Commit Graph

3496 Commits

Author SHA1 Message Date
Nick Mathewson
0279e48473 Add support for temporarily suppressing a warning
There are a few places where we want to disable a warning: for
example, when it's impossible to call a legacy API without
triggering it, or when it's impossible to include an external header
without triggering it.

This pile of macros uses GCC's c99 _Pragma support, plus the usual
macro trickery, to enable and disable warnings.
2016-05-27 11:23:52 -04:00
Nick Mathewson
44ea3dc331 Merge branch 'maint-0.2.8' 2016-05-25 10:21:15 -04:00
Nick Mathewson
6d375f17fc Merge branch 'bug19161_028_v2' into maint-0.2.8 2016-05-25 10:17:26 -04:00
Nick Mathewson
a3ec811c2e Merge branch 'maint-0.2.8' 2016-05-25 09:27:47 -04:00
Nick Mathewson
fdfc528f85 Merge branch 'bug19152_024_v2' into maint-0.2.8 2016-05-25 09:26:45 -04:00
Nick Mathewson
c4c4380a5e Fix a dangling pointer issue in our RSA keygen code
If OpenSSL fails to generate an RSA key, do not retain a dangling
pointer to the previous (uninitialized) key value. The impact here
should be limited to a difficult-to-trigger crash, if OpenSSL is
running an engine that makes key generation failures possible, or if
OpenSSL runs out of memory. Fixes bug 19152; bugfix on
0.2.1.10-alpha. Found by Yuan Jochen Kang, Suman Jana, and Baishakhi
Ray.

This is potentially scary stuff, so let me walk through my analysis.
I think this is a bug, and a backport candidate, but not remotely
triggerable in any useful way.

Observation 1a:

Looking over the OpenSSL code here, the only way we can really fail in
the non-engine case is if malloc() fails.  But if malloc() is failing,
then tor_malloc() calls should be tor_asserting -- the only way that an
attacker could do an exploit here would be to figure out some way to
make malloc() fail when openssl does it, but work whenever Tor does it.

(Also ordinary malloc() doesn't fail on platforms like Linux that
overcommit.)

Observation 1b:

Although engines are _allowed_ to fail in extra ways, I can't find much
evidence online  that they actually _do_ fail in practice. More evidence
would be nice, though.

Observation 2:

We don't call crypto_pk_generate*() all that often, and we don't do it
in response to external inputs. The only way to get it to happen
remotely would be by causing a hidden service to build new introduction
points.

Observation 3a:

So, let's assume that both of the above observations are wrong, and the
attacker can make us generate a crypto_pk_env_t with a dangling pointer
in its 'key' field, and not immediately crash.

This dangling pointer will point to what used to be an RSA structure,
with the fields all set to NULL.  Actually using this RSA structure,
before the memory is reused for anything else, will cause a crash.

In nearly every function where we call crypto_pk_generate*(), we quickly
use the RSA key pointer -- either to sign something, or to encode the
key, or to free the key.  The only exception is when we generate an
intro key in rend_consider_services_intro_points().  In that case, we
don't actually use the key until the intro circuit is opened -- at which
point we encode it, and use it to sign an introduction request.

So in order to exploit this bug to do anything besides crash Tor, the
attacker needs to make sure that by the time the introduction circuit
completes, either:
  * the e, d, and n BNs look valid, and at least one of the other BNs is
    still NULL.
OR
  * all 8 of the BNs must look valid.

To look like a valid BN, *they* all need to have their 'top' index plus
their 'd' pointer indicate an addressable region in memory.

So actually getting useful data of of this, rather than a crash, is
going to be pretty damn hard.  You'd have to force an introduction point
to be created (or wait for one to be created), and force that particular
crypto_pk_generate*() to fail, and then arrange for the memory that the
RSA points to to in turn point to 3...8 valid BNs, all by the time the
introduction circuit completes.

Naturally, the signature won't check as valid [*], so the intro point
will reject the ESTABLISH_INTRO cell.  So you need to _be_ the
introduction point, or you don't actually see this information.

[*] Okay, so if you could somehow make the 'rsa' pointer point to a
different valid RSA key, then you'd get a valid signature of an
ESTABLISH_INTRO cell using a key that was supposed to be used for
something else ... but nothing else looks like that, so you can't use
that signature elsewhere.

Observation 3b:

Your best bet as an attacker would be to make the dangling RSA pointer
actually contain a fake method, with a fake RSA_private_encrypt
function that actually pointed to code you wanted to execute.  You'd
still need to transit 3 or 4 pointers deep though in order to make that
work.

