The behavior appears to match monero core. monero core isn't
throwing an exception in the linked code, it's returning
boost::none (and logging an error) which is the same functional
behavior as finding that the output does not belong to the user.
* Use only the first input ring length for all RCT input signatures.
This is what Monero does:
ac02af9286/src/ringct/rctTypes.h (L422)https://github.com/monero-project/monero/blob/master/src/cryptonote_basic/cryptonote_basic.h#L308-L309
This isn't an issue for current transactions as from hf 12 Monero requires
all inputs to have the same number of decoys but for transactions before
that Monero would reject RCT txs with differing ring lengths. Monero would
deserialize each inputs signature using the ring length of the first so the
signatures for inputs other than the first would have a different
(wrong) number of elements for that input meaning the signature is invalid.
But as we are using the ring length of each input, which arguably is the
*correct* way, we would approve of transactions with inputs differing in
ring lengths.
* Check that there is more than one ring member for MLSAG signatures.
ac02af9286/src/ringct/rctSigs.cpp (L462)
* monero: require seed lang when decoding seed
- Require the seed language when decoding a Classic|Polyseed seed string
- As per https://github.com/monero-project/monero/issues/9089 and https://github.com/tevador/polyseed/issues/11
- Fixes#478
- Implementation note: I reused the `SeedType` enum and required it as a param to `Seed::from_string` because it seemed simplest, but perhaps there is a cleaner way to require the seed lang.
- Made sure the print statements from #487 print the seed as early as possible to help debug future issues
- A future PR could support deducing which languages a seed decodes to in order to support the UX @kayabaNerve suggested in https://github.com/monero-project/monero/issues/9089:
- "Wallets can also try to abstract [language specification], by decoding with all languages, and only asking the user if/when multiple valid options show up ("Is this seed Spanish or Italian?")."
* Lint
* Move logic for evaluating if a cosign should occur to its own file
Cleans it up and makes it more robust.
* Have expected_next_batch return an error instead of retrying
While convenient to offer an error-free implementation, it potentially caused
very long lived lock acquisitions in handle_processor_message.
* Unify and clean DkgConfirmer and DkgRemoval
Does so via adding a new file for the common code, SigningProtocol.
Modifies from_cache to return the preprocess with the machine, as there's no
reason not to. Also removes an unused Result around the type.
Clarifies the security around deterministic nonces, removing them for
saved-to-disk cached preprocesses. The cached preprocesses are encrypted as the
DB is not a proper secret store.
Moves arguments always present in the protocol from function arguments into the
struct itself.
Removes the horribly ugly code in DkgRemoval, fixing multiple issues present
with it which would cause it to fail on use.
* Set SeraiBlockNumber in cosign.rs as it's used by the cosigning protocol
* Remove unnecessary Clone from lambdas in coordinator
* Remove the EventDb from Tributary scanner
We used per-Transaction DB TXNs so on error, we don't have to rescan the entire
block yet only the rest of it. We prevented scanning multiple transactions by
tracking which we already had.
This is over-engineered and not worth it.
* Implement borsh for HasEvents, removing the manual encoding
* Merge DkgConfirmer and DkgRemoval into signing_protocol.rs
Fixes a bug in DkgConfirmer which would cause it to improperly handle indexes
if any validator had multiple key shares.
* Strictly type DataSpecification's Label
* Correct threshold_i_map_to_keys_and_musig_i_map
It didn't include the participant's own index and accordingly was offset.
* Create TributaryBlockHandler
This struct contains all variables prior passed to handle_block and stops them
from being passed around again and again.
This also ensures fatal_slash is only called while handling a block, as needed
as it expects to operate under perfect consensus.
* Inline accumulate, store confirmation nonces with shares
Inlining accumulate makes sense due to the amount of data accumulate needed to
be passed.
Storing confirmation nonces with shares ensures that both are available or
neither. Prior, one could be yet the other may not have been (requiring an
assert in runtime to ensure we didn't bungle it somehow).
* Create helper functions for handling DkgRemoval/SubstrateSign/Sign Tributary TXs
* Move Label into SignData
All of our transactions which use SignData end up with the same common usage
pattern for Label, justifying this.
Removes 3 transactions, explicitly de-duplicating their handlers.
* Remove CurrentlyCompletingKeyPair for the non-contextual DkgKeyPair
* Remove the manual read/write for TributarySpec for borsh
This struct doesn't have any optimizations booned by the manual impl. Using
borsh reduces our scope.
* Use temporary variables to further minimize LoC in tributary handler
* Remove usage of tuples for non-trivial Tributary transactions
* Remove serde from dkg
serde could be used to deserialize intenrally inconsistent objects which could
lead to panics or faults.
The BorshDeserialize derives have been replaced with a manual implementation
which won't produce inconsistent objects.
* Abstract Future generics using new trait definitions in coordinator
* Move published_signed_transaction to tributary/mod.rs to reduce the size of main.rs
* Split coordinator/src/tributary/mod.rs into spec.rs and transaction.rs
* Make it clear not providing a change address is fingerprintable
When no change address is provided, all change is shunted to the
fee. This PR makes it clear to the caller that it is fingerprintable
when the caller does this.
* Review comments
They're a bit more binding, smaller, provided by the Rust bitcoin library,
sane, and we don't have to worry about malleability since all of our inputs are
SegWit.
* Use debug builds in our Dockerfiles to reduce CI times
Also enables only spawning the mdns service when debug in the coordinator.
* Correct underflow in processor
Prior undetected due to relase builds not having bounds checks enabled.
