Error during boot: x86/PAT: Xorg...

User_2OY8H

Hi,

Edit 09.04.2021: The error messeages are not "hv_balloon", the error messeages are:

[  36.538399] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.569184] x86/PAT: Xorg:1703 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  36.763840] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.764575] x86/PAT: Xorg:1785 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  36.865129] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.865965] x86/PAT: Xorg:1803 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  36.953439] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.954553] x86/PAT: Xorg:1819 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  37.169868] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  37.170882] x86/PAT: Xorg:1853 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

Original post:

since the update to UEK 5.4.17-2102.200.13.el8uek.x86_64 our systems dont boot anymore with the error message: hv_balloon: Max. dynamic size

Downgrading to the previous kernel (5.4.17-2036.104.5.el8uek.x86_64) solved this issue!

For all affected: this this the command to downgrade:

sudo dnf remove kernel-uek-5.4.17-2102.200.13.el8uek.x86_64

Regards

VolneiCP

Hi,

It would be more interesting to get the attention of colleagues who better understand this type of problem, if you inform us about:

- What is your hardware? processor, RAM, etc.

- Is it a VM or not?

- If possible 'dmesg'

And any other relevant information on the subject.

Fortunately I didn't have this problem on my Xeon E3 123x systems with Intel S1200BTL Motherboard.

Good luck!

User_2OY8H

Hi,

thanks, I thought it was clear.

Only VMs via Windows Server 2019 -> Hyper-V.

Hardware: Different, we can reproduce it on multiple systems. Mostly Intel Xeon CPUs, between 2 and 16 cores, dpending on the VM.

RAM: dynamic, start: 2048MB, min: 1024MB, max: 10240MB

But same outcome with static too.

We have the problem in the company. But I can reproduce it on my private Win10 Professional with Hyper-V (running on an Intel i7) as well. Sameerror messages...

Downgrading (as written above) solves the problem. Maybe its a Hyper-V related problem?

Same error at home:

I can connect cia SSH, dmesg output:

User_2OY8H

And dmesg output with kernel 5.4.17-2036.104.5.el8uek.x86_64:

User_2OY8H

Maybe its related to this (x86/PAT...)?

[  36.538399] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.569184] x86/PAT: Xorg:1703 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  36.763840] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.764575] x86/PAT: Xorg:1785 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  36.865129] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.865965] x86/PAT: Xorg:1803 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  36.953439] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  36.954553] x86/PAT: Xorg:1819 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

[  37.169868] Lockdown: Xorg: raw io port access is restricted; see man kernel_lockdown.7

[  37.170882] x86/PAT: Xorg:1853 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back

VolneiCP

Hi,

I´m not an expert in this type of subject, but I would like to leave the suggestion that you use the RAM model of fixed size and do a test in case you haven´t done it yet. Maybe this possible bug could be related to that.

User_2OY8H

Hi, we tested fixed and dynamic RAM - same error.

Is there a way I can tell this Oracle? I tried via https://bugzilla.oracle.com but their response was:

 This isn't a customer support portal...

Well, I dont want support. I want to report an obvious bug in their most recent kernel. So I dont know how I can address this? Is here someone from Oracle who can fordward this to the right persons?

I think the problem is related to Hyper-V installations and specific to kernal "UEK 5.4.17-2102.200.13.el8uek.x86_64". With the "5.4.17-2036.104.5.el8uek.x86_64" absolutely no problems. As shown above.

VolneiCP

Well ... on my post https://community.oracle.com/tech/apps-infra/discussion/4481130/kernel-crashes-for-no-apparent-reason#latest
The Avi read and helped by passing the problem on to the right people, maybe you can send him a message, maybe he can help you. 

https://community.oracle.com/tech/apps-infra/messages/add/Avi%20Miller-Oracle

VolneiCP

https://community.oracle.com/tech/apps-infra/discussion/comment/16791331#Comment_16791331

Access this link and see if there is anything that makes any sense to you:

https://www.quora.com/What-does-Xorg-raw-io-port-access-is-restricted-see-man-kernel_lockdown-7-mean-in-Fedora-when-trying-to-install-the-proprietary-NVIDIA-driver-and-the-screen-just-blanks

User_2OY8H

Thanks! :)

I installed today a fresh 8.3 "Full ISO" from here in my Hyper-V:

https://yum.oracle.com/oracle-linux-isos.html

2 CPU cores, static 2048 MB RAM, " Secure Boot" disabled in Hyper-V.

After the install all I did was "sudo dnf update -y". It updated and installed the 2102-kernel.

Same outcome:

I can now rule out a service that we/I installed (docker or something).

andreas.dijkman

https://community.oracle.com/tech/apps-infra/discussion/comment/16791242#Comment_16791242

Oracle isn't used to getting feedback from users without them wanting support. But @Avi Miller-Oracle has been the best in passing these kind of bugs/feedback on to the right people.

But it as a rather strange error regarding Xorg without a X-environment? But is the system booting correctly after installation of the latest kernel or do you get these errors and then the system halts/crashes?

Avi Miller-Oracle

I'll pass this onto the UEK team.

User_XVVJD

hv_balloon: Max. dynamic size is not an error message but an informational message that the kernel hv_balloon driver has successfully connected to the Hyper-V Dynamic Memory protocol.

This message is to be expected when running under a Hyper-V host with DM enabled, few seconds after the system boot.

Due to this delay it's often the last message in the kernel log. This does not mean that the actual problem is with hv_balloon driver (but neither does it exclude it).

You can try booting the kernel with module_blacklist=hv_balloon,hv-balloon kernel command line parameter.

If this helps, then it's definitely an issue related to the hv_balloon driver.

User_2OY8H

Hi, thanks Avi!

@ User_XVVJD, I know this too - now. I am not a Linux specialist. For me, last message where the system hangs was the problematic or "error" message. I have learned here about the "dmesg" tool, which revealed the real error messages. I edited the original (first) post now.

I think the problem is not related to the "hv_balloon", as there are a lot "x86/PAT: Xorg:1703 map pfn expected mapping type uncached-minus for [mem 0xf8800000-0xf8ffffff], got write-back" errors.

But I did as you mentioned, now the "hv_balloon" message is gone, only a white-blinking cursor is there. But the system hangs as before at the same stage.

Regards