Viktor Malík

bpftrace 0.16.0 is more than 2 years old. Could you please try to reproduce this with the latest version? For instance, you can use the static AppImage for version 0.21.2:...

Just for the reference, I played a bit with bool types in #3436. This is a different situation as we're talking about *internal* types here while my work was on...

> I noticed that there is an API named `bpf_program__attach_usdt` in libbpf. Why don't we migrate the `bpf_attach_uprobe` to the `bpf_program__attach_usdt`? [@danobi](https://github.com/danobi) [@jordalgo](https://github.com/jordalgo) IIRC, there are some specifics on how...

Yeah, this is unfortunate. I think that until we have proper C interop (or at least a way to call BPF linker from inside bpftrace), we have no other choice...

> Maybe we can utilize probe attributes ([RFC](https://github.com/bpftrace/bpftrace/issues/3360)) for such a thing e.g. > > ``` > [required_arch = "x86_64", min_kernel_version = "6.12.0"] > kprobe:x86_only { ...} > ``` Without...

@amscanne I think one argument for using the attributes would be the case when you want to control the *attachment* of the probe. But for that, I think we can...

> One problem is kfunc probes require [libbpf/libbpf@dd589c3](https://github.com/libbpf/libbpf/commit/dd589c3b31c13164bdc61ed174fbae6fe76c8308). Unfortunately, it's not released, yet, so we must consider two things: > > 1. Do we want to have master depending on...

> > Once the patch is released (should be in the next libbpf version 1.5.0), do we want bpftrace to require the latest libbpf (which many distros do not package)?...

> Just curious, it seems like multiple subprogs now work with this, which is AWESOME! Did you want to revert [this commit](https://github.com/bpftrace/bpftrace/pull/3233/files) in this PR and update the tests? Absolutely!...

All of the above comments should be addressed now. The CI is also passing at last.