TODO.future 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329
  1. $Id: TODO 16258 2008-07-30 13:04:38Z nickm $
  2. Legend:
  3. SPEC!! - Not specified
  4. SPEC - Spec not finalized
  5. N - nick claims
  6. R - arma claims
  7. P - phobos claims
  8. S - Steven claims
  9. E - Matt claims
  10. M - Mike claims
  11. J - Jeff claims
  12. I - ioerror claims
  13. W - weasel claims
  14. K - Karsten claims
  15. - Not done
  16. * Top priority
  17. . Partially done
  18. o Done
  19. d Deferrable
  20. D Deferred
  21. X Abandoned
  22. =======================================================================
  23. Later, unless people want to implement them now:
  24. - tor as a socks proxy should accept (and ignore) password auth
  25. - Actually use SSL_shutdown to close our TLS connections.
  26. - Include "v" line in networkstatus getinfo values.
  27. [Nick: bridge authorities output a networkstatus that is missing
  28. version numbers. This is inconvenient if we want to make sure
  29. bridgedb gives out bridges with certain characteristics. -RD]
  30. [Okay. Is this a separate item, or is it the same issue as the lack of
  31. a "v" line in response to the controller GETINFO command? -NM]
  32. - Let tor dir mirrors proxy connections to the tor download site, so
  33. if you know a bridge you can fetch the tor software.
  34. - when somebody uses the controlport as an http proxy, give them
  35. a "tor isn't an http proxy" error too like we do for the socks port.
  36. - MAYBE kill stalled circuits rather than stalled connections. This is
  37. possible thanks to cell queues, but we need to consider the anonymity
  38. implications.
  39. - Make resolves no longer use edge_connection_t unless they are actually
  40. _on_ a socks connection: have edge_connection_t and (say)
  41. dns_request_t both extend an edge_stream_t, and have p_streams and
  42. n_streams both be linked lists of edge_stream_t.
  43. - Generate torrc.{complete|sample}.in, tor.1.in, the HTML manual, and the
  44. online config documentation from a single source.
  45. - It would be potentially helpful to respond to https requests on
  46. the OR port by acting like an HTTPS server.
  47. - Make the timestamp granularity on logs configurable, with default
  48. of "1 second". This might make some kinds of after-the-fact attack harder.
  49. Can anybody remember why we wanted to do this and/or what it means?
  50. - config option __ControllerLimit that hangs up if there are a limit
  51. of controller connections already.
  52. [This was mwenge's idea. The idea is that a Tor controller can
  53. "fill" Tor's controller slot quota, so jerks can't do cross-protocol
  54. attacks like the http form attack. -RD]
  55. - Bridge issues
  56. . Ask all directory questions to bridge via BEGIN_DIR.
  57. - use the bridges for dir fetches even when our dirport is open.
  58. - drop 'authority' queries if they're to our own identity key; accept
  59. them otherwise.
  60. - give extend_info_t a router_purpose again
  61. If somebody wants to do this in some version, they should:
  62. - Create packages for Nokia 800, requested by Chris Soghoian
  63. - More work on AvoidDiskWrites
  64. - Make DNSPort support TCP DNS.
  65. * * * * Roger, please sort these: * * * *
  66. - bridge communities with local bridge authorities:
  67. - clients who have a password configured decide to ask their bridge
  68. authority for a networkstatus
  69. - be able to have bridges that aren't in your torrc. save them in
  70. state file, etc.
  71. - Consider if we can solve: the Tor client doesn't know what flags
  72. its bridge has (since it only gets the descriptor), so it can't
  73. make decisions based on Fast or Stable.
  74. - Some mechanism for specifying that we want to stop using a cached
  75. bridge.
  76. =======================================================================
  77. Future versions:
  78. - Protocol
  79. - Our current approach to block attempts to use Tor as a single-hop proxy
  80. is pretty lame; we should get a better one.
  81. - Allow small cells and large cells on the same network?
  82. - Cell buffering and resending. This will allow us to handle broken
  83. circuits as long as the endpoints don't break, plus will allow
  84. connection (tls session key) rotation.
  85. - Implement Morphmix, so we can compare its behavior, complexity,
  86. etc. But see paper breaking morphmix.
  87. - Other transport. HTTP, udp, rdp, airhook, etc. May have to do our own
  88. link crypto, unless we can bully DTLS into it.
  89. - Need a relay teardown cell, separate from one-way ends.
  90. (Pending a user who needs this)
  91. - Handle half-open connections: right now we don't support all TCP
  92. streams, at least according to the protocol. But we handle all that
  93. we've seen in the wild.
  94. (Pending a user who needs this)
  95. - Directory system
  96. - BEGIN_DIR items
  97. - handle connect-dir streams that don't have a chosen_exit_name set.
  98. - Have a "Faster" status flag that means it. Fast2, Fast4, Fast8?
  99. - Add an option (related to AvoidDiskWrites) to disable directory
  100. caching. (Is this actually a good idea??)
