[PATCH v5 0/6] platform/x86/intel/vsec: Prep for ACPI PMT discovery

[PATCH v5 0/6] platform/x86/intel/vsec: Prep for ACPI PMT discovery

[David E. Box]

This series updates intel_vsec to improve const-correctness, decouple
helper APIs from PCI, enhance error handling, and plumb ACPI-based Intel
Platform Monitoring Technology (PMT) discovery through the vsec layer. It
is preparatory infrastructure for follow-on PMT core/telemetry and
PMC/SSRAM series that add ACPI discovery and support for new platforms.

The series is organized as follows:

Patches 1-2 refactor and improve const-correctness of base_addr handling.
Patch 1 makes base_addr an explicit parameter throughout the call chain,
clarifying ownership and removing conditional logic. Patch 2 then makes the
platform info data structure read-only, preventing unintended modifications
to shared driver data.

Patches 3-4 decouple the vsec layer from PCI-specific types, updating
helper APIs and data structures to use generic struct device instead of
struct pci_dev.  This enables vsec to work with both PCI and ACPI parent
devices.

Patch 5 enhances error visibility by returning meaningful error codes from
the registration path instead of collapsing to boolean success/failure.

Patch 6 adds infrastructure for ACPI-based PMT discovery, allowing client
drivers to consume discovery data from either PCI or ACPI sources.

Signed-off-by: David E. Box <david.e.box@linux.intel.com>

David E. Box (6):
      platform/x86/intel/vsec: Refactor base_addr handling
      platform/x86/intel/vsec: Make driver_data info const
      platform/x86/intel/vsec: Decouple add/link helpers from PCI
      platform/x86/intel/vsec: Switch exported helpers from pci_dev to device
      platform/x86/intel/vsec: Return real error codes from registration path
      platform/x86/intel/vsec: Plumb ACPI PMT discovery tables through vsec

 drivers/gpu/drm/xe/xe_debugfs.c                  |   2 +-
 drivers/gpu/drm/xe/xe_hwmon.c                    |   2 +-
 drivers/gpu/drm/xe/xe_vsec.c                     |   7 +-
 drivers/gpu/drm/xe/xe_vsec.h                     |   4 +-
 drivers/platform/x86/intel/pmc/core.c            |   4 +-
 drivers/platform/x86/intel/pmc/ssram_telemetry.c |   2 +-
 drivers/platform/x86/intel/pmt/class.c           |   8 +-
 drivers/platform/x86/intel/pmt/class.h           |   5 +-
 drivers/platform/x86/intel/pmt/discovery.c       |   4 +-
 drivers/platform/x86/intel/pmt/telemetry.c       |  13 +--
 drivers/platform/x86/intel/pmt/telemetry.h       |  12 +--
 drivers/platform/x86/intel/sdsi.c                |   5 +-
 drivers/platform/x86/intel/vsec.c                | 121 ++++++++++++++---------
 drivers/platform/x86/intel/vsec_tpmi.c           |  10 +-
 include/linux/intel_vsec.h                       |  39 ++++++--
 15 files changed, 144 insertions(+), 94 deletions(-)
---
base-commit: 6de23f81a5e08be8fbf5e8d7e9febc72a5b5f27f
-- 

[PATCH v5 1/6] platform/x86/intel/vsec: Refactor base_addr handling

[David E. Box]

The base_addr field in intel_vsec_platform_info was originally added to
support devices that emulate PCI VSEC capabilities in MMIO. Previously,
the code would check at registration time whether base_addr was set,
falling back to the PCI BAR if not.

Refactor this by making base_addr an explicit function parameter. This
clarifies ownership of the value and removes conditional logic from
intel_vsec_add_dev(). It also enables making intel_vsec_platform_info
const in a later patch, since the function no longer needs to write to
info->base_addr.

No functional change intended.

Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
Changes in v5:
- No change

Changes in v4:
- No change

Changes in v3:
- No change

Changes in v2:
- Use pci_resource_start() macro instead of direct pdev->resource array access
  (suggested by Ilpo)

 drivers/platform/x86/intel/vsec.c | 23 ++++++++++-------------
 1 file changed, 10 insertions(+), 13 deletions(-)

diff --git a/drivers/platform/x86/intel/vsec.c b/drivers/platform/x86/intel/vsec.c
index 5059d320edf8..46966edca03b 100644
--- a/drivers/platform/x86/intel/vsec.c
+++ b/drivers/platform/x86/intel/vsec.c
@@ -271,14 +271,13 @@ EXPORT_SYMBOL_NS_GPL(intel_vsec_add_aux, "INTEL_VSEC");
 
 static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *header,
 			      struct intel_vsec_platform_info *info,
-			      unsigned long cap_id)
+			      unsigned long cap_id, u64 base_addr)
 {
 	struct intel_vsec_device __free(kfree) *intel_vsec_dev = NULL;
 	struct resource __free(kfree) *res = NULL;
 	struct resource *tmp;
 	struct device *parent;
 	unsigned long quirks = info->quirks;
-	u64 base_addr;
 	int i;
 
 	if (info->parent)
@@ -310,11 +309,6 @@ static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *he
 	if (quirks & VSEC_QUIRK_TABLE_SHIFT)
 		header->offset >>= TABLE_OFFSET_SHIFT;
 
-	if (info->base_addr)
-		base_addr = info->base_addr;
-	else
-		base_addr = pdev->resource[header->tbir].start;
-
 	/*
 	 * The DVSEC/VSEC contains the starting offset and count for a block of
 	 * discovery tables. Create a resource array of these tables to the
@@ -412,7 +406,8 @@ static int get_cap_id(u32 header_id, unsigned long *cap_id)
 
 static int intel_vsec_register_device(struct pci_dev *pdev,
 				      struct intel_vsec_header *header,
-				      struct intel_vsec_platform_info *info)
+				      struct intel_vsec_platform_info *info,
+				      u64 base_addr)
 {
 	const struct vsec_feature_dependency *consumer_deps;
 	struct vsec_priv *priv;
@@ -428,7 +423,7 @@ static int intel_vsec_register_device(struct pci_dev *pdev,
 	 * For others using the exported APIs, add the device directly.
 	 */
 	if (!pci_match_id(intel_vsec_pci_ids, pdev))
-		return intel_vsec_add_dev(pdev, header, info, cap_id);
+		return intel_vsec_add_dev(pdev, header, info, cap_id, base_addr);
 
 	priv = pci_get_drvdata(pdev);
 	if (priv->state[cap_id] == STATE_REGISTERED ||
@@ -444,7 +439,7 @@ static int intel_vsec_register_device(struct pci_dev *pdev,
 
 	consumer_deps = get_consumer_dependencies(priv, cap_id);
 	if (!consumer_deps || suppliers_ready(priv, consumer_deps, cap_id)) {
-		ret = intel_vsec_add_dev(pdev, header, info, cap_id);
+		ret = intel_vsec_add_dev(pdev, header, info, cap_id, base_addr);
 		if (ret)
 			priv->state[cap_id] = STATE_SKIP;
 		else
@@ -464,7 +459,7 @@ static bool intel_vsec_walk_header(struct pci_dev *pdev,
 	int ret;
 
 	for ( ; *header; header++) {
-		ret = intel_vsec_register_device(pdev, *header, info);
+		ret = intel_vsec_register_device(pdev, *header, info, info->base_addr);
 		if (!ret)
 			have_devices = true;
 	}
@@ -512,7 +507,8 @@ static bool intel_vsec_walk_dvsec(struct pci_dev *pdev,
 		pci_read_config_dword(pdev, pos + PCI_DVSEC_HEADER2, &hdr);
 		header.id = PCI_DVSEC_HEADER2_ID(hdr);
 
-		ret = intel_vsec_register_device(pdev, &header, info);
+		ret = intel_vsec_register_device(pdev, &header, info,
+						 pci_resource_start(pdev, header.tbir));
 		if (ret)
 			continue;
 
@@ -557,7 +553,8 @@ static bool intel_vsec_walk_vsec(struct pci_dev *pdev,
 		header.tbir = INTEL_DVSEC_TABLE_BAR(table);
 		header.offset = INTEL_DVSEC_TABLE_OFFSET(table);
 
-		ret = intel_vsec_register_device(pdev, &header, info);
+		ret = intel_vsec_register_device(pdev, &header, info,
+						 pci_resource_start(pdev, header.tbir));
 		if (ret)
 			continue;
 

-- 
2.43.0

Claude review: platform/x86/intel/vsec: Refactor base_addr handling

[Claude Code Review Bot]

Patch Review

**Overall: Clean, straightforward refactor.**

This patch removes the conditional `info->base_addr` vs `pdev->resource[header->tbir].start` logic from `intel_vsec_add_dev()` and instead passes `base_addr` as an explicit parameter all the way down the call chain. The callers now resolve the address: `intel_vsec_walk_header()` passes `info->base_addr`, while `intel_vsec_walk_dvsec()` and `intel_vsec_walk_vsec()` use `pci_resource_start()`.

