4 SPEC - Spec not finalized
23 =======================================================================
25 Later, unless people want to implement them now:
26 - tor as a socks proxy should accept (and ignore) password auth
27 - Actually use SSL_shutdown to close our TLS connections.
28 - Include "v" line in networkstatus getinfo values.
29 [Nick: bridge authorities output a networkstatus that is missing
30 version numbers. This is inconvenient if we want to make sure
31 bridgedb gives out bridges with certain characteristics. -RD]
32 [Okay. Is this a separate item, or is it the same issue as the lack of
33 a "v" line in response to the controller GETINFO command? -NM]
34 - Let tor dir mirrors proxy connections to the tor download site, so
35 if you know a bridge you can fetch the tor software.
36 - when somebody uses the controlport as an http proxy, give them
37 a "tor isn't an http proxy" error too like we do for the socks port.
38 - MAYBE kill stalled circuits rather than stalled connections. This is
39 possible thanks to cell queues, but we need to consider the anonymity
41 - Make resolves no longer use edge_connection_t unless they are actually
42 _on_ a socks connection: have edge_connection_t and (say)
43 dns_request_t both extend an edge_stream_t, and have p_streams and
44 n_streams both be linked lists of edge_stream_t.
45 - Generate torrc.{complete|sample}.in, tor.1.in, the HTML manual, and the
46 online config documentation from a single source.
47 - It would be potentially helpful to respond to https requests on
48 the OR port by acting like an HTTPS server.
49 - Make the timestamp granularity on logs configurable, with default
50 of "1 second". This might make some kinds of after-the-fact attack harder.
52 - We should get smarter about handling address resolve failures, or
53 addresses that resolve to local IPs. It would be neat to retry
54 them, since right now we just close the stream. But we need to
55 make sure we don't retry them on the same exit as before. But if
56 we mark the circuit, then any user who types "localhost" will
57 cycle through circuits till they run out of retries. See bug 872.
59 Can anybody remember why we wanted to do this and/or what it means?
60 - config option __ControllerLimit that hangs up if there are a limit
61 of controller connections already.
62 [This was mwenge's idea. The idea is that a Tor controller can
63 "fill" Tor's controller slot quota, so jerks can't do cross-protocol
64 attacks like the http form attack. -RD]
66 . Ask all directory questions to bridge via BEGIN_DIR.
67 - use the bridges for dir fetches even when our dirport is open.
68 - drop 'authority' queries if they're to our own identity key; accept
70 - give extend_info_t a router_purpose again
74 If somebody wants to do this in some version, they should:
75 - Create packages for Maemo/Nokia 800/810, requested by Chris Soghoian
76 - debian already makes ARM-arch debs, can maemo use these asks
78 - More work on AvoidDiskWrites
79 - Make DNSPort support TCP DNS.
82 * * * * Roger, please sort these: * * * *
84 - bridge communities with local bridge authorities:
85 - clients who have a password configured decide to ask their bridge
86 authority for a networkstatus
87 - be able to have bridges that aren't in your torrc. save them in
89 - Consider if we can solve: the Tor client doesn't know what flags
90 its bridge has (since it only gets the descriptor), so it can't
91 make decisions based on Fast or Stable.
92 - Some mechanism for specifying that we want to stop using a cached
95 =======================================================================
100 - Our current approach to block attempts to use Tor as a single-hop proxy
101 is pretty lame; we should get a better one.
102 - Allow small cells and large cells on the same network?
103 - Cell buffering and resending. This will allow us to handle broken
104 circuits as long as the endpoints don't break, plus will allow
105 connection (tls session key) rotation.
106 - Implement Morphmix, so we can compare its behavior, complexity,
107 etc. But see paper breaking morphmix.
108 - Other transport. HTTP, udp, rdp, airhook, etc. May have to do our own
109 link crypto, unless we can bully DTLS into it.
110 - Need a relay teardown cell, separate from one-way ends.
111 (Pending a user who needs this)
112 - Handle half-open connections: right now we don't support all TCP
113 streams, at least according to the protocol. But we handle all that
114 we've seen in the wild.
115 (Pending a user who needs this)
119 - handle connect-dir streams that don't have a chosen_exit_name set.
120 - Have a "Faster" status flag that means it. Fast2, Fast4, Fast8?
121 - Add an option (related to AvoidDiskWrites) to disable directory
122 caching. (Is this actually a good idea??)
