[rhcos 4.2][s390x] kola run errors out with " Cluster failed starting machines: didn't find an address" about mantle

[arithx]

cc @barthy1 @alicefr

[Prashanth684]

looks like the localhost representation is switched from 0100007F to 7F000001 because s390x is big endian

[builder@0a00db6c41ec fd]$ cat /proc/net/tcp
  sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode                                                     
   0: 7F000001:AEC3 00000000:0000 0A 00000000:00000000 00:00000000 00000000  1000        0 4188471 1 00000000b89335ae 100 0 0 10 0                   
   1: 7F000001:A669 00000000:0000 0A 00000000:00000000 00:00000000 00000000  1000        0 4187723 1 000000008c1df680 100 0 0 10 0                   

and the check is:

isLocalPat := regexp.MustCompile("0100007F:[[:xdigit:]]{4}")
if !isLocalPat.MatchString(localAddress) || remoteAddress != "00000000:0000" {
	continue
}

[Prashanth684]

I took a stab at fixing this issue in the PR that is linked. I put an endian variable in the base flight object and used that when creating the qemu cluster for getting the port. This way, the endianness can be accessed by any machine through the base flight if needed. We could also store an arch variable there if necessary to distinguish various arches. Testing on s390x, this helped get past the problem of not getting the qemu port. Please let me know comments/improvements.