Commit Graph

22743 Commits

Author SHA1 Message Date
Nick Mathewson
d97fca16d0 Fix an integer overflow related to monotonic time on windows.
To maintain precision, to get nanoseconds, we were multiplying our
tick count by a billion, then dividing by ticks-per-second.  But
that apparently isn't such a great idea, since ticks-per-second is
sometimes a billion on its own, so our intermediate result was
giving us attoseconds.

When you're counting in attoseconds, you can only fit about 9
seconds into an int64_t, which is not so great for our purposes.

Instead, we now simplify the 1000000000/1000000000 fraction before
we start messing with nanoseconds.  This has potential to mess us
up if some future MS version declares that performance counters will
use 1,000,000,007 units per second, but let's burn that bridge when
we come to it.
2016-07-26 11:23:58 -04:00
Nick Mathewson
09c25697d7 Add a function to simplify a fraction.
Apparently remembering euclid's algorithm does pay off sooner or later.
2016-07-26 11:23:34 -04:00
Nick Mathewson
90ca446048 Remove windows debugging prints: it was an integer overflow hitting ftrapv 2016-07-26 11:07:53 -04:00
Nick Mathewson
019b7ddb9f fix identifier mistake :( 2016-07-26 10:44:51 -04:00
Nick Mathewson
160d2c6aab Redux^3: Temporarily add windows verbosity to track down jenkins failures 2016-07-26 10:36:44 -04:00
Nick Mathewson
0cef69713c Redux^2: Temporarily add windows verbosity to track down jenkins failures 2016-07-26 10:04:40 -04:00
Nick Mathewson
fb7f90c181 Tweaks on 19435 fix:
* Raise limit: 16k isn't all that high.
   * Don't log when limit exceded; log later on.
   * Say "over" when we log more than we say we log.
   * Add target version to changes file
2016-07-26 09:59:48 -04:00
Ivan Markin
77459b97aa Fix integer overflow in the rate-limiter (#19435). 2016-07-26 09:49:40 -04:00
Nick Mathewson
264fb7eb82 debugging: print ticks-per-second on windows. is it 0? 2016-07-26 09:44:41 -04:00
Nick Mathewson
e77c23e455 Redux: Temporarily add windows verbosity to track down jenkins failures 2016-07-26 09:24:53 -04:00
Nick Mathewson
1033713c9c Temporarily add some windows verbosity to track down unit test failure on jenkins. 2016-07-26 08:56:55 -04:00
Nick Mathewson
3f9c036821 Try a little harder to work around mingw clock_gettime weirdness 2016-07-26 08:22:37 -04:00
Roger Dingledine
81cc31d3fa three typo fixes i found in my sandbox 2016-07-25 20:17:35 -04:00
Peter Palfrader
518c8fe0ec Also ship compat_time.h in release tarballs. Fixes bug #19746 2016-07-25 09:07:29 +02:00
Nick Mathewson
53f9f71985 ug no, the RIGHT fix. 2016-07-21 15:29:56 +02:00
Nick Mathewson
9c210d0e81 Avoid infinite stack explosion in windows monotime.
[init calls get calls init calls get calls init.... ]
2016-07-21 15:26:05 +02:00
Nick Mathewson
1d0775684d Once more, 32-bit fixes on monotime mocking 2016-07-21 14:32:15 +02:00
Nick Mathewson
9c87869dde Merge branch 'maint-0.2.8' 2016-07-21 14:15:19 +02:00
Nick Mathewson
f1973e70a4 Coverity hates it when we do "E1 ? E2 : E2".
It says, 'Incorrect expression (IDENTICAL_BRANCHES)'

