Claude review: drm/vmwgfx: Add some checks to vmw_cursor_plane_atomic_update

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

Patch Review

**Status: Looks good, minor style nit**

The fix correctly addresses a null dereference when `old_state` is not present. The root cause is that the old code did:
```c
struct drm_crtc *crtc = new_state->crtc ?: old_state->crtc;
struct vmw_display_unit *du = vmw_crtc_to_du(crtc);
```
...which crashes if both are NULL or `old_state` itself is NULL.

The new approach of getting `du` via `container_of(plane, struct vmw_display_unit, cursor.base)` is correct — the cursor plane *is* `du->cursor.base`, so `container_of` is valid and avoids the dependency on any plane state. The early return on `!new_state` is also appropriate since there's nothing to update.

**Minor style nit**: The declarations of `dev_priv` and `du` appear after the `if (!new_state) return;` statement:
```c
new_state = drm_atomic_get_new_plane_state(state, plane);
if (!new_state)
    return;

struct vmw_private *dev_priv = vmw_priv(plane->dev);
struct vmw_display_unit *du = vmw_cursor_to_du(plane);
```
While this is valid C11 (which the kernel uses), it's a coding style preference issue. Some reviewers prefer all declarations before statements within a block. Not a correctness issue.

---

---
Generated by Claude Code Patch Reviewer