[net-next-2.6.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 <linux/slab.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, &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>
5a0e3ad6	23	#include <linux/slab.h>
1da177e4	24
457c4cbc	25	#include <net/net_namespace.h>
1da177e4 LT	26	#include <net/sock.h>
	27
	28	#include <net/ipv6.h>
	29	#include <net/ndisc.h>
	30	#include <net/protocol.h>
	31	#include <net/ip6_route.h>
	32	#include <net/addrconf.h>
	33	#include <net/rawv6.h>
	34	#include <net/icmp.h>
	35	#include <net/transp_v6.h>
	36
	37	#include <asm/uaccess.h>
	38
	39	#define FL_MIN_LINGER 6 /* Minimal linger. It is set to 6sec specified
	40	in old IPv6 RFC. Well, it was reasonable value.
	41	*/
	42	#define FL_MAX_LINGER 60 /* Maximal linger timeout */
	43
	44	/* FL hash table */
	45
	46	#define FL_MAX_PER_SOCK 32
	47	#define FL_MAX_SIZE 4096
	48	#define FL_HASH_MASK 255
	49	#define FL_HASH(l) (ntohl(l)&FL_HASH_MASK)
	50
	51	static atomic_t fl_size = ATOMIC_INIT(0);
	52	static struct ip6_flowlabel *fl_ht[FL_HASH_MASK+1];
	53
	54	static void ip6_fl_gc(unsigned long dummy);
8d06afab	55	static DEFINE_TIMER(ip6_fl_gc_timer, ip6_fl_gc, 0, 0);
1da177e4 LT	56
	57	/* FL hash table lock: it protects only of GC */
	58
	59	static DEFINE_RWLOCK(ip6_fl_lock);
	60
	61	/* Big socket sock */
	62
	63	static DEFINE_RWLOCK(ip6_sk_fl_lock);
	64
	65
60e8fbc4	66	static inline struct ip6_flowlabel __fl_lookup(struct net net, __be32 label)
1da177e4 LT	67	{
	68	struct ip6_flowlabel *fl;
	69
	70	for (fl=fl_ht[FL_HASH(label)]; fl; fl = fl->next) {
09ad9bc7	71	if (fl->label == label && net_eq(fl->fl_net, net))
1da177e4 LT	72	return fl;
	73	}
	74	return NULL;
	75	}
	76
60e8fbc4	77	static struct ip6_flowlabel fl_lookup(struct net net, __be32 label)
1da177e4 LT	78	{
	79	struct ip6_flowlabel *fl;
	80
	81	read_lock_bh(&ip6_fl_lock);
60e8fbc4	82	fl = __fl_lookup(net, label);
1da177e4 LT	83	if (fl)
	84	atomic_inc(&fl->users);
	85	read_unlock_bh(&ip6_fl_lock);
	86	return fl;
	87	}
	88
	89
	90	static void fl_free(struct ip6_flowlabel *fl)
	91	{
60e8fbc4 BT	92	if (fl) {
60e8fbc4 BT	93	release_net(fl->fl_net);
1da177e4	94	kfree(fl->opt);
60e8fbc4	95	}
1da177e4 LT	96	kfree(fl);
	97	}
	98
	99	static void fl_release(struct ip6_flowlabel *fl)
	100	{
	101	write_lock_bh(&ip6_fl_lock);
	102
	103	fl->lastuse = jiffies;
	104	if (atomic_dec_and_test(&fl->users)) {
	105	unsigned long ttd = fl->lastuse + fl->linger;
	106	if (time_after(ttd, fl->expires))
