Align MQA logits admission with CP split runtime shape

CP shared-KV batching previously estimated MQA logits from full request
extend/context rows, which overstated memory because CP in-seq split only
computes each rank's two zigzag segments. Add CP-size aware row accounting
that mirrors the fused CP MQA materialization path and take the worst local
rank peak for scheduler admission.

Expose SGLANG_NSA_MQA_LOGITS_CHUNK_MAX_GB as a more direct cap for one fp32
MQA logits chunk. Runtime and scheduler now both translate this GB cap into
chunk rows from the actual K rows, while keeping the old row cap as a
mutually-exclusive expert override.

Constraint: Scheduler admission must stay CUDA-sync-free and use static budget information only.
Rejected: Keep full-request q*k admission | it over-gates CP bs>1 batches because CP splits q rows per rank.
Rejected: Let rows and GB caps both apply | precedence would be ambiguous during tuning.
Confidence: medium
Scope-risk: moderate
Directive: Keep MQA logits admission tied to the fused CP MQA segment shape; do not revert to full request token counts.
Tested: Local py_compile for touched runtime, scheduler, estimator, and tests.
Tested: Local pytest test_cp_shared_kv_prefill_buffer_estimator.py: 8 passed.
Tested: Remote g0034 cjy-glm5-new py_compile and targeted estimator/runtime tests: 13 passed.
Not-tested: Full ETE high-cache-hit CP bs>1 load with SGLANG_NSA_MQA_LOGITS_CHUNK_MAX_GB.

Co-authored-by: OmX <omx@oh-my-codex.dev>

test/registered/unit/mem_cache/test_cp_shared_kv_runtime.py

@@ -5854,6 +5854,38 @@ class TestCpSharedKVTaiMaterializeIntegration(unittest.TestCase):
                 128,
             )
 
+    def test_nsa_mqa_logits_chunk_max_gb_caps_rows_by_bytes(self):
+        from sglang.srt.environ import envs
+        from sglang.srt.layers.attention.nsa.nsa_indexer import Indexer
+
+        indexer = object.__new__(Indexer)
+        with envs.SGLANG_NSA_MQA_LOGITS_CHUNK_MAX_ROWS.override(
+            0
+        ), envs.SGLANG_NSA_MQA_LOGITS_CHUNK_MAX_GB.override(1.0):
+            self.assertEqual(
+                indexer._mqa_logits_chunk_max_rows(
+                    num_q=8192,
+                    num_k=1_000_000,
+                    logits_budget_bytes=1,
+                ),
+                250,
+            )
+
+    def test_nsa_mqa_logits_chunk_rows_and_gb_conflict_fail_fast(self):
+        from sglang.srt.environ import envs
+        from sglang.srt.layers.attention.nsa.nsa_indexer import Indexer
+
+        indexer = object.__new__(Indexer)
+        with envs.SGLANG_NSA_MQA_LOGITS_CHUNK_MAX_ROWS.override(
+            128
+        ), envs.SGLANG_NSA_MQA_LOGITS_CHUNK_MAX_GB.override(1.0):
+            with self.assertRaisesRegex(RuntimeError, "mqa_logits_chunk"):
+                indexer._mqa_logits_chunk_max_rows(
+                    num_q=8192,
+                    num_k=1_000_000,
+                    logits_budget_bytes=1,
+                )
+
 
 if __name__ == "__main__":
     unittest.main()