diff options
| author | Saeed Mahameed | 2016-09-21 11:19:49 +0200 |
|---|---|---|
| committer | David S. Miller | 2016-09-22 08:51:41 +0200 |
| commit | 35b510e257f7516546a0a3f725f71dfbccc3f733 (patch) | |
| tree | eae31e97372200fe825241bba2481174730b3764 /drivers/net/ethernet/mellanox/mlx5/core/en.h | |
| parent | net/mlx5e: XDP TX forwarding support (diff) | |
| download | kernel-qcow2-linux-35b510e257f7516546a0a3f725f71dfbccc3f733.tar.gz kernel-qcow2-linux-35b510e257f7516546a0a3f725f71dfbccc3f733.tar.xz kernel-qcow2-linux-35b510e257f7516546a0a3f725f71dfbccc3f733.zip | |
net/mlx5e: XDP TX xmit more
Previously we rang XDP SQ doorbell on every forwarded XDP packet.
Here we introduce a xmit more like mechanism that will queue up more
than one packet into SQ (up to RX napi budget) w/o notifying the hardware.
Once RX napi budget is consumed and we exit napi RX loop, we will
flush (doorbell) all XDP looped packets in case there are such.
XDP forward packet rate:
Comparing XDP with and w/o xmit more (bulk transmit):
RX Cores XDP TX XDP TX (xmit more)
---------------------------------------------------
1 6.5Mpps 12.4Mpps
2 13.2Mpps 24.2Mpps
4 25.2Mpps 36.3Mpps*
8 36.3Mpps* 36.3Mpps*
*My xmitter was limited to 36.3Mpps, so it is the bottleneck.
It seems that receive side can handle more.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'drivers/net/ethernet/mellanox/mlx5/core/en.h')
| -rw-r--r-- | drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h index 82eededfc92a..346015407b70 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h @@ -433,6 +433,7 @@ struct mlx5e_sq { struct { struct mlx5e_sq_wqe_info *wqe_info; struct mlx5e_dma_info *di; + bool doorbell; } xdp; } db; |