  101. X Add d64 and fp64 along-side d and fp so people can paste status
  102. entries into a url. since + is a valid base64 char, only allow one
  103. at a time. Consider adding to controller as well.
  104. [abandoned for lack of demand]
  105. - Some back-out mechanism for auto-approval on authorities
  106. - a way of rolling back approvals to before a timestamp
  107. - Consider minion-like fingerprint file/log combination.
  108. X Have new people be in limbo and need to demonstrate usefulness
  109. before we approve them.
  110. - Hidden services:
  111. d Standby/hotswap/redundant hidden services: needs a proposal.
  112. - you can insert a hidserv descriptor via the controller.
  113. - auth mechanisms to let hidden service midpoint and responder filter
  114. connection requests: proposal 121.
  115. - Let each hidden service (or other thing) specify its own
  116. OutboundBindAddress?
  117. - Server operation
  118. - If the server is spewing complaints about raising your ulimit -n,
  119. we should add a note about this to the server descriptor so other
  120. people can notice too.
  121. - When we hit a funny error from a dir request (eg 403 forbidden),
  122. but tor is working and happy otherwise, and we haven't seen many
  123. such errors recently, then don't warn about it.
  124. - Controller
  125. - Implement missing status events and accompanying getinfos
  126. - DIR_REACHABLE
  127. - BAD_DIR_RESPONSE (Unexpected directory response; maybe we're behind
  128. a firewall.)
  129. - BAD_PROXY (Bad http or https proxy)
  130. - UNRECOGNIZED_ROUTER (a nickname we asked for is unavailable)
  131. - Status events related to hibernation
  132. - something about failing to parse our address?
  133. from resolve_my_address() in config.c
  134. - sketchy OS, sketchy threading
  135. - too many onions queued: threading problems or slow CPU?
  136. - Implement missing status event fields:
  137. - TIMEOUT on CHECKING_REACHABILITY
  138. - GETINFO status/client, status/server, status/general: There should be
  139. some way to learn which status events are currently "in effect."
  140. We should specify which these are, what format they appear in, and so
  141. on.
  142. - More information in events:
  143. - Include bandwidth breakdown by conn->type in BW events.
  144. - Change circuit status events to give more details, like purpose,
  145. whether they're internal, when they become dirty, when they become
  146. too dirty for further circuits, etc.
  147. - Change stream status events analogously.
  148. - Expose more information via getinfo:
  149. - import and export rendezvous descriptors
  150. - Review all static fields for additional candidates
  151. - Allow EXTENDCIRCUIT to unknown server.
  152. - We need some way to adjust server status, and to tell tor not to
  153. download directories/network-status, and a way to force a download.
  154. - Make everything work with hidden services
  155. - Performance/resources
  156. - per-conn write buckets
  157. - separate config options for read vs write limiting
  158. (It's hard to support read > write, since we need better
  159. congestion control to avoid overfull buffers there. So,
  160. defer the whole thing.)
  161. - Rate limit exit connections to a given destination -- this helps
  162. us play nice with websites when Tor users want to crawl them; it
  163. also introduces DoS opportunities.
  164. - Consider truncating rather than destroying failed circuits,
  165. in order to save the effort of restarting. There are security
  166. issues here that need thinking, though.
  167. - Handle full buffers without totally borking
  168. - Rate-limit OR and directory connections overall and per-IP and
  169. maybe per subnet.
  170. - Misc
  171. - Hold-open-until-flushed now works by accident; it should work by
  172. design.
  173. - Display the reasons in 'destroy' and 'truncated' cells under
  174. some circumstances?
  175. - Make router_is_general_exit() a bit smarter once we're sure what
  176. it's for.
  177. - Automatically determine what ports are reachable and start using
  178. those, if circuits aren't working and it's a pattern we
  179. recognize ("port 443 worked once and port 9001 keeps not
  180. working").
  181. - Security
  182. - some better fix for bug #516?
  183. - Directory guards
  184. - Mini-SoaT:
  185. - Servers might check certs for known-good ssl websites, and if
  186. they come back self-signed, declare themselves to be
  187. non-exits. Similar to how we test for broken/evil dns now.
  188. - Authorities should try using exits for http to connect to some
  189. URLS (specified in a configuration file, so as not to make the
  190. List Of Things Not To Censor completely obvious) and ask them
  191. for results. Exits that don't give good answers should have
  192. the BadExit flag set.
  193. - Alternatively, authorities should be able to import opinions
  194. from Snakes on a Tor.
  195. - Bind to random port when making outgoing connections to Tor servers,
  196. to reduce remote sniping attacks.
  197. - Audit everything to make sure rend and intro points are just as
  198. likely to be us as not.
  199. - Do something to prevent spurious EXTEND cells from making
  200. middleman nodes connect all over. Rate-limit failed
  201. connections, perhaps?
  202. - DoS protection: TLS puzzles, public key ops, bandwidth exhaustion.