* Restore Serai release due to CI/RPC failures caused by compiling it in debug mode
This is *probably* worth an issue filed upstream, if it can be tracked down.
* Correct failing debug asserts in Monero
These debug asserts assumed there was a change address to take the remainder.
If there's no change address, the remainder is shunted to the fee, causing the
fee to be distinct from the estimate.
We presumably need to modify monero-serai such that change: None isn't valid,
and users must use Change::Fingerprintable(None).
* Add v1 ring sig verifying
* allow calculating signature hash for v1 txs
* add unreduced scalar type with recovery
I have added this type for borromen sigs, the ee field can be a normal
scalar as in the verify function the ee
field is checked against a reduced scalar mean for it to verify as
correct ee must be reduced
* change block major/ minor versions to u8
this matches Monero
I have also changed a couple varint functions to accept the `VarInt`
trait
* expose `serialize_hashable` on `Block`
* add back MLSAG verifying functions
I still need to revert the commit removing support for >1 input MLSAG FULL
This adds a new rct type to separate Full and simple rct
* add back support for multiple inputs for RCT FULL
* comment `non_adjacent_form` function
also added `#[allow(clippy::needless_range_loop)]` around a loop as without a re-write satisfying clippy without it will make the function worse.
* Improve Mlsag verifying API
* fix rebase errors
* revert the changes on `reserialize_chain`
plus other misc changes
* fix no-std
* Reduce the amount of rpc calls needed for `get_block_by_number`.
This function was causing me problems, every now and then a node would return a block with a different number than requested.
* change `serialize_hashable` to give the POW hashing blob.
Monero calculates the POW hash and the block hash using *slightly* different blobs :/
* make ring_signatures public and add length check when verifying.
* Misc improvements and bug fixes
---------
Co-authored-by: Luke Parker <lukeparker5132@gmail.com>
ethers-solc was used for a type (now manually specified) and to call out to
solc. Since Foundry was already a documented dependency, a call to it now
handles building.
Removing this single crate removes a total of 17 crates from our dependency
tree. While these may still be around due to Foundry, they at least may not
be.
Further work to remove the requirement on Foundry for solc alone would be
appreciated.
I don't like blindly retrying in the Monero library. The amount of errors,
which weren't present with reqwest (well, the error rate was the same, yet due
to a distinct bug this code fixed), demand we do *something* though.
The trace log shows hyper is erroring with 0 bytes of the response read. My
guess is it's somehow a closed connection? A connection pool would detect this
and have created a new connection (as this does, except once finding out
there's an issue).
While we should be able to detect this with `ready()`, we do call ready and it
claims no error. We also can successfully write which makes this... a mess.
Hopefully, it either actually works as intended, yet it at least requires two
consecutive errors which should be much less frequent.
The prior system spawned a new connection per request to enable parallelism,
yet kept hitting hyper::IncompleteMessages I couldn't track down. This
attempts to resolve those by a long-lived socket.
Halves the amount of requests per-authenticated RPC call, and accordingly is
likely still better overall.
I don't believe this is resolved yet but this is still worth pushing.
reqwest was replaced with hyper and hyper-rustls within monero-serai due to
reqwest *solely* offering a connection pool API. In the process, it was
demonstrated how quickly we can achieve equivalent functionality to reqwest for
our use cases with a fraction of the code.
This adds our own reqwest alternative to the tree, applying it to both
bitcoin-serai and message-queue. By doing so, bitcoin-serai decreases its tree
by 21 packages and the processor by 18. Cargo.lock decreases by 8 dependencies,
solely adding simple-request. Notably removed is openssl-sys and openssl.
One noted decrease functionality is the requirement on the system having
installed CA certificates. While we could fallback to the rustls certificates
if the system doesn't have any, that's blocked by
https://github.com/rustls/hyper-rustls/pulls/228.
Removes bitcoin-serai's usage of sha2 for bitcoin-hashes. While sha2 is still
in play due to modular-frost (more specifically, due to ciphersuite), this
offers a bit more performance (assuming equivalency between sha2 and
bitcoin-hashes' impl) due to removing a static for a const.
Makes secp256k1 a dev dependency for bitcoin-serai. While secp256k1 is still
pulled in via bitcoin, it's hopefully slightly better to compile now and makes
usage of secp256k1 an implementation detail of bitcoin (letting it change it
freely).
Also offers slightly more efficient signing as we don't decode to a signature
just to re-encode for the transaction.
Removes a 20s sleep for a check every second, up to 20 times, for reduced test
times in the processor.
It *looks like* hyper will drop the connection once its request sender is
dropped, regardless of if the last request hasn't had its response completed.
This attempts to resolve some spurious connection errors.
If a crate has std set, it should enable std for all dependencies in order to
let them properly select which algorithms to use. Some crates fallback to
slower/worse algorithms on no-std.
Also more aggressively sets default-features = false leading to a *10%*
reduction in the amount of crates coordinator builds.
The lack of locking the connection when making an authenticated request, which
is actually two sequential requests, risked another caller making a request in
between, invalidating the state.
Now, only unauthenticated connections share a connection object.
* Design and document a multisig rotation flow
* Make Scanner::eventualities a HashMap so it's per-key
* Don't drop eventualities, always follow through on them
Technical improvements made along the way.
* Start creating an isolate object to manage multisigs, which doesn't require being a signer
Removes key from SubstrateBlock.
* Move Scanner/Scheduler under multisigs
* Move Batch construction into MultisigManager
* Clarify "should" in Multisig Rotation docs
* Add block_number to MultisigManager, as it controls the scanner
* Move sign_plans into MultisigManager
Removes ThresholdKeys from prepare_send.
