Claude review: drm/mode-config: Create drm_mode_config_create_initial_state()

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

Patch Review

This is the key patch introducing `drm_mode_config_create_initial_state()`.

**Main concern — no cleanup on partial failure:** If, say, plane state creation succeeds for all planes but a CRTC state allocation fails, the function returns an error but all the plane states that were already allocated are leaked. The caller gets an error with no way to know which objects got states. The `continue` for objects with existing states helps with re-entrance, but doesn't solve the forward-leak problem.

```c
drm_for_each_plane(plane, dev) {
    if (plane->state)
        continue;
    ret = drm_mode_config_plane_create_state(plane);
    if (ret)
        return ret;  // leaked: all privobj and colorop states allocated above
}
```

In practice, this is called during probe, so a failure here likely means the driver fails to load and the device gets torn down entirely (which would clean up via `drm_mode_config_cleanup()`). So this may be acceptable, but it would be cleaner to document this expectation or add cleanup.

The `if (privobj->state) continue` / `if (colorop->state) continue` guards are good for handling the case where `drm_atomic_private_obj_init()` already created the state, or where `drmm_connector_hdmi_init()` already set up a connector state.

The doc update to patch 1's lifetime documentation (splitting reset-time into init-time and resume-time) is correct.

---
Generated by Claude Code Patch Reviewer