Merge tag 'mlx5-fixes-2022-02-23' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2022-02-22

This series provides bug fixes to mlx5 driver.
Please pull and let me know if there is any problem.

* tag 'mlx5-fixes-2022-02-23' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Fix VF min/max rate parameters interchange mistake
  net/mlx5e: Add missing increment of count
  net/mlx5e: MPLSoUDP decap, fix check for unsupported matches
  net/mlx5e: Fix MPLSoUDP encap to use MPLS action information
  net/mlx5e: Add feature check for set fec counters
  net/mlx5e: TC, Skip redundant ct clear actions
  net/mlx5e: TC, Reject rules with forward and drop actions
  net/mlx5e: TC, Reject rules with drop and modify hdr action
  net/mlx5e: kTLS, Use CHECKSUM_UNNECESSARY for device-offloaded packets
  net/mlx5e: Fix wrong return value on ioctl EEPROM query failure
  net/mlx5: Fix possible deadlock on rule deletion
  net/mlx5: Fix tc max supported prio for nic mode
  net/mlx5: Fix wrong limitation of metadata match on ecpf
  net/mlx5: Update log_max_qp value to be 17 at most
  net/mlx5: DR, Fix the threshold that defines when pool sync is initiated
  net/mlx5: DR, Don't allow match on IP w/o matching on full ethertype/ip_version
  net/mlx5: DR, Fix slab-out-of-bounds in mlx5_cmd_dr_create_fte
  net/mlx5: DR, Cache STE shadow memory
  net/mlx5: Update the list of the PCI supported devices
====================

Link: https://lore.kernel.org/r/20220224001123.365265-1-saeed@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

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

@@ -16,11 +16,13 @@ struct mlx5e_tc_act_parse_state {
 	unsigned int num_actions;
 	struct mlx5e_tc_flow *flow;
 	struct netlink_ext_ack *extack;
+	bool ct_clear;
 	bool encap;
 	bool decap;
 	bool mpls_push;
 	bool ptype_host;
 	const struct ip_tunnel_info *tun_info;
+	struct mlx5e_mpls_info mpls_info;
 	struct pedit_headers_action hdrs[__PEDIT_CMD_MAX];
 	int ifindexes[MLX5_MAX_FLOW_FWD_VPORTS];
 	int if_count;

drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/ct.c

@@ -27,8 +27,13 @@ tc_act_parse_ct(struct mlx5e_tc_act_parse_state *parse_state,
 		struct mlx5e_priv *priv,
 		struct mlx5_flow_attr *attr)
 {
+	bool clear_action = act->ct.action & TCA_CT_ACT_CLEAR;
 	int err;
 
+	/* It's redundant to do ct clear more than once. */
+	if (clear_action && parse_state->ct_clear)
+		return 0;
+
 	err = mlx5_tc_ct_parse_action(parse_state->ct_priv, attr,
 				      &attr->parse_attr->mod_hdr_acts,
 				      act, parse_state->extack);
@@ -40,6 +45,8 @@ tc_act_parse_ct(struct mlx5e_tc_act_parse_state *parse_state,
 	if (mlx5e_is_eswitch_flow(parse_state->flow))
 		attr->esw_attr->split_count = attr->esw_attr->out_count;
 
+	parse_state->ct_clear = clear_action;
+
 	return 0;
 }
 

drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mirred.c

@@ -177,6 +177,12 @@ parse_mirred_encap(struct mlx5e_tc_act_parse_state *parse_state,
 		return -ENOMEM;
 
 	parse_state->encap = false;
+
+	if (parse_state->mpls_push) {
+		memcpy(&parse_attr->mpls_info[esw_attr->out_count],
+		       &parse_state->mpls_info, sizeof(parse_state->mpls_info));
+		parse_state->mpls_push = false;
+	}
 	esw_attr->dests[esw_attr->out_count].flags |= MLX5_ESW_DEST_ENCAP;
 	esw_attr->out_count++;
 	/* attr->dests[].rep is resolved when we handle encap */

drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/mpls.c

@@ -22,6 +22,16 @@ tc_act_can_offload_mpls_push(struct mlx5e_tc_act_parse_state *parse_state,
 	return true;
 }
 
