[ceph.git] / ceph / src / spdk / dpdk / lib / librte_node / ip4_lookup_sse.h

/* SPDX-License-Identifier: BSD-3-Clause
 * Copyright(C) 2020 Marvell International Ltd.
 */

#ifndef __INCLUDE_IP4_LOOKUP_SSE_H__
#define __INCLUDE_IP4_LOOKUP_SSE_H__

/* X86 SSE */
static uint16_t
ip4_lookup_node_process(struct rte_graph *graph, struct rte_node *node,
			void **objs, uint16_t nb_objs)
{
	struct rte_mbuf *mbuf0, *mbuf1, *mbuf2, *mbuf3, **pkts;
	rte_edge_t next0, next1, next2, next3, next_index;
	struct rte_ipv4_hdr *ipv4_hdr;
	uint32_t ip0, ip1, ip2, ip3;
	void **to_next, **from;
	uint16_t last_spec = 0;
	uint16_t n_left_from;
	struct rte_lpm *lpm;
	uint16_t held = 0;
	uint32_t drop_nh;
	rte_xmm_t dst;
	__m128i dip; /* SSE register */
	int rc, i;

	/* Speculative next */
	next_index = RTE_NODE_IP4_LOOKUP_NEXT_REWRITE;
	/* Drop node */
	drop_nh = ((uint32_t)RTE_NODE_IP4_LOOKUP_NEXT_PKT_DROP) << 16;

	/* Get socket specific LPM from ctx */
	lpm = *((struct rte_lpm **)node->ctx);

	pkts = (struct rte_mbuf **)objs;
	from = objs;
	n_left_from = nb_objs;

	if (n_left_from >= 4) {
		for (i = 0; i < 4; i++)
			rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[i], void *,
						sizeof(struct rte_ether_hdr)));
	}

	/* Get stream for the speculated next node */
	to_next = rte_node_next_stream_get(graph, node, next_index, nb_objs);
	while (n_left_from >= 4) {
		/* Prefetch next-next mbufs */
		if (likely(n_left_from > 11)) {
			rte_prefetch0(pkts[8]);
			rte_prefetch0(pkts[9]);
			rte_prefetch0(pkts[10]);
			rte_prefetch0(pkts[11]);
		}

		/* Prefetch next mbuf data */
		if (likely(n_left_from > 7)) {
			rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[4], void *,
						sizeof(struct rte_ether_hdr)));
			rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[5], void *,
						sizeof(struct rte_ether_hdr)));
			rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[6], void *,
						sizeof(struct rte_ether_hdr)));
			rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[7], void *,
						sizeof(struct rte_ether_hdr)));
		}

		mbuf0 = pkts[0];
		mbuf1 = pkts[1];
		mbuf2 = pkts[2];
		mbuf3 = pkts[3];

		pkts += 4;
		n_left_from -= 4;

		/* Extract DIP of mbuf0 */
		ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf0, struct rte_ipv4_hdr *,
						sizeof(struct rte_ether_hdr));
		ip0 = ipv4_hdr->dst_addr;
		/* Extract cksum, ttl as ipv4 hdr is in cache */
		node_mbuf_priv1(mbuf0)->cksum = ipv4_hdr->hdr_checksum;
		node_mbuf_priv1(mbuf0)->ttl = ipv4_hdr->time_to_live;

		/* Extract DIP of mbuf1 */
		ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf1, struct rte_ipv4_hdr *,
						sizeof(struct rte_ether_hdr));
		ip1 = ipv4_hdr->dst_addr;
		/* Extract cksum, ttl as ipv4 hdr is in cache */
		node_mbuf_priv1(mbuf1)->cksum = ipv4_hdr->hdr_checksum;
		node_mbuf_priv1(mbuf1)->ttl = ipv4_hdr->time_to_live;

		/* Extract DIP of mbuf2 */
		ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf2, struct rte_ipv4_hdr *,
						sizeof(struct rte_ether_hdr));
		ip2 = ipv4_hdr->dst_addr;
		/* Extract cksum, ttl as ipv4 hdr is in cache */
		node_mbuf_priv1(mbuf2)->cksum = ipv4_hdr->hdr_checksum;
		node_mbuf_priv1(mbuf2)->ttl = ipv4_hdr->time_to_live;

		/* Extract DIP of mbuf3 */
		ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf3, struct rte_ipv4_hdr *,
						sizeof(struct rte_ether_hdr));
		ip3 = ipv4_hdr->dst_addr;

		/* Prepare for lookup x4 */
		dip = _mm_set_epi32(ip3, ip2, ip1, ip0);

		/* Byte swap 4 IPV4 addresses. */
		const __m128i bswap_mask = _mm_set_epi8(
			12, 13, 14, 15, 8, 9, 10, 11, 4, 5, 6, 7, 0, 1, 2, 3);
		dip = _mm_shuffle_epi8(dip, bswap_mask);

