Merge tag 'mlx5-fixes-2022-01-06' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2022-01-06

This series provides bug fixes to mlx5 driver.
Please pull and let me know if there is any problem.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

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

@@ -148,8 +148,12 @@ static void cmd_ent_put(struct mlx5_cmd_work_ent *ent)
 	if (!refcount_dec_and_test(&ent->refcnt))
 		return;
 
-	if (ent->idx >= 0)
-		cmd_free_index(ent->cmd, ent->idx);
+	if (ent->idx >= 0) {
+		struct mlx5_cmd *cmd = ent->cmd;
+
+		cmd_free_index(cmd, ent->idx);
+		up(ent->page_queue ? &cmd->pages_sem : &cmd->sem);
+	}
 
 	cmd_free_ent(ent);
 }
@@ -900,25 +904,6 @@ static bool opcode_allowed(struct mlx5_cmd *cmd, u16 opcode)
 	return cmd->allowed_opcode == opcode;
 }
 
-static int cmd_alloc_index_retry(struct mlx5_cmd *cmd)
-{
-	unsigned long alloc_end = jiffies + msecs_to_jiffies(1000);
-	int idx;
-
-retry:
-	idx = cmd_alloc_index(cmd);
-	if (idx < 0 && time_before(jiffies, alloc_end)) {
-		/* Index allocation can fail on heavy load of commands. This is a temporary
-		 * situation as the current command already holds the semaphore, meaning that
-		 * another command completion is being handled and it is expected to release
-		 * the entry index soon.
-		 */
-		cpu_relax();
-		goto retry;
-	}
-	return idx;
-}
-
 bool mlx5_cmd_is_down(struct mlx5_core_dev *dev)
 {
 	return pci_channel_offline(dev->pdev) ||
@@ -946,7 +931,7 @@ static void cmd_work_handler(struct work_struct *work)
 	sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem;
 	down(sem);
 	if (!ent->page_queue) {
-		alloc_ret = cmd_alloc_index_retry(cmd);
+		alloc_ret = cmd_alloc_index(cmd);
 		if (alloc_ret < 0) {
 			mlx5_core_err_rl(dev, "failed to allocate command entry\n");
 			if (ent->callback) {
@@ -1602,8 +1587,6 @@ static void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool force
 	vector = vec & 0xffffffff;
 	for (i = 0; i < (1 << cmd->log_sz); i++) {
 		if (test_bit(i, &vector)) {
-			struct semaphore *sem;
-
 			ent = cmd->ent_arr[i];
 
 			/* if we already completed the command, ignore it */
@@ -1626,10 +1609,6 @@ static void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool force
 			    dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
 				cmd_ent_put(ent);
 
-			if (ent->page_queue)
-				sem = &cmd->pages_sem;
-			else
-				sem = &cmd->sem;
 			ent->ts2 = ktime_get_ns();
 			memcpy(ent->out->first.data, ent->lay->out, sizeof(ent->lay->out));
 			dump_command(dev, ent, 0);
@@ -1683,7 +1662,6 @@ static void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec, bool force
 				 */
 				complete(&ent->done);
 			}
-			up(sem);
 		}
 	}
 }

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

@@ -1579,6 +1579,8 @@ mlx5e_init_fib_work_ipv4(struct mlx5e_priv *priv,
 	struct net_device *fib_dev;
 
 	fen_info = container_of(info, struct fib_entry_notifier_info, info);
+	if (fen_info->fi->nh)
+		return NULL;
 	fib_dev = fib_info_nh(fen_info->fi, 0)->fib_nh_dev;
 	if (!fib_dev || fib_dev->netdev_ops != &mlx5e_netdev_ops ||
 	    fen_info->dst_len != 32)

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

@@ -11,13 +11,13 @@ static int mlx5e_xsk_map_pool(struct mlx5e_priv *priv,
 {
 	struct device *dev = mlx5_core_dma_dev(priv->mdev);
 
-	return xsk_pool_dma_map(pool, dev, 0);
+	return xsk_pool_dma_map(pool, dev, DMA_ATTR_SKIP_CPU_SYNC);
 }
 
 static void mlx5e_xsk_unmap_pool(struct mlx5e_priv *priv,
 				 struct xsk_buff_pool *pool)
 {
-	return xsk_pool_dma_unmap(pool, 0);
+	return xsk_pool_dma_unmap(pool, DMA_ATTR_SKIP_CPU_SYNC);
 }
 
 static int mlx5e_xsk_get_pools(struct mlx5e_xsk *xsk)

