Commit Graph

30620 Commits

Author SHA1 Message Date
Nick Mathewson
368413a321 Fix possible UB in an end-of-string check in get_next_token().
Remember, you can't check to see if there are N bytes left in a
buffer by doing (buf + N < end), since the buf + N computation might
take you off the end of the buffer and result in undefined behavior.

Fixes 28202; bugfix on 0.2.0.3-alpha.
2018-10-25 09:06:13 -04:00
Nick Mathewson
c78904b630 Merge branch 'maint-0.3.4' into maint-0.3.5
"ours" to avoid taking a second instance of 27797.
2018-10-24 12:57:14 -04:00
David Goulet
b063ca0604 node: Make node_supports_v3_rendezvous_point() also check for the key
It is not enough to look at protover for v3 rendezvous support but also we
need to make sure that the curve25519 onion key is present or in other words
that the descriptor has been fetched and does contain it.

Fixes #27797.

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-24 10:30:26 -04:00
Nick Mathewson
6851a08090 Merge branch 'ticket28115_035' into maint-0.3.5 2018-10-24 10:20:41 -04:00
cypherpunks
bfab7b9d35 rust/tor_log: fix C_RUST_COUPLED documentation
This file was moved in 97b15a1d7c
and moved again in e7f5f48d68.
2018-10-24 09:02:04 -04:00
Nick Mathewson
98e28efa8c Merge branch 'nss-travis' into maint-0.3.5 2018-10-24 08:57:22 -04:00
Nick Mathewson
131ab9d523 Merge branch 'maint-0.3.4' into maint-0.3.5
"ours" version to avoid removing rust use declaration.
2018-10-23 19:54:04 -04:00
Nick Mathewson
fd6078b33a Merge branch 'maint-0.3.3' into maint-0.3.4 2018-10-23 19:50:53 -04:00
Nick Mathewson
b2c52f5d6d Remove a now-unused rust "use". 2018-10-23 19:50:45 -04:00
Nick Mathewson
289a7dbac3 Merge branch 'maint-0.3.4' into maint-0.3.5 2018-10-23 19:21:44 -04:00
Nick Mathewson
7edc594ee7 Merge branch 'maint-0.3.3' into maint-0.3.4 2018-10-23 19:18:23 -04:00
Nick Mathewson
169bc670dd Merge remote-tracking branch 'onionk/rust-allsupportednull1' into maint-0.3.3 2018-10-23 19:17:48 -04:00
George Kadianakis
a614731144 Documentation: Move the hs_service_descriptor_t elements around.
Move the elements around to concentrate mutable and immutable elements
together. This commit changes no code, check with --color-moved.
2018-10-23 16:43:54 +03:00
George Kadianakis
df78fb2451 Documentation: Document which descriptor elements are (im)mutable. 2018-10-23 16:43:54 +03:00
George Kadianakis
29c194e022 Func rename: Make it clear that update_all_descriptors() does intro points.
With the new refresh_service_descriptor() function we had both
refresh_service_descriptor() and update_service_descriptor() which is basically
the same thing.

This commit renames update_service_descriptor() to
update_service_descriptor_intro_points() to make it clear it's not a generic
refresh and it's only about intro points.

Commit changes no code.
2018-10-23 16:43:54 +03:00
teor
939657771c
Treat backtrace test failures as expected on most BSD-derived systems
Treat backtrace test failures as expected on NetBSD, OpenBSD, and
macOS/Darwin, until we solve bug 17808.

(FreeBSD failures have been treated as expected since 18204 in 0.2.8.)

Fixes bug 27948; bugfix on 0.2.5.2-alpha.
2018-10-23 11:21:10 +10:00
David Goulet
81c466c34a hs-v3: Create desc signing key cert before uploading
Before this commit, we would create the descriptor signing key certificate
when first building the descriptor.

In some extreme cases, it lead to the expiry of the certificate which triggers
a BUG() when encoding the descriptor before uploading.

