bpf: Refactor bpf_link update handling
Make bpf_link update support more generic by making it into another
bpf_link_ops methods. This allows generic syscall handling code to be agnostic
to various conditionally compiled features (e.g., the case of
CONFIG_CGROUP_BPF). This also allows to keep link type-specific code to remain
static within respective code base. Refactor existing bpf_cgroup_link code and
take advantage of this.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20200429001614.1544-2-andriin@fb.com
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 7626b802..f5358e1 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -3645,13 +3645,10 @@ static int link_update(union bpf_attr *attr)
goto out_put_progs;
}
-#ifdef CONFIG_CGROUP_BPF
- if (link->ops == &bpf_cgroup_link_lops) {
- ret = cgroup_bpf_replace(link, old_prog, new_prog);
- goto out_put_progs;
- }
-#endif
- ret = -EINVAL;
+ if (link->ops->update_prog)
+ ret = link->ops->update_prog(link, new_prog, old_prog);
+ else
+ ret = EINVAL;
out_put_progs:
if (old_prog)
