net/mlx5: Adjust log_max_qp to be 18 at most
authorMaher Sanalla <msanalla@nvidia.com>
Sun, 24 Jul 2022 08:28:21 +0000 (11:28 +0300)
committerSaeed Mahameed <saeedm@nvidia.com>
Thu, 28 Jul 2022 20:44:35 +0000 (13:44 -0700)
The cited commit limited log_max_qp to be 17 due to FW capabilities.
Recently, it turned out that there are old FW versions that supported
more than 17, so the cited commit caused a degradation.

Thus, set the maximum log_max_qp back to 18 as it was before the
cited commit.

Fixes: 7f839965b2d7 ("net/mlx5: Update log_max_qp value to be 17 at most")
Signed-off-by: Maher Sanalla <msanalla@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
drivers/net/ethernet/mellanox/mlx5/core/main.c

index c9b4e50..95f2662 100644 (file)
@@ -524,7 +524,7 @@ static int handle_hca_cap(struct mlx5_core_dev *dev, void *set_ctx)
 
        /* Check log_max_qp from HCA caps to set in current profile */
        if (prof->log_max_qp == LOG_MAX_SUPPORTED_QPS) {
-               prof->log_max_qp = min_t(u8, 17, MLX5_CAP_GEN_MAX(dev, log_max_qp));
+               prof->log_max_qp = min_t(u8, 18, MLX5_CAP_GEN_MAX(dev, log_max_qp));
        } else if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < prof->log_max_qp) {
                mlx5_core_warn(dev, "log_max_qp value in current profile is %d, changing it to HCA capability limit (%d)\n",
                               prof->log_max_qp,