Re: [PATCH bpf-next] kernel/bpf: Fix an array-index-out-of-bounds issue in disasm.c

[Eduard Zingerman <eddyz87@gmail.com>]

On Mon, 2023-07-31 at 13:45 -0700, Yonghong Song wrote:
> syzbot reported an array-index-out-of-bounds when printing out bpf
> insns. Further investigation shows the insn is illegal but
> is printed out due to log level 1 or 2 before actual insn verification
> in do_check().
> 
> This particular illegal insn is a MOVSX insn with offset value 2.
> The legal offset value for MOVSX should be 8, 16 and 32.
> The disasm sign-extension-size array index is calculated as
>  (insn->off / 8) - 1
> and offset value 2 gives an out-of-bound index -1.
> 
> Tighten the checking for MOVSX insn in disasm.c to avoid
> array-index-out-of-bounds issue.
> 
> Reported-by: syzbot+3758842a6c01012aa73b@syzkaller.appspotmail.com
> Fixes: f835bb622299 ("bpf: Add kernel/bpftool asm support for new instructions")
> Signed-off-by: Yonghong Song <yonghong.song@linux.dev>

Acked-by: Eduard Zingerman <eddyz87@gmail.com>

> ---
>  kernel/bpf/disasm.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/bpf/disasm.c b/kernel/bpf/disasm.c
> index d7bff608f299..ef7c107c7b8f 100644
> --- a/kernel/bpf/disasm.c
> +++ b/kernel/bpf/disasm.c
> @@ -162,7 +162,8 @@ static bool is_sdiv_smod(const struct bpf_insn *insn)
>  
>  static bool is_movsx(const struct bpf_insn *insn)
>  {
> -	return BPF_OP(insn->code) == BPF_MOV && insn->off != 0;
> +	return BPF_OP(insn->code) == BPF_MOV &&
> +	       (insn->off == 8 || insn->off == 16 || insn->off == 32);
>  }
>  
>  void print_bpf_insn(const struct bpf_insn_cbs *cbs,