openssh bugs - Nov 2018 - [Bug 2924] New: Order a limited host keys list in client based on the known hosts

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

            Bug ID: 2924
           Summary: Order a limited host keys list in client based on the
                    known hosts
           Product: Portable OpenSSH
           Version: 7.7p1
          Hardware: Other
                OS: Linux
            Status: NEW
          Keywords: patch
          Severity: enhancement
          Priority: P5
         Component: ssh
          Assignee: unassigned-bugs at mindrot.org
          Reporter: jjelen at redhat.com

Created attachment 3198
  --> https://bugzilla.mindrot.org/attachment.cgi?id=3198&action=edit
possibility to order host keys in client

The HostKeyAlgorithms option in the client has a difference from all
the other algorithm limiting options that should be sorted according to
the list of known hosts available. This works fine out of the box with
default negotiated list, but when one tries to limit (or extend) the
algorithm list to something else than default, the ordering is turned
off and one can simply hit the hostkey changed, even though it did not
change at all (only different one is offered for a connection).

There is attached proposed patch to implement new configuration option
HostKeyAlgorithmsOrder, which will turn on sorting also on the
user-provided list of host keys.

Other possibility to resolve this problem would be to introduce another
configuration option HostKeyAlgorithmsAllow or similar, which would
have this semantics (would be ordered before the algorithm
negotiation).

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

Tomas Mraz <t8m at centrum.cz> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |t8m at centrum.cz

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org

--- Comment #1 from Damien Miller <djm at mindrot.org> ---
I'm kinda loath to add another option. Maybe another magic character,
e.g. HostKeyAlgorithms=:ssh-ed25519,ssh-rsa etc to specify a list
ordered by the known host keys?

(i mean, it's ugly but so it yet another option...)

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

--- Comment #2 from Jakub Jelen <jjelen at redhat.com> ---
Thank you for the suggestion. This would also solve the original
problem. Not sure about the character as ":", but "~" might
work as
"approximate" list?

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

Jakub Jelen <jjelen at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #3198|0                           |1
        is obsolete|                            |

--- Comment #3 from Jakub Jelen <jjelen at redhat.com> ---
Created attachment 3313
  --> https://bugzilla.mindrot.org/attachment.cgi?id=3313&action=edit
Introduce a new modifier for HostKeyAlgorithms to allow ordering

Damien, I rewrote the patch to use the colon prefix notation to
signalize the same. Would it work this way for you?

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

--- Comment #4 from Damien Miller <djm at mindrot.org> ---
OpenSSH 8.2 will enable UpdateHostKeys by default. IMO this goes some
way to avoiding this problem.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

--- Comment #5 from Jakub Jelen <jjelen at redhat.com> ---
Right. After the first successful authentication, the client will learn
all the server host keys and we should be able to validate whatever key
server provides according to our preference.

But getting over the first connection can still problem and it is hard
to guess how long it can take to make sure the users already connected
at least once to the particular host to be safe to roll out this
change.

Therefore I see the UpdateHostKeys as a good addition, but the attached
patch would still significantly simplify the migration path and
decrease user frustration when dealing with failed host key checking.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

jatjasjem at gmail.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jatjasjem at gmail.com

--- Comment #6 from jatjasjem at gmail.com ---
i might be hitting this issue, can someone confirm that this it the 
same one or i might want to open another one?

if i run this on default configuration and accept the RSA key:

    rm ~/.ssh/known_hosts
    ssh user at localhost -oHostKeyAlgorithms=rsa-sha2-512

then this works:

    ssh user at localhost

this also works:

    ssh user at localhost -oHostKeyAlgorithms=rsa-sha2-512,ssh-ed25519

this doesn't work:

    ssh user at localhost -oHostKeyAlgorithms=ssh-ed25519,rsa-sha2-512

now edit ssh_config, setting HostKeyAlgorithms to the *default* value 
from man ssh_config. then this also doesn't work:

    ssh user at localhost

this behavior is very unexpected and at the very least should be
mentioned in the manual?

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=2924

Kenyon Ralph <kenyon at kenyonralph.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kenyon at kenyonralph.com

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.