AMD : BSOD unsupported processor since Windows build 26100.4202+ ( update kb5060842 + its preview kb5058499 )

  • Like
Reactions: pxmxfan, dcain, _gabriel and 1 other person
Ran into the bsod issue this morning. 7950x3d set to host. Switched it to EPYC-v4 and it booted.
 
The reboot problem is fixed but there is still an issue with GPU passthrough being either slow (using an Epyc 7773X i have to use Epyc-milan-v2 profile to get it to work fine !!!!???!!!), or not loading the driver correctly (both on AMD and Geforce GPUs) on AMD 5xxxx or 7xxxx CPU.
 
butch2k said:
The reboot problem is fixed but there is still an issue with GPU passthrough being either slow (using an Epyc 7773X i have to use Epyc-milan-v2 profile to get it to work fine !!!!???!!!), or not loading the driver correctly (both on AMD and Geforce GPUs) on AMD 5xxxx or 7xxxx CPU.
Click to expand...
Anyone else having this issue? I am going to image my drive just in case.
 
Just drag a window on the desktop, there is an awful amount of lag. Tested with both 3090s and a 9070xt, with latest drivers, no special modification to the conf file, used to work fine with host before the windows patch. It even affect my 7773X for some reasons, it never crashed due to windows patch but is slow as hell if i do no change the CPU type to milan rather than host. The windows insider version is not affected. Tested on 4 different hosts.
 
butch2k said:
slow as hell if i do no change the CPU type to milan rather than host
Click to expand...
What about with first fix :
M

Post in thread 'AMD : BSOD unsupported processor since Windows build 26100.4202+ ( update kb5060842 + its preview kb5058499 )'

The following settings resolved the "Unsupported Processor" issue:

args: -cpu 'host,arch_capabilities=off'

The following flags seem to be set when the CPU is host.
They were not present when using EPYC-v4 or physical installation.
  • Like
 
Just a follow up.

After today's Patch Tues update, 2025-07 Cumulative Update for Windows 11 Version 24H2 for x64-based Systems (KB5062553), I was able to revert the CPU back to host and the system boots properly.
 
  • Like
Reactions: Taomyn
Nick Kooij said:
Just a follow up.

After today's Patch Tues update, 2025-07 Cumulative Update for Windows 11 Version 24H2 for x64-based Systems (KB5062553), I was able to revert the CPU back to host and the system boots properly.
Click to expand...
Thanks for posting, that fixed it for me too.
 
Last edited:
Nick Kooij said:
Just a follow up.

After today's Patch Tues update, 2025-07 Cumulative Update for Windows 11 Version 24H2 for x64-based Systems (KB5062553), I was able to revert the CPU back to host and the system boots properly.
Click to expand...
Confirmed for me as well - two different hosts with different AMD processors now all boot when set to host
 
Taomyn said:
Confirmed for me as well - two different hosts with different AMD processors now all boot when set to host
Click to expand...
Still have an issue with Code 43 on nvidia cards, whatever trick i use, romfile, grub mod, etc... I had no issue before the whole unsupported processor fiasco and had not to resort to anything. AMD cards, the 9070XT at least seem to work fiine.
 
_gabriel said:
What about with first fix :
M

Post in thread 'AMD : BSOD unsupported processor since Windows build 26100.4202+ ( update kb5060842 + its preview kb5058499 )'

The following settings resolved the "Unsupported Processor" issue:

args: -cpu 'host,arch_capabilities=off'

The following flags seem to be set when the CPU is host.
They were not present when using EPYC-v4 or physical installation.
  • Like
Click to expand...
Forgot to answer, but no it does not work, the system is still slow as hell.
I'm fairly certain it has to do with the nvidia driver not working properly anymore as i have Code 43 errors on non Epyc systems and just a huge slowdown on Epyc, and no issue with AMD cards whatever CPU i use.
 
