netdev - Re: [PATCH net-next] lib/win_minmax: export symbol of minmax_running

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: <20230414142029.6da5c1bfbf1da82cd0f1c085@linux-foundation.org>
Date:   Fri, 14 Apr 2023 14:20:29 -0700
From:   Andrew Morton <akpm@...ux-foundation.org>
To:     Eric Dumazet <edumazet@...gle.com>
Cc:     Yixin Shen <bobankhshen@...il.com>, leon@...nel.org,
        davem@...emloft.net, linux-kernel@...r.kernel.org,
        ncardwell@...gle.com, netdev@...r.kernel.org, rdunlap@...radead.org
Subject: Re: [PATCH net-next] lib/win_minmax: export symbol of
 minmax_running_min

On Fri, 14 Apr 2023 16:17:40 +0200 Eric Dumazet <edumazet@...gle.com> wrote:

> On Fri, Apr 14, 2023 at 4:27 AM Yixin Shen <bobankhshen@...il.com> wrote:
> >
> > > Please provide in-tree kernel user for that EXPORT_SYMBOL.
> >
> > It is hard to provide such an in-tree kernel user. We are trying to
> > implement newer congestion control algorithms as dynamically loaded modules.
> > For example, Copa(NSDI'18) which is adopted by Facebook needs to maintain
> > such windowed min filters. Althought it is true that we can just
> > copy-and-paste the code inside lib/win_minmax, it it more convenient to
> > give the same status of minmax_running_min as minmax_running_max.
> > It is confusing that only minmax_running_max is exported.
> 
> This is needed by net/ipv4/tcp_bbr.c , which can be a module.
> 
> > If this patch is rejected because the changes are too significant,
> 
> Well, this path would soon be reverted by people using bots/tools to
> detect unused functions,
> or unused EXPORT symbols.
> 
> So there is no point accepting it, before you submit the CC in the
> official linux tree.

It seems pretty darn screwy that we export minmax_running_max() but not
minmax_running_min().

I'd be OK taking the patch just so we aren't pretty darn screwy.  But
it would be perfectly OK to include that one-liner within the patchset
which adds a minmax_running_min() user.