DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] examples/l3fwd: optimised packet processing on powerpc
@ 2017-09-21  8:37 Gowrishankar
  2017-09-21 10:05 ` [dpdk-dev] [PATCH v2] " Gowrishankar
  0 siblings, 1 reply; 4+ messages in thread
From: Gowrishankar @ 2017-09-21  8:37 UTC (permalink / raw)
  To: dev; +Cc: Chao Zhu, Tomasz Kantecki, Gowrishankar Muthukrishnan

From: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>

This patch adds altivec support for lpm packet processing in powerpc.

Signed-off-by: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
---
 examples/l3fwd/l3fwd_altivec.h     | 279 +++++++++++++++++++++++++++++++++++++
 examples/l3fwd/l3fwd_lpm.c         |   4 +-
 examples/l3fwd/l3fwd_lpm_altivec.h | 155 +++++++++++++++++++++
 3 files changed, 437 insertions(+), 1 deletion(-)
 create mode 100644 examples/l3fwd/l3fwd_altivec.h
 create mode 100644 examples/l3fwd/l3fwd_lpm_altivec.h

diff --git a/examples/l3fwd/l3fwd_altivec.h b/examples/l3fwd/l3fwd_altivec.h
new file mode 100644
index 0000000..732a461
--- /dev/null
+++ b/examples/l3fwd/l3fwd_altivec.h
@@ -0,0 +1,279 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2016 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2017 IBM Corporation.
+ *   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.
+ */
+
+
+#ifndef _L3FWD_ALTIVEC_H_
+#define _L3FWD_ALTIVEC_H_
+
+#include "l3fwd.h"
+#include "l3fwd_common.h"
+
+/*
+ * Update source and destination MAC addresses in the ethernet header.
+ * Perform RFC1812 checks and updates for IPV4 packets.
+ */
+static inline void
+processx4_step3(struct rte_mbuf *pkt[FWDSTEP], uint16_t dst_port[FWDSTEP])
+{
+	vector unsigned int te[FWDSTEP];
+	vector unsigned int ve[FWDSTEP];
+	vector unsigned int *p[FWDSTEP];
+
+	p[0] = rte_pktmbuf_mtod(pkt[0], vector unsigned int *);
+	p[1] = rte_pktmbuf_mtod(pkt[1], vector unsigned int *);
+	p[2] = rte_pktmbuf_mtod(pkt[2], vector unsigned int *);
+	p[3] = rte_pktmbuf_mtod(pkt[3], vector unsigned int *);
+
+	ve[0] = (vector unsigned int)val_eth[dst_port[0]];
+	te[0] = *p[0];
+
+	ve[1] = (vector unsigned int)val_eth[dst_port[1]];
+	te[1] = *p[1];
+
+	ve[2] = (vector unsigned int)val_eth[dst_port[2]];
+	te[2] = *p[2];
+
+	ve[3] = (vector unsigned int)val_eth[dst_port[3]];
+	te[3] = *p[3];
+
+	/* Update first 12 bytes, keep rest bytes intact. */
+	te[0] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[0],
+			(vector unsigned short)te[0],
+			(vector unsigned short){0,0,0,0,0,0,0xffff,0xffff});
+
+	te[1] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[1],
+			(vector unsigned short)te[1],
+			(vector unsigned short){0,0,0,0,0,0,0xffff,0xffff});
+
+	te[2] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[2],
+			(vector unsigned short)te[2],
+			(vector unsigned short){0,0,0,0,0,0,0xffff,0xffff});
+
+	te[3] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[3],
+			(vector unsigned short)te[3],
+			(vector unsigned short){0,0,0,0,0,0,0xffff,0xffff});
+
+	*p[0] = te[0];
+	*p[1] = te[1];
+	*p[2] = te[2];
+	*p[3] = te[3];
+
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[0] + 1),
+		&dst_port[0], pkt[0]->packet_type);
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[1] + 1),
+		&dst_port[1], pkt[1]->packet_type);
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[2] + 1),
+		&dst_port[2], pkt[2]->packet_type);
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[3] + 1),
+		&dst_port[3], pkt[3]->packet_type);
+}
+
+/*
+ * Group consecutive packets with the same destination port in bursts of 4.
+ * Suppose we have array of destionation ports:
+ * dst_port[] = {a, b, c, d,, e, ... }
+ * dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
+ * We doing 4 comparisons at once and the result is 4 bit mask.
+ * This mask is used as an index into prebuild array of pnum values.
+ */
+static inline uint16_t *
+port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, vector unsigned short dp1, vector unsigned short dp2)
+{
+	union {
+		uint16_t u16[FWDSTEP + 1];
+		uint64_t u64;
+	} *pnum = (void *)pn;
+
+	int32_t v;
+
+	v = vec_any_eq(dp1, dp2);
+
+
+	/* update last port counter. */
+	lp[0] += gptbl[v].lpv;
+
+	/* if dest port value has changed. */
+	if (v != GRPMSK) {
+		pnum->u64 = gptbl[v].pnum;
+		pnum->u16[FWDSTEP] = 1;
+		lp = pnum->u16 + gptbl[v].idx;
+	}
+
+	return lp;
+}
+
+/**
+ * Process one packet:
+ * Update source and destination MAC addresses in the ethernet header.
+ * Perform RFC1812 checks and updates for IPV4 packets.
+ */
+static inline void
+process_packet(struct rte_mbuf *pkt, uint16_t *dst_port)
+{
+	struct ether_hdr *eth_hdr;
+	vector unsigned int te, ve;
+
+	eth_hdr = rte_pktmbuf_mtod(pkt, struct ether_hdr *);
+
+	te = *(vector unsigned int *)eth_hdr;
+	ve = (vector unsigned int)val_eth[dst_port[0]];
+
+	rfc1812_process((struct ipv4_hdr *)(eth_hdr + 1), dst_port,
+			pkt->packet_type);
+
+	/* dynamically vec_sel te and ve for MASK_ETH (0x3f) */
+	te = (vector unsigned int)vec_sel(
+		(vector unsigned short)ve,
+		(vector unsigned short)te,
+		(vector unsigned short){0,0,0,0,0,0,0xffff,0xffff});
+
+	*(vector unsigned int *)eth_hdr = te;
+}
+
+/**
+ * Send packets burst from pkts_burst to the ports in dst_port array
+ */
+static __rte_always_inline void
+send_packets_multi(struct lcore_conf *qconf, struct rte_mbuf **pkts_burst,
+		uint16_t dst_port[MAX_PKT_BURST], int nb_rx)
+{
+	int32_t k;
+	int j = 0;
+	uint16_t dlp;
+	uint16_t *lp;
+	uint16_t pnum[MAX_PKT_BURST + 1];
+
+	/*
+	 * Finish packet processing and group consecutive
+	 * packets with the same destination port.
+	 */
+	k = RTE_ALIGN_FLOOR(nb_rx, FWDSTEP);
+	if (k != 0) {
+		vector unsigned short dp1, dp2;
+
+		lp = pnum;
+		lp[0] = 1;
+
+		processx4_step3(pkts_burst, dst_port);
+
+		/* dp1: <d[0], d[1], d[2], d[3], ... > */
+		dp1 = *(vector unsigned short *)dst_port;
+
+		for (j = FWDSTEP; j != k; j += FWDSTEP) {
+			processx4_step3(&pkts_burst[j], &dst_port[j]);
+
+			/*
+			 * dp2:
+			 * <d[j-3], d[j-2], d[j-1], d[j], ... >
+			 */
+			dp2 = *((vector unsigned short *)
+					&dst_port[j - FWDSTEP + 1]);
+			lp  = port_groupx4(&pnum[j - FWDSTEP], lp, dp1, dp2);
+
+			/*
+			 * dp1:
+			 * <d[j], d[j+1], d[j+2], d[j+3], ... >
+			 */
+			dp1 = vec_sro(dp2,
+					0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
+					(vector unsigned char){
+						(FWDSTEP - 1) * sizeof(dst_port[0])
+					});
+		}
+
+		/*
+		 * dp2: <d[j-3], d[j-2], d[j-1], d[j-1], ... >
+		 */
+		dp2 = vec_perm(dp1, (vector unsigned short){}, (vector unsigned char){0xf9});
+		lp  = port_groupx4(&pnum[j - FWDSTEP], lp, dp1, dp2);
+
+		/*
+		 * remove values added by the last repeated
+		 * dst port.
+		 */
+		lp[0]--;
+		dlp = dst_port[j - 1];
+	} else {
+		/* set dlp and lp to the never used values. */
+		dlp = BAD_PORT - 1;
+		lp = pnum + MAX_PKT_BURST;
+	}
+
+	/* Process up to last 3 packets one by one. */
+	switch (nb_rx % FWDSTEP) {
+	case 3:
+		process_packet(pkts_burst[j], dst_port + j);
+		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
+		j++;
+		/* fall-through */
+	case 2:
+		process_packet(pkts_burst[j], dst_port + j);
+		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
+		j++;
+		/* fall-through */
+	case 1:
+		process_packet(pkts_burst[j], dst_port + j);
+		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
+		j++;
+	}
+
+	/*
+	 * Send packets out, through destination port.
+	 * Consecutive packets with the same destination port
+	 * are already grouped together.
+	 * If destination port for the packet equals BAD_PORT,
+	 * then free the packet without sending it out.
+	 */
+	for (j = 0; j < nb_rx; j += k) {
+
+		int32_t m;
+		uint16_t pn;
+
+		pn = dst_port[j];
+		k = pnum[j];
+
+		if (likely(pn != BAD_PORT))
+			send_packetsx4(qconf, pn, pkts_burst + j, k);
+		else
+			for (m = j; m != j + k; m++)
+				rte_pktmbuf_free(pkts_burst[m]);
+
+	}
+}
+
+#endif /* _L3FWD_ALTIVEC_H_ */
diff --git a/examples/l3fwd/l3fwd_lpm.c b/examples/l3fwd/l3fwd_lpm.c
index ff1e403..1440792 100644
--- a/examples/l3fwd/l3fwd_lpm.c
+++ b/examples/l3fwd/l3fwd_lpm.c
@@ -191,6 +191,8 @@ struct ipv6_l3fwd_lpm_route {
 #include "l3fwd_lpm_sse.h"
 #elif defined RTE_MACHINE_CPUFLAG_NEON
 #include "l3fwd_lpm_neon.h"
+#elif defined(RTE_ARCH_PPC_64)
+#include "l3fwd_lpm_altivec.h"
 #else
 #include "l3fwd_lpm.h"
 #endif
@@ -263,7 +265,7 @@ struct ipv6_l3fwd_lpm_route {
 			if (nb_rx == 0)
 				continue;
 