123 X Add d64 and fp64 along-side d and fp so people can paste status
124 entries into a url. since + is a valid base64 char, only allow one
125 at a time. Consider adding to controller as well.
126 [abandoned for lack of demand]
127 - Some back-out mechanism for auto-approval on authorities
128 - a way of rolling back approvals to before a timestamp
129 - Consider minion-like fingerprint file/log combination.
130 X Have new people be in limbo and need to demonstrate usefulness
131 before we approve them.
134 d Standby/hotswap/redundant hidden services: needs a proposal.
135 - you can insert a hidserv descriptor via the controller.
136 - auth mechanisms to let hidden service midpoint and responder filter
137 connection requests: proposal 121.
138 - Let each hidden service (or other thing) specify its own
142 - If the server is spewing complaints about raising your ulimit -n,
143 we should add a note about this to the server descriptor so other
144 people can notice too.
145 - When we hit a funny error from a dir request (eg 403 forbidden),
146 but tor is working and happy otherwise, and we haven't seen many
147 such errors recently, then don't warn about it.
150 - Implement missing status events and accompanying getinfos
152 - BAD_DIR_RESPONSE (Unexpected directory response; maybe we're behind
154 - BAD_PROXY (Bad http or https proxy)
155 - UNRECOGNIZED_ROUTER (a nickname we asked for is unavailable)
156 - Status events related to hibernation
157 - something about failing to parse our address?
158 from resolve_my_address() in config.c
159 - sketchy OS, sketchy threading
160 - too many onions queued: threading problems or slow CPU?
161 - Implement missing status event fields:
162 - TIMEOUT on CHECKING_REACHABILITY
163 - GETINFO status/client, status/server, status/general: There should be
164 some way to learn which status events are currently "in effect."
165 We should specify which these are, what format they appear in, and so
167 - More information in events:
168 - Include bandwidth breakdown by conn->type in BW events.
169 - Change circuit status events to give more details, like purpose,
170 whether they're internal, when they become dirty, when they become
171 too dirty for further circuits, etc.
172 - Change stream status events analogously.
173 - Expose more information via getinfo:
174 - import and export rendezvous descriptors
175 - Review all static fields for additional candidates
176 - Allow EXTENDCIRCUIT to unknown server.
177 - We need some way to adjust server status, and to tell tor not to
178 download directories/network-status, and a way to force a download.
179 - Make everything work with hidden services
181 - Performance/resources
182 - per-conn write buckets
183 - separate config options for read vs write limiting
184 (It's hard to support read > write, since we need better
185 congestion control to avoid overfull buffers there. So,
186 defer the whole thing.)
187 - Rate limit exit connections to a given destination -- this helps
188 us play nice with websites when Tor users want to crawl them; it
189 also introduces DoS opportunities.
190 - Consider truncating rather than destroying failed circuits,
191 in order to save the effort of restarting. There are security
192 issues here that need thinking, though.
193 - Handle full buffers without totally borking
194 - Rate-limit OR and directory connections overall and per-IP and
198 - Hold-open-until-flushed now works by accident; it should work by
200 - Display the reasons in 'destroy' and 'truncated' cells under
202 - Make router_is_general_exit() a bit smarter once we're sure what
204 - Automatically determine what ports are reachable and start using
205 those, if circuits aren't working and it's a pattern we
206 recognize ("port 443 worked once and port 9001 keeps not
210 - some better fix for bug #516?
213 - Servers might check certs for known-good ssl websites, and if
214 they come back self-signed, declare themselves to be
215 non-exits. Similar to how we test for broken/evil dns now.
216 - Authorities should try using exits for http to connect to some
217 URLS (specified in a configuration file, so as not to make the
218 List Of Things Not To Censor completely obvious) and ask them
219 for results. Exits that don't give good answers should have
220 the BadExit flag set.
221 - Alternatively, authorities should be able to import opinions
222 from Snakes on a Tor.
223 - Bind to random port when making outgoing connections to Tor servers,
224 to reduce remote sniping attacks.
225 - Audit everything to make sure rend and intro points are just as
226 likely to be us as not.
227 - Do something to prevent spurious EXTEND cells from making
228 middleman nodes connect all over. Rate-limit failed
229 connections, perhaps?
230 - DoS protection: TLS puzzles, public key ops, bandwidth exhaustion.
233 - Now that we're avoiding exits when picking non-exit positions,
234 we need to consider how to pick nodes for internal circuits. If
235 we avoid exits for all positions, we skew the load balancing. If
236 we accept exits for all positions, we leak whether it's an
237 internal circuit at every step. If we accept exits only at the
238 last hop, we reintroduce Lasse's attacks from the Oakland paper.
