kern/53616: dom0 crashes when dumping domU images with vnd+fss

[triaxx%NetBSD.org@localhost]

>Number:         53616
>Category:       kern
>Synopsis:       dom0 crashes when dumping domU images with vnd+fss
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Sep 18 14:35:01 +0000 2018
>Originator:     Frédéric Fauberteau
>Release:        netbsd-8-0-RELEASE (pkgsrc-2018Q2)
>Organization:
>Environment:
>Description:
I weekly backup my 3 VMs by snapshoting the associated image file:
1) vndconfig over the vm.img
2) mount the vnd device
3) fssconfig over the mountpoint
4) dd the fss device over the network

But after the first VM is backuped, the system panic. I set ddb.on_panic to 1. The system has just crash with the following informations:

db> show panic
Panic string: ufs_inactive: dirty filesystem?

db> bt
breakpoint() at netbsd:breakpoint+0x5
vpanic() at netbsd:vpanic+0x140
snprintf() at netbsd:snprintf
ufs_inactive() at netbsd:ufs_inactive+0x20d
VOP_INACTIVE() at netbsd:VOP_INACTIVE+0x2f
vrelel() at netbsd:vrelel+0x168
do_sys_statat() at netbsd:do_sys_statat+0x94
sys___lstat50() at netbsd:sys___lstat50+0x25
syscall() at netbsd:syscall+0x9c
--- syscall (number 441) ---
6ff9936a17ba:


>How-To-Repeat:
It is not deterministic. I have manually run my backup script and there was no problem.
>Fix:
Probably the filesystem of my image file is not clean, but one of the command I run should raise an error before the kernel panics.