Arch Linux

frix

Member

Registered: 2011-05-02

Posts: 37

[SOLVED-ish] "vgaarb: VGA decodes changed" spammed in kernel logs ?

Hello,

For a few weeks now (currently on kernel version 7.1.8) I see the following kernel logs every 10 seconds on one of my machines (a server that doesn't have a connected display and isn't running any kind of display server) running Arch:

Aug 13 21:54:25 xxxx kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io,decodes=io:owns=io+mem
Aug 13 21:54:25 xxxx kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io,decodes=io:owns=io+mem
Aug 13 21:54:35 xxxx kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io,decodes=io:owns=io+mem
Aug 13 21:54:35 xxxx kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io,decodes=io:owns=io+mem
Aug 13 21:54:45 xxxx kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io,decodes=io:owns=io+mem
Aug 13 21:54:45 xxxx kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io,decodes=io:owns=io+mem

This is a bit annoying even though it appears harmless, but I wasn't able to figure out what exactly is causing it and it never happened before in the 10 years I've been running Arch on this machine. Any idea?

Last edited by frix (Yesterday 19:57:53)

seth

Member

From: Won't reply 2 private help req

Registered: 2012-09-03

Posts: 77,550

Re: [SOLVED-ish] "vgaarb: VGA decodes changed" spammed in kernel logs ?

a server that doesn't have a connected display

I assume you could just blacklist i915
10s seems unlikely to be the chip to re-power (from S3 or so) and more likely desperate attempts to find a monitor (so plugging one would shut this down?)

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

frix

Member

Registered: 2011-05-02

Posts: 37

Re: [SOLVED-ish] "vgaarb: VGA decodes changed" spammed in kernel logs ?

Yes, blacklisting the module (+removing it from the initcpio modules because I had that enabled since forever on that machine) did the trick. I'm not sure what was causing this, but at least the workaround broke nothing else and I didn't need that module anyway. Thanks for the idea!