ACPI errors swamping jounralctl, messages, and syslog

Discussion:

(too old to reply)

William Unruh

2022-07-29 15:26:22 UTC

I just noticed that a whole raft of ACPI error messages swamped my
journal and messages and syslog (I run rsyslog)
About 2GB of error messages in less than a minute.
Jul 21 15:39:16 syzygy kernel: [69373.824163] ACPI BIOS Error (bug): Could not resolve symbol [\_GPE._L66.I*4], AE_NOT_FOUND (20200925/psargs-330)
Jul 21 15:39:16 syzygy kernel: [69373.824165] ACPI Error: Aborting method \_GPE._L66 due to previous error (AE_NOT_FOUND) (20200925/psparse-531)

Mageia 7
5.10.46-desktop-1.mga7 kernel.

About a minute before those errors started the sysmte apparently entered
sleep

Jul 21 15:38:44 syzygy systemd[1]: Reached target Sleep.
Jul 21 15:38:44 syzygy systemd[1]: Starting Hibernate...
Jul 21 15:38:44 syzygy kernel: [69371.050303] PM: hibernation: hibernation entry
Jul 21 15:38:44 syzygy systemd-sleep[19492]: Suspending system...
Jul 21 15:39:16 syzygy kernel: [69371.058790] Filesystems sync: 0.003 seconds

What is disturbing is that this could so rapidly fill the logs.
At about 15:54, the system seemed to reboot, or reload itself

Jul 21 15:54:02 syzygy fedora-loadmodules[510]: Loading modules: nvram rtsx_usb evdev acpi-cpufreq cpufreq_powersave cpufreq_conservative cpufreq_ondemand
Jul 21 15:54:02 syzygy kernel: [ 0.000000] Linux version 5.10.46-desktop-1.mga7 (***@ecosse.mageia.org) (gcc (Mageia 8.4.0-1.mga7) 8.4.0, GNU ld (GNU Binutils) 2.33.1) #1 SMP Thu Jun 24 14:55:57 UTC 2021
...

This swamp does not seem to have reoccured. But wht is disturbing is the
(Bug) item in those error messages. (NOt sure what it means the bug is
in, whether the kernel, or in the firmware)

Has anone else seen this? Is there any clue as to what could be going
on?

realise that Magea 7 is out of date, but this is worrysome
none-the-less.

Aragorn

2022-07-29 15:43:27 UTC

Permalink

Post by William Unruh
Could not resolve symbol [\_GPE._L66.I*4], AE_NOT_FOUND
(20200925/psargs-330) Jul 21 15:39:16 syzygy kernel: [69373.824165]
ACPI Error: Aborting method \_GPE._L66 due to previous error
(AE_NOT_FOUND) (20200925/psparse-531)
[...]
This swamp does not seem to have reoccured. But wht is disturbing is
the (Bug) item in those error messages. (NOt sure what it means the
bug is in, whether the kernel, or in the firmware)

Most likely the firmware, judging by the format of the error message.

--
With respect,
= Aragorn =

William Unruh

2022-07-29 16:16:24 UTC

Permalink

Some of the Web search responses seem to blame this on the Nvidia
graphics drivere. However my system is not Nvidia. It is an Asus Zenbook
13 with Intel graphics

lspci
00:02.0 VGA compatible controller: Intel Corporation WhiskeyLake-U GT2 [UHD Graphics 620]

This error message occured again over the next few days, but seems to
have stopped now. The repeats were also only a few lines, not 2GB of
error messages.
Something went into a semi-infinte loop.(12Megalines of errors 1.8GB )

