-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2010-2014 Intel Corporation
*/
#include <stdio.h>
#include <rte_branch_prediction.h>
#include <rte_mbuf.h>
#include <rte_bitmap.h>
+#include <rte_reciprocal.h>
#include "rte_sched.h"
#include "rte_sched_common.h"
#include "rte_approx.h"
-#include "rte_reciprocal.h"
#ifdef __INTEL_COMPILER
#pragma warning(disable:2259) /* conversion may lose significant bits */
#define RTE_SCHED_TB_RATE_CONFIG_ERR (1e-7)
#define RTE_SCHED_WRR_SHIFT 3
+#define RTE_SCHED_MAX_QUEUES_PER_TC RTE_SCHED_BE_QUEUES_PER_PIPE
#define RTE_SCHED_GRINDER_PCACHE_SIZE (64 / RTE_SCHED_QUEUES_PER_PIPE)
#define RTE_SCHED_PIPE_INVALID UINT32_MAX
#define RTE_SCHED_BMP_POS_INVALID UINT32_MAX
*/
#define RTE_SCHED_TIME_SHIFT 8
-struct rte_sched_subport {
- /* Token bucket (TB) */
- uint64_t tb_time; /* time of last update */
- uint32_t tb_period;
- uint32_t tb_credits_per_period;
- uint32_t tb_size;
- uint32_t tb_credits;
-
- /* Traffic classes (TCs) */
- uint64_t tc_time; /* time of next update */
- uint32_t tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
- uint32_t tc_credits[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
- uint32_t tc_period;
-
- /* TC oversubscription */
- uint32_t tc_ov_wm;
- uint32_t tc_ov_wm_min;
- uint32_t tc_ov_wm_max;
- uint8_t tc_ov_period_id;
- uint8_t tc_ov;
- uint32_t tc_ov_n;
- double tc_ov_rate;
-
- /* Statistics */
- struct rte_sched_subport_stats stats;
-};
-
struct rte_sched_pipe_profile {
/* Token bucket (TB) */
- uint32_t tb_period;
- uint32_t tb_credits_per_period;
- uint32_t tb_size;
+ uint64_t tb_period;
+ uint64_t tb_credits_per_period;
+ uint64_t tb_size;
/* Pipe traffic classes */
- uint32_t tc_period;
- uint32_t tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint64_t tc_period;
+ uint64_t tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
uint8_t tc_ov_weight;
- /* Pipe queues */
- uint8_t wrr_cost[RTE_SCHED_QUEUES_PER_PIPE];
+ /* Pipe best-effort traffic class queues */
+ uint8_t wrr_cost[RTE_SCHED_BE_QUEUES_PER_PIPE];
};
struct rte_sched_pipe {
/* Token bucket (TB) */
uint64_t tb_time; /* time of last update */
- uint32_t tb_credits;
+ uint64_t tb_credits;
/* Pipe profile and flags */
uint32_t profile;
/* Traffic classes (TCs) */
uint64_t tc_time; /* time of next update */
- uint32_t tc_credits[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint64_t tc_credits[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
/* Weighted Round Robin (WRR) */
- uint8_t wrr_tokens[RTE_SCHED_QUEUES_PER_PIPE];
+ uint8_t wrr_tokens[RTE_SCHED_BE_QUEUES_PER_PIPE];
/* TC oversubscription */
- uint32_t tc_ov_credits;
+ uint64_t tc_ov_credits;
uint8_t tc_ov_period_id;
- uint8_t reserved[3];
} __rte_cache_aligned;
struct rte_sched_queue {
e_GRINDER_READ_MBUF
};
-/*
- * Path through the scheduler hierarchy used by the scheduler enqueue
- * operation to identify the destination queue for the current
- * packet. Stored in the field pkt.hash.sched of struct rte_mbuf of
- * each packet, typically written by the classification stage and read
- * by scheduler enqueue.
- */
-struct rte_sched_port_hierarchy {
- uint16_t queue:2; /**< Queue ID (0 .. 3) */
- uint16_t traffic_class:2; /**< Traffic class ID (0 .. 3)*/
- uint32_t color:2; /**< Color */
- uint16_t unused:10;
- uint16_t subport; /**< Subport ID */
- uint32_t pipe; /**< Pipe ID */
-};
-
struct rte_sched_grinder {
/* Pipe cache */
uint16_t pcache_qmask[RTE_SCHED_GRINDER_PCACHE_SIZE];
struct rte_sched_pipe_profile *pipe_params;
/* TC cache */
- uint8_t tccache_qmask[4];
- uint32_t tccache_qindex[4];
+ uint8_t tccache_qmask[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint32_t tccache_qindex[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
uint32_t tccache_w;
uint32_t tccache_r;
/* Current TC */
uint32_t tc_index;
- struct rte_sched_queue *queue[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
- struct rte_mbuf **qbase[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
- uint32_t qindex[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ struct rte_sched_queue *queue[RTE_SCHED_MAX_QUEUES_PER_TC];
+ struct rte_mbuf **qbase[RTE_SCHED_MAX_QUEUES_PER_TC];
+ uint32_t qindex[RTE_SCHED_MAX_QUEUES_PER_TC];
uint16_t qsize;
uint32_t qmask;
uint32_t qpos;
struct rte_mbuf *pkt;
/* WRR */
- uint16_t wrr_tokens[RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS];
- uint16_t wrr_mask[RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS];
- uint8_t wrr_cost[RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS];
+ uint16_t wrr_tokens[RTE_SCHED_BE_QUEUES_PER_PIPE];
+ uint16_t wrr_mask[RTE_SCHED_BE_QUEUES_PER_PIPE];
+ uint8_t wrr_cost[RTE_SCHED_BE_QUEUES_PER_PIPE];
};
-struct rte_sched_port {
- /* User parameters */
- uint32_t n_subports_per_port;
- uint32_t n_pipes_per_subport;
- uint32_t rate;
- uint32_t mtu;
- uint32_t frame_overhead;
- uint16_t qsize[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+struct rte_sched_subport {
+ /* Token bucket (TB) */
+ uint64_t tb_time; /* time of last update */
+ uint64_t tb_period;
+ uint64_t tb_credits_per_period;
+ uint64_t tb_size;
+ uint64_t tb_credits;
+
+ /* Traffic classes (TCs) */
+ uint64_t tc_time; /* time of next update */
+ uint64_t tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint64_t tc_credits[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint64_t tc_period;
+
+ /* TC oversubscription */
+ uint64_t tc_ov_wm;
+ uint64_t tc_ov_wm_min;
+ uint64_t tc_ov_wm_max;
+ uint8_t tc_ov_period_id;
+ uint8_t tc_ov;
+ uint32_t tc_ov_n;
+ double tc_ov_rate;
+
+ /* Statistics */
+ struct rte_sched_subport_stats stats __rte_cache_aligned;
+
+ /* Subport pipes */
+ uint32_t n_pipes_per_subport_enabled;
uint32_t n_pipe_profiles;
- uint32_t pipe_tc3_rate_max;
+ uint32_t n_max_pipe_profiles;
+
+ /* Pipe best-effort TC rate */
+ uint64_t pipe_tc_be_rate_max;
+
+ /* Pipe queues size */
+ uint16_t qsize[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+
#ifdef RTE_SCHED_RED
- struct rte_red_config red_config[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE][e_RTE_METER_COLORS];
+ struct rte_red_config red_config[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE][RTE_COLORS];
#endif
- /* Timing */
- uint64_t time_cpu_cycles; /* Current CPU time measured in CPU cyles */
- uint64_t time_cpu_bytes; /* Current CPU time measured in bytes */
- uint64_t time; /* Current NIC TX time measured in bytes */
- struct rte_reciprocal inv_cycles_per_byte; /* CPU cycles per byte */
-
/* Scheduling loop detection */
uint32_t pipe_loop;
uint32_t pipe_exhaustion;
/* Grinders */
struct rte_sched_grinder grinder[RTE_SCHED_PORT_N_GRINDERS];
uint32_t busy_grinders;
- struct rte_mbuf **pkts_out;
- uint32_t n_pkts_out;
/* Queue base calculation */
uint32_t qsize_add[RTE_SCHED_QUEUES_PER_PIPE];
uint32_t qsize_sum;
- /* Large data structures */
- struct rte_sched_subport *subport;
struct rte_sched_pipe *pipe;
struct rte_sched_queue *queue;
struct rte_sched_queue_extra *queue_extra;
uint8_t memory[0] __rte_cache_aligned;
} __rte_cache_aligned;
-enum rte_sched_port_array {
- e_RTE_SCHED_PORT_ARRAY_SUBPORT = 0,
- e_RTE_SCHED_PORT_ARRAY_PIPE,
- e_RTE_SCHED_PORT_ARRAY_QUEUE,
- e_RTE_SCHED_PORT_ARRAY_QUEUE_EXTRA,
- e_RTE_SCHED_PORT_ARRAY_PIPE_PROFILES,
- e_RTE_SCHED_PORT_ARRAY_BMP_ARRAY,
- e_RTE_SCHED_PORT_ARRAY_QUEUE_ARRAY,
- e_RTE_SCHED_PORT_ARRAY_TOTAL,
-};
+struct rte_sched_port {
+ /* User parameters */
+ uint32_t n_subports_per_port;
+ uint32_t n_pipes_per_subport;
+ uint32_t n_pipes_per_subport_log2;
+ uint16_t pipe_queue[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint8_t pipe_tc[RTE_SCHED_QUEUES_PER_PIPE];
+ uint8_t tc_queue[RTE_SCHED_QUEUES_PER_PIPE];
+ uint64_t rate;
+ uint32_t mtu;
+ uint32_t frame_overhead;
+ int socket;
-#ifdef RTE_SCHED_COLLECT_STATS
+ /* Timing */
+ uint64_t time_cpu_cycles; /* Current CPU time measured in CPU cyles */
+ uint64_t time_cpu_bytes; /* Current CPU time measured in bytes */
+ uint64_t time; /* Current NIC TX time measured in bytes */
+ struct rte_reciprocal inv_cycles_per_byte; /* CPU cycles per byte */
-static inline uint32_t
-rte_sched_port_queues_per_subport(struct rte_sched_port *port)
-{
- return RTE_SCHED_QUEUES_PER_PIPE * port->n_pipes_per_subport;
-}
+ /* Grinders */
+ struct rte_mbuf **pkts_out;
+ uint32_t n_pkts_out;
+ uint32_t subport_id;
-#endif
+ /* Large data structures */
+ struct rte_sched_subport *subports[0] __rte_cache_aligned;
+} __rte_cache_aligned;
+
+enum rte_sched_subport_array {
+ e_RTE_SCHED_SUBPORT_ARRAY_PIPE = 0,
+ e_RTE_SCHED_SUBPORT_ARRAY_QUEUE,
+ e_RTE_SCHED_SUBPORT_ARRAY_QUEUE_EXTRA,
+ e_RTE_SCHED_SUBPORT_ARRAY_PIPE_PROFILES,
+ e_RTE_SCHED_SUBPORT_ARRAY_BMP_ARRAY,
+ e_RTE_SCHED_SUBPORT_ARRAY_QUEUE_ARRAY,
+ e_RTE_SCHED_SUBPORT_ARRAY_TOTAL,
+};
static inline uint32_t
-rte_sched_port_queues_per_port(struct rte_sched_port *port)
+rte_sched_subport_pipe_queues(struct rte_sched_subport *subport)
{
- return RTE_SCHED_QUEUES_PER_PIPE * port->n_pipes_per_subport * port->n_subports_per_port;
+ return RTE_SCHED_QUEUES_PER_PIPE * subport->n_pipes_per_subport_enabled;
}
static inline struct rte_mbuf **
-rte_sched_port_qbase(struct rte_sched_port *port, uint32_t qindex)
+rte_sched_subport_pipe_qbase(struct rte_sched_subport *subport, uint32_t qindex)
{
uint32_t pindex = qindex >> 4;
- uint32_t qpos = qindex & 0xF;
+ uint32_t qpos = qindex & (RTE_SCHED_QUEUES_PER_PIPE - 1);
- return (port->queue_array + pindex *
- port->qsize_sum + port->qsize_add[qpos]);
+ return (subport->queue_array + pindex *
+ subport->qsize_sum + subport->qsize_add[qpos]);
}
static inline uint16_t
-rte_sched_port_qsize(struct rte_sched_port *port, uint32_t qindex)
+rte_sched_subport_pipe_qsize(struct rte_sched_port *port,
+struct rte_sched_subport *subport, uint32_t qindex)
{
- uint32_t tc = (qindex >> 2) & 0x3;
+ uint32_t tc = port->pipe_tc[qindex & (RTE_SCHED_QUEUES_PER_PIPE - 1)];
- return port->qsize[tc];
+ return subport->qsize[tc];
}
-static int
-rte_sched_port_check_params(struct rte_sched_port_params *params)
+static inline uint32_t
+rte_sched_port_queues_per_port(struct rte_sched_port *port)
{
- uint32_t i, j;
+ uint32_t n_queues = 0, i;
- if (params == NULL)
- return -1;
+ for (i = 0; i < port->n_subports_per_port; i++)
+ n_queues += rte_sched_subport_pipe_queues(port->subports[i]);
- /* socket */
- if ((params->socket < 0) || (params->socket >= RTE_MAX_NUMA_NODES))
- return -3;
+ return n_queues;
+}
- /* rate */
- if (params->rate == 0)
- return -4;
+static inline uint16_t
+rte_sched_port_pipe_queue(struct rte_sched_port *port, uint32_t traffic_class)
+{
+ uint16_t pipe_queue = port->pipe_queue[traffic_class];
- /* mtu */
- if (params->mtu == 0)
- return -5;
+ return pipe_queue;
+}
- /* n_subports_per_port: non-zero, limited to 16 bits, power of 2 */
- if (params->n_subports_per_port == 0 ||
- params->n_subports_per_port > 1u << 16 ||
- !rte_is_power_of_2(params->n_subports_per_port))
- return -6;
+static inline uint8_t
+rte_sched_port_pipe_tc(struct rte_sched_port *port, uint32_t qindex)
+{
+ uint8_t pipe_tc = port->pipe_tc[qindex & (RTE_SCHED_QUEUES_PER_PIPE - 1)];
- /* n_pipes_per_subport: non-zero, power of 2 */
- if (params->n_pipes_per_subport == 0 ||
- !rte_is_power_of_2(params->n_pipes_per_subport))
- return -7;
+ return pipe_tc;
+}
- /* qsize: non-zero, power of 2,
- * no bigger than 32K (due to 16-bit read/write pointers)
- */
- for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
- uint16_t qsize = params->qsize[i];
+static inline uint8_t
+rte_sched_port_tc_queue(struct rte_sched_port *port, uint32_t qindex)
+{
+ uint8_t tc_queue = port->tc_queue[qindex & (RTE_SCHED_QUEUES_PER_PIPE - 1)];
- if (qsize == 0 || !rte_is_power_of_2(qsize))
- return -8;
- }
+ return tc_queue;
+}
- /* pipe_profiles and n_pipe_profiles */
- if (params->pipe_profiles == NULL ||
- params->n_pipe_profiles == 0 ||
- params->n_pipe_profiles > RTE_SCHED_PIPE_PROFILES_PER_PORT)
- return -9;
+static int
+pipe_profile_check(struct rte_sched_pipe_params *params,
+ uint32_t rate, uint16_t *qsize)
+{
+ uint32_t i;
- for (i = 0; i < params->n_pipe_profiles; i++) {
- struct rte_sched_pipe_params *p = params->pipe_profiles + i;
+ /* Pipe parameters */
+ if (params == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter params\n", __func__);
+ return -EINVAL;
+ }
- /* TB rate: non-zero, not greater than port rate */
- if (p->tb_rate == 0 || p->tb_rate > params->rate)
- return -10;
+ /* TB rate: non-zero, not greater than port rate */
+ if (params->tb_rate == 0 ||
+ params->tb_rate > rate) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tb rate\n", __func__);
+ return -EINVAL;
+ }
- /* TB size: non-zero */
- if (p->tb_size == 0)
- return -11;
