Hi Andy, Yes, the issue can be reproduced with the details in patch 2/3 [1]. [1]: https://lore.kernel.org/linux-arm-kernel/20241115162120.83990-3-detlev.casanova@xxxxxxxxxxxxx/T/#m82b38f4a83c4793bb82919bf736b2f6bd804a283 Detlev. On Monday, 25 November 2024 02:55:41 EST Andy Yan wrote: > Hello Detlev, > > At 2024-11-16 00:20:39, "Detlev Casanova" <detlev.casanova@xxxxxxxxxxxxx> wrote: > >The clock reset must be used when the VOP is configured. Skipping it can > >put the VOP in an unknown state where the HDMI signal is either lost or > >not matching the selected mode. > > Can you provide some detail about how to reproduce this issue ? > If we can reproduce this issue, we might be able to do a more deep analysis. > >This adds support for rk3588(s) based SoCs. > > > >Changes since v3: > >- Rebased on drm-misc-next > >- Reword first patch subject > >- Reorder commits for different trees > > > >Changes since v2: > >- Rebase on latest master > >- Add details on how to reproduce the issue > > > >Changes since v1: > >- Add AXI and AHB clock resets > >- Set maxItems for !rk3588 in vop2 bindings > > > >Detlev Casanova (3): > > dt-bindings: display: vop2: Add VP clock resets > > drm/rockchip: vop2: Add clock resets support > > arm64: dts: rockchip: Add VOP clock resets for rk3588s > > > > .../display/rockchip/rockchip-vop2.yaml | 40 +++++++++++++++++++ > > arch/arm64/boot/dts/rockchip/rk3588-base.dtsi | 12 ++++++ > > drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 30 ++++++++++++++ > > 3 files changed, 82 insertions(+)
