X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=lib%2Flibrte_sched%2Frte_sched.c;h=bc06bc3f4e3278b3d842bdb906b2a6b72335e119;hb=cfe3aeb170b2f6277e6f96173599da51eab0654f;hp=968c2b31215b242000697e5f7bfd1b09b51b687c;hpb=3031749c2df04a63cdcef186dcce3781e61436e8;p=dpdk.git

diff --git a/lib/librte_sched/rte_sched.c b/lib/librte_sched/rte_sched.c
index 968c2b3121..bc06bc3f4e 100644
--- a/lib/librte_sched/rte_sched.c
+++ b/lib/librte_sched/rte_sched.c
@@ -1,34 +1,5 @@
-/*-
- *   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>
@@ -42,9 +13,10 @@
 #include <rte_prefetch.h>
 #include <rte_branch_prediction.h>
 #include <rte_mbuf.h>
+#include <rte_bitmap.h>
+#include <rte_reciprocal.h>
 
 #include "rte_sched.h"
-#include "rte_bitmap.h"
 #include "rte_sched_common.h"
 #include "rte_approx.h"
 
@@ -52,56 +24,28 @@
 #pragma warning(disable:2259) /* conversion may lose significant bits */
 #endif
 
-#ifndef RTE_SCHED_DEBUG
-#define RTE_SCHED_DEBUG                       0
-#endif
+#ifdef RTE_SCHED_VECTOR
+#include <rte_vect.h>
 
-#ifndef RTE_SCHED_OPTIMIZATIONS
-#define RTE_SCHED_OPTIMIZATIONS		          0
+#ifdef RTE_ARCH_X86
+#define SCHED_VECTOR_SSE4
+#elif defined(RTE_MACHINE_CPUFLAG_NEON)
+#define SCHED_VECTOR_NEON
 #endif
 
-#if RTE_SCHED_OPTIMIZATIONS
-#include <immintrin.h>
 #endif
 
-#define RTE_SCHED_ENQUEUE                     1
-
-#define RTE_SCHED_TS                          1
-
-#if RTE_SCHED_TS == 0 /* Infinite credits. Traffic shaping disabled. */
-#define RTE_SCHED_TS_CREDITS_UPDATE           0
-#define RTE_SCHED_TS_CREDITS_CHECK            0
-#else                 /* Real Credits. Full traffic shaping implemented. */
-#define RTE_SCHED_TS_CREDITS_UPDATE           1
-#define RTE_SCHED_TS_CREDITS_CHECK            1
-#endif
-
-#ifndef RTE_SCHED_TB_RATE_CONFIG_ERR
 #define RTE_SCHED_TB_RATE_CONFIG_ERR          (1e-7)
-#endif
-
-#define RTE_SCHED_WRR                         1
-
-#ifndef RTE_SCHED_WRR_SHIFT
 #define RTE_SCHED_WRR_SHIFT                   3
-#endif
-
-#ifndef RTE_SCHED_PORT_N_GRINDERS
-#define RTE_SCHED_PORT_N_GRINDERS             8
-#endif
-#if (RTE_SCHED_PORT_N_GRINDERS == 0) || (RTE_SCHED_PORT_N_GRINDERS & (RTE_SCHED_PORT_N_GRINDERS - 1))
-#error Number of grinders must be non-zero and a power of 2
-#endif
-#if (RTE_SCHED_OPTIMIZATIONS && (RTE_SCHED_PORT_N_GRINDERS != 8))
-#error Number of grinders must be 8 when RTE_SCHED_OPTIMIZATIONS is set
-#endif
-
 #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
 
+/* Scaling for cycles_per_byte calculation
+ * Chosen so that minimum rate is 480 bit/sec
+ */
+#define RTE_SCHED_TIME_SHIFT		      8
+
 struct rte_sched_subport {
 	/* Token bucket (TB) */
 	uint64_t tb_time; /* time of last update */
@@ -225,6 +169,7 @@ 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;
 	uint32_t rate;
 	uint32_t mtu;
 	uint32_t frame_overhead;
@@ -232,14 +177,14 @@ struct rte_sched_port {
 	uint32_t n_pipe_profiles;
 	uint32_t pipe_tc3_rate_max;
 #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 */
-	double cycles_per_byte;       /* CPU cycles per byte */
+	struct rte_reciprocal inv_cycles_per_byte; /* CPU cycles per byte */
 
 	/* Scheduling loop detection */
 	uint32_t pipe_loop;
@@ -297,94 +242,123 @@ rte_sched_port_queues_per_port(struct rte_sched_port *port)
 	return RTE_SCHED_QUEUES_PER_PIPE * port->n_pipes_per_subport * port->n_subports_per_port;
 }
 
+static inline struct rte_mbuf **
+rte_sched_port_qbase(struct rte_sched_port *port, uint32_t qindex)
+{
+	uint32_t pindex = qindex >> 4;
+	uint32_t qpos = qindex & 0xF;
+
+	return (port->queue_array + pindex *
+		port->qsize_sum + port->qsize_add[qpos]);
+}
+
+static inline uint16_t
+rte_sched_port_qsize(struct rte_sched_port *port, uint32_t qindex)
+{
+	uint32_t tc = (qindex >> 2) & 0x3;
+
+	return port->qsize[tc];
+}
+
+static int
+pipe_profile_check(struct rte_sched_pipe_params *params,
+	uint32_t rate)
+{
+	uint32_t i;
+
+	/* Pipe parameters */
+	if (params == NULL)
+		return -10;
+
+	/* TB rate: non-zero, not greater than port rate */
+	if (params->tb_rate == 0 ||
+		params->tb_rate > rate)
+		return -11;
+
+	/* TB size: non-zero */
+	if (params->tb_size == 0)
+		return -12;
+
+	/* TC rate: non-zero, less than pipe rate */
+	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 -13;
+	}
+
+	/* TC period: non-zero */
+	if (params->tc_period == 0)
+		return -14;
+
+#ifdef RTE_SCHED_SUBPORT_TC_OV
+	/* TC3 oversubscription weight: non-zero */
+	if (params->tc_ov_weight == 0)
+		return -15;
+#endif
+
+	/* Queue WRR weights: non-zero */
+	for (i = 0; i < RTE_SCHED_QUEUES_PER_PIPE; i++) {
+		if (params->wrr_weights[i] == 0)
+			return -16;
+	}
+
+	return 0;
+}
+
 static int
 rte_sched_port_check_params(struct rte_sched_port_params *params)
 {
-	uint32_t i, j;
+	uint32_t i;
 
-	if (params == NULL) {
+	if (params == NULL)
 		return -1;
-	}
 
 	/* socket */
-	if ((params->socket < 0) || (params->socket >= RTE_MAX_NUMA_NODES)) {
+	if (params->socket < 0)
 		return -3;
-	}
 
 	/* rate */
-	if (params->rate == 0) {
+	if (params->rate == 0)
 		return -4;
-	}
 
 	/* mtu */
-	if (params->mtu == 0) {
+	if (params->mtu == 0)
 		return -5;
-	}
 
-	/* n_subports_per_port: non-zero, power of 2 */
-	if ((params->n_subports_per_port == 0) || (!rte_is_power_of_2(params->n_subports_per_port))) {
+	/* 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;
-	}
 
