Release 4.8 net/core/dst.c

/*
 * net/core/dst.c       Protocol independent destination cache.
 *
 * Authors:             Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
 *
 */

#include <linux/bitops.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/workqueue.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/string.h>
#include <linux/types.h>
#include <net/net_namespace.h>
#include <linux/sched.h>
#include <linux/prefetch.h>
#include <net/lwtunnel.h>

#include <net/dst.h>
#include <net/dst_metadata.h>

/*
 * Theory of operations:
 * 1) We use a list, protected by a spinlock, to add
 *    new entries from both BH and non-BH context.
 * 2) In order to keep spinlock held for a small delay,
 *    we use a second list where are stored long lived
 *    entries, that are handled by the garbage collect thread
 *    fired by a workqueue.
 * 3) This list is guarded by a mutex,
 *    so that the gc_task and dst_dev_event() can be synchronized.
 */

/*
 * We want to keep lock & list close together
 * to dirty as few cache lines as possible in __dst_free().
 * As this is not a very strong hint, we dont force an alignment on SMP.
 */
static struct {
	
spinlock_t		lock;
	
struct dst_entry	*list;
	
unsigned long		timer_inc;
	
unsigned long		timer_expires;
} 
dst_garbage = {
	.lock = __SPIN_LOCK_UNLOCKED(dst_garbage.lock),
	.timer_inc = DST_GC_MAX,
};
static void dst_gc_task(struct work_struct *work);
static void ___dst_free(struct dst_entry *dst);

static DECLARE_DELAYED_WORK(dst_gc_work, dst_gc_task);

static DEFINE_MUTEX(dst_gc_mutex);
/*
 * long lived entries are maintained in this list, guarded by dst_gc_mutex
 */

static struct dst_entry         *dst_busy_list;



static void dst_gc_task(struct work_struct *work)
{
	int    delayed = 0;
	int    work_performed = 0;
	unsigned long expires = ~0L;
	struct dst_entry *dst, *next, head;
	struct dst_entry *last = &head;

	mutex_lock(&dst_gc_mutex);
	next = dst_busy_list;

loop:
	while ((dst = next) != NULL) {
		next = dst->next;
		prefetch(&next->next);
		cond_resched();
		if (likely(atomic_read(&dst->__refcnt))) {
			last->next = dst;
			last = dst;
			delayed++;
			continue;
		}
		work_performed++;

		dst = dst_destroy(dst);
		if (dst) {
			/* NOHASH and still referenced. Unless it is already
                         * on gc list, invalidate it and add to gc list.
                         *
                         * Note: this is temporary. Actually, NOHASH dst's
                         * must be obsoleted when parent is obsoleted.
                         * But we do not have state "obsoleted, but
                         * referenced by parent", so it is right.
                         */
			if (dst->obsolete > 0)
				continue;

			___dst_free(dst);
			dst->next = next;
			next = dst;
		}
	}

	spin_lock_bh(&dst_garbage.lock);
	next = dst_garbage.list;
	if (next) {
		dst_garbage.list = NULL;
		spin_unlock_bh(&dst_garbage.lock);
		goto loop;
	}
	last->next = NULL;
	dst_busy_list = head.next;
	if (!dst_busy_list)
		dst_garbage.timer_inc = DST_GC_MAX;
	else {
		/*
                 * if we freed less than 1/10 of delayed entries,
                 * we can sleep longer.
                 */
		if (work_performed <= delayed/10) {
			dst_garbage.timer_expires += dst_garbage.timer_inc;
			if (dst_garbage.timer_expires > DST_GC_MAX)
				dst_garbage.timer_expires = DST_GC_MAX;
			dst_garbage.timer_inc += DST_GC_INC;
		} else {
			dst_garbage.timer_inc = DST_GC_INC;
			dst_garbage.timer_expires = DST_GC_MIN;
		}
		expires = dst_garbage.timer_expires;
		/*
                 * if the next desired timer is more than 4 seconds in the
                 * future then round the timer to whole seconds
                 */
		if (expires > 4*HZ)
			expires = round_jiffies_relative(expires);
		schedule_delayed_work(&dst_gc_work, expires);
	}

