答复: 答复: [PATCH 1/2] drm/amdgpu: reset the asic when hibernating

Qu, Jim Jim.Qu at amd.com
Tue Aug 23 05:34:36 UTC 2016


All right, from my previous experience. In thaw() and restore(), there are always ring test fail on GFX, SDMA , sometimes fails on VCE.
I will try your change. Hope it is benefit to the issue.


Thanks
JimQu

________________________________________
发件人: Alex Deucher <alexdeucher at gmail.com>
发送时间: 2016年8月23日 12:46:43
收件人: Qu, Jim
抄送: amd-gfx at lists.freedesktop.org; Deucher, Alexander
主题: Re: 答复: [PATCH 1/2] drm/amdgpu: reset the asic when hibernating

On Tue, Aug 23, 2016 at 12:18 AM, Qu, Jim <Jim.Qu at amd.com> wrote:
> Hi Alex:
>
> Are you working on S4 issue ?

No, not specifically.  Just seemed like a more reliable way to deal with it.

Alex

>
> Thanks
> JimQu
>
> ________________________________________
> 发件人: amd-gfx <amd-gfx-bounces at lists.freedesktop.org> 代表 Alex Deucher <alexdeucher at gmail.com>
> 发送时间: 2016年8月23日 0:05:24
> 收件人: amd-gfx at lists.freedesktop.org
> 抄送: Deucher, Alexander
> 主题: [PATCH 1/2] drm/amdgpu: reset the asic when hibernating
>
> Some blocks require a reset to properly resume if there
> is no power down of the asic like during various
> hibernation steps.
>
> Port of:
> 274ad65c9d02bdcbee9bae045517864c3521d530
> (drm/radeon: hard reset r600 and newer GPU when hibernating.)
> from radeon.
>
> Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu.h        |  3 ++-
>  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 10 +++++++---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    |  6 +++---
>  3 files changed, 12 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> index 6fa8268..d0d0588 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
> @@ -2436,7 +2436,8 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
>                                  struct drm_file *file_priv);
>  void amdgpu_driver_preclose_kms(struct drm_device *dev,
>                                 struct drm_file *file_priv);
> -int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon);
> +int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon,
> +                      bool freeze);
>  int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon);
>  u32 amdgpu_get_vblank_counter_kms(struct drm_device *dev, unsigned int pipe);
>  int amdgpu_enable_vblank_kms(struct drm_device *dev, unsigned int pipe);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 21f4b74..5747e37 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -1037,7 +1037,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
>                 printk(KERN_INFO "amdgpu: switched off\n");
>                 drm_kms_helper_poll_disable(dev);
>                 dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
> -               amdgpu_suspend_kms(dev, true, true);
> +               amdgpu_suspend_kms(dev, true, true, false);
>                 dev->switch_power_state = DRM_SWITCH_POWER_OFF;
>         }
>  }
> @@ -1780,7 +1780,8 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
>   * Returns 0 for success or an error on failure.
>   * Called at driver suspend.
>   */
> -int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
> +int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon,
> +                      bool freeze)
>  {
>         struct amdgpu_device *adev;
>         struct drm_crtc *crtc;
> @@ -1844,7 +1845,10 @@ int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
>         amdgpu_bo_evict_vram(adev);
>
>         pci_save_state(dev->pdev);
> -       if (suspend) {
> +       if (freeze) {
> +               /* reset the asic during freeze */
> +               r = amdgpu_asic_reset(adev);
> +       } else if (suspend) {
>                 /* Shut down the device */
>                 pci_disable_device(dev->pdev);
>                 pci_set_power_state(dev->pdev, PCI_D3hot);
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index 58b1db8..eb4c9b0 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -395,7 +395,7 @@ static int amdgpu_pmops_suspend(struct device *dev)
>  {
>         struct pci_dev *pdev = to_pci_dev(dev);
>         struct drm_device *drm_dev = pci_get_drvdata(pdev);
> -       return amdgpu_suspend_kms(drm_dev, true, true);
> +       return amdgpu_suspend_kms(drm_dev, true, true, false);
>  }
>
>  static int amdgpu_pmops_resume(struct device *dev)
> @@ -409,7 +409,7 @@ static int amdgpu_pmops_freeze(struct device *dev)
>  {
>         struct pci_dev *pdev = to_pci_dev(dev);
>         struct drm_device *drm_dev = pci_get_drvdata(pdev);
> -       return amdgpu_suspend_kms(drm_dev, false, true);
> +       return amdgpu_suspend_kms(drm_dev, false, true, true);
>  }
>
>  static int amdgpu_pmops_thaw(struct device *dev)
> @@ -434,7 +434,7 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
>         drm_kms_helper_poll_disable(drm_dev);
>         vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_OFF);
>
> -       ret = amdgpu_suspend_kms(drm_dev, false, false);
> +       ret = amdgpu_suspend_kms(drm_dev, false, false, false);
>         pci_save_state(pdev);
>         pci_disable_device(pdev);
>         pci_ignore_hotplug(pdev);
> --
> 2.5.5
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx


More information about the amd-gfx mailing list