From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Miller <davem@davemloft.net>
Subject: Re: [PATCH] net: arp, ipv6: handle special case of tap device
Date: Sat, 29 Sep 2018 11:40:23 -0700 (PDT)
Message-ID: <20180929.114023.1370133506922372520.davem@davemloft.net>
References: <20180926093018.6646-1-vdronov@redhat.com>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Cc: kuznet@ms2.inr.ac.ru, yoshfuji@linux-ipv6.org,
        netdev@vger.kernel.org, syzkaller@googlegroups.com,
        linux-kernel@vger.kernel.org
To: vdronov@redhat.com
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <20180926093018.6646-1-vdronov@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: netdev.vger.kernel.org

From: Vladis Dronov <vdronov@redhat.com>
Date: Wed, 26 Sep 2018 11:30:18 +0200

> @@ -187,7 +187,14 @@ EXPORT_SYMBOL(arp_tbl);
>  
>  int arp_mc_map(__be32 addr, u8 *haddr, struct net_device *dev, int dir)
>  {
> -	switch (dev->type) {
> +	unsigned short type = dev->type;
> +
> +#if IS_ENABLED(CONFIG_TAP)
> +	if (dev->rtnl_link_ops && !strcmp(dev->rtnl_link_ops->kind, "tun"))
> +		type = ARPHRD_ETHER;
> +#endif /* CONFIG_TAP */

This is insanely ugly.

dev->type determines the link layer header layout and size.

You can fix the kernel, but userspace AF_PACKET applications are still
going to be broken by this behavior.  And that is just the tip of the
iceberg.

I'm not applying this, sorry.  I think tun/tap should be prevented from
allowing the dev->type to be changed, unless it will make those changes
adjust the link layer headers properly as well.

Sorry.