netdev - Re: [PATCH] tcp: Fix tcp_max_syn_backlog limit on connection requests

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:   Fri, 3 Jan 2020 05:52:45 -0800
From:   Eric Dumazet <edumazet@...gle.com>
To:     Ttttabcd <ttttabcd@...tonmail.com>
Cc:     Netdev <netdev@...r.kernel.org>,
        David Miller <davem@...emloft.net>,
        "kuznet@....inr.ac.ru" <kuznet@....inr.ac.ru>,
        "yoshfuji@...ux-ipv6.org" <yoshfuji@...ux-ipv6.org>
Subject: Re: [PATCH] tcp: Fix tcp_max_syn_backlog limit on connection requests

On Fri, Jan 3, 2020 at 5:19 AM Ttttabcd <ttttabcd@...tonmail.com> wrote:
>
> However, I think that backward compatibility should not be too serious because sysctl_max_syn_backlog is only enabled when syn_cookies is turned off.
>

Yes, but not your new code :

       max_syn_backlog = min_t(u32, net->ipv4.sysctl_max_syn_backlog,
                             sk->sk_max_ack_backlog);


> If sysctl_max_syn_backlog is set small, there is no difference between the original code and the new code.

There is a difference though....

Set sysctl_max_syn_backlog to 1, and start a reasonable test (not a synflood)

As soon as one SYN_RECV request socket is inserted in the hash, other
SYN packets will generate a syncookie, even if the backlog of the
listener is 100,000

I think the intent of the code is to allow a heavy duty server to use
a huge backlog (say 1,000,000) to avoid syncookies if it has enough
RAM.

In this mode, people never had to tweak sysctl_max_syn_backlog.

>
> Only in the BUG scenarios I mentioned in the patch, the system behavior will change, but these are corrections that have no impact on users.
>
> It's just that the part of the request retention queue will not be mistakenly occupied, and earlier use of syn cookies instead of filling up the backlog.