From: Maxim Mikityanskiy <maximmi@xxxxxxxxxxxx>
[ Upstream commit f42139ba49791ab6b12443c60044872705b74a1e ]
async_icosq_lock may be taken from softirq and non-softirq contexts. It
requires protection with spin_lock_bh, otherwise a softirq may be
triggered in the middle of the critical section, and it may deadlock if
it tries to take the same lock. This patch fixes such a scenario by
using spin_lock_bh to disable softirqs on that CPU while inside the
critical section.
Fixes: 8d94b590f1e4 ("net/mlx5e: Turn XSK ICOSQ into a general asynchronous one")
Signed-off-by: Maxim Mikityanskiy <maximmi@xxxxxxxxxxxx>
Reviewed-by: Tariq Toukan <tariqt@xxxxxxxxxx>
Signed-off-by: Saeed Mahameed <saeedm@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
.../net/ethernet/mellanox/mlx5/core/en/xsk/setup.c | 4 ++--
.../net/ethernet/mellanox/mlx5/core/en/xsk/tx.c | 4 ++--
.../ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c | 14 +++++++-------
3 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c
index 55e65a438de70..fcaeb30778bc7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c
@@ -122,9 +122,9 @@ void mlx5e_activate_xsk(struct mlx5e_channel *c)
set_bit(MLX5E_RQ_STATE_ENABLED, &c->xskrq.state);
/* TX queue is created active. */
- spin_lock(&c->async_icosq_lock);
+ spin_lock_bh(&c->async_icosq_lock);
mlx5e_trigger_irq(&c->async_icosq);
- spin_unlock(&c->async_icosq_lock);
+ spin_unlock_bh(&c->async_icosq_lock);
}
void mlx5e_deactivate_xsk(struct mlx5e_channel *c)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c
index 4d892f6cecb3e..4de70cee80c0a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/tx.c
@@ -36,9 +36,9 @@ int mlx5e_xsk_wakeup(struct net_device *dev, u32 qid, u32 flags)
if (test_and_set_bit(MLX5E_SQ_STATE_PENDING_XSK_TX, &c->async_icosq.state))
return 0;
- spin_lock(&c->async_icosq_lock);
+ spin_lock_bh(&c->async_icosq_lock);
mlx5e_trigger_irq(&c->async_icosq);
- spin_unlock(&c->async_icosq_lock);
+ spin_unlock_bh(&c->async_icosq_lock);
}
return 0;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
index 6bbfcf18107d2..979ff5658a3f7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c
@@ -188,7 +188,7 @@ static int post_rx_param_wqes(struct mlx5e_channel *c,
err = 0;
sq = &c->async_icosq;
- spin_lock(&c->async_icosq_lock);
+ spin_lock_bh(&c->async_icosq_lock);
cseg = post_static_params(sq, priv_rx);
if (IS_ERR(cseg))
@@ -199,7 +199,7 @@ static int post_rx_param_wqes(struct mlx5e_channel *c,
mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, cseg);
unlock:
- spin_unlock(&c->async_icosq_lock);
+ spin_unlock_bh(&c->async_icosq_lock);
return err;
@@ -265,10 +265,10 @@ resync_post_get_progress_params(struct mlx5e_icosq *sq,
BUILD_BUG_ON(MLX5E_KTLS_GET_PROGRESS_WQEBBS != 1);
- spin_lock(&sq->channel->async_icosq_lock);
+ spin_lock_bh(&sq->channel->async_icosq_lock);
if (unlikely(!mlx5e_wqc_has_room_for(&sq->wq, sq->cc, sq->pc, 1))) {
- spin_unlock(&sq->channel->async_icosq_lock);
+ spin_unlock_bh(&sq->channel->async_icosq_lock);
err = -ENOSPC;
goto err_dma_unmap;
}
@@ -299,7 +299,7 @@ resync_post_get_progress_params(struct mlx5e_icosq *sq,
icosq_fill_wi(sq, pi, &wi);
sq->pc++;
mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, cseg);
- spin_unlock(&sq->channel->async_icosq_lock);
+ spin_unlock_bh(&sq->channel->async_icosq_lock);
return 0;
@@ -360,7 +360,7 @@ static int resync_handle_seq_match(struct mlx5e_ktls_offload_context_rx *priv_rx
err = 0;
sq = &c->async_icosq;
- spin_lock(&c->async_icosq_lock);
+ spin_lock_bh(&c->async_icosq_lock);
cseg = post_static_params(sq, priv_rx);
if (IS_ERR(cseg)) {
@@ -372,7 +372,7 @@ static int resync_handle_seq_match(struct mlx5e_ktls_offload_context_rx *priv_rx
mlx5e_notify_hw(&sq->wq, sq->pc, sq->uar_map, cseg);
priv_rx->stats->tls_resync_res_ok++;
unlock:
- spin_unlock(&c->async_icosq_lock);
+ spin_unlock_bh(&c->async_icosq_lock);
return err;
}
--
2.27.0
