[RFC PATCH bpf-next v5 07/12] libbpf: Support passing user-defined flags for struct_ops
From: Hui Zhu <hidden>
Date: 2026-01-27 09:46:43
Also in:
bpf, cgroups, linux-kselftest, linux-mm, lkml
Subsystem:
bpf [general] (safe dynamic programs and tools), bpf [library] (libbpf), the rest · Maintainers:
Alexei Starovoitov, Daniel Borkmann, Andrii Nakryiko, Eduard Zingerman, Kumar Kartikeya Dwivedi, Linus Torvalds
From: Hui Zhu <redacted> Building on the previous change that added flags to the kernel's link creation path, this patch exposes this functionality through libbpf. The `bpf_struct_ops_opts` struct is extended with a `flags` member, which is then passed to the `bpf_link_create` syscall within `bpf_map__attach_struct_ops_opts`. This enables userspace applications to pass flags, such as `BPF_F_ALLOW_OVERRIDE`, when attaching struct_ops to cgroups, providing more control over the attachment behavior in nested hierarchies. Signed-off-by: Geliang Tang <geliang@kernel.org> Signed-off-by: Hui Zhu <redacted> --- tools/lib/bpf/libbpf.c | 1 + 1 file changed, 1 insertion(+)
diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 70a00da54ff5..06c936bad211 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c@@ -13511,6 +13511,7 @@ struct bpf_link *bpf_map__attach_struct_ops_opts(const struct bpf_map *map, } link_opts.cgroup.relative_fd = OPTS_GET(opts, relative_fd, 0); + link_opts.flags = OPTS_GET(opts, flags, 0); fd = bpf_link_create(map->fd, 0, BPF_STRUCT_OPS, &link_opts); if (fd < 0) {
--
2.43.0