Merge branch 'rate-management-on-traffic-classes-misc'

Tariq Toukan says:

====================
mlx5 misc

Patches 1-3 by William reduce the memory consumption for representors to
achieve better scalability.

Patches 4-5 by Akiva expose ICM memory consumption per function.

Patches 6-8 expose helpful information on RSS resources in devlink RX
reporter diagnose.

Patches 9-10 are simple enhancements by Alex Lazar.
====================

Link: https://patch.msgid.link/20250209101716.112774-1-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Author: kuba-moo

Documentation/networking/devlink/mlx5.rst

@@ -280,6 +280,10 @@ Description of the vnic counters:
 	number of packets handled by the VNIC experiencing unexpected steering
 	failure (at any point in steering flow owned by the VNIC, including the FDB
 	for the eswitch owner).
+- icm_consumption
+        amount of Interconnect Host Memory (ICM) consumed by the vnic in
+        granularity of 4KB. ICM is host memory allocated by SW upon HCA request
+        and is used for storing data structures that control HCA operation.
 
 User commands examples:
 

drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.c

@@ -13,6 +13,50 @@ struct mlx5_vnic_diag_stats {
 	__be64 query_vnic_env_out[MLX5_ST_SZ_QW(query_vnic_env_out)];
 };
 
+static void mlx5_reporter_vnic_diagnose_counter_icm(struct mlx5_core_dev *dev,
+						    struct devlink_fmsg *fmsg,
+						    u16 vport_num, bool other_vport)
+{
+	u32 out_icm_reg[MLX5_ST_SZ_DW(vhca_icm_ctrl_reg)] = {};
+	u32 in_icm_reg[MLX5_ST_SZ_DW(vhca_icm_ctrl_reg)] = {};
+	u32 out_reg[MLX5_ST_SZ_DW(nic_cap_reg)] = {};
+	u32 in_reg[MLX5_ST_SZ_DW(nic_cap_reg)] = {};
+	u32 cur_alloc_icm;
+	int vhca_icm_ctrl;
+	u16 vhca_id;
+	int err;
+
+	err = mlx5_core_access_reg(dev, in_reg, sizeof(in_reg), out_reg,
+				   sizeof(out_reg), MLX5_REG_NIC_CAP, 0, 0);
+	if (err) {
+		mlx5_core_warn(dev, "Reading nic_cap_reg failed. err = %d\n", err);
+		return;
+	}
+	vhca_icm_ctrl = MLX5_GET(nic_cap_reg, out_reg, vhca_icm_ctrl);
+	if (!vhca_icm_ctrl)
+		return;
+
+	MLX5_SET(vhca_icm_ctrl_reg, in_icm_reg, vhca_id_valid, other_vport);
+	if (other_vport) {
+		err = mlx5_vport_get_vhca_id(dev, vport_num, &vhca_id);
+		if (err) {
+			mlx5_core_warn(dev, "vport to vhca_id failed. vport_num = %d, err = %d\n",
+				       vport_num, err);
+			return;
+		}
+		MLX5_SET(vhca_icm_ctrl_reg, in_icm_reg, vhca_id, vhca_id);
+	}
+	err = mlx5_core_access_reg(dev, in_icm_reg, sizeof(in_icm_reg),
+				   out_icm_reg, sizeof(out_icm_reg),
+				   MLX5_REG_VHCA_ICM_CTRL, 0, 0);
+	if (err) {
+		mlx5_core_warn(dev, "Reading vhca_icm_ctrl failed. err = %d\n", err);
+		return;
+	}
+	cur_alloc_icm = MLX5_GET(vhca_icm_ctrl_reg, out_icm_reg, cur_alloc_icm);
+	devlink_fmsg_u32_pair_put(fmsg, "icm_consumption", cur_alloc_icm);
+}
+
 void mlx5_reporter_vnic_diagnose_counters(struct mlx5_core_dev *dev,
 					  struct devlink_fmsg *fmsg,
 					  u16 vport_num, bool other_vport)
@@ -59,6 +103,8 @@ void mlx5_reporter_vnic_diagnose_counters(struct mlx5_core_dev *dev,
 		devlink_fmsg_u64_pair_put(fmsg, "handled_pkt_steering_fail",
 					  VNIC_ENV_GET64(&vnic, handled_pkt_steering_fail));
 	}
+	if (MLX5_CAP_GEN(dev, nic_cap_reg))
+		mlx5_reporter_vnic_diagnose_counter_icm(dev, fmsg, vport_num, other_vport);
 
 	devlink_fmsg_obj_nest_end(fmsg);
 	devlink_fmsg_pair_nest_end(fmsg);

