annotate libpurple/nat-pmp.c @ 25687:03541b3adb38

propagate from branch 'im.pidgin.pidgin' (head f3cf3ba10ebc911a27833496dcc4d4dca3de66c5) to branch 'im.pidgin.cpw.malu.xmpp.idle' (head 8a78a27a5123f05441fd5df8c423b57929973128)
author Marcus Lundblad <ml@update.uu.se>
date Sun, 01 Feb 2009 18:04:38 +0000
parents b175f6441bba
children 584063555949
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
1 /**
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
2 * @file nat-pmp.c NAT-PMP Implementation
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
3 * @ingroup core
20074
6bf32c9e15a7 remove gpl boilerplate from doxygen docs
Sean Egan <seanegan@gmail.com>
parents: 19654
diff changeset
4 */
6bf32c9e15a7 remove gpl boilerplate from doxygen docs
Sean Egan <seanegan@gmail.com>
parents: 19654
diff changeset
5
6bf32c9e15a7 remove gpl boilerplate from doxygen docs
Sean Egan <seanegan@gmail.com>
parents: 19654
diff changeset
6 /* purple
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
7 *
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
8 * Purple is the legal property of its developers, whose names are too numerous
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
9 * to list here. Please refer to the COPYRIGHT file distributed with this
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
10 * source distribution.
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
11 *
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
12 * Most code in nat-pmp.c copyright (C) 2007, R. Tyler Ballance, bleep, LLC.
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
13 * This file is distributed under the 3-clause (modified) BSD license:
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
14 * Redistribution and use in source and binary forms, with or without modification, are permitted
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
15 * provided that the following conditions are met:
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
16 *
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
17 * Redistributions of source code must retain the above copyright notice, this list of conditions and
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
18 * the following disclaimer.
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
19 * Neither the name of the bleep. LLC nor the names of its contributors may be used to endorse or promote
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
20 * products derived from this software without specific prior written permission.
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
21 *
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
22 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
23 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
24 * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
25 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
26 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
28 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
29 * OF SUCH DAMAGE.
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
30 */
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
31
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
32 #include "nat-pmp.h"
16908
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
33 #include "internal.h"
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
34 #include "debug.h"
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
35 #include "signals.h"
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
36 #include "network.h"
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
37
23375
b175f6441bba Some header inclusion fixes to help on NetBSD/FreeBSD.
Richard Laager <rlaager@wiktel.com>
parents: 23201
diff changeset
38 #ifdef HAVE_SYS_PARAM_H
b175f6441bba Some header inclusion fixes to help on NetBSD/FreeBSD.
Richard Laager <rlaager@wiktel.com>
parents: 23201
diff changeset
39 #include <sys/param.h>
b175f6441bba Some header inclusion fixes to help on NetBSD/FreeBSD.
Richard Laager <rlaager@wiktel.com>
parents: 23201
diff changeset
40 #endif
b175f6441bba Some header inclusion fixes to help on NetBSD/FreeBSD.
Richard Laager <rlaager@wiktel.com>
parents: 23201
diff changeset
41
16908
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
42 #ifdef HAVE_SYS_SYSCTL_H
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
43 #include <sys/sysctl.h>
15975
328ae2c041c3 fix win32 build for nat-pmp stuff. I haven't tested it, but it compiles.
Daniel Atallah <daniel.atallah@gmail.com>
parents: 15974
diff changeset
44 #endif
328ae2c041c3 fix win32 build for nat-pmp stuff. I haven't tested it, but it compiles.