+ /* TB size: non-zero */
+ if (params->tb_size == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tb size\n", __func__);
+ return -EINVAL;
+ }
- /* TC rate: non-zero, less than pipe rate */
- for (j = 0; j < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; j++) {
- if (p->tc_rate[j] == 0 || p->tc_rate[j] > p->tb_rate)
- return -12;
+ /* TC rate: non-zero if qsize non-zero, less than pipe rate */
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
+ if ((qsize[i] == 0 && params->tc_rate[i] != 0) ||
+ (qsize[i] != 0 && (params->tc_rate[i] == 0 ||
+ params->tc_rate[i] > params->tb_rate))) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for qsize or tc_rate\n", __func__);
+ return -EINVAL;
}
+ }
- /* TC period: non-zero */
- if (p->tc_period == 0)
- return -13;
+ if (params->tc_rate[RTE_SCHED_TRAFFIC_CLASS_BE] == 0 ||
+ qsize[RTE_SCHED_TRAFFIC_CLASS_BE] == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for be traffic class rate\n", __func__);
+ return -EINVAL;
+ }
-#ifdef RTE_SCHED_SUBPORT_TC_OV
- /* TC3 oversubscription weight: non-zero */
- if (p->tc_ov_weight == 0)
- return -14;
-#endif
+ /* TC period: non-zero */
+ if (params->tc_period == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tc period\n", __func__);
+ return -EINVAL;
+ }
+
+ /* Best effort tc oversubscription weight: non-zero */
+ if (params->tc_ov_weight == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tc ov weight\n", __func__);
+ return -EINVAL;
+ }
- /* Queue WRR weights: non-zero */
- for (j = 0; j < RTE_SCHED_QUEUES_PER_PIPE; j++) {
- if (p->wrr_weights[j] == 0)
- return -15;
+ /* Queue WRR weights: non-zero */
+ for (i = 0; i < RTE_SCHED_BE_QUEUES_PER_PIPE; i++) {
+ if (params->wrr_weights[i] == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for wrr weight\n", __func__);
+ return -EINVAL;
}
}
return 0;
}
+static int
+rte_sched_port_check_params(struct rte_sched_port_params *params)
+{
+ if (params == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter params\n", __func__);
+ return -EINVAL;
+ }
+
+ /* socket */
+ if (params->socket < 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for socket id\n", __func__);
+ return -EINVAL;
+ }
+
+ /* rate */
+ if (params->rate == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for rate\n", __func__);
+ return -EINVAL;
+ }
+
+ /* mtu */
+ if (params->mtu == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for mtu\n", __func__);
+ return -EINVAL;
+ }
+
+ /* n_subports_per_port: non-zero, limited to 16 bits, power of 2 */
+ if (params->n_subports_per_port == 0 ||
+ params->n_subports_per_port > 1u << 16 ||
+ !rte_is_power_of_2(params->n_subports_per_port)) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for number of subports\n", __func__);
+ return -EINVAL;
+ }
+
+ /* n_pipes_per_subport: non-zero, power of 2 */
+ if (params->n_pipes_per_subport == 0 ||
+ !rte_is_power_of_2(params->n_pipes_per_subport)) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for maximum pipes number\n", __func__);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
static uint32_t
-rte_sched_port_get_array_base(struct rte_sched_port_params *params, enum rte_sched_port_array array)
+rte_sched_subport_get_array_base(struct rte_sched_subport_params *params,
+ enum rte_sched_subport_array array)
{
- uint32_t n_subports_per_port = params->n_subports_per_port;
- uint32_t n_pipes_per_subport = params->n_pipes_per_subport;
- uint32_t n_pipes_per_port = n_pipes_per_subport * n_subports_per_port;
- uint32_t n_queues_per_port = RTE_SCHED_QUEUES_PER_PIPE * n_pipes_per_subport * n_subports_per_port;
+ uint32_t n_pipes_per_subport = params->n_pipes_per_subport_enabled;
+ uint32_t n_subport_pipe_queues =
+ RTE_SCHED_QUEUES_PER_PIPE * n_pipes_per_subport;
- uint32_t size_subport = n_subports_per_port * sizeof(struct rte_sched_subport);
- uint32_t size_pipe = n_pipes_per_port * sizeof(struct rte_sched_pipe);
- uint32_t size_queue = n_queues_per_port * sizeof(struct rte_sched_queue);
+ uint32_t size_pipe = n_pipes_per_subport * sizeof(struct rte_sched_pipe);
+ uint32_t size_queue =
+ n_subport_pipe_queues * sizeof(struct rte_sched_queue);
uint32_t size_queue_extra
- = n_queues_per_port * sizeof(struct rte_sched_queue_extra);
- uint32_t size_pipe_profiles
- = RTE_SCHED_PIPE_PROFILES_PER_PORT * sizeof(struct rte_sched_pipe_profile);
- uint32_t size_bmp_array = rte_bitmap_get_memory_footprint(n_queues_per_port);
+ = n_subport_pipe_queues * sizeof(struct rte_sched_queue_extra);
+ uint32_t size_pipe_profiles = params->n_max_pipe_profiles *
+ sizeof(struct rte_sched_pipe_profile);
+ uint32_t size_bmp_array =
+ rte_bitmap_get_memory_footprint(n_subport_pipe_queues);
uint32_t size_per_pipe_queue_array, size_queue_array;
uint32_t base, i;
size_per_pipe_queue_array = 0;
for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
- size_per_pipe_queue_array += RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS
- * params->qsize[i] * sizeof(struct rte_mbuf *);
+ if (i < RTE_SCHED_TRAFFIC_CLASS_BE)
+ size_per_pipe_queue_array +=
+ params->qsize[i] * sizeof(struct rte_mbuf *);
+ else
+ size_per_pipe_queue_array += RTE_SCHED_MAX_QUEUES_PER_TC *
+ params->qsize[i] * sizeof(struct rte_mbuf *);
}
- size_queue_array = n_pipes_per_port * size_per_pipe_queue_array;
+ size_queue_array = n_pipes_per_subport * size_per_pipe_queue_array;
base = 0;
- if (array == e_RTE_SCHED_PORT_ARRAY_SUBPORT)
- return base;
- base += RTE_CACHE_LINE_ROUNDUP(size_subport);
-
- if (array == e_RTE_SCHED_PORT_ARRAY_PIPE)
+ if (array == e_RTE_SCHED_SUBPORT_ARRAY_PIPE)
return base;
base += RTE_CACHE_LINE_ROUNDUP(size_pipe);
- if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE)
+ if (array == e_RTE_SCHED_SUBPORT_ARRAY_QUEUE)
return base;
base += RTE_CACHE_LINE_ROUNDUP(size_queue);
- if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE_EXTRA)
+ if (array == e_RTE_SCHED_SUBPORT_ARRAY_QUEUE_EXTRA)
return base;
base += RTE_CACHE_LINE_ROUNDUP(size_queue_extra);
- if (array == e_RTE_SCHED_PORT_ARRAY_PIPE_PROFILES)
+ if (array == e_RTE_SCHED_SUBPORT_ARRAY_PIPE_PROFILES)
return base;
base += RTE_CACHE_LINE_ROUNDUP(size_pipe_profiles);
- if (array == e_RTE_SCHED_PORT_ARRAY_BMP_ARRAY)
+ if (array == e_RTE_SCHED_SUBPORT_ARRAY_BMP_ARRAY)
return base;
base += RTE_CACHE_LINE_ROUNDUP(size_bmp_array);
- if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE_ARRAY)
+ if (array == e_RTE_SCHED_SUBPORT_ARRAY_QUEUE_ARRAY)
return base;
base += RTE_CACHE_LINE_ROUNDUP(size_queue_array);
return base;
}
-uint32_t
-rte_sched_port_get_memory_footprint(struct rte_sched_port_params *params)
-{
- uint32_t size0, size1;
- int status;
-
- status = rte_sched_port_check_params(params);
- if (status != 0) {
- RTE_LOG(NOTICE, SCHED,
- "Port scheduler params check failed (%d)\n", status);
-
- return 0;
- }
-
- size0 = sizeof(struct rte_sched_port);
- size1 = rte_sched_port_get_array_base(params, e_RTE_SCHED_PORT_ARRAY_TOTAL);
-
- return size0 + size1;
-}
-
static void
-rte_sched_port_config_qsize(struct rte_sched_port *port)
+rte_sched_subport_config_qsize(struct rte_sched_subport *subport)
{
- /* TC 0 */
- port->qsize_add[0] = 0;
- port->qsize_add[1] = port->qsize_add[0] + port->qsize[0];
- port->qsize_add[2] = port->qsize_add[1] + port->qsize[0];
- port->qsize_add[3] = port->qsize_add[2] + port->qsize[0];
-
- /* TC 1 */
- port->qsize_add[4] = port->qsize_add[3] + port->qsize[0];
- port->qsize_add[5] = port->qsize_add[4] + port->qsize[1];
- port->qsize_add[6] = port->qsize_add[5] + port->qsize[1];
- port->qsize_add[7] = port->qsize_add[6] + port->qsize[1];
-
- /* TC 2 */
- port->qsize_add[8] = port->qsize_add[7] + port->qsize[1];
- port->qsize_add[9] = port->qsize_add[8] + port->qsize[2];
- port->qsize_add[10] = port->qsize_add[9] + port->qsize[2];
- port->qsize_add[11] = port->qsize_add[10] + port->qsize[2];
-
- /* TC 3 */
- port->qsize_add[12] = port->qsize_add[11] + port->qsize[2];
- port->qsize_add[13] = port->qsize_add[12] + port->qsize[3];
- port->qsize_add[14] = port->qsize_add[13] + port->qsize[3];
- port->qsize_add[15] = port->qsize_add[14] + port->qsize[3];
+ uint32_t i;
- port->qsize_sum = port->qsize_add[15] + port->qsize[3];
+ subport->qsize_add[0] = 0;
+
+ /* Strict prority traffic class */
+ for (i = 1; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ subport->qsize_add[i] = subport->qsize_add[i-1] + subport->qsize[i-1];
+
+ /* Best-effort traffic class */
+ subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE + 1] =
+ subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE] +
+ subport->qsize[RTE_SCHED_TRAFFIC_CLASS_BE];
+ subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE + 2] =
+ subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE + 1] +
+ subport->qsize[RTE_SCHED_TRAFFIC_CLASS_BE];
+ subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE + 3] =
+ subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE + 2] +
+ subport->qsize[RTE_SCHED_TRAFFIC_CLASS_BE];
+
+ subport->qsize_sum = subport->qsize_add[RTE_SCHED_TRAFFIC_CLASS_BE + 3] +
+ subport->qsize[RTE_SCHED_TRAFFIC_CLASS_BE];
}
static void
-rte_sched_port_log_pipe_profile(struct rte_sched_port *port, uint32_t i)
+rte_sched_port_log_pipe_profile(struct rte_sched_subport *subport, uint32_t i)
{
- struct rte_sched_pipe_profile *p = port->pipe_profiles + i;
+ struct rte_sched_pipe_profile *p = subport->pipe_profiles + i;
RTE_LOG(DEBUG, SCHED, "Low level config for pipe profile %u:\n"
- " Token bucket: period = %u, credits per period = %u, size = %u\n"
- " Traffic classes: period = %u, credits per period = [%u, %u, %u, %u]\n"
- " Traffic class 3 oversubscription: weight = %hhu\n"
- " WRR cost: [%hhu, %hhu, %hhu, %hhu], [%hhu, %hhu, %hhu, %hhu], [%hhu, %hhu, %hhu, %hhu], [%hhu, %hhu, %hhu, %hhu]\n",
+ " Token bucket: period = %"PRIu64", credits per period = %"PRIu64", size = %"PRIu64"\n"
+ " Traffic classes: period = %"PRIu64",\n"
+ " credits per period = [%"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64
+ ", %"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64
+ ", %"PRIu64", %"PRIu64", %"PRIu64"]\n"
+ " Best-effort traffic class oversubscription: weight = %hhu\n"
+ " WRR cost: [%hhu, %hhu, %hhu, %hhu]\n",
i,
/* Token bucket */
p->tc_credits_per_period[1],
p->tc_credits_per_period[2],
p->tc_credits_per_period[3],
-
- /* Traffic class 3 oversubscription */
+ p->tc_credits_per_period[4],
+ p->tc_credits_per_period[5],
+ p->tc_credits_per_period[6],
+ p->tc_credits_per_period[7],
+ p->tc_credits_per_period[8],
+ p->tc_credits_per_period[9],
+ p->tc_credits_per_period[10],
+ p->tc_credits_per_period[11],
+ p->tc_credits_per_period[12],
+
+ /* Best-effort traffic class oversubscription */
p->tc_ov_weight,
/* WRR */
- p->wrr_cost[ 0], p->wrr_cost[ 1], p->wrr_cost[ 2], p->wrr_cost[ 3],
- p->wrr_cost[ 4], p->wrr_cost[ 5], p->wrr_cost[ 6], p->wrr_cost[ 7],
- p->wrr_cost[ 8], p->wrr_cost[ 9], p->wrr_cost[10], p->wrr_cost[11],
- p->wrr_cost[12], p->wrr_cost[13], p->wrr_cost[14], p->wrr_cost[15]);
+ p->wrr_cost[0], p->wrr_cost[1], p->wrr_cost[2], p->wrr_cost[3]);
}
static inline uint64_t
-rte_sched_time_ms_to_bytes(uint32_t time_ms, uint32_t rate)
+rte_sched_time_ms_to_bytes(uint64_t time_ms, uint64_t rate)
{
uint64_t time = time_ms;
}
static void
-rte_sched_port_config_pipe_profile_table(struct rte_sched_port *port, struct rte_sched_port_params *params)
+rte_sched_pipe_profile_convert(struct rte_sched_subport *subport,
+ struct rte_sched_pipe_params *src,
+ struct rte_sched_pipe_profile *dst,
+ uint64_t rate)
{
- uint32_t i, j;
+ uint32_t wrr_cost[RTE_SCHED_BE_QUEUES_PER_PIPE];
+ uint32_t lcd1, lcd2, lcd;
+ uint32_t i;
+
+ /* Token Bucket */
+ if (src->tb_rate == rate) {
+ dst->tb_credits_per_period = 1;
+ dst->tb_period = 1;
+ } else {
+ double tb_rate = (double) src->tb_rate
+ / (double) rate;
+ double d = RTE_SCHED_TB_RATE_CONFIG_ERR;
+
+ rte_approx_64(tb_rate, d, &dst->tb_credits_per_period,
+ &dst->tb_period);
+ }
+
+ dst->tb_size = src->tb_size;
+
+ /* Traffic Classes */
+ dst->tc_period = rte_sched_time_ms_to_bytes(src->tc_period,
+ rate);
+
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ if (subport->qsize[i])
+ dst->tc_credits_per_period[i]
+ = rte_sched_time_ms_to_bytes(src->tc_period,
+ src->tc_rate[i]);
+
+ dst->tc_ov_weight = src->tc_ov_weight;
+
+ /* WRR queues */
+ wrr_cost[0] = src->wrr_weights[0];
+ wrr_cost[1] = src->wrr_weights[1];
+ wrr_cost[2] = src->wrr_weights[2];
+ wrr_cost[3] = src->wrr_weights[3];
+
+ lcd1 = rte_get_lcd(wrr_cost[0], wrr_cost[1]);
+ lcd2 = rte_get_lcd(wrr_cost[2], wrr_cost[3]);
+ lcd = rte_get_lcd(lcd1, lcd2);
+
+ wrr_cost[0] = lcd / wrr_cost[0];
+ wrr_cost[1] = lcd / wrr_cost[1];
+ wrr_cost[2] = lcd / wrr_cost[2];
+ wrr_cost[3] = lcd / wrr_cost[3];
