Roger Dingledine
b384c5a18d
bugfix: do directory_send_command only after we've established a socket
...
also many comments, cleaning, etc
svn:r1823
2004-05-09 16:33:04 +00:00
Roger Dingledine
9bf9ca4d09
bugfix: add a circ->resolving_streams field, and make dns resolving streams not actually in the connection_array until they're done resolving.
...
bugfix: actually complain if we duplicate mark-for-close a circuit
add more logging for relay ends that claim dns resolve failed, so we can
find out why they're not being retried.
svn:r1798
2004-05-06 11:08:04 +00:00
Nick Mathewson
2ba0776b02
Only connection_add connections once they have conn->s sett; refactor code around this. Should make stuff more bulletproof.
...
svn:r1788
2004-05-05 01:26:57 +00:00
Roger Dingledine
bbbc063ed9
stop the circuit_receive_relay_cell warning
...
mark the circ for close if circuit_extend fails because the next
hop is not up yet; don't just return -1
svn:r1765
2004-05-02 03:32:00 +00:00
Nick Mathewson
ddb15b8f67
Remove IVs from cipher code, since AES-ctr has none.
...
svn:r1742
2004-04-28 20:31:32 +00:00
Nick Mathewson
7055f837ab
Make Tor build on win32 with VC6 without warnings.
...
svn:r1739
2004-04-28 20:13:21 +00:00
Roger Dingledine
9965bd8387
make my assumption explicit
...
svn:r1730
2004-04-27 11:28:45 +00:00
Roger Dingledine
505bb2f4a0
it was the second bug that was nailing us.
...
periodically we expire some circuits if we already have
enough, and we were counting rend circs in that number.
(the first one wasn't a bug after all)
svn:r1729
2004-04-27 11:23:56 +00:00
Roger Dingledine
8e8af6aaee
don't expire joined-rend circs that have an exit connection attached
...
and don't expire non-general excess circuits
svn:r1728
2004-04-27 10:16:31 +00:00
Nick Mathewson
f6dbe5a0d4
Refactor crypto error handling to be more like TLS error handling:
...
crypto_perror is a no-no, since an operation can set more than one
error.
Also, fix a bug in the unix crypto_seed_rng: mixing stdio with
/dev/urandom is a bad idea, since fopen can make all kinds of weird
extraneous syscalls (mmap, fcntl, stat64, etc.) and since fread tends
to buffer data in big chunks, thus depleting the entropy pool.
svn:r1717
2004-04-26 18:09:50 +00:00
Roger Dingledine
fb2279b90b
since we don't support truncateds much, don't bother sending them;
...
just close the circ.
(this wasn't relevant before, because we were mis-handling destroys.)
svn:r1711
2004-04-26 04:32:01 +00:00
Roger Dingledine
bd38df8472
fix a bug that's been lurking since 27 may 03 (!)
...
when passing back a destroy cell, we would use the wrong circ id.
how the heck did this work?
svn:r1710
2004-04-26 03:42:41 +00:00
Roger Dingledine
25d54257fb
use tor_assert
...
remove obsolete BUF_OK macro
svn:r1697
2004-04-25 20:37:37 +00:00
Nick Mathewson
0fca143ea1
Remove onion_pkey from connection, since onion keys can change more often than connections. Also add more log messages
...
svn:r1693
2004-04-25 19:04:11 +00:00
Roger Dingledine
c4b72a254f
tmp patch to also expire old clean rend_ready circs for now
...
will have to be fixed once we build-and-open rend circs preemptively
svn:r1668
2004-04-18 09:27:05 +00:00
Roger Dingledine
3fa2925a6a
if an intro circ waiting for an ack dies before getting one, then
...
count it as a nack
svn:r1665
2004-04-18 07:37:16 +00:00
Roger Dingledine
65d507b980
log n_circ_ids when attaching rend streams
...
svn:r1657
2004-04-17 10:25:38 +00:00
Roger Dingledine
26dd5c68b0
a ! goes a long way toward more helpful debug messages
...
svn:r1656
2004-04-17 10:04:00 +00:00
Roger Dingledine
f5427463f5
more debugging to hunt for a bug
...
svn:r1647
2004-04-17 01:33:04 +00:00
Roger Dingledine
9c1470840f
i was wrong about one of my invariants
...
now it's commented
this should fix asserts when you try an intro point, get a nack,
and try a second one and it works.
svn:r1637
2004-04-16 08:21:35 +00:00
Roger Dingledine
26c1c8f173
make unattached rend streams expire after a while (60s)
...
(they were darned persistent)
also make intro circs that are waiting for acks expire after a while (20s)
svn:r1628
2004-04-15 01:08:59 +00:00
Roger Dingledine
b51561f7fb
prune abandoned circs better
...
add debugging info for the bug weasel found
svn:r1626
2004-04-14 23:52:29 +00:00
Nick Mathewson
37255d24bc
Retry non-final-hop rendezvous failures
...
svn:r1625
2004-04-14 21:40:50 +00:00
Roger Dingledine
465f0e528a
call circuit_build_failed() even when the failure is immediate
...
svn:r1617
2004-04-14 04:32:49 +00:00
Roger Dingledine
1d7979b460
let introcircs receive acks. rend streams now deal correctly again.
