Turns out this has already been fixed in ACPICA so expect it in a mainline kernel near you soon. The fix is to simply not use the sleep registers unless the related reduced hardware flag is set, even if the sleep registers exist.
https://github.com/acpica/acpica/commit/34f226fa2643f1d2e6527ea4edb24947cfe1fb6a
http://article.gmane.org/gmane.linux.acpi.devel/61011
Turns out this has already been fixed in ACPICA so expect it in a mainline kernel near you soon. The fix is to simply not use the sleep registers unless the related reduced hardware flag is set, even if the sleep registers exist.
https:/ /github. com/acpica/ acpica/ commit/ 34f226fa2643f1d 2e6527ea4edb249 47cfe1fb6a
http:// article. gmane.org/ gmane.linux. acpi.devel/ 61011