Tue, 14 Aug 2007 18:31:01 +0000
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).
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
1 | /** |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
2 | * @file nat-pmp.c NAT-PMP Implementation |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
3 | * @ingroup core |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
4 | * |
| 15884 | 5 | * purple |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
6 | * |
| 15884 | 7 | * Purple is the legal property of its developers, whose names are too numerous |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
8 | * to list here. Please refer to the COPYRIGHT file distributed with this |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
9 | * source distribution. |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
10 | * |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
11 | * Most code in nat-pmp.c copyright (C) 2007, R. Tyler Ballance, bleep, LLC. |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
12 | * This file is distributed under the 3-clause (modified) BSD license: |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
13 | * Redistribution and use in source and binary forms, with or without modification, are permitted |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
14 | * provided that the following conditions are met: |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
15 | * |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
16 | * Redistributions of source code must retain the above copyright notice, this list of conditions and |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
17 | * the following disclaimer. |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
18 | * Neither the name of the bleep. LLC nor the names of its contributors may be used to endorse or promote |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
19 | * products derived from this software without specific prior written permission. |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
20 | * |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
21 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
22 | * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
23 | * PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
24 | * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
25 | * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
26 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
27 | * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
28 | * OF SUCH DAMAGE. |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
29 | */ |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
30 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
31 | #include "nat-pmp.h" |
|
16965
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
32 | #include "internal.h" |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
33 | #include "debug.h" |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
34 | #include "signals.h" |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
35 | #include "network.h" |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
36 | |
|
16965
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
37 | #ifdef HAVE_SYS_SYSCTL_H |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
38 | #include <sys/sysctl.h> |
|
16036
d74e95743ae1
fix win32 build for nat-pmp stuff. I haven't tested it, but it compiles.
Daniel Atallah <datallah@pidgin.im>
parents:
16035
diff
changeset
|
39 | #endif |
|
d74e95743ae1
fix win32 build for nat-pmp stuff. I haven't tested it, but it compiles.
Daniel Atallah <datallah@pidgin.im>
parents:
16035
diff
changeset
|
40 | |
|
16965
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
41 | #ifdef HAVE_SYS_SOCKET_H |
|
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
42 | #include <sys/socket.h> |
|
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
43 | #endif |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
44 | |
|
16965
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
45 | /* We will need sysctl() and NET_RT_DUMP, both of which are not present |
|
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
46 | * on all platforms, to continue. */ |
|
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
47 | #if defined(HAVE_SYS_SYSCTL_H) && defined(NET_RT_DUMP) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
48 | |
|
16965
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
49 | #include <sys/types.h> |
|
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
50 | #include <net/route.h> |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
51 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
52 | #define PMP_DEBUG 1 |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
53 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
54 | typedef struct { |
|
16964
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
55 | guint8 version; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
56 | guint8 opcode; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
57 | } PurplePmpIpRequest; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
58 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
59 | typedef struct { |
|
16964
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
60 | guint8 version; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
61 | guint8 opcode; /* 128 + n */ |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
62 | guint16 resultcode; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
63 | guint32 epoch; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
64 | guint32 address; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
65 | } PurplePmpIpResponse; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
66 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
67 | typedef struct { |
|
16964
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
68 | guint8 version; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
69 | guint8 opcode; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
70 | char reserved[2]; |
|
16964
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
71 | guint16 privateport; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
72 | guint16 publicport; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
73 | guint32 lifetime; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
74 | } PurplePmpMapRequest; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
75 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
76 | struct _PurplePmpMapResponse { |
|
16964
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
77 | guint8 version; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
