TODO 8.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176
  1. Legend:
  2. SPEC!! - Not specified
  3. SPEC - Spec not finalized
  4. NICK - nick claims
  5. ARMA - arma claims
  6. - Not done
  7. * Top priority
  8. . Partially done
  9. o Done
  10. D Deferred
  11. X Abandoned
  12. . streams / circuits
  13. o Implement streams
  14. o Rotate circuits after N minutes?
  15. X Circuits should expire when circuit->expire triggers
  16. NICK . Handle half-open connections
  17. o Figure out what causes connections to close, standardize
  18. when we mark a connection vs when we tear it down
  19. o Look at what ssl does to keep from mutating data streams
  20. ARMA - Reduce streamid footprint from 7 bytes to 3 bytes
  21. - Check for collisions in streamid (now possible with
  22. just 3 bytes), and back up & replace with padding if so
  23. - Use the 3 saved bytes to put pseudorandomness in each cell
  24. - Use the 4 reserved bytes in each cell header to keep 1/5
  25. of a sha1 of the payload
  26. - (Move these 4 bytes into the stream header)
  27. - Consider moving length into the stream header too
  28. - Spec the stream_id stuff. Clarify that nobody on the backward
  29. stream should look at stream_id.
  30. ARMA . Exit policies
  31. o Spec how to write the exit policies
  32. - Path selection algorithms
  33. - Let user request certain nodes
  34. - And disallow certain nodes
  35. D Choose path by jurisdiction, etc?
  36. - Make relay end cells have failure status and payload attached
  37. - Streams that fail due to exit policy must reextend to new node
  38. - Add extend_wait state to edge connections, thumb through them
  39. when the AP get an extended cell.
  40. SPEC!! D Non-clique topologies
  41. D Implement our own memory management, at least for common structs
  42. . Appropriate logging
  43. - Come up with convention for what log level means what
  44. - Make code follow convention
  45. . Put CPU workers in separate processes
  46. o Handle multiple cpu workers (one for each cpu, plus one)
  47. o Queue for pending tasks if all workers full
  48. o Support the 'process this onion' task
  49. - Support the 'decrypt this RSA blob' handshake1 task
  50. - Handle cpuworkers dying
  51. D Support later handshake parts
  52. . Directory servers
  53. D Automated reputation management
  54. o Include key in source; sign directories
  55. o Signed directory backend
  56. o Document
  57. o Integrate
  58. - Add versions to code
  59. . Have directories list recommended-versions
  60. o Include (unused) line in directories
  61. o Check for presence of line.
  62. - Quit if running the wrong version
  63. - Command-line option to override quit
  64. . Add more information to directory server entries
  65. o Exit policies
  66. D jurisdiction? others?
  67. SPEC!! D Figure out how to do threshold directory servers
  68. . Scrubbing proxies
  69. - Find an smtp proxy?
  70. - Check the old smtp proxy code
  71. o Find an ftp proxy? wget --passive
  72. D Wait until there are packet redirectors for Linux
  73. . Get socks4a support into Mozilla
  74. . Get tor to act like a socks server
  75. o socks4, socks4a
  76. D socks5
  77. SPEC!! - Handle socks commands other than connect, eg, bind?
  78. . Develop rendezvous points
  79. . Spec (still needs step-by-step instructions)
  80. - Implement
  81. D Deploy and manage open source development site.
  82. . Documentation
  83. o Discussion of socks, tsocks, etc
  84. o On-the-network protocol
  85. o Onions
  86. o Cells
  87. . Better comments for functions!
  88. - Tests
  89. o Testing harness/infrastructure
  90. NICK . Unit tests
  91. D System tests (how?)
  92. - Performance tests, so we know when we've improved
  93. . webload infrastructure (Bruce)
  94. . httperf infrastructure (easy to set up)
  95. . oprofile (installed in RH >8.0)
  96. D Deploy a widespread network
  97. . Router twins
  98. o Choose twin if primary is down, when laying circuit
  99. D Load balancing between twins
  100. - Keep track of load over links/nodes, to
  101. know who's hosed
  102. NICK . Daemonize and package
  103. o Teach it to fork and background
  104. - Red Hat spec file
  105. - Debian spec file equivalent
  106. . Autoconf
  107. . Which .h files are we actually using?
  108. . Port to:
  109. o Linux
  110. o BSD
  111. . Solaris
  112. o Cygwin
  113. . Win32
  114. o OS X
  115. o openssl randomness
  116. o inet_ntoa
  117. . stdint.h
  118. - Make a script to set up a local network on your machine
  119. D Move away from openssl
  120. o Abstract out crypto calls
  121. D Look at nss, others? Just include code?
  122. . Clearer bandwidth management
  123. - Do we want to remove bandwidth from OR handshakes?
  124. - What about OP handshakes?
  125. - More flexibility in node addressing
  126. D Support IPv6 rather than just 4
  127. - Handle multihomed servers (config variable to set IP)
  128. . Move from onions to ephemeral DH
  129. o incremental path building
  130. o transition circuit-level sendmes to hop-level sendmes
  131. o implement truncate, truncated
  132. o move from 192byte DH to 128byte DH, so it isn't so damn slow
  133. - exiting from not-last hop
  134. - OP logic to decide to extend/truncate a path
  135. - make sure exiting from the not-last hop works
  136. - logic to find last *open* hop, not last hop, in cpath
  137. - choose exit nodes by exit policies
  138. Older (done) todo stuff:
  139. o Use a stronger cipher
  140. o aes now, by including the code ourselves
  141. X On the fly compression of each stream
  142. o Clean up the event loop (optimize and sanitize)
  143. o Remove that awful concept of 'roles'
  144. o Terminology
  145. o Circuits, topics, cells stay named that
  146. o 'Connection' gets divided, or renamed, or something?
  147. o DNS farm
  148. o Distribute queries onto the farm, get answers
  149. o Preemptively grow a new worker before he's needed
  150. o Prune workers when too many are idle
  151. o DNS cache
  152. o Clear DNS cache over time
  153. D Honor DNS TTL info (how??)
  154. o Have strategy when all workers are busy
  155. o Keep track of which connections are in dns_wait
  156. o Need to cache positives/negatives on the tor side
  157. o Keep track of which queries have been asked
  158. o Better error handling when
  159. o An address doesn't resolve
  160. o We have max workers running
  161. o Consider taking the master out of the loop?
  162. D Implement reply onions
  163. o Total rate limiting
  164. o Look at OR handshake in more detail
  165. o Spec it
  166. o Merge OR and OP handshakes
  167. o rearrange connection_or so it doesn't suck so much to read
  168. D Periodic link key rotation. Spec?
  169. o wrap malloc with something that explodes when it fails
  170. o Clean up the number of places that get to look at prkey