Claude review: drm: writeback: Modify drm_writeback_signal_completion helper

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

Patch Review

Changes `drm_writeback_signal_completion()` to take `struct drm_connector *`.

**Bug:**

In `amdgpu_dm.c`:
```c
-			drm_writeback_signal_completion(acrtc->wb_conn, 0);
+			drm_writeback_signal_completion(acrtc->connector, 0);
```

**`acrtc->connector` does not exist.** The field is `acrtc->wb_conn` (type `struct drm_writeback_connector *`). After the API change, this should be:
```c
drm_writeback_signal_completion(drm_writeback_to_connector(acrtc->wb_conn), 0);
```

Notably the MSM driver gets this right in the same patch:
```c
+		drm_writeback_signal_completion(drm_writeback_to_connector(wb_enc->wb_conn), 0);
```

The amdgpu site should follow the same pattern. The `amdgpu_crtc` struct has `struct drm_writeback_connector *wb_conn` (confirmed in `amdgpu_mode.h:513`), so the correct call would use `drm_writeback_to_connector(acrtc->wb_conn)`. Alternatively, the `amdgpu_crtc` struct could be updated to store a `struct drm_connector *` pointer instead, but that's not done in this series.

**This will cause a build failure.**

The `malidp_hw.c` changes are also questionable -- the diff header shows reverting to the original file hash (`9b845d3f34e1`) before the patch 1 changes. This is likely a rebase issue but the final state after applying patches 1 and 4 should be correct (both change the same lines in the same way).

---
Generated by Claude Code Patch Reviewer