To break a potential read/write deadlock, gossipsub uses an unbounded
queue for writes - when peers are too slow to process this queue, it may
end up growing without bounds causing high memory usage.
Here, we introduce a maximum write queue length after which the peer is
disconnected - the queue is generous enough that any "normal" usage
should be fine - writes that are `await`:ed are not affected, only
writes that are launched in an `asyncSpawn` task or similar.
* avoid unnecessary copy of message when there are no send observers
* release message memory earlier in gossipsub
* simplify pubsubpeer logging
When messages can't be sent to peer, we try to establish a send
connection - this causes messages to stack up as more and more unsent
messages are blocked on the dial lock.
* remove dial lock
* run reconnection loop in background task
This change modifies how the backpressure algorithm in bufferstream
works - in particular, instead of working byte-by-byte, it will now work
seq-by-seq.
When data arrives, it usually does so in packets - in the current
bufferstream, the packet is read then split into bytes which are fed one
by one to the bufferstream. On the reading side, the bytes are popped of
the bufferstream, again byte by byte, to satisfy `readOnce` requests -
this introduces a lot of synchronization traffic because the checks for
full buffer and for async event handling must be done for every byte.
In this PR, a queue of length 1 is used instead - this means there will
at most exist one "packet" in `pushTo`, one in the queue and one in the
slush buffer that is used to store incomplete reads.
* avoid byte-by-byte copy to buffer, with synchronization in-between
* reuse AsyncQueue synchronization logic instead of rolling own
* avoid writeHandler callback - implement `write` method instead
* simplify EOF signalling by only setting EOF flag in queue reader (and
reset)
* remove BufferStream pipes (unused)
* fixes drainBuffer deadlock when drain is called from within read loop
and thus blocks draining
* fix lpchannel init order
* mcache fixes
* remove timed cache - the window shifting already removes old messages
* ref -> object
* avoid unnecessary allocations with `[]` operator
* simplify init
* fix several gossipsub/floodsub issues
* floodsub, gossipsub: don't rebroadcast messages that fail validation
(!)
* floodsub, gossipsub: don't crash when unsubscribing from unknown
topics (!)
* gossipsub: don't send message to peers that are not interested in the
topic, when messages don't share topic list
* floodsub: don't repeat all messages for each message when
rebroadcasting
* floodsub: allow sending empty data
* floodsub: fix inefficient unsubscribe
* sync floodsub/gossipsub logging
* gossipsub: include incoming messages in mcache (!)
* gossipsub: don't rebroadcast already-seen messages (!)
* pubsubpeer: remove incoming/outgoing seen caches - these are already
handled in gossipsub, floodsub and will cause trouble when peers try to
resubscribe / regraft topics (because control messages will have same
digest)
* timedcache: reimplement without timers (fixes timer leaks and extreme
inefficiency due to per-message closures, futures etc)
* timedcache: ref -> obj
* remove send lock
When mplex receives data it will block until a reader has processed the
data. Thus, when a large message is received, such as a gossipsub
subscription table, all of mplex will be blocked until all reading is
finished.
However, if at the same time a `dial` to establish a gossipsub send
connection is ongoing, that `dial` will be blocked because mplex is no
longer reading data - specifically, it might indeed be the connection
that's processing the previous data that is waiting for a send
connection.
There are other problems with the current code:
* If an exception is raised, it is not necessarily raised for the same
connection as `p.sendConn`, so resetting `p.sendConn` in the exception
handling is wrong
* `p.isConnected` is checked before taking the lock - thus, if it
returns false, a new dial will be started. If a new task enters `send`
before dial is finished, it will also determine `p.isConnected` is
false, then get stuck on the lock - when the previous task finishes and
releases the lock, the new task will _also_ dial and thus reset
`p.sendConn` causing a leak.
