Timo Beckers

Hey @rgo3, thanks for jumping in! Also hoping @qmonnet could provide us some perspective, as he wrote most of what's currently in bpftool. **Maintenance burden** Something I'd like to zoom...

Indeed, replacing the whole of bpftool in a few weeks is not realistic. I don't think that was ever on the table. :sweat_smile: The Cilium code linked above has undergone...

After offline discussion: - Cache flusing should be omitted for now, until there is a clearer use case for it. I think we can ignore the cap-drop TOCTOU catch for...

This was added in #673, #321 and others. Closing.

@mmat11 Is there a plan to move forward with this? If you don't have the resources to investigate this in the near future, maybe best to add a comment to...

I think we should look at something like [bingo](https://www.bwplotka.dev/2020/bingo/) instead of `internal/tooldeps`. (the package approach is mentioned specifically) Could we drop that change from the PR and address it separately?

Thanks for the writeup, I think I have a better understanding of the full picture. I'm still not sure why my suggestion without an explicit Merge() is bad. LoadSplitSpecFromReader should...

Here are a few extra constraints I had in mind that I don't find mentioned explicitly, although these might be described by your pseudocode above: - All `*Type(s)By*` methods should...

@lmb I think you missed this text in the OP: `5.10.112-108.499.amzn2.x86_64` My host kernel doesn't have any duplicates FUNCs that I can immediately spot, but a bunch of syscalls point...

We could always relicense bpf2go under the GPL. Since it's a CLI tool and not part of the build artifact, it should not be subject to the same linking restrictions...