[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-bugs] #3786 [Tor Relay]: Make clients and bridges use their IPv6 address



#3786: Make clients and bridges use their IPv6 address
-----------------------+----------------------------------------------------
 Reporter:  ln5        |          Owner:  ln5               
     Type:  task       |         Status:  needs_review      
 Priority:  normal     |      Milestone:  Tor: 0.2.3.x-final
Component:  Tor Relay  |        Version:                    
 Keywords:             |         Parent:  #3563             
   Points:             |   Actualpoints:                    
-----------------------+----------------------------------------------------

Comment(by ln5):

 Replying to [comment:12 nickm]:
 > In extend_info_from_* , I think that use_pref_addr is the wrong name for
 the argument; I think instead it should be "for_direct_connect" or
 something.  That way, the caller only needs to know what they want to do
 with the extend_info; not what their needs imply about how the extend_info
 needs to be generated.

 Good point.  Fixed in eed5d661.


 > The test in circuit_get_open_circ_or_launch looks right to me: in the
 case of a onehop circuit, the extend_info for the exit node you want
 should be the one you're using to connect to it.

 Great.  Comment removed in eadc5594.


 > If node_get_all_orports() is at all near the critical path, we should
 think about some way to make it faster: allocating and freeing little
 objects over and over can get expensive.

 Would you suggest node_get_all_orports() taking a smartlist pointer as an
 argument?  And even two (optional) tor_addr_t pointers?


 > Have a look at the callers of node_get_{prim,perf}_{addr,orport} : would
 it make more sense or less sense to make these functions yield a
 tor_addr_port_t instead?

 Yes, that's what I intend to do next.


 > In get_configured_bridge_by_routerinfo: if there is no bridge configured
 with the preferred address, should we fall back and check whether there is
 one configured with the primary address?  Or did I misunderstand
 something?

 We should check "the other" address.  I totally missed that.
 More generally, we should check all addresses, preferring "the preferred".


 > Otherwise looks okay, I think.  If you think I should merge after the
 next round, please include a changes file, and let me know how tested this
 is/isn't at that point.

 I'll squash {node,router}_get_{prim,alt,perf}_addr with their _orport
 counterparts and then ask for merge.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/3786#comment:13>
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