kernel/linux.git/drivers/platform/x86/intel/vsec.c, branch linux-7.1.y

platform/x86/intel/vsec: Fix enable_cnt imbalance on PCIe error recovery

2026-05-19T14:03:36+00:00

After a PCIe Uncorrectable Error has been reported by a device with Intel Vendor Specific Extended Capabilities and has been recovered through a Secondary Bus Reset, its driver calls intel_vsec_pci_probe() to rescan and reinitialize VSECs. intel_vsec_pci_probe() invokes pcim_enable_device() and thereby adds another devm action which calls pcim_disable_device() on driver unbind. So once the driver unbinds, pcim_disable_device() will be called as many times as an Uncorrectable Error occurred, plus one. This will lead to an enable_cnt imbalance on driver unbind. Additionally, since commit dc957ab6aa05 ("platform/x86/intel/vsec: Add private data for per-device data"), a devm_kzalloc() allocation is leaked on every Uncorrectable Error. Avoid by splitting the VSEC rescan out of intel_vsec_pci_probe() into a separate helper and calling that on PCIe error recovery. Fixes: 936874b77dd0 ("platform/x86/intel/vsec: Add PCI error recovery support to Intel PMT") Signed-off-by: Lukas Wunner Cc: stable@vger.kernel.org # v6.0+ Link: https://patch.msgid.link/bd594d09fa866dc51dddc9a447c3b23f9b1402cc.1778736835.git.lukas@wunner.de Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

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

2026-03-17T16:51:18+00:00

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 Link: https://patch.msgid.link/20260313015202.3660072-7-david.e.box@linux.intel.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

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

2026-03-17T16:51:17+00:00

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 Signed-off-by: David E. Box Link: https://patch.msgid.link/20260313015202.3660072-6-david.e.box@linux.intel.com Signed-off-by: Ilpo Järvinen

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

2026-03-17T16:51:15+00:00

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 Signed-off-by: David E. Box Link: https://patch.msgid.link/20260313015202.3660072-5-david.e.box@linux.intel.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

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

2026-03-17T16:51:13+00:00

This refactor 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. Signed-off-by: David E. Box Reviewed-by: Ilpo Järvinen Reviewed-by: Michael J. Ruhl Link: https://patch.msgid.link/20260313015202.3660072-4-david.e.box@linux.intel.com Signed-off-by: Ilpo Järvinen

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

2026-03-17T16:51:12+00:00

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 Reviewed-by: Michael J. Ruhl Link: https://patch.msgid.link/20260313015202.3660072-3-david.e.box@linux.intel.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

platform/x86/intel/vsec: Refactor base_addr handling

2026-03-17T16:51:09+00:00

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 Reviewed-by: Michael J. Ruhl Link: https://patch.msgid.link/20260313015202.3660072-2-david.e.box@linux.intel.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen

Convert 'alloc_obj' family to use the new default GFP_KERNEL argument

2026-02-22T01:09:51+00:00

This was done entirely with mindless brute force, using git grep -l '\

treewide: Replace kmalloc with kmalloc_obj for non-scalar types

2026-02-21T09:02:28+00:00

This is the result of running the Coccinelle script from scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to avoid scalar types (which need careful case-by-case checking), and instead replace kmalloc-family calls that allocate struct or union object instances: Single allocations: kmalloc(sizeof(TYPE), ...) are replaced with: kmalloc_obj(TYPE, ...) Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...) are replaced with: kmalloc_objs(TYPE, COUNT, ...) Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...) are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...) (where TYPE may also be *VAR) The resulting allocations no longer return "void *", instead returning "TYPE *". Signed-off-by: Kees Cook

platform/x86/intel/vsec: Add Nova Lake PUNIT support

2026-01-26T14:42:02+00:00

Add PCI ID for Nova Lake, supporting PUNIT telemetry. Signed-off-by: David E. Box Link: https://patch.msgid.link/20260122022110.3231344-1-david.e.box@linux.intel.com Signed-off-by: Ilpo Järvinen