Also eliminate some possible fatal errors on reconfig or internal error

If two arrays differ in size, make them the same size to
perform bit_or/and/and_not operations on the full array

create src/common/tres_frequency.[ch] module based upon cpu_frequency.[ch]
modify launch RPCs to pass the value from slurmctld to slurmstepd
validate --gpu-freq value from salloc, sbatch, and srun

Add v18.08 versions of un/pack functions for REQUEST_LAUNCH_TASKS
and REQUEST_BATCH_JOB_LAUNCH RPCs

Use tres_per_job/node/socket/task instead

Completely remove "gres" field from step record in slurmctld and
step info message. Use "tres_per_node", "tres_per_socket", etc.

This bug was introduced in commit 4a9bffe1
Test12.7 was resulting in the logging of error messages of this sort:
sacct: error: slurmdb_ave_tres_usage: couldn't make tres_list from ''
This was due to the tres_usage_in_ave and tres_usage_out_ave fields
being empty ('\0') if the job's cpu count is zero, which makes
calculation of averages impossible.
bug 2782

Bug 5174.

Since having 'nocreate' would override the following option:
create 640 slurm root

Bug 5174.

Add node_features plugin function "node_features_p_reboot_weight()" to
   return the node weight to be used for a compute node that requires reboot
   for use (e.g. to change the NUMA mode of a KNL node).
Add NodeRebootWeight parameter to knl.conf configuration file.

If ReturnToService=2 is configured, the test could generate an error
changing node state to resume after setting it to down. The reason
is if the node communicates with slurmctld, then its state will
automatically be changed from down to idle and resuming an idle
node triggers an error.

Bug 5168.

Previously the default paths continued to be tested even when new ones
were requested. This had as a consequence that if any of the new paths
was the same as any of the default ones (i.e. /usr or /usr/local), the
configure script was incorrectly erroring out specifying that a version
of PMIx was already found in a previous path.

Bug 5168.

Variable initialization plus cosmetic work

Rename gres_per_job for step to gres_per_step
Remove job gres gres_name_type_id field
Build step gres data structure

Prevent run-away jobs

This is not currently supported and no date for support has been set.

If burst_buffer.conf has GetSysState configured to a non-standard
location, but GetSysStatus is not configured that is likely indicative
of a bad configuration rather than a Slurm failure.

Gracefully fail if salloc does not get job allocation

Introduced in bf4cb0b1.

Support for AIX was removed before 17.02.

First issue was identified on multi partition requests. job_limits_check()
was overriding the original memory requests, so the next partition
Slurm validating limits against was not using the original values. The
solution consists in adding three members to job_details struct to
preserve the original requests. This issue is reported in bug 4895.

Second issue was memory enforcement behavior being different depending on
job the request issued against a reservation or not.

Third issue had to do with the automatic adjustments Slurm did underneath
when the memory request exceeded the limit. These adjustments included
increasing pn_min_cpus (even incorrectly beyond the number of cpus
available on the nodes) or different tricks increasing cpus_per_task and
decreasing mem_per_cpu.

Fourth issue was identified when requesting the special case of 0 memory,
which was handled inside the select plugin after the partition validations
and thus that could be used to incorrectly bypass the limits.

Issues 2-4 were identified in bug 4976.

Patch also includes an entire refactor on how and when job memory is
is both set to default values (if not requested initially) and how and
when limits are validated.

Co-authored-by: Dominik Bartkiewicz <bart@schedmd.com>

The slurmctld doesn't need to send the fini message, and actually if
it does things get messed up as the slurmdbd will close the database
connection prematurely.  Up till now we would print an error on the
slurmctld saying we couldn't send the FINI.

partition is removed then the slurmdbd comes up and we go refresh
the tres pointers and try to deference the part_ptr.

Related to commit de7eac9a.

Bug 5136

and move the agent into the accounting_storage/slurmdbd plugin.

This should be cleaner going forward and will be easier to maintain.