From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: Rust bindings for gem shmem Date: Thu, 23 Apr 2026 08:05:54 +1000 Message-ID: In-Reply-To: <20260421235346.672794-1-lyude@redhat.com> References: <20260421235346.672794-1-lyude@redhat.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: Rust bindings for gem shmem Author: Lyude Paul Patches: 6 Reviewed: 2026-04-23T08:05:54.209470 --- This is v12 of the Rust bindings for GEM shmem, adding SGTable and VMap sup= port on top of the existing shmem bindings. The series is well-structured: = two C-side preparatory patches (extracting `__drm_gem_shmem_free_sgt_locked= ` and exporting `drm_gem_shmem_get_pages_sgt_locked`), followed by three Ru= st patches building SGTable and VMap abstractions. The overall design is sound =E2=80=94 using `Devres` to ensure SGTable mapp= ings are revoked on driver unbind is the right approach for safety. The VMa= p abstraction with borrowed vs. owned variants and integration with the `Io= ` trait is clean. There are a few concerns worth discussing, primarily around locking pattern= s in the Rust code (non-interruptible `dma_resv_lock` calls, lock not relea= sed on error paths), a naming mismatch in the C-side kdoc, and some safety = reasoning that could be tightened. Nothing is a showstopper, but a couple o= f items warrant attention before merging. --- Generated by Claude Code Patch Reviewer