Windows VMs might fail with a BSOD screen with 0x5c signature due to platform delays

Products

• VMware vSphere ESXi

Issue/Introduction

The Windows VM fails with BSOD.

In vmware.log you may see records similar to: 

vmware.log
YYYY-MM-DDTHH:MM:SS Wa(03) vcpu-1 - WinBSOD: Synthetic MSR[0x40000100] 0x5c
YYYY-MM-DDTHH:MM:SS Wa(03)+ vcpu-1 -
YYYY-MM-DDTHH:MM:SS Wa(03) vcpu-1 - WinBSOD: Synthetic MSR[0x40000101] 0x115
YYYY-MM-DDTHH:MM:SS Wa(03)+ vcpu-1 -
YYYY-MM-DDTHH:MM:SS Wa(03) vcpu-1 - WinBSOD: Synthetic MSR[0x40000102] 0xfffff7cb40015d10
YYYY-MM-DDTHH:MM:SS Wa(03)+ vcpu-1 -
YYYY-MM-DDTHH:MM:SS Wa(03) vcpu-1 - WinBSOD: Synthetic MSR[0x40000103] 0x14555
YYYY-MM-DDTHH:MM:SS Wa(03)+ vcpu-1 -
YYYY-MM-DDTHH:MM:SS Wa(03) vcpu-1 - WinBSOD: Synthetic MSR[0x40000104] 0xffffffffc0000001

Environment

ESXi 7.0
ESXi 8.0 

Cause

Checks done by HPET timers in Windows might fail due to platform delays and cause the issue.

Resolution

Troubleshoot the underlying storage I/O latency issue.

Or, apply the mitigation patch.
Functionality has been improved to reduce the occurrence of BSODs.
This change is implemented in ESXi 7.0 U3o, ESXi 8.0, and later versions.

Additional Information

VMware ESXi 7.0 Update 3o Release Notes
https://techdocs.broadcom.com/us/en/vmware-cis/vsphere/vsphere/7-0/release-notes/esxi-update-and-patch-release-notes/vsphere-esxi-70u3o-release-notes.html

-----
PR 3209853: Windows VMs might fail with a blue diagnostic screen with 0x5c signature due to platform delays
Windows VMs might fail with a blue diagnostic screen with the following signature due to a combination of factors involving the Windows timer code and platform delays, such as storage or network delays:
HAL_INITIALIZATION_FAILED (0x5c) (This indicates that the HAL initialization failed.)
Arguments:
Arg1: 0000000000000115, HAL_TIMER_INITIALIZATION_FAILURE
Arg2: fffff7ea800153e0, Timer address
Arg3: 00000000000014da, Delta in QPC units (delta to program the timer with)
Arg4: ffffffffc0000001, invalid status code (STATUS_UNSUCCESSFUL)
Checks done by HPET timers in Windows might fail due to such platform delays and cause the issue.
-----