netdev - [net-next 05/15] net/mlx5e: RX, Restrict bulk size for small Striding RQs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20220217075632.831542-6-saeed@kernel.org>
Date:   Wed, 16 Feb 2022 23:56:22 -0800
From:   Saeed Mahameed <saeed@...nel.org>
To:     "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>
Cc:     netdev@...r.kernel.org, Tariq Toukan <tariqt@...dia.com>,
        Saeed Mahameed <saeedm@...dia.com>
Subject: [net-next 05/15] net/mlx5e: RX, Restrict bulk size for small Striding RQs

From: Tariq Toukan <tariqt@...dia.com>

In RQs of type multi-packet WQE (Striding RQ), each WQE is relatively
large (typically 256KB) but their number is relatively small (8 in
default).

Re-mapping the descriptors' buffers before re-posting them is done via
UMR (User-Mode Memory Registration) operations.

On the one hand, posting UMR WQEs in bulks reduces communication overhead
with the HW and better utilizes its processing units.
On the other hand, delaying the WQE repost operations for a small RQ
(say, of 4 WQEs) might drastically hit its performance, causing packet
drops due to no receive buffer, for high or bursty incoming packets rate.

Here we restrict the bulk size for too small RQs. Effectively, with the current
constants, RQ of size 4 (minimum allowed) would have no bulking, while larger
RQs will continue working with bulks of 2.

Signed-off-by: Tariq Toukan <tariqt@...dia.com>
Signed-off-by: Saeed Mahameed <saeedm@...dia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h        | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en/params.c | 6 ++++++
 drivers/net/ethernet/mellanox/mlx5/core/en/params.h | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c   | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c     | 3 +--
 5 files changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index e29ac77e9bec..2704c7537481 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -714,6 +714,7 @@ struct mlx5e_rq {
 			u8                     umr_in_progress;
 			u8                     umr_last_bulk;
 			u8                     umr_completed;
+			u8                     min_wqe_bulk;
 			struct mlx5e_shampo_hd *shampo;
 		} mpwqe;
 	};
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
index 2d9707c024b0..0bd8698f7226 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
@@ -178,6 +178,12 @@ u8 mlx5e_mpwqe_get_log_num_strides(struct mlx5_core_dev *mdev,
 		mlx5e_mpwqe_get_log_stride_size(mdev, params, xsk);
 }
 
+u8 mlx5e_mpwqe_get_min_wqe_bulk(unsigned int wq_sz)
+{
+#define UMR_WQE_BULK (2)
+	return min_t(unsigned int, UMR_WQE_BULK, wq_sz / 2 - 1);
+}
+
 u16 mlx5e_get_rq_headroom(struct mlx5_core_dev *mdev,
 			  struct mlx5e_params *params,
 			  struct mlx5e_xsk_param *xsk)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
index 433e6967692d..47a368112e31 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
@@ -129,6 +129,7 @@ u8 mlx5e_mpwqe_get_log_stride_size(struct mlx5_core_dev *mdev,
 u8 mlx5e_mpwqe_get_log_num_strides(struct mlx5_core_dev *mdev,
 				   struct mlx5e_params *params,
 				   struct mlx5e_xsk_param *xsk);
+u8 mlx5e_mpwqe_get_min_wqe_bulk(unsigned int wq_sz);
 u16 mlx5e_get_rq_headroom(struct mlx5_core_dev *mdev,
 			  struct mlx5e_params *params,
 			  struct mlx5e_xsk_param *xsk);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index b157c7aac4ca..b2ed2f6d4a92 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -595,6 +595,7 @@ static int mlx5e_alloc_rq(struct mlx5e_params *params,
 		rq->mpwqe.log_stride_sz = mlx5e_mpwqe_get_log_stride_size(mdev, params, xsk);
 		rq->mpwqe.num_strides =
 			BIT(mlx5e_mpwqe_get_log_num_strides(mdev, params, xsk));
+		rq->mpwqe.min_wqe_bulk = mlx5e_mpwqe_get_min_wqe_bulk(wq_sz);
 
 		rq->buff.frame0_sz = (1 << rq->mpwqe.log_stride_sz);
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 3fe4f06f3e71..4cb7c7135b6a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -960,8 +960,7 @@ INDIRECT_CALLABLE_SCOPE bool mlx5e_post_rx_mpwqes(struct mlx5e_rq *rq)
 	if (unlikely(rq->mpwqe.umr_in_progress > rq->mpwqe.umr_last_bulk))
 		rq->stats->congst_umr++;
 
-#define UMR_WQE_BULK (2)
-	if (likely(missing < UMR_WQE_BULK))
+	if (likely(missing < rq->mpwqe.min_wqe_bulk))
 		return false;
 
 	if (rq->page_pool)
-- 
2.34.1