 	/* 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))) {
+	if (params->n_pipes_per_subport == 0 ||
+	    !rte_is_power_of_2(params->n_pipes_per_subport))
 		return -7;
-	}
 
-	/* 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 ++) {
+	/* 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];
 
-		if ((qsize == 0) || (!rte_is_power_of_2(qsize))) {
+		if (qsize == 0 || !rte_is_power_of_2(qsize))
 			return -8;
-		}
 	}
 
 	/* 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)) {
+	if (params->pipe_profiles == NULL ||
+	    params->n_pipe_profiles == 0 ||
+	    params->n_pipe_profiles > RTE_SCHED_PIPE_PROFILES_PER_PORT)
 		return -9;
-	}
 
-	for (i = 0; i < params->n_pipe_profiles; i ++) {
+	for (i = 0; i < params->n_pipe_profiles; i++) {
 		struct rte_sched_pipe_params *p = params->pipe_profiles + i;
+		int status;
 
-		/* TB rate: non-zero, not greater than port rate */
-		if ((p->tb_rate == 0) || (p->tb_rate > params->rate)) {
-			return -10;
-		}
-
-		/* TB size: non-zero */
-		if (p->tb_size == 0) {
-			return -11;
-		}
-
-		/* 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 period: non-zero */
-		if (p->tc_period == 0) {
-			return -13;
-		}
-
-#ifdef RTE_SCHED_SUBPORT_TC_OV
-		/* TC3 oversubscription weight: non-zero */
-		if (p->tc_ov_weight == 0) {
-			return -14;
-		}
-#endif
-
-		/* Queue WRR weights: non-zero */
-		for (j = 0; j < RTE_SCHED_QUEUES_PER_PIPE; j ++) {
-			if (p->wrr_weights[j] == 0) {
-				return -15;
-			}
-		}
+		status = pipe_profile_check(p, params->rate);
+		if (status != 0)
+			return status;
 	}
 
 	return 0;
@@ -401,41 +375,51 @@ rte_sched_port_get_array_base(struct rte_sched_port_params *params, enum rte_sch
 	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_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_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);
 	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 *);
+	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 *);
 	}
 	size_queue_array = n_pipes_per_port * size_per_pipe_queue_array;
 
 	base = 0;
 
-	if (array == e_RTE_SCHED_PORT_ARRAY_SUBPORT) return base;
-	base += CACHE_LINE_ROUNDUP(size_subport);
+	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) return base;
-	base += CACHE_LINE_ROUNDUP(size_pipe);
+	if (array == e_RTE_SCHED_PORT_ARRAY_PIPE)
+		return base;
+	base += RTE_CACHE_LINE_ROUNDUP(size_pipe);
 
-	if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE) return base;
-	base += CACHE_LINE_ROUNDUP(size_queue);
+	if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE)
+		return base;
+	base += RTE_CACHE_LINE_ROUNDUP(size_queue);
 
-	if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE_EXTRA) return base;
-	base += CACHE_LINE_ROUNDUP(size_queue_extra);
+	if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE_EXTRA)
+		return base;
+	base += RTE_CACHE_LINE_ROUNDUP(size_queue_extra);
 
-	if (array == e_RTE_SCHED_PORT_ARRAY_PIPE_PROFILES) return base;
-	base += CACHE_LINE_ROUNDUP(size_pipe_profiles);
+	if (array == e_RTE_SCHED_PORT_ARRAY_PIPE_PROFILES)
+		return base;
+	base += RTE_CACHE_LINE_ROUNDUP(size_pipe_profiles);
 
-	if (array == e_RTE_SCHED_PORT_ARRAY_BMP_ARRAY) return base;
-	base += CACHE_LINE_ROUNDUP(size_bmp_array);
+	if (array == e_RTE_SCHED_PORT_ARRAY_BMP_ARRAY)
+		return base;
+	base += RTE_CACHE_LINE_ROUNDUP(size_bmp_array);
 
-	if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE_ARRAY) return base;
-	base += CACHE_LINE_ROUNDUP(size_queue_array);
+	if (array == e_RTE_SCHED_PORT_ARRAY_QUEUE_ARRAY)
+		return base;
+	base += RTE_CACHE_LINE_ROUNDUP(size_queue_array);
 
 	return base;
 }
@@ -448,7 +432,8 @@ rte_sched_port_get_memory_footprint(struct rte_sched_port_params *params)
 
 	status = rte_sched_port_check_params(params);
 	if (status != 0) {
-		RTE_LOG(INFO, SCHED, "Port scheduler params check failed (%d)\n", status);
+		RTE_LOG(NOTICE, SCHED,
+			"Port scheduler params check failed (%d)\n", status);
 
 		return 0;
 	}
@@ -456,7 +441,7 @@ rte_sched_port_get_memory_footprint(struct rte_sched_port_params *params)
 	size0 = sizeof(struct rte_sched_port);
 	size1 = rte_sched_port_get_array_base(params, e_RTE_SCHED_PORT_ARRAY_TOTAL);
 
-	return (size0 + size1);
+	return size0 + size1;
 }
 
 static void
@@ -494,11 +479,11 @@ rte_sched_port_log_pipe_profile(struct rte_sched_port *port, uint32_t i)
 {
 	struct rte_sched_pipe_profile *p = port->pipe_profiles + i;
 
-	RTE_LOG(INFO, SCHED, "Low level config for pipe profile %u:\n"
-		"\tToken bucket: period = %u, credits per period = %u, size = %u\n"
-		"\tTraffic classes: period = %u, credits per period = [%u, %u, %u, %u]\n"
-		"\tTraffic class 3 oversubscription: weight = %hhu\n"
-		"\tWRR cost: [%hhu, %hhu, %hhu, %hhu], [%hhu, %hhu, %hhu, %hhu], [%hhu, %hhu, %hhu, %hhu], [%hhu, %hhu, %hhu, %hhu]\n",
+	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",
 		i,
 
 		/* Token bucket */
@@ -527,80 +512,97 @@ static inline uint64_t
 rte_sched_time_ms_to_bytes(uint32_t time_ms, uint32_t rate)
 {
 	uint64_t time = time_ms;
+
 	time = (time * rate) / 1000;
 
 	return time;
 }
 
 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_pipe_params *src,
+	struct rte_sched_pipe_profile *dst,
+	uint32_t rate)
 {
-	uint32_t i, j;
+	uint32_t i;
 
-	for (i = 0; i < port->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 == 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;
 
-		/* 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);
+	}
 
-			rte_approx(tb_rate, d, &dst->tb_credits_per_period, &dst->tb_period);
-		}
-		dst->tb_size = src->tb_size;
+	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++)
+		dst->tc_credits_per_period[i]
+			= rte_sched_time_ms_to_bytes(src->tc_period,
+				src->tc_rate[i]);
 
