Commit Graph

4282 Commits

Author SHA1 Message Date
Nick Mathewson
273ee3e814 r7392@Kushana: nickm | 2006-08-14 23:50:32 -0400
Only do the expensive version of router_have_minimum_dir_info() when the dir info has changed.  Backport candidate, since oprofile suggests that this function and ones it calls account for 25-35% of oprofile samples.


svn:r7062
2006-08-15 03:54:09 +00:00
Nick Mathewson
bc21df6a5c r7386@Kushana: nickm | 2006-08-14 17:43:44 -0400
Patch from Adam Langley.
 
  * I meant getaddrinfo_a, not getaddrinfo_r - fixed
  * Added more checks to the parsing code.
  * It seems you switched an alloca to a malloc, but didn't add any frees
 


svn:r7061
2006-08-14 21:44:45 +00:00
Nick Mathewson
200e24981a r7046@Kushana: nickm | 2006-08-05 13:57:04 -0400
Make it possible for dns_init() to fail; note failure of eventdns configuratoin.


svn:r7059
2006-08-14 21:44:34 +00:00
Nick Mathewson
a5c0aa43a0 r7383@Kushana: nickm | 2006-08-14 16:16:03 -0400
Stop walking entire dns cache for every request, now that we found our hash table bug.


svn:r7057
2006-08-14 20:16:21 +00:00
Roger Dingledine
5194b91053 avoid complaining about our SOCKS proxy proxy.
svn:r7053
2006-08-14 10:00:15 +00:00
Roger Dingledine
4b94dabb53 note a compile warning that we should investigate one day.
svn:r7052
2006-08-14 09:44:54 +00:00
Roger Dingledine
9b10f11d39 fix typos and stuff
svn:r7049
2006-08-14 08:55:41 +00:00
Roger Dingledine
bf10d8a06a simplify compare_cached_resolves_by_expiry() to make it match
the idioms of other compare functions. hopefully i didn't
break it?


svn:r7048
2006-08-14 07:08:29 +00:00
Nick Mathewson
6da96f0ed7 r7369@Kushana: nickm | 2006-08-14 02:03:10 -0400
Shortening fields is only one part of making structs shorter. You must also consider alignment padding. Whee.


svn:r7045
2006-08-14 06:03:26 +00:00
Roger Dingledine
2c27cfc38e save 3-7 bytes per edge connection
svn:r7044
2006-08-14 05:53:57 +00:00
Roger Dingledine
6fdc63aed2 simplify the connection_write_to_buf_zlib() wrapper.
svn:r7042
2006-08-13 23:38:30 +00:00
Nick Mathewson
b68ffcc193 r7341@Kushana: nickm | 2006-08-11 00:51:05 -0700
Amazing how much difference adding a ! to all your asserts can make.


svn:r7029
2006-08-11 07:51:34 +00:00
Nick Mathewson
33fc829273 r7337@Kushana: nickm | 2006-08-11 00:42:04 -0700
Only use __builtin_offsetof with gcc 4 or later


svn:r7028
2006-08-11 07:42:11 +00:00
Roger Dingledine
f294575469 fix funny-looking assignment that crashes unit tests
svn:r7027
2006-08-11 07:41:21 +00:00
Roger Dingledine
5647e5c7e6 fix typo, add explanatory comment
svn:r7026
2006-08-11 07:31:16 +00:00
Nick Mathewson
42a5d7a4ba r7335@Kushana: nickm | 2006-08-11 00:13:03 -0700
fix wide lines


svn:r7025
2006-08-11 07:17:16 +00:00
Nick Mathewson
9b5a643475 r7328@Kushana: nickm | 2006-08-11 00:04:26 -0700
remove an assert that can never be false (an array in a non-null struct cannot be null).


svn:r7024
2006-08-11 07:09:52 +00:00
Nick Mathewson
04bf4b0fc4 r7327@Kushana: nickm | 2006-08-11 00:00:36 -0700
Fix a pedantic warning


svn:r7023
2006-08-11 07:09:45 +00:00
Nick Mathewson
ffab3b48a8 r7326@Kushana: nickm | 2006-08-10 23:50:49 -0700
And another GCC change: predict that tor_frees() are usually real frees, and tor_asserts() usually wont happen. Other test should wait till -fprofile-arcs


svn:r7022
2006-08-11 07:09:35 +00:00
Nick Mathewson
4eddbcf262 r7325@Kushana: nickm | 2006-08-10 23:37:31 -0700
Use gcc offsetof where available.