This is a good preparation for making `info` const later and for ACPI paths where BAR lookups don't apply.

No issues found.

---

---
Generated by Claude Code Patch Reviewer

[PATCH v5 2/6] platform/x86/intel/vsec: Make driver_data info const

[David E. Box]

Treat PCI id->driver_data (intel_vsec_platform_info) as read-only by making
vsec_priv->info a const pointer and updating all function signatures to
accept const intel_vsec_platform_info *.

This improves const-correctness and clarifies that the platform info data
from the driver_data table is not meant to be modified at runtime.

No functional changes intended.

Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
Changes in v5:
- No change

Changes in v4:
- No change

Changes in v3:
- No change

Changes in v2:
- New patch

 drivers/platform/x86/intel/vsec.c | 20 ++++++++++----------
 include/linux/intel_vsec.h        |  4 ++--
 2 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/drivers/platform/x86/intel/vsec.c b/drivers/platform/x86/intel/vsec.c
index 46966edca03b..e0096be605d9 100644
--- a/drivers/platform/x86/intel/vsec.c
+++ b/drivers/platform/x86/intel/vsec.c
@@ -42,7 +42,7 @@ enum vsec_device_state {
 };
 
 struct vsec_priv {
-	struct intel_vsec_platform_info *info;
+	const struct intel_vsec_platform_info *info;
 	struct device *suppliers[VSEC_FEATURE_COUNT];
 	struct oobmsm_plat_info plat_info;
 	enum vsec_device_state state[VSEC_FEATURE_COUNT];
@@ -270,7 +270,7 @@ int intel_vsec_add_aux(struct pci_dev *pdev, struct device *parent,
 EXPORT_SYMBOL_NS_GPL(intel_vsec_add_aux, "INTEL_VSEC");
 
 static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *header,
-			      struct intel_vsec_platform_info *info,
+			      const struct intel_vsec_platform_info *info,
 			      unsigned long cap_id, u64 base_addr)
 {
 	struct intel_vsec_device __free(kfree) *intel_vsec_dev = NULL;
@@ -406,7 +406,7 @@ static int get_cap_id(u32 header_id, unsigned long *cap_id)
 
 static int intel_vsec_register_device(struct pci_dev *pdev,
 				      struct intel_vsec_header *header,
-				      struct intel_vsec_platform_info *info,
+				      const struct intel_vsec_platform_info *info,
 				      u64 base_addr)
 {
 	const struct vsec_feature_dependency *consumer_deps;
@@ -452,7 +452,7 @@ static int intel_vsec_register_device(struct pci_dev *pdev,
 }
 
 static bool intel_vsec_walk_header(struct pci_dev *pdev,
-				   struct intel_vsec_platform_info *info)
+				   const struct intel_vsec_platform_info *info)
 {
 	struct intel_vsec_header **header = info->headers;
 	bool have_devices = false;
@@ -468,7 +468,7 @@ static bool intel_vsec_walk_header(struct pci_dev *pdev,
 }
 
 static bool intel_vsec_walk_dvsec(struct pci_dev *pdev,
-				  struct intel_vsec_platform_info *info)
+				  const struct intel_vsec_platform_info *info)
 {
 	bool have_devices = false;
 	int pos = 0;
@@ -519,7 +519,7 @@ static bool intel_vsec_walk_dvsec(struct pci_dev *pdev,
 }
 
 static bool intel_vsec_walk_vsec(struct pci_dev *pdev,
-				 struct intel_vsec_platform_info *info)
+				 const struct intel_vsec_platform_info *info)
 {
 	bool have_devices = false;
 	int pos = 0;
@@ -565,7 +565,7 @@ static bool intel_vsec_walk_vsec(struct pci_dev *pdev,
 }
 
 int intel_vsec_register(struct pci_dev *pdev,
-			 struct intel_vsec_platform_info *info)
+			const struct intel_vsec_platform_info *info)
 {
 	if (!pdev || !info || !info->headers)
 		return -EINVAL;
@@ -578,7 +578,7 @@ int intel_vsec_register(struct pci_dev *pdev,
 EXPORT_SYMBOL_NS_GPL(intel_vsec_register, "INTEL_VSEC");
 
 static bool intel_vsec_get_features(struct pci_dev *pdev,
-				    struct intel_vsec_platform_info *info)
+				    const struct intel_vsec_platform_info *info)
 {
 	bool found = false;
 
@@ -622,7 +622,7 @@ static void intel_vsec_skip_missing_dependencies(struct pci_dev *pdev)
 
 static int intel_vsec_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 {
-	struct intel_vsec_platform_info *info;
+	const struct intel_vsec_platform_info *info;
 	struct vsec_priv *priv;
 	int num_caps, ret;
 	int run_once = 0;
@@ -633,7 +633,7 @@ static int intel_vsec_pci_probe(struct pci_dev *pdev, const struct pci_device_id
 		return ret;
 
 	pci_save_state(pdev);
-	info = (struct intel_vsec_platform_info *)id->driver_data;
+	info = (const struct intel_vsec_platform_info *)id->driver_data;
 	if (!info)
 		return -EINVAL;
 
diff --git a/include/linux/intel_vsec.h b/include/linux/intel_vsec.h
index 1a0f357c2427..d551174b0049 100644
--- a/include/linux/intel_vsec.h
+++ b/include/linux/intel_vsec.h
@@ -200,13 +200,13 @@ static inline struct intel_vsec_device *auxdev_to_ivdev(struct auxiliary_device
 
 #if IS_ENABLED(CONFIG_INTEL_VSEC)
 int intel_vsec_register(struct pci_dev *pdev,
-			 struct intel_vsec_platform_info *info);
+			const struct intel_vsec_platform_info *info);
 int intel_vsec_set_mapping(struct oobmsm_plat_info *plat_info,
 			   struct intel_vsec_device *vsec_dev);
 struct oobmsm_plat_info *intel_vsec_get_mapping(struct pci_dev *pdev);
 #else
 static inline int intel_vsec_register(struct pci_dev *pdev,
-				       struct intel_vsec_platform_info *info)
+				      const struct intel_vsec_platform_info *info)
 {
 	return -ENODEV;
 }

-- 
2.43.0

Claude review: platform/x86/intel/vsec: Make driver_data info const

[Claude Code Review Bot]

Patch Review

**Overall: Clean const-correctness improvement.**

Makes `vsec_priv->info` a `const` pointer and updates all function signatures that accept `intel_vsec_platform_info *` to `const`. The cast in `intel_vsec_pci_probe()`:

```c
info = (const struct intel_vsec_platform_info *)id->driver_data;
```

is correct since `driver_data` is `unsigned long` and this is a standard PCI ID table pattern.

Minor nit: the alignment of `const` in the `intel_vsec_register` declaration differs slightly between the two hunks but this is cosmetic.

No issues found.

---

---
Generated by Claude Code Patch Reviewer

[PATCH v5 3/6] platform/x86/intel/vsec: Decouple add/link helpers from PCI

[David E. Box]

This refactor is prepares for adding ACPI-enumerated PMT endpoints. While
intel_vsec is bound to PCI today, some helpers are used by code that will
also register PMT endpoints from non-PCI (ACPI) paths. Clean up
PCI-specific plumbing where it isn’t strictly required and rely on
generic struct device where possible.

Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
Changes in v5:
- No change

Changes in v4:
- No change

Changes in v3:
- No change

Changes in v2:
- No chnage (previous patch 1)

 drivers/platform/x86/intel/vsec.c      | 13 +++++++++----
 drivers/platform/x86/intel/vsec_tpmi.c |  2 +-
 include/linux/intel_vsec.h             |  2 +-
 3 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/drivers/platform/x86/intel/vsec.c b/drivers/platform/x86/intel/vsec.c
index e0096be605d9..938648b9ef09 100644
--- a/drivers/platform/x86/intel/vsec.c
+++ b/drivers/platform/x86/intel/vsec.c
@@ -158,18 +158,23 @@ static bool vsec_driver_present(int cap_id)
  */
 static const struct pci_device_id intel_vsec_pci_ids[];
 
-static int intel_vsec_link_devices(struct pci_dev *pdev, struct device *dev,
+static int intel_vsec_link_devices(struct device *parent, struct device *dev,
 				   int consumer_id)
 {
 	const struct vsec_feature_dependency *deps;
 	enum vsec_device_state *state;
 	struct device **suppliers;
 	struct vsec_priv *priv;
+	struct pci_dev *pdev;
 	int supplier_id;
 
 	if (!consumer_id)
 		return 0;
 
+	if (!dev_is_pci(parent))
+		return 0;
+
+	pdev = to_pci_dev(parent);
 	if (!pci_match_id(intel_vsec_pci_ids, pdev))
 		return 0;
 
@@ -204,7 +209,7 @@ static int intel_vsec_link_devices(struct pci_dev *pdev, struct device *dev,
 	return 0;
 }
 