-#if defined RTE_ARCH_X86 || defined RTE_MACHINE_CPUFLAG_NEON
+#if defined RTE_ARCH_X86 || defined RTE_MACHINE_CPUFLAG_NEON || defined RTE_ARCH_PPC_64
 			l3fwd_lpm_send_packets(nb_rx, pkts_burst,
 						portid, qconf);
 #else
diff --git a/examples/l3fwd/l3fwd_lpm_altivec.h b/examples/l3fwd/l3fwd_lpm_altivec.h
new file mode 100644
index 0000000..cb29706
--- /dev/null
+++ b/examples/l3fwd/l3fwd_lpm_altivec.h
@@ -0,0 +1,155 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2017 IBM Corporation.
+ *   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.
+ */
+
+#ifndef __L3FWD_LPM_ALTIVEC_H__
+#define __L3FWD_LPM_ALTIVEC_H__
+
+#include "l3fwd_altivec.h"
+
+/*
+ * Read packet_type and destination IPV4 addresses from 4 mbufs.
+ */
+static inline void
+processx4_step1(struct rte_mbuf *pkt[FWDSTEP],
+		vector unsigned int *dip,
+		uint32_t *ipv4_flag)
+{
+	struct ipv4_hdr *ipv4_hdr;
+	struct ether_hdr *eth_hdr;
+	uint32_t x0, x1, x2, x3;
+
+	eth_hdr = rte_pktmbuf_mtod(pkt[0], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x0 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] = pkt[0]->packet_type & RTE_PTYPE_L3_IPV4;
+
+	rte_compiler_barrier();
+	eth_hdr = rte_pktmbuf_mtod(pkt[1], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x1 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] &= pkt[1]->packet_type;
+
+	rte_compiler_barrier();
+	eth_hdr = rte_pktmbuf_mtod(pkt[2], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x2 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] &= pkt[2]->packet_type;
+
+	rte_compiler_barrier();
+	eth_hdr = rte_pktmbuf_mtod(pkt[3], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x3 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] &= pkt[3]->packet_type;
+
+	rte_compiler_barrier();
+	dip[0] = (vector unsigned int){x0, x1, x2, x3};
+}
+
+/*
+ * Lookup into LPM for destination port.
+ * If lookup fails, use incoming port (portid) as destination port.
+ */
+static inline void
+processx4_step2(const struct lcore_conf *qconf,
+		vector unsigned int dip,
+		uint32_t ipv4_flag,
+		uint8_t portid,
+		struct rte_mbuf *pkt[FWDSTEP],
+		uint16_t dprt[FWDSTEP])
+{
+	rte_xmm_t dst;
+	const vector unsigned char bswap_mask = (vector unsigned char){3, 2, 1, 0, 7, 6, 5, 4,
+						11, 10, 9, 8, 15, 14, 13, 12};
+
+	/* Byte swap 4 IPV4 addresses. */
+	dip = (vector unsigned int)vec_perm(*(vector unsigned char *)&dip, (vector unsigned char){}, bswap_mask);
+
+	/* if all 4 packets are IPV4. */
+	if (likely(ipv4_flag)) {
+		rte_lpm_lookupx4(qconf->ipv4_lookup_struct, (xmm_t)dip, (uint32_t *)&dst,
+			portid);
+		/* get rid of unused upper 16 bit for each dport. */
+		dst.x = (xmm_t)vec_packs(dst.x, dst.x);
+		*(uint64_t *)dprt = dst.u64[0];
+	} else {
+		dst.x = (xmm_t)dip;
+		dprt[0] = lpm_get_dst_port_with_ipv4(qconf, pkt[0], dst.u32[0], portid);
+		dprt[1] = lpm_get_dst_port_with_ipv4(qconf, pkt[1], dst.u32[1], portid);
+		dprt[2] = lpm_get_dst_port_with_ipv4(qconf, pkt[2], dst.u32[2], portid);
+		dprt[3] = lpm_get_dst_port_with_ipv4(qconf, pkt[3], dst.u32[3], portid);
+	}
+}
+
+/*
+ * Buffer optimized handling of packets, invoked
+ * from main_loop.
+ */
+static inline void
+l3fwd_lpm_send_packets(int nb_rx, struct rte_mbuf **pkts_burst,
+			uint8_t portid, struct lcore_conf *qconf)
+{
+	int32_t j;
+	uint16_t dst_port[MAX_PKT_BURST];
+	vector unsigned int dip[MAX_PKT_BURST / FWDSTEP];
+	uint32_t ipv4_flag[MAX_PKT_BURST / FWDSTEP];
+	const int32_t k = RTE_ALIGN_FLOOR(nb_rx, FWDSTEP);
+
+	for (j = 0; j != k; j += FWDSTEP)
+		processx4_step1(&pkts_burst[j], &dip[j / FWDSTEP],
+				&ipv4_flag[j / FWDSTEP]);
+
+	for (j = 0; j != k; j += FWDSTEP)
+		processx4_step2(qconf, dip[j / FWDSTEP],
+				ipv4_flag[j / FWDSTEP], portid, &pkts_burst[j], &dst_port[j]);
+
+	/* Classify last up to 3 packets one by one */
+	switch (nb_rx % FWDSTEP) {
+	case 3:
+		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j], portid);
+		j++;
+		/* fall-through */
+	case 2:
+		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j], portid);
+		j++;
+		/* fall-through */
+	case 1:
+		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j], portid);
+		j++;
+		/* fall-through */
+	}
+
+	send_packets_multi(qconf, pkts_burst, dst_port, nb_rx);
+}
+
+#endif /* __L3FWD_LPM_ALTIVEC_H__ */
-- 
1.9.1