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

@@ -95,8 +95,6 @@ struct page_pool;
 #define MLX5_MPWRQ_DEF_LOG_STRIDE_SZ(mdev) \
 	MLX5_MPWRQ_LOG_STRIDE_SZ(mdev, order_base_2(MLX5E_RX_MAX_HEAD))
 
-#define MLX5_MPWRQ_MAX_LOG_WQE_SZ 18
-
 /* Keep in sync with mlx5e_mpwrq_log_wqe_sz.
  * These are theoretical maximums, which can be further restricted by
  * capabilities. These values are used for static resource allocations and
@@ -386,7 +384,6 @@ enum {
 	MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE,
 	MLX5E_SQ_STATE_PENDING_XSK_TX,
 	MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC,
-	MLX5E_SQ_STATE_XDP_MULTIBUF,
 	MLX5E_NUM_SQ_STATES, /* Must be kept last */
 };
 

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

@@ -10,6 +10,9 @@
 #include <net/page_pool/types.h>
 #include <net/xdp_sock_drv.h>
 
+#define MLX5_MPWRQ_MAX_LOG_WQE_SZ 18
+#define MLX5_REP_MPWRQ_MAX_LOG_WQE_SZ 17
+
 static u8 mlx5e_mpwrq_min_page_shift(struct mlx5_core_dev *mdev)
 {
 	u8 min_page_shift = MLX5_CAP_GEN_2(mdev, log_min_mkey_entity_size);
@@ -103,18 +106,22 @@ u8 mlx5e_mpwrq_log_wqe_sz(struct mlx5_core_dev *mdev, u8 page_shift,
 			  enum mlx5e_mpwrq_umr_mode umr_mode)
 {
 	u8 umr_entry_size = mlx5e_mpwrq_umr_entry_size(umr_mode);
-	u8 max_pages_per_wqe, max_log_mpwqe_size;
+	u8 max_pages_per_wqe, max_log_wqe_size_calc;
+	u8 max_log_wqe_size_cap;
 	u16 max_wqe_size;
 
 	/* Keep in sync with MLX5_MPWRQ_MAX_PAGES_PER_WQE. */
 	max_wqe_size = mlx5e_get_max_sq_aligned_wqebbs(mdev) * MLX5_SEND_WQE_BB;
 	max_pages_per_wqe = ALIGN_DOWN(max_wqe_size - sizeof(struct mlx5e_umr_wqe),
 				       MLX5_UMR_FLEX_ALIGNMENT) / umr_entry_size;
-	max_log_mpwqe_size = ilog2(max_pages_per_wqe) + page_shift;
+	max_log_wqe_size_calc = ilog2(max_pages_per_wqe) + page_shift;
+
+	WARN_ON_ONCE(max_log_wqe_size_calc < MLX5E_ORDER2_MAX_PACKET_MTU);
 
-	WARN_ON_ONCE(max_log_mpwqe_size < MLX5E_ORDER2_MAX_PACKET_MTU);
+	max_log_wqe_size_cap = mlx5_core_is_ecpf(mdev) ?
+			   MLX5_REP_MPWRQ_MAX_LOG_WQE_SZ : MLX5_MPWRQ_MAX_LOG_WQE_SZ;
 
