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/v3d: Fix indirect CSD jobs with zeroed workgroups Date: Thu, 04 Jun 2026 15:15:36 +1000 Message-ID: In-Reply-To: <20260530-v3d-fix-indirect-csd-v1-0-15533948663f@igalia.com> References: <20260530-v3d-fix-indirect-csd-v1-0-15533948663f@igalia.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/v3d: Fix indirect CSD jobs with zeroed workgroups Author: =3D?utf-8?q?Ma=3DC3=3DADra_Canal?=3D Patches: 4 Reviewed: 2026-06-04T15:15:36.533079 --- This is a clean, well-motivated two-patch series fixing real bugs in the V3= D indirect CSD (Compute Shader Dispatch) path. Indirect CSD defers workgrou= p counts to a GPU buffer that may legitimately contain zeros at runtime. Th= e series fixes two problems: (1) a vaddr mapping leak when the early-return= path skips cleanup, and (2) submitting zero-count dispatches to hardware w= here the V3D block interprets 0 as 65536, causing invalid execution. Both patches are small, well-targeted, and have appropriate `Cc: stable` an= d `Fixes:` tags. The series is correctly ordered =E2=80=94 patch 1 fixes th= e leak in the CPU-side rewrite function, and patch 2 adds the guard in the = hardware submission path. One minor concern worth raising on patch 2, detailed below. **Verdict: Series looks good. One minor style/consistency question on patch= 2.** --- --- Generated by Claude Code Patch Reviewer