[HPU][Bugfix] Don't use /dev/accel/accel0 for HPU autodetection in setup.py

setup.py

@@ -324,21 +324,26 @@ def run(self) -> None:
 
 
 def _is_hpu() -> bool:
-    is_hpu_available = True
+    # if VLLM_TARGET_DEVICE env var was set explicitly, skip HPU autodetection
+    if os.getenv("VLLM_TARGET_DEVICE", None) == VLLM_TARGET_DEVICE:
+        return VLLM_TARGET_DEVICE == "hpu"
+
+    # if VLLM_TARGET_DEVICE was not set explicitly, check if hl-smi succeeds,
+    # and if it doesn't, check if habanalabs driver is loaded
+    is_hpu_available = False
     try:
-        subprocess.run(["hl-smi"], capture_output=True, check=True)
+        out = subprocess.run(["hl-smi"], capture_output=True, check=True)
+        is_hpu_available = out.returncode == 0
     except (FileNotFoundError, PermissionError, subprocess.CalledProcessError):
-        if not os.path.exists('/dev/accel/accel0') and not os.path.exists(
-                '/dev/accel/accel_controlD0'):
-            # last resort...
+        if sys.platform.startswith("linux"):
             try:
                 output = subprocess.check_output(
                     'lsmod | grep habanalabs | wc -l', shell=True)
                 is_hpu_available = int(output) > 0
             except (ValueError, FileNotFoundError, PermissionError,
                     subprocess.CalledProcessError):
-                is_hpu_available = False
-    return is_hpu_available or VLLM_TARGET_DEVICE == "hpu"
+                pass
+    return is_hpu_available
 
 
 def _no_device() -> bool: