Commit Graph

105 Commits

Author SHA1 Message Date
Nick Mathewson
b7cdcf3462 Hopefully, this will make ORs much faster, and not break them: keep a big splay tree of (circid,orconn)->circuit mappings to make circuit_get_by_circid_conn much faster.
svn:r4020
2005-04-06 05:33:32 +00:00
Nick Mathewson
055ee7c323 Give better warnings if connection_close_unattached_ap gets called twice or called on a marked connection; rename it to connection_mark_unattached_ap.
svn:r3990
2005-04-02 22:11:24 +00:00
Nick Mathewson
0e81265359 update copyright notices.
svn:r3982
2005-04-01 20:15:56 +00:00
Roger Dingledine
1ff62dbd18 require exactly 0.0.9.7 for middle hops of testing circs, for now
svn:r3967
2005-04-01 09:23:51 +00:00
Roger Dingledine
e786e5dedb when building testing circuits, always use middle hops
running at least 0.0.9.7


svn:r3966
2005-04-01 08:42:26 +00:00
Roger Dingledine
2ff2db8d2a fix some log spacing problems
svn:r3942
2005-04-01 06:23:21 +00:00
Roger Dingledine
1f5c8335e8 still publish your descriptor if orport is reachable but dirport isn't
when building testing circs for orport testing, require high-bandwidth
nodes, so fewer circs fail. complain about unreachable orport separately
from unreachable dirport.


svn:r3935
2005-03-31 19:26:33 +00:00
Roger Dingledine
4f5192b280 ...and use it to make sure we only conclude reachability if
we didn't initiate the conn.


svn:r3932
2005-03-31 07:46:59 +00:00
Roger Dingledine
b8fe866a9d forward-port the fixing of the crash bug.
svn:r3900
2005-03-27 07:18:30 +00:00
Roger Dingledine
36baf7219d stop most cases of hanging up on a socks connection without sending
the socks reject. audit for remaining ones. also make things more
uniform so we always remember to hold-open-until-flushed, etc.


svn:r3891
2005-03-27 04:55:13 +00:00
Roger Dingledine
b8d38dc57b try harder to establish reachability, in the first 20 minutes
of uptime.


svn:r3884
2005-03-26 05:54:50 +00:00
Roger Dingledine
1328383482 make it clearer to the human that his server is testing
its reachability. tell him when it succeeds, or when 20
minutes pass and it hasn't succeeded yet.


svn:r3882
2005-03-26 01:43:39 +00:00
Nick Mathewson
631ab5c69b Add a magic value to cpath_layer_t to make sure that we can tell valid cpaths from freed ones. I audited this once; it could use another audit.
svn:r3831
2005-03-23 06:21:48 +00:00
Roger Dingledine
ad5f5ddf7e remove a duplicate circuit_mark_for_close() when picking
a path for the circuit fails.


svn:r3822
2005-03-22 23:20:33 +00:00
Roger Dingledine
f1ba4296c1 clean up and refactor some more
svn:r3798
2005-03-22 01:01:15 +00:00
Roger Dingledine
35953edae0 Implement controller's "extendcircuit" directive.
Also refactor circuit building so we plan the whole path ahead
of time.


svn:r3797
2005-03-22 00:42:38 +00:00
Roger Dingledine
4a497e5030 if our clock jumps forward by 100 seconds or more, assume something
has gone wrong with our network and abandon all not-yet-used circs.


svn:r3792
2005-03-19 23:58:42 +00:00
Roger Dingledine
856ab90ca8 stop logging at -l notice every single time a create cell successfully
gets processed


svn:r3791
2005-03-19 23:04:15 +00:00
Nick Mathewson
b88c4ba11c Turn addr_policy_compare from a tristate to a quadstate; this should help address our "Ah, you allow 1.2.3.4:80. You are a good choice for google.com" problem.
svn:r3786
2005-03-19 06:57:16 +00:00
Roger Dingledine
229761465c re-enable reachability testing stuff.
also, consider your ORPort reachable after you've processed a
create cell from any non-local address.


svn:r3763
2005-03-15 01:44:46 +00:00
Nick Mathewson
97dad670ea Renormalize whitespace
svn:r3757
2005-03-14 03:18:35 +00:00
Roger Dingledine
a1f1fa6ab2 Checkpoint in-progress fixes:
Add 'testing' circuit purpose, for reachability testing.
Notice when our IP changes, and reset stats.
Try to pull down a directory via Tor to see if our DirPort is working.
Try to extend a circuit back to us to see if our ORPort is working.
Only publish a descriptor if they're both reachable.