Conclusion:

By 1, you probably can't trigger this without Tor crashing from OOM.

By 2, you probably can't trigger this reliably.

By 3, even if I'm wrong about 1 and 2, you have to jump through a pretty
big array of hoops in order to get any kind of data leak or code
execution.

So I'm calling it a bug, but not a security hole. Still worth
patching.
2016-05-25 09:23:57 -04:00
Nick Mathewson
6abceca182 Merge branch 'memarea_overflow_027_squashed' into maint-0.2.8 2016-05-25 09:22:02 -04:00
Nick Mathewson
be2d37ad3c Fix a pointer arithmetic bug in memarea_alloc()
Fortunately, the arithmetic cannot actually overflow, so long as we
*always* check for the size of potentially hostile input before
copying it.  I think we do, though.  We do check each line against
MAX_LINE_LENGTH, and each object name or object against
MAX_UNPARSED_OBJECT_SIZE, both of which are 128k.  So to get this
overflow, we need to have our memarea allocated way way too high up
in RAM, which most allocators won't actually do.

Bugfix on 0.2.1.1-alpha, where memarea was introduced.

Found by Guido Vranken.
2016-05-25 09:20:37 -04:00
Nick Mathewson
0ef36626ea Use calloc, not malloc(a*b), in ed25519 batch signature check fn
[Not a triggerable bug unless somebody is going to go checking
millions+ of signatures in a single go.]
2016-05-25 08:59:08 -04:00
Nick Mathewson
be3875cda2 Make sure that libscrypt_scrypt actually exists before using it.
Previously, if the header was present, we'd proceed even if the
function wasn't there.

Easy fix for bug 19161.  A better fix would involve trying harder to
find libscrypt_scrypt.
2016-05-24 10:31:02 -04:00
Nick Mathewson
b53a2059c4 Expose crypto_digest_algorithm_get_length from crypto.c
Also, use it in routerparse.c
2016-05-23 10:58:27 -04:00
Nick Mathewson
f2205071f0 Remove round_int64_to_next_multiple_of: It is now unused. 2016-05-19 21:21:24 -04:00
Nick Mathewson
6d6c8287d5 Include __mulodi4 in libor_ctime when it fixes clang -m32 -ftrapv
We use a pretty specific pair of autoconf tests here to make sure
that we only add this code when:
   a) a 64-bit signed multiply fails to link,
 AND
   b) the same 64-bit signed multiply DOES link correctly when
      __mulodi4 is defined.

Closes ticket 19079.
2016-05-18 09:50:38 -04:00
Nick Mathewson
060e0d0a75 Merge branch 'crypto_unit_tests_v2_squashed' 2016-05-16 08:26:11 -04:00
Nick Mathewson
62c5a1fa45 Mark even more crypto lines (the fragile_assert ones) as unreachable 2016-05-16 08:26:00 -04:00
Nick Mathewson
b688945dfb Refactor digest allocation backend code
I'm doing this to simplify crypto_digest_smartlist_prefix, and make
it better covered by our tests.
2016-05-16 08:26:00 -04:00
Nick Mathewson
365d0fcc6d Cover all our DH code, and/or mark it unreachable. 2016-05-16 08:26:00 -04:00
Nick Mathewson
98a590577a Treat absent argument to crypto_log_errors as a bug. 2016-05-16 08:26:00 -04:00
Nick Mathewson
d88656ec06 Slight improvements to DH coverage. 2016-05-16 08:25:59 -04:00
Nick Mathewson
c395334879 Mark some unreachable lines in crypto.c 2016-05-16 08:25:59 -04:00
Nick Mathewson
7a5f15b6e0 Improve test coverage of our strongest-rng code. 2016-05-16 08:25:59 -04:00
Nick Mathewson
ec81329339 Do not leak the 'tag' when trying to read a truncated ed25519 key file
Fix for bug 18956.
2016-05-16 08:25:59 -04:00
Nick Mathewson
5b91e70a4f Mark unreachable lines in crypto_ed25519.c 2016-05-16 08:25:59 -04:00
Nick Mathewson
8a536be705 Mark unreachable lines in crypto_curve25519.c
Also, resolve a bug in test_ntor_cl.c
2016-05-16 08:25:53 -04:00
Nick Mathewson
820b1984ad Mark three lines unreachable, with extensive docs and use of BUG macros 2016-05-16 08:25:53 -04:00
Nick Mathewson
7bc9d1e002 Merge branch 'maint-0.2.8' 2016-05-12 15:33:56 -04:00
Nick Mathewson
e8cc9f3edf Merge branch 'maint-0.2.7' into maint-0.2.8 2016-05-12 15:33:47 -04:00
Nick Mathewson
4165b1a0da Merge branch 'bug18977_026_v2' into maint-0.2.7 2016-05-12 15:33:35 -04:00
Nick Mathewson
44cbd00dfa Fix a compiler warning on windows when sizeof(long)==sizeof(int) 2016-05-12 14:51:38 -04:00
Nick Mathewson
20b01cece8 Merge branch 'bug18977_024_v2' into bug18977_026_v2
Had conflicts related to other correct_tm bugs in 0.2.6.  Added wday
for another case.
2016-05-12 14:39:06 -04:00
Nick Mathewson
e57f26c135 Have correct_tm set tm_wday as well.
The tm_wday field had been left uninitialized, which was causing
some assertions to fail on Windows unit tests.

