Prevent UDP packet count and operations overflow

[davidBar-On]

• Version of iperf3 (or development branch, such as master or
  3.1-STABLE) to which this pull request applies:
  master

• Issues fixed (if any):

• UDP loss data not calculated correctly?  #1534

• Brief description of code changes (suitable for use as a commit message):

Make all variables related to UDP packets numbers and related calculations int64_t. Although --udp-counters-64bit options allows UDP packet number to be larger than 2^32 bits, several of these related variables were defined as int which is 32 nits in several environments. Note that int64_t is used and not uint64_t to keep sign compatibility with the replaced int (assuming that the 63 bits limit is large enough).

In addition, report_bw_udp/report_sum_bw_udp... definitions in iperf_locale.c were changed to use %PRIu64 instead of %d when printing packets counts. Note that the related reportCSV_bw_udp_format was not changed, as I am not sure how it is used.

[bmah888]

Thanks for the PR, this looks good to me. I've done some testing on it, but not enough to get the counters past 32-bit integers. Regarding your comment about using signed 64-bit integers, I think if we were writing this code from scratch using uint64_t variables would be best, but I agree with your rationale of a least-impact change and just sticking with signed variables for now.