bpf: lsm: Implement attach, detach and execution
authorKP Singh <kpsingh@google.com>
Sun, 29 Mar 2020 00:43:52 +0000 (01:43 +0100)
committerDaniel Borkmann <daniel@iogearbox.net>
Sun, 29 Mar 2020 23:34:00 +0000 (01:34 +0200)
commit9e4e01dfd3254c7f04f24b7c6b29596bc12332f3
tree5c98d39252c414dd308b372c07c0bd6bec4944ac
parent9d3fdea789c8fab51381c2d609932fabe94c0517
bpf: lsm: Implement attach, detach and execution

JITed BPF programs are dynamically attached to the LSM hooks
using BPF trampolines. The trampoline prologue generates code to handle
conversion of the signature of the hook to the appropriate BPF context.

The allocated trampoline programs are attached to the nop functions
initialized as LSM hooks.

BPF_PROG_TYPE_LSM programs must have a GPL compatible license and
and need CAP_SYS_ADMIN (required for loading eBPF programs).

Upon attachment:

* A BPF fexit trampoline is used for LSM hooks with a void return type.
* A BPF fmod_ret trampoline is used for LSM hooks which return an
  int. The attached programs can override the return value of the
  bpf LSM hook to indicate a MAC Policy decision.

Signed-off-by: KP Singh <kpsingh@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Brendan Jackman <jackmanb@google.com>
Reviewed-by: Florent Revest <revest@google.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: James Morris <jamorris@linux.microsoft.com>
Link: https://lore.kernel.org/bpf/20200329004356.27286-5-kpsingh@chromium.org
include/linux/bpf_lsm.h
kernel/bpf/bpf_lsm.c
kernel/bpf/btf.c
kernel/bpf/syscall.c
kernel/bpf/trampoline.c
kernel/bpf/verifier.c