]> The Tcpdump Group git mirrors - tcpdump/log
tcpdump
15 years agonlpid.h/c: mark several structs as const
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 07:48:23 +0000 (08:48 +0100)] 
nlpid.h/c: mark several structs as const

15 years agol2vpn.h/c: mark several structs as const
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 07:47:15 +0000 (08:47 +0100)] 
l2vpn.h/c: mark several structs as const

15 years agoipproto.h/c: mark several structs as const
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 07:45:56 +0000 (08:45 +0100)] 
ipproto.h/c: mark several structs as const

15 years agogmpls.h/c: mark several structs as const
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 07:44:31 +0000 (08:44 +0100)] 
gmpls.h/c: mark several structs as const

15 years agoPrecompute CRC-10 table.
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 07:40:44 +0000 (08:40 +0100)] 
Precompute CRC-10 table.

We now precompute the CRC-10 table allowing us to mark it as const. The table was precomputed using a Python snippet (included as comment)

15 years agoaf.h/c: make several structs const
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 06:44:56 +0000 (07:44 +0100)] 
af.h/c: make several structs const

15 years agoaddrtoname.c: make several structs static
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 06:31:43 +0000 (07:31 +0100)] 
addrtoname.c: make several structs static

15 years agoaddrtoname.c: constify several structs
Kovarththanan Rajaratnam [Sat, 20 Feb 2010 06:32:38 +0000 (07:32 +0100)] 
addrtoname.c: constify several structs

15 years agoRecognize -PRE-GIT as well as alpha/beta; clean up regexps a bit.
Guy Harris [Tue, 9 Feb 2010 00:28:42 +0000 (16:28 -0800)] 
Recognize -PRE-GIT as well as alpha/beta; clean up regexps a bit.

15 years agoGet rid of extra blank line.
Guy Harris [Mon, 8 Feb 2010 10:42:48 +0000 (02:42 -0800)] 
Get rid of extra blank line.

15 years agoWe're using Git, not CVS, now.
Guy Harris [Mon, 8 Feb 2010 10:35:30 +0000 (02:35 -0800)] 
We're using Git, not CVS, now.

15 years agoRecognize -b as a flag.
Guy Harris [Mon, 8 Feb 2010 10:33:12 +0000 (02:33 -0800)] 
Recognize -b as a flag.

15 years agoMention the b flag in the usage message.
Guy Harris [Mon, 8 Feb 2010 10:24:10 +0000 (02:24 -0800)] 
Mention the b flag in the usage message.

15 years agoDocument the -b flag.
Guy Harris [Mon, 8 Feb 2010 10:22:10 +0000 (02:22 -0800)] 
Document the -b flag.

15 years agoUse rlen rather than tll as the name of "remaining length" variables.
Guy Harris [Mon, 8 Feb 2010 02:02:41 +0000 (18:02 -0800)] 
Use rlen rather than tll as the name of "remaining length" variables.

15 years agoCheck the containing item length in some loops.
Guy Harris [Mon, 8 Feb 2010 01:47:31 +0000 (17:47 -0800)] 
Check the containing item length in some loops.

In some loops, don't loop just until we get an error, stop when we run
out of data to parse.

Also, add some comments to indicate what we know about lengths at
various points, before we do something with the length that happens to
assume what we know.  Add some checks that this auditing found
necessary.

15 years agoUse #defines rather than sizeof.
Guy Harris [Mon, 8 Feb 2010 00:05:46 +0000 (16:05 -0800)] 
Use #defines rather than sizeof.

Use ForCES_HDRL, TLV_HDRL, and ILV_HDRL instead of various sizeof's, to
make it clearer what certain tests ensure are true (as, in other cases,
the #defines are subtracted from values that should be, at that point,
ensured not to be less than the #define in question).

15 years agoRFC 5006 support.
David Horn [Sun, 7 Feb 2010 22:52:07 +0000 (14:52 -0800)] 
RFC 5006 support.

Reviewed-By: Guy Harris <[email protected]>
15 years agoPrint bit numbers in decimal, not hex - it's not the presence bit field,
Guy Harris [Sun, 7 Feb 2010 03:21:30 +0000 (19:21 -0800)] 
Print bit numbers in decimal, not hex - it's not the presence bit field,
its the index of a bit from that field.

