Some questions about attach_lsm

[wyEumenides]

I would like to know how to use the attach_lsm and detach_lsm methods, what should be passed as the "fn_name" parameter, and whether it is possible to pass the kernel call as event and the custom trace function name as fn_name, just like attach_kprobe.

[wyEumenides]

I wrote the following code：

from bpfcc import BPF

BLOCKED_INODE = 2930240  // test 

bpf_program = (
    """
#include <uapi/linux/ptrace.h>
#include <linux/fs.h>
#include <linux/dcache.h>

#define BLOCKED_INODE %d



LSM_PROBE(file_open, struct file *file) {
    struct inode *inode = file->f_inode;
    if (inode->i_ino == BLOCKED_INODE) {
        return -1; 
    }
    return 0;

}

"""
    % BLOCKED_INODE
)

if not BPF.support_lsm():
    print("do not spport LSM!")
b = BPF(text=bpf_program)

It's in line with expectations, but how to change to use attach_lsm?