From: Shay Drory shayd@nvidia.com
[ Upstream commit f88c4876347400a577598e06f1b230a7b19ee0e9 ]
When setting number of completion EQs of the SF, consider number of online CPUs. Without this consideration, when number of online cpus are less than 8, unnecessary 8 completion EQs are allocated.
Fixes: c36326d38d93 ("net/mlx5: Round-Robin EQs over IRQs") Signed-off-by: Shay Drory shayd@nvidia.com Reviewed-by: Parav Pandit parav@nvidia.com Signed-off-by: Saeed Mahameed saeedm@nvidia.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c b/drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c index 49e6f5003991..d9345c9ebbff 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c @@ -610,8 +610,9 @@ void mlx5_irq_table_destroy(struct mlx5_core_dev *dev) int mlx5_irq_table_get_sfs_vec(struct mlx5_irq_table *table) { if (table->sf_comp_pool) - return table->sf_comp_pool->xa_num_irqs.max - - table->sf_comp_pool->xa_num_irqs.min + 1; + return min_t(int, num_online_cpus(), + table->sf_comp_pool->xa_num_irqs.max - + table->sf_comp_pool->xa_num_irqs.min + 1); else return mlx5_irq_table_get_num_comp(table); }