DTLS cookies rendered useless by SSL_accept behavior

2011-07-23 Thread com...@gmx.ch
Hi, I got some code which uses nonblocking dtls via self fed bios. If a unknown session comes up, I create a new session, feed/drain the sessions bios from/to the wire until SSL_accept returns success. Now, I added DTLS cookies. I expected SSL_accept to return an error if a ClientHello packet

Re: DTLS cookies rendered useless by SSL_accept behavior

2011-07-23 Thread com...@gmx.ch
On 07/24/2011 03:55 AM, Michael Tuexen wrote: On Jul 24, 2011, at 12:40 AM, com...@gmx.ch wrote: I got some code which uses nonblocking dtls via self fed bios. You are supposed to call DTLSv1_listen() to handle the front state... #define DTLSv1_listen(ssl, peer) \ SSL_ctrl(ssl

Re: DTLS cookies rendered useless by SSL_accept behavior

2011-07-24 Thread com...@gmx.ch
Hi, good news, I got this working - but it requires either openssl 1.0.0d with patches from http://sctp.fh-muenster.de/dtls-patches.html - I think the Sequence Numbers (#2555) should be enough - or OpenSSL CVS, some of the dtls patches got merged recently :). I set a flag on the session in t

OpenSSL Engine - configurable ciphers/digests

2011-10-20 Thread com...@gmx.ch
Hi, I need some help with a special case: a dynamic engine with non-static or configureable ciphers. While I do not use cryptodev, the code provides a good example of the intial problem: http://cvs.openssl.org/fileview?f=openssl/crypto/engine/eng_cryptodev.c&v=1.23 I basically have it wor