	spin_unlock_bh(&dst_garbage.lock);
	mutex_unlock(&dst_gc_mutex);
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
eric dumazet
eric dumazet
168
52.83%
2
15.38%
pre-git
pre-git
79
24.84%
5
38.46%
alexey kuznetsov
alexey kuznetsov
35
11.01%
2
15.38%
denis v. lunev
denis v. lunev
21
6.60%
1
7.69%
arjan van de ven
arjan van de ven
13
4.09%
1
7.69%
laurent chavey
laurent chavey
1
0.31%
1
7.69%
david s. miller
david s. miller
1
0.31%
1
7.69%
Total
318
100.00%
13
100.00%



int dst_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	kfree_skb(skb);
	return 0;
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
pre-git
pre-git
15
53.57%
1
25.00%
eric w. biederman
eric w. biederman
6
21.43%
1
25.00%
eric dumazet
eric dumazet
5
17.86%
1
25.00%
herbert xu
herbert xu
2
7.14%
1
25.00%
Total
28
100.00%
4
100.00%


EXPORT_SYMBOL(dst_discard_out);


const u32 dst_default_metrics[RTAX_MAX + 1] = {
	/* This initializer is needed to force linker to place this variable
         * into const section. Otherwise it might end into bss section.
         * We really want to avoid false sharing on this variable, and catch
         * any writes on it.
         */
	[RTAX_MAX] = 0xdeadbeef,
};



void dst_init(struct dst_entry *dst, struct dst_ops *ops,
	      struct net_device *dev, int initial_ref, int initial_obsolete,
	      unsigned short flags)
{
	dst->child = NULL;
	dst->dev = dev;
	if (dev)
		dev_hold(dev);
	dst->ops = ops;
	dst_init_metrics(dst, dst_default_metrics, true);
	dst->expires = 0UL;
	dst->path = dst;
	dst->from = NULL;
#ifdef CONFIG_XFRM
	dst->xfrm = NULL;
#endif
	dst->input = dst_discard;
	dst->output = dst_discard_out;
	dst->error = 0;
	dst->obsolete = initial_obsolete;
	dst->header_len = 0;
	dst->trailer_len = 0;
#ifdef CONFIG_IP_ROUTE_CLASSID
	dst->tclassid = 0;
#endif
	dst->lwtstate = NULL;
	atomic_set(&dst->__refcnt, initial_ref);
	dst->__use = 0;
	dst->lastuse = jiffies;
	dst->flags = flags;
	dst->pending_confirm = 0;
	dst->next = NULL;
	if (!(flags & DST_NOCOUNT))
		dst_entries_add(ops, 1);
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
david s. miller
david s. miller
136
66.67%
7
46.67%
pre-git
pre-git
41
20.10%
2
13.33%
thomas graf
thomas graf
7
3.43%
1
6.67%
hideaki yoshifuji
hideaki yoshifuji
6
2.94%
1
6.67%
jiri benc
jiri benc
6
2.94%
1
6.67%
alexey kuznetsov
alexey kuznetsov
4
1.96%
1
6.67%
eric dumazet
eric dumazet
3
1.47%
1
6.67%
eric w. biederman
eric w. biederman
1
0.49%
1
6.67%
Total
204
100.00%
15
100.00%


EXPORT_SYMBOL(dst_init);



void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
		int initial_ref, int initial_obsolete, unsigned short flags)
{
	struct dst_entry *dst;

	if (ops->gc && dst_entries_get_fast(ops) > ops->gc_thresh) {
		if (ops->gc(ops))
			return NULL;
	}

	dst = kmem_cache_alloc(ops->kmem_cachep, GFP_ATOMIC);
	if (!dst)
		return NULL;

	dst_init(dst, ops, dev, initial_ref, initial_obsolete, flags);

	return dst;
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
thomas graf
thomas graf
93
95.88%
1
50.00%
pre-git
pre-git
4
4.12%
1
50.00%
Total
97
100.00%
2
100.00%