-		/* Traffic Classes */
-		dst->tc_period = (uint32_t) rte_sched_time_ms_to_bytes(src->tc_period, params->rate);
-		for (j = 0; j < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; j ++) {
-			dst->tc_credits_per_period[j] = (uint32_t) rte_sched_time_ms_to_bytes(src->tc_period, src->tc_rate[j]);
-		}
 #ifdef RTE_SCHED_SUBPORT_TC_OV
-		dst->tc_ov_weight = src->tc_ov_weight;
+	dst->tc_ov_weight = src->tc_ov_weight;
 #endif
 
-		/* 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];
-		}
+	/* WRR */
+	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
+		uint32_t wrr_cost[RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS];
+		uint32_t lcd, lcd1, lcd2;
+		uint32_t qindex;
+
+		qindex = i * 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];
+	}
+}
+
+static void
+rte_sched_port_config_pipe_profile_table(struct rte_sched_port *port,
+	struct rte_sched_port_params *params)
+{
+	uint32_t i;
+
+	for (i = 0; i < port->n_pipe_profiles; i++) {
+		struct rte_sched_pipe_params *src = params->pipe_profiles + i;
+		struct rte_sched_pipe_profile *dst = port->pipe_profiles + i;
+
+		rte_sched_pipe_profile_convert(src, dst, params->rate);
 		rte_sched_port_log_pipe_profile(port, i);
 	}
 
 	port->pipe_tc3_rate_max = 0;
-	for (i = 0; i < port->n_pipe_profiles; i ++) {
+	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];
 
-		if (port->pipe_tc3_rate_max < pipe_tc3_rate) {
+		if (port->pipe_tc3_rate_max < pipe_tc3_rate)
 			port->pipe_tc3_rate_max = pipe_tc3_rate;
-		}
 	}
 }
 
@@ -608,23 +610,28 @@ 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;
+	uint32_t mem_size, bmp_mem_size, n_queues_per_port, i, cycles_per_byte;
 
 	/* Check user parameters. Determine the amount of memory to allocate */
 	mem_size = rte_sched_port_get_memory_footprint(params);
-	if (mem_size == 0) {
+	if (mem_size == 0)
 		return NULL;
-	}
 
 	/* Allocate memory to store the data structures */
-	port = rte_zmalloc("qos_params", mem_size, CACHE_LINE_SIZE);
-	if (port == NULL) {
+	port = rte_zmalloc_socket("qos_params", mem_size, RTE_CACHE_LINE_SIZE,
+		params->socket);
+	if (port == NULL)
 		return NULL;
-	}
+
+	/* 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 */
 	port->n_subports_per_port = params->n_subports_per_port;
 	port->n_pipes_per_subport = params->n_pipes_per_subport;
+	port->n_pipes_per_subport_log2 =
+			__builtin_ctz(params->n_pipes_per_subport);
 	port->rate = params->rate;
 	port->mtu = params->mtu + params->frame_overhead;
 	port->frame_overhead = params->frame_overhead;
@@ -635,12 +642,19 @@ rte_sched_port_config(struct rte_sched_port_params *params)
 	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++) {
 		uint32_t j;
 
-		for (j = 0; j < e_RTE_METER_COLORS; 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(&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) {
+				rte_free(port);
 				return NULL;
 			}
 		}
@@ -651,7 +665,10 @@ rte_sched_port_config(struct rte_sched_port_params *params)
 	port->time_cpu_cycles = rte_get_tsc_cycles();
 	port->time_cpu_bytes = 0;
 	port->time = 0;
-	port->cycles_per_byte = ((double) rte_get_tsc_hz()) / ((double) params->rate);
+
+	cycles_per_byte = (rte_get_tsc_hz() << RTE_SCHED_TIME_SHIFT)
+		/ params->rate;
+	port->inv_cycles_per_byte = rte_reciprocal_value(cycles_per_byte);
 
 	/* Scheduling loop detection */
 	port->pipe_loop = RTE_SCHED_PIPE_INVALID;
@@ -666,13 +683,26 @@ rte_sched_port_config(struct rte_sched_port_params *params)
 	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));
+	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));
 
 	/* Pipe profile table */
 	rte_sched_port_config_pipe_profile_table(port, params);
@@ -680,14 +710,17 @@ rte_sched_port_config(struct rte_sched_port_params *params)
 	/* 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);
+	port->bmp = rte_bitmap_init(n_queues_per_port, port->bmp_array,
+				    bmp_mem_size);
 	if (port->bmp == NULL) {
-		RTE_LOG(INFO, SCHED, "Bitmap init error\n");
+		RTE_LOG(ERR, SCHED, "Bitmap init error\n");
+		rte_free(port);
 		return NULL;
 	}
-	for (i = 0; i < RTE_SCHED_PORT_N_GRINDERS; i ++) {
+
+	for (i = 0; i < RTE_SCHED_PORT_N_GRINDERS; i++)
 		port->grinder_base_bmp_pos[i] = RTE_SCHED_PIPE_INVALID;
-	}
+
 
 	return port;
 }
@@ -695,9 +728,25 @@ rte_sched_port_config(struct rte_sched_port_params *params)
 void
 rte_sched_port_free(struct rte_sched_port *port)
 {
+	uint32_t qindex;
+	uint32_t n_queues_per_port;
+
 	/* Check user parameters */
-	if (port == NULL){
+	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]);
 	}
 
 	rte_bitmap_free(port->bmp);
@@ -709,10 +758,10 @@ rte_sched_port_log_subport_config(struct rte_sched_port *port, uint32_t i)
 {
 	struct rte_sched_subport *s = port->subport + i;
 
-	RTE_LOG(INFO, SCHED, "Low level config for subport %u:\n"
-		"\tToken bucket: period = %u, credits per period = %u, size = %u\n"
-		"\tTraffic classes: period = %u, credits per period = [%u, %u, %u, %u]\n"
-		"\tTraffic class 3 oversubscription: wm min = %u, wm max = %u\n",
+	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",
 		i,
 
 		/* Token bucket */
@@ -741,29 +790,25 @@ rte_sched_subport_config(struct rte_sched_port *port,
 	uint32_t i;
 
 	/* Check user parameters */
-	if ((port == NULL) ||
-	    (subport_id >= port->n_subports_per_port) ||
-		(params == NULL)) {
+	if (port == NULL ||
+	    subport_id >= port->n_subports_per_port ||
+	    params == NULL)
 		return -1;
-	}
 
-	if ((params->tb_rate == 0) || (params->tb_rate > port->rate)) {
+	if (params->tb_rate == 0 || params->tb_rate > port->rate)
 		return -2;
-	}
 
-	if (params->tb_size == 0) {
+	if (params->tb_size == 0)
 		return -3;
-	}
 
-	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i ++) {
-		if ((params->tc_rate[i] == 0) || (params->tc_rate[i] > params->tb_rate)) {
+	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;
-		}
 	}
 
-	if (params->tc_period == 0) {
+	if (params->tc_period == 0)
 		return -5;
-	}
 
 	s = port->subport + subport_id;
 
