Commit Graph

5208 Commits

Author SHA1 Message Date
Nick Mathewson
3bd613b44d Rename connection_get_by_identity_digest to reflect that it is OR-only. Make it use a hashtable instead of a linear search.
svn:r5469
2005-11-30 03:01:16 +00:00
Nick Mathewson
6b49a93bd5 Move code to check for zero digests into util.c
svn:r5468
2005-11-30 02:47:59 +00:00
Nick Mathewson
447a3a7ff2 Funny. I could have sworn that it compiled.
svn:r5467
2005-11-29 17:17:02 +00:00
Nick Mathewson
94e8dcef44 Fix a bug weasel found: when we are an authority and somebody requests "all" statuses, cconsider regenerating our own.
svn:r5466
2005-11-29 17:14:10 +00:00
Nick Mathewson
c90d6c7e5f Make arguments to circuit_set_circid_orconn match the connection that is closing in circuit_about_to_close_connection. This accounts for our mysterious "arbitrary ram gets decremented!" bug.
svn:r5465
2005-11-29 17:13:34 +00:00
Nick Mathewson
7de0c33857 Avoid a segfault in routerlist_reset_warnings()
svn:r5464
2005-11-28 16:29:27 +00:00
Roger Dingledine
1da454a42d when returning a list of circuits, don't return a list of the same
conn over and over. that's really bad for memory.


svn:r5463
2005-11-28 02:52:17 +00:00
Roger Dingledine
f94f4c1686 when event_add or event_del fail, tell us why.
(nick, do i have my libevent strerror dance moves correct?)


svn:r5462
2005-11-26 09:37:00 +00:00
Roger Dingledine
97f748f680 clean up broken comment
svn:r5461
2005-11-26 09:27:36 +00:00
Nick Mathewson
652e1899ac "How about 'never'? Does 'never' work for you?"
Weasel says circuit_get_by_conn is his main timesink.  Most of its
users were just checking whether OR conns had circuits, so add a
circuit count to OR conns, and check that. One was
circuit_about_to_close_conn, which was doing an O(n^2) series of calls
to get all circs on an OR conn, so make an O(n) function for that.
Finally, circuit_get_by_edge_conn was using it as a sanity test that
has been around for a while but never found any actualy insanity, so
kill that.

circuit_get_by_conn is finally dead, which is good, since it was never
sane to begin with.


svn:r5460
2005-11-26 01:43:57 +00:00
Nick Mathewson
31d5d96739 Cut down a common call to circuit_get_by_conn by about half.
svn:r5459
2005-11-26 00:53:51 +00:00
Nick Mathewson
72cb64406a HT_NEXT invalidates the last iterator; fix rmv-related segfault.
svn:r5458
2005-11-26 00:42:25 +00:00
Roger Dingledine
fe221f3dff Start the process of treating internal circuits and exit circuits
separately. It's important to keep them separate because internal
circuits have their last hops picked like middle hops, rather than like
exit hops. So exiting on them will break the user's expectations.

- Stop cannibalizing internal circuits for general exits, and stop
  cannibalizing exit circuits for rendezvous stuff.

- Don't let new exit streams attach to internal circuits.

- When deciding if we have enough circuits for internal and for exit,
  don't count the wrong ones.

- Treat predicted resolves as predicted port 80 exits.


svn:r5457
2005-11-25 08:08:56 +00:00
Roger Dingledine
6452aecedb when calling connection_get_by_identity_digest, don't do a memcmp
for non-OR conns. this should save a bit of time.


svn:r5456
2005-11-25 06:43:53 +00:00
Roger Dingledine
b6b1b4f283 Bandaid workaround to make cvs not crash tor clients.
This is not a real fix. I didn't look at the rest of the code.
Nick?


svn:r5455
2005-11-25 02:16:10 +00:00
Roger Dingledine
2350e487a2 minor cleanup
svn:r5454
2005-11-24 06:44:13 +00:00
Roger Dingledine
ea7fe0c267 Fix a bug found by Lasse Overlier: when we were making internal circuits
(intended to be cannibalized later for rendezvous and introduction
circuits), we were picking them so that they had useful exit nodes. There
was no need for this, and it actually aids some statistical attacks.