EXPORT_SYMBOL(dst_alloc);



static void ___dst_free(struct dst_entry *dst)
{
	/* The first case (dev==NULL) is required, when
           protocol module is unloaded.
         */
	if (dst->dev == NULL || !(dst->dev->flags&IFF_UP)) {
		dst->input = dst_discard;
		dst->output = dst_discard_out;
	}
	dst->obsolete = DST_OBSOLETE_DEAD;
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
pre-git
pre-git
42
82.35%
2
33.33%
eric dumazet
eric dumazet
4
7.84%
1
16.67%
alexey kuznetsov
alexey kuznetsov
3
5.88%
1
16.67%
david s. miller
david s. miller
1
1.96%
1
16.67%
eric w. biederman
eric w. biederman
1
1.96%
1
16.67%
Total
51
100.00%
6
100.00%



void __dst_free(struct dst_entry *dst)
{
	spin_lock_bh(&dst_garbage.lock);
	___dst_free(dst);
	dst->next = dst_garbage.list;
	dst_garbage.list = dst;
	if (dst_garbage.timer_inc > DST_GC_INC) {
		dst_garbage.timer_inc = DST_GC_INC;
		dst_garbage.timer_expires = DST_GC_MIN;
		mod_delayed_work(system_wq, &dst_gc_work,
				 dst_garbage.timer_expires);
	}
	spin_unlock_bh(&dst_garbage.lock);
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
pre-git
pre-git
29
36.71%
2
33.33%
eric dumazet
eric dumazet
25
31.65%
1
16.67%
alexey kuznetsov
alexey kuznetsov
19
24.05%
1
16.67%
tejun heo
tejun heo
3
3.80%
1
16.67%
vinay k nallamothu
vinay k nallamothu
3
3.80%
1
16.67%
Total
79
100.00%
6
100.00%


EXPORT_SYMBOL(__dst_free);



struct dst_entry *dst_destroy(struct dst_entry * dst)
{
	struct dst_entry *child;

	smp_rmb();

again:
	child = dst->child;

	if (!(dst->flags & DST_NOCOUNT))
		dst_entries_add(dst->ops, -1);

	if (dst->ops->destroy)
		dst->ops->destroy(dst);
	if (dst->dev)
		dev_put(dst->dev);

	lwtstate_put(dst->lwtstate);

	if (dst->flags & DST_METADATA)
		metadata_dst_free((struct metadata_dst *)dst);
	else
		kmem_cache_free(dst->ops->kmem_cachep, dst);

	dst = child;
	if (dst) {
		int nohash = dst->flags & DST_NOHASH;

		if (atomic_dec_and_test(&dst->__refcnt)) {
			/* We were real parent of this dst, so kill child. */
			if (nohash)
				goto again;
		} else {
			/* Child is still referenced, return it for freeing. */
			if (nohash)
				return dst;
			/* Child is still in his hash table */
		}
	}
	return NULL;
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
pre-git
pre-git
59
34.71%
3
25.00%
alexey kuznetsov
alexey kuznetsov
58
34.12%
2
16.67%
herbert xu
herbert xu
14
8.24%
2
16.67%
thomas graf
thomas graf
13
7.65%
1
8.33%
david s. miller
david s. miller
9
5.29%
1
8.33%
americo wang
americo wang
7
4.12%
1
8.33%
paolo abeni
paolo abeni
6
3.53%
1
8.33%
eric dumazet
eric dumazet
4
2.35%
1
8.33%
Total
170
100.00%
12
100.00%


