Commit Graph

269 Commits

Author SHA1 Message Date
Michael Kaye
f85b9842f0 Don't encode object as UTF-8 string if not needed.
I believe that string formatting ~10-15 sized events will
take a proportion of CPU time.
2019-10-24 18:08:45 +01:00
Amber Brown
850dcfd2d3
Fix well-known lookups with the federation certificate whitelist (#5997) 2019-09-14 04:58:38 +10:00
Andrew Morgan
4548d1f87e
Remove unnecessary parentheses around return statements (#5931)
Python will return a tuple whether there are parentheses around the returned values or not.

I'm just sick of my editor complaining about this all over the place :)
2019-08-30 16:28:26 +01:00
Erik Johnston
f5b50d0871
Merge pull request #5895 from matrix-org/erikj/notary_key
Add config option to sign remote key query responses with a separate key.
2019-08-27 11:51:37 +01:00
Erik Johnston
7af5a63063 Fixup review comments 2019-08-23 15:36:28 +01:00
Richard van der Hoff
ef1c524bb3
Improve error msg when key-fetch fails (#5896)
There's no point doing a raise_from here, because the exception is always
logged at warn with no stacktrace in the caller. Instead, let's try to give
better messages to reduce confusion.

In particular, this means that we won't log 'Failed to connect to remote
server' when we don't even attempt to connect to the remote server due to
blacklisting.
2019-08-22 10:42:06 +01:00
Erik Johnston
97cbc96093 Only sign when we respond to remote key requests 2019-08-21 11:21:58 +01:00
Erik Johnston
5906be8589 Add config option for keys to use to sign keys
This allows servers to separate keys that are used to sign remote keys
when acting as a notary server.
2019-08-21 10:44:58 +01:00
Erik Johnston
a9bcae9f50 Share SSL options for well-known requests 2019-07-31 10:39:24 +01:00
Amber Brown
4806651744
Replace returnValue with return (#5736) 2019-07-23 23:00:55 +10:00
Richard van der Hoff
dcca56baba Add a delay to key lookup lock release to fix stack overflow
A tactical call_later here should fix #5723
2019-07-19 17:57:00 +01:00
Richard van der Hoff
c7095be913 Refactor Keyring._start_key_lookups
There's an awful lot of deferreds and dictionaries flying around here. The
whole thing can be made much simpler and achieve the same effect.
2019-07-19 17:49:19 +01:00
Amber Brown
463b072b12
Move logging utilities out of the side drawer of util/ and into logging/ (#5606) 2019-07-04 00:07:04 +10:00
Amber Brown
be3b901ccd
Update the TLS cipher string and provide configurability for TLS on outgoing federation (#5550) 2019-06-28 18:19:09 +10:00
Amber Brown
32e7c9e7f2
Run Black. (#5482) 2019-06-20 19:32:02 +10:00
Erik Johnston
a6b1817940
Merge pull request #5417 from matrix-org/rav/shared_ssl_context
Share an SSL context object between SSL connections
2019-06-10 18:16:12 +01:00
Richard van der Hoff
81b8fdedf2 rename gutwrenched attr 2019-06-10 17:51:11 +01:00
Richard van der Hoff
efe7b3176e Fix federation connections to literal IP addresses
turns out we need a shiny version of service_identity to enforce this
correctly.
2019-06-10 15:58:35 +01:00
Richard van der Hoff
d11c634ced clean up impl, and import idna directly 2019-06-10 15:55:12 +01:00
Richard van der Hoff
88d7182ada
Improve startup checks for insecure notary configs (#5392)
It's not really a problem to trust notary responses signed by the old key so
long as we are also doing TLS validation.

This commit adds a check to the config parsing code at startup to check that
we do not have the insecure matrix.org key without tls validation, and refuses
to start without it.

This allows us to remove the rather alarming-looking warning which happens at
runtime.
2019-06-10 10:33:00 +01:00
Richard van der Hoff
c2b6e945e1 Share an SSL context object between SSL connections
This involves changing how the info callbacks work.
2019-06-09 14:01:32 +01:00
Richard van der Hoff
9fbb20a531
Stop hardcoding trust of old matrix.org key (#5374)
There are a few changes going on here:

* We make checking the signature on a key server response optional: if no
  verify_keys are specified, we trust to TLS to validate the connection.