- for (i = 0; i < port->n_pipe_profiles; i++) {
+ dst->wrr_cost[0] = (uint8_t) wrr_cost[0];
+ dst->wrr_cost[1] = (uint8_t) wrr_cost[1];
+ dst->wrr_cost[2] = (uint8_t) wrr_cost[2];
+ dst->wrr_cost[3] = (uint8_t) wrr_cost[3];
+}
+
+static void
+rte_sched_subport_config_pipe_profile_table(struct rte_sched_subport *subport,
+ struct rte_sched_subport_params *params, uint32_t rate)
+{
+ uint32_t i;
+
+ for (i = 0; i < subport->n_pipe_profiles; i++) {
struct rte_sched_pipe_params *src = params->pipe_profiles + i;
- struct rte_sched_pipe_profile *dst = port->pipe_profiles + i;
-
- /* Token Bucket */
- if (src->tb_rate == params->rate) {
- dst->tb_credits_per_period = 1;
- dst->tb_period = 1;
- } else {
- double tb_rate = (double) src->tb_rate
- / (double) params->rate;
- double d = RTE_SCHED_TB_RATE_CONFIG_ERR;
-
- rte_approx(tb_rate, d,
- &dst->tb_credits_per_period, &dst->tb_period);
+ struct rte_sched_pipe_profile *dst = subport->pipe_profiles + i;
+
+ rte_sched_pipe_profile_convert(subport, src, dst, rate);
+ rte_sched_port_log_pipe_profile(subport, i);
+ }
+
+ subport->pipe_tc_be_rate_max = 0;
+ for (i = 0; i < subport->n_pipe_profiles; i++) {
+ struct rte_sched_pipe_params *src = params->pipe_profiles + i;
+ uint64_t pipe_tc_be_rate = src->tc_rate[RTE_SCHED_TRAFFIC_CLASS_BE];
+
+ if (subport->pipe_tc_be_rate_max < pipe_tc_be_rate)
+ subport->pipe_tc_be_rate_max = pipe_tc_be_rate;
+ }
+}
+
+static int
+rte_sched_subport_check_params(struct rte_sched_subport_params *params,
+ uint32_t n_max_pipes_per_subport,
+ uint64_t rate)
+{
+ uint32_t i;
+
+ /* Check user parameters */
+ if (params == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter params\n", __func__);
+ return -EINVAL;
+ }
+
+ if (params->tb_rate == 0 || params->tb_rate > rate) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tb rate\n", __func__);
+ return -EINVAL;
+ }
+
+ if (params->tb_size == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tb size\n", __func__);
+ return -EINVAL;
+ }
+
+ /* qsize: if non-zero, power of 2,
+ * no bigger than 32K (due to 16-bit read/write pointers)
+ */
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
+ uint16_t qsize = params->qsize[i];
+
+ if (qsize != 0 && !rte_is_power_of_2(qsize)) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for qsize\n", __func__);
+ return -EINVAL;
}
- dst->tb_size = src->tb_size;
+ }
- /* Traffic Classes */
- dst->tc_period = rte_sched_time_ms_to_bytes(src->tc_period,
- params->rate);
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
+ uint64_t tc_rate = params->tc_rate[i];
+ uint16_t qsize = params->qsize[i];
- for (j = 0; j < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; j++)
- dst->tc_credits_per_period[j]
- = rte_sched_time_ms_to_bytes(src->tc_period,
- src->tc_rate[j]);
+ if ((qsize == 0 && tc_rate != 0) ||
+ (qsize != 0 && tc_rate == 0) ||
+ (tc_rate > params->tb_rate)) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tc rate\n", __func__);
+ return -EINVAL;
+ }
+ }
-#ifdef RTE_SCHED_SUBPORT_TC_OV
- dst->tc_ov_weight = src->tc_ov_weight;
-#endif
+ if (params->qsize[RTE_SCHED_TRAFFIC_CLASS_BE] == 0 ||
+ params->tc_rate[RTE_SCHED_TRAFFIC_CLASS_BE] == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect qsize or tc rate(best effort)\n", __func__);
+ return -EINVAL;
+ }
- /* WRR */
- for (j = 0; j < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; j++) {
- uint32_t wrr_cost[RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS];
- uint32_t lcd, lcd1, lcd2;
- uint32_t qindex;
-
- qindex = j * RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS;
-
- wrr_cost[0] = src->wrr_weights[qindex];
- wrr_cost[1] = src->wrr_weights[qindex + 1];
- wrr_cost[2] = src->wrr_weights[qindex + 2];
- wrr_cost[3] = src->wrr_weights[qindex + 3];
-
- lcd1 = rte_get_lcd(wrr_cost[0], wrr_cost[1]);
- lcd2 = rte_get_lcd(wrr_cost[2], wrr_cost[3]);
- lcd = rte_get_lcd(lcd1, lcd2);
-
- wrr_cost[0] = lcd / wrr_cost[0];
- wrr_cost[1] = lcd / wrr_cost[1];
- wrr_cost[2] = lcd / wrr_cost[2];
- wrr_cost[3] = lcd / wrr_cost[3];
-
- dst->wrr_cost[qindex] = (uint8_t) wrr_cost[0];
- dst->wrr_cost[qindex + 1] = (uint8_t) wrr_cost[1];
- dst->wrr_cost[qindex + 2] = (uint8_t) wrr_cost[2];
- dst->wrr_cost[qindex + 3] = (uint8_t) wrr_cost[3];
+ if (params->tc_period == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tc period\n", __func__);
+ return -EINVAL;
+ }
+
+ /* n_pipes_per_subport: non-zero, power of 2 */
+ if (params->n_pipes_per_subport_enabled == 0 ||
+ params->n_pipes_per_subport_enabled > n_max_pipes_per_subport ||
+ !rte_is_power_of_2(params->n_pipes_per_subport_enabled)) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for pipes number\n", __func__);
+ return -EINVAL;
+ }
+
+ /* pipe_profiles and n_pipe_profiles */
+ if (params->pipe_profiles == NULL ||
+ params->n_pipe_profiles == 0 ||
+ params->n_max_pipe_profiles == 0 ||
+ params->n_pipe_profiles > params->n_max_pipe_profiles) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for pipe profiles\n", __func__);
+ return -EINVAL;
+ }
+
+ for (i = 0; i < params->n_pipe_profiles; i++) {
+ struct rte_sched_pipe_params *p = params->pipe_profiles + i;
+ int status;
+
+ status = pipe_profile_check(p, rate, ¶ms->qsize[0]);
+ if (status != 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Pipe profile check failed(%d)\n", __func__, status);
+ return -EINVAL;
}
+ }
+
+ return 0;
+}
+
+uint32_t
+rte_sched_port_get_memory_footprint(struct rte_sched_port_params *port_params,
+ struct rte_sched_subport_params **subport_params)
+{
+ uint32_t size0 = 0, size1 = 0, i;
+ int status;
+
+ status = rte_sched_port_check_params(port_params);
+ if (status != 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Port scheduler port params check failed (%d)\n",
+ __func__, status);
- rte_sched_port_log_pipe_profile(port, i);
+ return 0;
}
- port->pipe_tc3_rate_max = 0;
- for (i = 0; i < port->n_pipe_profiles; i++) {
- struct rte_sched_pipe_params *src = params->pipe_profiles + i;
- uint32_t pipe_tc3_rate = src->tc_rate[3];
+ for (i = 0; i < port_params->n_subports_per_port; i++) {
+ struct rte_sched_subport_params *sp = subport_params[i];
- if (port->pipe_tc3_rate_max < pipe_tc3_rate)
- port->pipe_tc3_rate_max = pipe_tc3_rate;
+ status = rte_sched_subport_check_params(sp,
+ port_params->n_pipes_per_subport,
+ port_params->rate);
+ if (status != 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Port scheduler subport params check failed (%d)\n",
+ __func__, status);
+
+ return 0;
+ }
}
+
+ size0 = sizeof(struct rte_sched_port);
+
+ for (i = 0; i < port_params->n_subports_per_port; i++) {
+ struct rte_sched_subport_params *sp = subport_params[i];
+
+ size1 += rte_sched_subport_get_array_base(sp,
+ e_RTE_SCHED_SUBPORT_ARRAY_TOTAL);
+ }
+
+ return size0 + size1;
}
struct rte_sched_port *
rte_sched_port_config(struct rte_sched_port_params *params)
{
struct rte_sched_port *port = NULL;
- uint32_t mem_size, bmp_mem_size, n_queues_per_port, i, cycles_per_byte;
+ uint32_t size0, size1;
+ uint32_t cycles_per_byte;
+ uint32_t i, j;
+ int status;
- /* Check user parameters. Determine the amount of memory to allocate */
- mem_size = rte_sched_port_get_memory_footprint(params);
- if (mem_size == 0)
+ status = rte_sched_port_check_params(params);
+ if (status != 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Port scheduler params check failed (%d)\n",
+ __func__, status);
return NULL;
+ }
- /* Allocate memory to store the data structures */
- port = rte_zmalloc("qos_params", mem_size, RTE_CACHE_LINE_SIZE);
- if (port == NULL)
- return NULL;
+ size0 = sizeof(struct rte_sched_port);
+ size1 = params->n_subports_per_port * sizeof(struct rte_sched_subport *);
- /* compile time checks */
- RTE_BUILD_BUG_ON(RTE_SCHED_PORT_N_GRINDERS == 0);
- RTE_BUILD_BUG_ON(RTE_SCHED_PORT_N_GRINDERS & (RTE_SCHED_PORT_N_GRINDERS - 1));
+ /* Allocate memory to store the data structures */
+ port = rte_zmalloc_socket("qos_params", size0 + size1, RTE_CACHE_LINE_SIZE,
+ params->socket);
+ if (port == NULL) {
+ RTE_LOG(ERR, SCHED, "%s: Memory allocation fails\n", __func__);
+
+ return NULL;
+ }
/* User parameters */
port->n_subports_per_port = params->n_subports_per_port;
port->n_pipes_per_subport = params->n_pipes_per_subport;
- port->rate = params->rate;
- port->mtu = params->mtu + params->frame_overhead;
- port->frame_overhead = params->frame_overhead;
- memcpy(port->qsize, params->qsize, sizeof(params->qsize));
- port->n_pipe_profiles = params->n_pipe_profiles;
+ port->n_pipes_per_subport_log2 =
+ __builtin_ctz(params->n_pipes_per_subport);
+ port->socket = params->socket;
-#ifdef RTE_SCHED_RED
- for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
- uint32_t j;
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ port->pipe_queue[i] = i;
- for (j = 0; j < e_RTE_METER_COLORS; j++) {
- /* if min/max are both zero, then RED is disabled */
- if ((params->red_params[i][j].min_th |
- params->red_params[i][j].max_th) == 0) {
- continue;
- }
+ for (i = 0, j = 0; i < RTE_SCHED_QUEUES_PER_PIPE; i++) {
+ port->pipe_tc[i] = j;
- if (rte_red_config_init(&port->red_config[i][j],
- params->red_params[i][j].wq_log2,
- params->red_params[i][j].min_th,
- params->red_params[i][j].max_th,
- params->red_params[i][j].maxp_inv) != 0) {
- return NULL;
- }
- }
+ if (j < RTE_SCHED_TRAFFIC_CLASS_BE)
+ j++;
}
-#endif
+
+ for (i = 0, j = 0; i < RTE_SCHED_QUEUES_PER_PIPE; i++) {
+ port->tc_queue[i] = j;
+
+ if (i >= RTE_SCHED_TRAFFIC_CLASS_BE)
+ j++;
+ }
+ port->rate = params->rate;
+ port->mtu = params->mtu + params->frame_overhead;
+ port->frame_overhead = params->frame_overhead;
/* Timing */
port->time_cpu_cycles = rte_get_tsc_cycles();
/ params->rate;
port->inv_cycles_per_byte = rte_reciprocal_value(cycles_per_byte);
- /* Scheduling loop detection */
- port->pipe_loop = RTE_SCHED_PIPE_INVALID;
- port->pipe_exhaustion = 0;
-
/* Grinders */
- port->busy_grinders = 0;
port->pkts_out = NULL;
port->n_pkts_out = 0;
+ port->subport_id = 0;
- /* Queue base calculation */
- rte_sched_port_config_qsize(port);
-
- /* Large data structures */
- port->subport = (struct rte_sched_subport *)
- (port->memory + rte_sched_port_get_array_base(params,
- e_RTE_SCHED_PORT_ARRAY_SUBPORT));
- port->pipe = (struct rte_sched_pipe *)
- (port->memory + rte_sched_port_get_array_base(params,
- e_RTE_SCHED_PORT_ARRAY_PIPE));
- port->queue = (struct rte_sched_queue *)
- (port->memory + rte_sched_port_get_array_base(params,
- e_RTE_SCHED_PORT_ARRAY_QUEUE));
- port->queue_extra = (struct rte_sched_queue_extra *)
- (port->memory + rte_sched_port_get_array_base(params,
- e_RTE_SCHED_PORT_ARRAY_QUEUE_EXTRA));
- port->pipe_profiles = (struct rte_sched_pipe_profile *)
- (port->memory + rte_sched_port_get_array_base(params,
- e_RTE_SCHED_PORT_ARRAY_PIPE_PROFILES));
- port->bmp_array = port->memory
- + rte_sched_port_get_array_base(params, e_RTE_SCHED_PORT_ARRAY_BMP_ARRAY);
- port->queue_array = (struct rte_mbuf **)
- (port->memory + rte_sched_port_get_array_base(params,
- e_RTE_SCHED_PORT_ARRAY_QUEUE_ARRAY));
+ return port;
+}
- /* Pipe profile table */
- rte_sched_port_config_pipe_profile_table(port, params);
+static inline void
+rte_sched_subport_free(struct rte_sched_port *port,
+ struct rte_sched_subport *subport)
+{
+ uint32_t n_subport_pipe_queues;
+ uint32_t qindex;
- /* Bitmap */
- n_queues_per_port = rte_sched_port_queues_per_port(port);
- bmp_mem_size = rte_bitmap_get_memory_footprint(n_queues_per_port);
- port->bmp = rte_bitmap_init(n_queues_per_port, port->bmp_array,
- bmp_mem_size);
- if (port->bmp == NULL) {
- RTE_LOG(ERR, SCHED, "Bitmap init error\n");
- return NULL;
- }
+ if (subport == NULL)
+ return;
- for (i = 0; i < RTE_SCHED_PORT_N_GRINDERS; i++)
- port->grinder_base_bmp_pos[i] = RTE_SCHED_PIPE_INVALID;
+ n_subport_pipe_queues = rte_sched_subport_pipe_queues(subport);
+ /* Free enqueued mbufs */
+ for (qindex = 0; qindex < n_subport_pipe_queues; qindex++) {
+ struct rte_mbuf **mbufs =
+ rte_sched_subport_pipe_qbase(subport, qindex);
+ uint16_t qsize = rte_sched_subport_pipe_qsize(port, subport, qindex);
+ if (qsize != 0) {
+ struct rte_sched_queue *queue = subport->queue + qindex;
+ uint16_t qr = queue->qr & (qsize - 1);
+ uint16_t qw = queue->qw & (qsize - 1);
+
+ for (; qr != qw; qr = (qr + 1) & (qsize - 1))
+ rte_pktmbuf_free(mbufs[qr]);
+ }
+ }
- return port;
+ rte_bitmap_free(subport->bmp);
}
void
rte_sched_port_free(struct rte_sched_port *port)
{
- uint32_t qindex;
- uint32_t n_queues_per_port;
+ uint32_t i;
/* Check user parameters */
if (port == NULL)
return;
- n_queues_per_port = rte_sched_port_queues_per_port(port);
-
- /* Free enqueued mbufs */
- for (qindex = 0; qindex < n_queues_per_port; qindex++) {