EXPORT_SYMBOL(dst_destroy);



static void dst_destroy_rcu(struct rcu_head *head)
{
	struct dst_entry *dst = container_of(head, struct dst_entry, rcu_head);

	dst = dst_destroy(dst);
	if (dst)
		__dst_free(dst);
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
eric dumazet
eric dumazet
42
100.00%
1
100.00%
Total
42
100.00%
1
100.00%



void dst_release(struct dst_entry *dst)
{
	if (dst) {
		int newrefcnt;
		unsigned short nocache = dst->flags & DST_NOCACHE;

		newrefcnt = atomic_dec_return(&dst->__refcnt);
		if (unlikely(newrefcnt < 0))
			net_warn_ratelimited("%s: dst:%p refcnt:%d\n",
					     __func__, dst, newrefcnt);
		if (!newrefcnt && unlikely(nocache))
			call_rcu(&dst->rcu_head, dst_destroy_rcu);
	}
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
eric dumazet
eric dumazet
32
40.51%
4
57.14%
ilpo jarvinen
ilpo jarvinen
22
27.85%
1
14.29%
konstantin khlebnikov
konstantin khlebnikov
14
17.72%
1
14.29%
francesco ruggeri
francesco ruggeri
11
13.92%
1
14.29%
Total
79
100.00%
7
100.00%


EXPORT_SYMBOL(dst_release);



u32 *dst_cow_metrics_generic(struct dst_entry *dst, unsigned long old)
{
	u32 *p = kmalloc(sizeof(u32) * RTAX_MAX, GFP_ATOMIC);

	if (p) {
		u32 *old_p = __DST_METRICS_PTR(old);
		unsigned long prev, new;

		memcpy(p, old_p, sizeof(u32) * RTAX_MAX);

		new = (unsigned long) p;
		prev = cmpxchg(&dst->_metrics, old, new);

		if (prev != old) {
			kfree(p);
			p = __DST_METRICS_PTR(prev);
			if (prev & DST_METRICS_READ_ONLY)
				p = NULL;
		}
	}
	return p;
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
david s. miller
david s. miller
121
100.00%
1
100.00%
Total
121
100.00%
1
100.00%


EXPORT_SYMBOL(dst_cow_metrics_generic);

/* Caller asserts that dst_metrics_read_only(dst) is false.  */


void __dst_destroy_metrics_generic(struct dst_entry *dst, unsigned long old)
{
	unsigned long prev, new;

	new = ((unsigned long) dst_default_metrics) | DST_METRICS_READ_ONLY;
	prev = cmpxchg(&dst->_metrics, old, new);
	if (prev == old)
		kfree(__DST_METRICS_PTR(old));
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
david s. miller
david s. miller
56
93.33%
1
50.00%
eric dumazet
eric dumazet
4
6.67%
1
50.00%
Total
60
100.00%
2
100.00%


EXPORT_SYMBOL(__dst_destroy_metrics_generic);


static struct dst_ops md_dst_ops = {
	.family =		AF_UNSPEC,
};



static int dst_md_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	WARN_ONCE(1, "Attempting to call output on metadata dst\n");
	kfree_skb(skb);
	return 0;
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
thomas graf
thomas graf
30
83.33%
1
50.00%
eric w. biederman
eric w. biederman
6
16.67%
1
50.00%
Total
36
100.00%
2
100.00%



static int dst_md_discard(struct sk_buff *skb)
{
	WARN_ONCE(1, "Attempting to call input on metadata dst\n");
	kfree_skb(skb);
	return 0;
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
thomas graf
thomas graf
26
100.00%
1
100.00%
Total
26
100.00%
1
100.00%



static void __metadata_dst_init(struct metadata_dst *md_dst, u8 optslen)
{
	struct dst_entry *dst;

	dst = &md_dst->dst;
	dst_init(dst, &md_dst_ops, NULL, 1, DST_OBSOLETE_NONE,
		 DST_METADATA | DST_NOCACHE | DST_NOCOUNT);

