| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Avoid using a pronoun where it makes comments unclear.
Avoid using gender for things that don't have it.
Avoid assigning gender to people unnecessarily.
|
|
I believe that the final SMARTLIST_DEL_CURRENT was sometimes
double-removing items that had already been removed by
connection_mark_unattached_ap or
connection_ap_handshake_attach_circuit().
The fix here is to prevent iteration over the list that other
functions might be modifying.
|
|
This might also be what #17752 needs.
|
|
Fixes #17876
|
|
It is AP-specific, so that's where it belongs. This shouldn't have
caused a bug, but due to #17876, we were never actually calling
connection_edge_about_to_close from connection_ap_about_to_close,
causing bug #17874 (aka bug #17752).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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.
|
|
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
|
|
|
|
When set, this limits the maximum number of simultaneous streams per
rendezvous circuit on the server side of a HS, with further RELAY_BEGIN
cells being silently ignored.
This can be modified via "HiddenServiceMaxStreamsCloseCircuit", which
if set will cause offending rendezvous circuits to be torn down instead.
Addresses part of #16052.
|
|
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>
|
|
Every callsite that use to allocate a rend_data_t object now use the
rend_data_client/service_create() function.
Signed-off-by: David Goulet <dgoulet@ev0ke.net>
|
|
|
|
|
|
Fixes bug 15064; bugfix on 0.2.5.4-alpha.
|
|
|
|
|
|
|
|
|
|
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>
|
|
|
|
|
|
|
|
Fixes 14280 bugfix on 1053af0b9c4127873034a935ce33 in 0.2.4.7-alpha.
|
|
Conflicts:
src/or/connection_edge.c
|
|
Also, do a little light refactoring to move some variable declarations
around and make a few things const
Also fix an obnoxious bug on checking for the DONE stream end reason.
It's not a flag; it's a possible value or a variable that needs to be
masked.
|
|
Spotted by asn. #14259. Bugfix on 368eb6a97 in 0.2.0.1-alpha.
|
|
Also rename some options for uniformity, and apply this script:
@@
entry_connection_t *conn;
@@
conn->
+entry_cfg.
\(
isolation_flags
\|
session_group
\|
socks_prefer_no_auth
\|
ipv4_traffic
\|
ipv6_traffic
\|
prefer_ipv6
\|
cache_ipv4_answers
\|
cache_ipv6_answers
\|
use_cached_ipv4_answers
\|
use_cached_ipv6_answers
\|
prefer_ipv6_virtaddr
\)
|
|
The trick here is to apply mapaddress first, and only then apply
automapping. Otherwise, the automap checks don't get done.
Fix for bug 7555; bugfix on all versions of Tor supporting both
MapAddress and AutoMap.
|
|
|
|
|
|
|
|
I'd also like to split out the part that sends early socks responses.
|
|
This allows hidden services to disable the anti-scanning feature
introduced in 0.2.6.2-alpha. With this option not set, a connection
to an unlisted port closes the circuit. With this option set, only
a RELAY_DONE cell is sent.
Closes ticket #14084.
|
|
|
|
|
|
Make hidden service port scanning harder by sending back REASON_DONE which
does not disclose that it was in fact an exit policy issue. After that, kill
the circuit immediately to avoid more bad requests on it.
This means that everytime an hidden service exit policy does match, the user
(malicious or not) needs to build a new circuit.
Fixes #13667.
Signed-off-by: David Goulet <dgoulet@ev0ke.net>
|
|
|
|
|
|
The original call to getsockopt to know the original address on transparently
proxyed sockets using REDIRECT in iptables failed with IPv6 addresses because
it assumed all sockets used IPv4.
This patch fixes this by using the appropriate options and adding the headers
containing the needed definitions for these.
This patch is released under the same license as the original file as
long as the author iscredited.
Signed-off-by: Francisco Blas Izquierdo Riera (klondike) <klondike@gentoo.org>
|
|
The address of an array in the middle of a structure will
always be non-NULL. clang recognises this and complains.
Disable the tautologous and redundant check to silence
this warning.
Fixes bug 14001.
|
