Nick Mathewson
0d946e1773
Merge remote-tracking branch 'arma/bug7029'
2012-10-14 23:06:43 -04:00
Roger Dingledine
38b7947386
fix trivial typo
2012-10-13 18:34:24 -04:00
Nick Mathewson
45d7fb44c4
typo in src/ext/README; caught by rransom (thanks!)
2012-10-12 19:48:26 -04:00
Nick Mathewson
e3c746384a
Fix whitespace in aes.c
2012-10-12 17:17:36 -04:00
Nick Mathewson
286b6df038
Simplify list of files that get whitespace-checked
...
The rule is simple: Everything not in src/ext!
2012-10-12 17:16:20 -04:00
Nick Mathewson
48b3ae8fe0
Move strlcpy and strlcat into src/ext too
2012-10-12 17:14:28 -04:00
Nick Mathewson
c3162d39bd
Distribute src/ext/README.
2012-10-12 17:00:35 -04:00
Nick Mathewson
f1cb14ae49
Add a README file for the src/ext directory.
2012-10-12 17:00:01 -04:00
Nick Mathewson
8586611718
Make tor_addr_is_internal log the calling function on error
...
This might make it a little easier to track down bug 7086.
2012-10-12 11:34:54 -04:00
Andrea Shepard
9c605ecb7e
Install correct incoming cell handlers on reachability testing channels
...
Fix for bug 7086.
2012-10-12 10:02:13 -04:00
Andrea Shepard
3a33b1fe3b
Merge branch 'move_contrib_source' of git://git.torproject.org/nickm/tor
2012-10-11 15:55:26 -07:00
Nick Mathewson
63f542a5c2
Move all externally maintained source files into src/ext
...
The rationale for treating these files differently is that we should
be checking upstream for changes as applicable, and merging changes
upstream as warranted.
2012-10-11 17:22:03 -04:00
Nick Mathewson
a45760b53b
Make very sure to handle cells in-order on channels.
...
Fix on code for 6465, not yet in any release.
2012-10-11 00:35:58 -04:00
Nick Mathewson
7ea904cbc0
Merge branch 'bug7011'
...
Conflicts:
src/or/circuitbuild.c
The conflict was trivial, since no line of code actually changed in
both branches: There was a fmt_addr() that turned into fmt_addrport()
in bug7011, and a "if (!n_conn)" that turned into "if (!n_chan)" in
master.
2012-10-10 22:31:06 -04:00
Nick Mathewson
e6d1ab3600
Fix changes/bug7011 slightly
2012-10-10 22:27:52 -04:00
David Fifield
6e31b9c503
Add changes file for bug7011.
2012-10-10 22:25:30 -04:00
David Fifield
8b3f3fb86e
Use fmt_addrport in pt_get_extra_info_descriptor_string.
...
This patch is by asn.
https://trac.torproject.org/projects/tor/ticket/7011#comment:11
2012-10-10 22:25:30 -04:00
David Fifield
8419d18441
Use fmt_addrport where appropriate.
...
This is mostly a conversion from this pattern:
log("... %s:%d ...", fmt_and_decorate_addr(&addr), port);
to this:
log("... %s ...", fmt_addrport(&addr, port));
The output is the same in all cases.
2012-10-10 22:25:30 -04:00
David Fifield
78e2d8c7a8
Add fmt_addrport.
...
This function formats an addr:port pair, and always decorates IPv6
addresses.
2012-10-10 22:25:30 -04:00
David Fifield
4b0a039cb8
Use fmt_and_decorate_addr in extra-info "transport" lines.
...
Apparently BridgeDB is already expecting transport lines to be formatted
thus; see https://trac.torproject.org/projects/tor/ticket/7011#comment:12 ff.
It may be that there are no extant IPv6 pluggable transport bridges yet,
so this didn't cause a problem.
2012-10-10 22:25:29 -04:00
David Fifield
5cbf0f2106
Use fmt_and_decorate_addr in TransportProxy statefile entry.
...
state_transport_line_is_valid calls tor_addr_port_lookup, which expects
brackets around an IPv6 address. Without this, cached transport
addresses can't be parsed later:
[warn] state: Could not parse addrport.
[warn] state: State file seems to be broken.
See #7011 .
2012-10-10 22:25:29 -04:00
David Fifield
41328c7009
Decorate addresses given to a HTTP CONNECT proxy.
...
This affects the Request-URI and the value of the Host header. RFC 2616
doesn't directly address the formatting of IPv6 addresses, but it
delegates some productions to RFC 2396 "Uniform Resource Identifiers
(URI): Generic Syntax," which is obsoleted by RFC 3986, which requires
square brackets for IPv6 addresses in both places.
I tested this with
HTTPSProxy 127.0.0.1:8000
Bridge <IPv6 bridge>
UseBridges 1
and an Ncat HTTP proxy:
ncat --proxy-type http -l 8000 -vvv
https://tools.ietf.org/html/rfc2616#section-3.2.1
https://tools.ietf.org/html/rfc2616#section-5.1.2
https://tools.ietf.org/html/rfc2616#section-14.23
https://tools.ietf.org/html/rfc3986#section-3.2.2
2012-10-10 22:25:29 -04:00
David Fifield
34c6ee7e9b
Use fmt_and_decorate_addr in log messages.
