port-arm/51063: diskless machine reboots early in test runs

[martin%NetBSD.org@localhost]

>Number:         51063
>Category:       port-arm
>Synopsis:       diskless machine reboots early in test runs
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    port-arm-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Apr 13 07:00:00 +0000 2016
>Originator:     Martin Husemann
>Release:        NetBSD 7.99.27
>Organization:
The NetBSD Foundation, Inc.
>Environment:
System: NetBSD unpluged.duskware.de 7.99.27 NetBSD 7.99.27 (UNPLUGED) #0: Tue Apr 12 20:47:17 CEST 2016 martin%martins.aprisoft.de@localhost:/ssd/src/sys/arch/evbarm/compile/UNPLUGED evbarm
Architecture: earm
Machine: evbarm
>Description:

Running the regular tests on evbarm fails this week (last weeks run was fine).
The machine just reboots, no ddb, no panic message.
It happens within the first two minutes of the test run, but not at fixed
tests.

I have

Available watchdog timers:
        mvsoctmr0, 21 second period [armed, user tickle, pid 22]

but the failure does not happen after a 21 second pause of output. Root
is NFS mounted on 

mvgbe0: flags=0x8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        capabilities=3700<IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,UDP4CSUM_Rx>
        capabilities=3700<UDP4CSUM_Tx>
        enabled=3700<IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,UDP4CSUM_Rx>
        enabled=3700<UDP4CSUM_Tx>
        ec_capabilities=5<VLAN_MTU,JUMBO_MTU>
        ec_enabled=0

and /tmp is on tmpfs.

>How-To-Repeat:

On a diskless machine do:

	cd /usr/tests
	atf-run | atf-report

>Fix:
n/a