240 - Windows server usability
241 - Solve the ENOBUFS problem.
242 - make tor's use of openssl operate on buffers rather than sockets,
243 so we can make use of libevent's buffer paradigm once it has one.
244 - make tor's use of libevent tolerate either the socket or the
245 buffer paradigm; includes unifying the functions in connect.c.
246 - We need a getrlimit equivalent on Windows so we can reserve some
247 file descriptors for saving files, etc. Otherwise we'll trigger
248 asserts when we're out of file descriptors and crash.
251 - a way to generate the website diagrams from source, so we can
252 translate them as utf-8 text rather than with gimp. (svg? or
254 . Flesh out options_description array in src/or/config.c
255 . multiple sample torrc files
256 - Refactor tor man page to divide generally useful options from
258 - Add a doxygen style checker to make check-spaces so nick doesn't drift
259 too far from arma's undocumented styleguide. Also, document that
260 styleguide in HACKING. (See r9634 for example.)
261 - exactly one space at beginning and at end of comments, except i
262 guess when there's line-length pressure.
263 - if we refer to a function name, put a () after it.
264 - only write <b>foo</b> when foo is an argument to this function.
265 - doxygen comments must always end in some form of punctuation.
266 - capitalize the first sentence in the doxygen comment, except
268 - avoid spelling errors and incorrect comments. ;)
271 - The Debian package now uses --verify-config when (re)starting,
272 to distinguish configuration errors from other errors. Perhaps
273 the RPM and other startup scripts should too?
274 - add a "default.action" file to the tor/vidalia bundle so we can
275 fix the https thing in the default configuration:
276 http://wiki.noreply.org/noreply/TheOnionRouter/TorFAQ#PrivoxyWeirdSSLPort
279 =======================================================================
281 Documentation, non-version-specific.
283 - Mark up spec; note unclear points about servers
284 NR - write a spec appendix for 'being nice with tor'
285 - Specify the keys and key rotation schedules and stuff
286 . Finish path-spec.txt
287 - Mention controller libs someplace.
288 - Remove need for HACKING file.
289 - document http://wiki.noreply.org/noreply/TheOnionRouter/TransparentProxy on freebsd and osx
290 P - figure out rpm spec files for bundles of vidalia-tor-polipo
291 P - figure out polipo install scripts for bundles of vidalia-tor-polipo on osx, win32
292 - figure out selinux policy for tor
293 P - change packaging system to more automated and specific for each
294 platform, suggested by Paul Wouter
295 P - Setup repos for redhat and suse rpms & start signing the rpms the
296 way package management apps prefer
299 J . tor-in-the-media page
300 P - Figure out licenses for website material.
301 (Phobos reccomends the Open Publication License with Option A at
302 http://opencontent.org/openpub/)
303 P - put the logo on the website, in source form, so people can put it on
304 stickers directly, etc.
305 P - put the source image for the stickers on the website, so people can
307 P - figure out a license for the logos and docs we publish (trademark
309 (Phobos reccomends the Open Publication License with Option A at
310 http://opencontent.org/openpub/)
311 I - add a page for localizing all tor's components.
312 - It would be neat if we had a single place that described _all_ the
313 tor-related tools you can use, and what they give you, and how well they
314 work. Right now, we don't give a lot of guidance wrt
315 torbutton/foxproxy/privoxy/polipo in any consistent place.
316 P - create a 'blog badge' for tor fans to link to and feature on their
317 blogs. A sample is at http://interloper.org/tmp/tor/tor-button.png
318 - More prominently, we should have a recommended apps list.
319 - recommend pidgin (gaim is renamed)
320 - unrecommend IE because of ftp:// bug.
321 - Addenda to tor-design
322 - we should add a preamble to tor-design saying it's out of date.
323 - we should add an appendix or errata on what's changed.
326 - make a mailing list with the mirror operators
327 o make an automated tool to check /project/trace/ at mirrors to
328 learn which ones are lagging behind.
329 - auto (or manually) cull the mirrors that are broken; and
330 contact their operator?
331 - a set of instructions for mirror operators to make their apaches
332 serve our charsets correctly, and bonus points for language
334 - figure out how to load-balance the downloads across mirrors?
335 - ponder how to get users to learn that they should google for
336 "tor mirrors" if the main site is blocked.
337 - find a mirror volunteer to coordinate all of this