svn:r5453
2005-11-24 06:43:39 +00:00
Roger Dingledine
5b583073a1 remove some more dead code (found while hunting lasse's bug)
svn:r5452
2005-11-24 06:06:50 +00:00
Roger Dingledine
23efc30776 get rid of some unused variables
svn:r5451
2005-11-24 01:47:05 +00:00
Roger Dingledine
276ca7ad28 link with -ldl if you need it.
svn:r5450
2005-11-23 20:15:39 +00:00
Roger Dingledine
4e77913c67 whitespace and copyright fixes on configure.in
svn:r5449
2005-11-23 20:15:23 +00:00
Nick Mathewson
05482a0899 Add functions to return number of elts in associative containers
svn:r5448
2005-11-23 07:48:13 +00:00
Roger Dingledine
0118a236a4 comment that scary-looking function.
svn:r5447
2005-11-23 07:30:44 +00:00
Nick Mathewson
eb0021b06e Fix dumb logic error that kept some old descripttors from getting the axe.
svn:r5446
2005-11-23 07:28:54 +00:00
Nick Mathewson
061ef582a1 Make dropping very old descriptors faster (remove old correctness check). Track bytes dropped that are still in our store or journal, and rebuild when it gets very high.
svn:r5445
2005-11-23 07:24:59 +00:00
Nick Mathewson
36109fea33 e more aggressive about calling routerlist_remove_old_routers()
svn:r5444
2005-11-23 07:06:36 +00:00
Nick Mathewson
fbf8a8c25f Dump bytes held to store descriptors.
svn:r5443
2005-11-23 06:00:58 +00:00
Nick Mathewson
7d7bb95f36 Track how many bytes we have in signed descriptors. Try to make sure we are tracking this right. Somebody should valgrind a dirserver on an example net. There should be code to dump this value.
svn:r5442
2005-11-23 05:47:51 +00:00
Nick Mathewson
a39269572f Replace balanced trees with hash tables: this should make stuff significantly faster.
svn:r5441
2005-11-23 04:18:45 +00:00
Peter Palfrader
ae67b87f9a Update debian/tor.docs to match which docs are now distributed
svn:r5440
2005-11-22 17:40:26 +00:00
Nick Mathewson
50ce5e1932 Avoid spurious re-launch of first pending requested descriptor
svn:r5439
2005-11-22 02:20:05 +00:00
Roger Dingledine
8182eb69cd a smidgen more hint when we fail to seed the rng
svn:r5438
2005-11-21 01:12:01 +00:00
Peter Palfrader
a48b722432 Ancient gcc does not like you when you declare variables in the middle of a block
svn:r5437
2005-11-20 16:53:49 +00:00
Peter Palfrader
b9d37a2d58 Add a new controller event type that allows controllers to get all
server descriptors that were uploaded to a router in its role as authoritative
dirserver.


svn:r5436
2005-11-19 18:35:43 +00:00
Roger Dingledine
7446c7e6d7 mark off the next todo item
svn:r5435
2005-11-19 11:07:40 +00:00
Roger Dingledine
f28770b639 be less eager to decide that the entire network is down
svn:r5434
2005-11-19 11:04:00 +00:00
Roger Dingledine
859d1c902a and another instance where i was trying to be too sneaky
svn:r5433
2005-11-19 10:38:23 +00:00
Roger Dingledine
98f353fa7c actually listen when we try to route our dir connection via tor.
svn:r5432
2005-11-19 10:33:34 +00:00
Roger Dingledine
5c949fce51 Start obeying our firewall options more rigorously:
- If we can't get to a dirserver directly, try going via Tor.
  - Don't ever try to connect (as a client) to a place our firewall
    options forbid.
  - If we specify a proxy and also firewall options, obey the firewall
    options even when we're using the proxy: some proxies can only proxy
    to certain destinations.


svn:r5431
2005-11-19 10:12:10 +00:00
Roger Dingledine
b819a53654 mark off some todo items
svn:r5430
2005-11-19 07:16:12 +00:00
Roger Dingledine
7aae63994f Recover better from TCP connections to Tor servers that are broken but
don't tell you (it happens!); and rotate TLS connections once a week.

1) If an OR conn becomes more than a week old, make it obsolete.
2) If it's obsolete and empty, kill it.
3) When an OR makes a second connection to you, allow it.
4) If we want to send a new create cell, but the best conn we've
   got is obsolete, and the router is 0.1.1.9-alpha-cvs or later, ask
   for a new conn instead.
5) When we time out on circuit building on the first hop, make that
   connection obsolete.


svn:r5429
2005-11-19 06:57:44 +00:00
Roger Dingledine
700c370a3b misc cleanups
svn:r5428
2005-11-19 01:56:58 +00:00
Roger Dingledine
f824fb5f83 Bugfix: it looks like some middleman circuits were spending the
first few moments of their existence in CIRCUIT_STATE_OPEN, then
when Alice sent an extend request for a Tor that they're not connected
to, they switched to CIRCUIT_STATE_OR_WAIT and spent the rest of
their sorry little lives in that state, even when the connection
was established and they were shuttling relay cells back and forth.

And I'm not going to backport this (yet), because somehow it worked!


svn:r5427
2005-11-19 01:55:58 +00:00
Roger Dingledine
0de8f4ee55 another case (i think) of redundant code.
svn:r5426
2005-11-19 01:03:34 +00:00
Roger Dingledine
409294153f remove some redundant (and misleading) code
svn:r5425
2005-11-19 00:44:24 +00:00
Peter Palfrader
6734371234 15*60 is 15 minutes, not 5
svn:r5424
2005-11-18 21:57:49 +00:00
Peter Palfrader
901712ee59 Remove ip address change flapping detection. It is not really needed and I do not think it works quite right.
svn:r5423
2005-11-18 19:28:34 +00:00
Roger Dingledine
4ddf2385a0 if it's not less than three does that mean it's more than two?
svn:r5422
2005-11-18 11:35:23 +00:00
Roger Dingledine
9d6d6da883 and code the part where redirectstream can take a port.
svn:r5421
2005-11-18 11:32:59 +00:00
Roger Dingledine
a32a63d50b document how the error values for closestream and closecircuit work.
also, let redirectstream take a port, in case we want to change that too.


svn:r5420
2005-11-18 11:17:24 +00:00