platform/x86/intel/vsec: Restore BAR fallback for header walk

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

Posted by David E. Box 1 week, 2 days ago

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@linux.intel.com>
---
 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);
+		}
+
+		ret = intel_vsec_register_device(dev, *header, info, base_addr);
 		if (ret)
 			return ret;
 	}

base-commit: a167ae8eace52dd6c80438b77d92450fe12cd4be
-- 
2.43.0

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

Posted by Ruhl, Michael J 4 days, 15 hours ago

>-----Original Message-----
>From: David E. Box <david.e.box@linux.intel.com>
>Sent: Friday, May 29, 2026 2:32 PM
>To: ilpo.jarvinen@linux.intel.com; Ruhl, Michael J <michael.j.ruhl@intel.com>;
>david.e.box@linux.intel.com
>Cc: Brost, Matthew <matthew.brost@intel.com>;
>thomas.hellstrom@linux.intel.com; intel-xe@lists.freedesktop.org; linux-
>kernel@vger.kernel.org; platform-driver-x86@vger.kernel.org
>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@linux.intel.com>
>---
> 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@intel.com>

M

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

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

Posted by David Box 14 hours ago

On Wed, Jun 03, 2026 at 06:32:32PM +0000, Ruhl, Michael J wrote:
> >-----Original Message-----
> >From: David E. Box <david.e.box@linux.intel.com>
> >Sent: Friday, May 29, 2026 2:32 PM
> >To: ilpo.jarvinen@linux.intel.com; Ruhl, Michael J <michael.j.ruhl@intel.com>;
> >david.e.box@linux.intel.com
> >Cc: Brost, Matthew <matthew.brost@intel.com>;
> >thomas.hellstrom@linux.intel.com; intel-xe@lists.freedesktop.org; linux-
> >kernel@vger.kernel.org; platform-driver-x86@vger.kernel.org
> >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@linux.intel.com>
> >---
> > 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);}
> }

Thanks. I'll do this in the next patch.

> 
> 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.

Agreed. But that would be separate from this fixup patch.

> 
> Either way, this looks reasonable to me.
> 
> Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>

Thanks

David

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