Commit Graph

373 Commits

Author SHA1 Message Date
Nick Mathewson
0a701e5377 More fixes/debugging attempts for 17659 2015-11-27 12:54:57 -05:00
Nick Mathewson
e5754c42d1 Merge branch 'bug17686_v2_027' 2015-11-25 22:33:49 -05:00
Nick Mathewson
ddcbe26474 Now that crypto_rand() cannot fail, it should return void. 2015-11-25 22:29:59 -05:00
Nick Mathewson
84b3350c83 Be more conservative in scanning the list of pending streams
Now we only re-scan the list in the cases we did before: when we
have a new circuit that we should try attaching to, or when we have
added a new stream that we haven't tried to attach yet.

This is part of 17590.
2015-11-17 09:04:25 -05:00
Nick Mathewson
b1d56fc589 Decouple ..attach_circuit() from most of its callers.
Long ago we used to call connection_ap_handshake_attach_circuit()
only in a few places, since connection_ap_attach_pending() attaches
all the pending connections, and does so regularly.  But this turned
out to have a performance problem: it would introduce a delay to
launching or connecting a stream.

We couldn't just call connection_ap_attach_pending() every time we
make a new connection, since it walks the whole connection list.  So
we started calling connection_ap_attach_pending all over, instead!
But that's kind of ugly and messes up our callgraph.

So instead, we now have connection_ap_attach_pending() use a list
only of the pending connections, so we can call it much more
frequently.  We have a separate function to scan the whole
connection array to see if we missed adding anything, and log a
warning if so.

Closes ticket #17590
2015-11-17 08:53:34 -05:00
Sebastian Hahn
5cf24ff3af Fix a bunch of check-spaces complaints 2015-08-21 10:36:53 -04:00
David Goulet
d06af95ba3 Purge client HS failure cache on NEWNYM
Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-08-11 09:34:41 -04:00
David Goulet
ab9a0e3407 Add rend failure cache
When fetching a descriptor, we know test every introduction points in it
against our rend failure cache to know if we keep it or not. For this to
work, now everytime an introduction points is discareded (ex: receiving a
NACK), we note it down in our introduction cache.

See rendcache.c for a detailed explanation of the cache's behavior.

Fixes #16389

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-08-11 09:34:41 -04:00
David Goulet
c1ffeadff4 Add missing descriptor ID to HS_DESC control event
For FAILED and RECEIVED action of the HS_DESC event, we now sends back the
descriptor ID at the end like specified in the control-spec section 4.1.25.

Fixes #15881

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-05-14 10:46:38 -04:00
Nick Mathewson
e8814816c7 whitespace fixes 2015-04-30 13:38:39 -04:00
David Goulet
a9b9f6d90f Update descriptor ID when it changes in rend_data
When we have a new descriptor ID for an onion address request, change it in
the rend_data_t object and purge the old one from the last hid serv request
cache.

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-04-30 12:35:21 -04:00
David Goulet
d33327ec22 Use descriptor ID when purging last hid fetch cache
Stop using an onion address since it's not indexed with that anymore in the
last hid serv request cache. Instead use a base32 encoded descriptor ID
contained in the rend_data_t object.

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-04-30 12:35:21 -04:00
David Goulet
e6a581f126 Add a create function for rend_data_t object
Ground works for fixing #15816. This adds the rend_data_create() function in
order to have a single place where we initialize that data structure.

Furthermore, an array of descriptor IDs is added (one per replica) so we can
keep a copy of the current id in the object. It will be used to purge the
last hid serv request cache using those descriptor IDs. When they change,
they will be replaced and the old ones will be purged from the cache.

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-04-30 12:35:20 -04:00
Nick Mathewson
b64eb6c47c Make a return value explicitly ignored.
This might make coverity happy
2015-04-23 13:05:06 -04:00
Nick Mathewson
54000d7ad9 Merge remote-tracking branch 'dgoulet/bug14847_027_06' 2015-04-23 12:24:35 -04:00
Nick Mathewson
3acee61422 Merge branch 'feature15652_squashed' 2015-04-23 09:09:33 -04:00
Yawning Angel
196499da73 Use a custom Base64 encoder with more control over the output format. 2015-04-23 09:06:58 -04:00
David Goulet
a4585405d6 Multiple fixes for the HSFETCH command
Ref:
https://trac.torproject.org/projects/tor/ticket/14847?replyto=31#comment:31

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-04-21 14:22:54 -04:00
David Goulet
59f8dced11 Refactor HS descriptor fetch to support descriptor ID
Big refactor of the HS client descriptor fetch functionnality. This allows
to fetch an HS descriptor using only a descriptor ID. Furthermore, it's also
possible to provide a list of HSDir(s) now that are used instead of the
automatically choosen one.