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [dpdk-dev] [PATCH v2] examples/l3fwd: optimised packet processing on powerpc
  2017-09-21  8:37 [dpdk-dev] [PATCH] examples/l3fwd: optimised packet processing on powerpc Gowrishankar
@ 2017-09-21 10:05 ` Gowrishankar
  2017-09-28  1:59   ` Chao Zhu
  0 siblings, 1 reply; 4+ messages in thread
From: Gowrishankar @ 2017-09-21 10:05 UTC (permalink / raw)
  To: dev; +Cc: Chao Zhu, Tomasz Kantecki, Gowrishankar Muthukrishnan

From: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>

This patch adds altivec support for lpm packet processing in powerpc.

Signed-off-by: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
---
v2:
 * coding style standards

 MAINTAINERS                        |   1 +
 examples/l3fwd/l3fwd_altivec.h     | 284 +++++++++++++++++++++++++++++++++++++
 examples/l3fwd/l3fwd_lpm.c         |   5 +-
 examples/l3fwd/l3fwd_lpm_altivec.h | 164 +++++++++++++++++++++
 4 files changed, 453 insertions(+), 1 deletion(-)
 create mode 100644 examples/l3fwd/l3fwd_altivec.h
 create mode 100644 examples/l3fwd/l3fwd_lpm_altivec.h

diff --git a/MAINTAINERS b/MAINTAINERS
index a0cd75e..0010b6c 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -168,6 +168,7 @@ M: Chao Zhu <chaozhu@linux.vnet.ibm.com>
 F: lib/librte_eal/common/arch/ppc_64/
 F: lib/librte_eal/common/include/arch/ppc_64/
 F: drivers/net/i40e/i40e_rxtx_vec_altivec.c
+F: examples/l3fwd/*altivec.h
 