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

@@ -4789,15 +4789,22 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev)
 	}
 
 	if (mlx5_vxlan_allowed(mdev->vxlan) || mlx5_geneve_tx_allowed(mdev)) {
-		netdev->hw_features     |= NETIF_F_GSO_UDP_TUNNEL;
-		netdev->hw_enc_features |= NETIF_F_GSO_UDP_TUNNEL;
-		netdev->vlan_features |= NETIF_F_GSO_UDP_TUNNEL;
+		netdev->hw_features     |= NETIF_F_GSO_UDP_TUNNEL |
+					   NETIF_F_GSO_UDP_TUNNEL_CSUM;
+		netdev->hw_enc_features |= NETIF_F_GSO_UDP_TUNNEL |
+					   NETIF_F_GSO_UDP_TUNNEL_CSUM;
+		netdev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM;
+		netdev->vlan_features |= NETIF_F_GSO_UDP_TUNNEL |
+					 NETIF_F_GSO_UDP_TUNNEL_CSUM;
 	}
 
 	if (mlx5e_tunnel_proto_supported_tx(mdev, IPPROTO_GRE)) {
-		netdev->hw_features     |= NETIF_F_GSO_GRE;
-		netdev->hw_enc_features |= NETIF_F_GSO_GRE;
-		netdev->gso_partial_features |= NETIF_F_GSO_GRE;
+		netdev->hw_features     |= NETIF_F_GSO_GRE |
+					   NETIF_F_GSO_GRE_CSUM;
+		netdev->hw_enc_features |= NETIF_F_GSO_GRE |
+					   NETIF_F_GSO_GRE_CSUM;
+		netdev->gso_partial_features |= NETIF_F_GSO_GRE |
+						NETIF_F_GSO_GRE_CSUM;
 	}
 
 	if (mlx5e_tunnel_proto_supported_tx(mdev, IPPROTO_IPIP)) {

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

@@ -50,6 +50,7 @@
 #include "fs_core.h"
 #include "lib/mlx5.h"
 #include "lib/devcom.h"
+#include "lib/vxlan.h"
 #define CREATE_TRACE_POINTS
 #include "diag/en_rep_tracepoint.h"
 #include "en_accel/ipsec.h"
@@ -1027,6 +1028,7 @@ static void mlx5e_uplink_rep_enable(struct mlx5e_priv *priv)
 	rtnl_lock();
 	if (netif_running(netdev))
 		mlx5e_open(netdev);
+	udp_tunnel_nic_reset_ntf(priv->netdev);
 	netif_device_attach(netdev);
 	rtnl_unlock();
 }
@@ -1048,6 +1050,7 @@ static void mlx5e_uplink_rep_disable(struct mlx5e_priv *priv)
 	mlx5_notifier_unregister(mdev, &priv->events_nb);
 	mlx5e_rep_tc_disable(priv);
 	mlx5_lag_remove_netdev(mdev, priv->netdev);
+	mlx5_vxlan_reset_to_default(mdev->vxlan);
 }
 
 static MLX5E_DEFINE_STATS_GRP(sw_rep, 0);

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

@@ -278,8 +278,8 @@ static inline int mlx5e_page_alloc_pool(struct mlx5e_rq *rq,
 	if (unlikely(!dma_info->page))
 		return -ENOMEM;
 
-	dma_info->addr = dma_map_page(rq->pdev, dma_info->page, 0,
-				      PAGE_SIZE, rq->buff.map_dir);
+	dma_info->addr = dma_map_page_attrs(rq->pdev, dma_info->page, 0, PAGE_SIZE,
+					    rq->buff.map_dir, DMA_ATTR_SKIP_CPU_SYNC);
 	if (unlikely(dma_mapping_error(rq->pdev, dma_info->addr))) {
 		page_pool_recycle_direct(rq->page_pool, dma_info->page);
 		dma_info->page = NULL;
@@ -300,7 +300,8 @@ static inline int mlx5e_page_alloc(struct mlx5e_rq *rq,
 
 void mlx5e_page_dma_unmap(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info)
 {
-	dma_unmap_page(rq->pdev, dma_info->addr, PAGE_SIZE, rq->buff.map_dir);
+	dma_unmap_page_attrs(rq->pdev, dma_info->addr, PAGE_SIZE, rq->buff.map_dir,
+			     DMA_ATTR_SKIP_CPU_SYNC);
 }
 
 void mlx5e_page_release_dynamic(struct mlx5e_rq *rq,

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

@@ -1949,6 +1949,111 @@ u8 mlx5e_tc_get_ip_version(struct mlx5_flow_spec *spec, bool outer)
 	return ip_version;
 }
 
