xref: /openbmc/linux/include/net/dst.h (revision 93df8a1e)
1 /*
2  * net/dst.h	Protocol independent destination cache definitions.
3  *
4  * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
5  *
6  */
7 
8 #ifndef _NET_DST_H
9 #define _NET_DST_H
10 
11 #include <net/dst_ops.h>
12 #include <linux/netdevice.h>
13 #include <linux/rtnetlink.h>
14 #include <linux/rcupdate.h>
15 #include <linux/bug.h>
16 #include <linux/jiffies.h>
17 #include <net/neighbour.h>
18 #include <asm/processor.h>
19 
20 #define DST_GC_MIN	(HZ/10)
21 #define DST_GC_INC	(HZ/2)
22 #define DST_GC_MAX	(120*HZ)
23 
24 /* Each dst_entry has reference count and sits in some parent list(s).
25  * When it is removed from parent list, it is "freed" (dst_free).
26  * After this it enters dead state (dst->obsolete > 0) and if its refcnt
27  * is zero, it can be destroyed immediately, otherwise it is added
28  * to gc list and garbage collector periodically checks the refcnt.
29  */
30 
31 struct sk_buff;
32 
33 struct dst_entry {
34 	struct rcu_head		rcu_head;
35 	struct dst_entry	*child;
36 	struct net_device       *dev;
37 	struct  dst_ops	        *ops;
38 	unsigned long		_metrics;
39 	unsigned long           expires;
40 	struct dst_entry	*path;
41 	struct dst_entry	*from;
42 #ifdef CONFIG_XFRM
43 	struct xfrm_state	*xfrm;
44 #else
45 	void			*__pad1;
46 #endif
47 	int			(*input)(struct sk_buff *);
48 	int			(*output)(struct sock *sk, struct sk_buff *skb);
49 
50 	unsigned short		flags;
51 #define DST_HOST		0x0001
52 #define DST_NOXFRM		0x0002
53 #define DST_NOPOLICY		0x0004
54 #define DST_NOHASH		0x0008
55 #define DST_NOCACHE		0x0010
56 #define DST_NOCOUNT		0x0020
57 #define DST_FAKE_RTABLE		0x0040
58 #define DST_XFRM_TUNNEL		0x0080
59 #define DST_XFRM_QUEUE		0x0100
60 
61 	unsigned short		pending_confirm;
62 
63 	short			error;
64 
65 	/* A non-zero value of dst->obsolete forces by-hand validation
66 	 * of the route entry.  Positive values are set by the generic
67 	 * dst layer to indicate that the entry has been forcefully
68 	 * destroyed.
69 	 *
70 	 * Negative values are used by the implementation layer code to
71 	 * force invocation of the dst_ops->check() method.
72 	 */
73 	short			obsolete;
74 #define DST_OBSOLETE_NONE	0
75 #define DST_OBSOLETE_DEAD	2
76 #define DST_OBSOLETE_FORCE_CHK	-1
77 #define DST_OBSOLETE_KILL	-2
78 	unsigned short		header_len;	/* more space at head required */
79 	unsigned short		trailer_len;	/* space to reserve at tail */
80 #ifdef CONFIG_IP_ROUTE_CLASSID
81 	__u32			tclassid;
82 #else
83 	__u32			__pad2;
84 #endif
85 
86 	/*
87 	 * Align __refcnt to a 64 bytes alignment
88 	 * (L1_CACHE_SIZE would be too much)
89 	 */
90 #ifdef CONFIG_64BIT
91 	long			__pad_to_align_refcnt[2];
92 #endif
93 	/*
94 	 * __refcnt wants to be on a different cache line from
95 	 * input/output/ops or performance tanks badly
96 	 */
97 	atomic_t		__refcnt;	/* client references	*/
98 	int			__use;
99 	unsigned long		lastuse;
100 	union {
101 		struct dst_entry	*next;
102 		struct rtable __rcu	*rt_next;
103 		struct rt6_info		*rt6_next;
104 		struct dn_route __rcu	*dn_next;
105 	};
106 };
107 
108 u32 *dst_cow_metrics_generic(struct dst_entry *dst, unsigned long old);
109 extern const u32 dst_default_metrics[];
110 
111 #define DST_METRICS_READ_ONLY		0x1UL
112 #define DST_METRICS_FLAGS		0x3UL
113 #define __DST_METRICS_PTR(Y)	\
114 	((u32 *)((Y) & ~DST_METRICS_FLAGS))
115 #define DST_METRICS_PTR(X)	__DST_METRICS_PTR((X)->_metrics)
116 
117 static inline bool dst_metrics_read_only(const struct dst_entry *dst)
118 {
119 	return dst->_metrics & DST_METRICS_READ_ONLY;
120 }
121 
122 void __dst_destroy_metrics_generic(struct dst_entry *dst, unsigned long old);
123 
124 static inline void dst_destroy_metrics_generic(struct dst_entry *dst)
125 {
126 	unsigned long val = dst->_metrics;
127 	if (!(val & DST_METRICS_READ_ONLY))
128 		__dst_destroy_metrics_generic(dst, val);
129 }
130 
131 static inline u32 *dst_metrics_write_ptr(struct dst_entry *dst)
132 {
133 	unsigned long p = dst->_metrics;
134 
135 	BUG_ON(!p);
136 
137 	if (p & DST_METRICS_READ_ONLY)
138 		return dst->ops->cow_metrics(dst, p);
139 	return __DST_METRICS_PTR(p);
140 }
141 
142 /* This may only be invoked before the entry has reached global
143  * visibility.
