linux-kernel - Re: [PATCH net-next 06/14] ipvlan: Support GSO for port -> ipvlan

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <dfad18c7-0721-486a-bd6e-75107bb54920@huawei.com>
Date: Wed, 5 Nov 2025 19:58:21 +0300
From: Dmitry Skorodumov <skorodumov.dmitry@...wei.com>
To: Eric Dumazet <edumazet@...gle.com>
CC: <netdev@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
	<andrey.bokhanko@...wei.com>, Andrew Lunn <andrew+netdev@...n.ch>, "David S.
 Miller" <davem@...emloft.net>, Jakub Kicinski <kuba@...nel.org>, Paolo Abeni
	<pabeni@...hat.com>
Subject: Re: [PATCH net-next 06/14] ipvlan: Support GSO for port -> ipvlan

On 05.11.2025 19:29, Eric Dumazet wrote:
> On Wed, Nov 5, 2025 at 8:15 AM Dmitry Skorodumov
> <skorodumov.dmitry@...wei.com> wrote:
>> If main port interface supports GSO, we need manually segment
>> the skb before forwarding it to ipvlan interface.
> Why ?
>
> I think you need to explain much more than a neutral sentence,

Ok, got it. Will resend the patch with more description: I expect there will be v4 anyway.

The reason is that this function is a protocol handler, installed on main port (with dev_add_pack()), so if main port supports GSO/checksum offload, OS will send us big/non-checksummed packets (tested with scp to IP of some child port). This packet is forwarded to child. I believe we may not expect child ipvlan-iface be prepared to RX big packet, without checksum. But I agree, that I should investigate behaviour in more details. May be I missed something and it is possible to force corresponding  TAP to somehow do this.

> Also I do not see any tests, for the whole series ?
Ok, If modules like this have some kind of unit-tests, I should study it and provide it. I haven't seen this as a common practice for most of the modules here. So far all testing is made manually (likely this should be described anyway)
>
> I have not seen the cover letter.
Cover letter was sent to netdev@...r.kernel.org. Wasn't sure that it is a good practice to add to CC every maintainer to each email of the series.
> Also you sent the series twice today :/

Well, I've sent just 000* by mistake. And immediately resent patches (as v3 in cover letter) after noticing this

Dmitry