Commit Graph

262 Commits

Author SHA1 Message Date
Nick Mathewson
58a6761056 r12330@catbus: nickm | 2007-04-09 19:15:42 -0400
Split type of "packed cell" from "parsed cell"; pack cells before queueing them on circuits.  This will help us avoid dumb errors when we confuse the two types.


svn:r9935
2007-04-09 23:15:46 +00:00
Nick Mathewson
2a07919198 r12326@catbus: nickm | 2007-04-09 17:36:41 -0400
Clean up crap from 9930.  I can hardly wait till I trust SVK 2.x enough to upgrade.


svn:r9933
2007-04-09 21:36:45 +00:00
Nick Mathewson
32506ad282 svn:r9930 2007-04-09 21:33:49 +00:00
Nick Mathewson
98ae4bbd0f Comment out the "clear queue when sending a destroy" logic: it seems potentially dangerous. (see comment for details)
svn:r9913
2007-03-29 02:32:00 +00:00
Nick Mathewson
a25c5d6994 r12654@Kushana: nickm | 2007-03-25 19:03:44 -0400
Add documentation for cell queue functions; make destroy cells result in cell queues getting cleared before the destroy gets sent.


svn:r9907
2007-03-26 14:08:35 +00:00
Nick Mathewson
d1381aef82 r12652@Kushana: nickm | 2007-03-25 15:01:48 -0400
A surprisingly simple patch to stop reading on edge connections when their circuits get too full, and start again when they empty out.  This lets us remove the logic to block begin_dir conns when the corresponding or conns get full: it was already broken by cell queues anyway.


svn:r9905
2007-03-26 14:08:18 +00:00
Nick Mathewson
38c0bb3a99 r12651@Kushana: nickm | 2007-03-24 18:26:42 -0400
Initial version of circuit-based cell queues.  Instead of hammering or_conns with piles of cells, queue cells on their corresponding circuits, and append them to the or_conn as needed.  This seems to work so far, but needs a bit more work.  This will break the memory-use-limitation patch for begin_dir conns: the solution will be a fun but fiddly.


svn:r9904
2007-03-26 14:07:59 +00:00
Roger Dingledine
0724bc113f correct a function comment that's become wrong
svn:r9879
2007-03-19 23:32:30 +00:00
Nick Mathewson
7fcceb2c25 r12074@catbus: nickm | 2007-03-04 15:11:43 -0500
Make all LD_BUG log messsages get prefixed with "Bug: ".  Remove manually-generated "Bug: "s from log-messages.  (Apparently, we remembered to add them about 40% of the time.)


svn:r9733
2007-03-04 20:11:46 +00:00
Nick Mathewson
5f9f62daa2 r12053@catbus: nickm | 2007-03-03 16:45:38 -0500
Remove some XXX012s: leave a check in connection_or_set_identity_digest it costs basically nothing to do on average.  Forget about reinstating the user check on NT Service creation.  Note that a notice message in main.c ("Is something wrong with your network connection?") is really useless.


svn:r9721
2007-03-03 21:55:31 +00:00
Roger Dingledine
33430d3a9e bugfix on r9568: we were throwing around an uninitialized
nickname array, and only treating a server as reachable if
the array magically has the right nickname in it. this was
causing the authorities to label only themselves as running,
and clients were "mysteriously" failing.


svn:r9628
2007-02-24 01:26:09 +00:00
Roger Dingledine
9946bb7fbd cosmetic, comment, and todo fixes
svn:r9627
2007-02-24 01:12:53 +00:00
Nick Mathewson
759c58151e r11775@catbus: nickm | 2007-02-12 16:39:09 -0500
Update copyright dates.


svn:r9570
2007-02-12 21:39:53 +00:00
Nick Mathewson
0c40a080a4 r11773@catbus: nickm | 2007-02-12 15:18:48 -0500
Implement proposal 106: stop requiring clients to have certificates, and stop checking for nicknames in certificates.  [See proposal 106 for rationale.]  Also improve messages when checking TLS handshake, to re-resolve bug 382.