15 years agoConstify a bunch of stuff.
Guy Harris [Wed, 27 Jan 2010 04:14:23 +0000 (20:14 -0800)] 
Constify a bunch of stuff.

15 years agoPut PRIu64 on the same line as the rest of the format string.
Guy Harris [Wed, 27 Jan 2010 02:50:53 +0000 (18:50 -0800)] 
Put PRIu64 on the same line as the rest of the format string.

15 years agoGet rid of some no-longer-necessary casts.
Guy Harris [Wed, 27 Jan 2010 02:45:58 +0000 (18:45 -0800)] 
Get rid of some no-longer-necessary casts.

15 years agoAdd TCHECKs, make length variables 32 bits.
Guy Harris [Wed, 27 Jan 2010 02:43:22 +0000 (18:43 -0800)] 
Add TCHECKs, make length variables 32 bits.

Add TCHECK/TCHECK2 calls to make sure we are within the packet boundary
before fetching data.  Make some length variables that could in theory
have values that don't fit in 16 bits 32 bits long.

Get rid of blanks before newlines.

15 years agoUse EXTRACT_16BITS and EXTRACT_32BITS to extract values from a packet.
Guy Harris [Wed, 27 Jan 2010 02:37:41 +0000 (18:37 -0800)] 
Use EXTRACT_16BITS and EXTRACT_32BITS to extract values from a packet.

There's no guarantee that those values are aligned on the appropriate
boundary, nor is there a guarantee that accessing misaligned values will
behave as desired.

15 years agoUse EXTRACT_16BITS to extract big-endian 16-bit quantities from the packet.
Guy Harris [Wed, 27 Jan 2010 02:35:53 +0000 (18:35 -0800)] 
Use EXTRACT_16BITS to extract big-endian 16-bit quantities from the packet.

There's no guarantee that those quantities will be properly aligned, nor
is there any guarantee that fetching an improperly-aligned quantity will
work as desired.

15 years agoUse EXTRACT_16BITS() and EXTRACT_32BITS() to fetch integral values.
Guy Harris [Wed, 27 Jan 2010 01:56:36 +0000 (17:56 -0800)] 
Use EXTRACT_16BITS() and EXTRACT_32BITS() to fetch integral values.

Use EXTRACT_16BITS() and EXTRACT_32BITS() to fetch 16-bit and 32-bit
big-endian quantities from the packet, as there's no guarantee that a
given 16-bit quantity will be aligned on a 2-byte boundary or that a
given 32-bit quantity will be aligned on a 4-byte boundary, nor is there
a guarantee that unaligned accesses will succeed (they might trap, or
they might not fetch an unaligned quantity).

15 years agoAdd some information about the symbolic references to TCP flags.
Randy Sofia [Tue, 26 Jan 2010 23:46:37 +0000 (15:46 -0800)] 
Add some information about the symbolic references to TCP flags.

Reviewed-by: Guy Harris <[email protected]>
15 years agoEnhance NFS access request messages.
Lucas C. Villa Real [Tue, 26 Jan 2010 20:09:24 +0000 (12:09 -0800)] 
Enhance NFS access request messages.

When debugging NFS operations one may find it easier to get the actual access
flags decoded rather than having to look up the NFS access flags to find which
permissions were requested by the client.

Reviewed-by: Guy Harris <[email protected]>
15 years agoHandle the 48-byte and 64-byte Linux USB headers differently; for now,
Guy Harris [Mon, 25 Jan 2010 23:25:33 +0000 (15:25 -0800)] 
Handle the 48-byte and 64-byte Linux USB headers differently; for now,
the difference isn't major, but it might be more important later.

15 years agoSquelch some int-vs-long errors.
Guy Harris [Mon, 11 Jan 2010 02:36:45 +0000 (18:36 -0800)] 
Squelch some int-vs-long errors.

15 years agoRemove more non-C89 features.
Guy Harris [Mon, 11 Jan 2010 02:23:11 +0000 (18:23 -0800)] 
Remove more non-C89 features.

No array initialization with [index] =, no // comments; some compilers
that might be used to compile tcpdump don't support them.

15 years agoGet rid of BIG_ENDIAN and LITTLE_ENDIAN definitions.
Guy Harris [Mon, 11 Jan 2010 02:00:08 +0000 (18:00 -0800)] 
Get rid of BIG_ENDIAN and LITTLE_ENDIAN definitions.

