* streamline socket read/write hot path
This avoids some unnecessary memory copying on the hot path of noise /
mplex, as well as getting rid of a few futures - profiling shows that
this is one of the main culprits of small memory allocations, which
makes sense - this is where gossip fan-out happens.
* fewer futures (and corresponding closures) when sending lpchannel
messages
* avoid data copies when encrypting and framing noise messages
* avoid copying tuple when reading noise data (poor c codegen)
* fix setting eof flag in secure read
* write noise frames in one go
...and closing secure socket once is enough
* peer hooks -> events
* peerinfo -> peerid
* include connection direction in event
* check connection status after event
* lock connmanager lookup also when dialling peer
* clean up un-upgraded connection when upgrade fails
* await peer eventing
* remove join/lifetime future from peerinfo
Peerinfo instances are not unique per peer so the lifetime future is
misleading - it fires when a random connection is closed, not the "last"
one
* document switch values
* naming
* peerevent->conneevent
* prefer PeerID in switch api
This avoids ref issues like ref identity and nil
* use existing peerinfo instance if possible
* remove secureCodec
there may be multiple connections per peerinfo with different codecs
* avoid some extra async::
* Minprotobuf initial commit
* Fix noise.
* Add signed integers support.
Add checks for field number value.
Remove some casts.
* Fix compile errors.
* Fix comments and constants.
* reuse single RNG instance for all crypto key generation
* use foolproof rng
* initRng -> newRng (because it's ref)
* fix test
* imports/exports, chat fix
* fix rsa
* imports and exports
* work around threadvar issue
* fixup
* mac workaround test
* Peer resultification and defect only
* Fixing some tests
* test fixes
* Rename peer into peerid
* better result error message in identify
* further merge fixes
* consolidate reading in lpstream
* remove debug echo
* throw if not enough bytes where read
* tune log level
* set eof flag
* test readExactly to fail on not enough bytes
* 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
* Remove noise padding payload (spec removed it)
* add log scope in secure
* avoid defect array out of range in switch secure when "na"
* improve identify traces
* wip noise fixes
* noise protobuf adjustments (trying)
* add more debugging messages/traces, improve their actual contents
* re-enable ID check in noise
* bump go daemon tag version
* bump go daemon tag version
* enable noise in daemonapi
* interop testing, (both secio and noise will be tested)
* azure cache bump (p2pd)
* CI changes
- Travis: use Go 1.14
- azure-pipelines.yml: big cleanup
- Azure: bump cache keys
- build 64-bit p2pd on 32-bit Windows
- install both Mingw-w64 architectures
* noise logging fixes
* alternate testing between noise and secio
* increase timeout to avoid VM errors in CI (multistream tests)
* refactor heartbeat management in gossipsub
* remove locking within heartbeat
* refactor heartbeat management in gossipsub
* remove locking within heartbeat
Co-authored-by: Ștefan Talpalaru <stefantalpalaru@yahoo.com>
* 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
* handle a few exceptions
Some of these are maybe too aggressive, but in return, they'll log
their exception - more refactoring needed to sort this out - right now
we get crashes on unhandled exceptions of unknown origin
* during connection setup
* while closing channels
* while processing pubsubs
* catch exceptions that are raised and don't try to catch exceptions that are not raised
* propagate cancellederror
* one more
* more
* more
* make interop tests less fragile
* Raise expiration time in gossipsub fanout test for slow CI
Co-authored-by: Dmitriy Ryajov <dryajov@gmail.com>
Co-authored-by: Giovanni Petrantoni <giovanni@fragcolor.xyz>
* Add chronos trackers and used them to sanitize resource disposal
* Chronos trackers for transport tests wip
* No more chronos leaks in testtransport
* Make tcp transport and test more robust when closing
* Test async leaking tracking wip
* Fix a regression in wire connect
* Add chronos trackers to more tests and sanitize resource closure
* Wip fixing floodsub tests
* Floodsub wip
* Made floodsub basically deterministic, hit a nim bug with captures tho
* Wrap up floodsub tests refactor
* Wrapping up
* Add allFuturesThrowing utility
* Fix missing allFuturesThrowing in noise tests!
* Make tests green
* attempt fixing gossipsub failing cases
* Make sure to check also fanout in waitSub
* More verbose traces
* Gossipsub test improvments
* Refactor TcpTransport remove asyncCheck
* Add Connection trackers
* Add stricter connection tracking, wip mplex fix
* More asynccheck removal, in order to avoid connection leaks
* bump chronicles requirement
* Enable tracker dump to check CI output
* Wait for more futures in testmplex
* Remove tracker dump messages
* add tryAndWarn utility, fix mplex issue with go interop
* All allFuturesThrowing to directchat too
* make sure to cleanup on transport close
* Make traces less verbose with shortHexDump utility
* Rename shortHexDump into shortLog
* Improve shortLog, add shortLog for crypto keys
* Add proper shortLog implementations in messages
* Start ChaCha20Poly1305 integration (BearSSL)
* Add Curve25519 (BearSSL) required operations for noise
* Fix curve mulgen iterate/derive
* Fix misleading header
* Add chachapoly proper test
* Curve25519 integration tests (failing, something is wrong)
* Add few converters, finish c25519 integration tests
* Remove implicit converters
* removed internal globals
* Start noise implementation
* Fix public() using proper bear mulgen
* Noise protocol WIP
* Noise progress
* Add a quick nim version of HKDF
* Converted hkdf to iterator, useful for noise
* Noise protocol implementation progress
* Noise progress
* XX handshake almost there
* noise progress
* Noise, testing handshake with test vectors
* Noise handshake progress, still wrong somewhere!
* Noise handshake success!
* Full verified noise XX handshake completed
* Fix and rewrite test to be similar to switch one
* Start with connection upgrade
* Switch chachapoly to CT implementations
* Improve HKDF implementation
* Use a type insted of tuple for HandshakeResult
* Remove unnecessary Let
* More cosmetic fixes
* Properly check randomBytes result
* Fix chachapoly signature
* Noise full circle (altho dispatcher is nil cursed)
* Allow nil aads in chachapoly routines
* Noise implementation up to running full test
* Use bearssl HKDF as well
* Directly use bearssl rng for curve25519 keys
* Add a (disabled/no CI) noise interop test server
* WIP on fixing interop issues
* More fixes in noise implementation for interop
* bump chronos requirement (nimble)
* Add a chachapoly test for very small size payloads
* Noise, more tracing
* Add 2 properly working noise tests
* Fix payload packing, following the spec properly (and not go version but
rather rust)
* Sanity, replace discard with asyncCheck
* Small fixes and optimization
* Use stew endian2 rather then system endian module
* Update nimble deps (chronos)
* Minor cosmetic/code sanity fixes
* Noise, handle Nonce max
* Noise tests, make sure to close secured conns
* More polish, improve code readability too
* More polish and testing again which test fails
* Further polishing
* Restore noise tests
* Remove useless Future[void]
* Remove useless CipherState initializer
* add a proper read wait future in second noise test
* Remove noise generic secure implementation for now
* Few fixes to run eth2 sim
* Add more debug info in noise traces
* Merge size + payload write in sendEncryptedMessage
* Revert secure interface, add outgoing property directly in newNoise
* remove sendEncrypted and receiveEncrypted
* Use openarray in chachapoly and curve25519 helpers
Dmitriy Ryajov
Jacek Sieka
Eugene Kabanov
Giovanni Petrantoni
Viktor Kirilov
Giovanni Petrantoni