78 | guint8 opcode; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
79 | guint16 resultcode; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
80 | guint32 epoch; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
81 | guint16 privateport; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
82 | guint16 publicport; |
|
564274a64f4b
Change the nat-pmp code to use glib's integer types instead of those from
Richard Laager <rlaager@pidgin.im>
parents:
16413
diff
changeset
|
83 | guint32 lifetime; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
84 | }; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
85 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
86 | typedef struct _PurplePmpMapResponse PurplePmpMapResponse; |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
87 | |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
88 | typedef enum { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
89 | PURPLE_PMP_STATUS_UNDISCOVERED = -1, |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
90 | PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER, |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
91 | PURPLE_PMP_STATUS_DISCOVERING, |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
92 | PURPLE_PMP_STATUS_DISCOVERED |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
93 | } PurpleUPnPStatus; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
94 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
95 | typedef struct { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
96 | PurpleUPnPStatus status; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
97 | gchar *publicip; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
98 | } PurplePmpInfo; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
99 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
100 | static PurplePmpInfo pmp_info = {PURPLE_PMP_STATUS_UNDISCOVERED, NULL}; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
101 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
102 | /* |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
103 | * Thanks to R. Matthew Emerson for the fixes on this |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
104 | */ |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
105 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
106 | #define PMP_MAP_OPCODE_UDP 1 |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
107 | #define PMP_MAP_OPCODE_TCP 2 |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
108 | |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
109 | #define PMP_VERSION 0 |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
110 | #define PMP_PORT 5351 |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
111 | #define PMP_TIMEOUT 250000 /* 250000 useconds */ |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
112 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
113 | /* alignment constraint for routing socket */ |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
114 | #define ROUNDUP(a) ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
115 | #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
116 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
117 | static void |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
118 | get_rtaddrs(int bitmask, struct sockaddr *sa, struct sockaddr *addrs[]) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
119 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
120 | int i; |
| 15970 | 121 | |
| 122 | for (i = 0; i < RTAX_MAX; i++) | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
123 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
124 | if (bitmask & (1 << i)) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
125 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
126 | addrs[i] = sa; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
127 | sa = (struct sockaddr *)(ROUNDUP(sa->sa_len) + (char *)sa); |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
128 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
129 | else |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
130 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
131 | addrs[i] = NULL; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
132 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
133 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
134 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
135 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
136 | static int |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
137 | is_default_route(struct sockaddr *sa, struct sockaddr *mask) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
138 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
139 | struct sockaddr_in *sin; |
| 15970 | 140 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
141 | if (sa->sa_family != AF_INET) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
142 | return 0; |
| 15970 | 143 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
144 | sin = (struct sockaddr_in *)sa; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
145 | if ((sin->sin_addr.s_addr == INADDR_ANY) && |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
146 | mask && |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
147 | (ntohl(((struct sockaddr_in *)mask)->sin_addr.s_addr) == 0L || |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
148 | mask->sa_len == 0)) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
149 | return 1; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
150 | else |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
151 | return 0; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
152 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
153 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
154 | /*! |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
155 | * The return sockaddr_in must be g_free()'d when no longer needed |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
156 | */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
157 | static struct sockaddr_in * |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
158 | default_gw() |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
159 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
160 | int mib[6]; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
161 | size_t needed; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
162 | char *buf, *next, *lim; |
|
16014
9ee607a29a8c
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 <evands@pidgin.im>
parents:
15984
diff
changeset
|
163 | struct rt_msghdr *rtm; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
164 | struct sockaddr *sa; |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
