4 SPEC - Spec not finalized
16 Things we'd like to do in 0.2.0.x:
17 - See also Flyspray tasks.
18 - See also all items marked XXXX020 and DOCDOC in the code
19 - http://tor.eff.org/eff/legal-faq.html#License doesn't mention
20 licenses for other components of the bundles.
23 - Bug reports Roger has heard along that way that don't have enough
24 details/attention to solve them yet.
25 - tup said that when he set FetchUselessDescriptors, after
26 24 or 48 hours he wasn't fetching any descriptors at all
27 anymore. This was in 0.2.0 but worked fine in 0.1.2.
28 - arma noticed that when his network went away and he tried
29 a new guard node and the connect() syscall failed to it,
30 the guard wasn't being marked as down. 0.2.0.x.
31 - after being without network for 12 hours, arma's tor decided
32 it couldn't fetch any network statuses, and never tried again
33 even when the network came back and arma clicked on things.
35 - phobos says relaybandwidth* sometimes don't do what we expect.
36 http://interloper.org/tmp/2007-06-bw-usage.png
37 - this notion of authorities notifying servers that they're
38 unreachable is bunk -- it's leftover from the time when all
39 servers ran 24/7. now it triggers every time a server goes
40 away and then returns before the old descriptor has expired.
41 o 0.2.0.x dir authorities have stopped giving people Authority
43 - add a --quiet commandline option that suppresses logs. useful
44 for --hashed-password and maybe others.
45 - Tor logs the libevent version on startup, for debugging purposes.
46 This is great. But it does this before configuring the logs, so
47 it only goes to stdout and is then lost.
48 o Aug 30 [...] wmtbf 4:6:16
49 The wmtbf time value is missing some zero-pads.
52 . 101: Voting on the Tor Directory System (plus 103)
54 * Describe schedule in copious detail.
55 - Get authorities voting
56 o Code to manage key certificates
59 . Code to retry download.
60 o Code to generate consensus from a list of votes
61 o Detect whether votes are really all for the same period.
62 . Push/pull documents as appropriate.
63 - Pull votes and signatures if we don't get them.
64 - Cache votes and signatures on disk?
65 o Code to keep consensus docs in limbo if they don't have
66 have enough signatures.
67 o Have clients know which authorities are v3 authorities, and what
69 - While we're at it, let v3 authorities have fqdns lines.
70 - Fix all XXXX020s in vote code
71 . Validate information properly.
72 o Warn if we get a vote with different authorities than we know.
73 o Don't count votes with a different valid-after when generating
75 - Dump certificates with the wrong time. Or just warn?
76 o Warn authority ops when their certs are nearly invalid.
77 - When checking a consensus, make sure that its times are plausible.
78 o Add a function that will eventually tell us about our clock skew.
79 For now, just require that authorities not be skewed.
80 - Start caching consensus documents once authorities make them
81 - Start downloading and using consensus documents once caches serve them
83 - GETINFO to get consensus
84 - Event when new consensus arrives
85 o 104: Long and Short Router Descriptors
86 o Drop bandwidth history from router-descriptors around September 10,
87 once torstatus and weasel have upgraded.
88 - 105: Version negotiation for the Tor protocol
89 . 111: Prioritize local traffic over relayed.
91 - Merge into tor-spec.txt.
94 . Make cells get buffered on circuit, not on the or_conn.
95 . Switch to pool-allocation for cells?
96 - Benchmark pool-allocation vs straightforward malloc.
97 - Adjust memory allocation logic in pools to favor a little less
99 . Remove socketpair-based bridges conns, and the word "bridge". (Use
100 shared (or connected) buffers for communication, rather than sockets.)
102 - Handle rate-limiting on directory writes to linked directory
103 connections in a more sensible manner.
104 - Find more ways to test this.
105 - Have clients do TLS connection rotation less often than "every 10
106 minutes" in the thrashy case, and more often than "once a week" in the
108 - Streamline how we pick entry nodes: Make choose_random_entry() have
109 less magic and less control logic.
110 o Maybe move NT services into their own module.
111 - Refactor networkstatus generation:
112 - Include "v" line in getinfo values.
115 . Bridges users (rudimentary version)
116 o Ability to specify bridges manually
117 o Config option 'UseBridges' that bridge users can turn on.
118 o uses bridges as first hop rather than entry guards.
119 o if you don't have any routerinfos for your bridges, or you don't
120 like the ones you have, ask a new bridge for its server/authority.
