PSOD at function fnic_internal_io_timeout with fnic driver version 5.0.0.15

Products

• VMware vSphere ESXi

Issue/Introduction

ESXi Server might fail with a PSOD having backtrace similar to the following

@BlueScreen: #PF Exception 14 in world 2099206:tq:tq-iport IP 0x42001eba2b6b addr 0x1350 PTEs:0x0;
Code start: 0x42001dc00000 VMK uptime: 113:16:05:21.489
0x453a3b01bf30:[0x42001eba2b6b]fnic_internal_io_timeout@(nfnic)#<None>+0x27 stack: 0x36
0x453a3b01bf50:[0x42001dc2c4cb]VmkTimerQueueWorldFunc@vmkernel#nover+0x258 stack: 0x1b9d43128703a5fd
0x453a3b01bfe0:[0x42001dfb4d55]CpuSched_StartWorld@vmkernel#nover+0x86 stack: 0x0
0x453a3b01c000:[0x42001dcc4ddf]Debug_IsInitialized@vmkernel#nover+0xc stack: 0x0

VMkernel logs show an event suggesting NULL command during io timeout

fnic_internal_io_timeout: cmd: NULL, io_req: 0x45ca48403300

Environment

ESXi Server 7.x with fnic driver version 5.0.0.15

Cause

PSOD occurs due to a possible race condition between removal of a target port with outstanding IO or Report LUN command. For more details please see CSCwa77324

Resolution

This issue is reported to be fixed with fnic driver version 5.0.0.45. Please update ESXi Servers. 

Note: it is advised to update the firmware as recommended by Broadcom Compatibility Guide along with driver. Though the fix is available in version 5.0.0.45, consider updating to the newer if available.

Additional Information

Please follow KB 366755 to find the IO driver downloads on Broadcom Support Portal.