Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
MIME-Version: 1.0
References: <20220407121230.132627-1-kai.heng.feng@canonical.com>
In-Reply-To: <20220407121230.132627-1-kai.heng.feng@canonical.com>
From:   Alex Deucher <alexdeucher@gmail.com>
Date:   Thu, 7 Apr 2022 12:04:22 -0400
Message-ID: <CADnq5_P7jM8-FCW1i3Lm7MBrWYd2RKsFvBdTJJp1rp6tPdn2tQ@mail.gmail.com>
Subject: Re: [PATCH] drm/amdgpu: Ensure HDA function is suspended before ASIC reset
To:     Kai-Heng Feng <kai.heng.feng@canonical.com>
Cc:     "Deucher, Alexander" <alexander.deucher@amd.com>,
        Christian Koenig <christian.koenig@amd.com>,
        xinhui pan <Xinhui.Pan@amd.com>,
        David Airlie <airlied@linux.ie>,
        Maling list - DRI developers 
        <dri-devel@lists.freedesktop.org>,
        LKML <linux-kernel@vger.kernel.org>,
        amd-gfx list <amd-gfx@lists.freedesktop.org>,
        Solomon Chiu <solomon.chiu@amd.com>,
        Mario Limonciello <mario.limonciello@amd.com>,
        Evan Quan <evan.quan@amd.com>,
        Luben Tuikov <luben.tuikov@amd.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Applied.  Thanks!

Alex

On Thu, Apr 7, 2022 at 8:21 AM Kai-Heng Feng
<kai.heng.feng@canonical.com> wrote:
>
> DP/HDMI audio on AMD PRO VII stops working after S3:
> [  149.450391] amdgpu 0000:63:00.0: amdgpu: MODE1 reset
> [  149.450395] amdgpu 0000:63:00.0: amdgpu: GPU mode1 reset
> [  149.450494] amdgpu 0000:63:00.0: amdgpu: GPU psp mode1 reset
> [  149.983693] snd_hda_intel 0000:63:00.1: refused to change power state from D0 to D3hot
> [  150.003439] amdgpu 0000:63:00.0: refused to change power state from D0 to D3hot
> ...
> [  155.432975] snd_hda_intel 0000:63:00.1: CORB reset timeout#2, CORBRP = 65535
>
> The offending commit is daf8de0874ab5b ("drm/amdgpu: always reset the asic in
> suspend (v2)"). Commit 34452ac3038a7 ("drm/amdgpu: don't use BACO for
> reset in S3 ") doesn't help, so the issue is something different.
>
> Assuming that to make HDA resume to D0 fully realized, it needs to be
> successfully put to D3 first. And this guesswork proves working, by
> moving amdgpu_asic_reset() to noirq callback, so it's called after HDA
> function is in D3.
>
> Fixes: daf8de0874ab5b ("drm/amdgpu: always reset the asic in suspend (v2)")
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 18 ++++++++++++------
>  1 file changed, 12 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index bb1c025d90019..31f7229e7ea89 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -2323,18 +2323,23 @@ static int amdgpu_pmops_suspend(struct device *dev)
>  {
>         struct drm_device *drm_dev = dev_get_drvdata(dev);
>         struct amdgpu_device *adev = drm_to_adev(drm_dev);
> -       int r;
>
>         if (amdgpu_acpi_is_s0ix_active(adev))
>                 adev->in_s0ix = true;
>         else
>                 adev->in_s3 = true;
> -       r = amdgpu_device_suspend(drm_dev, true);
> -       if (r)
> -               return r;
> +       return amdgpu_device_suspend(drm_dev, true);
> +}
> +
> +static int amdgpu_pmops_suspend_noirq(struct device *dev)
> +{
> +       struct drm_device *drm_dev = dev_get_drvdata(dev);
> +       struct amdgpu_device *adev = drm_to_adev(drm_dev);
> +
>         if (!adev->in_s0ix)
> -               r = amdgpu_asic_reset(adev);
> -       return r;
> +               return amdgpu_asic_reset(adev);
> +
> +       return 0;
>  }
>
>  static int amdgpu_pmops_resume(struct device *dev)
> @@ -2575,6 +2580,7 @@ static const struct dev_pm_ops amdgpu_pm_ops = {
>         .prepare = amdgpu_pmops_prepare,
>         .complete = amdgpu_pmops_complete,
>         .suspend = amdgpu_pmops_suspend,
> +       .suspend_noirq = amdgpu_pmops_suspend_noirq,
>         .resume = amdgpu_pmops_resume,
>         .freeze = amdgpu_pmops_freeze,
>         .thaw = amdgpu_pmops_thaw,
> --
> 2.34.1
>