+static void
+copy_mpls_info(struct mlx5e_mpls_info *mpls_info,
+	       const struct flow_action_entry *act)
+{
+	mpls_info->label = act->mpls_push.label;
+	mpls_info->tc = act->mpls_push.tc;
+	mpls_info->bos = act->mpls_push.bos;
+	mpls_info->ttl = act->mpls_push.ttl;
+}
+
 static int
 tc_act_parse_mpls_push(struct mlx5e_tc_act_parse_state *parse_state,
 		       const struct flow_action_entry *act,
@@ -29,6 +39,7 @@ tc_act_parse_mpls_push(struct mlx5e_tc_act_parse_state *parse_state,
 		       struct mlx5_flow_attr *attr)
 {
 	parse_state->mpls_push = true;
+	copy_mpls_info(&parse_state->mpls_info, act);
 
 	return 0;
 }

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

@@ -35,6 +35,7 @@ enum {
 
 struct mlx5e_tc_flow_parse_attr {
 	const struct ip_tunnel_info *tun_info[MLX5_MAX_FLOW_FWD_VPORTS];
+	struct mlx5e_mpls_info mpls_info[MLX5_MAX_FLOW_FWD_VPORTS];
 	struct net_device *filter_dev;
 	struct mlx5_flow_spec spec;
 	struct mlx5e_tc_mod_hdr_acts mod_hdr_acts;

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

@@ -750,6 +750,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
 	struct mlx5e_tc_flow_parse_attr *parse_attr;
 	struct mlx5_flow_attr *attr = flow->attr;
 	const struct ip_tunnel_info *tun_info;
+	const struct mlx5e_mpls_info *mpls_info;
 	unsigned long tbl_time_before = 0;
 	struct mlx5e_encap_entry *e;
 	struct mlx5e_encap_key key;
@@ -760,6 +761,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
 
 	parse_attr = attr->parse_attr;
 	tun_info = parse_attr->tun_info[out_index];
+	mpls_info = &parse_attr->mpls_info[out_index];
 	family = ip_tunnel_info_af(tun_info);
 	key.ip_tun_key = &tun_info->key;
 	key.tc_tunnel = mlx5e_get_tc_tun(mirred_dev);
@@ -810,6 +812,7 @@ int mlx5e_attach_encap(struct mlx5e_priv *priv,
 		goto out_err_init;
 	}
 	e->tun_info = tun_info;
+	memcpy(&e->mpls_info, mpls_info, sizeof(*mpls_info));
 	err = mlx5e_tc_tun_init_encap_attr(mirred_dev, priv, e, extack);
 	if (err)
 		goto out_err_init;

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

@@ -30,16 +30,15 @@ static int generate_ip_tun_hdr(char buf[],
 			       struct mlx5e_encap_entry *r)
 {
 	const struct ip_tunnel_key *tun_key = &r->tun_info->key;
+	const struct mlx5e_mpls_info *mpls_info = &r->mpls_info;
 	struct udphdr *udp = (struct udphdr *)(buf);
 	struct mpls_shim_hdr *mpls;
-	u32 tun_id;
 
-	tun_id = be32_to_cpu(tunnel_id_to_key32(tun_key->tun_id));
 	mpls = (struct mpls_shim_hdr *)(udp + 1);
 	*ip_proto = IPPROTO_UDP;
 
 	udp->dest = tun_key->tp_dst;
-	*mpls = mpls_entry_encode(tun_id, tun_key->ttl, tun_key->tos, true);
+	*mpls = mpls_entry_encode(mpls_info->label, mpls_info->ttl, mpls_info->tc, mpls_info->bos);
 
 	return 0;
 }
@@ -60,37 +59,31 @@ static int parse_tunnel(struct mlx5e_priv *priv,
 			void *headers_v)
 {
 	struct flow_rule *rule = flow_cls_offload_flow_rule(f);
-	struct flow_match_enc_keyid enc_keyid;
 	struct flow_match_mpls match;
 	void *misc2_c;
 	void *misc2_v;
 
