The bug arises with the following NSK test :-

nsk/jdwp/Event/VM_DEATH/vmdeath002

The test requests a VM_DEATH event with SuspendPolicy=ALL. When the
VM_DEATH event is posted all threads are suspended. Then the tests
sents a Resume command which resumes all threads. Resuming all threads
creates a race condition whereby the connection may be closed before
the Resume reply is enqueued to the transport. Here's the tail of
the log file for when the failure arises :-

#Packet header (11 bytes):
# 0000 (length) : 0x0000001a (26)
# 0004 (id) : 0x00000001 (1)
# 0008 (flags) : 0x00 (0)
# 0009 (cmd set): 0x40 (64)
# 000a (command): 0x64 (100)
# 000b (policy): 0x02 (2)
# 000c (events): 0x00000002 (2)
# 0010 (kind): 0x63 (99)
#Entire packet (26 bytes):
# 0000: 00 00 00 1a 00 00 00 01 00 40 64 02 00 00 00 02 .........@d.....
# 0010: 63 00 00 00 02 63 00 00 00 00 c....c....
#Packet end
#Checking header of event packet
# ... packet header is OK
#Parsing event packet:
# suspendPolicy: 2
# events: 2
# event #0:
# eventKind: 99
# requestID: 2
# event #1:
# eventKind: 99
# requestID: 0
# ... event packet parsed
# ... event received
#
#Resuming debuggee
## ERROR: TEST FAILED: Cought IOException while receiving reply packet:
## ERROR: java.io.IOException: JDWP socket connection closed by remote host