net/mlx5: remove redundant operations in NEON Rx

[ upstream commit ff6fcd4 ]

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: 570acdb ("net/mlx5: add vectorized Rx/Tx burst for ARM")

Signed-off-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>

drivers/net/mlx5/mlx5_rxtx_vec_neon.h

@@ -671,16 +671,15 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n,
 		comp_idx = __builtin_clzl(vget_lane_u64(vreinterpret_u64_u16(
 					  comp_mask), 0)) /
 					  (sizeof(uint16_t) * 8);
-		/* D.6 mask out entries after the compressed CQE. */
-		mask = vcreate_u16(comp_idx < MLX5_VPMD_DESCS_PER_LOOP ?
-				   -1UL >> (comp_idx * sizeof(uint16_t) * 8) :
-				   0);
-		invalid_mask = vorr_u16(invalid_mask, mask);
+		invalid_mask = vorr_u16(invalid_mask, comp_mask);
 		/* D.7 count non-compressed valid CQEs. */
 		n = __builtin_clzl(vget_lane_u64(vreinterpret_u64_u16(
 				   invalid_mask), 0)) / (sizeof(uint16_t) * 8);
 		nocmp_n += n;
-		/* D.2 get the final invalid mask. */
+		/*
+		 * D.2 mask out entries after the compressed CQE.
+		 *     get the final invalid mask.
+		 */
 		mask = vcreate_u16(n < MLX5_VPMD_DESCS_PER_LOOP ?
 				   -1UL >> (n * sizeof(uint16_t) * 8) : 0);
 		invalid_mask = vorr_u16(invalid_mask, mask);