	dst->input = dst_md_discard;
	dst->output = dst_md_discard_out;

	memset(dst + 1, 0, sizeof(*md_dst) + optslen - sizeof(*dst));
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
thomas graf
thomas graf
73
90.12%
1
33.33%
alexei starovoitov
alexei starovoitov
7
8.64%
1
33.33%
eric w. biederman
eric w. biederman
1
1.23%
1
33.33%
Total
81
100.00%
3
100.00%



struct metadata_dst *metadata_dst_alloc(u8 optslen, gfp_t flags)
{
	struct metadata_dst *md_dst;

	md_dst = kmalloc(sizeof(*md_dst) + optslen, flags);
	if (!md_dst)
		return NULL;

	__metadata_dst_init(md_dst, optslen);

	return md_dst;
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
alexei starovoitov
alexei starovoitov
47
92.16%
1
50.00%
thomas graf
thomas graf
4
7.84%
1
50.00%
Total
51
100.00%
2
100.00%


EXPORT_SYMBOL_GPL(metadata_dst_alloc);



void metadata_dst_free(struct metadata_dst *md_dst)
{
#ifdef CONFIG_DST_CACHE
	dst_cache_destroy(&md_dst->u.tun_info.dst_cache);
#endif
	kfree(md_dst);
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
paolo abeni
paolo abeni
32
100.00%
1
100.00%
Total
32
100.00%
1
100.00%



struct metadata_dst __percpu *metadata_dst_alloc_percpu(u8 optslen, gfp_t flags)
{
	int cpu;
	struct metadata_dst __percpu *md_dst;

	md_dst = __alloc_percpu_gfp(sizeof(struct metadata_dst) + optslen,
				    __alignof__(struct metadata_dst), flags);
	if (!md_dst)
		return NULL;

	for_each_possible_cpu(cpu)
		__metadata_dst_init(per_cpu_ptr(md_dst, cpu), optslen);

	return md_dst;
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
alexei starovoitov
alexei starovoitov
71
100.00%
1
100.00%
Total
71
100.00%
1
100.00%


EXPORT_SYMBOL_GPL(metadata_dst_alloc_percpu);

/* Dirty hack. We did it in 2.2 (in __dst_free),
 * we have _very_ good reasons not to repeat
 * this mistake in 2.3, but we have no choice
 * now. _It_ _is_ _explicit_ _deliberate_
 * _race_ _condition_.
 *
 * Commented and originally written by Alexey.
 */


static void dst_ifdown(struct dst_entry *dst, struct net_device *dev,
		       int unregister)
{
	if (dst->ops->ifdown)
		dst->ops->ifdown(dst, dev, unregister);

	if (dev != dst->dev)
		return;

	if (!unregister) {
		dst->input = dst_discard;
		dst->output = dst_discard_out;
	} else {
		dst->dev = dev_net(dst->dev)->loopback_dev;
		dev_hold(dst->dev);
		dev_put(dev);
	}
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
herbert xu
herbert xu
50
52.08%
3
25.00%
pre-git
pre-git
28
29.17%
3
25.00%
hideaki yoshifuji
hideaki yoshifuji
8
8.33%
2
16.67%
denis v. lunev
denis v. lunev
4
4.17%
1
8.33%
daniel lezcano
daniel lezcano
3
3.12%
1
8.33%
eric dumazet
eric dumazet
2
2.08%
1
8.33%
eric w. biederman
eric w. biederman
1
1.04%
1
8.33%
Total
96
100.00%
12
100.00%



static int dst_dev_event(struct notifier_block *this, unsigned long event,
			 void *ptr)
{
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
	struct dst_entry *dst, *last = NULL;

	switch (event) {
	case NETDEV_UNREGISTER_FINAL:
	case NETDEV_DOWN:
		mutex_lock(&dst_gc_mutex);
		for (dst = dst_busy_list; dst; dst = dst->next) {
			last = dst;
			dst_ifdown(dst, dev, event != NETDEV_DOWN);
		}

		spin_lock_bh(&dst_garbage.lock);
		dst = dst_garbage.list;
		dst_garbage.list = NULL;
		spin_unlock_bh(&dst_garbage.lock);

		if (last)
			last->next = dst;
		else
			dst_busy_list = dst;
		for (; dst; dst = dst->next)
			dst_ifdown(dst, dev, event != NETDEV_DOWN);
		mutex_unlock(&dst_gc_mutex);
		break;
	}
	return NOTIFY_DONE;
}

