{5} Accepted, Active Tickets by Owner (Full Description) (112 matches)

The existing GoodBadISP tables need sorting into the new format. All opinions, feedback and communications to that ISP must go in the correct section on ISPCorrespondence page to keep the primary page clean and to the point since it will grow substantially over time.

The new format should be available soon after this ticket is posted as it will be done for the US hosts (good experiences).

The Tor Browser User Manual on ​https://tb-manual.torproject.org/windows/en-US/ needs a section covering

a) where to find the source, b) how it is licensed, and c) where and how to report bugs.

Cheers,

The two text files currently in the doc directory are doc/HACKING and doc/TUNING. The latter is the only one that deals with relay operation, and its subject is oddly specific: increasing the maximum number of file descriptors. If we're going to put critical documentation in the codebase, I think it would also be worthwhile to have a basic hardening guide. It could include suggestions like:

• allowing only public key non-root SSH login
• using a firewall
• keeping your system up-to-date
• not running any other programs (especially networked ones)
• considering hardened or security-focused OS choices

Nick suggested that most of the actual information be contained in referenced links, which I agree with. There's no good reason to duplicate effort when there are, for example, so many good SSH hardening guides.

Let me know what you think, or if you have any contributions. If this is generally considered a good idea, I can start writing a draft.

When trying to execute

tor.exe --service install --options -f C:\absolute\path\to\torrc

an <unformattable error> is given:

Running on a Post-Win2K OS, so we'll assume that the LocalService account exists.
Done with CreateService.
Service installed successfully
Service failed to start : <unformattable error>

This bug happens when there is something written in the torrc file, but on other attempts it was happening when using

HiddenServiceAuthorizeClient basic PCNAME

and so deleting this line it has been starting again.

If I leave the torrc file empty the error doesn't show up and the service is installed and working.

I tried to differently formatting torrc file, so using LF and CRLF, but the result is the same.

Have a look at the tor_spawn_background unit tests. That's sure a lot of #ifdefs! It would be nice if our portability code actually let us write code to be portable across platforms: we should fix tor_spawn_background and tor_read_all_handle to act the same across platforms.

While testing a fix for #11879, Kathy and I noticed that if the bootstrap process is interrupted by setting DisableNetwork=1 via the control port, Tor waits about a minute after DisableNetwork is set back to 0 before continuing network activity. We observed this problem on a Mac OS 10.8.5 system. Possibly related tickets: #9229, #11069.

Once release candidates for Tor Browser 4.5 are available, this should be reproducible by following these steps:

1. Start Tor Browser and click "Connect".
2. Click "Open Settings" in the connection progress window to interrupt the bootstrap process.
3. Click "Connect" again. Notice that there is a delay before the bootstrap makes more progress.

We are also able to reproduce it using Tor 0.2.6.6 and a manual (telnet) control port connection. Follow these steps (control port authentication is up to you):

1. Remove all cached Tor data and start Tor like this:

./tor --defaults-torrc torrc-defaults -f torrc DisableNetwork 1

2. Make a control port connection and issue this command:

SETCONF DisableNetwork=0

3. Wait for bootstrapping to reach 25-50% and then do:

SETCONF DisableNetwork=1

4. Re-enable network access:

SETCONF DisableNetwork=0 Notice that there is a delay before the bootstrap makes more progress.

We used the torrc-defaults file that ships with Tor Browser 4.5a5:

# If non-zero, try to write to disk less frequently than we would otherwise.
AvoidDiskWrites 1
# Where to send logging messages.  Format is minSeverity[-maxSeverity]
# (stderr|stdout|syslog|file FILENAME).
Log notice stdout
# Bind to this address to listen to connections from SOCKS-speaking
# applications.
SocksPort 9150
ControlPort 9151
CookieAuthentication 1
## fteproxy configuration
ClientTransportPlugin fte exec PluggableTransports/fteproxy.bin --managed

## obfs4proxy configuration
ClientTransportPlugin obfs2,obfs3,obfs4,scramblesuit exec PluggableTransports/obfs4proxy

## flash proxy configuration
#
# Change the second number here (9000) to the number of a port that can
# receive connections from the Internet (the port for which you
# configured port forwarding).
ClientTransportPlugin flashproxy exec PluggableTransports/flashproxy-client --register :0 :9000

## meek configuration
ClientTransportPlugin meek exec PluggableTransports/meek-client-torbrowser -- PluggableTransports/meek-client

Our torrc is also from Tor Browser and it just contains a few paths:

DataDirectory /Users/.../tb-11879.app/TorBrowser/Data/Tor
GeoIPFile /Users/.../tb-11879.app/TorBrowser/Data/Tor/geoip
GeoIPv6File /Users/.../tb-11879.app/TorBrowser/Data/Tor/geoip6

I will attach some log output.