Tested them as well, along with a few other related to virt and i can confirm the issue on 3 different Proxmox server configurations, Zen 3, 4 and Epyc. Adjusted several CPU flags to no avail.
Could someone else test the Nvidia GPU passthrough on W11H2 patched on Proxmox 8.4.1 on an AMD platform ? I tried different passthrough method as well (blackklist, no blacklist, vfio, grub settings etc...) to no avail.
 
It worked normally with PVE8.4.1+9950X+RTX4060Ti.
The FFXVI benchmark also completed with a comfortable rating.

/etc/default/grub
-----
GRUB_CMDLINE_LINUX_DEFAULT="pcie_aspm=off acpi_enforce_resources=lax iommu=pt pcie_acs_override=downstream,multifunction vfio-pci.ids=10de:2803,10de:22bd"
-----

/etc/modprobe.d/blacklist-vfio.conf
-----
blacklist snd_hda_intel
blacklist nouveau
blacklist radeon
-----

/etc/modprobe.d/vfio.conf
-----
options vfio-pci disable_vga=y disable_idle_d3=y enable_sriov=y
options vfio_iommu_type1 allow_unsafe_interrupts=1
softdep radeon pre: vfio-pci
softdep amdgpu pre: vfio-pci
softdep nouveau pre: vfio-pci
softdep nvidiafb pre: vfio-pci
softdep snd_hda_intel pre: vfio-pci
-----

etc/modprobe.d/kvm.conf
-----
options kvm nested=1
-----

/etc/pve/qemu-server/1033001.conf
-----
agents: 1
args: -fw_cfg 'opt/ovmf/X-PciMmio64Mb,string=65536' -smp 'cpus=32,sockets=1,threads=2,cores=16,maxcpus=32'
bios: ovmf
boot: order=scsi0;sata0;net0
cores: 32
cpu: custom-host-fix
efidisk0: local-zfs:vm-1033001-disk-0,efitype=4m,pre-enrolled-keys=1,size=4M
hookscript: local:snippets/pcie_hostdev.sh
hostpci0: 0000:13:00.0,pcie=1
hostpci1: 0000:7b:00.3,pcie=1
hostpci2: 0000:7b:00.4,pcie=1
hostpci3: 0000:01:00.0,pcie=1,x-vga=1
hostpci4: 0000:01:00.1,pcie=1
machine: pc-q35-9.2+pve1
memory: 16384
meta: creation-qemu=9.2.0,ctime=1749808390
name: Win11
net0:virtio=BC:24:11:DF:82:CC,bridge=vmbr0
numa: 0
ostype: win11
parent: init_setup
sata0: local:iso/virtio-win-0.1.271.iso,media=cdrom,size=709474K
scsi0: local-zfs:vm-1033001-disk-1,size=1000G,ssd=1
scsihw: virtio-scsi-single
smbios1: uuid=f56e328f-f1fe-4b46-a074-67df8448e164
sockets: 1
tpmstate0:local-zfs:vm-1033001-disk-2,size=4M,version=v2.0
vcpus: 32
vga: none
vmgenid: c85a2eac-bdb9-4f93-a4d7-1a96e2607913
-----
 
Unfortunately still a no go, tested it same results.
What kernel version do you use ? Do you have rebar enabled in Bios ? Nvidia drivers version ?

My winver is 26100.4652, same as yours ?
 
Last edited:
butch2k said:
That's what i'm testing right now. Let you know once i'm done.
Click to expand...
Capture d’écran 2025-07-12 201911.jpg
As expected the RTX 3090 is working properly, no issue.

Step 2: Updating to the latest KB...
Capture d’écran 2025-07-12 205039.jpg
Still working fine, unexpectedly

Step 3: Updating to the latest Nvidia Drivers, as stuff like rebar is not enabled with the default MS provided drivers.
Capture d’écran 2025-07-12 210610.jpg
And still working fine.

Step 4: install the QA-Agent and the virtio 271 stuff.


Step 5: check the windows stuff like vm in vm support
 
Last edited:
You must log in or register to reply here.
Top Bottom