Commit Graph

6246 Commits

Author SHA1 Message Date
Roger Dingledine
85d408a573 Bandaid for a seg fault i just got in 0.1.1.20.
More generally, i reopened bug 222. Whee.


svn:r6571
2006-06-09 06:35:45 +00:00
Roger Dingledine
a91be49402 fix a bootstrapping check we ignored that prevents us from running
with only one dir authority.


svn:r6570
2006-06-09 02:45:39 +00:00
Roger Dingledine
11dcb7f4ae try a better string at the top of torrc's autogenerated torrc.
svn:r6569
2006-06-09 02:20:42 +00:00
Roger Dingledine
bb18168fae ok, ok, maybe *this* time my rep_hist_circbuilding_dormant() will work.
svn:r6568
2006-06-08 22:36:13 +00:00
Roger Dingledine
18eb8192db remove a bit more obsolete code
svn:r6566
2006-06-08 09:35:20 +00:00
Roger Dingledine
a18766e50c ah, that explains why we weren't going dormant with
respect to descriptor fetches. maybe now it will work.


svn:r6565
2006-06-08 09:20:58 +00:00
Roger Dingledine
91bd12c20d re-enable per-connection rate limiting. get rid of the "OP bandwidth"
concept. lay groundwork for "bandwidth classes" -- separate global
buckets that apply depending on what sort of conn it is.


svn:r6563
2006-06-07 09:18:53 +00:00
Roger Dingledine
741b11df45 put a bandaid in place so servers will have an easier time
believing that they're reachable. this may help resolve the
servers-on-dynamic-ip-addresses problem.


svn:r6561
2006-06-07 08:42:24 +00:00
Roger Dingledine
1e9b35aee3 make connection_or_nonopen_was_started_here() based on something less
voodooey. it turns out we already do keep a flag like that around.


svn:r6560
2006-06-07 07:11:42 +00:00
Roger Dingledine
003561fa64 the CookieAuthentication section in our spec seems to assume we're
still using the v0 control protocol.


svn:r6559
2006-06-07 06:53:43 +00:00
Roger Dingledine
8cf45df230 and now the exciting part: there is now no such thing as doing
a client-only tls, that is, one with no certs.


svn:r6558
2006-06-07 06:21:11 +00:00
Roger Dingledine
0bfef523df simplify the tortls api: we only support being a "server", that
is, even tor clients do the same sort of handshake.

this has been true for years, so it's best to get rid of the
stale code.


svn:r6557
2006-06-07 06:10:54 +00:00
Roger Dingledine
7512be0b65 looks like we missed a piece of the 0.1.1.9 paranoia code.
hopefully this change is a no-op.


svn:r6556
2006-06-07 02:57:23 +00:00
Roger Dingledine
96914760bc fix the bug where we sometimes would fail to send some create cells
once we'd connected to a(nother) tor server.


svn:r6552
2006-06-06 03:33:24 +00:00
Nick Mathewson
5e4b4451ec Also, add a temporary hack to make sure eventdns.c is distributd.
svn:r6551
2006-06-06 00:12:22 +00:00
Roger Dingledine
5c57348acd We got an obscure report of an assert error on a windows Tor server with
connection_add being called with socket = -1. The only places I can see
where that happen would be if our tor_socketpair succeeds but it hands
back negative values for some of its file descriptors. Perhaps this will
produce more useful assert errors next time.


svn:r6550
2006-06-06 00:06:52 +00:00
Roger Dingledine
411f37d534 fix typo
svn:r6549
2006-06-06 00:05:39 +00:00
Roger Dingledine
07c35b4f52 ship the event*.h files too. perhaps this will make my 'make dist'
produce a tarball that i can build.


svn:r6548
2006-06-06 00:04:52 +00:00
Nick Mathewson
2bd7c19b73 Try to log useful messages at info and debug about what we are resolving and what answers we are getting wrt eventdns.
svn:r6547
2006-06-05 23:01:22 +00:00
Roger Dingledine
542b8af592 don't tell people that the testing circuit failed if we already
consider ourselves reachable. this just confuses them.


svn:r6546
2006-06-05 10:01:52 +00:00
Roger Dingledine
8705db6c2c whoops, add a man page entry for ProtocolWarnings
svn:r6545
2006-06-05 09:51:29 +00:00
Roger Dingledine
a2fec314b6 scream louder if you've got a pending circuit for a given
addr/port but the intended n_conn digest is wrong.


svn:r6544
2006-06-05 09:47:19 +00:00
Roger Dingledine
45065f1466 simplify code now that libevent considers all sockets pollable.
what we really mean now is ">= 0", which is clearer to test for.


svn:r6543
2006-06-05 09:08:10 +00:00
Roger Dingledine
4e773352c2 bugfix: if we are making our first ever connection to any entry
guard, then don't mark it down at first. we had this implemented
but it was disabled due to a bug.


svn:r6541
2006-06-05 08:58:18 +00:00
Roger Dingledine
8f3bb66ebc simplify some code, since circuit_build_failed() is only called
on non-open circuits.


