From e4d81e48c91354d442f235102ae84e6be66fe266 Mon Sep 17 00:00:00 2001
From: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com>
Date: Tue, 19 May 2026 09:27:06 +0800
Subject: [PATCH] Pull the max-prefix-len computation into its own helper and
 rename the matched-token argument (#25728)

---
 python/sglang/srt/managers/schedule_batch.py | 21 +++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/python/sglang/srt/managers/schedule_batch.py b/python/sglang/srt/managers/schedule_batch.py
index 1830c90c9..ca8215909 100755
--- a/python/sglang/srt/managers/schedule_batch.py
+++ b/python/sglang/srt/managers/schedule_batch.py
@@ -1026,25 +1026,21 @@ class Req(ReqDllmMixin):
             )
             self.logprob_start_len = -1
 
-        # NOTE: the matched length is at most 1 less than the input length to enable logprob computation
-        max_prefix_len = input_len - 1
-        if self.return_logprob and self.logprob_start_len >= 0:
-            max_prefix_len = min(max_prefix_len, self.logprob_start_len)
-        max_prefix_len = max(max_prefix_len, 0)
-        token_ids = self.fill_ids[:max_prefix_len]
-        del max_prefix_len
+        token_ids_to_match = self.fill_ids[: self._compute_max_prefix_len(input_len)]
 
         # Disable prefix caching when embed overrides are present: same token IDs
         # with different override vectors must not share cached KV values.
         if self.positional_embed_overrides is not None:
-            token_ids = []
+            token_ids_to_match = []
 
         if tree_cache is not None:
             if cow_mamba is None:
                 cow_mamba = tree_cache.supports_mamba()
             match_result = tree_cache.match_prefix(
                 MatchPrefixParams(
-                    key=RadixKey(token_ids=token_ids, extra_key=self.extra_key),
+                    key=RadixKey(
+                        token_ids=token_ids_to_match, extra_key=self.extra_key
+                    ),
                     req=self,
                     cow_mamba=cow_mamba,
                 )
@@ -1091,6 +1087,13 @@ class Req(ReqDllmMixin):
 
         self.set_extend_input_len(len(self.fill_ids) - len(self.prefix_indices))
 
+    def _compute_max_prefix_len(self, input_len: int) -> int:
+        # NOTE: the matched length is at most 1 less than the input length to enable logprob computation
+        max_prefix_len = input_len - 1
+        if self.return_logprob and self.logprob_start_len >= 0:
+            max_prefix_len = min(max_prefix_len, self.logprob_start_len)
+        return max(max_prefix_len, 0)
+
     # Based on https://github.com/vllm-project/vllm/blob/7a64d24aad69e4d2548aa0bf528d9fe63428ab01/vllm/transformers_utils/detokenizer.py#L194-L313
     def init_incremental_detokenize(self):
         first_iter = self.surr_offset is None or self.read_offset is None