Commit Graph

497 Commits

Author SHA1 Message Date
Mathieu Velten
6cddf24e36
Faster joins: don't stall when a user joins during a fast join (#14606)
Fixes #12801.
Complement tests are at
https://github.com/matrix-org/complement/pull/567.

Avoid blocking on full state when handling a subsequent join into a
partial state room.

Also always perform a remote join into partial state rooms, since we do
not know whether the joining user has been banned and want to avoid
leaking history to banned users.

Signed-off-by: Mathieu Velten <mathieuv@matrix.org>
Co-authored-by: Sean Quah <seanq@matrix.org>
Co-authored-by: David Robertson <davidr@element.io>
2023-02-10 23:31:05 +00:00
Shay
03bccd542b
Add a class UnpersistedEventContext to allow for the batching up of storing state groups (#14675)
* add class UnpersistedEventContext

* modify create new client event to create unpersistedeventcontexts

* persist event contexts after creation

* fix tests to persist unpersisted event contexts

* cleanup

* misc lints + cleanup

* changelog + fix comments

* lints

* fix batch insertion?

* reduce redundant calculation

* add unpersisted event classes

* rework compute_event_context, split into function that returns unpersisted event context and then persists it

* use calculate_context_info to create unpersisted event contexts

* update typing

* $%#^&*

* black

* fix comments and consolidate classes, use attr.s for class

* requested changes

* lint

* requested changes

* requested changes

* refactor to be stupidly explicit

* clearer renaming and flow

* make partial state non-optional

* update docstrings

---------

Co-authored-by: Erik Johnston <erik@matrix.org>
2023-02-09 13:05:02 -08:00
Shay
55e4d27b36
Limit concurrent event creation for a room to avoid state resolution when sending bursts of events to a local room (#14977) 2023-02-08 11:25:11 -08:00
Sean Quah
805b641fb6
Fix "Re-starting finished log context" spam when creating events (#14947)
`run_in_background` calls re-use the current logging context. When they
are not awaited, they can complete after the current logging context has
been marked as finished, which leads to log spam. Use
`run_as_background_process` instead.

Fixes one of the instances of #13090.

Signed-off-by: Sean Quah <seanq@matrix.org>
2023-01-31 11:31:52 +00:00
David Robertson
796a4b7482
Prefer type(x) is int to isinstance(x, int) (#14945)
* Perfer `type(x) is int` to `isinstance(x, int)`

This covered all additional instances I could see where `x` was
user-controlled.
The remaining cases are

```
$ rg -s 'isinstance.*[^_]int'
tests/replication/_base.py
576:        if isinstance(obj, int):

synapse/util/caches/stream_change_cache.py
136:        assert isinstance(stream_pos, int)
214:        assert isinstance(stream_pos, int)
246:        assert isinstance(stream_pos, int)
267:        assert isinstance(stream_pos, int)

synapse/replication/tcp/external_cache.py
133:        if isinstance(result, int):

synapse/metrics/__init__.py
100:        if isinstance(calls, (int, float)):

synapse/handlers/appservice.py
262:        assert isinstance(new_token, int)

synapse/config/_util.py
62:        if isinstance(p, int):
```

which cover metrics, logic related to `jsonschema`, and replication and
data streams. AFAICS these are all internal to Synapse

* Changelog
2023-01-31 10:33:07 +00:00
reivilibre
5172c8c403
Faster remote room joins (worker mode): do not populate external hosts-in-room cache when sending events as this requires blocking for full state. (#14749)
Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
Co-authored-by: Sean Quah <seanq@matrix.org>
2023-01-11 13:21:53 +00:00
Mathieu Velten
54c012c5a8
Make handle_new_client_event throws PartialStateConflictError (#14665)
Then adapts calling code to retry when needed so it doesn't 500
to clients.

