| … | |
… | |
| 3 | TODO: bridge mode, finally? |
3 | TODO: bridge mode, finally? |
| 4 | TODO: gcm mode? |
4 | TODO: gcm mode? |
| 5 | TODO: replace ripemd160 as the only authentication hash. |
5 | TODO: replace ripemd160 as the only authentication hash. |
| 6 | TODO: increase rsa size. |
6 | TODO: increase rsa size. |
| 7 | TODO: replace transport bits by transport endpoint structs? |
7 | TODO: replace transport bits by transport endpoint structs? |
| 8 | TODO: ecdh to avoid session replay attacks |
|
|
| 9 | TODO: http://incog-izick.blogspot.de/2011/08/using-openssl-aes-gcm.html |
8 | TODO: http://incog-izick.blogspot.de/2011/08/using-openssl-aes-gcm.html |
| 10 | TODO: http://stackoverflow.com/questions/12153009/openssl-c-example-of-aes-gcm-using-evp-interfaces |
9 | TODO: http://stackoverflow.com/questions/12153009/openssl-c-example-of-aes-gcm-using-evp-interfaces |
| 11 | |
10 | |
| 12 | proposed: 3 types, req, resreq, res |
11 | TODO: verify |
| 13 | |
12 | TODO: make sense of overhead calculation |
| 14 | req (hmac1) rsa(seqno1 hmac1 aes1 seqno2 hmac2 aes2 auth) ecdh1 |
13 | TODO: if-up &c should not be scripts? |
| 15 | res (hmac1) hash(rsa-contents) ecdh2 |
14 | TODO: ipv6 |
| 16 | |
15 | TODO: gvpectrl should not use default privatekey,. or maybe document it better |
| 17 | req hmac0(*) rsa(seqno hmac0 hmac aes auth) hkdf-salt ecdh1 |
|
|
| 18 | res hmac0(rsa-contents ecdh2) |
|
|
| 19 | hmac_key = hkdf(hkdf-salt, hmac | ecdh) |
|
|
| 20 | aes_key = hkdf(hkdf-salt, aes | ecdh) |
|
|
| 21 | |
|
|
| 22 | - INCOMPATIBLE CHANGE: core protocol version 1.0. |
16 | - INCOMPATIBLE CHANGE: core protocol version 1.0. |
|
|
17 | - INCOMPATIBLE CHANGE: node sections are now introduced |
|
|
18 | with "node nodename", not "node = nodename". |
|
|
19 | - add workaround for temporary/rare ENOBUFS condition. |
| 23 | - while individual packets couldn't be replayed, a whole session |
20 | - while individual packets couldn't be replayed, a whole session |
| 24 | could be replayed - this has been fixed by an extra key exchange. |
21 | could be replayed - this has been fixed by an extra key exchange. |
|
|
22 | - fix a delete vs. delete [] mismatch in the central logging function. |
| 25 | - in addition to rsa key exchange and authentication, the handshake now |
23 | - in addition to rsa key exchange and authentication, the handshake now |
| 26 | adds a diffie-hellman key exchange (using curve25119) for perfect |
24 | adds a diffie-hellman key exchange (using curve25119) for perfect |
| 27 | forward secrecy. mac and cipher keys are derived using HKDF. |
25 | forward secrecy. mac and cipher keys are derived using HKDF. |
| 28 | - rsa key sizes are now configurable and larger (default is 3072). |
26 | - rsa key sizes are now configurable and larger (default is 3072). |
| 29 | correspondingly, the minimum mtu is no longer 296 but 576. |
27 | correspondingly, the minimum mtu is no longer 296 but 576. |
| 30 | - fixed a potential (unverified) buffer overrun on rsa decryption. |
28 | - fixed a potential (unverified) buffer overrun on rsa decryption. |
| 31 | - RAND_bytes was used when generating session keys, potentially |
29 | - new per-node low-power setting that tries to reduce cpu/network usage. |
| 32 | leading to temporary freezes when entropy was low. |
30 | - router reconnects could cause excessive rekeying on other connections. |
| 33 | - gvpectrl no longer generates all missing public keys, but |
31 | - gvpectrl no longer generates all missing public keys, but |
| 34 | only missing private keys. private keys are also put |
32 | only missing private keys. private keys are also put |
| 35 | into the configured location. |
33 | into the configured location. |
| 36 | - the pid-file now accepts %s as nodename as elsewhere. |
34 | - the pid-file now accepts %s as nodename as elsewhere. |
| 37 | |
35 | - switch to counter mode (only aes supported at the moment in |
|
|
36 | openssl). this gets rid of the need to generate a random iv, |
|
|
37 | is likely more secure (and, as a side effect, gets rid of |
|
|
38 | slow randomness generation. counter mode is often faster |
|
|
39 | then cbc mode as well, and packets are smaller). |
|
|
40 | - no longer use RAND_bytes to generate session keys - you NEED |
|
|
41 | a real source of entropy now (e.g. egd or /dev/random - see the |
|
|
42 | openssl documentation). |
|
|
