Commit Graph

11682 Commits

Author SHA1 Message Date
Nick Mathewson
40c13240c0 When computing performance thresholds, ignore omitted-as-sybil nodes.
Fixes bug 8146.
2013-02-04 10:19:26 -05:00
Nick Mathewson
ddf2c36ab1 Whoops: evdns_log, not tor_log, in src/ext/eventdns.c 2013-02-04 10:07:02 -05:00
Nick Mathewson
acb43c0735 Merge remote-tracking branch 'public/feature7706' 2013-02-01 17:24:08 -05:00
Nick Mathewson
c93f66b103 Merge remote-tracking branch 'public/bug7708_023_v3_squashed' 2013-02-01 17:11:46 -05:00
Nick Mathewson
690ea9e8cf Clarify documentation of connection_finished_flushing 2013-02-01 17:10:15 -05:00
Nick Mathewson
b442930789 Fix serious breakage in connection_handle_write_impl
When we first implemented TLS, we assumed in conneciton_handle_write
that a TOR_TLS_WANT_WRITE from flush_buf_tls meant that nothing had
been written. But when we moved our buffers to a ring buffer
implementation back in 0.1.0.5-rc (!), we broke that invariant: it's
possible that some bytes have been written but nothing.

That's bad.  It means that if we do a sequence of TLS writes that ends
with a WANTWRITE, we don't notice that we flushed any bytes, and we
don't (I think) decrement buckets.

Fixes bug 7708; bugfix on 0.1.0.5-rc
2013-02-01 17:10:15 -05:00
Nick Mathewson
996db755c2 Fix a couple of warnings on the 8081 branch. 2013-02-01 17:03:00 -05:00
Mike Perry
fed7f01377 Add EntryGuardPathUseBias to state file keyword list. 2013-02-01 17:01:26 -05:00
Mike Perry
b3e57b760e Increment an informational counter for use failed state.
This informational counter is probably now redundant, but might as well keep
it consistent I guess.
2013-02-01 17:01:26 -05:00
Mike Perry
da5817772d Rename and relocate the bw weight scale param getter.
It had nothing to do with circuit build times.
2013-02-01 17:01:22 -05:00
Mike Perry
6e4610de02 Fix a log typo found by sysrqb. 2013-02-01 17:01:22 -05:00
Mike Perry
95d272f5d8 Bounds-check path bias rate parameters.
The other remaining parameters don't really need range checks.
2013-02-01 17:01:22 -05:00
Mike Perry
bce6714f99 Refactor code that rolls back the use state
Also document it better.

Mention this refactoring in the comments for the path state machine.
2013-02-01 17:01:16 -05:00
Mike Perry
3a63e5ef42 Refactor and rename pathbias rate evaluation. 2013-02-01 17:01:12 -05:00
Mike Perry
dfcfb5d17d Refactor the scaling parameter fetching into a single function.
Also, deprecate the torrc options for the scaling values. It's unlikely anyone
but developers will ever tweak them, even if we provided a single ratio value.
2013-02-01 17:01:12 -05:00
Mike Perry
2b2c7f23f5 Mark entry guard state dirty everwhere the pathbias code touches it. 2013-02-01 17:01:12 -05:00
Mike Perry
6828a19670 Add a tristate to guard against unexpected circ purpose transitions 2013-02-01 17:01:12 -05:00
Mike Perry
173ed05d2f Clarify state transition and related pathbias comments 2013-02-01 17:01:12 -05:00
Nick Mathewson
ec90ed4f6d Merge branch 'rename_log_7599' 2013-02-01 16:23:26 -05:00
Roger Dingledine
fd49226385 Help us track bug 8093:
Improve the log message when "Bug/attack: unexpected sendme cell
from client" occurs.
2013-02-01 16:22:34 -05:00
Nick Mathewson
7301339e33 fix wide lines from tor_log rename 2013-02-01 16:19:02 -05:00
Nick Mathewson
1dd03fbc77 Fix a silly mistake in the tor_mathlog() documentation. Give it a unit test. 2013-02-01 16:09:16 -05:00
Nick Mathewson
b034d07acd Remove old wrapper code and defines for keeping log() and log(3) apart
This is the non-automated portion of bug 7599.
2013-02-01 15:49:51 -05:00
Nick Mathewson
a141430ec3 Rename log() to tor_log() for logging
This is meant to avoid conflict with the built-in log() function in
math.h.  It resolves ticket 7599.  First reported by dhill.

This was generated with the following perl script:

 #!/usr/bin/perl -w -i -p

 s/\blog\(LOG_(ERR|WARN|NOTICE|INFO|DEBUG)\s*,\s*/log_\L$1\(/g;