Daniel Atallah <daniel.atallah@gmail.com>
parents: 15974
diff changeset
45
16908
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
46 #ifdef HAVE_SYS_SOCKET_H
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
47 #include <sys/socket.h>
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
48 #endif
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
49
16908
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
50 /* We will need sysctl() and NET_RT_DUMP, both of which are not present
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
51 * on all platforms, to continue. */
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
52 #if defined(HAVE_SYS_SYSCTL_H) && defined(NET_RT_DUMP)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
53
16908
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
54 #include <sys/types.h>
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
55 #include <net/route.h>
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
56
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
57 #define PMP_DEBUG 1
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
58
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
59 typedef struct {
16907
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
60 guint8 version;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
61 guint8 opcode;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
62 } PurplePmpIpRequest;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
63
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
64 typedef struct {
16907
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
65 guint8 version;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
66 guint8 opcode; /* 128 + n */
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
67 guint16 resultcode;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
68 guint32 epoch;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
69 guint32 address;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
70 } PurplePmpIpResponse;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
71
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
72 typedef struct {
16907
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
73 guint8 version;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
74 guint8 opcode;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
75 char reserved[2];
16907
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
76 guint16 privateport;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
77 guint16 publicport;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
78 guint32 lifetime;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
79 } PurplePmpMapRequest;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
80
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
81 struct _PurplePmpMapResponse {
16907
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
82 guint8 version;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
83 guint8 opcode;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
84 guint16 resultcode;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
85 guint32 epoch;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
86 guint16 privateport;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
87 guint16 publicport;
533ebd404184 Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@wiktel.com>
parents: 16345
diff changeset
88 guint32 lifetime;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
89 };
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
90
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
91 typedef struct _PurplePmpMapResponse PurplePmpMapResponse;
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
92
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
93 typedef enum {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
94 PURPLE_PMP_STATUS_UNDISCOVERED = -1,
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
95 PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER,
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
96 PURPLE_PMP_STATUS_DISCOVERING,
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
97 PURPLE_PMP_STATUS_DISCOVERED
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
98 } PurpleUPnPStatus;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
99
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
100 typedef struct {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
101 PurpleUPnPStatus status;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
102 gchar *publicip;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
103 } PurplePmpInfo;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
104
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
105 static PurplePmpInfo pmp_info = {PURPLE_PMP_STATUS_UNDISCOVERED, NULL};
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
106
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
107 /*
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
108 * Thanks to R. Matthew Emerson for the fixes on this
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
109 */
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
110
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
111 #define PMP_MAP_OPCODE_UDP 1
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
112 #define PMP_MAP_OPCODE_TCP 2
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
113
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
114 #define PMP_VERSION 0
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
115 #define PMP_PORT 5351
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
116 #define PMP_TIMEOUT 250000 /* 250000 useconds */
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
117
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
118 /* alignment constraint for routing socket */
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
119 #define ROUNDUP(a) ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long))
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
120 #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len))
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
121
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
122 static void
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
123 get_rtaddrs(int bitmask, struct sockaddr *sa, struct sockaddr *addrs[])
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
124 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
125 int i;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
126
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
127 for (i = 0; i < RTAX_MAX; i++)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
128 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
129 if (bitmask & (1 << i))
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
130 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
131 addrs[i] = sa;
23201
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
132 #ifdef HAVE_STRUCT_SOCKADDR_SA_LEN
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
133 sa = (struct sockaddr *)(ROUNDUP(sa->sa_len) + (char *)sa);
23201
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
134 #else
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
135 if (sa->sa_family == AF_INET)
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
136 sa = (struct sockaddr*)(sizeof(struct sockaddr_in) + (char *)sa);
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
137 #ifdef AF_INET6
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
138 else if (sa->sa_family == AF_INET6)
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
139 sa = (struct sockaddr*)(sizeof(struct sockaddr_in6) + (char *)sa);
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
140 #endif
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
141 #endif
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
142 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
143 else
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
144 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
145 addrs[i] = NULL;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
146 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
147 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
148 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
149
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
150 static int
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
151 is_default_route(struct sockaddr *sa, struct sockaddr *mask)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
152 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
153 struct sockaddr_in *sin;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
154
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
155 if (sa->sa_family != AF_INET)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
156 return 0;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
157
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
158 sin = (struct sockaddr_in *)sa;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
159 if ((sin->sin_addr.s_addr == INADDR_ANY) &&
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
160 mask &&
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
161 (ntohl(((struct sockaddr_in *)mask)->sin_addr.s_addr) == 0L ||
23201
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
162 #ifdef HAVE_STRUCT_SOCKADDR_SA_LEN
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
163 mask->sa_len == 0
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
164 #else
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
165 0
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
166 #endif
313b87adb730 A patch from Peter O'Gorman at The Written Word, Inc. to fix various
Peter O'Gorman <pogma@thewrittenword.com>
parents: 22533
diff changeset
167 ))
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
168 return 1;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
169 else
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
170 return 0;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
171 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
172
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
173 /*!
