Re: [PATCH bpf-next 01/11] bpf, perf: fix bpftool compilation with !CONFIG_PERF_EVENTS

[Song Liu]

On Fri, Apr 15, 2022 at 4:07 PM Song Liu <song@xxxxxxxxxx> wrote:
>
> On Thu, Apr 14, 2022 at 3:45 PM Alexander Lobakin <alobakin@xxxxx> wrote:
> >
> > When CONFIG_PERF_EVENTS is not set, struct perf_event remains empty.
> > However, the structure is being used by bpftool indirectly via BTF.
> > This leads to:
> >
> > skeleton/pid_iter.bpf.c:49:30: error: no member named 'bpf_cookie' in 'struct perf_event'
> >         return BPF_CORE_READ(event, bpf_cookie);
> >                ~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~
> >
> > ...
> >
> > skeleton/pid_iter.bpf.c:49:9: error: returning 'void' from a function with incompatible result type '__u64' (aka 'unsigned long long')
> >         return BPF_CORE_READ(event, bpf_cookie);
> >                ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >
> > Tools and samples can't use any CONFIG_ definitions, so the fields
> > used there should always be present.
> > Move CONFIG_BPF_SYSCALL block out of the CONFIG_PERF_EVENTS block
> > to make it available unconditionally.
> >
> > Fixes: cbdaf71f7e65 ("bpftool: Add bpf_cookie to link output")
> > Signed-off-by: Alexander Lobakin <alobakin@xxxxx>
>
> While I can't think of a real failure with this approach, it does feel
> weird to me. Can we fix this with bpf_core_field_exists()?

Hmm.. the error happens at compile time, so I guess it is not very easy.

Andrii,
Do you have some recommendation on this?

Song