[tor-bugs] #17254 [Tor]: Scalable HSes by splitting intro/rendezvous

Tor Bug Tracker & Wiki blackhole at torproject.org
Mon Nov 23 18:25:24 UTC 2015


#17254: Scalable HSes by splitting intro/rendezvous
-------------------------+------------------------------------
 Reporter:  TvdW         |          Owner:
     Type:  enhancement  |         Status:  needs_revision
 Priority:  High         |      Milestone:  Tor: 0.2.8.x-final
Component:  Tor          |        Version:
 Severity:  Normal       |     Resolution:
 Keywords:  tor-hs       |  Actual Points:
Parent ID:               |         Points:  medium
  Sponsor:  SponsorR     |
-------------------------+------------------------------------

Comment (by TvdW):

 > I took a small peak at the code, and it looks nice!

 Thanks!

 > Here are some things that crossed my mind in this initial review:
 >
 > - Some places are lacking documentation. e.g.
 `rend_service_handoff_introduce()`

 You're right. I'll add some

 > - It would be really great if there was some unittest that verified the
 correctness of the handoff-then-perform procedure.

 Haven't figured out how to do this yet :-( I suppose I could make
 `rend_service_perform_rendezvous()` mock-able, and then call the
 `rend_service_handoff_introduce()`/`handle_control_perform_rendezvous()`
 combo. That wouldn't really test the behavior though, which is something I
 think only Chutney could test.

 > - As you noted with your XXX, in case of perform failure it would be
 better if we send a 550 or 551 control failure instead of
 `send_control_done(conn)`.

 Patched and pushed!

 > - `uint8_t buffer[2048];` I wonder if this stack buffer needs to be so
 big. Or maybe this is premature optimization and we should just move on.

 Allocating 2k on the stack is essentially free. `memwipe()` is not (I
 forgot a few; patched those just now) but it's still cheap compared to the
 other code. I've changed the stack size to 1k, as that should also be
 fine, though I'd rather not 'optimize' that further as it seems premature.

 > Will review more seriously after the spec branch is also updated!

 Thanks! I'll push something later this week.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/17254#comment:21>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list