hs_client_purge_state() and hs_cache_clean_as_client() can remove a descriptor
from the client cache with a NEWNYM or simply when the descriptor expires.
Which means that for an INTRO circuit being established during that time, once
it opens, we lookup the descriptor to get the IP object but hey surprised, no
more descriptor.
The approach here is minimalist that is accept the race and close the circuit
since we can not continue. Before that, the circuit would stay opened and the
client wait the SockTimeout.
Fixers #28970.
Signed-off-by: David Goulet <dgoulet@torproject.org>
* actually sleep when tor has not logged anything
* log at debug level when waiting for tor to log something
* backslash-replace bad UTF-8 characters in logs
* format control messages as ASCII: tor does not accept UTF-8 control commands
Fixes bug 31837; bugfix on 0.3.5.1-alpha.
When tearing down all periodic events during shutdown, disable them first so
their enable flag is updated.
This allows the tor_api.h to relaunch tor properly after a clean shutdown.
Fixes#32058
Signed-off-by: David Goulet <dgoulet@torproject.org>
Code adapted from Rob's proposed patch in #30344.
Also add a comment in connection_mark_for_close_internal_() on why we should
not be adding extra code there without a very good reason.
When encoding introduction points, we were not checking if that intro points
had an established circuit.
When botting up, the service will pick, by default, 3 + 2 intro points and the
first 3 that establish, we use them and upload the descriptor.
However, the intro point is removed from the service descriptor list only when
the circuit has opened and we see that we have already enough intro points, it
is then removed.
But it is possible that the service establishes 3 intro points successfully
before the other(s) have even opened yet.
This lead to the service encoding extra intro points in the descriptor even
though the circuit is not opened or might never establish (#31561).
Fixes#31548
Signed-off-by: David Goulet <dgoulet@torproject.org>
When processing a %included folder, a bug caused the pointer to
the last element of the options list to be set to NULL when
processing a file with only comments or whitepace. This could
cause options from other files on the same folder to be
discarded depending on the lines after the affected %include.
This warning would previously be given every time we tried to open a
connection to a foo.exit address, which could potentially be used to
flood the logs. Now, we don't allow this warning to appear more
than once every 15 minutes.
Fixes bug 31466; bugfix on 0.2.2.1-alpha, when .exit was first
deprecated.
Our dimap code asserts if you try to add the same key twice; this
can't happen if everything is running smoothly, but it's possible if
you try to start a relay where secret_onion_key_ntor is the same as
secret_onion_key_ntor.old.
Fixes bug 30916; bugfix on 0.2.4.8-alpha when ntor keys were
introduced.
We previously used tor_fragile_assert() to declare that this case
could not happen: VERSIONS cells are always supposed to be
variable-sized, right?
This is incorrect, though. On a v1 link protocol connection, all
cells are fixed-sized. There aren't supposed to be any VERSIONS
cells with this version of the protocol, but apparently, somebody
was messing up. (The v1 link protocol is obsolete, so probably the
implementer responsible didn't mean to be using it.)
Fixes bug 31107. Bugfix on 0.2.4.4-alpha, when we introduced a
tor_fragile_assert() for this case.
Our code assumes that when we're configured to get IPv6 addresses
out of a TRANS_PF transparent proxy connection, we actually will.
But we didn't check that, and so FreeBSD started warning us about a
potential NULL pointer dereference.
Fixes part of bug 31687; bugfix on 0.2.3.4-alpha when this code was
added.
We used to do this on Windows only, but it appears to affect
multiple platforms when building with certain versions of GCC, and a
common pattern for defining the floating-point classifier functions.
Fixes part of 31687. I'm calling this a bugfux on 31687, when we
started suppressing these warnings on Windows.
When tor is missing descriptors for some primary entry guards, make the
log message less alarming. It's normal for descriptors to expire, as long
as tor fetches new ones soon after.
Fixes bug 31657; bugfix on 0.3.3.1-alpha.
We can't use strlcat() or strlcpy() in torerr, because they are defined
in string/compat_string.h on some platforms, and string uses torerr.
Part of 31571.
Some platforms (macOS, maybe others?) can swallow the last write before an
abort. This issue is probably caused by a race condition between write
buffer cache flushing, and process termination. So we write an extra
newline, to make sure that the message always gets through.