	107	fl->expires = ttd;
	108	ttd = fl->expires;
	109	if (fl->opt && fl->share == IPV6_FL_S_EXCL) {
	110	struct ipv6_txoptions *opt = fl->opt;
	111	fl->opt = NULL;
	112	kfree(opt);
	113	}
	114	if (!timer_pending(&ip6_fl_gc_timer) \|\|
	115	time_after(ip6_fl_gc_timer.expires, ttd))
	116	mod_timer(&ip6_fl_gc_timer, ttd);
	117	}
1da177e4 LT	118	write_unlock_bh(&ip6_fl_lock);
	119	}
	120
	121	static void ip6_fl_gc(unsigned long dummy)
	122	{
	123	int i;
	124	unsigned long now = jiffies;
	125	unsigned long sched = 0;
	126
	127	write_lock(&ip6_fl_lock);
	128
	129	for (i=0; i<=FL_HASH_MASK; i++) {
	130	struct ip6_flowlabel fl, *flp;
	131	flp = &fl_ht[i];
	132	while ((fl=*flp) != NULL) {
	133	if (atomic_read(&fl->users) == 0) {
	134	unsigned long ttd = fl->lastuse + fl->linger;
	135	if (time_after(ttd, fl->expires))
	136	fl->expires = ttd;
	137	ttd = fl->expires;
	138	if (time_after_eq(now, ttd)) {
	139	*flp = fl->next;
	140	fl_free(fl);
	141	atomic_dec(&fl_size);
	142	continue;
	143	}
	144	if (!sched \|\| time_before(ttd, sched))
	145	sched = ttd;
	146	}
	147	flp = &fl->next;
	148	}
	149	}
	150	if (!sched && atomic_read(&fl_size))
	151	sched = now + FL_MAX_LINGER;
	152	if (sched) {
60e8fbc4	153	mod_timer(&ip6_fl_gc_timer, sched);
1da177e4 LT	154	}
	155	write_unlock(&ip6_fl_lock);
	156	}
	157
2c8c1e72	158	static void __net_exit ip6_fl_purge(struct net *net)
60e8fbc4 BT	159	{
	160	int i;
	161
	162	write_lock(&ip6_fl_lock);
	163	for (i = 0; i <= FL_HASH_MASK; i++) {
	164	struct ip6_flowlabel fl, *flp;
	165	flp = &fl_ht[i];
	166	while ((fl = *flp) != NULL) {
09ad9bc7 OP	167	if (net_eq(fl->fl_net, net) &&
09ad9bc7 OP	168	atomic_read(&fl->users) == 0) {
60e8fbc4 BT	169	*flp = fl->next;
	170	fl_free(fl);
	171	atomic_dec(&fl_size);
	172	continue;
	173	}
	174	flp = &fl->next;
	175	}
	176	}
	177	write_unlock(&ip6_fl_lock);
	178	}
	179
	180	static struct ip6_flowlabel fl_intern(struct net net,
	181	struct ip6_flowlabel *fl, __be32 label)
1da177e4	182	{
78c2e502 PE	183	struct ip6_flowlabel *lfl;
78c2e502 PE	184
1da177e4 LT	185	fl->label = label & IPV6_FLOWLABEL_MASK;
	186
	187	write_lock_bh(&ip6_fl_lock);
	188	if (label == 0) {
	189	for (;;) {
	190	fl->label = htonl(net_random())&IPV6_FLOWLABEL_MASK;
	191	if (fl->label) {
60e8fbc4	192	lfl = __fl_lookup(net, fl->label);
1da177e4 LT	193	if (lfl == NULL)
	194	break;
	195	}
	196	}
78c2e502 PE	197	} else {
	198	/*
	199	* we dropper the ip6_fl_lock, so this entry could reappear
	200	* and we need to recheck with it.