144  */
145 static inline void dst_init_metrics(struct dst_entry *dst,
146 				    const u32 *src_metrics,
147 				    bool read_only)
148 {
149 	dst->_metrics = ((unsigned long) src_metrics) |
150 		(read_only ? DST_METRICS_READ_ONLY : 0);
151 }
152 
153 static inline void dst_copy_metrics(struct dst_entry *dest, const struct dst_entry *src)
154 {
155 	u32 *dst_metrics = dst_metrics_write_ptr(dest);
156 
157 	if (dst_metrics) {
158 		u32 *src_metrics = DST_METRICS_PTR(src);
159 
160 		memcpy(dst_metrics, src_metrics, RTAX_MAX * sizeof(u32));
161 	}
162 }
163 
164 static inline u32 *dst_metrics_ptr(struct dst_entry *dst)
165 {
166 	return DST_METRICS_PTR(dst);
167 }
168 
169 static inline u32
170 dst_metric_raw(const struct dst_entry *dst, const int metric)
171 {
172 	u32 *p = DST_METRICS_PTR(dst);
173 
174 	return p[metric-1];
175 }
176 
177 static inline u32
178 dst_metric(const struct dst_entry *dst, const int metric)
179 {
180 	WARN_ON_ONCE(metric == RTAX_HOPLIMIT ||
181 		     metric == RTAX_ADVMSS ||
182 		     metric == RTAX_MTU);
183 	return dst_metric_raw(dst, metric);
184 }
185 
186 static inline u32
187 dst_metric_advmss(const struct dst_entry *dst)
188 {
189 	u32 advmss = dst_metric_raw(dst, RTAX_ADVMSS);
190 
191 	if (!advmss)
192 		advmss = dst->ops->default_advmss(dst);
193 
194 	return advmss;
195 }
196 
197 static inline void dst_metric_set(struct dst_entry *dst, int metric, u32 val)
198 {
199 	u32 *p = dst_metrics_write_ptr(dst);
200 
201 	if (p)
202 		p[metric-1] = val;
203 }
204 
205 static inline u32
206 dst_feature(const struct dst_entry *dst, u32 feature)
207 {
208 	return dst_metric(dst, RTAX_FEATURES) & feature;
209 }
210 
211 static inline u32 dst_mtu(const struct dst_entry *dst)
212 {
213 	return dst->ops->mtu(dst);
214 }
215 
216 /* RTT metrics are stored in milliseconds for user ABI, but used as jiffies */
217 static inline unsigned long dst_metric_rtt(const struct dst_entry *dst, int metric)
218 {
219 	return msecs_to_jiffies(dst_metric(dst, metric));
220 }
221 
222 static inline u32
223 dst_allfrag(const struct dst_entry *dst)
224 {
225 	int ret = dst_feature(dst,  RTAX_FEATURE_ALLFRAG);
226 	return ret;
227 }
228 
229 static inline int
230 dst_metric_locked(const struct dst_entry *dst, int metric)
231 {
232 	return dst_metric(dst, RTAX_LOCK) & (1<<metric);
233 }
234 
235 static inline void dst_hold(struct dst_entry *dst)
236 {
237 	/*
238 	 * If your kernel compilation stops here, please check
239 	 * __pad_to_align_refcnt declaration in struct dst_entry
240 	 */
241 	BUILD_BUG_ON(offsetof(struct dst_entry, __refcnt) & 63);
242 	atomic_inc(&dst->__refcnt);
243 }
244 
245 static inline void dst_use(struct dst_entry *dst, unsigned long time)
246 {
247 	dst_hold(dst);
248 	dst->__use++;
249 	dst->lastuse = time;
250 }
251 
252 static inline void dst_use_noref(struct dst_entry *dst, unsigned long time)
253 {
254 	dst->__use++;
255 	dst->lastuse = time;
256 }
257 
258 static inline struct dst_entry *dst_clone(struct dst_entry *dst)
259 {
260 	if (dst)
261 		atomic_inc(&dst->__refcnt);
262 	return dst;
263 }
264 
265 void dst_release(struct dst_entry *dst);
266 
267 static inline void refdst_drop(unsigned long refdst)
268 {
269 	if (!(refdst & SKB_DST_NOREF))
270 		dst_release((struct dst_entry *)(refdst & SKB_DST_PTRMASK));
271 }
272 
273 /**
274  * skb_dst_drop - drops skb dst
275  * @skb: buffer
276  *
277  * Drops dst reference count if a reference was taken.
