Commit Graph

20542 Commits

Author SHA1 Message Date
Nick Mathewson
8375663adf Wrap two long lines 2015-02-17 08:30:28 -05:00
Roger Dingledine
d6dbd55849 trivial whitespace fix 2015-02-17 06:01:33 -05:00
Nick Mathewson
0b3ae7d9e5 Maybe this will make coverity not care that this check is dead. CID 1268063 2015-02-16 15:44:18 -05:00
Nick Mathewson
2fe139473d No, coverity, this is not a memory leak. 1268065 2015-02-16 15:42:48 -05:00
Nick Mathewson
5d2a23397a Fix a few coverity "Use after NULL check" warnings
Also remove the unit test mocks that allowed get_options() to be
NULL; that's an invariant violation for get_options().
2015-02-16 15:40:15 -05:00
Nick Mathewson
7117959199 Fix deadcode warning in get_interface_addresses_raw().
CID  1268070
2015-02-16 15:28:36 -05:00
Nick Mathewson
2af7bc5b46 Fix a trivial double-close in address.c. CID 1268071 2015-02-16 15:24:13 -05:00
Sina Rabbani
8e61d38cf1 Faravahar's New IP Address as of 2/20/2015 2015-02-16 11:51:36 -05:00
Sebastian Hahn
0c11d8b2d2 Implement status/fresh-relay-descs command
The idea here is that a controller should be able to make Tor produce a
new relay descriptor on demand, without that descriptor actually being
uploaded to the dirauths (they would likely reject it anyway due to
freshness concerns).

Implements #14784.
2015-02-15 12:36:07 +01:00
Sebastian Hahn
8feaf3846d Refactor router_rebuild_descriptor
Allow building a router descriptor without storing it to global state.
This is in preparation of a patch to export the created descriptors via
the control port.
2015-02-15 12:33:20 +01:00
Sebastian Hahn
3bcdb26267 Call cpu_init if we change to being a relay
The issue is that we use the cpuworker system with relays only, so if we
start up as a client and transition to being a relay later, we'll be
sad.

This fixes bug 14901; not in any released version of Tor.
2015-02-15 11:49:19 +01:00
Sebastian Hahn
424edd5710 Don't leak a cond var when starting threads in a pool 2015-02-15 11:21:54 +01:00
Nick Mathewson
5644d92dd7 Merge remote-tracking branch 'sebastian/bug14875' 2015-02-12 14:50:13 -05:00
Sebastian Hahn
c8ce973dc7 Avoid undefined behaviour in ifreq_to_smartlist
This could trigger where _SIZEOF_ADDR_IFREQ() might not return a
multiple of sizeof(void *). Fixes bug 14875; not in any released version
of Tor.
2015-02-12 14:56:47 +01:00
Sebastian Hahn
87a95b0236 Actually get all interface addresses
If we guessed a buffer size too small, we never increased the buffer and
tried again

Also simplify the interface of ifreq_to_smartlist a little
2015-02-12 14:54:39 +01:00
Nick Mathewson
f794a09164 Merge branch 'bug14478_squashed' 2015-02-11 15:59:41 -05:00
Sebastian Hahn
dfb409b8b0 Put error messages to stderr
Thanks weasel :)

Also remove stray debug output
2015-02-11 15:59:34 -05:00
Sebastian Hahn
c5f176b9b0 Change exit code of zero_length_keys script
Weasel notes that exit codes above 128 are for when a script dies from a
signal.
2015-02-11 15:59:34 -05:00
Sebastian Hahn
6ecd6e27da Avoid undefined behaviour of test
Fixes part of bug #14478, patch idea suggested by an anonymous
contributor. Thanks!
2015-02-11 15:59:34 -05:00
Sebastian Hahn
0a0c5d7c0e Check that mktemp result is a directory
Fixes part of bug #14478, patch idea suggested by an anonymous
contributor. Thanks!
2015-02-11 15:59:34 -05:00
Sebastian Hahn
f84a54c159 Clean up after zero length key test 2015-02-11 15:59:34 -05:00
Nick Mathewson
99e915dbfe Merge remote-tracking branch 'public/bug14759' 2015-02-11 15:15:24 -05:00
Nick Mathewson
b223b7c22d fix a compilation warning 2015-02-11 15:10:35 -05:00
Nick Mathewson
d7a1e83f50 Merge remote-tracking branch 'public/remove_freelist' 2015-02-11 15:09:01 -05:00
Nick Mathewson
caf28519d9 Merge branch 'bug12844'
Conflicts:
	src/or/circuituse.c
	src/test/include.am
	src/test/test_entrynodes.c
2015-02-11 15:06:04 -05:00
Nick Mathewson
0899f51bc6 Additional paranoia: do not even build tor2web stuff if not using.
(That is, don't build it unless we're building for tor2web, or we
are building for tests.)
2015-02-11 14:54:16 -05:00
Nick Mathewson
f5f6d13e4c Fix crash on glibc __libc_message()
__libc_message() tries to open /dev/tty with O_RDWR, but the sandbox
catches that and calls it a crash.  Instead, I'm making the sandbox
setenv LIBC_FATAL_STDERR_, so that glibc uses stderr instead.

Fix for 14759, bugfix on 0.2.5.1-alpha
2015-02-11 09:46:29 -05:00
Nick Mathewson
6f331645c7 Remove mempools and buf freelists
They have been off-by-default since 0.2.5 and nobody has complained. :)

Also remove the buf_shrink() function, which hasn't done anything
since we first stopped using contiguous memory to store buffers.

Closes ticket 14848.
2015-02-11 09:03:50 -05:00
Nick Mathewson
0c81dfa848 Merge remote-tracking branch 'public/feature_13555' 2015-02-11 08:42:00 -05:00
Nick Mathewson
37c9c268c8 Merge remote-tracking branch 'sebastian/bug14819' 2015-02-11 08:40:48 -05:00
Nick Mathewson
5c820def99 Merge remote-tracking branch 'sysrqb/bug14802_025' 2015-02-09 22:39:55 -08:00
Nick Mathewson
f4c0735933 Merge remote-tracking branch 'teor/avoid-NULL-smartlist-foreach' 2015-02-09 22:35:40 -08:00
Sebastian Hahn
9667b2b88a Add some comments to or_circuit_t.workqueue_entry
These were suggested by dgoulet, thanks!
2015-02-09 16:13:08 +01:00
Sebastian Hahn
7337510090 Avoid use-after-free of circ belonging to cancelled job
This fixes a bug where we decide to free the circuit because it isn't on
any workqueue anymore, and then the job finishes and the circuit gets
freed again.

Fixes bug #14815, not in any released version of Tor.
2015-02-09 16:12:47 +01:00
Sebastian Hahn
353d2fe7e1 Re-remove the --disable-threads configure option 2015-02-09 10:36:59 +01:00
Roger Dingledine
56061976db Recover better when our clock jumps back many hours
like might happen for Tails or Whonix users who start with a very wrong
hardware clock, use Tor to discover a more accurate time, and then
fix their clock.

Resolves part of ticket 8766.

(There are still some timers in various places that aren't addressed yet.)
2015-02-09 01:05:31 -05:00
Damian Johnson
44abbf5ab6 Dropping test_cmdline_args.py
Before a couple weeks ago didn't know Tor had these tests, interesting! Stem
already has tests for spawning tor processes but lacked any with this targeted
focus on its arguments.

I've added our own counterpart for these tests. Many are direct copies but
there were others I improved a little...

  https://trac.torproject.org/projects/tor/ticket/14109
  https://gitweb.torproject.org/stem.git/commit/?id=137d193a026638f066e817e3396cebbbb6ace012

Now that Tor uses Stem to supplement its tests no reason for these to live
separately. Tested by simply building tor and confirming test_cmdline_args.py
is no longer in the generated Makefile.
2015-02-08 21:34:36 -08:00
Roger Dingledine
1cb9064d7d shift all the static times into a struct
no actual behavior changes yet
2015-02-09 00:07:15 -05:00
Sebastian Hahn
37d16c3cc7 Reserve enough space for rend_service_port_config_t
In #14803, Damian noticed that his Tor sometimes segfaults. Roger noted
that his valgrind gave an invalid write of size one here. Whenever we
use FLEXIBLE_ARRAY_MEMBER, we have to make sure to actually malloc a
thing that's large enough.

Fixes bug #14803, not in any released version of Tor.
2015-02-09 04:48:16 +01:00
Matthew Finkel
9ae321db66 Return 0 when detecting the amount of memory fails
Fixes bug 14802;  bugfix on 0.2.5.4-alpha.
2015-02-09 02:06:18 +00:00
teor
d0759da14e Avoid calling SMARTLIST_FOREACH on a NULL smartlist in tests
Check if each smartlist is NULL before calling SMARTLIST_FOREACH on it.

Bug discovered by the clang static analyzer.
Apple clang 600.0.56 (LLVM 3.5svn) on x86_64-apple-darwin14.1.0.
2015-02-08 23:41:37 +11:00
Nick Mathewson
8b82f6261e Search-and-replace to regain coding style in wake of 13822.
(all-caps should be reserved for global or module-global constants and
2015-02-07 08:33:36 -05:00
Nick Mathewson
4beb830953 Split ROUTER_REQUIRED_MIN_BANDWIDTH into RELAY_ and BRIDGE_ variants
Also raise those minima from 20 to 75 and 50 respectively.

Closes ticket 13822.
2015-02-07 08:33:23 -05:00
David Goulet
b101f4e98c Control: getinfo entry-guards report down-since
If the guard unreachable_since variable was set, the status "up" was
reported which is wrong. This adds the "down" status followed by the
unreachable_since time value.

Fixes #14184

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-02-06 16:06:20 -05:00
Sebastian Hahn
136b1d8ed0 Remove useless export in zero length key test
Fixes part of bug #14478, patch idea suggested by an anonymous
contributor. Thanks!
2015-02-06 21:56:26 +01:00
Sebastian Hahn
d695f7b423 Don't use invalid exit values in zero length key test
Shell exit values must fall into the range of [0-255], so let's honour
this. In practice, the "exit -1" from the old code set an exit value of
255 on most systems, so let's pick that.

Fixes part of bug #14478, patch idea suggested by an anonymous
contributor. Thanks!
2015-02-06 21:55:59 +01:00
Nick Mathewson
e36faeec1d Merge remote-tracking branch 'sebastian/bug13993' 2015-02-06 15:42:53 -05:00
Sebastian Hahn
e0c3de40ad Fix check-spaces complaints 2015-02-06 21:36:40 +01:00
Sebastian Hahn
6d8b614729 Avoid logging startup messages twice 2015-02-06 21:34:21 +01:00
Sebastian Hahn
b3bc871214 Add unit test for #13290 2015-02-06 21:04:05 +01:00