netdev - Re: [PATCH v5 bpf 1/2] xdp: introduce xdp_convert_skb_to

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: <49edb9fa-56be-4853-aced-429f5b40b4be@kernel.org>
Date: Wed, 5 Nov 2025 19:39:46 +0100
From: Jesper Dangaard Brouer <hawk@...nel.org>
To: Maciej Fijalkowski <maciej.fijalkowski@...el.com>,
 Jakub Kicinski <kuba@...nel.org>
Cc: bpf@...r.kernel.org, ast@...nel.org, daniel@...earbox.net,
 netdev@...r.kernel.org, magnus.karlsson@...el.com,
 aleksander.lobakin@...el.com, ilias.apalodimas@...aro.org, toke@...hat.com,
 lorenzo@...nel.org, syzbot+ff145014d6b0ce64a173@...kaller.appspotmail.com,
 Ihor Solodrai <ihor.solodrai@...ux.dev>, Octavian Purdila <tavip@...gle.com>
Subject: Re: [PATCH v5 bpf 1/2] xdp: introduce xdp_convert_skb_to_buff()



On 03/11/2025 13.16, Maciej Fijalkowski wrote:
> On Fri, Oct 31, 2025 at 11:49:52AM -0700, Jakub Kicinski wrote:
>> On Fri, 31 Oct 2025 12:37:37 +0100 Maciej Fijalkowski wrote:
>>>>> would be fine for you? Plus AI reviewer has kicked me in the nuts on veth
>>>>> patch so have to send v6 anyways.
>>>>
>>>> The veth side unfortunately needs more work than Mr Robot points out.
>>>> For some reason veth tries to turn skb into an xdp_frame..
>>>
>>> That is beyond the scope of the fix that I started doing as you're
>>> undermining overall XDP support in veth, IMHO.
>>>
>>> I can follow up on this on some undefined future but right now I will
>>> have to switch to some other work.
>>>
>>> If you disagree and insist on addressing skb->xdp_frame in veth within
>>> this patchset then I'm sorry but I will have to postpone my activities
>>> here.
>>
>> Yeah, I understand. A lot of the skb<>XDP integration is a steaming
>> pile IMO, as mentioned elsewhere. I'd like to keep the core clean
>> tho, so if there's some corner cases in veth after your changes
>> I'll live. But I'm worried that the bugs in veth will make you
>> want to preserve the conditional in xdp_convert_skb_to_buff() :(
> 
> Probably the conditional would cover these corner cases, but I am not
> insisting on it. skb_pp_cow_data() is called under certain circumstances
> in veth but I have never seen a case on my side where it was skipped,
> mostly because of headroom size being too small.

This reminds me that we should change/increase veth
net_device.needed_headroom when there is an XDP prog attached.
As you said, today all SKB will basically get mem reallocated.

--Jesper