[tor-bugs] #21302 [Core Tor/Tor]: hs: Multiple service issues

[Tor Bug Tracker & Wiki]

#21302: hs: Multiple service issues
--------------------------+------------------------------------
 Reporter:  dgoulet       |          Owner:  dgoulet
     Type:  defect        |         Status:  needs_review
 Priority:  Medium        |      Milestone:  Tor: 0.3.0.x-final
Component:  Core Tor/Tor  |        Version:  Tor: 0.2.9.8
 Severity:  Normal        |     Resolution:
 Keywords:  tor-hs        |  Actual Points:
Parent ID:                |         Points:
 Reviewer:                |        Sponsor:
--------------------------+------------------------------------

Comment (by dgoulet):

 Replying to [comment:3 nickm]:
 > In a82e572e9ff5994d2ba14dcc1a25995f6bdda9b7, how do we know that
 count_intro_point_circuits() is always greater than expiring_nodes_len?
 In other words, how can we be sure that this subtraction never underflows:
 > {{{
 > +  num_ip_circuits = count_intro_point_circuits(service) -
 expiring_nodes_len;
 > }}}

 Great observation. In *theory*, if a node is in the expiring node list it
 means it has a circuit. This is why the third commit actually improves on
 that and could be a cause of the unsolved #21297. However, I don't believe
 we have a good guarantee of this. Real guarantee would be that when we
 close an HS circuits, we also make sure anything expecting a circuit is
 cleaned up.

 How is this fixup commit looking? `c1f6238c`.

 >
 > The other commits look okay to me, if you've tested them.  (We've
 actually removed something with the loop in
 0342c1a82c56ad1f9ae9b8a536f32355640dc4bd, yeah?)

 Yes it has been tested. Would be nice to have asn's eyes on that as well.
 It is quite core in the HS subsystem but worth fixing!

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/21302#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online