Merge tag 'mlx5-fixes-2021-12-22' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2021-12-22

This series provides bug fixes to mlx5 driver.

* tag 'mlx5-fixes-2021-12-22' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5: Fix some error handling paths in 'mlx5e_tc_add_fdb_flow()'
  net/mlx5e: Delete forward rule for ct or sample action
  net/mlx5e: Fix ICOSQ recovery flow for XSK
  net/mlx5e: Fix interoperability between XSK and ICOSQ recovery flow
  net/mlx5e: Fix skb memory leak when TC classifier action offloads are disabled
  net/mlx5e: Wrap the tx reporter dump callback to extract the sq
  net/mlx5: Fix tc max supported prio for nic mode
  net/mlx5: Fix SF health recovery flow
  net/mlx5: Fix error print in case of IRQ request failed
  net/mlx5: Use first online CPU instead of hard coded CPU
  net/mlx5: DR, Fix querying eswitch manager vport for ECPF
  net/mlx5: DR, Fix NULL vs IS_ERR checking in dr_domain_init_resources
====================

Link: https://lore.kernel.org/r/20211223190441.153012-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Author: kuba-moo

drivers/net/ethernet/mellanox/mlx5/core/en.h

@@ -783,6 +783,8 @@ struct mlx5e_channel {
 	DECLARE_BITMAP(state, MLX5E_CHANNEL_NUM_STATES);
 	int                        ix;
 	int                        cpu;
+	/* Sync between icosq recovery and XSK enable/disable. */
+	struct mutex               icosq_recovery_lock;
 };
 
 struct mlx5e_ptp;
@@ -1014,9 +1016,6 @@ int mlx5e_create_rq(struct mlx5e_rq *rq, struct mlx5e_rq_param *param);
 void mlx5e_destroy_rq(struct mlx5e_rq *rq);
 
 struct mlx5e_sq_param;
-int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
-		     struct mlx5e_sq_param *param, struct mlx5e_icosq *sq);
-void mlx5e_close_icosq(struct mlx5e_icosq *sq);
 int mlx5e_open_xdpsq(struct mlx5e_channel *c, struct mlx5e_params *params,
 		     struct mlx5e_sq_param *param, struct xsk_buff_pool *xsk_pool,
 		     struct mlx5e_xdpsq *sq, bool is_redirect);

drivers/net/ethernet/mellanox/mlx5/core/en/health.h

@@ -30,6 +30,8 @@ void mlx5e_reporter_rx_destroy(struct mlx5e_priv *priv);
 void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq);
 void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq);
 void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq);
+void mlx5e_reporter_icosq_suspend_recovery(struct mlx5e_channel *c);
+void mlx5e_reporter_icosq_resume_recovery(struct mlx5e_channel *c);
 
 #define MLX5E_REPORTER_PER_Q_MAX_LEN 256
 

drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.h

@@ -66,7 +66,7 @@ mlx5e_rep_setup_tc(struct net_device *dev, enum tc_setup_type type,
 
 static inline void
 mlx5e_rep_tc_receive(struct mlx5_cqe64 *cqe, struct mlx5e_rq *rq,
-		     struct sk_buff *skb) {}
+		     struct sk_buff *skb) { napi_gro_receive(rq->cq.napi, skb); }
 
 #endif /* CONFIG_MLX5_CLS_ACT */
 

drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c

@@ -62,6 +62,7 @@ static void mlx5e_reset_icosq_cc_pc(struct mlx5e_icosq *icosq)
 
 static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx)
 {
+	struct mlx5e_rq *xskrq = NULL;
 	struct mlx5_core_dev *mdev;
 	struct mlx5e_icosq *icosq;
 	struct net_device *dev;
@@ -70,7 +71,13 @@ static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx)
 	int err;
 
 	icosq = ctx;
+
+	mutex_lock(&icosq->channel->icosq_recovery_lock);
+
+	/* mlx5e_close_rq cancels this work before RQ and ICOSQ are killed. */
 	rq = &icosq->channel->rq;
+	if (test_bit(MLX5E_RQ_STATE_ENABLED, &icosq->channel->xskrq.state))
+		xskrq = &icosq->channel->xskrq;
 	mdev = icosq->channel->mdev;
 	dev = icosq->channel->netdev;
 	err = mlx5_core_query_sq_state(mdev, icosq->sqn, &state);
@@ -84,6 +91,9 @@ static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx)
 		goto out;
 
 	mlx5e_deactivate_rq(rq);
+	if (xskrq)
+		mlx5e_deactivate_rq(xskrq);
+
 	err = mlx5e_wait_for_icosq_flush(icosq);
 	if (err)
 		goto out;
@@ -97,15 +107,28 @@ static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx)
 		goto out;
 
 	mlx5e_reset_icosq_cc_pc(icosq);
+
 	mlx5e_free_rx_in_progress_descs(rq);
+	if (xskrq)
+		mlx5e_free_rx_in_progress_descs(xskrq);
+
 	clear_bit(MLX5E_SQ_STATE_RECOVERING, &icosq->state);
 	mlx5e_activate_icosq(icosq);
-	mlx5e_activate_rq(rq);
 
