From 57e852429ded674ff27ed03eb3dc6b3d6f66f393 Mon Sep 17 00:00:00 2001
From: Bruce Richardson <bruce.richardson@intel.com>
Date: Wed, 12 Feb 2014 16:09:58 +0000
Subject: [PATCH] app/testpmd: added new macfwd-retry mode

This adds a new forwarding mode to testpmd, whereby it will forward
packets, but on a transmit failure, it will wait a small amount of time
before retrying the transmit again. This mode can be useful in some
cases to set a zero-packet loss rate at the cost of overall throughput,
especially in cases where TX ring size/buffering is constrained, such as
with virtio.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
 app/test-pmd/Makefile       |   1 +
 app/test-pmd/cmdline.c      |  83 +++++++++++++++---
 app/test-pmd/config.c       |   3 +
 app/test-pmd/macfwd-retry.c | 165 ++++++++++++++++++++++++++++++++++++
 app/test-pmd/testpmd.c      |   1 +
 app/test-pmd/testpmd.h      |   4 +
 6 files changed, 247 insertions(+), 10 deletions(-)
 create mode 100644 app/test-pmd/macfwd-retry.c

diff --git a/app/test-pmd/Makefile b/app/test-pmd/Makefile
index 9882c6145e..4f80442cd4 100644
--- a/app/test-pmd/Makefile
+++ b/app/test-pmd/Makefile
@@ -52,6 +52,7 @@ SRCS-$(CONFIG_RTE_TEST_PMD) += cmdline.c
 SRCS-$(CONFIG_RTE_TEST_PMD) += config.c
 SRCS-$(CONFIG_RTE_TEST_PMD) += iofwd.c
 SRCS-$(CONFIG_RTE_TEST_PMD) += macfwd.c
+SRCS-$(CONFIG_RTE_TEST_PMD) += macfwd-retry.c
 SRCS-$(CONFIG_RTE_TEST_PMD) += rxonly.c
 SRCS-$(CONFIG_RTE_TEST_PMD) += txonly.c
 SRCS-$(CONFIG_RTE_TEST_PMD) += csumonly.c
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 4bf0720eb2..8c47c01f0b 100644
--- a/app/test-pmd/cmdline.c
+++ b/app/test-pmd/cmdline.c
@@ -224,6 +224,10 @@ static void cmd_help_long_parsed(void *parsed_result,
 			"set burst (num)\n"
 			"    Set number of packets per burst.\n\n"
 
+			"set burst tx delay (microseconds) retry (num)\n"
+			"    Set the transmit delay time and number of retries"
+			" in mac_retry forwarding mode.\n\n"
+
 			"set txpkts (x[,y]*)\n"
 			"    Set the length of each segment of TXONLY"
 			" packets.\n\n"
@@ -289,13 +293,13 @@ static void cmd_help_long_parsed(void *parsed_result,
 			"    Please check the NIC datasheet for HW limits.\n\n"
 
 #ifdef RTE_LIBRTE_IEEE1588
-			"set fwd (io|mac|rxonly|txonly|csum|ieee1588)\n"
-			"    Set IO, MAC, RXONLY, CSUM or TXONLY or ieee1588"
+			"set fwd (io|mac|mac_retry|rxonly|txonly|csum|ieee1588)\n"
+			"    Set IO, MAC, MAC_RETRY, RXONLY, CSUM or TXONLY or ieee1588"
 			" packet forwarding mode.\n\n"
 
 #else
-			"set fwd (io|mac|rxonly|txonly|csum)\n"
-			"    Set IO, MAC, RXONLY, CSUM or TXONLY packet"
+			"set fwd (io|mac|mac_retry|rxonly|txonly|csum)\n"
+			"    Set IO, MAC, MAC_RETRY, RXONLY, CSUM or TXONLY packet"
 			" forwarding mode.\n\n"
 
