From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: Add support for DisplayPort link training information report Date: Sun, 12 Apr 2026 10:51:23 +1000 Message-ID: In-Reply-To: <20260409-feat_link_cap-v1-0-7069e8199ce2@bootlin.com> References: <20260409-feat_link_cap-v1-0-7069e8199ce2@bootlin.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: Add support for DisplayPort link training information report Author: Kory Maincent Patches: 18 Reviewed: 2026-04-12T10:51:23.842996 --- This is a 12-patch RFC series introducing a generic DRM framework for exposing DisplayPort link training state to userspace via connector properties. The approach is structurally sound and addresses a real gap -- DP link parameters are currently only accessible through driver-specific debugfs, with no standard interface. The series is organized into three phases: bugfixes (1-3), i915 drmm conversion (4-8), and the new DP connector framework + driver adoption (9-12). The structural decomposition is sensible. **Key concerns:** 1. **No userspace notification mechanism.** Properties are updated via `drm_object_property_set_value()` which does not generate a uevent. Userspace has no way to know when link training completes or values change -- it must poll. This fundamentally limits the usefulness of the feature and should be addressed in the design. 2. **Bitmask semantics are confusing.** The same bitmask property type is used both for capabilities (supported values) and runtime state (current value). Setting `nlanes` to `BIT(1)` to mean "2 lanes active" is unintuitive. An enum property for current state and a bitmask for capabilities would be clearer. 3. **Several bugs in the driver adoption patches** (10, 12), including uninitialized stack variables, unchecked return values, and an ops overwrite bug. 4. **The i915 drmm conversion (patches 4-8) is large and largely AI-generated** ("Assisted-by: Claude Code"). While the approach is correct, this is safety-critical code (resource lifetime management) that needs careful per-platform testing before merging. 5. **Missing UAPI documentation.** New connector properties visible to userspace should be documented and ideally have IGT test coverage. --- --- Generated by Claude Code Patch Reviewer