165 | struct sockaddr_in *sin = NULL; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
166 | gboolean found = FALSE; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
167 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
168 | mib[0] = CTL_NET; |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
169 | mib[1] = PF_ROUTE; /* entire routing table or a subset of it */ |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
170 | mib[2] = 0; /* protocol number - always 0 */ |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
171 | mib[3] = 0; /* address family - 0 for all addres families */ |
|
16014
9ee607a29a8c
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 <evands@pidgin.im>
parents:
15984
diff
changeset
|
172 | mib[4] = NET_RT_DUMP; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
173 | mib[5] = 0; |
| 15970 | 174 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
175 | /* Determine the buffer side needed to get the full routing table */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
176 | if (sysctl(mib, 6, NULL, &needed, NULL, 0) < 0) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
177 | { |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
178 | purple_debug_warning("nat-pmp", "sysctl: net.route.0.0.dump estimate"); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
179 | return NULL; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
180 | } |
| 15970 | 181 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
182 | if (!(buf = malloc(needed))) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
183 | { |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
184 | purple_debug_warning("nat-pmp", "malloc"); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
185 | return NULL; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
186 | } |
| 15970 | 187 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
188 | /* Read the routing table into buf */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
189 | if (sysctl(mib, 6, buf, &needed, NULL, 0) < 0) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
190 | { |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
191 | purple_debug_warning("nat-pmp", "sysctl: net.route.0.0.dump"); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
192 | return NULL; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
193 | } |
| 15970 | 194 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
195 | lim = buf + needed; |
| 15970 | 196 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
197 | for (next = buf; next < lim; next += rtm->rtm_msglen) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
198 | { |
|
16014
9ee607a29a8c
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 <evands@pidgin.im>
parents:
15984
diff
changeset
|
199 | rtm = (struct rt_msghdr *)next; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
200 | sa = (struct sockaddr *)(rtm + 1); |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
201 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
202 | if (sa->sa_family == AF_INET) |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
203 | { |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
204 | sin = (struct sockaddr_in*) sa; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
205 | |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
206 | if ((rtm->rtm_flags & RTF_GATEWAY) && sin->sin_addr.s_addr == INADDR_ANY) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
207 | { |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
208 | /* We found the default route. Now get the destination address and netmask. */ |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
209 | struct sockaddr *rti_info[RTAX_MAX]; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
210 | struct sockaddr addr, mask; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
211 | |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
212 | get_rtaddrs(rtm->rtm_addrs, sa, rti_info); |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
213 | memset(&addr, 0, sizeof(addr)); |
| 15970 | 214 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
215 | if (rtm->rtm_addrs & RTA_DST) |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
216 | memcpy(&addr, rti_info[RTAX_DST], sizeof(addr)); |
| 15970 | 217 | |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
218 | memset(&mask, 0, sizeof(mask)); |
| 15970 | 219 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
220 | if (rtm->rtm_addrs & RTA_NETMASK) |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
221 | memcpy(&mask, rti_info[RTAX_NETMASK], sizeof(mask)); |
| 15970 | 222 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
223 | if (rtm->rtm_addrs & RTA_GATEWAY && |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
224 | is_default_route(&addr, &mask)) |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
225 | { |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
226 | if (rti_info[RTAX_GATEWAY]) { |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
227 | struct sockaddr_in *rti_sin = (struct sockaddr_in *)rti_info[RTAX_GATEWAY]; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
228 | sin = g_new0(struct sockaddr_in, 1); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
229 | sin->sin_family = rti_sin->sin_family; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
230 | sin->sin_port = rti_sin->sin_port; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
231 | sin->sin_addr.s_addr = rti_sin->sin_addr.s_addr; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
232 | memcpy(sin, rti_info[RTAX_GATEWAY], sizeof(struct sockaddr_in)); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
233 | |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
234 | purple_debug_info("nat-pmp", "found a default gateway"); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
235 | found = TRUE; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
236 | break; |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
237 | } |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
238 | } |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
239 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
240 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
241 | } |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
242 | |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
243 | return (found ? sin : NULL); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
244 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
245 | |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
246 | /*! |
| 15884 | 247 | * purple_pmp_get_public_ip() will return the publicly facing IP address of the |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