-int intel_vsec_add_aux(struct pci_dev *pdev, struct device *parent,
+int intel_vsec_add_aux(struct device *parent,
 		       struct intel_vsec_device *intel_vsec_dev,
 		       const char *name)
 {
@@ -252,7 +257,7 @@ int intel_vsec_add_aux(struct pci_dev *pdev, struct device *parent,
 	if (ret)
 		goto cleanup_aux;
 
-	ret = intel_vsec_link_devices(pdev, &auxdev->dev, intel_vsec_dev->cap_id);
+	ret = intel_vsec_link_devices(parent, &auxdev->dev, intel_vsec_dev->cap_id);
 	if (ret)
 		goto cleanup_aux;
 
@@ -343,7 +348,7 @@ static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *he
 	 * Pass the ownership of intel_vsec_dev and resource within it to
 	 * intel_vsec_add_aux()
 	 */
-	return intel_vsec_add_aux(pdev, parent, no_free_ptr(intel_vsec_dev),
+	return intel_vsec_add_aux(parent, no_free_ptr(intel_vsec_dev),
 				  intel_vsec_name(header->id));
 }
 
diff --git a/drivers/platform/x86/intel/vsec_tpmi.c b/drivers/platform/x86/intel/vsec_tpmi.c
index 98846e88d3d0..e08af071fe87 100644
--- a/drivers/platform/x86/intel/vsec_tpmi.c
+++ b/drivers/platform/x86/intel/vsec_tpmi.c
@@ -655,7 +655,7 @@ static int tpmi_create_device(struct intel_tpmi_info *tpmi_info,
 	 * feature_vsec_dev and res memory are also freed as part of
 	 * device deletion.
 	 */
-	return intel_vsec_add_aux(vsec_dev->pcidev, &vsec_dev->auxdev.dev,
+	return intel_vsec_add_aux(&vsec_dev->pcidev->dev,
 				  feature_vsec_dev, feature_id_name);
 }
 
diff --git a/include/linux/intel_vsec.h b/include/linux/intel_vsec.h
index d551174b0049..49a746ec0128 100644
--- a/include/linux/intel_vsec.h
+++ b/include/linux/intel_vsec.h
@@ -184,7 +184,7 @@ struct pmt_feature_group {
 	struct telemetry_region	regions[];
 };
 
-int intel_vsec_add_aux(struct pci_dev *pdev, struct device *parent,
+int intel_vsec_add_aux(struct device *parent,
 		       struct intel_vsec_device *intel_vsec_dev,
 		       const char *name);
 

-- 
2.43.0

Claude review: platform/x86/intel/vsec: Decouple add/link helpers from PCI

[Claude Code Review Bot]

Patch Review

**Overall: Good refactor, clean separation.**

This removes the `struct pci_dev *pdev` parameter from `intel_vsec_add_aux()` and `intel_vsec_link_devices()`, replacing them with `struct device *parent`. The key safety addition is:

```c
if (!dev_is_pci(parent))
    return 0;
```

in `intel_vsec_link_devices()`, which correctly short-circuits the dependency-tracking logic (which relies on PCI ID matching) for non-PCI devices. The dependency/linking machinery is only relevant to the VSEC PCI driver's own probe path, so returning 0 for non-PCI parents is correct.

The `vsec_tpmi.c` caller update from `vsec_dev->pcidev` to `&vsec_dev->pcidev->dev` is correct given that at this point in the series, `pcidev` still exists.

No issues found.

---

---
Generated by Claude Code Patch Reviewer

[PATCH v5 4/6] platform/x86/intel/vsec: Switch exported helpers from pci_dev to device

[David E. Box]

Preparatory refactor for ACPI-enumerated PMT endpoints. Several exported
PMT/VSEC interfaces and structs carried struct pci_dev * even though
callers only need a generic struct device. Move those to struct device * so
the same APIs work for PCI and ACPI parents.

Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
Changes in v5:
- xe_vsec.h: Remove forward declaration for pci_dev and add for device

Changes in v4:
- No change

Changes in v3:
- No change

Changes in v2:
- Add forward declarations for struct device in class.h, telemetry.h
- Restore struct pci_dev forward declaration in intel_vsec.h that was removed
- Remove base_addr parameter changes (moved to separate patch)
- Remove erroneous hunk in intel_vsec_pci_probe()
  (review comments by Ilpo Järvinen)

 drivers/gpu/drm/xe/xe_debugfs.c                  |  2 +-
 drivers/gpu/drm/xe/xe_hwmon.c                    |  2 +-
 drivers/gpu/drm/xe/xe_vsec.c                     |  7 ++--
 drivers/gpu/drm/xe/xe_vsec.h                     |  4 +--
 drivers/platform/x86/intel/pmc/core.c            |  4 +--
 drivers/platform/x86/intel/pmc/ssram_telemetry.c |  2 +-
 drivers/platform/x86/intel/pmt/class.c           |  8 ++---
 drivers/platform/x86/intel/pmt/class.h           |  5 +--
 drivers/platform/x86/intel/pmt/discovery.c       |  4 +--
 drivers/platform/x86/intel/pmt/telemetry.c       | 13 +++----
 drivers/platform/x86/intel/pmt/telemetry.h       | 12 +++----
 drivers/platform/x86/intel/sdsi.c                |  5 +--
 drivers/platform/x86/intel/vsec.c                | 44 ++++++++++++++----------
 drivers/platform/x86/intel/vsec_tpmi.c           | 10 +++---
 include/linux/intel_vsec.h                       | 13 +++----
 15 files changed, 73 insertions(+), 62 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_debugfs.c b/drivers/gpu/drm/xe/xe_debugfs.c
index 844cfafe1ec7..ad2d8f179eb6 100644
--- a/drivers/gpu/drm/xe/xe_debugfs.c
+++ b/drivers/gpu/drm/xe/xe_debugfs.c
@@ -45,7 +45,7 @@ static void read_residency_counter(struct xe_device *xe, struct xe_mmio *mmio,
 	u64 residency = 0;
 	int ret;
 
-	ret = xe_pmt_telem_read(to_pci_dev(xe->drm.dev),
+	ret = xe_pmt_telem_read(xe->drm.dev,
 				xe_mmio_read32(mmio, PUNIT_TELEMETRY_GUID),
 				&residency, offset, sizeof(residency));
 	if (ret != sizeof(residency)) {
diff --git a/drivers/gpu/drm/xe/xe_hwmon.c b/drivers/gpu/drm/xe/xe_hwmon.c
index 0fd4d4f1014a..92e423a339f1 100644
--- a/drivers/gpu/drm/xe/xe_hwmon.c
+++ b/drivers/gpu/drm/xe/xe_hwmon.c
@@ -506,7 +506,7 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, int channel, long *energy)
 	if (hwmon->xe->info.platform == XE_BATTLEMAGE) {
 		u64 pmt_val;
 
-		ret = xe_pmt_telem_read(to_pci_dev(hwmon->xe->drm.dev),
+		ret = xe_pmt_telem_read(hwmon->xe->drm.dev,
 					xe_mmio_read32(mmio, PUNIT_TELEMETRY_GUID),
 					&pmt_val, BMG_ENERGY_STATUS_PMT_OFFSET,	sizeof(pmt_val));
 		if (ret != sizeof(pmt_val)) {
diff --git a/drivers/gpu/drm/xe/xe_vsec.c b/drivers/gpu/drm/xe/xe_vsec.c
index 4ebb4dbe1c9b..a9baf0bfe572 100644
--- a/drivers/gpu/drm/xe/xe_vsec.c
+++ b/drivers/gpu/drm/xe/xe_vsec.c
@@ -140,10 +140,10 @@ static int xe_guid_decode(u32 guid, int *index, u32 *offset)
 	return 0;
 }
 
-int xe_pmt_telem_read(struct pci_dev *pdev, u32 guid, u64 *data, loff_t user_offset,
+int xe_pmt_telem_read(struct device *dev, u32 guid, u64 *data, loff_t user_offset,
 		      u32 count)
 {
-	struct xe_device *xe = pdev_to_xe_device(pdev);
+	struct xe_device *xe = kdev_to_xe_device(dev);
 	void __iomem *telem_addr = xe->mmio.regs + BMG_TELEMETRY_OFFSET;
 	u32 mem_region;
 	u32 offset;
@@ -198,7 +198,6 @@ void xe_vsec_init(struct xe_device *xe)
 {
 	struct intel_vsec_platform_info *info;
 	struct device *dev = xe->drm.dev;
-	struct pci_dev *pdev = to_pci_dev(dev);
 	enum xe_vsec platform;
 
 	platform = get_platform_info(xe);
@@ -221,6 +220,6 @@ void xe_vsec_init(struct xe_device *xe)
 	 * Register a VSEC. Cleanup is handled using device managed
 	 * resources.
 	 */
-	intel_vsec_register(pdev, info);
+	intel_vsec_register(dev, info);
 }
 MODULE_IMPORT_NS("INTEL_VSEC");
diff --git a/drivers/gpu/drm/xe/xe_vsec.h b/drivers/gpu/drm/xe/xe_vsec.h
index dabfb4e02d70..a25b4e6e681b 100644
--- a/drivers/gpu/drm/xe/xe_vsec.h
+++ b/drivers/gpu/drm/xe/xe_vsec.h
@@ -6,10 +6,10 @@
 