* Make SubstrateMutable an alias for MultisigManager
* Rewrite Multisig Rotation
The prior scheme had an exploit possible where funds were sent to the old
multisig, then burnt on Serai to send from the new multisig, locking liquidity
for 6 hours. While a fee could be applied to stragglers, to make this attack
unprofitable, the newly described scheme avoids all this.
* Add mini
mini is a miniature version of Serai, emphasizing Serai's nature as a
collection of independent clocks. The intended use is to identify race
conditions and prove protocols are comprehensive regarding when certain clocks
tick.
This uses loom, a prior candidate for evaluating the processor/coordinator as
free of race conditions (#361).
* Use mini to prove a race condition in the current multisig rotation docs, and prove safety of alternatives
Technically, the prior commit had mini prove the race condition.
The docs currently say the activation block of the new multisig is the block
after the next Batch's. If the two next Batches had already entered the
mempool, prior to set_keys being called, the second next Batch would be
expected to contain the new key's data yet fail to as the key wasn't public
when the Batch was actually created.
The naive solution is to create a Batch, publish it, wait until it's included,
and only then scan the next block. This sets a bound of
`Batch publication time < block time`. Optimistically, we can publish a Batch
in 24s while our shortest block time is 2m. Accordingly, we should be fine with
the naive solution which doesn't take advantage of throughput. #333 may
significantly change latency however and require an algorithm whose throughput
exceeds the rate of blocks created.
In order to re-introduce parallelization, enabling throughput, we need to
define a safe range of blocks to scan without Serai ordering the first one.
mini demonstrates safety of scanning n blocks Serai hasn't acknowledged, so
long as the first is scanned before block n+1 is (shifting the n-block window).
The docs will be updated next, to reflect this.
* Fix Multisig Rotation
I believe this is finally good enough to be final.
1) Fixes the race condition present in the prior document, as demonstrated by
mini.
`Batch`s for block `n` and `n+1`, may have been in the mempool when a
multisig's activation block was set to `n`. This would cause a potentially
distinct `Batch` for `n+1`, despite `n+1` already having a signed `Batch`.
2) Tightens when UIs should use the new multisig to prevent eclipse attacks,
and protection against `Batch` publication delays.
3) Removes liquidity fragmentation by tightening flow/handling of latency.
4) Several clarifications and documentation of reasoning.
5) Correction of "prior multisig" to "all prior multisigs" regarding historical
verification, with explanation why.
* Clarify terminology in mini
Synchronizes it from my original thoughts on potential schema to the design
actually created.
* Remove most of processor's README for a reference to docs/processor
This does drop some misc commentary, though none too beneficial. The section on
scanning, deemed sufficiently beneficial, has been moved to a document and
expanded on.
* Update scanner TODOs in line with new docs
* Correct documentation on Bitcoin::Block::time, and Block::time
* Make the scanner in MultisigManager no longer public
* Always send ConfirmKeyPair, regardless of if in-set
* Cargo.lock changes from a prior commit
* Add a policy document on defining a Canonical Chain
I accidentally committed a version of this with a few headers earlier, and this
is a proper version.
* Competent MultisigManager::new
* Update processor's comments
* Add mini to copied files
* Re-organize Scanner per multisig rotation document
* Add RUST_LOG trace targets to e2e tests
* Have the scanner wait once it gets too far ahead
Also bug fixes.
* Add activation blocks to the scanner
* Split received outputs into existing/new in MultisigManager
* Select the proper scheduler
* Schedule multisig activation as detailed in documentation
* Have the Coordinator assert if multiple `Batch`s occur within a block
While the processor used to have ack_up_to_block, enabling skips in the block
acked, support for this was removed while reworking it for multiple multisigs.
It should happen extremely infrequently.
While it would still be beneficial to have, if multiple `Batch`s could occur
within a block (with the complexity here not being worth adding that ban as a
policy), multiple `Batch`s were blocked for DoS reasons.
* Schedule payments to the proper multisig
* Correct >= to <
* Use the new multisig's key for change on schedule
* Don't report External TXs to prior multisig once deprecated
* Forward from the old multisig to the new one at all opportunities
* Move unfulfilled payments in queue from prior to new multisig
* Create MultisigsDb, splitting it out of MainDb
Drops the call to finish_signing from the Signer. While this will cause endless
re-attempts, the Signer will still consider them completed and drop them,
making this an O(n) cost at boot even if we did nothing from here.
The MultisigManager should call finish_signing once the Scanner completes the
Eventuality.
* Don't check Scanner-emitted completions, trust they are completions
Prevents needing to use async code to mark the completion and creates a
fault-free model. The current model, on fault, would cause a lack of marked
completion in the signer.
* Fix a possible panic in the processor
A shorter-chain reorg could cause this assert to trip. It's fixed by
de-duplicating the data, as the assertion checked consistency. Without the
potential for inconsistency, it's unnecessary.
* Document why an existing TODO isn't valid
* Change when we drop payments for being to the change address
The earlier timing prevents creating Plans solely to the branch address,
causing the payments to be dropped, and the TX to become an effective
aggregation TX.
* Extensively document solutions to Eventualities being potentially created after having already scanned their resolutions
* When closing, drop External/Branch outputs which don't cause progress
* Properly decide if Change outputs should be forward or not when closing
This completes all code needed to make the old multisig have a finite lifetime.