2012-10-10 22:25:29 -04:00
Andrea Shepard
8b36d4cc2a
Merge branch 'bug6816_squashed_nowarn' of git://git.torproject.org/nickm/tor
2012-10-10 18:53:38 -07:00
Nick Mathewson
bd28322d38
Remove variables; fix gcc 4.7 warnings
...
My GCC warns when variables are assigned to but never used. There
were a few like that in the 6816/6465 branches.
2012-10-10 21:25:52 -04:00
Andrea Shepard
cf994f7ad7
Add changes file for bug6816
2012-10-10 00:51:55 -07:00
Andrea Shepard
5543c5b202
Fix formatting in various places after 6465/6816 work
2012-10-10 00:48:36 -07:00
Andrea Shepard
217352c362
Make channel_flush_some_cells() compile cleanly on machines with ssize_t larger than int per sjumrdoch comment
2012-10-10 00:44:47 -07:00
Andrea Shepard
bec776480d
Don't remove circuitmux hash table entries in circuitmux_detach_circuit() until after circuitmux_make_circuit_inactive()
2012-10-10 00:44:47 -07:00
Andrea Shepard
1bc9a040f7
Fix 'warning: circuit was already inactive' and assert in circuitmux_make_circuit_inactive() during circuitmux_detach_all_circuits()
2012-10-10 00:44:47 -07:00
Andrea Shepard
c9e48ded5d
Bring summary comment block in circuitmux.c up to date
2012-10-10 00:44:47 -07:00
Andrea Shepard
49d534e524
New and improved circuitmux_detach_all_circuits(), now without the stupid
2012-10-10 00:44:47 -07:00
Andrea Shepard
c9607694c9
Correctly set magic numbers on ewma policy data/circuit data when allocating
2012-10-10 00:44:47 -07:00
Andrea Shepard
9d615cc5c0
Set circuitmux policy on existing active channels when ewma_enabled changes
2012-10-10 00:44:47 -07:00
Andrea Shepard
bb62281ba4
Set circuitmux policy on new channels in channeltls.c
2012-10-10 00:44:47 -07:00
Andrea Shepard
0c4f717b3e
Implement cell_ewma_enabled()
2012-10-10 00:44:47 -07:00
Andrea Shepard
8afe41b481
Implement channel_set_cmux_policy_everywhere()
2012-10-10 00:44:46 -07:00
Andrea Shepard
13972aee78
Fix broken circuitmux_move_active_circ_to_tail(); don't assume n_chan is not NULL in circuitmux_detach_circuit()
2012-10-10 00:44:46 -07:00
Andrea Shepard
903cc8acd1
Allow n_chan to be NULL in circuitmux_find_map_entry(); it can be but with non-NULL p_chan when extending a circuit
2012-10-10 00:44:46 -07:00
Andrea Shepard
a0200c1f6e
Use circuit_get_by_circid_channel_even_if_marked() and fix some asserts in circuitmux.c
2012-10-10 00:44:46 -07:00
Andrea Shepard
1498a6e84a
Fix some circuitmux-related asserts in relay.c
2012-10-10 00:44:46 -07:00
Andrea Shepard
b28119e6a8
Fix detach when setting circuit ID to 0 bug in circuit_set_circid_chan_helper() and add circuit_get_by_circid_channel_even_if_marked()
2012-10-10 00:44:46 -07:00
Andrea Shepard
c097fb33fc
Detach circuits from circuitmux early when freeing channel so they can find the channel by ID
2012-10-10 00:44:46 -07:00
Andrea Shepard
3d092ffbdd
Handle closing circuits correctly with circuitmux_t
2012-10-10 00:44:46 -07:00
Andrea Shepard
6830c9c232
Add a bunch of paranoid-mode expensive asserts incircuitmux.c
2012-10-10 00:44:46 -07:00
Andrea Shepard
7598e669e8
Adjust the circuitmux_t counter correctly in circuitmux_notify_xmit_cells()
2012-10-10 00:44:46 -07:00
Andrea Shepard
96a6eff8fe
Fix circuitmux attach/detach logic in circuit_set_circid_chan_helper(); it's possible for id to be zero (not assigned yet) and shouldn't be attached then
2012-10-10 00:44:46 -07:00
Andrea Shepard
60ec46a2ec
Update circuitmux-related TODOs in channeltls.c
2012-10-10 00:44:46 -07:00
Andrea Shepard
2565710fd7
Fix comment on circuit_t.n_hop and remove circuitmux TODO in or.h
2012-10-10 00:44:46 -07:00
Andrea Shepard
bbb2c31d26
Move all EWMA code into circuitmux.c and implement circuitmux_policy_t methods using it
2012-10-10 00:44:46 -07:00