The configure script doesn't define BYTE_ORDER, so they can't validly be
compared against BYTE_ORDER; don't leave them around for people to think
they can.

15 years agoDon't use bitfields in on-the-wire structures.
Guy Harris [Mon, 11 Jan 2010 01:56:49 +0000 (17:56 -0800)] 
Don't use bitfields in on-the-wire structures.

C doesn't guarantee that bit fields will be in any particular order.

15 years agoPut forces.h into the release tarball.
Guy Harris [Sun, 10 Jan 2010 22:44:35 +0000 (14:44 -0800)] 
Put forces.h into the release tarball.

15 years agoFix a bunch of warnings in the ForCES printer.
Guy Harris [Sun, 10 Jan 2010 22:16:41 +0000 (14:16 -0800)] 
Fix a bunch of warnings in the ForCES printer.

Get rid of C++/C99-style comments while we're at it; some C compilers
reject them.

15 years agoraw_if_print() isn't yet a new-style print routine.
Guy Harris [Sun, 10 Jan 2010 21:34:22 +0000 (13:34 -0800)] 
raw_if_print() isn't yet a new-style print routine.

15 years agoadded -vv/-vvv and -vvvv test output
Michael Richardson [Sun, 10 Jan 2010 19:59:18 +0000 (14:59 -0500)] 
added -vv/-vvv and -vvvv test output

15 years agoIETF FORCES protocol printer, from Jamal Hadi Salim <hadi at cyberus.ca.
Michael Richardson [Sun, 10 Jan 2010 19:49:46 +0000 (14:49 -0500)] 
IETF FORCES protocol printer, from Jamal Hadi Salim <hadi at cyberus.ca.

15 years agoThis commit makes ipnet_if_print the first NDO aware top-level
Michael Richardson [Sun, 10 Jan 2010 19:36:07 +0000 (14:36 -0500)] 
This commit makes ipnet_if_print the first NDO aware top-level
printer.
Merge commit 'origin/master'

Conflicts:
netdissect.h

15 years agoether_encap_print() is a utility for use by printers for protocols that
Guy Harris [Wed, 25 Nov 2009 21:26:52 +0000 (13:26 -0800)] 
ether_encap_print() is a utility for use by printers for protocols that
use Ethernet types, to call the appropriate printer routine for the
encapsulated type.  IPNET is the only protocol using IPNET types, so
ipnet_encap_print() isn't needed.

15 years agoLook for core dumps, and fail test if some are found
Michael Richardson [Sun, 10 Jan 2010 19:28:59 +0000 (14:28 -0500)] 
Look for core dumps, and fail test if some are found

15 years agotest cases for Solaris print-ipnet. Confirms ndo printer code works
Michael Richardson [Sun, 10 Jan 2010 19:28:33 +0000 (14:28 -0500)] 
test cases for Solaris print-ipnet. Confirms ndo printer code works

15 years agomake new list of printers that take ndo argument, and use it as appropriate
Michael Richardson [Sun, 10 Jan 2010 19:27:33 +0000 (14:27 -0500)] 
make new list of printers that take ndo argument, and use it as appropriate

15 years agoUpdate config.guess and config.sub to the latest versions.
Guy Harris [Fri, 8 Jan 2010 18:37:35 +0000 (10:37 -0800)] 
Update config.guess and config.sub to the latest versions.

15 years agoDon't warn about the non-use of some arguments to rsvp_obj_print() if we
Guy Harris [Wed, 6 Jan 2010 04:24:44 +0000 (20:24 -0800)] 
Don't warn about the non-use of some arguments to rsvp_obj_print() if we
don't have libcrypto, as we can't use them if we don't.

15 years agoInclude ipnet.h in the distribution.
Guy Harris [Wed, 6 Jan 2010 04:17:26 +0000 (20:17 -0800)] 
Include ipnet.h in the distribution.

15 years agoAdd support for DLT_IPV4 and DLT_IPV6.
Guy Harris [Mon, 4 Jan 2010 20:11:04 +0000 (12:11 -0800)] 
Add support for DLT_IPV4 and DLT_IPV6.

While we're at it, clean up some whitespace.

15 years agoWhen removing conftest*, use "-r", to clean up OS X .dSYM turds.
Guy Harris [Sat, 2 Jan 2010 23:59:43 +0000 (15:59 -0800)] 
When removing conftest*, use "-r", to clean up OS X .dSYM turds.

