:Well that was quick:
:
:panic: assertion: (msg->ms_flags & (MSGF_DONE|MSGF_QUEUED|MSGF_INTRANSIT)) =
:=3D=3D 0 in lwkt_thread_replyport
Doh! And there's a bug in netmsg_so_notify_abort() right there in
plain sight. MSGF_DONE is not set until the message reaches the
target cpu, so there is a window of opportunity where the conditional
in netmsg_so_notify_abort() will try to reply the message twice.
I'll get that fixed in a sec and we'll see whether it was responsible
for all your problems or not.
-Matt
