> -----Original Message-----
> From: Vivi, Rodrigo <rodrigo.vivi@xxxxxxxxx>
> Sent: Tuesday, August 15, 2023 7:28 PM
> To: Gupta, Anshuman <anshuman.gupta@xxxxxxxxx>
> Cc: intel-gfx@xxxxxxxxxxxxxxxxxxxxx; Nilawar, Badal
> <badal.nilawar@xxxxxxxxx>; Tauro, Riana <riana.tauro@xxxxxxxxx>; Wang,
> Lidong <lidong.wang@xxxxxxxxx>; Yu, Jianshui <jianshui.yu@xxxxxxxxx>
> Subject: Re: [PATCH] drm/i915/dgfx: Enable d3cold at s2idle
>
> On Tue, Aug 15, 2023 at 03:29:12AM -0400, Gupta, Anshuman wrote:
> >
> >
> > > -----Original Message-----
> > > From: Vivi, Rodrigo <rodrigo.vivi@xxxxxxxxx>
> > > Sent: Monday, August 14, 2023 9:33 PM
> > > To: Gupta, Anshuman <anshuman.gupta@xxxxxxxxx>
> > > Cc: intel-gfx@xxxxxxxxxxxxxxxxxxxxx; Nilawar, Badal
> > > <badal.nilawar@xxxxxxxxx>; Tauro, Riana <riana.tauro@xxxxxxxxx>
> > > Subject: Re: [PATCH] drm/i915/dgfx: Enable d3cold at s2idle
> > >
> > > On Mon, Aug 14, 2023 at 04:34:18PM +0530, Anshuman Gupta wrote:
> > > > System wide suspend already has support for lmem save/restore
> > > > during suspend therefore enabling d3cold for s2idle and keepng it
> > > > disable for runtime PM.(Refer below commit for d3cold runtime PM
> > > > disable
> > > > justification) 'commit 66eb93e71a7a ("drm/i915/dgfx: Keep PCI
> > > > autosuspend control 'on' by default on all dGPU")'
> > > >
> > > > It will reduce the DG2 Card power consumption to ~0 Watt for
> > > > s2idle power KPI.
> > > >
> > > > Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/8755
> > > > Cc: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx>
> > > > Signed-off-by: Anshuman Gupta <anshuman.gupta@xxxxxxxxx>
> > >
> > > Reviewed-by: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx>
> > Hi Rodrigo,
> > Thanks for review, will this be good candidate for Linux stable tree ?
>
> Yes, I think so.
>
> do we have a good commit to mark for 'Fixes:'?
>
> but we can add the cc: stable even without the Fixes already in advance
> anyway.
Thanks I will add "Cc: stable@xxxxxxxxxxxxxxx" tag to v2.
Thanks,
Anshuman Gupta.
>
> otherwise we need to wait for this to land in Linus tree and then send
> directly again to the stable mailing list.
>
> > Thanks,
> > Anshuman Gupta.
> > >
> > > > ---
> > > > drivers/gpu/drm/i915/i915_driver.c | 33
> > > > ++++++++++++++++--------------
> > > > 1 file changed, 18 insertions(+), 15 deletions(-)
> > > >
> > > > diff --git a/drivers/gpu/drm/i915/i915_driver.c
> > > > b/drivers/gpu/drm/i915/i915_driver.c
> > > > index b870c0df081a..ec4d26b3c17c 100644
> > > > --- a/drivers/gpu/drm/i915/i915_driver.c
> > > > +++ b/drivers/gpu/drm/i915/i915_driver.c
> > > > @@ -443,7 +443,6 @@ static int i915_pcode_init(struct
> > > > drm_i915_private
> > > > *i915) static int i915_driver_hw_probe(struct drm_i915_private
> > > > *dev_priv) {
> > > > struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
> > > > - struct pci_dev *root_pdev;
> > > > int ret;
> > > >
> > > > if (i915_inject_probe_failure(dev_priv))
> > > > @@ -557,15 +556,6 @@ static int i915_driver_hw_probe(struct
> > > > drm_i915_private *dev_priv)
> > > >
> > > > intel_bw_init_hw(dev_priv);
> > > >
> > > > - /*
> > > > - * FIXME: Temporary hammer to avoid freezing the machine on our
> > > DGFX
> > > > - * This should be totally removed when we handle the pci states
> > > properly
> > > > - * on runtime PM and on s2idle cases.
> > > > - */
> > > > - root_pdev = pcie_find_root_port(pdev);
> > > > - if (root_pdev)
> > > > - pci_d3cold_disable(root_pdev);
> > > > -
> > > > return 0;
> > > >
> > > > err_opregion:
> > > > @@ -591,7 +581,6 @@ static int i915_driver_hw_probe(struct
> > > > drm_i915_private *dev_priv) static void
> > > > i915_driver_hw_remove(struct drm_i915_private *dev_priv) {
> > > > struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
> > > > - struct pci_dev *root_pdev;
> > > >
> > > > i915_perf_fini(dev_priv);
> > > >
> > > > @@ -599,10 +588,6 @@ static void i915_driver_hw_remove(struct
> > > > drm_i915_private *dev_priv)
> > > >
> > > > if (pdev->msi_enabled)
> > > > pci_disable_msi(pdev);
> > > > -
> > > > - root_pdev = pcie_find_root_port(pdev);
> > > > - if (root_pdev)
> > > > - pci_d3cold_enable(root_pdev);
> > > > }
> > > >
> > > > /**
> > > > @@ -1519,6 +1504,8 @@ static int intel_runtime_suspend(struct
> > > > device
> > > > *kdev) {
> > > > struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
> > > > struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
> > > > + struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
> > > > + struct pci_dev *root_pdev;
> > > > struct intel_gt *gt;
> > > > int ret, i;
> > > >
> > > > @@ -1570,6 +1557,15 @@ static int intel_runtime_suspend(struct
> > > > device
> > > *kdev)
> > > > drm_err(&dev_priv->drm,
> > > > "Unclaimed access detected prior to suspending\n");
> > > >
> > > > + /*
> > > > + * FIXME: Temporary hammer to avoid freezing the machine on our
> > > DGFX
> > > > + * This should be totally removed when we handle the pci states
> > > properly
> > > > + * on runtime PM.
> > > > + */
> > > > + root_pdev = pcie_find_root_port(pdev);
> > > > + if (root_pdev)
> > > > + pci_d3cold_disable(root_pdev);
> > > > +
> > > > rpm->suspended = true;
> > > >
> > > > /*
> > > > @@ -1608,6 +1604,8 @@ static int intel_runtime_resume(struct
> > > > device
> > > > *kdev) {
> > > > struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
> > > > struct intel_runtime_pm *rpm = &dev_priv->runtime_pm;
> > > > + struct pci_dev *pdev = to_pci_dev(dev_priv->drm.dev);
> > > > + struct pci_dev *root_pdev;
> > > > struct intel_gt *gt;
> > > > int ret, i;
> > > >
> > > > @@ -1621,6 +1619,11 @@ static int intel_runtime_resume(struct
> > > > device
> > > > *kdev)
> > > >
> > > > intel_opregion_notify_adapter(dev_priv, PCI_D0);
> > > > rpm->suspended = false;
> > > > +
> > > > + root_pdev = pcie_find_root_port(pdev);
> > > > + if (root_pdev)
> > > > + pci_d3cold_enable(root_pdev);
> > > > +
> > > > if (intel_uncore_unclaimed_mmio(&dev_priv->uncore))
> > > > drm_dbg(&dev_priv->drm,
> > > > "Unclaimed access during suspend, bios?\n");
> > > > --
> > > > 2.25.1
> > > >
