Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.34 as permitted sender) client-ip=23.128.96.34;
MIME-Version: 1.0
References: <20231210130001.2050847-1-menglong8.dong@gmail.com> <4457e84f-4417-4a60-a814-9288b0756d91@linux.dev>
In-Reply-To: <4457e84f-4417-4a60-a814-9288b0756d91@linux.dev>
From:   Menglong Dong <menglong8.dong@gmail.com>
Date:   Mon, 11 Dec 2023 17:39:40 +0800
Message-ID: <CADxym3bNJXWZRfcGWpD7YW1rFe93vSOastmGrLvAcG3U2SaUdg@mail.gmail.com>
Subject: Re: [PATCH bpf-next] bpf: make the verifier trace the "not qeual" for regs
To:     Yonghong Song <yonghong.song@linux.dev>
Cc:     andrii@kernel.org, ast@kernel.org, daniel@iogearbox.net,
        john.fastabend@gmail.com, martin.lau@linux.dev, song@kernel.org,
        kpsingh@kernel.org, sdf@google.com, haoluo@google.com,
        jolsa@kernel.org, bpf@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hello,

On Mon, Dec 11, 2023 at 1:09=E2=80=AFPM Yonghong Song <yonghong.song@linux.=
dev> wrote:
>
>
> On 12/10/23 5:00 AM, Menglong Dong wrote:
> > We can derive some new information for BPF_JNE in regs_refine_cond_op()=
.
> > Take following code for example:
> >
> >    /* The type of "a" is u16 */
> >    if (a > 0 && a < 100) {
> >      /* the range of the register for a is [0, 99], not [1, 99],
> >       * and will cause the following error:
> >       *
> >       *   invalid zero-sized read
> >       *
> >       * as a can be 0.
> >       */
> >      bpf_skb_store_bytes(skb, xx, xx, a, 0);
> >    }
>
> Could you have a C test to demonstrate this example?
> Also, you should have a set of inline asm code (progs/verifier*.c)
> to test various cases as in mark_reg32_not_equal() and
> mark_reg_not_equal().
>

Yeah! I found that this part is tested in the test_progs/reg_bounds_crafted
too, and this commit failed that test case, which I should fix in the next
version.

> >
> > In the code above, "a > 0" will be compiled to "jmp xxx if a =3D=3D 0".=
 In the
> > TRUE branch, the dst_reg will be marked as known to 0. However, in the
> > fallthrough(FALSE) branch, the dst_reg will not be handled, which makes
> > the [min, max] for a is [0, 99], not [1, 99].
> >
> > For BPF_JNE, we can reduce the range of the dst reg if the src reg is a
> > const and is exactly the edge of the dst reg.
> >
> > Signed-off-by: Menglong Dong <menglong8.dong@gmail.com>
> > ---
> >   kernel/bpf/verifier.c | 45 ++++++++++++++++++++++++++++++++++++++++++=
-
> >   1 file changed, 44 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> > index 727a59e4a647..7b074ac93190 100644
> > --- a/kernel/bpf/verifier.c
> > +++ b/kernel/bpf/verifier.c
> > @@ -1764,6 +1764,40 @@ static void __mark_reg_const_zero(struct bpf_reg=
_state *reg)
> >       reg->type =3D SCALAR_VALUE;
> >   }
> >
> > +#define CHECK_REG_MIN(value)                 \
> > +do {                                         \
> > +     if ((value) =3D=3D (typeof(value))imm)      \
> > +             value++;                        \
> > +} while (0)
> > +
> > +#define CHECK_REG_MAX(value)                 \
> > +do {                                         \
> > +     if ((value) =3D=3D (typeof(value))imm)      \
> > +             value--;                        \
> > +} while (0)
> > +
> > +static void mark_reg32_not_equal(struct bpf_reg_state *reg, u64 imm)
> > +{
>
> What if reg->s32_min_value =3D=3D imm and reg->s32_max_value =3D=3D imm?
> Has this been handled in previous verifier logic?

Will such a case happen? In current code path, the src reg is a const,
and the is_branch_taken() will return 0 or 1 if the
dst_reg->s32_min_value =3D=3D dst_reg->s32_max_value.

Enn......maybe we can do more checking here in case that someone
calls this function in another place.

Thanks!
Menglong Dong

>
> > +             CHECK_REG_MIN(reg->s32_min_value);
> > +             CHECK_REG_MAX(reg->s32_max_value);
> > +             CHECK_REG_MIN(reg->u32_min_value);
> > +             CHECK_REG_MAX(reg->u32_max_value);
> > +}
> > +
> > +static void mark_reg_not_equal(struct bpf_reg_state *reg, u64 imm)
> > +{
> > +             CHECK_REG_MIN(reg->smin_value);
> > +             CHECK_REG_MAX(reg->smax_value);
> > +
> > +             CHECK_REG_MIN(reg->umin_value);
> > +             CHECK_REG_MAX(reg->umax_value);
> > +
> > +             CHECK_REG_MIN(reg->s32_min_value);
> > +             CHECK_REG_MAX(reg->s32_max_value);
> > +             CHECK_REG_MIN(reg->u32_min_value);
> > +             CHECK_REG_MAX(reg->u32_max_value);
> > +}
> > +
> >   static void mark_reg_known_zero(struct bpf_verifier_env *env,
> >                               struct bpf_reg_state *regs, u32 regno)
> >   {
> > @@ -14332,7 +14366,16 @@ static void regs_refine_cond_op(struct bpf_reg=
_state *reg1, struct bpf_reg_state
> >               }
> >               break;
> >       case BPF_JNE:
> > -             /* we don't derive any new information for inequality yet=
 */
> > +             /* try to recompute the bound of reg1 if reg2 is a const =
and
> > +              * is exactly the edge of reg1.
> > +              */
> > +             if (is_reg_const(reg2, is_jmp32)) {
> > +                     val =3D reg_const_value(reg2, is_jmp32);
> > +                     if (is_jmp32)
> > +                             mark_reg32_not_equal(reg1, val);
> > +                     else
> > +                             mark_reg_not_equal(reg1, val);
> > +             }
> >               break;
> >       case BPF_JSET:
> >               if (!is_reg_const(reg2, is_jmp32))