* Commentary on forwarding schemes
* Provide a 1 block window, with liquidity fragmentation risks, due to latency
On Bitcoin, this will be 10 minutes for the relevant Batch to be confirmed. On
Monero, 2 minutes. On Ethereum, ~6 minutes.
Also updates the Multisig Rotation document with the new forwarding plan.
* Implement transaction forwarding from old multisig to new multisig
Identifies a fault where Branch outputs which shouldn't be dropped may be, if
another output fulfills their next step. Locking Branch fulfillment down to
only Branch outputs is not done in this commit, but will be in the next.
* Only let Branch outputs fulfill branches
* Update TODOs
* Move the location of handling signer events to avoid a race condition
* Avoid a deadlock by using a RwLock on a single txn instead of two txns
* Move Batch ID out of the Scanner
* Increase from one block of latency on new keys activation to two
For Monero, this offered just two minutes when our latency to publish a Batch
is around a minute already. This does increase the time our liquidity can be
fragmented by up to 20 minutes (Bitcoin), yet it's a stupid attack only
possible once a week (when we rotate). Prioritizing normal users' transactions
not being subject to forwarding is more important here.
Ideally, we'd not do +2 blocks yet plus `time`, such as +10 minutes, making
this agnostic of the underlying network's block scheduling. This is a
complexity not worth it.
* Split MultisigManager::substrate_block into multiple functions
* Further tweaks to substrate_block
* Acquire a lock on all Scanner operations after calling ack_block
Gives time to call register_eventuality and initiate signing.
* Merge sign_plans into substrate_block
Also ensure the Scanner's lock isn't prematurely released.
* Use a HashMap to pass to-be-forwarded instructions, not the DB
* Successfully determine in ClosingExisting
* Move from 2 blocks of latency when rotating to 10 minutes
Superior as noted in 6d07af92ce10cfd74c17eb3400368b0150eb36d7, now trivial to
implement thanks to prior commit.
* Add note justifying measuring time in blocks when rotating
* Implement delaying of outputs received early to the new multisig per specification
* Documentation on why Branch outputs don't have the race condition concerns Change do
Also ensures 6 hours is at least N::CONFIRMATIONS, for sanity purposes.
* Remove TODO re: sanity checking Eventualities
We sanity check the Plan the Eventuality is derived from, and the Eventuality
is handled moments later (in the same file, with a clear call path). There's no
reason to add such APIs to Eventualities for a sanity check given that.
* Add TODO(now) for TODOs which must be done in this branch
Also deprecates a pair of TODOs to TODO2, and accepts the flow of the Signer
having the Eventuality.
* Correct errors in potential/future flow descriptions
* Accept having a single Plan Vec
Per the following code consuming it, there's no benefit to bifurcating it by
key.
* Only issue sign_transaction on boot for the proper signer
* Only set keys when participating in their construction
* Misc progress
Only send SubstrateBlockAck when we have a signer, as it's only used to tell
the Tributary of what Plans are being signed in response to this block.
Only immediately sets substrate_signer if session is 0.
On boot, doesn't panic if we don't have an active key (as we wouldn't if only
joining the next multisig). Continues.
* Correctly detect and set retirement block
Modifies the retirement block from first block meeting requirements to block
CONFIRMATIONS after.
Adds an ack flow to the Scanner's Confirmed event and Block event to accomplish
this, which may deadlock at this time (will be fixed shortly).
Removes an invalid await (after a point declared unsafe to use await) from
MultisigsManager::next_event.
* Remove deadlock in multisig_completed and document alternative
The alternative is simpler, albeit less efficient. There's no reason to adopt
it now, yet perhaps if it benefits modeling?
* Handle the final step of retirement, dropping the old key and setting new to existing
* Remove TODO about emitting a Block on every step
If we emit on NewAsChange, we lose the purpose of the NewAsChange period.
The only concern is if we reach ClosingExisting, and nothing has happened, then
all coins will still be in the old multisig until something finally does. This
isn't a problem worth solving, as it's latency under exceptional dead time.
* Add TODO about potentially not emitting a Block event for the reitrement block
* Restore accidentally deleted CI file
* Pair of slight tweaks
* Add missing if statement
* Disable an assertion when testing
One of the test flows currently abuses the Scanner in a way triggering it.
Eventualities need to be binding not just to a plan, yet to the execution of
the plan (the outputs). Bitcoin's Eventuality definition short-cutted this
under a honest multisig assumption, causing the following issue:
If multisig n+1 is verifying multisig n's actions, as detailed in
multi-multisig's document on multisig rotation, it'll check no outstanding
eventualities exist. If we solely bind to the plan, a malicious multisig n
could steal outbound payments yet cause the plan to be marked as successfully
completed.
By modifying the eventuality to also include the expected outputs, this is no
longer possible. Binding to the expected input is preserved in order to remain
binding to the plan (allowing two plans with the same output-set to co-exist).
Fixes where ram_scanned is updated in processor. The prior version, while safe,
would redo massive amounts of work during periods of inactivity. It also hit an
undocumented invariant where get_eventuality_completions assumes new blocks,
yet redone work wouldn't have new blocks.
Modifies Monero's generate_blocks to return the hashes of the generated blocks.
* Add in an implementation of BP+ based off the paper, intended for clarity and review
This was done as part of my work on FCMPs from Monero, and is copied from https://github.com/kayabaNerve/full-chain-membership-proofs
* Remove crate structure of BP+
* Remove arithmetic circuit code
* Remove AC/VC generators code
* Remove generator transcript
Monero uses non-transcripted static generators.
* Further trimming of generators
* Remove the single range proof
It's unused by Monero and accordingly unhelpful.
