diff options
author | Mike Perry <mikeperry-git@fscked.org> | 2013-01-30 17:01:10 -0400 |
---|---|---|
committer | Nick Mathewson <nickm@torproject.org> | 2013-02-01 17:01:12 -0500 |
commit | 173ed05d2f7233371dfcb1ef32a4d95f5096c435 (patch) | |
tree | 946d9b12e5de64060f968083f852ab1d29718618 /src/or/circuituse.c | |
parent | ec90ed4f6d1155d28fde91097e76f574195583c6 (diff) | |
download | tor-173ed05d2f7233371dfcb1ef32a4d95f5096c435.tar.gz tor-173ed05d2f7233371dfcb1ef32a4d95f5096c435.zip |
Clarify state transition and related pathbias comments
Diffstat (limited to 'src/or/circuituse.c')
-rw-r--r-- | src/or/circuituse.c | 8 |
1 files changed, 5 insertions, 3 deletions
diff --git a/src/or/circuituse.c b/src/or/circuituse.c index 48a774352e..cfd41be792 100644 --- a/src/or/circuituse.c +++ b/src/or/circuituse.c @@ -1498,15 +1498,17 @@ circuit_launch_by_extend_info(uint8_t purpose, purpose == CIRCUIT_PURPOSE_C_INTRODUCING) && circ->path_state == PATH_STATE_BUILD_SUCCEEDED) { /* Path bias: Cannibalized rends pre-emptively count as a - * successfully used circ. We don't wait until the extend, - * because the rend point could be malicious. + * successfully built but unused closed circuit. We don't + * wait until the extend (or the close) because the rend + * point could be malicious. * * Same deal goes for client side introductions. Clients * can be manipulated to connect repeatedly to them * (especially web clients). * * If we decide to probe the initial portion of these circs, - * (up to the adversaries final hop), we need to remove this. + * (up to the adversary's final hop), we need to remove this, + * or somehow mark the circuit with a special path state. */ /* This must be called before the purpose change */ |