@@ -777,24 +822,27 @@ rte_sched_subport_config(struct rte_sched_port *port,
 
 		rte_approx(tb_rate, d, &s->tb_credits_per_period, &s->tb_period);
 	}
+
 	s->tb_size = params->tb_size;
 	s->tb_time = port->time;
 	s->tb_credits = s->tb_size / 2;
 
 	/* Traffic Classes (TCs) */
-	s->tc_period = (uint32_t) 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] = (uint32_t) rte_sched_time_ms_to_bytes(params->tc_period, params->tc_rate[i]);
+	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]);
 	}
 	s->tc_time = port->time + s->tc_period;
-	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i ++) {
+	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i++)
 		s->tc_credits[i] = s->tc_credits_per_period[i];
-	}
 
 #ifdef RTE_SCHED_SUBPORT_TC_OV
 	/* TC oversubscription */
 	s->tc_ov_wm_min = port->mtu;
-	s->tc_ov_wm_max = (uint32_t) rte_sched_time_ms_to_bytes(params->tc_period, port->pipe_tc3_rate_max);
+	s->tc_ov_wm_max = rte_sched_time_ms_to_bytes(params->tc_period,
+						     port->pipe_tc3_rate_max);
 	s->tc_ov_wm = s->tc_ov_wm_max;
 	s->tc_ov_period_id = 0;
 	s->tc_ov = 0;
@@ -821,18 +869,18 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 	/* 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))) {
+
+	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;
-	}
+
 
 	/* Check that subport configuration is valid */
 	s = port->subport + subport_id;
-	if (s->tb_period == 0) {
+	if (s->tb_period == 0)
 		return -2;
-	}
 
 	p = port->pipe + (subport_id * port->n_pipes_per_subport + pipe_id);
 
@@ -841,8 +889,10 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 		params = port->pipe_profiles + p->profile;
 
 #ifdef RTE_SCHED_SUBPORT_TC_OV
-		double subport_tc3_rate = ((double) s->tc_credits_per_period[3]) / ((double) s->tc_period);
-		double pipe_tc3_rate = ((double) params->tc_credits_per_period[3]) / ((double) params->tc_period);
+		double subport_tc3_rate = (double) s->tc_credits_per_period[3]
+			/ (double) s->tc_period;
+		double pipe_tc3_rate = (double) params->tc_credits_per_period[3]
+			/ (double) params->tc_period;
 		uint32_t tc3_ov = s->tc_ov;
 
 		/* Unplug pipe from its subport */
@@ -851,7 +901,8 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 		s->tc_ov = s->tc_ov_rate > subport_tc3_rate;
 
 		if (s->tc_ov != tc3_ov) {
-			RTE_LOG(INFO, SCHED, "Subport %u TC3 oversubscription is OFF (%.4lf >= %.4lf)\n",
+			RTE_LOG(DEBUG, SCHED,
+				"Subport %u TC3 oversubscription is OFF (%.4lf >= %.4lf)\n",
 				subport_id, subport_tc3_rate, s->tc_ov_rate);
 		}
 #endif
@@ -860,9 +911,8 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 		memset(p, 0, sizeof(struct rte_sched_pipe));
 	}
 
-	if (deactivate) {
+	if (deactivate)
 		return 0;
-	}
 
 	/* Apply the new pipe configuration */
 	p->profile = profile;
@@ -874,15 +924,16 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 
 	/* Traffic Classes (TCs) */
 	p->tc_time = port->time + params->tc_period;
-	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; i ++) {
+	for (i = 0; i < RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE; 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]) / ((double) s->tc_period);
-		double pipe_tc3_rate = ((double) params->tc_credits_per_period[3]) / ((double) params->tc_period);
+		double subport_tc3_rate = (double) s->tc_credits_per_period[3]
+			/ (double) s->tc_period;
+		double pipe_tc3_rate = (double) params->tc_credits_per_period[3]
+			/ (double) params->tc_period;
 		uint32_t tc3_ov = s->tc_ov;
 
 		s->tc_ov_n += params->tc_ov_weight;
@@ -890,7 +941,8 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 		s->tc_ov = s->tc_ov_rate > subport_tc3_rate;
 
 		if (s->tc_ov != tc3_ov) {
-			RTE_LOG(INFO, SCHED, "Subport %u TC3 oversubscription is ON (%.4lf < %.4lf)\n",
+			RTE_LOG(DEBUG, SCHED,
+				"Subport %u TC3 oversubscription is ON (%.4lf < %.4lf)\n",
 				subport_id, subport_tc3_rate, s->tc_ov_rate);
 		}
 		p->tc_ov_period_id = s->tc_ov_period_id;
@@ -901,28 +953,116 @@ rte_sched_pipe_config(struct rte_sched_port *port,
 	return 0;
 }
 
+int
+rte_sched_port_pipe_profile_add(struct rte_sched_port *port,
+	struct rte_sched_pipe_params *params,
+	uint32_t *pipe_profile_id)
+{
+	struct rte_sched_pipe_profile *pp;
+	uint32_t i;
+	int status;
+
+	/* Port */
+	if (port == NULL)
+		return -1;
+
+	/* Pipe profiles not exceeds the max limit */
+	if (port->n_pipe_profiles >= RTE_SCHED_PIPE_PROFILES_PER_PORT)
+		return -2;
+
+	/* Pipe params */
+	status = pipe_profile_check(params, port->rate);
+	if (status != 0)
+		return status;
+
+	pp = &port->pipe_profiles[port->n_pipe_profiles];
+	rte_sched_pipe_profile_convert(params, pp, port->rate);
+
+	/* Pipe profile not exists */
+	for (i = 0; i < port->n_pipe_profiles; i++)
+		if (memcmp(port->pipe_profiles + i, pp, sizeof(*pp)) == 0)
+			return -3;
+
+	/* Pipe profile commit */
+	*pipe_profile_id = port->n_pipe_profiles;
+	port->n_pipe_profiles++;
+
+	if (port->pipe_tc3_rate_max < params->tc_rate[3])
+		port->pipe_tc3_rate_max = params->tc_rate[3];
+
+	rte_sched_port_log_pipe_profile(port, *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)
+{
+	return ((subport & (port->n_subports_per_port - 1)) <<
+			(port->n_pipes_per_subport_log2 + 4)) |
+			((pipe & (port->n_pipes_per_subport - 1)) << 4) |
+			((traffic_class &
+			    (RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE - 1)) << 2) |
+			(queue & (RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS - 1));
+}
+
+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);
+	rte_mbuf_sched_set(pkt, queue_id, traffic_class, (uint8_t)color);
+}
+
+void
+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)
+{
+	uint32_t queue_id = rte_mbuf_sched_queue_get(pkt);
+
+	*subport = queue_id >> (port->n_pipes_per_subport_log2 + 4);
+	*pipe = (queue_id >> 4) & (port->n_pipes_per_subport - 1);
+	*traffic_class = (queue_id >> 2) &
+				(RTE_SCHED_TRAFFIC_CLASSES_PER_PIPE - 1);
+	*queue = queue_id & (RTE_SCHED_QUEUES_PER_TRAFFIC_CLASS - 1);
+}
+
+enum rte_color
+rte_sched_port_pkt_read_color(const struct rte_mbuf *pkt)
+{
+	return (enum rte_color)rte_mbuf_sched_color_get(pkt);
+}
+
 int
 rte_sched_subport_read_stats(struct rte_sched_port *port,
-	uint32_t subport_id,
-	struct rte_sched_subport_stats *stats,
-	uint32_t *tc_ov)
+			     uint32_t subport_id,
+			     struct rte_sched_subport_stats *stats,
+			     uint32_t *tc_ov)
 {
 	struct rte_sched_subport *s;
 
