xref: /openbmc/linux/net/xdp/xsk_queue.h (revision 3557b3fd)
1 /* SPDX-License-Identifier: GPL-2.0 */
2 /* XDP user-space ring structure
3  * Copyright(c) 2018 Intel Corporation.
4  */
5 
6 #ifndef _LINUX_XSK_QUEUE_H
7 #define _LINUX_XSK_QUEUE_H
8 
9 #include <linux/types.h>
10 #include <linux/if_xdp.h>
11 #include <net/xdp_sock.h>
12 
13 #define RX_BATCH_SIZE 16
14 #define LAZY_UPDATE_THRESHOLD 128
15 
16 struct xdp_ring {
17 	u32 producer ____cacheline_aligned_in_smp;
18 	u32 consumer ____cacheline_aligned_in_smp;
19 };
20 
21 /* Used for the RX and TX queues for packets */
22 struct xdp_rxtx_ring {
23 	struct xdp_ring ptrs;
24 	struct xdp_desc desc[0] ____cacheline_aligned_in_smp;
25 };
26 
27 /* Used for the fill and completion queues for buffers */
28 struct xdp_umem_ring {
29 	struct xdp_ring ptrs;
30 	u64 desc[0] ____cacheline_aligned_in_smp;
31 };
32 
33 struct xsk_queue {
34 	u64 chunk_mask;
35 	u64 size;
36 	u32 ring_mask;
37 	u32 nentries;
38 	u32 prod_head;
39 	u32 prod_tail;
40 	u32 cons_head;
41 	u32 cons_tail;
42 	struct xdp_ring *ring;
43 	u64 invalid_descs;
44 };
45 
46 /* The structure of the shared state of the rings are the same as the
47  * ring buffer in kernel/events/ring_buffer.c. For the Rx and completion
48  * ring, the kernel is the producer and user space is the consumer. For
49  * the Tx and fill rings, the kernel is the consumer and user space is
50  * the producer.
51  *
52  * producer                         consumer
53  *
54  * if (LOAD ->consumer) {           LOAD ->producer
55  *                    (A)           smp_rmb()       (C)
56  *    STORE $data                   LOAD $data
57  *    smp_wmb()       (B)           smp_mb()        (D)
58  *    STORE ->producer              STORE ->consumer
59  * }
60  *
61  * (A) pairs with (D), and (B) pairs with (C).
62  *
63  * Starting with (B), it protects the data from being written after
64  * the producer pointer. If this barrier was missing, the consumer
65  * could observe the producer pointer being set and thus load the data
66  * before the producer has written the new data. The consumer would in
67  * this case load the old data.
68  *
69  * (C) protects the consumer from speculatively loading the data before
70  * the producer pointer actually has been read. If we do not have this
71  * barrier, some architectures could load old data as speculative loads
72  * are not discarded as the CPU does not know there is a dependency
73  * between ->producer and data.
74  *
75  * (A) is a control dependency that separates the load of ->consumer
76  * from the stores of $data. In case ->consumer indicates there is no
77  * room in the buffer to store $data we do not. So no barrier is needed.
78  *
79  * (D) protects the load of the data to be observed to happen after the
80  * store of the consumer pointer. If we did not have this memory
81  * barrier, the producer could observe the consumer pointer being set
82  * and overwrite the data with a new value before the consumer got the
83  * chance to read the old value. The consumer would thus miss reading
84  * the old entry and very likely read the new entry twice, once right
85  * now and again after circling through the ring.
