[PATCH v2] app/testpmd: txonly multiflow port change support

[Joshua Washington <joshwash@google.com>]

Google cloud routes traffic using IP addresses without the support of MAC
addresses, so changing source IP address for txonly-multi-flow can have
negative performance implications for net/gve when using testpmd. This
change adds a new flag --txonly-alter-port, which allows the alteration
of source port instead of source IP address for txonly multiflow mode.

The new flag can be tested with the following command:
dpdk-testpmd -- --forward-mode=txonly --txonly-multi-flow \
  --txonly-alter-port --txip=<SRC>,<DST>

Signed-off-by: Joshua Washington <joshwash@google.com>
Reviewed-by: Rushil Gupta <rushilg@google.com>
---
 app/test-pmd/parameters.c             |  4 +++
 app/test-pmd/testpmd.c                |  6 ++++
 app/test-pmd/testpmd.h                |  1 +
 app/test-pmd/txonly.c                 | 51 ++++++++++++++++++---------
 doc/guides/testpmd_app_ug/run_app.rst |  6 ++++
 5 files changed, 52 insertions(+), 16 deletions(-)

diff --git a/app/test-pmd/parameters.c b/app/test-pmd/parameters.c
index 3b37809baf..a99f04b4f0 100644
--- a/app/test-pmd/parameters.c
+++ b/app/test-pmd/parameters.c
@@ -158,6 +158,7 @@ usage(char* progname)
 		" or total packet length.\n");
 	printf("  --multi-rx-mempool: enable multi-rx-mempool support\n");
 	printf("  --txonly-multi-flow: generate multiple flows in txonly mode\n");
+	printf("  --txonly-alter-port: alter source port when generating multiple flows in txonly mode. Source IP address is changed by default\n");
 	printf("  --tx-ip=src,dst: IP addresses in Tx-only mode\n");
 	printf("  --tx-udp=src[,dst]: UDP ports in Tx-only mode\n");
 	printf("  --eth-link-speed: force link speed.\n");
@@ -678,6 +679,7 @@ launch_args_parse(int argc, char** argv)
 		{ "txpkts",			1, 0, 0 },
 		{ "multi-rx-mempool",           0, 0, 0 },
 		{ "txonly-multi-flow",		0, 0, 0 },
+		{ "txonly-alter-port",		0, 0, 0 },
 		{ "rxq-share",			2, 0, 0 },
 		{ "eth-link-speed",		1, 0, 0 },
 		{ "disable-link-check",		0, 0, 0 },
@@ -1307,6 +1309,8 @@ launch_args_parse(int argc, char** argv)
 				multi_rx_mempool = 1;
 			if (!strcmp(lgopts[opt_idx].name, "txonly-multi-flow"))
 				txonly_multi_flow = 1;
+			if (!strcmp(lgopts[opt_idx].name, "txonly-alter-port"))
+				txonly_alter_port = 1;
 			if (!strcmp(lgopts[opt_idx].name, "rxq-share")) {
 				if (optarg == NULL) {
 					rxq_share = UINT32_MAX;
diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
index 5cb6f92523..b78bc4c26d 100644
--- a/app/test-pmd/testpmd.c
+++ b/app/test-pmd/testpmd.c
@@ -267,6 +267,12 @@ enum tx_pkt_split tx_pkt_split = TX_PKT_SPLIT_OFF;
 uint8_t txonly_multi_flow;
 /**< Whether multiple flows are generated in TXONLY mode. */
 
+uint8_t txonly_alter_port;
+/*
+ * Whether source port should be altered instead of IP address when generating
+ * multiple flows in TXONLY mode.
+ */
+
 uint32_t tx_pkt_times_inter;
 /**< Timings for send scheduling in TXONLY mode, time between bursts. */
 
diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
index bdfbfd36d3..8926b51595 100644
--- a/app/test-pmd/testpmd.h
+++ b/app/test-pmd/testpmd.h
@@ -619,6 +619,7 @@ enum tx_pkt_split {
 extern enum tx_pkt_split tx_pkt_split;
 
 extern uint8_t txonly_multi_flow;
+extern uint8_t txonly_alter_port;
 
 extern uint32_t rxq_share;
 
diff --git a/app/test-pmd/txonly.c b/app/test-pmd/txonly.c
index b3d6873104..58b8834f08 100644
--- a/app/test-pmd/txonly.c
+++ b/app/test-pmd/txonly.c
@@ -56,7 +56,7 @@ uint32_t tx_ip_dst_addr = (198U << 24) | (18 << 16) | (0 << 8) | 2;
 #define IP_DEFTTL  64   /* from RFC 1340. */
 
 static struct rte_ipv4_hdr pkt_ip_hdr; /**< IP header of transmitted packets. */
-RTE_DEFINE_PER_LCORE(uint8_t, _ip_var); /**< IP address variation */
+RTE_DEFINE_PER_LCORE(uint16_t, _src_var); /**< Source address/port variation */
 static struct rte_udp_hdr pkt_udp_hdr; /**< UDP header of tx packets. */
 