Fixes bug 18977.
2016-05-12 14:37:27 -04:00
Nick Mathewson
6bc052365a Use a much less clever scan_signed no-overflow hack 2016-05-12 14:33:26 -04:00
Nick Mathewson
445e05a015 Fix inconsistent tab/space mixing in include.am files.
This is a whitespace only, cosmetic fix.

There is still some inconsistency between lists, but less
inconsistency inside individual lists.
2016-05-12 13:06:58 -04:00
Nick Mathewson
607a9056d4 Merge branch 'ftrapv_v3'
There were some conflicts here, and some breakage to fix concerning
library link order in newer targets.
2016-05-12 13:00:45 -04:00
Nick Mathewson
ef01109932 Rename SOURCES to SRC for things in include.am 2016-05-12 11:21:28 -04:00
Nick Mathewson
e40cfc4425 Move the ctime part of choose_array_element_by_weight into di_ops
This way it gets the ctime options.
2016-05-12 11:21:28 -04:00
Nick Mathewson
ce854a8d22 Add -ftrapv to gcc-hardening ... mostly!
We know there are overflows in curve25519-donna-c32, so we'll have
to have that one be fwrapv.

Only apply the asan, ubsan, and trapv options to the code that does
not need to run in constant time.  Those options introduce branches
to the code they instrument.

(These introduced branches should never actually be taken, so it
might _still_ be constant time after all, but branch predictors are
complicated enough that I'm not really confident here. Let's aim for
safety.)

Closes 17983.
2016-05-12 11:21:28 -04:00
Nick Mathewson
99c0e1bd5b Fix bad allocation in pubsub.c
Closes 19038.  Bug not in any released Tor.
2016-05-12 09:56:42 -04:00
Nick Mathewson
00ee62b8a5 Merge branch 'pubsub_squashed' 2016-05-11 13:26:29 -04:00
Nick Mathewson
80a6c8caa3 Basic work on a publish/subscribe abstraction
The goal here is to provide a way to decouple pieces of the code
that want to learn "when something happens" from those that realize
that it has happened.

The implementation here consists of a generic backend, plus a set of
macros to define and implement a set of type-safe frontends.
2016-05-11 13:25:11 -04:00
Nick Mathewson
69380033d6 Merge branch 'timeouts_v2_squashed' 2016-05-09 14:06:10 -04:00
Nick Mathewson
11a09778d6 Test coverage for timers. 2016-05-09 14:04:54 -04:00
Nick Mathewson
118556e4b3 Quick-and-dirty test for timers code. 2016-05-09 14:04:53 -04:00
Nick Mathewson
dcf948da06 Add wrappers to tie the new timeouts into libevent. 2016-05-09 14:04:06 -04:00
Nick Mathewson
2384256a37 Merge branch 'maint-0.2.8' 2016-05-04 15:12:20 -04:00
Nick Mathewson
b8e8910d60 Merge branch 'bug18686_025' into maint-0.2.8 2016-05-04 15:12:11 -04:00
Nick Mathewson
c7b9e0b8ed Report success when not terminating an already terminated process.
Also, document the actual behavior and return values of
tor_terminate_process.

Fixes bug18686; bugfix on 0.2.3.9-alpha.
2016-05-04 15:10:36 -04:00
Nick Mathewson
bff53aabce Remove redundant declarations of MIN
Apparently somewhere along the line we decided that MIN might be
missing.

But we already defined it (if it was missing) in compat.h, which
everybody includes.

