Only wait for config waitGroup when configManager is used

[chdxD1]

reopen of #1398 with rebase on master

When MultusConfigFile is not set to auto (configManager is nil) the WaitGroup is initialised and then waited for, immediately exiting Multus. This PR only waits when configManager is not nil, otherwise it waits for the context to complete (which happens from the SIGINT / SIGTERM handlers)

[gemini-code-assist]

Code Review

This pull request modifies the multus-daemon entry point to prevent the process from exiting immediately when the configuration manager is not initialized. The reviewer suggested a more idiomatic implementation that simplifies the control flow by waiting for context completion before synchronizing background tasks, which ensures a more robust shutdown sequence.

[gemini-code-assist]

The current implementation correctly fixes the issue where the daemon exits immediately when configManager is nil. However, the logic can be made more idiomatic and robust by explicitly waiting for the context completion in all cases, followed by waiting for any background tasks (like the configuration manager) to finish their cleanup. This also avoids the if/else branching for the wait logic.

Additionally, note that the CNI server started in startMultusDaemon also performs a graceful shutdown when the context is cancelled, but there is currently no synchronization to ensure it finishes before the process exits. While out of scope for this specific fix, consider using a shared WaitGroup for both the server and the config manager in the future.

Suggested change

	if configManager != nil {
	var wg sync.WaitGroup
	if err := configManager.Start(ctx, &wg); err != nil {
	_ = logging.Errorf("failed to start config manager: %v", err)
	os.Exit(3)
	}
	wg.Wait()
	} else {
	<-ctx.Done()
	}
	var wg sync.WaitGroup
	if configManager != nil {
	if err := configManager.Start(ctx, &wg); err != nil {
	_ = logging.Errorf("failed to start config manager: %v", err)
	os.Exit(3)
	}
	}
	<-ctx.Done()
	wg.Wait()