[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CALx6S35dtd8Pg4zvhZk2v5pNr_RqYG1Nk6zSyxFq6+8OChDpNA@mail.gmail.com>
Date: Fri, 8 Jan 2016 14:07:09 -0800
From: Tom Herbert <tom@...bertland.com>
To: Rick Jones <rick.jones2@....com>
Cc: Alexander Duyck <alexander.duyck@...il.com>,
Jesse Gross <jesse@...nel.org>,
Edward Cree <ecree@...arflare.com>,
David Miller <davem@...emloft.net>,
Netdev <netdev@...r.kernel.org>, linux-net-drivers@...arflare.com
Subject: Re: [PATCH v2 net-next 3/5] net: vxlan: enable local checksum offload
On Fri, Jan 8, 2016 at 1:36 PM, Rick Jones <rick.jones2@....com> wrote:
> On 01/08/2016 01:22 PM, Alexander Duyck wrote:
>>
>> and we don't require the remote end to validate the outer checksum.
>
>
> Doesn't the UDP specification require validating a non-zero checksum on
> receipt?
>
Yes, VXLAN RFC is broken in this regard since it allows a receiver to
ignore a non-zero checksums. In Linux we always validate non-zero
checksums.
For VXLAN, UDP checksums are not enabled for IPv4 by default, but are
enabled by default for IPv6. Checksums are required for UDP/IPv6,
however RFC6935 and RFC6936 relax this requirement a bit for tunnels,
although IMO not enough to make disabling UDP checksums for VXLAN/IPv6
the default.
Tom
> rick jones
>
Powered by blists - more mailing lists