On Tue, Mar 19, 2024 at 9:24 AM Eduard Zingerman <eddyz87@xxxxxxxxx> wrote:
>
> On Tue, 2024-03-19 at 09:16 -0700, Andrii Nakryiko wrote:
> [...]
>
> > So I considered `bpf_raw_tracepoint_open_opts()`, but it felt so
> > verbose that I decided to shorten it to `bpf_raw_tp_open()`, given we
> > do have SEC("raw_tp") and that's very recognizable contraction.
> >
> > Having said that, I'm not opposed to going with
> > bpf_raw_tracepoint_open_opts(), as I don't think many users will ever
> > need to call it directly, so verboseness doesn't matter all that much.
> >
> > Let me know if you still prefer the `bpf_raw_tracepoint_open_opts()` variant.
>
> I'd prefer the longer variant if you don't mind.
> I'm a relative beginner to libbpf internals and seeing bpf_raw_tp_open
> instead bpf_raw_tracepoint_open_opts kinda broke my intuitive
> expectations based on other APIs, so we can use it as a test :)
Ok, sounds good. You can see I didn't deviate in high-level API
(bpf_program__attach_raw_tracepoint_opts()), even though this
verboseness is breaking my heart, because that's way more
user-visible. But I guess it's not worth it, I'll post v3 with
bpf_raw_tracepoint_open_opts() then.
