Skip to content

Skip oow slot freeing under eagle #34823

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ispobock merged 1 commit into from
Aug 14, 2026
Merged

Skip oow slot freeing under eagle #34823

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion python/sglang/srt/mem_cache/unified_radix_cache.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -773,7 +773,12 @@ def cache_unfinished_req(self, req: Req, chunked: bool = False, **kwargs) -> Non
if cl is not None:
effective_cache_len = min(effective_cache_len, cl)

if envs.SGLANG_OPT_UNIFIED_CACHE_FREE_OUT_OF_WINDOW_SLOTS.get():
# swa_evicted_seqlen is a raw-token length, but under EAGLE the insert key is
# bigram-indexed, so SWA would carve tombstones at the wrong offset (#34653).
if (
envs.SGLANG_OPT_UNIFIED_CACHE_FREE_OUT_OF_WINDOW_SLOTS.get()
and not self.tree_core.is_eagle
):
for comp in self._components_tuple:
comp.free_out_of_window_slots(
req, effective_cache_len - 1, insert_params
Expand Down
Loading