Re: [PATCH v3 4/5] pci: check if VGA decoding was really activated

[Bjorn Helgaas <helgaas@kernel.org>]

On Sun, Mar 08, 2026 at 02:35:37AM +0900, Simon Richter wrote:
> PCI bridges are allowed to refuse activating VGA decoding, by simply
> ignoring attempts to set the bit that enables it, so after setting the bit,
> read it back to verify.
> 
> One example of such a bridge is the root bridge in IBM PowerNV, but this is
> also useful for GPU passthrough into virtual machines, where it is
> difficult to set up routing for legacy IO through IOMMU.

Based on this:
https://lore.kernel.org/all/9f297568-5616-40b6-b401-df1af57d5e14@hogyros.de

I *guess* this is the critical patch, and together with
https://patchwork.freedesktop.org/patch/709684/?series=161721&rev=4,
it fixes an issue?

Without these two patches, i915 assumes its
vga_get_uninterruptible(pdev, VGA_RSRC_LEGACY_IO) always succeeds, and
I guess the problem is that if it *didn't* succeed, i915's subsequent
ioport accesses may go to the wrong device or to no device at all,
which might corrupt another device or cause a PCI error?

We could include the i915 patch in this series.  Wouldn't be a
problem from the PCI side, I dunno about the DRM side.

> Signed-off-by: Simon Richter <Simon.Richter@hogyros.de>
> ---
>  drivers/pci/pci.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 8479c2e1f74f..e60b948f8576 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -6197,6 +6197,12 @@ int pci_set_vga_state(struct pci_dev *dev, bool decode,
>  				cmd &= ~PCI_BRIDGE_CTL_VGA;
>  			pci_write_config_word(bridge, PCI_BRIDGE_CONTROL,
>  					      cmd);
> +			if (decode) {
> +				pci_read_config_word(bridge, PCI_BRIDGE_CONTROL,
> +						     &cmd);
> +				if(!(cmd & PCI_BRIDGE_CTL_VGA))
> +					return -EIO;
> +			}
>  		}
>  		bus = bus->parent;
>  	}
> -- 
> 2.47.3
>