[PATCH bpf-next 2/6] fs: wire in BPF_CGROUP_FILE_OPEN hook

[Alexei Starovoitov]

enable cgroup-bpf BPF_CGROUP_FILE_OPEN hook after security_file_open() LSM hook.
Similarly to other cgroup-bpf hooks it's gated by static key 'cgroup_bpf_enabled'
and has zero overhead until bpf prog is attached to that hook.

Signed-off-by: Alexei Starovoitov <ast@xxxxxxxxxx>
---
 fs/open.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/fs/open.c b/fs/open.c
index 0285ce7dbd51..7e1170863f40 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -759,6 +759,10 @@ static int do_dentry_open(struct file *f,
 	if (error)
 		goto cleanup_all;
 
+	error = BPF_CGROUP_RUN_PROG_FILE_FILTER(f);
+	if (error)
+		goto cleanup_all;
+
 	error = break_lease(locks_inode(f), f->f_flags);
 	if (error)
 		goto cleanup_all;
-- 
2.17.1