diff mbox series

[net-next,3/8] net/mlx5e: Expand XPS cpumask to cover all online cpus

Message ID 20190125201818.9973-4-saeedm@mellanox.com
State Accepted
Delegated to: David Miller
Headers show
Series [net-next,1/8] net/mlx5e: RX, Make sure packet header does not cross page boundary | expand

Commit Message

Saeed Mahameed Jan. 25, 2019, 8:18 p.m. UTC 
From: Moshe Shemesh <moshe@mellanox.com>

Currently we have one cpu in XPS cpumask per tx queue, this is good
enough for default configuration where there is a tx queue per cpu.
However, once configuration changes to use less tx queues, part of the
cpus are not XPS-mapped and so the select queue decision falls back to
hash calculation and balancing is not guaranteed.

Expand XPS cpumask to enable using all cpus even when number of tx
queues is smaller than number of cpus.

Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |  1 +
 .../net/ethernet/mellanox/mlx5/core/en_main.c | 34 +++++++++++++++++--
 2 files changed, 33 insertions(+), 2 deletions(-)
diff mbox series

Patch

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 587f94d68dc9..3e35a90cf21a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -638,6 +638,7 @@  struct mlx5e_channel {
 	struct hwtstamp_config    *tstamp;
 	int                        ix;
 	int                        cpu;
+	cpumask_var_t              xps_cpumask;
 };
 
 struct mlx5e_channels {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 1bf547a2b905..2cf6816f1426 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1949,6 +1949,29 @@  static int mlx5e_set_tx_maxrate(struct net_device *dev, int index, u32 rate)
 	return err;
 }
 
+static int mlx5e_alloc_xps_cpumask(struct mlx5e_channel *c,
+				   struct mlx5e_params *params)
+{
+	int num_comp_vectors = mlx5_comp_vectors_count(c->mdev);
+	int irq;
+
+	if (!zalloc_cpumask_var(&c->xps_cpumask, GFP_KERNEL))
+		return -ENOMEM;
+
+	for (irq = c->ix; irq < num_comp_vectors; irq += params->num_channels) {
+		int cpu = cpumask_first(mlx5_comp_irq_get_affinity_mask(c->mdev, irq));
+
+		cpumask_set_cpu(cpu, c->xps_cpumask);
+	}
+
+	return 0;
+}
+
+static void mlx5e_free_xps_cpumask(struct mlx5e_channel *c)
+{
+	free_cpumask_var(c->xps_cpumask);
+}
+
 static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
 			      struct mlx5e_params *params,
 			      struct mlx5e_channel_param *cparam,
@@ -1981,9 +2004,12 @@  static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
 	c->num_tc   = params->num_tc;
 	c->xdp      = !!params->xdp_prog;
 	c->stats    = &priv->channel_stats[ix].ch;
-
 	c->irq_desc = irq_to_desc(irq);
 
+	err = mlx5e_alloc_xps_cpumask(c, params);
+	if (err)
+		goto err_free_channel;
+
 	netif_napi_add(netdev, &c->napi, mlx5e_napi_poll, 64);
 
 	err = mlx5e_open_cq(c, icocq_moder, &cparam->icosq_cq, &c->icosq.cq);
@@ -2066,6 +2092,9 @@  static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
 
 err_napi_del:
 	netif_napi_del(&c->napi);
+	mlx5e_free_xps_cpumask(c);
+
+err_free_channel:
 	kvfree(c);
 
 	return err;
@@ -2078,7 +2107,7 @@  static void mlx5e_activate_channel(struct mlx5e_channel *c)
 	for (tc = 0; tc < c->num_tc; tc++)
 		mlx5e_activate_txqsq(&c->sq[tc]);
 	mlx5e_activate_rq(&c->rq);
-	netif_set_xps_queue(c->netdev, get_cpu_mask(c->cpu), c->ix);
+	netif_set_xps_queue(c->netdev, c->xps_cpumask, c->ix);
 }
 
 static void mlx5e_deactivate_channel(struct mlx5e_channel *c)
@@ -2106,6 +2135,7 @@  static void mlx5e_close_channel(struct mlx5e_channel *c)
 	mlx5e_close_tx_cqs(c);
 	mlx5e_close_cq(&c->icosq.cq);
 	netif_napi_del(&c->napi);
+	mlx5e_free_xps_cpumask(c);
 
 	kvfree(c);
 }