 #include <linux/types.h>
 
-struct pci_dev;
+struct device;
 struct xe_device;
 
 void xe_vsec_init(struct xe_device *xe);
-int xe_pmt_telem_read(struct pci_dev *pdev, u32 guid, u64 *data, loff_t user_offset, u32 count);
+int xe_pmt_telem_read(struct device *dev, u32 guid, u64 *data, loff_t user_offset, u32 count);
 
 #endif
diff --git a/drivers/platform/x86/intel/pmc/core.c b/drivers/platform/x86/intel/pmc/core.c
index 02b303418d18..d91e1ab842d6 100644
--- a/drivers/platform/x86/intel/pmc/core.c
+++ b/drivers/platform/x86/intel/pmc/core.c
@@ -1315,7 +1315,7 @@ static struct telem_endpoint *pmc_core_register_endpoint(struct pci_dev *pcidev,
 	unsigned int i;
 
 	for (i = 0; guids[i]; i++) {
-		ep = pmt_telem_find_and_register_endpoint(pcidev, guids[i], 0);
+		ep = pmt_telem_find_and_register_endpoint(&pcidev->dev, guids[i], 0);
 		if (!IS_ERR(ep))
 			return ep;
 	}
@@ -1600,7 +1600,7 @@ static int pmc_core_get_telem_info(struct pmc_dev *pmcdev, struct pmc_dev_info *
 		if (!pmc->map->lpm_req_guid)
 			return -ENXIO;
 
-		ep = pmt_telem_find_and_register_endpoint(pcidev, pmc->map->lpm_req_guid, 0);
+		ep = pmt_telem_find_and_register_endpoint(&pcidev->dev, pmc->map->lpm_req_guid, 0);
 		if (IS_ERR(ep)) {
 			dev_dbg(&pmcdev->pdev->dev, "couldn't get telem endpoint %pe", ep);
 			return -EPROBE_DEFER;
diff --git a/drivers/platform/x86/intel/pmc/ssram_telemetry.c b/drivers/platform/x86/intel/pmc/ssram_telemetry.c
index 03fad9331fc0..6f6e83e70fc5 100644
--- a/drivers/platform/x86/intel/pmc/ssram_telemetry.c
+++ b/drivers/platform/x86/intel/pmc/ssram_telemetry.c
@@ -60,7 +60,7 @@ pmc_ssram_telemetry_add_pmt(struct pci_dev *pcidev, u64 ssram_base, void __iomem
 	info.base_addr = ssram_base;
 	info.parent = &pcidev->dev;
 
-	return intel_vsec_register(pcidev, &info);
+	return intel_vsec_register(&pcidev->dev, &info);
 }
 
 static inline u64 get_base(void __iomem *addr, u32 offset)
diff --git a/drivers/platform/x86/intel/pmt/class.c b/drivers/platform/x86/intel/pmt/class.c
index be3c8d9e4fff..b4c9964df807 100644
--- a/drivers/platform/x86/intel/pmt/class.c
+++ b/drivers/platform/x86/intel/pmt/class.c
@@ -60,11 +60,11 @@ pmt_memcpy64_fromio(void *to, const u64 __iomem *from, size_t count)
 	return count;
 }
 
-int pmt_telem_read_mmio(struct pci_dev *pdev, struct pmt_callbacks *cb, u32 guid, void *buf,
+int pmt_telem_read_mmio(struct device *dev, struct pmt_callbacks *cb, u32 guid, void *buf,
 			void __iomem *addr, loff_t off, u32 count)
 {
 	if (cb && cb->read_telem)
-		return cb->read_telem(pdev, guid, buf, off, count);
+		return cb->read_telem(dev, guid, buf, off, count);
 
 	addr += off;
 
@@ -99,7 +99,7 @@ intel_pmt_read(struct file *filp, struct kobject *kobj,
 	if (count > entry->size - off)
 		count = entry->size - off;
 
-	count = pmt_telem_read_mmio(entry->pcidev, entry->cb, entry->header.guid, buf,
+	count = pmt_telem_read_mmio(entry->ep->dev, entry->cb, entry->header.guid, buf,
 				    entry->base, off, count);
 
 	return count;
@@ -208,7 +208,7 @@ static int intel_pmt_populate_entry(struct intel_pmt_entry *entry,
 				    struct intel_vsec_device *ivdev,
 				    struct resource *disc_res)
 {
-	struct pci_dev *pci_dev = ivdev->pcidev;
+	struct pci_dev *pci_dev = to_pci_dev(ivdev->dev);
 	struct device *dev = &ivdev->auxdev.dev;
 	struct intel_pmt_header *header = &entry->header;
 	u8 bir;
diff --git a/drivers/platform/x86/intel/pmt/class.h b/drivers/platform/x86/intel/pmt/class.h
index 3c5ad5f52bca..1ae56a5baad2 100644
--- a/drivers/platform/x86/intel/pmt/class.h
+++ b/drivers/platform/x86/intel/pmt/class.h
@@ -19,11 +19,12 @@
 #define GET_BIR(v)		((v) & GENMASK(2, 0))
 #define GET_ADDRESS(v)		((v) & GENMASK(31, 3))
 
+struct device;
 struct pci_dev;
 extern struct class intel_pmt_class;
 
 struct telem_endpoint {
-	struct pci_dev		*pcidev;
+	struct device		*dev;
 	struct telem_header	header;
 	struct pmt_callbacks	*cb;
 	void __iomem		*base;
@@ -65,7 +66,7 @@ struct intel_pmt_namespace {
 				struct intel_pmt_entry *entry);
 };
 
-int pmt_telem_read_mmio(struct pci_dev *pdev, struct pmt_callbacks *cb, u32 guid, void *buf,
+int pmt_telem_read_mmio(struct device *dev, struct pmt_callbacks *cb, u32 guid, void *buf,
 			void __iomem *addr, loff_t off, u32 count);
 bool intel_pmt_is_early_client_hw(struct device *dev);
 int intel_pmt_dev_create(struct intel_pmt_entry *entry,
diff --git a/drivers/platform/x86/intel/pmt/discovery.c b/drivers/platform/x86/intel/pmt/discovery.c
index e500aa327d23..c482368bfaae 100644
--- a/drivers/platform/x86/intel/pmt/discovery.c
+++ b/drivers/platform/x86/intel/pmt/discovery.c
@@ -542,7 +542,7 @@ static int pmt_features_probe(struct auxiliary_device *auxdev, const struct auxi
 	if (!priv)
 		return -ENOMEM;
 
-	priv->parent = &ivdev->pcidev->dev;
+	priv->parent = ivdev->dev;
 	auxiliary_set_drvdata(auxdev, priv);
 
 	priv->dev = device_create(&intel_pmt_class, &auxdev->dev, MKDEV(0, 0), priv,
@@ -609,7 +609,7 @@ void intel_pmt_get_features(struct intel_pmt_entry *entry)
 
 	mutex_lock(&feature_list_lock);
 	list_for_each_entry(feature, &pmt_feature_list, list) {
-		if (feature->priv->parent != &entry->ep->pcidev->dev)
+		if (feature->priv->parent != entry->ep->dev)
 			continue;
 
 		pmt_get_features(entry, feature);
diff --git a/drivers/platform/x86/intel/pmt/telemetry.c b/drivers/platform/x86/intel/pmt/telemetry.c
index a52803bfe124..bdc7c24a3678 100644
--- a/drivers/platform/x86/intel/pmt/telemetry.c
+++ b/drivers/platform/x86/intel/pmt/telemetry.c
@@ -112,7 +112,7 @@ static int pmt_telem_add_endpoint(struct intel_vsec_device *ivdev,
 		return -ENOMEM;
 
 	ep = entry->ep;
-	ep->pcidev = ivdev->pcidev;
+	ep->dev = ivdev->dev;
 	ep->header.access_type = entry->header.access_type;
 	ep->header.guid = entry->header.guid;
 	ep->header.base_offset = entry->header.base_offset;
@@ -204,7 +204,7 @@ int pmt_telem_get_endpoint_info(int devid, struct telem_endpoint_info *info)
 		goto unlock;
 	}
 
-	info->pdev = entry->ep->pcidev;
+	info->dev = entry->ep->dev;
 	info->header = entry->ep->header;
 
 unlock:
@@ -218,9 +218,10 @@ static int pmt_copy_region(struct telemetry_region *region,
 			   struct intel_pmt_entry *entry)
 {
 
+	struct pci_dev *pdev = to_pci_dev(entry->ep->dev);
 	struct oobmsm_plat_info *plat_info;
 