		/* Extract cksum, ttl as ipv4 hdr is in cache */
		node_mbuf_priv1(mbuf3)->cksum = ipv4_hdr->hdr_checksum;
		node_mbuf_priv1(mbuf3)->ttl = ipv4_hdr->time_to_live;

		/* Perform LPM lookup to get NH and next node */
		rte_lpm_lookupx4(lpm, dip, dst.u32, drop_nh);

		/* Extract next node id and NH */
		node_mbuf_priv1(mbuf0)->nh = dst.u32[0] & 0xFFFF;
		next0 = (dst.u32[0] >> 16);

		node_mbuf_priv1(mbuf1)->nh = dst.u32[1] & 0xFFFF;
		next1 = (dst.u32[1] >> 16);

		node_mbuf_priv1(mbuf2)->nh = dst.u32[2] & 0xFFFF;
		next2 = (dst.u32[2] >> 16);

		node_mbuf_priv1(mbuf3)->nh = dst.u32[3] & 0xFFFF;
		next3 = (dst.u32[3] >> 16);

		/* Enqueue four to next node */
		rte_edge_t fix_spec =
			(next_index ^ next0) | (next_index ^ next1) |
			(next_index ^ next2) | (next_index ^ next3);

		if (unlikely(fix_spec)) {
			/* Copy things successfully speculated till now */
			rte_memcpy(to_next, from, last_spec * sizeof(from[0]));
			from += last_spec;
			to_next += last_spec;
			held += last_spec;
			last_spec = 0;

			/* Next0 */
			if (next_index == next0) {
				to_next[0] = from[0];
				to_next++;
				held++;
			} else {
				rte_node_enqueue_x1(graph, node, next0,
						    from[0]);
			}

			/* Next1 */
			if (next_index == next1) {
				to_next[0] = from[1];
				to_next++;
				held++;
			} else {
				rte_node_enqueue_x1(graph, node, next1,
						    from[1]);
			}

			/* Next2 */
			if (next_index == next2) {
				to_next[0] = from[2];
				to_next++;
				held++;
			} else {
				rte_node_enqueue_x1(graph, node, next2,
						    from[2]);
			}

			/* Next3 */
			if (next_index == next3) {
				to_next[0] = from[3];
				to_next++;
				held++;
			} else {
				rte_node_enqueue_x1(graph, node, next3,
						    from[3]);
			}

			from += 4;

		} else {
			last_spec += 4;
		}
	}

	while (n_left_from > 0) {
		uint32_t next_hop;

		mbuf0 = pkts[0];

		pkts += 1;
		n_left_from -= 1;

		/* Extract DIP of mbuf0 */
		ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf0, struct rte_ipv4_hdr *,
						sizeof(struct rte_ether_hdr));
		/* Extract cksum, ttl as ipv4 hdr is in cache */
		node_mbuf_priv1(mbuf0)->cksum = ipv4_hdr->hdr_checksum;
		node_mbuf_priv1(mbuf0)->ttl = ipv4_hdr->time_to_live;

		rc = rte_lpm_lookup(lpm, rte_be_to_cpu_32(ipv4_hdr->dst_addr),
				    &next_hop);
		next_hop = (rc == 0) ? next_hop : drop_nh;

		node_mbuf_priv1(mbuf0)->nh = next_hop & 0xFFFF;
		next0 = (next_hop >> 16);

		if (unlikely(next_index ^ next0)) {
			/* Copy things successfully speculated till now */
			rte_memcpy(to_next, from, last_spec * sizeof(from[0]));
			from += last_spec;
			to_next += last_spec;
			held += last_spec;
			last_spec = 0;

			rte_node_enqueue_x1(graph, node, next0, from[0]);
			from += 1;
		} else {
			last_spec += 1;
		}
	}

	/* !!! Home run !!! */
	if (likely(last_spec == nb_objs)) {
		rte_node_next_stream_move(graph, node, next_index);
		return nb_objs;
	}

	held += last_spec;
	/* Copy things successfully speculated till now */
	rte_memcpy(to_next, from, last_spec * sizeof(from[0]));
	rte_node_next_stream_put(graph, node, next_index, held);