The approach taken was to add a descriptor_id field to the rend_data_t
structure so it can be used, if available, by the HS client. The onion
address field however has priority over it that is if both are set, the
onion address is used to fetch the descriptor.

A new public function is introduced called rend_client_fetch_v2_desc(...)
that does NOT lookup the client cache before fetching and can take a list of
HSDirs as a parameter.

The HSFETCH control command now uses this new function thus making it work
and final.

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-04-21 14:22:48 -04:00
David Goulet
e9782043c8 Remove onion address usage in lookup_last_hid_serv_request
Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-04-21 14:15:02 -04:00
Nick Mathewson
8837cc266e Merge remote-tracking branch 'dgoulet/bug14391_026_v2' 2015-04-15 10:33:04 -04:00
Nick Mathewson
86105a4009 Check ENABLE_TOR2WEB_MODE before any tor2webmode code 2015-02-19 09:09:25 -05:00
David Goulet
91009dce97 Refactor rend_cache_lookup_entry() and how it's used
Here is why:

1) v0 descriptors are deprecated since 0.2.2.1 and not suppose to be alive
in the network anymore. This function should only serve v2 version for now
as the default.

2) It should return different error code depending on what's the actual
error is. Right now, there is no distinction between a cache entry not found
and an invalid query.

3) This function should NOT test if the intro points are usable or not. This
adds some load on a function that should be "O(1)" and do one job.
Furthermore, multiple callsites actually already test that doing twice the
job...

4) While adding control event, it would be useful to be able to lookup a
cache entry without having it checking the intro points. There are also
places in the code that do want to lookup the cache entry without doing
that.

Fixes #14391

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-01-27 12:04:40 -05:00
Nick Mathewson
9ddc1fb10c Merge remote-tracking branch 'dgoulet/bug14224_025_v1' 2015-01-20 14:02:07 -05:00
David Goulet
b5525476f5 Fix: close intro circuit if no more intro points are usable
Once a NACK is received on the intro circuit, tor tries an other usable one
by extending the current circuit to it. If no more intro points are usable,
now close the circuit. Also, it's reason is changed before closing it so we
don't report again an intro point failure and trigger an extra HS fetch.

Fixes #14224

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-01-19 10:23:58 -05:00
David Goulet
b3c1152bae Fix: close intro circuit if no more intro points are usable
Once a NACK is received on the intro circuit, tor tries an other usable one
by extending the current circuit to it. If no more intro points are usable,
now close the circuit.

Fixes #14224

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2015-01-19 09:44:11 -05:00
Roger Dingledine
660a35d97c fix typo 2015-01-15 11:03:13 -05:00
Andrea Shepard
f7bb60e202 Add comments in rendclient.c noting that certain functions involved in handling ended HS connection attempts must be idempotent 2015-01-15 15:19:31 +00:00
Nick Mathewson
f54e54b0b4 Bump copyright dates to 2015, in case someday this matters. 2015-01-02 14:27:39 -05:00
Nick Mathewson
b3bd7a736c Remove Support022HiddenServices
This has been already disabled in the directory consensus for a while;
it didn't seem to break anything.

Finally closes #7803.
2014-11-17 11:52:10 -05:00
Nick Mathewson
a87c697fb1 Merge remote-tracking branch 'public/bug13698_024_v1' 2014-11-12 10:23:55 -05:00
David Goulet
34eb007d22 Fix: don't report timeout when closing parallel intro points
When closing parallel introduction points, the given reason (timeout)
was actually changed to "no reason" thus when the circuit purpose was
CIRCUIT_PURPOSE_C_INTRODUCE_ACK_WAIT, we were reporting an introduction
point failure and flagging it "unreachable". After three times, that
intro point gets removed from the rend cache object.

In the case of CIRCUIT_PURPOSE_C_INTRODUCING, the intro point was
flagged has "timed out" and thus not used until the connection to the HS
is closed where that flag gets reset.

This commit adds an internal circuit reason called
END_CIRC_REASON_IP_NOW_REDUNDANT which tells the closing circuit
mechanism to not report any intro point failure.

This has been observed while opening hundreds of connections to an HS on
different circuit for each connection. This fix makes this use case to
work like a charm.

Fixes #13698.

