Claude review: clk: qcom: common: ensure runtime PM suspend completes on probe

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

Patch Review

```c
-		pm_runtime_put(dev);
+		pm_runtime_put_sync(dev);
```

**Concern: This is a broad change affecting all `use_rpm` clock controllers.** The `qcom_cc_really_probe()` function is shared infrastructure used by every Qualcomm clock controller with `use_rpm = true`. The commit message motivates this change specifically for gxclkctl/IFPC, but the change applies globally.

For most clock controllers, the async `pm_runtime_put()` is preferable because it avoids blocking probe unnecessarily. The synchronous variant adds latency to every such probe call. If the requirement is specific to gxclkctl (which Patch 4 then disables RPM on), consider whether this should be scoped more narrowly -- perhaps by adding a flag to `qcom_cc_desc` to opt into synchronous put, or by doing the sync in the gxclkctl probe itself.

Also, no Fixes tag is present. Is this a standalone fix or only needed as a prerequisite for Patch 4?

---
Generated by Claude Code Patch Reviewer