git: 89491b1edb34 - main - mlx5en: stop arbitrary limiting max wqe size
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Thu, 13 Mar 2025 16:03:23 UTC
The branch main has been updated by kib: URL: https://cgit.FreeBSD.org/src/commit/?id=89491b1edb3438ebce84e17fdb6c1acab35df601 commit 89491b1edb3438ebce84e17fdb6c1acab35df601 Author: Konstantin Belousov <kib@FreeBSD.org> AuthorDate: 2025-02-27 01:54:31 +0000 Commit: Konstantin Belousov <kib@FreeBSD.org> CommitDate: 2025-03-13 16:01:17 +0000 mlx5en: stop arbitrary limiting max wqe size Since the times the driver accepts s/g receive buffers, there is no sense in trying to use pre-existing mbuf clusters sizes. The only possible optimization is to use full page size if wqe size is greater than MCLBYTES. Reviewed by: Ariel Ehrenberg <aehrenberg@nvidia.com>, Slava Shwartsman <slavash@nvidia.com> Sponsored by: NVidia networking MFC after: 1 week --- sys/dev/mlx5/mlx5_en/mlx5_en_main.c | 12 +----------- 1 file changed, 1 insertion(+), 11 deletions(-) diff --git a/sys/dev/mlx5/mlx5_en/mlx5_en_main.c b/sys/dev/mlx5/mlx5_en/mlx5_en_main.c index d69d2e433c9c..8b2acddb869a 100644 --- a/sys/dev/mlx5/mlx5_en/mlx5_en_main.c +++ b/sys/dev/mlx5/mlx5_en/mlx5_en_main.c @@ -2321,17 +2321,7 @@ mlx5e_get_wqe_sz(struct mlx5e_priv *priv, u32 *wqe_sz, u32 *nsegs) r = priv->params.hw_lro_en ? priv->params.lro_wqe_sz : MLX5E_SW2MB_MTU(if_getmtu(priv->ifp)); - if (r > MJUM16BYTES) - return (-ENOMEM); - - if (r > MJUM9BYTES) - r = MJUM16BYTES; - else if (r > MJUMPAGESIZE) - r = MJUM9BYTES; - else if (r > MCLBYTES) - r = MJUMPAGESIZE; - else - r = MCLBYTES; + r = r > MCLBYTES ? MJUMPAGESIZE : MCLBYTES; /* * n + 1 must be a power of two, because stride size must be.