-	misc2_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
-			       misc_parameters_2);
-	misc2_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
-			       misc_parameters_2);
-
-	if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_MPLS))
-		return 0;
-
-	if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID))
-		return 0;
-
-	flow_rule_match_enc_keyid(rule, &enc_keyid);
-
-	if (!enc_keyid.mask->keyid)
-		return 0;
-
 	if (!MLX5_CAP_ETH(priv->mdev, tunnel_stateless_mpls_over_udp) &&
 	    !(MLX5_CAP_GEN(priv->mdev, flex_parser_protocols) & MLX5_FLEX_PROTO_CW_MPLS_UDP))
 		return -EOPNOTSUPP;
 
+	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID))
+		return -EOPNOTSUPP;
+
+	if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_MPLS))
+		return 0;
+
 	flow_rule_match_mpls(rule, &match);
 
 	/* Only support matching the first LSE */
 	if (match.mask->used_lses != 1)
 		return -EOPNOTSUPP;
 
+	misc2_c = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+			       misc_parameters_2);
+	misc2_v = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+			       misc_parameters_2);
+
 	MLX5_SET(fte_match_set_misc2, misc2_c,
 		 outer_first_mpls_over_udp.mpls_label,
 		 match.mask->ls[0].mpls_label);

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

@@ -1792,7 +1792,7 @@ static int mlx5e_get_module_eeprom(struct net_device *netdev,
 		if (size_read < 0) {
 			netdev_err(priv->netdev, "%s: mlx5_query_eeprom failed:0x%x\n",
 				   __func__, size_read);
-			return 0;
+			return size_read;
 		}
 
 		i += size_read;

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

@@ -183,6 +183,13 @@ struct mlx5e_decap_entry {
 	struct rcu_head rcu;
 };
 
+struct mlx5e_mpls_info {
+	u32             label;
+	u8              tc;
+	u8              bos;
+	u8              ttl;
+};
+
 struct mlx5e_encap_entry {
 	/* attached neigh hash entry */
 	struct mlx5e_neigh_hash_entry *nhe;
@@ -196,6 +203,7 @@ struct mlx5e_encap_entry {
 	struct list_head route_list;
 	struct mlx5_pkt_reformat *pkt_reformat;
 	const struct ip_tunnel_info *tun_info;
+	struct mlx5e_mpls_info mpls_info;
 	unsigned char h_dest[ETH_ALEN];	/* destination eth addr	*/
 
 	struct net_device *out_dev;

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

@@ -1349,7 +1349,8 @@ static inline void mlx5e_handle_csum(struct net_device *netdev,
 	}
 
 	/* True when explicitly set via priv flag, or XDP prog is loaded */
-	if (test_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &rq->state))
+	if (test_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &rq->state) ||
+	    get_cqe_tls_offload(cqe))
 		goto csum_unnecessary;
 
 	/* CQE csum doesn't cover padding octets in short ethernet

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

@@ -334,6 +334,7 @@ void mlx5e_self_test(struct net_device *ndev, struct ethtool_test *etest,
 		netdev_info(ndev, "\t[%d] %s start..\n", i, st.name);
 		buf[count] = st.st_func(priv);
 		netdev_info(ndev, "\t[%d] %s end: result(%lld)\n", i, st.name, buf[count]);
+		count++;
 	}
 
 	mutex_unlock(&priv->state_lock);

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

@@ -1254,9 +1254,6 @@ static void fec_set_corrected_bits_total(struct mlx5e_priv *priv,
 	u32 in[MLX5_ST_SZ_DW(ppcnt_reg)] = {};
 	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
 
-	if (!MLX5_CAP_PCAM_FEATURE(mdev, ppcnt_statistical_group))
-		return;
-
 	MLX5_SET(ppcnt_reg, in, local_port, 1);
 	MLX5_SET(ppcnt_reg, in, grp, MLX5_PHYSICAL_LAYER_STATISTICAL_GROUP);
 	if (mlx5_core_access_reg(mdev, in, sz, ppcnt_phy_statistical,
@@ -1272,6 +1269,9 @@ static void fec_set_corrected_bits_total(struct mlx5e_priv *priv,
 void mlx5e_stats_fec_get(struct mlx5e_priv *priv,
 			 struct ethtool_fec_stats *fec_stats)
 {
+	if (!MLX5_CAP_PCAM_FEATURE(priv->mdev, ppcnt_statistical_group))
+		return;
+
 	fec_set_corrected_bits_total(priv, fec_stats);
 	fec_set_block_stats(priv, fec_stats);
 }