- struct rte_mbuf **mbufs = rte_sched_port_qbase(port, qindex);
- uint16_t qsize = rte_sched_port_qsize(port, qindex);
- struct rte_sched_queue *queue = port->queue + qindex;
- uint16_t qr = queue->qr & (qsize - 1);
- uint16_t qw = queue->qw & (qsize - 1);
-
- for (; qr != qw; qr = (qr + 1) & (qsize - 1))
- rte_pktmbuf_free(mbufs[qr]);
- }
+ for (i = 0; i < port->n_subports_per_port; i++)
+ rte_sched_subport_free(port, port->subports[i]);
- rte_bitmap_free(port->bmp);
rte_free(port);
}
static void
rte_sched_port_log_subport_config(struct rte_sched_port *port, uint32_t i)
{
- struct rte_sched_subport *s = port->subport + i;
+ struct rte_sched_subport *s = port->subports[i];
RTE_LOG(DEBUG, SCHED, "Low level config for subport %u:\n"
- " Token bucket: period = %u, credits per period = %u, size = %u\n"
- " Traffic classes: period = %u, credits per period = [%u, %u, %u, %u]\n"
- " Traffic class 3 oversubscription: wm min = %u, wm max = %u\n",
+ " Token bucket: period = %"PRIu64", credits per period = %"PRIu64
+ ", size = %"PRIu64"\n"
+ " Traffic classes: period = %"PRIu64"\n"
+ " credits per period = [%"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64
+ ", %"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64", %"PRIu64
+ ", %"PRIu64", %"PRIu64", %"PRIu64"]\n"
+ " Best effort traffic class oversubscription: wm min = %"PRIu64
+ ", wm max = %"PRIu64"\n",
i,
/* Token bucket */
s->tc_credits_per_period[1],
s->tc_credits_per_period[2],
s->tc_credits_per_period[3],
-
- /* Traffic class 3 oversubscription */
+ s->tc_credits_per_period[4],
+ s->tc_credits_per_period[5],
+ s->tc_credits_per_period[6],
+ s->tc_credits_per_period[7],
+ s->tc_credits_per_period[8],
+ s->tc_credits_per_period[9],
+ s->tc_credits_per_period[10],
+ s->tc_credits_per_period[11],
+ s->tc_credits_per_period[12],
+
+ /* Best effort traffic class oversubscription */
s->tc_ov_wm_min,
s->tc_ov_wm_max);
}
+static void
+rte_sched_free_memory(struct rte_sched_port *port, uint32_t n_subports)
+{
+ uint32_t i;
+
+ for (i = 0; i < n_subports; i++) {
+ struct rte_sched_subport *subport = port->subports[i];
+
+ rte_sched_subport_free(port, subport);
+ }
+
+ rte_free(port);
+}
+
int
rte_sched_subport_config(struct rte_sched_port *port,
uint32_t subport_id,
struct rte_sched_subport_params *params)
{
- struct rte_sched_subport *s;
- uint32_t i;
+ struct rte_sched_subport *s = NULL;
+ uint32_t n_subports = subport_id;
+ uint32_t n_subport_pipe_queues, i;
+ uint32_t size0, size1, bmp_mem_size;
+ int status;
/* Check user parameters */
- if (port == NULL ||
- subport_id >= port->n_subports_per_port ||
- params == NULL)
- return -1;
+ if (port == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter port\n", __func__);
+ return 0;
+ }
- if (params->tb_rate == 0 || params->tb_rate > port->rate)
- return -2;
+ if (subport_id >= port->n_subports_per_port) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for subport id\n", __func__);
- if (params->tb_size == 0)
- return -3;
+ rte_sched_free_memory(port, n_subports);
+ return -EINVAL;
+ }
- for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
- if (params->tc_rate[i] == 0 ||
- params->tc_rate[i] > params->tb_rate)
- return -4;
+ status = rte_sched_subport_check_params(params,
+ port->n_pipes_per_subport,
+ port->rate);
+ if (status != 0) {
+ RTE_LOG(NOTICE, SCHED,
+ "%s: Port scheduler params check failed (%d)\n",
+ __func__, status);
+
+ rte_sched_free_memory(port, n_subports);
+ return -EINVAL;
+ }
+
+ /* Determine the amount of memory to allocate */
+ size0 = sizeof(struct rte_sched_subport);
+ size1 = rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_TOTAL);
+
+ /* Allocate memory to store the data structures */
+ s = rte_zmalloc_socket("subport_params", size0 + size1,
+ RTE_CACHE_LINE_SIZE, port->socket);
+ if (s == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Memory allocation fails\n", __func__);
+
+ rte_sched_free_memory(port, n_subports);
+ return -ENOMEM;
}
- if (params->tc_period == 0)
- return -5;
+ n_subports++;
- s = port->subport + subport_id;
+ /* Port */
+ port->subports[subport_id] = s;
/* Token Bucket (TB) */
if (params->tb_rate == port->rate) {
double tb_rate = ((double) params->tb_rate) / ((double) port->rate);
double d = RTE_SCHED_TB_RATE_CONFIG_ERR;
- rte_approx(tb_rate, d, &s->tb_credits_per_period, &s->tb_period);
+ rte_approx_64(tb_rate, d, &s->tb_credits_per_period, &s->tb_period);
}
s->tb_size = params->tb_size;
/* Traffic Classes (TCs) */
s->tc_period = rte_sched_time_ms_to_bytes(params->tc_period, port->rate);
for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
- s->tc_credits_per_period[i]
- = rte_sched_time_ms_to_bytes(params->tc_period,
- params->tc_rate[i]);
+ if (params->qsize[i])
+ s->tc_credits_per_period[i]
+ = rte_sched_time_ms_to_bytes(params->tc_period,
+ params->tc_rate[i]);
}
s->tc_time = port->time + s->tc_period;
for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
- s->tc_credits[i] = s->tc_credits_per_period[i];
+ if (params->qsize[i])
+ s->tc_credits[i] = s->tc_credits_per_period[i];
+
+ /* compile time checks */
+ RTE_BUILD_BUG_ON(RTE_SCHED_PORT_N_GRINDERS == 0);
+ RTE_BUILD_BUG_ON(RTE_SCHED_PORT_N_GRINDERS &
+ (RTE_SCHED_PORT_N_GRINDERS - 1));
+
+ /* User parameters */
+ s->n_pipes_per_subport_enabled = params->n_pipes_per_subport_enabled;
+ memcpy(s->qsize, params->qsize, sizeof(params->qsize));
+ s->n_pipe_profiles = params->n_pipe_profiles;
+ s->n_max_pipe_profiles = params->n_max_pipe_profiles;
+
+#ifdef RTE_SCHED_RED
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
+ uint32_t j;
+
+ for (j = 0; j < RTE_COLORS; j++) {
+ /* if min/max are both zero, then RED is disabled */
+ if ((params->red_params[i][j].min_th |
+ params->red_params[i][j].max_th) == 0) {
+ continue;
+ }
+
+ if (rte_red_config_init(&s->red_config[i][j],
+ params->red_params[i][j].wq_log2,
+ params->red_params[i][j].min_th,
+ params->red_params[i][j].max_th,
+ params->red_params[i][j].maxp_inv) != 0) {
+ rte_sched_free_memory(port, n_subports);
+
+ RTE_LOG(NOTICE, SCHED,
+ "%s: RED configuration init fails\n", __func__);
+ return -EINVAL;
+ }
+ }
+ }
+#endif
+
+ /* Scheduling loop detection */
+ s->pipe_loop = RTE_SCHED_PIPE_INVALID;
+ s->pipe_exhaustion = 0;
+
+ /* Grinders */
+ s->busy_grinders = 0;
+
+ /* Queue base calculation */
+ rte_sched_subport_config_qsize(s);
+
+ /* Large data structures */
+ s->pipe = (struct rte_sched_pipe *)
+ (s->memory + rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_PIPE));
+ s->queue = (struct rte_sched_queue *)
+ (s->memory + rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_QUEUE));
+ s->queue_extra = (struct rte_sched_queue_extra *)
+ (s->memory + rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_QUEUE_EXTRA));
+ s->pipe_profiles = (struct rte_sched_pipe_profile *)
+ (s->memory + rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_PIPE_PROFILES));
+ s->bmp_array = s->memory + rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_BMP_ARRAY);
+ s->queue_array = (struct rte_mbuf **)
+ (s->memory + rte_sched_subport_get_array_base(params,
+ e_RTE_SCHED_SUBPORT_ARRAY_QUEUE_ARRAY));
+
+ /* Pipe profile table */
+ rte_sched_subport_config_pipe_profile_table(s, params, port->rate);
+
+ /* Bitmap */
+ n_subport_pipe_queues = rte_sched_subport_pipe_queues(s);
+ bmp_mem_size = rte_bitmap_get_memory_footprint(n_subport_pipe_queues);
+ s->bmp = rte_bitmap_init(n_subport_pipe_queues, s->bmp_array,
+ bmp_mem_size);
+ if (s->bmp == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Subport bitmap init error\n", __func__);
+
+ rte_sched_free_memory(port, n_subports);
+ return -EINVAL;
+ }
+
+ for (i = 0; i < RTE_SCHED_PORT_N_GRINDERS; i++)
+ s->grinder_base_bmp_pos[i] = RTE_SCHED_PIPE_INVALID;
#ifdef RTE_SCHED_SUBPORT_TC_OV
/* TC oversubscription */
s->tc_ov_wm_min = port->mtu;
s->tc_ov_wm_max = rte_sched_time_ms_to_bytes(params->tc_period,
- port->pipe_tc3_rate_max);
+ s->pipe_tc_be_rate_max);
s->tc_ov_wm = s->tc_ov_wm_max;
s->tc_ov_period_id = 0;
s->tc_ov = 0;
struct rte_sched_subport *s;
struct rte_sched_pipe *p;
struct rte_sched_pipe_profile *params;
+ uint32_t n_subports = subport_id + 1;
uint32_t deactivate, profile, i;
/* Check user parameters */
profile = (uint32_t) pipe_profile;
deactivate = (pipe_profile < 0);
- if (port == NULL ||
- subport_id >= port->n_subports_per_port ||
- pipe_id >= port->n_pipes_per_subport ||
- (!deactivate && profile >= port->n_pipe_profiles))
- return -1;
+ if (port == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter port\n", __func__);
+ return -EINVAL;
+ }
+
+ if (subport_id >= port->n_subports_per_port) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter subport id\n", __func__);
+
+ rte_sched_free_memory(port, n_subports);
+ return -EINVAL;
+ }
+
+ s = port->subports[subport_id];
+ if (pipe_id >= s->n_pipes_per_subport_enabled) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter pipe id\n", __func__);
+ rte_sched_free_memory(port, n_subports);
+ return -EINVAL;
+ }
- /* Check that subport configuration is valid */
- s = port->subport + subport_id;
- if (s->tb_period == 0)
- return -2;
+ if (!deactivate && profile >= s->n_pipe_profiles) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter pipe profile\n", __func__);
- p = port->pipe + (subport_id * port->n_pipes_per_subport + pipe_id);
+ rte_sched_free_memory(port, n_subports);
+ return -EINVAL;
+ }
/* Handle the case when pipe already has a valid configuration */
+ p = s->pipe + pipe_id;
if (p->tb_time) {
- params = port->pipe_profiles + p->profile;
+ params = s->pipe_profiles + p->profile;
-#ifdef RTE_SCHED_SUBPORT_TC_OV
- double subport_tc3_rate = (double) s->tc_credits_per_period[3]
+ double subport_tc_be_rate =
+ (double) s->tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASS_BE]
/ (double) s->tc_period;
- double pipe_tc3_rate = (double) params->tc_credits_per_period[3]
+ double pipe_tc_be_rate =
+ (double) params->tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASS_BE]
/ (double) params->tc_period;
- uint32_t tc3_ov = s->tc_ov;
+ uint32_t tc_be_ov = s->tc_ov;
/* Unplug pipe from its subport */
s->tc_ov_n -= params->tc_ov_weight;
- s->tc_ov_rate -= pipe_tc3_rate;
- s->tc_ov = s->tc_ov_rate > subport_tc3_rate;
+ s->tc_ov_rate -= pipe_tc_be_rate;
+ s->tc_ov = s->tc_ov_rate > subport_tc_be_rate;
- if (s->tc_ov != tc3_ov) {
+ if (s->tc_ov != tc_be_ov) {
RTE_LOG(DEBUG, SCHED,
- "Subport %u TC3 oversubscription is OFF (%.4lf >= %.4lf)\n",
- subport_id, subport_tc3_rate, s->tc_ov_rate);
+ "Subport %u Best-effort TC oversubscription is OFF (%.4lf >= %.4lf)\n",
+ subport_id, subport_tc_be_rate, s->tc_ov_rate);
}
-#endif
/* Reset the pipe */
memset(p, 0, sizeof(struct rte_sched_pipe));
/* Apply the new pipe configuration */
p->profile = profile;
- params = port->pipe_profiles + p->profile;
+ params = s->pipe_profiles + p->profile;
/* Token Bucket (TB) */
p->tb_time = port->time;
/* Traffic Classes (TCs) */
p->tc_time = port->time + params->tc_period;
+
for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
- p->tc_credits[i] = params->tc_credits_per_period[i];
+ if (s->qsize[i])
+ p->tc_credits[i] = params->tc_credits_per_period[i];
-#ifdef RTE_SCHED_SUBPORT_TC_OV
{
- /* Subport TC3 oversubscription */
- double subport_tc3_rate = (double) s->tc_credits_per_period[3]
+ /* Subport best effort tc oversubscription */
+ double subport_tc_be_rate =
+ (double) s->tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASS_BE]
/ (double) s->tc_period;
- double pipe_tc3_rate = (double) params->tc_credits_per_period[3]
+ double pipe_tc_be_rate =
+ (double) params->tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASS_BE]
/ (double) params->tc_period;
- uint32_t tc3_ov = s->tc_ov;
+ uint32_t tc_be_ov = s->tc_ov;
s->tc_ov_n += params->tc_ov_weight;
- s->tc_ov_rate += pipe_tc3_rate;
- s->tc_ov = s->tc_ov_rate > subport_tc3_rate;
+ s->tc_ov_rate += pipe_tc_be_rate;
+ s->tc_ov = s->tc_ov_rate > subport_tc_be_rate;
- if (s->tc_ov != tc3_ov) {
+ if (s->tc_ov != tc_be_ov) {
RTE_LOG(DEBUG, SCHED,
- "Subport %u TC3 oversubscription is ON (%.4lf < %.4lf)\n",
- subport_id, subport_tc3_rate, s->tc_ov_rate);
+ "Subport %u Best effort TC oversubscription is ON (%.4lf < %.4lf)\n",
+ subport_id, subport_tc_be_rate, s->tc_ov_rate);
}
p->tc_ov_period_id = s->tc_ov_period_id;
p->tc_ov_credits = s->tc_ov_wm;
}
-#endif
return 0;
}
-void
-rte_sched_port_pkt_write(struct rte_mbuf *pkt,
- uint32_t subport, uint32_t pipe, uint32_t traffic_class,
- uint32_t queue, enum rte_meter_color color)
+int
+rte_sched_subport_pipe_profile_add(struct rte_sched_port *port,
+ uint32_t subport_id,
+ struct rte_sched_pipe_params *params,
+ uint32_t *pipe_profile_id)
