Claude review: drm/plane: Add new atomic_create_state callback

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

Patch Review

Adds the `atomic_create_state` callback to `drm_plane_funcs`, creates `__drm_atomic_helper_plane_create_state()` and `drm_atomic_helper_plane_create_state()` helpers. Also adds fallback logic in `drm_mode_config_reset()` to call `atomic_create_state` if `reset` isn't set.

**Issue**: `__drm_atomic_helper_plane_create_state()` is a trivial one-line wrapper around `__drm_atomic_helper_plane_state_init()`:
```c
void __drm_atomic_helper_plane_create_state(struct drm_plane *plane,
                                            struct drm_plane_state *state)
{
    __drm_atomic_helper_plane_state_init(state, plane);
}
```
This is unlike the crtc version (patch 10) which adds `drm_crtc_vblank_reset()`. For planes there's no extra init needed, so this wrapper adds indirection with no value. Consider whether `__drm_atomic_helper_plane_state_init` could just be called directly, or document why the wrapper exists (future-proofing for additional init steps?).

**Issue in drm_mode_config_reset()**: The return value of `drm_mode_config_plane_create_state()` is silently ignored:
```c
else if (plane->funcs->atomic_create_state)
    drm_mode_config_plane_create_state(plane);
```
Since `drm_mode_config_reset()` returns void, there's nothing to do with the error, but a `WARN_ON` or `drm_err` on failure would help debugging. This same issue applies to the crtc and connector versions added in patches 10 and 13.

---
Generated by Claude Code Patch Reviewer