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/vblank: Deferred Enable and Disable Date: Wed, 25 Mar 2026 07:22:01 +1000 Message-ID: In-Reply-To: <20260323202755.315929-1-sunpeng.li@amd.com> References: <20260323202755.315929-1-sunpeng.li@amd.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Overall Series Review Subject: drm/vblank: Deferred Enable and Disable Author: Patches: 7 Reviewed: 2026-03-25T07:22:01.297418 --- This series introduces deferred vblank enable/disable for DRM, motivated by= AMD's IPS (Idle Power States) feature where the DCN block can be power-gat= ed and re-enabling requires acquiring a mutex (blocking operation) that can= not be done from atomic context. The approach defers `drm_vblank_enable()` = and `vblank_disable_fn()` to a workqueue, with pre/post hooks for drivers t= o do blocking work. The design is reasonable for the stated problem, but there are several conc= urrency concerns in patch 2 that need attention =E2=80=94 particularly arou= nd races between deferred enable, deferred disable, and `drm_crtc_vblank_of= f()`. The series also unconditionally allocates a workqueue for all DRM dev= ices even when no driver uses deferred vblank. --- Generated by Claude Code Patch Reviewer