From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jarek Poplawski <jarkao2@gmail.com>
Subject: Re: [PATCH] net: fix setsockopt() locking errors
Date: Mon, 26 Jan 2009 11:50:12 +0000
Message-ID: <20090126115012.GA5620@ff.dom.local>
References: <20090124224930.GA4456@localhost.localdomain>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: "David S. Miller" <davem@davemloft.net>,
	=?us-ascii?B?PT9JU08tODg1OS0yP1E/TWFydGluX01PS1I/PSA9P0lTTy04ODU5LTI/?=
	 =?us-ascii?B?UT9FSj1BOT89?= <mmokrejs@ribosome.natur.cuni.cz>,
	netdev@vger.kernel.org
To: Vegard Nossum <vegard.nossum@gmail.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from fg-out-1718.google.com ([72.14.220.152]:25931 "EHLO
	fg-out-1718.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1751512AbZAZLuX (ORCPT
	<rfc822;netdev@vger.kernel.org>); Mon, 26 Jan 2009 06:50:23 -0500
Received: by fg-out-1718.google.com with SMTP id 19so3485711fgg.17
        for <netdev@vger.kernel.org>; Mon, 26 Jan 2009 03:50:18 -0800 (PST)
Content-Disposition: inline
In-Reply-To: <20090124224930.GA4456@localhost.localdomain>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On 24-01-2009 23:49, Vegard Nossum wrote:
> Hi,
> 
> This survives basic testing here, but I don't know what that counts for
> when I couldn't reproduce the lockdep report in the first place. Please
> review.
> 
> 
> Vegard
> 
> 
> From cc8bcd1c4fd219a31d6d191aefa4b4b57dadb9b0 Mon Sep 17 00:00:00 2001
> From: Vegard Nossum <vegard.nossum@gmail.com>
> Date: Sat, 24 Jan 2009 22:44:16 +0100
> Subject: [PATCH] net: fix setsockopt() locking errors
> MIME-Version: 1.0
> Content-Type: text/plain; charset=utf-8
> Content-Transfer-Encoding: 8bit
> 
> Martin MOKREJ.  <mmokrejs@ribosome.natur.cuni.cz> reported:
>> =======================================================
>> [ INFO: possible circular locking dependency detected ]
>> 2.6.29-rc2-git1 #1
>> -------------------------------------------------------
>> tcpdump/3734 is trying to acquire lock:
>>  (&mm->mmap_sem){----}, at: [<c1053294>] might_fault+0x30/0x6b
>>
>> but task is already holding lock:
>>  (sk_lock-AF_PACKET){--..}, at: [<c12798c8>] sock_setsockopt+0x12b/0x4a4
>>
>> which lock already depends on the new lock.
> 
> It turns out that sock_setsockopt() is calling copy_from_user() while
> holding the lock on the socket.

I guess it has been like this for some time, so it would be nice to
mention what scenario happens here, or IOW what exactly needs to get
these locks in reverse order.

> We fix it by splitting the ioctl code
> so that one switch handles the ioctls that have their own code for
> reading from userspace, and one switch handles the cases that require
> no additional reading.
> 
> Reported-by: Martin MOKREJ.  <mmokrejs@ribosome.natur.cuni.cz>
> Signed-off-by: Vegard Nossum <vegard.nossum@gmail.com>
> ---
>  net/core/sock.c |  134 +++++++++++++++++++++++++++++++++++-------------------
>  1 files changed, 87 insertions(+), 47 deletions(-)
> 
> diff --git a/net/core/sock.c b/net/core/sock.c
> index f3a0d08..6bd618d 100644
> --- a/net/core/sock.c
> +++ b/net/core/sock.c
> @@ -424,6 +424,80 @@ out:
>  	return ret;
>  }
>  
> +static int sock_linger(struct sock *sk, char __user *optval, int optlen)
...
> +static int sock_set_rcvtimeo(struct sock *sk, char __user *optval, int optlen)
> +{
> +	int ret;
> +	long rcvtimeo;
> +
> +	ret = sock_set_timeout(&rcvtimeo, optval, optlen);

A check for error is needed here and below.

> +
> +	lock_sock(sk);
> +	sk->sk_rcvtimeo = rcvtimeo;
> +	release_sock(sk);
> +
> +	return ret;
> +}
> +
> +static int sock_set_sndtimeo(struct sock *sk, char __user *optval, int optlen)
> +{
> +	int ret;
> +	long sndtimeo;
> +
> +	ret = sock_set_timeout(&sndtimeo, optval, optlen);
> +
> +	lock_sock(sk);
> +	sk->sk_sndtimeo = sndtimeo;
> +	release_sock(sk);
> +
> +	return ret;
> +}
...

Jarek P.