15 years agoFix some problems that show up in autoconf 2.64 and later.
Guy Harris [Sat, 2 Jan 2010 23:29:06 +0000 (15:29 -0800)] 
Fix some problems that show up in autoconf 2.64 and later.

15 years agoUpdate CHANGES for 4.0.2 release
Ken Bantoft [Tue, 22 Dec 2009 01:38:10 +0000 (20:38 -0500)] 
Update CHANGES for 4.0.2 release

15 years agoWhen printing a list of link-layer types for -L, give the name of the
Guy Harris [Tue, 1 Dec 2009 08:39:54 +0000 (00:39 -0800)] 
When printing a list of link-layer types for -L, give the name of the
device for which they're being reported, so if no -i flag was given
people don't think it's something such as the union of of all link-layer
types supported by all devices, and, if we can determine whether the
device supports monitor mode (with libpcap 1.0.0 and later), indicate
whether the link-layer types are the ones supported when not in monitor
mode (if -I wasn't specified) or when in monitor mode (if -I was
specified), so people don't wonder why some types aren't available (as
some are available only in monitor mode, and others are available only
when not in monitor mode, on some platforms).

15 years agoSquelch a couple of compiler warnings.
Guy Harris [Fri, 27 Nov 2009 20:03:25 +0000 (12:03 -0800)] 
Squelch a couple of compiler warnings.

15 years agoInclude "netdissect.h" in print-ip6.c, and move icmp6_print's
Guy Harris [Thu, 26 Nov 2009 23:07:47 +0000 (15:07 -0800)] 
Include "netdissect.h" in print-ip6.c, and move icmp6_print's
declaration out from inside #if 0/#endif, to get it declared in
print-ip6.c.

15 years agoXXX_if_print() routines are not passed a "struct netdissect_options *",
Guy Harris [Thu, 26 Nov 2009 22:57:38 +0000 (14:57 -0800)] 
XXX_if_print() routines are not passed a "struct netdissect_options *",
so ipnet_if_print() can't expect one.

ipnet_print() isn't called from outside print-ipnet.c, so make it static.

15 years agoether_encap_print() is a utility for use by printers for protocols that
Guy Harris [Wed, 25 Nov 2009 21:26:52 +0000 (13:26 -0800)] 
ether_encap_print() is a utility for use by printers for protocols that
use Ethernet types, to call the appropriate printer routine for the
encapsulated type.  IPNET is the only protocol using IPNET types, so
ipnet_encap_print() isn't needed.

15 years agonetdissect.h did not actually declare ipnet_if_print, make it visible
Michael Richardson [Wed, 25 Nov 2009 02:38:56 +0000 (21:38 -0500)] 
netdissect.h did not actually declare ipnet_if_print, make it visible

15 years agoupdated print-ipnet.c to use netdissect options structure
Michael Richardson [Wed, 25 Nov 2009 02:31:04 +0000 (21:31 -0500)] 
updated print-ipnet.c to use netdissect options structure

15 years agoThe patches attached to this email are required to get a fully working tcpdump
Darren Reed [Wed, 25 Nov 2009 02:20:25 +0000 (21:20 -0500)] 
The patches attached to this email are required to get a fully working tcpdump
on OpenSolaris, or Solaris Express Community Edition, build 125 and later.

The attached patch introduces support for printing out the IPNET headers used
for packet capture inside of zones that share their networking with the global
zone and for packets "transmitted" between zones.

tcpdump 4.0.0 will ship with builds 129 and later of OpenSolaris/SXCE and
when run as root with the '-L' option, should behave as below to indicate that
the system is fully functional:

Data link types (use option -y to set):
 DOCSIS (DOCSIS) (printing not supported)
 IPNET (Solaris IPNET)
 EN10MB (Ethernet)

Cheers,
Darren

15 years agoRemove print-X.out, as it collides with print-x.out on systems with
Guy Harris [Tue, 17 Nov 2009 00:36:37 +0000 (16:36 -0800)] 
Remove print-X.out, as it collides with print-x.out on systems with
case-insensitive file systems; print-capX.out has the same contents as
the most recent version of print-X.out, and doesn't collide with
print-x.out.