In #8243 we started requiring Stable flag for becoming HSDirs, but this is still not hard enough for motivated adversaries. Hence we need to make it even harder for a relay to become HSDir, so that only relays that have been around for long get the flag. After prop224 gets deployed, there will be less incentive for adversaries to become HSDirs since they won't be able to harvest onion addresses.

Until then, our current plan is to increase the bandwidth and uptime required to become an HSDir to something almost unreasonable. For example requiring an uptime of over 6 months, or maybe requiring that the relay is in the top 1/4th of uptimes on the network.

The parameter WarnUnsafeSocks does not work as specified in the documentation, no warning is logged in the log file when a connection is done to an ip address.

If WarnUnsafeSocks 1 (default) is set there is no warning in the log file. If you look at the code for log_unsafe_socks_warning, the only case where an error is logged is when safe_socks is true. safe_socks is true only when SafeSocks parameter is set, but not when WarnUnsafeSocks is set.

The code should be

if (safe_socks || options->WarnUnsafeSocks) {

instead of

if (safe_socks) {

During the last beta release we realized that some translators translate "about:tor" which breaks it. We should write a test that checks this crucial page is working in built bundles.

If i specify SessionGroup as described in manual. tor stops with error.

torrc:

SocksPort 9051 SessionGroup=1

Jun 15 16:46:24.700 [warn] Invalid SocksPort option '"SessionGroup=INT"'

Jun 15 16:46:24.700 [warn] Failed to parse/validate config: Invalid SocksPort/SocksListenAddress configuration

Jun 15 16:46:24.701 [err] Reading config failed--see warnings above.

second try

SocksPort 9051 SessionGroup=INT

Results into:

Jun 15 16:46:35.677 [warn] Invalid SocksPort option '"SessionGroup=1"'

Jun 15 16:46:35.678 [warn] Failed to parse/validate config: Invalid SocksPort/SocksListenAddress configuration

Jun 15 16:46:35.677 [warn] Invalid SocksPort option '"SessionGroup=1"'

Jun 15 16:46:35.678 [warn] Failed to parse/validate config: Invalid SocksPort/SocksListenAddress configuration

Jun 15 16:46:35.678 [err] Reading config failed--see warnings above.

Tor increases data usage, and most of the demographics that need Tor the most have very limited data, so please make PTs available for zero-rated services.

For example FreedomPop zero-rates WhatsApp (​https://9to5mac.com/2016/08/17/freedompop-whatsapp-sim-free-iphone/) in 30 countries on a free ($0/month) plan.

A WhatsApp plugable transport would therefor allow users in poverty-stricken, massively oppressed countries to have unlimited free access to information through Tor.

I'm not sure if the component should be Orbot or Tor; the WhatsApp example is for mobile but there is also zero-rating for house internet, or someone might want to use a WhatsApp PT on a desktop/laptop/notebook with conventional Tor, tethered to a smartphone for the zero-rated protocol.

As an example, KIST controls how much and when channel data is pushed on the network which means this wants to write event used by the Vanilla scheduler as "queued cell but no space on the outbuf" is not something that is coherent with KIST.

A channel is scheduled in when it has cells in the queue, then they are flushed one by one by the KIST scheduler until the kernel says "no more space". Then, that channel is put back in the channel pending list and will get handled at the next tick of KIST.

So, we really don't need KIST to be notified of this event from connection_flushed_some() which is used by Vanilla to try to flush more cells in the outbuf because the outbuf has room for it.

Where it is useful is the second callsite of that even in channel_tls_handle_state_change_on_orconn() which notifies the scheduler that it might be in need of flushing some stuff. In the case of a brand new channel just opening, its state is "IDLE" and that even will then put it in "waiting for cells" after.

That being said, what needs to happened:

1. Make the notification event a per-scheduler thing because KIST and Vanilla have different semantic for those events really. We should of course avoid as much as we can of code duplication and thus some "generic event handler" do make sense if they share the same semantic.