 Intel x86
 M: Bruce Richardson <bruce.richardson@intel.com>
diff --git a/examples/l3fwd/l3fwd_altivec.h b/examples/l3fwd/l3fwd_altivec.h
new file mode 100644
index 0000000..29c2627
--- /dev/null
+++ b/examples/l3fwd/l3fwd_altivec.h
@@ -0,0 +1,284 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2016 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2017 IBM Corporation.
+ *   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.
+ */
+
+
+#ifndef _L3FWD_ALTIVEC_H_
+#define _L3FWD_ALTIVEC_H_
+
+#include "l3fwd.h"
+#include "l3fwd_common.h"
+
+/*
+ * Update source and destination MAC addresses in the ethernet header.
+ * Perform RFC1812 checks and updates for IPV4 packets.
+ */
+static inline void
+processx4_step3(struct rte_mbuf *pkt[FWDSTEP], uint16_t dst_port[FWDSTEP])
+{
+	vector unsigned int te[FWDSTEP];
+	vector unsigned int ve[FWDSTEP];
+	vector unsigned int *p[FWDSTEP];
+
+	p[0] = rte_pktmbuf_mtod(pkt[0], vector unsigned int *);
+	p[1] = rte_pktmbuf_mtod(pkt[1], vector unsigned int *);
+	p[2] = rte_pktmbuf_mtod(pkt[2], vector unsigned int *);
+	p[3] = rte_pktmbuf_mtod(pkt[3], vector unsigned int *);
+
+	ve[0] = (vector unsigned int)val_eth[dst_port[0]];
+	te[0] = *p[0];
+
+	ve[1] = (vector unsigned int)val_eth[dst_port[1]];
+	te[1] = *p[1];
+
+	ve[2] = (vector unsigned int)val_eth[dst_port[2]];
+	te[2] = *p[2];
+
+	ve[3] = (vector unsigned int)val_eth[dst_port[3]];
+	te[3] = *p[3];
+
+	/* Update first 12 bytes, keep rest bytes intact. */
+	te[0] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[0],
+			(vector unsigned short)te[0],
+			(vector unsigned short) {0, 0, 0, 0,
+						0, 0, 0xffff, 0xffff});
+
+	te[1] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[1],
+			(vector unsigned short)te[1],
+			(vector unsigned short) {0, 0, 0, 0,
+						0, 0, 0xffff, 0xffff});
+
+	te[2] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[2],
+			(vector unsigned short)te[2],
+			(vector unsigned short) {0, 0, 0, 0, 0,
+						0, 0xffff, 0xffff});
+
+	te[3] = (vector unsigned int)vec_sel(
+			(vector unsigned short)ve[3],
+			(vector unsigned short)te[3],
+			(vector unsigned short) {0, 0, 0, 0,
+						0, 0, 0xffff, 0xffff});
+
+	*p[0] = te[0];
+	*p[1] = te[1];
+	*p[2] = te[2];
+	*p[3] = te[3];
+
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[0] + 1),
+		&dst_port[0], pkt[0]->packet_type);
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[1] + 1),
+		&dst_port[1], pkt[1]->packet_type);
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[2] + 1),
+		&dst_port[2], pkt[2]->packet_type);
+	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[3] + 1),
+		&dst_port[3], pkt[3]->packet_type);
+}
+
+/*
+ * Group consecutive packets with the same destination port in bursts of 4.
+ * Suppose we have array of destionation ports:
+ * dst_port[] = {a, b, c, d,, e, ... }
+ * dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
+ * We doing 4 comparisons at once and the result is 4 bit mask.
+ * This mask is used as an index into prebuild array of pnum values.
+ */
+static inline uint16_t *
+port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, vector unsigned short dp1,
+	vector unsigned short dp2)
+{
+	union {
+		uint16_t u16[FWDSTEP + 1];
+		uint64_t u64;
+	} *pnum = (void *)pn;
+
+	int32_t v;
+
+	v = vec_any_eq(dp1, dp2);
+
+
+	/* update last port counter. */
+	lp[0] += gptbl[v].lpv;
+
+	/* if dest port value has changed. */
+	if (v != GRPMSK) {
+		pnum->u64 = gptbl[v].pnum;
+		pnum->u16[FWDSTEP] = 1;
+		lp = pnum->u16 + gptbl[v].idx;
+	}
+
+	return lp;
+}
+
+/**
+ * Process one packet:
+ * Update source and destination MAC addresses in the ethernet header.
+ * Perform RFC1812 checks and updates for IPV4 packets.
+ */
+static inline void
+process_packet(struct rte_mbuf *pkt, uint16_t *dst_port)
+{
+	struct ether_hdr *eth_hdr;
+	vector unsigned int te, ve;
+
+	eth_hdr = rte_pktmbuf_mtod(pkt, struct ether_hdr *);
+
+	te = *(vector unsigned int *)eth_hdr;
+	ve = (vector unsigned int)val_eth[dst_port[0]];
+
+	rfc1812_process((struct ipv4_hdr *)(eth_hdr + 1), dst_port,
+			pkt->packet_type);
+
+	/* dynamically vec_sel te and ve for MASK_ETH (0x3f) */
+	te = (vector unsigned int)vec_sel(
+		(vector unsigned short)ve,
+		(vector unsigned short)te,
+		(vector unsigned short){0, 0, 0, 0,
+					0, 0, 0xffff, 0xffff});
+
+	*(vector unsigned int *)eth_hdr = te;
+}
+
+/**
+ * Send packets burst from pkts_burst to the ports in dst_port array
+ */
+static __rte_always_inline void
+send_packets_multi(struct lcore_conf *qconf, struct rte_mbuf **pkts_burst,
+		uint16_t dst_port[MAX_PKT_BURST], int nb_rx)
+{
+	int32_t k;
+	int j = 0;
+	uint16_t dlp;
+	uint16_t *lp;
+	uint16_t pnum[MAX_PKT_BURST + 1];
+
+	/*
+	 * Finish packet processing and group consecutive
+	 * packets with the same destination port.
+	 */
+	k = RTE_ALIGN_FLOOR(nb_rx, FWDSTEP);
+	if (k != 0) {
+		vector unsigned short dp1, dp2;
+
+		lp = pnum;
+		lp[0] = 1;
+
+		processx4_step3(pkts_burst, dst_port);
+
+		/* dp1: <d[0], d[1], d[2], d[3], ... > */
+		dp1 = *(vector unsigned short *)dst_port;
+
+		for (j = FWDSTEP; j != k; j += FWDSTEP) {
+			processx4_step3(&pkts_burst[j], &dst_port[j]);
+
+			/*
+			 * dp2:
+			 * <d[j-3], d[j-2], d[j-1], d[j], ... >
+			 */
+			dp2 = *((vector unsigned short *)
+					&dst_port[j - FWDSTEP + 1]);
+			lp  = port_groupx4(&pnum[j - FWDSTEP], lp, dp1, dp2);
+
+			/*
+			 * dp1:
+			 * <d[j], d[j+1], d[j+2], d[j+3], ... >
+			 */
+			dp1 = vec_sro(dp2, (vector unsigned char) {
+				0, 0, 0, 0, 0, 0, 0, 0,
+				0, 0, 0, (FWDSTEP - 1) * sizeof(dst_port[0])});
+		}
+
+		/*
+		 * dp2: <d[j-3], d[j-2], d[j-1], d[j-1], ... >
+		 */
+		dp2 = vec_perm(dp1, (vector unsigned short){},
+				(vector unsigned char){0xf9});
+		lp  = port_groupx4(&pnum[j - FWDSTEP], lp, dp1, dp2);
+
+		/*
+		 * remove values added by the last repeated
+		 * dst port.
+		 */
+		lp[0]--;
+		dlp = dst_port[j - 1];
+	} else {
+		/* set dlp and lp to the never used values. */
+		dlp = BAD_PORT - 1;
+		lp = pnum + MAX_PKT_BURST;
+	}
+
+	/* Process up to last 3 packets one by one. */
+	switch (nb_rx % FWDSTEP) {
+	case 3:
+		process_packet(pkts_burst[j], dst_port + j);
+		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
+		j++;
+		/* fall-through */
+	case 2:
+		process_packet(pkts_burst[j], dst_port + j);
+		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
+		j++;
+		/* fall-through */
+	case 1:
+		process_packet(pkts_burst[j], dst_port + j);
+		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
+		j++;
+	}
+
+	/*
+	 * Send packets out, through destination port.
+	 * Consecutive packets with the same destination port
+	 * are already grouped together.
+	 * If destination port for the packet equals BAD_PORT,
+	 * then free the packet without sending it out.
+	 */
+	for (j = 0; j < nb_rx; j += k) {
+
+		int32_t m;
+		uint16_t pn;
+
+		pn = dst_port[j];
+		k = pnum[j];
+
+		if (likely(pn != BAD_PORT))
+			send_packetsx4(qconf, pn, pkts_burst + j, k);
+		else
+			for (m = j; m != j + k; m++)
+				rte_pktmbuf_free(pkts_burst[m]);
+
+	}
+}
+
+#endif /* _L3FWD_ALTIVEC_H_ */
diff --git a/examples/l3fwd/l3fwd_lpm.c b/examples/l3fwd/l3fwd_lpm.c
index ff1e403..b508886 100644
--- a/examples/l3fwd/l3fwd_lpm.c
+++ b/examples/l3fwd/l3fwd_lpm.c
@@ -191,6 +191,8 @@ struct ipv6_l3fwd_lpm_route {
 #include "l3fwd_lpm_sse.h"
 #elif defined RTE_MACHINE_CPUFLAG_NEON
 #include "l3fwd_lpm_neon.h"
+#elif defined(RTE_ARCH_PPC_64)
+#include "l3fwd_lpm_altivec.h"
 #else
 #include "l3fwd_lpm.h"
 #endif
@@ -263,7 +265,8 @@ struct ipv6_l3fwd_lpm_route {
 			if (nb_rx == 0)
 				continue;
 
