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/panthor: Fix the "done_fence is initialized" detection logic Date: Tue, 10 Mar 2026 12:25:51 +1000 Message-ID: In-Reply-To: <20260309103053.211415-1-boris.brezillon@collabora.com> References: <20260309103053.211415-1-boris.brezillon@collabora.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/panthor: Fix the "done_fence is initialized" detection logic Author: Boris Brezillon Patches: 8 Reviewed: 2026-03-10T12:25:51.999443 --- This is a single-patch fix for the panthor GPU scheduler addressing a regre= ssion introduced by commit 541c8f2468b9 ("dma-buf: detach fence ops on sign= al v3"). That dma-buf change nullifies `dma_fence::ops` on signal, which br= oke panthor's `job_release()` logic that relied on `ops !=3D NULL` to deter= mine whether a fence was initialized (via `dma_fence_init()`) vs merely all= ocated (via `kzalloc`). The approach =E2=80=94 encoding an "initialized" flag in the low bit of the= pointer =E2=80=94 is a well-known kernel pattern and is sound given `dma_f= ence` alignment guarantees. The patch is generally well-structured, but I h= ave concerns about correctness in certain paths. **Verdict:** The fix addresses a real bug and the approach is reasonable, b= ut there are issues that should be addressed before merging. --- Generated by Claude Code Patch Reviewer