 	/* Check user parameters */
-	if ((port == NULL) ||
-	    (subport_id >= port->n_subports_per_port) ||
-		(stats == NULL) ||
-		(tc_ov == NULL)) {
+	if (port == NULL || subport_id >= port->n_subports_per_port ||
+	    stats == NULL || tc_ov == NULL)
 		return -1;
-	}
+
 	s = port->subport + 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;
@@ -957,53 +1097,14 @@ rte_sched_queue_read_stats(struct rte_sched_port *port,
 	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;
-}
-
-static inline struct rte_mbuf **
-rte_sched_port_qbase(struct rte_sched_port *port, uint32_t qindex)
-{
-	uint32_t pindex = qindex >> 4;
-	uint32_t qpos = qindex & 0xF;
-
-	return (port->queue_array + pindex * port->qsize_sum + port->qsize_add[qpos]);
-}
-
-static inline uint16_t
-rte_sched_port_qsize(struct rte_sched_port *port, uint32_t qindex)
-{
-	uint32_t tc = (qindex >> 2) & 0x3;
-
-	return port->qsize[tc];
-}
-
-#if RTE_SCHED_DEBUG
+#ifdef RTE_SCHED_DEBUG
 
 static inline int
 rte_sched_port_queue_is_empty(struct rte_sched_port *port, uint32_t qindex)
 {
 	struct rte_sched_queue *queue = port->queue + qindex;
 
-	return (queue->qr == queue->qw);
-}
-
-static inline int
-rte_sched_port_queue_is_full(struct rte_sched_port *port, uint32_t qindex)
-{
-	struct rte_sched_queue *queue = port->queue + qindex;
-	uint16_t qsize = rte_sched_port_qsize(port, qindex);
-	uint16_t qlen = queue->qw - queue->qr;
-
-	return (qlen >= qsize);
+	return queue->qr == queue->qw;
 }
 
 #endif /* RTE_SCHED_DEBUG */
@@ -1015,41 +1116,65 @@ rte_sched_port_update_subport_stats(struct rte_sched_port *port, uint32_t qindex
 {
 	struct rte_sched_subport *s = port->subport + (qindex / rte_sched_port_queues_per_subport(port));
 	uint32_t tc_index = (qindex >> 2) & 0x3;
-	uint32_t pkt_len = pkt->pkt.pkt_len;
+	uint32_t pkt_len = pkt->pkt_len;
 
 	s->stats.n_pkts_tc[tc_index] += 1;
 	s->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)
+rte_sched_port_update_subport_stats_on_drop(struct rte_sched_port *port,
+						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)
+#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 pkt_len = pkt->pkt.pkt_len;
+	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;
+#ifdef RTE_SCHED_RED
+	s->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)
 {
 	struct rte_sched_queue_extra *qe = port->queue_extra + qindex;
-	uint32_t pkt_len = pkt->pkt.pkt_len;
+	uint32_t pkt_len = pkt->pkt_len;
 
 	qe->stats.n_pkts += 1;
 	qe->stats.n_bytes += pkt_len;
 }
 
+#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)
+#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_sched_port_update_queue_stats_on_drop(struct rte_sched_port *port,
+						uint32_t qindex,
+						struct rte_mbuf *pkt, __rte_unused uint32_t red)
+#endif
 {
 	struct rte_sched_queue_extra *qe = port->queue_extra + qindex;
-	uint32_t pkt_len = pkt->pkt.pkt_len;
+	uint32_t pkt_len = pkt->pkt_len;
 
 	qe->stats.n_pkts_dropped += 1;
 	qe->stats.n_bytes_dropped += pkt_len;
+#ifdef RTE_SCHED_RED
+	qe->stats.n_pkts_red_dropped += red;
+#endif
 }
 
 #endif /* RTE_SCHED_COLLECT_STATS */
@@ -1061,14 +1186,17 @@ rte_sched_port_red_drop(struct rte_sched_port *port, struct rte_mbuf *pkt, uint3
 {
 	struct rte_sched_queue_extra *qe;
 	struct rte_red_config *red_cfg;
-    struct rte_red *red;
+	struct rte_red *red;
 	uint32_t tc_index;
-	enum rte_meter_color color;
+	enum rte_color color;
 
 	tc_index = (qindex >> 2) & 0x3;
 	color = rte_sched_port_pkt_read_color(pkt);
 	red_cfg = &port->red_config[tc_index][color];
 
+	if ((red_cfg->min_th | red_cfg->max_th) == 0)
+		return 0;
+
 	qe = port->queue_extra + qindex;
 	red = &qe->red;
 
@@ -1078,11 +1206,8 @@ rte_sched_port_red_drop(struct rte_sched_port *port, struct rte_mbuf *pkt, uint3
 static inline void
 rte_sched_port_set_queue_empty_timestamp(struct rte_sched_port *port, uint32_t qindex)
 {
-	struct rte_sched_queue_extra *qe;
-    struct rte_red *red;
-
-	qe = port->queue_extra + qindex;
-	red = &qe->red;
+	struct rte_sched_queue_extra *qe = port->queue_extra + qindex;
+	struct rte_red *red = &qe->red;
 
 	rte_red_mark_queue_empty(red, port->time);
 }
@@ -1095,44 +1220,21 @@ rte_sched_port_set_queue_empty_timestamp(struct rte_sched_port *port, uint32_t q
 
 #endif /* RTE_SCHED_RED */
 
-#if RTE_SCHED_DEBUG
-
-static inline int
-debug_pipe_is_empty(struct rte_sched_port *port, uint32_t pindex)
-{
-	uint32_t qindex, i;
-
-	qindex = pindex << 4;
-
-	for (i = 0; i < 16; i ++){
-		uint32_t queue_empty = rte_sched_port_queue_is_empty(port, qindex + i);
-		uint32_t bmp_bit_clear = (rte_bitmap_get(port->bmp, qindex + i) == 0);
-
-		if (queue_empty != bmp_bit_clear){
-			rte_panic("Queue status mismatch for queue %u of pipe %u\n", i, pindex);
-		}
-
-		if (!queue_empty){
-			return 0;
-		}
-	}
-
-	return 1;
-}
+#ifdef RTE_SCHED_DEBUG
 
 static inline void
-debug_check_queue_slab(struct rte_sched_port *port, uint32_t bmp_pos, uint64_t bmp_slab)
+debug_check_queue_slab(struct rte_sched_port *port, uint32_t bmp_pos,
+		       uint64_t bmp_slab)
 {
 	uint64_t mask;
 	uint32_t i, panic;
 
