Commit Graph

15586 Commits

Author SHA1 Message Date
Roger Dingledine
ca525db02d fix typos from 783f705d 2012-06-15 17:08:25 -04:00
Nick Mathewson
97555f4537 fix a compiler warning added in one of my XXX023 fixes. 2012-06-15 16:43:59 -04:00
Nick Mathewson
30c4653780 Whitespace fix 2012-06-15 16:12:24 -04:00
Nick Mathewson
eab75d2c36 Fix a typo found by Mike. 2012-06-15 16:12:04 -04:00
Nick Mathewson
bdfb399867 Merge remote-tracking branch 'public/xxx023' 2012-06-15 16:10:59 -04:00
Nick Mathewson
87409771c4 Clarify some messages about publishing hidden service descriptors
Fix for bug 3311.
2012-06-15 15:25:46 -04:00
Nick Mathewson
cc21e56ed4 Check the correct consensus before giving it to the client
Previously, a directory would check the latest NS consensus for
having the signatures the client wanted, and use that consensus's
valid_until time to set the HTTP lifetime.  With this patch, the
directory looks at NS consensus or the microdesc consensus,
depending on what the client asked for.
2012-06-15 15:07:54 -04:00
Nick Mathewson
32bf258881 Change a silent ignore-the-bug in microdesc.c to a LOG_INFO
I don't believe this bug occurs, but there was an XXX023 to make
sure it doesn't.
2012-06-15 15:07:54 -04:00
Nick Mathewson
e62104a7d2 Move tor_gettimeofday_cached() into compat_libevent 2012-06-15 15:07:53 -04:00
Nick Mathewson
1755f792ed Refactor GETINFO process/descriptor-limit
Previously it duplicated some getrlimit code and content from compat.c;
now it doesn't.
2012-06-15 15:07:53 -04:00
Nick Mathewson
783f705ddc Document that we are unlikely to underflow session group IDs. 2012-06-15 15:07:53 -04:00
Nick Mathewson
2491fff5a6 Triage the XXX023 and XXX022 comments: postpone many. 2012-06-15 15:07:52 -04:00
Nick Mathewson
879b1e1010 Merge remote-tracking branch 'public/bug5932' 2012-06-15 14:44:32 -04:00
Nick Mathewson
8030ec4f27 Downgrade log messages about cbt enabled/disabled. Bug 6169. 2012-06-15 09:57:18 -04:00
Roger Dingledine
a7efedc322 whitespace fix 2012-06-15 04:41:57 -04:00
Roger Dingledine
afa07b4f87 bump to 0.2.3.17-beta-dev 2012-06-15 04:29:32 -04:00
Roger Dingledine
1e1c7d779e give 0.2.3.17-beta a release blurb 2012-06-15 04:16:40 -04:00
Roger Dingledine
a1caa96f9b another little step at making debugging 5458 easier 2012-06-15 03:58:47 -04:00
Roger Dingledine
5625812f9a tab-man returneth (this time using the name 'rob') 2012-06-15 03:28:18 -04:00
Roger Dingledine
427d4dd148 fix typos, logic error, default in man page 2012-06-15 03:25:27 -04:00
Roger Dingledine
4c87e82c6c bump to 0.2.3.17-beta 2012-06-15 03:13:00 -04:00
Roger Dingledine
f56e3122d8 fold in next changes items 2012-06-15 03:10:07 -04:00
Nick Mathewson
4a9498d682 Expand on bug5458 changes file 2012-06-15 00:45:48 -04:00
Mike Perry
eb2a4be2f7 Add a changes file for bug5458. 2012-06-14 21:34:33 -07:00
Mike Perry
daedae4115 Lower the default path bias notice rate to 40%.
I saw 72% on a test run with 26 circuits. 70% might be a little close to the
line. That, or min_circs is too low and we need to be more patient. We still
need to test/simulate more.
2012-06-14 21:20:10 -07:00
Mike Perry
61a5730392 For now, never disable any guards. 2012-06-14 13:20:01 -07:00
Mike Perry
8d59690033 Defend against entry node path bias attacks
The defense counts the circuit failure rate for each guard for the past N
circuits. Failure is defined as the ability to complete a first hop, but not
finish completing the circuit all the way to the exit.

If the failure rate exceeds a certain amount, a notice is emitted.

If it exceeds a greater amount, a warn is emitted and the guard is disabled.

