From http://archives.seul.org/tor/relays/Mar-2010/msg00006.html :
As I understand it, the bug should show up on relays that don't set
Address to an IP address (so they need to resolve their Address
line or their hostname to guess their IP address), and their
hostname or Address line fails to resolve -- at that point they'll
pick a random 4 bytes out of memory and call that their address. At
the same time, relays that *do* successfully resolve their address
will ignore the result, and only come up with a useful address if
their interface address happens to be a public IP address.
When the bandwidth-weights branch added the "directory-footer"
token, and began parsing the directory footer at the first
occurrence of "directory-footer", it made it possible to fool the
parsing algorithm into accepting unsigned data at the end of a
consensus or vote. This patch fixes that bug by treating the footer
as starting with the first "directory-footer" or the first
"directory-signature", whichever comes first.
Treat strings returned from signed_descriptor_get_body_impl() as not
NUL-terminated. Since the length of the strings is available, this is
not a big problem.
Discovered by rieo.
Another dereference-then-NULL-check sequence. No reports of this bug
triggered in the wild. Fixes bugreport 1256.
Thanks to ekir for discovering and reporting this bug.
Fix a dereference-then-NULL-check sequence. This bug wasn't triggered
in the wild, but we should fix it anyways in case it ever happens.
Also make sure users get a note about this being a bug when they
see it in their log.
Thanks to ekir for discovering and reporting this bug.
We used to only zero the first ptrsize bytes of the cipher. Since
cipher is large enough, we didn't zero too many bytes. Discovered
and fixed by ekir. Fixes bug 1254.
All other bandwidthrate settings are restricted to INT32_MAX, but
this check was forgotten for PerConnBWRate and PerConnBWBurst. Also
update the manpage to reflect the fact that specifying a bandwidth
in terabytes does not make sense, because that value will be too
large.
Fix a dereference-then-NULL-check sequence. This bug wasn't triggered
in the wild, but we should fix it anyways in case it ever happens.
Also make sure users get a note about this being a bug when they
see it in their log.
Thanks to ekir for discovering and reporting this bug.
We used to only zero the first ptrsize bytes of the cipher. Since
cipher is large enough, we didn't zero too many bytes. Discovered
and fixed by ekir. Fixes bug 1254.
Spec conformance issue: The code didn't force the network-status-version
token to be the first token in a v3 vote or consensus.
Problem discovered by Parakeep.
We need to use evdns_add_server_port_with_base() when configuring
our DNS listener, because libevent segfaults otherwise. Add a macro
in compat_libevent.h to pick the correct implementation depending
on the libevent version.
Fixes bug 1143, found by SwissTorExit
We implemented ratelimiting for warnings going into the logfile, but didn't
rate-limit controller events. Now both log warnings and controller events
are rate-limited.
Tor has tor_lookup_hostname(), which prefers ipv4 addresses automatically.
Bug 1244 occured because gethostbyname() returned an ipv6 address, which
Tor cannot handle currently. Fixes bug 1244; bugfix on 0.0.2pre25.
Reported by Mike Mestnik.
The problem was that we didn't allocate enough memory on 32-bit
platforms with 64-bit time_t. The memory leak occured every time
we fetched a hidden service descriptor we've fetched before.
exit_policy_is_general_exit() assumed that there are no redundancies
in the passed policy, in the sense that we actively combine entries
in the policy to really get rid of any redundancy. Since we cannot
do that without massively rewriting the policy lines the relay
operators set, fix exit_policy_is_general_exit().
Fixes bug 1238, discovered by Martin Kowalczyk.
We need to do this because Apple doesn't update its dev-tools headers
when it updates its libraries in a security patch. On the bright
side, this might get us out of shipping a statically linked OpenSSL on
OSX.
May fix bug 1225.
[backported]
We need to do this because Apple doesn't update its dev-tools headers
when it updates its libraries in a security patch. On the bright
side, this might get us out of shipping a statically linked OpenSSL on
OSX.
May fix bug 1225.
We accidentally freed the internal buffer for bridge stats when we
were writing the bridge stats file or honoring a control port
request for said data. Change the interfaces for
geoip_get_bridge_stats* to prevent these problems, and remove the
offending free/add a tor_strdup.