* We change the default config so that it does not require responses to be
  signed by the old key.

* We replace the old 'perspectives' config with 'trusted_key_servers', which
  is also formatted slightly differently.

* We emit a warning to the logs every time we trust a key server response
  signed by the old key.
2019-06-06 17:33:11 +01:00
Richard van der Hoff
cea9750d11 Associate a request_name with each verify request, for logging
Also:
* rename VerifyKeyRequest->VerifyJsonRequest
* calculate key_ids on VerifyJsonRequest construction
* refactor things to pass around VerifyJsonRequests instead of 4-tuples
2019-06-05 10:46:26 +01:00
Richard van der Hoff
2615c6bd9e
Clean up debug logging (#5347)
Remove some spurious stuff, clarify some other stuff
2019-06-05 10:35:40 +01:00
Richard van der Hoff
016af01598
Rename VerifyKeyRequest.deferred field (#5343)
it's a bit confusing
2019-06-05 10:35:13 +01:00
Richard van der Hoff
4d08b8f30c
Don't do long retries when calling the key notary server. (#5334)
It takes at least 20 minutes to work through the long_retries schedule (11
attempts, each with a 60 second timeout, and 60 seconds between each request),
so if the notary server isn't returning within the timeout, we'll just end up
blocking whatever request is happening for 20 minutes.

Ain't nobody got time for that.
2019-06-04 11:53:07 +01:00
Richard van der Hoff
c5d60eadd5 Notary server: make requests to origins in parallel
... else we're guaranteed to time out.
2019-06-04 00:16:56 +01:00
Richard van der Hoff
def5ea4062 Don't bomb out on direct key fetches as soon as one fails 2019-06-04 00:16:56 +01:00
Richard van der Hoff
06a1f3e207 Reduce timeout for outbound /key/v2/server requests. 2019-06-03 23:17:38 +01:00
Richard van der Hoff
fec2dcb1a5
Enforce validity period on server_keys for fed requests. (#5321)
When handling incoming federation requests, make sure that we have an
up-to-date copy of the signing key.

We do not yet enforce the validity period for event signatures.
2019-06-03 22:59:51 +01:00
Richard van der Hoff
862b2f9ad5
Merge pull request #5307 from matrix-org/rav/server_keys/07-fix-notary-cache-poison
Stop overwriting server keys with other keys
2019-06-03 13:19:20 +01:00
Richard van der Hoff
3600f5568b Stop overwriting server keys with other keys
Fix a bug where we would discard a key result which the origin server is no
longer returning. Fixes #5305.
2019-05-31 15:58:35 +01:00
Richard van der Hoff
c605da97bf Merge remote-tracking branch 'origin/develop' into rav/server_keys/05-rewrite-gsvk-again 2019-05-31 11:38:13 +01:00
Richard van der Hoff
8ea2f756a9 Remove some pointless exception handling
The verify_request deferred already returns a suitable SynapseError, so I don't
really know what we expect to achieve by doing more wrapping, other than log
spam.

Fixes #4278.
2019-05-30 18:29:56 +01:00
Richard van der Hoff
a82c96b87f Rewrite get_server_verify_keys, again.
Attempt to simplify the logic in get_server_verify_keys by splitting it into
two methods.
2019-05-30 18:20:40 +01:00
Richard van der Hoff
099829d5a9 use attr.s for VerifyKeyRequest
because namedtuple is awful
2019-05-30 17:39:28 +01:00
Richard van der Hoff
540f40f0cd
Merge pull request #5251 from matrix-org/rav/server_keys/01-check_sig
Ensure that server_keys fetched via a notary server are correctly signed.
2019-05-28 21:32:17 +01:00
Richard van der Hoff
fa1b293da2
Simplification to Keyring.wait_for_previous_lookups. (#5250)
The list of server names was redundant, since it was equivalent to the keys on
the server_to_deferred map. This reduces the number of large lists being passed
around, and has the benefit of deduplicating the entries in `wait_on`.
2019-05-24 22:17:18 +01:00
Richard van der Hoff
b825d1c800 Improve error handling/logging for perspectives-key fetching.
In particular, don't give up on the first failure.
2019-05-24 15:46:25 +01:00
Richard van der Hoff
753b1270da Require sig from origin server on perspectives responses 2019-05-23 15:01:09 +01:00
Richard van der Hoff
895b79ac2e Factor out KeyFetchers from KeyRing
Rather than have three methods which have to have the same interface,
factor out a separate interface which is provided by three implementations.

