[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <586D7E6A.5080009@iogearbox.net>
Date: Wed, 04 Jan 2017 23:59:54 +0100
From: Daniel Borkmann <daniel@...earbox.net>
To: Sowmini Varadhan <sowmini.varadhan@...cle.com>
CC: linux-kselftest@...r.kernel.org, netdev@...r.kernel.org,
willemb@...gle.com, davem@...emloft.net, shuah@...nel.org
Subject: Re: [PATCH v3 net-next 1/2] tools: psock_lib: tighten conditions
checked in sock_setfilter
On 01/04/2017 11:48 PM, Sowmini Varadhan wrote:
> On (01/04/17 23:26), Daniel Borkmann wrote:
[...]
>>>> As it stands it makes it a bit harder to parse / less readable with macros
>>>> actually. Rest seems fine, thanks.
>
> Usually macros are there (a) as an abstraction so you
> dont have to hard-code things, and, (b) to make things
> more readable. (maybe that's why the 1992 VJ paper on
> BPF came up with these macros?)
>
> I think we differ on code-aesthetics (not correctness) here.
> It was not immediately obvious to me that "0x15 is actually
> BPF_JMP + BPF_JEQ + BPF_K" etc, when I wanted to extend
> the bpf_prog to harden the checks in the existing code.
>
> Would it be ok to leave the extremely subjective
> "make this more readable" part for you to tackle later?
> Or I can just drop patch1, and you can fix it to your
> satisfaction later.
I think we're talking past each other (?), my suggestion
from my original email was to use bpf_asm and paste the
(human readable) program as a comment above as done also
elsewhere. But just leave it as it is then, no big deal
either.
Powered by blists - more mailing lists