-	plat_info = intel_vsec_get_mapping(entry->ep->pcidev);
+	plat_info = intel_vsec_get_mapping(pdev);
 	if (IS_ERR(plat_info))
 		return PTR_ERR(plat_info);
 
@@ -308,7 +309,7 @@ int pmt_telem_read(struct telem_endpoint *ep, u32 id, u64 *data, u32 count)
 	if (offset + NUM_BYTES_QWORD(count) > size)
 		return -EINVAL;
 
-	pmt_telem_read_mmio(ep->pcidev, ep->cb, ep->header.guid, data, ep->base, offset,
+	pmt_telem_read_mmio(ep->dev, ep->cb, ep->header.guid, data, ep->base, offset,
 			    NUM_BYTES_QWORD(count));
 
 	return ep->present ? 0 : -EPIPE;
@@ -335,7 +336,7 @@ int pmt_telem_read32(struct telem_endpoint *ep, u32 id, u32 *data, u32 count)
 EXPORT_SYMBOL_NS_GPL(pmt_telem_read32, "INTEL_PMT_TELEMETRY");
 
 struct telem_endpoint *
-pmt_telem_find_and_register_endpoint(struct pci_dev *pcidev, u32 guid, u16 pos)
+pmt_telem_find_and_register_endpoint(struct device *dev, u32 guid, u16 pos)
 {
 	int devid = 0;
 	int inst = 0;
@@ -348,7 +349,7 @@ pmt_telem_find_and_register_endpoint(struct pci_dev *pcidev, u32 guid, u16 pos)
 		if (err)
 			return ERR_PTR(err);
 
-		if (ep_info.header.guid == guid && ep_info.pdev == pcidev) {
+		if (ep_info.header.guid == guid && ep_info.dev == dev) {
 			if (inst == pos)
 				return pmt_telem_register_endpoint(devid);
 			++inst;
diff --git a/drivers/platform/x86/intel/pmt/telemetry.h b/drivers/platform/x86/intel/pmt/telemetry.h
index d45af5512b4e..0f88c5e7d90e 100644
--- a/drivers/platform/x86/intel/pmt/telemetry.h
+++ b/drivers/platform/x86/intel/pmt/telemetry.h
@@ -6,8 +6,8 @@
 #define PMT_TELEM_TELEMETRY	0
 #define PMT_TELEM_CRASHLOG	1
 
+struct device;
 struct telem_endpoint;
-struct pci_dev;
 
 struct telem_header {
 	u8	access_type;
@@ -17,7 +17,7 @@ struct telem_header {
 };
 
 struct telem_endpoint_info {
-	struct pci_dev		*pdev;
+	struct device		*dev;
 	struct telem_header	header;
 };
 
@@ -71,8 +71,8 @@ int pmt_telem_get_endpoint_info(int devid, struct telem_endpoint_info *info);
 
 /**
  * pmt_telem_find_and_register_endpoint() - Get a telemetry endpoint from
- * pci_dev device, guid and pos
- * @pdev:   PCI device inside the Intel vsec
+ * device, guid and pos
+ * @dev:    device inside the Intel vsec
  * @guid:   GUID of the telemetry space
  * @pos:    Instance of the guid
  *
@@ -80,8 +80,8 @@ int pmt_telem_get_endpoint_info(int devid, struct telem_endpoint_info *info);
  * * endpoint    - On success returns pointer to the telemetry endpoint
  * * -ENXIO      - telemetry endpoint not found
  */
-struct telem_endpoint *pmt_telem_find_and_register_endpoint(struct pci_dev *pcidev,
-				u32 guid, u16 pos);
+struct telem_endpoint *
+pmt_telem_find_and_register_endpoint(struct device *dev, u32 guid, u16 pos);
 
 /**
  * pmt_telem_read() - Read qwords from counter sram using sample id
diff --git a/drivers/platform/x86/intel/sdsi.c b/drivers/platform/x86/intel/sdsi.c
index da75f53d0bcc..d7e37d4ace23 100644
--- a/drivers/platform/x86/intel/sdsi.c
+++ b/drivers/platform/x86/intel/sdsi.c
@@ -599,13 +599,14 @@ static int sdsi_get_layout(struct sdsi_priv *priv, struct disc_table *table)
 	return 0;
 }
 
-static int sdsi_map_mbox_registers(struct sdsi_priv *priv, struct pci_dev *parent,
+static int sdsi_map_mbox_registers(struct sdsi_priv *priv, struct device *dev,
 				   struct disc_table *disc_table, struct resource *disc_res)
 {
 	u32 access_type = FIELD_GET(DT_ACCESS_TYPE, disc_table->access_info);
 	u32 size = FIELD_GET(DT_SIZE, disc_table->access_info);
 	u32 tbir = FIELD_GET(DT_TBIR, disc_table->offset);
 	u32 offset = DT_OFFSET(disc_table->offset);
+	struct pci_dev *parent = to_pci_dev(dev);
 	struct resource res = {};
 
 	/* Starting location of SDSi MMIO region based on access type */
@@ -681,7 +682,7 @@ static int sdsi_probe(struct auxiliary_device *auxdev, const struct auxiliary_de
 		return ret;
 
 	/* Map the SDSi mailbox registers */
-	ret = sdsi_map_mbox_registers(priv, intel_cap_dev->pcidev, &disc_table, disc_res);
+	ret = sdsi_map_mbox_registers(priv, intel_cap_dev->dev, &disc_table, disc_res);
 	if (ret)
 		return ret;
 
diff --git a/drivers/platform/x86/intel/vsec.c b/drivers/platform/x86/intel/vsec.c
index 938648b9ef09..a547e4b98245 100644
--- a/drivers/platform/x86/intel/vsec.c
+++ b/drivers/platform/x86/intel/vsec.c
@@ -274,7 +274,7 @@ int intel_vsec_add_aux(struct device *parent,
 }
 EXPORT_SYMBOL_NS_GPL(intel_vsec_add_aux, "INTEL_VSEC");
 
-static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *header,
+static int intel_vsec_add_dev(struct device *dev, struct intel_vsec_header *header,
 			      const struct intel_vsec_platform_info *info,
 			      unsigned long cap_id, u64 base_addr)
 {
@@ -288,18 +288,18 @@ static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *he
 	if (info->parent)
 		parent = info->parent;
 	else
-		parent = &pdev->dev;
+		parent = dev;
 
 	if (!intel_vsec_supported(header->id, info->caps))
 		return -EINVAL;
 
 	if (!header->num_entries) {
-		dev_dbg(&pdev->dev, "Invalid 0 entry count for header id %d\n", header->id);
+		dev_dbg(dev, "Invalid 0 entry count for header id %d\n", header->id);
 		return -EINVAL;
 	}
 
 	if (!header->entry_size) {
-		dev_dbg(&pdev->dev, "Invalid 0 entry size for header id %d\n", header->id);
+		dev_dbg(dev, "Invalid 0 entry size for header id %d\n", header->id);
 		return -EINVAL;
 	}
 
@@ -331,7 +331,7 @@ static int intel_vsec_add_dev(struct pci_dev *pdev, struct intel_vsec_header *he
 		release_mem_region(tmp->start, resource_size(tmp));
 	}
 
-	intel_vsec_dev->pcidev = pdev;
+	intel_vsec_dev->dev = dev;
 	intel_vsec_dev->resource = no_free_ptr(res);
 	intel_vsec_dev->num_resources = header->num_entries;
 	intel_vsec_dev->quirks = info->quirks;
@@ -409,13 +409,14 @@ static int get_cap_id(u32 header_id, unsigned long *cap_id)
 	return 0;
 }
 
-static int intel_vsec_register_device(struct pci_dev *pdev,
+static int intel_vsec_register_device(struct device *dev,
 				      struct intel_vsec_header *header,
 				      const struct intel_vsec_platform_info *info,
 				      u64 base_addr)
 {
 	const struct vsec_feature_dependency *consumer_deps;
 	struct vsec_priv *priv;
+	struct pci_dev *pdev;
 	unsigned long cap_id;
 	int ret;
 
@@ -427,8 +428,12 @@ static int intel_vsec_register_device(struct pci_dev *pdev,
 	 * Only track dependencies for devices probed by the VSEC driver.
 	 * For others using the exported APIs, add the device directly.
 	 */
+	if (!dev_is_pci(dev))
+		return intel_vsec_add_dev(dev, header, info, cap_id, base_addr);
+
+	pdev = to_pci_dev(dev);
 	if (!pci_match_id(intel_vsec_pci_ids, pdev))
-		return intel_vsec_add_dev(pdev, header, info, cap_id, base_addr);
+		return intel_vsec_add_dev(dev, header, info, cap_id, base_addr);
 
 	priv = pci_get_drvdata(pdev);
 	if (priv->state[cap_id] == STATE_REGISTERED ||
@@ -444,7 +449,7 @@ static int intel_vsec_register_device(struct pci_dev *pdev,
 
 	consumer_deps = get_consumer_dependencies(priv, cap_id);
 	if (!consumer_deps || suppliers_ready(priv, consumer_deps, cap_id)) {
-		ret = intel_vsec_add_dev(pdev, header, info, cap_id, base_addr);
+		ret = intel_vsec_add_dev(dev, header, info, cap_id, base_addr);
 		if (ret)
 			priv->state[cap_id] = STATE_SKIP;
 		else
@@ -456,7 +461,7 @@ static int intel_vsec_register_device(struct pci_dev *pdev,
 	return -EAGAIN;
 }
 
