Created attachment 294197 [details] dmesg When the Lenovo Ideapad 330s is running on battery, the following error message is printed in the kernel log every second or so. [ 0.967136] ACPI BIOS Error (bug): Could not resolve symbol [^^^GFX0.AFN2], AE_NOT_FOUND (20200717/psargs-330) [ 0.967173] ACPI Error: Aborting method \_SB.PCI0.LPCB.H_EC._QC9 due to previous error (AE_NOT_FOUND) (20200717/psparse-529) This does not happen when the AC adapter is plugged in, and the laptop does not have other symptoms. Does not seem specific to a Linux kernel version.
Created attachment 294199 [details] acpidump
$ grep AFN2 *.dsl dsdt.dsl: External (_SB_.PCI0.GFX0.AFN2, MethodObj) // 1 Arguments dsdt.dsl: ^^^GFX0.AFN2 (Local1) dsdt.dsl: ^^^GFX0.AFN2 (Local1) This is apparently a BIOS bug. The buggy AML invokes an ACPI control method named AFN2, which is never defined. Sorry that I will close this one as we can do nothing in kernel for this issue.
"The fact that ACPI was designed by a group of monkeys high on LSD, and is some of the worst designs in the industry obviously makes running it at any point pretty damn ugly." - Linus Torvalds Most computers have ACPI bugs and the kernel contains workarounds for many of them. So, why should this particular issue get closed?
I close this because we don't observe any functionality issue so far. And the error messages are caused by buggy BIOS code we can not fix. I can propose a customized DSDT and you can rebuild your kernel with the customized DSDT, and you will not see these errors. Another potential issue is that why it fires every second. please try following command and attach the int-before and int-after log files. grep . /sys/firmware/acpi/interrupts/* > int-before; sleep 10; grep . /sys/firmware/acpi/interrupts/* > int-after
Created attachment 295991 [details] int-before
Created attachment 295993 [details] int-after
The EC event indeed happens around every second. To me, the handling for the EC event (0xC9) is broken as it references the AFN2 method which does not exist at all. And this results in the EC event not handled properly and it keeps on firing. As I mentioned earlier, this is a BIOS issue. For some of the ACPI BIOS bugs, we can workaround them in Linux kernel, but for issues like this, we can do nothing in kernel because we don't know what is missing in the BIOS and it can only get fixed in ACPI BIOS code by the OEM Vendor. So I will close this bug report, sorry that we can not help on this issue. You can check if there is any new BIOS release or not.