Received-SPF: pass (google.com: domain of linux-kernel+bounces-109983-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
From: Puranjay Mohan <puranjay12@gmail.com>
To: Alexei Starovoitov <alexei.starovoitov@gmail.com>, Ilya Leoshkevich
 <iii@linux.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>, David Ahern
 <dsahern@kernel.org>, Alexei Starovoitov <ast@kernel.org>, Daniel Borkmann
 <daniel@iogearbox.net>, Andrii Nakryiko <andrii@kernel.org>, Martin KaFai
 Lau <martin.lau@linux.dev>, Eduard Zingerman <eddyz87@gmail.com>, Song Liu
 <song@kernel.org>, Yonghong Song <yonghong.song@linux.dev>, John Fastabend
 <john.fastabend@gmail.com>, KP Singh <kpsingh@kernel.org>, Stanislav
 Fomichev <sdf@google.com>, Hao Luo <haoluo@google.com>, Jiri Olsa
 <jolsa@kernel.org>, Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar
 <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>, Dave Hansen
 <dave.hansen@linux.intel.com>, X86 ML <x86@kernel.org>, "H. Peter Anvin"
 <hpa@zytor.com>, Jean-Philippe Brucker <jean-philippe@linaro.org>, Network
 Development <netdev@vger.kernel.org>, bpf <bpf@vger.kernel.org>, LKML
 <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH bpf v2] bpf: verifier: prevent userspace memory access
In-Reply-To: <CAADnVQLHrmkJ5p2gEUJkf_CRxq9gv8rcSuBm5GeZ_nUJxQOE0Q@mail.gmail.com>
References: <20240321101058.68530-1-puranjay12@gmail.com>
 <CAADnVQLhwLgq=QuXD-Ls=t9Scr_4Zn9JwdkXfZQfZkT=ysx64Q@mail.gmail.com>
 <CAADnVQLHrmkJ5p2gEUJkf_CRxq9gv8rcSuBm5GeZ_nUJxQOE0Q@mail.gmail.com>
Date: Thu, 21 Mar 2024 11:48:52 +0000
Message-ID: <mb61pwmpvst3v.fsf@gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

Alexei Starovoitov <alexei.starovoitov@gmail.com> writes:

> On Thu, Mar 21, 2024 at 4:05=E2=80=AFAM Alexei Starovoitov
> <alexei.starovoitov@gmail.com> wrote:
>>
>> On Thu, Mar 21, 2024 at 3:11=E2=80=AFAM Puranjay Mohan <puranjay12@gmail=
com> wrote:
>> >
>> > diff --git a/arch/s390/net/bpf_jit_comp.c b/arch/s390/net/bpf_jit_comp=
c
>> > index e613eebfd349..e61a51a5b4be 100644
>> > --- a/arch/s390/net/bpf_jit_comp.c
>> > +++ b/arch/s390/net/bpf_jit_comp.c
>> > @@ -2691,3 +2691,8 @@ bool bpf_jit_supports_subprog_tailcalls(void)
>> >  {
>> >         return true;
>> >  }
>> > +
>> > +u64 bpf_arch_uaddress_limit(void)
>> > +{
>> > +       return -ENOTSUPP;
>> > +}
>>
>> Looks good and should work, but s390 CI is still not happy.
>> Ideas?
>> sock tests were not failing before. So something is going on.
>
> I think I have an explanation.
> -ENOTSUPP and u64... and later:
> u64 uaddress_limit =3D bpf_arch_uaddress_limit()
> if (uaddress_limit < 0)
>
> I bet the compiler simply removes this check since unsigned cannot
> be negative.
> Odd that there is no compiler warning.
>
> pw-bot: cr
>

Yes, I verified that the compiler is removing this:


                if (BPF_CLASS(insn->code) =3D=3D BPF_LDX &&
    a944:       7100047f        cmp     w3, #0x1
    a948:       540013e1        b.ne    abc4 <do_misc_fixups+0x66c>  // b.a=
ny
    a94c:       721a041f        tst     w0, #0xc0
    a950:       54fff4e1        b.ne    a7ec <do_misc_fixups+0x294>  // b.a=
ny
                        u64 uaddress_limit =3D bpf_arch_uaddress_limit();
    a954:       b90003e6        str     w6, [sp]
    a958:       94000000        bl      0 <bpf_arch_uaddress_limit>
                        *patch++ =3D BPF_MOV64_REG(BPF_REG_AX, insn->src_re=
g);


We should do:
   if (!uaddress_limit)
        goto next_insn;

and in the disabled case return 0 in place of -ENOSUPP.

Doing this adds the check:


             if (BPF_CLASS(insn->code) =3D=3D BPF_LDX &&
    a944:       7100047f        cmp     w3, #0x1
    a948:       54001401        b.ne    abc8 <do_misc_fixups+0x670>  // b.a=
ny
    a94c:       721a041f        tst     w0, #0xc0
    a950:       54fff4e1        b.ne    a7ec <do_misc_fixups+0x294>  // b.a=
ny
                        u64 uaddress_limit =3D bpf_arch_uaddress_limit();
    a954:       b90003e6        str     w6, [sp]
    a958:       94000000        bl      0 <bpf_arch_uaddress_limit>
                        if (!uaddress_limit)
    a95c:       b4fff020        cbz     x0, a760 <do_misc_fixups+0x208>
                        *patch++ =3D BPF_MOV64_REG(BPF_REG_AX, insn->src_re=
g);


I will send v3 with this approach.


Thanks,
Puranjay