net/mlx5e: Use a memory barrier to enforce PTP WQ xmit submission tracking occurs...

Just simply reordering the functions mlx5e_ptp_metadata_map_put and
mlx5e_ptpsq_track_metadata in the mlx5e_txwqe_complete context is not good
enough since both the compiler and CPU are free to reorder these two
functions. If reordering does occur, the issue that was supposedly fixed by
7e3f3ba97e6c ("net/mlx5e: Track xmit submission to PTP WQ after populating
metadata map") will be seen. This will lead to NULL pointer dereferences in
mlx5e_ptpsq_mark_ts_cqes_undelivered in the NAPI polling context due to the
tracking list being populated before the metadata map.

Fixes: 7e3f3ba97e6c ("net/mlx5e: Track xmit submission to PTP WQ after populating metadata map")
Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
CC: Vadim Fedorenko <vadfed@meta.com>

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
index 5c166d9d2dca62a8db671c7cb62476781a8d1b97..2fa076b23fbead06bceb6697e0ebb0238bb5be7e 100644 (file)
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -401,6 +401,8 @@ mlx5e_txwqe_complete(struct mlx5e_txqsq *sq, struct sk_buff *skb,
                mlx5e_skb_cb_hwtstamp_init(skb);
                mlx5e_ptp_metadata_map_put(&sq->ptpsq->metadata_map, skb,
                                           metadata_index);
+               /* ensure skb is put on metadata_map before tracking the index */
+               wmb();
                mlx5e_ptpsq_track_metadata(sq->ptpsq, metadata_index);
                if (!netif_tx_queue_stopped(sq->txq) &&
                    mlx5e_ptpsq_metadata_freelist_empty(sq->ptpsq)) {