	return nb_objs;
}

#endif /* __INCLUDE_IP4_LOOKUP_SSE_H__ */
Commit	Line	Data
f67539c2 TL	1	/* SPDX-License-Identifier: BSD-3-Clause
	2	* Copyright(C) 2020 Marvell International Ltd.
	3	*/
	4
	5	#ifndef __INCLUDE_IP4_LOOKUP_SSE_H__
	6	#define __INCLUDE_IP4_LOOKUP_SSE_H__
	7
	8	/* X86 SSE */
	9	static uint16_t
	10	ip4_lookup_node_process(struct rte_graph graph, struct rte_node node,
	11	void **objs, uint16_t nb_objs)
	12	{
	13	struct rte_mbuf mbuf0, mbuf1, mbuf2, mbuf3, **pkts;
	14	rte_edge_t next0, next1, next2, next3, next_index;
	15	struct rte_ipv4_hdr *ipv4_hdr;
	16	uint32_t ip0, ip1, ip2, ip3;
	17	void to_next, from;
	18	uint16_t last_spec = 0;
	19	uint16_t n_left_from;
	20	struct rte_lpm *lpm;
	21	uint16_t held = 0;
	22	uint32_t drop_nh;
	23	rte_xmm_t dst;
	24	__m128i dip; /* SSE register */
	25	int rc, i;
	26
	27	/* Speculative next */
	28	next_index = RTE_NODE_IP4_LOOKUP_NEXT_REWRITE;
	29	/* Drop node */
	30	drop_nh = ((uint32_t)RTE_NODE_IP4_LOOKUP_NEXT_PKT_DROP) << 16;
	31
	32	/* Get socket specific LPM from ctx */
	33	lpm = ((struct rte_lpm *)node->ctx);
	34
	35	pkts = (struct rte_mbuf **)objs;
	36	from = objs;
	37	n_left_from = nb_objs;
	38
	39	if (n_left_from >= 4) {
	40	for (i = 0; i < 4; i++)
	41	rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[i], void *,
	42	sizeof(struct rte_ether_hdr)));
	43	}
	44
	45	/* Get stream for the speculated next node */
	46	to_next = rte_node_next_stream_get(graph, node, next_index, nb_objs);
	47	while (n_left_from >= 4) {
	48	/* Prefetch next-next mbufs */
	49	if (likely(n_left_from > 11)) {
	50	rte_prefetch0(pkts[8]);
	51	rte_prefetch0(pkts[9]);
	52	rte_prefetch0(pkts[10]);
	53	rte_prefetch0(pkts[11]);
	54	}
	55
	56	/* Prefetch next mbuf data */
	57	if (likely(n_left_from > 7)) {
	58	rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[4], void *,
	59	sizeof(struct rte_ether_hdr)));
	60	rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[5], void *,
	61	sizeof(struct rte_ether_hdr)));
	62	rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[6], void *,
	63	sizeof(struct rte_ether_hdr)));
	64	rte_prefetch0(rte_pktmbuf_mtod_offset(pkts[7], void *,
65	sizeof(struct rte_ether_hdr)));
66	}
67
68	mbuf0 = pkts[0];
69	mbuf1 = pkts[1];
70	mbuf2 = pkts[2];
71	mbuf3 = pkts[3];
72
73	pkts += 4;
74	n_left_from -= 4;
75
76	/* Extract DIP of mbuf0 */
77	ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf0, struct rte_ipv4_hdr *,
78	sizeof(struct rte_ether_hdr));
79	ip0 = ipv4_hdr->dst_addr;
80	/* Extract cksum, ttl as ipv4 hdr is in cache */
81	node_mbuf_priv1(mbuf0)->cksum = ipv4_hdr->hdr_checksum;
82	node_mbuf_priv1(mbuf0)->ttl = ipv4_hdr->time_to_live;
83
84	/* Extract DIP of mbuf1 */
85	ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf1, struct rte_ipv4_hdr *,
86	sizeof(struct rte_ether_hdr));
87	ip1 = ipv4_hdr->dst_addr;
88	/* Extract cksum, ttl as ipv4 hdr is in cache */
89	node_mbuf_priv1(mbuf1)->cksum = ipv4_hdr->hdr_checksum;
90	node_mbuf_priv1(mbuf1)->ttl = ipv4_hdr->time_to_live;
91
92	/* Extract DIP of mbuf2 */
93	ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf2, struct rte_ipv4_hdr *,
94	sizeof(struct rte_ether_hdr));
95	ip2 = ipv4_hdr->dst_addr;
96	/* Extract cksum, ttl as ipv4 hdr is in cache */
97	node_mbuf_priv1(mbuf2)->cksum = ipv4_hdr->hdr_checksum;
98	node_mbuf_priv1(mbuf2)->ttl = ipv4_hdr->time_to_live;
99
100	/* Extract DIP of mbuf3 */
101	ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf3, struct rte_ipv4_hdr *,
102	sizeof(struct rte_ether_hdr));
103	ip3 = ipv4_hdr->dst_addr;
104
105	/* Prepare for lookup x4 */