 #endif
@@ -2683,20 +2687,20 @@ cmdline_parse_token_string_t cmd_setfwd_fwd =
 cmdline_parse_token_string_t cmd_setfwd_mode =
 	TOKEN_STRING_INITIALIZER(struct cmd_set_fwd_mode_result, mode,
 #ifdef RTE_LIBRTE_IEEE1588
-				 "io#mac#rxonly#txonly#csum#ieee1588");
+				 "io#mac#mac_retry#rxonly#txonly#csum#ieee1588");
 #else
-				 "io#mac#rxonly#txonly#csum");
+				 "io#mac#mac_retry#rxonly#txonly#csum");
 #endif
 
 cmdline_parse_inst_t cmd_set_fwd_mode = {
 	.f = cmd_set_fwd_mode_parsed,
 	.data = NULL,
 #ifdef RTE_LIBRTE_IEEE1588
-	.help_str = "set fwd io|mac|rxonly|txonly|csum|ieee1588 - set IO, MAC,"
-	" RXONLY, TXONLY, CSUM or IEEE1588 packet forwarding mode",
+	.help_str = "set fwd io|mac|mac_retry|rxonly|txonly|csum|ieee1588 - set IO, MAC,"
+	" MAC_RETRY, RXONLY, TXONLY, CSUM or IEEE1588 packet forwarding mode",
 #else
-	.help_str = "set fwd io|mac|rxonly|txonly|csum - set IO, MAC,"
-	" RXONLY, CSUM or TXONLY packet forwarding mode",
+	.help_str = "set fwd io|mac|mac_retry|rxonly|txonly|csum - set IO, MAC,"
+	" MAC_RETRY, RXONLY, CSUM or TXONLY packet forwarding mode",
 #endif
 	.tokens = {
 		(void *)&cmd_setfwd_set,
@@ -2706,6 +2710,64 @@ cmdline_parse_inst_t cmd_set_fwd_mode = {
 	},
 };
 
+/* *** SET BURST TX DELAY TIME RETRY NUMBER *** */
+struct cmd_set_burst_tx_retry_result {
+	cmdline_fixed_string_t set;
+	cmdline_fixed_string_t burst;
+	cmdline_fixed_string_t tx;
+	cmdline_fixed_string_t delay;
+	uint32_t time;
+	cmdline_fixed_string_t retry;
+	uint32_t retry_num;
+};
+
+static void cmd_set_burst_tx_retry_parsed(void *parsed_result,
+					__attribute__((unused)) struct cmdline *cl,
+					__attribute__((unused)) void *data)
+{
+	struct cmd_set_burst_tx_retry_result *res = parsed_result;
+
+	if (!strcmp(res->set, "set") && !strcmp(res->burst, "burst")
+		&& !strcmp(res->tx, "tx")) {
+		if (!strcmp(res->delay, "delay"))
+			burst_tx_delay_time = res->time;	
+		if (!strcmp(res->retry, "retry"))
+			burst_tx_retry_num = res->retry_num;	
+	}
+
+}
+
+cmdline_parse_token_string_t cmd_set_burst_tx_retry_set =
+	TOKEN_STRING_INITIALIZER(struct cmd_set_burst_tx_retry_result, set, "set");
+cmdline_parse_token_string_t cmd_set_burst_tx_retry_burst =
+	TOKEN_STRING_INITIALIZER(struct cmd_set_burst_tx_retry_result, burst,
+				 "burst");
+cmdline_parse_token_string_t cmd_set_burst_tx_retry_tx =
+	TOKEN_STRING_INITIALIZER(struct cmd_set_burst_tx_retry_result, tx, "tx");
+cmdline_parse_token_string_t cmd_set_burst_tx_retry_delay =
+	TOKEN_STRING_INITIALIZER(struct cmd_set_burst_tx_retry_result, delay, "delay");
+cmdline_parse_token_num_t cmd_set_burst_tx_retry_time =
+	TOKEN_NUM_INITIALIZER(struct cmd_set_burst_tx_retry_result, time, UINT32);
+cmdline_parse_token_string_t cmd_set_burst_tx_retry_retry =
+	TOKEN_STRING_INITIALIZER(struct cmd_set_burst_tx_retry_result, retry, "retry");
+cmdline_parse_token_num_t cmd_set_burst_tx_retry_retry_num =
+	TOKEN_NUM_INITIALIZER(struct cmd_set_burst_tx_retry_result, retry_num, UINT32);
+
+cmdline_parse_inst_t cmd_set_burst_tx_retry = {
+	.f = cmd_set_burst_tx_retry_parsed,
+	.help_str = "set burst tx delay (time_by_useconds) retry (retry_num)",
+	.tokens = {
+		(void *)&cmd_set_burst_tx_retry_set,
+		(void *)&cmd_set_burst_tx_retry_burst,
+		(void *)&cmd_set_burst_tx_retry_tx,
+		(void *)&cmd_set_burst_tx_retry_delay,
+		(void *)&cmd_set_burst_tx_retry_time,
+		(void *)&cmd_set_burst_tx_retry_retry,
+		(void *)&cmd_set_burst_tx_retry_retry_num,
+		NULL,
+	},
+};
+
 /* *** SET PROMISC MODE *** */
 struct cmd_set_promisc_mode_result {
 	cmdline_fixed_string_t set;
@@ -4952,6 +5014,7 @@ cmdline_parse_ctx_t main_ctx[] = {
 	(cmdline_parse_inst_t *)&cmd_set_fwd_list,
 	(cmdline_parse_inst_t *)&cmd_set_fwd_mask,
 	(cmdline_parse_inst_t *)&cmd_set_fwd_mode,
+	(cmdline_parse_inst_t *)&cmd_set_burst_tx_retry,
 	(cmdline_parse_inst_t *)&cmd_set_promisc_mode_one,
 	(cmdline_parse_inst_t *)&cmd_set_promisc_mode_all,
 	(cmdline_parse_inst_t *)&cmd_set_allmulti_mode_one,
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index a7a125b66f..3a244f5473 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -980,6 +980,9 @@ pkt_fwd_config_display(struct fwd_config *cfg)
 		numa_support == 1 ? "enabled" : "disabled",
 		mp_anon != 0 ? "enabled" : "disabled");
 
