From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Miller <davem@davemloft.net>
Subject: Re: [patch net] net: correct check in dev_addr_del()
Date: Thu, 15 Nov 2012 17:58:35 -0500 (EST)
Message-ID: <20121115.175835.2209521940651702184.davem@davemloft.net>
References: <1352897464-832-1-git-send-email-jiri@resnulli.us>
	<20121114.215254.973344009331632746.davem@davemloft.net>
	<20121115101238.GA2807@minipsycho.orion>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Cc: netdev@vger.kernel.org, eric.dumazet@gmail.com,
	shemminger@vyatta.com, john.r.fastabend@intel.com
To: jiri@resnulli.us
Return-path: <netdev-owner@vger.kernel.org>
Received: from shards.monkeyblade.net ([149.20.54.216]:49240 "EHLO
	shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1750965Ab2KOW6j (ORCPT
	<rfc822;netdev@vger.kernel.org>); Thu, 15 Nov 2012 17:58:39 -0500
In-Reply-To: <20121115101238.GA2807@minipsycho.orion>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

From: Jiri Pirko <jiri@resnulli.us>
Date: Thu, 15 Nov 2012 11:12:38 +0100

> Thu, Nov 15, 2012 at 03:52:54AM CET, davem@davemloft.net wrote:
>>From: Jiri Pirko <jiri@resnulli.us>
>>Date: Wed, 14 Nov 2012 13:51:04 +0100
>>
>>> Check (ha->addr == dev->dev_addr) is always true because dev_addr_init()
>>> sets this. Correct the check to behave properly on addr removal.
>>> 
>>> Signed-off-by: Jiri Pirko <jiri@resnulli.us>
>>
>>I'm pretty sure this is very intentional.
>>
>>It's trying to prevent deletion of the implicit dev->dev_addr
>>entry.  But it will allow decementing the reference count to
>>1, but no further.
>>
>>I'm not applying this.
> 
> Please look at dev_addr_init(), line 266:
> dev->dev_addr = ha->addr;
> 
> and that is never changed.
> Therefore check (ha->addr == dev->dev_addr) in dev_addr_del() is always
> true and has no meaning.
> 
> dev_addr_del() should check if the address passed in "const unsigned
> char *addr" function arg is the same as in ha->addr (dev->dev_addr) and
> prevent to remove it in that case. And that is what this patch does.

Thanks for explaining.  I misread the code and thought that 'ha'
was already the object looked up using 'addr' as the key.

Applied and queued up for -stable, thanks.