106	dip = _mm_set_epi32(ip3, ip2, ip1, ip0);
107
108	/* Byte swap 4 IPV4 addresses. */
109	const __m128i bswap_mask = _mm_set_epi8(
110	12, 13, 14, 15, 8, 9, 10, 11, 4, 5, 6, 7, 0, 1, 2, 3);
111	dip = _mm_shuffle_epi8(dip, bswap_mask);
112
113	/* Extract cksum, ttl as ipv4 hdr is in cache */
114	node_mbuf_priv1(mbuf3)->cksum = ipv4_hdr->hdr_checksum;
115	node_mbuf_priv1(mbuf3)->ttl = ipv4_hdr->time_to_live;
116
117	/* Perform LPM lookup to get NH and next node */
118	rte_lpm_lookupx4(lpm, dip, dst.u32, drop_nh);
119
120	/* Extract next node id and NH */
121	node_mbuf_priv1(mbuf0)->nh = dst.u32[0] & 0xFFFF;
122	next0 = (dst.u32[0] >> 16);
123
124	node_mbuf_priv1(mbuf1)->nh = dst.u32[1] & 0xFFFF;
125	next1 = (dst.u32[1] >> 16);
126
127	node_mbuf_priv1(mbuf2)->nh = dst.u32[2] & 0xFFFF;
128	next2 = (dst.u32[2] >> 16);
129
130	node_mbuf_priv1(mbuf3)->nh = dst.u32[3] & 0xFFFF;
131	next3 = (dst.u32[3] >> 16);
132
133	/* Enqueue four to next node */
134	rte_edge_t fix_spec =
135	(next_index ^ next0) \| (next_index ^ next1) \|
136	(next_index ^ next2) \| (next_index ^ next3);
137
138	if (unlikely(fix_spec)) {
139	/* Copy things successfully speculated till now */
140	rte_memcpy(to_next, from, last_spec * sizeof(from[0]));
141	from += last_spec;
142	to_next += last_spec;
143	held += last_spec;
144	last_spec = 0;
145
146	/* Next0 */
147	if (next_index == next0) {
148	to_next[0] = from[0];
149	to_next++;
150	held++;
151	} else {
152	rte_node_enqueue_x1(graph, node, next0,
153	from[0]);
154	}
155
156	/* Next1 */
157	if (next_index == next1) {
158	to_next[0] = from[1];
159	to_next++;
160	held++;
161	} else {
162	rte_node_enqueue_x1(graph, node, next1,
163	from[1]);
164	}
165
166	/* Next2 */
167	if (next_index == next2) {
168	to_next[0] = from[2];
169	to_next++;
170	held++;
171	} else {
172	rte_node_enqueue_x1(graph, node, next2,
173	from[2]);
174	}
175
176	/* Next3 */
177	if (next_index == next3) {
178	to_next[0] = from[3];
179	to_next++;
180	held++;
181	} else {
182	rte_node_enqueue_x1(graph, node, next3,
183	from[3]);
184	}
185
186	from += 4;
187
188	} else {
189	last_spec += 4;
190	}
191	}
192
193	while (n_left_from > 0) {
194	uint32_t next_hop;
195
196	mbuf0 = pkts[0];
197
198	pkts += 1;
199	n_left_from -= 1;
200
201	/* Extract DIP of mbuf0 */
202	ipv4_hdr = rte_pktmbuf_mtod_offset(mbuf0, struct rte_ipv4_hdr *,
203	sizeof(struct rte_ether_hdr));
204	/* Extract cksum, ttl as ipv4 hdr is in cache */
205	node_mbuf_priv1(mbuf0)->cksum = ipv4_hdr->hdr_checksum;
206	node_mbuf_priv1(mbuf0)->ttl = ipv4_hdr->time_to_live;
207
208	rc = rte_lpm_lookup(lpm, rte_be_to_cpu_32(ipv4_hdr->dst_addr),
209	&next_hop);
210	next_hop = (rc == 0) ? next_hop : drop_nh;
211
212	node_mbuf_priv1(mbuf0)->nh = next_hop & 0xFFFF;
213	next0 = (next_hop >> 16);
214
215	if (unlikely(next_index ^ next0)) {
216	/* Copy things successfully speculated till now */
217	rte_memcpy(to_next, from, last_spec * sizeof(from[0]));
218	from += last_spec;
219	to_next += last_spec;
220	held += last_spec;
221	last_spec = 0;
222
223	rte_node_enqueue_x1(graph, node, next0, from[0]);
224	from += 1;
225	} else {
226	last_spec += 1;
227	}
228	}
229
230	/* !!! Home run !!! */
231	if (likely(last_spec == nb_objs)) {
232	rte_node_next_stream_move(graph, node, next_index);
233	return nb_objs;
234	}
235
236	held += last_spec;
237	/* Copy things successfully speculated till now */
238	rte_memcpy(to_next, from, last_spec * sizeof(from[0]));
239	rte_node_next_stream_put(graph, node, next_index, held);
240
241	return nb_objs;
242	}
243
244	#endif /* __INCLUDE_IP4_LOOKUP_SSE_H__ */