Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Toke =?utf-8?Q?H=C3=B8iland-J=C3=B8rgensen?= <toke@redhat.com>
To:     Sami Tolvanen <samitolvanen@google.com>
Cc:     =?utf-8?B?QmrDtnJuIFTDtnBlbA==?= <bjorn.topel@intel.com>,
        Yonghong Song <yhs@fb.com>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Kees Cook <keescook@chromium.org>, Martin Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>,
        "netdev\@vger.kernel.org" <netdev@vger.kernel.org>,
        "bpf\@vger.kernel.org" <bpf@vger.kernel.org>,
        "linux-kernel\@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Jesper Dangaard Brouer <brouer@redhat.com>
Subject: Re: [PATCH] bpf: validate bpf_func when BPF_JIT is enabled
In-Reply-To: <CABCJKufCwjXQ6a4oLjywDmxY2apUZ1yop-5+qty82bfwV-QTAA@mail.gmail.com>
References: <20190909223236.157099-1-samitolvanen@google.com> <4f4136f5-db54-f541-2843-ccb35be25ab4@fb.com> <20190910172253.GA164966@google.com> <c7c7668e-6336-0367-42b3-2f6026c466dd@fb.com> <fd8b6f04-3902-12e9-eab1-fa85b7e44dd5@intel.com> <87impzt4pu.fsf@toke.dk> <CABCJKufCwjXQ6a4oLjywDmxY2apUZ1yop-5+qty82bfwV-QTAA@mail.gmail.com>
Date:   Thu, 12 Sep 2019 11:46:39 +0100
Message-ID: <87sgp1ssfk.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8BIT
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Sami Tolvanen <samitolvanen@google.com> writes:

> On Wed, Sep 11, 2019 at 5:09 AM Toke Høiland-Jørgensen <toke@redhat.com> wrote:
>>
>> Björn Töpel <bjorn.topel@intel.com> writes:
>> > I ran the "xdp_rxq_info" sample with and without Sami's patch:
>>
>> Thanks for doing this!
>
> Yes, thanks for testing this Björn!
>
>> Or (1/22998700 - 1/23923874) * 10**9 == 1.7 nanoseconds of overhead.
>>
>> I guess that is not *too* bad; but it's still chipping away at
>> performance; anything we could do to lower the overhead?
>
> The check is already rather minimal, but I could move this to a static
> inline function to help ensure the compiler doesn't generate an
> additional function call for this. I'm also fine with gating this
> behind a separate config option, but I'm not sure if that's worth it.
> Any thoughts?

I think it would be good if you do both. I'm a bit worried that XDP
performance will end up in a "death by a thousand paper cuts" situation,
so I'd rather push back on even relatively small overheads like this; so
being able to turn it off in the config would be good.

Can you share more details about what the "future CFI checking" is
likely to look like?

-Toke