+{
+ struct rte_sched_subport *s;
+ struct rte_sched_pipe_profile *pp;
+ uint32_t i;
+ int status;
+
+ /* Port */
+ if (port == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter port\n", __func__);
+ return -EINVAL;
+ }
+
+ /* Subport id not exceeds the max limit */
+ if (subport_id > port->n_subports_per_port) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for subport id\n", __func__);
+ return -EINVAL;
+ }
+
+ s = port->subports[subport_id];
+
+ /* Pipe profiles exceeds the max limit */
+ if (s->n_pipe_profiles >= s->n_max_pipe_profiles) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Number of pipe profiles exceeds the max limit\n", __func__);
+ return -EINVAL;
+ }
+
+ /* Pipe params */
+ status = pipe_profile_check(params, port->rate, &s->qsize[0]);
+ if (status != 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Pipe profile check failed(%d)\n", __func__, status);
+ return -EINVAL;
+ }
+
+ pp = &s->pipe_profiles[s->n_pipe_profiles];
+ rte_sched_pipe_profile_convert(s, params, pp, port->rate);
+
+ /* Pipe profile should not exists */
+ for (i = 0; i < s->n_pipe_profiles; i++)
+ if (memcmp(s->pipe_profiles + i, pp, sizeof(*pp)) == 0) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Pipe profile exists\n", __func__);
+ return -EINVAL;
+ }
+
+ /* Pipe profile commit */
+ *pipe_profile_id = s->n_pipe_profiles;
+ s->n_pipe_profiles++;
+
+ if (s->pipe_tc_be_rate_max < params->tc_rate[RTE_SCHED_TRAFFIC_CLASS_BE])
+ s->pipe_tc_be_rate_max = params->tc_rate[RTE_SCHED_TRAFFIC_CLASS_BE];
+
+ rte_sched_port_log_pipe_profile(s, *pipe_profile_id);
+
+ return 0;
+}
+
+static inline uint32_t
+rte_sched_port_qindex(struct rte_sched_port *port,
+ uint32_t subport,
+ uint32_t pipe,
+ uint32_t traffic_class,
+ uint32_t queue)
{
- struct rte_sched_port_hierarchy *sched
- = (struct rte_sched_port_hierarchy *) &pkt->hash.sched;
+ return ((subport & (port->n_subports_per_port - 1)) <<
+ (port->n_pipes_per_subport_log2 + 4)) |
+ ((pipe &
+ (port->subports[subport]->n_pipes_per_subport_enabled - 1)) << 4) |
+ ((rte_sched_port_pipe_queue(port, traffic_class) + queue) &
+ (RTE_SCHED_QUEUES_PER_PIPE - 1));
+}
- RTE_BUILD_BUG_ON(sizeof(*sched) > sizeof(pkt->hash.sched));
+void
+rte_sched_port_pkt_write(struct rte_sched_port *port,
+ struct rte_mbuf *pkt,
+ uint32_t subport, uint32_t pipe,
+ uint32_t traffic_class,
+ uint32_t queue, enum rte_color color)
+{
+ uint32_t queue_id =
+ rte_sched_port_qindex(port, subport, pipe, traffic_class, queue);
- sched->color = (uint32_t) color;
- sched->subport = subport;
- sched->pipe = pipe;
- sched->traffic_class = traffic_class;
- sched->queue = queue;
+ rte_mbuf_sched_set(pkt, queue_id, traffic_class, (uint8_t)color);
}
void
-rte_sched_port_pkt_read_tree_path(const struct rte_mbuf *pkt,
+rte_sched_port_pkt_read_tree_path(struct rte_sched_port *port,
+ const struct rte_mbuf *pkt,
uint32_t *subport, uint32_t *pipe,
uint32_t *traffic_class, uint32_t *queue)
{
- const struct rte_sched_port_hierarchy *sched
- = (const struct rte_sched_port_hierarchy *) &pkt->hash.sched;
+ uint32_t queue_id = rte_mbuf_sched_queue_get(pkt);
- *subport = sched->subport;
- *pipe = sched->pipe;
- *traffic_class = sched->traffic_class;
- *queue = sched->queue;
+ *subport = queue_id >> (port->n_pipes_per_subport_log2 + 4);
+ *pipe = (queue_id >> 4) &
+ (port->subports[*subport]->n_pipes_per_subport_enabled - 1);
+ *traffic_class = rte_sched_port_pipe_tc(port, queue_id);
+ *queue = rte_sched_port_tc_queue(port, queue_id);
}
-enum rte_meter_color
+enum rte_color
rte_sched_port_pkt_read_color(const struct rte_mbuf *pkt)
{
- const struct rte_sched_port_hierarchy *sched
- = (const struct rte_sched_port_hierarchy *) &pkt->hash.sched;
-
- return (enum rte_meter_color) sched->color;
+ return (enum rte_color)rte_mbuf_sched_color_get(pkt);
}
int
struct rte_sched_subport *s;
/* Check user parameters */
- if (port == NULL || subport_id >= port->n_subports_per_port ||
- stats == NULL || tc_ov == NULL)
- return -1;
+ if (port == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter port\n", __func__);
+ return -EINVAL;
+ }
+
+ if (subport_id >= port->n_subports_per_port) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for subport id\n", __func__);
+ return -EINVAL;
+ }
+
+ if (stats == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter stats\n", __func__);
+ return -EINVAL;
+ }
- s = port->subport + subport_id;
+ if (tc_ov == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for tc_ov\n", __func__);
+ return -EINVAL;
+ }
+
+ s = port->subports[subport_id];
/* Copy subport stats and clear */
memcpy(stats, &s->stats, sizeof(struct rte_sched_subport_stats));
memset(&s->stats, 0, sizeof(struct rte_sched_subport_stats));
- /* Subport TC ovesubscription status */
+ /* Subport TC oversubscription status */
*tc_ov = s->tc_ov;
return 0;
struct rte_sched_queue_stats *stats,
uint16_t *qlen)
{
+ struct rte_sched_subport *s;
struct rte_sched_queue *q;
struct rte_sched_queue_extra *qe;
+ uint32_t subport_id, subport_qmask, subport_qindex;
/* Check user parameters */
- if ((port == NULL) ||
- (queue_id >= rte_sched_port_queues_per_port(port)) ||
- (stats == NULL) ||
- (qlen == NULL)) {
- return -1;
+ if (port == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter port\n", __func__);
+ return -EINVAL;
+ }
+
+ if (queue_id >= rte_sched_port_queues_per_port(port)) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for queue id\n", __func__);
+ return -EINVAL;
+ }
+
+ if (stats == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter stats\n", __func__);
+ return -EINVAL;
+ }
+
+ if (qlen == NULL) {
+ RTE_LOG(ERR, SCHED,
+ "%s: Incorrect value for parameter qlen\n", __func__);
+ return -EINVAL;
}
- q = port->queue + queue_id;
- qe = port->queue_extra + queue_id;
+ subport_qmask = port->n_pipes_per_subport_log2 + 4;
+ subport_id = (queue_id >> subport_qmask) & (port->n_subports_per_port - 1);
+
+ s = port->subports[subport_id];
+ subport_qindex = ((1 << subport_qmask) - 1) & queue_id;
+ q = s->queue + subport_qindex;
+ qe = s->queue_extra + subport_qindex;
/* Copy queue stats and clear */
memcpy(stats, &qe->stats, sizeof(struct rte_sched_queue_stats));
return 0;
}
-static inline uint32_t
-rte_sched_port_qindex(struct rte_sched_port *port, uint32_t subport, uint32_t pipe, uint32_t traffic_class, uint32_t queue)
-{
- uint32_t result;
-
- result = subport * port->n_pipes_per_subport + pipe;
- result = result * RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE + traffic_class;
- result = result * RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS + queue;
-
- return result;
-}
-
#ifdef RTE_SCHED_DEBUG
static inline int
-rte_sched_port_queue_is_empty(struct rte_sched_port *port, uint32_t qindex)
+rte_sched_port_queue_is_empty(struct rte_sched_subport *subport,
+ uint32_t qindex)
{
- struct rte_sched_queue *queue = port->queue + qindex;
+ struct rte_sched_queue *queue = subport->queue + qindex;
return queue->qr == queue->qw;
}
#ifdef RTE_SCHED_COLLECT_STATS
static inline void
-rte_sched_port_update_subport_stats(struct rte_sched_port *port, uint32_t qindex, struct rte_mbuf *pkt)
+rte_sched_port_update_subport_stats(struct rte_sched_port *port,
+ struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf *pkt)
{
- struct rte_sched_subport *s = port->subport + (qindex / rte_sched_port_queues_per_subport(port));
- uint32_t tc_index = (qindex >> 2) & 0x3;
+ uint32_t tc_index = rte_sched_port_pipe_tc(port, qindex);
uint32_t pkt_len = pkt->pkt_len;
- s->stats.n_pkts_tc[tc_index] += 1;
- s->stats.n_bytes_tc[tc_index] += pkt_len;
+ subport->stats.n_pkts_tc[tc_index] += 1;
+ subport->stats.n_bytes_tc[tc_index] += pkt_len;
}
#ifdef RTE_SCHED_RED
static inline void
rte_sched_port_update_subport_stats_on_drop(struct rte_sched_port *port,
- uint32_t qindex,
- struct rte_mbuf *pkt, uint32_t red)
+ struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf *pkt,
+ uint32_t red)
#else
static inline void
rte_sched_port_update_subport_stats_on_drop(struct rte_sched_port *port,
- uint32_t qindex,
- struct rte_mbuf *pkt, __rte_unused uint32_t red)
+ struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf *pkt,
+ __rte_unused uint32_t red)
#endif
{
- struct rte_sched_subport *s = port->subport + (qindex / rte_sched_port_queues_per_subport(port));
- uint32_t tc_index = (qindex >> 2) & 0x3;
+ uint32_t tc_index = rte_sched_port_pipe_tc(port, qindex);
uint32_t pkt_len = pkt->pkt_len;
- s->stats.n_pkts_tc_dropped[tc_index] += 1;
- s->stats.n_bytes_tc_dropped[tc_index] += pkt_len;
+ subport->stats.n_pkts_tc_dropped[tc_index] += 1;
+ subport->stats.n_bytes_tc_dropped[tc_index] += pkt_len;
#ifdef RTE_SCHED_RED
- s->stats.n_pkts_red_dropped[tc_index] += red;
+ subport->stats.n_pkts_red_dropped[tc_index] += red;
#endif
}
static inline void
-rte_sched_port_update_queue_stats(struct rte_sched_port *port, uint32_t qindex, struct rte_mbuf *pkt)
+rte_sched_port_update_queue_stats(struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf *pkt)
{
- struct rte_sched_queue_extra *qe = port->queue_extra + qindex;
+ struct rte_sched_queue_extra *qe = subport->queue_extra + qindex;
uint32_t pkt_len = pkt->pkt_len;
qe->stats.n_pkts += 1;
#ifdef RTE_SCHED_RED
static inline void
-rte_sched_port_update_queue_stats_on_drop(struct rte_sched_port *port,
- uint32_t qindex,
- struct rte_mbuf *pkt, uint32_t red)
+rte_sched_port_update_queue_stats_on_drop(struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf *pkt,
+ uint32_t red)
#else
static inline void
-rte_sched_port_update_queue_stats_on_drop(struct rte_sched_port *port,
- uint32_t qindex,
- struct rte_mbuf *pkt, __rte_unused uint32_t red)
+rte_sched_port_update_queue_stats_on_drop(struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf *pkt,
+ __rte_unused uint32_t red)
#endif
{
- struct rte_sched_queue_extra *qe = port->queue_extra + qindex;
+ struct rte_sched_queue_extra *qe = subport->queue_extra + qindex;
uint32_t pkt_len = pkt->pkt_len;
qe->stats.n_pkts_dropped += 1;
#ifdef RTE_SCHED_RED
static inline int
-rte_sched_port_red_drop(struct rte_sched_port *port, struct rte_mbuf *pkt, uint32_t qindex, uint16_t qlen)
+rte_sched_port_red_drop(struct rte_sched_port *port,
+ struct rte_sched_subport *subport,
+ struct rte_mbuf *pkt,
+ uint32_t qindex,
+ uint16_t qlen)
{
struct rte_sched_queue_extra *qe;
struct rte_red_config *red_cfg;
struct rte_red *red;
uint32_t tc_index;
- enum rte_meter_color color;
+ enum rte_color color;
- tc_index = (qindex >> 2) & 0x3;
+ tc_index = rte_sched_port_pipe_tc(port, qindex);
color = rte_sched_port_pkt_read_color(pkt);
- red_cfg = &port->red_config[tc_index][color];
+ red_cfg = &subport->red_config[tc_index][color];
if ((red_cfg->min_th | red_cfg->max_th) == 0)
return 0;
- qe = port->queue_extra + qindex;
+ qe = subport->queue_extra + qindex;
red = &qe->red;
return rte_red_enqueue(red_cfg, red, qlen, port->time);
}
static inline void
-rte_sched_port_set_queue_empty_timestamp(struct rte_sched_port *port, uint32_t qindex)
+rte_sched_port_set_queue_empty_timestamp(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t qindex)
{
- struct rte_sched_queue_extra *qe = port->queue_extra + qindex;
+ struct rte_sched_queue_extra *qe = subport->queue_extra + qindex;
struct rte_red *red = &qe->red;
rte_red_mark_queue_empty(red, port->time);
#else
-#define rte_sched_port_red_drop(port, pkt, qindex, qlen) 0
+static inline int rte_sched_port_red_drop(struct rte_sched_port *port __rte_unused,
+ struct rte_sched_subport *subport __rte_unused,
+ struct rte_mbuf *pkt __rte_unused,
+ uint32_t qindex __rte_unused,
+ uint16_t qlen __rte_unused)
+{
+ return 0;
+}
-#define rte_sched_port_set_queue_empty_timestamp(port, qindex)
+#define rte_sched_port_set_queue_empty_timestamp(port, subport, qindex)
#endif /* RTE_SCHED_RED */
#ifdef RTE_SCHED_DEBUG
static inline void
-debug_check_queue_slab(struct rte_sched_port *port, uint32_t bmp_pos,
+debug_check_queue_slab(struct rte_sched_subport *subport, uint32_t bmp_pos,
uint64_t bmp_slab)
{
uint64_t mask;
panic = 0;
for (i = 0, mask = 1; i < 64; i++, mask <<= 1) {
if (mask & bmp_slab) {
- if (rte_sched_port_queue_is_empty(port, bmp_pos + i)) {
+ if (rte_sched_port_queue_is_empty(subport, bmp_pos + i)) {
printf("Queue %u (slab offset %u) is empty\n", bmp_pos + i, i);
panic = 1;
}
#endif /* RTE_SCHED_DEBUG */
+static inline struct rte_sched_subport *
+rte_sched_port_subport(struct rte_sched_port *port,
+ struct rte_mbuf *pkt)
+{
+ uint32_t queue_id = rte_mbuf_sched_queue_get(pkt);
+ uint32_t subport_id = queue_id >> (port->n_pipes_per_subport_log2 + 4);
+
+ return port->subports[subport_id];
+}
+
static inline uint32_t
-rte_sched_port_enqueue_qptrs_prefetch0(struct rte_sched_port *port,
