PCIPassthru devices enabledModule(s) involved in panic: [vmkernel Version Releasebuild-24755229]cr0=0x80050033 cr2=0x38 cr3=0x81536db000 cr4=0x152660FMS=06/ad/1 uCode=0x1000404frame=0x4539e75dbc60 ip=0x42002d2a01aa err=0x0 rf lags=0x10002rax=0x1 rbx=0x42005280a380 rcx=0x400rdx=0x4b rbp=0x1 rsi=0x0rdi=0x4539eb95f000 r8=0x452184780000 r9=0x3r10=0x42002d2a0154 r11=0x0 r12=0x42005280a390r13=0x4313abc01430 r14=0x0 r15=0x0*PCPU74:2298688/vmm15:rhe19.7 (VMM)PCPU 0: UUVIVSVUVSV IVSVUVSVUVUVUVSVSVSVSUVVSVSVS ISVIVIVIVSVUVIVSVIVIVIVIPCPU 64: VIVISIVIVSVIVIVIVIVSVIVIVIVISVVSVIVIVIVIVIVIVIVIVIVIVIVIVIISIIIUIIIIIIIIIIIIIIUSSSSUUISUUUUSSUUUUIPCPU192: SISIUSUUUISUUUUUISSUIUUUUIIUIIUUUIUIIUIIIUIIIUIIIUIIIIIISIII
Code start: 0x42002d200000 VMK uptime: 6:08:47:02.9000x4539e75dbd20: [0x42002d2a01aa ]VAPICIntrVirtHandler@vmkernel#nover+0x56 stack: 0x00x4539e75dbd50 : [0x42002d376069]IntrCookie_Do Interrupt@vmkernel#nover+0x1a2 stack: 0x00x4539e75dbe00: [0x42002d376578]IntrCookie_Vmkernel Interrupt@vmkernel#nover+0x3d stack: 0xf30x4539e75dbe20: [0x42002d8dd395]IDT_IntrHandler@vmkernel#nover+0x9a stack: 0x00x4539e75dbe40 : [0x42002d8d60ce]gate_entry@vmkernel#nover+Oxaf stack: 0x00x4539e75dbf00:[0x42002d8e91eb]VMMVMKCall_Call@vmkernel#nover+0x277 stack: 0x00x4539e75dbfd0: [0x42002d8e579d]VMKVMM_ArchEnterVMKerne1@vmkernel#nover+0x26 stack: 0x0
passthru devices are being unregistered, while the devices un-registration is happening, if another device is hot removed.
Esx 9.0
This issue occurs due to a race condition while the VM is shutting down, all the passthru devices device are being un-registered.
Broadcom Engineering is aware of this and would be releasing a fix in the upcoming release.
Workaround:
pciPassthru.allowP2P = false