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
174 * The return sockaddr_in must be g_free()'d when no longer needed
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
175 */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
176 static struct sockaddr_in *
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
177 default_gw()
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
178 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
179 int mib[6];
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
180 size_t needed;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
181 char *buf, *next, *lim;
15950
45cbb80bfd54 Using rt_msghdr instead of rt_msghdr2 compiles on a greater number of systems -- specifically Mac OS X 10.3.x, in addition to the previous compilation on OS X 10.4.x and above -- and seems to work just as well. If anyone knows how the use of these seemingly very similar structs differs, enlightenment would be appreciated.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15921
diff changeset
182 struct rt_msghdr *rtm;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
183 struct sockaddr *sa;
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
184 struct sockaddr_in *sin = NULL;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
185 gboolean found = FALSE;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
186
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
187 mib[0] = CTL_NET;
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
188 mib[1] = PF_ROUTE; /* entire routing table or a subset of it */
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
189 mib[2] = 0; /* protocol number - always 0 */
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
190 mib[3] = 0; /* address family - 0 for all addres families */
15950
45cbb80bfd54 Using rt_msghdr instead of rt_msghdr2 compiles on a greater number of systems -- specifically Mac OS X 10.3.x, in addition to the previous compilation on OS X 10.4.x and above -- and seems to work just as well. If anyone knows how the use of these seemingly very similar structs differs, enlightenment would be appreciated.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15921
diff changeset
191 mib[4] = NET_RT_DUMP;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
192 mib[5] = 0;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
193
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
194 /* Determine the buffer side needed to get the full routing table */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
195 if (sysctl(mib, 6, NULL, &needed, NULL, 0) < 0)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
196 {
19654
941965d6fd88 Patch from QuLogic. Fixes #2903 ('Missing newlines in debug messages.')
Sadrul Habib Chowdhury <imadil@gmail.com>
parents: 19268
diff changeset
197 purple_debug_warning("nat-pmp", "sysctl: net.route.0.0.dump estimate\n");
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
198 return NULL;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
199 }
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
200
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
201 if (!(buf = malloc(needed)))
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
202 {
22533
8972581c2326 Improved some of the debug output from nat-pmp
Evan Schoenberg <evan.s@dreskin.net>
parents: 21121
diff changeset
203 purple_debug_warning("nat-pmp", "Failed to malloc %i\n", needed);
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
204 return NULL;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
205 }
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
206
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
207 /* Read the routing table into buf */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
208 if (sysctl(mib, 6, buf, &needed, NULL, 0) < 0)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
209 {
19654
941965d6fd88 Patch from QuLogic. Fixes #2903 ('Missing newlines in debug messages.')
Sadrul Habib Chowdhury <imadil@gmail.com>
parents: 19268
diff changeset
210 purple_debug_warning("nat-pmp", "sysctl: net.route.0.0.dump\n");
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
211 return NULL;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
212 }
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
213
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
214 lim = buf + needed;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
215
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
216 for (next = buf; next < lim; next += rtm->rtm_msglen)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
217 {
15950
45cbb80bfd54 Using rt_msghdr instead of rt_msghdr2 compiles on a greater number of systems -- specifically Mac OS X 10.3.x, in addition to the previous compilation on OS X 10.4.x and above -- and seems to work just as well. If anyone knows how the use of these seemingly very similar structs differs, enlightenment would be appreciated.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15921
diff changeset
218 rtm = (struct rt_msghdr *)next;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
219 sa = (struct sockaddr *)(rtm + 1);
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
220
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
221 if (sa->sa_family == AF_INET)
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
222 {
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
223 sin = (struct sockaddr_in*) sa;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
224
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
225 if ((rtm->rtm_flags & RTF_GATEWAY) && sin->sin_addr.s_addr == INADDR_ANY)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
226 {
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
227 /* We found the default route. Now get the destination address and netmask. */
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
228 struct sockaddr *rti_info[RTAX_MAX];
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
229 struct sockaddr addr, mask;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
230
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
231 get_rtaddrs(rtm->rtm_addrs, sa, rti_info);
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
232 memset(&addr, 0, sizeof(addr));
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
233
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
234 if (rtm->rtm_addrs & RTA_DST)
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
235 memcpy(&addr, rti_info[RTAX_DST], sizeof(addr));
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
236
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
237 memset(&mask, 0, sizeof(mask));
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
238
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
239 if (rtm->rtm_addrs & RTA_NETMASK)
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
240 memcpy(&mask, rti_info[RTAX_NETMASK], sizeof(mask));
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