43 | - multiple node statements for the same node are now supported |
|
|
44 | and will be merged. |
|
|
45 | - a new directive "global" switches back to the global section |
|
|
46 | of the config file. |
|
|
47 | - if-up scripts can now be specified with absolute paths. |
|
|
48 | - new global option: serial, to detect configuration mismatches. |
|
|
49 | - use HKDF as authentication proof, not HMAC or a plain hash |
|
|
50 | (hint by Ilmari Karonen). |
|
|
51 | - during rekeying or conenction establishments, hmac authentication |
|
|
52 | errors could occur and reset the connection. Transient hmac |
|
|
53 | authentication errors are now being ignored for 3 seconds. |
|
|
54 | - log the reason for a conneciton loss. |
|
|
55 | - use a (hopefully) constant time memcmp to compare internal secrets. |
|
|
56 | - fix a (harmless) errornous out of bounds stack read that would trigger |
|
|
57 | gcc's -fsanitize=address. |
|
|
58 | |
| 38 | 2.25 Sat Jul 13 06:42:33 CEST 2013 |
59 | 2.25 Sat Jul 13 06:42:33 CEST 2013 |
| 39 | - INCOMPATIBLE CHANGE: no longer enable udp protocol if no other |
60 | - INCOMPATIBLE CHANGE: no longer enable udp protocol if no other |
| 40 | protocols are enabled - this is necessary when you have nodes with |
61 | protocols are enabled - this is necessary when you have nodes with |
| 41 | completely unknown protocols, to force mediated connection requests. |
62 | completely unknown protocols, to force mediated connection requests. |
| 42 | - INCOMPATIBLE CHANGE: dns transport protocol bumped to version 2. |
63 | - INCOMPATIBLE CHANGE: dns transport protocol bumped to version 2. |
| … | |
… | |
| 51 | the protocol not being configured locally. |
72 | the protocol not being configured locally. |
| 52 | - new global configuration options: chroot, chuser, chuid, chgid, |
73 | - new global configuration options: chroot, chuser, chuid, chgid, |
| 53 | to chroot to a specified or anonymous new root, and change user id. |
74 | to chroot to a specified or anonymous new root, and change user id. |
| 54 | - new global configuration options seed_device and seed_interval, |
75 | - new global configuration options seed_device and seed_interval, |
| 55 | to configure another device than /dev/urandom for random seeds, |
76 | to configure another device than /dev/urandom for random seeds, |
| 56 | and to configure a regular imterval to reseed the rng. |
77 | and to configure a regular interval to reseed the rng. |
| 57 | - prefer inet_aton over gethostbyname, as the latter is not guaranteed |
78 | - prefer inet_aton over gethostbyname, as the latter is not guaranteed |
| 58 | to "resolve" literal ip addresses. |
79 | to "resolve" literal ip addresses. |
| 59 | - configure didn't detect openssl 1.0 because SHA1_version became private |
80 | - configure didn't detect openssl 1.0 because SHA1_version became private |
| 60 | (patch by TANIGUCHI Takaki). |
81 | (patch by TANIGUCHI Takaki). |
| 61 | - fix a bug where nodes would tell the other side that it supports |
82 | - fix a bug where nodes would tell the other side that it supports |
| … | |
… | |
| 66 | - gvpectrl no longer evaluates any "on" directives. |
87 | - gvpectrl no longer evaluates any "on" directives. |
| 67 | - icmp and rawip protocols weren't upgradable to each other. |
88 | - icmp and rawip protocols weren't upgradable to each other. |
| 68 | - major, but incremental, dns transport improvements: |
89 | - major, but incremental, dns transport improvements: |
| 69 | - do not simply abort in some error cases in the dns transport, |
90 | - do not simply abort in some error cases in the dns transport, |
| 70 | but try to recover. |
91 | but try to recover. |
| 71 | - allow lowercase/uppercase alises for base-n encodings that do |
92 | - allow lowercase/uppercase aliases for base-n encodings that do |
| 72 | not rely on case. |
93 | not rely on case. |
| 73 | - use base26 instead of base22 encoding for dns syn's, and |
94 | - use base26 instead of base22 encoding for dns syn's, and |
| 74 | base36 instead of base22 for headers (saves one byte/packet). |
95 | base36 instead of base22 for headers (saves one byte/packet). |
| 75 | - back off far quicker in dns tunnel when idling - increases |
96 | - back off far quicker in dns tunnel when idling - increases |
| 76 | latency on an idle link somewhat, but avoids hundreds of |
97 | latency on an idle link somewhat, but avoids hundreds of |
