linux-kernel - Re: [PATCH v2 1/3] bpf: Allow NULL buffers in bpf_dynptr_slice(

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

Date:   Tue, 18 Jul 2023 16:17:24 -0700
From:   Alexei Starovoitov <alexei.starovoitov@...il.com>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     Daniel Rosenberg <drosen@...gle.com>, bpf <bpf@...r.kernel.org>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        John Fastabend <john.fastabend@...il.com>,
        Andrii Nakryiko <andrii@...nel.org>,
        Martin KaFai Lau <martin.lau@...ux.dev>,
        Song Liu <song@...nel.org>, Yonghong Song <yhs@...com>,
        KP Singh <kpsingh@...nel.org>,
        Stanislav Fomichev <sdf@...gle.com>,
        Hao Luo <haoluo@...gle.com>, Jiri Olsa <jolsa@...nel.org>,
        Shuah Khan <shuah@...nel.org>,
        Jonathan Corbet <corbet@....net>,
        Joanne Koong <joannelkoong@...il.com>,
        Mykola Lysenko <mykolal@...com>,
        LKML <linux-kernel@...r.kernel.org>,
        "open list:KERNEL SELFTEST FRAMEWORK" 
        <linux-kselftest@...r.kernel.org>,
        Android Kernel Team <kernel-team@...roid.com>
Subject: Re: [PATCH v2 1/3] bpf: Allow NULL buffers in bpf_dynptr_slice(_rw)

On Tue, Jul 18, 2023 at 4:06 PM Jakub Kicinski <kuba@...nel.org> wrote:
>
> On Tue, 18 Jul 2023 13:34:06 -0700 Alexei Starovoitov wrote:
> > > Direct packet access via skb->data is there for those who want high
> > > speed 🤷️
> >
> > skb->data/data_end approach unfortunately doesn't work that well.
> > Too much verifier fighting. That's why dynptr was introduced.
>
> I wish Daniel told us more about the use case.
>
> > > My worry is that people will think that whether the buffer is needed or
> > > not depends on _their program_, rather than on the underlying platform.
> > > So if it works in testing without the buffer - the buffer must not be
> > > required for their use case.
> >
> > Are you concerned about bpf progs breaking this way?
>
> Both, BPF progs breaking and netdev code doing things which don't make
> sense. But I won't argue too hard about the former, i.e. the BPF API.
>
> > I thought you're worried about the driver misusing
> > skb_header_pointer() with buffer==NULL.
> >
> > We can remove !buffer check as in the attached patch,
> > but I don't quite see how it would improve driver quality.
>
> The drivers may not be pretty but they aren't buggy AFAICT.
>
> > [0001-bpf-net-Introduce-skb_pointer_if_linear.patch  application/octet-stream (2873 bytes)]
>
> Or we can simply pretend we don't have the skb:
>
> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
> index 91ed66952580..217447f01d56 100644
> --- a/include/linux/skbuff.h
> +++ b/include/linux/skbuff.h
> @@ -4023,7 +4023,7 @@ __skb_header_pointer(const struct sk_buff *skb, int offset, int len,
>         if (likely(hlen - offset >= len))
>                 return (void *)data + offset;
>
> -       if (!skb || !buffer || unlikely(skb_copy_bits(skb, offset, buffer, len) < 0))
> +       if (!skb || unlikely(skb_copy_bits(skb, offset, buffer, len) < 0))
>                 return NULL;
>
>         return buffer;
> diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
> index 9e80efa59a5d..8bc4622cc1df 100644
> --- a/kernel/bpf/helpers.c
> +++ b/kernel/bpf/helpers.c
> @@ -2239,7 +2239,13 @@ __bpf_kfunc void *bpf_dynptr_slice(const struct bpf_dynptr_kern *ptr, u32 offset
>         case BPF_DYNPTR_TYPE_RINGBUF:
>                 return ptr->data + ptr->offset + offset;
>         case BPF_DYNPTR_TYPE_SKB:
> -               return skb_header_pointer(ptr->data, ptr->offset + offset, len, buffer__opt);
> +       {
> +               const struct sk_buff *skb = ptr->data;
> +
> +               return __skb_header_pointer(NULL, ptr->offset + offset, len,
> +                                           skb->data, skb_headlen(skb),
> +                                           buffer__opt);
> +       }

Which would encourage bnxt-like hacks.
I don't like it tbh.
At least skb_pointer_if_linear() has a clear meaning.
It's more run-time overhead, since buffer__opt is checked early,
but that's ok.