* Work on getting BP+ to compile in its new env
* Correct BP+ folder name
* Further tweaks to get closer to compiling
* Remove the ScalarMatrix file
It's only used for AC proofs
* Compiles, with tests passing
* Lock BP+ to Ed25519 instead of the generic Ciphersuite
* Resolve most warnings in BP+
* Make existing bulletproofs test easier to read
* Further strip generators
* Swap G/H as Monero did
* Replace RangeCommitment with Commitment
* Hard-code BP+ h to Ed25519's generator
* Use pub(crate) for BP+, not pub
* Replace initial_transcript with hash_plus
* Rename hash_plus to initial_transcript
* Finish integrating the FCMP BP+ impl
* Move BP+ folder
* Correct no-std support
* Rename "long_n" to eta
* Add note on non-prime order dfg points
Arguably not meaningful, as it adds the scanner yet not the RPC, and no signing
code since modular-frost doesn't support no-std yet. It's a step in the right
direction though.
Only recently (I believe the most recent HF) were output keys checked to be
valid. This means returned keys may be invalid points, despite being the
legitimate keys for the specified outputs.
Does still label the node as invalid if it doesn't return 32 bytes,
hex-encoded.
Allows running `cargo build` in monero-serai and message-queue without
erroring, since it'd automatically try to build the binaries which require
additional features.
While we could make those features not optional, it'd increase time to build
and disk space required, which is why the features exist for monero-serai and
message-queue in the first place (since both are frequently used as libs).
This is technically over-agressive, as a dropped output will reduce the fee,
yet this edge case is so minor the flow for it to not be over-aggressive (over
a few fractions of a cent) is by no means worth it.
Fixes the crash causable by the WIP send_test.
All uses were safe due to addresses being converted to script_pubkeys which
don't embed their network. The only risk of there being an issue is if a
future address spec did embed the net ID into the script_pubkey and that was
moved to.
This resolves the audit note and does offer that tightening.
* add polyseed support
* fix pr comments
* fix tests
* Embed the mempool into the Blockchain
* Plan scheduled payments whenever outputs are received
The scheduler prior waited for the next series of payments to be added.
* Replace Tendermint step with sync_block
Step moved a step forward after an externally synced/added block. This created
a race condition to add the block between the sync process and the Tendermint
machine. Now that the block routes through Tendermint, there is no such race
condition.
* Finish binding Tendermint into Tributary and define a Tributary master object
* Add correction the last commit missed
* Add DoS limits to tributary and require provided transactions be ordered
* Fix the scheduler from dropping UTXOs when there weren't any payments
* Documentation and cargo update
* Add a dedicated db crate with a basic DB trait
It's needed by the processor and tributary (coordinator).
* Add a DB to Tributary
Adds support for reloading most of the blockchain.
* Reloaded provided transactions from the disk
Also resolves a race condition by asserting provided transactions must be
unique, allowing them to be safely provided multiple times.
* must_use annotations on DbTxn
* Support reloading the mempool from disk
* Add a NewSet event to validator-sets
Updates to the latest serai-dex/substrate due to depending on
10ccaca0eb498a2316bbf627d419b29b1a75933a.
* Add basic getters to tributary
* cargo update
* Update to the latest subxt
Writes a custom unsigned extrinic creator due to subxt having an internal error
with the scale metadata. While the code in our scope increased, it's much more
ergonomic to our usage. We may end up rewriting most of subxt, eventually.
* Make unsigned private due to unsafe calling potential
* Start defining the coordinator
* Merge AckBlock with Burns
Offers greater efficiency while reducing concerns re: atomicity.
* Correct processor flow to have the coordinator decide signing set/re-attempts
The signing set should be the first group to submit preprocesses to Tributary.
Re-attempts shouldn't be once every 30s, yet n blocks since the last relevant
message.
Removes the use of an async task/channel in the signer (and Substrate signer).
Also removes the need to be able to get the time from a coin's block, which was
a fragile system marked with a TODO already.
* cargo +nightly fmt
* cargo update
Since p256 now pulls in an extra crate with this update, the {k,p}256 imports
disable default-features to prevent growing the tree.
* Support extracting timestamps from blocks
* Make progres on handling NewSet events
Further bones out the coordinator.
* Resolve#245
* Have InInstructions track the latest block for a network in storage
* Fill out code for the rest of the Substrate events
* Clean up the Substrate block processing code
* Rename transaction file to tributary, add function for genesis
* Add a processor API to the coordinator
* Add extensive commentary on mutable to the processor's main file
Clearly establishes why consistency is guaranteed from a Rust borrow-checker
mindset. While there are plenty of... 'violations', they're clearly explained.
Hopefully, this method of thinking helps promote/ensure consistency in the
future.
* Move ConfirmKeyPair from key_gen to substrate
Clarifies the emitter and accordingly why its mutations are justified.
* Remove BatchSigned
SubstrateBlock's provision of the most recently acknowledged block has
equivalent information with the same latency. Accordingly, there's no need for
it.
* Add note to processor_messages
* Use a single txn for an entire coordinator message
Removes direct DB accesses whre possible. Documents the safety of the rest.
Does uncover one case of unsafety not previously noted.
* cargo update to remove usage of yanked crate
* Clarify safety of Scanner::block_number and KeyGen::keys
* Tweak ConfirmKeyPair to alleviate database requirements of coordinator
* Use an enum for Coin/NetworkId
It originally wasn't an enum so software which had yet to update before an
integration wouldn't error (as now enums are strictly typed). The strict typing
is preferable though.
