[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <jjrg5bbt5ivalb3pd64z26rs5zcdjupaoct4f4vyybcn6bjrce@m677v5ggt6t3>
Date: Wed, 7 Jun 2023 11:59:23 +0100
From: Max Tottenham <mtottenh@...mai.com>
To: Pedro Tammela <pctammela@...atatu.com>
CC: Josh Hunt <johunt@...mai.com>, <netdev@...r.kernel.org>,
Jamal Hadi Salim
<jhs@...atatu.com>,
Cong Wang <xiyou.wangcong@...il.com>, Jiri Pirko
<jiri@...nulli.us>,
Amir Vadai <amir@...ai.me>, kernel test robot
<lkp@...el.com>
Subject: Re: [PATCH v2] net/sched: act_pedit: Parse L3 Header for L4 offset
On 05/27, Pedro Tammela wrote:
> On 26/05/2023 18:54, Josh Hunt wrote:
> > On 5/26/23 7:03 AM, Pedro Tammela wrote:
> >> On 26/05/2023 10:47, Pedro Tammela wrote:
> >>>
> >>>> +
> >>>> + switch (skb->protocol) {
> >>>> + case htons(ETH_P_IP):
> >>>> + if (!pskb_may_pull(skb, sizeof(*iph) + noff))
> >>>> + goto out;
> >>>
> >>> I might have missed something but is this really needed?
> >>> https://elixir.bootlin.com/linux/latest/source/net/ipv4/ip_input.c#L456
> >>
> >> Yes this obviously happens before the mentioned function.
> >> Now I'm wondering if it's not better to use skb_header_pointer()
> >> instead...
> >
> > Can you elaborate on why you think it would be better?
> >
>
> I don't have a strong argument for one over the other and I believe it's
> fine as is.
> It just looks like 'skb_header_pointer()' is a more conservative
> approach as ithas a smaller margin for errorwhen compared to
> 'pskb_may_pull()'.
> But I shall admit that the errors conditions for 'pskb_may_pull()' are
> extreme.
>
Okay, I'm happy to re-spin a v3 addressing the above, and the other nits
you identified.
--
Max Tottenham | Senior Software Engineer
/(* Akamai Technologies
Powered by blists - more mailing lists