fix: separate grain processing thread from LWS event loop

The fundamental issue: mxlSleepUntil/mxlSleepForNs blocks the entire
thread, making LWS unresponsive and causing control commands to hang.
Meanwhile, LWS poll(1) adds latency that makes grain timing unreliable.

Solution: run node->process() in a dedicated thread that can freely use
mxlSleepUntil for precise grain timing, while the main thread runs
control_server->poll(10) for LWS event handling.

Passthrough now uses mxlSleepUntil(deadline) matching the mxl-gst-sink
Cursor pattern, with 2-grain read delay for buffering.

libs/dmf-node/src/node_runner.cpp

@@ -170,20 +170,19 @@ int NodeRunner::exec(std::unique_ptr<Node> node) {
     running_ = true;
     spdlog::info("Node '{}' entering process loop", node_id_);
 
-    while (g_running && running_) {
-        control_server->poll(1);
-        node->process();
-
-        static int status_counter = 0;
-        if (++status_counter % 100 == 0) {
-            nlohmann::json status_event;
-            status_event["event"] = "status";
-            status_event["node_id"] = node_id_;
-            status_event["data"] = node->status();
-            control_server->send_event(status_event);
+    std::thread process_thread([&]() {
+        while (g_running && running_) {
+            node->process();
         }
+    });
+
+    while (g_running && running_) {
+        control_server->poll(10);
     }
 
+    running_ = false;
+    process_thread.join();
+
     spdlog::info("Node '{}' shutting down", node_id_);
     mxlDestroyInstance(mxl_instance_);
     return 0;