What changes, and why?

−# Framework pseudocode, not an executable implementation.
−for request in inputs:
−    llm_engine.add_request(request)
+# Framework pseudocode for request-owned cache state.
+state = sequence_table[request_id]
+blocks = block_manager.resolve(state.block_table)
+metadata = build_input_metadata(
+    query_tokens=planned_tokens,
+    cached_positions=state.computed_positions,
+    cache_blocks=blocks,
+)
+worker.execute_model_step(metadata, cache_engine.cache)
  
−while llm_engine.has_unfinished_requests():
−    ready_outputs = llm_engine.step()
−    collect(ready_outputs)
− 
−collect(llm_engine.flush_inflight_reqs())
− 
−# Inside a scheduled round:
−# Scheduler -> Worker -> ModelRunner -> sampling
−# -> ready state updates and request outputs
−# A round may cover several requests and token positions.
+# Execution consumes input positions and updates cache state.
+# A selected token becomes input only if the request continues.
+# Reclaim blocks only when their references are no longer needed.