> On Apr 15, 2021, at 7:55 PM, Dave Marchevsky <davemarchevsky@xxxxxx> wrote: > > Verifier can constrain the min/max bounds of bpf_get_task_stack's return > value more tightly than the default tnum_unknown. Like bpf_get_stack, > return value is num bytes written into a caller-supplied buf, or error, > so do_refine_retval_range will work. > > Signed-off-by: Dave Marchevsky <davemarchevsky@xxxxxx> Acked-by: Song Liu <songliubraving@xxxxxx> > --- > kernel/bpf/verifier.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c > index 852541a435ef..348e97f77003 100644 > --- a/kernel/bpf/verifier.c > +++ b/kernel/bpf/verifier.c > @@ -5767,6 +5767,7 @@ static void do_refine_retval_range(struct bpf_reg_state *regs, int ret_type, > > if (ret_type != RET_INTEGER || > (func_id != BPF_FUNC_get_stack && > + func_id != BPF_FUNC_get_task_stack && > func_id != BPF_FUNC_probe_read_str && > func_id != BPF_FUNC_probe_read_kernel_str && > func_id != BPF_FUNC_probe_read_user_str)) > -- > 2.30.2 >
