Re: [PATCH v2 bpf-next 4/5] selftests/bpf: Add tests for bpf_cgroup_read_xattr

[Eduard Zingerman <eddyz87@gmail.com>]

On Fri, 2025-06-20 at 11:11 -0700, Alexei Starovoitov wrote:
> On Thu, Jun 19, 2025 at 3:02 PM Song Liu <song@kernel.org> wrote:
> > +       bpf_dynptr_from_mem(xattr_value, sizeof(xattr_value), 0, &value_ptr);
> 
> https://github.com/kernel-patches/bpf/actions/runs/15767046528/job/44445539248
> 
> progs/cgroup_read_xattr.c:19:9: error: ‘bpf_dynptr_from_mem’ is static
> but used in inline function ‘read_xattr’ which is not static [-Werror]
> 19 | bpf_dynptr_from_mem(value, sizeof(value), 0, &value_ptr);
> > ^~~~~~~~~~~~~~~~~~~
> 
> 
> Jose,
> 
> Could you please help us understand this gcc-bpf error ?
> What does it mean?

Not Jose, but was curious.
Some googling lead to the following C99 wording [1]:

  > An inline definition of a function with external linkage shall not
  > contain a definition of a modifiable object with static storage
  > duration, and shall not contain a reference to an identifier with
  > internal linkage

[1] https://www.open-std.org/jtc1/sc22/wg14/www/docs/n1256.pdf
    6.7.4 Function specifiers, paragraph 3

The helper is defined as `static`:

  static long (* const bpf_dynptr_from_mem)(...) = (void *) 197;

While `read_xattr` has external linkage:

  __always_inline void read_xattr(struct cgroup *cgroup)
  {
	...
	bpf_dynptr_from_mem(value, sizeof(value), 0, &value_ptr);
	...
  }

I think that declaring `read_xattr` as `static` should help with gcc.