Re: [PATCH bpf-next v5 4/8] bpf: Introduce cgroup iter

[Tejun Heo]

On Thu, Jul 28, 2022 at 10:20:46AM -0700, Hao Luo wrote:
...
> is a good feature of this convention. My thoughts: It seems that ID
> may be better, for two reasons. First, because ID is stateless, the
> userspace doesn't have to remember closing the FD. Second, using
> different identifications in two directions (userspace specifies
> cgroup using FD, while kernel reports cgroup using ID) introduces a
> little complexity when connecting them together.

Yeah, you can pass the IDs around different processes, print and log them in
meaningful ways and so on because they're actual IDs, so my preference is
towards using them for anything new.

Thanks.

-- 
tejun