From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: drm: Improve logic behind damage handling Date: Sat, 16 May 2026 15:34:43 +1000 Message-ID: In-Reply-To: <20260511122421.114014-1-tzimmermann@suse.de> References: <20260511122421.114014-1-tzimmermann@suse.de> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit MIME-Version: 1.0 Overall Series Review Subject: drm: Improve logic behind damage handling Author: Thomas Zimmermann Patches: 11 Reviewed: 2026-05-16T15:34:43.959666 --- This is a well-structured cleanup series from Thomas Zimmermann that consolidates and simplifies the DRM damage clipping logic. The core problem is real: damage evaluation was happening at the wrong point in the atomic check pipeline (before driver atomic_check could trigger modesets), leading to inconsistent state visible to different pipeline stages and drivers working around the framework. The approach is sound: 1. Stop mutating `fb_damage_clips`; use the `ignore_damage_clips` flag instead (patch 1) 2. Move damage evaluation to after all atomic_check helpers (patches 2, 4, 8) 3. Clean up drivers that no longer need old_state (patches 3, 5, 6, 7, 10) 4. Cosmetic rename (patch 9) The series is generally clean and correct. There are a few observations below, one being a potential resource leak bug in the appletbdrm patch. The removal of unit tests in patch 4 is justified since the src-moved tests now exercise a different code path (the check happens at check time, not iter_init time). **Recommendation**: Mostly looks good. One bug to fix in patch 5 (appletbdrm error path leak), and a minor nit on patch ordering. Otherwise, ready for merge after addressing those. --- Generated by Claude Code Patch Reviewer