TODO 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266
  1. - Rotate tls-level connections -- make new ones, expire old ones.
  2. Nick, can you remember why we wanted to do this?
  3. Legend:
  4. SPEC!! - Not specified
  5. SPEC - Spec not finalized
  6. NICK - nick claims
  7. ARMA - arma claims
  8. - Not done
  9. * Top priority
  10. . Partially done
  11. o Done
  12. D Deferred
  13. X Abandoned
  14. Short-term:
  15. - Make tls connections tls_close intentionally
  16. o Rename ACI to circID
  17. . integrate rep_ok functions, see what breaks
  18. - update tor faq
  19. o obey SocksBindAddress, ORBindAddress
  20. - warn if we're running as root
  21. o make connection_flush_buf() more obviously obsolete
  22. .* let hup reread the config file, eg so we can get new exit
  23. policies without restarting
  24. - use times(2) rather than gettimeofday to measure how long it
  25. takes to process a cell
  26. - Separate trying to rebuild a circuit because you have none from trying
  27. to rebuild a circuit because the current one is stale
  28. - Continue reading from socks port even while waiting for connect.
  29. .* Exit policies
  30. o Spec how to write the exit policies
  31. -* More flexible policies (18.*, 18.0.0.0/8)
  32. -* Path selection algorithms
  33. -* Choose path more incrementally
  34. -* Let user request first/last node
  35. - And disallow certain nodes
  36. D Choose path by jurisdiction, etc?
  37. . Make relay end cells have failure status and payload attached
  38. -* Streams that fail due to exit policy must reextend to new node
  39. -* Add extend_wait state to edge connections, thumb through them
  40. when the AP get an extended cell.
  41. -* let non-approved routers handshake.
  42. -* just list approved routers in directory.
  43. - Dirserver shouldn't put you in running-routers list if you haven't
  44. uploaded a descriptor recently
  45. . migrate to using nickname rather than addr:port for routers
  46. o decide_aci_type
  47. - generate onion skins
  48. - circuit_send_next_onion_skin
  49. - circuit_extend
  50. - onion_generate_cpath
  51. - get_unique_aci_by_addr_port
  52. - circ->n_addr and circ->n_port
  53. - circuit_enumerate_by_naddr_nport
  54. - cpath layers
  55. - connection_or_connect
  56. - connection_exact_get_by_addr_port
  57. - connection_twin_get_by_addr_port
  58. - router_get_by_addr_port
  59. - connection_or_init_conn_from_router
  60. - tag_pack, tag_unpack, connection_cpu_process_inbuf
  61. - directory_initiate_command
  62. .* Move from onions to ephemeral DH
  63. o incremental path building
  64. o transition circuit-level sendmes to hop-level sendmes
  65. o implement truncate, truncated
  66. o move from 192byte DH to 128byte DH, so it isn't so damn slow
  67. -* exiting from not-last hop
  68. -* OP logic to decide to extend/truncate a path
  69. -* make sure exiting from the not-last hop works
  70. -* logic to find last *open* hop, not last hop, in cpath
  71. -* choose exit nodes by exit policies
  72. o Remember address and port when beginning.
  73. - Extend by nickname/hostname/something, not by IP.
  74. - Need a relay teardown cell, separate from one-way ends.
  75. On-going
  76. . Better comments for functions!
  77. . Go through log messages, reduce confusing error messages.
  78. . make the logs include more info (fd, etc)
  79. . Unit tests
  80. Mid-term:
  81. - Are there anonymity issues with sequential streamIDs? Sequential
  82. circIDs? Eg an attacker can learn how many there have been.
  83. - Look at having smallcells and largecells
  84. . Redo scheduler
  85. o fix SSL_read bug for buffered records
  86. - make round-robining more fair
  87. - What happens when a circuit's length is 1? What breaks?
  88. . streams / circuits
  89. o Implement streams
  90. o Rotate circuits after N minutes?
  91. X Circuits should expire when circuit->expire triggers
  92. NICK . Handle half-open connections
  93. o openssh is an application that uses half-open connections
  94. o Figure out what causes connections to close, standardize
  95. when we mark a connection vs when we tear it down
  96. o Look at what ssl does to keep from mutating data streams
  97. ARMA - Reduce streamid footprint from 7 bytes to 2 bytes
  98. - Check for collisions in streamid (now possible with
  99. just 2 bytes), and back up & replace with padding if so
  100. - Use the 3 saved bytes to put pseudorandomness in each relay cell
  101. - Use the 4 reserved bytes in each cell header to keep 1/5
  102. of a sha1 of the relay payload (move into stream header)
  103. - Move length into the stream header too
  104. - Spec the stream_id stuff. Clarify that nobody on the backward
  105. stream should look at stream_id.
  106. . Put CPU workers in separate processes
  107. o Handle multiple cpu workers (one for each cpu, plus one)
  108. o Queue for pending tasks if all workers full
  109. o Support the 'process this onion' task
  110. D Merge dnsworkers and cpuworkers to some extent
  111. - Handle cpuworkers dying
  112. . Scrubbing proxies
  113. - Find an smtp proxy?