2. Add a "channel state is open" notification event instead of "wants to write" which is really only true in very specific cases in channel_tls_handle_state_change_on_orconn(). That way, the scheduler can take a decision on the status of the channel instead of blind guessing it is waiting for cells.

3. Nullify the "wants to write" event for KIST considering (2) so it stops possibly scheduling channels that do not need at all to be scheduled.

Recently Tor added PT_LOG and PT_STATUS events to the spec...

​https://gitweb.torproject.org/torspec.git/commit/?id=3028cf1 ​https://gitweb.torproject.org/torspec.git/commit/?id=b38257e

Unfortunately the 'pt-spec.txt section 3.3.5' section they mention does not exist, and in looking around I can't find anything that describes what these event fields are defined as ('PT=' 'TYPE=', 'CONNECT=', etc).

I started to write a stem parser for these but can't continue until this is done (I can't parse events without knowing what fields they include).

David is aware of this and plans to has kindly offered to add the missing info...

22:24 <+atagar> dgoulet: Your control-spec addition to descript PT_LOG and PT_STATUS
                cite a pt-spec section 3.3.4 which does not exist.
22:24 <+atagar> s/descript/describe
22:29 <+atagar> dgoulet: Huh. I'm not spotting anything that lists the keyword
                arguments ('PT=' and 'SEVERITY=') so guess the sections simply
                missing from the spec. I need that for stem support so please
                give me a nudge when the event spec's done. :)
22:59 <+dgoulet> atagar: oh hmmm I'll fix that sorry
23:17 <+atagar> Thanks! Much appreciated. :)

the Linux 'capabilities' library for allowing non-root users to perform tasks which normally require elevated privileges.

at present the torsocks wrappers have checked for setuid and setgid flags on the binaries it executes and failed closed, throwing an error if this occurs, however there is currently no check to see if the binaries have capabilities applied.

in the case where they do, the LD_PRELOAD set by torsocks is stripped and the program will execute with no warning and without the torsocks wrapper.

as an example of this, the current 'ping' command on my Linux is setcap:

$ getcap which ping /usr/bin/ping = cap_net_raw+ep $ torsocks ping -c 1 torproject.org PING torproject.org (82.195.75.101) 56(84) bytes of data. 64 bytes from 82.195.75.101: icmp_seq=1 ttl=50 time=38.1 ms

​https://projects.archlinux.org/svntogit/packages.git/tree/trunk/iputils.install?h=packages/iputils

We have a branch bug15618_030-testing at ​https://gitweb.torproject.org/user/dgoulet/tor.git that logs a warning when we call more than 2 times circuit has_opened() on a circuit. The patch was made to confirm the theory on #15618 related to a warning we see on relays. While until now I never saw the warning on relay side, I see it quite often on client side.

Based on the log messages, I think this is related to introduction point circuits and not rendezvous circuits as we thought. Basically it happens when we open a 4 hop circuit with last hop X, and immediately we extend that circuit to a 5 hop one with last hop Y, where Y is (I think) the introduction point. I got the warning 2 times in a row and the similarity between them is that they are 5 hop circuits with the last two hops 4 and 5 the same, but different hops 2 and 3.

Dec 25 13:15:15.000 [info] internal circ (length 5, last hop $D7316BF7FD633DD7474B18C33E1D5FDEB04D26A7): $A7C411B809D0AA98C4264917BB701ABF17B2181E(open) $B5212DB685A2A0FCFBAE425738E478D12361710D(open) $123F403DA94A74F959B7FE0E5B27FA57EFA12925(open) $1F4105C688E835A56AF3D66C787677B57240FFA2(open) $D7316BF7FD633DD7474B18C33E1D5FDEB04D26A7(open)
Dec 25 13:15:15.000 [info] entry_guards_note_guard_success(): Recorded success for primary confirmed guard sakujakira ($A7C411B809D0AA98C4264917BB701ABF17B2181E)
Dec 25 13:15:15.000 [info] circuit_send_next_onion_skin(): circuit built!
Dec 25 13:15:15.000 [warn] BLAM. Circuit has_opened() called 3 times.
Dec 25 13:15:15.000 [info] rend_client_introcirc_has_opened(): introcirc is open
Dec 25 13:15:15.000 [info] connection_ap_handshake_attach_circuit(): pending-join circ 3419634369 already here, with intro ack. Stalling. (stream 26 sec old)
Dec 25 13:15:15.000 [info] connection_ap_handshake_attach_circuit(): ready rend circ 2404371907 already here (no intro-ack yet on intro 2259155742). (stream 26 sec old)
Dec 25 13:15:15.000 [info] connection_ap_handshake_attach_circuit(): found open intro circ 2259155742 (rend 2404371907); sending introduction. (stream 26 sec old)
Dec 25 13:15:15.000 [info] rend_client_send_introduction(): Sending an INTRODUCE1 cell
Dec 25 13:15:15.000 [info] pathbias_count_use_attempt(): Used circuit 53293 is already in path state use succeeded. Circuit is a Hidden service client: Pending rendezvous point currently open.