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

@@ -3204,6 +3204,18 @@ actions_match_supported(struct mlx5e_priv *priv,
 		return false;
 	}
 
+	if (!(~actions &
+	      (MLX5_FLOW_CONTEXT_ACTION_FWD_DEST | MLX5_FLOW_CONTEXT_ACTION_DROP))) {
+		NL_SET_ERR_MSG_MOD(extack, "Rule cannot support forward+drop action");
+		return false;
+	}
+
+	if (actions & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR &&
+	    actions & MLX5_FLOW_CONTEXT_ACTION_DROP) {
+		NL_SET_ERR_MSG_MOD(extack, "Drop with modify header action is not supported");
+		return false;
+	}
+
 	if (actions & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR &&
 	    !modify_header_match_supported(priv, &parse_attr->spec, flow_action,
 					   actions, ct_flow, ct_clear, extack))

drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c

@@ -697,7 +697,7 @@ void mlx5_esw_qos_vport_disable(struct mlx5_eswitch *esw, struct mlx5_vport *vpo
 }
 
 int mlx5_esw_qos_set_vport_rate(struct mlx5_eswitch *esw, struct mlx5_vport *vport,
-				u32 min_rate, u32 max_rate)
+				u32 max_rate, u32 min_rate)
 {
 	int err;
 

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

@@ -2838,10 +2838,6 @@ bool mlx5_esw_vport_match_metadata_supported(const struct mlx5_eswitch *esw)
 	if (!MLX5_CAP_ESW_FLOWTABLE(esw->dev, flow_source))
 		return false;
 
-	if (mlx5_core_is_ecpf_esw_manager(esw->dev) ||
-	    mlx5_ecpf_vport_exists(esw->dev))
-		return false;
-
 	return true;
 }
 

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

@@ -2074,6 +2074,8 @@ void mlx5_del_flow_rules(struct mlx5_flow_handle *handle)
 		fte->node.del_hw_func = NULL;
 		up_write_ref_node(&fte->node, false);
 		tree_put_node(&fte->node, false);
+	} else {
+		up_write_ref_node(&fte->node, false);
 	}
 	kfree(handle);
 }

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

@@ -526,7 +526,7 @@ static int handle_hca_cap(struct mlx5_core_dev *dev, void *set_ctx)
 
 	/* Check log_max_qp from HCA caps to set in current profile */
 	if (prof->log_max_qp == LOG_MAX_SUPPORTED_QPS) {
-		prof->log_max_qp = MLX5_CAP_GEN_MAX(dev, log_max_qp);
+		prof->log_max_qp = min_t(u8, 17, MLX5_CAP_GEN_MAX(dev, log_max_qp));
 	} else if (MLX5_CAP_GEN_MAX(dev, log_max_qp) < prof->log_max_qp) {
 		mlx5_core_warn(dev, "log_max_qp value in current profile is %d, changing it to HCA capability limit (%d)\n",
 			       prof->log_max_qp,
@@ -1840,10 +1840,12 @@ static const struct pci_device_id mlx5_core_pci_table[] = {
 	{ PCI_VDEVICE(MELLANOX, 0x101e), MLX5_PCI_DEV_IS_VF},	/* ConnectX Family mlx5Gen Virtual Function */
 	{ PCI_VDEVICE(MELLANOX, 0x101f) },			/* ConnectX-6 LX */
 	{ PCI_VDEVICE(MELLANOX, 0x1021) },			/* ConnectX-7 */
+	{ PCI_VDEVICE(MELLANOX, 0x1023) },			/* ConnectX-8 */
 	{ PCI_VDEVICE(MELLANOX, 0xa2d2) },			/* BlueField integrated ConnectX-5 network controller */
 	{ PCI_VDEVICE(MELLANOX, 0xa2d3), MLX5_PCI_DEV_IS_VF},	/* BlueField integrated ConnectX-5 network controller VF */
 	{ PCI_VDEVICE(MELLANOX, 0xa2d6) },			/* BlueField-2 integrated ConnectX-6 Dx network controller */
 	{ PCI_VDEVICE(MELLANOX, 0xa2dc) },			/* BlueField-3 integrated ConnectX-7 network controller */
+	{ PCI_VDEVICE(MELLANOX, 0xa2df) },			/* BlueField-4 integrated ConnectX-8 network controller */
 	{ 0, }
 };
 

