Linux Fixes Hosts Randomly Rebooting During Virtualization With Ryzen 7000/8000 CPUs

[phoronix]

Phoronix: Linux Fixes Hosts Randomly Rebooting During Virtualization With Ryzen 7000/8000 CPUs

Ahead of the Linux 6.12 kernel release expected today there is a last minute "x86/urgent" pull request. Notable with this last minute x86 urgent fixes for Linux 6.12 -- and also to be back-ported to prior kernel versions -- is working around an issue with AMD Ryzen Zen 4 client processors such as the Ryzen 7000/8000 series processors when making use of virtualization that could lead to the host randomly being rebooted...

https://www.phoronix.com/news/Linux-Clear-VMLOAD-VMSAVE-Zen4

[mrpops2ko]

this unfortunately didn't fix the issue for me :/

theres a thread on the proxmox forums about this, lots of us with 7950x's have had this since the launch and it seems specifically the 7950x that are affected.

i did some in place modifications to add this namely in

/etc/modprobe.d/kvm-amd.conf

put this

options kvm-amd nested=1 vls=0

hard reboots without any time between once a week to once a month

[sophisticles]

I thought only Intel CPUs had stability issues.

Imagine all the ignoramuses that built systems around these AMD CPUs to run VMs only to find that they are not stable.

There is no kernel panic, no log entries, no relevant output to serial console. It is as if platform is simply hard reset. It seems time to reproduce it varies from system to system and can be dependent on workload and even specific CPU model."

I can confirm this behavior with Ryzen 1600 running different distros bare metal and doing long encodes that last hours.

Sometimes the system can complete the work, sometimes after hours of going strong it will just shutdown as if it got tired of working.

This does not happen on any Intel platform I have tested.

This is why I don't buy AMD or trust AMD CPUs any longer.

Of course it could just be a Linux issue.

[gustavoar]

I thought only Intel CPUs had stability issues.

Imagine all the ignoramuses that built systems around these AMD CPUs to run VMs only to find that they are not stable.




I can confirm this behavior with Ryzen 1600 running different distros bare metal and doing long encodes that last hours.

Sometimes the system can complete the work, sometimes after hours of going strong it will just shutdown as if it got tired of working.

This does not happen on any Intel platform I have tested.

This is why I don't buy AMD or trust AMD CPUs any longer.

Of course it could just be a Linux issue.

I have an AMD 7950x and I run VMs using QEMU constantly in Linux and never experienced this issue. So, it doesn't affect everyone.

[intelfx]

I have an AMD 7950x and I run VMs using QEMU constantly in Linux and never experienced this issue. So, it doesn't affect everyone.

This issue is about nested virtualization, i.e. VMSAVE/VMLOAD in the guest. You won't ever experience this issue unless you really go out of your way (and get unlucky with the BIOS/ucode).

[dayone]

I thought only Intel CPUs had stability issues.

Imagine all the ignoramuses that built systems around these AMD CPUs to run VMs only to find that they are not stable.




I can confirm this behavior with Ryzen 1600 running different distros bare metal and doing long encodes that last hours.

Sometimes the system can complete the work, sometimes after hours of going strong it will just shutdown as if it got tired of working.

This does not happen on any Intel platform I have tested.

This is why I don't buy AMD or trust AMD CPUs any longer.

Of course it could just be a Linux issue.

That was an issue with Early Ryzen 1000 CPUs, you shouldve RMAd your CPU like everyone else at that time.

This Post is about nested VMs aka VMs running in VMs. Its an Software issue and not a hardware issue like with Intel CPUs.

[intelfx]

Imagine all the ignoramuses that built systems around these AMD CPUs to run VMs only to find that they are not stable.

Oh boy, someone is really salty.

I can confirm this behavior with Ryzen 1600 running different distros bare metal and doing long encodes that last hours.

It's not this behavior. You should really learn some reading comprehension skills one of these days.

[domih]

Damn, another exercise in sophistry in a thread. What an idiotic post, attempting to equate a software bug to a hardware damaging bug under the fallacious use of "stability". When you see this commenter name, you know you're going to read something mostly deranged. Sometimes it is so blatant that it is funny. Most of the times, it's tragic.

[caligula]

I can confirm this behavior with Ryzen 1600 running different distros bare metal and doing long encodes that last hours.

Sometimes the system can complete the work, sometimes after hours of going strong it will just shutdown as if it got tired of working.

First gen Ryzen had some stability issues in idle mode. There were fixes. Some had HW bugs, some could be updated with a bios patch or Linux service poking some port. I've used several generations of ryzens since then without any hardware lockups. Typically uptime is 20+ days according to uprecords. Then it's time to update the kernel.

[WannaBeOCer]

That was an issue with Early Ryzen 1000 CPUs, you shouldve RMAd your CPU like everyone else at that time.

This Post is about nested VMs aka VMs running in VMs. Its an Software issue and not a hardware issue like with Intel CPUs.

Intel's issue is also a software, just like AMD's Ryzen 7000X3D chips that were overvolting too high due to their microcode causing them to instantly fry. Unlike AMD, Intel's was slow roasting.