Re: ssh-hostkey assertion

[Daniel Miller <bonsaiviking () gmail com>]

On Mon, Aug 25, 2014 at 10:32 PM, Kent Fritz <kfritz () wolfman devio us>
wrote:

> On Mon, Aug 25, 2014 at 06:06:59PM -0500, Daniel Miller wrote:
> > The assertion is triggered by a failure to extract the payload and
> padding
> > lengths from an SSH2 packet. I can't reproduce this, so could you provide
> > either the output with --packet-trace or (preferably) a pcap of the
> > transaction?
>
> I'll send a pcap off-list.
>
> > I couldn't reproduce this with OpenSSH 6.6 on OpenBSD 5.5, which is the
> > latest released version on  http://www.openssh.com. Did you install this
> > from the development tree or is something else going on?
>
> I'm running a recent snapshot, which is probably very close to what will
> be released as 5.6 in November.  I don't have 5.5 installed anywhere (never
> bothered due to OpenSSL bugs).  I suspect that some key exchange algorithm
> got deprecated.
Kent,

Thanks again for this bug report. I fixed it in r33615; you can reproduce
with any OpenSSH server by setting:

KexAlgorithms diffie-hellman-group14-sha1

in your sshd_config. We were assuming (based on the RFC 4253 "MUST be
supported" language) that diffie-hellman-group1-sha1 would always be
supported. I just added the group14 kex method, but this may be solved in
the future by moving to libssh2 for these things instead of doing our own
in Lua. Commit method follows, patch is attached if you are not using our
latest development branch.

r33615 | dmiller | 2014-08-26 08:35:25 -0500 (Tue, 26 Aug 2014) | 9 lines

Support diffie-hellman-group14-sha1 in ssh2.lua

ssh-hostkey was hitting an assertion failure when scanning OpenSSH 6.7.
The cause was a failure to check for a shared KEX algorithm. We just
assumed diffie-hellman-group1-sha1 would be supported, since RFC 4253
says it "MUST be supported". This adds support for group14, which is a
2048-bit modulus; key strength was likely the reason for dropping
group1. A more complete solution would be to support more KEX
algorithms, but that's beyond the scope of the bug report.

Attachment: ssh2.patch
Description:

_______________________________________________
Sent through the dev mailing list
http://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/