Dec 25 13:15:16.000 [info] internal circ (length 5, last hop $D7316BF7FD633DD7474B18C33E1D5FDEB04D26A7): $A7C411B809D0AA98C4264917BB701ABF17B2181E(open) $5C01D5518D1F5A071C6F07D1F4630F577AB5B60A(open) $DA9DA02A7B0565DF5F3E961CA911E750072DCBBD(open) $1F4105C688E835A56AF3D66C787677B57240FFA2(open) $D7316BF7FD633DD7474B18C33E1D5FDEB04D26A7(open)
Dec 25 13:15:16.000 [info] entry_guards_note_guard_success(): Recorded success for primary confirmed guard sakujakira ($A7C411B809D0AA98C4264917BB701ABF17B2181E)
Dec 25 13:15:16.000 [info] circuit_send_next_onion_skin(): circuit built!
Dec 25 13:15:16.000 [warn] BLAM. Circuit has_opened() called 3 times.
Dec 25 13:15:16.000 [info] rend_client_introcirc_has_opened(): introcirc is open
Dec 25 13:15:16.000 [info] connection_ap_handshake_attach_circuit(): pending-join circ 3419634369 already here, with intro ack. Stalling. (stream 26 sec old)
Dec 25 13:15:16.000 [info] connection_ap_handshake_attach_circuit(): Intro circ 2259155742 present and awaiting ack (rend 2404371907). Stalling. (stream 26 sec old)
Dec 25 13:15:16.000 [info] connection_ap_handshake_attach_circuit(): Intro circ 3200472182 present and awaiting ack (rend 4286776399). Stalling. (stream 26 sec old)
Dec 25 13:15:16.000 [info] connection_ap_handshake_attach_circuit(): ready rend circ 2334828448 already here (no intro-ack yet on intro 4002828471). (stream 27 sec old)
Dec 25 13:15:16.000 [info] connection_ap_handshake_attach_circuit(): found open intro circ 4002828471 (rend 2334828448); sending introduction. (stream 27 sec old)
Dec 25 13:15:16.000 [info] rend_client_send_introduction(): Sending an INTRODUCE1 cell
Dec 25 13:15:16.000 [info] pathbias_count_use_attempt(): Used circuit 53301 is already in path state use succeeded. Circuit is a Hidden service client: Pending rendezvous point currently open.

Here is how we make IPv6 (and other unreachable addresses) work with single-hop client and service connections to intro and rend points. It works for v2 single onion services. We talked about it for v3, but it never made it into the prop224 spec.

Here are the steps:

0. The service chooses and connects to the intro point (possibly using a 3-hop path if it is a single onion service and can't reach it directly)
1. The service always puts IPv4 and IPv6 in its descriptor link specifiers (if they are available in directory documents)
2. If the link specifier has a reachable address, and the service is not a single onion service, a Tor2web client (currently v2 only) can use it to make a direct connection to the intro point
3. Otherwise, the client connects over a 3-hop path via one of its reachable entry nodes

The process for client rendezvous is similar, but if the client knows that the service is a single onion service, it *must* connect to the rend point using a 3-hop path. (Again, this only matters for Tor2web, which is v2 only).

If you look at circuitmux_append_destroy_cell(), it is the one appending a DESTROY cell to the cmux queue and then calls channel_flush_from_first_active_circuit() if no writes are pending that is if the outbuf is empty (also looks at the out queue but that is always empty #23709).

In the case the flush is triggered, the cell is immediately put in the outbuf and written to kernel by libevent which completely bypasses the scheduler. Maybe it is what we want that is go as fast as we can in destroying a circuit? Don't know but it has this effect on the scheduler where the channel is scheduled with a "wants_to_write" event from the connection subsystem and ultimately the channel gets scheduled with nothing in the queue because it is already on the outbuf. For KIST, this is not ideal because KIST should control the flow of data to the kernel.