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c

@@ -4,7 +4,6 @@
 #include "dr_types.h"
 
 #define DR_ICM_MODIFY_HDR_ALIGN_BASE 64
-#define DR_ICM_SYNC_THRESHOLD_POOL (64 * 1024 * 1024)
 
 struct mlx5dr_icm_pool {
 	enum mlx5dr_icm_type icm_type;
@@ -136,37 +135,35 @@ static void dr_icm_pool_mr_destroy(struct mlx5dr_icm_mr *icm_mr)
 	kvfree(icm_mr);
 }
 
-static int dr_icm_chunk_ste_init(struct mlx5dr_icm_chunk *chunk)
+static int dr_icm_buddy_get_ste_size(struct mlx5dr_icm_buddy_mem *buddy)
 {
-	chunk->ste_arr = kvzalloc(chunk->num_of_entries *
-				  sizeof(chunk->ste_arr[0]), GFP_KERNEL);
-	if (!chunk->ste_arr)
-		return -ENOMEM;
-
-	chunk->hw_ste_arr = kvzalloc(chunk->num_of_entries *
-				     DR_STE_SIZE_REDUCED, GFP_KERNEL);
-	if (!chunk->hw_ste_arr)
-		goto out_free_ste_arr;
-
-	chunk->miss_list = kvmalloc(chunk->num_of_entries *
-				    sizeof(chunk->miss_list[0]), GFP_KERNEL);
-	if (!chunk->miss_list)
-		goto out_free_hw_ste_arr;
+	/* We support only one type of STE size, both for ConnectX-5 and later
+	 * devices. Once the support for match STE which has a larger tag is
+	 * added (32B instead of 16B), the STE size for devices later than
+	 * ConnectX-5 needs to account for that.
+	 */
+	return DR_STE_SIZE_REDUCED;
+}
 
-	return 0;
+static void dr_icm_chunk_ste_init(struct mlx5dr_icm_chunk *chunk, int offset)
+{
+	struct mlx5dr_icm_buddy_mem *buddy = chunk->buddy_mem;
+	int index = offset / DR_STE_SIZE;
 
-out_free_hw_ste_arr:
-	kvfree(chunk->hw_ste_arr);
-out_free_ste_arr:
-	kvfree(chunk->ste_arr);
-	return -ENOMEM;
+	chunk->ste_arr = &buddy->ste_arr[index];
+	chunk->miss_list = &buddy->miss_list[index];
+	chunk->hw_ste_arr = buddy->hw_ste_arr +
+			    index * dr_icm_buddy_get_ste_size(buddy);
 }
 
 static void dr_icm_chunk_ste_cleanup(struct mlx5dr_icm_chunk *chunk)
 {
-	kvfree(chunk->miss_list);
-	kvfree(chunk->hw_ste_arr);
-	kvfree(chunk->ste_arr);
+	struct mlx5dr_icm_buddy_mem *buddy = chunk->buddy_mem;
+
+	memset(chunk->hw_ste_arr, 0,
+	       chunk->num_of_entries * dr_icm_buddy_get_ste_size(buddy));
+	memset(chunk->ste_arr, 0,
+	       chunk->num_of_entries * sizeof(chunk->ste_arr[0]));
 }
 
 static enum mlx5dr_icm_type
@@ -189,6 +186,44 @@ static void dr_icm_chunk_destroy(struct mlx5dr_icm_chunk *chunk,
 	kvfree(chunk);
 }
 