248 | * default NAT gateway. The function will return NULL if: |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
249 | * - The gateway doesn't support NAT-PMP |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
250 | * - The gateway errors in some other spectacular fashion |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
251 | */ |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
252 | char * |
| 15884 | 253 | purple_pmp_get_public_ip() |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
254 | { |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
255 | struct sockaddr_in addr, *gateway, *publicsockaddr = NULL; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
256 | struct timeval req_timeout; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
257 | socklen_t len; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
258 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
259 | PurplePmpIpRequest req; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
260 | PurplePmpIpResponse resp; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
261 | int sendfd; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
262 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
263 | if (pmp_info.status == PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER) |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
264 | return NULL; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
265 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
266 | if ((pmp_info.status == PURPLE_PMP_STATUS_DISCOVERED) && (pmp_info.publicip != NULL)) |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
267 | { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
268 | #ifdef PMP_DEBUG |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
269 | purple_debug_info("nat-pmp", "Returning cached publicip %s",pmp_info.publicip); |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
270 | #endif |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
271 | return pmp_info.publicip; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
272 | } |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
273 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
274 | gateway = default_gw(); |
| 15970 | 275 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
276 | if (!gateway) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
277 | { |
| 15884 | 278 | purple_debug_info("nat-pmp", "Cannot request public IP from a NULL gateway!\n"); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
279 | /* If we get a NULL gateway, don't try again next time */ |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
280 | pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
281 | return NULL; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
282 | } |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
283 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
284 | /* Default port for NAT-PMP is 5351 */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
285 | if (gateway->sin_port != PMP_PORT) |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
286 | gateway->sin_port = htons(PMP_PORT); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
287 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
288 | req_timeout.tv_sec = 0; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
289 | req_timeout.tv_usec = PMP_TIMEOUT; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
290 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
291 | sendfd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); |
| 15970 | 292 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
293 | /* Clean out both req and resp structures */ |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
294 | memset(&req, 0, sizeof(PurplePmpIpRequest)); |
|
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
295 | memset(&resp, 0, sizeof(PurplePmpIpResponse)); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
296 | req.version = 0; |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
297 | req.opcode = 0; |
| 15970 | 298 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
299 | /* The NAT-PMP spec says we should attempt to contact the gateway 9 times, doubling the time we wait each time. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
300 | * Even starting with a timeout of 0.1 seconds, that means that we have a total waiting of 204.6 seconds. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
301 | * With the recommended timeout of 0.25 seconds, we're talking 511.5 seconds (8.5 minutes). |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
302 | * |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
303 | * This seems really silly... if this were nonblocking, a couple retries might be in order, but it's not at present. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
304 | */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
305 | #ifdef PMP_DEBUG |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
306 | purple_debug_info("nat-pmp", "Attempting to retrieve the public ip address for the NAT device at: %s\n", inet_ntoa(gateway->sin_addr)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
307 | purple_debug_info("nat-pmp", "\tTimeout: %ds %dus\n", req_timeout.tv_sec, req_timeout.tv_usec); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
308 | #endif |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
309 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
310 | /* TODO: Non-blocking! */ |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
311 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
312 | if (sendto(sendfd, &req, sizeof(req), 0, (struct sockaddr *)(gateway), sizeof(struct sockaddr)) < 0) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
313 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
314 | purple_debug_info("nat-pmp", "There was an error sending the NAT-PMP public IP request! (%s)\n", strerror(errno)); |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
315 | g_free(gateway); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
316 | pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
317 | return NULL; |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
318 | } |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
319 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
320 | if (setsockopt(sendfd, SOL_SOCKET, SO_RCVTIMEO, &req_timeout, sizeof(req_timeout)) < 0) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
321 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
322 | purple_debug_info("nat-pmp", "There was an error setting the socket's options! (%s)\n", strerror(errno)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