+	if (strcmp(cfg->fwd_eng->fwd_mode_name, "mac_retry") == 0)
+		printf("TX retry num: %u, delay between TX retries: %uus\n",
+			burst_tx_retry_num, burst_tx_delay_time);
 	for (lc_id = 0; lc_id < cfg->nb_fwd_lcores; lc_id++) {
 		printf("Logical Core %u (socket %u) forwards packets on "
 		       "%d streams:",
diff --git a/app/test-pmd/macfwd-retry.c b/app/test-pmd/macfwd-retry.c
new file mode 100644
index 0000000000..98fc037e6e
--- /dev/null
+++ b/app/test-pmd/macfwd-retry.c
@@ -0,0 +1,165 @@
+/*-
+ *   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.
+ */
+
+#include <stdarg.h>
+#include <string.h>
+#include <stdio.h>
+#include <errno.h>
+#include <stdint.h>
+#include <unistd.h>
+#include <inttypes.h>
+
+#include <sys/queue.h>
+#include <sys/stat.h>
+
+#include <rte_common.h>
+#include <rte_byteorder.h>
+#include <rte_log.h>
+#include <rte_debug.h>
+#include <rte_cycles.h>
+#include <rte_memory.h>
+#include <rte_memcpy.h>
+#include <rte_memzone.h>
+#include <rte_launch.h>
+#include <rte_tailq.h>
+#include <rte_eal.h>
+#include <rte_per_lcore.h>
+#include <rte_lcore.h>
+#include <rte_atomic.h>
+#include <rte_branch_prediction.h>
+#include <rte_ring.h>
+#include <rte_memory.h>
+#include <rte_mempool.h>
+#include <rte_mbuf.h>
+#include <rte_interrupts.h>
+#include <rte_pci.h>
+#include <rte_ether.h>
+#include <rte_ethdev.h>
+#include <rte_ip.h>
+#include <rte_string_fns.h>
+
+#include "testpmd.h"
+
+#define BURST_TX_WAIT_US 10
+#define BURST_TX_RETRIES 5
+
+/*
+ * Global variables that control number of retires and
+ * timeout (in us) between retires.
+ */
+uint32_t burst_tx_delay_time = BURST_TX_WAIT_US;
+uint32_t burst_tx_retry_num = BURST_TX_RETRIES;
+
+/*
+ * Forwarding of packets in MAC mode with a wait and retry on TX to reduce packet loss.
+ * Change the source and the destination Ethernet addressed of packets
+ * before forwarding them.
+ */
+static void
+pkt_burst_mac_retry_forward(struct fwd_stream *fs)
+{
+	struct rte_mbuf  *pkts_burst[MAX_PKT_BURST];
+	struct rte_mbuf  *mb;
+	struct ether_hdr *eth_hdr;
+	uint32_t retry;
+	uint16_t nb_rx;
+	uint16_t nb_tx;
+	uint16_t i;
+#ifdef RTE_TEST_PMD_RECORD_CORE_CYCLES
+	uint64_t start_tsc;
+	uint64_t end_tsc;
+	uint64_t core_cycles;
+#endif
+
+#ifdef RTE_TEST_PMD_RECORD_CORE_CYCLES
+	start_tsc = rte_rdtsc();
+#endif
+
+	/*
+	 * Receive a burst of packets and forward them.
+	 */
+	nb_rx = rte_eth_rx_burst(fs->rx_port, fs->rx_queue, pkts_burst,
+				 nb_pkt_per_burst);
+	if (unlikely(nb_rx == 0))
+		return;
+
+#ifdef RTE_TEST_PMD_RECORD_BURST_STATS
+	fs->rx_burst_stats.pkt_burst_spread[nb_rx]++;
+#endif