svn:r9568
2007-02-12 21:39:33 +00:00
Nick Mathewson
85d3c2fbaf r11702@catbus: nickm | 2007-02-08 02:10:19 -0500
Wow. Remember when I wrote that script to tell me about macros that were never used?  Good times.


svn:r9525
2007-02-08 07:24:15 +00:00
Roger Dingledine
9aa8e490f8 cleanups based on looking through svn commits
svn:r9491
2007-02-06 00:27:03 +00:00
Nick Mathewson
03ef2156c9 r11637@catbus: nickm | 2007-02-05 12:41:51 -0500
Fix an XXXX012, and make circuits_pending_or_conns a static variable.  In addition to cleaning up the code, this may also resolve Bug 386 if Roger has the right intuition there.


svn:r9482
2007-02-05 17:42:40 +00:00
Nick Mathewson
fefba95363 r11629@catbus: nickm | 2007-02-02 15:06:17 -0500
Removing the last DOCDOC comment hurt so much that I had to use Doxygen to identify undocumented macros and comments, and add 150 more DOCDOCs to point out where they were.  Oops.  Hey, kids!  Fixing some of these could be your first Tor patch!


svn:r9477
2007-02-02 20:06:43 +00:00
Nick Mathewson
9984cad6e8 r11552@catbus: nickm | 2007-01-27 03:55:02 -0500
This one is a little tricky.  Our BEGIN_DIR implementation has a
 problem: the dirserv conns will decide they can flush all their data
 immediately, since the edge_conns will read greedily.
 
 For our 0.1.2 workaround, we track which or_conn a bridged dirserv
 conn is attached to, and stop writing when its outbuf is too full, and
 start writing again when the or_conn's outbuf empties out a little.
 
 This requires a bit of pointer management.  Let's hope it works.
 


svn:r9432
2007-01-27 08:55:06 +00:00
Nick Mathewson
ba53e0029f r11277@catbus: nickm | 2007-01-23 14:08:08 -0500
When logging TLS certificate and identity-related errors, scrub IP addresses for incoming connections, and distinguish between incoming and outgoing connections in log messages.  Backport candidate.


svn:r9391
2007-01-23 19:22:49 +00:00
Roger Dingledine
ab838bddb8 Flush local controller connection buffers periodically as we're
writing to them, so we avoid queueing 4+ megabytes of data before
trying to flush.

Also add a new XXX012.


svn:r9382
2007-01-22 06:07:51 +00:00
Nick Mathewson
23c18f1dc2 svn:r9357 2007-01-15 21:21:21 +00:00
Nick Mathewson
380f8983c7 r11966@Kushana: nickm | 2007-01-15 16:12:17 -0500
Tidy up ORCONN reason patch from Mike Perry.  Changes: make some of the handling of TLS error codes less error prone.  Enforce house style wrt spaces.  Make it compile with --enable-gcc-warnings.  Only set or_conn->tls_error in the case of an actual error.  Add a changelog entry.


svn:r9355
2007-01-15 21:21:05 +00:00
Nick Mathewson
ead35ef944 r11957@Kushana: nickm | 2007-01-15 15:25:57 -0500
Patch from Mike Perry: Track reasons for OR connection failure; display them in control events. Needs review and revision.


svn:r9354
2007-01-15 21:13:37 +00:00
Nick Mathewson
26c0d6ec58 r11743@Kushana: nickm | 2006-12-28 23:13:21 -0500
Refactor _connection_controller_force_write back into conncetion_handle_write.  Again, the line count goes down: groovy!


svn:r9208
2006-12-29 05:06:47 +00:00
Nick Mathewson
b59573949e r11736@Kushana: nickm | 2006-12-28 17:46:23 -0500
Mark a pile of XXXX comments as targetted for 0.1.2: some because they look important, some because they look easy.


svn:r9205
2006-12-29 02:47:51 +00:00
Roger Dingledine
b3b75a3bef write-limiting fix: avoid a rare bug where we don't remember
that there's stuff on the outbuf we want to flush.