278  */
279 static inline void skb_dst_drop(struct sk_buff *skb)
280 {
281 	if (skb->_skb_refdst) {
282 		refdst_drop(skb->_skb_refdst);
283 		skb->_skb_refdst = 0UL;
284 	}
285 }
286 
287 static inline void skb_dst_copy(struct sk_buff *nskb, const struct sk_buff *oskb)
288 {
289 	nskb->_skb_refdst = oskb->_skb_refdst;
290 	if (!(nskb->_skb_refdst & SKB_DST_NOREF))
291 		dst_clone(skb_dst(nskb));
292 }
293 
294 /**
295  * skb_dst_force - makes sure skb dst is refcounted
296  * @skb: buffer
297  *
298  * If dst is not yet refcounted, let's do it
299  */
300 static inline void skb_dst_force(struct sk_buff *skb)
301 {
302 	if (skb_dst_is_noref(skb)) {
303 		WARN_ON(!rcu_read_lock_held());
304 		skb->_skb_refdst &= ~SKB_DST_NOREF;
305 		dst_clone(skb_dst(skb));
306 	}
307 }
308 
309 
310 /**
311  *	__skb_tunnel_rx - prepare skb for rx reinsert
312  *	@skb: buffer
313  *	@dev: tunnel device
314  *	@net: netns for packet i/o
315  *
316  *	After decapsulation, packet is going to re-enter (netif_rx()) our stack,
317  *	so make some cleanups. (no accounting done)
318  */
319 static inline void __skb_tunnel_rx(struct sk_buff *skb, struct net_device *dev,
320 				   struct net *net)
321 {
322 	skb->dev = dev;
323 
324 	/*
325 	 * Clear hash so that we can recalulate the hash for the
326 	 * encapsulated packet, unless we have already determine the hash
327 	 * over the L4 4-tuple.
328 	 */
329 	skb_clear_hash_if_not_l4(skb);
330 	skb_set_queue_mapping(skb, 0);
331 	skb_scrub_packet(skb, !net_eq(net, dev_net(dev)));
332 }
333 
334 /**
335  *	skb_tunnel_rx - prepare skb for rx reinsert
336  *	@skb: buffer
337  *	@dev: tunnel device
338  *
339  *	After decapsulation, packet is going to re-enter (netif_rx()) our stack,
340  *	so make some cleanups, and perform accounting.
341  *	Note: this accounting is not SMP safe.
