[PATCH] PCI/ACPI: Whitelist hotplug ports for D3 if power managed by ACPI

[Rafael J. Wysocki]

On Fri, Oct 2, 2020 at 7:17 AM Lukas Wunner <lukas at wunner.de> wrote:
>
> Recent laptops with dual AMD GPUs fail to suspend the discrete GPU, thus
> causing lockups on system sleep and high power consumption at runtime.
> The discrete GPU would normally be suspended to D3cold by turning off
> ACPI _PR3 Power Resources of the Root Port above the GPU.
>
> However on affected systems, the Root Port is hotplug-capable and
> pci_bridge_d3_possible() only allows hotplug ports to go to D3 if they
> belong to a Thunderbolt device or if the Root Port possesses a
> "HotPlugSupportInD3" ACPI property.  Neither is the case on affected
> laptops.  The reason for whitelisting only specific, known to work
> hotplug ports for D3 is that there have been reports of SkyLake Xeon-SP
> systems raising Hardware Error NMIs upon suspending their hotplug ports:
> https://lore.kernel.org/linux-pci/20170503180426.GA4058@otc-nc-03/
>
> But if a hotplug port is power manageable by ACPI (as can be detected
> through presence of Power Resources and corresponding _PS0 and _PS3
> methods) then it ought to be safe to suspend it to D3.  To this end,
> amend acpi_pci_bridge_d3() to whitelist such ports for D3.
>
> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1222
> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1252
> Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1304
> Reported-and-tested-by: Arthur Borsboom <arthurborsboom at gmail.com>
> Reported-and-tested-by: matoro <matoro at airmail.cc>
> Reported-by: Aaron Zakhrov <aaron.zakhrov at gmail.com>
> Reported-by: Michal Rostecki <mrostecki at suse.com>
> Reported-by: Shai Coleman <git at shaicoleman.com>
> Signed-off-by: Lukas Wunner <lukas at wunner.de>
> Cc: stable at vger.kernel.org
> Cc: Alex Deucher <alexander.deucher at amd.com>
> Cc: Rafael J. Wysocki <rafael.j.wysocki at intel.com>
> Cc: Mika Westerberg <mika.westerberg at linux.intel.com>
> ---
>  drivers/pci/pci-acpi.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
>
> diff --git a/drivers/pci/pci-acpi.c b/drivers/pci/pci-acpi.c
> index d5869a0..d9aa551 100644
> --- a/drivers/pci/pci-acpi.c
> +++ b/drivers/pci/pci-acpi.c
> @@ -944,6 +944,16 @@ static bool acpi_pci_bridge_d3(struct pci_dev *dev)
>         if (!dev->is_hotplug_bridge)
>                 return false;
>
> +       /* Assume D3 support if the bridge is power-manageable by ACPI. */
> +       adev = ACPI_COMPANION(&dev->dev);
> +       if (!adev && !pci_dev_is_added(dev)) {
> +               adev = acpi_pci_find_companion(&dev->dev);
> +               ACPI_COMPANION_SET(&dev->dev, adev);
> +       }
> +
> +       if (adev && acpi_device_power_manageable(adev))
> +               return true;
> +
>         /*
>          * Look for a special _DSD property for the root port and if it
>          * is set we know the hierarchy behind it supports D3 just fine.
> --

I'm going to apply this patch for 5.10 unless Bjorn would rather route
it through the PCI tree.

Thanks!