-static bool intel_vsec_walk_header(struct pci_dev *pdev,
+static bool intel_vsec_walk_header(struct device *dev,
 				   const struct intel_vsec_platform_info *info)
 {
 	struct intel_vsec_header **header = info->headers;
@@ -464,7 +469,7 @@ static bool intel_vsec_walk_header(struct pci_dev *pdev,
 	int ret;
 
 	for ( ; *header; header++) {
-		ret = intel_vsec_register_device(pdev, *header, info, info->base_addr);
+		ret = intel_vsec_register_device(dev, *header, info, info->base_addr);
 		if (!ret)
 			have_devices = true;
 	}
@@ -512,7 +517,7 @@ static bool intel_vsec_walk_dvsec(struct pci_dev *pdev,
 		pci_read_config_dword(pdev, pos + PCI_DVSEC_HEADER2, &hdr);
 		header.id = PCI_DVSEC_HEADER2_ID(hdr);
 
-		ret = intel_vsec_register_device(pdev, &header, info,
+		ret = intel_vsec_register_device(&pdev->dev, &header, info,
 						 pci_resource_start(pdev, header.tbir));
 		if (ret)
 			continue;
@@ -558,7 +563,7 @@ static bool intel_vsec_walk_vsec(struct pci_dev *pdev,
 		header.tbir = INTEL_DVSEC_TABLE_BAR(table);
 		header.offset = INTEL_DVSEC_TABLE_OFFSET(table);
 
-		ret = intel_vsec_register_device(pdev, &header, info,
+		ret = intel_vsec_register_device(&pdev->dev, &header, info,
 						 pci_resource_start(pdev, header.tbir));
 		if (ret)
 			continue;
@@ -569,13 +574,13 @@ static bool intel_vsec_walk_vsec(struct pci_dev *pdev,
 	return have_devices;
 }
 
-int intel_vsec_register(struct pci_dev *pdev,
+int intel_vsec_register(struct device *dev,
 			const struct intel_vsec_platform_info *info)
 {
-	if (!pdev || !info || !info->headers)
+	if (!dev || !info || !info->headers)
 		return -EINVAL;
 
-	if (!intel_vsec_walk_header(pdev, info))
+	if (!intel_vsec_walk_header(dev, info))
 		return -ENODEV;
 	else
 		return 0;
@@ -601,7 +606,7 @@ static bool intel_vsec_get_features(struct pci_dev *pdev,
 		found = true;
 
 	if (info && (info->quirks & VSEC_QUIRK_NO_DVSEC) &&
-	    intel_vsec_walk_header(pdev, info))
+	    intel_vsec_walk_header(&pdev->dev, info))
 		found = true;
 
 	return found;
@@ -673,7 +678,10 @@ int intel_vsec_set_mapping(struct oobmsm_plat_info *plat_info,
 {
 	struct vsec_priv *priv;
 
-	priv = pci_get_drvdata(vsec_dev->pcidev);
+	if (!dev_is_pci(vsec_dev->dev))
+		return -ENODEV;
+
+	priv = pci_get_drvdata(to_pci_dev(vsec_dev->dev));
 	if (!priv)
 		return -EINVAL;
 
@@ -821,7 +829,7 @@ static pci_ers_result_t intel_vsec_pci_slot_reset(struct pci_dev *pdev)
 
 	xa_for_each(&auxdev_array, index, intel_vsec_dev) {
 		/* check if pdev doesn't match */
-		if (pdev != intel_vsec_dev->pcidev)
+		if (&pdev->dev != intel_vsec_dev->dev)
 			continue;
 		devm_release_action(&pdev->dev, intel_vsec_remove_aux,
 				    &intel_vsec_dev->auxdev);
diff --git a/drivers/platform/x86/intel/vsec_tpmi.c b/drivers/platform/x86/intel/vsec_tpmi.c
index e08af071fe87..259addf664e8 100644
--- a/drivers/platform/x86/intel/vsec_tpmi.c
+++ b/drivers/platform/x86/intel/vsec_tpmi.c
@@ -530,7 +530,7 @@ static const struct file_operations mem_write_ops = {
 	.release        = single_release,
 };
 
-#define tpmi_to_dev(info)	(&info->vsec_dev->pcidev->dev)
+#define tpmi_to_dev(info)	((info)->vsec_dev->dev)
 
 static void tpmi_dbgfs_register(struct intel_tpmi_info *tpmi_info)
 {
@@ -642,7 +642,7 @@ static int tpmi_create_device(struct intel_tpmi_info *tpmi_info,
 		tmp->flags = IORESOURCE_MEM;
 	}
 
-	feature_vsec_dev->pcidev = vsec_dev->pcidev;
+	feature_vsec_dev->dev = vsec_dev->dev;
 	feature_vsec_dev->resource = res;
 	feature_vsec_dev->num_resources = pfs->pfs_header.num_entries;
 	feature_vsec_dev->priv_data = &tpmi_info->plat_info;
@@ -655,8 +655,8 @@ static int tpmi_create_device(struct intel_tpmi_info *tpmi_info,
 	 * feature_vsec_dev and res memory are also freed as part of
 	 * device deletion.
 	 */
-	return intel_vsec_add_aux(&vsec_dev->pcidev->dev,
-				  feature_vsec_dev, feature_id_name);
+	return intel_vsec_add_aux(vsec_dev->dev, feature_vsec_dev,
+				  feature_id_name);
 }
 
 static int tpmi_create_devices(struct intel_tpmi_info *tpmi_info)
@@ -742,7 +742,7 @@ static int tpmi_fetch_pfs_header(struct intel_tpmi_pm_feature *pfs, u64 start, i
 static int intel_vsec_tpmi_init(struct auxiliary_device *auxdev)
 {
 	struct intel_vsec_device *vsec_dev = auxdev_to_ivdev(auxdev);
-	struct pci_dev *pci_dev = vsec_dev->pcidev;
+	struct pci_dev *pci_dev = to_pci_dev(vsec_dev->dev);
 	struct intel_tpmi_info *tpmi_info;
 	u64 pfs_start = 0;
 	int ret, i;
diff --git a/include/linux/intel_vsec.h b/include/linux/intel_vsec.h
index 49a746ec0128..4eecb2a6bac4 100644
--- a/include/linux/intel_vsec.h
+++ b/include/linux/intel_vsec.h
@@ -29,6 +29,7 @@
 #define INTEL_DVSEC_TABLE_OFFSET(x)	((x) & GENMASK(31, 3))
 #define TABLE_OFFSET_SHIFT		3
 
+struct device;
 struct pci_dev;
 struct resource;
 
@@ -82,14 +83,14 @@ enum intel_vsec_quirks {
  * struct pmt_callbacks - Callback infrastructure for PMT devices
  * @read_telem: when specified, called by client driver to access PMT
  * data (instead of direct copy).
- * * pdev:  PCI device reference for the callback's use
+ * * dev:   device reference for the callback's use
  * * guid:  ID of data to acccss
  * * data:  buffer for the data to be copied
  * * off:   offset into the requested buffer
  * * count: size of buffer
  */
 struct pmt_callbacks {
-	int (*read_telem)(struct pci_dev *pdev, u32 guid, u64 *data, loff_t off, u32 count);
+	int (*read_telem)(struct device *dev, u32 guid, u64 *data, loff_t off, u32 count);
 };
 
 struct vsec_feature_dependency {
@@ -122,7 +123,7 @@ struct intel_vsec_platform_info {
 /**
  * struct intel_vsec_device - Auxbus specific device information
  * @auxdev:        auxbus device struct for auxbus access
- * @pcidev:        pci device associated with the device
+ * @dev:           struct device associated with the device
  * @resource:      any resources shared by the parent
  * @ida:           id reference
  * @num_resources: number of resources
@@ -135,7 +136,7 @@ struct intel_vsec_platform_info {
  */
 struct intel_vsec_device {
 	struct auxiliary_device auxdev;
-	struct pci_dev *pcidev;
+	struct device *dev;
 	struct resource *resource;
 	struct ida *ida;
 	int num_resources;
@@ -199,13 +200,13 @@ static inline struct intel_vsec_device *auxdev_to_ivdev(struct auxiliary_device
 }
 
 #if IS_ENABLED(CONFIG_INTEL_VSEC)
-int intel_vsec_register(struct pci_dev *pdev,
+int intel_vsec_register(struct device *dev,
 			const struct intel_vsec_platform_info *info);
 int intel_vsec_set_mapping(struct oobmsm_plat_info *plat_info,
 			   struct intel_vsec_device *vsec_dev);
 struct oobmsm_plat_info *intel_vsec_get_mapping(struct pci_dev *pdev);
 #else
-static inline int intel_vsec_register(struct pci_dev *pdev,
+static inline int intel_vsec_register(struct device *dev,
 				      const struct intel_vsec_platform_info *info)
 {
 	return -ENODEV;

-- 
2.43.0

Claude review: platform/x86/intel/vsec: Switch exported helpers from pci_dev to device

[Claude Code Review Bot]

Patch Review

**Overall: The largest patch, mostly mechanical. Has a potential safety concern.**

This is the core conversion: `struct intel_vsec_device` gains `struct device *dev` replacing `struct pci_dev *pcidev`. All consumers are updated.

