Fix radix cache duplicate prefix remapping - #142
Open
Rapheal-zc wants to merge 1 commit into
Open
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
This PR fixes radix prefix cache handling for unfinished requests.
When
CacheManager.cache_req()inserts a request prefix into the radix cache, part of the current request's computed prefix may already exist in the cache. The previous logic freed those duplicate request-local pages directly, but the request'spage_tablecould still point to the freed pages. Later decode steps could then read KV pages that had already been returned to the free pool and potentially reused.This PR remaps the unfinished request's
page_tableentries to the canonical radix-cache page indices before freeing the duplicate request-local pages.Validation
Validated with Qwen3-32B AIME26 repeat testing using radix cache.