to operate the same way salloc or sbatch did and assign a task per cpu
by default instead of task per node.

correctly.  Before the errno wasn't being checked correctly

switch/nrt: add test for ability of job to be suspended

Adds most of the infrastructure to the switch plugin for suspend APIs

still need to flesh out the functions for switch/nrt and add support to
new RPC in slurmd.

switch/nrt: and un/pack and use of switch info for job suspend/resume

Major re-write of switch plugin web page for new job preemption functions

switch/nrt: add more logging for job suspend/resume

switch/nrt: document long-term job suspension support on web page

handler correctly.

setup signal handling correctly.

appear this function call happens when PMD calls it.

to node address to be used on the PMD fanout

Conflicts:
	META
	NEWS

This patch adds logic to work around advanced reservations better.

Change node_req field in struct job_resources from 8 to 32 bits so we can
run more than 256 jobs per node.