+	mlx5e_activate_rq(rq);
 	rq->stats->recover++;
+
+	if (xskrq) {
+		mlx5e_activate_rq(xskrq);
+		xskrq->stats->recover++;
+	}
+
+	mutex_unlock(&icosq->channel->icosq_recovery_lock);
+
 	return 0;
 out:
 	clear_bit(MLX5E_SQ_STATE_RECOVERING, &icosq->state);
+	mutex_unlock(&icosq->channel->icosq_recovery_lock);
 	return err;
 }
 
@@ -706,6 +729,16 @@ void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq)
 	mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
 }
 
+void mlx5e_reporter_icosq_suspend_recovery(struct mlx5e_channel *c)
+{
+	mutex_lock(&c->icosq_recovery_lock);
+}
+
+void mlx5e_reporter_icosq_resume_recovery(struct mlx5e_channel *c)
+{
+	mutex_unlock(&c->icosq_recovery_lock);
+}
+
 static const struct devlink_health_reporter_ops mlx5_rx_reporter_ops = {
 	.name = "rx",
 	.recover = mlx5e_rx_reporter_recover,

drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c

@@ -466,6 +466,14 @@ static int mlx5e_tx_reporter_dump_sq(struct mlx5e_priv *priv, struct devlink_fms
 	return mlx5e_health_fmsg_named_obj_nest_end(fmsg);
 }
 
+static int mlx5e_tx_reporter_timeout_dump(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg,
+					  void *ctx)
+{
+	struct mlx5e_tx_timeout_ctx *to_ctx = ctx;
+
+	return mlx5e_tx_reporter_dump_sq(priv, fmsg, to_ctx->sq);
+}
+
 static int mlx5e_tx_reporter_dump_all_sqs(struct mlx5e_priv *priv,
 					  struct devlink_fmsg *fmsg)
 {
@@ -561,7 +569,7 @@ int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq)
 	to_ctx.sq = sq;
 	err_ctx.ctx = &to_ctx;
 	err_ctx.recover = mlx5e_tx_reporter_timeout_recover;
-	err_ctx.dump = mlx5e_tx_reporter_dump_sq;
+	err_ctx.dump = mlx5e_tx_reporter_timeout_dump;
 	snprintf(err_str, sizeof(err_str),
 		 "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u",
 		 sq->ch_ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,

drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c

@@ -4,6 +4,7 @@
 #include "setup.h"
 #include "en/params.h"
 #include "en/txrx.h"
+#include "en/health.h"
 
 /* It matches XDP_UMEM_MIN_CHUNK_SIZE, but as this constant is private and may
  * change unexpectedly, and mlx5e has a minimum valid stride size for striding
@@ -170,7 +171,13 @@ void mlx5e_close_xsk(struct mlx5e_channel *c)
 
 void mlx5e_activate_xsk(struct mlx5e_channel *c)
 {
+	/* ICOSQ recovery deactivates RQs. Suspend the recovery to avoid
+	 * activating XSKRQ in the middle of recovery.
+	 */
+	mlx5e_reporter_icosq_suspend_recovery(c);
 	set_bit(MLX5E_RQ_STATE_ENABLED, &c->xskrq.state);
+	mlx5e_reporter_icosq_resume_recovery(c);
+
 	/* TX queue is created active. */
 
 	spin_lock_bh(&c->async_icosq_lock);
@@ -180,6 +187,13 @@ void mlx5e_activate_xsk(struct mlx5e_channel *c)
 
 void mlx5e_deactivate_xsk(struct mlx5e_channel *c)
 {
-	mlx5e_deactivate_rq(&c->xskrq);
+	/* ICOSQ recovery may reactivate XSKRQ if clear_bit is called in the
+	 * middle of recovery. Suspend the recovery to avoid it.
+	 */
+	mlx5e_reporter_icosq_suspend_recovery(c);
+	clear_bit(MLX5E_RQ_STATE_ENABLED, &c->xskrq.state);
+	mlx5e_reporter_icosq_resume_recovery(c);
+	synchronize_net(); /* Sync with NAPI to prevent mlx5e_post_rx_wqes. */
+
 	/* TX queue is disabled on close. */
 }