121 . Ask all directory questions to bridge via BEGIN_DIR.
122 - use the bridges for dir fetches even when our dirport is open.
123 R - drop 'authority' queries if they're to our own identity key; accept
125 N - Design/implement the "local-status" or something like it, from the
126 "Descriptor purposes: how to tell them apart" section of
127 http://archives.seul.org/or/dev/May-2007/msg00008.html
128 - cache of bridges that we've learned about and use but aren't
129 manually listed in the torrc.
130 o timeout and retry schedules for fetching bridge descriptors
131 - give extend_info_t a router_purpose again
132 o react faster to download networkstatuses after the first bridge
134 o be more robust to bridges being marked as down and leaving us
135 stranded without any known "running" bridges.
136 - Bridges operators (rudimentary version)
137 - Ability to act as dir cache without a dir port.
138 o Bridges publish to bridge authorities
139 o Fix BEGIN_DIR so that you connect to bridge of which you only
140 know IP (and optionally fingerprint), and then use BEGIN_DIR to learn
142 - look at server_mode() and decide if it always applies to bridges too.
143 - Bridges authorities (rudimentary version)
144 o Rudimentary "do not publish networkstatus" option for bridge
146 - Clients can ask bridge authorities for more bridges.
148 o Clients can ask bridge authorities for updates on known bridges.
149 - More TLS normalization work: make Tor less easily
151 - Directory system improvements
153 - Make BEGIN_DIR mandatory for asking questions of bridge authorities?
155 - Features (other than bridges):
156 - Blocking-resistance.
157 - Write a proposal; make this part of 105.
158 - Audit how much RAM we're using for buffers and cell pools; try to
160 o Accept \n as end of lines in the control protocol in addition to \r\n.
161 o Use fetch_from_buf_line_lf in control.c instead of fetch_from_buf_line.
162 o Fix up read escaped_data to accept LF instead of CRLF, and to
163 always translate_newlines (since that's the only way it's called).
164 - Base relative control socket paths on datadir.
165 - We should ship with a list of stable dir mirrors -- they're not
166 trusted like the authorities, but they'll provide more robustness
167 and diversity for bootstrapping clients.
168 - Better estimates in the directory of whether servers have good uptime
169 (high expected time to failure) or good guard qualities (high
171 - AKA Track uptime as %-of-time-up, as well as time-since-last-down
173 - Make uptime info persist too.
174 - Base Guard on weighted fractional uptime.
175 - Make TrackHostExits expire TrackHostExitsExpire seconds after their
176 *last* use, not their *first* use.
177 - Limit to 2 dir, 2 OR, N SOCKS connections per IP.
178 - Or maybe close connections from same IP when we get a lot from one.
179 - Or maybe block IPs that connect too many times at once.
180 - add an AuthDirBadexit torrc option if we decide we want one.
181 o Add a GuardsSelectedByVersion line to the state file so we know
182 not to drop guards we added.
183 o Have it include the date too.
186 N - Hack up a client that gives out weird/no certificates, so we can
187 test to make sure that this doesn't cause servers to crash.
190 - can we deprecate 'getinfo network-status'?
191 - can we deprecate the FastFirstHopPK config option?
195 - Tell people about OSX Uninstaller
196 - Quietly document NT Service options
197 - More prominently, we should have a recommended apps list.
199 - unrecommend IE because of ftp:// bug.
200 - we should add a preamble to tor-design saying it's out of date.
201 . Document transport and natdport in a good HOWTO.
202 - Publicize torel. (What else?
203 . Finish path-spec.txt
206 P - Can we switch to polipo? Please?
207 - Make documentation realize that location of system configuration file
208 will depend on location of system defaults, and isn't always /etc/torrc.
209 P - If we haven't replaced privoxy, lock down its configuration in all
210 packages, as documented in tor-doc-unix.html
211 P - Figure out why dll's compiled in mingw don't work right in WinXP.
212 P - Figure out why openssl 0.9.8e "make test" fails at sha256t test.
213 P - Figure out if including RSA and IDEA are bad for Tor from a legal
215 P - Create packages for Nokia 800, requested by Chris Soghoian
216 P - Consider creating special Tor-Polipo-Vidalia test packages,
217 requested by Dmitri Vitalev
218 - Get Vidalia supporting protocolinfo and using auth by default.
219 - Try to get newer Torbuttons into the windows bundle.