-	return min_t(u8, max_log_mpwqe_size, MLX5_MPWRQ_MAX_LOG_WQE_SZ);
+	return min_t(u8, max_log_wqe_size_calc, max_log_wqe_size_cap);
 }
 
 u8 mlx5e_mpwrq_pages_per_wqe(struct mlx5_core_dev *mdev, u8 page_shift,
@@ -1240,7 +1247,6 @@ void mlx5e_build_xdpsq_param(struct mlx5_core_dev *mdev,
 	mlx5e_build_sq_param_common(mdev, param);
 	MLX5_SET(wq, wq, log_wq_sz, params->log_sq_size);
 	param->is_mpw = MLX5E_GET_PFLAG(params, MLX5E_PFLAG_XDP_TX_MPWQE);
-	param->is_xdp_mb = !mlx5e_rx_is_linear_skb(mdev, params, xsk);
 	mlx5e_build_tx_cq_param(mdev, params, &param->cqp);
 }
 

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

@@ -33,7 +33,6 @@ struct mlx5e_sq_param {
 	struct mlx5_wq_param       wq;
 	bool                       is_mpw;
 	bool                       is_tls;
-	bool                       is_xdp_mb;
 	u16                        stop_room;
 };
 

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

@@ -317,10 +317,8 @@ mlx5e_rx_reporter_diagnose_common_ptp_config(struct mlx5e_priv *priv, struct mlx
 }
 
 static void
-mlx5e_rx_reporter_diagnose_common_config(struct devlink_health_reporter *reporter,
-					 struct devlink_fmsg *fmsg)
+mlx5e_rx_reporter_diagnose_common_config(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg)
 {
-	struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
 	struct mlx5e_rq *generic_rq = &priv->channels.c[0]->rq;
 	struct mlx5e_ptp *ptp_ch = priv->channels.ptp;
 
@@ -340,20 +338,100 @@ static void mlx5e_rx_reporter_build_diagnose_output_ptp_rq(struct mlx5e_rq *rq,
 	devlink_fmsg_obj_nest_end(fmsg);
 }
 
-static int mlx5e_rx_reporter_diagnose(struct devlink_health_reporter *reporter,
-				      struct devlink_fmsg *fmsg,
-				      struct netlink_ext_ack *extack)
+static void mlx5e_rx_reporter_diagnose_rx_res_dir_tirns(struct mlx5e_rx_res *rx_res,
+							struct devlink_fmsg *fmsg)
 {
-	struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
-	struct mlx5e_ptp *ptp_ch = priv->channels.ptp;
+	unsigned int max_nch = mlx5e_rx_res_get_max_nch(rx_res);
 	int i;
 
-	mutex_lock(&priv->state_lock);
+	devlink_fmsg_arr_pair_nest_start(fmsg, "Direct TIRs");
 