drivers/net/ethernet/mellanox/mlx5/core/en_main.c

@@ -1087,8 +1087,6 @@ void mlx5e_deactivate_rq(struct mlx5e_rq *rq)
 void mlx5e_close_rq(struct mlx5e_rq *rq)
 {
 	cancel_work_sync(&rq->dim.work);
-	if (rq->icosq)
-		cancel_work_sync(&rq->icosq->recover_work);
 	cancel_work_sync(&rq->recover_work);
 	mlx5e_destroy_rq(rq);
 	mlx5e_free_rx_descs(rq);
@@ -1216,9 +1214,20 @@ static void mlx5e_icosq_err_cqe_work(struct work_struct *recover_work)
 	mlx5e_reporter_icosq_cqe_err(sq);
 }
 
+static void mlx5e_async_icosq_err_cqe_work(struct work_struct *recover_work)
+{
+	struct mlx5e_icosq *sq = container_of(recover_work, struct mlx5e_icosq,
+					      recover_work);
+
+	/* Not implemented yet. */
+
+	netdev_warn(sq->channel->netdev, "async_icosq recovery is not implemented\n");
+}
+
 static int mlx5e_alloc_icosq(struct mlx5e_channel *c,
 			     struct mlx5e_sq_param *param,
-			     struct mlx5e_icosq *sq)
+			     struct mlx5e_icosq *sq,
+			     work_func_t recover_work_func)
 {
 	void *sqc_wq               = MLX5_ADDR_OF(sqc, param->sqc, wq);
 	struct mlx5_core_dev *mdev = c->mdev;
@@ -1239,7 +1248,7 @@ static int mlx5e_alloc_icosq(struct mlx5e_channel *c,
 	if (err)
 		goto err_sq_wq_destroy;
 
-	INIT_WORK(&sq->recover_work, mlx5e_icosq_err_cqe_work);
+	INIT_WORK(&sq->recover_work, recover_work_func);
 
 	return 0;
 
@@ -1575,13 +1584,14 @@ void mlx5e_tx_err_cqe_work(struct work_struct *recover_work)
 	mlx5e_reporter_tx_err_cqe(sq);
 }
 
-int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
-		     struct mlx5e_sq_param *param, struct mlx5e_icosq *sq)
+static int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
+			    struct mlx5e_sq_param *param, struct mlx5e_icosq *sq,
+			    work_func_t recover_work_func)
 {
 	struct mlx5e_create_sq_param csp = {};
 	int err;
 
-	err = mlx5e_alloc_icosq(c, param, sq);
+	err = mlx5e_alloc_icosq(c, param, sq, recover_work_func);
 	if (err)
 		return err;
 
@@ -1620,7 +1630,7 @@ void mlx5e_deactivate_icosq(struct mlx5e_icosq *icosq)
 	synchronize_net(); /* Sync with NAPI. */
 }
 
-void mlx5e_close_icosq(struct mlx5e_icosq *sq)
+static void mlx5e_close_icosq(struct mlx5e_icosq *sq)
 {
 	struct mlx5e_channel *c = sq->channel;
 
@@ -2084,11 +2094,15 @@ static int mlx5e_open_queues(struct mlx5e_channel *c,
 
 	spin_lock_init(&c->async_icosq_lock);
 
-	err = mlx5e_open_icosq(c, params, &cparam->async_icosq, &c->async_icosq);
+	err = mlx5e_open_icosq(c, params, &cparam->async_icosq, &c->async_icosq,
+			       mlx5e_async_icosq_err_cqe_work);
 	if (err)
 		goto err_close_xdpsq_cq;
 
-	err = mlx5e_open_icosq(c, params, &cparam->icosq, &c->icosq);
+	mutex_init(&c->icosq_recovery_lock);
+
+	err = mlx5e_open_icosq(c, params, &cparam->icosq, &c->icosq,
+			       mlx5e_icosq_err_cqe_work);
 	if (err)
 		goto err_close_async_icosq;
 
