scheduler: Re-compute time-average nr_running on read

Re-compute time-average nr_running when it is read. This would
prevent reading stalled average value if there were no run-queue
changes for a long time. New average value is returned to the reader,
but not stored to avoid concurrent writes. Light-weight sequential
counter synchronization is used to assure data consistency for
re-computing average.

Change-Id: I08de26087414d6663633c723ddfcbf1099a01ce2
Signed-off-by: Alex Frid <afrid@nvidia.com>
Reviewed-on: http://git-master/r/108072


Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Ryan Wong <ryanw@nvidia.com>
Tested-by: Ryan Wong <ryanw@nvidia.com>

Original commit: https://github.com/halaszk/Perseus-halaszk-universal5433/commit/adf51e6f6fa7478ef27854c5fc67aaaaddb5406e