Claude review: drm/colorop: make lut(1/3)d_interpolation mutable

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

Patch Review

**Purpose:** Moves `lut1d_interpolation` and `lut3d_interpolation` from `struct drm_colorop` (immutable HW object) to `struct drm_colorop_state` (per-atomic-state), making them properly mutable.

**Review:**

The state reset initialization is correctly done through `drm_object_property_get_default_value`:

```c
if (colorop->lut1d_interpolation_property) {
    drm_object_property_get_default_value(&colorop->base,
                                          colorop->lut1d_interpolation_property,
                                          &val);
    colorop_state->lut1d_interpolation = val;
}
```

**Minor concern:** The return value of `drm_object_property_get_default_value` is not checked here. Looking at the existing `curve_1d_type` pattern just above:

```c
if (colorop->curve_1d_type_property) {
    drm_object_property_get_default_value(&colorop->base,
                                          colorop->curve_1d_type_property,
                                          &val))
```

Wait — in the existing code, the return value IS checked as part of an `if` condition. But in the new code for interpolation properties, it is NOT:

```c
if (colorop->lut1d_interpolation_property) {
    drm_object_property_get_default_value(&colorop->base,
                                          colorop->lut1d_interpolation_property,
                                          &val);
    colorop_state->lut1d_interpolation = val;
}
```

Looking at the existing code for `curve_1d_type`, the pattern is:
```c
if (colorop->curve_1d_type_property) {
    if (!drm_object_property_get_default_value(&colorop->base,
                                               colorop->curve_1d_type_property,
                                               &val))
        colorop_state->curve_1d_type = val;
}
```

**The new code doesn't guard the assignment with the return value check.** If `drm_object_property_get_default_value` fails, `val` would be stale from a previous call (the `curve_1d_type` lookup). This is probably benign since the property is guaranteed to exist when the property pointer is non-NULL, but it's inconsistent with the existing pattern. This should match the existing defensive style.

The removal of `colorop->lut1d_interpolation = interpolation` from the init functions is correct since the value now lives in state.

The print_state and get/set property changes are mechanical and correct.

---
Generated by Claude Code Patch Reviewer