netdev - Re: [bug] stuck localhost TCP connections, v2.6.26-rc3+

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:	Tue, 27 May 2008 00:20:05 +0300 (EEST)
From:	"Ilpo Järvinen" <ilpo.jarvinen@...sinki.fi>
To:	Ingo Molnar <mingo@...e.hu>, Pavel Emelyanov <xemul@...nvz.org>,
	"Denis V. Lunev" <den@...nvz.org>
cc:	LKML <linux-kernel@...r.kernel.org>,
	Netdev <netdev@...r.kernel.org>,
	"David S. Miller" <davem@...emloft.net>,
	"Rafael J. Wysocki" <rjw@...k.pl>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [bug] stuck localhost TCP connections, v2.6.26-rc3+

...Added Pavel & Denis.

On Mon, 26 May 2008, Ingo Molnar wrote:
> * Ingo Molnar <mingo@...e.hu> wrote:
> 
> > > It's well possible that e.g., net namespaces have some bug in 
> > > handling of orphaned tcp.
> > 
> > yes, that would match the symptoms i think. I half-assumed that it's a 
> > state machine problem so i didnt even check what the reader does - and 
> > in this case it appears to not exist at all anymore ;-)

I keep wondering why distcc in the first place has a connection orphaned
before it has read all data from it, I guess it should not yet be
speculatively executing anything or so :-). Anyway, there should be RST
due to that data because nobody is going read it after the process in gone 
but since state is still ESTABLISHED, I doubt that tcp_close() was ever 
executed which adds some strangeness.

> after ~7 hours of uptime the networking code produced this assertion:
> 
>  [25441.140000] KERNEL: assertion (!sk->sk_ack_backlog) failed at 
>  net/ipv4/inet_connection_sock.c (642)

...I checked /proc/net/tcp output you gave earlier, and indeed, it shows 
1 in sk->sk_ack_backlog of the LISTENing socket at that time.

> no further information in the logs. I kept the failing system booted up 
> for 8 hours and will reboot it now.



-- 
 i.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html