test: recover the config-namespace-migration deferrals (#33171)

The module-skipped tests injected config by faking get_server_args (a
SimpleNamespace stand-in patched onto the module) or by writing fields
onto a ServerArgs instance post-publish — both invisible to the namespace
accessors the production code now reads. Re-enable them by publishing the
config they need (get_context().override_server_args seeding, scoped per
test), asserting bag state where the old assertions checked instance
write-through (declare_load_time_override is bag-only), and extending the
per-runner stubs the code genuinely reads (kv_cache_dtype_str,
max_total_tokens, context_len).

The unified-radix-cache file (which grew a large hicache/insert-walk suite
while skipped) is recovered in the same change:

- test_cache_finished_req_strips_thinking (19 parametrized classes) wrote
  strip_thinking_cache onto the ServerArgs instance; the cache reads
  get_serving().strip_thinking_cache — use the serving bag's scoped
  override.
- test_shallower_crossing_backs_up_above_backuped_middle staged its
  broken-backup-continuity setup through insert_host, which now
  deliberately drops refills below an un-backed-up node under
  write-through (host_insert_dropped). Build the same tree state through
  an explicit backup + device eviction.

Every config-namespace-migration deferral is recovered, so the deferral
ratchet (test_migration_deferral_ratchet.py) has done its job and is
retired.

test/registered/unit/spec/test_eagle_worker_v2_topk1_fastpath.py

@@ -13,6 +13,7 @@ from unittest.mock import MagicMock, patch
 import torch
 
 from sglang.srt.model_executor.forward_batch_info import ForwardMode
+from sglang.srt.runtime_context import get_context
 from sglang.srt.speculative.adaptive_runtime_state import SpecRuntimeState
 from sglang.srt.speculative.eagle_utils import organize_draft_results
 from sglang.srt.speculative.eagle_worker_v2 import EagleDraftWorker, EAGLEWorkerV2
@@ -26,20 +27,6 @@ from sglang.test.test_utils import CustomTestCase
 register_cuda_ci(est_time=20, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=20, stage="stage-b", runner_config="1-gpu-small-amd")
 
-import pytest as _pytest_defer
-
-_DEFER_REASON = (
-    "Temporarily skipped during the ServerArgs config-namespace migration; "
-    "re-enabled once the runtime-config accessor API stabilizes."
-)
-pytestmark = _pytest_defer.mark.skip(reason=_DEFER_REASON)
-
-
-def setUpModule():
-    import unittest
-
-    raise unittest.SkipTest(_DEFER_REASON)
-
 
 register_cpu_ci(est_time=20, suite="base-a-test-cpu")
 
@@ -109,6 +96,15 @@ def _make_backend_factory(decode_backend, draft_extend_backend, captured_kwargs=
 
 
 class TestEagleWorkerV2Topk1FastPath(CustomTestCase):
+    def setUp(self):
+        # _rebuild_topk1_chain_buffers sizes its preallocation from the
+        # published config: get_exec().graph.cuda_graph_config stays None on
+        # the dummy-boundary publish (no resolution), so
+        # get_schedule().max_running_requests alone sizes the buffers.
+        override = get_context().override_server_args(max_running_requests=8)
+        override.install()
+        self.addCleanup(override.restore)
+
     def test_fast_path_matches_slow_path(self):
         bs = 3
         for num_steps in (1, 2, 3, 4):
@@ -145,6 +141,13 @@ class TestEagleWorkerV2Topk1FastPath(CustomTestCase):
 
 
 class TestEagleWorkerV2BackendFallback(CustomTestCase):
+    def setUp(self):
+        # The adaptive state-machine paths write live spec switches through
+        # get_context().override, which needs a published config.
+        override = get_context().override_server_args()
+        override.install()
+        self.addCleanup(override.restore)
+
     def test_missing_seed_cuda_graph_fallback(self):
         graph_result = (
             [],