323 | g_free(gateway); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
324 | pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
325 | return NULL; |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
326 | } |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
327 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
328 | /* TODO: Non-blocking! */ |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
329 | len = sizeof(struct sockaddr_in); |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
330 | if (recvfrom(sendfd, &resp, sizeof(PurplePmpIpResponse), 0, (struct sockaddr *)(&addr), &len) < 0) |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
331 | { |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
332 | if (errno != EAGAIN) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
333 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
334 | purple_debug_info("nat-pmp", "There was an error receiving the response from the NAT-PMP device! (%s)\n", strerror(errno)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
335 | g_free(gateway); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
336 | pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
337 | return NULL; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
338 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
339 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
340 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
341 | if (addr.sin_addr.s_addr == gateway->sin_addr.s_addr) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
342 | publicsockaddr = &addr; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
343 | else |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
344 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
345 | purple_debug_info("nat-pmp", "Response was not received from our gateway! Instead from: %s\n", inet_ntoa(addr.sin_addr)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
346 | g_free(gateway); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
347 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
348 | pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
349 | return NULL; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
350 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
351 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
352 | if (!publicsockaddr) { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
353 | g_free(gateway); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
354 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
355 | pmp_info.status = PURPLE_PMP_STATUS_UNABLE_TO_DISCOVER; |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
356 | return NULL; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
357 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
358 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
359 | #ifdef PMP_DEBUG |
| 15884 | 360 | purple_debug_info("nat-pmp", "Response received from NAT-PMP device:\n"); |
| 361 | purple_debug_info("nat-pmp", "version: %d\n", resp.version); | |
| 362 | purple_debug_info("nat-pmp", "opcode: %d\n", resp.opcode); | |
| 363 | purple_debug_info("nat-pmp", "resultcode: %d\n", ntohs(resp.resultcode)); | |
| 364 | purple_debug_info("nat-pmp", "epoch: %d\n", ntohl(resp.epoch)); | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
365 | struct in_addr in; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
366 | in.s_addr = resp.address; |
| 15884 | 367 | purple_debug_info("nat-pmp", "address: %s\n", inet_ntoa(in)); |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
368 | #endif |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
369 | |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
370 | publicsockaddr->sin_addr.s_addr = resp.address; |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
371 | |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
372 | g_free(gateway); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
373 | |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
374 | g_free(pmp_info.publicip); |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
375 | pmp_info.publicip = g_strdup(inet_ntoa(publicsockaddr->sin_addr)); |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
376 | pmp_info.status = PURPLE_PMP_STATUS_DISCOVERED; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
377 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
378 | return inet_ntoa(publicsockaddr->sin_addr); |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
379 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
380 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
381 | gboolean |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
382 | purple_pmp_create_map(PurplePmpType type, unsigned short privateport, unsigned short publicport, int lifetime) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
383 | { |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
384 | struct sockaddr_in *gateway; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
385 | gboolean success = TRUE; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
386 | int sendfd; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
387 | struct timeval req_timeout; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
388 | PurplePmpMapRequest req; |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
389 | PurplePmpMapResponse *resp; |
| 15970 | 390 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
391 | gateway = default_gw(); |
| 15970 | 392 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
393 | if (!gateway) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
394 | { |
| 15884 | 395 | purple_debug_info("nat-pmp", "Cannot create mapping on a NULL gateway!\n"); |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
396 | return FALSE; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
397 | } |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
398 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
399 | /* Default port for NAT-PMP is 5351 */ |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
400 | if (gateway->sin_port != PMP_PORT) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
401 | gateway->sin_port = htons(PMP_PORT); |
| 15970 | 402 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
403 | resp = g_new0(PurplePmpMapResponse, 1); |
| 15970 | 404 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
405 | req_timeout.tv_sec = 0; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
406 | req_timeout.tv_usec = PMP_TIMEOUT; |
| 15970 | 407 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
408 | sendfd = socket(AF_INET, SOCK_DGRAM, IPPROTO_UDP); |