Signed-off-by: Mathieu Velten <mathieuv@matrix.org>
Co-authored-by: Sean Quah <8349537+squahtx@users.noreply.github.com>
2022-12-15 16:04:23 +00:00
David Robertson
e2a1adbf5d
Allow selecting "prejoin" events by state keys (#14642)
* Declare new config

* Parse new config

* Read new config

* Don't use trial/our TestCase where it's not needed

Before:

```
$ time trial tests/events/test_utils.py > /dev/null

real	0m2.277s
user	0m2.186s
sys	0m0.083s
```

After:
```
$ time trial tests/events/test_utils.py > /dev/null

real	0m0.566s
user	0m0.508s
sys	0m0.056s
```

* Helper to upsert to event fields

without exceeding size limits.

* Use helper when adding invite/knock state

Now that we allow admins to include events in prejoin room state with
arbitrary state keys, be a good Matrix citizen and ensure they don't
accidentally create an oversized event.

* Changelog

* Move StateFilter tests

should have done this in #14668

* Add extra methods to StateFilter

* Use StateFilter

* Ensure test file enforces typed defs; alphabetise

* Workaround surprising get_current_state_ids

* Whoops, fix mypy
2022-12-13 00:54:46 +00:00
David Robertson
b5b5f66084
Move StateFilter to synapse.types (#14668)
* Move `StateFilter` to `synapse.types`

* Changelog
2022-12-12 16:19:30 +00:00
Shay
72f3e38137
Fix possible variable shadow in create_new_client_event (#14575) 2022-11-28 19:18:12 -08:00
Brendan Abolivier
86c5a710d8
Implement MSC3912: Relation-based redactions (#14260)
Co-authored-by: Sean Quah <8349537+squahtx@users.noreply.github.com>
2022-11-03 16:21:31 +00:00
Shay
b7a7ff6ee3
Add initial power level event to batch of bulk persisted events when creating a new room. (#14228) 2022-10-21 10:46:22 -07:00
Shay
847e2393f3
Prepatory work for adding power level event to batched events (#14214) 2022-10-18 09:58:47 -07:00
Shay
b6baa46db0
Fix a bug where the joined hosts for a given event were not being properly cached (#14125) 2022-10-12 11:01:00 -07:00
Shay
7b7478e8b6
Batch up notifications after event persistence (#14033) 2022-10-05 10:12:48 -07:00
Shay
8ab16a92ed
Persist CreateRoom events to DB in a batch (#13800) 2022-09-28 10:11:48 +00:00
Shay
a2cf66a94d
Prepatory work for batching events to send (#13487)
This PR begins work on batching up events during the creation of a room. The PR splits out the creation and sending/persisting of the events. The first three events in the creation of the room-creating the room, joining the creator to the room, and the power levels event are sent sequentially, while the subsequent events are created and collected to be sent at the end of the function. This is currently done by appending them to a list and then iterating over the list to send, the next step (after this PR) would be to send and persist the collected events as a batch.
2022-09-28 10:39:03 +01:00
reivilibre
6302753012
Deduplicate is_server_notices_room. (#13780) 2022-09-14 15:53:18 +00:00
Eric Eastwood
d58615c82c
Directly lookup local membership instead of getting all members in a room first (get_users_in_room mis-use) (#13608)
See https://github.com/matrix-org/synapse/pull/13575#discussion_r953023755
2022-08-24 14:13:12 -05:00
Quentin Gliech
3dd175b628
synapse.api.auth.Auth cleanup: make permission-related methods use Requester instead of the UserID (#13024)
Part of #13019

This changes all the permission-related methods to rely on the Requester instead of the UserID. This is a first step towards enabling scoped access tokens at some point, since I expect the Requester to have scope-related informations in it.

It also changes methods which figure out the user/device/appservice out of the access token to return a Requester instead of something else. This avoids having store-related objects in the methods signatures.
2022-08-22 14:17:59 +01:00
reivilibre
c3516e9dec
Faster room joins: make /joined_members block whilst the room is partial stated. (#13514) 2022-08-16 13:16:56 +01:00
Shay
a648a06d52
Add some tracing spans to give insight into local joins (#13439) 2022-08-03 10:19:34 -07:00
andrew do
78a3111c41
Return 404 or member list when getting joined_members after leaving (#13374)
Signed-off-by: Andrew Doh <andrewddo@gmail.com>
Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com>
Co-authored-by: Andrew Morgan <andrewm@element.io>
Co-authored-by: Brendan Abolivier <babolivier@matrix.org>
2022-08-03 14:26:31 +02:00
Will Hunt
502f075e96
Implement MSC3848: Introduce errcodes for specific event sending failures (#13343)
Implements MSC3848
2022-07-27 13:44:40 +01:00
Sean Quah
335ebb21cc
Faster room joins: avoid blocking when pulling events with missing prevs (#13355)
Avoid blocking on full state in `_resolve_state_at_missing_prevs` and
return a new flag indicating whether the resolved state is partial.
Thread that flag around so that it makes it into the event context.

Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-07-26 12:39:23 +01:00
David Robertson
b977867358
Rate limit joins per-room (#13276) 2022-07-19 11:45:17 +00:00
Erik Johnston
cf5fa5063d
Don't pull out full state when sending dummy events (#13310) 2022-07-18 14:19:11 +01:00
Erik Johnston
0ca4172b5d
Don't pull out state in compute_event_context for unconflicted state (#13267) 2022-07-14 13:57:02 +00:00
Sean Quah
68db233f0c
Handle race between persisting an event and un-partial stating a room (#13100)
Whenever we want to persist an event, we first compute an event context,
which includes the state at the event and a flag indicating whether the
state is partial. After a lot of processing, we finally try to store the
event in the database, which can fail for partial state events when the
containing room has been un-partial stated in the meantime.

We detect the race as a foreign key constraint failure in the data store
layer and turn it into a special `PartialStateConflictError` exception,
which makes its way up to the method in which we computed the event
context.

To make things difficult, the exception needs to cross a replication
request: `/fed_send_events` for events coming over federation and
`/send_event` for events from clients. We transport the
`PartialStateConflictError` as a `409 Conflict` over replication and
turn `409`s back into `PartialStateConflictError`s on the worker making
the request.

All client events go through
`EventCreationHandler.handle_new_client_event`, which is called in
*a lot* of places. Instead of trying to update all the code which
creates client events, we turn the `PartialStateConflictError` into a
`429 Too Many Requests` in
`EventCreationHandler.handle_new_client_event` and hope that clients
take it as a hint to retry their request.

On the federation event side, there are 7 places which compute event
contexts. 4 of them use outlier event contexts:
`FederationEventHandler._auth_and_persist_outliers_inner`,
`FederationHandler.do_knock`, `FederationHandler.on_invite_request` and
`FederationHandler.do_remotely_reject_invite`. These events won't have
the partial state flag, so we do not need to do anything for then.

The remaining 3 paths which create events are
`FederationEventHandler.process_remote_join`,
`FederationEventHandler.on_send_membership_event` and
`FederationEventHandler._process_received_pdu`.

We can't experience the race in `process_remote_join`, unless we're
handling an additional join into a partial state room, which currently
blocks, so we make no attempt to handle it correctly.

`on_send_membership_event` is only called by
`FederationServer._on_send_membership_event`, so we catch the
`PartialStateConflictError` there and retry just once.

`_process_received_pdu` is called by `on_receive_pdu` for incoming
events and `_process_pulled_event` for backfill. The latter should never
try to persist partial state events, so we ignore it. We catch the
`PartialStateConflictError` in `on_receive_pdu` and retry just once.

Refering to the graph of code paths in
https://github.com/matrix-org/synapse/issues/12988#issuecomment-1156857648
may make the above make more sense.

Signed-off-by: Sean Quah <seanq@matrix.org>
2022-07-05 16:12:52 +01:00
Shay
046a6513bc
Don't process /send requests for users who have hit their ratelimit (#13134) 2022-06-30 09:22:40 -07:00
Quentin Gliech
92103cb2c8
Decouple synapse.api.auth_blocking.AuthBlocking from synapse.api.auth.Auth. (#13021) 2022-06-14 09:51:15 +01:00
David Teller
a164a46038
Uniformize spam-checker API, part 4: port other spam-checker callbacks to return Union[Allow, Codes]. (#12857)
Co-authored-by: Brendan Abolivier <babolivier@matrix.org>
2022-06-13 18:16:16 +00:00
Richard van der Hoff
f68b5e5773 Merge branch 'rav/simplify_event_auth_interface' into develop 2022-06-13 11:34:59 +01:00
Richard van der Hoff
c1b28b8842 Remove redundant room_version param from check_auth_rules_from_context
It's now implied by the room_version property on the event.
2022-06-12 23:13:10 +01:00
Richard van der Hoff
68be42f6b6 Remove room_version param from validate_event_for_room_version
Instead, use the `room_version` property of the event we're validating.

