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 CPU job teardown bugs Date: Sat, 16 May 2026 08:58:36 +1000 Message-ID: In-Reply-To: <20260515-v3d-cpu-job-leaks-v1-0-7f147cbbf935@igalia.com> References: <20260515-v3d-cpu-job-leaks-v1-0-7f147cbbf935@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 CPU job teardown bugs Author: =3D?utf-8?q?Ma=3DC3=3DADra_Canal?=3D Patches: 3 Reviewed: 2026-05-16T08:58:36.842986 --- This is a clean, well-motivated 2-patch series that fixes real teardown bug= s in V3D's CPU job submission path. Both patches are correct, minimal, and = well-structured. **Patch 1** consolidates the CPU job teardown into the kref destructor (`v3= d_cpu_job_free`), mirroring the existing `v3d_render_job_free` pattern. Thi= s eliminates a use-after-free on the ioctl error path where `cpu_job` was a= ccessed after `v3d_job_cleanup()` had already released it, and also fixes a= syncobj leak since the old manual `kvfree()` calls didn't call `drm_syncob= j_put()` on each query entry. **Patch 2** adds the missing `drm_gem_object_put()` for the indirect CSD BO= reference acquired by `drm_gem_object_lookup()` in `v3d_get_cpu_indirect_c= sd_params()`. The series is correctly ordered =E2=80=94 patch 2 builds on the `v3d_cpu_jo= b_free` kref destructor introduced in patch 1. The commit messages are clea= r and provide good analysis of the bugs. **Verdict: Both patches look good for merging.** No issues found. --- --- Generated by Claude Code Patch Reviewer