netdev - Re: GRO disabled with IPv4 options

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

Message-ID: <CAF=yD-K3w_PaaDNh12v6yPLO2SHq_32jP2weV38ZGe531LWBtw@mail.gmail.com>
Date:   Fri, 17 Nov 2017 08:51:57 -0500
From:   Willem de Bruijn <willemdebruijn.kernel@...il.com>
To:     Tom Herbert <tom@...bertland.com>
Cc:     Herbert Xu <herbert@...dor.apana.org.au>,
        Cristian Klein <cklein@...umu.se>,
        Linux Kernel Network Developers <netdev@...r.kernel.org>,
        Ahmed Ali-Eldin <ahmeda@...umu.se>
Subject: Re: GRO disabled with IPv4 options

On Thu, Nov 16, 2017 at 5:47 PM, Tom Herbert <tom@...bertland.com> wrote:
> On Thu, Nov 16, 2017 at 1:40 PM, Herbert Xu <herbert@...dor.apana.org.au> wrote:
>> On Thu, Nov 16, 2017 at 04:12:43PM +0100, Cristian Klein wrote:
>>>
>>> Does somebody know the rationale for this? Is it because IPv4
>>> options are rarely used, hence implementing GRO in that case does
>>> not pay off or are there some caveats? Specifically would it make
>>
>> Precisely.  GRO is about optimising for the common case.  At the
>> time there was no impetus to support IP options.
>>
>>> sense to do GRO when the IPv4 options are byte-identical in
>>> consecutive packets?
>>
>> Yes there is no reason why we can't do this.  As long as it doesn't
>> penalise the non-IP-option case too much.
>>
> Of course it would also be nice to have GRO support for various IPv6
> extension headers, at this point we're more likely to see those rather
> than IP options in real deployment!

ipv6_gro_receive already pulls common ones with ipv6_gso_pull_exthdrs.
And to add a counterpoint: GRO has to be resilient to malformed packets,
so there is value in keeping it simple and limited to the common case.