port-evbarm/56944: ZFS heavy usage on NetBSD running in Mac M1 VM results in kernel thread running away and filesystem hang

[pjledge%me.com@localhost]

>Number:         56944
>Category:       port-evbarm
>Synopsis:       ZFS heavy usage on NetBSD running in Mac M1 VM results in kernel thread running away and filesystem hang
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-evbarm-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Jul 27 12:50:00 +0000 2022
>Originator:     Paul J Lavoie
>Release:        NetBSD 9.99.97
>Organization:
>Environment:
System: NetBSD plucky 9.99.97 NetBSD 9.99.97 (GENERIC64) #0: Wed Jun 8 10:17:55
 EDT 2022 sysbuild@plucky:/home/sysbuild/evbarm/obj/home/sysbuild/src/sys/arch/
evbarm/compile/GENERIC64 evbarm
Architecture: aarch64
Machine: evbarm

>Description:
When trying to copy a large amount of data (300GB) to a fresh ZFS file system, the copy will hang after a few minutes. Looking at the output of top, a [system] process can be seen using a full CPU constantly, only going away on a system restart.

The instance of NetBSD is running on virtualization software under a Apple MacBook Pro M1 Max chip. Both UVM and VMWare Fusion result in the same issue.

This works fine using ZFS on the amd64 platform (same NetBSD version)
>How-To-Repeat:
1) Create zpool on NetBSD image running as a virtual instance under a MacOS M1 virtualization tool
2) Start a copy (tar cfj - | ssh <sys> 'cd /; tar xvfj -), where <sys> is the Mac NetBSD instance with a large amount of data which will require a significant amount of time to transfer
3) Wait a few minutes and check for stalling
>Fix:
unknown