ESX Host Crashes with PSOD: PF Exception 14 HELPER_VMXNE
Article ID: 420576
Updated On:
Products
Issue/Introduction
An ESX host fails with a purple diagnostic screen (PSOD) showing a Page Fault exception.
The backtrace on the screen includes references to VVTDLogDMARFault, VVTDRemapDMA, similar backtrace to the one shown below may be recorded on the impacted host:
World: 3400: PRDA 0x420040c00000 ss 0x0 ds 0x750 es 0x750 fs 0x0 gs 0x0
World: 3402: TR 0x768 GDT 0xfffffffffca02888 (0xffff) IDT 0xfffffffffc408000 (0xffff)
World: 3404: CR0 0x80050033 CR3 0xd465e85000 CR4 0x152660
Backtrace for current CPU #3, worldID=209####, fp=0xfcfe5000
0x453bc789bc08:[0x420019311797]VVTDLogDMARFault@vmkernel#nover+0x7b stack: 0x453bc789bd80, 0x269c737e740208, 0x1a916e439a00, 0x453bc789bd00, 0x100000001
0x453bc789bc10:[0x42001931222a]VVTDRemapDMA@vmkernel#nover+0x17b stack: 0x269c737e740208, 0x1a916e439a00, 0x453bc789bd00, 0x100000001, 0x452300000005
0x453bc789bd10:[0x4200193fbd46]PhysMemRemapDMA@vmkernel#nover+0x13 stack: 0x453c2069f000, 0x4200193fca1c, 0x453bc789bd78, 0x4200193fcd53, 0x5b2
0x453bc789bd30:[0x4200193fca1b]PhysMemRemapSinglePPN@vmkernel#nover+0xe4 stack: 0x5b2, 0x117c789f480, 0x1, 0x11700000000, 0x1
0x453bc789bdd0:[0x4200193fcebb]PhysMem_GetMPNList@vmkernel#nover+0x80 stack: 0x453c2069f000, 0x4200193fcdec, 0x20200000000, 0x430a11c0ff10, 0x453c2069f000
0x453bc789be40:[0x4200194d4973]Vmxnet3VMKDevDoPageIn@vmkernel#nover+0x2a4 stack: 0x0, 0xd193765e5, 0x4309ff816780, 0x4308ef1e0c40, 0x2
0x453bc789bf60:[0x4200193704db]HelperQueueFunc@vmkernel#nover+0x1b0 stack: 0x430e1a601238, 0x453bc789f000, 0x453bc785f000, 0x453bc789f100, 0x0
0x453bc789bfe0:[0x420019913ec7]CpuSched_StartWorld@vmkernel#nover+0xc4 stack: 0x0, 0x420019357dcc, 0x0, 0x0, 0x0
0x453bc789c000:[0x420019357dcb]Debug_IsInitialized@vmkernel#nover+0x10 stack: 0x0, 0x0, 0x0, 0x0, 0x0
ESC[45mESC[33;1mVMware ESXi 9.0.0 [Releasebuild-24755229 x86_64]ESC[0m
#PF Exception 14 in world 209####:HELPER_VMXNE IP 0x420019311797 addr 0x38
PTEs:0x0;Environment
VMware vSphere ESX 9.0.0
Cause
This issue occurs where the helper world incorrectly initiated a memory action that triggered a DMA fault during remapping. This critical error occurred because a non-VMM/VMX world was passed to the Action_Post function during the DMA remapping process
The PSOD backtrace indicates the HELPER_ world, not the dedicated VMX world was active when the page fault occurred. This behavior aligns with a scenario where a memory function was erroneously called by a helper world, resulting in the incorrect memory mapping context and the fatal PSOD.
Resolution
This issue is resolved in VMware ESX 9.0.1.0 and future releases. To fix the issue, upgrade the ESX host to version 9.0.1.0
Additional Information
If the issue persists after upgrading, please contact Broadcom Support and reference this article. Be sure to collect a PSOD screenshot and a Log Bundle from the affected ESX host beforehand.
Step to generate log file is mentioned here: Collecting diagnostic information for ESXi hosts
Feedback