It seems there are two places we queue cells into a cmux queue: circuitmux_append_destroy_cell() and append_cell_to_circuit_queue(). The latter triggers a "has waiting cells" for the scheduler which is what we want but the former just bypasses it.

I think it should simply trigger that notify to the scheduler instead of flushing it by itself.

Prop224 says:

The hidden service SHOULD NOT reject any LSTYPE fields which it
doesn't recognize; instead, it should use them verbatim in its EXTEND
request to the rendezvous point.

​https://gitweb.torproject.org/torspec.git/tree/proposals/224-rend-spec-ng.txt#n1689

We should either remove this from the spec, or we should:

• add a similar sentence for client descriptor lspecs
• put unrecognised lspecs in descriptors in client intro EXTEND requests
• put unrecognised lspecs in INTRODUCE cells in service rend EXTEND requests

recvmsg() is supported as of now. A full exit should be done here because Torsocks can't handle this inet socket with Tor.

This warning is possible for anything socket trying to connect to a localhost address.

WARNING torsocks[12360]: [connect] Connection to a local address are denied since it might be a TCP DNS query to a local DNS server. Rejecting it for safety reasons. (in tsocks_connect() at connect.c:177)

We should implement a whitelist mechanism so the user can tell which local network is allowed such as localhost.

Multiple issues need FD passing through a Unix socket to work: #8585, #16183

It's maybe possible to support this safely. My intuition is that we might be able to get it work by passing some cookies in the ancillary data so we can recognize the sendmsg() with the recvmsg(). Maybe!?...

Clients should put IPv6 link specifiers in the EXTEND cells to relays they choose from directory documents.

We need to do this in the same releases as #24181, to avoid adding an v3 onion service distinguisher.

What torsocks does: Routes all traffic through Tor.

What it should do: It shoud have an option to route .onion domains through Tor, while normal traffic is not routed through Tor.

Advantages This would allow Mail/XMPP servers to connect to .onion domains, without any configuration hassle.

Original discussion ​http://tor-talk.torproject.narkive.com/j7MtPG5T/torsocks-usewithtor-only-for-onion

For several years everyone was able to post on ​https://blog.torproject.org without enabling JavaScript and other dangerous things.

Observed behaviour: can not post unless slider set to medium or low Expected behaviour: high security supported Steps to reproduce: try to post at ​https://blog.torproject.org with security slider on high

During the meeting in CDMX some of us chatted about the possibility to build and maintain a sort of Tor web docs (like the mozilla web docs [1]).

The idea is to collect all the techniques we use to make websites tor-browser and privacy friendly in a single place. We could also include the reasoning behind doing certain things in pure css vs js. Or why we decide to do things in a certain way.

Some topic that I have been thinking about are:

• Why tor browser is slightly different from Firefox (or another browser)

• Why does my app work differently in tor browser?

• How can I make my app compatible for people that do not use JS?

• Code examples for css and js

• Server side website programming, what to keep in mind...

While all these topics are documented in various articles around the interwebs, I kinda think we (tor community) should own something like this, since it would also help to push forward the idea that the web as we know it needs to change.

Up to now I have always thought this sort of content belonged to the styleguide. Recently I have been thinking that while the individual implementation details of what use in our websites can be included in the styleguide, the reasoning behind those and the general implementations should be put somewhere else.

Also I do not think these topics belong directly to the dev portal, as I tend to think that should be about developing on tor rather than considering how the web works a bit differently when using tor browser. Unless we would rather do a specific section regarding all this in the web portal.

Further things to consider:

• Kevin (on cc) is working on a Tor friendliness scanner tool for websites.
• We see often threads like this [2] on our mailing lists. This one is about how to avoid privacy leaks for onion services, but we might identify more similar best practice tips being discussed.

[1] ​https://developer.mozilla.org/en-US/ [2] ​https://lists.torproject.org/pipermail/tor-onions/2018-August/000295.html

I know this is basically just Debian instructions, but framing this differently could really help for those that don't already know that Raspbian is essentially just Debian underneath.

This could be deferred until after the big website changes.

This will allow easy update to conform to the styleguide and make it easier to edit the content, and provide translations if desired.

RS uses "AS0" as a placeholder for "unknown AS number", but "AS0" has a special meaning in BGP (​https://tools.ietf.org/html/rfc7607 ), lets use something else?

maybe "ASXY"?