- struct rte_mbuf *pkt)
+rte_sched_port_enqueue_qptrs_prefetch0(struct rte_sched_subport *subport,
+ struct rte_mbuf *pkt, uint32_t subport_qmask)
{
struct rte_sched_queue *q;
#ifdef RTE_SCHED_COLLECT_STATS
struct rte_sched_queue_extra *qe;
#endif
- uint32_t subport, pipe, traffic_class, queue, qindex;
+ uint32_t qindex = rte_mbuf_sched_queue_get(pkt);
+ uint32_t subport_queue_id = subport_qmask & qindex;
- rte_sched_port_pkt_read_tree_path(pkt, &subport, &pipe, &traffic_class, &queue);
-
- qindex = rte_sched_port_qindex(port, subport, pipe, traffic_class, queue);
- q = port->queue + qindex;
+ q = subport->queue + subport_queue_id;
rte_prefetch0(q);
#ifdef RTE_SCHED_COLLECT_STATS
- qe = port->queue_extra + qindex;
+ qe = subport->queue_extra + subport_queue_id;
rte_prefetch0(qe);
#endif
- return qindex;
+ return subport_queue_id;
}
static inline void
rte_sched_port_enqueue_qwa_prefetch0(struct rte_sched_port *port,
- uint32_t qindex, struct rte_mbuf **qbase)
+ struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf **qbase)
{
struct rte_sched_queue *q;
struct rte_mbuf **q_qw;
uint16_t qsize;
- q = port->queue + qindex;
- qsize = rte_sched_port_qsize(port, qindex);
+ q = subport->queue + qindex;
+ qsize = rte_sched_subport_pipe_qsize(port, subport, qindex);
q_qw = qbase + (q->qw & (qsize - 1));
rte_prefetch0(q_qw);
- rte_bitmap_prefetch0(port->bmp, qindex);
+ rte_bitmap_prefetch0(subport->bmp, qindex);
}
static inline int
-rte_sched_port_enqueue_qwa(struct rte_sched_port *port, uint32_t qindex,
- struct rte_mbuf **qbase, struct rte_mbuf *pkt)
+rte_sched_port_enqueue_qwa(struct rte_sched_port *port,
+ struct rte_sched_subport *subport,
+ uint32_t qindex,
+ struct rte_mbuf **qbase,
+ struct rte_mbuf *pkt)
{
struct rte_sched_queue *q;
uint16_t qsize;
uint16_t qlen;
- q = port->queue + qindex;
- qsize = rte_sched_port_qsize(port, qindex);
+ q = subport->queue + qindex;
+ qsize = rte_sched_subport_pipe_qsize(port, subport, qindex);
qlen = q->qw - q->qr;
/* Drop the packet (and update drop stats) when queue is full */
- if (unlikely(rte_sched_port_red_drop(port, pkt, qindex, qlen) ||
+ if (unlikely(rte_sched_port_red_drop(port, subport, pkt, qindex, qlen) ||
(qlen >= qsize))) {
rte_pktmbuf_free(pkt);
#ifdef RTE_SCHED_COLLECT_STATS
- rte_sched_port_update_subport_stats_on_drop(port, qindex, pkt,
- qlen < qsize);
- rte_sched_port_update_queue_stats_on_drop(port, qindex, pkt,
- qlen < qsize);
+ rte_sched_port_update_subport_stats_on_drop(port, subport,
+ qindex, pkt, qlen < qsize);
+ rte_sched_port_update_queue_stats_on_drop(subport, qindex, pkt,
+ qlen < qsize);
#endif
return 0;
}
qbase[q->qw & (qsize - 1)] = pkt;
q->qw++;
- /* Activate queue in the port bitmap */
- rte_bitmap_set(port->bmp, qindex);
+ /* Activate queue in the subport bitmap */
+ rte_bitmap_set(subport->bmp, qindex);
/* Statistics */
#ifdef RTE_SCHED_COLLECT_STATS
- rte_sched_port_update_subport_stats(port, qindex, pkt);
- rte_sched_port_update_queue_stats(port, qindex, pkt);
+ rte_sched_port_update_subport_stats(port, subport, qindex, pkt);
+ rte_sched_port_update_queue_stats(subport, qindex, pkt);
#endif
return 1;
*pkt30, *pkt31, *pkt_last;
struct rte_mbuf **q00_base, **q01_base, **q10_base, **q11_base,
**q20_base, **q21_base, **q30_base, **q31_base, **q_last_base;
+ struct rte_sched_subport *subport00, *subport01, *subport10, *subport11,
+ *subport20, *subport21, *subport30, *subport31, *subport_last;
uint32_t q00, q01, q10, q11, q20, q21, q30, q31, q_last;
uint32_t r00, r01, r10, r11, r20, r21, r30, r31, r_last;
+ uint32_t subport_qmask;
uint32_t result, i;
result = 0;
+ subport_qmask = (1 << (port->n_pipes_per_subport_log2 + 4)) - 1;
/*
* Less then 6 input packets available, which is not enough to
* feed the pipeline
*/
if (unlikely(n_pkts < 6)) {
+ struct rte_sched_subport *subports[5];
struct rte_mbuf **q_base[5];
uint32_t q[5];
for (i = 0; i < n_pkts; i++)
rte_prefetch0(pkts[i]);
+ /* Prefetch the subport structure for each packet */
+ for (i = 0; i < n_pkts; i++)
+ subports[i] = rte_sched_port_subport(port, pkts[i]);
+
/* Prefetch the queue structure for each queue */
for (i = 0; i < n_pkts; i++)
- q[i] = rte_sched_port_enqueue_qptrs_prefetch0(port,
- pkts[i]);
+ q[i] = rte_sched_port_enqueue_qptrs_prefetch0(subports[i],
+ pkts[i], subport_qmask);
/* Prefetch the write pointer location of each queue */
for (i = 0; i < n_pkts; i++) {
- q_base[i] = rte_sched_port_qbase(port, q[i]);
- rte_sched_port_enqueue_qwa_prefetch0(port, q[i],
- q_base[i]);
+ q_base[i] = rte_sched_subport_pipe_qbase(subports[i], q[i]);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subports[i],
+ q[i], q_base[i]);
}
/* Write each packet to its queue */
for (i = 0; i < n_pkts; i++)
- result += rte_sched_port_enqueue_qwa(port, q[i],
- q_base[i], pkts[i]);
+ result += rte_sched_port_enqueue_qwa(port, subports[i],
+ q[i], q_base[i], pkts[i]);
return result;
}
rte_prefetch0(pkt10);
rte_prefetch0(pkt11);
- q20 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt20);
- q21 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt21);
+ subport20 = rte_sched_port_subport(port, pkt20);
+ subport21 = rte_sched_port_subport(port, pkt21);
+ q20 = rte_sched_port_enqueue_qptrs_prefetch0(subport20,
+ pkt20, subport_qmask);
+ q21 = rte_sched_port_enqueue_qptrs_prefetch0(subport21,
+ pkt21, subport_qmask);
pkt00 = pkts[4];
pkt01 = pkts[5];
rte_prefetch0(pkt00);
rte_prefetch0(pkt01);
- q10 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt10);
- q11 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt11);
+ subport10 = rte_sched_port_subport(port, pkt10);
+ subport11 = rte_sched_port_subport(port, pkt11);
+ q10 = rte_sched_port_enqueue_qptrs_prefetch0(subport10,
+ pkt10, subport_qmask);
+ q11 = rte_sched_port_enqueue_qptrs_prefetch0(subport11,
+ pkt11, subport_qmask);
- q20_base = rte_sched_port_qbase(port, q20);
- q21_base = rte_sched_port_qbase(port, q21);
- rte_sched_port_enqueue_qwa_prefetch0(port, q20, q20_base);
- rte_sched_port_enqueue_qwa_prefetch0(port, q21, q21_base);
+ q20_base = rte_sched_subport_pipe_qbase(subport20, q20);
+ q21_base = rte_sched_subport_pipe_qbase(subport21, q21);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport20, q20, q20_base);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport21, q21, q21_base);
/* Run the pipeline */
for (i = 6; i < (n_pkts & (~1)); i += 2) {
q31 = q21;
q20 = q10;
q21 = q11;
+ subport30 = subport20;
+ subport31 = subport21;
+ subport20 = subport10;
+ subport21 = subport11;
q30_base = q20_base;
q31_base = q21_base;
rte_prefetch0(pkt00);
rte_prefetch0(pkt01);
- /* Stage 1: Prefetch queue structure storing queue pointers */
- q10 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt10);
- q11 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt11);
+ /* Stage 1: Prefetch subport and queue structure storing queue pointers */
+ subport10 = rte_sched_port_subport(port, pkt10);
+ subport11 = rte_sched_port_subport(port, pkt11);
+ q10 = rte_sched_port_enqueue_qptrs_prefetch0(subport10,
+ pkt10, subport_qmask);
+ q11 = rte_sched_port_enqueue_qptrs_prefetch0(subport11,
+ pkt11, subport_qmask);
/* Stage 2: Prefetch queue write location */
- q20_base = rte_sched_port_qbase(port, q20);
- q21_base = rte_sched_port_qbase(port, q21);
- rte_sched_port_enqueue_qwa_prefetch0(port, q20, q20_base);
- rte_sched_port_enqueue_qwa_prefetch0(port, q21, q21_base);
+ q20_base = rte_sched_subport_pipe_qbase(subport20, q20);
+ q21_base = rte_sched_subport_pipe_qbase(subport21, q21);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport20, q20, q20_base);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport21, q21, q21_base);
/* Stage 3: Write packet to queue and activate queue */
- r30 = rte_sched_port_enqueue_qwa(port, q30, q30_base, pkt30);
- r31 = rte_sched_port_enqueue_qwa(port, q31, q31_base, pkt31);
+ r30 = rte_sched_port_enqueue_qwa(port, subport30,
+ q30, q30_base, pkt30);
+ r31 = rte_sched_port_enqueue_qwa(port, subport31,
+ q31, q31_base, pkt31);
result += r30 + r31;
}
pkt_last = pkts[n_pkts - 1];
rte_prefetch0(pkt_last);
- q00 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt00);
- q01 = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt01);
-
- q10_base = rte_sched_port_qbase(port, q10);
- q11_base = rte_sched_port_qbase(port, q11);
- rte_sched_port_enqueue_qwa_prefetch0(port, q10, q10_base);
- rte_sched_port_enqueue_qwa_prefetch0(port, q11, q11_base);
-
- r20 = rte_sched_port_enqueue_qwa(port, q20, q20_base, pkt20);
- r21 = rte_sched_port_enqueue_qwa(port, q21, q21_base, pkt21);
+ subport00 = rte_sched_port_subport(port, pkt00);
+ subport01 = rte_sched_port_subport(port, pkt01);
+ q00 = rte_sched_port_enqueue_qptrs_prefetch0(subport00,
+ pkt00, subport_qmask);
+ q01 = rte_sched_port_enqueue_qptrs_prefetch0(subport01,
+ pkt01, subport_qmask);
+
+ q10_base = rte_sched_subport_pipe_qbase(subport10, q10);
+ q11_base = rte_sched_subport_pipe_qbase(subport11, q11);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport10, q10, q10_base);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport11, q11, q11_base);
+
+ r20 = rte_sched_port_enqueue_qwa(port, subport20,
+ q20, q20_base, pkt20);
+ r21 = rte_sched_port_enqueue_qwa(port, subport21,
+ q21, q21_base, pkt21);
result += r20 + r21;
- q_last = rte_sched_port_enqueue_qptrs_prefetch0(port, pkt_last);
+ subport_last = rte_sched_port_subport(port, pkt_last);
+ q_last = rte_sched_port_enqueue_qptrs_prefetch0(subport_last,
+ pkt_last, subport_qmask);
- q00_base = rte_sched_port_qbase(port, q00);
- q01_base = rte_sched_port_qbase(port, q01);
- rte_sched_port_enqueue_qwa_prefetch0(port, q00, q00_base);
- rte_sched_port_enqueue_qwa_prefetch0(port, q01, q01_base);
+ q00_base = rte_sched_subport_pipe_qbase(subport00, q00);
+ q01_base = rte_sched_subport_pipe_qbase(subport01, q01);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport00, q00, q00_base);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport01, q01, q01_base);
- r10 = rte_sched_port_enqueue_qwa(port, q10, q10_base, pkt10);
- r11 = rte_sched_port_enqueue_qwa(port, q11, q11_base, pkt11);
+ r10 = rte_sched_port_enqueue_qwa(port, subport10, q10,
+ q10_base, pkt10);
+ r11 = rte_sched_port_enqueue_qwa(port, subport11, q11,
+ q11_base, pkt11);
result += r10 + r11;
- q_last_base = rte_sched_port_qbase(port, q_last);
- rte_sched_port_enqueue_qwa_prefetch0(port, q_last, q_last_base);
+ q_last_base = rte_sched_subport_pipe_qbase(subport_last, q_last);
+ rte_sched_port_enqueue_qwa_prefetch0(port, subport_last,
+ q_last, q_last_base);
- r00 = rte_sched_port_enqueue_qwa(port, q00, q00_base, pkt00);
- r01 = rte_sched_port_enqueue_qwa(port, q01, q01_base, pkt01);
+ r00 = rte_sched_port_enqueue_qwa(port, subport00, q00,
+ q00_base, pkt00);
+ r01 = rte_sched_port_enqueue_qwa(port, subport01, q01,
+ q01_base, pkt01);
result += r00 + r01;
if (n_pkts & 1) {
- r_last = rte_sched_port_enqueue_qwa(port, q_last, q_last_base, pkt_last);
+ r_last = rte_sched_port_enqueue_qwa(port, subport_last,
+ q_last, q_last_base, pkt_last);
result += r_last;
}
#ifndef RTE_SCHED_SUBPORT_TC_OV
static inline void
-grinder_credits_update(struct rte_sched_port *port, uint32_t pos)
+grinder_credits_update(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- struct rte_sched_subport *subport = grinder->subport;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_pipe *pipe = grinder->pipe;
struct rte_sched_pipe_profile *params = grinder->pipe_params;
uint64_t n_periods;
+ uint32_t i;
/* Subport TB */
n_periods = (port->time - subport->tb_time) / subport->tb_period;
subport->tb_credits += n_periods * subport->tb_credits_per_period;
- subport->tb_credits = rte_sched_min_val_2_u32(subport->tb_credits, subport->tb_size);
+ subport->tb_credits = RTE_MIN(subport->tb_credits, subport->tb_size);
subport->tb_time += n_periods * subport->tb_period;
/* Pipe TB */
n_periods = (port->time - pipe->tb_time) / params->tb_period;
pipe->tb_credits += n_periods * params->tb_credits_per_period;
- pipe->tb_credits = rte_sched_min_val_2_u32(pipe->tb_credits, params->tb_size);
+ pipe->tb_credits = RTE_MIN(pipe->tb_credits, params->tb_size);
pipe->tb_time += n_periods * params->tb_period;
/* Subport TCs */
if (unlikely(port->time >= subport->tc_time)) {
- subport->tc_credits[0] = subport->tc_credits_per_period[0];
- subport->tc_credits[1] = subport->tc_credits_per_period[1];
- subport->tc_credits[2] = subport->tc_credits_per_period[2];
- subport->tc_credits[3] = subport->tc_credits_per_period[3];
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ subport->tc_credits[i] = subport->tc_credits_per_period[i];
+
subport->tc_time = port->time + subport->tc_period;
}
