net/mlx5: remove redundant operations in NEON Rx
authorRuifeng Wang <ruifeng.wang@arm.com>
Wed, 7 Jul 2021 09:03:06 +0000 (17:03 +0800)
committerRaslan Darawsheh <rasland@nvidia.com>
Thu, 15 Jul 2021 13:16:26 +0000 (15:16 +0200)
commitff6fcd415f4b2be7572abffa5284978c176d3de4
tree537c76a1da95990b985beb9b12523941e5815bc9
parentdbd8e4102d49ba89f5b4ae7a9557e48693da4609
net/mlx5: remove redundant operations in NEON Rx

Mask of entries after the compressed CQE is covered by invalid mask of
non-compressed valid CQEs. Hence remove redundant calculation on mask.
The change showed slight performance uplift on N1SDP.

Fixes: 570acdb1da8a ("net/mlx5: add vectorized Rx/Tx burst for ARM")
Cc: stable@dpdk.org
Signed-off-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
drivers/net/mlx5/mlx5_rxtx_vec_neon.h