241
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
242 if (rtm->rtm_addrs & RTA_GATEWAY &&
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
243 is_default_route(&addr, &mask))
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
244 {
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
245 if (rti_info[RTAX_GATEWAY]) {
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
246 struct sockaddr_in *rti_sin = (struct sockaddr_in *)rti_info[RTAX_GATEWAY];
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
247 sin = g_new0(struct sockaddr_in, 1);
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
248 sin->sin_family = rti_sin->sin_family;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
249 sin->sin_port = rti_sin->sin_port;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
250 sin->sin_addr.s_addr = rti_sin->sin_addr.s_addr;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
251 memcpy(sin, rti_info[RTAX_GATEWAY], sizeof(struct sockaddr_in));
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
252
22533
8972581c2326 Improved some of the debug output from nat-pmp
Evan Schoenberg <evan.s@dreskin.net>
parents: 21121
diff changeset
253 purple_debug_info("nat-pmp", "Found a default gateway\n");
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
254 found = TRUE;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
255 break;
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
256 }
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
257 }
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
258 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
259 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
260 }
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
261
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
262 return (found ? sin : NULL);
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
263 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
264
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
265 /*!
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
266 * purple_pmp_get_public_ip() will return the publicly facing IP address of the
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
267 * default NAT gateway. The function will return NULL if:
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
268 * - The gateway doesn't support NAT-PMP
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
269 * - The gateway errors in some other spectacular fashion
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
270 */
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
271 char *
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
272 purple_pmp_get_public_ip()
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
273 {
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
274 struct sockaddr_in addr, *gateway, *publicsockaddr = NULL;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
275 struct timeval req_timeout;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
276 socklen_t len;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
277
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
278 PurplePmpIpRequest req;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
279 PurplePmpIpResponse resp;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
280 int sendfd;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
281
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
282 if (pmp_info.status == PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER)
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
283 return NULL;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
284
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
285 if ((pmp_info.status == PURPLE_PMP_STATUS_DISCOVERED) && (pmp_info.publicip != NULL))
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
286 {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
287 #ifdef PMP_DEBUG
19654
941965d6fd88 Patch from QuLogic. Fixes #2903 ('Missing newlines in debug messages.')
Sadrul Habib Chowdhury <imadil@gmail.com>
parents: 19268
diff changeset
288 purple_debug_info("nat-pmp", "Returning cached publicip %s\n",pmp_info.publicip);
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
289 #endif
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
290 return pmp_info.publicip;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
291 }
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
292
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
293 gateway = default_gw();
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
294
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
295 if (!gateway)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
296 {
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
297 purple_debug_info("nat-pmp", "Cannot request public IP from a NULL gateway!\n");
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
298 /* If we get a NULL gateway, don't try again next time */
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
299 pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
300 return NULL;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
301 }
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
302
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
303 /* Default port for NAT-PMP is 5351 */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
304 if (gateway->sin_port != PMP_PORT)
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
305 gateway->sin_port = htons(PMP_PORT);
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
306
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
307 req_timeout.tv_sec = 0;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
308 req_timeout.tv_usec = PMP_TIMEOUT;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
309
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
310 sendfd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP);
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
311
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
312 /* Clean out both req and resp structures */
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
313 memset(&req, 0, sizeof(PurplePmpIpRequest));
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
314 memset(&resp, 0, sizeof(PurplePmpIpResponse));
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
315 req.version = 0;
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
316 req.opcode = 0;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
317
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
318 /* The NAT-PMP spec says we should attempt to contact the gateway 9 times, doubling the time we wait each time.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
319 * Even starting with a timeout of 0.1 seconds, that means that we have a total waiting of 204.6 seconds.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
320 * With the recommended timeout of 0.25 seconds, we're talking 511.5 seconds (8.5 minutes).