342  */
343 static inline void skb_tunnel_rx(struct sk_buff *skb, struct net_device *dev,
344 				 struct net *net)
345 {
346 	/* TODO : stats should be SMP safe */
347 	dev->stats.rx_packets++;
348 	dev->stats.rx_bytes += skb->len;
349 	__skb_tunnel_rx(skb, dev, net);
350 }
351 
352 int dst_discard_sk(struct sock *sk, struct sk_buff *skb);
353 static inline int dst_discard(struct sk_buff *skb)
354 {
355 	return dst_discard_sk(skb->sk, skb);
356 }
357 void *dst_alloc(struct dst_ops *ops, struct net_device *dev, int initial_ref,
358 		int initial_obsolete, unsigned short flags);
359 void __dst_free(struct dst_entry *dst);
360 struct dst_entry *dst_destroy(struct dst_entry *dst);
361 
362 static inline void dst_free(struct dst_entry *dst)
363 {
364 	if (dst->obsolete > 0)
365 		return;
366 	if (!atomic_read(&dst->__refcnt)) {
367 		dst = dst_destroy(dst);
368 		if (!dst)
369 			return;
370 	}
371 	__dst_free(dst);
372 }
373 
374 static inline void dst_rcu_free(struct rcu_head *head)
375 {
376 	struct dst_entry *dst = container_of(head, struct dst_entry, rcu_head);
377 	dst_free(dst);
378 }
379 
380 static inline void dst_confirm(struct dst_entry *dst)
381 {
382 	dst->pending_confirm = 1;
383 }
384 
385 static inline int dst_neigh_output(struct dst_entry *dst, struct neighbour *n,
386 				   struct sk_buff *skb)
387 {
388 	const struct hh_cache *hh;
389 
390 	if (dst->pending_confirm) {
391 		unsigned long now = jiffies;
392 
393 		dst->pending_confirm = 0;
394 		/* avoid dirtying neighbour */
395 		if (n->confirmed != now)
396 			n->confirmed = now;
397 	}
398 
399 	hh = &n->hh;
400 	if ((n->nud_state & NUD_CONNECTED) && hh->hh_len)
401 		return neigh_hh_output(hh, skb);
402 	else
403 		return n->output(n, skb);
404 }
405 
406 static inline struct neighbour *dst_neigh_lookup(const struct dst_entry *dst, const void *daddr)
407 {
408 	struct neighbour *n = dst->ops->neigh_lookup(dst, NULL, daddr);
409 	return IS_ERR(n) ? NULL : n;
410 }
411 
412 static inline struct neighbour *dst_neigh_lookup_skb(const struct dst_entry *dst,
413 						     struct sk_buff *skb)
414 {
415 	struct neighbour *n =  dst->ops->neigh_lookup(dst, skb, NULL);
416 	return IS_ERR(n) ? NULL : n;
417 }
418 
419 static inline void dst_link_failure(struct sk_buff *skb)
420 {
421 	struct dst_entry *dst = skb_dst(skb);
422 	if (dst && dst->ops && dst->ops->link_failure)
423 		dst->ops->link_failure(skb);
424 }
425 
426 static inline void dst_set_expires(struct dst_entry *dst, int timeout)
427 {
428 	unsigned long expires = jiffies + timeout;
429 
430 	if (expires == 0)
431 		expires = 1;
432 
433 	if (dst->expires == 0 || time_before(expires, dst->expires))
434 		dst->expires = expires;
435 }
436 
437 /* Output packet to network from transport.  */
438 static inline int dst_output_sk(struct sock *sk, struct sk_buff *skb)
439 {
440 	return skb_dst(skb)->output(sk, skb);
441 }
442 static inline int dst_output(struct sk_buff *skb)
443 {
444 	return dst_output_sk(skb->sk, skb);
445 }
446 
447 /* Input packet from network to transport.  */
448 static inline int dst_input(struct sk_buff *skb)
449 {
450 	return skb_dst(skb)->input(skb);
451 }
452 
453 static inline struct dst_entry *dst_check(struct dst_entry *dst, u32 cookie)
454 {
455 	if (dst->obsolete)
456 		dst = dst->ops->check(dst, cookie);
457 	return dst;
458 }
459 
460 void dst_init(void);
461 
462 /* Flags for xfrm_lookup flags argument. */
463 enum {
464 	XFRM_LOOKUP_ICMP = 1 << 0,
465 	XFRM_LOOKUP_QUEUE = 1 << 1,
466 	XFRM_LOOKUP_KEEP_DST_REF = 1 << 2,
467 };
468 
469 struct flowi;
470 #ifndef CONFIG_XFRM
471 static inline struct dst_entry *xfrm_lookup(struct net *net,
472 					    struct dst_entry *dst_orig,
473 					    const struct flowi *fl, struct sock *sk,
474 					    int flags)
475 {
476 	return dst_orig;
477 }
478 
479 static inline struct dst_entry *xfrm_lookup_route(struct net *net,
480 						  struct dst_entry *dst_orig,
481 						  const struct flowi *fl,
482 						  struct sock *sk,
483 						  int flags)
484 {
485 	return dst_orig;
486 }
487 
488 static inline struct xfrm_state *dst_xfrm(const struct dst_entry *dst)
489 {
490 	return NULL;
491 }
492 
493 #else
494 struct dst_entry *xfrm_lookup(struct net *net, struct dst_entry *dst_orig,
495 			      const struct flowi *fl, struct sock *sk,
496 			      int flags);
497 
498 struct dst_entry *xfrm_lookup_route(struct net *net, struct dst_entry *dst_orig,
499 				    const struct flowi *fl, struct sock *sk,
500 				    int flags);
501 
502 /* skb attached with this dst needs transformation if dst->xfrm is valid */
503 static inline struct xfrm_state *dst_xfrm(const struct dst_entry *dst)
504 {
505 	return dst->xfrm;
506 }
507 #endif
508 
509 #endif /* _NET_DST_H */
510