netdev - [net-next 10/15] net/mlx5: Handle SF IRQ request in the absence of SF IRQ pool

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: <20230807175642.20834-11-saeed@kernel.org>
Date: Mon,  7 Aug 2023 10:56:37 -0700
From: Saeed Mahameed <saeed@...nel.org>
To: "David S. Miller" <davem@...emloft.net>,
	Jakub Kicinski <kuba@...nel.org>,
	Paolo Abeni <pabeni@...hat.com>,
	Eric Dumazet <edumazet@...gle.com>
Cc: Saeed Mahameed <saeedm@...dia.com>,
	netdev@...r.kernel.org,
	Tariq Toukan <tariqt@...dia.com>,
	Maher Sanalla <msanalla@...dia.com>,
	Shay Drory <shayd@...dia.com>,
	Moshe Shemesh <moshe@...dia.com>
Subject: [net-next 10/15] net/mlx5: Handle SF IRQ request in the absence of SF IRQ pool

From: Maher Sanalla <msanalla@...dia.com>

In case the SF IRQ pool is not available due to setup limitations,
SF currently relies on the already allocated PF IRQs to fulfill
its IRQ vector requests.

However, with the dynamic EQ allocation introduced in the next patch,
it is possible that not all IRQs of PF will be allocated after the driver
is loaded. In such case, if a SF requests a completion IRQ without having
its own independent IRQ pool, SF will lack a PF IRQ to utilize.

To address this scenario, allocate an IRQ for the SF from the PF's IRQ pool
on demand. The new IRQ will be shared between the SF and it's PF.

Signed-off-by: Maher Sanalla <msanalla@...dia.com>
Reviewed-by: Shay Drory <shayd@...dia.com>
Reviewed-by: Moshe Shemesh <moshe@...dia.com>
Signed-off-by: Saeed Mahameed <saeedm@...dia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/eq.c  | 24 +++++++++++++++++--
 .../mellanox/mlx5/core/irq_affinity.c         | 12 ++++------
 2 files changed, 26 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
index 6272962ea077..6e6e0a1c12b5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
@@ -850,14 +850,29 @@ static int mlx5_cpumask_default_spread(int numa_node, int index)
 	return found_cpu;
 }
 
+static struct cpu_rmap *mlx5_eq_table_get_pci_rmap(struct mlx5_core_dev *dev)
+{
+#ifdef CONFIG_RFS_ACCEL
+#ifdef CONFIG_MLX5_SF
+	if (mlx5_core_is_sf(dev))
+		return dev->priv.parent_mdev->priv.eq_table->rmap;
+#endif
+	return dev->priv.eq_table->rmap;
+#else
+	return NULL;
+#endif
+}
+
 static int comp_irq_request_pci(struct mlx5_core_dev *dev, u16 vecidx)
 {
 	struct mlx5_eq_table *table = dev->priv.eq_table;
+	struct cpu_rmap *rmap;
 	struct mlx5_irq *irq;
 	int cpu;
 
+	rmap = mlx5_eq_table_get_pci_rmap(dev);
 	cpu = mlx5_cpumask_default_spread(dev->priv.numa_node, vecidx);
-	irq = mlx5_irq_request_vector(dev, cpu, vecidx, &table->rmap);
+	irq = mlx5_irq_request_vector(dev, cpu, vecidx, &rmap);
 	if (IS_ERR(irq))
 		return PTR_ERR(irq);
 
@@ -883,8 +898,13 @@ static int comp_irq_request_sf(struct mlx5_core_dev *dev, u16 vecidx)
 	struct mlx5_irq *irq;
 
 	irq = mlx5_irq_affinity_irq_request_auto(dev, &table->used_cpus, vecidx);
-	if (IS_ERR(irq))
+	if (IS_ERR(irq)) {
+		/* In case SF irq pool does not exist, fallback to the PF irqs*/
+		if (PTR_ERR(irq) == -ENOENT)
+			return comp_irq_request_pci(dev, vecidx);
+
 		return PTR_ERR(irq);
+	}
 
 	return xa_err(xa_store(&table->comp_irqs, vecidx, irq, GFP_KERNEL));
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/irq_affinity.c b/drivers/net/ethernet/mellanox/mlx5/core/irq_affinity.c
index ed51800f9f67..047d5fed5f89 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/irq_affinity.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/irq_affinity.c
@@ -191,17 +191,13 @@ struct mlx5_irq *mlx5_irq_affinity_irq_request_auto(struct mlx5_core_dev *dev,
 	struct irq_affinity_desc af_desc = {};
 	struct mlx5_irq *irq;
 
+	if (!mlx5_irq_pool_is_sf_pool(pool))
+		return ERR_PTR(-ENOENT);
+
 	af_desc.is_managed = 1;
 	cpumask_copy(&af_desc.mask, cpu_online_mask);
 	cpumask_andnot(&af_desc.mask, &af_desc.mask, used_cpus);
-	if (mlx5_irq_pool_is_sf_pool(pool))
-		irq = mlx5_irq_affinity_request(pool, &af_desc);
-	else
-		/* In case SF pool doesn't exists, fallback to the PF IRQs.
-		 * The PF IRQs are already allocated and binded to CPU
-		 * at this point. Hence, only an index is needed.
-		 */
-		irq = mlx5_irq_request(dev, vecidx, NULL, NULL);
+	irq = mlx5_irq_affinity_request(pool, &af_desc);
 
 	if (IS_ERR(irq))
 		return irq;
-- 
2.41.0