According to ​https://svn.torproject.org/cgi-bin/viewvc.cgi/Tor/ , it's been at least six months since anything changed on the public SVN repository.

Most of the repositories there have been migrated to git.

We could migrate all the rest to git (if appropriate) or archive them (if appropriate), and turn off the lights on public svn.

The ultimate direction we want to go is towards an adblock plus model, where people can subscribe to rule feeds that are relevant to them, maintained by third parties. This involves both altering our XML schema to include a 'rulefeed' envelope tag, and adding a bit of UI to add and manage subscription urls.

It also depends upon a few enhancements being completed first. These are in the child ticket list below:

A user (DEplan on #guardianproject) reported that Gibberbot was using his real IP despite Orbot's transproxy being turned on; further research led to the conclusion that recent releases of Android seem to use IPv4-mapped IPv6 adresses for a large portion of connections. For examples, please see ​http://pastebin.com/Z4KDDq40. These connections completely bypass transproxy.

I am not yet sure about the circumstances under which Android employs these addresses.

The problems in finding a solution are that Android usually does not include ip6tables (though Orbot could simply package that) and kernels do usually not include IPv6 netfilter modules. The latter is a major issue, since Orbot can't package modules for every single kernel a user might be running.

As a side note, IPv6 does not support NAT (which is what transproxying is based on).

I'll try to figure out what triggers this behaviour of Android and find possible solutions (using sysctl to disable IPv6 does not solve it).

Behaviour: When closing tor network with big Button and exiting Orbot after tor is "deactivated", privoxy is still running and the Orbot service is not stopped.

Actions:

• Killing Privoxy from shell stops the privoxy process (OK)
• Killing Orbot process simply restarts the process (BAD)

Env:

• Running Orbot v1.0.4.1
• Android Froyo 2.2.1 speedmod kernel
• Samsung Galaxy

I'm using Orbot (v0.2.3.10-alpha-1.0.7-FINAL, on Android ICS v4.0.1) and I can't seem to get the exit node I request. In the Exit and Entrance Node fields I have "{us}" entered, yet sometimes I get IP's outside the US. Yesterday I got a UK ip.

Also, at random (usually after 30 minutes or so) I seem to lose connection to the Tor network without Orbot notifying me. I'm using Pandora from Canada.

Our existing OOS code kills low-priority OR connections. But really, we need to look at all connections that an adversary might be able to create (especially dir and exit connections), or else an adversary will be able to open a bunch of those, and force us to kill as many OR connections as they want.

This problem is the reason that DisableOOSCheck is now on-by-default.

It would be handy if the following roles were split up:

1) The list of IP:Orport:Identity to which every relay should upload every descriptor. 2) The list of IP:Orport:Identity from which caches should expect to find canonical consensuses and descriptors. 3) The list of IP:Orport:Identity from which non-caches should expect to bootstrap consensuses and descriptors. (See 'fallbackdir') 4) The list of keys that must sign a vote or a consensus. 5) The list of IP:Orport:Identity that authorities use when sending and receiving votes.

Splitting roles up in this way would better prepare us for an implementation of prop#257 down the road.

Once upon a time, the tor spec files and proposals were in the Tor tarball, so they clearly were covered under Tor's copyright license (3-clause BSD).

But now they're in their own git repository.

I think maybe they technically have no copyright license now?

We should pick one (I vote cc-by) and try to apply it. This plan could become tricky for proposals, because there are a lot of authors on proposals by now.

Parent ticket for transitioning current and future client versions off the tor network with a minimal amount of pain.

From my prop271 branch:

 * XXXX prop271 this function is used in four ways: picking out guards for
 *   the old (pre-prop271) guard algorithm; picking out guards for circuits;
 *   picking out guards for testing circuits on non-bridgees;
 *   picking out entries when entry guards are disabled.  These options
 *   should be disentangled.

I'm running tor on Gentoo Hardened. The bug exists in 0.2.6.7 and 0.2.7.1-alpha. tor crashes within seconds of starting, before any clients can connect I think.