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
321 *
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
322 * This seems really silly... if this were nonblocking, a couple retries might be in order, but it's not at present.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
323 */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
324 #ifdef PMP_DEBUG
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
325 purple_debug_info("nat-pmp", "Attempting to retrieve the public ip address for the NAT device at: %s\n", inet_ntoa(gateway->sin_addr));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
326 purple_debug_info("nat-pmp", "\tTimeout: %ds %dus\n", req_timeout.tv_sec, req_timeout.tv_usec);
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
327 #endif
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
328
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
329 /* TODO: Non-blocking! */
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
330
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
331 if (sendto(sendfd, &req, sizeof(req), 0, (struct sockaddr *)(gateway), sizeof(struct sockaddr)) < 0)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
332 {
21121
35b4f1dc4c8d replace most calls to strerror with calls to g_strerror. strerror will return
Nathan Walp <nwalp@pidgin.im>
parents: 20074
diff changeset
333 purple_debug_info("nat-pmp", "There was an error sending the NAT-PMP public IP request! (%s)\n", g_strerror(errno));
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
334 g_free(gateway);
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
335 pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
336 return NULL;
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
337 }
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
338
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
339 if (setsockopt(sendfd, SOL_SOCKET, SO_RCVTIMEO, &req_timeout, sizeof(req_timeout)) < 0)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
340 {
21121
35b4f1dc4c8d replace most calls to strerror with calls to g_strerror. strerror will return
Nathan Walp <nwalp@pidgin.im>
parents: 20074
diff changeset
341 purple_debug_info("nat-pmp", "There was an error setting the socket's options! (%s)\n", g_strerror(errno));
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
342 g_free(gateway);
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
343 pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
344 return NULL;
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
345 }
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
346
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
347 /* TODO: Non-blocking! */
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
348 len = sizeof(struct sockaddr_in);
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
349 if (recvfrom(sendfd, &resp, sizeof(PurplePmpIpResponse), 0, (struct sockaddr *)(&addr), &len) < 0)
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
350 {
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
351 if (errno != EAGAIN)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
352 {
21121
35b4f1dc4c8d replace most calls to strerror with calls to g_strerror. strerror will return
Nathan Walp <nwalp@pidgin.im>
parents: 20074
diff changeset
353 purple_debug_info("nat-pmp", "There was an error receiving the response from the NAT-PMP device! (%s)\n", g_strerror(errno));
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
354 g_free(gateway);
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
355 pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
356 return NULL;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
357 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
358 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
359
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
360 if (addr.sin_addr.s_addr == gateway->sin_addr.s_addr)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
361 publicsockaddr = &addr;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
362 else
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
363 {
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
364 purple_debug_info("nat-pmp", "Response was not received from our gateway! Instead from: %s\n", inet_ntoa(addr.sin_addr));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
365 g_free(gateway);
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
366
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
367 pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
368 return NULL;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
369 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
370
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
371 if (!publicsockaddr) {
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
372 g_free(gateway);
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
373
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
374 pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER;
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
375 return NULL;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
376 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
377
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
378 #ifdef PMP_DEBUG
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
379 purple_debug_info("nat-pmp", "Response received from NAT-PMP device:\n");
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
380 purple_debug_info("nat-pmp", "version: %d\n", resp.version);
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
381 purple_debug_info("nat-pmp", "opcode: %d\n", resp.opcode);
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
382 purple_debug_info("nat-pmp", "resultcode: %d\n", ntohs(resp.resultcode));
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
383 purple_debug_info("nat-pmp", "epoch: %d\n", ntohl(resp.epoch));
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
384 struct in_addr in;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
385 in.s_addr = resp.address;
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
386 purple_debug_info("nat-pmp", "address: %s\n", inet_ntoa(in));
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
387 #endif
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
388
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
389 publicsockaddr->sin_addr.s_addr = resp.address;
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
390
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
391 g_free(gateway);
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
392
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
393 g_free(pmp_info.publicip);
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
394 pmp_info.publicip = g_strdup(inet_ntoa(publicsockaddr->sin_addr));
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
395 pmp_info.status = PURPLE_PMP_STATUS_DISCOVERED;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
396
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
397 return inet_ntoa(publicsockaddr->sin_addr);
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