+static int dr_icm_buddy_init_ste_cache(struct mlx5dr_icm_buddy_mem *buddy)
+{
+	int num_of_entries =
+		mlx5dr_icm_pool_chunk_size_to_entries(buddy->pool->max_log_chunk_sz);
+
+	buddy->ste_arr = kvcalloc(num_of_entries,
+				  sizeof(struct mlx5dr_ste), GFP_KERNEL);
+	if (!buddy->ste_arr)
+		return -ENOMEM;
+
+	/* Preallocate full STE size on non-ConnectX-5 devices since
+	 * we need to support both full and reduced with the same cache.
+	 */
+	buddy->hw_ste_arr = kvcalloc(num_of_entries,
+				     dr_icm_buddy_get_ste_size(buddy), GFP_KERNEL);
+	if (!buddy->hw_ste_arr)
+		goto free_ste_arr;
+
+	buddy->miss_list = kvmalloc(num_of_entries * sizeof(struct list_head), GFP_KERNEL);
+	if (!buddy->miss_list)
+		goto free_hw_ste_arr;
+
+	return 0;
+
+free_hw_ste_arr:
+	kvfree(buddy->hw_ste_arr);
+free_ste_arr:
+	kvfree(buddy->ste_arr);
+	return -ENOMEM;
+}
+
+static void dr_icm_buddy_cleanup_ste_cache(struct mlx5dr_icm_buddy_mem *buddy)
+{
+	kvfree(buddy->ste_arr);
+	kvfree(buddy->hw_ste_arr);
+	kvfree(buddy->miss_list);
+}
+
 static int dr_icm_buddy_create(struct mlx5dr_icm_pool *pool)
 {
 	struct mlx5dr_icm_buddy_mem *buddy;
@@ -208,11 +243,19 @@ static int dr_icm_buddy_create(struct mlx5dr_icm_pool *pool)
 	buddy->icm_mr = icm_mr;
 	buddy->pool = pool;
 
+	if (pool->icm_type == DR_ICM_TYPE_STE) {
+		/* Reduce allocations by preallocating and reusing the STE structures */
+		if (dr_icm_buddy_init_ste_cache(buddy))
+			goto err_cleanup_buddy;
+	}
+
 	/* add it to the -start- of the list in order to search in it first */
 	list_add(&buddy->list_node, &pool->buddy_mem_list);
 
 	return 0;
 
+err_cleanup_buddy:
+	mlx5dr_buddy_cleanup(buddy);
 err_free_buddy:
 	kvfree(buddy);
 free_mr:
@@ -234,6 +277,9 @@ static void dr_icm_buddy_destroy(struct mlx5dr_icm_buddy_mem *buddy)
 
 	mlx5dr_buddy_cleanup(buddy);
 
+	if (buddy->pool->icm_type == DR_ICM_TYPE_STE)
+		dr_icm_buddy_cleanup_ste_cache(buddy);
+
 	kvfree(buddy);
 }
 
@@ -261,34 +307,30 @@ dr_icm_chunk_create(struct mlx5dr_icm_pool *pool,
 	chunk->byte_size =
 		mlx5dr_icm_pool_chunk_size_to_byte(chunk_size, pool->icm_type);
 	chunk->seg = seg;
+	chunk->buddy_mem = buddy_mem_pool;
 
-	if (pool->icm_type == DR_ICM_TYPE_STE && dr_icm_chunk_ste_init(chunk)) {
-		mlx5dr_err(pool->dmn,
-			   "Failed to init ste arrays (order: %d)\n",
-			   chunk_size);
-		goto out_free_chunk;
-	}
+	if (pool->icm_type == DR_ICM_TYPE_STE)
+		dr_icm_chunk_ste_init(chunk, offset);
 
 	buddy_mem_pool->used_memory += chunk->byte_size;
-	chunk->buddy_mem = buddy_mem_pool;
 	INIT_LIST_HEAD(&chunk->chunk_list);
 
 	/* chunk now is part of the used_list */
 	list_add_tail(&chunk->chunk_list, &buddy_mem_pool->used_list);
 
 	return chunk;
-
-out_free_chunk:
-	kvfree(chunk);
-	return NULL;
 }
 
 static bool dr_icm_pool_is_sync_required(struct mlx5dr_icm_pool *pool)
 {
-	if (pool->hot_memory_size > DR_ICM_SYNC_THRESHOLD_POOL)
-		return true;
+	int allow_hot_size;
+
+	/* sync when hot memory reaches half of the pool size */
+	allow_hot_size =
+		mlx5dr_icm_pool_chunk_size_to_byte(pool->max_log_chunk_sz,
+						   pool->icm_type) / 2;
 
-	return false;
+	return pool->hot_memory_size > allow_hot_size;
 }
 
 static int dr_icm_pool_sync_all_buddy_pools(struct mlx5dr_icm_pool *pool)

