[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #7802 [Tor]: Remove PathBias use of timestamp_dirty and record usage accounting
#7802: Remove PathBias use of timestamp_dirty and record usage accounting
-----------------------------+----------------------------------------------
Reporter: mikeperry | Owner:
Type: enhancement | Status: needs_review
Priority: normal | Milestone: Tor: 0.2.4.x-final
Component: Tor | Version:
Keywords: MikePerry201301 | Parent: #7755
Points: | Actualpoints: 15
-----------------------------+----------------------------------------------
Comment(by mikeperry):
Replying to [comment:7 nickm]:
> Is there any subset of these that could be done for a minimal version?
This is a 1300 line diff and a 1700 line patch series. It's a lot of
changed code.
Most of the changes are refactoring that should both improve readability
and future-proof the state machine against bitrot.
> If not, what is the most rigorous, specific explanation of what this
code is supposed to do?
The only substantial functional change is to provide separate tunable
limits for the acceptable rate of circuit use failure before we emit log
messages. A circuit use failure is defined as a built circuit for which
any streams detach and that we are also unable to probe successfully.
The separate use failure accounting is actually limited to changes to just
a couple of functions: pathbias_check_use_rate(), and writing two counter
variables to the state file as EntryGuardPathUseBias.
All of the other circuit use state transitions and probing code were
already there. I just cleaned things up to remove dependence upon
timestamp_dirty and improve/guard the state machine transitions, and
refactor some common code into helper functions.
#7804 is my ticket to document all of the path bias stuff in path-
spec.txt.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/7802#comment:8>
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