Received-SPF: pass (google.com: domain of linux-kernel+bounces-110036-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Feedback-ID: i85e1472c:Fastmail
Content-Type: text/plain;
	charset=utf-8
Precedence: bulk
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3774.300.61.1.2\))
Subject: Re: [EXTERNAL] [PATCH] netpoll: support sending over raw IP
 interfaces
From: Mark <mark@yotsuba.nl>
In-Reply-To: <MWHPR1801MB191845C94DD80E4CD6C8C962D32D2@MWHPR1801MB1918.namprd18.prod.outlook.com>
Date: Thu, 21 Mar 2024 13:33:17 +0100
Cc: "netdev@vger.kernel.org" <netdev@vger.kernel.org>,
 Hans de Goede <hdegoede@redhat.com>,
 Eric Dumazet <edumazet@google.com>,
 Jakub Kicinski <kuba@kernel.org>,
 Breno Leitao <leitao@debian.org>,
 Ingo Molnar <mingo@redhat.com>,
 "David S. Miller" <davem@davemloft.net>,
 Paolo Abeni <pabeni@redhat.com>,
 "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <4ACBB9C7-EFF5-47CA-9A8A-114209AE7953@yotsuba.nl>
References: <20240313124613.51399-1-mark@yotsuba.nl>
 <20240313133602.GA1263314@maili.marvell.com>
 <7C42FC4B-D803-4194-8FBB-19A432D37124@yotsuba.nl>
 <MWHPR1801MB1918F15413BA4766F29A8581D3292@MWHPR1801MB1918.namprd18.prod.outlook.com>
 <57AE2A31-257C-4702-A571-C590A5DD234A@yotsuba.nl>
 <MWHPR1801MB191845C94DD80E4CD6C8C962D32D2@MWHPR1801MB1918.namprd18.prod.outlook.com>
To: Ratheesh Kannoth <rkannoth@marvell.com>

Hi Ratheesh,

> Op 18 mrt 6 Reiwa, om 15:06 heeft Ratheesh Kannoth =
<rkannoth@marvell.com> het volgende geschreven:
>=20
>> [=E2=80=A6]
> Is below network topology possible ? =20
> Netpoll()- ------> netdev A ----> raw interface=20
> Where netdev A's  netdev->header_ops !=3D NULL

I believe so, this is not uncommon in tunnel devices like gretap.
However, those fully encapsulate the link layer header in the packet
to the lower interface. I am not aware of a interface driver that =
removes
a header upon xmit, so to speak. However, I have just posted a v2 that
instead uses the documented `dev_has_header()` API, which seems to fit =
the
check exactly, here:

https://lore.kernel.org/netdev/20240321122003.20089-1-mark@yotsuba.nl/T/

I hope this change manages to mitigate your concerns. :-)

> [=E2=80=A6]

Thanks and regards,
Mark