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_matcher.c

@@ -13,18 +13,6 @@ static bool dr_mask_is_dmac_set(struct mlx5dr_match_spec *spec)
 	return (spec->dmac_47_16 || spec->dmac_15_0);
 }
 
-static bool dr_mask_is_src_addr_set(struct mlx5dr_match_spec *spec)
-{
-	return (spec->src_ip_127_96 || spec->src_ip_95_64 ||
-		spec->src_ip_63_32 || spec->src_ip_31_0);
-}
-
-static bool dr_mask_is_dst_addr_set(struct mlx5dr_match_spec *spec)
-{
-	return (spec->dst_ip_127_96 || spec->dst_ip_95_64 ||
-		spec->dst_ip_63_32 || spec->dst_ip_31_0);
-}
-
 static bool dr_mask_is_l3_base_set(struct mlx5dr_match_spec *spec)
 {
 	return (spec->ip_protocol || spec->frag || spec->tcp_flags ||
@@ -503,11 +491,11 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher,
 						    &mask, inner, rx);
 
 		if (outer_ipv == DR_RULE_IPV6) {
-			if (dr_mask_is_dst_addr_set(&mask.outer))
+			if (DR_MASK_IS_DST_IP_SET(&mask.outer))
 				mlx5dr_ste_build_eth_l3_ipv6_dst(ste_ctx, &sb[idx++],
 								 &mask, inner, rx);
 
-			if (dr_mask_is_src_addr_set(&mask.outer))
+			if (DR_MASK_IS_SRC_IP_SET(&mask.outer))
 				mlx5dr_ste_build_eth_l3_ipv6_src(ste_ctx, &sb[idx++],
 								 &mask, inner, rx);
 
@@ -610,11 +598,11 @@ static int dr_matcher_set_ste_builders(struct mlx5dr_matcher *matcher,
 						    &mask, inner, rx);
 
 		if (inner_ipv == DR_RULE_IPV6) {
-			if (dr_mask_is_dst_addr_set(&mask.inner))
+			if (DR_MASK_IS_DST_IP_SET(&mask.inner))
 				mlx5dr_ste_build_eth_l3_ipv6_dst(ste_ctx, &sb[idx++],
 								 &mask, inner, rx);
 
-			if (dr_mask_is_src_addr_set(&mask.inner))
+			if (DR_MASK_IS_SRC_IP_SET(&mask.inner))
 				mlx5dr_ste_build_eth_l3_ipv6_src(ste_ctx, &sb[idx++],
 								 &mask, inner, rx);
 

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c

@@ -602,12 +602,34 @@ int mlx5dr_ste_set_action_decap_l3_list(struct mlx5dr_ste_ctx *ste_ctx,
 						 used_hw_action_num);
 }
 
+static int dr_ste_build_pre_check_spec(struct mlx5dr_domain *dmn,
+				       struct mlx5dr_match_spec *spec)
+{
+	if (spec->ip_version) {
+		if (spec->ip_version != 0xf) {
+			mlx5dr_err(dmn,
+				   "Partial ip_version mask with src/dst IP is not supported\n");
+			return -EINVAL;
+		}
+	} else if (spec->ethertype != 0xffff &&
+		   (DR_MASK_IS_SRC_IP_SET(spec) || DR_MASK_IS_DST_IP_SET(spec))) {
+		mlx5dr_err(dmn,
+			   "Partial/no ethertype mask with src/dst IP is not supported\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 int mlx5dr_ste_build_pre_check(struct mlx5dr_domain *dmn,
 			       u8 match_criteria,
 			       struct mlx5dr_match_param *mask,
 			       struct mlx5dr_match_param *value)
 {
-	if (!value && (match_criteria & DR_MATCHER_CRITERIA_MISC)) {
+	if (value)
+		return 0;
+
+	if (match_criteria & DR_MATCHER_CRITERIA_MISC) {
 		if (mask->misc.source_port && mask->misc.source_port != 0xffff) {
 			mlx5dr_err(dmn,
 				   "Partial mask source_port is not supported\n");
@@ -621,6 +643,14 @@ int mlx5dr_ste_build_pre_check(struct mlx5dr_domain *dmn,
 		}
 	}
 