-#if defined RTE_ARCH_X86 || defined RTE_MACHINE_CPUFLAG_NEON
+#if defined RTE_ARCH_X86 || defined RTE_MACHINE_CPUFLAG_NEON \
+			 || defined RTE_ARCH_PPC_64
 			l3fwd_lpm_send_packets(nb_rx, pkts_burst,
 						portid, qconf);
 #else
diff --git a/examples/l3fwd/l3fwd_lpm_altivec.h b/examples/l3fwd/l3fwd_lpm_altivec.h
new file mode 100644
index 0000000..36ca983
--- /dev/null
+++ b/examples/l3fwd/l3fwd_lpm_altivec.h
@@ -0,0 +1,164 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
+ *   Copyright(c) 2017 IBM Corporation.
+ *   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.
+ */
+
+#ifndef __L3FWD_LPM_ALTIVEC_H__
+#define __L3FWD_LPM_ALTIVEC_H__
+
+#include "l3fwd_altivec.h"
+
+/*
+ * Read packet_type and destination IPV4 addresses from 4 mbufs.
+ */
+static inline void
+processx4_step1(struct rte_mbuf *pkt[FWDSTEP],
+		vector unsigned int *dip,
+		uint32_t *ipv4_flag)
+{
+	struct ipv4_hdr *ipv4_hdr;
+	struct ether_hdr *eth_hdr;
+	uint32_t x0, x1, x2, x3;
+
+	eth_hdr = rte_pktmbuf_mtod(pkt[0], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x0 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] = pkt[0]->packet_type & RTE_PTYPE_L3_IPV4;
+
+	rte_compiler_barrier();
+	eth_hdr = rte_pktmbuf_mtod(pkt[1], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x1 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] &= pkt[1]->packet_type;
+
+	rte_compiler_barrier();
+	eth_hdr = rte_pktmbuf_mtod(pkt[2], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x2 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] &= pkt[2]->packet_type;
+
+	rte_compiler_barrier();
+	eth_hdr = rte_pktmbuf_mtod(pkt[3], struct ether_hdr *);
+	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
+	x3 = ipv4_hdr->dst_addr;
+	ipv4_flag[0] &= pkt[3]->packet_type;
+
+	rte_compiler_barrier();
+	dip[0] = (vector unsigned int){x0, x1, x2, x3};
+}
+
+/*
+ * Lookup into LPM for destination port.
+ * If lookup fails, use incoming port (portid) as destination port.
+ */
+static inline void
+processx4_step2(const struct lcore_conf *qconf,
+		vector unsigned int dip,
+		uint32_t ipv4_flag,
+		uint8_t portid,
+		struct rte_mbuf *pkt[FWDSTEP],
+		uint16_t dprt[FWDSTEP])
+{
+	rte_xmm_t dst;
+	const vector unsigned char bswap_mask = (vector unsigned char){
+							3, 2, 1, 0,
+							7, 6, 5, 4,
+							11, 10, 9, 8,
+							15, 14, 13, 12};
+
+	/* Byte swap 4 IPV4 addresses. */
+	dip = (vector unsigned int)vec_perm(*(vector unsigned char *)&dip,
+					(vector unsigned char){}, bswap_mask);
+
+	/* if all 4 packets are IPV4. */
+	if (likely(ipv4_flag)) {
+		rte_lpm_lookupx4(qconf->ipv4_lookup_struct, (xmm_t)dip,
+			(uint32_t *)&dst, portid);
+		/* get rid of unused upper 16 bit for each dport. */
+		dst.x = (xmm_t)vec_packs(dst.x, dst.x);
+		*(uint64_t *)dprt = dst.u64[0];
+	} else {
+		dst.x = (xmm_t)dip;
+		dprt[0] = lpm_get_dst_port_with_ipv4(qconf, pkt[0],
+							dst.u32[0], portid);
+		dprt[1] = lpm_get_dst_port_with_ipv4(qconf, pkt[1],
+							dst.u32[1], portid);
+		dprt[2] = lpm_get_dst_port_with_ipv4(qconf, pkt[2],
+							dst.u32[2], portid);
+		dprt[3] = lpm_get_dst_port_with_ipv4(qconf, pkt[3],
+							dst.u32[3], portid);
+	}
+}
+
+/*
+ * Buffer optimized handling of packets, invoked
+ * from main_loop.
+ */
+static inline void
+l3fwd_lpm_send_packets(int nb_rx, struct rte_mbuf **pkts_burst,
+			uint8_t portid, struct lcore_conf *qconf)
+{
+	int32_t j;
+	uint16_t dst_port[MAX_PKT_BURST];
+	vector unsigned int dip[MAX_PKT_BURST / FWDSTEP];
+	uint32_t ipv4_flag[MAX_PKT_BURST / FWDSTEP];
+	const int32_t k = RTE_ALIGN_FLOOR(nb_rx, FWDSTEP);
+
+	for (j = 0; j != k; j += FWDSTEP)
+		processx4_step1(&pkts_burst[j], &dip[j / FWDSTEP],
+				&ipv4_flag[j / FWDSTEP]);
+
+	for (j = 0; j != k; j += FWDSTEP)
+		processx4_step2(qconf, dip[j / FWDSTEP],
+				ipv4_flag[j / FWDSTEP],
+				portid, &pkts_burst[j], &dst_port[j]);
+
+	/* Classify last up to 3 packets one by one */
+	switch (nb_rx % FWDSTEP) {
+	case 3:
+		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j], portid);
+		j++;
+		/* fall-through */
+	case 2:
+		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j], portid);
+		j++;
+		/* fall-through */
+	case 1:
+		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j], portid);
+		j++;
+		/* fall-through */
+	}
+
+	send_packets_multi(qconf, pkts_burst, dst_port, nb_rx);
+}
+
+#endif /* __L3FWD_LPM_ALTIVEC_H__ */
-- 
1.9.1

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH v2] examples/l3fwd: optimised packet processing on powerpc
  2017-09-21 10:05 ` [dpdk-dev] [PATCH v2] " Gowrishankar
@ 2017-09-28  1:59   ` Chao Zhu
  2017-10-13 22:52     ` Thomas Monjalon
  0 siblings, 1 reply; 4+ messages in thread
From: Chao Zhu @ 2017-09-28  1:59 UTC (permalink / raw)
  To: 'Gowrishankar', dev; +Cc: 'Tomasz Kantecki'

