From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: Add splash DRM client Date: Sat, 16 May 2026 16:02:57 +1000 Message-ID: In-Reply-To: <20260510-drm_client_splash-v3-0-a9aee9f0b2fc@valla.it> References: <20260510-drm_client_splash-v3-0-a9aee9f0b2fc@valla.it> 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: Add splash DRM client Author: Francesco Valla Patches: 6 Reviewed: 2026-05-16T16:02:57.224131 --- This is a v3 RFC series adding a DRM-based splash screen client (3 patches)= . The concept is sound and addresses a long-standing TODO item. The impleme= ntation follows existing DRM client patterns (modeled after `drm_log.c`) an= d supports three image sources: plain color background, BMP firmware, and E= FI BGRT. The Kconfig structure is reasonable and the code is generally well= -structured. However, there are several issues ranging from potential crashes to data in= tegrity bugs that should be addressed before this moves past RFC stage. **Key concerns:** 1. **Uninitialized thread pointer crash** in unregister path when hotplug n= ever fires. 2. **Out-of-bounds memory read** in BMP pixel blitting (reading 4 bytes fro= m 3-byte pixel data). 3. **Render thread sleep/wake race condition** =E2=80=94 the classic `set_c= urrent_state` vs `wake_up_process` pattern has a window for lost wakeups. 4. **Tiled display dimension accumulation** looks swapped (hdisplay vs vdis= play). 5. **`modeset_mask` / scanout indexing** is broken =E2=80=94 `j` doesn't tr= ack scanout indices correctly. --- --- Generated by Claude Code Patch Reviewer