Contributors
Person
Tokens
Prop
Commits
CommitProp
eric dumazet
eric dumazet
79
48.47%
2
28.57%
herbert xu
herbert xu
59
36.20%
2
28.57%
pre-git
pre-git
12
7.36%
1
14.29%
hideaki yoshifuji
hideaki yoshifuji
10
6.13%
1
14.29%
jiri pirko
jiri pirko
3
1.84%
1
14.29%
Total
163
100.00%
7
100.00%


static struct notifier_block dst_dev_notifier = {
	.notifier_call	= dst_dev_event,
	.priority = -10, /* must be called after other network notifiers */
};



void __init dst_subsys_init(void)
{
	register_netdevice_notifier(&dst_dev_notifier);
}
Contributors
Person
Tokens
Prop
Commits
CommitProp
pre-git
pre-git
13
92.86%
2
66.67%
thomas graf
thomas graf
1
7.14%
1
33.33%
Total
14
100.00%
3
100.00%

Overall Contributors
Person
Tokens
Prop
Commits
CommitProp
eric dumazet
eric dumazet
448
21.77%
12
16.44%
pre-git
pre-git
366
17.78%
11
15.07%
david s. miller
david s. miller
343
16.67%
9
12.33%
thomas graf
thomas graf
272
13.22%
1
1.37%
herbert xu
herbert xu
130
6.32%
7
9.59%
alexei starovoitov
alexei starovoitov
130
6.32%
1
1.37%
alexey kuznetsov
alexey kuznetsov
129
6.27%
3
4.11%
paolo abeni
paolo abeni
38
1.85%
1
1.37%
ilpo jarvinen
ilpo jarvinen
27
1.31%
1
1.37%
denis v. lunev
denis v. lunev
25
1.21%
2
2.74%
hideaki yoshifuji
hideaki yoshifuji
24
1.17%
3
4.11%
eric w. biederman
eric w. biederman
20
0.97%
2
2.74%
konstantin khlebnikov
konstantin khlebnikov
14
0.68%
1
1.37%
arjan van de ven
arjan van de ven
13
0.63%
1
1.37%
arnaldo carvalho de melo
arnaldo carvalho de melo
12
0.58%
2
2.74%
laurent chavey
laurent chavey
11
0.53%
1
1.37%
francesco ruggeri
francesco ruggeri
11
0.53%
1
1.37%
jiri benc
jiri benc
9
0.44%
1
1.37%
americo wang
americo wang
7
0.34%
1
1.37%
tejun heo
tejun heo
6
0.29%
2
2.74%
nicolas dichtel
nicolas dichtel
5
0.24%
1
1.37%
linus torvalds
linus torvalds
3
0.15%
1
1.37%
jiri pirko
jiri pirko
3
0.15%
1
1.37%
daniel lezcano
daniel lezcano
3
0.15%
1
1.37%
vinay k nallamothu
vinay k nallamothu
3
0.15%
1
1.37%
andrew morton
andrew morton
2
0.10%
1
1.37%
adrian bunk
adrian bunk
1
0.05%
1
1.37%
alex pankratov
alex pankratov
1
0.05%
1
1.37%
ingo molnar
ingo molnar
1
0.05%
1
1.37%
christoph hellwig
christoph hellwig
1
0.05%
1
1.37%
Total
2058
100.00%
73
100.00%
  
Directory: net/core

Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.