| 15970 | 409 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
410 | /* Set up the req */ |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
411 | memset(&req, 0, sizeof(PurplePmpMapRequest)); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
412 | req.version = 0; |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
413 | req.opcode = ((type == PURPLE_PMP_TYPE_UDP) ? PMP_MAP_OPCODE_UDP : PMP_MAP_OPCODE_TCP); |
| 16413 | 414 | req.privateport = htons(privateport); /* What a difference byte ordering makes...d'oh! */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
415 | req.publicport = htons(publicport); |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
416 | req.lifetime = htonl(lifetime); |
| 15970 | 417 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
418 | /* The NAT-PMP spec says we should attempt to contact the gateway 9 times, doubling the time we wait each time. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
419 | * Even starting with a timeout of 0.1 seconds, that means that we have a total waiting of 204.6 seconds. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
420 | * With the recommended timeout of 0.25 seconds, we're talking 511.5 seconds (8.5 minutes). |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
421 | * |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
422 | * This seems really silly... if this were nonblocking, a couple retries might be in order, but it's not at present. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
423 | * XXX Make this nonblocking. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
424 | * XXX This code looks like the pmp_get_public_ip() code. Can it be consolidated? |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
425 | */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
426 | #ifdef PMP_DEBUG |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
427 | purple_debug_info("nat-pmp", "Attempting to create a NAT-PMP mapping the private port %d, and the public port %d\n", privateport, publicport); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
428 | purple_debug_info("nat-pmp", "\tTimeout: %ds %dus\n", req_timeout.tv_sec, req_timeout.tv_usec); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
429 | #endif |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
430 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
431 | /* TODO: Non-blocking! */ |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
432 | success = (sendto(sendfd, &req, sizeof(req), 0, (struct sockaddr *)(gateway), sizeof(struct sockaddr)) >= 0); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
433 | if (!success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
434 | purple_debug_info("nat-pmp", "There was an error sending the NAT-PMP mapping request! (%s)\n", strerror(errno)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
435 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
436 | if (success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
437 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
438 | success = (setsockopt(sendfd, SOL_SOCKET, SO_RCVTIMEO, &req_timeout, sizeof(req_timeout)) >= 0); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
439 | if (!success) |
| 15884 | 440 | purple_debug_info("nat-pmp", "There was an error setting the socket's options! (%s)\n", strerror(errno)); |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
441 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
442 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
443 | if (success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
444 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
445 | /* The original code treats EAGAIN as a reason to iterate.. but I've removed iteration. This may be a problem */ |
|
15961
a3ef09c8ca24
More complete namespacing of nat-pmp
Mark Doliner <markdoliner@pidgin.im>
parents:
15933
diff
changeset
|
446 | /* TODO: Non-blocking! */ |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
447 | success = ((recvfrom(sendfd, resp, sizeof(PurplePmpMapResponse), 0, NULL, NULL) >= 0) || |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
448 | (errno == EAGAIN)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
449 | if (!success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
450 | purple_debug_info("nat-pmp", "There was an error receiving the response from the NAT-PMP device! (%s)\n", strerror(errno)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
451 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
452 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
453 | if (success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
454 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
455 | success = (resp->opcode == (req.opcode + 128)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
456 | if (!success) |
| 15884 | 457 | purple_debug_info("nat-pmp", "The opcode for the response from the NAT device does not match the request opcode!\n"); |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
458 | } |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
459 | |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
460 | #ifdef PMP_DEBUG |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
461 | if (success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
462 | { |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
463 | purple_debug_info("nat-pmp", "Response received from NAT-PMP device:\n"); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
464 | purple_debug_info("nat-pmp", "version: %d\n", resp->version); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
465 | purple_debug_info("nat-pmp", "opcode: %d\n", resp->opcode); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
466 | purple_debug_info("nat-pmp", "resultcode: %d\n", ntohs(resp->resultcode)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
467 | purple_debug_info("nat-pmp", "epoch: %d\n", ntohl(resp->epoch)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
468 | purple_debug_info("nat-pmp", "privateport: %d\n", ntohs(resp->privateport)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
469 | purple_debug_info("nat-pmp", "publicport: %d\n", ntohs(resp->publicport)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
470 | purple_debug_info("nat-pmp", "lifetime: %d\n", ntohl(resp->lifetime)); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
471 | } |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
472 | #endif |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
473 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
474 | g_free(resp); |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
