netdev - Re: [RFC] [PATCH] Don't destroy TCP sockets twice

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]

Date:	Tue, 10 Aug 2010 06:39:31 -0400
From:	Herbert Xu <herbert@...dor.apana.org.au>
To:	Andi Kleen <andi@...stfloor.org>
Cc:	netdev@...r.kernel.org
Subject: Re: [RFC] [PATCH] Don't destroy TCP sockets twice

On Tue, Aug 10, 2010 at 12:32:06PM +0200, Andi Kleen wrote:
>
> Yes I stored the backtrace of the first caller in the ugly debug
> patches and dumped that on the second destroy. It was tcp_done the 
> first time.
> 
> Also did the same for tcp_sk() and there it was the fin sending.
> 
> I agree that tcp_close() should skip it in theory but I saw
> it anyways :/

So I presume the second caller was tcp_close? That means we have
a serious bug in our stack, as if the socket is already in the
CLOSE state then tcp_close should have short-circuited.

This means that something is changing the TCP socket state after
going into CLOSE.  Can you try adding your debug backtrace patch to
tcp_set_state to see if anybody is indeed changing the socket
state after going into CLOSE (and more importantly who is changing
it)?

Thanks,
-- 
Email: Herbert Xu <herbert@...dor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
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