Test RAM loaded GGUF without pipe deadlock

app.py

@@ -323,7 +323,7 @@ def _run_cpu_smoke(model_filter):
         "engine": "llama-cli",
         "gpu_layers": 0,
         "threads": 1,
-        "no_mmap": False,
+        "no_mmap": True,
         "phases": {},
     }
 
@@ -354,7 +354,7 @@ def _run_cpu_smoke(model_filter):
     except Exception as exc:
         result["phases"]["help"] = {"ok": False, "elapsed_sec": round(time.time() - help_t0, 2), "error": str(exc)}
 
-    command = [cli, "-m", path, "-t", "1", "-ngl", "0", "-c", "256", "-n", "1", "-p", "Say hi.", "--no-display-prompt", "--no-warmup", "--log-disable", "-fa", "0"]
+    command = [cli, "-m", path, "-t", "1", "-ngl", "0", "-c", "256", "-n", "1", "-p", "Say hi.", "--no-display-prompt", "--no-warmup", "--no-mmap", "--log-disable", "-fa", "0"]
     t0 = time.time()
     try:
         run = subprocess.run(command, stdout=subprocess.DEVNULL, stderr=subprocess.PIPE, text=True, timeout=180, env=env)
@@ -396,7 +396,7 @@ def _run_cpu_process_probe(model_filter):
     env["OPENBLAS_NUM_THREADS"] = "1"
     env["VECLIB_MAXIMUM_THREADS"] = "1"
     env["NUMEXPR_NUM_THREADS"] = "1"
-    command = [cli, "-m", path, "-t", "1", "-ngl", "0", "-c", "256", "-n", "1", "-p", "Say hi.", "--no-display-prompt", "--no-warmup", "--log-disable", "-fa", "0"]
+    command = [cli, "-m", path, "-t", "1", "-ngl", "0", "-c", "256", "-n", "1", "-p", "Say hi.", "--no-display-prompt", "--no-warmup", "--no-mmap", "--log-disable", "-fa", "0"]
     started = time.time()
     child = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, env=env)
     time.sleep(8)
@@ -456,7 +456,7 @@ def _run_cpu_diagnose(model_filter):
             p = subprocess.run(
                 ["strace", "-f", "-tt", "-o", trace_path, cli, "-m", path,
                  "-t", "1", "-ngl", "0", "-c", "256", "-n", "1", "-p", "Say hi.",
-                 "--no-display-prompt", "--no-warmup", "-fa", "0"],
+                 "--no-display-prompt", "--no-warmup", "--no-mmap", "-fa", "0"],
                 capture_output=True, text=True, timeout=25, env=env,
             )
             result["load_trace"] = {"returncode": p.returncode, "stdout": (p.stdout or "")[-1000:], "stderr": (p.stderr or "")[-2000:]}