netdev - Re: [PATCH bpf 0/2] bpf: mark registers as safe or unknown in all frames

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20190422183014.GA12006@Nover>
Date:   Mon, 22 Apr 2019 20:30:15 +0200
From:   Paul Chaignon <paul.chaignon@...nge.com>
To:     Yonghong Song <yhs@...com>, Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>, netdev@...r.kernel.org,
        bpf@...r.kernel.org
Cc:     Xiao Han <xiao.han@...nge.com>, Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>
Subject: Re: [PATCH bpf 0/2] bpf: mark registers as safe or unknown in all
 frames

On Mon, Apr 22, 2019 04:57PM, Yonghong Song wrote: 
> On 4/20/19 5:38 AM, Paul Chaignon wrote:
> > In case of a null check on a pointer inside a subprog, we should mark all
> > registers with this pointer as either safe or unknown, in both the current
> > and previous frames.  Currently, only spilled registers and registers in
> > the current frame are marked.  This first patch also marks registers in
> > previous frames.
> > 
> > A good reproducer looks as follow:
> > 
> > 1: ptr = bpf_map_lookup_elem(map, &key);
> > 2: ret = subprog(ptr) {
> > 3:   return ptr != NULL;
> > 4: }
> > 5: if (ret)
> > 6:   value = *ptr;
> > 
> > With the above, the verifier will complain on line 6 because it sees ptr
> > as map_value_or_null despite the null check in subprog 1.  The second
> > patch implements the above as a new test case.
> > 
> > Note that this patch fixes another resulting bug when using
> > bpf_sk_release():
> > 
> > 1: sk = bpf_sk_lookup_tcp();
> > 2: subprog(sk) {
> > 3:   if (sk)
> > 4:     bpf_sk_release(sk, 0);
> 
> The specification for bpf_sk_release() in uapi/linux/bpf.h is:
>   int bpf_sk_release(struct bpf_sock *sock)
> 
> Do you explain what is bpf_sk_release(sk, 0)?

Thanks for the review Yonghong.  I think I took this extra argument by
mistake from the description of the commit introducing bpf_sk_release
(6acc9b432 ("bpf: Add helper to retrieve socket in BPF")).  I'll send a v2
with a fixed commit message.  Of course, the helper on line 1 also takes
arguments, so it might be better to write it as bpf_sk_lookup_tcp(...).

> 
> > 5: }
> > 6: if (!sk)
> > 7:   return 0;
> > 8: return sk;
> 
> If sk has been released, the program should not really return sk, right?

I'll change in v2.  I don't think it matters to reproduce the warning
though.  The verifier won't complain as the return value won't be
dereferenced and the register holding sk is readable.

> 
> > 
> > In the above, mark_ptr_or_null_regs will warn on line 6 because it will
> > try to free the reference state, even though it was already freed on
> > line 3.
> > 
> > Paul Chaignon (2):
> >    bpf: mark registers as safe or unknown in all frames
> >    selftests/bpf: test case for pointer null check in subprog
> > 
> >   kernel/bpf/verifier.c                        |  6 ++---
> >   tools/testing/selftests/bpf/verifier/calls.c | 25 ++++++++++++++++++++
> >   2 files changed, 28 insertions(+), 3 deletions(-)
> >