 static uint64_t timestamp_mask; /**< Timestamp dynamic flag mask */
@@ -230,28 +230,47 @@ pkt_burst_prepare(struct rte_mbuf *pkt, struct rte_mempool *mbp,
 	copy_buf_to_pkt(eth_hdr, sizeof(*eth_hdr), pkt, 0);
 	copy_buf_to_pkt(&pkt_ip_hdr, sizeof(pkt_ip_hdr), pkt,
 			sizeof(struct rte_ether_hdr));
+	copy_buf_to_pkt(&pkt_udp_hdr, sizeof(pkt_udp_hdr), pkt,
+			sizeof(struct rte_ether_hdr) +
+			sizeof(struct rte_ipv4_hdr));
 	if (txonly_multi_flow) {
-		uint8_t  ip_var = RTE_PER_LCORE(_ip_var);
-		struct rte_ipv4_hdr *ip_hdr;
-		uint32_t addr;
+		uint16_t src_var = RTE_PER_LCORE(_src_var);
 
-		ip_hdr = rte_pktmbuf_mtod_offset(pkt,
-				struct rte_ipv4_hdr *,
-				sizeof(struct rte_ether_hdr));
 		/*
-		 * Generate multiple flows by varying IP src addr. This
-		 * enables packets are well distributed by RSS in
+		 * Generate multiple flows by varying IP source address/port.
+		 * This enables packets are well distributed by RSS in
 		 * receiver side if any and txonly mode can be a decent
 		 * packet generator for developer's quick performance
 		 * regression test.
 		 */
-		addr = (tx_ip_dst_addr | (ip_var++ << 8)) + rte_lcore_id();
-		ip_hdr->src_addr = rte_cpu_to_be_32(addr);
-		RTE_PER_LCORE(_ip_var) = ip_var;
+		if (txonly_alter_port) {
+			struct rte_udp_hdr *udp_hdr;
+			uint16_t port;
+
+			udp_hdr = rte_pktmbuf_mtod_offset(pkt,
+					struct rte_udp_hdr *,
+					sizeof(struct rte_ether_hdr) +
+					sizeof(struct rte_ipv4_hdr));
+			/*
+			 *  Vary source port.
+			 */
+			port = src_var++;
+			udp_hdr->src_port = rte_cpu_to_be_16(port);
+		} else {
+			struct rte_ipv4_hdr *ip_hdr;
+			uint32_t addr;
+
+			ip_hdr = rte_pktmbuf_mtod_offset(pkt,
+					struct rte_ipv4_hdr *,
+					sizeof(struct rte_ether_hdr));
+			/*
+			 * Vary IP source address.
+			 */
+			addr = (tx_ip_dst_addr | ((src_var++ & 0xFF) << 8)) + rte_lcore_id();
+			ip_hdr->src_addr = rte_cpu_to_be_32(addr);
+		}
+		RTE_PER_LCORE(_src_var) = src_var;
 	}
-	copy_buf_to_pkt(&pkt_udp_hdr, sizeof(pkt_udp_hdr), pkt,
-			sizeof(struct rte_ether_hdr) +
-			sizeof(struct rte_ipv4_hdr));
 
 	if (unlikely(tx_pkt_split == TX_PKT_SPLIT_RND) || txonly_multi_flow)
 		update_pkt_header(pkt, pkt_len);
@@ -393,7 +412,7 @@ pkt_burst_transmit(struct fwd_stream *fs)
 	nb_tx = common_fwd_stream_transmit(fs, pkts_burst, nb_pkt);
 
 	if (txonly_multi_flow)
-		RTE_PER_LCORE(_ip_var) -= nb_pkt - nb_tx;
+		RTE_PER_LCORE(_src_var) -= nb_pkt - nb_tx;
 
 	if (unlikely(nb_tx < nb_pkt)) {
 		if (verbose_level > 0 && fs->fwd_dropped == 0)
diff --git a/doc/guides/testpmd_app_ug/run_app.rst b/doc/guides/testpmd_app_ug/run_app.rst
index 57b23241cf..89fac9e964 100644
--- a/doc/guides/testpmd_app_ug/run_app.rst
+++ b/doc/guides/testpmd_app_ug/run_app.rst
@@ -373,6 +373,12 @@ The command line options are:
 
     Generate multiple flows in txonly mode.
 
+*   ``--txonly-alter-port``
+
+    Alter source port when generating multiple flows in txonly mode. The default
+    behavior is to alter source IP address. This flag must be used in
+    conjunction with --txonly-multi-flow.
+
 *   ``--rxq-share=[X]``
 
     Create queues in shared Rx queue mode if device supports.
-- 
2.40.0.577.gac1e443424-goog