net/mlx5e: Fix features check of IPv6 traffic
authorGal Pressman <galp@mellanox.com>
Tue, 21 Nov 2017 15:49:36 +0000 (17:49 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 2 Jan 2018 19:35:13 +0000 (20:35 +0100)
commit9424a79ec11d5aac7fcb73583afc448be1ca6848
treea88bc5551fc77c137fdcb35edf9a71d87e1d7380
parent138723912343f74b23170a2f86728f70e65e4d6b
net/mlx5e: Fix features check of IPv6 traffic

[ Upstream commit 2989ad1ec03021ee6d2193c35414f1d970a243de ]

The assumption that the next header field contains the transport
protocol is wrong for IPv6 packets with extension headers.
Instead, we should look the inner-most next header field in the buffer.
This will fix TSO offload for tunnels over IPv6 with extension headers.

Performance testing: 19.25x improvement, cool!
Measuring bandwidth of 16 threads TCP traffic over IPv6 GRE tap.
CPU: Intel(R) Xeon(R) CPU E5-2660 v2 @ 2.20GHz
NIC: Mellanox Technologies MT28800 Family [ConnectX-5 Ex]
TSO: Enabled
Before: 4,926.24  Mbps
Now   : 94,827.91 Mbps

Fixes: b3f63c3d5e2c ("net/mlx5e: Add netdev support for VXLAN tunneling")
Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/net/ethernet/mellanox/mlx5/core/en_main.c