[PATCH] drm/amd: Disable S/G for APUs when 64GB or more host memory

[Alex Deucher]

On Thu, Jul 27, 2023 at 1:06 PM Mario Limonciello
<mario.limonciello at amd.com> wrote:
>
> Users report a white flickering screen on multiple systems that
> is tied to having 64GB or more memory.  When S/G is enabled pages
> will get pinned to both VRAM carve out and system RAM leading to
> this.
>
> Until it can be fixed properly, disable S/G when 64GB of memory or
> more is detected.  This will force pages to be pinned into VRAM.
> This should fix white screen flickers but if VRAM pressure is
> encountered may lead to black screens.  It's a trade-off for now.
>
> Fixes: 81d0bcf990093 ("drm/amdgpu: make display pinning more flexible (v2)")
> Cc: Hamza Mahfooz <Hamza.Mahfooz at amd.com>
> Cc: Roman Li <roman.li at amd.com>
> Cc: <stable at vger.kernel.org> # 6.1.y: bf0207e172703 ("drm/amdgpu: add S/G display parameter")
> Cc: <stable at vger.kernel.org> # 6.4.y
> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2735
> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2354
> Signed-off-by: Mario Limonciello <mario.limonciello at amd.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu.h           |  1 +
>  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c    | 25 +++++++++++++++++++
>  .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c |  6 +----
>  3 files changed, 27 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 93d0f4c7b560e..2e3c7c15cb8e3 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -1313,6 +1313,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
>  void amdgpu_device_pci_config_reset(struct amdgpu_device *adev);
>  int amdgpu_device_pci_reset(struct amdgpu_device *adev);
>  bool amdgpu_device_need_post(struct amdgpu_device *adev);
> +bool amdgpu_sg_display_supported(struct amdgpu_device *adev);
>  bool amdgpu_device_pcie_dynamic_switching_supported(void);
>  bool amdgpu_device_should_use_aspm(struct amdgpu_device *adev);
>  bool amdgpu_device_aspm_support_quirk(void);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index dc0e5227119b1..a4e36b178d86c 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -1296,6 +1296,31 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev)
>         return true;
>  }
>
> +/*
> + * On APUs with >= 64GB white flickering has been observed w/ SG enabled.
> + * Disable S/G on such systems until we have a proper fix.
> + * https://gitlab.freedesktop.org/drm/amd/-/issues/2354
> + * https://gitlab.freedesktop.org/drm/amd/-/issues/2735
> + */
> +bool amdgpu_sg_display_supported(struct amdgpu_device *adev)
> +{
> +       switch (amdgpu_sg_display) {
> +       case -1:
> +               break;
> +       case 0:
> +               return false;
> +       case 1:
> +               return true;
> +       default:
> +               return false;
> +       }
> +       if (totalram_pages() << (PAGE_SHIFT - 10) >= 64000000) {
> +               DRM_WARN("Disabling S/G due to >=64GB RAM\n");
> +               return false;
> +       }
> +       return true;
> +}
> +
>  /*
>   * Intel hosts such as Raptor Lake and Sapphire Rapids don't support dynamic
>   * speed switching. Until we have confirmation from Intel that a specific host
> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> index 77d970a2ee693..f81c6c249c02a 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> @@ -1639,12 +1639,8 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
>                 }
>                 break;
>         }
> -       if (init_data.flags.gpu_vm_support &&
> -           (amdgpu_sg_display == 0))
> -               init_data.flags.gpu_vm_support = false;
> -
>         if (init_data.flags.gpu_vm_support)
> -               adev->mode_info.gpu_vm_support = true;
> +               init_data.flags.gpu_vm_support = amdgpu_sg_display_supported(adev);
>

You miss updating adev->mode_info.gpu_vm_support here.

Alex

>         if (amdgpu_dc_feature_mask & DC_FBC_MASK)
>                 init_data.flags.fbc_support = true;
> --
> 2.25.1
>