Jul 14 13:13:07.000 [notice] Tor 0.2.7.1-alpha (git-df76da0f3bfd6897) opening log file.
Jul 14 13:13:07.182 [notice] Tor v0.2.7.1-alpha (git-df76da0f3bfd6897) running on Linux with Libevent 2.0.22-stable, OpenSSL 1.0.1p and Zlib 1.2.8.
Jul 14 13:13:07.182 [notice] Tor can't help you if you use it wrong! Learn how to be safe at https://www.torproject.org/download/download#warning
Jul 14 13:13:07.182 [notice] This version is not a stable Tor release. Expect more bugs than usual.
Jul 14 13:13:07.182 [notice] Read configuration file "/etc/tor/torrc".
Jul 14 13:13:07.187 [notice] Opening Socks listener on 127.0.0.1:9050
Jul 14 13:13:07.187 [notice] Opening Socks listener on 127.0.0.1:9056
Jul 14 13:13:07.187 [notice] Opening Socks listener on 127.0.0.1:9055
Jul 14 13:13:07.187 [notice] Opening Control listener on 127.0.0.1:9015
Jul 14 13:13:07.000 [notice] Parsing GEOIP IPv4 file /usr/share/tor/geoip.
Jul 14 13:13:07.000 [notice] Parsing GEOIP IPv6 file /usr/share/tor/geoip6.
Jul 14 13:13:07.000 [notice] Bootstrapped 0%: Starting

============================================================ T= 1436875987
(Sandbox) Caught a bad syscall attempt (syscall socket)
/usr/bin/tor(+0x142148)[0x4bb7bc8148]
/lib64/libc.so.6(socket+0x7)[0x3adc706ea07]
/lib64/libc.so.6(socket+0x7)[0x3adc706ea07]
/lib64/libc.so.6(+0xf16a0)[0x3adc70686a0]
/lib64/libc.so.6(__vsyslog_chk+0x3ef)[0x3adc7068aff]
/lib64/libc.so.6(__syslog_chk+0x89)[0x3adc7068df9]
/usr/bin/tor(+0x135bb0)[0x4bb7bbbbb0]
/usr/bin/tor(tor_log+0xd0)[0x4bb7bbc680]
/usr/bin/tor(control_event_bootstrap+0x1e4)[0x4bb7b7ba74]
/usr/bin/tor(do_main_loop+0x84)[0x4bb7abe234]
/usr/bin/tor(tor_main+0x16c5)[0x4bb7ac1225]
/lib64/libc.so.6(__libc_start_main+0x114)[0x3adc6f97134]
/usr/bin/tor(+0x34519)[0x4bb7aba519]

$ uname -r
3.18.9-hardened

This bug has been reported downstream: ​https://bugs.gentoo.org/show_bug.cgi?id=550302. It occurs with the following torrc:

#
# Minimal torrc so tor will work out of the box
#
User tor
PIDFile /var/run/tor/tor.pid
Log notice syslog
Log notice file /var/log/tor/log
DataDirectory /var/lib/tor/data
SandBox 1

SocksPort 9050
SocksPort 9056 IsolateDestAddr IsolateDestPort
SocksPort 9055

ControlPort 9015
CookieAuthentication 1

By commenting out "Sandbox 1" or unsetting it, tor will obviously run without crashing.

We're creating a vote that is invalid, but try to make a consensus anyway like nothing's wrong. Then we fail doing that as described above.

The following bug warnings should probably be protocol warnings, or be caught earlier:

./src/test/fuzz_static_testcases.sh
Running tests for consensus
Feb 16 02:17:26.012 [warn] tor_timegm: Bug: Out-of-range argument to tor_timegm (on Tor 0.3.0.3-alpha-dev d633c4757c1392fb)
Feb 16 02:17:30.133 [warn] sr_parse_commit: Bug: SR: Commit algorithm "sha6-256" is not recognized. (on Tor 0.3.0.3-alpha-dev d633c4757c1392fb)
Feb 16 02:17:34.625 [warn] commit_decode: Bug: SR: Commit from authority 20EE989200EF98A75102B461DF62F01B2932C0D6 decoded length doesn't match the expected length (36 vs 40). (on Tor 0.3.0.3-alpha-dev d633c4757c1392fb)
Feb 16 02:17:40.758 [warn] commit_decode: Bug: SR: Commit from authority B0F141F4B8CCBCC328572C71E5590BBA19775594 can't be decoded. (on Tor 0.3.0.3-alpha-dev d633c4757c1392fb)
Running tests for descriptor
Feb 16 02:18:08.780 [warn] tor_timegm: Bug: Out-of-range argument to tor_timegm (on Tor 0.3.0.3-alpha-dev d633c4757c1392fb)
Running tests for extrainfo
Feb 16 02:18:18.548 [warn] tor_timegm: Bug: Out-of-range argument to tor_timegm (on Tor 0.3.0.3-alpha-dev d633c4757c1392fb)
...