398 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
399
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
400 gboolean
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
401 purple_pmp_create_map(PurplePmpType type, unsigned short privateport, unsigned short publicport, int lifetime)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
402 {
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
403 struct sockaddr_in *gateway;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
404 gboolean success = TRUE;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
405 int sendfd;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
406 struct timeval req_timeout;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
407 PurplePmpMapRequest req;
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
408 PurplePmpMapResponse *resp;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
409
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
410 gateway = default_gw();
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
411
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
412 if (!gateway)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
413 {
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
414 purple_debug_info("nat-pmp", "Cannot create mapping on a NULL gateway!\n");
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
415 return FALSE;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
416 }
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
417
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
418 /* Default port for NAT-PMP is 5351 */
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
419 if (gateway->sin_port != PMP_PORT)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
420 gateway->sin_port = htons(PMP_PORT);
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
421
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
422 resp = g_new0(PurplePmpMapResponse, 1);
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
423
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
424 req_timeout.tv_sec = 0;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
425 req_timeout.tv_usec = PMP_TIMEOUT;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
426
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
427 sendfd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP);
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
428
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
429 /* Set up the req */
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
430 memset(&req, 0, sizeof(PurplePmpMapRequest));
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
431 req.version = 0;
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
432 req.opcode = ((type == PURPLE_PMP_TYPE_UDP) ? PMP_MAP_OPCODE_UDP : PMP_MAP_OPCODE_TCP);
16345
6b4e778ee4b4 Death to // comments.
Richard Laager <rlaager@wiktel.com>
parents: 15975
diff changeset
433 req.privateport = htons(privateport); /* What a difference byte ordering makes...d'oh! */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
434 req.publicport = htons(publicport);
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
435 req.lifetime = htonl(lifetime);
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
436
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
437 /* The NAT-PMP spec says we should attempt to contact the gateway 9 times, doubling the time we wait each time.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
438 * Even starting with a timeout of 0.1 seconds, that means that we have a total waiting of 204.6 seconds.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
439 * With the recommended timeout of 0.25 seconds, we're talking 511.5 seconds (8.5 minutes).
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
440 *
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
441 * This seems really silly... if this were nonblocking, a couple retries might be in order, but it's not at present.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
442 * XXX Make this nonblocking.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
443 * XXX This code looks like the pmp_get_public_ip() code. Can it be consolidated?
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
444 */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
445 #ifdef PMP_DEBUG
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
446 purple_debug_info("nat-pmp", "Attempting to create a NAT-PMP mapping the private port %d, and the public port %d\n", privateport, publicport);
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
447 purple_debug_info("nat-pmp", "\tTimeout: %ds %dus\n", req_timeout.tv_sec, req_timeout.tv_usec);
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
448 #endif
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
449
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
450 /* TODO: Non-blocking! */
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
451 success = (sendto(sendfd, &req, sizeof(req), 0, (struct sockaddr *)(gateway), sizeof(struct sockaddr)) >= 0);
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
452 if (!success)
21121
35b4f1dc4c8d replace most calls to strerror with calls to g_strerror. strerror will return
Nathan Walp <nwalp@pidgin.im>
parents: 20074
diff changeset
453 purple_debug_info("nat-pmp", "There was an error sending the NAT-PMP mapping request! (%s)\n", g_strerror(errno));
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
454
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
455 if (success)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
456 {
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
457 success = (setsockopt(sendfd, SOL_SOCKET, SO_RCVTIMEO, &req_timeout, sizeof(req_timeout)) >= 0);
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
458 if (!success)
21121
35b4f1dc4c8d replace most calls to strerror with calls to g_strerror. strerror will return
Nathan Walp <nwalp@pidgin.im>
parents: 20074
diff changeset
459 purple_debug_info("nat-pmp", "There was an error setting the socket's options! (%s)\n", g_strerror(errno));
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
460 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
461
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
462 if (success)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
463 {
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
464 /* The original code treats EAGAIN as a reason to iterate.. but I've removed iteration. This may be a problem */
15898
9cfe41743c65 More complete namespacing of nat-pmp
Mark Doliner <mark@kingant.net>
parents: 15855
diff changeset
465 /* TODO: Non-blocking! */
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
466 success = ((recvfrom(sendfd, resp, sizeof(PurplePmpMapResponse), 0, NULL, NULL) >= 0) ||
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
467 (errno == EAGAIN));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
468 if (!success)
21121
35b4f1dc4c8d replace most calls to strerror with calls to g_strerror. strerror will return
Nathan Walp <nwalp@pidgin.im>
parents: 20074