**Concern — unchecked `to_pci_dev()` in downstream code after ACPI support is added:**

Several downstream consumers now do `to_pci_dev(ivdev->dev)` or `to_pci_dev(entry->ep->dev)` without first checking `dev_is_pci()`. While this is safe *today* because only PCI parents exist, patch 6 adds ACPI discovery support, and future ACPI-enumerated PMT endpoints could reach this code. Specifically:

1. In `class.c`, `intel_pmt_populate_entry()`:
   ```c
   struct pci_dev *pci_dev = to_pci_dev(ivdev->dev);
   ```
   This function uses `pci_resource_start()` which will crash on a non-PCI device.

2. In `telemetry.c`, `pmt_copy_region()`:
   ```c
   struct pci_dev *pdev = to_pci_dev(entry->ep->dev);
   ```

3. In `sdsi.c`, `sdsi_map_mbox_registers()`:
   ```c
   struct pci_dev *parent = to_pci_dev(dev);
   ```

4. In `vsec_tpmi.c`, `intel_vsec_tpmi_init()`:
   ```c
   struct pci_dev *pci_dev = to_pci_dev(vsec_dev->dev);
   ```

While the commit message says "same APIs work for PCI and ACPI parents", these unconverted sites will crash with non-PCI devices. It would be prudent to add `dev_is_pci()` guards or at minimum a comment at each `to_pci_dev()` site asserting that only PCI devices reach that path. The `intel_vsec_set_mapping()` function correctly adds such a guard:

```c
if (!dev_is_pci(vsec_dev->dev))
    return -ENODEV;
```

This pattern should be applied more consistently.

**Minor: `xe_pmt_telem_read` uses `kdev_to_xe_device()`** — this needs `kdev_to_xe_device()` to exist in the xe driver. It does (it's a standard xe helper), so this is fine.

---

---
Generated by Claude Code Patch Reviewer

[PATCH v5 5/6] platform/x86/intel/vsec: Return real error codes from registration path

[David E. Box]

Stop collapsing registration results into booleans. Make
intel_vsec_walk_header() return int and propagate the first non-zero error
from intel_vsec_register_device(). intel_vsec_register() now returns that
error directly and 0 on success.

This preserves success behavior while surfacing meaningful errors instead
of hiding them behind a bool/-ENODEV, which makes debugging and probe
ordering issues clearer.

Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
Changes in v5:
- No change

Changes in v4:
- No change

Changes in v3:
- No change

Changes in v2:
- No chnage (previous patch 3)

 drivers/platform/x86/intel/vsec.c | 16 ++++++----------
 1 file changed, 6 insertions(+), 10 deletions(-)

diff --git a/drivers/platform/x86/intel/vsec.c b/drivers/platform/x86/intel/vsec.c
index a547e4b98245..34b2c19ecff0 100644
--- a/drivers/platform/x86/intel/vsec.c
+++ b/drivers/platform/x86/intel/vsec.c
@@ -461,20 +461,19 @@ static int intel_vsec_register_device(struct device *dev,
 	return -EAGAIN;
 }
 
-static bool intel_vsec_walk_header(struct device *dev,
-				   const struct intel_vsec_platform_info *info)
+static int intel_vsec_walk_header(struct device *dev,
+				  const struct intel_vsec_platform_info *info)
 {
 	struct intel_vsec_header **header = info->headers;
-	bool have_devices = false;
 	int ret;
 
 	for ( ; *header; header++) {
 		ret = intel_vsec_register_device(dev, *header, info, info->base_addr);
-		if (!ret)
-			have_devices = true;
+		if (ret)
+			return ret;
 	}
 
-	return have_devices;
+	return 0;
 }
 
 static bool intel_vsec_walk_dvsec(struct pci_dev *pdev,
@@ -580,10 +579,7 @@ int intel_vsec_register(struct device *dev,
 	if (!dev || !info || !info->headers)
 		return -EINVAL;
 
-	if (!intel_vsec_walk_header(dev, info))
-		return -ENODEV;
-	else
-		return 0;
+	return intel_vsec_walk_header(dev, info);
 }
 EXPORT_SYMBOL_NS_GPL(intel_vsec_register, "INTEL_VSEC");
 

-- 
2.43.0

Claude review: platform/x86/intel/vsec: Return real error codes from registration path

[Claude Code Review Bot]

Patch Review

**Overall: Good intent but introduces a semantic change in error behavior.**

The patch converts `intel_vsec_walk_header()` from returning `bool` (any device registered successfully) to `int` (first error encountered). This is a meaningful behavioral change:

```c
 for ( ; *header; header++) {
     ret = intel_vsec_register_device(dev, *header, info, info->base_addr);
-    if (!ret)
-        have_devices = true;
+    if (ret)
+        return ret;
 }
-return have_devices;
+return 0;
```

**The old behavior**: iterate all headers, register as many devices as possible, return success if *any* device was registered. Individual failures were silently ignored.

**The new behavior**: return the *first* error encountered and *stop processing remaining headers*.

This is a significant semantic change that the commit message describes as "preserves success behavior" but it doesn't fully:
- If the first header fails but the second would succeed, the old code would register the second device and return success. The new code fails immediately.
- For cases where `intel_vsec_register_device()` returns `-EAGAIN` (deferred dependency), the old code would continue and register other devices. The new code aborts.

Consider whether this short-circuit on first error is intentional. If only partial registration is acceptable, the old approach of continuing was more robust. If the intent is truly "all-or-nothing", the already-registered devices from earlier iterations should be cleaned up (which isn't done here).

**Suggestion**: Either restore the iterate-all-and-track-first-error approach (return first error only if *no* devices succeeded), or document that all-or-nothing semantics is the new intended behavior. The `-EAGAIN` case seems particularly problematic with the new short-circuit.

---

---
Generated by Claude Code Patch Reviewer

[PATCH v5 6/6] platform/x86/intel/vsec: Plumb ACPI PMT discovery tables through vsec

[David E. Box]

Some platforms expose PMT discovery via ACPI instead of PCI BARs. Add a
generic discovery source flag and carry ACPI discovery entries alongside
the existing PCI resource path so PMT clients can consume either.

Changes:
  - Add enum intel_vsec_disc_source { _PCI, _ACPI }.
  - Extend intel_vsec_platform_info and intel_vsec_device with source enum
    and ACPI discovery table pointer/
  - When src==ACPI, skip BAR resource setup and copy the ACPI discovery
    entries into the aux device.

No user-visible behavior change yet; this only wires ACPI data through vsec
in preparation for ACPI-enumerated PMT clients.

Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
Changes in v5:
- No change

Changes in v4:
- Use check_mul_overflow() instead of array_size() which was
  incorrectly checking for 0 anyway.

Changes in v3:
- Re-send with all changes intended for v2 which was sent without them
  being applied.