I find it easier to grok the code this way.
2019-05-23 13:46:47 +01:00
Richard van der Hoff
b75537beaf Store key validity time in the storage layer
This is a first step to checking that the key is valid at the required moment.

The idea here is that, rather than passing VerifyKey objects in and out of the
storage layer, we instead pass FetchKeyResult objects, which simply wrap the
VerifyKey and add a valid_until_ts field.
2019-05-23 11:52:22 +01:00
Richard van der Hoff
84660d91b2
Simplify process_v2_response (#5236)
* Pass time_added_ms into process_v2_response

* Simplify process_v2_response

We can merge old_verify_keys into verify_keys, and reduce the number of dicts
flying around.
2019-05-23 11:51:39 +01:00
Richard van der Hoff
cc187f9337
Remove unused VerifyKey.expired and .time_added fields (#5235)
These were never used, and poking arbitary data into objects from other
packages seems confusing at best.
2019-05-23 11:46:05 +01:00
Richard van der Hoff
2e052110ee
Rewrite store_server_verify_key to store several keys at once (#5234)
Storing server keys hammered the database a bit. This replaces the
implementation which stored a single key, with one which can do many updates at
once.
2019-05-23 11:45:39 +01:00
Richard van der Hoff
1a94de60e8
Run black on synapse.crypto.keyring (#5232) 2019-05-22 18:39:33 +01:00
Richard van der Hoff
fd8fb32bdd remove extraneous exception logging 2019-04-25 22:02:03 +01:00
Richard van der Hoff
7ca638c761 Clarify logging when PDU signature checking fails 2019-04-25 20:55:12 +01:00
Andrew Morgan
6824ddd93d Config option for verifying federation certificates (MSC 1711) (#4967) 2019-04-25 14:22:49 +01:00
Andrew Morgan
caa76e6021
Remove periods from copyright headers (#5046) 2019-04-11 17:08:13 +01:00
Richard van der Hoff
18b69be00f Rewrite Datastore.get_server_verify_keys
Rewrite this so that it doesn't hammer the database.
2019-04-09 00:00:10 +01:00
Richard van der Hoff
f88a9e6323 Remove redundant merged_keys dict
There's no point in collecting a merged dict of keys: it is sufficient to
consider just the new keys which have been fetched by the most recent
key_fetch_fns.
2019-04-08 22:36:18 +01:00
Richard van der Hoff
7d2a0c848e Fix from_server buglet in get_keys_from_perspectives
make sure we store the name of the server the keys came from, rather than the
origin server, after doing a fetch-from-perspectives.
2019-04-08 12:51:16 +01:00
Richard van der Hoff
6ae9361510 Hoist server_name check out of process_v2_response
It's easier to check it in the caller than to complicate the interface with an
extra param.
2019-04-04 19:12:54 +01:00
Richard van der Hoff
ef27d434d1 Clean up Keyring.process_v2_response
Make this just return the key dict, rather than a single-entry dict mapping the
server name to the key dict. It's easy for the caller to get the server name
from from the response object anyway.
2019-04-04 19:12:54 +01:00
Erik Johnston
78c563b77c Correctly log expected errors when fetching server keys 2019-03-11 14:11:10 +00:00
Erik Johnston
65d1003d01 raise_from already raises 2019-02-25 14:34:03 +00:00
Erik Johnston
41285ffe5b Handle errors when fetching remote server keys 2019-02-23 15:09:39 +00:00
Amber Brown
561eebe170 fix to use makeContext so that we don't need to rebuild the certificateoptions each time 2019-02-19 16:18:05 +11:00
Erik Johnston
7fc1196a36 Correctly handle RequestSendFailed exceptions
This mainly reduces the number of exceptions we log.
2019-02-14 14:01:04 +00:00
Richard van der Hoff
9645728619 Don't create server contexts when TLS is disabled
we aren't going to use them anyway.
2019-02-11 21:32:01 +00:00
Erik Johnston
554ca58ea1 Make add_hashes_and_signatures operate on dicts 2019-01-29 11:12:38 +00:00
Erik Johnston
855a151015 Refactor event signing to work on dicts
This is in preparation for making EventBuilder format agnostic, which
means event signing should be done against the event dict rather than
the EventBuilder object.
2019-01-28 16:42:10 +00:00
Richard van der Hoff
97fd29c019
Don't send IP addresses as SNI (#4452)
The problem here is that we have cut-and-pasted an impl from Twisted, and then
failed to maintain it. It was fixed in Twisted in
https://github.com/twisted/twisted/pull/1047/files; let's do the same here.
2019-01-24 09:34:44 +00:00
Richard van der Hoff
6bfa735a69
Make key fetches use regular federation client (#4426)
All this magic is redundant.
2019-01-22 11:04:20 +00:00
Amber Brown
23b0813599
Require ECDH key exchange & remove dh_params (#4429)
* remove dh_params and set better cipher string
2019-01-22 21:58:50 +11:00
Amber Brown
916efc8249
Remove fetching keys via the deprecated v1 kex method (#4120) 2018-10-31 23:14:39 +11:00
Richard van der Hoff
ef771cc4c2 Fix a number of flake8 errors
Broadly three things here:

* disable W504 which seems a bit whacko
* remove a bunch of `as e` expressions from exception handlers that don't use
  them
* use `r""` for strings which include backslashes

Also, we don't use pep8 any more, so we can get rid of the duplicate config
there.
2018-10-24 10:39:03 +01:00
Amber Brown
33716c4aea
Merge pull request #3826 from matrix-org/rav/logging_for_keyring
add some logging for the keyring queue
2018-09-12 20:43:47 +10:00
Amber Brown
8fd93b5eea
Port crypto/ to Python 3 (#3822) 2018-09-12 20:16:31 +10:00
Richard van der Hoff
806964b5de add some logging for the keyring queue
why is it so damn slow?
2018-09-06 18:51:06 +01:00
Erik Johnston
808d8e06aa Don't log exceptions when failing to fetch server keys
Not being able to resolve or connect to remote servers is an expected
error, so we shouldn't log at ERROR with stacktraces.
2018-08-21 11:19:26 +01:00
Jeroen
2e9c73e8ca more generic conversion of str/bytes to unicode 2018-08-09 21:31:26 +02:00
Jeroen
64899341dc include private functions from twisted 2018-08-09 21:04:22 +02:00
Jeroen
d5c0ce4cad updated docstring for ServerContextFactory 2018-08-08 19:25:01 +02:00
Jeroen
2903e65aff fix isort 2018-07-29 19:47:08 +02:00
Jeroen
8e3f75b39a fix accidental removal of hs 2018-07-27 12:17:31 +02:00
Jeroen
505530f36a Merge remote-tracking branch 'upstream/develop' into send_sni_for_federation_requests
# Conflicts:
#	synapse/crypto/context_factory.py
2018-07-14 20:24:46 +02:00
Jeroen
b5e157d895 Merge branch 'develop' into send_sni_for_federation_requests
# Conflicts:
#	synapse/http/endpoint.py
2018-07-09 08:51:11 +02:00
Amber Brown
49af402019 run isort 2018-07-09 16:09:20 +10:00
Amber Brown
6350bf925e
Attempt to be more performant on PyPy (#3462) 2018-06-28 14:49:57 +01:00
Jeroen
95341a8f6f take idna implementation from twisted 2018-06-26 21:15:14 +02:00
Jeroen
b7f34ee348 allow self-signed certificates 2018-06-26 20:41:05 +02:00
Jeroen
07b4f88de9 formatting changes for pep8 2018-06-25 12:31:16 +02:00
Jeroen
3d605853c8 send SNI for federation requests 2018-06-24 22:38:43 +02:00
Richard van der Hoff
e82db24a0e Try to log more helpful info when a sig verification fails
Firstly, don't swallow the reason for the failure

Secondly, don't assume all exceptions are verification failures

Thirdly, log a bit of info about the key being used if debug is enabled
2018-06-08 12:13:08 +01:00
Will Hunt
2ad3fc36e6 Fixes #3135 - Replace _OpenSSLECCurve with crypto.get_elliptic_curve (#3157)
fixes #3135