	201	*
	202	* OTOH no need to search the active socket first, like it is
	203	* done in ipv6_flowlabel_opt - sock is locked, so new entry
	204	* with the same label can only appear on another sock
	205	*/
60e8fbc4	206	lfl = __fl_lookup(net, fl->label);
78c2e502 PE	207	if (lfl != NULL) {
	208	atomic_inc(&lfl->users);
	209	write_unlock_bh(&ip6_fl_lock);
	210	return lfl;
	211	}
1da177e4 LT	212	}
	213
	214	fl->lastuse = jiffies;
	215	fl->next = fl_ht[FL_HASH(fl->label)];
	216	fl_ht[FL_HASH(fl->label)] = fl;
	217	atomic_inc(&fl_size);
	218	write_unlock_bh(&ip6_fl_lock);
78c2e502	219	return NULL;
1da177e4 LT	220	}
	221
	222
	223
	224	/* Socket flowlabel lists */
	225
90bcaf7b	226	struct ip6_flowlabel * fl6_sock_lookup(struct sock *sk, __be32 label)
1da177e4 LT	227	{
	228	struct ipv6_fl_socklist *sfl;
	229	struct ipv6_pinfo *np = inet6_sk(sk);
	230
	231	label &= IPV6_FLOWLABEL_MASK;
	232
bd0bf577	233	read_lock_bh(&ip6_sk_fl_lock);
1da177e4 LT	234	for (sfl=np->ipv6_fl_list; sfl; sfl = sfl->next) {
	235	struct ip6_flowlabel *fl = sfl->fl;
	236	if (fl->label == label) {
	237	fl->lastuse = jiffies;
	238	atomic_inc(&fl->users);
52f095ee	239	read_unlock_bh(&ip6_sk_fl_lock);
1da177e4 LT	240	return fl;
	241	}
	242	}
bd0bf577	243	read_unlock_bh(&ip6_sk_fl_lock);
1da177e4 LT	244	return NULL;
	245	}
	246
3cf3dc6c ACM	247	EXPORT_SYMBOL_GPL(fl6_sock_lookup);
3cf3dc6c ACM	248
1da177e4 LT	249	void fl6_free_socklist(struct sock *sk)
	250	{
	251	struct ipv6_pinfo *np = inet6_sk(sk);
	252	struct ipv6_fl_socklist *sfl;
	253
	254	while ((sfl = np->ipv6_fl_list) != NULL) {
	255	np->ipv6_fl_list = sfl->next;
	256	fl_release(sfl->fl);
	257	kfree(sfl);
	258	}
	259	}
	260
	261	/* Service routines */
	262
	263
	264	/*
	265	It is the only difficult place. flowlabel enforces equal headers
	266	before and including routing header, however user may supply options
	267	following rthdr.
	268	*/
	269
	270	struct ipv6_txoptions fl6_merge_options(struct ipv6_txoptions opt_space,
	271	struct ip6_flowlabel * fl,
	272	struct ipv6_txoptions * fopt)
	273	{
df9890c3	274	struct ipv6_txoptions * fl_opt = fl->opt;
1ab1457c	275
df9890c3 YH	276	if (fopt == NULL \|\| fopt->opt_flen == 0)
df9890c3 YH	277	return fl_opt;
1ab1457c	278
1da177e4 LT	279	if (fl_opt != NULL) {
1da177e4 LT	280	opt_space->hopopt = fl_opt->hopopt;
df9890c3	281	opt_space->dst0opt = fl_opt->dst0opt;
1da177e4 LT	282	opt_space->srcrt = fl_opt->srcrt;
	283	opt_space->opt_nflen = fl_opt->opt_nflen;
	284	} else {
	285	if (fopt->opt_nflen == 0)
	286	return fopt;
	287	opt_space->hopopt = NULL;
	288	opt_space->dst0opt = NULL;
	289	opt_space->srcrt = NULL;
	290	opt_space->opt_nflen = 0;
	291	}
	292	opt_space->dst1opt = fopt->dst1opt;
1da177e4 LT	293	opt_space->opt_flen = fopt->opt_flen;
	294	return opt_space;
	295	}
	296
