[tor-bugs] #1795 [Tor Relay]: Prop 174: Optimistic Data for Tor: Server Side

[Tor Bug Tracker & Wiki]

#1795: Prop 174: Optimistic Data for Tor: Server Side
-------------------------+--------------------------------------------------
 Reporter:  nickm        |          Owner:  iang               
     Type:  enhancement  |         Status:  needs_review       
 Priority:  normal       |      Milestone:  Deliverable-Mar2011
Component:  Tor Relay    |        Version:                     
 Keywords:  prop174      |         Parent:  #1849              
   Points:               |   Actualpoints:                     
-------------------------+--------------------------------------------------

Comment(by arma):

 See the optimistic_server branch in my git for a whitespace cleanup
 commit.

 The overall patch looks plausible. I had looked at it a few months ago and
 identified the same issue Nick patched ("if you send 1000 cells to a
 circuit optimistically, it will never send a circuit-level sendme back"),
 but my patch was different -- I had thought to call
 circuit_consider_sending_sendme() from inside
 connection_edge_finished_flushing() (and I guess also from inside
 connection_edge_flushed_some() now that #2756 exists).

 The benefit of the approach Nick took is that we can have optimistic data
 queued for an unbounded number of not-yet-connected streams before any of
 them connects. The benefit of my approach is that we can't. But that said,
 you can queue a bunch of data cells at exit relays in plenty of other
 ways, such as by having actual open streams that never flush, so my
 "benefit" isn't really much use.

 Fine by me to merge.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/1795#comment:30>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online