Refactor tor to support the new tor-fw-helper protocol.
Add handle_fw_helper_output(), a function responsible for parsing the
output of tor-fw-helper. Refactor tor_check_port_forwarding() and
run_scheduled_events() accordingly too.
We now issue warnings when we get control output from tor-fw-helper,
and we log the verbose output of tor-fw-helper in LOG_INFO.
get_lines_from_handle() is a multiplatform function which drains lines
from a stream and stuffs it into a smartlist. It's useful for
line-based protocols, like the one managed proxy and the tor-fw-helper
protocols.
Linus Nordberg [Fri, 31 Aug 2012 21:02:19 +0000 (23:02 +0200)]
Use preferred OR for nodes with routerstatus and microdesc too.
extend_info_from_node() used to use the primary OR port (i.e. IPv4)
unless the node had routerinfo. Now that we have IPv6 addresses in
microdescs we may want to use them.
Note that this patch changes using r->cache_info.identity_digest into
using node->identity. I count on these being well synchronised, or
things would break in other ways. Right?
Linus Nordberg [Mon, 27 Aug 2012 13:03:34 +0000 (15:03 +0200)]
Clients connect to public relays over IPv6.
Add ClientUseIPv6 and ClientPreferIPv6ORPort configuration options.
Use "preferred OR port" for all entry nodes, not only for bridges.
Mark bridges with "prefer IPv6 OR port" if an IPv6 address is
configured in Bridge line and ClientPreferIPv6ORPort is set.
Mark relays with "prefer IPv6 OR port" if an IPv6 address is found in
descriptor and ClientPreferIPv6ORPort is set.
Filter "preferred OR port" through the ClientUseIPv6 config option. We
might want to move this test to where actual connection is being set
up once we have a fall back mechanism in place.
Have only non-servers pick an IPv6 address for the first hop: We
don't want relays to connect over IPv6 yet. (IPv6 has never been used
for second or third hops.)
Linus Nordberg [Fri, 31 Aug 2012 21:02:19 +0000 (23:02 +0200)]
Generate microdescriptors with "a" lines.
Generate and store all supported microdescriptor formats. Generate
votes with one "m" line for each format. Only "m" lines with version
info matching chosen consensus method will be voted upon.
An optimisation would be to combine "m" lines with identical hashes,
i.e. instead of "m 1,2,3 H1" and "m 4,5 H1", say "m 1,2,3,4,5 H1".
Linus Nordberg [Tue, 14 Aug 2012 12:03:58 +0000 (14:03 +0200)]
Have directory authorities vote on IPv6 OR ports according to the spec
Define new new consensus method 14 adding "a" lines to vote and
consensus documents.
From proposal 186:
As with other data in the vote derived from the descriptor, the
consensus will include whichever set of "a" lines are given by the
most authorities who voted for the descriptor digest that will be
used for the router.
Allow one-hop directory fetching circuits the full "circuit build timeout"
period, rather than just half of it, before failing them and marking
the relay down. This fix should help reduce cases where clients declare
relays (or worse, bridges) unreachable because the TLS handshake takes
a few seconds to complete.
Fixes bug 6743 (one piece of bug 3443); bugfix on 0.2.2.2-alpha, where
we changed the timeout from a static 30 seconds.
We are doing this merge to avoid taking the patch from
"nickm/disable_pathbias_warnings_v2" into 0.2.4 -- we're going to try
to fix the pathbias code for real here.
Nick Mathewson [Mon, 27 Aug 2012 14:46:17 +0000 (10:46 -0400)]
Downgrade path-bias warning messages to INFO for now.
We've had over two months to fix them, and didn't. Now we need
0.2.3.x stable. Yes, it would be cool to get this working in
0.2.3.x, but not at the expense of delaying every other feature that
_does_ work in 0.2.3.x. We can do a real fix in 0.2.4.
Linus Nordberg [Thu, 23 Aug 2012 10:23:00 +0000 (12:23 +0200)]
Move ipv6_preferred from routerinfo_t to node_t.
Move extend_info_from_router() from circuitbuild.c to router.c and
make it static.
Add get_configured_bridge_by_orports_digest() and have
get_configured_bridge_by_routerinfo() and
node_is_a_configured_bridge() use it. We now consider all OR ports of
a bridge when looking for it.
Move node_get_*_orport to nodelist.c.
Fix a cut'n'paste error in header of nodelist.h.
Add node_assert_ok().
Add router_get_all_orports(). It's duplicating code from
node_get_all_orports(). Worth fixing at the cost of complicating the
API slightly?
Nick Mathewson [Tue, 21 Aug 2012 14:15:52 +0000 (10:15 -0400)]
Avoid spurious warnings in rend_service_introduce
There was some code in the "err:" block that would always log a
warning, reporting an "unknown error" if we hadn't set err_msg. But
there were also plenty of "goto err" blocks that did their own
logging, and never set err_msg at all. Now we should only log when
we have an error message to log.
This fixes bug 6638, from no released Tor version.
Nick Mathewson [Fri, 17 Aug 2012 20:46:11 +0000 (16:46 -0400)]
When iterating over connections pending DNS, skip marked ones
Failure to do this would lead to double-free cases and similar,
especially when the exit's DNS was broken. See bug 6472 for full
details; this is a fix for 6472.