On 8/11/20 3:55 PM, Ramamurthy, Harshitha wrote: > is a pre-cursor to potentially calling the in-kernel flow dissector from > a helper function. >
That is going to be a challenge to do in a way that does not impact the kernel's ability to change as needed, and I suspect it will be much slower than doing in ebpf. Modular ebpf code that can be adapted to use cases is going to be more appropriate for XDP for example.