/* Pipe TCs */
if (unlikely(port->time >= pipe->tc_time)) {
- pipe->tc_credits[0] = params->tc_credits_per_period[0];
- pipe->tc_credits[1] = params->tc_credits_per_period[1];
- pipe->tc_credits[2] = params->tc_credits_per_period[2];
- pipe->tc_credits[3] = params->tc_credits_per_period[3];
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ pipe->tc_credits[i] = params->tc_credits_per_period[i];
+
pipe->tc_time = port->time + params->tc_period;
}
}
#else
-static inline uint32_t
-grinder_tc_ov_credits_update(struct rte_sched_port *port, uint32_t pos)
+static inline uint64_t
+grinder_tc_ov_credits_update(struct rte_sched_port *port,
+ struct rte_sched_subport *subport)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- struct rte_sched_subport *subport = grinder->subport;
- uint32_t tc_ov_consumption[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
- uint32_t tc_ov_consumption_max;
- uint32_t tc_ov_wm = subport->tc_ov_wm;
+ uint64_t tc_ov_consumption[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint64_t tc_consumption = 0, tc_ov_consumption_max;
+ uint64_t tc_ov_wm = subport->tc_ov_wm;
+ uint32_t i;
if (subport->tc_ov == 0)
return subport->tc_ov_wm_max;
- tc_ov_consumption[0] = subport->tc_credits_per_period[0] - subport->tc_credits[0];
- tc_ov_consumption[1] = subport->tc_credits_per_period[1] - subport->tc_credits[1];
- tc_ov_consumption[2] = subport->tc_credits_per_period[2] - subport->tc_credits[2];
- tc_ov_consumption[3] = subport->tc_credits_per_period[3] - subport->tc_credits[3];
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASS_BE; i++) {
+ tc_ov_consumption[i] =
+ subport->tc_credits_per_period[i] - subport->tc_credits[i];
+ tc_consumption += tc_ov_consumption[i];
+ }
+
+ tc_ov_consumption[RTE_SCHED_TRAFFIC_CLASS_BE] =
+ subport->tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASS_BE] -
+ subport->tc_credits[RTE_SCHED_TRAFFIC_CLASS_BE];
- tc_ov_consumption_max = subport->tc_credits_per_period[3] -
- (tc_ov_consumption[0] + tc_ov_consumption[1] + tc_ov_consumption[2]);
+ tc_ov_consumption_max =
+ subport->tc_credits_per_period[RTE_SCHED_TRAFFIC_CLASS_BE] -
+ tc_consumption;
- if (tc_ov_consumption[3] > (tc_ov_consumption_max - port->mtu)) {
+ if (tc_ov_consumption[RTE_SCHED_TRAFFIC_CLASS_BE] >
+ (tc_ov_consumption_max - port->mtu)) {
tc_ov_wm -= tc_ov_wm >> 7;
if (tc_ov_wm < subport->tc_ov_wm_min)
tc_ov_wm = subport->tc_ov_wm_min;
}
static inline void
-grinder_credits_update(struct rte_sched_port *port, uint32_t pos)
+grinder_credits_update(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- struct rte_sched_subport *subport = grinder->subport;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_pipe *pipe = grinder->pipe;
struct rte_sched_pipe_profile *params = grinder->pipe_params;
uint64_t n_periods;
+ uint32_t i;
/* Subport TB */
n_periods = (port->time - subport->tb_time) / subport->tb_period;
subport->tb_credits += n_periods * subport->tb_credits_per_period;
- subport->tb_credits = rte_sched_min_val_2_u32(subport->tb_credits, subport->tb_size);
+ subport->tb_credits = RTE_MIN(subport->tb_credits, subport->tb_size);
subport->tb_time += n_periods * subport->tb_period;
/* Pipe TB */
n_periods = (port->time - pipe->tb_time) / params->tb_period;
pipe->tb_credits += n_periods * params->tb_credits_per_period;
- pipe->tb_credits = rte_sched_min_val_2_u32(pipe->tb_credits, params->tb_size);
+ pipe->tb_credits = RTE_MIN(pipe->tb_credits, params->tb_size);
pipe->tb_time += n_periods * params->tb_period;
/* Subport TCs */
if (unlikely(port->time >= subport->tc_time)) {
- subport->tc_ov_wm = grinder_tc_ov_credits_update(port, pos);
+ subport->tc_ov_wm = grinder_tc_ov_credits_update(port, subport);
- subport->tc_credits[0] = subport->tc_credits_per_period[0];
- subport->tc_credits[1] = subport->tc_credits_per_period[1];
- subport->tc_credits[2] = subport->tc_credits_per_period[2];
- subport->tc_credits[3] = subport->tc_credits_per_period[3];
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ subport->tc_credits[i] = subport->tc_credits_per_period[i];
subport->tc_time = port->time + subport->tc_period;
subport->tc_ov_period_id++;
/* Pipe TCs */
if (unlikely(port->time >= pipe->tc_time)) {
- pipe->tc_credits[0] = params->tc_credits_per_period[0];
- pipe->tc_credits[1] = params->tc_credits_per_period[1];
- pipe->tc_credits[2] = params->tc_credits_per_period[2];
- pipe->tc_credits[3] = params->tc_credits_per_period[3];
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ pipe->tc_credits[i] = params->tc_credits_per_period[i];
pipe->tc_time = port->time + params->tc_period;
}
#ifndef RTE_SCHED_SUBPORT_TC_OV
static inline int
-grinder_credits_check(struct rte_sched_port *port, uint32_t pos)
+grinder_credits_check(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- struct rte_sched_subport *subport = grinder->subport;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_pipe *pipe = grinder->pipe;
struct rte_mbuf *pkt = grinder->pkt;
uint32_t tc_index = grinder->tc_index;
- uint32_t pkt_len = pkt->pkt_len + port->frame_overhead;
- uint32_t subport_tb_credits = subport->tb_credits;
- uint32_t subport_tc_credits = subport->tc_credits[tc_index];
- uint32_t pipe_tb_credits = pipe->tb_credits;
- uint32_t pipe_tc_credits = pipe->tc_credits[tc_index];
+ uint64_t pkt_len = pkt->pkt_len + port->frame_overhead;
+ uint64_t subport_tb_credits = subport->tb_credits;
+ uint64_t subport_tc_credits = subport->tc_credits[tc_index];
+ uint64_t pipe_tb_credits = pipe->tb_credits;
+ uint64_t pipe_tc_credits = pipe->tc_credits[tc_index];
int enough_credits;
/* Check queue credits */
#else
static inline int
-grinder_credits_check(struct rte_sched_port *port, uint32_t pos)
+grinder_credits_check(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- struct rte_sched_subport *subport = grinder->subport;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_pipe *pipe = grinder->pipe;
struct rte_mbuf *pkt = grinder->pkt;
uint32_t tc_index = grinder->tc_index;
- uint32_t pkt_len = pkt->pkt_len + port->frame_overhead;
- uint32_t subport_tb_credits = subport->tb_credits;
- uint32_t subport_tc_credits = subport->tc_credits[tc_index];
- uint32_t pipe_tb_credits = pipe->tb_credits;
- uint32_t pipe_tc_credits = pipe->tc_credits[tc_index];
- uint32_t pipe_tc_ov_mask1[] = {UINT32_MAX, UINT32_MAX, UINT32_MAX, pipe->tc_ov_credits};
- uint32_t pipe_tc_ov_mask2[] = {0, 0, 0, UINT32_MAX};
- uint32_t pipe_tc_ov_credits = pipe_tc_ov_mask1[tc_index];
+ uint64_t pkt_len = pkt->pkt_len + port->frame_overhead;
+ uint64_t subport_tb_credits = subport->tb_credits;
+ uint64_t subport_tc_credits = subport->tc_credits[tc_index];
+ uint64_t pipe_tb_credits = pipe->tb_credits;
+ uint64_t pipe_tc_credits = pipe->tc_credits[tc_index];
+ uint64_t pipe_tc_ov_mask1[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE];
+ uint64_t pipe_tc_ov_mask2[RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE] = {0};
+ uint64_t pipe_tc_ov_credits;
+ uint32_t i;
int enough_credits;
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
+ pipe_tc_ov_mask1[i] = ~0LLU;
+
+ pipe_tc_ov_mask1[RTE_SCHED_TRAFFIC_CLASS_BE] = pipe->tc_ov_credits;
+ pipe_tc_ov_mask2[RTE_SCHED_TRAFFIC_CLASS_BE] = ~0LLU;
+ pipe_tc_ov_credits = pipe_tc_ov_mask1[tc_index];
+
/* Check pipe and subport credits */
enough_credits = (pkt_len <= subport_tb_credits) &&
(pkt_len <= subport_tc_credits) &&
static inline int
-grinder_schedule(struct rte_sched_port *port, uint32_t pos)
+grinder_schedule(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_queue *queue = grinder->queue[grinder->qpos];
struct rte_mbuf *pkt = grinder->pkt;
uint32_t pkt_len = pkt->pkt_len + port->frame_overhead;
+ uint32_t be_tc_active;
- if (!grinder_credits_check(port, pos))
+ if (!grinder_credits_check(port, subport, pos))
return 0;
/* Advance port time */
/* Send packet */
port->pkts_out[port->n_pkts_out++] = pkt;
queue->qr++;
- grinder->wrr_tokens[grinder->qpos] += pkt_len * grinder->wrr_cost[grinder->qpos];
+
+ be_tc_active = (grinder->tc_index == RTE_SCHED_TRAFFIC_CLASS_BE) ? ~0x0 : 0x0;
+ grinder->wrr_tokens[grinder->qpos] +=
+ (pkt_len * grinder->wrr_cost[grinder->qpos]) & be_tc_active;
+
if (queue->qr == queue->qw) {
uint32_t qindex = grinder->qindex[grinder->qpos];
- rte_bitmap_clear(port->bmp, qindex);
+ rte_bitmap_clear(subport->bmp, qindex);
grinder->qmask &= ~(1 << grinder->qpos);
- grinder->wrr_mask[grinder->qpos] = 0;
- rte_sched_port_set_queue_empty_timestamp(port, qindex);
+ if (be_tc_active)
+ grinder->wrr_mask[grinder->qpos] = 0;
+ rte_sched_port_set_queue_empty_timestamp(port, subport, qindex);
}
/* Reset pipe loop detection */
- port->pipe_loop = RTE_SCHED_PIPE_INVALID;
+ subport->pipe_loop = RTE_SCHED_PIPE_INVALID;
grinder->productive = 1;
return 1;
#ifdef SCHED_VECTOR_SSE4
static inline int
-grinder_pipe_exists(struct rte_sched_port *port, uint32_t base_pipe)
+grinder_pipe_exists(struct rte_sched_subport *subport, uint32_t base_pipe)
{
__m128i index = _mm_set1_epi32(base_pipe);
- __m128i pipes = _mm_load_si128((__m128i *)port->grinder_base_bmp_pos);
+ __m128i pipes = _mm_load_si128((__m128i *)subport->grinder_base_bmp_pos);
__m128i res = _mm_cmpeq_epi32(pipes, index);
- pipes = _mm_load_si128((__m128i *)(port->grinder_base_bmp_pos + 4));
+ pipes = _mm_load_si128((__m128i *)(subport->grinder_base_bmp_pos + 4));
pipes = _mm_cmpeq_epi32(pipes, index);
res = _mm_or_si128(res, pipes);
#elif defined(SCHED_VECTOR_NEON)
static inline int
-grinder_pipe_exists(struct rte_sched_port *port, uint32_t base_pipe)
+grinder_pipe_exists(struct rte_sched_subport *subport, uint32_t base_pipe)
{
uint32x4_t index, pipes;
- uint32_t *pos = (uint32_t *)port->grinder_base_bmp_pos;
+ uint32_t *pos = (uint32_t *)subport->grinder_base_bmp_pos;
index = vmovq_n_u32(base_pipe);
pipes = vld1q_u32(pos);
#else
static inline int
-grinder_pipe_exists(struct rte_sched_port *port, uint32_t base_pipe)
+grinder_pipe_exists(struct rte_sched_subport *subport, uint32_t base_pipe)
{
uint32_t i;
for (i = 0; i < RTE_SCHED_PORT_N_GRINDERS; i++) {
- if (port->grinder_base_bmp_pos[i] == base_pipe)
+ if (subport->grinder_base_bmp_pos[i] == base_pipe)
return 1;
}
#endif /* RTE_SCHED_OPTIMIZATIONS */
static inline void
-grinder_pcache_populate(struct rte_sched_port *port, uint32_t pos, uint32_t bmp_pos, uint64_t bmp_slab)
+grinder_pcache_populate(struct rte_sched_subport *subport,
+ uint32_t pos, uint32_t bmp_pos, uint64_t bmp_slab)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
uint16_t w[4];
grinder->pcache_w = 0;
}
static inline void
-grinder_tccache_populate(struct rte_sched_port *port, uint32_t pos, uint32_t qindex, uint16_t qmask)
+grinder_tccache_populate(struct rte_sched_subport *subport,
+ uint32_t pos, uint32_t qindex, uint16_t qmask)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- uint8_t b[4];
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
+ uint8_t b, i;
grinder->tccache_w = 0;
grinder->tccache_r = 0;
- b[0] = (uint8_t) (qmask & 0xF);
- b[1] = (uint8_t) ((qmask >> 4) & 0xF);
- b[2] = (uint8_t) ((qmask >> 8) & 0xF);
- b[3] = (uint8_t) ((qmask >> 12) & 0xF);
-
- grinder->tccache_qmask[grinder->tccache_w] = b[0];
- grinder->tccache_qindex[grinder->tccache_w] = qindex;
- grinder->tccache_w += (b[0] != 0);
-
- grinder->tccache_qmask[grinder->tccache_w] = b[1];
- grinder->tccache_qindex[grinder->tccache_w] = qindex + 4;
- grinder->tccache_w += (b[1] != 0);
-
- grinder->tccache_qmask[grinder->tccache_w] = b[2];
- grinder->tccache_qindex[grinder->tccache_w] = qindex + 8;
- grinder->tccache_w += (b[2] != 0);
+ for (i = 0; i < RTE_SCHED_TRAFFIC_CLASS_BE; i++) {
+ b = (uint8_t) ((qmask >> i) & 0x1);
+ grinder->tccache_qmask[grinder->tccache_w] = b;
+ grinder->tccache_qindex[grinder->tccache_w] = qindex + i;
+ grinder->tccache_w += (b != 0);
+ }
- grinder->tccache_qmask[grinder->tccache_w] = b[3];
- grinder->tccache_qindex[grinder->tccache_w] = qindex + 12;
- grinder->tccache_w += (b[3] != 0);
+ b = (uint8_t) (qmask >> (RTE_SCHED_TRAFFIC_CLASS_BE));
+ grinder->tccache_qmask[grinder->tccache_w] = b;
+ grinder->tccache_qindex[grinder->tccache_w] = qindex +
+ RTE_SCHED_TRAFFIC_CLASS_BE;
+ grinder->tccache_w += (b != 0);
}
static inline int
-grinder_next_tc(struct rte_sched_port *port, uint32_t pos)
+grinder_next_tc(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_mbuf **qbase;
uint32_t qindex;
uint16_t qsize;
return 0;
qindex = grinder->tccache_qindex[grinder->tccache_r];
- qbase = rte_sched_port_qbase(port, qindex);
- qsize = rte_sched_port_qsize(port, qindex);
+ qbase = rte_sched_subport_pipe_qbase(subport, qindex);
+ qsize = rte_sched_subport_pipe_qsize(port, subport, qindex);
- grinder->tc_index = (qindex >> 2) & 0x3;
