On Tue, May 24, 2016 at 7:59 AM, Alexander Duyck
<alexander.du...@gmail.com> wrote:
> On Mon, May 23, 2016 at 3:48 PM, Tom Herbert <t...@herbertland.com> wrote:
>> Add SKB_GSO_TOU. In udp[64]_ufo_fragment check for SKB_GSO_TOU. If this
>> is set call skb_udp_tou_segment. skb_udp_tou_segment is very similar
>> to skb_udp_tunnel_segment except that we only need to deal with the
>> L4 headers.
>>
>> Signed-off-by: Tom Herbert <t...@herbertland.com>
>> ---
>> include/linux/skbuff.h | 2 +
>> include/net/udp.h | 2 +
>> net/ipv4/fou.c | 2 +
>> net/ipv4/ip_output.c | 2 +
>> net/ipv4/udp_offload.c | 164
>> +++++++++++++++++++++++++++++++++++++--
>> net/ipv6/inet6_connection_sock.c | 3 +
>> net/ipv6/udp_offload.c | 128 +++++++++++++++---------------
>> 7 files changed, 236 insertions(+), 67 deletions(-)
>>
>> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
>> index 65968a9..b57e484 100644
>> --- a/include/linux/skbuff.h
>> +++ b/include/linux/skbuff.h
>> @@ -482,6 +482,8 @@ enum {
>> SKB_GSO_PARTIAL = 1 << 13,
>>
>> SKB_GSO_TUNNEL_REMCSUM = 1 << 14,
>> +
>> + SKB_GSO_TOU = 1 << 15,
>> };
>>
>
> So where do you add the netdev feature bit? From what I can tell that
> was overlooked and as a result devices that support FCoE CRC will end
> up corrupting TOU frames because netif_gso_ok currently ands the two
> together.
>
An obvious omission, thanks for pointing it out.
> Also I am pretty sure we can offload this on the Intel NICs using the
> GSO partial approach as we can just stuff the UDP header into the
> space that we would use for IPv4 options or IPv6 extension headers and
> it shouldn't complain.
>
That would be cool!
> - Alex
