Claude review: drm/msm/dpu: enable virtual planes by default

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

Patch Review

**Code change:**
```c
-bool dpu_use_virtual_planes;
+bool dpu_use_virtual_planes = true;
```

**Observations:**

1. **Correct and minimal change.** The diff does exactly what the commit message describes. The explicit `= true` initializer is the right approach (rather than, say, inverting the logic and the parameter name).

2. **Boot parameter documentation.** The commit message correctly documents the fallback: `msm.dpu_use_virtual_planes=false`. This matches the `module_param(dpu_use_virtual_planes, bool, 0)` declaration at `dpu_kms.c:56`.

3. **Minor nit — no MODULE_PARM_DESC.** The parameter has no `MODULE_PARM_DESC()`, so `modinfo` won't show any description for it. This pre-dates this patch, but given that the parameter is now the escape hatch for a default-on feature, it might be worth adding one in a follow-up:
   ```c
   MODULE_PARM_DESC(dpu_use_virtual_planes,
                    "Use virtual planes for improved pipeline utilization (default: true)");
   ```

4. **Parameter permissions are 0.** The `module_param(..., 0)` means the parameter is not exposed in sysfs at all — it's boot-time only. This is fine for this kind of fundamental mode switch, since changing it at runtime would likely be unsafe without a full modeset teardown/rebuild.

5. **Usage sites are clean.** The variable is checked in two places:
   - `dpu_kms.c:852` — choosing between `dpu_plane_init_virtual()` and `dpu_plane_init()` during hardware init.
   - `dpu_crtc.c:1537` — conditionally calling `dpu_crtc_reassign_planes()` during atomic check.
   
   Both are one-shot init or atomic-check paths, so the default flip doesn't introduce any new races or lifecycle concerns.

**Verdict:** The patch is correct and ready. No blocking issues. The only suggestion is the optional `MODULE_PARM_DESC` addition for better self-documentation, but that can be a separate cleanup.

---
Generated by Claude Code Patch Reviewer