On Sat, 2013-11-23 at 19:26 +0000, David Woodhouse wrote: > > This change removes the protocol-weakening options (e.g., the disabling of > > secure renegotiation, the removal of ECDHE ciphersuites, and the restriction > > to the known to be weak TLS 1.0). > > I don't think we can do that; certainly not unconditionally. Some > servers (or their firewalls) are very picky about what they allow, and > secure renegotiation and TLS > 1.0 have definitely been known to give > immediately connection failures. Could that be conditionally? (or relate that to --pfs option?) btw. Are the issues you refer to related to the bigip firewall and http://www.ietf.org/mail-archive/web/tls/current/msg10423.html There it seems that bigip is interpreting TLS1.0 messages as SSL2.0 if a number of extensions is present. His suggestion was to add a very large TLS extension to avoid that bug (In 3.2.7 I added the %DUMBFW option in priority strings to do just that). regards, Nikos