86  */
87 
88 /* Common functions operating for both RXTX and umem queues */
89 
90 static inline u64 xskq_nb_invalid_descs(struct xsk_queue *q)
91 {
92 	return q ? q->invalid_descs : 0;
93 }
94 
95 static inline u32 xskq_nb_avail(struct xsk_queue *q, u32 dcnt)
96 {
97 	u32 entries = q->prod_tail - q->cons_tail;
98 
99 	if (entries == 0) {
100 		/* Refresh the local pointer */
101 		q->prod_tail = READ_ONCE(q->ring->producer);
102 		entries = q->prod_tail - q->cons_tail;
103 	}
104 
105 	return (entries > dcnt) ? dcnt : entries;
106 }
107 
108 static inline u32 xskq_nb_free(struct xsk_queue *q, u32 producer, u32 dcnt)
109 {
110 	u32 free_entries = q->nentries - (producer - q->cons_tail);
111 
112 	if (free_entries >= dcnt)
113 		return free_entries;
114 
115 	/* Refresh the local tail pointer */
116 	q->cons_tail = READ_ONCE(q->ring->consumer);
117 	return q->nentries - (producer - q->cons_tail);
118 }
119 
120 /* UMEM queue */
121 
122 static inline bool xskq_is_valid_addr(struct xsk_queue *q, u64 addr)
123 {
124 	if (addr >= q->size) {
125 		q->invalid_descs++;
126 		return false;
127 	}
128 
129 	return true;
130 }
131 
132 static inline u64 *xskq_validate_addr(struct xsk_queue *q, u64 *addr)
133 {
134 	while (q->cons_tail != q->cons_head) {
135 		struct xdp_umem_ring *ring = (struct xdp_umem_ring *)q->ring;
136 		unsigned int idx = q->cons_tail & q->ring_mask;
137 
138 		*addr = READ_ONCE(ring->desc[idx]) & q->chunk_mask;
139 		if (xskq_is_valid_addr(q, *addr))
140 			return addr;
141 
142 		q->cons_tail++;
143 	}
144 
145 	return NULL;
146 }
147 
148 static inline u64 *xskq_peek_addr(struct xsk_queue *q, u64 *addr)
149 {
150 	if (q->cons_tail == q->cons_head) {
151 		smp_mb(); /* D, matches A */
152 		WRITE_ONCE(q->ring->consumer, q->cons_tail);
153 		q->cons_head = q->cons_tail + xskq_nb_avail(q, RX_BATCH_SIZE);
154 
155 		/* Order consumer and data */
156 		smp_rmb();
157 	}
158 
159 	return xskq_validate_addr(q, addr);
160 }
161 
162 static inline void xskq_discard_addr(struct xsk_queue *q)
163 {
164 	q->cons_tail++;
165 }
166 
167 static inline int xskq_produce_addr(struct xsk_queue *q, u64 addr)
168 {
169 	struct xdp_umem_ring *ring = (struct xdp_umem_ring *)q->ring;
170 
171 	if (xskq_nb_free(q, q->prod_tail, 1) == 0)
172 		return -ENOSPC;
173 
174 	/* A, matches D */
175 	ring->desc[q->prod_tail++ & q->ring_mask] = addr;
176 
177 	/* Order producer and data */
178 	smp_wmb(); /* B, matches C */
179 
180 	WRITE_ONCE(q->ring->producer, q->prod_tail);
181 	return 0;
182 }
183 
184 static inline int xskq_produce_addr_lazy(struct xsk_queue *q, u64 addr)
185 {
186 	struct xdp_umem_ring *ring = (struct xdp_umem_ring *)q->ring;
187 
188 	if (xskq_nb_free(q, q->prod_head, LAZY_UPDATE_THRESHOLD) == 0)
189 		return -ENOSPC;
190 
191 	/* A, matches D */
192 	ring->desc[q->prod_head++ & q->ring_mask] = addr;
193 	return 0;
194 }
195 
196 static inline void xskq_produce_flush_addr_n(struct xsk_queue *q,
197 					     u32 nb_entries)
198 {
199 	/* Order producer and data */