* Code a method to determine the activation block before any block has consensus
[0; 32] is a magic for no block has been set yet due to this being the first
key pair. If [0; 32] is the latest finalized block, the processor determines
an activation block based on timestamps.
This doesn't use an Option for ergonomic reasons.
* automate whitespace & trimming test cases
* Save keys by their tweaked group_key
Keys are referred to by their tweaked versions. If a tweak was needed, keys
would fail to confirm.
* Use crypto-bigint's reduction in ed448
Achieves feasible performance in the ed448 which makes it potentially viable
for real world usage.
Accordingly prepares a new release, updating the README.
* Move the entirety of ed448 to Residue, offering a further 2-4x speedup
* Resolve#68
Notably speeds up monero-serai's build and CLSAG performance.
* Make MainDB into SubstrateDB
* Initial Tributary handling
* Add additional checks to key_gen/sign
There is the ability to cause state bloat by flooding Tributary.
KeyGen/Sign specifically shouldn't allow bloat since we check the
commitments/preprocesses/shares for validity. Accordingly, any invalid data
(such as bloat) should be detected.
It was posssible to place bloat after the valid data. Doing so would be
considered a valid KeyGen/Sign message, yet could add up to 50k kB per sign.
* Apply DKG TX handling code to all sign TXs
The existing code was almost entirely applicable. It just needed to be scoped
with an ID. While the handle function is now a bit convoluted, I don't see a
better option.
* Split FinalizedBlock into ExternalBlock and SeraiBlock
Also re-arranges their orders.
* Add support for multiple orderings in Provided
Necessary as our Tributary chains needed to agree when a Serai block has
occurred, and when a Monero block has occurred. Since those could happen at the
same time, some validators may put SeraiBlock before ExternalBlock and vice
versa, causing a chain halt. Now they can have distinct ordering queues.
* Slash on unrecognized ID
* ExternalBlock handler
* Add a SubstrateBlockAck message to the processor
When a Substrate block occurs, the coordinator is expected to emit
SubstrateBlock. This causes the processor to begin a variety of plans. The
processor now emits SubstrateBlockAck, explicitly listing all plan IDs, before
starting signing.
This lets the coordinator provide a SubstrateBlock transaction, and with it,
recognize all plan IDs as valid.
Prior, we would've had to have a spotty algorithm based upon the upcoming
Preprocess messages, or if we immediately provided the SubstrateBlock
transaction, then wait for the processor to inform us of the contained plans.
This creates an explicitly proper async flow not reliant on waiting for data
availability.
Alternatively, we could've replaced Preprocess with (Block, Vec<Preprocess>).
This would've been more efficient, yet also clunky due to the multiple usages
of the Preprocess message.
* Route the SubstrateBlock message, which is the last Tributary transaction type
* Add recent bloat checks added to signer to substrate_signer as well
* Add no_std support to transcript, dalek-ff-group, ed448, ciphersuite, multiexp, schnorr, and monero-generators
transcript, dalek-ff-group, ed449, and ciphersuite are all usable with no_std
alone. The rest additionally require alloc.
Part of #279.
* Add a test to the coordinator for running a Tributary
Impls a LocalP2p for testing.
Moves rebroadcasting into Tendermint, since it's what knows if a message is
fully valid + original.
Removes TributarySpec::validators() HashMap, as its non-determinism caused
different instances to have different round robin schedules. It was already
prior moved to a Vec for this issue, so I'm unsure why this remnant existed.
Also renames the GH no-std workflow from the prior commit.
* Add a test for Tributary
Further fleshes out the Tributary testing code.
* Test handling of DKG commitments transactions
* Add Transaction::sign.
While I don't love the introduction of empty_signed, it's practically fine.
* Tributary test wait_for_tx_inclusion function
* Additionally test DKGShares
* Handle adding new Tributaries
Removes last_block as an argument from Tendermint. It now loads from the DB as
needed. While slightly less performant, it's easiest and should be fine.
* Reload Tributaries
add_active_tributary writes the spec to disk before it returns, so even if the
VecDeque it pushes to isn't popped, the tributary will still be loaded on boot.
* Start handling P2P messages
This defines the tart of a very complex series of locks I'm really unhappy
with. At the same time, there's not immediately a better solution. This also
should work without issue.
* Clarify Arc RwLocks and sleeps in coordinator
* Send a heartbeat message when a Tributary falls behind
* cargo fmt
* cargo update
* Move json word lists to rs
Allows building the seed code without serde_json.
* Break coordinator main into multiple functions
Also moves from std::sync::RwLock to tokio::sync::RwLock to prevent wasting
cycles on spinning.
* Remove reliance on a blockchain read lock from block/commit
* Implement Tributary syncing
Also adds a forwards-lookup to the Tributary blockchain.
* Don't return from sync_block until the Tendermint machine returns if it's valid or not
We had a race condition where'd we be informed of blocks 1 .. 3, and
immediately add 1 .. 3. Because we immediately tried to add 2 after 1, it'd
fail since the tip was still the genesis, yet 2 needs the tip to be 1.
Adding a channel, while ugly, was the simplest way to accomplish this.
Also has any added block be broadcasted. Else there's a race condition where a
node which syncs up to the most recent block does so, yet fails to add the next
block when it's committed to.
* Test handle_p2p and Tributary syncing
Includes bug fixes.
* Tweak tests workflow
* Add a TributaryReader which doesn't require a borrow to operate
Reduces lock contention.