	297	static unsigned long check_linger(unsigned long ttl)
	298	{
	299	if (ttl < FL_MIN_LINGER)
	300	return FL_MIN_LINGER*HZ;
	301	if (ttl > FL_MAX_LINGER && !capable(CAP_NET_ADMIN))
	302	return 0;
	303	return ttl*HZ;
	304	}
	305
	306	static int fl6_renew(struct ip6_flowlabel *fl, unsigned long linger, unsigned long expires)
	307	{
	308	linger = check_linger(linger);
	309	if (!linger)
	310	return -EPERM;
	311	expires = check_linger(expires);
	312	if (!expires)
	313	return -EPERM;
	314	fl->lastuse = jiffies;
	315	if (time_before(fl->linger, linger))
	316	fl->linger = linger;
	317	if (time_before(expires, fl->linger))
	318	expires = fl->linger;
	319	if (time_before(fl->expires, fl->lastuse + expires))
	320	fl->expires = fl->lastuse + expires;
	321	return 0;
	322	}
	323
	324	static struct ip6_flowlabel *
60e8fbc4 BT	325	fl_create(struct net net, struct in6_flowlabel_req freq, char __user *optval,
60e8fbc4 BT	326	int optlen, int *err_p)
1da177e4	327	{
684de409	328	struct ip6_flowlabel *fl = NULL;
1da177e4 LT	329	int olen;
	330	int addr_type;
	331	int err;
	332
684de409 DM	333	olen = optlen - CMSG_ALIGN(sizeof(*freq));
	334	err = -EINVAL;
	335	if (olen > 64 * 1024)
	336	goto done;
	337
1da177e4	338	err = -ENOMEM;
0c600eda	339	fl = kzalloc(sizeof(*fl), GFP_KERNEL);
1da177e4 LT	340	if (fl == NULL)
1da177e4 LT	341	goto done;
1da177e4	342
1da177e4 LT	343	if (olen > 0) {
	344	struct msghdr msg;
	345	struct flowi flowi;
	346	int junk;
	347
	348	err = -ENOMEM;
	349	fl->opt = kmalloc(sizeof(*fl->opt) + olen, GFP_KERNEL);
	350	if (fl->opt == NULL)
	351	goto done;
	352
	353	memset(fl->opt, 0, sizeof(*fl->opt));
	354	fl->opt->tot_len = sizeof(*fl->opt) + olen;
	355	err = -EFAULT;
	356	if (copy_from_user(fl->opt+1, optval+CMSG_ALIGN(sizeof(*freq)), olen))
	357	goto done;
	358
	359	msg.msg_controllen = olen;
	360	msg.msg_control = (void*)(fl->opt+1);
	361	flowi.oif = 0;
	362
13b52cd4 BH	363	err = datagram_send_ctl(net, &msg, &flowi, fl->opt, &junk,
13b52cd4 BH	364	&junk, &junk);
1da177e4 LT	365	if (err)
	366	goto done;
	367	err = -EINVAL;
	368	if (fl->opt->opt_flen)
	369	goto done;
	370	if (fl->opt->opt_nflen == 0) {
	371	kfree(fl->opt);
	372	fl->opt = NULL;
	373	}
	374	}
	375
60e8fbc4	376	fl->fl_net = hold_net(net);
1da177e4 LT	377	fl->expires = jiffies;
	378	err = fl6_renew(fl, freq->flr_linger, freq->flr_expires);
	379	if (err)
	380	goto done;
	381	fl->share = freq->flr_share;
	382	addr_type = ipv6_addr_type(&freq->flr_dst);
35700212 JP	383	if ((addr_type & IPV6_ADDR_MAPPED) \|\|
35700212 JP	384	addr_type == IPV6_ADDR_ANY) {
c6817e4c	385	err = -EINVAL;
1da177e4	386	goto done;
c6817e4c	387	}
1da177e4 LT	388	ipv6_addr_copy(&fl->dst, &freq->flr_dst);
	389	atomic_set(&fl->users, 1);
	390	switch (fl->share) {
	391	case IPV6_FL_S_EXCL:
	392	case IPV6_FL_S_ANY:
	393	break;
	394	case IPV6_FL_S_PROCESS:
	395	fl->owner = current->pid;
	396	break;
	397	case IPV6_FL_S_USER:
f82b3590	398	fl->owner = current_euid();
1da177e4 LT	399	break;
	400	default:
	401	err = -EINVAL;
	402	goto done;
	403	}
	404	return fl;
	405
	406	done:
	407	fl_free(fl);
	408	*err_p = err;
	409	return NULL;
	410	}
	411
	412	static int mem_check(struct sock *sk)
	413	{
	414	struct ipv6_pinfo *np = inet6_sk(sk);
	415	struct ipv6_fl_socklist *sfl;
	416	int room = FL_MAX_SIZE - atomic_read(&fl_size);
	417	int count = 0;
	418
	419	if (room > FL_MAX_SIZE - FL_MAX_PER_SOCK)
	420	return 0;
	421
	422	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next)
	423	count++;
	424
	425	if (room <= 0 \|\|
	426	((count >= FL_MAX_PER_SOCK \|\|
35700212 JP	427	(count > 0 && room < FL_MAX_SIZE/2) \|\| room < FL_MAX_SIZE/4) &&
35700212 JP	428	!capable(CAP_NET_ADMIN)))
1da177e4 LT	429	return -ENOBUFS;
	430
	431	return 0;
	432	}
	433
	434	static int ipv6_hdr_cmp(struct ipv6_opt_hdr h1, struct ipv6_opt_hdr h2)
	435	{
	436	if (h1 == h2)
	437	return 0;
	438	if (h1 == NULL \|\| h2 == NULL)
	439	return 1;
	440	if (h1->hdrlen != h2->hdrlen)
	441	return 1;
	442	return memcmp(h1+1, h2+1, ((h1->hdrlen+1)<<3) - sizeof(*h1));
	443	}
	444
	445	static int ipv6_opt_cmp(struct ipv6_txoptions o1, struct ipv6_txoptions o2)
	446	{
	447	if (o1 == o2)
	448	return 0;
	449	if (o1 == NULL \|\| o2 == NULL)
	450	return 1;
	451	if (o1->opt_nflen != o2->opt_nflen)
	452	return 1;
	453	if (ipv6_hdr_cmp(o1->hopopt, o2->hopopt))
	454	return 1;
	455	if (ipv6_hdr_cmp(o1->dst0opt, o2->dst0opt))
	456	return 1;
	457	if (ipv6_hdr_cmp((struct ipv6_opt_hdr )o1->srcrt, (struct ipv6_opt_hdr )o2->srcrt))
	458	return 1;
	459	return 0;
	460	}
	461
04028045 PE	462	static inline void fl_link(struct ipv6_pinfo np, struct ipv6_fl_socklist sfl,
	463	struct ip6_flowlabel *fl)
	464	{
	465	write_lock_bh(&ip6_sk_fl_lock);
	466	sfl->fl = fl;
	467	sfl->next = np->ipv6_fl_list;
	468	np->ipv6_fl_list = sfl;
	469	write_unlock_bh(&ip6_sk_fl_lock);
	470	}
	471
1da177e4 LT	472	int ipv6_flowlabel_opt(struct sock sk, char __user optval, int optlen)
1da177e4 LT	473	{
55205d40	474	int uninitialized_var(err);
60e8fbc4	475	struct net *net = sock_net(sk);
1da177e4 LT	476	struct ipv6_pinfo *np = inet6_sk(sk);
	477	struct in6_flowlabel_req freq;
	478	struct ipv6_fl_socklist *sfl1=NULL;
	479	struct ipv6_fl_socklist sfl, *sflp;
78c2e502 PE	480	struct ip6_flowlabel fl, fl1 = NULL;
78c2e502 PE	481
1da177e4 LT	482
	483	if (optlen < sizeof(freq))
	484	return -EINVAL;
	485
	486	if (copy_from_user(&freq, optval, sizeof(freq)))
	487	return -EFAULT;
	488
	489	switch (freq.flr_action) {
	490	case IPV6_FL_A_PUT:
	491	write_lock_bh(&ip6_sk_fl_lock);
	492	for (sflp = &np->ipv6_fl_list; (sfl=*sflp)!=NULL; sflp = &sfl->next) {
	493	if (sfl->fl->label == freq.flr_label) {