  203. - Needs thinking
  204. - Now that we're avoiding exits when picking non-exit positions,
  205. we need to consider how to pick nodes for internal circuits. If
  206. we avoid exits for all positions, we skew the load balancing. If
  207. we accept exits for all positions, we leak whether it's an
  208. internal circuit at every step. If we accept exits only at the
  209. last hop, we reintroduce Lasse's attacks from the Oakland paper.
  210. - Windows server usability
  211. - Solve the ENOBUFS problem.
  212. - make tor's use of openssl operate on buffers rather than sockets,
  213. so we can make use of libevent's buffer paradigm once it has one.
  214. - make tor's use of libevent tolerate either the socket or the
  215. buffer paradigm; includes unifying the functions in connect.c.
  216. - We need a getrlimit equivalent on Windows so we can reserve some
  217. file descriptors for saving files, etc. Otherwise we'll trigger
  218. asserts when we're out of file descriptors and crash.
  219. - Documentation
  220. - a way to generate the website diagrams from source, so we can
  221. translate them as utf-8 text rather than with gimp. (svg? or
  222. imagemagick?)
  223. . Flesh out options_description array in src/or/config.c
  224. . multiple sample torrc files
  225. - Refactor tor man page to divide generally useful options from
  226. less useful ones?
  227. - Add a doxygen style checker to make check-spaces so nick doesn't drift
  228. too far from arma's undocumented styleguide. Also, document that
  229. styleguide in HACKING. (See r9634 for example.)
  230. - exactly one space at beginning and at end of comments, except i
  231. guess when there's line-length pressure.
  232. - if we refer to a function name, put a () after it.
  233. - only write <b>foo</b> when foo is an argument to this function.
  234. - doxygen comments must always end in some form of punctuation.
  235. - capitalize the first sentence in the doxygen comment, except
  236. when you shouldn't.
  237. - avoid spelling errors and incorrect comments. ;)
  238. - Packaging
  239. - The Debian package now uses --verify-config when (re)starting,
  240. to distinguish configuration errors from other errors. Perhaps
  241. the RPM and other startup scripts should too?
  242. - add a "default.action" file to the tor/vidalia bundle so we can
  243. fix the https thing in the default configuration:
  244. http://wiki.noreply.org/noreply/TheOnionRouter/TorFAQ#PrivoxyWeirdSSLPort
  245. =======================================================================
  246. Documentation, non-version-specific.
  247. - Specs
  248. - Mark up spec; note unclear points about servers
  249. NR - write a spec appendix for 'being nice with tor'
  250. - Specify the keys and key rotation schedules and stuff
  251. . Finish path-spec.txt
  252. - Mention controller libs someplace.
  253. - Remove need for HACKING file.
  254. - document http://wiki.noreply.org/noreply/TheOnionRouter/TransparentProxy on freebsd and osx
  255. P - figure out rpm spec files for bundles of vidalia-tor-polipo
  256. P - figure out polipo install scripts for bundles of vidalia-tor-polipo on osx, win32
  257. - figure out selinux policy for tor
  258. P - change packaging system to more automated and specific for each
  259. platform, suggested by Paul Wouter
  260. P - Setup repos for redhat and suse rpms & start signing the rpms the
  261. way package management apps prefer
  262. Website:
  263. J . tor-in-the-media page
  264. P - Figure out licenses for website material.
  265. (Phobos reccomends the Open Publication License with Option A at
  266. http://opencontent.org/openpub/)
  267. P - put the logo on the website, in source form, so people can put it on
  268. stickers directly, etc.
  269. P - put the source image for the stickers on the website, so people can
  270. print their own
  271. P - figure out a license for the logos and docs we publish (trademark
  272. figures into this)
  273. (Phobos reccomends the Open Publication License with Option A at
  274. http://opencontent.org/openpub/)
  275. I - add a page for localizing all tor's components.
  276. - It would be neat if we had a single place that described _all_ the
  277. tor-related tools you can use, and what they give you, and how well they
  278. work. Right now, we don't give a lot of guidance wrt
  279. torbutton/foxproxy/privoxy/polipo in any consistent place.
  280. P - create a 'blog badge' for tor fans to link to and feature on their
  281. blogs. A sample is at http://interloper.org/tmp/tor/tor-button.png
  282. - More prominently, we should have a recommended apps list.
  283. - recommend pidgin (gaim is renamed)
  284. - unrecommend IE because of ftp:// bug.
  285. - Addenda to tor-design
  286. - we should add a preamble to tor-design saying it's out of date.
  287. - we should add an appendix or errata on what's changed.
  288. - Tor mirrors
  289. - make a mailing list with the mirror operators
  290. o make an automated tool to check /project/trace/ at mirrors to
  291. learn which ones are lagging behind.
  292. - auto (or manually) cull the mirrors that are broken; and
  293. contact their operator?
  294. - a set of instructions for mirror operators to make their apaches
  295. serve our charsets correctly, and bonus points for language
  296. negotiation.
  297. - figure out how to load-balance the downloads across mirrors?
  298. - ponder how to get users to learn that they should google for
  299. "tor mirrors" if the main site is blocked.
  300. - find a mirror volunteer to coordinate all of this