15 years agoPropagate
Guy Harris [Tue, 17 Nov 2009 00:35:24 +0000 (16:35 -0800)] 
Propagate

commit 0a029bab087e2b66ba35aa144f8993cb26056f5d
Author: Michael Richardson <[email protected]>
Date:   Sat Nov 14 20:51:44 2009 -0500

    updated print-flags test case to have new TCP flags

to print-capXX.out, prior to removing print-XX.out (again - see previous
commits; it collides with print-xx.out on systems with case-insensitive
file systems).

15 years agoprint-XX.out collides with print-xx.out on systems with case-insensitive
Guy Harris [Tue, 17 Nov 2009 00:34:14 +0000 (16:34 -0800)] 
print-XX.out collides with print-xx.out on systems with case-insensitive
file systems; we've propagated the latest version of print-XX.out to
print-capXX.out, so get rid of print-XX.out.

15 years agoPropagate
Guy Harris [Tue, 17 Nov 2009 00:32:21 +0000 (16:32 -0800)] 
Propagate

commit 0a029bab087e2b66ba35aa144f8993cb26056f5d
Author: Michael Richardson <[email protected]>
Date:   Sat Nov 14 20:51:44 2009 -0500

    updated print-flags test case to have new TCP flags

to print-capXX.out, prior to removing print-XX.out (again - see previous
commits; it collides with print-xx.out on systems with case-insensitive
file systems).

15 years agoignore lmp.new file
Michael Richardson [Sun, 15 Nov 2009 03:17:25 +0000 (22:17 -0500)] 
ignore lmp.new file

15 years agoinitial printing of RPL DIO messages
Michael Richardson [Sun, 15 Nov 2009 03:17:12 +0000 (22:17 -0500)] 
initial printing of RPL DIO messages

15 years agonetdisect converted icmp6_print()
Michael Richardson [Sun, 15 Nov 2009 02:36:05 +0000 (21:36 -0500)] 
netdisect converted icmp6_print()

15 years agoinitial test case for decoding RPL messages
Michael Richardson [Sun, 15 Nov 2009 02:21:58 +0000 (21:21 -0500)] 
initial test case for decoding RPL messages

15 years agoignore generated files from tests
Michael Richardson [Sun, 15 Nov 2009 01:58:41 +0000 (20:58 -0500)] 
ignore generated files from tests

15 years agoupdated ikev2 test cases to show initiator/responder flag
Michael Richardson [Sun, 15 Nov 2009 01:55:24 +0000 (20:55 -0500)] 
updated ikev2 test cases to show initiator/responder flag

15 years agoupdate bgp tests with new tcp seq/ack flags
Michael Richardson [Sun, 15 Nov 2009 01:54:59 +0000 (20:54 -0500)] 
update bgp tests with new tcp seq/ack flags

15 years agoupdated print-flags test case to have new TCP flags
Michael Richardson [Sun, 15 Nov 2009 01:51:44 +0000 (20:51 -0500)] 
updated print-flags test case to have new TCP flags

15 years agoupdated lmp test case to have no-cksum flag
Michael Richardson [Sun, 15 Nov 2009 01:51:28 +0000 (20:51 -0500)] 
updated lmp test case to have no-cksum flag

15 years agomake check should depend uponi presence of built tcpdump
Michael Richardson [Sun, 15 Nov 2009 01:50:39 +0000 (20:50 -0500)] 
make check should depend uponi presence of built tcpdump

15 years agoWhen parsing IEs in a management frame, save the *first* IE of a given
Guy Harris [Wed, 4 Nov 2009 22:55:05 +0000 (14:55 -0800)] 
When parsing IEs in a management frame, save the *first* IE of a given
type we see, not the *last*, so the first one gets printed.  If,
however, a rates IE has a zero-length list of rates, don't save that
one; some devices send out frames with two rates IEs, one zero-length
and a later one containing the actual rates.

Have parse_elements() return an indication of whether truncation
occurred - and pass it the frame length, so it doesn't just keep parsing
until an end-of-captured-data test fails, as that would make *every*
frame look truncated.

15 years agoFix typo.
Guy Harris [Mon, 12 Oct 2009 23:58:56 +0000 (16:58 -0700)] 
Fix typo.

15 years agomodularize ip6 nextlayer checksum generation and
Hannes Gredler [Mon, 12 Oct 2009 08:16:58 +0000 (10:16 +0200)] 
modularize ip6 nextlayer checksum generation and
change tcp6 and pim6 to use it.