+ grinder->tc_index = rte_sched_port_pipe_tc(port, qindex);
grinder->qmask = grinder->tccache_qmask[grinder->tccache_r];
grinder->qsize = qsize;
- grinder->qindex[0] = qindex;
- grinder->qindex[1] = qindex + 1;
- grinder->qindex[2] = qindex + 2;
- grinder->qindex[3] = qindex + 3;
+ if (grinder->tc_index < RTE_SCHED_TRAFFIC_CLASS_BE) {
+ grinder->queue[0] = subport->queue + qindex;
+ grinder->qbase[0] = qbase;
+ grinder->qindex[0] = qindex;
+ grinder->tccache_r++;
+
+ return 1;
+ }
- grinder->queue[0] = port->queue + qindex;
- grinder->queue[1] = port->queue + qindex + 1;
- grinder->queue[2] = port->queue + qindex + 2;
- grinder->queue[3] = port->queue + qindex + 3;
+ grinder->queue[0] = subport->queue + qindex;
+ grinder->queue[1] = subport->queue + qindex + 1;
+ grinder->queue[2] = subport->queue + qindex + 2;
+ grinder->queue[3] = subport->queue + qindex + 3;
grinder->qbase[0] = qbase;
grinder->qbase[1] = qbase + qsize;
grinder->qbase[2] = qbase + 2 * qsize;
grinder->qbase[3] = qbase + 3 * qsize;
+ grinder->qindex[0] = qindex;
+ grinder->qindex[1] = qindex + 1;
+ grinder->qindex[2] = qindex + 2;
+ grinder->qindex[3] = qindex + 3;
+
grinder->tccache_r++;
return 1;
}
static inline int
-grinder_next_pipe(struct rte_sched_port *port, uint32_t pos)
+grinder_next_pipe(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
uint32_t pipe_qindex;
uint16_t pipe_qmask;
uint32_t bmp_pos = 0;
/* Get another non-empty pipe group */
- if (unlikely(rte_bitmap_scan(port->bmp, &bmp_pos, &bmp_slab) <= 0))
+ if (unlikely(rte_bitmap_scan(subport->bmp, &bmp_pos, &bmp_slab) <= 0))
return 0;
#ifdef RTE_SCHED_DEBUG
- debug_check_queue_slab(port, bmp_pos, bmp_slab);
+ debug_check_queue_slab(subport, bmp_pos, bmp_slab);
#endif
/* Return if pipe group already in one of the other grinders */
- port->grinder_base_bmp_pos[pos] = RTE_SCHED_BMP_POS_INVALID;
- if (unlikely(grinder_pipe_exists(port, bmp_pos)))
+ subport->grinder_base_bmp_pos[pos] = RTE_SCHED_BMP_POS_INVALID;
+ if (unlikely(grinder_pipe_exists(subport, bmp_pos)))
return 0;
- port->grinder_base_bmp_pos[pos] = bmp_pos;
+ subport->grinder_base_bmp_pos[pos] = bmp_pos;
/* Install new pipe group into grinder's pipe cache */
- grinder_pcache_populate(port, pos, bmp_pos, bmp_slab);
+ grinder_pcache_populate(subport, pos, bmp_pos, bmp_slab);
pipe_qmask = grinder->pcache_qmask[0];
pipe_qindex = grinder->pcache_qindex[0];
/* Install new pipe in the grinder */
grinder->pindex = pipe_qindex >> 4;
- grinder->subport = port->subport + (grinder->pindex / port->n_pipes_per_subport);
- grinder->pipe = port->pipe + grinder->pindex;
+ grinder->subport = subport;
+ grinder->pipe = subport->pipe + grinder->pindex;
grinder->pipe_params = NULL; /* to be set after the pipe structure is prefetched */
grinder->productive = 0;
- grinder_tccache_populate(port, pos, pipe_qindex, pipe_qmask);
- grinder_next_tc(port, pos);
+ grinder_tccache_populate(subport, pos, pipe_qindex, pipe_qmask);
+ grinder_next_tc(port, subport, pos);
/* Check for pipe exhaustion */
- if (grinder->pindex == port->pipe_loop) {
- port->pipe_exhaustion = 1;
- port->pipe_loop = RTE_SCHED_PIPE_INVALID;
+ if (grinder->pindex == subport->pipe_loop) {
+ subport->pipe_exhaustion = 1;
+ subport->pipe_loop = RTE_SCHED_PIPE_INVALID;
}
return 1;
static inline void
-grinder_wrr_load(struct rte_sched_port *port, uint32_t pos)
+grinder_wrr_load(struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_pipe *pipe = grinder->pipe;
struct rte_sched_pipe_profile *pipe_params = grinder->pipe_params;
- uint32_t tc_index = grinder->tc_index;
uint32_t qmask = grinder->qmask;
- uint32_t qindex;
- qindex = tc_index * 4;
-
- grinder->wrr_tokens[0] = ((uint16_t) pipe->wrr_tokens[qindex]) << RTE_SCHED_WRR_SHIFT;
- grinder->wrr_tokens[1] = ((uint16_t) pipe->wrr_tokens[qindex + 1]) << RTE_SCHED_WRR_SHIFT;
- grinder->wrr_tokens[2] = ((uint16_t) pipe->wrr_tokens[qindex + 2]) << RTE_SCHED_WRR_SHIFT;
- grinder->wrr_tokens[3] = ((uint16_t) pipe->wrr_tokens[qindex + 3]) << RTE_SCHED_WRR_SHIFT;
+ grinder->wrr_tokens[0] =
+ ((uint16_t) pipe->wrr_tokens[0]) << RTE_SCHED_WRR_SHIFT;
+ grinder->wrr_tokens[1] =
+ ((uint16_t) pipe->wrr_tokens[1]) << RTE_SCHED_WRR_SHIFT;
+ grinder->wrr_tokens[2] =
+ ((uint16_t) pipe->wrr_tokens[2]) << RTE_SCHED_WRR_SHIFT;
+ grinder->wrr_tokens[3] =
+ ((uint16_t) pipe->wrr_tokens[3]) << RTE_SCHED_WRR_SHIFT;
grinder->wrr_mask[0] = (qmask & 0x1) * 0xFFFF;
grinder->wrr_mask[1] = ((qmask >> 1) & 0x1) * 0xFFFF;
grinder->wrr_mask[2] = ((qmask >> 2) & 0x1) * 0xFFFF;
grinder->wrr_mask[3] = ((qmask >> 3) & 0x1) * 0xFFFF;
- grinder->wrr_cost[0] = pipe_params->wrr_cost[qindex];
- grinder->wrr_cost[1] = pipe_params->wrr_cost[qindex + 1];
- grinder->wrr_cost[2] = pipe_params->wrr_cost[qindex + 2];
- grinder->wrr_cost[3] = pipe_params->wrr_cost[qindex + 3];
+ grinder->wrr_cost[0] = pipe_params->wrr_cost[0];
+ grinder->wrr_cost[1] = pipe_params->wrr_cost[1];
+ grinder->wrr_cost[2] = pipe_params->wrr_cost[2];
+ grinder->wrr_cost[3] = pipe_params->wrr_cost[3];
}
static inline void
-grinder_wrr_store(struct rte_sched_port *port, uint32_t pos)
+grinder_wrr_store(struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
struct rte_sched_pipe *pipe = grinder->pipe;
- uint32_t tc_index = grinder->tc_index;
- uint32_t qindex;
-
- qindex = tc_index * 4;
- pipe->wrr_tokens[qindex] = (grinder->wrr_tokens[0] & grinder->wrr_mask[0])
- >> RTE_SCHED_WRR_SHIFT;
- pipe->wrr_tokens[qindex + 1] = (grinder->wrr_tokens[1] & grinder->wrr_mask[1])
- >> RTE_SCHED_WRR_SHIFT;
- pipe->wrr_tokens[qindex + 2] = (grinder->wrr_tokens[2] & grinder->wrr_mask[2])
- >> RTE_SCHED_WRR_SHIFT;
- pipe->wrr_tokens[qindex + 3] = (grinder->wrr_tokens[3] & grinder->wrr_mask[3])
- >> RTE_SCHED_WRR_SHIFT;
+ pipe->wrr_tokens[0] =
+ (grinder->wrr_tokens[0] & grinder->wrr_mask[0]) >>
+ RTE_SCHED_WRR_SHIFT;
+ pipe->wrr_tokens[1] =
+ (grinder->wrr_tokens[1] & grinder->wrr_mask[1]) >>
+ RTE_SCHED_WRR_SHIFT;
+ pipe->wrr_tokens[2] =
+ (grinder->wrr_tokens[2] & grinder->wrr_mask[2]) >>
+ RTE_SCHED_WRR_SHIFT;
+ pipe->wrr_tokens[3] =
+ (grinder->wrr_tokens[3] & grinder->wrr_mask[3]) >>
+ RTE_SCHED_WRR_SHIFT;
}
static inline void
-grinder_wrr(struct rte_sched_port *port, uint32_t pos)
+grinder_wrr(struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
uint16_t wrr_tokens_min;
grinder->wrr_tokens[0] |= ~grinder->wrr_mask[0];
}
-#define grinder_evict(port, pos)
+#define grinder_evict(subport, pos)
static inline void
-grinder_prefetch_pipe(struct rte_sched_port *port, uint32_t pos)
+grinder_prefetch_pipe(struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
rte_prefetch0(grinder->pipe);
rte_prefetch0(grinder->queue[0]);
}
static inline void
-grinder_prefetch_tc_queue_arrays(struct rte_sched_port *port, uint32_t pos)
+grinder_prefetch_tc_queue_arrays(struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
- uint16_t qsize, qr[4];
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
+ uint16_t qsize, qr[RTE_SCHED_MAX_QUEUES_PER_TC];
qsize = grinder->qsize;
+ grinder->qpos = 0;
+
+ if (grinder->tc_index < RTE_SCHED_TRAFFIC_CLASS_BE) {
+ qr[0] = grinder->queue[0]->qr & (qsize - 1);
+
+ rte_prefetch0(grinder->qbase[0] + qr[0]);
+ return;
+ }
+
qr[0] = grinder->queue[0]->qr & (qsize - 1);
qr[1] = grinder->queue[1]->qr & (qsize - 1);
qr[2] = grinder->queue[2]->qr & (qsize - 1);
rte_prefetch0(grinder->qbase[0] + qr[0]);
rte_prefetch0(grinder->qbase[1] + qr[1]);
- grinder_wrr_load(port, pos);
- grinder_wrr(port, pos);
+ grinder_wrr_load(subport, pos);
+ grinder_wrr(subport, pos);
rte_prefetch0(grinder->qbase[2] + qr[2]);
rte_prefetch0(grinder->qbase[3] + qr[3]);
}
static inline void
-grinder_prefetch_mbuf(struct rte_sched_port *port, uint32_t pos)
+grinder_prefetch_mbuf(struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
uint32_t qpos = grinder->qpos;
struct rte_mbuf **qbase = grinder->qbase[qpos];
uint16_t qsize = grinder->qsize;
}
static inline uint32_t
-grinder_handle(struct rte_sched_port *port, uint32_t pos)
+grinder_handle(struct rte_sched_port *port,
+ struct rte_sched_subport *subport, uint32_t pos)
{
- struct rte_sched_grinder *grinder = port->grinder + pos;
+ struct rte_sched_grinder *grinder = subport->grinder + pos;
switch (grinder->state) {
case e_GRINDER_PREFETCH_PIPE:
{
- if (grinder_next_pipe(port, pos)) {
- grinder_prefetch_pipe(port, pos);
- port->busy_grinders++;
+ if (grinder_next_pipe(port, subport, pos)) {
+ grinder_prefetch_pipe(subport, pos);
+ subport->busy_grinders++;
grinder->state = e_GRINDER_PREFETCH_TC_QUEUE_ARRAYS;
return 0;
{
struct rte_sched_pipe *pipe = grinder->pipe;
- grinder->pipe_params = port->pipe_profiles + pipe->profile;
- grinder_prefetch_tc_queue_arrays(port, pos);
- grinder_credits_update(port, pos);
+ grinder->pipe_params = subport->pipe_profiles + pipe->profile;
+ grinder_prefetch_tc_queue_arrays(subport, pos);
+ grinder_credits_update(port, subport, pos);
grinder->state = e_GRINDER_PREFETCH_MBUF;
return 0;
case e_GRINDER_PREFETCH_MBUF:
{
- grinder_prefetch_mbuf(port, pos);
+ grinder_prefetch_mbuf(subport, pos);
grinder->state = e_GRINDER_READ_MBUF;
return 0;
case e_GRINDER_READ_MBUF:
{
- uint32_t result = 0;
+ uint32_t wrr_active, result = 0;
+
+ result = grinder_schedule(port, subport, pos);
- result = grinder_schedule(port, pos);
+ wrr_active = (grinder->tc_index == RTE_SCHED_TRAFFIC_CLASS_BE);
/* Look for next packet within the same TC */
if (result && grinder->qmask) {
- grinder_wrr(port, pos);
- grinder_prefetch_mbuf(port, pos);
+ if (wrr_active)
+ grinder_wrr(subport, pos);
+
+ grinder_prefetch_mbuf(subport, pos);
return 1;
}
- grinder_wrr_store(port, pos);
+
+ if (wrr_active)
+ grinder_wrr_store(subport, pos);
/* Look for another active TC within same pipe */
- if (grinder_next_tc(port, pos)) {
- grinder_prefetch_tc_queue_arrays(port, pos);
+ if (grinder_next_tc(port, subport, pos)) {
+ grinder_prefetch_tc_queue_arrays(subport, pos);
grinder->state = e_GRINDER_PREFETCH_MBUF;
return result;
}
if (grinder->productive == 0 &&
- port->pipe_loop == RTE_SCHED_PIPE_INVALID)
- port->pipe_loop = grinder->pindex;
+ subport->pipe_loop == RTE_SCHED_PIPE_INVALID)
+ subport->pipe_loop = grinder->pindex;
- grinder_evict(port, pos);
+ grinder_evict(subport, pos);
/* Look for another active pipe */
- if (grinder_next_pipe(port, pos)) {
- grinder_prefetch_pipe(port, pos);
+ if (grinder_next_pipe(port, subport, pos)) {
+ grinder_prefetch_pipe(subport, pos);
grinder->state = e_GRINDER_PREFETCH_TC_QUEUE_ARRAYS;
return result;
}
/* No active pipe found */
- port->busy_grinders--;
+ subport->busy_grinders--;
grinder->state = e_GRINDER_PREFETCH_PIPE;
return result;
uint64_t cycles = rte_get_tsc_cycles();
uint64_t cycles_diff = cycles - port->time_cpu_cycles;
uint64_t bytes_diff;
+ uint32_t i;
/* Compute elapsed time in bytes */
bytes_diff = rte_reciprocal_divide(cycles_diff << RTE_SCHED_TIME_SHIFT,
port->time = port->time_cpu_bytes;
/* Reset pipe loop detection */
- port->pipe_loop = RTE_SCHED_PIPE_INVALID;
+ for (i = 0; i < port->n_subports_per_port; i++)
+ port->subports[i]->pipe_loop = RTE_SCHED_PIPE_INVALID;
}
static inline int
-rte_sched_port_exceptions(struct rte_sched_port *port, int second_pass)
+rte_sched_port_exceptions(struct rte_sched_subport *subport, int second_pass)
{
int exceptions;
/* Check if any exception flag is set */
- exceptions = (second_pass && port->busy_grinders == 0) ||
- (port->pipe_exhaustion == 1);
+ exceptions = (second_pass && subport->busy_grinders == 0) ||
+ (subport->pipe_exhaustion == 1);
/* Clear exception flags */
- port->pipe_exhaustion = 0;
+ subport->pipe_exhaustion = 0;
return exceptions;
}
int
rte_sched_port_dequeue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint32_t n_pkts)
{
- uint32_t i, count;
+ struct rte_sched_subport *subport;
+ uint32_t subport_id = port->subport_id;
+ uint32_t i, n_subports = 0, count;
port->pkts_out = pkts;
port->n_pkts_out = 0;
/* Take each queue in the grinder one step further */
for (i = 0, count = 0; ; i++) {
- count += grinder_handle(port, i & (RTE_SCHED_PORT_N_GRINDERS - 1));
- if ((count == n_pkts) ||
- rte_sched_port_exceptions(port, i >= RTE_SCHED_PORT_N_GRINDERS)) {
+ subport = port->subports[subport_id];
+
+ count += grinder_handle(port, subport,
+ i & (RTE_SCHED_PORT_N_GRINDERS - 1));
+
+ if (count == n_pkts) {
+ subport_id++;
+
+ if (subport_id == port->n_subports_per_port)
+ subport_id = 0;
+
+ port->subport_id = subport_id;
+ break;
+ }
+
+ if (rte_sched_port_exceptions(subport, i >= RTE_SCHED_PORT_N_GRINDERS)) {
+ i = 0;
+ subport_id++;
+ n_subports++;
+ }
+
+ if (subport_id == port->n_subports_per_port)
+ subport_id = 0;
+
+ if (n_subports == port->n_subports_per_port) {
+ port->subport_id = subport_id;
break;
}
}