-	if (bmp_slab == 0){
+	if (bmp_slab == 0)
 		rte_panic("Empty slab at position %u\n", bmp_pos);
-	}
 
 	panic = 0;
-	for (i = 0, mask = 1; i < 64; i ++, mask <<= 1) {
-		if (mask & bmp_slab){
+	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)) {
 				printf("Queue %u (slab offset %u) is empty\n", bmp_pos + i, i);
 				panic = 1;
@@ -1140,26 +1242,23 @@ debug_check_queue_slab(struct rte_sched_port *port, uint32_t bmp_pos, uint64_t b
 		}
 	}
 
-	if (panic){
+	if (panic)
 		rte_panic("Empty queues in slab 0x%" PRIx64 "starting at position %u\n",
 			bmp_slab, bmp_pos);
-	}
 }
 
 #endif /* RTE_SCHED_DEBUG */
 
 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_port *port,
+				       struct rte_mbuf *pkt)
 {
 	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;
-
-	rte_sched_port_pkt_read_tree_path(pkt, &subport, &pipe, &traffic_class, &queue);
+	uint32_t qindex = rte_mbuf_sched_queue_get(pkt);
 
-	qindex = rte_sched_port_qindex(port, subport, pipe, traffic_class, queue);
 	q = port->queue + qindex;
 	rte_prefetch0(q);
 #ifdef RTE_SCHED_COLLECT_STATS
@@ -1171,7 +1270,8 @@ rte_sched_port_enqueue_qptrs_prefetch0(struct rte_sched_port *port, struct rte_m
 }
 
 static inline void
-rte_sched_port_enqueue_qwa_prefetch0(struct rte_sched_port *port, uint32_t qindex, struct rte_mbuf **qbase)
+rte_sched_port_enqueue_qwa_prefetch0(struct rte_sched_port *port,
+				     uint32_t qindex, struct rte_mbuf **qbase)
 {
 	struct rte_sched_queue *q;
 	struct rte_mbuf **q_qw;
@@ -1186,7 +1286,8 @@ rte_sched_port_enqueue_qwa_prefetch0(struct rte_sched_port *port, uint32_t qinde
 }
 
 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, uint32_t qindex,
+			   struct rte_mbuf **qbase, struct rte_mbuf *pkt)
 {
 	struct rte_sched_queue *q;
 	uint16_t qsize;
@@ -1197,18 +1298,21 @@ rte_sched_port_enqueue_qwa(struct rte_sched_port *port, uint32_t qindex, struct
 	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) || (qlen >= qsize))) {
+	if (unlikely(rte_sched_port_red_drop(port, 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);
-		rte_sched_port_update_queue_stats_on_drop(port, qindex, pkt);
+		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);
 #endif
 		return 0;
 	}
 
 	/* Enqueue packet */
 	qbase[q->qw & (qsize - 1)] = pkt;
-	q->qw ++;
+	q->qw++;
 
 	/* Activate queue in the port bitmap */
 	rte_bitmap_set(port->bmp, qindex);
@@ -1222,40 +1326,12 @@ rte_sched_port_enqueue_qwa(struct rte_sched_port *port, uint32_t qindex, struct
 	return 1;
 }
 
-#if RTE_SCHED_ENQUEUE == 0
-
-int
-rte_sched_port_enqueue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint32_t n_pkts)
-{
-	uint32_t result, i;
-
-	result = 0;
-
-	for (i = 0; i < n_pkts; i ++) {
-		struct rte_mbuf *pkt;
-		struct rte_mbuf **q_base;
-		uint32_t subport, pipe, traffic_class, queue, qindex;
-
-		pkt = pkts[i];
-
-		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_base = rte_sched_port_qbase(port, qindex);
-
-		result += rte_sched_port_enqueue_qwa(port, qindex, q_base, pkt);
-	}
-
-	return result;
-}
-
-#else
 
-/* The enqueue function implements a 4-level pipeline with each stage processing
- * two different packets. The purpose of using a pipeline is to hide the latency
- * of prefetching the data structures. The naming convention is presented in the
- * diagram below:
+/*
+ * The enqueue function implements a 4-level pipeline with each stage
+ * processing two different packets. The purpose of using a pipeline
+ * is to hide the latency of prefetching the data structures. The
+ * naming convention is presented in the diagram below:
  *
  *   p00  _______   p10  _______   p20  _______   p30  _______
  * ----->|       |----->|       |----->|       |----->|       |----->
@@ -1263,43 +1339,49 @@ rte_sched_port_enqueue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint
  * ----->|_______|----->|_______|----->|_______|----->|_______|----->
  *   p01            p11            p21            p31
  *
- ***/
+ */
 int
-rte_sched_port_enqueue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint32_t n_pkts)
+rte_sched_port_enqueue(struct rte_sched_port *port, struct rte_mbuf **pkts,
+		       uint32_t n_pkts)
 {
-	struct rte_mbuf *pkt00, *pkt01, *pkt10, *pkt11, *pkt20, *pkt21, *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_mbuf *pkt00, *pkt01, *pkt10, *pkt11, *pkt20, *pkt21,
+		*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;
 	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 result, i;
 
 	result = 0;
 
-	/* Less then 6 input packets available, which is not enough to feed the pipeline */
+	/*
+	 * Less then 6 input packets available, which is not enough to
+	 * feed the pipeline
+	 */
 	if (unlikely(n_pkts < 6)) {
 		struct rte_mbuf **q_base[5];
 		uint32_t q[5];
 
 		/* Prefetch the mbuf structure of each packet */
-		for (i = 0; i < n_pkts; i ++) {
+		for (i = 0; i < n_pkts; i++)
 			rte_prefetch0(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]);
-		}
+		for (i = 0; i < n_pkts; i++)
+			q[i] = rte_sched_port_enqueue_qptrs_prefetch0(port,
+								      pkts[i]);
 
 		/* Prefetch the write pointer location of each queue */
-		for (i = 0; i < n_pkts; i ++) {
+		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]);
+			rte_sched_port_enqueue_qwa_prefetch0(port, 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]);
-		}
+		for (i = 0; i < n_pkts; i++)
+			result += rte_sched_port_enqueue_qwa(port, q[i],
+							     q_base[i], pkts[i]);
 
 		return result;
 	}
@@ -1369,8 +1451,11 @@ rte_sched_port_enqueue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint
 		result += r30 + r31;
 	}
 
-	/* Drain the pipeline (exactly 6 packets). Handle the last packet in the case
-	of an odd number of input packets. */
+	/*
+	 * Drain the pipeline (exactly 6 packets).
+	 * Handle the last packet in the case
+	 * of an odd number of input packets.
+	 */
 	pkt_last = pkts[n_pkts - 1];
 	rte_prefetch0(pkt_last);
 
@@ -1412,13 +1497,7 @@ rte_sched_port_enqueue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint
 	return result;
 }
 
-#endif /* RTE_SCHED_ENQUEUE */
-
-#if RTE_SCHED_TS_CREDITS_UPDATE == 0
-
-#define grinder_credits_update(port, pos)
-
-#elif !defined(RTE_SCHED_SUBPORT_TC_OV)
+#ifndef RTE_SCHED_SUBPORT_TC_OV
 
 static inline void
 grinder_credits_update(struct rte_sched_port *port, uint32_t pos)
@@ -1471,9 +1550,8 @@ grinder_tc_ov_credits_update(struct rte_sched_port *port, uint32_t pos)
 	uint32_t tc_ov_consumption_max;
 	uint32_t tc_ov_wm = subport->tc_ov_wm;
 
-	if (subport->tc_ov == 0) {
+	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];
@@ -1485,16 +1563,16 @@ grinder_tc_ov_credits_update(struct rte_sched_port *port, uint32_t pos)
 
 	if (tc_ov_consumption[3] > (tc_ov_consumption_max - port->mtu)) {
 		tc_ov_wm  -= tc_ov_wm >> 7;
-		if (tc_ov_wm < subport->tc_ov_wm_min) {
+		if (tc_ov_wm < subport->tc_ov_wm_min)
 			tc_ov_wm = subport->tc_ov_wm_min;
-		}
+
 		return tc_ov_wm;
 	}
 
 	tc_ov_wm += (tc_ov_wm >> 7) + 1;
-	if (tc_ov_wm > subport->tc_ov_wm_max) {
+	if (tc_ov_wm > subport->tc_ov_wm_max)
 		tc_ov_wm = subport->tc_ov_wm_max;
-	}
+
 	return tc_ov_wm;
 }
 
@@ -1529,7 +1607,7 @@ grinder_credits_update(struct rte_sched_port *port, uint32_t pos)
 		subport->tc_credits[3] = subport->tc_credits_per_period[3];
 
 		subport->tc_time = port->time + subport->tc_period;
-		subport->tc_ov_period_id ++;
+		subport->tc_ov_period_id++;
 	}
 
 	/* Pipe TCs */
@@ -1551,7 +1629,6 @@ grinder_credits_update(struct rte_sched_port *port, uint32_t pos)
 
 #endif /* RTE_SCHED_TS_CREDITS_UPDATE, RTE_SCHED_SUBPORT_TC_OV */
 
-#if RTE_SCHED_TS_CREDITS_CHECK
 
 #ifndef RTE_SCHED_SUBPORT_TC_OV
 
@@ -1563,7 +1640,7 @@ grinder_credits_check(struct rte_sched_port *port, uint32_t 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.pkt_len + port->frame_overhead;
+	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;
@@ -1576,9 +1653,8 @@ grinder_credits_check(struct rte_sched_port *port, uint32_t pos)
 		(pkt_len <= pipe_tb_credits) &&
 		(pkt_len <= pipe_tc_credits);
 
-	if (!enough_credits) {
+	if (!enough_credits)
 		return 0;
-	}
 
 	/* Update port credits */
 	subport->tb_credits -= pkt_len;
@@ -1599,7 +1675,7 @@ grinder_credits_check(struct rte_sched_port *port, uint32_t 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.pkt_len + port->frame_overhead;
+	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;
@@ -1616,9 +1692,8 @@ grinder_credits_check(struct rte_sched_port *port, uint32_t pos)
 		(pkt_len <= pipe_tc_credits) &&
 		(pkt_len <= pipe_tc_ov_credits);
 
-	if (!enough_credits) {
+	if (!enough_credits)
 		return 0;
-	}
 
 	/* Update pipe and subport credits */
 	subport->tb_credits -= pkt_len;
@@ -1632,7 +1707,6 @@ grinder_credits_check(struct rte_sched_port *port, uint32_t pos)
 
 #endif /* RTE_SCHED_SUBPORT_TC_OV */
 
-#endif /* RTE_SCHED_TS_CREDITS_CHECK */
 
 static inline int
 grinder_schedule(struct rte_sched_port *port, uint32_t pos)
@@ -1640,20 +1714,17 @@ grinder_schedule(struct rte_sched_port *port, uint32_t pos)
 	struct rte_sched_grinder *grinder = port->grinder + pos;
 	struct rte_sched_queue *queue = grinder->queue[grinder->qpos];
 	struct rte_mbuf *pkt = grinder->pkt;
-	uint32_t pkt_len = pkt->pkt.pkt_len + port->frame_overhead;
+	uint32_t pkt_len = pkt->pkt_len + port->frame_overhead;
 
-#if RTE_SCHED_TS_CREDITS_CHECK
-	if (!grinder_credits_check(port, pos)) {
+	if (!grinder_credits_check(port, pos))
 		return 0;
-	}
-#endif
 
 	/* Advance port time */
 	port->time += pkt_len;
 
 	/* Send packet */
-	port->pkts_out[port->n_pkts_out ++] = pkt;
-	queue->qr ++;
+	port->pkts_out[port->n_pkts_out++] = pkt;
+	queue->qr++;
 	grinder->wrr_tokens[grinder->qpos] += pkt_len * grinder->wrr_cost[grinder->qpos];
 	if (queue->qr == queue->qw) {
 		uint32_t qindex = grinder->qindex[grinder->qpos];
@@ -1671,14 +1742,15 @@ grinder_schedule(struct rte_sched_port *port, uint32_t pos)
 	return 1;
 }
 
-#if RTE_SCHED_OPTIMIZATIONS
+#ifdef SCHED_VECTOR_SSE4
 
 static inline int
 grinder_pipe_exists(struct rte_sched_port *port, uint32_t base_pipe)
 {
-	__m128i index = _mm_set1_epi32 (base_pipe);
+	__m128i index = _mm_set1_epi32(base_pipe);
 	__m128i pipes = _mm_load_si128((__m128i *)port->grinder_base_bmp_pos);
 	__m128i res = _mm_cmpeq_epi32(pipes, index);
+
 	pipes = _mm_load_si128((__m128i *)(port->grinder_base_bmp_pos + 4));
 	pipes = _mm_cmpeq_epi32(pipes, index);
 	res = _mm_or_si128(res, pipes);
@@ -1689,6 +1761,26 @@ grinder_pipe_exists(struct rte_sched_port *port, uint32_t base_pipe)
 	return 1;
 }
 
+#elif defined(SCHED_VECTOR_NEON)
+
+static inline int
+grinder_pipe_exists(struct rte_sched_port *port, uint32_t base_pipe)
+{
+	uint32x4_t index, pipes;
+	uint32_t *pos = (uint32_t *)port->grinder_base_bmp_pos;
+
+	index = vmovq_n_u32(base_pipe);
+	pipes = vld1q_u32(pos);
+	if (!vminvq_u32(veorq_u32(pipes, index)))
+		return 1;
+
+	pipes = vld1q_u32(pos + 4);
+	if (!vminvq_u32(veorq_u32(pipes, index)))
+		return 1;
+
+	return 0;
+}
+
 #else
 
 static inline int
@@ -1696,10 +1788,9 @@ grinder_pipe_exists(struct rte_sched_port *port, 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) {
+	for (i = 0; i < RTE_SCHED_PORT_N_GRINDERS; i++) {
+		if (port->grinder_base_bmp_pos[i] == base_pipe)
 			return 1;
-		}
 	}
 
 	return 0;
@@ -1777,9 +1868,8 @@ grinder_next_tc(struct rte_sched_port *port, uint32_t pos)
 	uint32_t qindex;
 	uint16_t qsize;
 