-	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
-		goto unlock;
+	for (i = 0; i < max_nch; i++) {
+		devlink_fmsg_obj_nest_start(fmsg);
+
+		devlink_fmsg_u32_pair_put(fmsg, "ix", i);
+		devlink_fmsg_u32_pair_put(fmsg, "tirn", mlx5e_rx_res_get_tirn_direct(rx_res, i));
+		devlink_fmsg_u32_pair_put(fmsg, "rqtn", mlx5e_rx_res_get_rqtn_direct(rx_res, i));
+
+		devlink_fmsg_obj_nest_end(fmsg);
+	}
+
+	devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+
+static void mlx5e_rx_reporter_diagnose_rx_res_rss_tirn(struct mlx5e_rss *rss, bool inner,
+						       struct devlink_fmsg *fmsg)
+{
+	bool found_valid_tir = false;
+	int tt;
+
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		if (!mlx5e_rss_valid_tir(rss, tt, inner))
+			continue;
+
+		if (!found_valid_tir) {
+			char *tir_msg = inner ? "Inner TIRs Numbers" : "TIRs Numbers";
+
+			found_valid_tir = true;
+			devlink_fmsg_arr_pair_nest_start(fmsg, tir_msg);
+		}
+
+		devlink_fmsg_obj_nest_start(fmsg);
+		devlink_fmsg_string_pair_put(fmsg, "tt", mlx5_ttc_get_name(tt));
+		devlink_fmsg_u32_pair_put(fmsg, "tirn", mlx5e_rss_get_tirn(rss, tt, inner));
+		devlink_fmsg_obj_nest_end(fmsg);
+	}
+
+	if (found_valid_tir)
+		devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+
+static void mlx5e_rx_reporter_diagnose_rx_res_rss_ix(struct mlx5e_rx_res *rx_res, u32 rss_idx,
+						     struct devlink_fmsg *fmsg)
+{
+	struct mlx5e_rss *rss = mlx5e_rx_res_rss_get(rx_res, rss_idx);
+
+	if (!rss)
+		return;
+
+	devlink_fmsg_obj_nest_start(fmsg);
+
+	devlink_fmsg_u32_pair_put(fmsg, "Index", rss_idx);
+	devlink_fmsg_u32_pair_put(fmsg, "rqtn", mlx5e_rss_get_rqtn(rss));
+	mlx5e_rx_reporter_diagnose_rx_res_rss_tirn(rss, false, fmsg);
+	if (mlx5e_rss_get_inner_ft_support(rss))
+		mlx5e_rx_reporter_diagnose_rx_res_rss_tirn(rss, true, fmsg);
+
+	devlink_fmsg_obj_nest_end(fmsg);
+}
+
+static void mlx5e_rx_reporter_diagnose_rx_res_rss(struct mlx5e_rx_res *rx_res,
+						  struct devlink_fmsg *fmsg)
+{
+	int rss_ix;
+
+	devlink_fmsg_arr_pair_nest_start(fmsg, "RSS");
+	for (rss_ix = 0; rss_ix < MLX5E_MAX_NUM_RSS; rss_ix++)
+		mlx5e_rx_reporter_diagnose_rx_res_rss_ix(rx_res, rss_ix, fmsg);
+	devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+
+static void mlx5e_rx_reporter_diagnose_rx_res(struct mlx5e_priv *priv,
+					      struct devlink_fmsg *fmsg)
+{
+	struct mlx5e_rx_res *rx_res = priv->rx_res;
+
+	mlx5e_health_fmsg_named_obj_nest_start(fmsg, "RX resources");
+	mlx5e_rx_reporter_diagnose_rx_res_dir_tirns(rx_res, fmsg);
+	mlx5e_rx_reporter_diagnose_rx_res_rss(rx_res, fmsg);
+	mlx5e_health_fmsg_named_obj_nest_end(fmsg);
+}
+
+static void mlx5e_rx_reporter_diagnose_rqs(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg)
+{
+	struct mlx5e_ptp *ptp_ch = priv->channels.ptp;
+	int i;
 
-	mlx5e_rx_reporter_diagnose_common_config(reporter, fmsg);
 	devlink_fmsg_arr_pair_nest_start(fmsg, "RQs");
 
 	for (i = 0; i < priv->channels.num; i++) {
@@ -367,7 +445,24 @@ static int mlx5e_rx_reporter_diagnose(struct devlink_health_reporter *reporter,
 	}
 	if (ptp_ch && test_bit(MLX5E_PTP_STATE_RX, ptp_ch->state))
 		mlx5e_rx_reporter_build_diagnose_output_ptp_rq(&ptp_ch->rq, fmsg);
+
 	devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+
+static int mlx5e_rx_reporter_diagnose(struct devlink_health_reporter *reporter,
+				      struct devlink_fmsg *fmsg,
+				      struct netlink_ext_ack *extack)
+{
+	struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
+
+	mutex_lock(&priv->state_lock);
+
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		goto unlock;
+
+	mlx5e_rx_reporter_diagnose_common_config(priv, fmsg);
+	mlx5e_rx_reporter_diagnose_rqs(priv, fmsg);
+	mlx5e_rx_reporter_diagnose_rx_res(priv, fmsg);
 unlock:
 	mutex_unlock(&priv->state_lock);
 	return 0;

