Comments on: SSL MitM Attack, Part 2

By: securitywhispers » Blog Archive » Coverage story of TLS blind prefix injection attack

Tue, 10 Nov 2009 12:48:52 +0000

[...] SSL MitM Attack, Part 2 [...]

By: Pat

Pat — Mon, 09 Nov 2009 09:27:06 +0000

do you have a defence to the mopnkey comment?

By: David-Sarah Hopwood

David-Sarah Hopwood — Sat, 07 Nov 2009 04:10:04 +0000

https://svn.resiprocate.org/rep/ietf-drafts/ekr/draft-rescorla-tls-renegotiate.txt

What irony that it takes three clicks to dismiss the silly dialog telling me that the connection is insecure (which I did pretty much on autopilot, given that broken certs are so common).

Incidentally, the attacker can have control over some headers in some CSRF attacks. The restrictions are complicated.

By: m

Fri, 06 Nov 2009 21:05:51 +0000

i’m guessing that the guy referring to ISA and OWA are referring to Microsoft Internet Security & Acceleration Server and Microsoft Outlook Web Access.

By: Alex Lam

Alex Lam — Fri, 06 Nov 2009 19:38:25 +0000

Thanks for confirming that session resumption is not affected by this patch.

More of an implementation question, are we vulnerable if the attacker hijacks the client’s session resumption handshakes instead of the initial one?
Does OpenSSL mandate a ClientHello.session_id check against the existing session’s, or does it blindly retrieve the master secret based on the
ClientHello.session_id?

By: Benson

Benson — Fri, 06 Nov 2009 17:42:10 +0000

I just want to say this because you’re probably mostly hearing people bitching right now: You rock. OpenSSL is a really important piece of software, and I really appreciate the work you and the rest of the team do to make it work.

By: Mike

Mike — Fri, 06 Nov 2009 14:21:41 +0000

“Though the fact that this attack doesn’t actually make HTTP much worse is a pretty damning indictment of HTTP (and HTML)!”

what does that even mean? please explain how any of what you were just talking about is an indictment of HTTP..

By: Clement

Clement — Fri, 06 Nov 2009 12:53:28 +0000

Since when does CSRF require clicking on links? All I need is for your browser to render out the image tag that I insert into some straight-HTTP connection that I MITM, or I can insert JS if I really want POST data.

And are there any other protocols that use SSL which are vulnerable to this particular corner case?