aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2008-02-20if you have more than 40k FDs on your system let Tor use up to 16k of them ↵Peter Palfrader
by default svn:r13627
2008-02-20 r18269@catbus: nickm | 2008-02-20 17:28:24 -0500Nick Mathewson
Apply patch from Sebastian Hahn: stop imposing an arbitrary maximum on the number of file descriptors used for busy servers. Bug reported by Olaf Selke. svn:r13626
2008-02-20Change the default for MAX_FILEDESCRIPTORS in our init script to depend on thePeter Palfrader
number of system-wide available file descriptors: /proc/sys/fs/file-max is bigger than 100k, set ulimit -n to 32k, if it's smaller than 20k set it to 1024. Big servers at the moment regularly use more than 10k FDs, so our old default of 8k no longer is sufficient. On the other hand we don't want lower end systems to run out of FDs on Tor's account. svn:r13625
2008-02-20 r18267@catbus: nickm | 2008-02-20 16:57:45 -0500Nick Mathewson
Suppress signed/unsigned comparison warnings in buffers.c svn:r13624
2008-02-20 r18264@catbus: nickm | 2008-02-20 12:48:21 -0500Nick Mathewson
fix bufs in buf_pos_t implementation. svn:r13623
2008-02-20 r18263@catbus: nickm | 2008-02-20 12:45:15 -0500Nick Mathewson
Oops. not supposed to use tor_assert() in eventdns.c svn:r13622
2008-02-20One last tweak on debugging code.Nick Mathewson
svn:r13621
2008-02-20 r18259@catbus: nickm | 2008-02-20 12:03:15 -0500Nick Mathewson
Make the last of the veracode warning issues get fixed. svn:r13620
2008-02-20 r18256@catbus: nickm | 2008-02-20 11:57:31 -0500Nick Mathewson
Simplify rounding logic in bitarray; fix a bug in bitarray_expand(). svn:r13619
2008-02-20 r18255@catbus: nickm | 2008-02-20 11:44:55 -0500Nick Mathewson
Add asserts and refactor some comparisons in order to fix some veracode-identified issues. Note a bug in buffers.c svn:r13618
2008-02-20pass bucket sizes num_read and num_written around as size_tsRoger Dingledine
rather than ints svn:r13610
2008-02-20We were comparing the raw BridgePassword entry with a base64'edRoger Dingledine
version of it, when handling a "/tor/networkstatus-bridges" directory request. Now compare correctly. Noticed by Veracode. svn:r13609
2008-02-20similarly, don't throw around an int for the uint8_t circ->stateRoger Dingledine
svn:r13608
2008-02-20don't throw around an int for relay_command if we're going toRoger Dingledine
squeeze it into a uint8_t later. also reported by veracode. svn:r13607
2008-02-20notice an integer overflow if we ever get source files with moreRoger Dingledine
than 65k lines in them. also pointed out by veracode. svn:r13606
2008-02-20if you thought r13604 was a kludge, how about this.Roger Dingledine
svn:r13605
2008-02-20be more clear about a possible integer underflow problem. thisRoger Dingledine
should probably be fixed better somehow. svn:r13604
2008-02-19check return value for crypto_pk_asn1_encode here too. thanks veracode.Roger Dingledine
svn:r13603
2008-02-19 r18236@catbus: nickm | 2008-02-19 18:55:21 -0500Nick Mathewson
Explain why I am right and veracode is wrong in routerparse.c line 1141. Using math! svn:r13602
2008-02-19rearrange our assert so we don't possibly overflow rh.length beforeRoger Dingledine
triggering the assert. reported by veracode. svn:r13601
2008-02-19 r18233@catbus: nickm | 2008-02-19 18:46:07 -0500Nick Mathewson
Count sockets returned from socketpair() too. This is probably not the socket counting bug. svn:r13600
2008-02-19start checking the return value of crypto_pk_asn1_encode() whileRoger Dingledine
we're trying to establish an introduction point. bug reported by veracode. svn:r13599
2008-02-19 r18230@catbus: nickm | 2008-02-19 18:29:43 -0500Nick Mathewson
Add a few asserts to catch possible errors found by veracode. svn:r13598
2008-02-19is_reverse in our struct is uint8_t. don't make it be an intRoger Dingledine
elsewhere. bug reported by veracode. svn:r13597
