[Spec] Anchor GLM-5.2 MTP IndexShare topk on the draft-extend step (#29787)

Co-authored-by: kpham-sgl <264503018+kpham-sgl@users.noreply.github.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_extend_runner.py

@@ -20,6 +20,7 @@ from sglang.srt.speculative.eagle_draft_extend_cuda_graph_runner import (
     EAGLEDraftExtendCudaGraphRunner,
 )
 from sglang.srt.speculative.eagle_info import EagleDraftExtendInput
+from sglang.srt.speculative.eagle_worker_v2 import EagleDraftWorker
 from sglang.srt.speculative.spec_info import SpeculativeAlgorithm
 from sglang.srt.speculative.spec_utils import fast_topk
 
@@ -496,6 +497,7 @@ class _EagleDraftExtendV2WorkerHarness:
         self.eagle_use_aux_hidden_state = False
         self.hot_token_id = None
         self.draft_runner.model = model_forward
+        EagleDraftWorker._init_dsa_index_share_state(self)
 
 
 def _build_eagle_draft_extend_fixture(

python/sglang/test/kits/attention_unittest/runner_modes/speculative_draft_runner.py

@@ -188,15 +188,7 @@ class _EagleDraftWorkerHarness:
         self._topk1_parents_prealloc = None
         self._topk1_score_indices_prealloc = None
         EagleDraftWorker._rebuild_topk1_chain_buffers(self)
-        # draft_forward reads this (set in EagleDraftWorker.__init__, skipped here).
-        self.index_share_for_mtp_iteration = (
-            getattr(
-                self.model_config.hf_config,
-                "index_share_for_mtp_iteration",
-                False,
-            )
-            and self.topk == 1
-        )
+        EagleDraftWorker._init_dsa_index_share_state(self)
 
     @property
     def draft_model_runner(self):