On Thu, 13 Dec 2018 16:17:37 +0100, Daniel Borkmann wrote: > Could we rather extend the test_verifier infrastructure in order to > be able to define data input for bpf_prog_test_run()? I think this > would be very useful for future tests there as well and avoid having > to duplicate or split functionality into test_progs.c instead.
No strong feelings but if it's called test_verifier, and the sample puts no stress on the verifier it feels weird to put it there.. But okay, I will respin.. at some point :)
