5

today my nas debian 9 based started to write out this error in the startup phase, the same is reported in red by calling journalctl -xe:

ACPI Error: SMBus/IPMI/Generic write requires Buffer of length 66, found length 32 (20160831/exfield-427)
ACPI Error: Method parse/execution failed [\SB.PMIO._PMM] (Node ffff8a71878aeaf0), AE_AML_BUFFER_LIMIT (20160831/psparse-543)
ACPI Exception: AE_AML_BUFFER_LIMIT, Evaluating _PMM (20160831/power_meter-338)

I have a double raid1 (sda/sdb and sdc/sdd) inside this nas, could it be that one of the disk is going to be defective ?

Should I be worried ? What could have caused this error and how can I fix it?

Could it be an error given by the fact that I pressed the power off sometimes instead of login and write manually shutdown -h now ?

Thanks

user3450548
  • 2,714
  • 11
  • 29
  • 46

2 Answers2

8

ACPI is the subsystem that uses information from the BIOS to control hardware, mostly for power management, temperature sensing, and related issues. SMBus is a simple two-wire communications protocol, used as side channel to access temperature sensors and other hardware.

So your BIOS contains sloppy ACPI data that specifies the wrong buffer size for a write action on that channel. _PMM seems to indicate that it is related to some chip that measures something power related. Which means it probably fails to initialize some chip that monitors voltage levels somewhere. Which is usually not a problem (unless you want to measure voltage levels, and shut down your computer if there's something odd, which is a feature you have to install and set up, and is usually only used on servers).

You can investigate that by looking at the ACPI data, but that requires a bit of expertise. Sloppy BIOS data is nothing unusual (unfortunately), vendors suck at setting up the BIOS properly, as they only test with the pre-installed Windows drivers which may work even with faulty data.

dirkt
  • 31,679
  • 3
  • 40
  • 73
4

HP's \\SB._OSC method violates the ACPI spec. \\SB._OSC is documented as taking an 8-byte argument, while the HP firmware is attempting to interpret it as a 12-byte one. HP have ignored the spec for this method and implemented their own as part of the Processor Clocking Control specification.

We can't work around this without risking breaking spec-compliant machines. Nothing will actually be broken as a result of printing this error. This error indicates that kernel was not able to read ACPI power meter values through invocation of ACPI method.

Consider blacklisting acpi_power_meter kernel module.

modprobe -r acpi_power_meter
echo "blacklist acpi_power_meter" >> /etc/modprobe.d/hwmon.conf
echo "install acpi_power_meter /bin/false" >> /etc/modprobe.d/hwmon.conf

vi /etc/sensors3.conf

Go to the bottom of the file and add:

chip “power_meter-acpi-0”
ignore power1

Then reboot.

Stephen Kitt
  • 411,918
  • 54
  • 1,065
  • 1,164
Navin Kumar
  • 41
  • 2
  • This seems to be borrowing from https://partner-bugzilla.redhat.com/show_bug.cgi?id=616449 without attributing properly. The original comment was made by Matthew Garrett on 2010-07-20. – Richlv Apr 27 '21 at 10:25