The `room_version` was originally added as a parameter somewhere around #4482,
but really it's been redundant since #6875 added a `room_version` field to `EventBase`.
2022-06-12 23:13:09 +01:00
Richard van der Hoff
7c6b2204d1
Faster joins: add issue links to the TODOs (#13004)
... to help us keep track of these things
2022-06-09 10:13:03 +00:00
Erik Johnston
e3163e2e11
Reduce the amount of state we pull from the DB (#12811) 2022-06-06 09:24:12 +01:00
Erik Johnston
888a29f412
Wait for lazy join to complete when getting current state (#12872) 2022-06-01 16:02:53 +01:00
Richard van der Hoff
79dadf7216
Fix 404 on /sync when the last event is a redaction of an unknown/purged event (#12905)
Currently, we try to pull the event corresponding to a sync token from the database. However, when
we fetch redaction events, we check the target of that redaction (because we aren't allowed to send
redactions to clients without validating them). So, if the sync token points to a redaction of an event
that we don't have, we have a problem.

It turns out we don't really need that event, and can just work with its ID and metadata, which
sidesteps the whole problem.
2022-06-01 11:29:51 +00:00
Erik Johnston
3594f6c1f3 Merge branch 'master' into develop 2022-05-31 14:48:22 +01:00
Erik Johnston
1e453053cb
Rename storage classes (#12913) 2022-05-31 12:17:50 +00:00
Brendan Abolivier
8fd87739bf
Fix import in module_api module and docs on the new check_event_for_spam signature (#12918)
Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>
2022-05-31 12:04:53 +02:00
David Teller
af7db19e1e
Uniformize spam-checker API, part 3: Expand check_event_for_spam with the ability to return additional fields (#12846)
Signed-off-by: David Teller <davidt@element.io>
2022-05-30 18:24:56 +02:00
Erik Johnston
b83bc5fab5
Pull out less state when handling gaps mk2 (#12852) 2022-05-26 09:48:12 +00:00
Erik Johnston
4660d9fdcf
Fix up state_store naming (#12871) 2022-05-25 12:59:04 +01:00
David Teller
28199e9357
Uniformize spam-checker API, part 2: check_event_for_spam (#12808)
Signed-off-by: David Teller <davidt@element.io>
2022-05-23 17:27:39 +00:00
Shay
71e8afe34d
Update EventContext get_current_event_ids and get_prev_event_ids to accept state filters and update calls where possible (#12791) 2022-05-20 09:54:12 +01:00
Patrick Cloke
86a515ccbf
Consolidate logic for parsing relations. (#12693)
Parse the `m.relates_to` event content field (which describes relations)
in a single place, this is used during:

* Event persistence.
* Validation of the Client-Server API.
* Fetching bundled aggregations.
* Processing of push rules.

Each of these separately implement the logic and each made slightly
different assumptions about what was valid. Some had minor / potential
bugs.
2022-05-16 12:42:45 +00:00
Patrick Cloke
a4c75918b3
Remove unneeded ActionGenerator class. (#12691)
It simply passes through to `BulkPushRuleEvaluator`, which can be
called directly instead.
2022-05-11 07:15:21 -04:00
Erik Johnston
c72d26c1e1
Refactor EventContext (#12689)
Refactor how the `EventContext` class works, with the intention of reducing the amount of state we fetch from the DB during event processing.

The idea here is to get rid of the cached `current_state_ids` and `prev_state_ids` that live in the `EventContext`, and instead defer straight to the database (and its caching). 

One change that may have a noticeable effect is that we now no longer prefill the `get_current_state_ids` cache on a state change. However, that query is relatively light, since its just a case of reading a table from the DB (unlike fetching state at an event which is more heavyweight). For deployments with workers this cache isn't even used.


Part of #12684
2022-05-10 19:43:13 +00:00