Post by William Unruh
I just noticed that a whole raft of ACPI error messages swamped my
journal and messages and syslog (I run rsyslog)
About 2GB of error messages in less than a minute.
Jul 21 15:39:16 syzygy kernel: [69373.824163] ACPI BIOS Error (bug): Could not resolve symbol [\_GPE._L66.I*4], AE_NOT_FOUND (20200925/psargs-330)
Jul 21 15:39:16 syzygy kernel: [69373.824165] ACPI Error: Aborting method \_GPE._L66 due to previous error (AE_NOT_FOUND) (20200925/psparse-531)
Mageia 7
5.10.46-desktop-1.mga7 kernel.
About a minute before those errors started the sysmte apparently entered
sleep
Jul 21 15:38:44 syzygy systemd[1]: Reached target Sleep.
Jul 21 15:38:44 syzygy systemd[1]: Starting Hibernate...
Jul 21 15:38:44 syzygy kernel: [69371.050303] PM: hibernation: hibernation entry
Jul 21 15:38:44 syzygy systemd-sleep[19492]: Suspending system...
Jul 21 15:39:16 syzygy kernel: [69371.058790] Filesystems sync: 0.003 seconds
What is disturbing is that this could so rapidly fill the logs.
At about 15:54, the system seemed to reboot, or reload itself
Jul 21 15:54:02 syzygy fedora-loadmodules[510]: Loading modules: nvram rtsx_usb evdev acpi-cpufreq cpufreq_powersave cpufreq_conservative cpufreq_ondemand
...
This swamp does not seem to have reoccured. But wht is disturbing is the
(Bug) item in those error messages. (NOt sure what it means the bug is
in, whether the kernel, or in the firmware)
Has anone else seen this? Is there any clue as to what could be going
on?
realise that Magea 7 is out of date, but this is worrysome
none-the-less.

David W. Hodgins

2022-07-29 16:51:50 UTC

Permalink

It's a bug in the firmware being noticed by the kernel and reported. Something
being displayed triggered the bug.

Make sure you limit the max size of the journal ...
$ grep -v -e ^'#' -e ^$ /etc/systemd/journald.conf
[Journal]
SystemMaxUse=200M
RuntimeMaxUse=200M
ForwardToSyslog=yes
ForwardToConsole=yes
TTYPath=/dev/tty12

Also, see https://bbs.archlinux.org/viewtopic.php?id=266085
Upgrade to Mageia 8.

Regards, Dave Hodgins

faeychild

2022-07-29 22:31:40 UTC

Permalink

Post by David W. Hodgins
Make sure you limit the max size of the journal ...
$ grep -v -e ^'#' -e ^$ /etc/systemd/journald.conf
[Journal]
SystemMaxUse=200M
RuntimeMaxUse=200M
ForwardToSyslog=yes
ForwardToConsole=yes
TTYPath=/dev/tty12
Also, see https://bbs.archlinux.org/viewtopic.php?id=266085
Upgrade to Mageia 8.
Regards, Dave Hodgins

Even as root I don't get this. Did it change for MGA 8?

[***@unimatrix ~]# grep -v -e ^'#' -e ^$ /etc/systemd/journald.conf
[Journal]
ForwardToSyslog=yes
ForwardToConsole=yes
TTYPath=/dev/tty12
[***@unimatrix ~]#

--
faeychild
Running plasmashell 5.20.4 on 5.15.55-desktop-2.mga8 kernel.
Mageia release 8 (Official) for x86_64 installed via Mageia-8-x86_64-DVD.iso

David W. Hodgins

2022-07-29 22:47:54 UTC

Permalink

Post by faeychild

Even as root I don't get this. Did it change for MGA 8?
[Journal]
ForwardToSyslog=yes
ForwardToConsole=yes
TTYPath=/dev/tty12

No, it hasn't changed. The point of my message is that it's a change
each administrator should make based on how much space they have and
how long they want the messages to be retained.

With 200M, I usually have about a month of journal messages. The default
is 15% of the file system the journal is stored in (not 15% of free space).

With the default, if the file system has 15% of it's size or less available
for the journal, then since journald runs as root it can fill the file system.
A full file system can stop the system from booting if it can't write files
normally written during boot.

Regards, Dave Hodgins

Bit Twister

2022-07-30 03:32:04 UTC

Permalink

Post by faeychild

Even as root I don't get this. Did it change for MGA 8?
[Journal]
ForwardToSyslog=yes
ForwardToConsole=yes
TTYPath=/dev/tty12

Looking like you have been playing around with some settings somewhere.

# grep -v -e ^'#' -e ^$ /etc/systemd/journald.conf
[Journal]
ForwardToConsole=yes
TTYPath=/dev/tty12

David W. Hodgins

2022-07-30 04:51:12 UTC

Permalink

Post by Bit Twister

Post by faeychild
Even as root I don't get this. Did it change for MGA 8?
[Journal]
ForwardToSyslog=yes
ForwardToConsole=yes
TTYPath=/dev/tty12

Looking like you have been playing around with some settings somewhere.
# grep -v -e ^'#' -e ^$ /etc/systemd/journald.conf
[Journal]
ForwardToConsole=yes
TTYPath=/dev/tty12

The ForwardToSyslog is added when rsyslog is installed.

Regards, Dave Hodgins