[PATCH] ACPI: Fix selecting the wrong ACPI fwnode for the iGPU on some Dell laptops

[Hans de Goede]

p.s.

This fixes a regression in 6.1, adding the regressions list to the Cc.

Once we figure out the best way to fix this (this patch is more of a proposal
how to fix this rather then a definitive fix), we should also backport
the fix to 6.1.y.


On 1/9/23 21:57, Hans de Goede wrote:
> The Dell Latitude E6430 both with and without the optional NVidia dGPU
> has a bug in its ACPI tables which is causing Linux to assign the wrong
> ACPI fwnode / companion to the pci_device for the i915 iGPU.
> 
> Specifically under the PCI root bridge there are these 2 ACPI Device()s :
> 
>  Scope (_SB.PCI0)
>  {
>      Device (GFX0)
>      {
>          Name (_ADR, 0x00020000)  // _ADR: Address
>      }
> 
>      ...
> 
>      Device (VID)
>      {
>          Name (_ADR, 0x00020000)  // _ADR: Address
>          ...
> 
>          Method (_DOS, 1, NotSerialized)  // _DOS: Disable Output Switching
>          {
>              VDP8 = Arg0
>              VDP1 (One, VDP8)
>          }
> 
>          Method (_DOD, 0, NotSerialized)  // _DOD: Display Output Devices
>          {
>              ...
>          }
>          ...
>      }
>  }
> 
> The non-functional GFX0 ACPI device is a problem, because this gets
> returned as ACPI companion-device by acpi_find_child_device() for the iGPU.
> 
> This is a long standing problem and the i915 driver does use the ACPI
> companion for some things, but works fine without it.
> 
> However since commit 63f534b8bad9 ("ACPI: PCI: Rework acpi_get_pci_dev()")
> acpi_get_pci_dev() relies on the physical-node pointer in the acpi_device
> and that is set on the wrong acpi_device because of the wrong
> acpi_find_child_device() return. This breaks the ACPI video code, leading
> to non working backlight control in some cases.
> 
> Make find_child_checks() return a higher score for children which have
> pnp-ids set by various scan helpers like acpi_is_video_device(), so
> that it picks the right companion-device.
> 
> An alternative approach would be to directly call acpi_is_video_device()
> from find_child_checks() but that would be somewhat computationally
> expensive given that acpi_find_child_device() iterates over all the
> PCI0 children every time it is called.
> 
> Fixes: 63f534b8bad9 ("ACPI: PCI: Rework acpi_get_pci_dev()")
> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
> ---
>  drivers/acpi/glue.c | 14 ++++++++++++--
>  1 file changed, 12 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/acpi/glue.c b/drivers/acpi/glue.c
> index 204fe94c7e45..2055dfd7678b 100644
> --- a/drivers/acpi/glue.c
> +++ b/drivers/acpi/glue.c
> @@ -75,7 +75,7 @@ static struct acpi_bus_type *acpi_get_bus_type(struct device *dev)
>  }
>  
>  #define FIND_CHILD_MIN_SCORE	1
> -#define FIND_CHILD_MAX_SCORE	2
> +#define FIND_CHILD_MAX_SCORE	3
>  
>  static int match_any(struct acpi_device *adev, void *not_used)
>  {
> @@ -89,15 +89,25 @@ static bool acpi_dev_has_children(struct acpi_device *adev)
>  
>  static int find_child_checks(struct acpi_device *adev, bool check_children)
>  {
> +	int score = FIND_CHILD_MIN_SCORE;
>  	unsigned long long sta;
>  	acpi_status status;
>  
>  	if (check_children && !acpi_dev_has_children(adev))
>  		return -ENODEV;
>  
> +	/*
> +	 * For devices without a _STA method, prefer devices without a _HID
> +	 * (which conflicts with having an _ADR) but which have been matched
> +	 * in some other way, like e.g. by acpi_is_video_device() over devices
> +	 * with no ids at all.
> +	 */
> +	if (!adev->pnp.type.platform_id && adev->pnp.type.hardware_id)
> +		score++;
> +
>  	status = acpi_evaluate_integer(adev->handle, "_STA", NULL, &sta);
>  	if (status == AE_NOT_FOUND)
> -		return FIND_CHILD_MIN_SCORE;
> +		return score;
>  
>  	if (ACPI_FAILURE(status) || !(sta & ACPI_STA_DEVICE_ENABLED))
>  		return -ENODEV;