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/nouveau/disp/r535: Add scanline position support + head state support Date: Tue, 05 May 2026 12:01:33 +1000 Message-ID: In-Reply-To: <20260429030348.3930866-1-lyude@redhat.com> References: <20260429030348.3930866-1-lyude@redhat.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Overall Series Review Subject: drm/nouveau/disp/r535: Add scanline position support + head state = support Author: Lyude Paul Patches: 1 Reviewed: 2026-05-05T12:01:33.971366 --- This is a single-patch fix from Lyude Paul that adds scanline position (`.r= gpos`) and head state (`.state`) support to the r535 GSP-RM display path in= nouveau. The r535 code path previously had an empty `.state` stub and no `= .rgpos` at all, causing `drm_crtc_vblank_helper_get_vblank_timestamp_intern= al` to fall back to the less accurate DRM core fallback for vblank timestam= ps. The fix is clean and minimal: it removes the `static` keyword from `gv100_h= ead_state()` and `gv100_head_rgpos()` in `gv100.c`, adds their declarations= to `head.h`, and wires them into the `r535_head` function table. This is t= he same pattern used by the non-GSP gv100 head, so it's well-established co= de being reused. The approach of reading the MMIO registers directly (bypassing GSP RPC) is = well-justified in the commit message =E2=80=94 vblank/scanline queries are = latency-sensitive and don't need to go through the GSP firmware. **Verdict: Looks good.** This is a correct, well-reasoned bug fix suitable = for stable backport. --- Generated by Claude Code Patch Reviewer