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/panfrost: Move shrinker initialization and unplug one level down Date: Fri, 05 Jun 2026 06:16:24 +1000 Message-ID: In-Reply-To: <20260604-claude-fixes-v2-3-57c6bd4c1655@collabora.com> References: <20260604-claude-fixes-v2-0-57c6bd4c1655@collabora.com> <20260604-claude-fixes-v2-3-57c6bd4c1655@collabora.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit MIME-Version: 1.0 Patch Review **Verdict: Looks correct.** Moves `panfrost_gem_shrinker_init/fini` into `panfrost_device_init/fini` (mirroring Panthor), and renames `panfrost_gem_shrinker_cleanup` to `panfrost_gem_shrinker_fini` for consistency. The rationale is that the shrinker should be ready before `drm_dev_register()` is called. The init/fini ordering looks correct: - Init: `...perfcnt_init -> shrinker_init -> gem_init` - Fini: `shrinker_fini -> perfcnt_fini -> jm_fini -> ...` (reverse order) The function signatures are changed from `struct drm_device *` to `struct panfrost_device *` which makes sense since `panfrost_device_init` already works with `panfrost_device`. **Note:** The error path in `panfrost_probe` loses the `drm_dev_unregister` call in `err_out2`. That's fine because `err_out2` was only for `panfrost_gem_shrinker_init` failure, which now happens inside `panfrost_device_init` (before `drm_dev_register` is called), so `drm_dev_unregister` would be wrong there. --- --- Generated by Claude Code Patch Reviewer