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

@@ -16,7 +16,6 @@ static const char * const sq_sw_state_type_name[] = {
 	[MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE] = "vlan_need_l2_inline",
 	[MLX5E_SQ_STATE_PENDING_XSK_TX] = "pending_xsk_tx",
 	[MLX5E_SQ_STATE_PENDING_TLS_RX_RESYNC] = "pending_tls_rx_resync",
-	[MLX5E_SQ_STATE_XDP_MULTIBUF] = "xdp_multibuf",
 };
 
 static int mlx5e_wait_for_sq_flush(struct mlx5e_txqsq *sq)

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

@@ -81,6 +81,11 @@ struct mlx5e_rss {
 	refcount_t refcnt;
 };
 
+bool mlx5e_rss_get_inner_ft_support(struct mlx5e_rss *rss)
+{
+	return rss->inner_ft_support;
+}
+
 void mlx5e_rss_params_indir_modify_actual_size(struct mlx5e_rss *rss, u32 num_channels)
 {
 	rss->indir.actual_table_size = mlx5e_rqt_size(rss->mdev, num_channels);
@@ -449,6 +454,16 @@ u32 mlx5e_rss_get_tirn(struct mlx5e_rss *rss, enum mlx5_traffic_types tt,
 	return mlx5e_tir_get_tirn(tir);
 }
 
+u32 mlx5e_rss_get_rqtn(struct mlx5e_rss *rss)
+{
+	return mlx5e_rqt_get_rqtn(&rss->rqt);
+}
+
+bool mlx5e_rss_valid_tir(struct mlx5e_rss *rss, enum mlx5_traffic_types tt, bool inner)
+{
+	return !!rss_get_tir(rss, tt, inner);
+}
+
 /* Fill the "tirn" output parameter.
  * Create the requested TIR if it's its first usage.
  */

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

@@ -32,8 +32,11 @@ void mlx5e_rss_refcnt_inc(struct mlx5e_rss *rss);
 void mlx5e_rss_refcnt_dec(struct mlx5e_rss *rss);
 unsigned int mlx5e_rss_refcnt_read(struct mlx5e_rss *rss);
 
+bool mlx5e_rss_get_inner_ft_support(struct mlx5e_rss *rss);
 u32 mlx5e_rss_get_tirn(struct mlx5e_rss *rss, enum mlx5_traffic_types tt,
 		       bool inner);
+bool mlx5e_rss_valid_tir(struct mlx5e_rss *rss, enum mlx5_traffic_types tt, bool inner);
+u32 mlx5e_rss_get_rqtn(struct mlx5e_rss *rss);
 int mlx5e_rss_obtain_tirn(struct mlx5e_rss *rss,
 			  enum mlx5_traffic_types tt,
 			  const struct mlx5e_packet_merge_param *init_pkt_merge_param,

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

@@ -5,8 +5,6 @@
 #include "channels.h"
 #include "params.h"
 
-#define MLX5E_MAX_NUM_RSS 16
-
 struct mlx5e_rx_res {
 	struct mlx5_core_dev *mdev; /* primary */
 	enum mlx5e_rx_res_features features;
@@ -497,6 +495,11 @@ void mlx5e_rx_res_destroy(struct mlx5e_rx_res *res)
 	mlx5e_rx_res_free(res);
 }
 
+unsigned int mlx5e_rx_res_get_max_nch(struct mlx5e_rx_res *res)
+{
+	return res->max_nch;
+}
+
 u32 mlx5e_rx_res_get_tirn_direct(struct mlx5e_rx_res *res, unsigned int ix)
 {
 	return mlx5e_tir_get_tirn(&res->channels[ix].direct_tir);
@@ -522,7 +525,7 @@ u32 mlx5e_rx_res_get_tirn_ptp(struct mlx5e_rx_res *res)
 	return mlx5e_tir_get_tirn(&res->ptp.tir);
 }
 
-static u32 mlx5e_rx_res_get_rqtn_direct(struct mlx5e_rx_res *res, unsigned int ix)
+u32 mlx5e_rx_res_get_rqtn_direct(struct mlx5e_rx_res *res, unsigned int ix)
 {
 	return mlx5e_rqt_get_rqtn(&res->channels[ix].direct_rqt);
 }