From e8e55bb19bcc07ad7977577fd6ace7be283bd01b Mon Sep 17 00:00:00 2001
From: fzyzcjy <5236035+fzyzcjy@users.noreply.github.com>
Date: Tue, 19 May 2026 09:12:47 +0800
Subject: [PATCH] Split the request-reception loop into smaller phases (#25703)

---
 .../scheduler_components/request_receiver.py  | 25 +++++++++++++++----
 1 file changed, 20 insertions(+), 5 deletions(-)

diff --git a/python/sglang/srt/managers/scheduler_components/request_receiver.py b/python/sglang/srt/managers/scheduler_components/request_receiver.py
index fe29bd690..28ec4f098 100644
--- a/python/sglang/srt/managers/scheduler_components/request_receiver.py
+++ b/python/sglang/srt/managers/scheduler_components/request_receiver.py
@@ -71,6 +71,20 @@ class SchedulerRequestReceiver:
             if not self.recv_skipper.handle(self.get_last_forward_mode()):
                 return []
 
+        recv_reqs = self._pull_raw_reqs()
+
+        if self.input_blocker is not None:
+            recv_reqs = self.input_blocker.handle(recv_reqs)
+
+        recv_reqs = self._broadcast_reqs_across_ranks(recv_reqs)
+
+        recv_reqs = self._apply_mm_receiver(recv_reqs)
+
+        self._finalize_shm_features(recv_reqs)
+
+        return recv_reqs
+
+    def _pull_raw_reqs(self) -> Optional[List]:
         if self.ps.pp_rank == 0:
             if self.ps.attn_tp_rank == 0 and self.ps.attn_cp_rank == 0:
                 recv_reqs = []
@@ -106,10 +120,9 @@ class SchedulerRequestReceiver:
                 )
             else:
                 recv_reqs = None
+        return recv_reqs
 
-        if self.input_blocker is not None:
-            recv_reqs = self.input_blocker.handle(recv_reqs)
-
+    def _broadcast_reqs_across_ranks(self, recv_reqs: Optional[List]) -> List:
         if self.server_args.enable_dp_attention:
             if self.ps.attn_tp_rank == 0 and self.ps.attn_cp_rank == 0:
                 work_reqs, control_reqs = self._split_work_and_control_reqs(recv_reqs)
@@ -169,7 +182,9 @@ class SchedulerRequestReceiver:
                 self.tp_cpu_group,
                 src=self.tp_group.ranks[0],
             )
+        return recv_reqs
 
+    def _apply_mm_receiver(self, recv_reqs: List) -> List:
         # Process MM requests under EPD-disaggregation mode
         if (
             self.ps.pp_rank == 0
@@ -185,7 +200,9 @@ class SchedulerRequestReceiver:
                 )
                 prepare_abort(req, error_msg, status_code=status_code)
                 self.stream_output([req], req.return_logprob)
+        return recv_reqs
 
+    def _finalize_shm_features(self, recv_reqs: Optional[List]) -> None:
         # Unwrap shared memory features AFTER all broadcasts complete,
         # so that ShmPointerMMData metadata (not full tensor data) is what
         # gets serialized during broadcast_pyobj.
@@ -214,8 +231,6 @@ class SchedulerRequestReceiver:
             for req in recv_reqs:
                 unwrap_shm_features(req)
 
-        return recv_reqs
-
     def _split_work_and_control_reqs(self, recv_reqs: List):
         work_reqs = [
             req