Claude review: drm/imagination: Rename FW booted to FW initialised

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

Patch Review

**Verdict: Build breakage**

This patch is described as a pure rename (`booted` → `initialised`) but it silently changes function call arguments in `pvr_power.c`, creating a **compile error**.

In `pvr_power_device_suspend()`, the diff changes:
```c
-		err = pvr_power_fw_disable(pvr_dev, false, true);
+		err = pvr_power_fw_disable(pvr_dev, false);
```

In `pvr_power_device_resume()`:
```c
-		err = pvr_power_fw_enable(pvr_dev, true);
+		err = pvr_power_fw_enable(pvr_dev);
```

But the function signatures are *not* updated:
```c
pvr_power_fw_disable(struct pvr_device *pvr_dev, bool hard_reset, bool rpm_suspend)  // 3 params
pvr_power_fw_enable(struct pvr_device *pvr_dev, bool rpm_resume)  // 2 params
```

And other call sites in `pvr_power_reset()` still use the original argument count:
```c
err = pvr_power_fw_disable(pvr_dev, hard_reset, false);  // 3 args
err = pvr_power_fw_enable(pvr_dev, false);  // 2 args
```

This means the patch will fail to compile. It looks like the patch was generated against a different tree where `pvr_power_fw_disable`/`pvr_power_fw_enable` already had reduced parameter lists, or unrelated changes were accidentally folded in.

**Fix:** Either keep the original argument counts (just do the `booted` → `initialised` rename), or include the function signature changes and update all call sites consistently. If the intent is to remove the `rpm_suspend`/`rpm_resume` parameters entirely, that should be a separate patch with its own justification.

---
Generated by Claude Code Patch Reviewer