task/affinity plugin: Honor a job's --ntasks-per-socket and
    --ntasks-per-core options in task binding.
bug 3118

Bug 3007

e.g.
brian@lappy:~/slurm/master/lappy$ scontrol show res
ReservationName=root_167 StartTime=2016-10-12T10:24:35 EndTime=2016-10-12T10:54:35 Duration=00:30:00
   Nodes=lappy[1-10] NodeCnt=10 CoreCnt=40 Features=(null) PartitionName=(null) Flags=OVERLAP,SPEC_NODES,ALL_NODES
   TRES=cpu=80
   Users=root Accounts=(null) Licenses=(null) State=ACTIVE BurstBuffer=(null) Watts=n/a

ReservationName=root_168 StartTime=2016-10-12T10:24:37 EndTime=2016-10-12T11:24:37 Duration=01:00:00
   Nodes=lappy[1-10] NodeCnt=10 CoreCnt=40 Features=(null) PartitionName=(null) Flags=OVERLAP,SPEC_NODES,ALL_NODES
   TRES=cpu=80
   Users=root Accounts=(null) Licenses=(null) State=ACTIVE BurstBuffer=(null) Watts=n/a

brian@lappy:~/slurm/master/lappy$ sbatch --wrap="sleep 60" -t1 --test-only
sbatch: Job 14435 to start at 2016-10-12T11:24:37 using 1 processors on lappy1

The for loop wasn't resetting the job_start_time to the reservation's
end_time. This was taken out in 06dc5fa7.

Changed in df70b651

Preserve non-KNL node features when updating the KNL node features for a
    multi-node job in which the non-KNL node features vary by node.

node_features/knl_cray plugin: If the reconfiguration of nodes for an
    interactive job fails, kill the job (it can't be requeued like a batch job).

Execute "capmc node_status" at more frequent intervals to handle
  nodes getting added or removed from the system using Cray tools
  (i.e. try to keep Slurm and Cray software better synchronized).

node_features/knl_cray plugin: Add separate thread to interact with capmc
    in response to unexpected node reboots.
bug 3153

bug 3091

bug 3155

Prevent possible divide by zero in select/cons_res if a node's board count
    is higher than it's socket count.
bug 3155

If a node's socket or core count are changed at registration time (e.g. a
    KNL node's NUMA mode is changed), change it's board count to match.
bug 3155

Cray: The slurmd can manipulate the socket/core/thread values reported based
    upon the configuration. The logic failed to consider select/cray with
    SelectTypeParameters=other_cons_res as equivalent to select/cons_res.
bug 3155

abs() should not be used on long long variables as it would
truncate if strictly confirming to C99. Use llabs() instead.

Fix to commit 2aefc66b.

Inadvertently broken by commit 3251406c.

These tests fail with a configuration of CR_ONE_TASK_PER_CORE.
  The tests are being disabled now for such a configuration, but
  could be modified to run in such an environment in the future.

No functional change.

Expand test to run on various allocation sizes

Replaced by test1.114

Previous logic would result in sub-optimal job allocations and
  under some conditions invalid memory references resulting in
  the slurmctld daemon crashing or aborting.
bug 2732

Correct SchedulerParameters=bf_busy_nodes logic with respect to the job's
    minimum node count. Previous logic would not decremement counter in
    two locations and reject valid job request for not reaching minimum
    node count.

always just AccountingPolicy.

daf91a80