+/* Tunnel device follows RFC 6040, see include/net/inet_ecn.h.
+ * And changes inner ip_ecn depending on inner and outer ip_ecn as follows:
+ *      +---------+----------------------------------------+
+ *      |Arriving |         Arriving Outer Header          |
+ *      |   Inner +---------+---------+---------+----------+
+ *      |  Header | Not-ECT | ECT(0)  | ECT(1)  |   CE     |
+ *      +---------+---------+---------+---------+----------+
+ *      | Not-ECT | Not-ECT | Not-ECT | Not-ECT | <drop>   |
+ *      |  ECT(0) |  ECT(0) | ECT(0)  | ECT(1)  |   CE*    |
+ *      |  ECT(1) |  ECT(1) | ECT(1)  | ECT(1)* |   CE*    |
+ *      |    CE   |   CE    |  CE     | CE      |   CE     |
+ *      +---------+---------+---------+---------+----------+
+ *
+ * Tc matches on inner after decapsulation on tunnel device, but hw offload matches
+ * the inner ip_ecn value before hardware decap action.
+ *
+ * Cells marked are changed from original inner packet ip_ecn value during decap, and
+ * so matching those values on inner ip_ecn before decap will fail.
+ *
+ * The following helper allows offload when inner ip_ecn won't be changed by outer ip_ecn,
+ * except for the outer ip_ecn = CE, where in all cases inner ip_ecn will be changed to CE,
+ * and such we can drop the inner ip_ecn=CE match.
+ */
+
+static int mlx5e_tc_verify_tunnel_ecn(struct mlx5e_priv *priv,
+				      struct flow_cls_offload *f,
+				      bool *match_inner_ecn)
+{
+	u8 outer_ecn_mask = 0, outer_ecn_key = 0, inner_ecn_mask = 0, inner_ecn_key = 0;
+	struct flow_rule *rule = flow_cls_offload_flow_rule(f);
+	struct netlink_ext_ack *extack = f->common.extack;
+	struct flow_match_ip match;
+
+	*match_inner_ecn = true;
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IP)) {
+		flow_rule_match_enc_ip(rule, &match);
+		outer_ecn_key = match.key->tos & INET_ECN_MASK;
+		outer_ecn_mask = match.mask->tos & INET_ECN_MASK;
+	}
+
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_IP)) {
+		flow_rule_match_ip(rule, &match);
+		inner_ecn_key = match.key->tos & INET_ECN_MASK;
+		inner_ecn_mask = match.mask->tos & INET_ECN_MASK;
+	}
+
+	if (outer_ecn_mask != 0 && outer_ecn_mask != INET_ECN_MASK) {
+		NL_SET_ERR_MSG_MOD(extack, "Partial match on enc_tos ecn bits isn't supported");
+		netdev_warn(priv->netdev, "Partial match on enc_tos ecn bits isn't supported");
+		return -EOPNOTSUPP;
+	}
+
+	if (!outer_ecn_mask) {
+		if (!inner_ecn_mask)
+			return 0;
+
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Matching on tos ecn bits without also matching enc_tos ecn bits isn't supported");
+		netdev_warn(priv->netdev,
+			    "Matching on tos ecn bits without also matching enc_tos ecn bits isn't supported");
+		return -EOPNOTSUPP;
+	}
+
+	if (inner_ecn_mask && inner_ecn_mask != INET_ECN_MASK) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Partial match on tos ecn bits with match on enc_tos ecn bits isn't supported");
+		netdev_warn(priv->netdev,
+			    "Partial match on tos ecn bits with match on enc_tos ecn bits isn't supported");
+		return -EOPNOTSUPP;
+	}
+
+	if (!inner_ecn_mask)
+		return 0;
+
+	/* Both inner and outer have full mask on ecn */
+
+	if (outer_ecn_key == INET_ECN_ECT_1) {
+		/* inner ecn might change by DECAP action */
+
+		NL_SET_ERR_MSG_MOD(extack, "Match on enc_tos ecn = ECT(1) isn't supported");
+		netdev_warn(priv->netdev, "Match on enc_tos ecn = ECT(1) isn't supported");
+		return -EOPNOTSUPP;
+	}
+
+	if (outer_ecn_key != INET_ECN_CE)
+		return 0;
+
+	if (inner_ecn_key != INET_ECN_CE) {
+		/* Can't happen in software, as packet ecn will be changed to CE after decap */
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Match on tos enc_tos ecn = CE while match on tos ecn != CE isn't supported");
+		netdev_warn(priv->netdev,
+			    "Match on tos enc_tos ecn = CE while match on tos ecn != CE isn't supported");
+		return -EOPNOTSUPP;
+	}
+
+	/* outer ecn = CE, inner ecn = CE, as decap will change inner ecn to CE in anycase,
+	 * drop match on inner ecn
+	 */
+	*match_inner_ecn = false;
+
+	return 0;
+}
+
 static int parse_tunnel_attr(struct mlx5e_priv *priv,
 			     struct mlx5e_tc_flow *flow,
 			     struct mlx5_flow_spec *spec,
@@ -2144,6 +2249,7 @@ static int __parse_cls_flower(struct mlx5e_priv *priv,
 	struct flow_rule *rule = flow_cls_offload_flow_rule(f);
 	struct flow_dissector *dissector = rule->match.dissector;
 	enum fs_flow_table_type fs_type;
+	bool match_inner_ecn = true;
 	u16 addr_type = 0;
 	u8 ip_proto = 0;
 	u8 *match_level;
@@ -2197,6 +2303,10 @@ static int __parse_cls_flower(struct mlx5e_priv *priv,
 			headers_c = get_match_inner_headers_criteria(spec);
 			headers_v = get_match_inner_headers_value(spec);
 		}
+
+		err = mlx5e_tc_verify_tunnel_ecn(priv, f, &match_inner_ecn);
+		if (err)
+			return err;
 	}
 
