RE: [PATCH] platform/x86/intel/vsec: Restore BAR fallback for header walk

[Ruhl, Michael J]

>-----Original Message-----
>From: David E. Box <david.e.box@xxxxxxxxxxxxxxx>
>Sent: Friday, May 29, 2026 2:32 PM
>To: ilpo.jarvinen@xxxxxxxxxxxxxxx; Ruhl, Michael J <michael.j.ruhl@xxxxxxxxx>;
>david.e.box@xxxxxxxxxxxxxxx
>Cc: Brost, Matthew <matthew.brost@xxxxxxxxx>;
>thomas.hellstrom@xxxxxxxxxxxxxxx; intel-xe@xxxxxxxxxxxxxxxxxxxxx; linux-
>kernel@xxxxxxxxxxxxxxx; platform-driver-x86@xxxxxxxxxxxxxxx
>Subject: [PATCH] platform/x86/intel/vsec: Restore BAR fallback for header walk
>
>The base_addr refactor changed intel_vsec_walk_header() to pass
>info->base_addr as the discovery-table base address. For the PCI VSEC
>driver this info comes from driver_data, but exported callers may provide
>their own static headers and leave base_addr unset.
>
>For xe, this made the discovery-table base address zero instead of the BAR
>selected by header->tbir, preventing PMT endpoints from being created.
>
>Restore the previous behavior for the header-walk path by falling back to
>pci_resource_start(pdev, header->tbir) when base_addr is not specified.
>Keep explicit base_addr override behavior unchanged.
>
>This preserves the refactor structure while fixing the functional
>regression in manual-header users.
>
>Fixes: 904b333fc51c ("platform/x86/intel/vsec: Refactor base_addr handling")
>Assisted-by: Claude:claude-sonnet-4-6
>Signed-off-by: David E. Box <david.e.box@xxxxxxxxxxxxxxx>
>---
> drivers/platform/x86/intel/vsec.c | 17 ++++++++++++++++-
> 1 file changed, 16 insertions(+), 1 deletion(-)
>
>diff --git a/drivers/platform/x86/intel/vsec.c
>b/drivers/platform/x86/intel/vsec.c
>index 1657834fd275..3c6d8a1928c0 100644
>--- a/drivers/platform/x86/intel/vsec.c
>+++ b/drivers/platform/x86/intel/vsec.c
>@@ -482,10 +482,25 @@ static int intel_vsec_walk_header(struct device
>*dev,
> 				  const struct intel_vsec_platform_info *info)
> {
> 	struct intel_vsec_header **header = info->headers;
>+	u64 base_addr;
> 	int ret;
>
> 	for ( ; *header; header++) {
>-		ret = intel_vsec_register_device(dev, *header, info, info->base_addr);
>+		if (info->base_addr) {
>+			base_addr = info->base_addr;
>+		} else {
>+			struct pci_dev *pdev;
>+
>+			if (!dev_is_pci(dev)) {
>+				dev_err(dev, "non-PCI device without a base address\n");
>+				return -EINVAL;
>+			}
>+
>+			pdev = to_pci_dev(dev);
>+			base_addr = pci_resource_start(pdev, (*header)->tbir);

Alternate:

if (!info->base_addr) {
	struct pci_dev *pdev;

	if (!dev_is_pci(dev)) {
		dev_err(dev, "non-PCI device without a base address\n");
		return -EINVAL;
	}

	pdev = to_pci_dev(dev);
	info->base_addr = pci_resource_start(pdev, (*header)->tbir);}
}

Would it make sense to require the caller to fill this in?

i.e. if (!info->base_addr) return EINVAL?

Change of behavior, but forcing the caller to provide the right info seems reasonable.

Either way, this looks reasonable to me.

Reviewed-by: Michael J. Ruhl <michael.j.ruhl@xxxxxxxxx>

M

>+		}
>+
>+		ret = intel_vsec_register_device(dev, *header, info, base_addr);
> 		if (ret)
> 			return ret;
> 	}
>
>base-commit: a167ae8eace52dd6c80438b77d92450fe12cd4be
>--
>2.43.0