[PATCH net-next 09/15] net/mlx5e: Set page_pool order based on calculated page_shift
From: Tariq Toukan <tariqt@nvidia.com>
Date: 2026-02-23 20:43:59
Also in:
bpf, linux-rdma, lkml
Subsystem:
mellanox ethernet driver (mlx5e), mellanox mlx5 core vpi driver, networking drivers, the rest · Maintainers:
Saeed Mahameed, Tariq Toukan, Mark Bloch, Leon Romanovsky, Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
From: Dragos Tatulea <dtatulea@nvidia.com> Instead of unconditionally setting the page_pool to 0, calculate it from page_shift for MPWQE case. Signed-off-by: Dragos Tatulea <dtatulea@nvidia.com> Reviewed-by: Cosmin Ratiu <redacted> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index aca88fed2ac7..6344dbb6335e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c@@ -857,6 +857,7 @@ static int mlx5e_alloc_rq(struct mlx5e_params *params, { void *rqc_wq = MLX5_ADDR_OF(rqc, rq_param->rqc, wq); struct mlx5_core_dev *mdev = rq->mdev; + u32 pool_order = 0; u32 pool_size; int wq_sz; int err;
@@ -905,6 +906,7 @@ static int mlx5e_alloc_rq(struct mlx5e_params *params, pool_size = rq->mpwqe.pages_per_wqe << mlx5e_mpwqe_get_log_rq_size(mdev, params, rqo); + pool_order = rq->mpwqe.page_shift - PAGE_SHIFT; if (!mlx5e_rx_mpwqe_is_linear_skb(mdev, params, rqo) && params->xdp_prog)
@@ -960,7 +962,7 @@ static int mlx5e_alloc_rq(struct mlx5e_params *params, /* Create a page_pool and register it with rxq */ struct page_pool_params pp_params = { 0 }; - pp_params.order = 0; + pp_params.order = pool_order; pp_params.flags = PP_FLAG_DMA_MAP | PP_FLAG_DMA_SYNC_DEV; pp_params.pool_size = pool_size; pp_params.nid = node;
@@ -968,7 +970,7 @@ static int mlx5e_alloc_rq(struct mlx5e_params *params, pp_params.napi = rq->cq.napi; pp_params.netdev = rq->netdev; pp_params.dma_dir = rq->buff.map_dir; - pp_params.max_len = PAGE_SIZE; + pp_params.max_len = BIT(PAGE_SHIFT + pool_order); pp_params.queue_idx = rq->ix; /* Shampo header data split allow for unreadable netmem */
--
2.44.0