@@ -2156,9 +2170,12 @@ static void mlx5e_close_queues(struct mlx5e_channel *c)
 	mlx5e_close_xdpsq(&c->xdpsq);
 	if (c->xdp)
 		mlx5e_close_xdpsq(&c->rq_xdpsq);
+	/* The same ICOSQ is used for UMRs for both RQ and XSKRQ. */
+	cancel_work_sync(&c->icosq.recover_work);
 	mlx5e_close_rq(&c->rq);
 	mlx5e_close_sqs(c);
 	mlx5e_close_icosq(&c->icosq);
+	mutex_destroy(&c->icosq_recovery_lock);
 	mlx5e_close_icosq(&c->async_icosq);
 	if (c->xdp)
 		mlx5e_close_cq(&c->rq_xdpsq.cq);

drivers/net/ethernet/mellanox/mlx5/core/en_tc.c

@@ -1196,21 +1196,16 @@ void mlx5e_tc_unoffload_fdb_rules(struct mlx5_eswitch *esw,
 	if (attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH)
 		goto offload_rule_0;
 
-	if (flow_flag_test(flow, CT)) {
-		mlx5_tc_ct_delete_flow(get_ct_priv(flow->priv), flow, attr);
-		return;
-	}
-
-	if (flow_flag_test(flow, SAMPLE)) {
-		mlx5e_tc_sample_unoffload(get_sample_priv(flow->priv), flow->rule[0], attr);
-		return;
-	}
-
 	if (attr->esw_attr->split_count)
 		mlx5_eswitch_del_fwd_rule(esw, flow->rule[1], attr);
 
+	if (flow_flag_test(flow, CT))
+		mlx5_tc_ct_delete_flow(get_ct_priv(flow->priv), flow, attr);
+	else if (flow_flag_test(flow, SAMPLE))
+		mlx5e_tc_sample_unoffload(get_sample_priv(flow->priv), flow->rule[0], attr);
+	else
 offload_rule_0:
-	mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], attr);
+		mlx5_eswitch_del_offloaded_rule(esw, flow->rule[0], attr);
 }
 
 struct mlx5_flow_handle *
@@ -1445,7 +1440,7 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
 							MLX5_FLOW_NAMESPACE_FDB, VPORT_TO_REG,
 							metadata);
 			if (err)
-				return err;
+				goto err_out;
 		}
 	}
 
@@ -1461,22 +1456,26 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
 		if (attr->chain) {
 			NL_SET_ERR_MSG_MOD(extack,
 					   "Internal port rule is only supported on chain 0");
-			return -EOPNOTSUPP;
+			err = -EOPNOTSUPP;
+			goto err_out;
 		}
 
 		if (attr->dest_chain) {
 			NL_SET_ERR_MSG_MOD(extack,
 					   "Internal port rule offload doesn't support goto action");
-			return -EOPNOTSUPP;
+			err = -EOPNOTSUPP;
+			goto err_out;
 		}
 
 		int_port = mlx5e_tc_int_port_get(mlx5e_get_int_port_priv(priv),
 						 parse_attr->filter_dev->ifindex,
 						 flow_flag_test(flow, EGRESS) ?
 						 MLX5E_TC_INT_PORT_EGRESS :
 						 MLX5E_TC_INT_PORT_INGRESS);
-		if (IS_ERR(int_port))
-			return PTR_ERR(int_port);
+		if (IS_ERR(int_port)) {
+			err = PTR_ERR(int_port);
+			goto err_out;
+		}
 
 		esw_attr->int_port = int_port;
 	}

drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c

@@ -121,6 +121,9 @@ u32 mlx5_chains_get_nf_ft_chain(struct mlx5_fs_chains *chains)
 
 u32 mlx5_chains_get_prio_range(struct mlx5_fs_chains *chains)
 {
+	if (!mlx5_chains_prios_supported(chains))
+		return 1;
+
 	if (mlx5_chains_ignore_flow_level_supported(chains))
 		return UINT_MAX;
 

drivers/net/ethernet/mellanox/mlx5/core/main.c

@@ -1809,12 +1809,13 @@ void mlx5_disable_device(struct mlx5_core_dev *dev)
 
 int mlx5_recover_device(struct mlx5_core_dev *dev)
 {
-	int ret = -EIO;
+	if (!mlx5_core_is_sf(dev)) {
+		mlx5_pci_disable_device(dev);
+		if (mlx5_pci_slot_reset(dev->pdev) != PCI_ERS_RESULT_RECOVERED)
+			return -EIO;
+	}
 
-	mlx5_pci_disable_device(dev);
-	if (mlx5_pci_slot_reset(dev->pdev) == PCI_ERS_RESULT_RECOVERED)
-		ret = mlx5_load_one(dev);
-	return ret;
+	return mlx5_load_one(dev);
 }
 
 static struct pci_driver mlx5_core_driver = {