diff changeset
469 purple_debug_info("nat-pmp", "There was an error receiving the response from the NAT-PMP device! (%s)\n", g_strerror(errno));
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
470 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
471
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
472 if (success)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
473 {
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
474 success = (resp->opcode == (req.opcode + 128));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
475 if (!success)
22533
8972581c2326 Improved some of the debug output from nat-pmp
Evan Schoenberg <evan.s@dreskin.net>
parents: 21121
diff changeset
476 purple_debug_info("nat-pmp", "The opcode for the response from the NAT device (%i) does not match the request opcode (%i + 128 = %i)!\n",
8972581c2326 Improved some of the debug output from nat-pmp
Evan Schoenberg <evan.s@dreskin.net>
parents: 21121
diff changeset
477 resp->opcode, req.opcode, req.opcode + 128);
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
478 }
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
479
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
480 #ifdef PMP_DEBUG
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
481 if (success)
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
482 {
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
483 purple_debug_info("nat-pmp", "Response received from NAT-PMP device:\n");
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
484 purple_debug_info("nat-pmp", "version: %d\n", resp->version);
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
485 purple_debug_info("nat-pmp", "opcode: %d\n", resp->opcode);
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
486 purple_debug_info("nat-pmp", "resultcode: %d\n", ntohs(resp->resultcode));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
487 purple_debug_info("nat-pmp", "epoch: %d\n", ntohl(resp->epoch));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
488 purple_debug_info("nat-pmp", "privateport: %d\n", ntohs(resp->privateport));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
489 purple_debug_info("nat-pmp", "publicport: %d\n", ntohs(resp->publicport));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
490 purple_debug_info("nat-pmp", "lifetime: %d\n", ntohl(resp->lifetime));
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
491 }
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
492 #endif
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
493
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
494 g_free(resp);
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
495 g_free(gateway);
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
496
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
497 /* XXX The private port may actually differ from the one we requested, according to the spec.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
498 * We don't handle that situation at present.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
499 *
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
500 * TODO: Look at the result and verify it matches what we wanted; either return a failure if it doesn't,
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
501 * or change network.c to know what to do if the desired private port shifts as a result of the nat-pmp operation.
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
502 */
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
503 return success;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
504 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
505
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
506 gboolean
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
507 purple_pmp_destroy_map(PurplePmpType type, unsigned short privateport)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
508 {
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
509 gboolean success;
15907
23c57ef2cf9f Whitespace fixes
Evan Schoenberg <evan.s@dreskin.net>
parents: 15905
diff changeset
510
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
511 success = purple_pmp_create_map(((type == PURPLE_PMP_TYPE_UDP) ? PMP_MAP_OPCODE_UDP : PMP_MAP_OPCODE_TCP),
15855
9b875f0ecb86 nat-pmp is now functional with a compatible router. I'm not enabling this code yet because it hasn't been tested with a router which doesn't support nat-pmp.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15822
diff changeset
512 privateport, 0, 0);
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
513 if (!success)
22533
8972581c2326 Improved some of the debug output from nat-pmp
Evan Schoenberg <evan.s@dreskin.net>
parents: 21121
diff changeset
514 purple_debug_warning("nat-pmp", "Failed to properly destroy mapping for %s port %d!\n",
8972581c2326 Improved some of the debug output from nat-pmp
Evan Schoenberg <evan.s@dreskin.net>
parents: 21121
diff changeset
515 ((type == PURPLE_PMP_TYPE_UDP) ? "UDP" : "TCP"), privateport);
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
516
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
517 return success;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
518 }
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
519
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
520 static void
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
521 purple_pmp_network_config_changed_cb(void *data)
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
522 {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
523 pmp_info.status = PURPLE_PMP_STATUS_UNDISCOVERED;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
524 g_free(pmp_info.publicip);
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
525 pmp_info.publicip = NULL;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
526 }
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
527
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
528 static void*
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
529 purple_pmp_get_handle(void)
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
530 {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
531 static int handle;
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
532
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
533 return &handle;
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
534 }
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
535
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
536 void
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
537 purple_pmp_init()
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
538 {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
539 purple_signal_connect(purple_network_get_handle(), "network-configuration-changed",
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
540 purple_pmp_get_handle(), PURPLE_CALLBACK(purple_pmp_network_config_changed_cb),
19268
db6218615be4 bcopy and bzero shouldn't be used. This allows Alver to compile nat-pmp.c on one of his obscure systems (and I fixed some other stuff that I noticed at the same time).