Additionally changes block_key to include the genesis. While not technically
needed, the lack of genesis introduced a side effect where any Tributary on the
the database could return the block of any other Tributary. While that wasn't a
security issue, returning it suggested it was on-chain when it wasn't. This may
have been usable to create issues.
* Document panic in FROST
* Document a pair of panics requiring 256 GB of RAM/4 GB of a context
* Add a UID function to messages
When we receive messages, we're provided with a message ID we can use to
prevent handling an item multiple times. That doesn't prevent us from *sending*
an item multiple times though. Thanks to the UID system, we can now not send if
already present.
Alternatively, we can remove the ordered message ID for just the UID, allowing
duplicates to be sent without issue, and handled on the receiving end.
* Initial code to handle messages from processors
* Document the processor/tributary/coordinator/serai flow
* Have Coordinator MainDb take a mutable borrow
* Update to substrate polkadot-v0.9.42
* Correct error message in ff-group-tests
* Update to May's nightly
Doesn't use the PR due to the needed changes.
* Support arbitrary RPC providers in monero-serai
Sets a clean path for no-std premised RPCs (buffers to an external RPC impl)/
Tor-based RPCs/client-side load balancing/...
* Correct processor's handling of the new Monero RPC code
* Correct Serai Dockerfile
* Publish ExternablBlock/SubstrateBlock, delay *Preprocess until ID acknowledged
Adds a channel for the Tributary scanner to communicate when an ID has been
acknowledged.
* Rename uid to intent
* Use U448 for Ed448 instead of U512
* Spawn a new async task for each block message
This probably should be done with n-long lived tasks, one per Tributary. While
this may not be suitably performant long-term (potential DoS vector), this at
least resolves the halting concerns.
* Move the coordinator to a n-processor design
* Ensure Tributary commits are minimal
* Properly get genesis for a Processor message
* Create a vote transaction upon GeneratedKeyPair
* Remove TODO about code de-duplication
It's infeasible to write a macro/function there. Does add a type alias which
makes things cleaner.
* Have coordinator publish batches to Substrate
* Implement MuSig key aggregation into DKG
Isn't spec compliant due to the lack of a spec to be compliant too.
Slight deviation from the paper by using a unique list instead of a multiset.
Closes#186, progresses #277.
* Correct 2/3rds definitions throughout the codebase
The prior formula failed for some values, such as 20.
20 / 3 = 6, * 2 = 12, + 1 = 13. 13 is 65%, not >= 67.
* cargo update
Resolves a yanked crate and removes some duplicated dependencies.
* Add a dedicated function to get a MuSig key
* Do the minimal amount of work for dkg to compile under no-std
The Substrate runtime requires access to the MuSig key aggregation function.
\#279 related.
* Use a MuSig signature to publish validator set key pairs to Serai
The processor/coordinator flow still has to be rewritten.
* Correct various no_std definitions
* Add a context to MuSig key aggregation
* Use proper messages for ValidatorSets/InInstructions pallet
Provides a DST, and associated metadata as beneficial.
Also utilizes MuSig's context to session-bind. Since set_keys_messages also
binds to set, this is semi-redundant, yet that's appreciated.
* Remove signed Substrate TXs from Coordinator
* Only scan v2 Monero TXs
* Fix for prior commit
* Ensure canonical points in the cross-group DLEq proof
* Fix incorrect sig_hash generation
sig_hash was used as a challenge. challenges should be of the form H(R, A, m).
These sig hashes were solely H(A, m), allowing trivial forgeries.
* cargo update
Resolves an openssl advisory and nets ~-8 crates.
* Build no-std tests with RISC-V 32 IMAC
Turns out wasm still has std, making it suboptimal to use here.
* Pin setup-protoc to v2.0.0
* Update to substrate polkadot-v0.9.43
* fix tributary sync test
* Slight terminology correction in sync test
Also correct a mistake from merging the most recent polkadot version.
* Update nightly
* Replace lazy_static with OnceLock inside monero-serai
lazy_static, if no_std environments were used, effectively required always
using spin locks. This resolves the ergonomics of that while adopting Rust std
code.
no_std does still use a spin based solution. Theoretically, we could use
atomics, yet writing our own Mutex wasn't a priority.
* no-std support for monero-serai (#311)
* Move monero-serai from std to std-shims, where possible
* no-std fixes
* Make the HttpRpc its own feature, thiserror only on std
* Drop monero-rs's epee for a homegrown one
We only need it for a single function. While I tried jeffro's, it didn't work
out of the box, had three unimplemented!s, and is no where near viable for
no_std.
Fixes#182, though should be further tested.
* no-std monero-serai
* Allow base58-monero via git
* cargo fmt
* Represent RCT amounts with None, not 0.
Fixes#282.
Does allow any v1 TXs which exist, and v2 miner-TXs, to specify Some(0). As far
as I can tell, both were/are theoreitcally possible.
* Add a message queue
This is intended to be a reliable transport between the processors and
coordinator. Since it'll be intranet only, it's written as never fail.
Primarily needs testing and a proper ID.
* cargo update
Resolves https://github.com/serai-dex/serai/security/dependabot/29
* Correct deny.toml with inclusion of message-queue
* Update nightly
* std-shims: six `Read` for &[u8]
* Use serai- prefixes on Serai-specific packages
Fixes deny.toml, also runs a minor cargo update shrinking the tree.
* Update monero-tests workflow to new name for the processor
* Correct depends for processor-messages
* Disable Rust caching
We hit the cache limit after just one or two builds, making it infeasible.
