feat(s2n-quic-dc): expanded replay protection #2462
Draft
+1,301
−402
Conversation
This file contains 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
A key id flows through the following structures: * max seen key ID * 32-bit bitset * variable sized set (bitset or sorted list, tracking back 65k)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Release Summary:
Improves replay protection tracking to guarantee 2^16 replay window tracking in the absence of memory pressure, unlike the previous implementation which was dependent on hash collisions not happening (and so had no strong guarantee outside of ~128 entries).
Description of changes:
This is a pretty large set of changes, and may still evolve in response to review -- the overall idea is to allocate from a shared memory region (essentially a mixed-type slab) for either "sorted lists" (1-3700 entries, roughly) or bitsets (3700-65k entries) with a maximum per-path-secret usage of around 8kb. The resulting structure provides the release summary guarantee that if we don't have sufficient reordering across all peers that we use up our general allocation, we will be be tracking reliably up to 2^16 entries out. Our modeling suggests that at current reordering rates, we shouldn't ever use up close to the full map, but we'll prove that out once this gets deployed.
Opening for preliminary review, especially looking for high-level feedback: can we simplify something at low cost, ideas for different tradeoffs, etc.
Call-outs:
TODO
Testing:
TODO
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.