These mostly work, and I'd better get them in before I cause conflicts.


svn:r3703
2005-02-27 09:47:01 +00:00
Nick Mathewson
6e6d95b3db Change from inet_ntoa to a threadproof tor_inet_ntoa.
svn:r3656
2005-02-22 08:18:36 +00:00
Roger Dingledine
797419a62c remove unused code
svn:r3629
2005-02-16 02:06:54 +00:00
Roger Dingledine
2e97593d25 define TOR_FRAGILE if you want tor to give you a core when
something goes wrong. this should only be used by people actively
tracking bugs.


svn:r3487
2005-02-01 00:37:16 +00:00
Roger Dingledine
75d83827e0 fix another bug where general exit circs were looking at rendezvous
streams. consolidate the looking into one function so it doesn't
happen again.


svn:r3372
2005-01-19 17:59:37 +00:00
Roger Dingledine
ea24dc31bc remove some extra-verbose logs
svn:r3361
2005-01-17 18:49:13 +00:00
Roger Dingledine
d2400a5afd Introduce a notion of 'internal' circs, which are chosen without regard
to the exit policy of the last hop. Intro and rendezvous circs must
be internal circs, to avoid leaking information. Resolve and connect
streams can use internal circs if they want.

New circuit pooling algorithm: make sure to have enough circs around
to satisfy any predicted ports, and also make sure to have 2 internal
circs around if we've required internal circs lately (with high uptime
if we've seen that lately).

Split NewCircuitPeriod config option into NewCircuitPeriod (30 secs),
which describes how often we retry making new circuits if current ones
are dirty, and MaxCircuitDirtiness (10 mins), which describes how long
we're willing to make use of an already-dirty circuit.

Once rendezvous circuits are established, keep using the same circuit as
long as you attach a new stream to it at least every 10 minutes. (So web
browsing doesn't require you to build new rend circs every 30 seconds.)

Cannibalize GENERAL circs to be C_REND, C_INTRO, S_INTRO, and S_REND
circ as necessary, if there are any completed ones lying around when
we try to launch one.

Re-instate the ifdef's to use version-0 style introduce cells, since
there was yet another bug in handling version-1 style. We'll try switching
over again after 0.0.9 is obsolete.

Bugfix: when choosing an exit node for a new non-internal circ, don't take
into account whether it'll be useful for any pending x.onion addresses --
it won't.

Bugfix: we weren't actually publishing the hidden service descriptor when
it became dirty. So we only published it every 20 minutes or so, which
means when you first start your Tor, the hidden service will seem broken.


svn:r3360
2005-01-17 18:13:09 +00:00
Roger Dingledine
84c81e0783 Annotate circuits w/ whether they aim to contain high uptime nodes and/or
high capacity nodes. When building circuits, choose appropriate nodes.

New config option LongLivedPorts to indicate application streams
that will want high uptime circuits.

When attaching a stream to a circuit, pay attention to its requirements.

This means that every single node in an intro rend circuit, not just
the last one, will have a minimum uptime.

Boost the min uptime from an hour to 24 hours.


svn:r3339
2005-01-12 04:58:23 +00:00
Roger Dingledine
308ffa8a05 When listing circuits that are pending on an opening OR connection,
if we're an OR we were listing circuits that *end* at us as
being pending on every listener, dns/cpu worker, etc. Stop that.


svn:r3171
2004-12-17 23:17:00 +00:00
Roger Dingledine
011ad3cba5 clean up logging,
make it clearer which warns are bugs,
make the control log event match its specification,
point out a bug in how we deal with failure when renewing the tls context.


svn:r3138
2004-12-13 00:44:39 +00:00
Roger Dingledine
802d374a99 I'm a bad person.
Stop treating the uint16_t's as null-terminated strings,
and stop looking at the byte after them to see if it's null,
because sometimes you're not allowed to look there.