svn:r7021
2006-08-11 07:09:28 +00:00
Nick Mathewson
09a895e222 r7324@Kushana: nickm | 2006-08-10 23:23:15 -0700
Add more warnings to the list of those we tolerate. Start using GCC attributes more, for better error checking and better code generation.


svn:r7020
2006-08-11 07:09:17 +00:00
Nick Mathewson
02d42d9138 Fix crash in first-time option validation. Oops.
svn:r7018
2006-08-10 19:56:10 +00:00
Nick Mathewson
7ddd9e8cd9 r7304@Kushana: nickm | 2006-08-10 01:58:05 -0700
Fix verbose compilation errors; make sure transparent proxy fails when no method is configured.


svn:r7012
2006-08-10 09:02:26 +00:00
Nick Mathewson
3da737ac98 r7303@Kushana: nickm | 2006-08-10 01:52:19 -0700
whitespace fixes


svn:r7011
2006-08-10 09:02:12 +00:00
Nick Mathewson
74df271e00 r7302@Kushana: nickm | 2006-08-10 01:48:44 -0700
Warn about open TransListenAddress values.


svn:r7010
2006-08-10 09:02:02 +00:00
Nick Mathewson
cd560d8528 r7301@Kushana: nickm | 2006-08-10 01:41:27 -0700
Only open /dev/pf once.


svn:r7009
2006-08-10 09:01:54 +00:00
Nick Mathewson
d893d8c52e r7300@Kushana: nickm | 2006-08-10 01:36:40 -0700
Distinguish netfilter vs pf at configure time based on headers, not on OS.


svn:r7008
2006-08-10 09:01:46 +00:00
Nick Mathewson
5cff4164a0 r7299@Kushana: nickm | 2006-08-10 01:08:58 -0700
Patch from Tup to add support for transparent AP connections: this basically bundles the functionality of trans-proxy-tor into the tor mainline.  Now hosts with compliant pf/netfilter implementations can redirect TCP connections straight to Tor without diverting through SOCKS.


svn:r7007
2006-08-10 09:01:37 +00:00
Roger Dingledine
8075928b2a a way to make tor more stable in crummy situations
svn:r7005
2006-08-10 08:00:54 +00:00
Nick Mathewson
35f0881802 Experimentally re-enable kqueue on OSX when using libevent 1.1b or later. Log when we are doing this, so we can diagnose it when it fails.
svn:r7004
2006-08-10 08:00:13 +00:00
Nick Mathewson
f2a0df4d02 Recommend libevent 1.1b for kqueue and win32 methods; deprecate libevent 1.0b harder; make libevent recommendation system saner.
svn:r7003
2006-08-10 07:39:47 +00:00
Nick Mathewson
1b7ad5aed3 r7030@Kushana: nickm | 2006-08-04 14:46:52 -0700
Close an fd leak on failed mmap()


svn:r6988
2006-08-05 17:53:32 +00:00
Nick Mathewson
7c596c166b r7029@Kushana: nickm | 2006-08-04 14:08:41 -0700
Remove now-spurious size and data arguments from tor_mmap_file


svn:r6987
2006-08-05 17:53:21 +00:00
Nick Mathewson
dbac3fb481 r7028@Kushana: nickm | 2006-08-04 13:10:16 -0700
Make data and size fields visible in tor_mmap_t; hide win magic differently.


svn:r6986
2006-08-05 17:53:08 +00:00
Nick Mathewson
714d1b66aa r7027@Kushana: nickm | 2006-08-04 13:06:48 -0700
Oops. Fix downcast macro.


svn:r6985
2006-08-05 17:52:51 +00:00
Roger Dingledine
bc2e040669 make svn trunk link again. nick can fix this if it's wrong.
svn:r6982
2006-08-04 20:30:45 +00:00
Nick Mathewson
db72fb9cdc r7025@Kushana: nickm | 2006-08-04 12:03:22 -0700
Finish (I hope) windows mmap impl.


