From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: arm64: qcom: Enable additional hardware on Radxa Dragon Q6A Date: Mon, 25 May 2026 19:16:36 +1000 Message-ID: In-Reply-To: <20260522060645.4399-1-graham.oconnor@gmail.com> References: <20260522060645.4399-1-graham.oconnor@gmail.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit MIME-Version: 1.0 Overall Series Review Subject: arm64: qcom: Enable additional hardware on Radxa Dragon Q6A Author: Graham O'Connor Patches: 14 Reviewed: 2026-05-25T19:16:36.146330 --- This is a 6-patch series enabling additional hardware on the Radxa Dragon Q6A (QCS6490) when booting via UEFI. The series covers RSC firmware coexistence, EFI variable access, DP link training, and DTS additions for GPU/display/USB3. The DTS patches (4-6) are generally well-structured and demonstrate real hardware testing. However, there are two significant issues: 1. **Patch 3 (DP_TRAIN_LEVEL_MAX)** changes a global constant affecting *all* MSM DP connections, not just the RA620 bridge. This will regress other platforms that require voltage swing level 3. This needs to be reworked as a per-bridge or per-platform quirk. 2. **Patch 1 (rpmh-rsc)** early-returns from probe, skipping critical initialization including `devm_of_platform_populate()`, IRQ setup, and client cache initialization. Child devices won't probe, and the `rsc_drv` structure left in driver data is only partially initialized, risking use of uninitialized fields if any code later references it. The DTS display pipeline modeling in patch 5 uses `dp-connector` to represent what is described as a bridge, which is semantically questionable and misuses port@1 of the binding. --- --- Generated by Claude Code Patch Reviewer