svn:r3108
2004-12-07 15:29:54 +00:00
Roger Dingledine
a2d80ec767 were we on crack??
svn:r3107
2004-12-07 09:18:25 +00:00
Roger Dingledine
d9f86e2782 and bust another memory leak
svn:r3105
2004-12-07 08:58:26 +00:00
Nick Mathewson
a6aa5eebd6 Fix some memory leaks and unlikely segfaults
svn:r3103
2004-12-07 07:48:16 +00:00
Roger Dingledine
5ba3328172 fix a minor leak in my recent commit
svn:r3081
2004-12-05 12:35:00 +00:00
Roger Dingledine
ef6c9d18e7 New circuit building strategy: keep a list of ports that we've used in the past 6 hours, and always try to have 2 circuits open or on the way
that will handle each such port. (We can extend this to include addresses
if exit policies shift to require that.) Seed us with port 80 so web
browsers won't complain that Tor is "slow to start up".

This was necessary because our old circuit building strategy just involved
counting circuits, and as time went by we would build up a big pile of
circuits that had peculiar exit policies (e.g. only exit to 9001-9100)
which would take up space in the circuit pile but never get used.

Fix router_compare_addr_to_addr_policy: it was not treating a port of *
as always matching, so we were picking reject *:* nodes as exit nodes too.

If you haven't used a clean circuit in an hour, throw it away, just to
be on the safe side.

This means after 6 hours a totally unused Tor client will have no
circuits open.


svn:r3078
2004-12-05 07:10:08 +00:00
Roger Dingledine
c644886c38 bugfix: When we were checking to see if an ap_conn should time out
waiting for its connected cell, we were calculating time from when the
ap_conn was created. So if it waited say 20 seconds before being attached,
then we would immediately decide that the circuit had timed out.

Also, make circuit_dump_by_conn() display actual circuit progress,
including circuits that haven't been attached to the conn yet but
hope to when it finishes connecting.


svn:r3072
2004-12-04 02:51:11 +00:00
Roger Dingledine
8d69dc472b make circuit building and router parsing less noisy
svn:r3068
2004-12-04 00:25:54 +00:00
Roger Dingledine
3929f58018 clean obsolete circuit_log_path code
svn:r3046
2004-11-30 10:47:48 +00:00
Nick Mathewson
7fbd297532 Suggestion from weasel: Make tor --version --version dump the cvs Id of every file.
svn:r3019
2004-11-29 22:25:31 +00:00
Roger Dingledine
22727b4edc wrong is ok, and right is fine, but in between is apparently
totally unacceptable to me.


svn:r3005
2004-11-28 11:39:53 +00:00
Nick Mathewson
6f5dbefa7e Normalize space: add one between every control keyword and control clause.
svn:r3003
2004-11-28 09:05:49 +00:00
Roger Dingledine
7c9a707900 remove emacs droppings, since nick says he doesn't need them anymore
svn:r2989
2004-11-26 04:00:55 +00:00
Nick Mathewson
f5319a010d Dr. Seuss on iterating circular lists: "It's fun to have fun, but you
have to know how."


svn:r2948
2004-11-23 06:08:08 +00:00
Nick Mathewson
0a58bbe0dd Fix critical bug in circuit_list_path: cpath is a circular list! (Also reimplement circuit_log_cpath using circuit_list_cpath).
svn:r2946
2004-11-23 00:11:36 +00:00
Roger Dingledine
6a516dfdd3 be more greedy about filling up all relay cells.
this may have some bugs in it still.
and it may end up not being what we want to do.


svn:r2928
2004-11-21 07:43:12 +00:00
Nick Mathewson
647c076c8f add a TODO item and some comment changes.
svn:r2904
2004-11-16 03:32:01 +00:00
Nick Mathewson
60880cda84 Resolve a bunch of FIXME items; mark a lot more for attention; ask for clarification on some. Turn all XXXX008 ("showstopper for 0.0.8 release") items into XXXX009 or XXXX, since plainly they were not showstoppers for 0.0.8. Add/clean some docs.
svn:r2808
2004-11-12 16:39:03 +00:00
Nick Mathewson
5d92fbe30e Use autoconf to enable largefile support where necessary. Use ftello where available, since ftell can fail at 2GB.
svn:r2806
2004-11-12 05:05:41 +00:00