Fixes bug 1208.
The OutboundBindAddress option is useful for making sure that all of
your outbond connections use a given interface. But when connecting
to 127.0.0.1 (or ::1 even) it's important to actually have the
connection come _from_ localhost, since lots of programs running on
localhost use the source address to authenticate that the connection
is really coming from the same host.
Our old code always bound to OutboundBindAddress, whether connecting
to localhost or not. This would potentially break DNS servers on
localhost, and socks proxies on localhost. This patch changes the
behavior so that we only look at OutboundBindAddress when connecting
to a non-loopback address.
...to let us
rate-limit client connections as they enter the network. It's
controlled in the consensus so we can turn it on and off for
experiments. It's starting out off. Based on proposal 163.
Fix statistics on client numbers by country as seen by bridges that were
broken in 0.2.2.1-alpha. Also switch to reporting full 24-hour intervals
instead of variable 12-to-48-hour intervals.
The HSAuthorityRecordStats option was used to track statistics of overall
hidden service usage on the version 0 hidden service authorities. With the
version 2 hidden service directories being deployed and version 0
descriptors being phased out, these statistics are not as useful anymore.
Goodbye, you fine piece of software; my first major code contribution to
Tor.
Avoid crashing if the client is trying to upload many bytes and the
circuit gets torn down at the same time, or if the flip side
happens on the exit relay. Bugfix on 0.2.0.1-alpha; fixes bug 1150.
New config option "CircuitStreamTimeout" to override our internal
timeout schedule for how many seconds until we detach a stream from
a circuit and try a new circuit. If your network is particularly
slow, you might want to set this to a number like 60.
On this OSX version, there is a stub mlockall() function
that doesn't work, *and* the declaration for it is hidden by
an '#ifdef _P1003_1B_VISIBLE'. This would make autoconf
successfully find the function, but our code fail to build
when no declaration was found.
This patch adds an additional test for the declaration.
To fix a major security problem related to incorrect use of
SSL/TLS renegotiation, OpenSSL has turned off renegotiation by
default. We are not affected by this security problem, however,
since we do renegotiation right. (Specifically, we never treat a
renegotiated credential as authenticating previous communication.)
Nevertheless, OpenSSL's new behavior requires us to explicitly
turn renegotiation back on in order to get our protocol working
again.
Amusingly, this is not so simple as "set the flag when you create
the SSL object" , since calling connect or accept seems to clear
the flags.
For belt-and-suspenders purposes, we clear the flag once the Tor
handshake is done. There's no way to exploit a second handshake
either, but we might as well not allow it.
This commit implements a new config option: 'DisableAllSwap'
This option probably only works properly when Tor is started as root.
We added two new functions: tor_mlockall() and tor_set_max_memlock().
tor_mlockall() attempts to mlock() all current and all future memory pages.
For tor_mlockall() to work properly we set the process rlimits for memory to
RLIM_INFINITY (and beyond) inside of tor_set_max_memlock().
We behave differently from mlockall() by only allowing tor_mlockall() to be
called one single time. All other calls will result in a return code of 1.
It is not possible to change DisableAllSwap while running.
A sample configuration item was added to the torrc.complete.in config file.
A new item in the man page for DisableAllSwap was added.
Thanks to Moxie Marlinspike and Chris Palmer for their feedback on this patch.
Please note that we make no guarantees about the quality of your OS and its
mlock/mlockall implementation. It is possible that this will do nothing at all.
It is also possible that you can ulimit the mlock properties of a given user
such that root is not required. This has not been extensively tested and is
unsupported. I have included some comments for possible ways we can handle
this on win32.
If your relay can't keep up with the number of incoming create cells, it
would log one warning per failure into your logs. Limit warnings to 1 per
minute.
If all authorities restart at once right before a consensus vote, nobody
will vote about "Running", and clients will get a consensus with no usable
relays. Instead, authorities refuse to build a consensus if this happens.
When we tried to use the deprecated non-threadsafe evdns
interfaces in Libevent 2 without using the also-deprecated
event_init() interface, Libevent 2 would sensibly crash, since it
has no guess where to find the Libevent library.
Here we use the evdns_base_*() functions instead if they're
present, and fake them if they aren't.
