Use Keccak256 when hashing for signing #3138
Conversation
deuszx
force-pushed
the
keccak-hashing
branch
from
b93dd87 to
3988cf1
Compare
deuszx
force-pushed
the
keccak-hashing
branch
from
3988cf1 to
5fa8dd4
Compare
deuszx
force-pushed
the
keccak-hashing
branch
from
0db975d to
034f0a0
Compare
deuszx
force-pushed
the
keccak-hashing
branch
from
034f0a0 to
7fade4e
Compare
|
Testnet |
| @@ -167,7 +166,7 @@ impl Serialize for CryptoHash { | |||
| if serializer.is_human_readable() { | |||
| serializer.serialize_str(&self.to_string()) | |||
| } else { | |||
| serializer.serialize_newtype_struct("CryptoHash", &self.0) | |||
| serializer.serialize_newtype_struct("CryptoHash", &self.0 .0) | |||
There was a problem hiding this comment.
Maybe CryptoHash deserves an as_bytes method to avoid the .0 .0.
| @@ -396,6 +393,20 @@ where | |||
| } | |||
| } | |||
|
|
|||
| // Temporary struct to extend `Keccak256` with `io::Write`. | |||
There was a problem hiding this comment.
| // Temporary struct to extend `Keccak256` with `io::Write`. | |
| /// Temporary struct to extend `Keccak256` with `io::Write`. |
| @@ -396,6 +393,20 @@ where | |||
| } | |||
| } | |||
|
|
|||
| // Temporary struct to extend `Keccak256` with `io::Write`. | |||
| struct Keccak256Ext(Keccak256); | |||
There was a problem hiding this comment.
Is this used in places other than CryptoHash::new? Otherwise we might as well just call update directly there.
|
|
||
| integers | ||
| } | ||
|
|
||
| /// Returns the bytes that represent the `integers` in little-endian. | ||
| fn u64_array_to_le_bytes(integers: [u64; 4]) -> [u8; 32] { |
There was a problem hiding this comment.
Are we still using this? (And should we be using this at all?)
| @@ -1026,15 +1026,15 @@ mod tests { | |||
| fn chain_ids() { | |||
There was a problem hiding this comment.
It might be helpful to add root chains 1 and 2 here as well: They are probably used in a few READMEs, so this serves as documentation where one can look up the hash. It might also simplify any future changes to how chain IDs are computed.
| @@ -1333,22 +1333,23 @@ async fn test_multiple_messages_from_different_applications() -> anyhow::Result< | |||
|
|
|||
| // The registration message for the first destination chain | |||
There was a problem hiding this comment.
| // The registration message for the first destination chain | |
| // The registration message for the second destination chain |
There was a problem hiding this comment.
Or we could rename first_message and second_message to message0 and message1 and the destination chains similarly. (But keep the *_registration_messages.)
| }, | ||
| }; | ||
|
|
||
| // The registration message for the second destination chain |
There was a problem hiding this comment.
| // The registration message for the second destination chain | |
| // The registration message for the first destination chain |
Motivation
We want to verify Linera blocks on Ethereum (and other EVMs) but they don't have opcodes/precompiles for Sha256 that we're using .
Proposal
Replace Sha256 with Keccak256.
Notice that I'm only changing hashing scheme used in
CryptoHashstruct - i.e. there are still places in the code where we use sha266 (like for example linera views and cache).Test Plan
Release Plan
This is backwards incompatible change.
Links
Closes #3125