From mboxrd@z Thu Jan  1 00:00:00 1970
From: David Miller <davem@davemloft.net>
Subject: Re: [PATCH] net: ipv4: Remove MTU check in IP_HDRINCL send flow
Date: Sat, 06 Jan 2018 20:41:36 -0500 (EST)
Message-ID: <20180106.204136.1027674698469454142.davem@davemloft.net>
References: <20180106232647.GA8203@jong.localdomain>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Cc: netdev@vger.kernel.org
To: dev.jongy@gmail.com
Return-path: <netdev-owner@vger.kernel.org>
Received: from shards.monkeyblade.net ([184.105.139.130]:50268 "EHLO
        shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751105AbeAGBli (ORCPT
        <rfc822;netdev@vger.kernel.org>); Sat, 6 Jan 2018 20:41:38 -0500
In-Reply-To: <20180106232647.GA8203@jong.localdomain>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

From: Yonatan Goldschmidt <dev.jongy@gmail.com>
Date: Sun, 7 Jan 2018 01:26:48 +0200

> IP fragmentation can be performed as expected down the stack, without touching
> irrelevant fields in the included header besides fragment offset, setting
> IP_MF and header checksum.
> If the included header has IP_DF set, EMSGSIZE is returned.
> 
> This allows users of IP_HDRINCL to have the kernel perform fragmentation.
> 
> Signed-off-by: Yonatan Goldschmidt <dev.jongy@gmail.com>

This has been brought up before and we decided that for RAW sockets
the user is responsible for fragmentation, and the kernel should not
"manage" the packet given to it in any way whatsoever.