[PATCH] drm/amd/amdgpu: S4 issue for amdgpu

Deucher, Alexander Alexander.Deucher at amd.com
Fri Sep 2 13:29:51 UTC 2016


> -----Original Message-----
> From: amd-gfx [mailto:amd-gfx-bounces at lists.freedesktop.org] On Behalf
> Of Qu, Jim
> Sent: Thursday, September 01, 2016 9:07 PM
> To: Alex Deucher
> Cc: amd-gfx list
> Subject: 答复: [PATCH] drm/amd/amdgpu: S4 issue for amdgpu
> 
> Hi Alex:
> 
> Weird, last day, I used the fiji to check the S4 issue, there are some atom
> errors, but it did not observed if add atombios_fini/init.
> today, I change another fiji and  also check the issue with the patch which
> comment out the atombios_fini/init. but could not see the atombios error.
> I will confirm it and update the patch.
> 
> Error like:
> 
> [   81.167263] [drm:atom_op_jump [amdgpu]] *ERROR* atombios stuck in
> loop for more than 5secs aborting
> [   81.167276] [drm:amdgpu_atom_execute_table_locked [amdgpu]]
> *ERROR* atombios stuck executing D208 (len 239, WS 4, PS 0) @ 0xD2AB
> [   81.167288] [drm:amdgpu_atom_execute_table_locked [amdgpu]]
> *ERROR* atombios stuck executing B586 (len 96, WS 0, PS 4) @ 0xB5C2
> [   81.167299] [drm:amdgpu_atom_execute_table_locked [amdgpu]]
> *ERROR* atombios stuck executing B104 (len 183, WS 0, PS 8) @ 0xB1AC

Errors like that are usually a symptom of the GPU being is a bad state rather than some software state.

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9月1日 21:42:46
> 收件人: Qu, Jim
> 抄送: amd-gfx list
> 主题: Re: [PATCH] drm/amd/amdgpu: S4 issue for amdgpu
> 
> On Thu, Sep 1, 2016 at 2:22 AM, jimqu <Jim.Qu at amd.com> wrote:
> > reset the asic if adapter is not powerdown when doing freeze()
> > thaw() and restore(), in order to get a valid state of adapter.
> >
> > Change-Id: Id2e4587c06b556eb651ab9277ffc903f780dea56
> > Signed-off-by: JimQu <Jim.Qu at amd.com>
> > ---
> >  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 23
> +++++++++++++++++------
> >  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    |  4 ++--
> >  2 files changed, 19 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> > index 52cec16..5fc76bbf 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> > @@ -1936,8 +1936,14 @@ int amdgpu_device_suspend(struct drm_device
> *dev, bool suspend, bool fbcon)
> >                 /* Shut down the device */
> >                 pci_disable_device(dev->pdev);
> >                 pci_set_power_state(dev->pdev, PCI_D3hot);
> > +       } else {
> > +               r = amdgpu_asic_reset(adev);
> > +               if (r)
> > +                       DRM_ERROR("amdgpu asic reset failed\n");
> >         }
> >
> > +       amdgpu_atombios_fini(adev);
> > +
> 
> Why do you need atombios_fini() here?  All it does it free the driver
> allocations.
> 
> >         if (fbcon) {
> >                 console_lock();
> >                 amdgpu_fbdev_set_suspend(adev, 1);
> > @@ -1965,21 +1971,23 @@ int amdgpu_device_resume(struct drm_device
> *dev, bool resume, bool fbcon)
> >         if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
> >                 return 0;
> >
> > -       if (fbcon) {
> > +       if (fbcon)
> >                 console_lock();
> > -       }
> > +
> > +       amdgpu_atombios_init(adev);
> > +
> 
> Same question here.  This doesn't actually change any hw state.
> 
> >         if (resume) {
> >                 pci_set_power_state(dev->pdev, PCI_D0);
> >                 pci_restore_state(dev->pdev);
> > -               if (pci_enable_device(dev->pdev)) {
> > +               if (r = pci_enable_device(dev->pdev)) {
> >                         if (fbcon)
> >                                 console_unlock();
> > -                       return -1;
> > +                       return r;
> >                 }
> >         }
> >
> >         /* post card */
> > -       if (!amdgpu_card_posted(adev))
> > +       if (!amdgpu_card_posted(adev) || !resume)
> >                 amdgpu_atom_asic_init(adev->mode_info.atom_context);
> >
> >         r = amdgpu_resume(adev);
> > @@ -1998,8 +2006,11 @@ int amdgpu_device_resume(struct drm_device
> *dev, bool resume, bool fbcon)
> >         }
> >
> >         r = amdgpu_late_init(adev);
> > -       if (r)
> > +       if (r) {
> > +               if (fbcon)
> > +                       console_unlock();
> >                 return r;
> > +       }
> >
> >         /* pin cursors */
> >         list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> > index 06816cf..02941db3 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> > @@ -601,8 +601,8 @@ static const struct dev_pm_ops amdgpu_pm_ops =
> {
> >         .resume = amdgpu_pmops_resume,
> >         .freeze = amdgpu_pmops_freeze,
> >         .thaw = amdgpu_pmops_thaw,
> > -       .poweroff = amdgpu_pmops_freeze,
> > -       .restore = amdgpu_pmops_resume,
> > +       .poweroff = amdgpu_pmops_suspend,
> > +       .restore = amdgpu_pmops_thaw,
> >         .runtime_suspend = amdgpu_pmops_runtime_suspend,
> >         .runtime_resume = amdgpu_pmops_runtime_resume,
> >         .runtime_idle = amdgpu_pmops_runtime_idle,
> > --
> > 1.9.1
> >
> > _______________________________________________
> > amd-gfx mailing list
> > amd-gfx at lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/amd-gfx
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
> _______________________________________________
> 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