* WIP constant-time implementation of the ec-divisors library
* Fix misc logic errors in poly.rs
* Remove accidentally committed test statements
* Fix ConstantTimeEq for CoefficientIndex
* Correct the iterations formula
x**3 / (0 y + x**1) would prior be considered indivisible with iterations = 0.
It is divisible however. The amount of iterations should be the amount of
coefficients within the numerator *excluding the coefficient for y**0 x**0*.
* Poly PartialEq, conditional_select_poly which checks poly structure equivalence
If the first passed argument is smaller than the latter, it's padded to the
necessary length.
Also adds code to trim the remainder as the remainder is the value modulo, so
it's very important it remains concise and workable.
* Fix the line function
It selected the case if both were identity before selecting the case if either
were identity, the latter overwriting the former.
* Final fixes re: ct_get
1) Our quotient structure does need to be of size equal to the numerator
entirely to prevent out-of-bounds reads on it
2) We need to get from yx_coefficients if of length >=, so if the length is 1
we can read y_pow=1 from it. If y_pow=0, and its length is 0 so it has no
inner Vecs, we need to fall back with the guard y_pow != 0.
* Add a trim algorithm to lib.rs to prevent Polys from becoming unbearably gigantic
Our Poly algorithm is incredibly leaky. While it presumably should be improved,
we can take advantage of our known structure while constructing divisors (and
the small modulus) to simply trim out the zero coefficients leaked. This
maintains Polys in a manageable size.
* Move constant-time scalar mul gadget divisor creation from dkg to ec-divisors
Anyone creating a divisor for the scalar mul gadget should use constant time
code, so this code should at least be in the EC gadgets crate It's of
non-trivial complexity to deal with otherwise.
* Remove unsafe, cache timing attacks from ec-divisors
This technically has a TOCTOU where we sync an Epoch's metadata (signifying we
did sync to that point), then check if the Router was deployed, yet at that
very moment the node resets to genesis. By ensuring the Router is deployed, we
avoid this (and don't need to track the deployment block in-contract).
Also uses a JoinSet to sync the 32 blocks in parallel.
The caller is paid a fixed fee per unit of gas spent. That arguably
incentivizes the publisher to raise the gas used by internal calls, yet this
doesn't effect the user UX as they'll have flatly paid the worst-case fee
already. It does pose a risk where callers are arguably incentivized to cause
transaction failures which consume all the gas, not just increased gas, yet:
1) Modern smart contracts don't error by consuming all the gas
2) This is presumably infeasible
3) Even if it was feasible, the gas fees gained presumably exceed the gas fees
spent causing the failure
The benefit to only paying the callers for the gas used, not the gas alotted,
is it allows Serai to build up a buffer. While this should be minor, a few
cents on every transaction at best, if we ever do have any costs slip through
the cracks, it ideally is sufficient to handle those.
contracts was smashed out of ethereum-serai. Both have now been smashed into
individual crates.
Creates a TODO directory with left-over test code yet to be moved.
The router will now match the top-level transfer so it isn't used as the
justification for the InInstruction it's handling. This allows the theoretical
case where a top-level transfer occurs (to any entity) and an internal call
performs a transfer to Serai.
Also uses a JoinSet for fetching transactions' top-level transfers in the ERC20
crate. This does add a dependency on tokio yet improves performance, and it's
scoped under serai-processor (which is always presumed to be tokio-based).
While we could instead import futures for join_all,
https://github.com/smol-rs/futures-lite/issues/6 summarizes why that wouldn't
be a good idea. While we could prefer async-executor over tokio's JoinSet,
JoinSet doesn't share the same issues as FuturesUnordered. That means our
question is solely if we want the async-executor executor or the tokio
executor, when we've already established the Serai processor is always presumed
to be tokio-based.
It makes sense for networks which support arbitrary data to do as part of their
address. This reduces the ability to perform DoSs, achieves better performance,
and better uses the type system (as now networks we don't support data on don't
have a data field).
Updates the Ethereum address definition in serai-client accordingly
If instead of intaking calls, we intake code, we can deploy a fresh contract
which makes arbitrary calls *without* attempting to build our abstraction
layer over the concept.
This should have the same gas costs, as we still have one contract deployment.
The new contract only has a constructor, so it should have no actual code and
beat the Sandbox in that regard? We do have to call into ourselves to meter the
gas, yet we already had to call into the deployed Sandbox to achieve that.
Also re-defines the OutInstruction to include tokens, implements
OutInstruction-specified gas amounts, bumps the Solidity version, and other
such misc changes.