This is a possible fix for bug 1023, where if we vote (or make a v2
consensus networkstatus) right after we come online, we can call
rep_hist_note_router_unreachable() on every router we haven't connected
to yet, and thereby make all their uptime values reset.
Don't count one-hop circuits when we're estimating how long it
takes circuits to build on average. Otherwise we'll set our circuit
build timeout lower than we should. Bugfix on 0.2.2.2-alpha.
Directory authorities now reject Tor relays with versions less than
0.1.2.14. This step cuts out four relays from the current network,
none of which are very big.
- Avoid memmoving 0 bytes which might lead to compiler warnings.
- Don't require relays to be entry node AND bridge at the same to time to
record clients.
- Fix a memory leak when writing dirreq-stats.
- Don't say in the stats files that measurement intervals are twice as long
as they really are.
- Reduce minimum observation time for requests to 12 hours, or we might
never record usage.
- Clear exit stats correctly after writing them, or we accumulate old stats
over time.
- Reset interval start for buffer stats, too.
The big change is to add a function to display the current SSL handshake
state, and to log it everywhere reasonable. (A failure in
SSL23_ST_CR_SRVR_HELLO_A is different from one in
SSL3_ST_CR_SESSION_TICKET_A.)
This patch also adds a new log domain for OR handshaking, so you can pull out
all the handshake log messages without having to run at debug for everything.
For example, you'd just say "log notice-err [handshake]debug-err file
tor.log".
If the networkstatus consensus tells us that we should use a
negative circuit package window, ignore it. Otherwise we'll
believe it and then trigger an assert.
Also, change the interface for networkstatus_get_param() so we
don't have to lookup the consensus beforehand.
To further attempt to fix bug 1090, make sure connection_ap_can_use_exit
always returns 0 when the chosen exit router is excluded. This should fix
bug1090.
When we excluded some Exits, we were sometimes warning the user that we
were going to use the node regardless. Many of those warnings were in
fact bogus, because the relay in question was not used to connect to
the outside world.
Based on patch by Rotor, thanks!
Tor now reads the "circwindow" parameter out of the consensus,
and uses that value for its circuit package window rather than the
default of 1000 cells. Begins the implementation of proposal 168.
This code adds a new field to vote on: "params". It consists of a list of
sorted key=int pairs. The output is computed as the median of all the
integers for any key on which anybody voted.
Improved with input from Roger.
Adding the same vote to a networkstatus consensus leads to a memory leak
on the client side. Fix that by only using the first vote from any given
voter, and ignoring the others.
Problem found by Rotor, who also helped writing the patch. Thanks!
A vote may only contain exactly one signature. Make sure we reject
votes that violate this.
Problem found by Rotor, who also helped writing the patch. Thanks!
Fix an obscure bug where hidden services on 64-bit big-endian
systems might mis-read the timestamp in v3 introduce cells, and
refuse to connect back to the client. Discovered by "rotor".
Bugfix on 0.2.1.6-alpha.
Add a "getinfo status/accepted-server-descriptor" controller
command, which is the recommended way for controllers to learn
whether our server descriptor has been successfully received by at
least on directory authority. Un-recommend good-server-descriptor
getinfo and status events until we have a better design for them.
We were telling the controller about CHECKING_REACHABILITY and
REACHABILITY_FAILED status events whenever we launch a testing
circuit or notice that one has failed. Instead, only tell the
controller when we want to inform the user of overall success or
overall failure. Bugfix on 0.1.2.6-alpha. Fixes bug 1075. Reported
by SwissTorExit.
When we added support for fractional units (like 1.5 MB) I broke
support for giving units with no space (like 2MB). This patch should
fix that. It also adds a propoer tor_parse_double().
Fix for bug 1076. Bugfix on 0.2.2.1-alpha.
We were triggering a CLOCK_SKEW controller status event whenever
we connect via the v2 connection protocol to any relay that has
a wrong clock. Instead, we should only inform the controller when
it's a trusted authority that claims our clock is wrong. Bugfix
on 0.2.0.20-rc; starts to fix bug 1074. Reported by SwissTorExit.
Note that unlike subversion revision numbers, it isn't meaningful to
compare these for anything but equality. We define a sort-order anyway,
in case one of these accidentally slips into a recommended-versions
list.