Fixes bug 31571; bugfix on 0.3.5.1-alpha.
We want to report the tor version, even on platforms that don't have
backtrace support (like Android).
This commit stores the backtrace Tor version, regardless of USE_BACKTRACE.
Preparation for 31571.
Previously, v3 single onion services failed when all intro nodes were
unreachable via a 1-hop path. Now, we select intros that are only available
via a 3-hop path, and use a 3-hop path to connect to them.
Fixes bug 23507; bugfix on 0.3.2.1-alpha.
Previously, we used a 1-hop path when a single onion rend failed
immediately, and a 3-hop path when it failed after trying to build
a circuit.
Fixes bug 23818; bugfix on 0.3.2.1-alpha.
Previously, we used a 1-hop path when a single onion rend failed
immediately, and a 3-hop path when it failed after trying to build
a circuit.
Fixes bug 23818; bugfix on 0.2.9.3-alpha.
New IP address from 194.109.206.212 to 45.66.33.45.
Signed request from Alex de Joode, operator of dizum:
https://trac.torproject.org/projects/tor/ticket/31406
Published descriptor by dizum on August 12th, 2019:
--
r dizum fqbq1v2DCDxTj0QDi7+gd1h911U GZmZtCLaPDQNxkhIFj8UcgTRAuA 2019-08-12 15:28:40 45.66.33.45 443 80
s Authority Fast Running Stable V2Dir Valid
v Tor 0.4.0.5
pr Cons=1-2 Desc=1-2 DirCache=1-2 HSDir=1-2 HSIntro=3-4 HSRend=1-2 Link=1-5 LinkAuth=1,3 Microdesc=1-2 Relay=1-2 Padding=1
w Bandwidth=20 Unmeasured=1
p reject 1-65535
--
Finally, confirmed by DNS:
$ dig +short tor.dizum.com
45.66.33.45
Closes#31406
Signed-off-by: David Goulet <dgoulet@torproject.org>
On some windows builds, time_t is 64 bits but long is not. This is
causing appveyor builds to fail.
Also, one of our uses of labs() on time_t was logically incorrect:
it was telling us to accept NETINFO cells up to three minutes
_before_ the message they were responding to, which doesn't make
sense.
This patch adds a time_abs() function that we should eventually move
to intmath.h or something. For now, though, it will make merges
easier to have it file-local in channeltls.c.
Fixes bug 31343; bugfix on 0.2.4.4-alpha.
There is other code that uses this value, and some of it is
apparently reachable from inside router_dir_info_changed(), which
routerlist_free() apparently calls. (ouch!) This is a minimal fix
to try to resolve the issue without causing other problems.
Fixes bug 31003. I'm calling this a bugfix on 0.1.2.2-alpha, where
the call to router_dir_info_changed() was added to routerlist_free().
Overflowing a signed integer in C is an undefined behaviour.
It is possible to trigger this undefined behaviour in tor_asprintf on
Windows or systems lacking vasprintf.
On these systems, eiter _vscprintf or vsnprintf is called to retrieve
the required amount of bytes to hold the string. These functions can
return INT_MAX. The easiest way to recreate this is the use of a
specially crafted configuration file, e.g. containing the line:
FirewallPorts AAAAA<in total 2147483610 As>
This line triggers the needed tor_asprintf call which eventually
leads to an INT_MAX return value from _vscprintf or vsnprintf.
The needed byte for \0 is added to the result, triggering the
overflow and therefore the undefined behaviour.
Casting the value to size_t before addition fixes the behaviour.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
Update the fallback directory mirrors by merging the current list with:
fallback_dirs_2019-06-28-08-58-39_AU_f0437a39ddbc8459.inc
Part of 28795, see that ticket for logs.
Update the fallback directory mirrors by replacing the old list with:
fallback_dirs_2019-06-25-11-49-10_AU_a37adb956fbb5cd2.inc
Part of 28795, see that ticket for logs.
When we consider all circuits in "waiting for guard" state to be promoted to
an "open" state, we were considering all circuits, even the one marked for
close.
This ultiamtely triggers a "circuit_has_opened()" called on the circuit that
is marked for close which then leads to possible undesirable behaviors within
a subsystem.
For instance, the HS subsystem would be unable to find the authentication key
of the introduction point circuit leading to a BUG() warning and a duplicate
mark for close on the circuit.