* prefer existing connection
simplifies flow
* move pubsub of off switch, pass switch into pubsub
* use join on lpstreams
* properly cleanup up failed peers
* fix tests
* fix peertable hasPeerId
* fix tests
* rework sending, remove helpers from pubsubpeer, unify in broadcast
* further split broadcast into send
* use send where appropriate
* use formatIt
* improve trace
Co-authored-by: Giovanni Petrantoni <giovanni@fragcolor.xyz>
Every Chronicles log record has an existing `msg` property matching
the static string supplied in the log statement. Thus, it's currently
not possible to use `msg` as the name of a user property:
https://github.com/status-im/nim-chronicles/issues/86
* add finegrained timeouts to pubsub
* use 10 millis timeout in tests
* finalization
* revert timeouts
* use `atEof` for reads
* adjust timeouts and use atEof for reads
* use atEof for reads
* set isEof flag
* no backoff for pubsub streams
* temp timer increase, make macos finalize
* don't call `subscribePeer` in libp2p anymore
* more traces
* leak tests
* lower timeouts
* handle exceptions in control message
* don't use `cancelAndWait`
* handle exceptions in helpers
* wip
* don't send empty messages
* check for leaks properly
* don't use cancelAndWait
* don't await subscribption sends
* remove subscrivePeer calls from switch
* trying without the hooks again
* Fix gossip messages seqno according to spec
* Add peers back to gossipsub table, slow down heartbeat
* Revert "Add peers back to gossipsub table, slow down heartbeat"
This reverts commit 01e2e62172.
* make seqno a threadvar, remove from peerinfo
* seqno refactor, into pubsub
* gossipsub is a function of subscription messages only
* graft/prune work with mesh, get filled up from gossipsub
* fix race conditions with await
* fix exception unsafety when grafting/pruning
* fix allowing up to DHi peers in mesh on incoming graft
* fix metrics in several places
* Peer resultification and defect only
* Fixing some tests
* test fixes
* Rename peer into peerid
* better result error message in identify
* further merge fixes
* don't send public key in message when not signing (information leak)
* don't run rebalance if there are peers in gossip (see #242)
* don't crash randomly on bad peer id from remote
* count published messages
* don't call `switch.dial` in `subscribeToPeer`
* add secureconn constructor
* close in the correct order
* concurent dial lock and track in/out conns better
* make tests pass
* add todo comment
* disconect peers that open too many connections
* wip
* do connection and muxer tracking in one place
* prevent nil pointer in observers
* drop connections when peers is over max
* prevent channel leaks
* don't use closure to handle channel
* count published messages
* don't call `switch.dial` in `subscribeToPeer`
* don't use delegation in connection
* move connection out to own file
* don't breakout on reset
* make sure to call close on secured conn
* add lpstream tracing
* don't breackdown by conn id
* fix import
* remove unused lable
* reset connection on exception
* add additional metrics for skipped messages
* check for nil in secure.close
* Start adding some metrics to pubsub
In order to visualize it's functionality
Still WIP
* more metrics
* add per topic metrics
* finishup with requested metrics
* add a metrisServer define to start local server
* PR fixes and cleanup
* call write until all is written out
* add comments to lpchannel fields
* add an eof flag to signal which end closed
* wip: rework with proper half-closed
* add eof and closed handling
* propagate closes to piped
* call parent close
* moving bufferstream trackers out
* move writeLock to bufferstream
* move writeLock out
* remove unused call
* wip
* rebasing master
* fix mplex tests
* wip
* fix bufferstream after backport
* wip
* rename to differentiate from chronos tracker
* close connection on chronos close
* make reset request asyncCheck
* fix channel cleanup
* misc
* don't use read
* fix backports
* make noise work again
* proper exception handling
* don't reraise just yet
* add convenience templates
* dont double wrap
* use async pragma
* fixes after backporting
* muxer owns connection
* remove on transport close cleanup
* revert back allread
* adding some todos
* read from stream
* inc count before closing
* rebasing master
* rebase master
* use correct exception type
* use try/finally insted of defer
* fix compile in trace mode
* reset channels on mplex close
Jacek Sieka
Giovanni Petrantoni
Eugene Kabanov
Zahary Karadjov
Dmitriy Ryajov
Ștefan Talpalaru
Viktor Kirilov
Giovanni Petrantoni