If any the v3 certs we download are unparseable, we should actually
notice the failure so we don't retry indefinitely. Bugfix on 0.2.0.x;
reported by "rotator".
Once we had called log_free_all(), anything that tried to log a
message (like a failed tor_assert()) would fail like this:
1. The logging call eventually invokes the _log() function.
2. _log() calls tor_mutex_lock(log_mutex).
3. tor_mutex_lock(m) calls tor_assert(m).
4. Since we freed the log_mutex, tor_assert() fails, and tries to
log its failure.
5. GOTO 1.
Now we allocate the mutex statically, and never destroy it on
shutdown.
Bugfix on 0.2.0.16-alpha, which introduced the log mutex.
This bug was found by Matt Edman.
The more verbose logs that were added in ee58153 also include a string
that might not have been initialized. This can lead to segfaults, e.g.,
when setting up private Tor networks. Initialize this string with NULL.
Send circuit or stream sendme cells when our window has decreased
by 100 cells, not when it has decreased by 101 cells. Bug uncovered
by Karsten when testing the "reduce circuit window" performance
patch. Bugfix on the 54th commit on Tor -- from July 2002,
before the release of Tor 0.0.0. This is the new winner of the
oldest-bug prize.
Relays no longer publish a new server descriptor if they change
their MaxAdvertisedBandwidth config option but it doesn't end up
changing their advertised bandwidth numbers. Bugfix on 0.2.0.28-rc;
fixes bug 1026. Patch from Sebastian.
Specifically, every time we get a create cell but we have so many already
queued that we refuse it.
Bugfix on 0.2.0.19-alpha; fixes bug 1034. Reported by BarkerJr.
The problem is that clients and hidden services are receiving
relay_early cells, and they tear down the circuit.
Hack #1 is for rendezvous points to rewrite relay_early cells to
relay cells. That way there are never any incoming relay_early cells.
Hack #2 is for clients and hidden services to never send a relay_early
cell on an established rendezvous circuit. That works around rendezvous
points that haven't upgraded yet.
Hack #3 is for clients and hidden services to not tear down the circuit
when they receive an inbound relay_early cell. We already refuse extend
cells at clients.
Changes to directory request statistics:
- Rename GEOIP statistics to DIRREQ statistics, because they now include
more than only GeoIP-based statistics, whereas other statistics are
GeoIP-dependent, too.
- Rename output file from geoip-stats to dirreq-stats.
- Add new config option DirReqStatistics that is required to measure
directory request statistics.
- Clean up ChangeLog.
Also ensure that entry guards statistics have access to a local GeoIP
database.
This new option will allow clients to download the newest fresh consensus
much sooner than they normally would do so, even if they previously set
FetchDirInfoEarly. This includes a proper ChangeLog entry and an updated man
page.
The internal error "could not find intro key" occurs when we want to send
an INTRODUCE1 cell over a recently finished introduction circuit and think
we built the introduction circuit with a v2 hidden service descriptor, but
cannot find the introduction key in our descriptor.
My first guess how we can end up in this situation is that we are wrong in
thinking that we built the introduction circuit based on a v2 hidden
service descriptor. This patch checks if we have a v0 descriptor, too, and
uses that instead.
arma's rationale: "I think this is a bug, since people intentionally
set DirPortFrontPage, so they really do want their relay to serve that
page when it's asked for. Having it appear only sometimes (or roughly
never in Sebastian's case) makes it way less useful."
Fixes bug 1013; bugfix on 0.2.1.8-alpha.
If the Tor is running with AutomapHostsOnResolve set, it _is_
reasonable to do a DNS lookup on a .onion address. So instead we make
tor-resolve willing to try to resolve anything. Only if Tor refuses
to resolve it do we suggest to the user that resolving a .onion
address may not work.
Fix for bug 1005.
Fix an edge case where a malicious exit relay could convince a
controller that the client's DNS question resolves to an internal IP
address. Bug found and fixed by "optimist"; bugfix on 0.1.2.8-beta.
Fix an edge case where a malicious exit relay could convince a
controller that the client's DNS question resolves to an internal IP
address. Bug found and fixed by "optimist"; bugfix on 0.1.2.8-beta.