Closes ticket 18889.
2016-04-25 15:28:58 -04:00
Nick Mathewson
26db1b65b9 Remove trunnel files from libor/libcrypto, since they are in libtrunnel. Found with modularity tool. 2016-04-20 13:39:07 -04:00
Nick Mathewson
520799f084 Merge branch 'handles_squashed' 2016-04-19 14:08:05 -04:00
Nick Mathewson
e015f7c9cc Basic 'handle' implementation and tests.
This abstraction covers the case where one part of the program needs
to refer to another object that is allowed to disappear.
2016-04-19 14:07:43 -04:00
Nick Mathewson
94e3555187 Merge remote-tracking branch 'public/lcov_excl' 2016-04-19 14:05:51 -04:00
Nick Mathewson
12e26a6e76 Disambiguate: Avoid defining two static functions called chunk_free_unchecked 2016-04-15 12:20:14 -04:00
Nick Mathewson
381dae43b6 Add branch prediction to util_bug.h, and fix a bug. 2016-04-15 09:12:03 -04:00
Nick Mathewson
05499b6ded Add timeouts to libor-event.a 2016-04-15 09:03:22 -04:00
Nick Mathewson
0e354ad459 Merge branch 'assert_nonfatal_squashed' 2016-04-14 16:25:21 -04:00
Nick Mathewson
a86ed1d717 Add an IF_BUG_ONCE macro, since that's a pretty common pattern too. 2016-04-14 16:25:07 -04:00
Nick Mathewson
532820b11c Add a BUG macro for usage in if checks. 2016-04-14 16:25:06 -04:00
Nick Mathewson
a885271c08 Add new tor_assert_nonfatal*() macros.
Unlike tor_assert(), these macros don't abort the process.  They're
good for checking conditions we want to warn about, but which don't
warrant a full crash.

This commit also changes the default implementation for
tor_fragile_assert() to tor_assert_nonfatal_unreached_once().

Closes ticket 18613.
2016-04-14 16:24:28 -04:00
Nick Mathewson
0630f1982d Add LCOV_EXCL* markers to crypto.c and crypto_s2k.c
This marks some lines as unreachable by the unit tests, and as
therefore excluded from test coverage.

(Note: This convention is only for lines that are absolutely
unreachable.  Don't use it anywhere you wouldn't add a
tor_fragile_assert().)
2016-04-12 21:13:33 -04:00
Nick Mathewson
eafcd7b0fc Merge branch 'maint-0.2.8' 2016-04-12 13:02:37 -04:00
Nick Mathewson
7babf33239 Merge remote-tracking branch 'public/bug18716_027' into maint-0.2.8 2016-04-12 13:02:02 -04:00
Nick Mathewson
39c057d45a memarea: Don't assume that sizeof(ulong) >= sizeof(void*).
Fixes bug 18716; bugfix on 0.2.1.1-alpha where memarea.c was
introduced.  Found by wbenny.
2016-04-07 11:10:14 -04:00
Nick Mathewson
b46d126e64 Merge branch 'maint-0.2.8' 2016-04-05 10:38:53 -04:00
Nick Mathewson
967491f156 Only define NEW_THREAD_API when not building with LibreSSL. 2016-04-05 10:38:15 -04:00
Nick Mathewson
16f7851807 Merge remote-tracking branch 'teor/bug18720' 2016-04-05 10:08:11 -04:00
Nick Mathewson
6720628c97 Merge branch 'maint-0.2.8' 2016-04-05 10:06:18 -04:00
Yawning Angel
5db21f8f81 OpenSSL 1.1.0-pre5-dev and later made BIO opaque.
Detect newer versions and fix our TLS code to use the new API.
2016-04-05 10:03:24 -04:00
Yawning Angel
6729d7328c OpenSSL 1.1.0-pre4 and later(?) have a new "thread API".
It appears that setting the various callbacks is no longer required, so
don't.
2016-04-05 10:03:24 -04:00
teor (Tim Wilson-Brown)
6a2b4db4f9 Fix a comment typo in compat.h 2016-04-05 13:45:37 +10:00
Nick Mathewson
7865402106 Move tor_assert implementation into its own header/module. 2016-04-04 11:06:04 -04:00
Nick Mathewson
055a7a198a Rename tor_dup_addr to tor_addr_to_str_dup.
Patch from icanhasaccount; closes 18462.
2016-03-28 16:36:51 -04:00
Nick Mathewson
4895d8288c Do not treat "DOCDOC" as doxygen. 2016-03-26 10:11:45 -04:00
Nick Mathewson
cc90b57b04 add a little documentation to memarea. (I have been testing a tool.) 2016-03-26 10:09:19 -04:00
Nick Mathewson
c0568a89d9 Whitespace fixes 2016-03-26 09:54:31 -04:00
Nick Mathewson
dd572dac34 Fix all doxygen warnings (other than missing docs) 2016-03-26 09:53:12 -04:00
Nick Mathewson
4bb44f2c15 Only check in-boundsness of seconds when time_t is smaller than i64
Otherwise coverity complains that we're checking an whether an int64 is
less than INT64_MIN, which of course it isn't.

