From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: Revert "accel/amdxdna: Support read-only user-pointer BO mappings" Date: Mon, 25 May 2026 19:46:43 +1000 Message-ID: In-Reply-To: <20260521162930.1451042-1-lizhi.hou@amd.com> References: <20260521162930.1451042-1-lizhi.hou@amd.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: Revert "accel/amdxdna: Support read-only user-pointer BO mappings" Author: Lizhi Hou Patches: 2 Reviewed: 2026-05-25T19:46:43.479862 --- This is a single-patch series that reverts commit `f649e63d4a64` ("accel/am= dxdna: Support read-only user-pointer BO mappings") from the amdxdna accele= rator driver. The patch is mechanically correct as a revert =E2=80=94 it re= moves the `readonly_va_entry()` function, the `readonly` tracking variable,= and changes `pin_user_pages_fast()` and `dma_buf_export()` to always use w= ritable/RDWR semantics. The commit message is minimal: *"The read-only feature requires further con= sideration."* This is vague. For a revert of a feature that was presumably = reviewed and merged, the rationale should explain **what** needs further co= nsideration =E2=80=94 whether there's a correctness bug, a security issue, = an API concern, or a design disagreement. Reviewers and future readers of `= git log` deserve more than this. That said, the reverted code did have real problems worth reverting for, wh= ich makes this a reasonable change on technical grounds even if the commit = message doesn't articulate them. **Verdict:** The revert is mechanically sound and likely justified, but the= commit message needs improvement. --- --- Generated by Claude Code Patch Reviewer