221 Nice-to-have items for 0.2.0.x, time permitting:
223 - 113: Simplifying directory authority administration
224 - 110: prevent infinite-length circuits (phase one)
225 . Robust decentralized storage for hidden service descriptors.
226 (Karsten is working on this; proposal 114.)
227 - 118: Listen on and advertise multiple ports:
228 - Tor should be able to have a pool of outgoing IP addresses that it is
229 able to rotate through. (maybe. Possible overlap with proposal 118.)
230 - config option to publish what ports you listen on, beyond
231 ORPort/DirPort. It should support ranges and bit prefixes (?) too.
232 (This is very similar to proposal 118.)
234 - Internal code support for ipv6:
235 o Clone ipv6 functions (inet_ntop, inet_pton) where they don't exist.
236 - Most address variables need to become tor_addr_t
237 - Teach resolving code how to handle ipv6.
238 - Teach exit policies about ipv6 (consider ipv4/ipv6 interaction!)
241 - Let controller set router flags for authority to transmit, and for
243 - add an 'exit-address' line in the descriptor for servers that exit
244 from something that isn't their published address.
245 - Clients should estimate their skew as median of skew from servers
247 - More work on AvoidDiskWrites?
250 - MAYBE kill stalled circuits rather than stalled connections. This is
251 possible thanks to cell queues, but we need to consider the anonymity
253 - Implement TLS shutdown properly when possible.
256 - we try to build 4 test circuits to break them over different
257 servers. but sometimes our entry node is the same for multiple
258 test circuits. this defeats the point.
259 - If the client's clock is too far in the past, it will drop (or just not
260 try to get) descriptors, so it'll never build circuits.
263 - Move all status info out of routerinfo into local_routerstatus. Make
264 "who can change what" in local_routerstatus explicit. Make
265 local_routerstatus (or equivalent) subsume all places to go for "what
269 - Detect correct version of libraries from autoconf script.
272 - Review torrc.sample to make it more discursive.
274 Deferred from 0.2.0.x:
278 - Make resolves no longer use edge_connection_t unless they are actually
279 _on_ a socks connection: have edge_connection_t and (say)
280 dns_request_t both extend an edge_stream_t, and have p_streams and
281 n_streams both be linked lists of edge_stream_t.
282 - Generate torrc.{complete|sample}.in, tor.1.in, the HTML manual, and the
283 online config documentation from a single source.
284 - Blocking/scanning-resistance
285 - It would be potentially helpful to https requests on the OR port by
286 acting like an HTTPS server.
287 - Do we want to maintain our own set of entryguards that we use as
288 next hop after the bridge? Open research question; let's say no
289 for 0.2.0 unless we learn otherwise.
290 - Should do reachability testing but only on the purpose==bridge
292 - Some mechanism for specifying that we want to stop using a cached
297 - See also Flyspray tasks.
298 - See also all OPEN/ACCEPTED proposals.
299 - See also all items marked XXXX and FFFF in the code.
302 - Our current approach to block attempts to use Tor as a single-hop proxy
303 is pretty lame; we should get a better one.
304 - Allow small cells and large cells on the same network?
305 - Cell buffering and resending. This will allow us to handle broken
306 circuits as long as the endpoints don't break, plus will allow
307 connection (tls session key) rotation.
308 - Implement Morphmix, so we can compare its behavior, complexity,
309 etc. But see paper breaking morphmix.
310 - Other transport. HTTP, udp, rdp, airhook, etc. May have to do our own
311 link crypto, unless we can bully DTLS into it.
312 - Need a relay teardown cell, separate from one-way ends.
313 (Pending a user who needs this)
314 - Handle half-open connections: right now we don't support all TCP
315 streams, at least according to the protocol. But we handle all that
316 we've seen in the wild.
317 (Pending a user who needs this)
321 - turn the received socks addr:port into a digest for setting .exit
322 - handle connect-dir streams that don't have a chosen_exit_name set.
323 - Have a "Faster" status flag that means it. Fast2, Fast4, Fast8?
324 - Add an option (related to AvoidDiskWrites) to disable directory
325 caching. (Is this actually a good idea??)
326 - Add d64 and fp64 along-side d and fp so people can paste status
327 entries into a url. since + is a valid base64 char, only allow one
328 at a time. Consider adding to controller as well.
329 - Some back-out mechanism for auto-approval on authorities
330 - a way of rolling back approvals to before a timestamp
331 - Consider minion-like fingerprint file/log combination.
332 - Have new people be in limbo and need to demonstrate usefulness
333 before we approve them.
