[llvm] [BPF] Handle nested wrapper structs in BPF map definition traversal (PR #144097)

[via llvm-commits]

eddyz87 wrote:

@vadorovsky,

> On a side note, I know that such nested definition doesn't work on libbpf. I made it work in Aya by modyfing the logic for parsing BTF maps. It works fine and the kernel is able to work with such program.

Noted, I think it's fine, this is a unit test for some backend functionality it doesn't need to be something that fully works end-to-end from Rust/C to kernel.

> I'm happy to port that to libbpf eventually, if there is interest.

libbpf is C oriented and I don't think there is a use-case, but you can always ask on the mailing list.

> I see. Would you be still open to relaxing `print_btf.py` (which I already pushed as a prerequisite commit)

The name is printed anyways, so signal is not lost => removing the warning from `print_btf.py` is fine.

> and me sending a patch relaxing that requirement in the kernel as well?

As I said, there is no functional reason for pointer types to have names in BTF, any strings encoding such names would unnecessarily use space in the .BTF strings section. Hence, I think upstream kernel might object, but who knows.

By the way, I encourage you to make a small test for yourself, there might be some differences between BTF validation rules applied to kernel/module BTF and program BTF.

https://github.com/llvm/llvm-project/pull/144097