I'm not sure whether we need this in 030, but it would be nice to fix them eventually.

I have a bridge in my recent Tor (post prop#271), and my state file now says:

Guard in=bridges rsa_id=C7BE8154678E7537CCAC60B097D51A8A7EF8BCDF bridge_addr=94.242.249.2:94 sampled_on=2017-05-04T10:39:09 sampled_by=0.3.1.0-alpha-dev unlisted_since=2017-05-11T05:47:41 listed=0 confirmed_on=2017-05-02T01:15:03 confirmed_idx=0 pb_use_attempts=1.000000 pb_use_successes=1.000000 pb_circ_attempts=11.000000 pb_circ_successes=11.000000 pb_successful_circuits_closed=11.000000

The unlisted_since and listed=0 are not surprising, since bridges will never be listed, right? Does that make my Tor do anything I shouldn't want it to do, like throw out the bridge line sooner than I'd want? I think no, because hey if it's in torrc it'll go back in, but maybe throwing out the path bias info early was not what we expect?

When upgrading Briar's Tor binaries from 0.2.9.16 to 0.3.4.8, we noticed a difference in Tor's startup time on older Android phones.

Measuring the startup time of several recent Tor versions revealed an interesting pattern. The time that elapses between starting the Tor process and the creation of the authentication cookie file hasn't changed across versions, but the time between the creation of the cookie file and the response to the AUTHENTICATE command has changed substantially. (Briar sends the AUTHENTICATE command as soon as the cookie file's created.)

I measured five runs of each version on a Motorola Moto G 4G running Android 5.1. Here are the min and max times in seconds for each version:

The min and max have both increased substantially since 0.2.9, and the distribution has widened. This is having a noticeable impact on how long it takes for Briar to connect to contacts when the app's started.

I'll repeat these experiments on Linux x64 to see whether this is Android-specific.

By Nov 2016, we have a deliverable to get our route selection much more right than today, and to have it very tested. We should get this done significantly earlier.

https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/SecurityPolicy

In #22948, we discovered that the relay integrity digest was easier to guess than it should be. This makes the following classes of attacks easier:

• sending bandwidth and guessing the integrity digest, and
• modifying cells and manipulating the integrity digest.

Outright hostname lookup failures for previously configured hidden services prevent other options being set while DNS is down.

For example, I configure a hidden service redirecting to google.com while DNS is working. DNS subsequently stops working, e.g. nameserver becomes completely unreachable. If I then attempt to set a config option using the controller, it will not get set as long as tor cannot resolve the hidden service name.

Rejection of hidden service configurations (and hence any subsequent or unrelated config change) made while tor is running needs to be more tolerant of lookup failures.

The following attempts to validate the hidden service config currently in use (and previously validated when DNS was working). If the validation fails, it must be because DNS is down, so the existing config is retained. If the user was attempting to add a new hidden service config, then it doesn't get added.

Index: src/or/config.c
===================================================================
--- src/or/config.c     (revision 10545)
+++ src/or/config.c     (working copy)
@@ -963,10 +963,15 @@
     }
   }

-  if (running_tor && rend_config_services(options, 0)<0) {
-    log_warn(LD_BUG,
-       "Previously validated hidden services line could not be added!");
-    return -1;
+  if (running_tor && rend_config_services(options, 1)<0) {
+    log_warn(LD_CONFIG,
+       "Previously validated hidden services line no longer valid! Retaining existing hidden services config if there is one.");
+  }else{
+    if (rend_config_services(options, 0)<0){
+        log_warn(LD_BUG,
+           "Previously validated hidden services line could not be added!");
+        return -1;
+    }
   }

   if (running_tor) {
@@ -2920,9 +2925,10 @@
     }
   }

+/*
   if (rend_config_services(options, 1) < 0)
     REJECT("Failed to configure rendezvous options. See logs for details.");
-
+*/
   if (parse_virtual_addr_network(options->VirtualAddrNetwork, 1, NULL)<0)
     return -1;

[Automatically added by flyspray2trac: Operating System: All]

Weasel says this is a good idea, and IMO it can't hurt.

Reported here: ​https://bugzilla.mozilla.org/show_bug.cgi?id=694611

Test case:

Clicking on the fold-out tabs on the left of this page produces no results:

​http://msdn.microsoft.com/en-ca/subscriptions/downloads/default.aspx