This commit also adds a unit test to make sure we never select marked for
close circuits when upgrading its guard state from waiting for guard to open.
Fixes#30871
Signed-off-by: David Goulet <dgoulet@torproject.org>
Skip test_rebind when the TOR_SKIP_TEST_REBIND environmental variable
is set.
Skip test_rebind on macOS in Travis builds, because it is unreliable
on macOS on Travis.
Fixes bug 30713; bugfix on 0.3.5.1-alpha.
Leave the other rend and hs_ident data around until circuit free, since code
may still try to inspect it after marking the circuit for close. The
circuitmap is the important thing to clean up, since repurposed
intropoints must be removed from this map to ensure validity.
When this function was implemented, it counted all the entry guards
in the bridge set. But this included previously configured bridges,
as well as currently configured ones! Instead, only count the
_filtered_ bridges (ones that are configured and possibly reachable)
as maybe usable.
Fixes bug 29875; bugfix on 0.3.0.1-alpha.
Previously, we were looking at our global settings to see what kind
of proxy we had. But doing this would sometimes give us the wrong
results when we had ClientTransportPlugin configured but we weren't
using it for a particular connection. In several places in the
code, we had added checks to see if we were _really_ using a PT or
whether we were using a socks proxy, but we had forgotten to do so
in at least once case. Instead, since every time we call this
function we are asking about a single connection, it is probably
best just to make this function connection-specific.
Fixes bug 29670; bugfix on 0.2.6.2-alpha.
If tor is compiled on a system with neither vasprintf nor _vscprintf,
the fallback implementation exposes a logic flaw which prevents
proper usage of strings longer than 127 characters:
* tor_vsnprintf returns -1 if supplied buffer is not large enough,
but tor_vasprintf uses this function to retrieve required length
* the result of tor_vsnprintf is not properly checked for negative
return values
Both aspects together could in theory lead to exposure of uninitialized
stack memory in the resulting string. This requires an invalid format
string or data that exceeds integer limitations.
Fortunately tor is not even able to run with this implementation because
it runs into asserts early on during startup. Also the unit tests fail
during a "make check" run.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
[backported to 0.2.9 by nickm]
We want to support parsing a cell with unknown status code so we are forward
compatible.
Part of #30454
Signed-off-by: David Goulet <dgoulet@torproject.org>
Like the previous commit about the INTRODUCE_ACK status code, change all auth
key type to use the one defined in the trunnel file.
Standardize the use of these auth type to a common ABI.
Part of #30454
Signed-off-by: David Goulet <dgoulet@torproject.org>
This enum was the exact same as hs_intro_ack_status_t that was removed at the
previous commit. It was used client side when parsing the INTRODUCE_ACK cell.
Now, the entire code dealing with the INTRODUCE_ACK cell (both sending and
receiving) have been modified to all use the same ABI defined in the trunnel
introduce1 file.
Finally, the client will default to the normal behavior when receiving an
unknown NACK status code which is to note down that we've failed and re-extend
to the next intro point. This way, unknown status code won't trigger a
different behavior client side.
Part of #30454.
Signed-off-by: David Goulet <dgoulet@torproject.org>
Remove the hs_intro_ack_status_t enum and move the value into trunnel. Only
use these values from now on in the intro point code.
Interestingly enough, the client side also re-define these values in hs_cell.h
with the hs_cell_introd_ack_status_t enum. Next commit will fix that and force
to use the trunnel ABI.
Part of #30454
Signed-off-by: David Goulet <dgoulet@torproject.org>
Previously we purged it in 1-hour increments -- but one-hour is the
maximum TTL for the cache! Now we do it in 25%-TTL increments.
Fixes bug 29617; bugfix on 0.3.5.1-alpha.
Fortunately, in 0.3.5.1-alpha we improved logging for various
failure cases involved with onion service client auth.
Unfortunately, for this one, we freed the file right before logging
its name.
Fortunately, tor_free() sets its pointer to NULL, so we didn't have
a use-after-free bug.
Unfortunately, passing NULL to %s is not defined.
Fortunately, GCC 9.1.1 caught the issue!
Unfortunately, nobody has actually tried building Tor with GCC 9.1.1
before. Or if they had, they didn't report the warning.