These values are governed by consensus parameters which we intend to tune as
we perform experiments and statistical simulations.
2012-06-14 13:19:56 -07:00
Nick Mathewson
75706527c1 Document --hush; fix documentation for --quiet. 2012-06-14 14:58:51 -04:00
Nick Mathewson
4fdce6b091 Merge remote-tracking branch 'asn-mytor/bug5589_take2' 2012-06-14 13:05:16 -04:00
George Kadianakis
aa212b173c Remove validate_pluggable_transports_config(): redundant since 9d9b5ed0.
The warning message of validate_pluggable_transports_config() is
superseded by the changes in the warning message of
connection_or_connect() when the proxy credentials can't be found.
2012-06-14 18:01:22 +03:00
Roger Dingledine
7e57609a9f demote two entries that don't affect most users 2012-06-14 05:50:32 -04:00
Roger Dingledine
8c044af300 fold in further changes files 2012-06-14 05:47:55 -04:00
Nick Mathewson
e5beb82e04 Merge remote-tracking branch 'public/bug4663' 2012-06-13 17:01:53 -04:00
Nick Mathewson
e3a6e01fe1 Merge branch 'trac-5049-squashed' 2012-06-13 16:55:39 -04:00
Nick Mathewson
5b0977df31 One more fix for bug 5049. 2012-06-13 16:45:13 -04:00
Andrea Shepard
d98590d3b7 Satisfy make check-spaces 2012-06-13 16:45:13 -04:00
Andrea Shepard
aa284561c8 Move cbt->liveness.timeouts_after_firsthop free code into its own function 2012-06-13 16:45:13 -04:00
Andrea Shepard
825fb149cd Add change file for bug 5049 2012-06-13 16:45:13 -04:00
Andrea Shepard
39a9178ba7 Early exit from circuit_build_times_set_timeout() if adaptive timeouts are disabled 2012-06-13 16:45:13 -04:00
Andrea Shepard
0c3c0b1ddd Don't poll to see if we need to build circuits for timeout data if LearnCircuitBuildTimeout is disabled 2012-06-13 16:45:12 -04:00
Andrea Shepard
a0f76289fd Use K&R style 2012-06-13 16:45:12 -04:00
Andrea Shepard
7df26de948 Unconditionally use config CircuitBuildTimeout if LearnCircuitBuildTimeout is disabled 2012-06-13 16:44:33 -04:00
Andrea Shepard
5177ab9e47 Don't track circuit timeout history unless we're actually using adaptive timeouts 2012-06-13 16:44:33 -04:00
Andrea Shepard
41a458ece1 Add debug logging to circuit_build_times_* of circuitbuild.c to trace queries of consensus parameters for bug 5049 2012-06-13 16:44:33 -04:00
Nick Mathewson
7826eef93e Only use -Qunused-arguments when building with clang. fix on 5210 fix. 2012-06-13 16:37:23 -04:00
Nick Mathewson
54ef039ba5 Merge branch 'bug5263_023' 2012-06-13 16:23:16 -04:00
Nick Mathewson
9c58873059 Add changes file for bug5263 2012-06-13 16:21:27 -04:00
Nick Mathewson
9282c88998 Add rate-limited log message to bug5263 fix
Initially I said, "I claim that we shouldn't be reading and marked;
let's see if I'm right."  But Rob finds that it does.
2012-06-13 16:21:06 -04:00
Rob G. Jansen
03b48352c6 Fix busy Libevent loops (infinite loops in Shadow)
There is a bug causing busy loops in Libevent and infinite loops in
the Shadow simulator. A connection that is marked for close, wants
to flush, is held open to flush, but is rate limited (the token
bucket is empty) triggers the bug.

This commit fixes the bug. Details are below.

This currently happens on read and write callbacks when the active
socket is marked for close. In this case, Tor doesn't actually try
to complete the read or write (it returns from those methods when
marked), but instead tries to clear the connection with
conn_close_if_marked(). Tor will not close a marked connection that
contains data: it must be flushed first. The bug occurs when this
flush operation on the marked connection can not occur because the
connection is rate-limited (its write token bucket is empty).

The fix is to detect when rate limiting is preventing a marked
connection from properly flushing. In this case, it should be
flagged as read/write_blocked_on_bandwidth and the read/write events
de-registered from Libevent. When the token bucket gets refilled, it
will check the associated read/write_blocked_on_bandwidth flag, and
add the read/write event back to Libevent, which will cause it to
fire. This time, it will be properly flushed and closed.

The reason that both read and write events are both de-registered
when the marked connection can not flush is because both result in
the same behavior. Both read/write events on marked connections will
never again do any actual reads/writes, and are only useful to
trigger the flush and close the connection. By setting the
associated read/write_blocked_on_bandwidth flag, we ensure that the
event will get added back to Libevent, properly flushed, and closed.

Why is this important? Every Shadow event occurs at a discrete time
instant. If Tor does not properly deregister Libevent events that
fire but result in Tor essentially doing nothing, Libevent will
repeatedly fire the event. In Shadow this means infinite loop,
outside of Shadow this means wasted CPU cycles.
2012-06-13 16:04:07 -04:00
Nick Mathewson
37ef4f1689 Change smartlist_create->smartlist_new in bug4744 branch as merged to master 2012-06-13 12:16:02 -04:00