Claude review: drm/vkms: Backport generic vblank timer to fix ABBA deadlock

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

Overall Series Review

Subject: drm/vkms: Backport generic vblank timer to fix ABBA deadlock
Author: w15303746062@163.com
Patches: 6
Reviewed: 2026-05-27T14:49:44.274178

---

This is a **stable backport (6.18.y)** of 5 upstream commits that introduce the generic DRM vblank timer infrastructure and convert vkms to use it, fixing an ABBA deadlock between `vblank_time_lock` and `hrtimer_cancel()` in vkms. The original author is Thomas Zimmermann; Mingyu Wang is cherry-picking with appropriate `Signed-off-by` trailers.

All 5 patches are confirmed present in drm-next already (verified by inspecting the kernel tree), so this is a legitimate cherry-pick series. The commit ordering is correct and the dependency chain is sound: patches 1-2 introduce the infrastructure, patch 3 converts vkms, patch 4 fixes a timeout regression from the conversion, and patch 5 is a trivial doc fix.

**Procedural concern:** This series is addressed to `stable@vger.kernel.org` targeting **6.18.y** but 6.18 has not been released yet at time of submission (the -rc cycle would still be ongoing). Stable backports typically target released kernels. These commits should land via the normal merge path through drm-next into 6.18-rc/6.18, not as stable backports. If the author wants them in an already-released stable kernel (e.g., 6.12.y or 6.15.y), the target branch is wrong. If the patches are for the not-yet-released 6.18, they should go through the normal drm tree merge process rather than stable.

**The patches themselves are clean cherry-picks of reviewed/tested upstream code and are technically sound.**

---
Generated by Claude Code Patch Reviewer