Unify bpf_sock_tuple constructtion in tcp probe and sockops auth report

[hzxuzhonghu]

What would you like to be added:

In tcp probe, we construct sock tuple with https://github.com/kmesh-net/kmesh/blob/main/bpf/kmesh/probes/tcp_probe.h#L118

But in sokops, we use https://github.com/kmesh-net/kmesh/blob/main/bpf/kmesh/workload/sockops.c#L132

They are much similar except the port byte order, this lead to two different ways of parsing the tuple in usersapce go implemention.

Why is this needed:

To reduce developers burden on understanding the byte order.

[harisudarsan1]

@hzxuzhonghu Seems like in case of tcp_probe we are converting to host byte and in case of sockops we are converting to network byte order. So which order is preferred for unifying the port byte order?

[hzxuzhonghu]

To be clear, i think we can always use host byte order when we need to pass the info to userspace.