Claude review: drm/atomic: track individual colorop updates

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

Patch Review

**Author:** Melissa Wen

This is the largest patch and the core behavioral change. It ensures `plane_state->color_mgmt_changed` is set whenever any colorop property in the plane's color pipeline actually changes value.

**Changes reviewed:**

- **`drm_atomic_set_colorop_for_plane()`** now returns `bool` instead of `void`. It returns `false` (no change) when the same colorop pointer is set again, `true` when it actually changes. The early-return for no-change avoids unnecessary debug logging and change tracking. Only one caller exists in-tree, at line 615, which uses `|=` to combine the result into `color_mgmt_changed`. Correct.

- **`drm_atomic_colorop_set_property()`** gains a `bool *replaced` output parameter. Each property branch now compares old vs new value before writing, and only sets `*replaced = true` when the value actually changed:

  ```c
  if (state->bypass != val) {
      state->bypass = val;
      *replaced = true;
  }
  ```

  This pattern is applied to `bypass`, `lut1d_interpolation`, `curve_1d_type`, `multiplier`, and `lut3d_interpolation`. The `data` property delegates to `drm_atomic_color_set_data_property()` which passes the pointer through to `drm_property_replace_blob_from_id()` — that function already has a `bool *replaced` parameter with the correct semantics.

- **`drm_atomic_set_property()` (DRM_MODE_OBJECT_COLOROP case)**: After calling `drm_atomic_colorop_set_property()`, if `ret || !replaced` it breaks early (no plane state needed). Otherwise it fetches the plane state and sets `plane_state->color_mgmt_changed |= replaced`. The `drm_atomic_get_plane_state()` call may fail (returns ERR_PTR), which is correctly handled.

- **`include/drm/drm_atomic_uapi.h`**: Adds `#include <linux/types.h>` for the `bool` return type, and updates the declaration. Clean.

**One minor observation:** At line 1325, `plane_state->color_mgmt_changed |= replaced;` — at this point `replaced` is guaranteed `true` (the `if (ret || !replaced) break;` on line 1317 already filtered the false case), so the `|=` is equivalent to `= true`. This is fine — using `|=` is defensive and matches the convention used elsewhere (e.g., line 615).

**Verdict:** Correct and well-implemented. No issues.

---

---
Generated by Claude Code Patch Reviewer