svn:r6981
2006-08-04 19:03:40 +00:00
Nick Mathewson
bf72878cad r7012@Kushana: nickm | 2006-08-03 19:21:25 -0700
Add an "mmap handle" type to encapsulate bookkeeping elements of mmap issues; add prelim win32 impl


svn:r6980
2006-08-04 18:32:43 +00:00
Nick Mathewson
abe27b807e svn:r6979 2006-08-04 18:31:13 +00:00
Nick Mathewson
2c33f385f7 r7011@Kushana: nickm | 2006-08-03 13:26:34 -0700
eventdns: Document functions added to API; make suspended requests go to the front of the queue; check (or explicitly ignore) return values on libevent functions.


svn:r6978
2006-08-04 18:27:10 +00:00
Nick Mathewson
e8f28a71e8 r7010@Kushana: nickm | 2006-08-03 10:47:36 -0700
Enable log message format checking in eventdns.c when __GNUC__ is defined.


svn:r6977
2006-08-04 18:26:40 +00:00
Nick Mathewson
3fa1cdafce r7009@Kushana: nickm | 2006-08-03 10:44:58 -0700
Add functions to eventdns to allow detecting whether we have any nameservers configured, and to change the list of nameservers after initial configuration.


svn:r6976
2006-08-04 18:26:13 +00:00
Nick Mathewson
87939813c3 r7008@Kushana: nickm | 2006-08-03 10:03:39 -0700
Oops. We shouldnt initialize eventdns when we are not being a server and not resolving anything.


svn:r6975
2006-08-04 18:24:41 +00:00
Nick Mathewson
a122bfb439 r7007@Kushana: nickm | 2006-08-03 09:58:30 -0700
Export and use eventdns_config_windows_nameservers(); clean up some comments and log messages.


svn:r6974
2006-08-04 18:24:25 +00:00
Nick Mathewson
19cf66c0f1 Patch from Frediano Ziglio: Windows compilation fixes on eventdns.c.
svn:r6971
2006-08-03 07:46:25 +00:00
Roger Dingledine
e64825126b turn future seg faults into asserts
svn:r6970
2006-08-03 04:23:45 +00:00
Roger Dingledine
3152d44c3a fix a seg fault on exit for clients; and fix a comment.
svn:r6969
2006-08-03 04:22:25 +00:00
Roger Dingledine
74bf8b1289 man, our sample torrc sucked.
svn:r6964
2006-08-01 04:08:15 +00:00
Roger Dingledine
ed567c6feb more compile options mean more codepaths
svn:r6963
2006-07-31 20:25:57 +00:00
Nick Mathewson
80282d8f55 r6993@Kushana: nickm | 2006-07-31 16:19:21 -0400
Interesting how much a ! can change the behavior of an assert.


svn:r6962
2006-07-31 20:19:58 +00:00
Nick Mathewson
b9082682fd r6982@Kushana: nickm | 2006-07-31 13:47:19 -0400
documentation and naming tweaks in dns.c


svn:r6961
2006-07-31 18:01:49 +00:00
Nick Mathewson
58c0ac77be r6981@Kushana: nickm | 2006-07-31 13:23:26 -0400
More asserts for cache correctness.


svn:r6960
2006-07-31 18:01:45 +00:00
Nick Mathewson
ea0f5ec8ad r6980@Kushana: nickm | 2006-07-31 13:18:22 -0400
Add a utility function to verify that a string has been through strlower.


svn:r6959
2006-07-31 18:01:37 +00:00
Nick Mathewson
d934607069 r6979@Kushana: nickm | 2006-07-31 13:16:58 -0400
Add assert_ok functions for strmap and digestmap; use them in unit test code.


svn:r6958
2006-07-31 18:01:27 +00:00
Nick Mathewson
3843b1b3d0 r6978@Kushana: nickm | 2006-07-31 13:16:14 -0400
Add isupper and islower wrappers to compat.h


svn:r6957
2006-07-31 18:01:22 +00:00
Nick Mathewson
fd0bafb0de r6977@Kushana: nickm | 2006-07-31 13:01:28 -0400
Solve timing-out pending connections. Add pending resolves to expiry queue; when we find an answer, change the pending resolve to "done" and stick the actual answer in the expiry queue as a new entry.  This uses a little more memory, but makes the code simpler than other solutions.