336 - Standby/hotswap/redundant hidden services.
337 . Update the hidden service stuff for the new dir approach. (Much
338 of this will be superseded by 114.)
339 - switch to an ascii format, maybe sexpr?
340 - authdirservers publish blobs of them.
341 - other authdirservers fetch these blobs.
342 - hidserv people have the option of not uploading their blobs.
343 - you can insert a blob via the controller.
344 - and there's some amount of backwards compatibility.
345 - teach clients, intro points, and hidservs about auth mechanisms.
346 - come up with a few more auth mechanisms.
347 - auth mechanisms to let hidden service midpoint and responder filter
349 - Let each hidden service (or other thing) specify its own
351 - Hidserv offerers shouldn't need to define a SocksPort
354 - When we notice a 'Rejected: There is already a named server with
355 this nickname' message... or maybe instead when we see in the
356 networkstatuses that somebody else is Named with the name we
357 want: warn the user, send a STATUS_SERVER message, and fall back
359 - If the server is spewing complaints about raising your ulimit -n,
360 we should add a note about this to the server descriptor so other
361 people can notice too.
362 - When we hit a funny error from a dir request (eg 403 forbidden),
363 but tor is working and happy otherwise, and we haven't seen many
364 such errors recently, then don't warn about it.
367 - A way to adjust router flags from the controller. (How do we
368 prevent the authority from clobbering them soon afterward?)
369 - Implement missing status events and accompanying getinfos
371 - BAD_DIR_RESPONSE (Unexpected directory response; maybe we're behind
373 - BAD_PROXY (Bad http or https proxy)
374 - UNRECOGNIZED_ROUTER (a nickname we asked for is unavailable)
375 - Status events related to hibernation
376 - something about failing to parse our address?
377 from resolve_my_address() in config.c
378 - sketchy OS, sketchy threading
379 - too many onions queued: threading problems or slow CPU?
380 - Implement missing status event fields:
381 - TIMEOUT on CHECKING_REACHABILITY
382 - GETINFO status/client, status/server, status/general: There should be
383 some way to learn which status events are currently "in effect."
384 We should specify which these are, what format they appear in, and so
386 - More information in events:
387 - Include bandwidth breakdown by conn->type in BW events.
388 - Change circuit status events to give more details, like purpose,
389 whether they're internal, when they become dirty, when they become
390 too dirty for further circuits, etc.
391 - Change stream status events analogously.
392 - Expose more information via getinfo:
393 - import and export rendezvous descriptors
394 - Review all static fields for additional candidates
395 - Allow EXTENDCIRCUIT to unknown server.
396 - We need some way to adjust server status, and to tell tor not to
397 download directories/network-status, and a way to force a download.
398 - Make everything work with hidden services
400 - Performance/resources
401 - per-conn write buckets
402 - separate config options for read vs write limiting
403 (It's hard to support read > write, since we need better
404 congestion control to avoid overfull buffers there. So,
405 defer the whole thing.)
406 - Investigate RAM use in directory authorities.
407 - Look into pulling serverdescs off buffers as they arrive.
408 - Rate limit exit connections to a given destination -- this helps
409 us play nice with websites when Tor users want to crawl them; it
410 also introduces DoS opportunities.
411 - Consider truncating rather than destroying failed circuits,
412 in order to save the effort of restarting. There are security
413 issues here that need thinking, though.
414 - Handle full buffers without totally borking
415 - Rate-limit OR and directory connections overall and per-IP and
419 - Hold-open-until-flushed now works by accident; it should work by
421 - Display the reasons in 'destroy' and 'truncated' cells under
423 - Make router_is_general_exit() a bit smarter once we're sure what
425 - Automatically determine what ports are reachable and start using
426 those, if circuits aren't working and it's a pattern we
427 recognize ("port 443 worked once and port 9001 keeps not
431 - don't do dns hijacking tests if we're reject *:* exit policy?
432 (deferred until 0.1.1.x is less common)
435 - Servers might check certs for known-good ssl websites, and if
436 they come back self-signed, declare themselves to be
437 non-exits. Similar to how we test for broken/evil dns now.
438 - Authorities should try using exits for http to connect to some
439 URLS (specified in a configuration file, so as not to make the
440 List Of Things Not To Censor completely obvious) and ask them
441 for results. Exits that don't give good answers should have
442 the BadExit flag set.
443 - Alternatively, authorities should be able to import opinions
444 from Snakes on a Tor.