Changes in v2:
- Improve comment to clarify BAR resource setup doesn't apply to ACPI
  discovery
- Add missing #include for kmemdup()
- Use array_size() for overflow protection
  (review comments by Ilpo Järvinen)

 drivers/platform/x86/intel/vsec.c | 23 +++++++++++++++++++++++
 include/linux/intel_vsec.h        | 20 +++++++++++++++++++-
 2 files changed, 42 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/intel/vsec.c b/drivers/platform/x86/intel/vsec.c
index 34b2c19ecff0..7907505a365c 100644
--- a/drivers/platform/x86/intel/vsec.c
+++ b/drivers/platform/x86/intel/vsec.c
@@ -24,7 +24,9 @@
 #include <linux/intel_vsec.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
+#include <linux/overflow.h>
 #include <linux/pci.h>
+#include <linux/string.h>
 #include <linux/types.h>
 
 #define PMT_XA_START			0
@@ -109,6 +111,7 @@ static void intel_vsec_dev_release(struct device *dev)
 
 	ida_free(intel_vsec_dev->ida, intel_vsec_dev->auxdev.id);
 
+	kfree(intel_vsec_dev->acpi_disc);
 	kfree(intel_vsec_dev->resource);
 	kfree(intel_vsec_dev);
 }
@@ -320,6 +323,13 @@ static int intel_vsec_add_dev(struct device *dev, struct intel_vsec_header *head
 	 * auxiliary device driver.
 	 */
 	for (i = 0, tmp = res; i < header->num_entries; i++, tmp++) {
+		/*
+		 * Skip resource mapping check for ACPI-based discovery
+		 * since those tables are read from _DSD, not MMIO.
+		 */
+		if (info->src == INTEL_VSEC_DISC_ACPI)
+			break;
+
 		tmp->start = base_addr + header->offset + i * (header->entry_size * sizeof(u32));
 		tmp->end = tmp->start + (header->entry_size * sizeof(u32)) - 1;
 		tmp->flags = IORESOURCE_MEM;
@@ -338,6 +348,19 @@ static int intel_vsec_add_dev(struct device *dev, struct intel_vsec_header *head
 	intel_vsec_dev->base_addr = info->base_addr;
 	intel_vsec_dev->priv_data = info->priv_data;
 	intel_vsec_dev->cap_id = cap_id;
+	intel_vsec_dev->src = info->src;
+
+	if (info->src == INTEL_VSEC_DISC_ACPI) {
+		size_t bytes;
+
+		if (check_mul_overflow(intel_vsec_dev->num_resources,
+			sizeof(*info->acpi_disc), &bytes))
+			return -EOVERFLOW;
+
+		intel_vsec_dev->acpi_disc = kmemdup(info->acpi_disc, bytes, GFP_KERNEL);
+		if (!intel_vsec_dev->acpi_disc)
+			return -ENOMEM;
+	}
 
 	if (header->id == VSEC_ID_SDSI)
 		intel_vsec_dev->ida = &intel_vsec_sdsi_ida;
diff --git a/include/linux/intel_vsec.h b/include/linux/intel_vsec.h
index 4eecb2a6bac4..1fe5665a9d02 100644
--- a/include/linux/intel_vsec.h
+++ b/include/linux/intel_vsec.h
@@ -33,6 +33,11 @@ struct device;
 struct pci_dev;
 struct resource;
 
+enum intel_vsec_disc_source {
+	INTEL_VSEC_DISC_PCI,	/* PCI, default */
+	INTEL_VSEC_DISC_ACPI,	/* ACPI */
+};
+
 enum intel_vsec_id {
 	VSEC_ID_TELEMETRY	= 2,
 	VSEC_ID_WATCHER		= 3,
@@ -103,6 +108,10 @@ struct vsec_feature_dependency {
  * @parent:    parent device in the auxbus chain
  * @headers:   list of headers to define the PMT client devices to create
  * @deps:      array of feature dependencies
+ * @acpi_disc: ACPI discovery tables, each entry is two QWORDs
+ *             in little-endian format as defined by the PMT ACPI spec.
+ *             Valid only when @provider == INTEL_VSEC_DISC_ACPI.
+ * @src:       source of discovery table data
  * @priv_data: private data, usable by parent devices, currently a callback
  * @caps:      bitmask of PMT capabilities for the given headers
  * @quirks:    bitmask of VSEC device quirks
@@ -113,6 +122,8 @@ struct intel_vsec_platform_info {
 	struct device *parent;
 	struct intel_vsec_header **headers;
 	const struct vsec_feature_dependency *deps;
+	u32 (*acpi_disc)[4];
+	enum intel_vsec_disc_source src;
 	void *priv_data;
 	unsigned long caps;
 	unsigned long quirks;
@@ -124,7 +135,12 @@ struct intel_vsec_platform_info {
  * struct intel_vsec_device - Auxbus specific device information
  * @auxdev:        auxbus device struct for auxbus access
  * @dev:           struct device associated with the device
- * @resource:      any resources shared by the parent
+ * @resource:      PCI discovery resources (BAR windows), one per discovery
+ *                 instance. Valid only when @src == INTEL_VSEC_DISC_PCI
+ * @acpi_disc:     ACPI discovery tables, each entry is two QWORDs
+ *                 in little-endian format as defined by the PMT ACPI spec.
+ *                 Valid only when @src == INTEL_VSEC_DISC_ACPI.
+ * @src:           source of discovery table data
  * @ida:           id reference
  * @num_resources: number of resources
  * @id:            xarray id
@@ -138,6 +154,8 @@ struct intel_vsec_device {
 	struct auxiliary_device auxdev;
 	struct device *dev;
 	struct resource *resource;
+	u32 (*acpi_disc)[4];
+	enum intel_vsec_disc_source src;
 	struct ida *ida;
 	int num_resources;
 	int id; /* xa */

-- 
2.43.0

Claude review: platform/x86/intel/vsec: Plumb ACPI PMT discovery tables through vsec

[Claude Code Review Bot]

Patch Review

**Overall: Core feature patch. Has a resource leak issue and a minor doc typo.**

This adds the ACPI discovery infrastructure: `enum intel_vsec_disc_source`, `acpi_disc` arrays in both `intel_vsec_platform_info` and `intel_vsec_device`, and the `src` field.

**Issue 1 — Potential resource leak when `kmemdup` succeeds but device registration fails:**

In `intel_vsec_add_dev()`:
```c
intel_vsec_dev->acpi_disc = kmemdup(info->acpi_disc, bytes, GFP_KERNEL);
if (!intel_vsec_dev->acpi_disc)
    return -ENOMEM;
```

The `intel_vsec_dev` is declared with `__free(kfree)` cleanup, but `acpi_disc` is only freed in `intel_vsec_dev_release()` (the device release callback). If `intel_vsec_add_aux()` fails later in this function *before* the device is fully registered (before `device_add` succeeds), the `__free(kfree)` cleanup will free `intel_vsec_dev` but not `acpi_disc`, since `intel_vsec_dev_release` won't be called. This mirrors the existing pattern for `resource` though — `resource` has the same issue with `no_free_ptr(res)` — so this may be acceptable if the existing pattern is considered safe. But it's worth verifying.

Actually, looking more carefully: `intel_vsec_dev` uses `__free(kfree)` and is later passed via `no_free_ptr()` to `intel_vsec_add_aux()`. If `no_free_ptr()` has already been called, ownership has been transferred. But the `acpi_disc` allocation happens *before* `no_free_ptr()`, so if the function returns early between the `kmemdup` and `no_free_ptr`, `acpi_disc` will leak since `kfree(intel_vsec_dev)` doesn't also free `intel_vsec_dev->acpi_disc`.

**Issue 2 — `check_mul_overflow` alignment:**

```c
if (check_mul_overflow(intel_vsec_dev->num_resources,
    sizeof(*info->acpi_disc), &bytes))
```

The `sizeof(*info->acpi_disc)` evaluates to `sizeof(u32[4])` = 16 bytes, and `num_resources` is `int`. The `check_mul_overflow` macro should handle the type promotion correctly, but `bytes` is `size_t` while `num_resources` is `int` — if `num_resources` were somehow negative, this could behave unexpectedly. It won't happen in practice since `num_resources` is validated earlier, but using `(size_t)intel_vsec_dev->num_resources` would be more defensive.

**Issue 3 — Comment typo in the header:**

```c
 * @acpi_disc: ACPI discovery tables, each entry is two QWORDs
 *             in little-endian format as defined by the PMT ACPI spec.
 *             Valid only when @provider == INTEL_VSEC_DISC_ACPI.
```

`@provider` should be `@src` to match the actual field name.

**Issue 4 — Breaking the resource loop early:**

```c
for (i = 0, tmp = res; i < header->num_entries; i++, tmp++) {
    if (info->src == INTEL_VSEC_DISC_ACPI)
        break;
```

This `break` on ACPI effectively leaves `res` allocated but empty (all zeroed). The `num_resources` is still set to `header->num_entries`, so downstream code might expect valid resources. It might be cleaner to skip the allocation and loop entirely for ACPI, or set `num_resources` to 0 for ACPI devices.

---
Generated by Claude Code Patch Reviewer

Claude review: platform/x86/intel/vsec: Prep for ACPI PMT discovery

[Claude Code Review Bot]

Overall Series Review

Subject: platform/x86/intel/vsec: Prep for ACPI PMT discovery
Author: "David E. Box" <david.e.box@linux.intel.com>
Patches: 7
Reviewed: 2026-02-27T14:12:35.971128

---

This is a well-structured 6-patch series by David E. Box that prepares the Intel VSEC subsystem for ACPI-based PMT discovery. The series progressively decouples the vsec layer from PCI-specific types, improves const-correctness and error handling, and finally adds ACPI discovery table plumbing. The patch ordering is logical and each patch has a clear purpose.

The series is mostly clean with good incremental refactoring. However, there are a few issues worth addressing, particularly around safety in patch 4's PCI-to-device conversion (unchecked `to_pci_dev()` calls in downstream consumers), a semantic change in error handling in patch 5, and a potential resource leak in patch 6.

---

---
Generated by Claude Code Patch Reviewer