diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 1f3891fde2eb171be76294b45aefcfab57f9990e..d365551d2f10aa377458320bc6a75a52a594a4db 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -1933,6 +1933,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev)
 
 	hash_init(esw->offloads.encap_tbl);
 	hash_init(esw->offloads.mod_hdr_tbl);
+	atomic64_set(&esw->offloads.num_flows, 0);
 	mutex_init(&esw->state_lock);
 
 	mlx5_esw_for_all_vports(esw, i, vport) {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index a38e8a3c7c9a3b71b6a51c9f440b25878508340f..60f0c62b447b9a4a9af2dcfa90754f69b3467a0e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -35,6 +35,7 @@
 
 #include <linux/if_ether.h>
 #include <linux/if_link.h>
+#include <linux/atomic.h>
 #include <net/devlink.h>
 #include <linux/mlx5/device.h>
 #include <linux/mlx5/eswitch.h>
@@ -179,7 +180,7 @@ struct mlx5_esw_offload {
 	struct mutex termtbl_mutex; /* protects termtbl hash */
 	const struct mlx5_eswitch_rep_ops *rep_ops[NUM_REP_TYPES];
 	u8 inline_mode;
-	u64 num_flows;
+	atomic64_t num_flows;
 	enum devlink_eswitch_encap_mode encap;
 };
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 089ae4d48a82dda15e47c49530f9a4e513577acb..244ad1893691242e9efaaf2b7cb0efc9d2d14546 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -233,7 +233,7 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
 	if (IS_ERR(rule))
 		goto err_add_rule;
 	else
-		esw->offloads.num_flows++;
+		atomic64_inc(&esw->offloads.num_flows);
 
 	return rule;
 
@@ -298,7 +298,7 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
 	if (IS_ERR(rule))
 		goto add_err;
 
-	esw->offloads.num_flows++;
+	atomic64_inc(&esw->offloads.num_flows);
 
 	return rule;
 add_err:
@@ -326,7 +326,7 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw,
 			mlx5_eswitch_termtbl_put(esw, attr->dests[i].termtbl);
 	}
 
-	esw->offloads.num_flows--;
+	atomic64_dec(&esw->offloads.num_flows);
 
 	if (fwd_rule)  {
 		esw_put_prio_table(esw, attr->chain, attr->prio, 1);
@@ -2349,7 +2349,7 @@ int mlx5_devlink_eswitch_inline_mode_set(struct devlink *devlink, u8 mode,
 		break;
 	}
 
-	if (esw->offloads.num_flows > 0) {
+	if (atomic64_read(&esw->offloads.num_flows) > 0) {
 		NL_SET_ERR_MSG_MOD(extack,
 				   "Can't set inline mode when flows are configured");
 		return -EOPNOTSUPP;
@@ -2459,7 +2459,7 @@ int mlx5_devlink_eswitch_encap_mode_set(struct devlink *devlink,
 	if (esw->offloads.encap == encap)
 		return 0;
 
-	if (esw->offloads.num_flows > 0) {
+	if (atomic64_read(&esw->offloads.num_flows) > 0) {
 		NL_SET_ERR_MSG_MOD(extack,
 				   "Can't set encapsulation when flows are configured");
 		return -EOPNOTSUPP;