 s/\blog\(/tor_log\(/g;
2013-02-01 15:43:37 -05:00
Nick Mathewson
abb5519d93 typo in crypto_curve25519.c comment, spotted by rransom 2013-01-31 13:53:29 -05:00
Nick Mathewson
294ce2ea87 whitespace fix 2013-01-31 13:26:25 -05:00
Nick Mathewson
b35b4d5a9a Merge branch 'double-0-check' 2013-01-31 13:19:27 -05:00
Nick Mathewson
b0dd355891 Use %d, not %02d, for decimal percentages
Cosmetic tweak on 5956; not in any released tor.
2013-01-30 17:35:28 -05:00
Nick Mathewson
35daf6f602 Rename all of the macros in tor_queue.h to start with TOR_ 2013-01-30 12:58:49 -05:00
Nick Mathewson
29136bd7e4 Merge branch 'bug5956_squashed' 2013-01-30 11:59:51 -05:00
Nick Mathewson
02c320916e Parameterize FRAC_USABLE_NEEDED for fraction of circuits
Instead of hardcoding the minimum fraction of possible paths to 0.6, we
take it from the user, and failing that from the consensus, and
failing that we fall back to 0.6.
2013-01-30 11:58:17 -05:00
Nick Mathewson
813a0f8c40 Compute whether we're ready to build circuits based on fraction of paths
Previously we did this based on the fraction of descriptors we
had. But really, we should be going based on what fraction of paths
we're able to build based on weighted bandwidth, since otherwise a
directory guard or two could make us behave quite oddly.

Implementation for feature 5956
2013-01-30 11:58:17 -05:00
Nick Mathewson
bc52e0488b Add an optional out-arg to count_usable_descriptors
This way we get the usable nodes themselves, so we can feed them into
frac_nodes_with_descriptors
2013-01-30 11:58:17 -05:00
Nick Mathewson
fcf906ec73 Add a function to compute fraction of nodes (by weighted bw) with descriptors 2013-01-30 11:58:17 -05:00
Nick Mathewson
73d605b0f7 Detect platforms where memset(0) doesn't set doubles to 0.0.
This is allowed by the C statndard, which permits you to represent
doubles any way you like, but in practice we have some code that
assumes that memset() clears doubles in structs.  Noticed as part of
7802 review; see 8081 for more info.
2013-01-29 17:38:15 -05:00
Nick Mathewson
a0351311ae Fix unit test to not expect v2 directory request geoip info
When we implemented #5823 and removed v2 directory request info, we
never actually changed the unit tests not to expect it.

Fixes bug 8084; bug not in any released version of Tor.
2013-01-29 12:01:41 -05:00
Andrea Shepard
123daffb60 Merge branch 'bug7802' of ssh://git-rw.torproject.org/mikeperry/tor 2013-01-28 16:16:45 -08:00
Andrea Shepard
dfbd19df41 Merge branch 'time_based_onionqueue_v2' of ssh://git-rw.torproject.org/nickm/tor 2013-01-24 08:10:12 -08:00
Mike Perry
a78542f0c3 Bug 8024: Check for null/closed channel before probing. 2013-01-22 21:03:28 -08:00
Nick Mathewson
b415aba5fa Merge remote-tracking branch 'public/bug8012' 2013-01-22 14:54:30 +00:00
Nick Mathewson
dd77b652f2 More of b30d06255c for #6826: fix compat_libevent compilation
It looks like there was a compilation error for 6826 on some
platforms.  Removing even more now-uncallable code to handle detecting
libevent versions before 1.3e.

Fixes bug 8012; bug not in any released Tor.
2013-01-21 18:24:10 +00:00
Mike Perry
b810d322bf squash! Remove a source of error during path bias scaling
Improve debug logs and fix a state fencepost error.
2013-01-20 14:32:56 -08:00
Mike Perry
06a1d0b044 squash! Implement Path use bias accounting.
Make a debug log more informative.
2013-01-20 14:32:56 -08:00
Mike Perry
f858370233 Prevent early close of path bias testing circuits.
We need to let them live long enough to perform the test.
2013-01-20 14:32:56 -08:00
Mike Perry
fb711e6d77 squash! Remove a source of error during path bias scaling
Move a log message about scaling to after we scale
2013-01-20 14:32:27 -08:00
Nick Mathewson
c71b7db8f3 Merge remote-tracking branch 'karsten/bug5823' 2013-01-19 09:36:55 -05:00
Mike Perry
d80b881a52 Remove a source of error during path bias scaling
If any circuits were opened during a scaling event, we were scaling attempts
and successes by different amounts. This leads to rounding error.

The fix is to record how many circuits are in a state that hasn't been fully
counted yet, and subtract that before scaling, and add it back afterwords.
2013-01-18 21:23:33 -08:00
Mike Perry
a2db17a1aa Don't immediately count cannibalized circs as used.
Since they use RELAY_EARLY (which can be seen by all hops on the path),
it's not safe to say they actually count as a successful use.

There are also problems with trying to allow them to finish extending due to
the circuit purpose state machine logic. It is way less complicated (and
possibly more semantically coherent) to simply wait until we actually try to
do something with them before claiming we 'used' them.

Also, we shouldn't call timed out circuits 'used' either, for semantic
consistency.
2013-01-18 19:46:29 -08:00
Mike Perry
24b9b9f791 Roll back the path_state for circs if we detatch a stream.
An adversary could let the first stream request succeed (ie the resolve), but
then tag and timeout the remainder (via cell dropping), forcing them on new
circuits.

Rolling back the state will cause us to probe such circuits, which should lead
to probe failures in the event of such tagging due to either unrecognized
cells coming in while we wait for the probe, or the cipher state getting out
of sync in the case of dropped cells.
2013-01-18 19:46:28 -08:00
Mike Perry
e13e30221e Implement Path use bias accounting.
Path use bias measures how often we can actually succeed using the circuits we
actually try to use. It is a subset of path bias accounting, but it is
computed as a separate statistic because the rate of client circuit use may
vary depending on use case.
2013-01-18 19:46:21 -08:00