Ticket #27838 details a possible scenario in which this can happen. It is an
edge case where tor losts internet connectivity, notices it and closes all
circuits. When it came back up, the HS subsystem noticed that it had no
introduction circuits, created them and tried to upload the descriptor.

However, in the meantime, if tor did lack a live consensus because it is
currently seeking to download one, we would consider that we don't need to
rotate the descriptors leading to using the expired signing key certificate.

That being said, this commit does a bit more to make this process cleaner.
There are a series of things that we need to "refresh" before uploading a
descriptor: signing key cert, intro points and revision counter.

A refresh function is added to deal with all mutable descriptor fields. It in
turn simplified a bit the code surrounding the creation of the plaintext data.

We keep creating the cert when building the descriptor in order to accomodate
the unit tests. However, it is replaced every single time the descriptor is
uploaded.

Fixes #27838

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-22 16:34:41 -04:00
Nick Mathewson
1ae9116601 Merge remote-tracking branch 'onionk/rust-protocommas1' into maint-0.3.5 2018-10-19 14:20:57 -04:00
Nick Mathewson
d1d66866b2 Rename a function; it is used to convert a value _From_ le.
We can't use htons()/ntohs() -- those are no-ops on exactly the
wrong platforms.
2018-10-19 08:42:28 -04:00
Nick Mathewson
f8a1dc64f9 Fix a misspelled macro test that was breaking big-endian OPE
Fixes bug 28115; bugfix on 0.3.5.1-alpha.
2018-10-19 08:41:25 -04:00
David Goulet
8b2e72106a test: Add test for closing intro circuits when storing a new descriptor
This is client side and related to 27471 for previous commit that fixes this
issue.

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-18 12:56:51 -04:00
David Goulet
9ba16c4d03 hs-v3: Close client intro circuits if the descriptor is replaced
When storing a descriptor in the client cache, if we are about to replace an
existing descriptor, make sure to close every introduction circuits of the old
descriptor so we don't have leftovers lying around.

Ticket 27471 describes a situation where tor is sending an INTRODUCE1 cell on
an introduction circuit for which it doesn't have a matching intro point
object (taken from the descriptor).

The main theory is that, after a new descriptor showed up, the introduction
points changed which led to selecting an introduction circuit not used by the
service anymore thus for which we are unable to find the corresponding
introduction point within the descriptor we just fetched.

Closes #27471.

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-18 12:56:51 -04:00
David Goulet
56f713b8a4 hs-v3: Always generate the descriptor cookie
It won't be used if there are no authorized client configured. We do that so
we can easily support the addition of a client with a HUP signal which allow
us to avoid more complex code path to generate that cookie if we have at least
one client auth and we had none before.

Fixes #27995

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-18 11:46:07 -04:00
Nick Mathewson
d1eac7830f Merge branch 'maint-0.3.3' into maint-0.3.4 2018-10-18 09:12:58 -04:00
Nick Mathewson
a5599fb71c Merge branch 'maint-0.3.4' into maint-0.3.5 2018-10-18 09:12:58 -04:00
Taylor Yu
7f6c0fce46 Merge branch 'bug27800-034' into bug27800-035 2018-10-17 16:00:11 -05:00
Taylor Yu
4e7f65ee5d Merge branch 'bug27800-033' into tor-034 2018-10-17 15:42:58 -05:00
Taylor Yu
93fd924bdb Log more info for duplicate ed25519 IDs
Occasionally, key pinning doesn't catch a relay that shares an ed25519
ID with another relay.  Log the identity fingerprints and the shared
ed25519 ID when this happens, instead of making a BUG() warning.