  114. - Check the old smtp proxy code
  115. o Find an ftp proxy? wget --passive
  116. D Wait until there are packet redirectors for Linux
  117. . Get socks4a support into Mozilla
  118. . Develop rendezvous points
  119. SPEC!! - Handle socks commands other than connect, eg, bind?
  120. o Design
  121. - Spec
  122. - Implement
  123. - Tests
  124. o Testing harness/infrastructure
  125. D System tests (how?)
  126. - Performance tests, so we know when we've improved
  127. . webload infrastructure (Bruce)
  128. . httperf infrastructure (easy to set up)
  129. . oprofile (installed in RH >8.0)
  130. NICK . Daemonize and package
  131. o Teach it to fork and background
  132. - Red Hat spec file
  133. - Debian spec file equivalent
  134. . Portability
  135. . Which .h files are we actually using?
  136. . Port to:
  137. o Linux
  138. o BSD
  139. . Solaris
  140. o Cygwin
  141. . Win32
  142. o OS X
  143. - deal with pollhup / reached_eof on all platforms
  144. o openssl randomness
  145. o inet_ntoa
  146. . stdint.h
  147. - Make a script to set up a local network on your machine
  148. - More flexibility in node addressing
  149. D Support IPv6 rather than just 4
  150. - Handle multihomed servers (config variable to set IP)
  151. In the distant future:
  152. D Load balancing between router twins
  153. D Keep track of load over links/nodes, to
  154. know who's hosed
  155. SPEC!! D Non-clique topologies
  156. D Implement our own memory management, at least for common structs
  157. (Not ever necessary?)
  158. D Advanced directory servers
  159. D Automated reputation management
  160. SPEC!! D Figure out how to do threshold directory servers
  161. D jurisdiction info in dirserver entries? other info?
  162. Older (done) todo stuff:
  163. o Get tor to act like a socks server
  164. o socks4, socks4a
  165. o socks5
  166. o routers have identity key, link key, onion key.
  167. o link key certs are
  168. D signed by identity key
  169. D not in descriptor
  170. o not in config
  171. D not on disk
  172. o identity and onion keys are in descriptor (and disk)
  173. o upon boot, if it doesn't find identity key, generate it and write it.
  174. o also write a file with the identity key fingerprint in it
  175. o router generates descriptor: flesh out router_get_my_descriptor()
  176. o Routers sign descriptors with identity key
  177. o routers put version number in descriptor
  178. o routers should maybe have `uname -a` in descriptor?
  179. o Give nicknames to routers
  180. o in config
  181. o in descriptors
  182. o router posts descriptor
  183. o when it boots
  184. D when it changes
  185. o change tls stuff so certs don't get written to disk, or read from disk
  186. o make directory.c 'thread'safe
  187. o dirserver parses descriptor
  188. o dirserver checks signature
  189. D client checks signature?
  190. o dirserver writes directory to file
  191. o reads that file upon boot
  192. o directory includes all routers, up and down
  193. o add "up" line to directory, listing nicknames
  194. o instruments ORs to report stats
  195. o average cell fullness
  196. o average bandwidth used
  197. o configure log files. separate log file, separate severities.
  198. o what assumptions break if we fclose(0) when we daemonize?
  199. o make buffer struct elements opaque outside buffers.c
  200. o add log convention to the HACKING file
  201. o make 'make install' do the right thing
  202. o change binary name to tor
  203. o change config files so you look at commandline, else look in
  204. /etc/torrc. no cascading.
  205. o have an absolute datadir with fixed names for files, and fixed-name
  206. keydir under that with fixed names
  207. o Move (most of) the router/directory code out of main.c
  208. o Simple directory servers
  209. o Include key in source; sign directories
  210. o Signed directory backend
  211. o Document
  212. o Integrate
  213. o Add versions to code
  214. o Have directories list recommended-versions
  215. o Include line in directories
  216. o Check for presence of line.
  217. o Quit if running the wrong version
  218. o Command-line option to override quit
  219. o Add more information to directory server entries
  220. o Exit policies
  221. o Clearer bandwidth management
  222. o Do we want to remove bandwidth from OR handshakes?
  223. o What about OP handshakes?
  224. X Move away from openssl
  225. o Abstract out crypto calls
  226. X Look at nss, others? Just include code?
  227. o Use a stronger cipher
  228. o aes now, by including the code ourselves
  229. X On the fly compression of each stream
  230. o Clean up the event loop (optimize and sanitize)
  231. o Remove that awful concept of 'roles'
  232. o Terminology
  233. o Circuits, topics, cells stay named that
  234. o 'Connection' gets divided, or renamed, or something?
  235. o DNS farm
  236. o Distribute queries onto the farm, get answers
  237. o Preemptively grow a new worker before he's needed
  238. o Prune workers when too many are idle
  239. o DNS cache
  240. o Clear DNS cache over time
  241. D Honor DNS TTL info (how??)
  242. o Have strategy when all workers are busy
  243. o Keep track of which connections are in dns_wait
  244. o Need to cache positives/negatives on the tor side
  245. o Keep track of which queries have been asked
  246. o Better error handling when
  247. o An address doesn't resolve
  248. o We have max workers running
  249. o Consider taking the master out of the loop?
  250. X Implement reply onions
  251. o Total rate limiting
  252. o Look at OR handshake in more detail
  253. o Spec it
  254. o Merge OR and OP handshakes
  255. o rearrange connection_or so it doesn't suck so much to read
  256. D Periodic link key rotation. Spec?
  257. o wrap malloc with something that explodes when it fails
  258. o Clean up the number of places that get to look at prkey