Re: [PATCH] PCI: vmd: Handle BUS_RESTRICT_CFG value 3 for Arrow Lake-HX

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Unchecked PCI configuration reads can cause device removal or read failures to be treated as a valid configuration (case 3), falsely continuing probe.
--

commit fde973a5965da08fe76816c224c875b5ef1e4a2e
Author: Ali Alaei <ali.alaei.tabatabaei@gmail.com>

PCI: vmd: Handle BUS_RESTRICT_CFG value 3 for Arrow Lake-HX

This commit adds support for Intel Arrow Lake-HX platforms where the VMD
controller's VMCONFIG register returns 3 for the BUS_RESTRICT_CFG field,
mapping it to a bus number start of 224.

> diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c
> index d4ae250d4bc6d..9cf5ea2e1e20f 100644
> --- a/drivers/pci/controller/vmd.c
> +++ b/drivers/pci/controller/vmd.c

[ ... ]

> @@ -648,6 +648,7 @@ static int vmd_get_bus_number_start(struct vmd_dev *vmd)
>  		case 1:
>  			vmd->busn_start = 128;
>  			break;
> +		case 3:

[Severity: Medium]
Does this change unintentionally mask device removal errors?

If the device is removed during probe, pci_read_config_word() does not
necessarily return an error code and the hardware typically returns 0xFFFF.
Because BUS_RESTRICT_CFG(0xFFFF) evaluates to 3, the driver will now treat a
read failure as a valid configuration instead of aborting the probe.

Before this patch, the value 3 hit the default case and safely returned
-ENODEV. Should there be a check to ensure the register value is not 0xFFFF
before evaluating the configuration?

>  		case 2:
>  			vmd->busn_start = 224;
>  			break;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260604185953.17177-1-ali.alaei.tabatabaei@gmail.com?part=1