Fixes CID 1357176. Not in any released Tor.
2016-03-25 16:46:02 -04:00
Nick Mathewson
6256c61d95 Merge branch 'timegm_overflow_squashed' 2016-03-24 10:18:00 -04:00
teor (Tim Wilson-Brown)
e71e8e005a Avoid overflow in tor_timegm on 32 bit platforms due to year 2038 2016-03-24 10:17:48 -04:00
Nick Mathewson
7123e9706e Repair build when no sandbox support is enabled. 2016-03-22 13:18:18 -04:00
Nick Mathewson
ca8423a703 Merge remote-tracking branch 'public/bug18253' 2016-03-22 10:08:50 -04:00
Nick Mathewson
a17537a238 Fix an fd leak in check_private_dir().
The fd would leak when the User wasn't recogniezed by
getpwnam(). Since we'd then go on to exit, this wasn't a terribad
leak, but it's still not as nice as no leak at all.

CID 1355640; bugfix on no released Tor.
2016-03-22 08:29:51 -04:00
Nick Mathewson
b48f8a8114 Fix whitespace. 2016-03-15 09:21:29 -04:00
Nick Mathewson
c9899ee640 Merge remote-tracking branch 'weasel/bug18458' 2016-03-15 09:18:24 -04:00
Nick Mathewson
4b02af452d Merge branch 'bug15221_027' 2016-03-14 14:10:47 -04:00
Nick Mathewson
dd7c999617 Make unix sockets work with the linux seccomp2 sandbox again
I didn't want to grant blanket permissions for chmod() and chown(),
so here's what I had to do:
   * Grant open() on all parent directories of a unix socket
   * Write code to allow chmod() and chown() on a given file only.
   * Grant chmod() and chown() on the unix socket.
2016-03-14 14:07:02 -04:00
Nick Mathewson
725e0c76e3 Permit setrlimit, prlimit, prlimit64 calls.
We call setrlimit under some circumstances, and it can call prlimit
and prlimit64 under the hood.

Fixes bug 15221.
2016-03-14 13:21:16 -04:00
Nick Mathewson
36ad65a7d1 When using open() to make sure we created a dir, close the fd afterwards
Found by coverity. Not in any released Tor. Fixes CID 1355640.

Also, don't check for fd correctness with assert(fd).  You need to
assert (fd >= 0).
2016-03-14 13:03:44 -04:00
teor (Tim Wilson-Brown)
9f98e6535a Correctly duplicate addresses in get_interface_address6_list 2016-03-04 18:42:27 +01:00
teor (Tim Wilson-Brown)
2627299ef0 Avoid freeing an uninitialised pointer in get_interface_addresses_ioctl 2016-03-04 18:41:49 +01:00
Peter Palfrader
1ef7df551d First RelaxDirModeCheck implementation 2016-03-01 17:08:14 +01:00
Nick Mathewson
9fc472e1a8 clean/extend some module docs, including fix from #18403 2016-02-28 17:57:47 +01:00
Nick Mathewson
57699de005 Update the copyright year. 2016-02-27 18:48:19 +01:00
Nick Mathewson
f4864d37ec It appears I added an empty file by mistake. 2016-02-27 18:25:51 +01:00
Nick Mathewson
23f8c9b32f Add a brief file-level description for everything in src/common 2016-02-27 18:19:57 +01:00
Nick Mathewson
d5cbc21ad1 Fix an unused-variable warning 2016-02-27 10:20:15 +01:00
Nick Mathewson
0a276947ba Merge branch 'bug18392' 2016-02-27 10:17:51 +01:00
Nick Mathewson
7255b1121d Fix check_private_dir() to work on Windows again.
On windows, you cannot open() a directory.  So for Windows we should
just take our previous stat-based approach.

Closes bug 18392; bug not in any released Tor.
2016-02-25 13:34:12 -05:00
Nick Mathewson
7a782820e9 Make the sandbox work again with chutney.
Previously, we had a problem due to the check_private_dir() rewrite.

Bug not in any released Tor.
2016-02-24 16:01:24 -05:00
Nick Mathewson
73c433a48a Remove the freelist from memarea.c
This is in accordance with our usual policy against freelists,
now that working allocators are everywhere.

It should also make memarea.c's coverage higher.

I also doubt that this code ever helped performance.
2016-02-24 14:32:09 -05:00