netdev - Re: [PATCH v2 1/3] bpf: Use 1<<16 as ceiling for immediate alignment in verifier.

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:   Thu, 18 May 2017 15:10:03 +0100
From:   Edward Cree <ecree@...arflare.com>
To:     David Miller <davem@...emloft.net>
CC:     <ast@...com>, <daniel@...earbox.net>,
        <alexei.starovoitov@...il.com>, <netdev@...r.kernel.org>
Subject: Re: [PATCH v2 1/3] bpf: Use 1<<16 as ceiling for immediate alignment
 in verifier.

On 18/05/17 01:16, David Miller wrote:
> So, in C, addition (a += b) is something like:
>
> struct bpf_reg_bits {
>         u64 zero_bits;
>         u64 one_bits;
> };
>
> static void add_update_bits(struct bpf_reg_bits *a, struct bpf_reg_bits *b)
> {
>         u64 m_zeros, m_ones, m_all;
>
>         m_zeros = a->zero_bits ^ b->zero_bits;
>         m_ones = a->one_bits ^ b->one_bits;
>         m_all = m_zeros | m_ones;
No, this should be
    u64 m_a, m_b, m_all;

    m_a = a->zero_bits ^ a->one_bits; /* unknown bits in a */
    m_b = b->zero_bits ^ b->one_bits; /* unknown bits in b */
    m_all = m_a | m_b; /* unknown bits in result */
>         a->zero_bits = (a->zero_bits + b->zero_bits) | m_all;
>         a->one_bits = (a->one_bits + b->zero_bits) & ~m_all;
> }
>
> Then, is subtraction merely:
>
> static void sub_update_bits(struct bpf_reg_bits *a, struct bpf_reg_bits *b)
> {
>         u64 m_zeros, m_ones, m_all;
>
>         m_zeros = a->zero_bits ^ b->zero_bits;
>         m_ones = a->one_bits ^ b->one_bits;
>         m_all = m_zeros | m_ones;
>
>         a->zero_bits = (a->zero_bits - b->zero_bits) | m_all;
>         a->one_bits = (a->one_bits - b->zero_bits) & ~m_all;
> }
>
> Or is something different needed?
I suspect it's something different, just because I worry about what
 carries will do.

But I think Alexei's idea (mask and value) is better anyway; at the
 least it's easier to think about.

-Ed