Signed-off-by: Will Hunt will@half-shot.uk
2018-04-30 16:21:11 +01:00
Richard van der Hoff
fc149b4eeb Merge remote-tracking branch 'origin/develop' into rav/use_run_in_background 2018-04-27 14:31:23 +01:00
Richard van der Hoff
2a13af23bc Use run_in_background in preference to preserve_fn
While I was going through uses of preserve_fn for other PRs, I converted places
which only use the wrapped function once to use run_in_background, to avoid
creating the function object.
2018-04-27 12:55:51 +01:00
Richard van der Hoff
9255a6cb17 Improve exception handling for background processes
There were a bunch of places where we fire off a process to happen in the
background, but don't have any exception handling on it - instead relying on
the unhandled error being logged when the relevent deferred gets
garbage-collected.

This is unsatisfactory for a number of reasons:
 - logging on garbage collection is best-effort and may happen some time after
   the error, if at all
 - it can be hard to figure out where the error actually happened.
 - it is logged as a scary CRITICAL error which (a) I always forget to grep for
   and (b) it's not really CRITICAL if a background process we don't care about
   fails.

So this is an attempt to add exception handling to everything we fire off into
the background.
2018-04-27 11:07:40 +01:00
Adrian Tschira
1515560f5c Use str(e) instead of e.message
Doing this I learned e.message was pretty shortlived, added in 2.6,
they realized it was a bad idea and deprecated it in 2.7

Signed-off-by: Adrian Tschira <nota@notafile.com>
2018-04-15 20:32:42 +02:00
Richard van der Hoff
7303ed65e1 Fix 500 when joining matrix-dev
matrix-dev has an event (`$/6ANj/9QWQyd71N6DpRQPf+SDUu11+HVMeKSpMzBCwM:zemos.net`)
which has no `hashes` member.

Check for missing `hashes` element in events.
2017-11-29 16:00:46 +00:00
Richard van der Hoff
eaaabc6c4f replace 'except:' with 'except Exception:'
what could possibly go wrong
2017-10-23 15:52:32 +01:00
Richard van der Hoff
94133d7ce8 Merge branch 'develop' into develop 2017-09-25 11:50:11 +01:00
Richard van der Hoff
79b3cf3e02 Fix logcontxt leak in keyclient (#2465)
preserve_context_over_function doesn't do what you want it to do.
2017-09-25 09:51:39 +01:00
Richard van der Hoff
c5c24c239b Fix logcontext handling in verify_json_objects_for_server
preserve_context_over_fn is essentially broken, because (a) it pointlessly
drops the current logcontext before calling its wrapped function, which means
we don't get any useful logcontexts for _handle_key_deferred; (b) it wraps the
resulting deferred in a _PreservingContextDeferred, which is very dangerous
because you then can't yield on it without leaking context back into the
reactor.

Instead, let's specify that the resultant deferreds call their callbacks with
no logcontext.
2017-09-20 01:32:42 +01:00
Richard van der Hoff
c5b0e9f485 Turn _start_key_lookups into an inlineCallbacks function
... which means that logcontexts can be correctly preserved for the stuff it
does.

get_server_verify_keys is now called with the logcontext, so needs to
preserve_fn when it fires off its nested inlineCallbacks function.

Also renames get_server_verify_keys to reflect the fact it's meant to be
private.
2017-09-20 01:32:42 +01:00
Richard van der Hoff
abdefb8a01 Fix potential race in _start_key_lookups
If the verify_request.deferred has already completed, then `remove_deferreds`
will be called immediately. It therefore might resolve the server_to_deferred
deferred while there are still other requests for that server in flight.

To avoid that, we should build the complete list of requests, and *then* add the
callbacks.
2017-09-20 01:32:42 +01:00
Richard van der Hoff
afbd773dc6 Add some comments to _start_key_lookups 2017-09-20 01:32:42 +01:00
Richard van der Hoff
2a4b9ea233 Consistency for how verify_request.deferred is called
Define that it is run with no log context, and make sure that happens.

If we aren't careful to reset the logcontext, we can't bung the deferreds into
defer.gatherResults etc. We don't actually do that directly, but we *do*
resolve other deferreds from affected callbacks (notably the server_to_deferred
map in _start_key_lookups), and those *do* get passed into
defer.gatherResults. It turns out that this way ends up being least confusing.
2017-09-20 01:32:42 +01:00