Fixes bug 30475; bugfix on 0.3.5.1-alpha.
The INTRODUCE1 trunnel definition file doesn't support that value so it can
not be used else it leads to an assert on the intro point side if ever tried.
Fortunately, it was impossible to reach that code path.
Part of #30454
Signed-off-by: David Goulet <dgoulet@torproject.org>
In 0.3.4 and later, these functions are declared in rephist.h:
STATIC uint64_t find_largest_max(bw_array_t *b);
STATIC void commit_max(bw_array_t *b);
STATIC void advance_obs(bw_array_t *b);
But in 0.2.9, they are declared in rephist.c and test_relay.c.
So compilers fail with a "must use 'struct' tag" error.
We add the missing struct typedef in test_relay.c, to match the
declarations in rephist.c.
(Merge commit 813019cc57 moves these functions into rephist.h instead.)
Fixes bug 30184; not in any released version of Tor.
When releasing OpenSSL patch-level maintenance updates,
we do not want to rebuild binaries using it.
And since they guarantee ABI stability, we do not have to.
Without this patch, warning messages were produced
that confused users:
https://bugzilla.opensuse.org/show_bug.cgi?id=1129411
Fixes bug 30190; bugfix on 0.2.4.2-alpha commit 7607ad2bec
Signed-off-by: Bernhard M. Wiedemann <bwiedemann@suse.de>
In 0.3.4 and later, we declare write_array as:
extern struct bw_array_t *write_array;
...
typedef struct bw_array_t bw_array_t;
But in 0.2.9, we declare write_array as:
typedef struct bw_array_t bw_array_t;
extern bw_array_t *write_array;
And then again in rephist.c:
typedef struct bw_array_t bw_array_t;
So some compilers fail with a duplicate declaration error.
We backport 684b396ce5, which removes the duplicate declaration.
And this commit deals with the undeclared type error.
Backports a single line from merge commit 813019cc57.
Fixes bug 30184; not in any released version of Tor.
Previously, our use of abort() would break anywhere that we didn't
include stdlib.h. This was especially troublesome in case where
tor_assert_nonfatal() was used with ALL_BUGS_ARE_FATAL, since that
one seldom gets tested.
As an alternative, we could have just made this header include
stdlib.h. But that seems bloaty.
Fixes bug 30189; bugfix on 0.3.4.1-alpha.
The function compat_getdelim_ is used for tor_getline if tor is compiled
on a system that lacks getline and getdelim. These systems should be
very rare, considering that getdelim is POSIX.
If this system is further a 32 bit architecture, it is possible to
trigger a double free with huge files.
If bufsiz has been already increased to 2 GB, the next chunk would
be 4 GB in size, which wraps around to 0 due to 32 bit limitations.
A realloc(*buf, 0) could be imagined as "free(*buf); return malloc(0);"
which therefore could return NULL. The code in question considers
that an error, but will keep the value of *buf pointing to already
freed memory.
The caller of tor_getline() would free the pointer again, therefore
leading to a double free.
This code can only be triggered in dirserv_read_measured_bandwidths
with a huge measured bandwith list file on a system that actually
allows to reach 2 GB of space through realloc.
It is not possible to trigger this on Linux with glibc or other major
*BSD systems even on unit tests, because these systems cannot reach
so much memory due to memory fragmentation.
This patch is effectively based on the penetration test report of
cure53 for curl available at https://cure53.de/pentest-report_curl.pdf
and explained under section "CRL-01-007 Double-free in aprintf() via
unsafe size_t multiplication (Medium)".
If the concatenation of connection buffer and the buffer of linked
connection exceeds INT_MAX bytes, then buf_move_to_buf returns -1 as an
error value.
This value is currently casted to size_t (variable n_read) and will
erroneously lead to an increasement of variable "max_to_read".
This in turn can be used to call connection_buf_read_from_socket to
store more data inside the buffer than expected and clogging the
connection buffer.
