cuda: check for gpu instead of /dev/nvidiactl

The check for `/dev/nvidiactl` to determine if the CUDA plugin can be
used is unreliable because in some cases the default path for driver
installation is different [1]. This patch changes the logic to check
if a GPU device is available in `/proc/driver/nvidia/gpus/`. This is
a more accurate indicator, and the subsequent check for `--action`
option would confirm if the NVIDIA driver supports checkpoint/restore.

[1] https://github.com/NVIDIA/gpu-operator

Signed-off-by: Radostin Stoyanov <rstoyanov@fedoraproject.org>

plugins/cuda/cuda_plugin.c

@@ -470,6 +470,22 @@ int cuda_plugin_resume_devices_late(int pid)
 }
 CR_PLUGIN_REGISTER_HOOK(CR_PLUGIN_HOOK__RESUME_DEVICES_LATE, cuda_plugin_resume_devices_late)
 
+static bool has_gpu()
+{
+	DIR *dir = opendir("/proc/driver/nvidia/gpus/");
+	if (dir != NULL) {
+		struct dirent *entry;
+		while ((entry = readdir(dir)) != NULL) {
+			if (entry->d_type == DT_DIR && entry->d_name[0] != '.') {
+				closedir(dir);
+				return true;
+			}
+		}
+		closedir(dir);
+	}
+	return false;
+}
+
 int cuda_plugin_init(int stage)
 {
 	int ret;
@@ -481,10 +497,12 @@ int cuda_plugin_init(int stage)
 		}
 	}
 
-	if (!fault_injected(FI_PLUGIN_CUDA_FORCE_ENABLE) && access("/dev/nvidiactl", F_OK)) {
-		pr_info("/dev/nvidiactl doesn't exist. The CUDA plugin is disabled.\n");
-		plugin_disabled = true;
-		return 0;
+	if (!fault_injected(FI_PLUGIN_CUDA_FORCE_ENABLE)) {
+		if (!has_gpu()) {
+			pr_info("No GPU device found; CUDA plugin is disabled\n");
+			plugin_disabled = true;
+			return 0;
+		}
 	}
 
 	ret = cuda_checkpoint_supports_flag("--action");