Fix for CID 1364127. Not in any released Tor.
2016-07-21 14:14:33 +02:00
Nick Mathewson
22314f9050 loony mingwcross bug: insist we dont have clock_gettime. 2016-07-21 14:09:00 +02:00
Nick Mathewson
852cff043b fix monotime test mocking on 32-bit systems 2016-07-21 14:05:29 +02:00
Nick Mathewson
61ce6dcb40 Make monotime test mocking work with oom tests. 2016-07-21 07:02:34 -04:00
Nick Mathewson
50463524a9 Use new mock functions in buffer/time_tracking test 2016-07-21 07:02:33 -04:00
Nick Mathewson
2d26b1a549 Actually make monotonic time functions mockable.
This is different from making the functions mockable, since
monotime_t is opaque and so providing mocks for the functions is
really hard.
2016-07-21 07:02:33 -04:00
Nick Mathewson
72a1f0180d Revert "Make the monotonic{_coarse,}_get() functions mockable."
This reverts commit 2999f0b33f.
2016-07-21 10:30:21 +02:00
Nick Mathewson
2999f0b33f Make the monotonic{_coarse,}_get() functions mockable. 2016-07-21 10:25:23 +02:00
Nick Mathewson
a31078a581 Merge branch 'maint-0.2.8' 2016-07-19 12:34:37 +02:00
Nick Mathewson
4d5b252f0f Merge branch 'maint-0.2.7' into maint-0.2.8 2016-07-19 12:34:27 +02:00
Nick Mathewson
4d70ed7be0 Merge branch 'maint-0.2.6' into maint-0.2.7 2016-07-19 12:32:14 +02:00
Nick Mathewson
210928f66a Merge branch 'maint-0.2.5' into maint-0.2.6 2016-07-19 12:31:54 +02:00
Nick Mathewson
d95c2809b3 Merge branch 'maint-0.2.4' into maint-0.2.5 2016-07-19 12:31:20 +02:00
Nick Mathewson
558f7d3701 Merge branch 'monotonic_v2_squashed' 2016-07-19 11:42:26 +02:00
Nick Mathewson
1e3cf1cc83 Be sure to call monotime_init() at startup. 2016-07-19 11:40:47 +02:00
Nick Mathewson
6ba415d400 Make sure initialized_at is initialized before use. 2016-07-19 11:40:47 +02:00
Nick Mathewson
493142d91f Changes file for monotonic time branch. 2016-07-19 11:40:47 +02:00
Nick Mathewson
abcb8ce25d Unit tests for monotonic time 2016-07-19 11:40:47 +02:00
Nick Mathewson
2a217ef723 Expose monotonic time ratchet functions for testing. 2016-07-19 11:40:47 +02:00
Nick Mathewson
7bc4ca7de9 Remove tor_gettimeofday_cached_monotonic as broken and unneeded 2016-07-19 11:40:47 +02:00
Nick Mathewson
c7558c906a Use coarse monotonic timer instead of cached monotonized libevent time. 2016-07-19 11:40:46 +02:00
Nick Mathewson
6a2002fc09 convert timers.c to use real monotonic time. 2016-07-19 11:40:46 +02:00
Nick Mathewson
dc6f5d1dc1 Basic portable monotonic timer implementation
This code uses QueryPerformanceCounter() [**] on Windows,
mach_absolute_time() on OSX, clock_gettime() where available, and
gettimeofday() [*] elsewhere.

Timer types are stored in an opaque OS-specific format; the only
supported operation is to compute the difference between two timers.

[*] As you know, gettimeofday() isn't monotonic, so we include
a simple ratchet function to ensure that it only moves forward.

[**] As you may not know, QueryPerformanceCounter() isn't actually
always as monotonic as you might like it to be, so we ratchet that
one too.

We also include a "coarse monotonic timer" for cases where we don't
actually need high-resolution time.  This is GetTickCount{,64}() on
Windows, clock_gettime(CLOCK_MONOTONIC_COARSE) on Linux, and falls
back to regular monotonic time elsewhere.
2016-07-19 11:40:46 +02:00
Karsten Loesing
79939c6f11 Update geoip and geoip6 to the July 6 2016 database. 2016-07-18 08:40:22 +02:00
teor (Tim Wilson-Brown)
d8cd994ef6
Allow clients to retry HSDirs much faster in test networks 2016-07-18 13:25:15 +10:00
teor (Tim Wilson-Brown)
6afd5506e9
Rewrite test-network.sh so out-of-tree and $PATH binaries work 2016-07-18 12:15:49 +10:00
Andrea Shepard
94c27d4e8f Keep make check-spaces happy 2016-07-17 23:22:29 +00:00
Nick Mathewson
c138c9a2be Merge branch 'maint-0.2.8' 2016-07-17 13:55:04 -04:00
Nick Mathewson
fbae15a856 Merge remote-tracking branch 'weasel/bug19660' into maint-0.2.8 2016-07-17 13:54:40 -04:00
Nick Mathewson
bec4e41f4b Fix warnings in test_util_formats.
Storing 255 into a char gives a warning when char is signed.

Fixes bug 19682; bugfix on 0.2.8.1-alpha, where these tests were added.
2016-07-17 13:51:45 -04:00
teor (Tim Wilson-Brown)
579a80d4ae
Clients avoid choosing nodes that can't do ntor
If we know a node's version, and it can't do ntor, consider it not running.
If we have a node's descriptor, and it doesn't have a valid ntor key,
consider it not running.

Refactor these checks so they're consistent between authorities and clients.
2016-07-15 09:55:49 +10:00
teor (Tim Wilson-Brown)
a76d528bec
Clients no longer download descriptors for relays without ntor 2016-07-15 09:55:49 +10:00