Re: [tor-bugs] #9925 [Tor]: Directory Authorities can crash client/relay

["Tor Bug Tracker & Wiki" <blackhole@xxxxxxxxxxxxxx>]

#9925: Directory Authorities can crash client/relay
------------------------+--------------------------------------------
     Reporter:  sysrqb  |      Owner:
         Type:  defect  |     Status:  needs_review
     Priority:  normal  |  Milestone:  Tor: 0.2.5.x-final
    Component:  Tor     |    Version:
   Resolution:          |   Keywords:  025-triaged andrea-review-0255
Actual Points:          |  Parent ID:
       Points:          |
------------------------+--------------------------------------------

Comment (by cypherpunks):

 About defense-in-depth by 016c4dcbcddc391515e2fdd0f28832679ab2fae5
 What happen if someday two different sd for different relays will have the
 same digest? Hash collision, non-zero chance?
 What about digest for md, digestmap functions about DIGEST_LEN while
 descriptor_digest for md about DIGEST256_LEN. And descriptor_digest is
 under control of attacker actually, they no need to proof they had onion
 keys. Can attacker to drop (by client) every md-consensus such way?

 And related to this thoughts but another problem perhaps.
 What about two rs with the same digest for md-consensus, in general? Can
 code handle it? (it's another problem perhaps) Attacker could to generate
 descriptor with onion and ntor keys from victim relay, and choose self id
 so every new client (without cached documents) will stuck without chance
 to get victim's keys (md will be assigned to attacker's rs by
 nodelist_add_microdesc) till next consensus update.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/9925#comment:11>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs