[linux-block.git] / net / ipv6 / ip6_flowlabel.c

/*
 *	ip6_flowlabel.c		IPv6 flowlabel manager.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 *
 *	Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
 */

#include <linux/capability.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/net.h>
#include <linux/netdevice.h>
#include <linux/if_arp.h>
#include <linux/in6.h>
#include <linux/route.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>

#include <net/net_namespace.h>
#include <net/sock.h>

#include <net/ipv6.h>
#include <net/ndisc.h>
#include <net/protocol.h>
#include <net/ip6_route.h>
#include <net/addrconf.h>
#include <net/rawv6.h>
#include <net/icmp.h>
#include <net/transp_v6.h>

#include <asm/uaccess.h>

#define FL_MIN_LINGER	6	/* Minimal linger. It is set to 6sec specified
				   in old IPv6 RFC. Well, it was reasonable value.
				 */
#define FL_MAX_LINGER	60	/* Maximal linger timeout */

/* FL hash table */

#define FL_MAX_PER_SOCK	32
#define FL_MAX_SIZE	4096
#define FL_HASH_MASK	255
#define FL_HASH(l)	(ntohl(l)&FL_HASH_MASK)

static atomic_t fl_size = ATOMIC_INIT(0);
static struct ip6_flowlabel *fl_ht[FL_HASH_MASK+1];

static void ip6_fl_gc(unsigned long dummy);
static DEFINE_TIMER(ip6_fl_gc_timer, ip6_fl_gc, 0, 0);

/* FL hash table lock: it protects only of GC */

static DEFINE_RWLOCK(ip6_fl_lock);

/* Big socket sock */

static DEFINE_RWLOCK(ip6_sk_fl_lock);


static inline struct ip6_flowlabel *__fl_lookup(struct net *net, __be32 label)
{
	struct ip6_flowlabel *fl;

	for (fl=fl_ht[FL_HASH(label)]; fl; fl = fl->next) {
		if (fl->label == label && net_eq(fl->fl_net, net))
			return fl;
	}
	return NULL;
}

static struct ip6_flowlabel *fl_lookup(struct net *net, __be32 label)
{
	struct ip6_flowlabel *fl;

	read_lock_bh(&ip6_fl_lock);
	fl = __fl_lookup(net, label);
	if (fl)
		atomic_inc(&fl->users);
	read_unlock_bh(&ip6_fl_lock);
	return fl;
}


static void fl_free(struct ip6_flowlabel *fl)
{
	if (fl) {
		release_net(fl->fl_net);
		kfree(fl->opt);
	}
	kfree(fl);
}

static void fl_release(struct ip6_flowlabel *fl)
{
	write_lock_bh(&ip6_fl_lock);

	fl->lastuse = jiffies;
	if (atomic_dec_and_test(&fl->users)) {
		unsigned long ttd = fl->lastuse + fl->linger;
		if (time_after(ttd, fl->expires))
			fl->expires = ttd;
		ttd = fl->expires;
		if (fl->opt && fl->share == IPV6_FL_S_EXCL) {
			struct ipv6_txoptions *opt = fl->opt;
			fl->opt = NULL;
			kfree(opt);
		}
		if (!timer_pending(&ip6_fl_gc_timer) ||
		    time_after(ip6_fl_gc_timer.expires, ttd))
			mod_timer(&ip6_fl_gc_timer, ttd);
	}
	write_unlock_bh(&ip6_fl_lock);
}

static void ip6_fl_gc(unsigned long dummy)
{
	int i;
	unsigned long now = jiffies;
	unsigned long sched = 0;

	write_lock(&ip6_fl_lock);

	for (i=0; i<=FL_HASH_MASK; i++) {
		struct ip6_flowlabel *fl, **flp;
		flp = &fl_ht[i];
		while ((fl=*flp) != NULL) {
			if (atomic_read(&fl->users) == 0) {
				unsigned long ttd = fl->lastuse + fl->linger;
				if (time_after(ttd, fl->expires))
					fl->expires = ttd;
				ttd = fl->expires;
				if (time_after_eq(now, ttd)) {
					*flp = fl->next;
					fl_free(fl);
					atomic_dec(&fl_size);
					continue;
				}
				if (!sched || time_before(ttd, sched))
					sched = ttd;
			}
			flp = &fl->next;
		}
	}
	if (!sched && atomic_read(&fl_size))
		sched = now + FL_MAX_LINGER;
	if (sched) {
		mod_timer(&ip6_fl_gc_timer, sched);
	}
	write_unlock(&ip6_fl_lock);
}

static void __net_exit ip6_fl_purge(struct net *net)
{
	int i;

	write_lock(&ip6_fl_lock);
	for (i = 0; i <= FL_HASH_MASK; i++) {
		struct ip6_flowlabel *fl, **flp;
		flp = &fl_ht[i];
		while ((fl = *flp) != NULL) {
			if (net_eq(fl->fl_net, net) &&
			    atomic_read(&fl->users) == 0) {
				*flp = fl->next;
				fl_free(fl);
				atomic_dec(&fl_size);
				continue;
			}
			flp = &fl->next;
		}
	}
	write_unlock(&ip6_fl_lock);
}

static struct ip6_flowlabel *fl_intern(struct net *net,
				       struct ip6_flowlabel *fl, __be32 label)
{
	struct ip6_flowlabel *lfl;

	fl->label = label & IPV6_FLOWLABEL_MASK;

	write_lock_bh(&ip6_fl_lock);
	if (label == 0) {
		for (;;) {
			fl->label = htonl(net_random())&IPV6_FLOWLABEL_MASK;
			if (fl->label) {
				lfl = __fl_lookup(net, fl->label);
				if (lfl == NULL)
					break;
			}
		}
	} else {
		/*
		 * we dropper the ip6_fl_lock, so this entry could reappear
		 * and we need to recheck with it.
		 *
		 * OTOH no need to search the active socket first, like it is
		 * done in ipv6_flowlabel_opt - sock is locked, so new entry
		 * with the same label can only appear on another sock
		 */
		lfl = __fl_lookup(net, fl->label);
		if (lfl != NULL) {
			atomic_inc(&lfl->users);
			write_unlock_bh(&ip6_fl_lock);
			return lfl;
		}
	}

	fl->lastuse = jiffies;
	fl->next = fl_ht[FL_HASH(fl->label)];
	fl_ht[FL_HASH(fl->label)] = fl;
	atomic_inc(&fl_size);
	write_unlock_bh(&ip6_fl_lock);
	return NULL;
}


/* Socket flowlabel lists */

struct ip6_flowlabel * fl6_sock_lookup(struct sock *sk, __be32 label)
{
	struct ipv6_fl_socklist *sfl;
	struct ipv6_pinfo *np = inet6_sk(sk);

	label &= IPV6_FLOWLABEL_MASK;

	read_lock_bh(&ip6_sk_fl_lock);
	for (sfl=np->ipv6_fl_list; sfl; sfl = sfl->next) {
		struct ip6_flowlabel *fl = sfl->fl;
		if (fl->label == label) {
			fl->lastuse = jiffies;
			atomic_inc(&fl->users);
			read_unlock_bh(&ip6_sk_fl_lock);
			return fl;
		}
	}
	read_unlock_bh(&ip6_sk_fl_lock);
	return NULL;
}

EXPORT_SYMBOL_GPL(fl6_sock_lookup);

void fl6_free_socklist(struct sock *sk)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct ipv6_fl_socklist *sfl;

	while ((sfl = np->ipv6_fl_list) != NULL) {
		np->ipv6_fl_list = sfl->next;
		fl_release(sfl->fl);
		kfree(sfl);
	}
}

/* Service routines */


/*
   It is the only difficult place. flowlabel enforces equal headers
   before and including routing header, however user may supply options
   following rthdr.
 */

struct ipv6_txoptions *fl6_merge_options(struct ipv6_txoptions * opt_space,
					 struct ip6_flowlabel * fl,
					 struct ipv6_txoptions * fopt)
{
	struct ipv6_txoptions * fl_opt = fl->opt;

	if (fopt == NULL || fopt->opt_flen == 0)
		return fl_opt;

	if (fl_opt != NULL) {
		opt_space->hopopt = fl_opt->hopopt;
		opt_space->dst0opt = fl_opt->dst0opt;
		opt_space->srcrt = fl_opt->srcrt;
		opt_space->opt_nflen = fl_opt->opt_nflen;
	} else {
		if (fopt->opt_nflen == 0)
			return fopt;
		opt_space->hopopt = NULL;
		opt_space->dst0opt = NULL;
		opt_space->srcrt = NULL;
		opt_space->opt_nflen = 0;
	}
	opt_space->dst1opt = fopt->dst1opt;
	opt_space->opt_flen = fopt->opt_flen;
	return opt_space;
}

static unsigned long check_linger(unsigned long ttl)
{
	if (ttl < FL_MIN_LINGER)
		return FL_MIN_LINGER*HZ;
	if (ttl > FL_MAX_LINGER && !capable(CAP_NET_ADMIN))
		return 0;
	return ttl*HZ;
}

static int fl6_renew(struct ip6_flowlabel *fl, unsigned long linger, unsigned long expires)
{
	linger = check_linger(linger);
	if (!linger)
		return -EPERM;
	expires = check_linger(expires);
	if (!expires)
		return -EPERM;
	fl->lastuse = jiffies;
	if (time_before(fl->linger, linger))
		fl->linger = linger;
	if (time_before(expires, fl->linger))
		expires = fl->linger;
	if (time_before(fl->expires, fl->lastuse + expires))
		fl->expires = fl->lastuse + expires;
	return 0;
}

static struct ip6_flowlabel *
fl_create(struct net *net, struct in6_flowlabel_req *freq, char __user *optval,
	  int optlen, int *err_p)
{
	struct ip6_flowlabel *fl = NULL;
	int olen;
	int addr_type;
	int err;

	olen = optlen - CMSG_ALIGN(sizeof(*freq));
	err = -EINVAL;
	if (olen > 64 * 1024)
		goto done;

	err = -ENOMEM;
	fl = kzalloc(sizeof(*fl), GFP_KERNEL);
	if (fl == NULL)
		goto done;

	if (olen > 0) {
		struct msghdr msg;
		struct flowi flowi;
		int junk;

		err = -ENOMEM;
		fl->opt = kmalloc(sizeof(*fl->opt) + olen, GFP_KERNEL);
		if (fl->opt == NULL)
			goto done;

		memset(fl->opt, 0, sizeof(*fl->opt));
		fl->opt->tot_len = sizeof(*fl->opt) + olen;
		err = -EFAULT;
		if (copy_from_user(fl->opt+1, optval+CMSG_ALIGN(sizeof(*freq)), olen))
			goto done;

		msg.msg_controllen = olen;
		msg.msg_control = (void*)(fl->opt+1);
		flowi.oif = 0;

		err = datagram_send_ctl(net, &msg, &flowi, fl->opt, &junk, &junk);
		if (err)
			goto done;
		err = -EINVAL;
		if (fl->opt->opt_flen)
			goto done;
		if (fl->opt->opt_nflen == 0) {
			kfree(fl->opt);
			fl->opt = NULL;
		}
	}

	fl->fl_net = hold_net(net);
	fl->expires = jiffies;
	err = fl6_renew(fl, freq->flr_linger, freq->flr_expires);
	if (err)
		goto done;
	fl->share = freq->flr_share;
	addr_type = ipv6_addr_type(&freq->flr_dst);
	if ((addr_type & IPV6_ADDR_MAPPED) ||
	    addr_type == IPV6_ADDR_ANY) {
		err = -EINVAL;
		goto done;
	}
	ipv6_addr_copy(&fl->dst, &freq->flr_dst);
	atomic_set(&fl->users, 1);
	switch (fl->share) {
	case IPV6_FL_S_EXCL:
	case IPV6_FL_S_ANY:
		break;
	case IPV6_FL_S_PROCESS:
		fl->owner = current->pid;
		break;
	case IPV6_FL_S_USER:
		fl->owner = current_euid();
		break;
	default:
		err = -EINVAL;
		goto done;
	}
	return fl;

done:
	fl_free(fl);
	*err_p = err;
	return NULL;
}

static int mem_check(struct sock *sk)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct ipv6_fl_socklist *sfl;
	int room = FL_MAX_SIZE - atomic_read(&fl_size);
	int count = 0;

	if (room > FL_MAX_SIZE - FL_MAX_PER_SOCK)
		return 0;

	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next)
		count++;

	if (room <= 0 ||
	    ((count >= FL_MAX_PER_SOCK ||
	      (count > 0 && room < FL_MAX_SIZE/2) || room < FL_MAX_SIZE/4) &&
	     !capable(CAP_NET_ADMIN)))
		return -ENOBUFS;

	return 0;
}

static int ipv6_hdr_cmp(struct ipv6_opt_hdr *h1, struct ipv6_opt_hdr *h2)
{
	if (h1 == h2)
		return 0;
	if (h1 == NULL || h2 == NULL)
		return 1;
	if (h1->hdrlen != h2->hdrlen)
		return 1;
	return memcmp(h1+1, h2+1, ((h1->hdrlen+1)<<3) - sizeof(*h1));
}

static int ipv6_opt_cmp(struct ipv6_txoptions *o1, struct ipv6_txoptions *o2)
{
	if (o1 == o2)
		return 0;
	if (o1 == NULL || o2 == NULL)
		return 1;
	if (o1->opt_nflen != o2->opt_nflen)
		return 1;
	if (ipv6_hdr_cmp(o1->hopopt, o2->hopopt))
		return 1;
	if (ipv6_hdr_cmp(o1->dst0opt, o2->dst0opt))
		return 1;
	if (ipv6_hdr_cmp((struct ipv6_opt_hdr *)o1->srcrt, (struct ipv6_opt_hdr *)o2->srcrt))
		return 1;
	return 0;
}

static inline void fl_link(struct ipv6_pinfo *np, struct ipv6_fl_socklist *sfl,
		struct ip6_flowlabel *fl)
{
	write_lock_bh(&ip6_sk_fl_lock);
	sfl->fl = fl;
	sfl->next = np->ipv6_fl_list;
	np->ipv6_fl_list = sfl;
	write_unlock_bh(&ip6_sk_fl_lock);
}

int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen)
{
	int uninitialized_var(err);
	struct net *net = sock_net(sk);
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct in6_flowlabel_req freq;
	struct ipv6_fl_socklist *sfl1=NULL;
	struct ipv6_fl_socklist *sfl, **sflp;
	struct ip6_flowlabel *fl, *fl1 = NULL;


	if (optlen < sizeof(freq))
		return -EINVAL;

	if (copy_from_user(&freq, optval, sizeof(freq)))
		return -EFAULT;

	switch (freq.flr_action) {
	case IPV6_FL_A_PUT:
		write_lock_bh(&ip6_sk_fl_lock);
		for (sflp = &np->ipv6_fl_list; (sfl=*sflp)!=NULL; sflp = &sfl->next) {
			if (sfl->fl->label == freq.flr_label) {
				if (freq.flr_label == (np->flow_label&IPV6_FLOWLABEL_MASK))
					np->flow_label &= ~IPV6_FLOWLABEL_MASK;
				*sflp = sfl->next;
				write_unlock_bh(&ip6_sk_fl_lock);
				fl_release(sfl->fl);
				kfree(sfl);
				return 0;
			}
		}
		write_unlock_bh(&ip6_sk_fl_lock);
		return -ESRCH;

	case IPV6_FL_A_RENEW:
		read_lock_bh(&ip6_sk_fl_lock);
		for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) {
			if (sfl->fl->label == freq.flr_label) {
				err = fl6_renew(sfl->fl, freq.flr_linger, freq.flr_expires);
				read_unlock_bh(&ip6_sk_fl_lock);
				return err;
			}
		}
		read_unlock_bh(&ip6_sk_fl_lock);

		if (freq.flr_share == IPV6_FL_S_NONE && capable(CAP_NET_ADMIN)) {
			fl = fl_lookup(net, freq.flr_label);
			if (fl) {
				err = fl6_renew(fl, freq.flr_linger, freq.flr_expires);
				fl_release(fl);
				return err;
			}
		}
		return -ESRCH;

	case IPV6_FL_A_GET:
		if (freq.flr_label & ~IPV6_FLOWLABEL_MASK)
			return -EINVAL;

		fl = fl_create(net, &freq, optval, optlen, &err);
		if (fl == NULL)
			return err;
		sfl1 = kmalloc(sizeof(*sfl1), GFP_KERNEL);

		if (freq.flr_label) {
			err = -EEXIST;
			read_lock_bh(&ip6_sk_fl_lock);
			for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) {
				if (sfl->fl->label == freq.flr_label) {
					if (freq.flr_flags&IPV6_FL_F_EXCL) {
						read_unlock_bh(&ip6_sk_fl_lock);
						goto done;
					}
					fl1 = sfl->fl;
					atomic_inc(&fl1->users);
					break;
				}
			}
			read_unlock_bh(&ip6_sk_fl_lock);

			if (fl1 == NULL)
				fl1 = fl_lookup(net, freq.flr_label);
			if (fl1) {
recheck:
				err = -EEXIST;
				if (freq.flr_flags&IPV6_FL_F_EXCL)
					goto release;
				err = -EPERM;
				if (fl1->share == IPV6_FL_S_EXCL ||
				    fl1->share != fl->share ||
				    fl1->owner != fl->owner)
					goto release;

				err = -EINVAL;
				if (!ipv6_addr_equal(&fl1->dst, &fl->dst) ||
				    ipv6_opt_cmp(fl1->opt, fl->opt))
					goto release;

				err = -ENOMEM;
				if (sfl1 == NULL)
					goto release;
				if (fl->linger > fl1->linger)
					fl1->linger = fl->linger;
				if ((long)(fl->expires - fl1->expires) > 0)
					fl1->expires = fl->expires;
				fl_link(np, sfl1, fl1);
				fl_free(fl);
				return 0;

release:
				fl_release(fl1);
				goto done;
			}
		}
		err = -ENOENT;
		if (!(freq.flr_flags&IPV6_FL_F_CREATE))
			goto done;

		err = -ENOMEM;
		if (sfl1 == NULL || (err = mem_check(sk)) != 0)
			goto done;

		fl1 = fl_intern(net, fl, freq.flr_label);
		if (fl1 != NULL)
			goto recheck;

		if (!freq.flr_label) {
			if (copy_to_user(&((struct in6_flowlabel_req __user *) optval)->flr_label,
					 &fl->label, sizeof(fl->label))) {
				/* Intentionally ignore fault. */
			}
		}

		fl_link(np, sfl1, fl);
		return 0;

	default:
		return -EINVAL;
	}

done:
	fl_free(fl);
	kfree(sfl1);
	return err;
}

#ifdef CONFIG_PROC_FS

struct ip6fl_iter_state {
	struct seq_net_private p;
	int bucket;
};

#define ip6fl_seq_private(seq)	((struct ip6fl_iter_state *)(seq)->private)

static struct ip6_flowlabel *ip6fl_get_first(struct seq_file *seq)
{
	struct ip6_flowlabel *fl = NULL;
	struct ip6fl_iter_state *state = ip6fl_seq_private(seq);
	struct net *net = seq_file_net(seq);

	for (state->bucket = 0; state->bucket <= FL_HASH_MASK; ++state->bucket) {
		fl = fl_ht[state->bucket];

		while (fl && !net_eq(fl->fl_net, net))
			fl = fl->next;
		if (fl)
			break;
	}
	return fl;
}

static struct ip6_flowlabel *ip6fl_get_next(struct seq_file *seq, struct ip6_flowlabel *fl)
{
	struct ip6fl_iter_state *state = ip6fl_seq_private(seq);
	struct net *net = seq_file_net(seq);

	fl = fl->next;
try_again:
	while (fl && !net_eq(fl->fl_net, net))
		fl = fl->next;

	while (!fl) {
		if (++state->bucket <= FL_HASH_MASK) {
			fl = fl_ht[state->bucket];
			goto try_again;
		} else
			break;
	}
	return fl;
}

static struct ip6_flowlabel *ip6fl_get_idx(struct seq_file *seq, loff_t pos)
{
	struct ip6_flowlabel *fl = ip6fl_get_first(seq);
	if (fl)
		while (pos && (fl = ip6fl_get_next(seq, fl)) != NULL)
			--pos;
	return pos ? NULL : fl;
}

static void *ip6fl_seq_start(struct seq_file *seq, loff_t *pos)
	__acquires(ip6_fl_lock)
{
	read_lock_bh(&ip6_fl_lock);
	return *pos ? ip6fl_get_idx(seq, *pos - 1) : SEQ_START_TOKEN;
}

static void *ip6fl_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct ip6_flowlabel *fl;

	if (v == SEQ_START_TOKEN)
		fl = ip6fl_get_first(seq);
	else
		fl = ip6fl_get_next(seq, v);
	++*pos;
	return fl;
}

static void ip6fl_seq_stop(struct seq_file *seq, void *v)
	__releases(ip6_fl_lock)
{
	read_unlock_bh(&ip6_fl_lock);
}

static int ip6fl_seq_show(struct seq_file *seq, void *v)
{
	if (v == SEQ_START_TOKEN)
		seq_printf(seq, "%-5s %-1s %-6s %-6s %-6s %-8s %-32s %s\n",
			   "Label", "S", "Owner", "Users", "Linger", "Expires", "Dst", "Opt");
	else {
		struct ip6_flowlabel *fl = v;
		seq_printf(seq,
			   "%05X %-1d %-6d %-6d %-6ld %-8ld %pi6 %-4d\n",
			   (unsigned)ntohl(fl->label),
			   fl->share,
			   (unsigned)fl->owner,
			   atomic_read(&fl->users),
			   fl->linger/HZ,
			   (long)(fl->expires - jiffies)/HZ,
			   &fl->dst,
			   fl->opt ? fl->opt->opt_nflen : 0);
	}
	return 0;
}

static const struct seq_operations ip6fl_seq_ops = {
	.start	=	ip6fl_seq_start,
	.next	=	ip6fl_seq_next,
	.stop	=	ip6fl_seq_stop,
	.show	=	ip6fl_seq_show,
};

static int ip6fl_seq_open(struct inode *inode, struct file *file)
{
	return seq_open_net(inode, file, &ip6fl_seq_ops,
			    sizeof(struct ip6fl_iter_state));
}

static const struct file_operations ip6fl_seq_fops = {
	.owner		=	THIS_MODULE,
	.open		=	ip6fl_seq_open,
	.read		=	seq_read,
	.llseek		=	seq_lseek,
	.release	=	seq_release_net,
};

static int __net_init ip6_flowlabel_proc_init(struct net *net)
{
	if (!proc_net_fops_create(net, "ip6_flowlabel",
				  S_IRUGO, &ip6fl_seq_fops))
		return -ENOMEM;
	return 0;
}

static void __net_exit ip6_flowlabel_proc_fini(struct net *net)
{
	proc_net_remove(net, "ip6_flowlabel");
}
#else
static inline int ip6_flowlabel_proc_init(struct net *net)
{
	return 0;
}
static inline void ip6_flowlabel_proc_fini(struct net *net)
{
}
#endif

static void __net_exit ip6_flowlabel_net_exit(struct net *net)
{
	ip6_fl_purge(net);
	ip6_flowlabel_proc_fini(net);
}

static struct pernet_operations ip6_flowlabel_net_ops = {
	.init = ip6_flowlabel_proc_init,
	.exit = ip6_flowlabel_net_exit,
};

int ip6_flowlabel_init(void)
{
	return register_pernet_subsys(&ip6_flowlabel_net_ops);
}

void ip6_flowlabel_cleanup(void)
{
	del_timer(&ip6_fl_gc_timer);
	unregister_pernet_subsys(&ip6_flowlabel_net_ops);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* ip6_flowlabel.c IPv6 flowlabel manager.
	3	*
	4	* This program is free software; you can redistribute it and/or
	5	* modify it under the terms of the GNU General Public License
	6	* as published by the Free Software Foundation; either version
	7	* 2 of the License, or (at your option) any later version.
	8	*
	9	* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
	10	*/
	11
4fc268d2	12	#include <linux/capability.h>
1da177e4 LT	13	#include <linux/errno.h>
	14	#include <linux/types.h>
	15	#include <linux/socket.h>
	16	#include <linux/net.h>
	17	#include <linux/netdevice.h>
	18	#include <linux/if_arp.h>
	19	#include <linux/in6.h>
	20	#include <linux/route.h>
	21	#include <linux/proc_fs.h>
	22	#include <linux/seq_file.h>
	23
457c4cbc	24	#include <net/net_namespace.h>
1da177e4 LT	25	#include <net/sock.h>
	26
	27	#include <net/ipv6.h>
	28	#include <net/ndisc.h>
	29	#include <net/protocol.h>
	30	#include <net/ip6_route.h>
	31	#include <net/addrconf.h>
	32	#include <net/rawv6.h>
	33	#include <net/icmp.h>
	34	#include <net/transp_v6.h>
	35
	36	#include <asm/uaccess.h>
	37
	38	#define FL_MIN_LINGER 6 /* Minimal linger. It is set to 6sec specified
	39	in old IPv6 RFC. Well, it was reasonable value.
	40	*/
	41	#define FL_MAX_LINGER 60 /* Maximal linger timeout */
	42
	43	/* FL hash table */
	44
	45	#define FL_MAX_PER_SOCK 32
	46	#define FL_MAX_SIZE 4096
	47	#define FL_HASH_MASK 255
	48	#define FL_HASH(l) (ntohl(l)&FL_HASH_MASK)
	49
	50	static atomic_t fl_size = ATOMIC_INIT(0);
	51	static struct ip6_flowlabel *fl_ht[FL_HASH_MASK+1];
	52
	53	static void ip6_fl_gc(unsigned long dummy);
8d06afab	54	static DEFINE_TIMER(ip6_fl_gc_timer, ip6_fl_gc, 0, 0);
1da177e4 LT	55
	56	/* FL hash table lock: it protects only of GC */
	57
	58	static DEFINE_RWLOCK(ip6_fl_lock);
	59
	60	/* Big socket sock */
	61
	62	static DEFINE_RWLOCK(ip6_sk_fl_lock);
	63
	64
60e8fbc4	65	static inline struct ip6_flowlabel __fl_lookup(struct net net, __be32 label)
1da177e4 LT	66	{
	67	struct ip6_flowlabel *fl;
	68
	69	for (fl=fl_ht[FL_HASH(label)]; fl; fl = fl->next) {
09ad9bc7	70	if (fl->label == label && net_eq(fl->fl_net, net))
1da177e4 LT	71	return fl;
	72	}
	73	return NULL;
	74	}
	75
60e8fbc4	76	static struct ip6_flowlabel fl_lookup(struct net net, __be32 label)
1da177e4 LT	77	{
	78	struct ip6_flowlabel *fl;
	79
	80	read_lock_bh(&ip6_fl_lock);
60e8fbc4	81	fl = __fl_lookup(net, label);
1da177e4 LT	82	if (fl)
	83	atomic_inc(&fl->users);
	84	read_unlock_bh(&ip6_fl_lock);
	85	return fl;
	86	}
	87
	88
	89	static void fl_free(struct ip6_flowlabel *fl)
	90	{
60e8fbc4 BT	91	if (fl) {
60e8fbc4 BT	92	release_net(fl->fl_net);
1da177e4	93	kfree(fl->opt);
60e8fbc4	94	}
1da177e4 LT	95	kfree(fl);
	96	}
	97
	98	static void fl_release(struct ip6_flowlabel *fl)
	99	{
	100	write_lock_bh(&ip6_fl_lock);
	101
	102	fl->lastuse = jiffies;
	103	if (atomic_dec_and_test(&fl->users)) {
	104	unsigned long ttd = fl->lastuse + fl->linger;
	105	if (time_after(ttd, fl->expires))
	106	fl->expires = ttd;
	107	ttd = fl->expires;
	108	if (fl->opt && fl->share == IPV6_FL_S_EXCL) {
	109	struct ipv6_txoptions *opt = fl->opt;
	110	fl->opt = NULL;
	111	kfree(opt);
	112	}
	113	if (!timer_pending(&ip6_fl_gc_timer) \|\|
	114	time_after(ip6_fl_gc_timer.expires, ttd))
	115	mod_timer(&ip6_fl_gc_timer, ttd);
	116	}
1da177e4 LT	117	write_unlock_bh(&ip6_fl_lock);
	118	}
	119
	120	static void ip6_fl_gc(unsigned long dummy)
	121	{
	122	int i;
	123	unsigned long now = jiffies;
	124	unsigned long sched = 0;
	125
	126	write_lock(&ip6_fl_lock);
	127
	128	for (i=0; i<=FL_HASH_MASK; i++) {
	129	struct ip6_flowlabel fl, *flp;
	130	flp = &fl_ht[i];
	131	while ((fl=*flp) != NULL) {
	132	if (atomic_read(&fl->users) == 0) {
	133	unsigned long ttd = fl->lastuse + fl->linger;
	134	if (time_after(ttd, fl->expires))
	135	fl->expires = ttd;
	136	ttd = fl->expires;
	137	if (time_after_eq(now, ttd)) {
	138	*flp = fl->next;
	139	fl_free(fl);
	140	atomic_dec(&fl_size);
	141	continue;
	142	}
	143	if (!sched \|\| time_before(ttd, sched))
	144	sched = ttd;
	145	}
	146	flp = &fl->next;
	147	}
	148	}
	149	if (!sched && atomic_read(&fl_size))
	150	sched = now + FL_MAX_LINGER;
	151	if (sched) {
60e8fbc4	152	mod_timer(&ip6_fl_gc_timer, sched);
1da177e4 LT	153	}
	154	write_unlock(&ip6_fl_lock);
	155	}
	156
2c8c1e72	157	static void __net_exit ip6_fl_purge(struct net *net)
60e8fbc4 BT	158	{
	159	int i;
	160
	161	write_lock(&ip6_fl_lock);
	162	for (i = 0; i <= FL_HASH_MASK; i++) {
	163	struct ip6_flowlabel fl, *flp;
	164	flp = &fl_ht[i];
	165	while ((fl = *flp) != NULL) {
09ad9bc7 OP	166	if (net_eq(fl->fl_net, net) &&
09ad9bc7 OP	167	atomic_read(&fl->users) == 0) {
60e8fbc4 BT	168	*flp = fl->next;
	169	fl_free(fl);
	170	atomic_dec(&fl_size);
	171	continue;
	172	}
	173	flp = &fl->next;
	174	}
	175	}
	176	write_unlock(&ip6_fl_lock);
	177	}
	178
	179	static struct ip6_flowlabel fl_intern(struct net net,
	180	struct ip6_flowlabel *fl, __be32 label)
1da177e4	181	{
78c2e502 PE	182	struct ip6_flowlabel *lfl;
78c2e502 PE	183
1da177e4 LT	184	fl->label = label & IPV6_FLOWLABEL_MASK;
	185
	186	write_lock_bh(&ip6_fl_lock);
	187	if (label == 0) {
	188	for (;;) {
	189	fl->label = htonl(net_random())&IPV6_FLOWLABEL_MASK;
	190	if (fl->label) {
60e8fbc4	191	lfl = __fl_lookup(net, fl->label);
1da177e4 LT	192	if (lfl == NULL)
	193	break;
	194	}
	195	}
78c2e502 PE	196	} else {
	197	/*
	198	* we dropper the ip6_fl_lock, so this entry could reappear
	199	* and we need to recheck with it.
	200	*
	201	* OTOH no need to search the active socket first, like it is
	202	* done in ipv6_flowlabel_opt - sock is locked, so new entry
	203	* with the same label can only appear on another sock
	204	*/
60e8fbc4	205	lfl = __fl_lookup(net, fl->label);
78c2e502 PE	206	if (lfl != NULL) {
	207	atomic_inc(&lfl->users);
	208	write_unlock_bh(&ip6_fl_lock);
	209	return lfl;
	210	}
1da177e4 LT	211	}
	212
	213	fl->lastuse = jiffies;
	214	fl->next = fl_ht[FL_HASH(fl->label)];
	215	fl_ht[FL_HASH(fl->label)] = fl;
	216	atomic_inc(&fl_size);
	217	write_unlock_bh(&ip6_fl_lock);
78c2e502	218	return NULL;
1da177e4 LT	219	}
	220
	221
	222
	223	/* Socket flowlabel lists */
	224
90bcaf7b	225	struct ip6_flowlabel * fl6_sock_lookup(struct sock *sk, __be32 label)
1da177e4 LT	226	{
	227	struct ipv6_fl_socklist *sfl;
	228	struct ipv6_pinfo *np = inet6_sk(sk);
	229
	230	label &= IPV6_FLOWLABEL_MASK;
	231
bd0bf577	232	read_lock_bh(&ip6_sk_fl_lock);
1da177e4 LT	233	for (sfl=np->ipv6_fl_list; sfl; sfl = sfl->next) {
	234	struct ip6_flowlabel *fl = sfl->fl;
	235	if (fl->label == label) {
	236	fl->lastuse = jiffies;
	237	atomic_inc(&fl->users);
52f095ee	238	read_unlock_bh(&ip6_sk_fl_lock);
1da177e4 LT	239	return fl;
	240	}
	241	}
bd0bf577	242	read_unlock_bh(&ip6_sk_fl_lock);
1da177e4 LT	243	return NULL;
	244	}
	245
3cf3dc6c ACM	246	EXPORT_SYMBOL_GPL(fl6_sock_lookup);
3cf3dc6c ACM	247
1da177e4 LT	248	void fl6_free_socklist(struct sock *sk)
	249	{
	250	struct ipv6_pinfo *np = inet6_sk(sk);
	251	struct ipv6_fl_socklist *sfl;
	252
	253	while ((sfl = np->ipv6_fl_list) != NULL) {
	254	np->ipv6_fl_list = sfl->next;
	255	fl_release(sfl->fl);
	256	kfree(sfl);
	257	}
	258	}
	259
	260	/* Service routines */
	261
	262
	263	/*
	264	It is the only difficult place. flowlabel enforces equal headers
	265	before and including routing header, however user may supply options
	266	following rthdr.
	267	*/
	268
	269	struct ipv6_txoptions fl6_merge_options(struct ipv6_txoptions opt_space,
	270	struct ip6_flowlabel * fl,
	271	struct ipv6_txoptions * fopt)
	272	{
df9890c3	273	struct ipv6_txoptions * fl_opt = fl->opt;
1ab1457c	274
df9890c3 YH	275	if (fopt == NULL \|\| fopt->opt_flen == 0)
df9890c3 YH	276	return fl_opt;
1ab1457c	277
1da177e4 LT	278	if (fl_opt != NULL) {
1da177e4 LT	279	opt_space->hopopt = fl_opt->hopopt;
df9890c3	280	opt_space->dst0opt = fl_opt->dst0opt;
1da177e4 LT	281	opt_space->srcrt = fl_opt->srcrt;
	282	opt_space->opt_nflen = fl_opt->opt_nflen;
	283	} else {
	284	if (fopt->opt_nflen == 0)
	285	return fopt;
	286	opt_space->hopopt = NULL;
	287	opt_space->dst0opt = NULL;
	288	opt_space->srcrt = NULL;
	289	opt_space->opt_nflen = 0;
	290	}
	291	opt_space->dst1opt = fopt->dst1opt;
1da177e4 LT	292	opt_space->opt_flen = fopt->opt_flen;
	293	return opt_space;
	294	}
	295
	296	static unsigned long check_linger(unsigned long ttl)
	297	{
	298	if (ttl < FL_MIN_LINGER)
	299	return FL_MIN_LINGER*HZ;
	300	if (ttl > FL_MAX_LINGER && !capable(CAP_NET_ADMIN))
	301	return 0;
	302	return ttl*HZ;
	303	}
	304
	305	static int fl6_renew(struct ip6_flowlabel *fl, unsigned long linger, unsigned long expires)
	306	{
	307	linger = check_linger(linger);
	308	if (!linger)
	309	return -EPERM;
	310	expires = check_linger(expires);
	311	if (!expires)
	312	return -EPERM;
	313	fl->lastuse = jiffies;
	314	if (time_before(fl->linger, linger))
	315	fl->linger = linger;
	316	if (time_before(expires, fl->linger))
	317	expires = fl->linger;
	318	if (time_before(fl->expires, fl->lastuse + expires))
	319	fl->expires = fl->lastuse + expires;
	320	return 0;
	321	}
	322
	323	static struct ip6_flowlabel *
60e8fbc4 BT	324	fl_create(struct net net, struct in6_flowlabel_req freq, char __user *optval,
60e8fbc4 BT	325	int optlen, int *err_p)
1da177e4	326	{
684de409	327	struct ip6_flowlabel *fl = NULL;
1da177e4 LT	328	int olen;
	329	int addr_type;
	330	int err;
	331
684de409 DM	332	olen = optlen - CMSG_ALIGN(sizeof(*freq));
	333	err = -EINVAL;
	334	if (olen > 64 * 1024)
	335	goto done;
	336
1da177e4	337	err = -ENOMEM;
0c600eda	338	fl = kzalloc(sizeof(*fl), GFP_KERNEL);
1da177e4 LT	339	if (fl == NULL)
1da177e4 LT	340	goto done;
1da177e4	341
1da177e4 LT	342	if (olen > 0) {
	343	struct msghdr msg;
	344	struct flowi flowi;
	345	int junk;
	346
	347	err = -ENOMEM;
	348	fl->opt = kmalloc(sizeof(*fl->opt) + olen, GFP_KERNEL);
	349	if (fl->opt == NULL)
	350	goto done;
	351
	352	memset(fl->opt, 0, sizeof(*fl->opt));
	353	fl->opt->tot_len = sizeof(*fl->opt) + olen;
	354	err = -EFAULT;
	355	if (copy_from_user(fl->opt+1, optval+CMSG_ALIGN(sizeof(*freq)), olen))
	356	goto done;
	357
	358	msg.msg_controllen = olen;
	359	msg.msg_control = (void*)(fl->opt+1);
	360	flowi.oif = 0;
	361
91e1908f	362	err = datagram_send_ctl(net, &msg, &flowi, fl->opt, &junk, &junk);
1da177e4 LT	363	if (err)
	364	goto done;
	365	err = -EINVAL;
	366	if (fl->opt->opt_flen)
	367	goto done;
	368	if (fl->opt->opt_nflen == 0) {
	369	kfree(fl->opt);
	370	fl->opt = NULL;
	371	}
	372	}
	373
60e8fbc4	374	fl->fl_net = hold_net(net);
1da177e4 LT	375	fl->expires = jiffies;
	376	err = fl6_renew(fl, freq->flr_linger, freq->flr_expires);
	377	if (err)
	378	goto done;
	379	fl->share = freq->flr_share;
	380	addr_type = ipv6_addr_type(&freq->flr_dst);
35700212 JP	381	if ((addr_type & IPV6_ADDR_MAPPED) \|\|
35700212 JP	382	addr_type == IPV6_ADDR_ANY) {
c6817e4c	383	err = -EINVAL;
1da177e4	384	goto done;
c6817e4c	385	}
1da177e4 LT	386	ipv6_addr_copy(&fl->dst, &freq->flr_dst);
	387	atomic_set(&fl->users, 1);
	388	switch (fl->share) {
	389	case IPV6_FL_S_EXCL:
	390	case IPV6_FL_S_ANY:
	391	break;
	392	case IPV6_FL_S_PROCESS:
	393	fl->owner = current->pid;
	394	break;
	395	case IPV6_FL_S_USER:
f82b3590	396	fl->owner = current_euid();
1da177e4 LT	397	break;
	398	default:
	399	err = -EINVAL;
	400	goto done;
	401	}
	402	return fl;
	403
	404	done:
	405	fl_free(fl);
	406	*err_p = err;
	407	return NULL;
	408	}
	409
	410	static int mem_check(struct sock *sk)
	411	{
	412	struct ipv6_pinfo *np = inet6_sk(sk);
	413	struct ipv6_fl_socklist *sfl;
	414	int room = FL_MAX_SIZE - atomic_read(&fl_size);
	415	int count = 0;
	416
	417	if (room > FL_MAX_SIZE - FL_MAX_PER_SOCK)
	418	return 0;
	419
	420	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next)
	421	count++;
	422
	423	if (room <= 0 \|\|
	424	((count >= FL_MAX_PER_SOCK \|\|
35700212 JP	425	(count > 0 && room < FL_MAX_SIZE/2) \|\| room < FL_MAX_SIZE/4) &&
35700212 JP	426	!capable(CAP_NET_ADMIN)))
1da177e4 LT	427	return -ENOBUFS;
	428
	429	return 0;
	430	}
	431
	432	static int ipv6_hdr_cmp(struct ipv6_opt_hdr h1, struct ipv6_opt_hdr h2)
	433	{
	434	if (h1 == h2)
	435	return 0;
	436	if (h1 == NULL \|\| h2 == NULL)
	437	return 1;
	438	if (h1->hdrlen != h2->hdrlen)
	439	return 1;
	440	return memcmp(h1+1, h2+1, ((h1->hdrlen+1)<<3) - sizeof(*h1));
	441	}
	442
	443	static int ipv6_opt_cmp(struct ipv6_txoptions o1, struct ipv6_txoptions o2)
	444	{
	445	if (o1 == o2)
	446	return 0;
	447	if (o1 == NULL \|\| o2 == NULL)
	448	return 1;
	449	if (o1->opt_nflen != o2->opt_nflen)
	450	return 1;
	451	if (ipv6_hdr_cmp(o1->hopopt, o2->hopopt))
	452	return 1;
	453	if (ipv6_hdr_cmp(o1->dst0opt, o2->dst0opt))
	454	return 1;
	455	if (ipv6_hdr_cmp((struct ipv6_opt_hdr )o1->srcrt, (struct ipv6_opt_hdr )o2->srcrt))
	456	return 1;
	457	return 0;
	458	}
	459
04028045 PE	460	static inline void fl_link(struct ipv6_pinfo np, struct ipv6_fl_socklist sfl,
	461	struct ip6_flowlabel *fl)
	462	{
	463	write_lock_bh(&ip6_sk_fl_lock);
	464	sfl->fl = fl;
	465	sfl->next = np->ipv6_fl_list;
	466	np->ipv6_fl_list = sfl;
	467	write_unlock_bh(&ip6_sk_fl_lock);
	468	}
	469
1da177e4 LT	470	int ipv6_flowlabel_opt(struct sock sk, char __user optval, int optlen)
1da177e4 LT	471	{
55205d40	472	int uninitialized_var(err);
60e8fbc4	473	struct net *net = sock_net(sk);
1da177e4 LT	474	struct ipv6_pinfo *np = inet6_sk(sk);
	475	struct in6_flowlabel_req freq;
	476	struct ipv6_fl_socklist *sfl1=NULL;
	477	struct ipv6_fl_socklist sfl, *sflp;
78c2e502 PE	478	struct ip6_flowlabel fl, fl1 = NULL;
78c2e502 PE	479
1da177e4 LT	480
	481	if (optlen < sizeof(freq))
	482	return -EINVAL;
	483
	484	if (copy_from_user(&freq, optval, sizeof(freq)))
	485	return -EFAULT;
	486
	487	switch (freq.flr_action) {
	488	case IPV6_FL_A_PUT:
	489	write_lock_bh(&ip6_sk_fl_lock);
	490	for (sflp = &np->ipv6_fl_list; (sfl=*sflp)!=NULL; sflp = &sfl->next) {
	491	if (sfl->fl->label == freq.flr_label) {
	492	if (freq.flr_label == (np->flow_label&IPV6_FLOWLABEL_MASK))
	493	np->flow_label &= ~IPV6_FLOWLABEL_MASK;
	494	*sflp = sfl->next;
	495	write_unlock_bh(&ip6_sk_fl_lock);
	496	fl_release(sfl->fl);
	497	kfree(sfl);
	498	return 0;
	499	}
	500	}
	501	write_unlock_bh(&ip6_sk_fl_lock);
	502	return -ESRCH;
	503
	504	case IPV6_FL_A_RENEW:
	505	read_lock_bh(&ip6_sk_fl_lock);
	506	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) {
	507	if (sfl->fl->label == freq.flr_label) {
	508	err = fl6_renew(sfl->fl, freq.flr_linger, freq.flr_expires);
	509	read_unlock_bh(&ip6_sk_fl_lock);
	510	return err;
	511	}
	512	}
	513	read_unlock_bh(&ip6_sk_fl_lock);
	514
	515	if (freq.flr_share == IPV6_FL_S_NONE && capable(CAP_NET_ADMIN)) {
60e8fbc4	516	fl = fl_lookup(net, freq.flr_label);
1da177e4 LT	517	if (fl) {
	518	err = fl6_renew(fl, freq.flr_linger, freq.flr_expires);
	519	fl_release(fl);
	520	return err;
	521	}
	522	}
	523	return -ESRCH;
	524
	525	case IPV6_FL_A_GET:
	526	if (freq.flr_label & ~IPV6_FLOWLABEL_MASK)
	527	return -EINVAL;
	528
60e8fbc4	529	fl = fl_create(net, &freq, optval, optlen, &err);
1da177e4 LT	530	if (fl == NULL)
	531	return err;
	532	sfl1 = kmalloc(sizeof(*sfl1), GFP_KERNEL);
	533
	534	if (freq.flr_label) {
1da177e4 LT	535	err = -EEXIST;
	536	read_lock_bh(&ip6_sk_fl_lock);
	537	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) {
	538	if (sfl->fl->label == freq.flr_label) {
	539	if (freq.flr_flags&IPV6_FL_F_EXCL) {
	540	read_unlock_bh(&ip6_sk_fl_lock);
	541	goto done;
	542	}
	543	fl1 = sfl->fl;
4ea6a804	544	atomic_inc(&fl1->users);
1da177e4 LT	545	break;
	546	}
	547	}
	548	read_unlock_bh(&ip6_sk_fl_lock);
	549
	550	if (fl1 == NULL)
60e8fbc4	551	fl1 = fl_lookup(net, freq.flr_label);
1da177e4	552	if (fl1) {
78c2e502	553	recheck:
1da177e4 LT	554	err = -EEXIST;
	555	if (freq.flr_flags&IPV6_FL_F_EXCL)
	556	goto release;
	557	err = -EPERM;
	558	if (fl1->share == IPV6_FL_S_EXCL \|\|
	559	fl1->share != fl->share \|\|
	560	fl1->owner != fl->owner)
	561	goto release;
	562
	563	err = -EINVAL;
	564	if (!ipv6_addr_equal(&fl1->dst, &fl->dst) \|\|
	565	ipv6_opt_cmp(fl1->opt, fl->opt))
	566	goto release;
	567
	568	err = -ENOMEM;
	569	if (sfl1 == NULL)
	570	goto release;
	571	if (fl->linger > fl1->linger)
	572	fl1->linger = fl->linger;
	573	if ((long)(fl->expires - fl1->expires) > 0)
	574	fl1->expires = fl->expires;
04028045	575	fl_link(np, sfl1, fl1);
1da177e4 LT	576	fl_free(fl);
	577	return 0;
	578
	579	release:
	580	fl_release(fl1);
	581	goto done;
	582	}
	583	}
	584	err = -ENOENT;
	585	if (!(freq.flr_flags&IPV6_FL_F_CREATE))
	586	goto done;
	587
	588	err = -ENOMEM;
	589	if (sfl1 == NULL \|\| (err = mem_check(sk)) != 0)
	590	goto done;
	591
60e8fbc4	592	fl1 = fl_intern(net, fl, freq.flr_label);
78c2e502 PE	593	if (fl1 != NULL)
78c2e502 PE	594	goto recheck;
1da177e4	595
6c94d361 DM	596	if (!freq.flr_label) {
	597	if (copy_to_user(&((struct in6_flowlabel_req __user *) optval)->flr_label,
	598	&fl->label, sizeof(fl->label))) {
	599	/* Intentionally ignore fault. */
	600	}
	601	}
1da177e4	602
04028045	603	fl_link(np, sfl1, fl);
1da177e4 LT	604	return 0;
	605
	606	default:
	607	return -EINVAL;
	608	}
	609
	610	done:
	611	fl_free(fl);
	612	kfree(sfl1);
	613	return err;
	614	}
	615
	616	#ifdef CONFIG_PROC_FS
	617
	618	struct ip6fl_iter_state {
5983a3df	619	struct seq_net_private p;
1da177e4 LT	620	int bucket;
	621	};
	622
	623	#define ip6fl_seq_private(seq) ((struct ip6fl_iter_state *)(seq)->private)
	624
	625	static struct ip6_flowlabel ip6fl_get_first(struct seq_file seq)
	626	{
	627	struct ip6_flowlabel *fl = NULL;
	628	struct ip6fl_iter_state *state = ip6fl_seq_private(seq);
5983a3df	629	struct net *net = seq_file_net(seq);
1da177e4 LT	630
1da177e4 LT	631	for (state->bucket = 0; state->bucket <= FL_HASH_MASK; ++state->bucket) {
5983a3df BT	632	fl = fl_ht[state->bucket];
5983a3df BT	633
09ad9bc7	634	while (fl && !net_eq(fl->fl_net, net))
5983a3df BT	635	fl = fl->next;
5983a3df BT	636	if (fl)
1da177e4	637	break;
1da177e4 LT	638	}
	639	return fl;
	640	}
	641
	642	static struct ip6_flowlabel ip6fl_get_next(struct seq_file seq, struct ip6_flowlabel *fl)
	643	{
	644	struct ip6fl_iter_state *state = ip6fl_seq_private(seq);
5983a3df	645	struct net *net = seq_file_net(seq);
1da177e4 LT	646
1da177e4 LT	647	fl = fl->next;
5983a3df	648	try_again:
09ad9bc7	649	while (fl && !net_eq(fl->fl_net, net))
5983a3df BT	650	fl = fl->next;
5983a3df BT	651
1da177e4	652	while (!fl) {
5983a3df	653	if (++state->bucket <= FL_HASH_MASK) {
1da177e4	654	fl = fl_ht[state->bucket];
5983a3df BT	655	goto try_again;
5983a3df BT	656	} else
bcd62075	657	break;
1da177e4 LT	658	}
	659	return fl;
	660	}
	661
	662	static struct ip6_flowlabel ip6fl_get_idx(struct seq_file seq, loff_t pos)
	663	{
	664	struct ip6_flowlabel *fl = ip6fl_get_first(seq);
	665	if (fl)
	666	while (pos && (fl = ip6fl_get_next(seq, fl)) != NULL)
	667	--pos;
	668	return pos ? NULL : fl;
	669	}
	670
	671	static void ip6fl_seq_start(struct seq_file seq, loff_t *pos)
9a429c49	672	__acquires(ip6_fl_lock)
1da177e4 LT	673	{
	674	read_lock_bh(&ip6_fl_lock);
	675	return pos ? ip6fl_get_idx(seq, pos - 1) : SEQ_START_TOKEN;
	676	}
	677
	678	static void ip6fl_seq_next(struct seq_file seq, void v, loff_t pos)
	679	{
	680	struct ip6_flowlabel *fl;
	681
	682	if (v == SEQ_START_TOKEN)
	683	fl = ip6fl_get_first(seq);
	684	else
	685	fl = ip6fl_get_next(seq, v);
	686	++*pos;
	687	return fl;
	688	}
	689
	690	static void ip6fl_seq_stop(struct seq_file seq, void v)
9a429c49	691	__releases(ip6_fl_lock)
1da177e4 LT	692	{
	693	read_unlock_bh(&ip6_fl_lock);
	694	}
	695
1b7c2dbc	696	static int ip6fl_seq_show(struct seq_file seq, void v)
1da177e4	697	{
1b7c2dbc JM	698	if (v == SEQ_START_TOKEN)
	699	seq_printf(seq, "%-5s %-1s %-6s %-6s %-6s %-8s %-32s %s\n",
	700	"Label", "S", "Owner", "Users", "Linger", "Expires", "Dst", "Opt");
	701	else {
	702	struct ip6_flowlabel *fl = v;
1da177e4	703	seq_printf(seq,
4b7a4274	704	"%05X %-1d %-6d %-6d %-6ld %-8ld %pi6 %-4d\n",
1da177e4 LT	705	(unsigned)ntohl(fl->label),
	706	fl->share,
	707	(unsigned)fl->owner,
	708	atomic_read(&fl->users),
	709	fl->linger/HZ,
	710	(long)(fl->expires - jiffies)/HZ,
b071195d	711	&fl->dst,
1da177e4	712	fl->opt ? fl->opt->opt_nflen : 0);
1da177e4	713	}
1da177e4 LT	714	return 0;
	715	}
	716
56b3d975	717	static const struct seq_operations ip6fl_seq_ops = {
1da177e4 LT	718	.start = ip6fl_seq_start,
	719	.next = ip6fl_seq_next,
	720	.stop = ip6fl_seq_stop,
	721	.show = ip6fl_seq_show,
	722	};
	723
	724	static int ip6fl_seq_open(struct inode inode, struct file file)
	725	{
5983a3df BT	726	return seq_open_net(inode, file, &ip6fl_seq_ops,
5983a3df BT	727	sizeof(struct ip6fl_iter_state));
1da177e4 LT	728	}
1da177e4 LT	729
9a32144e	730	static const struct file_operations ip6fl_seq_fops = {
1da177e4 LT	731	.owner = THIS_MODULE,
	732	.open = ip6fl_seq_open,
	733	.read = seq_read,
	734	.llseek = seq_lseek,
5983a3df	735	.release = seq_release_net,
1da177e4	736	};
1da177e4	737
2c8c1e72	738	static int __net_init ip6_flowlabel_proc_init(struct net *net)
0a3e78ac	739	{
5983a3df BT	740	if (!proc_net_fops_create(net, "ip6_flowlabel",
5983a3df BT	741	S_IRUGO, &ip6fl_seq_fops))
0a3e78ac DL	742	return -ENOMEM;
	743	return 0;
	744	}
1da177e4	745
2c8c1e72	746	static void __net_exit ip6_flowlabel_proc_fini(struct net *net)
1da177e4	747	{
0a3e78ac DL	748	proc_net_remove(net, "ip6_flowlabel");
	749	}
	750	#else
	751	static inline int ip6_flowlabel_proc_init(struct net *net)
	752	{
	753	return 0;
	754	}
	755	static inline void ip6_flowlabel_proc_fini(struct net *net)
	756	{
0a3e78ac	757	}
1da177e4	758	#endif
0a3e78ac	759
2c8c1e72	760	static void __net_exit ip6_flowlabel_net_exit(struct net *net)
60e8fbc4 BT	761	{
60e8fbc4 BT	762	ip6_fl_purge(net);
5983a3df	763	ip6_flowlabel_proc_fini(net);
60e8fbc4 BT	764	}
	765
	766	static struct pernet_operations ip6_flowlabel_net_ops = {
5983a3df	767	.init = ip6_flowlabel_proc_init,
60e8fbc4 BT	768	.exit = ip6_flowlabel_net_exit,
	769	};
	770
0a3e78ac DL	771	int ip6_flowlabel_init(void)
0a3e78ac DL	772	{
5983a3df	773	return register_pernet_subsys(&ip6_flowlabel_net_ops);
1da177e4 LT	774	}
	775
	776	void ip6_flowlabel_cleanup(void)
	777	{
	778	del_timer(&ip6_fl_gc_timer);
60e8fbc4	779	unregister_pernet_subsys(&ip6_flowlabel_net_ops);
1da177e4	780	}