2008-02-19 r18228@catbus: nickm | 2008-02-19 18:05:53 -0500Nick Mathewson
Chris Palmer notes that almost nobody is using smartlist_set_capacity(). Chris Palmer is right. Remove this basically pointless function. svn:r13596
2008-02-19 r18226@catbus: nickm | 2008-02-19 18:01:01 -0500Nick Mathewson
Brown-paper-bag time. We were failing to count all the sockets from accept(). svn:r13595
2008-02-19call a "number of characters" a size_t, not an intRoger Dingledine
svn:r13594
2008-02-19 r18221@catbus: nickm | 2008-02-19 17:46:16 -0500Nick Mathewson
New debugging code to figure out what is happending with socket counts. svn:r13593
2008-02-19be more explicit about a harmless thing that freaked veracode outRoger Dingledine
svn:r13592
2008-02-19 r18218@catbus: nickm | 2008-02-19 17:27:40 -0500Nick Mathewson
When SafeLogging is off, have TLS errors and messages logged with their associated addresses. svn:r13591
2008-02-19make explicit that we don't care if a rename() call fails.Roger Dingledine
potential bug reported by veracode. svn:r13590
2008-02-19 r18208@catbus: nickm | 2008-02-19 17:02:30 -0500Nick Mathewson
Add some checks in torgzip.c to make sure we never overflow size_t there. Also make sure we do not realloc(list,0) in container.c. Backport candidate. svn:r13587
2008-02-19resolve another edge case in staying dormantRoger Dingledine
svn:r13586
2008-02-19maybe it should compile too.Roger Dingledine
svn:r13585
2008-02-19If we only ever used Tor for hidden service lookups or posts, weRoger Dingledine
would stop building circuits and start refusing connections after 24 hours, since we false believed that Tor was dormant. Reported by nwf; bugfix on 0.1.2.x. svn:r13583
2008-02-19clear up another connlimit lieRoger Dingledine
svn:r13582
2008-02-19clean up the socket counting thing. third time's a charm.Roger Dingledine
svn:r13581
2008-02-19 r18198@catbus: nickm | 2008-02-19 14:30:30 -0500Nick Mathewson
Try to *fix* the socket counting problem, and add an info log to detect whether we really fixed it svn:r13580
2008-02-19 r18195@catbus: nickm | 2008-02-19 14:11:15 -0500Nick Mathewson
TOR_PERF has not not been a sane thing for ages. Remove it. svn:r13579
2008-02-19hunt for killerchicken's socket counting problemRoger Dingledine
svn:r13578
2008-02-19document that if you enable TOR_PERF, you won't be compatible with theRoger Dingledine
main Tor network. svn:r13577
2008-02-19 r14246@tombo: nickm | 2008-02-17 19:33:02 -0500Nick Mathewson
Fix a wide line in control.c; make the check-spaces target ignore the openbsd malloc file. svn:r13566
2008-02-18 r18142@catbus: nickm | 2008-02-18 14:24:31 -0500Nick Mathewson
Our preferred spelling for a bitfield is "unsigned int", not "unsigned" or "uint8_t". svn:r13560
2008-02-18 r18139@catbus: nickm | 2008-02-18 13:14:05 -0500Nick Mathewson
Clarify logic in trusted_dirs_load_certs_from_string(); avoid a maybe-impossible maybe-not double-free spotted by lodger. svn:r13558
2008-02-18 r18138@catbus: nickm | 2008-02-18 13:13:18 -0500Nick Mathewson
Try to make conditional include logic for openbsd malloc not warn on arma's computer. May fix bug 610. svn:r13557
2008-02-17update thoughts on documentation license.Andrew Lewman
svn:r13549
2008-02-17 r14243@tombo: nickm | 2008-02-17 14:51:26 -0500Nick Mathewson
Better debugging to track down bug 606. Also remove a break; that may have been the possible culprit svn:r13547
2008-02-17 r14240@tombo: nickm | 2008-02-17 14:26:52 -0500Nick Mathewson
Add a --with-tcmalloc argument to configure. svn:r13546
2008-02-17 r18123@catbus: nickm | 2008-02-17 14:08:45 -0500Nick Mathewson
Fix capitalization error svn:r13545
2008-02-17 r14238@tombo: nickm | 2008-02-17 14:03:44 -0500Nick Mathewson
Add openbsd memory allocator discussed in bug 468, to make it easier for linux users to get non-awful allocation patterns. Use --enable-openbsd-malloc to turn it on. Needs more testing. svn:r13544