svn:r9142
2006-12-16 11:36:36 +00:00
Roger Dingledine
0cc7eaf302 remove some dead comments, and resolve some XXXs
svn:r9141
2006-12-16 11:34:12 +00:00
Roger Dingledine
04409f202d reenable write limiting. nick finally convinced me this was
a smart move.

more todo sub-items remain for it.


svn:r9101
2006-12-13 07:08:36 +00:00
Nick Mathewson
1bf1f9d2fc r9449@Kushana: nickm | 2006-10-31 00:12:02 -0500
Dump breakdown of PK operations when we get a USR2 signal.  This should help us figure out of we are doing some of them for stupid reasons.


svn:r8881
2006-10-31 19:17:07 +00:00
Nick Mathewson
e618c15aff r8967@totoro: nickm | 2006-10-08 23:38:50 -0400
Fix some test and warn failures in last commit


svn:r8665
2006-10-09 03:39:06 +00:00
Nick Mathewson
ea00437f74 r8933@totoro: nickm | 2006-10-07 20:33:16 -0400
whitespace fixes; add changelog for last commit.


svn:r8644
2006-10-08 00:33:43 +00:00
Roger Dingledine
1cf37f90c4 bugfix and cleanups re: entry guards.
svn:r8620
2006-10-06 23:37:07 +00:00
Nick Mathewson
9988112c87 r8973@Kushana: nickm | 2006-09-28 16:53:19 -0400
Refactor entry guard status logic a lot; allow more factors [like not
 having a Guard flag or being listed in ExcludeNodes] to render a guard
 "unlisted" (now called "unusable"); track guard down status (now
 called "unreachable") separately from is_running.


svn:r8519
2006-09-28 23:57:44 +00:00
Nick Mathewson
ec82db6995 r6909@Kushana: nickm | 2006-07-26 13:05:58 -0400
Clean up wide lines from last patch.


svn:r6907
2006-07-26 19:07:37 +00:00
Nick Mathewson
4ff4577beb r6908@Kushana: nickm | 2006-07-26 12:38:52 -0400
Refactor connection_t into edge, or, dir, control, and base subtypes.  This might save some RAM on busy exit servers, but really matters most in terms of correctness.


svn:r6906
2006-07-26 19:07:26 +00:00
Roger Dingledine
dc79dd6a46 minor fixes
svn:r6709
2006-07-04 03:19:59 +00:00
Roger Dingledine
a20a384b09 harmless typo
svn:r6613
2006-06-13 05:50:24 +00:00
Roger Dingledine
ab61c40661 Fix the bug that was causing servers to not find themselves
reachable if they changed IP addresses. This happened because
middle servers knew the old descriptor, and kept swapping the
addr/port we asked for with the one they thought was right.

So the create cell never got sent, because it was asking for a
different addr/port than we believed we had connected to.


svn:r6610
2006-06-13 05:36:35 +00:00
Roger Dingledine
14687fc443 take out the reachability bandaid in 0.1.2.x as well.
maybe we will actually be able to fix it, instead.


svn:r6585
2006-06-10 00:57:12 +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
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
152d6d13b4 remove some unused code (i think)
svn:r6539
2006-06-05 08:02:04 +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
7f611f4732 if we're a server and some peer has a broken tls certificate, don't
shout about it unless we want to hear about protocol violations.


svn:r6507
2006-05-26 16:32:16 +00:00
Roger Dingledine
581795f41d auth dir servers were only modifying a server's is_running field
when they created a network status. so if nobody asked for a
network status, they would never discover that any servers are
is_running, so they could never build a circuit.


svn:r6183
2006-03-18 01:24:04 +00:00
Nick Mathewson
bd8ffccae7 More cleanups noticed by weasel; also, remove macros that nobody uses.
svn:r6143
2006-03-12 23:31:16 +00:00
Nick Mathewson
5777ee0e1a Add some functions to escape values from the network before sending them to the log. Use them everywhere except for routerinfo->plaftorm, routerinfo->contact_info, and rend*.c. (need sleep now)
svn:r6087
2006-03-05 09:50:26 +00:00