Hidden service clients didn't use a cached service descriptor that
was older than 15 minutes, but wouldn't fetch a new one either. Now,
use a cached descriptor no matter how old it is and only fetch a new
one when all introduction points fail. Fix for bug 997. Patch from
Marcus Griep.
Apparently all the stuff that does a linear scan over all the DNS
cache entries can get really expensive when your DNS cache is very
large. It's hard to say how much this will help performance, since
gprof doesn't count time spent in OpenSSL or zlib, but I'd guess 10%.
Also, this patch removes calls to assert_connection_ok() from inside
the read and write callbacks, which are similarly unneeded, and a
little costlier than I'm happy with.
This is probably worth backporting to 0.2.0.
Provide a useful warning when launch_circuit tries to make us use a
node we don't want to use. Just give an info message when this is a
normal and okay situation. Fix for logging issues in bug 984.
This patch adds a function to determine whether we're in the main
thread, and changes control_event_logmsg() to return immediately if
we're in a subthread. This is necessary because otherwise we will
call connection_write_to_buf, which modifies non-locked data
structures.
Bugfix on 0.2.0.x; fix for at least one of the things currently
called "bug 977".
Tas (thanks!) noticed that when *ListenAddress is set, Tor would
still warn on startup when *Port is low and hibernation is active.
The patch parses all the *ListenAddress lines, and checks the
ports. Bugfix on 0.2.1.15-rc
- Write geoip stats to disk every 24 hours, not every hour.
- Remove configuration options and define reasonable defaults.
- Clear history of client requests every 24 hours (which wasn't done at
all before).
If we ever add an event, then set it, then add it again, there will be
now two pointers to the event in the event base. If we delete one and
free it, the first pointer will still be there, and possibly cause a
crash later.
This patch adds detection for this case to the code paths in
eventdns.c, and works around it. If the warning message ever
displays, then a cleverer fix is in order.
{I am not too confident that this *is* the fix, since bug 957 is very
tricky. If it is, it is a bugfix on 0.2.0.}
When we got a descriptor that we (as an authority) rejected as totally
bad, we were freeing it, then using the digest in its RAM to look up its
download status. Caught by arma with valgrind. Bugfix on 0.2.1.9-alpha.
The trick is that we should assert that our next_mem pointer has not
run off the end of the array _before_ we realign the pointer, since
doing that could take us over the end... but only if we're on a system
where malloc() gives us ram in increments smaller than sizeof(void*).
This matters because a cpuworker can close its socket when it
finishes. Cpuworker typically runs in another thread, so without a
lock here, we can have a race condition and get confused about how
many sockets are open. Possible fix for bug 939.
This might detect some possible causes of bug 930, and will at least
make sure we aren't doing some dumb memory-corruption stuff with the heap
and router-parsing.
(Don't crash immediately if we have leftover chunks to free after
freeing chunks in a buffer freelist; instead log a debugging message
that might help.)
Now, when you call tor --digests, it dumps the SHA1 digest of each
source file that Tor was built with. We support both 'sha1sum' and
'openssl sha1'. If the user is building from a tarball and they
haven't edited anything, they don't need any program that calculates
SHA1. If they _have_ modified a file but they don't have a program to
calculate SHA1, we try to build so we do not output digests.
bytes (aka 20KB/s), to match our documentation. Also update
directory authorities so they always assign the Fast flag to relays
with 20KB/s of capacity. Now people running relays won't suddenly
find themselves not seeing any use, if the network gets faster
on average.
svn:r19305
IP address changes: directory mirrors were mistakenly telling them
their old address if they asked via begin_dir, so they never got
an accurate answer about their new address, so they just vanished
after a day. Should fix bugs 827, 883, and 900 -- but alas, only
after every directory mirror has upgraded.
svn:r19291
ago. This change should significantly improve client performance,
especially once more people upgrade, since relays that have been
a guard for a long time are currently overloaded.
svn:r19287
The directory authorities were refusing v3 consensus votes from
other authorities, since the votes are now 504K. Fixes bug 959;
bugfix on 0.0.2pre17 (where we raised it from 50K to 500K ;).
