linux-kernel - Re: [PATCH bpf-next v7 00/11] Atomics for eBPF

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: <CAADnVQJVvwoZsE1K+6qRxzF7+6CvZNzygnoBW9tZNWJELk5c=Q@mail.gmail.com>
Date:   Thu, 14 Jan 2021 19:23:32 -0800
From:   Alexei Starovoitov <alexei.starovoitov@...il.com>
To:     Brendan Jackman <jackmanb@...gle.com>
Cc:     bpf <bpf@...r.kernel.org>, Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii.nakryiko@...il.com>,
        KP Singh <kpsingh@...omium.org>,
        Florent Revest <revest@...omium.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Björn Töpel <bjorn.topel@...il.com>,
        John Fastabend <john.fastabend@...il.com>,
        Yonghong Song <yhs@...com>
Subject: Re: [PATCH bpf-next v7 00/11] Atomics for eBPF

On Thu, Jan 14, 2021 at 10:18 AM Brendan Jackman <jackmanb@...gle.com> wrote:
>
> There's still one unresolved review comment from John[3] which I
> will resolve with a followup patch.
>
> Differences from v6->v7 [1]:
>
> * Fixed riscv build error detected by 0-day robot.

Applied.
Thanks a lot.

Please address John's request in a followup and these few issues:

- rst doesn't look correct. Example:
rst2man Documentation/networking/filter.rst >/dev/null
Documentation/networking/filter.rst:1053: (WARNING/2) Inline emphasis
start-string without end-string.

> Except ``BPF_ADD`` _without_ ``BPF_FETCH`` (for legacy reasons), all 4 byte
> atomic operations require alu32 mode. Clang enables this mode by default in
> architecture v3 (``-mcpu=v3``). For older versions it can be enabled with
> ``-Xclang -target-feature -Xclang +alu32``.

It reads confusing to me.
I would rephrase 'clang enables this mode by default' into
'clang can generate new atomic instruction when -mcpu=v3 is enabled'.

'For older versions...'
This part I didn't get. The users need clang 12 that is capable to
emit these insns.
What 'older versions' you're talking about?