> -----Original Message-----
> From: Gowrishankar [mailto:gowrishankar.m@linux.vnet.ibm.com]
> Sent: 2017年9月21日 18:05
> To: dev@dpdk.org
> Cc: Chao Zhu <chaozhu@linux.vnet.ibm.com>; Tomasz Kantecki
> <tomasz.kantecki@intel.com>; Gowrishankar Muthukrishnan
> <gowrishankar.m@linux.vnet.ibm.com>
> Subject: [PATCH v2] examples/l3fwd: optimised packet processing on powerpc
> 
> From: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
> 
> This patch adds altivec support for lpm packet processing in powerpc.
> 
> Signed-off-by: Gowrishankar Muthukrishnan
> <gowrishankar.m@linux.vnet.ibm.com>
> ---
> v2:
>  * coding style standards
> 
>  MAINTAINERS                        |   1 +
>  examples/l3fwd/l3fwd_altivec.h     | 284
> +++++++++++++++++++++++++++++++++++++
>  examples/l3fwd/l3fwd_lpm.c         |   5 +-
>  examples/l3fwd/l3fwd_lpm_altivec.h | 164 +++++++++++++++++++++
>  4 files changed, 453 insertions(+), 1 deletion(-)  create mode 100644
> examples/l3fwd/l3fwd_altivec.h  create mode 100644
> examples/l3fwd/l3fwd_lpm_altivec.h
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index a0cd75e..0010b6c 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -168,6 +168,7 @@ M: Chao Zhu <chaozhu@linux.vnet.ibm.com>
>  F: lib/librte_eal/common/arch/ppc_64/
>  F: lib/librte_eal/common/include/arch/ppc_64/
>  F: drivers/net/i40e/i40e_rxtx_vec_altivec.c
> +F: examples/l3fwd/*altivec.h
> 
>  Intel x86
>  M: Bruce Richardson <bruce.richardson@intel.com> diff --git
> a/examples/l3fwd/l3fwd_altivec.h b/examples/l3fwd/l3fwd_altivec.h new file
> mode 100644 index 0000000..29c2627
> --- /dev/null
> +++ b/examples/l3fwd/l3fwd_altivec.h
> @@ -0,0 +1,284 @@
> +/*-
> + *   BSD LICENSE
> + *
> + *   Copyright(c) 2016 Intel Corporation. All rights reserved.
> + *   Copyright(c) 2017 IBM Corporation.
> + *   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.
> + */
> +
> +
> +#ifndef _L3FWD_ALTIVEC_H_
> +#define _L3FWD_ALTIVEC_H_
> +
> +#include "l3fwd.h"
> +#include "l3fwd_common.h"
> +
> +/*
> + * Update source and destination MAC addresses in the ethernet header.
> + * Perform RFC1812 checks and updates for IPV4 packets.
> + */
> +static inline void
> +processx4_step3(struct rte_mbuf *pkt[FWDSTEP], uint16_t
> +dst_port[FWDSTEP]) {
> +	vector unsigned int te[FWDSTEP];
> +	vector unsigned int ve[FWDSTEP];
> +	vector unsigned int *p[FWDSTEP];
> +
> +	p[0] = rte_pktmbuf_mtod(pkt[0], vector unsigned int *);
> +	p[1] = rte_pktmbuf_mtod(pkt[1], vector unsigned int *);
> +	p[2] = rte_pktmbuf_mtod(pkt[2], vector unsigned int *);
> +	p[3] = rte_pktmbuf_mtod(pkt[3], vector unsigned int *);
> +
> +	ve[0] = (vector unsigned int)val_eth[dst_port[0]];
> +	te[0] = *p[0];
> +
> +	ve[1] = (vector unsigned int)val_eth[dst_port[1]];
> +	te[1] = *p[1];
> +
> +	ve[2] = (vector unsigned int)val_eth[dst_port[2]];
> +	te[2] = *p[2];
> +
> +	ve[3] = (vector unsigned int)val_eth[dst_port[3]];
> +	te[3] = *p[3];
> +
> +	/* Update first 12 bytes, keep rest bytes intact. */
> +	te[0] = (vector unsigned int)vec_sel(
> +			(vector unsigned short)ve[0],
> +			(vector unsigned short)te[0],
> +			(vector unsigned short) {0, 0, 0, 0,
> +						0, 0, 0xffff, 0xffff});
> +
> +	te[1] = (vector unsigned int)vec_sel(
> +			(vector unsigned short)ve[1],
> +			(vector unsigned short)te[1],
> +			(vector unsigned short) {0, 0, 0, 0,
> +						0, 0, 0xffff, 0xffff});
> +
> +	te[2] = (vector unsigned int)vec_sel(
> +			(vector unsigned short)ve[2],
> +			(vector unsigned short)te[2],
> +			(vector unsigned short) {0, 0, 0, 0, 0,
> +						0, 0xffff, 0xffff});
> +
> +	te[3] = (vector unsigned int)vec_sel(
> +			(vector unsigned short)ve[3],
> +			(vector unsigned short)te[3],
> +			(vector unsigned short) {0, 0, 0, 0,
> +						0, 0, 0xffff, 0xffff});
> +
> +	*p[0] = te[0];
> +	*p[1] = te[1];
> +	*p[2] = te[2];
> +	*p[3] = te[3];
> +
> +	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[0] + 1),
> +		&dst_port[0], pkt[0]->packet_type);
> +	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[1] + 1),
> +		&dst_port[1], pkt[1]->packet_type);
> +	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[2] + 1),
> +		&dst_port[2], pkt[2]->packet_type);
> +	rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[3] + 1),
> +		&dst_port[3], pkt[3]->packet_type);
> +}
> +
> +/*
> + * Group consecutive packets with the same destination port in bursts of
4.
> + * Suppose we have array of destionation ports:
> + * dst_port[] = {a, b, c, d,, e, ... }
> + * dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
> + * We doing 4 comparisons at once and the result is 4 bit mask.
> + * This mask is used as an index into prebuild array of pnum values.
> + */
> +static inline uint16_t *
> +port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, vector unsigned
short
> dp1,
> +	vector unsigned short dp2)
> +{
> +	union {
> +		uint16_t u16[FWDSTEP + 1];
> +		uint64_t u64;
> +	} *pnum = (void *)pn;
> +
> +	int32_t v;
> +
> +	v = vec_any_eq(dp1, dp2);
> +
> +
> +	/* update last port counter. */
> +	lp[0] += gptbl[v].lpv;
> +
> +	/* if dest port value has changed. */
> +	if (v != GRPMSK) {
> +		pnum->u64 = gptbl[v].pnum;
