Re: [PATCH] platform/x86: toshiba_acpi: validate ACPI object type in tci_raw()

[Tristan Madani]

On Mon, 9 Nov 2026, Ilpo Järvinen wrote:
> Your ACPI validation patches seem to lack information whether these
> actually have been observed to occur on some device.

No, this was found by code audit, not observed on hardware.  The
tci_raw() function accesses the returned ACPI object as a package
without checking the type field first.  If the firmware method ever
returns something other than a package (e.g. an integer or buffer),
the code interprets whatever the union holds as package.count and
package.elements, which is undefined.

The sibling hp-bioscfg driver already validates the return type in
the same way (biosattr-interface.c line 159), but toshiba_acpi
predates that and never picked up the check.

I can resend as v2 with the changelog explicitly noting this is a
hardening fix found by code inspection, if that helps.