svn:r6540
2006-06-05 08:25:02 +00:00
Roger Dingledine
152d6d13b4 remove some unused code (i think)
svn:r6539
2006-06-05 08:02:04 +00:00
Roger Dingledine
bf09898f79 Note a bug that causes servers to sometimes never send the pending
create cell. Nick, is this a bug? If so, is my fix right?


svn:r6538
2006-06-05 07:27:48 +00:00
Roger Dingledine
266afcd7d5 bandaid for bug 299. this is still a bug, since we don't
initialize for hardware acceleration in certain configurations;
but not critical until that is supported.


svn:r6536
2006-06-05 04:29:03 +00:00
Nick Mathewson
b7ca697faf More eventdns.c patches: use HAVE_ALLOCA_H; print IP addrs as dotted quads.
svn:r6535
2006-06-05 01:59:12 +00:00
Nick Mathewson
3db785ef73 Hm. Where did we put that ntohl the last time we were juggling it? (hoop-lah).
svn:r6534
2006-06-05 00:32:31 +00:00
Nick Mathewson
008a0c7c46 Some eventdns.c fixes for windows correctness. More will doubtless be needed, especially around the #includes.
svn:r6533
2006-06-04 23:23:53 +00:00
Nick Mathewson
853e2d99b6 Add a new warning to our "warn a lot" list: unused parameters. This means we have to explicitly "use" unuseds, but it can catch bugs. (It caught two coding mistakes so far.)
svn:r6532
2006-06-04 22:42:13 +00:00
Roger Dingledine
c4647545a0 if we insist on printing pointer values, at least make it stop
complaining on (my particular) 64 bit platform.


svn:r6530
2006-06-04 06:17:32 +00:00
Roger Dingledine
603b09b59b punctuation and spelling
svn:r6529
2006-06-04 06:16:20 +00:00
Nick Mathewson
c18fab5366 Make eventdns.[ch] into good C90; remove signed/unsigned comparisons.
svn:r6528
2006-06-03 22:05:23 +00:00
Nick Mathewson
2700e5a1ab Oops. When we dont get a TTL, we should default to the default, not to the minimum.
svn:r6527
2006-06-03 21:47:26 +00:00
Nick Mathewson
c660a0f6a2 More DNS fixes. Send meaningful TTLs back to the client when
possible.  Cache at the server side independently from the TTL, to
prevent attackers from probing the server to see who has been asking
for what hostnames.  (Hi, Dan Kaminski!)

Also, clean some whitespace.


svn:r6526
2006-06-03 21:41:14 +00:00
Nick Mathewson
85a25c8a29 Make dns.c use eventdns.c -- but only when you pass the --enable-eventdns argument to configure.in. This will *so* not work on Windows yet.
svn:r6525
2006-06-03 20:52:24 +00:00
Nick Mathewson
7cb3232ade Add async dns code from Adam Langley, tweaked to build on OSX. Long-term, we may want to switch to libevnet/c-ares, if they ever handle 10k fd situations properly. This one still needs work too, but at least it is small. This code is disabled by default, and not integrated with dns.c.
svn:r6524
2006-06-03 19:49:42 +00:00
Nick Mathewson
86da3e0a0a Patch based on post by Mike C to or-dev; special-case based on use of MSVC, rather than on MS_WINDOWS, so that mingw builds.
svn:r6523
2006-06-03 18:52:31 +00:00
Roger Dingledine
8f902e4b58 don't stop fetching server descriptors if we're a server and
haven't found ourselves reachable yet.


svn:r6522
2006-06-03 02:56:44 +00:00
Roger Dingledine
2431b2d8b0 update the explanation for deprecating v0 control spec.
svn:r6521
2006-06-01 08:43:56 +00:00
Nick Mathewson
bbe3f587aa Rearrange TODO.
svn:r6520
2006-05-30 20:41:22 +00:00
Roger Dingledine
4d76b4351e mark off a todo item. i'll put it back if it turns out it
doesn't work.


svn:r6519
2006-05-30 06:36:32 +00:00
Roger Dingledine
1bf0f99b66 connection_t kept the identity_pkey but all it did was store it and
free it. perhaps we don't need it after all?


svn:r6518
2006-05-30 06:23:44 +00:00
Roger Dingledine
9000a5aff1 and get grammar right
svn:r6517
2006-05-30 06:19:48 +00:00
Roger Dingledine
ae2648fb1c stop fetching descriptors if we're not a dir mirror and we
haven't tried to establish any circuits lately.


svn:r6516
2006-05-30 06:19:06 +00:00
Roger Dingledine
1f0e20806b END_CIRC_REASON_OR_IDENTITY apparently means that we were told to
connect to a different OR than lives on the addr:port we connected
to. we don't actually remember whether that was the case, currently.
so call it END_CIRC_REASON_OR_CONN_CLOSED as a compromise.


svn:r6515
2006-05-30 06:17:28 +00:00
Roger Dingledine
64b1f2fc75 and clarify the spec to say this too.
svn:r6514
2006-05-30 06:11:46 +00:00
Roger Dingledine
60abe23158 tentative change: if you have your dirport set, you are a directory
mirror, whether or not your orport is set.


svn:r6513
2006-05-30 06:11:36 +00:00