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/rockchip: dw_dp: Add runtime PM support Date: Tue, 05 May 2026 09:49:47 +1000 Message-ID: In-Reply-To: <20260501-synopsys-dw-dp-improvements-v2-10-d7e7f6bac77f@collabora.com> References: <20260501-synopsys-dw-dp-improvements-v2-0-d7e7f6bac77f@collabora.com> <20260501-synopsys-dw-dp-improvements-v2-10-d7e7f6bac77f@collabora.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Patch Review Rockchip glue for runtime PM. Restores HPD GRF state on resume. **Issue =E2=80=94 `ACQUIRE(pm_runtime_active_auto, pm)` in hpd_sw_sel/hpd_s= w_cfg:** ```c + ACQUIRE(pm_runtime_active_auto, pm)(dp->dev); + regmap_write(dp->vo_grf, hpd_reg, ...); ``` The `ACQUIRE` guard will `pm_runtime_get_sync` then `pm_runtime_put_autosus= pend` when the function scope ends. But since `hpd_sw_sel` can be called du= ring `dw_dp_bind` *before* `pm_runtime_enable` is called (looking at patch = 8, `hpd_sw_sel` is called in `dw_dp_bind` when USB-C is detected, and patch= 10 puts `pm_runtime_enable` *after* `dw_dp_bind`), the runtime PM call wou= ld operate on a device that hasn't had PM enabled yet. This could cause iss= ues depending on the PM subsystem's behavior with uninitialized state. The = ordering in `dw_dp_rockchip_bind` should be verified =E2=80=94 PM should be= enabled before anything that might call these handlers. --- Generated by Claude Code Patch Reviewer