Changed fqrate from uint to uint64 to allow pacing above 32G

[marcosfsch]

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

• Issues fixed (if any):
  support for 64bit fq-rate #1719

• Brief description of code changes (suitable for use as a commit message):
  Starting with Linux 5.x, fq pacing is now supported for speeds > 32G.
  see: https://patchwork.ozlabs.org/project/netdev/patch/[email protected]/
  and https://kernelnewbies.org/Linux_5.1
  To fix this we only need to change fqrate variable from uint to uint64 on both perf_tcp.c and iperf_udp.c

[swlars]

Hi, we like this change, but we're concerned about compatibility on supported devices with a Linux kernel version that's less than 5, like Rocky8. Do you know of a way to detect at run or compile time if 64-bit values are accepted?

[marcosfsch]

I don't know a way to test the values at runtime, but I just compiled and runned iperf3 with this patch and it silently trims the value to 32-bits on rocky8, and mantains the behaviour without tha patch. i.e. if you set above 32G it overloads and start from zero again.

[swlars]

We also tested it and it seems to work as expected. We have not tested on any big-endian systems, so we'll probably make a note of that somewhere. Thanks for the merge request!