* [6.18.y] Please pick ACPI platform parent lifetime fix
@ 2026-11-20 19:23 Artem Dinaburg
0 siblings, 0 replies; only message in thread
From: Artem Dinaburg @ 2026-11-20 19:23 UTC (permalink / raw)
To: stable; +Cc: andriy.shevchenko, lenb, linux-acpi, rafael.j.wysocki, rafael
Hi,
Please pick commit a9ba4dd2f18bf3f439d9ef0d8f375f90360ba1bd ("ACPI:
platform: Use acpi_bus_get_primary_device()") for 6.18.y. It fixes
CVE-2026-100072.
acpi_create_platform_device() looks up the parent's physical device with
acpi_get_first_physical_node(), which does not take a reference, so in
theory that device may be freed while it is still in use. The fix takes
one reference with acpi_bus_get_primary_device() and holds it through
resource inspection and platform-device registration.
The fix is in 6.6.160, 6.12.114 and 7.2.6, and in mainline since 7.3.
6.18.y received only its prerequisite, upstream commit 72530e1f72b0
("ACPI: bus: Introduce acpi_bus_get_primary_device()"), in 6.18.57; the
fix itself did not apply to 6.18.y at that time. It applies unchanged
since 6.18.60, which picked up the ACPI scan changes f234fdaae1ca,
7617cc05df28 and ab06eb920401 that it depends on for context.
Compiled drivers/acpi/acpi_platform.o on 6.18.64 from a cumulative
x86_64 tree using allmodconfig. No kernel boot or runtime test was
performed.
An LLM was used to check applicability and stable-queue history and to
draft this request; I reviewed the result.
Thanks,
Artem Dinaburg
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2026-11-20 19:23 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-20 19:23 [6.18.y] Please pick ACPI platform parent lifetime fix Artem Dinaburg
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox