(almost) silent hang in j1 buildworld on rpi2

[bob prohaska]

An rpi2 running 12.0-CURRENT (GENERIC) #9 r332826 almost but
didn't quite stall during a j1 buildworld. It eventually responded
to the debugger escape, allowing capture of a backtrace. There is
also a record of mvstat -w 5 output during the buildworld session,
along with the log of buildworld output. 

No explicit errors were reported, hopefully wiser eyes might find some
hints as to what happened at
http://www.zefox.net/~fbsd/rpi2/crashes/20180427/

Thanks for reading,

bob prohaska