* cargo update
Resolves a yanked crate
* Move location of serai-client in Cargo.toml
* Monero: support for legacy transactions (#308)
* add mlsag
* fix last commit
* fix miner v1 txs
* fix non-miner v1 txs
* add borromean + fix mlsag
* add block hash calculations
* fix for the jokester that added unreduced scalars
to the borromean signature of
2368d846e671bf79a1f84c6d3af9f0bfe296f043f50cf17ae5e485384a53707b
* Add Borromean range proof verifying functionality
* Add MLSAG verifying functionality
* fmt & clippy :)
* update MLSAG, ss2_elements will always be 2
* Add MgSig proving
* Tidy block.rs
* Tidy Borromean, fix bugs in last commit, replace todo! with unreachable!
* Mark legacy EcdhInfo amount decryption as experimental
* Correct comments
* Write a new impl of the merkle algorithm
This one tries to be understandable.
* Only pull in things only needed for experimental when experimental
* Stop caching the Monero block hash now in processor that we have Block::hash
* Corrections for recent processor commit
* Use a clearer algorithm for the merkle
Should also be more efficient due to not shifting as often.
* Tidy Mlsag
* Remove verify_rct_* from Mlsag
Both methods were ports from Monero, overtly specific without clear
documentation. They need to be added back in, with documentation, or included
in a node which provides the necessary further context for them to be naturally
understandable.
* Move mlsag/mod.rs to mlsag.rs
This should only be a folder if it has multiple files.
* Replace EcdhInfo terminology
The ECDH encrypted the amount, yet this struct contained the encrypted amount,
not some ECDH.
Also corrects the types on the original EcdhInfo struct.
* Correct handling of commitment masks when scanning
* Route read_array through read_raw_vec
* Misc lint
* Make a proper RctType enum
No longer caches RctType in the RctSignatures as well.
* Replace Vec<Bulletproofs> with Bulletproofs
Monero uses aggregated range proofs, so there's only ever one Bulletproof. This
is enforced with a consensus rule as well, making this safe.
As for why Monero uses a vec, it's probably due to the lack of variadic typing
used. Its effectively an Option for them, yet we don't need an Option since we
do have variadic typing (enums).
* Add necessary checks to Eventuality re: supported protocols
* Fix for block 202612 and fix merkel root calculations
* MLSAG (de)serialisation fix
ss_2_elements will not always be 2 as rct type 1 transactions are not enforced to have one input
* Revert "MLSAG (de)serialisation fix"
This reverts commit 5e710e0c96.
here it checks number of MGs == number of inputs:
0a1eaf26f9/src/cryptonote_core/tx_verification_utils.cpp (L60-59)
and here it checks for RctTypeFull number of MGs == 1:
0a1eaf26f9/src/ringct/rctSigs.cpp (L1325)
so number of inputs == 1
so ss_2_elements == 2
* update `MlsagAggregate` comment
* cargo update
Resolves a yanked crate
* Move location of serai-client in Cargo.toml
---------
Co-authored-by: Luke Parker <lukeparker5132@gmail.com>
* Fix the known issue with the DSA
I wrote it to only select TXs with a timelock, not only TXs which are unlocked.
This most likely explains why it so heavily selected coinbases.
Also moves an InternalError which would've never been hit on mainnet, yet
technically isn't an invariant, to only exist when cfg(test).
* Add a bin to download a chain, over RPC, reserializing and hashing every item
Parallelized. Doesn't check the deserialization is correct. Does use distinct,
persistent HTTP clients.
* Correct how Monero integration tests are run
* Support multiple RPCs in the reserialize_chain bin
* Don't call get_height every block
* Modify get_transactions to split requests as to not hit the restricted RPC limits
* Meaningful changes from aggressive-clippy
I do want to enable a few specific lints, yet aggressive-clippy as a whole
isn't worthwhile.
* Extend reserialize_chain with CLSAG/BP(+) verification
* Remove spammy println from reserialize_chain
* Update reserialize_chain for v1 and migration TXs
Also always marks 0-amount inputs as RCT due to impossibility of non-RCT
0-amount outputs.
* Only deserialize RctSignatures where's there at least one input
This is only enforced by the Monero protocol due to a single check the mixRing
isn't empty in get_pre_mlsag_hash. The value in ensuring there's a least one
input is to ensure the safety of our rct_type functions, which determines the
RctType based off structural analysis (specifically, input data if
MlsagBorromean).
rct_type was technically safe without this. A 0-input transaction would be
mis-classified as RctFull/MlsagAggregate, which would then make the
RctSignatures invalid for being RctFull (requiring exactly one input) yet not
having inputs, meaning an invalid RctSignatures would be mis-classified yet
still invalid.
This just removes the risk of mis-classification in the first place, tightening
the library's safety.
* docs/Getting Started.md: cargo build --release --all-features
* Fix the known instance of #295
* Bind RocksDB into serai-db
* Split up tests in CI to avoid node storage limits
* Corrections to prior commit
* Again
I called git commit --amend without calling git add . again :(
* Update the flow for completed signing processes
Now, an on-chain transaction exists. This resolves some ambiguities and
provides greater coordination.
* Clean Polyseed code
* Final tweaks
* Correct no-std builds for Polyseed
* Again correct no-std
---------
Co-authored-by: Luke Parker <lukeparker5132@gmail.com>
Co-authored-by: GitHub Actions <unknown>
Co-authored-by: Boog900 <54e72d8a-345f-4599-bd90-c6b9bc7d0ec5@aleeas.com>
Co-authored-by: Boog900 <108027008+Boog900@users.noreply.github.com>
Co-authored-by: Steven Chang <stevenchang5000@gmail.com>
