside.

make *-current merge easier

simplify gif attachment, by using encap_attach_func().
we no longer need to detach-reattach on gifconfig.

added support of SIOCDIFPHYADDR to enable "gifconfig delete"

nuke spaces end of line.

add $KAME$

fix bug when multiple gif interfaces are configured.
(

re-activate gif IFF_LINK0 (multi destination mode).

revise gif input logic.

- split outer address match logic as encap[46]_input, so that other guys
  (like mobile-ip6, 6to4, 6over4, whatever) can match packets based on src/dst.
  see comment on ip_encap.c for detail - I hate tunnelling.  really.
- allow raw ip{,6} output for proto #4 and #41.
TODO: ordering of packet matching in encap[46] must be prefix length-based.
	(we have src/dst prefix length, so it is not trivial to define total
	order among entries)
TODO: change ipip_input() to conform to this framework?
CAVEAT: IFF_LINK0 for gif is not working with this change, i'll try to cleanup
	afterwards.

changes for other OSes will be committed shortly.

check for 0.0.0.0 or :: duplicate, only after checking outer AF.
(previous code did not check AF, which may result in strange behavior)

on second thought IFF_RUNNING is not necessary.
we have no resource allocated for interface operation.

#define IFF_RUNNING     0x40            /* resources allocated */

bring IFF_RUNNING down, if interface becomes !IFF_UP.
noted by: deraadt@openbsd

- changed the (default) upper limit of how many times gif_output can be called
  from 9 to 1, since some systems do not allow so many recursive calls in
  kernel...

do not call in6_ifattach() explicitly, it is not necessary.
it will automatically be called from in6_if_up() in net/if.c.

OpenBSD portability fixes (not complete).
- OpenBSD needs struct proc on xx_control(), xx_ioctl() and other places.
- do not include if.h twice in in6.c.
- OpenBSD has SIOCSIFMTU defined but does not support it.

-added more mapped addr related defs
-include malloc.h for FreeBSD3

clearify prototype for if_ioctl() and xxattach().

sync with today's KAME/FreeBSD228 kernel.
GENERIC.v6 compiles fine but not tested.
need more synchronization.

dropped items: MAPPED_ADDR_ENABLED RADISH
need integration: PM ALTQ

bring in portability #ifdef from unified-ipv6.
not really tested and incomplete, incomplete but compiles fine on NetBSD.

Initial revision

KAME/NetBSD 1.4, as of 1999/7/30.
for unified-KAME code.
no cross-OS portability is introduced.
works fine on NetBSD as no change from KAME/NetBSD.