475 | g_free(gateway); |
|
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
476 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
477 | /* XXX The private port may actually differ from the one we requested, according to the spec. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
478 | * We don't handle that situation at present. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
479 | * |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
480 | * TODO: Look at the result and verify it matches what we wanted; either return a failure if it doesn't, |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
481 | * or change network.c to know what to do if the desired private port shifts as a result of the nat-pmp operation. |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
482 | */ |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
483 | return success; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
484 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
485 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
486 | gboolean |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
487 | purple_pmp_destroy_map(PurplePmpType type, unsigned short privateport) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
488 | { |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
489 | gboolean success; |
| 15970 | 490 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
491 | success = purple_pmp_create_map(((type == PURPLE_PMP_TYPE_UDP) ? PMP_MAP_OPCODE_UDP : PMP_MAP_OPCODE_TCP), |
|
15933
913ba1d4b81c
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 <evands@pidgin.im>
parents:
15884
diff
changeset
|
492 | privateport, 0, 0); |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
493 | if (!success) |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
494 | purple_debug_warning("nat-pmp", "Failed to properly destroy mapping for %d!\n", privateport); |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
495 | |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
496 | return success; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
497 | } |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
498 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
499 | static void |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
500 | purple_pmp_network_config_changed_cb(void *data) |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
501 | { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
502 | pmp_info.status = PURPLE_PMP_STATUS_UNDISCOVERED; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
503 | g_free(pmp_info.publicip); |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
504 | pmp_info.publicip = NULL; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
505 | } |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
506 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
507 | static void* |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
508 | purple_pmp_get_handle(void) |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
509 | { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
510 | static int handle; |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
511 | |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
512 | return &handle; |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
513 | } |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
514 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
515 | void |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
516 | purple_pmp_init() |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
517 | { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
518 | purple_signal_connect(purple_network_get_handle(), "network-configuration-changed", |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
519 | purple_pmp_get_handle(), PURPLE_CALLBACK(purple_pmp_network_config_changed_cb), |
|
19387
8dc4480381f3
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 <datallah@pidgin.im>
parents:
16965
diff
changeset
|
520 | GINT_TO_POINTER(0)); |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
521 | } |
|
16014
9ee607a29a8c
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 <evands@pidgin.im>
parents:
15984
diff
changeset
|
522 | #else /* #ifdef NET_RT_DUMP */ |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
523 | char * |
| 15884 | 524 | purple_pmp_get_public_ip() |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
525 | { |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
526 | return NULL; |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
527 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
528 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
529 | gboolean |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
530 | purple_pmp_create_map(PurplePmpType type, unsigned short privateport, unsigned short publicport, int lifetime) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
531 | { |
|
15984
45e6793c221c
These should return FALSE, not NULL, when NET_RT_DUMP2 isn't available; they return a gboolean, not a pointer.
Evan Schoenberg <evands@pidgin.im>
parents:
15970
diff
changeset
|
532 | return FALSE; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
533 | } |
|
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
534 | |
|
15968
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
535 | gboolean |
|
01dcc5c7cf9e
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 <evands@pidgin.im>
parents:
15961
diff
changeset
|
536 | purple_pmp_destroy_map(PurplePmpType type, unsigned short privateport) |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
537 | { |
|
15984
45e6793c221c
These should return FALSE, not NULL, when NET_RT_DUMP2 isn't available; they return a gboolean, not a pointer.
Evan Schoenberg <evands@pidgin.im>
parents:
15970
diff
changeset
|
538 | return FALSE; |
|
15708
979f43512119
Added nat-pmp implementation and #ifdef'd out changes to network.c which would utilize it.
Evan Schoenberg <evands@pidgin.im>
parents:
diff
changeset
|
539 | } |
|
16035
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
540 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
541 | void |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
542 | purple_pmp_init() |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
543 | { |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
544 | |
|
eb4eda79a147
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 <evands@pidgin.im>
parents:
16014
diff
changeset
|
545 | } |
|
16965
d3df125b3196
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 <evands@pidgin.im>
parents:
16964
diff
changeset
|
546 | #endif /* #if !(defined(HAVE_SYS_SYCTL_H) && defined(NET_RT_DUMP)) */ |