Claude review: drm/amd/display: Refactor amdgpu_dm_crtc_set_vblank

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

Patch Review

Clean refactoring. The split is logical: `amdgpu_dm_crtc_set_vblank()` does the HW interrupt work, `amdgpu_dm_crtc_queue_vblank_work()` handles the PSR/panel optimization work, and the enable/disable wrappers tie them together.

One minor issue: in `amdgpu_dm_crtc_set_vblank()`, the condition changed from `if (enable)` to `if (enable && crtc->enabled)`, but the `crtc->enabled` check was moved to `amdgpu_dm_crtc_enable_vblank()` which returns `-EINVAL` early. So in the non-deferred path, `amdgpu_dm_crtc_set_vblank()` should never be called with `enable=true` and `!crtc->enabled`. But patch 4 calls `amdgpu_dm_crtc_set_vblank(crtc, true)` from `amdgpu_dm_crtc_deferred_enable_vblank()` *without* the `crtc->enabled` guard. That means the `enable && crtc->enabled` check inside `set_vblank` becomes the only guard for the deferred path, which is fine — but it means a deferred enable on an unconfigured CRTC silently succeeds (returns 0) without actually enabling anything. That seems fragile; it might be better to return an error.

---
Generated by Claude Code Patch Reviewer