...
(i think)
svn:r1614
2004-04-13 22:56:24 +00:00
Nick Mathewson
a2ece0fc08
Separate "start-establishing-introduction" from "upload descriptors"; only upload descriptors containing fully established intro points.
...
svn:r1607
2004-04-13 17:16:47 +00:00
Roger Dingledine
e0ce205a76
clean up circuit rebuilding some; fix some bugs we hadn't hit yet
...
svn:r1606
2004-04-13 05:20:52 +00:00
Nick Mathewson
8483bace65
Get the rest of intro-ack/nak handling in ... except for the tricky "pick a new intro and re-extend and re-try" part.
...
svn:r1601
2004-04-13 01:41:39 +00:00
Roger Dingledine
f4ff5a75c1
break circuit_build_needed_circs into its own func
...
svn:r1596
2004-04-12 22:47:12 +00:00
Roger Dingledine
48e0228f1e
expire old CIRCUIT_PURPOSE_C_REND_READY circs too
...
svn:r1585
2004-04-09 20:08:13 +00:00
Roger Dingledine
eb9f5a561e
circuit-level sendmes weren't working, because the exit
...
thought it was an exit.
i overhauled the edge_type mechanism. hopefully this will help.
svn:r1570
2004-04-08 09:41:28 +00:00
Nick Mathewson
2a01e3ad10
Launch intro services after launching connections; allow ORs to build circuits through as-yet-unconnectd ORs
...
svn:r1562
2004-04-08 03:53:39 +00:00
Nick Mathewson
8351847afd
Nothing is good enough for you, is it?
...
svn:r1556
2004-04-08 02:50:34 +00:00
Nick Mathewson
14e31fbd71
Use CIRCUIT_IS_ORIGIN in favor of boolean circ->cpath
...
svn:r1555
2004-04-08 02:24:06 +00:00
Roger Dingledine
ad2192c90f
only call circuit_build_failed when we're circ's origin
...
svn:r1554
2004-04-08 02:22:26 +00:00
Nick Mathewson
60772ec90c
Split out major functionality of circuit_get_best; rephrase a few booleans
...
svn:r1547
2004-04-08 01:55:21 +00:00
Roger Dingledine
93801ac40d
don't pick myself or my twins as hops when building circuits
...
svn:r1531
2004-04-07 21:12:54 +00:00
Nick Mathewson
2633d353d7
Carefully increment failure counts when appropriate
...
svn:r1530
2004-04-07 20:59:38 +00:00
Roger Dingledine
505797e998
expire open-but-not-yet-acked circs if they're too old
...
svn:r1526
2004-04-07 20:38:23 +00:00
Roger Dingledine
650d483bea
convince ourselves that circuit_failed doesn't have to be any smarter
...
svn:r1525
2004-04-07 20:23:59 +00:00
Nick Mathewson
3d60cd1d61
Delete trailing whitespace and expand tabs; remove unused aes decrypt code
...
svn:r1522
2004-04-07 19:57:40 +00:00
Roger Dingledine
3c683a9e7c
let alice reuse recent rendjoin circs
...
svn:r1519
2004-04-07 19:22:55 +00:00
Roger Dingledine
cf67520e95
when counting how many being-built circs will handle a stream,
...
don't count ones that are too old
svn:r1517
2004-04-07 19:14:33 +00:00
Roger Dingledine
95f0e36f08
alice just fetched bob's hidden webpage.
...
yay.
svn:r1512
2004-04-06 21:52:01 +00:00
Roger Dingledine
d473cf7ee9
set the circ windows on the middle hop of the rend circs
...
svn:r1511
2004-04-06 21:25:11 +00:00
Roger Dingledine
5eca39cdb4
checkpoint commit: rend closer to working, still not there yet
...
svn:r1506
2004-04-06 20:25:18 +00:00
Nick Mathewson
6290d027c9
Continue attack on magic numbers; use new crypto wrappers where possible
...
svn:r1504
2004-04-06 20:16:12 +00:00
Roger Dingledine
2bb18e62cb
make rend apconn send to the right cpath layer
...
and fix circuit_log_path to know about rend circs
svn:r1496
2004-04-05 22:01:35 +00:00
Nick Mathewson
b1a8b208ca
Make init_cpath_crypto able to handle both sides of handshake, by adding a "reverse" flag
...
svn:r1489
2004-04-05 20:53:04 +00:00
Roger Dingledine
7793078dff
alice can now look up bob's service descriptor,
...
choose an intro point, connect to it,
choose a rend point, connect to it and establish a cookie,
get an ack from the rendezvous point,
and know when both circs are ready for her.
APConns don't use conn->purpose anymore
don't initiate a renddesc lookup if one is already in progress
also fix a buffer overflow in nickname parsing (only exploitable
by the operator though)
svn:r1471
2004-04-05 00:47:48 +00:00
