crypto/secp256k1: fix undefined behavior in BitCurve.Add #22621
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit fixes the behavior of (BitCurve).Add() to be more inline with btcd. It changes two different bugs that occured. 1) When adding a point at infinity to another point, the other point should be returned. While this is undefined behavior, it is better to be more inline with the go standard library. Thus (0,0) + (a, b) = (a,b) 2) Adding the same point to itself produced the point at infinity. This is incorrect, now doubleJacobian is used to correctly calculate it. This is also similar to the go standard library. Thus (a,b) + (a,b) == 2* (a,b) and not (0,0) anymore.
Contributor
|
We need to find a way to not copy the code into the fuzzer. If it's absolutely not possible to build the fuzzer with cgo, we can do it with build tags in crypto/secp256k1. |
Member
Author
|
@fjl It's absolutely not possible to build it with cgo, see also dvyukov/go-fuzz#101 dvyukov/go-fuzz#171 We would need to either modify |
fjl
changed the title
atif-konasl
pushed a commit
to frozeman/pandora-execution-engine
that referenced
this pull request
Oct 15, 2021
) This commit changes the behavior of BitCurve.Add to be more inline with btcd. It fixes two different bugs: 1) When adding a point at infinity to another point, the other point should be returned. While this is undefined behavior, it is better to be more inline with the go standard library. Thus (0,0) + (a, b) = (a,b) 2) Adding the same point to itself produced the point at infinity. This is incorrect, now doubleJacobian is used to correctly calculate it. Thus (a,b) + (a,b) == 2* (a,b) and not (0,0) anymore. The change also adds a differential fuzzer for Add, testing it against btcd. Co-authored-by: Felix Lange <[email protected]>
This was referenced Sep 23, 2022
This was referenced Mar 19, 2024
19 tasks
gzliudan
added a commit
to gzliudan/XDPoSChain
that referenced
this pull request
Dec 9, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR fixes the behavior of (BitCurve).Add() to be more inline
with btcd and the standard library. It changes two different bugs that occured.
When adding a point at infinity to another point, the other point
should be returned. While this is undefined behavior, it is better
to be more inline with the go standard library.
Thus (0,0) + (a, b) = (a,b)
Adding the same point to itself produced the point at infinity.
This is incorrect, now doubleJacobian is used to correctly calculate it.
This is also similar to the go standard library.
Thus (a,b) + (a,b) == 2* (a,b) and not (0,0) anymore
It also adds a differential fuzzer to test it against btcd.