Convert capable.py to CO-RE

[eiffel-fl]

Hi.

The tools to get capabilities used does not exist to CO-RE.
I will try to translate it (I cannot self-assigned this issue and I just open it to inform people I will try it so we do not end with several people trying to convert it at the same time).

Best regards.

[eiffel-fl]

I thought I would be able to start today but it will delayed until tomorrow or beginning of next week I think.

[KentaTada]

@eiffel-fl
Thank you for the review in #3890
We are also trying to convert python-based tools to libbpf tools.
Especially, we want to use libbpf-tools on aarm64.

• Implement the cgroup-based filter for the libbpf tools like below.
  Filter cgroup2 process to support tracing container #3741
• Support the custom path of BTF for capable and other tools for the old environment without CONFIG_DEBUG_INFO_BTF.
  • I just started to investigate it. So, the plan is subject to change.
    https://github.com/KentaTada/minimum-btf-bootstrap

and so on.

Currently, we are summarizing the comparison table between python-based tools to libbpf tools.
When we prepare for that, we will share it in the another issue.

Thanks!

[eiffel-fl]

Hi.

You are welcome and sorry for it, I planned to convert this gadget but I lacked time when I opened the issue.
Anyway, the important thing is that this tool is converted.

I think I will also to work on arm64, for the moment I was using bcc tools for this architecture and I had no problem.
But when I quickly tested CO-RE ones I got some troubles (I did not investigate about it).
I will let you know when I work on it but I cannot say you when it will occur.

Best regards.

[mauriciovasquezbernal]

@KentaTada perhaps you want to take a look at #3889.

[KentaTada]

@mauriciovasquezbernal
Thanks.
As feasibility study, we are trying to deploy the container including libbpf-based tools with the reduced BTF files on the environment which does not support BTF to reduce the runtime overhead.
I'll take a look at #3889 at first.

[eiffel-fl]

Closing as done in #3890!