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: Define user readable error codes for atomic ioctl Date: Wed, 01 Apr 2026 07:55:27 +1000 Message-ID: In-Reply-To: <20260331-atomic-v11-1-6a1df7ec5af8@intel.com> References: <20260331-atomic-v11-0-6a1df7ec5af8@intel.com> <20260331-atomic-v11-1-6a1df7ec5af8@intel.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit MIME-Version: 1.0 Patch Review **Typos in UAPI enum names (permanent ABI breakage):** ```c DRM_MODE_ATTOMIC_CONNECTOR_BW, DRM_MODE_ATTOMIC_PIPE_BW, DRM_MODE_ATOMIC_SPEC_VIOLOATION, ``` These should be `DRM_MODE_ATOMIC_CONNECTOR_BW`, `DRM_MODE_ATOMIC_PIPE_BW`, and `DRM_MODE_ATOMIC_SPEC_VIOLATION`. Once merged, these misspellings are permanent UAPI. **Typos in kernel-doc:** ```c * @DRM_MODE_ATOMIC_INVALID_API_USAGE: invallid API usage(DRM_ATOMIC not * enabled, invalid falg, page_flip event ``` "invallid" should be "invalid", "falg" should be "flag". **Kernel-doc for `drm_mode_atomic_err_code`:** ```c * @failure_code: error codes defined in enum drm_moide_atomic_failure_code ``` "drm_moide_atomic_failure_code" should be "drm_mode_atomic_failure_codes". **Design concern with `failure_objs_ptr`:** This field is a `__u64` pointer to DRM objects, but there's no documentation on what the pointed-to data looks like. Is it an array of `__u32` object IDs? What allocates/frees this memory? None of the patches in this series actually populate it, so it's dead UAPI surface area. **`__u64` for `failure_code`:** The enum only has 9 values. Using `__u32` would be more natural and leave the `__u64 reserved` for actual expansion. The current layout wastes space. **`failure_string` as a fixed 128-byte char array inside a struct that gets `copy_to_user`'d:** This is 128 bytes copied on every failed atomic commit. Consider whether this is the right tradeoff vs. a userspace-provided buffer. --- --- Generated by Claude Code Patch Reviewer