445 - More consistent error checking in router_parse_entry_from_string().
446 I can say "banana" as my bandwidthcapacity, and it won't even squeak.
447 - Bind to random port when making outgoing connections to Tor servers,
448 to reduce remote sniping attacks.
449 - Audit everything to make sure rend and intro points are just as
450 likely to be us as not.
451 - Do something to prevent spurious EXTEND cells from making
452 middleman nodes connect all over. Rate-limit failed
453 connections, perhaps?
454 - DoS protection: TLS puzzles, public key ops, bandwidth exhaustion.
457 - Tolerate clock skew on bridge relays.
460 - Now that we're avoiding exits when picking non-exit positions,
461 we need to consider how to pick nodes for internal circuits. If
462 we avoid exits for all positions, we skew the load balancing. If
463 we accept exits for all positions, we leak whether it's an
464 internal circuit at every step. If we accept exits only at the
465 last hop, we reintroduce Lasse's attacks from the Oakland paper.
467 - Windows server usability
468 - Solve the ENOBUFS problem.
469 - make tor's use of openssl operate on buffers rather than sockets,
470 so we can make use of libevent's buffer paradigm once it has one.
471 - make tor's use of libevent tolerate either the socket or the
472 buffer paradigm; includes unifying the functions in connect.c.
473 - We need a getrlimit equivalent on Windows so we can reserve some
474 file descriptors for saving files, etc. Otherwise we'll trigger
475 asserts when we're out of file descriptors and crash.
476 - Merge code from Urz into libevent
477 - Make Tor use evbuffers.
480 - a way to generate the website diagrams from source, so we can
481 translate them as utf-8 text rather than with gimp.
482 . Flesh out options_description array in src/or/config.c
483 . multiple sample torrc files
484 . figure out how to make nt service stuff work?
486 - Refactor tor man page to divide generally useful options from
488 - Add a doxygen style checker to make check-spaces so nick doesn't drift
489 too far from arma's undocumented styleguide. Also, document that
490 styleguide in HACKING. (See r9634 for example.)
491 - exactly one space at beginning and at end of comments, except i
492 guess when there's line-length pressure.
493 - if we refer to a function name, put a () after it.
494 - only write <b>foo</b> when foo is an argument to this function.
495 - doxygen comments must always end in some form of punctuation.
496 - capitalize the first sentence in the doxygen comment, except
498 - avoid spelling errors and incorrect comments. ;)
501 - The Debian package now uses --verify-config when (re)starting,
502 to distinguish configuration errors from other errors. Perhaps
503 the RPM and other startup scripts should too?
504 - add a "default.action" file to the tor/vidalia bundle so we can
505 fix the https thing in the default configuration:
506 http://wiki.noreply.org/noreply/TheOnionRouter/TorFAQ#PrivoxyWeirdSSLPort
509 - Patch privoxy and socks protocol to pass strings to the browser.
512 Documentation, non-version-specific.
514 - Mark up spec; note unclear points about servers
515 NR - write a spec appendix for 'being nice with tor'
516 - Specify the keys and key rotation schedules and stuff
517 - Mention controller libs someplace.
518 - Remove need for HACKING file.
519 P - document http://wiki.noreply.org/noreply/TheOnionRouter/TransparentProxy on freebsd and osx
520 P - figure out why x86_64 won't build rpms from tor.spec
521 P - figure out spec files for bundles of vidalia-tor-polipo
522 P - figure out polipo install scripts for bundles of vidalia-tor-polipo on osx, win32
523 P - evaluate https://sourceforge.net/projects/kleanup/
525 P - figure out selinux policy for tor
526 P - change packaging system to more automated and specific for each
527 platform, suggested by Paul Wouter
530 - tor-in-the-media page
531 . more pictures from ren. he wants to describe the tor handshake
532 - Figure out licenses for website material.
533 - and remove home and make the "Tor" picture be the link to home.
534 - put the logo on the website, in source form, so people can put it on
535 stickers directly, etc.
536 - put the source image for the stickers on the website, so people can
538 R - make a page with the hidden service diagrams.
539 - ask Jan to be the translation coordinator? add to volunteer page.
540 - add a page for localizing all tor's components.
541 - It would be neat if we had a single place that described _all_ the
542 tor-related tools you can use, and what they give you, and how well they
543 work. Right now, we don't give a lot of guidance wrt
544 torbutton/foxproxy/privoxy/polipo in any consistent place.