Claude review: drm/tegra: Switch private_obj initialization to atomic_create_state

[Claude Code Review Bot <claude-review@example.com>]

Patch Review

**Commit message**: Clear and well-written.

**Code review**:

The new `tegra_display_hub_create_state()` is structurally identical to the omapdrm version and is correct:

```c
static struct drm_private_state *
tegra_display_hub_create_state(struct drm_private_obj *obj)
{
	struct tegra_display_hub_state *hub_state;

	hub_state = kzalloc_obj(*hub_state);
	if (!hub_state)
		return ERR_PTR(-ENOMEM);

	__drm_atomic_helper_private_obj_create_state(obj, &hub_state->base);

	return &hub_state->base;
}
```

**Same issue: Missing return value check:**

```c
	drm_atomic_private_obj_init(drm, &hub->base, NULL,
				    &tegra_display_hub_state_funcs);

	tegra->hub = hub;

	return 0;
```

Should check the return value and propagate errors.

**Note**: I realize this same return-value-not-checked pattern exists in many other already-converted callers (komeda, ingenic, dpu, mdp5, vc4, drm_bridge, drm_dp_mst, drm_dp_tunnel). This is a wider pre-existing issue, but patches 1-3 should not perpetuate it. At minimum the three patches being submitted should get it right.

---

---
Generated by Claude Code Patch Reviewer