-	if (grinder->tccache_r == grinder->tccache_w) {
+	if (grinder->tccache_r == grinder->tccache_w)
 		return 0;
-	}
 
 	qindex = grinder->tccache_qindex[grinder->tccache_r];
 	qbase = rte_sched_port_qbase(port, qindex);
@@ -1804,7 +1894,7 @@ grinder_next_tc(struct rte_sched_port *port, uint32_t pos)
 	grinder->qbase[2] = qbase + 2 * qsize;
 	grinder->qbase[3] = qbase + 3 * qsize;
 
-	grinder->tccache_r ++;
+	grinder->tccache_r++;
 	return 1;
 }
 
@@ -1818,25 +1908,24 @@ grinder_next_pipe(struct rte_sched_port *port, uint32_t pos)
 	if (grinder->pcache_r < grinder->pcache_w) {
 		pipe_qmask = grinder->pcache_qmask[grinder->pcache_r];
 		pipe_qindex = grinder->pcache_qindex[grinder->pcache_r];
-		grinder->pcache_r ++;
+		grinder->pcache_r++;
 	} else {
 		uint64_t bmp_slab = 0;
 		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(port->bmp, &bmp_pos, &bmp_slab) <= 0))
 			return 0;
-		}
 
-#if RTE_SCHED_DEBUG
+#ifdef RTE_SCHED_DEBUG
 		debug_check_queue_slab(port, 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))) {
+		if (unlikely(grinder_pipe_exists(port, bmp_pos)))
 			return 0;
-		}
+
 		port->grinder_base_bmp_pos[pos] = bmp_pos;
 
 		/* Install new pipe group into grinder's pipe cache */
@@ -1866,24 +1955,6 @@ grinder_next_pipe(struct rte_sched_port *port, uint32_t pos)
 	return 1;
 }
 
-#if RTE_SCHED_WRR == 0
-
-#define grinder_wrr_load(a,b)
-
-#define grinder_wrr_store(a,b)
-
-static inline void
-grinder_wrr(struct rte_sched_port *port, uint32_t pos)
-{
-	struct rte_sched_grinder *grinder = port->grinder + pos;
-	uint64_t slab = grinder->qmask;
-
-	if (rte_bsf64(slab, &grinder->qpos) == 0) {
-		rte_panic("grinder wrr\n");
-	}
-}
-
-#elif RTE_SCHED_WRR == 1
 
 static inline void
 grinder_wrr_load(struct rte_sched_port *port, uint32_t pos)
@@ -1923,10 +1994,14 @@ grinder_wrr_store(struct rte_sched_port *port, uint32_t pos)
 
 	qindex = tc_index * 4;
 
-	pipe->wrr_tokens[qindex] = (uint8_t) ((grinder->wrr_tokens[0] & grinder->wrr_mask[0]) >> RTE_SCHED_WRR_SHIFT);
-	pipe->wrr_tokens[qindex + 1] = (uint8_t) ((grinder->wrr_tokens[1] & grinder->wrr_mask[1]) >> RTE_SCHED_WRR_SHIFT);
-	pipe->wrr_tokens[qindex + 2] = (uint8_t) ((grinder->wrr_tokens[2] & grinder->wrr_mask[2]) >> RTE_SCHED_WRR_SHIFT);
-	pipe->wrr_tokens[qindex + 3] = (uint8_t) ((grinder->wrr_tokens[3] & grinder->wrr_mask[3]) >> RTE_SCHED_WRR_SHIFT);
+	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;
 }
 
 static inline void
@@ -1949,11 +2024,6 @@ grinder_wrr(struct rte_sched_port *port, uint32_t pos)
 	grinder->wrr_tokens[3] -= wrr_tokens_min;
 }
 
-#else
-
-#error Invalid value for RTE_SCHED_WRR
-
-#endif /* RTE_SCHED_WRR */
 
 #define grinder_evict(port, pos)
 
@@ -2017,7 +2087,7 @@ grinder_handle(struct rte_sched_port *port, uint32_t pos)
 	{
 		if (grinder_next_pipe(port, pos)) {
 			grinder_prefetch_pipe(port, pos);
-			port->busy_grinders ++;
+			port->busy_grinders++;
 
 			grinder->state = e_GRINDER_PREFETCH_TC_QUEUE_ARRAYS;
 			return 0;
@@ -2068,9 +2138,11 @@ grinder_handle(struct rte_sched_port *port, uint32_t pos)
 			grinder->state = e_GRINDER_PREFETCH_MBUF;
 			return result;
 		}
-		if ((grinder->productive == 0) && (port->pipe_loop == RTE_SCHED_PIPE_INVALID)) {
+
+		if (grinder->productive == 0 &&
+		    port->pipe_loop == RTE_SCHED_PIPE_INVALID)
 			port->pipe_loop = grinder->pindex;
-		}
+
 		grinder_evict(port, pos);
 
 		/* Look for another active pipe */
@@ -2082,7 +2154,7 @@ grinder_handle(struct rte_sched_port *port, uint32_t pos)
 		}
 
 		/* No active pipe found */
-		port->busy_grinders --;
+		port->busy_grinders--;
 
 		grinder->state = e_GRINDER_PREFETCH_PIPE;
 		return result;
@@ -2099,14 +2171,17 @@ rte_sched_port_time_resync(struct rte_sched_port *port)
 {
 	uint64_t cycles = rte_get_tsc_cycles();
 	uint64_t cycles_diff = cycles - port->time_cpu_cycles;
-	double bytes_diff = ((double) cycles_diff) / port->cycles_per_byte;
+	uint64_t bytes_diff;
+
+	/* Compute elapsed time in bytes */
+	bytes_diff = rte_reciprocal_divide(cycles_diff << RTE_SCHED_TIME_SHIFT,
+					   port->inv_cycles_per_byte);
 
 	/* Advance port time */
 	port->time_cpu_cycles = cycles;
-	port->time_cpu_bytes += (uint64_t) bytes_diff;
-	if (port->time < port->time_cpu_bytes) {
+	port->time_cpu_bytes += bytes_diff;
+	if (port->time < port->time_cpu_bytes)
 		port->time = port->time_cpu_bytes;
-	}
 
 	/* Reset pipe loop detection */
 	port->pipe_loop = RTE_SCHED_PIPE_INVALID;
@@ -2138,7 +2213,7 @@ rte_sched_port_dequeue(struct rte_sched_port *port, struct rte_mbuf **pkts, uint
 	rte_sched_port_time_resync(port);
 
 	/* Take each queue in the grinder one step further */
-	for (i = 0, count = 0; ; i ++)  {
+	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)) {