Throwing this over the wall, since I'm still mostly off-line.
I'm not sure if this is some eSATA issue (apparently it happens on
eSATA hotplug), or just ACPI.
But it sounds like an endless loop in an error condition, considering
that the printk's apparently happen 2000x a second.
Or possibly a screaming level-triggered interrupt that doesn't get
shut up because of the error?
There's no traceback, so it's hard to guess.
It's apparently been going on for years, and back when Aaron
originally opened the bugzilla he was responsive to debug queries, so
I'd appreciate it if somebody took a look..
>From the dmesg and lspci in the launchpad report, the machine itself
looks fairly normal (ASRock Z170M Pro4S, mostly intel chips with a
nVidia GTX 770 and Atheros WiFi). Nothing strange stands out.
My only reaction is that the BIOS is ancient and should probably be
updated and that might fix it (at the time of the original report it
was "BIOS P1.40 08/13/2015", and accroding to
https://www.asrock.com/mb/Intel/Z170%20Pro4S/#BIOS
the current version is 7.50).
So Aaron, I'd suggest you try a BIOS update since it does seem to be
related to ACPI tables, but otherwise I'm throwing it over to Rafael,
Len and Jens..
Linus
---------- Forwarded message ---------
From: Aaron Franke <arnfranke@xxxxxxxxx>
Date: Fri, Sep 28, 2018 at 1:59 AM
Subject: Kernel spamming errors to log file at about 1.2 MB per
second, rendering my system inoperable
To: torvalds@xxxxxxxxxxxxxxxxxxxx <torvalds@xxxxxxxxxxxxxxxxxxxx>
Hello Mr. Torvalds,
I know that you are a busy man, but I am writing this E-mail to you
because I believe I've found a big problem in Linux:
* When there are errors, there is no limit on how many times, or how
frequently, these errors are logged.
My "/var/log/kern.log" file gains a 631 byte error message about once
every 0.5 milliseconds (or 0.0005 seconds / 2000 times per second)
resulting in about 1.2 MB of data being written to my
"/var/log/kern.log" file every second.
This is very bad, my entire system becomes inoperable if the root
partition fills up. I have temporarily worked around this issue by
mounting /var/log elsewhere, but this is still a serious issue.
Please consider adding a limit to the amount of content that is logged
in the "/var/log/kern.log" file. I would appreciate it very much.
I have reported an issue here
https://bugzilla.kernel.org/show_bug.cgi?id=188331 but the previous
person who tried to solve the issue seems to have been unsuccessful.
Sincerely,
Aaron Franke,
A novice programmer and Linux user experiencing a problem with Linux.