15 years agoUse tok2strbuf() rather than doing it by hand (with potentially-unsafe
Guy Harris [Sun, 11 Oct 2009 19:49:05 +0000 (12:49 -0700)] 
Use tok2strbuf() rather than doing it by hand (with potentially-unsafe
routines).  Thanks and a tip of the Hatlo hat to OpenBSD's linker for
warning about that.

15 years agoOn some platforms we appear to get a warning because dnsname_print() is
Guy Harris [Mon, 21 Sep 2009 17:16:14 +0000 (10:16 -0700)] 
On some platforms we appear to get a warning because dnsname_print() is
declared static and later defined without static.  Define it with
static.

15 years agoDon't use "uint"; it's not guaranteed to be defined.
Guy Harris [Mon, 21 Sep 2009 17:02:05 +0000 (10:02 -0700)] 
Don't use "uint"; it's not guaranteed to be defined.

Don't define a variable in the middle of a block; not all C compilers
support that.

15 years agoReport ps_ifdrop if it's non-zero.
Guy Harris [Mon, 7 Sep 2009 23:52:15 +0000 (16:52 -0700)] 
Report ps_ifdrop if it's non-zero.

15 years agoFrom Fang Wang:
Guy Harris [Tue, 18 Aug 2009 22:02:37 +0000 (15:02 -0700)] 
From Fang Wang:

RFC 5482 specifies a new TCP option -- the TCP User Timeout
Option (UTO) -- that allows one end of a TCP connection to
advertise its current user timeout value.  The Kind of UTO
option is 28 and the Length is 4.  The tcpdump currently does
not recognize UTO option.

15 years agoJust use a u_char * to go through the packet data; don't try to be
Guy Harris [Tue, 18 Aug 2009 21:56:06 +0000 (14:56 -0700)] 
Just use a u_char * to go through the packet data; don't try to be
clever and use a u_int16_t * - the generated code will be the same, and
this code works even if the Offset Size field value is odd.

15 years agoCheck that bytes exist in the captured data before fetching them.
Guy Harris [Tue, 18 Aug 2009 21:47:13 +0000 (14:47 -0700)] 
Check that bytes exist in the captured data before fetching them.

15 years agoAdd support for IEEE 802.15.4 packets
Dmitry Eremin-Solenikov [Sun, 9 Aug 2009 14:06:27 +0000 (18:06 +0400)] 
Add support for IEEE 802.15.4 packets

Add a dummy printer for ieee 802.15.4 pakets.

Signed-off-by: Dmitry Eremin-Solenikov <[email protected]>
15 years agoFrom Karl Norby: print the unsigned packet statistics with %u, not %d.
Guy Harris [Sat, 1 Aug 2009 01:08:12 +0000 (18:08 -0700)] 
From Karl Norby: print the unsigned packet statistics with %u, not %d.

15 years agoMake it look more like the FreeBSD tcpdump ethertype.h.
Guy Harris [Sat, 25 Jul 2009 17:48:32 +0000 (10:48 -0700)] 
Make it look more like the FreeBSD tcpdump ethertype.h.

15 years agoFrom FreeBSD: handle the non-standard Ethertypes for PPPoE that some
Guy Harris [Sat, 25 Jul 2009 17:47:00 +0000 (10:47 -0700)] 
From FreeBSD: handle the non-standard Ethertypes for PPPoE that some
3Com hardware uses.

15 years agoFrom Toshihiro Kanda, via FreeBSD: fix printing of address in AARP
Guy Harris [Sat, 25 Jul 2009 17:37:47 +0000 (10:37 -0700)] 
From Toshihiro Kanda, via FreeBSD: fix printing of address in AARP
replies (the relevant address is the source address, not the destination
address).

15 years agoMerge in stuff from the current top-of-tree FreeBSD tcpdump; this picks
Guy Harris [Wed, 15 Jul 2009 01:23:06 +0000 (18:23 -0700)] 
Merge in stuff from the current top-of-tree FreeBSD tcpdump; this picks
up changes from Sam Leffler for 802.11n and 802.11s.

Clean up ieee802_11_radio.h a bit.

15 years agoGet rid of some autoconf macros not used, or no longer used, with
Guy Harris [Wed, 8 Jul 2009 08:22:27 +0000 (01:22 -0700)] 
Get rid of some autoconf macros not used, or no longer used, with
tcpdump.