+	if ((match_criteria & DR_MATCHER_CRITERIA_OUTER) &&
+	    dr_ste_build_pre_check_spec(dmn, &mask->outer))
+		return -EINVAL;
+
+	if ((match_criteria & DR_MATCHER_CRITERIA_INNER) &&
+	    dr_ste_build_pre_check_spec(dmn, &mask->inner))
+		return -EINVAL;
+
 	return 0;
 }
 

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h

@@ -798,6 +798,16 @@ struct mlx5dr_match_param {
 				       (_misc3)->icmpv4_code || \
 				       (_misc3)->icmpv4_header_data)
 
+#define DR_MASK_IS_SRC_IP_SET(_spec) ((_spec)->src_ip_127_96 || \
+				      (_spec)->src_ip_95_64  || \
+				      (_spec)->src_ip_63_32  || \
+				      (_spec)->src_ip_31_0)
+
+#define DR_MASK_IS_DST_IP_SET(_spec) ((_spec)->dst_ip_127_96 || \
+				      (_spec)->dst_ip_95_64  || \
+				      (_spec)->dst_ip_63_32  || \
+				      (_spec)->dst_ip_31_0)
+
 struct mlx5dr_esw_caps {
 	u64 drop_icm_address_rx;
 	u64 drop_icm_address_tx;

drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c

@@ -233,7 +233,11 @@ static bool contain_vport_reformat_action(struct mlx5_flow_rule *dst)
 		dst->dest_attr.vport.flags & MLX5_FLOW_DEST_VPORT_REFORMAT_ID;
 }
 
-#define MLX5_FLOW_CONTEXT_ACTION_MAX  32
+/* We want to support a rule with 32 destinations, which means we need to
+ * account for 32 destinations plus usually a counter plus one more action
+ * for a multi-destination flow table.
+ */
+#define MLX5_FLOW_CONTEXT_ACTION_MAX  34
 static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 				  struct mlx5_flow_table *ft,
 				  struct mlx5_flow_group *group,
@@ -403,9 +407,9 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 			enum mlx5_flow_destination_type type = dst->dest_attr.type;
 			u32 id;
 
-			if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX ||
-			    num_term_actions >= MLX5_FLOW_CONTEXT_ACTION_MAX) {
-				err = -ENOSPC;
+			if (fs_dr_num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX ||
+			    num_term_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
+				err = -EOPNOTSUPP;
 				goto free_actions;
 			}
 
@@ -478,8 +482,9 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 			    MLX5_FLOW_DESTINATION_TYPE_COUNTER)
 				continue;
 
-			if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
-				err = -ENOSPC;
+			if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX ||
+			    fs_dr_num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
+				err = -EOPNOTSUPP;
 				goto free_actions;
 			}
 
@@ -499,14 +504,28 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 	params.match_sz = match_sz;
 	params.match_buf = (u64 *)fte->val;
 	if (num_term_actions == 1) {
-		if (term_actions->reformat)
+		if (term_actions->reformat) {
+			if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
+				err = -EOPNOTSUPP;
+				goto free_actions;
+			}
 			actions[num_actions++] = term_actions->reformat;
+		}
 
+		if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
+			err = -EOPNOTSUPP;
+			goto free_actions;
+		}
 		actions[num_actions++] = term_actions->dest;
 	} else if (num_term_actions > 1) {
 		bool ignore_flow_level =
 			!!(fte->action.flags & FLOW_ACT_IGNORE_FLOW_LEVEL);
 
+		if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX ||
+		    fs_dr_num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
+			err = -EOPNOTSUPP;
+			goto free_actions;
+		}
 		tmp_action = mlx5dr_action_create_mult_dest_tbl(domain,
 								term_actions,
 								num_term_actions,

drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h

@@ -160,6 +160,11 @@ struct mlx5dr_icm_buddy_mem {
 	 * sync_ste command sets them free.
 	 */
 	struct list_head	hot_list;
+
+	/* Memory optimisation */
+	struct mlx5dr_ste	*ste_arr;
+	struct list_head	*miss_list;
+	u8			*hw_ste_arr;
 };
 
 int mlx5dr_buddy_init(struct mlx5dr_icm_buddy_mem *buddy,