svn:r19194
When we used smartlist_free to free the list of succesful uploads
because we had succeeded in uploading everywhere, we did not actually
set the successful_uploads field to NULL, so later it would get freed
again in rend_service_descriptor_free. Fix for bug 948; bug
introduced in 0.2.1.6-alpha.
svn:r19073
It seems that 64-bit Sparc Solaris demands 64-bit-aligned access to
uint64_t, but does not 64-bit-align the stack-allocated char array we
use for cpuworker tags. So this patch adds a set/get_uint64 pair, and
uses them to access the conn_id field in the tag.
svn:r18743
It turns out that we weren't updating the _ExcludeExitNodesUnion set's
country numbers when we reloaded (or first loaded!) the IP-to-country
file. Spotted by Lark. Bugfix on 0.2.1.6-alpha.
svn:r18575
stream never finished making its connection, it would live
forever in circuit_wait state. Now we close it after SocksTimeout
seconds. Bugfix on 0.1.2.7-alpha; reported by Mike Perry.
svn:r18516
Previously, when we had the chosen_exit set but marked optional, and
we failed because we couldn't find an onion key for it, we'd just give
up on the circuit. But what we really want to do is try again, without
the forced exit node.
Spotted by rovv. Another case of bug 752. I think this might be
unreachable in our current code, but proposal 158 could change that.
svn:r18451
GCC's interpretation of the C99 aliasing rules, to be charitable,
creates a dialect of C intended for a better programmers than I am
certain of my ability to be in all times. I just spent 2 hours
tracking down a platform-hyperspecific libevent bug that turned out to
be because of this, and darned if I ever want to do *that* again.
One of Linus's recent rants will give you a picture of why GCC's
behavior here can lead to fun surprises in your binaries:
http://lwn.net/Articles/316126/
svn:r18351
This resolves bug 526, wherein we would crash if the following
events occurred in this order:
A: We're an OR, and one of our nameservers goes down.
B: We launch a probe to it to see if it's up again. (We do this hourly
in steady-state.)
C: Before the probe finishes, we reconfigure our nameservers,
usually because we got a SIGHUP and the resolve.conf file changed.
D: The probe reply comes back, or times out. (There is a five-second
window for this, after B has happens).
IOW, if one of our nameservers is down and our nameserver
configuration has changed, there were 5 seconds per hour where HUPing
the server was unsafe.
Bugfix on 0.1.2.1-alpha. Too obscure to backport.
svn:r18306
This fixes the last known case of bug 891, which could happen if two
hosts, A and B, disagree about how long a circuit has been open,
because of clock drift of some kind. Host A would then mark the
connection as is_bad_for_new_circs when it got too old and open a new
connection. In between when B receives a NETINFO cell on the new
conn, and when B receives a conn cell on the new circuit, the new
circuit will seem worse to B than the old one, and so B will mark it
as is_bad_for_new_circs in the second or third loop of
connection_or_group_set_badness().
Bugfix on 0.1.1.13-alpha. Bug found by rovv.
Not a backport candidate: the bug is too obscure and the fix too tricky.
svn:r18303
crypto_global_init gets called. Also have it be crypto_global_init
that calls crypto_seed_rng, so we are not dependent on OpenSSL's
RAND_poll in these fiddly cases.
Should fix bug 907. Bugfix on 0.0.9pre6. Backport candidate.
svn:r18210
are stored when the --enable-local-appdata option is configured. This
changes the Windows path from %APPDATA% to a host local
%USERPROFILE%\Local Settings\Application Data\ path (aka,
LOCAL_APPDATA).
Patch from coderman.
svn:r18122
Unfortunately, old Libevents don't _put_ a version in their headers, so
this can get a little tricky. Fortunately, the only binary-compatibility
issue we care about is the size of struct event. Even more fortunately,
Libevent 2.0 will let us keep binary compatiblity forever by letting us
decouple ourselves from the structs, if we like.
svn:r18014
five days old. Otherwise if Tor is off for a long time and then
starts with cached descriptors, it will try to use the onion
keys in those obsolete descriptors when building circuits. Bugfix
on 0.2.0.x. Fixes bug 887.
svn:r17993
cell back), avoid using that OR connection anymore, and also
tell all the one-hop directory requests waiting for it that they
should fail. Bugfix on 0.2.1.3-alpha.
svn:r17984