Check whether __attribute__((format)) can be applied to function
pointers and, if not, don't apply it to function pointers; some older
versions of GCC appear to support applying it to functions, but not
pointers to functions.

As we've gotten rid of missing/getaddrinfo.c, don't use it.

15 years agoNothing in missing/getaddrinfo.c is used in tcpdump, and it's a pain to
Guy Harris [Wed, 8 Jul 2009 08:18:11 +0000 (01:18 -0700)] 
Nothing in missing/getaddrinfo.c is used in tcpdump, and it's a pain to
make it build on Tru64 UNIX, so get rid of it.

15 years agoIn AC_LBL_C_INIT, don't explicitly set V_CCOPT, just set $1, as is done
Guy Harris [Sun, 5 Jul 2009 00:57:01 +0000 (17:57 -0700)] 
In AC_LBL_C_INIT, don't explicitly set V_CCOPT, just set $1, as is done
elsewhere in the macro.

For the Alpha C compiler, don't set -std1 unless it's necessary for ANSI
mode - we don't want to limit the compiler only to C89 features and,
more importantly, C89 headers (<inttypes.h> doesn't work, on at least
some versions of Digital UNIX and some C compiler versions, with -std1).

15 years agoCopy a comment from the libpcap configure.in.
Guy Harris [Fri, 3 Jul 2009 02:49:53 +0000 (19:49 -0700)] 
Copy a comment from the libpcap configure.in.

15 years agoInclude Makefile-devel-adds in the release tarball, just in case
Guy Harris [Fri, 3 Jul 2009 02:48:54 +0000 (19:48 -0700)] 
Include Makefile-devel-adds in the release tarball, just in case
somebody creates .devel.

15 years agoAdd a comment to the AC_DEFINE call for const, just as we did for
Guy Harris [Fri, 3 Jul 2009 02:28:34 +0000 (19:28 -0700)] 
Add a comment to the AC_DEFINE call for const, just as we did for
libpcap; that eliminates the need to list it in acconfig.h, so get rid
of that.

15 years agoExpand another AC_DEFINE comment.
Guy Harris [Fri, 3 Jul 2009 02:01:20 +0000 (19:01 -0700)] 
Expand another AC_DEFINE comment.

15 years agoPick up some AC_DEFINE comments from libpcap's aclocal.m4.
Guy Harris [Fri, 3 Jul 2009 01:59:24 +0000 (18:59 -0700)] 
Pick up some AC_DEFINE comments from libpcap's aclocal.m4.

Remove unused items, and items for which we now have AC_DEFINE() calls
that supply a comment, from acconfig.h.

Note that a host OS that matches "osf*" is assumed to be DEC
OSF1/Digital UNIX/Tru64 UNIX.

15 years agoAt least on HP-UX:
Guy Harris [Fri, 3 Jul 2009 01:43:40 +0000 (18:43 -0700)] 
At least on HP-UX:

1) getrpcbynumber() is declared in <netdb.h>, not any of the RPC
   header files

and

2) if _XOPEN_SOURCE_EXTENDED is defined, <netdb.h> doesn't declare
   it

so we undefine it.

15 years agoCast argument to isdigit to unsigned char, so if the characters 8th bit
Guy Harris [Thu, 2 Jul 2009 19:00:55 +0000 (12:00 -0700)] 
Cast argument to isdigit to unsigned char, so if the characters 8th bit
is set, it doesn't get treated as a negative number.

15 years agoAs with libpcap, so with tcpdump; when testing whether "inline" works,
Guy Harris [Thu, 2 Jul 2009 18:55:55 +0000 (11:55 -0700)] 
As with libpcap, so with tcpdump; when testing whether "inline" works,
use the V_CCOPT flags, because some of those might disable inlining;
otherwise, the test for inlining will succeed, but inlining won't work
with the options we're using when compiling.

15 years agoAs with libpcap, so with tcpdump; when testing whether "inline" works,
Guy Harris [Thu, 2 Jul 2009 18:55:00 +0000 (11:55 -0700)] 
As with libpcap, so with tcpdump; when testing whether "inline" works,
use the V_CCOPT flags, because some of those might disable inlining;
otherwise, the test for inlining will succeed, but inlining won't work
with the options we're using when compiling.