 	err = mlx5e_flower_parse_meta(filter_dev, f);
@@ -2420,10 +2530,12 @@ static int __parse_cls_flower(struct mlx5e_priv *priv,
 		struct flow_match_ip match;
 
 		flow_rule_match_ip(rule, &match);
-		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_ecn,
-			 match.mask->tos & 0x3);
-		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ecn,
-			 match.key->tos & 0x3);
+		if (match_inner_ecn) {
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_ecn,
+				 match.mask->tos & 0x3);
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_ecn,
+				 match.key->tos & 0x3);
+		}
 
 		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_dscp,
 			 match.mask->tos >> 2);

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

@@ -295,26 +295,28 @@ esw_setup_chain_src_port_rewrite(struct mlx5_flow_destination *dest,
 				 int *i)
 {
 	struct mlx5_esw_flow_attr *esw_attr = attr->esw_attr;
-	int j, err;
+	int err;
 
 	if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SRC_REWRITE))
 		return -EOPNOTSUPP;
 
-	for (j = esw_attr->split_count; j < esw_attr->out_count; j++, (*i)++) {
-		err = esw_setup_chain_dest(dest, flow_act, chains, attr->dest_chain, 1, 0, *i);
-		if (err)
-			goto err_setup_chain;
+	/* flow steering cannot handle more than one dest with the same ft
+	 * in a single flow
+	 */
+	if (esw_attr->out_count - esw_attr->split_count > 1)
+		return -EOPNOTSUPP;
 
-		if (esw_attr->dests[j].pkt_reformat) {
-			flow_act->action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
-			flow_act->pkt_reformat = esw_attr->dests[j].pkt_reformat;
-		}
+	err = esw_setup_chain_dest(dest, flow_act, chains, attr->dest_chain, 1, 0, *i);
+	if (err)
+		return err;
+
+	if (esw_attr->dests[esw_attr->split_count].pkt_reformat) {
+		flow_act->action |= MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
+		flow_act->pkt_reformat = esw_attr->dests[esw_attr->split_count].pkt_reformat;
 	}
-	return 0;
+	(*i)++;
 
-err_setup_chain:
-	esw_put_dest_tables_loop(esw, attr, esw_attr->split_count, j);
-	return err;
+	return 0;
 }
 
 static void esw_cleanup_chain_src_port_rewrite(struct mlx5_eswitch *esw,

drivers/net/ethernet/mellanox/mlx5/core/lag/mp.c

@@ -268,10 +268,8 @@ static int mlx5_lag_fib_event(struct notifier_block *nb,
 		fen_info = container_of(info, struct fib_entry_notifier_info,
 					info);
 		fi = fen_info->fi;
-		if (fi->nh) {
-			NL_SET_ERR_MSG_MOD(info->extack, "IPv4 route with nexthop objects is not supported");
-			return notifier_from_errno(-EINVAL);
-		}
+		if (fi->nh)
+			return NOTIFY_DONE;
 		fib_dev = fib_info_nh(fen_info->fi, 0)->fib_nh_dev;
 		if (fib_dev != ldev->pf[MLX5_LAG_P1].netdev &&
 		    fib_dev != ldev->pf[MLX5_LAG_P2].netdev) {

drivers/net/ethernet/mellanox/mlx5/core/sf/dev/dev.c

@@ -30,10 +30,7 @@ bool mlx5_sf_dev_allocated(const struct mlx5_core_dev *dev)
 {
 	struct mlx5_sf_dev_table *table = dev->priv.sf_dev_table;
 
-	if (!mlx5_sf_dev_supported(dev))
-		return false;
-
-	return !xa_empty(&table->devices);
+	return table && !xa_empty(&table->devices);
 }
 
 static ssize_t sfnum_show(struct device *dev, struct device_attribute *attr, char *buf)