Fixes bug 27800; bugfix on 0.3.2.1-alpha.
2018-10-17 15:39:55 -05:00
Nick Mathewson
389bae0e8b Bump to 0.3.5.3-alpha-dev 2018-10-17 14:46:00 -04:00
Nick Mathewson
444e9b37c5 expand the blurb 2018-10-17 14:20:55 -04:00
Nick Mathewson
cd6e75a6b4 Reflow the changelog 2018-10-17 14:19:33 -04:00
Roger Dingledine
df78a2730c merge in some fixes i found in a sandbox 2018-10-17 13:56:41 -04:00
Roger Dingledine
1a3fbba92b 0.3.5.3-alpha changelog touch-ups 2018-10-17 13:42:59 -04:00
Nick Mathewson
498a852bba Fold in the entry for 28089 2018-10-17 13:12:24 -04:00
Nick Mathewson
b1891068f1 Merge branch 'maint-0.3.4' 2018-10-17 13:02:23 -04:00
David Goulet
2000d04cb6 conn: Stop writing when our write bandwidth limist is exhausted
Commit 488e2b00bf introduced an issue, most
likely introduced by a bad copy paste, that made us stop reading on the
connection if our write bandwidth limit was reached.

The problem is that because "read_blocked_on_bw" was never set, the connection
was never reenabled for reading.

This is most likely the cause of #27813 where bytes were accumulating in the
kernel TCP bufers because tor was not doing reads. Only relays with
RelayBandwidthRate would suffer from this but affecting all relays connecting
to them. And using that tor option is recommended and best practice so many
many relays have it enabled.

Fixes #28089.
2018-10-17 12:25:12 -04:00
Nick Mathewson
0ba05092ad Light changelog edits 2018-10-17 09:40:20 -04:00
Nick Mathewson
d70ca3554e Bump to 0.3.5.3-alpha 2018-10-17 09:26:32 -04:00
Nick Mathewson
4b4233e0bc Run format-changelog 2018-10-17 09:26:04 -04:00
Nick Mathewson
e2bfffefb0 Fold changes file into changelog 2018-10-17 09:25:43 -04:00
Nick Mathewson
34cd1fc523 Merge remote-tracking branch 'tor-github/pr/406' 2018-10-16 21:33:38 -04:00
Nick Mathewson
63c5e09a40 Merge remote-tracking branch 'tor-github/pr/408' 2018-10-16 19:10:05 -04:00
Nick Mathewson
e97adaf8dc Argh. The unset value for OwningControllerFD is NOT -1. 2018-10-16 17:57:04 -04:00
Nick Mathewson
56a3cef4d7 Merge branch 'bug27849_redux' 2018-10-16 17:33:58 -04:00
Nick Mathewson
698629f5a9 Merge remote-tracking branch 'tor-github/pr/404' 2018-10-16 17:29:50 -04:00
Nick Mathewson
8a0b741487 Add a tor_free() in tor_gencert to fix a coverity warning 2018-10-16 17:18:46 -04:00
Neel Chauhan
f93ee8e4c4 Fix typo in comment for hs_cell_parse_introduce2() 2018-10-16 10:59:42 -04:00
David Goulet
b5731cdf2e mainloop: Set client role if ControlPort is set
It turns out that if _only_ the ControlPort is set and nothing else, tor would
simply not bootstrap and thus not start properly. Commit 67a41b6306
removed that requirement for tor to be considered a "client".

Unfortunately, this made the mainloop enable basically nothing if only the
ControlPort is set in the torrc.

This commit now makes it that we also consider the ControlPort when deciding
if we are a Client or not. It does not revert 67a41b6306 meaning
options_any_client_port_set() stays the same, not looking at the control port.

Fixes #27849.

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-16 08:11:49 -04:00
Nick Mathewson
b7fbd1f329 Merge branch 'maint-0.3.4' 2018-10-16 08:04:34 -04:00
David Goulet
38599de2dd test: Update approx_time before the test starts
This way we have the same time source when the IP is created and tested later.

Fixes #27810

Signed-off-by: David Goulet <dgoulet@torproject.org>
2018-10-16 08:03:55 -04:00