svn:r6956
2006-07-31 18:01:18 +00:00
Nick Mathewson
ccf9c1345e svn:r6955 2006-07-31 18:00:47 +00:00
Nick Mathewson
46574319a2 r6959@Kushana: nickm | 2006-07-29 22:33:18 -0400
start restructuring dns to use priority queues for expiring entries.


svn:r6954
2006-07-31 18:00:18 +00:00
Nick Mathewson
2fe537c57a r6958@Kushana: nickm | 2006-07-29 18:54:15 -0400
Looks like we might need a priority queue.


svn:r6953
2006-07-31 17:59:37 +00:00
Nick Mathewson
8ba913c660 r6952@Kushana: nickm | 2006-07-28 11:09:37 -0400
Add completely untested find-my-nameservers code for win32.


svn:r6952
2006-07-31 17:59:11 +00:00
Roger Dingledine
798b580695 fix assert found by DreadWingKnight: now that rendezvous streams
are attached to p_streams, the p_streams list can consist of both AP
and EXIT conns.


svn:r6945
2006-07-30 04:54:13 +00:00
Roger Dingledine
2bcb081cb1 defense in depth
svn:r6939
2006-07-30 04:32:58 +00:00
Roger Dingledine
c3b79ebf13 a bit more debugging for phobos
svn:r6937
2006-07-30 03:34:44 +00:00
Nick Mathewson
84c2a6c16b r6950@Kushana: nickm | 2006-07-28 10:32:08 -0400
Document split fields better. Now, I think we can take a break from type splitting for a bit.


svn:r6935
2006-07-28 15:11:28 +00:00
Nick Mathewson
85f381153b r6949@Kushana: nickm | 2006-07-28 10:17:38 -0400
Shave another 8 bytes from connection_t: turn inbuf_reached_eof into a bit, and lower timestamp_lastempty to or_connection_t


svn:r6934
2006-07-28 15:11:20 +00:00
Nick Mathewson
6dc13cdbeb r6948@Kushana: nickm | 2006-07-28 10:10:35 -0400
Identify some likely target fields for lowering; lower global_identifier (since we only use it for AP streams and origin circs).


svn:r6933
2006-07-28 15:11:11 +00:00
Roger Dingledine
a5e5074807 resolve typo
svn:r6927
2006-07-27 19:35:11 +00:00
Nick Mathewson
c2f28615ed More asserts in dns.c
svn:r6926
2006-07-27 18:35:56 +00:00
Nick Mathewson
8786b4b88e Fix a bug in HT_REMOVE.
svn:r6925
2006-07-27 18:35:25 +00:00
Nick Mathewson
45c3a4365f Get better numbers out of HT_REP_OK
svn:r6924
2006-07-27 17:37:37 +00:00
Nick Mathewson
dc2b7e9af0 Add more asserts in dns_found_answer. This may confirm my theory that dns_purge_resolve is the culprit.
svn:r6922
2006-07-27 17:16:10 +00:00
Nick Mathewson
5227395aba r6922@Kushana: nickm | 2006-07-26 16:32:24 -0400
Rename some fields, compress a bitfield, and document some structs and fields


svn:r6919
2006-07-27 05:03:57 +00:00
Nick Mathewson
dd779d99ea Fix comments that implied that only dir connections had a purpose field, and the code that believed in those comments.
svn:r6918
2006-07-27 04:10:51 +00:00
Nick Mathewson
0bc741ba7f r6918@Kushana: nickm | 2006-07-26 15:22:28 -0400
Fix compilation for eventdns dns.c with split structs.


svn:r6908
2006-07-26 19:29:30 +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
Nick Mathewson
18771e851f r6907@Kushana: nickm | 2006-07-25 19:03:43 -0400
Realign circuit structs to avoid wasted space.


svn:r6905
2006-07-26 19:07:23 +00:00
Nick Mathewson
e61d28dfb8 r6903@Kushana: nickm | 2006-07-25 18:22:48 -0400
No circuit can be both an intro point and a rend point, so we can merge both the cookie and the pk digest into one "rend_token" field for or circuits. This saves another 20 bytes per or circuit.