> +		pnum->u16[FWDSTEP] = 1;
> +		lp = pnum->u16 + gptbl[v].idx;
> +	}
> +
> +	return lp;
> +}
> +
> +/**
> + * Process one packet:
> + * Update source and destination MAC addresses in the ethernet header.
> + * Perform RFC1812 checks and updates for IPV4 packets.
> + */
> +static inline void
> +process_packet(struct rte_mbuf *pkt, uint16_t *dst_port) {
> +	struct ether_hdr *eth_hdr;
> +	vector unsigned int te, ve;
> +
> +	eth_hdr = rte_pktmbuf_mtod(pkt, struct ether_hdr *);
> +
> +	te = *(vector unsigned int *)eth_hdr;
> +	ve = (vector unsigned int)val_eth[dst_port[0]];
> +
> +	rfc1812_process((struct ipv4_hdr *)(eth_hdr + 1), dst_port,
> +			pkt->packet_type);
> +
> +	/* dynamically vec_sel te and ve for MASK_ETH (0x3f) */
> +	te = (vector unsigned int)vec_sel(
> +		(vector unsigned short)ve,
> +		(vector unsigned short)te,
> +		(vector unsigned short){0, 0, 0, 0,
> +					0, 0, 0xffff, 0xffff});
> +
> +	*(vector unsigned int *)eth_hdr = te;
> +}
> +
> +/**
> + * Send packets burst from pkts_burst to the ports in dst_port array
> +*/ static __rte_always_inline void send_packets_multi(struct lcore_conf
> +*qconf, struct rte_mbuf **pkts_burst,
> +		uint16_t dst_port[MAX_PKT_BURST], int nb_rx) {
> +	int32_t k;
> +	int j = 0;
> +	uint16_t dlp;
> +	uint16_t *lp;
> +	uint16_t pnum[MAX_PKT_BURST + 1];
> +
> +	/*
> +	 * Finish packet processing and group consecutive
> +	 * packets with the same destination port.
> +	 */
> +	k = RTE_ALIGN_FLOOR(nb_rx, FWDSTEP);
> +	if (k != 0) {
> +		vector unsigned short dp1, dp2;
> +
> +		lp = pnum;
> +		lp[0] = 1;
> +
> +		processx4_step3(pkts_burst, dst_port);
> +
> +		/* dp1: <d[0], d[1], d[2], d[3], ... > */
> +		dp1 = *(vector unsigned short *)dst_port;
> +
> +		for (j = FWDSTEP; j != k; j += FWDSTEP) {
> +			processx4_step3(&pkts_burst[j], &dst_port[j]);
> +
> +			/*
> +			 * dp2:
> +			 * <d[j-3], d[j-2], d[j-1], d[j], ... >
> +			 */
> +			dp2 = *((vector unsigned short *)
> +					&dst_port[j - FWDSTEP + 1]);
> +			lp  = port_groupx4(&pnum[j - FWDSTEP], lp, dp1,
dp2);
> +
> +			/*
> +			 * dp1:
> +			 * <d[j], d[j+1], d[j+2], d[j+3], ... >
> +			 */
> +			dp1 = vec_sro(dp2, (vector unsigned char) {
> +				0, 0, 0, 0, 0, 0, 0, 0,
> +				0, 0, 0, (FWDSTEP - 1) *
sizeof(dst_port[0])});
> +		}
> +
> +		/*
> +		 * dp2: <d[j-3], d[j-2], d[j-1], d[j-1], ... >
> +		 */
> +		dp2 = vec_perm(dp1, (vector unsigned short){},
> +				(vector unsigned char){0xf9});
> +		lp  = port_groupx4(&pnum[j - FWDSTEP], lp, dp1, dp2);
> +
> +		/*
> +		 * remove values added by the last repeated
> +		 * dst port.
> +		 */
> +		lp[0]--;
> +		dlp = dst_port[j - 1];
> +	} else {
> +		/* set dlp and lp to the never used values. */
> +		dlp = BAD_PORT - 1;
> +		lp = pnum + MAX_PKT_BURST;
> +	}
> +
> +	/* Process up to last 3 packets one by one. */
> +	switch (nb_rx % FWDSTEP) {
> +	case 3:
> +		process_packet(pkts_burst[j], dst_port + j);
> +		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
> +		j++;
> +		/* fall-through */
> +	case 2:
> +		process_packet(pkts_burst[j], dst_port + j);
> +		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
> +		j++;
> +		/* fall-through */
> +	case 1:
> +		process_packet(pkts_burst[j], dst_port + j);
> +		GROUP_PORT_STEP(dlp, dst_port, lp, pnum, j);
> +		j++;
> +	}
> +
> +	/*
> +	 * Send packets out, through destination port.
> +	 * Consecutive packets with the same destination port
> +	 * are already grouped together.
> +	 * If destination port for the packet equals BAD_PORT,
> +	 * then free the packet without sending it out.
> +	 */
> +	for (j = 0; j < nb_rx; j += k) {
> +
> +		int32_t m;
> +		uint16_t pn;
> +
> +		pn = dst_port[j];
> +		k = pnum[j];
> +
> +		if (likely(pn != BAD_PORT))
> +			send_packetsx4(qconf, pn, pkts_burst + j, k);
> +		else
> +			for (m = j; m != j + k; m++)
> +				rte_pktmbuf_free(pkts_burst[m]);
> +
> +	}
> +}
> +
> +#endif /* _L3FWD_ALTIVEC_H_ */
> diff --git a/examples/l3fwd/l3fwd_lpm.c b/examples/l3fwd/l3fwd_lpm.c index
> ff1e403..b508886 100644
> --- a/examples/l3fwd/l3fwd_lpm.c
> +++ b/examples/l3fwd/l3fwd_lpm.c
> @@ -191,6 +191,8 @@ struct ipv6_l3fwd_lpm_route {  #include
> "l3fwd_lpm_sse.h"
>  #elif defined RTE_MACHINE_CPUFLAG_NEON
>  #include "l3fwd_lpm_neon.h"
> +#elif defined(RTE_ARCH_PPC_64)
> +#include "l3fwd_lpm_altivec.h"
>  #else
>  #include "l3fwd_lpm.h"
>  #endif
> @@ -263,7 +265,8 @@ struct ipv6_l3fwd_lpm_route {
>  			if (nb_rx == 0)
>  				continue;
> 
> -#if defined RTE_ARCH_X86 || defined RTE_MACHINE_CPUFLAG_NEON
> +#if defined RTE_ARCH_X86 || defined RTE_MACHINE_CPUFLAG_NEON \
> +			 || defined RTE_ARCH_PPC_64
>  			l3fwd_lpm_send_packets(nb_rx, pkts_burst,
>  						portid, qconf);
>  #else
> diff --git a/examples/l3fwd/l3fwd_lpm_altivec.h
> b/examples/l3fwd/l3fwd_lpm_altivec.h
> new file mode 100644
> index 0000000..36ca983
> --- /dev/null
> +++ b/examples/l3fwd/l3fwd_lpm_altivec.h
> @@ -0,0 +1,164 @@
> +/*-
> + *   BSD LICENSE
> + *
> + *   Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> + *   Copyright(c) 2017 IBM Corporation.
> + *   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.
> + */
> +
> +#ifndef __L3FWD_LPM_ALTIVEC_H__