Daniel Atallah <daniel.atallah@gmail.com>
parents: 16908
diff changeset
541 GINT_TO_POINTER(0));
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
542 }
15950
45cbb80bfd54 Using rt_msghdr instead of rt_msghdr2 compiles on a greater number of systems -- specifically Mac OS X 10.3.x, in addition to the previous compilation on OS X 10.4.x and above -- and seems to work just as well. If anyone knows how the use of these seemingly very similar structs differs, enlightenment would be appreciated.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15921
diff changeset
543 #else /* #ifdef NET_RT_DUMP */
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
544 char *
15822
32c366eeeb99 sed -ie 's/gaim/purple/g'
Sean Egan <seanegan@gmail.com>
parents: 15646
diff changeset
545 purple_pmp_get_public_ip()
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
546 {
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
547 return NULL;
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
548 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
549
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
550 gboolean
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
551 purple_pmp_create_map(PurplePmpType type, unsigned short privateport, unsigned short publicport, int lifetime)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
552 {
15921
8bea220c0997 These should return FALSE, not NULL, when NET_RT_DUMP2 isn't available; they return a gboolean, not a pointer.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15907
diff changeset
553 return FALSE;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
554 }
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
555
15905
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
556 gboolean
2cf21661f828 Cleanup and fixes for nat-pmp. We no longer do the whole 'try 10 times, doubling the delay in a blocking manner' ttempt from the original code; as I note in the comments above where the attempt is made a single time, this leads to about 8 minutes of nonresponsiveness if the router both doesn't support nat-pmp and doesn't send back a response to let the program know.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15898
diff changeset
557 purple_pmp_destroy_map(PurplePmpType type, unsigned short privateport)
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
558 {
15921
8bea220c0997 These should return FALSE, not NULL, when NET_RT_DUMP2 isn't available; they return a gboolean, not a pointer.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15907
diff changeset
559 return FALSE;
15646
552be3958d6a Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evan.s@dreskin.net>
parents:
diff changeset
560 }
15974
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
561
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
562 void
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
563 purple_pmp_init()
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
564 {
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
565
1a12ce76c4f4 The network module now registers the signal 'network-configuration-changed' and emits it when a network change is detected via libnm or the win32 network monitor. The UI could also emit this signal if it knows something network.c doesn't. UPnP and NAT-PMP respond to the signal by clearing their IP address caches; changing networks without quitting/relaunching will now lead to the new IP address being (lazily) determined. This commit also enables nat-pmp and adds nat-pmp.[h|c] to the build process; please let me know if there are any problems building, as I only have OS X test machines.
Evan Schoenberg <evan.s@dreskin.net>
parents: 15950
diff changeset
566 }
16908
d0e74d0be43f I expanded your patch, Richard, to also do appropriate header checking in the configure script and then use that information. I removed some other #includes which don't appear to be needed.
Evan Schoenberg <evan.s@dreskin.net>
parents: 16907
diff changeset
567 #endif /* #if !(defined(HAVE_SYS_SYCTL_H) && defined(NET_RT_DUMP)) */