If the linked connection buffer was able to overflow INT_MAX, the call
of buf_move_to_buf would have previously internally triggered an integer
overflow, corrupting the state of the connection buffer.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
Many buffer functions have a hard limit of INT_MAX for datalen, but
this limitation is not enforced in all functions:
- buf_move_all may exceed that limit with too many chunks
- buf_move_to_buf exceeds that limit with invalid buf_flushlen argument
- buf_new_with_data may exceed that limit (unit tests only)
This patch adds some annotations in some buf_pos_t functions to
guarantee that no out of boundary access could occur even if another
function lacks safe guards against datalen overflows.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
If the concatenation of connection buffer and the buffer of linked
connection exceeds INT_MAX bytes, then buf_move_to_buf returns -1 as an
error value.
This value is currently casted to size_t (variable n_read) and will
erroneously lead to an increasement of variable "max_to_read".
This in turn can be used to call connection_buf_read_from_socket to
store more data inside the buffer than expected and clogging the
connection buffer.
If the linked connection buffer was able to overflow INT_MAX, the call
of buf_move_to_buf would have previously internally triggered an integer
overflow, corrupting the state of the connection buffer.
Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
Many buffer functions have a hard limit of INT_MAX for datalen, but
this limitation is not enforced in all functions:
- buf_move_all may exceed that limit with too many chunks
- buf_move_to_buf exceeds that limit with invalid buf_flushlen argument
- buf_new_with_data may exceed that limit (unit tests only)
This patch adds some annotations in some buf_pos_t functions to
guarantee that no out of boundary access could occur even if another
function lacks safe guards against datalen overflows.
[This is a backport of the submitted patch to 0.2.9, where the
buf_move_to_buf and buf_new_with_data functions did not exist.]
Fixes bug 29922; bugfix on 0.2.9.3-alpha when we tried to capture
all these warnings. No need to backport any farther than 0.3.5,
though -- these warnings don't cause test failures before then.
This one was tricky to find because apparently it only happened on
_some_ windows builds.
In current NSS versions, these ciphersuites don't work with
SSL_ExportKeyingMaterial(), which was causing relays to fail when
they tried to negotiate the v3 link protocol authentication.
Fixes bug 29241; bugfix on 0.4.0.1-alpha.
When classifying a client's selection of TLS ciphers, if the client
ciphers are not yet available, do not cache the result. Previously,
we had cached the unavailability of the cipher list and never looked
again, which in turn led us to assume that the client only supported
the ancient V1 link protocol. This, in turn, was causing Stem
integration tests to stall in some cases. Fixes bug 30021; bugfix
on 0.2.4.8-alpha.
Let's use the same function exit point for BUG() codepath that we're using
for every other exit condition. That way, we're not forgetting to clean up
the memarea.
so that the relays that would be "excluded" from the bandwidth
file because of something failed can be included to diagnose what
failed, without still including these relays in the bandwidth
authorities vote.
Closes#29806.
Allow connections to single onion services to remain idle without being
disconnected.
Relays acting as rendezvous points for single onion services were
mistakenly closing idle established rendezvous circuits after 60 seconds,
thinking that they are unused directory-fetching circuits that had served
their purpose.
Fixes bug 29665; bugfix on 0.2.1.26.
Allow connections to single onion services to remain idle without being
disconnected.
Relays acting as rendezvous points for single onion services were
mistakenly closing idle established rendezvous circuits after 60 seconds,
thinking that they are unused directory-fetching circuits that had served
their purpose.
Fixes bug 29665; bugfix on 0.2.1.26.
Stop leaking parts of the shared random state in the shared-random unit
tests. The previous fix in 29599 was incomplete.
Fixes bug 29706; bugfix on 0.2.9.1-alpha.
Turns out that when reloading a tor configured with hidden service(s), we
weren't copying all the needed information between the old service object to
the new one.
For instance, the desc_is_dirty timestamp wasn't which could lead to the
service uploading its descriptor much later than it would need to.
The replaycache wasn't also moved over and some intro point information as
well.
Fixes#23790
Signed-off-by: David Goulet <dgoulet@torproject.org>
Prior to #23100, we were not counting HS circuit build times in our
calculation of the timeout. This could lead to a condition where our timeout
was set too low, based on non HS circuit build times, and then we would
abandon all HS circuits, storing no valid timeouts in the histogram.
This commit avoids the assert.
When "auto" was used for the port number for a listening socket, the
message logged after opening the socket would incorrectly say port 0
instead of the actual port used.
Fixes bug 29144; bugfix on 0.3.5.1-alpha
Signed-off-by: Kris Katterjohn <katterjohn@gmail.com>