> +#define __L3FWD_LPM_ALTIVEC_H__
> +
> +#include "l3fwd_altivec.h"
> +
> +/*
> + * Read packet_type and destination IPV4 addresses from 4 mbufs.
> + */
> +static inline void
> +processx4_step1(struct rte_mbuf *pkt[FWDSTEP],
> +		vector unsigned int *dip,
> +		uint32_t *ipv4_flag)
> +{
> +	struct ipv4_hdr *ipv4_hdr;
> +	struct ether_hdr *eth_hdr;
> +	uint32_t x0, x1, x2, x3;
> +
> +	eth_hdr = rte_pktmbuf_mtod(pkt[0], struct ether_hdr *);
> +	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
> +	x0 = ipv4_hdr->dst_addr;
> +	ipv4_flag[0] = pkt[0]->packet_type & RTE_PTYPE_L3_IPV4;
> +
> +	rte_compiler_barrier();
> +	eth_hdr = rte_pktmbuf_mtod(pkt[1], struct ether_hdr *);
> +	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
> +	x1 = ipv4_hdr->dst_addr;
> +	ipv4_flag[0] &= pkt[1]->packet_type;
> +
> +	rte_compiler_barrier();
> +	eth_hdr = rte_pktmbuf_mtod(pkt[2], struct ether_hdr *);
> +	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
> +	x2 = ipv4_hdr->dst_addr;
> +	ipv4_flag[0] &= pkt[2]->packet_type;
> +
> +	rte_compiler_barrier();
> +	eth_hdr = rte_pktmbuf_mtod(pkt[3], struct ether_hdr *);
> +	ipv4_hdr = (struct ipv4_hdr *)(eth_hdr + 1);
> +	x3 = ipv4_hdr->dst_addr;
> +	ipv4_flag[0] &= pkt[3]->packet_type;
> +
> +	rte_compiler_barrier();
> +	dip[0] = (vector unsigned int){x0, x1, x2, x3}; }
> +
> +/*
> + * Lookup into LPM for destination port.
> + * If lookup fails, use incoming port (portid) as destination port.
> + */
> +static inline void
> +processx4_step2(const struct lcore_conf *qconf,
> +		vector unsigned int dip,
> +		uint32_t ipv4_flag,
> +		uint8_t portid,
> +		struct rte_mbuf *pkt[FWDSTEP],
> +		uint16_t dprt[FWDSTEP])
> +{
> +	rte_xmm_t dst;
> +	const vector unsigned char bswap_mask = (vector unsigned char){
> +							3, 2, 1, 0,
> +							7, 6, 5, 4,
> +							11, 10, 9, 8,
> +							15, 14, 13, 12};
> +
> +	/* Byte swap 4 IPV4 addresses. */
> +	dip = (vector unsigned int)vec_perm(*(vector unsigned char *)&dip,
> +					(vector unsigned char){},
bswap_mask);
> +
> +	/* if all 4 packets are IPV4. */
> +	if (likely(ipv4_flag)) {
> +		rte_lpm_lookupx4(qconf->ipv4_lookup_struct, (xmm_t)dip,
> +			(uint32_t *)&dst, portid);
> +		/* get rid of unused upper 16 bit for each dport. */
> +		dst.x = (xmm_t)vec_packs(dst.x, dst.x);
> +		*(uint64_t *)dprt = dst.u64[0];
> +	} else {
> +		dst.x = (xmm_t)dip;
> +		dprt[0] = lpm_get_dst_port_with_ipv4(qconf, pkt[0],
> +							dst.u32[0], portid);
> +		dprt[1] = lpm_get_dst_port_with_ipv4(qconf, pkt[1],
> +							dst.u32[1], portid);
> +		dprt[2] = lpm_get_dst_port_with_ipv4(qconf, pkt[2],
> +							dst.u32[2], portid);
> +		dprt[3] = lpm_get_dst_port_with_ipv4(qconf, pkt[3],
> +							dst.u32[3], portid);
> +	}
> +}
> +
> +/*
> + * Buffer optimized handling of packets, invoked
> + * from main_loop.
> + */
> +static inline void
> +l3fwd_lpm_send_packets(int nb_rx, struct rte_mbuf **pkts_burst,
> +			uint8_t portid, struct lcore_conf *qconf) {
> +	int32_t j;
> +	uint16_t dst_port[MAX_PKT_BURST];
> +	vector unsigned int dip[MAX_PKT_BURST / FWDSTEP];
> +	uint32_t ipv4_flag[MAX_PKT_BURST / FWDSTEP];
> +	const int32_t k = RTE_ALIGN_FLOOR(nb_rx, FWDSTEP);
> +
> +	for (j = 0; j != k; j += FWDSTEP)
> +		processx4_step1(&pkts_burst[j], &dip[j / FWDSTEP],
> +				&ipv4_flag[j / FWDSTEP]);
> +
> +	for (j = 0; j != k; j += FWDSTEP)
> +		processx4_step2(qconf, dip[j / FWDSTEP],
> +				ipv4_flag[j / FWDSTEP],
> +				portid, &pkts_burst[j], &dst_port[j]);
> +
> +	/* Classify last up to 3 packets one by one */
> +	switch (nb_rx % FWDSTEP) {
> +	case 3:
> +		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j],
portid);
> +		j++;
> +		/* fall-through */
> +	case 2:
> +		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j],
portid);
> +		j++;
> +		/* fall-through */
> +	case 1:
> +		dst_port[j] = lpm_get_dst_port(qconf, pkts_burst[j],
portid);
> +		j++;
> +		/* fall-through */
> +	}
> +
> +	send_packets_multi(qconf, pkts_burst, dst_port, nb_rx); }
> +
> +#endif /* __L3FWD_LPM_ALTIVEC_H__ */
> --
> 1.9.1
Acked-by: Chao Zhu <chaozhu@linux.vnet.ibm.com>
>

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [dpdk-dev] [PATCH v2] examples/l3fwd: optimised packet processing on powerpc
  2017-09-28  1:59   ` Chao Zhu
@ 2017-10-13 22:52     ` Thomas Monjalon
  0 siblings, 0 replies; 4+ messages in thread
From: Thomas Monjalon @ 2017-10-13 22:52 UTC (permalink / raw)
  To: 'Gowrishankar'; +Cc: dev, Chao Zhu, 'Tomasz Kantecki'

> > From: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
> > 
> > This patch adds altivec support for lpm packet processing in powerpc.
> > 
> > Signed-off-by: Gowrishankar Muthukrishnan <gowrishankar.m@linux.vnet.ibm.com>
> Acked-by: Chao Zhu <chaozhu@linux.vnet.ibm.com>

Applied, thanks

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2017-10-13 22:52 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-09-21  8:37 [dpdk-dev] [PATCH] examples/l3fwd: optimised packet processing on powerpc Gowrishankar
2017-09-21 10:05 ` [dpdk-dev] [PATCH v2] " Gowrishankar
2017-09-28  1:59   ` Chao Zhu
2017-10-13 22:52     ` Thomas Monjalon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).