On Fri, Oct 2, 2020 at 5:26 PM Stanislav Fomichev <sdf@xxxxxxxxxx> wrote:
>
> We are missing a deref for the case when we are doing BPF_PROG_BIND_MAP
> on a map that's being already held by the program.
> There is 'if (ret) bpf_map_put(map)' below which doesn't trigger
> because we don't consider this an error.
> Let's add missing bpf_map_put() for this specific condition.
>
> Fixes: ef15314aa5de ("bpf: Add BPF_PROG_BIND_MAP syscall")
> Reported-by: Alexei Starovoitov <ast@xxxxxxxxxx>
> Signed-off-by: Stanislav Fomichev <sdf@xxxxxxxxxx>
> ---
Acked-by: Andrii Nakryiko <andriin@xxxxxx>
> kernel/bpf/syscall.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> index f1528c2a6927..1110ecd7d1f3 100644
> --- a/kernel/bpf/syscall.c
> +++ b/kernel/bpf/syscall.c
> @@ -4323,8 +4323,10 @@ static int bpf_prog_bind_map(union bpf_attr *attr)
> used_maps_old = prog->aux->used_maps;
>
> for (i = 0; i < prog->aux->used_map_cnt; i++)
> - if (used_maps_old[i] == map)
> + if (used_maps_old[i] == map) {
> + bpf_map_put(map);
> goto out_unlock;
> + }
>
> used_maps_new = kmalloc_array(prog->aux->used_map_cnt + 1,
> sizeof(used_maps_new[0]),
> --
> 2.28.0.806.g8561365e88-goog
>
