An ESXi 8.x host experiences a Purple Diagnostic Screen (PSOD) with the following characteristics:
@BlueScreen: #PF Exception 14 in world <World_ID>:<Process> IP 0x<Address> addr 0x<Address>Port_TryAcqNonexclWithVersion function.Port_TryAcqNonexclWithVersion@vmkernel#nover+0x62
Net_VMMHasPendingWork@vmkernel#nover+0x2bc
CpuSchedVcpuHaltWork@vmkernel#nover+0x6c
VMMVMKCall_Call@vmkernel#nover+0x103
VMKVMM_ArchEnterVMKernel@vmkernel#nover+0x21
vSphere ESXi 8.0
This issue is caused by a faulty physical CPU core miscalculating internal register indices during high-frequency kernel operations.
Technical analysis of the Port_TryAcqNonexclWithVersion function shows that the hypervisor attempts to derive an index from the portsetArray. In affected cases, the processor returns logically incorrect data (e.g., a register value like 0xbe300003 when the expected value is 3), leading to a page fault exception. Because the CPU is electrically functional but returning incorrect computational results, standard hardware sensors often fail to detect the defect Host PSOD with PF Exception 14.
To resolve this issue, the faulty hardware component must be replaced:
/var/run/log/vmkernel.log or /var/run/log/LogEFI.log to determine if the crashes are consistently occurring on the same PCPU.