diff options
author | 2014-11-08 15:21:02 +0000 | |
---|---|---|
committer | 2014-11-08 15:21:02 +0000 | |
commit | 46f2cc82e9416c20e7f171a0d5603239f3575dcc (patch) | |
tree | 017bb52dd6b7d1233bf524527a14624bb3942f9e /lib/libssl/ssl_sess.c | |
parent | some systems install sensors where numbering introduces a gap which (diff) | |
download | wireguard-openbsd-46f2cc82e9416c20e7f171a0d5603239f3575dcc.tar.xz wireguard-openbsd-46f2cc82e9416c20e7f171a0d5603239f3575dcc.zip |
Clean up more SSLv2 remnants.
Diffstat (limited to 'lib/libssl/ssl_sess.c')
-rw-r--r-- | lib/libssl/ssl_sess.c | 19 |
1 files changed, 9 insertions, 10 deletions
diff --git a/lib/libssl/ssl_sess.c b/lib/libssl/ssl_sess.c index d76fb8b9c85..8f6032a491c 100644 --- a/lib/libssl/ssl_sess.c +++ b/lib/libssl/ssl_sess.c @@ -1,4 +1,4 @@ -/* $OpenBSD: ssl_sess.c,v 1.42 2014/10/18 16:13:16 jsing Exp $ */ +/* $OpenBSD: ssl_sess.c,v 1.43 2014/11/08 15:21:02 jsing Exp $ */ /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com) * All rights reserved. * @@ -238,15 +238,14 @@ SSL_SESSION_get_compress_id(const SSL_SESSION *s) } /* - * Even with SSLv2, we have 16 bytes (128 bits) of session ID space. - * SSLv3/TLSv1 has 32 bytes (256 bits). As such, filling the ID with random - * gunk repeatedly until we have no conflict is going to complete in one - * iteration pretty much "most" of the time (btw: understatement). So, if it - * takes us 10 iterations and we still can't avoid a conflict - well that's a - * reasonable point to call it quits. Either the RAND code is broken or someone - * is trying to open roughly very close to 2^128 (or 2^256) SSL sessions to our - * server. How you might store that many sessions is perhaps a more interesting - * question... + * SSLv3/TLSv1 has 32 bytes (256 bits) of session ID space. As such, filling + * the ID with random gunk repeatedly until we have no conflict is going to + * complete in one iteration pretty much "most" of the time (btw: + * understatement). So, if it takes us 10 iterations and we still can't avoid + * a conflict - well that's a reasonable point to call it quits. Either the + * arc4random code is broken or someone is trying to open roughly very close to + * 2^128 (or 2^256) SSL sessions to our server. How you might store that many + * sessions is perhaps a more interesting question... */ #define MAX_SESS_ID_ATTEMPTS 10 |