[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <d8262b12-3ac9-8608-fc6f-d48c33a4225e@linux.alibaba.com>
Date: Wed, 29 Nov 2023 23:02:01 +0800
From: "D. Wythe" <alibuda@...ux.alibaba.com>
To: Florian Westphal <fw@...len.de>
Cc: pablo@...filter.org, kadlec@...filter.org, bpf@...r.kernel.org,
linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
coreteam@...filter.org, netfilter-devel@...r.kernel.org,
davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
pabeni@...hat.com, ast@...nel.org
Subject: Re: [PATCH net] net/netfilter: bpf: avoid leakage of skb
On 11/29/23 10:47 PM, Florian Westphal wrote:
> D. Wythe <alibuda@...ux.alibaba.com> wrote:
>> And my origin intention was to allow ebpf progs to return NF_STOLEN, we are
>> trying to modify some netfilter modules via ebpf,
>> and some scenarios require the use of NF_STOLEN, but from your description,
> NF_STOLEN can only be supported via a trusted helper, as least as far as
> I understand.
>
> Otherwise verifier would have to guarantee that any branch that returns
> NF_STOLEN has released the skb, or passed it to a function that will
> release the skb in the near future.
Thank you very much for your help. I now understand the difficulty here.
The verifier cannot determine whether the consume_skb() was executed or not,
when the return value goes to NF_STOLEN.
We may use NF_DROP at first, it won't be make much difference for us now.
Also, do you have any plans to support this helper?
Best wishes,
D. Wythe
Powered by blists - more mailing lists