Signed-off-by: David Goulet <dgoulet@ev0ke.net>
2014-11-10 15:02:54 -05:00
Nick Mathewson
fcdcb377a4 Add another year to our copyright dates.
Because in 95 years, we or our successors will surely care about
enforcing the BSD license terms on this code.  Right?
2014-10-28 15:30:16 -04:00
Nick Mathewson
feee445771 Merge remote-tracking branch 'origin/maint-0.2.5' 2014-09-16 11:11:48 -04:00
Nick Mathewson
be0e26272b Merge remote-tracking branch 'origin/maint-0.2.4' into maint-0.2.5 2014-09-16 11:10:02 -04:00
Roger Dingledine
0c3b3650aa clients now send correct address for rendezvous point
Clients now send the correct address for their chosen rendezvous point
when trying to access a hidden service. They used to send the wrong
address, which would still work some of the time because they also
sent the identity digest of the rendezvous point, and if the hidden
service happened to try connecting to the rendezvous point from a relay
that already had a connection open to it, the relay would reuse that
connection. Now connections to hidden services should be more robust
and faster. Also, this bug meant that clients were leaking to the hidden
service whether they were on a little-endian (common) or big-endian (rare)
system, which for some users might have reduced their anonymity.

Fixes bug 13151; bugfix on 0.2.1.5-alpha.
2014-09-16 11:05:36 -04:00
Nick Mathewson
6969bd9a02 Autoconvert most circuit-list iterations to smartlist iterations
Breaks compilation.

Used this coccinelle script:

@@
identifier c;
typedef circuit_t;
iterator name TOR_LIST_FOREACH;
iterator name SMARTLIST_FOREACH_BEGIN;
statement S;
@@
- circuit_t *c;
   ...
- TOR_LIST_FOREACH(c, \(&global_circuitlist\|circuit_get_global_list()\), head)
+ SMARTLIST_FOREACH_BEGIN(circuit_get_global_list(), circuit_t *, c)
  S
+ SMARTLIST_FOREACH_END(c);
2014-08-15 16:23:22 -04:00
Nick Mathewson
3dfed0806c Merge remote-tracking branch 'public/bug10722' 2014-02-16 12:13:12 -05:00
Nick Mathewson
b5d6e47002 Warning message when bug 10722 would trigger
If somebody's excludenodes settings are keeping their hidden service
connections from working, they should probably get notified about it.
2014-02-16 12:11:07 -05:00
Karsten Loesing
74c2bff781 Remove remaining v0 hidden service descriptor code.
Fixes the rest of #10841 after #10881 already removed some hidden service
authority code.
2014-02-12 14:36:08 +01:00
Nick Mathewson
cb28fe486f Merge remote-tracking branch 'public/bug10722' 2014-02-11 10:08:58 -05:00
Qingping Hou
bf66ff915a fix longname returned in HS_DESC control events
According to control spec, longname should not contain any spaces and is
consists only of identy_digest + nickname

added two functions:
* node_get_verbose_nickname_by_id()
* node_describe_longname_by_id()
2014-02-06 16:13:55 -05:00
Qingping Hou
29c18f5b71 add hidden service descriptor async control event 2014-01-29 22:23:57 -05:00
Nick Mathewson
bb21d14255 Apply StrictNodes to hidden service directories early
Previously, we would sometimes decide in directory_get_from_hs_dir()
to connect to an excluded node, and then later in
directory_initiate_command_routerstatus_rend() notice that it was
excluded and strictnodes was set, and catch it as a stopgap.

Additionally, this patch preferentially tries to fetch from
non-excluded nodes even when StrictNodes is off.

Fix for bug #10722. Bugfix on 0.2.0.10-alpha (the v2 hidserv directory
system was introduced in e136f00ca). Reported by "mr-4".
2014-01-24 12:56:10 -05:00
Nick Mathewson
761ee93c69 Add missing includes for circpathbias.h 2013-10-31 14:33:34 -04:00
Nick Mathewson
5b1c257d9a Merge remote-tracking branch 'origin/maint-0.2.4' 2013-09-23 01:48:30 -04:00
Nick Mathewson
116e6af7a7 Fix a compilation warning with older gcc 2013-09-23 01:47:52 -04:00
Nick Mathewson
6178aaea06 Merge remote-tracking branch 'origin/maint-0.2.4' 2013-09-20 11:01:10 -04:00
Nick Mathewson
f8b44eedf7 Get ready to stop sending timestamps in INTRODUCE cells
For now, round down to the nearest 10 minutes.  Later, eliminate entirely by
setting a consensus parameter.

(This rounding is safe because, in 0.2.2, where the timestamp mattered,
REND_REPLAY_TIME_INTERVAL was a nice generous 60 minutes.)
2013-09-20 11:00:27 -04:00
Nick Mathewson
a3ffa1f76e Rename circuit_get_global_list to remove trailing _ 2013-08-15 15:37:23 -04:00