+	fs->rx_packets += nb_rx;
+	for (i = 0; i < nb_rx; i++) {
+		mb = pkts_burst[i];
+		eth_hdr = (struct ether_hdr *) mb->pkt.data;
+		ether_addr_copy(&peer_eth_addrs[fs->peer_addr],
+				&eth_hdr->d_addr);
+		ether_addr_copy(&ports[fs->tx_port].eth_addr,
+				&eth_hdr->s_addr);
+	}
+	nb_tx = rte_eth_tx_burst(fs->tx_port, fs->tx_queue, pkts_burst, nb_rx);
+
+	/*
+	 * If not all packets have been TX'd then wait and retry.
+	 */
+	if (unlikely(nb_tx < nb_rx)) {
+		for (retry = 0; retry < burst_tx_retry_num; retry++) {
+			rte_delay_us(burst_tx_delay_time);
+			nb_tx += rte_eth_tx_burst(fs->tx_port, fs->tx_queue,
+				&pkts_burst[nb_tx], nb_rx - nb_tx);
+			if (nb_tx == nb_rx)
+				break;
+		}
+	}
+
+	fs->tx_packets += nb_tx;
+#ifdef RTE_TEST_PMD_RECORD_BURST_STATS
+	fs->tx_burst_stats.pkt_burst_spread[nb_tx]++;
+#endif
+	if (unlikely(nb_tx < nb_rx)) {
+		fs->fwd_dropped += (nb_rx - nb_tx);
+		do {
+			rte_pktmbuf_free(pkts_burst[nb_tx]);
+		} while (++nb_tx < nb_rx);
+	}
+#ifdef RTE_TEST_PMD_RECORD_CORE_CYCLES
+	end_tsc = rte_rdtsc();
+	core_cycles = (end_tsc - start_tsc);
+	fs->core_cycles = (uint64_t) (fs->core_cycles + core_cycles);
+#endif
+}
+
+struct fwd_engine mac_retry_fwd_engine = {
+	.fwd_mode_name  = "mac_retry",
+	.port_fwd_begin = NULL,
+	.port_fwd_end   = NULL,
+	.packet_fwd     = pkt_burst_mac_retry_forward,
+};
diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 0b376ccc4d..28440e5a43 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -143,6 +143,7 @@ streamid_t nb_fwd_streams;       /**< Is equal to (nb_ports * nb_rxq). */
 struct fwd_engine * fwd_engines[] = {
 	&io_fwd_engine,
 	&mac_fwd_engine,
+	&mac_retry_fwd_engine,
 	&rx_only_engine,
 	&tx_only_engine,
 	&csum_fwd_engine,
diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
index 834277a611..5b4ee6f929 100644
--- a/app/test-pmd/testpmd.h
+++ b/app/test-pmd/testpmd.h
@@ -198,6 +198,7 @@ struct fwd_engine {
 
 extern struct fwd_engine io_fwd_engine;
 extern struct fwd_engine mac_fwd_engine;
+extern struct fwd_engine mac_retry_fwd_engine;
 extern struct fwd_engine rx_only_engine;
 extern struct fwd_engine tx_only_engine;
 extern struct fwd_engine csum_fwd_engine;
@@ -363,6 +364,9 @@ extern struct fwd_stream **fwd_streams;
 extern portid_t nb_peer_eth_addrs; /**< Number of peer ethernet addresses. */
 extern struct ether_addr peer_eth_addrs[RTE_MAX_ETHPORTS];
 
+extern uint32_t burst_tx_delay_time; /**< Burst tx delay time(us) for mac-retry. */
+extern uint32_t burst_tx_retry_num;  /**< Burst tx retry number for mac-retry. */
+
 static inline unsigned int
 lcore_num(void)
 {
-- 
2.39.5