200 	smp_wmb(); /* B, matches C */
201 
202 	q->prod_tail += nb_entries;
203 	WRITE_ONCE(q->ring->producer, q->prod_tail);
204 }
205 
206 static inline int xskq_reserve_addr(struct xsk_queue *q)
207 {
208 	if (xskq_nb_free(q, q->prod_head, 1) == 0)
209 		return -ENOSPC;
210 
211 	/* A, matches D */
212 	q->prod_head++;
213 	return 0;
214 }
215 
216 /* Rx/Tx queue */
217 
218 static inline bool xskq_is_valid_desc(struct xsk_queue *q, struct xdp_desc *d)
219 {
220 	if (!xskq_is_valid_addr(q, d->addr))
221 		return false;
222 
223 	if (((d->addr + d->len) & q->chunk_mask) != (d->addr & q->chunk_mask) ||
224 	    d->options) {
225 		q->invalid_descs++;
226 		return false;
227 	}
228 
229 	return true;
230 }
231 
232 static inline struct xdp_desc *xskq_validate_desc(struct xsk_queue *q,
233 						  struct xdp_desc *desc)
234 {
235 	while (q->cons_tail != q->cons_head) {
236 		struct xdp_rxtx_ring *ring = (struct xdp_rxtx_ring *)q->ring;
237 		unsigned int idx = q->cons_tail & q->ring_mask;
238 
239 		*desc = READ_ONCE(ring->desc[idx]);
240 		if (xskq_is_valid_desc(q, desc))
241 			return desc;
242 
243 		q->cons_tail++;
244 	}
245 
246 	return NULL;
247 }
248 
249 static inline struct xdp_desc *xskq_peek_desc(struct xsk_queue *q,
250 					      struct xdp_desc *desc)
251 {
252 	if (q->cons_tail == q->cons_head) {
253 		smp_mb(); /* D, matches A */
254 		WRITE_ONCE(q->ring->consumer, q->cons_tail);
255 		q->cons_head = q->cons_tail + xskq_nb_avail(q, RX_BATCH_SIZE);
256 
257 		/* Order consumer and data */
258 		smp_rmb(); /* C, matches B */
259 	}
260 
261 	return xskq_validate_desc(q, desc);
262 }
263 
264 static inline void xskq_discard_desc(struct xsk_queue *q)
265 {
266 	q->cons_tail++;
267 }
268 
269 static inline int xskq_produce_batch_desc(struct xsk_queue *q,
270 					  u64 addr, u32 len)
271 {
272 	struct xdp_rxtx_ring *ring = (struct xdp_rxtx_ring *)q->ring;
273 	unsigned int idx;
274 
275 	if (xskq_nb_free(q, q->prod_head, 1) == 0)
276 		return -ENOSPC;
277 
278 	/* A, matches D */
279 	idx = (q->prod_head++) & q->ring_mask;
280 	ring->desc[idx].addr = addr;
281 	ring->desc[idx].len = len;
282 
283 	return 0;
284 }
285 
286 static inline void xskq_produce_flush_desc(struct xsk_queue *q)
287 {
288 	/* Order producer and data */
289 	smp_wmb(); /* B, matches C */
290 
291 	q->prod_tail = q->prod_head,
292 	WRITE_ONCE(q->ring->producer, q->prod_tail);
293 }
294 
295 static inline bool xskq_full_desc(struct xsk_queue *q)
296 {
297 	return xskq_nb_avail(q, q->nentries) == q->nentries;
298 }
299 
300 static inline bool xskq_empty_desc(struct xsk_queue *q)
301 {
302 	return xskq_nb_free(q, q->prod_tail, q->nentries) == q->nentries;
303 }
304 
305 void xskq_set_umem(struct xsk_queue *q, u64 size, u64 chunk_mask);
306 struct xsk_queue *xskq_create(u32 nentries, bool umem_queue);
307 void xskq_destroy(struct xsk_queue *q_ops);
308 
309 /* Executed by the core when the entire UMEM gets freed */
310 void xsk_reuseq_destroy(struct xdp_umem *umem);
311 
312 #endif /* _LINUX_XSK_QUEUE_H */
313