svn:r6904
2006-07-26 19:05:41 +00:00
Nick Mathewson
a88ec48a39 r6902@Kushana: nickm | 2006-07-25 17:30:27 -0400
Move rend_query to origin_circuit_t where it belongs; save another 17 bytes per OR circuit.


svn:r6903
2006-07-26 19:05:34 +00:00
Nick Mathewson
96bcbb1e1c Remove code to special-case "-cvs" ending, since it has not actually mattered since 0.0.9. Perhaps we can special-case even more...
svn:r6898
2006-07-25 04:34:14 +00:00
Peter Palfrader
521c006fd5 Remove .cvsignore files from trunk
svn:r6891
2006-07-25 01:00:48 +00:00
Roger Dingledine
c9b8458f59 publish a new descriptor after we hup. this is important if our
config has changed such that we'll want to start advertising our
dirport now, etc.


svn:r6819
2006-07-23 11:50:03 +00:00
Nick Mathewson
d27cbf5d4b Fix another segfault in assert_circuit_ok.
svn:r6818
2006-07-23 08:13:45 +00:00
Nick Mathewson
7239262f71 Don't tell anybody, but we're going OO here. This patch splits
circuit_t into origin_circuit_t and or_circuit_t.  I fixed some
segaults; there may be more.  We still need to move more rendezvous
stuff into subtypes.

This is a trial run for splitting up connection_t; if the approach is
insane, please say so soon so we can do something smarter.

Also, this discards the old HALF_OPEN code, which nobody seems to
want.


svn:r6817
2006-07-23 07:37:35 +00:00
Nick Mathewson
6d2eb77555 Whitespace fix
svn:r6816
2006-07-23 07:19:49 +00:00
Roger Dingledine
40cc536307 rewrite conn->address for GET commands as well as POST commands.
svn:r6815
2006-07-23 06:41:02 +00:00
Roger Dingledine
0b5ad1bcfb use tor_mem_is_zero() in more places.
svn:r6814
2006-07-23 05:52:27 +00:00
Nick Mathewson
4280218a42 Delete trailing whitespace in eventdns.c
svn:r6813
2006-07-23 05:40:24 +00:00
Nick Mathewson
1c0f02f657 Add (void) lines for unused parameters in eventdns.c
svn:r6812
2006-07-23 05:39:37 +00:00
Nick Mathewson
f4e506f423 Remove STRUCT_OFFSET from config.c
svn:r6811
2006-07-23 05:33:10 +00:00
Nick Mathewson
35960e1162 Add a mem_is_zero function (I think we will need this) and a STRUCT_OFFSET macro (we already need this).
svn:r6810
2006-07-23 05:32:35 +00:00
Roger Dingledine
ee789052b8 whitespace and docs
svn:r6808
2006-07-23 05:18:29 +00:00
Roger Dingledine
00aefaab6b think harder about my logic
svn:r6806
2006-07-22 07:19:11 +00:00
Roger Dingledine
2d6a4d283b i lied, that won't work at all. maybe this will.
svn:r6805
2006-07-22 07:15:34 +00:00
Roger Dingledine
444f096d2d more bulletproof reachability testing
svn:r6804
2006-07-22 05:29:31 +00:00
Nick Mathewson
64c14d8ce3 These asserts will either cause spurious crashes or help debug the pend->conn->s == -1 issue.
svn:r6803
2006-07-21 22:02:58 +00:00
Nick Mathewson
08a4114abf Another MSVC6 fix. Grnk.
svn:r6802
2006-07-21 14:53:23 +00:00
Roger Dingledine
2249180b60 bugfix: if you find yourself reachable, then don't ever make any client
requests (so you stop predicting circuits), then hup, then later your
IP changes, you won't think circuits are working so you won't try to
test reachability, so you won't publish.


svn:r6800
2006-07-21 07:55:35 +00:00
Roger Dingledine
4d5fab4f9a tweak
svn:r6799
2006-07-21 07:53:21 +00:00
Roger Dingledine
9283ddb760 an entry guard that is "unlisted", as well as not known to be "down", is
not therefore "up".


svn:r6798
2006-07-21 07:06:18 +00:00