netdev - Re: tcp hang when socket fills up ?

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:   Wed, 18 Apr 2018 18:36:22 +0900
From:   Dominique Martinet <asmadeus@...ewreck.org>
To:     Jozsef Kadlecsik <kadlec@...ckhole.kfki.hu>,
        Florian Westphal <fw@...len.de>,
        Michal Kubecek <mkubecek@...e.cz>, netdev@...r.kernel.org,
        Marcelo Ricardo Leitner <marcelo.leitner@...il.com>,
        Eric Dumazet <eric.dumazet@...il.com>
Subject: Re: tcp hang when socket fills up ?

Dominique Martinet wrote on Wed, Apr 18, 2018:
> Jozsef Kadlecsik wrote on Wed, Apr 18, 2018:
> > Yes, the state transition is wrong for simultaneous open, because the 
> > tcp_conntracks table is not (cannot be) smart enough. Could you verify the 
> > next untested patch?
> 
> Thanks for the patch; I'll give it a try (probably won't make it today
> so will report tomorrow)

Actually had time; I can confirm (added printks) we did get in that if
that was pointed at, and we no longer get there now.
The connection no longer gets in invalid state, so that looks like it
nailed it.

I'm now confused what this has to do with tcp_timestamp though, since
setting that off also seemed to work around the issue, but if we get
something like that in I'll be happy anyway.

Big props to everyone involved, I would have taken much longer alone,
-- 
Dominique Martinet | Asmadeus