Virtual Machine crashes when the OS attempts to access AHCI PCI advanced feature.

Products

• VMware vSphere ESXi

Issue/Introduction

• The Virtual Machine crashes multiple times and the vSphere HA continuously restarts the Virtual Machine as a result.
  
  
• A vmx-dump is generated every time the Virtual Machine crashes.
  
  
• The error stack of the dump looks like below.

    #0  0x00000022d190f11d in VMKernel_LiveCoreDump (coreFileNameLen=4196, coreFileName=0x2319574fb0 "") at bora/vmkernel/public/x86/uwvmk.h:781
    #1  Sig_CoreDump () at bora/lib/sig/sigPosix.c:2326
    #2  0x00000022d103e4ea in Panic (fmt=fmt@entry=0x22d1ddd4b4 "Unexpected signal: %d.\n") at bora/vmx/main/vmx.c:1667
    #3  0x00000022d190e054 in SigNoHandler (s=s@entry=11, cp=cp@entry=0x2319576840, info=<optimized out>) at bora/lib/sig/sigPosix.c:1069
    #4  0x00000022d190e9b1 in SigCatcherCommon (context=0x0, ucontext=<optimized out>, info=0x23195767c0, s=11) at bora/lib/sig/sigPosix.c:813
    #5  SigCatcherPosix (s=<optimized out>, info=0x23195767c0, uap=0x2319576840) at bora/lib/sig/sigPosix.c:713
    #6  <signal handler called>
    #7  0x00000022d129b2e3 in PCIAF_Access (advFeat=0x0, devHdr=0x22d2ccbc50, reg=0, flags=<optimized out>, data=0x231647c1b8) at bora/devices/pci/pciaf.c:79
    #8  0x00000022d128d2c9 in PCIAHCIConf (addr=<optimized out>, flags=242, dword=<optimized out>, clientData=0x22d2ccbc50) at bora/devices/ahci/pciAhci.c:123
    #9  0x00000022d11a4a20 in ConfData (func=<optimized out>, func=<optimized out>, data=0x231647c1b8, flags=<optimized out>, addr=2130032)
        at bora/devices/pci/pci.c:1488
    #10 PCI_RPCHandler (rpcBlock=0x231647c1a0) at bora/devices/pci/pci.c:1488
    #11 0x00000022d15f5481 in MonitorLoop_ProcessUserCall (vm=0x22d2a76040, rpcblock=rpcblock@entry=0x231647c1a0, callId=USERCALL_PCI)
        at bora/vmcore/vmx/main/monitor_loop.c:398
    #12 0x00000022d15ff65c in MonitorLoopHostVCPULoop (vcpuid=4) at bora/vmcore/vmx/main/monitorLoopVmkernel.c:72
    #13 0x00000022d15f5625 in MonitorLoopVCPUThreadBase (voidvcpuid=0x4) at bora/vmcore/vmx/main/monitor_loop.c:540
    #14 0x00000022d1707169 in VThreadThreadWrapper (clientData=0x0) at bora/lib/thread/vthreadPosix.c:406
    #15 0x0000002313fd3d3b in start_thread (arg=0x231957e700) at pthread_create.c:308
    #16 0x00000023142d617d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113

Environment

VMware vSphere ESXi
VMware vCenter Server

Cause

This is caused when the AHCI PCI advanced feature is not enabled but your guest operating system is still trying to access it. This causes a NULL pointer de-referencing in PCIAF_Access and hence the kernel crashes. This is a misbehavior from guest OS.

Resolution

Contact the OS vendor for this.

Also ensure that you are on a compatible Virtual Machine Hardware version.