	494	if (freq.flr_label == (np->flow_label&IPV6_FLOWLABEL_MASK))
	495	np->flow_label &= ~IPV6_FLOWLABEL_MASK;
	496	*sflp = sfl->next;
	497	write_unlock_bh(&ip6_sk_fl_lock);
	498	fl_release(sfl->fl);
	499	kfree(sfl);
	500	return 0;
	501	}
	502	}
	503	write_unlock_bh(&ip6_sk_fl_lock);
	504	return -ESRCH;
	505
	506	case IPV6_FL_A_RENEW:
	507	read_lock_bh(&ip6_sk_fl_lock);
	508	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) {
	509	if (sfl->fl->label == freq.flr_label) {
	510	err = fl6_renew(sfl->fl, freq.flr_linger, freq.flr_expires);
	511	read_unlock_bh(&ip6_sk_fl_lock);
	512	return err;
	513	}
	514	}
	515	read_unlock_bh(&ip6_sk_fl_lock);
	516
	517	if (freq.flr_share == IPV6_FL_S_NONE && capable(CAP_NET_ADMIN)) {
60e8fbc4	518	fl = fl_lookup(net, freq.flr_label);
1da177e4 LT	519	if (fl) {
	520	err = fl6_renew(fl, freq.flr_linger, freq.flr_expires);
	521	fl_release(fl);
	522	return err;
	523	}
	524	}
	525	return -ESRCH;
	526
	527	case IPV6_FL_A_GET:
	528	if (freq.flr_label & ~IPV6_FLOWLABEL_MASK)
	529	return -EINVAL;
	530
60e8fbc4	531	fl = fl_create(net, &freq, optval, optlen, &err);
1da177e4 LT	532	if (fl == NULL)
	533	return err;
	534	sfl1 = kmalloc(sizeof(*sfl1), GFP_KERNEL);
	535
	536	if (freq.flr_label) {
1da177e4 LT	537	err = -EEXIST;
	538	read_lock_bh(&ip6_sk_fl_lock);
	539	for (sfl = np->ipv6_fl_list; sfl; sfl = sfl->next) {
	540	if (sfl->fl->label == freq.flr_label) {
	541	if (freq.flr_flags&IPV6_FL_F_EXCL) {
	542	read_unlock_bh(&ip6_sk_fl_lock);
	543	goto done;
	544	}
	545	fl1 = sfl->fl;
4ea6a804	546	atomic_inc(&fl1->users);
1da177e4 LT	547	break;
	548	}
	549	}
	550	read_unlock_bh(&ip6_sk_fl_lock);
	551
	552	if (fl1 == NULL)
60e8fbc4	553	fl1 = fl_lookup(net, freq.flr_label);
1da177e4	554	if (fl1) {
78c2e502	555	recheck:
1da177e4 LT	556	err = -EEXIST;
	557	if (freq.flr_flags&IPV6_FL_F_EXCL)
	558	goto release;
	559	err = -EPERM;
	560	if (fl1->share == IPV6_FL_S_EXCL \|\|
	561	fl1->share != fl->share \|\|
	562	fl1->owner != fl->owner)
	563	goto release;
	564
	565	err = -EINVAL;
	566	if (!ipv6_addr_equal(&fl1->dst, &fl->dst) \|\|
	567	ipv6_opt_cmp(fl1->opt, fl->opt))
	568	goto release;
	569
	570	err = -ENOMEM;
	571	if (sfl1 == NULL)
	572	goto release;
	573	if (fl->linger > fl1->linger)
	574	fl1->linger = fl->linger;
	575	if ((long)(fl->expires - fl1->expires) > 0)
	576	fl1->expires = fl->expires;
04028045	577	fl_link(np, sfl1, fl1);
1da177e4 LT	578	fl_free(fl);
	579	return 0;
	580
	581	release:
	582	fl_release(fl1);
	583	goto done;
	584	}
	585	}
	586	err = -ENOENT;
	587	if (!(freq.flr_flags&IPV6_FL_F_CREATE))
	588	goto done;
	589
	590	err = -ENOMEM;
	591	if (sfl1 == NULL \|\| (err = mem_check(sk)) != 0)
	592	goto done;
	593
60e8fbc4	594	fl1 = fl_intern(net, fl, freq.flr_label);
78c2e502 PE	595	if (fl1 != NULL)
78c2e502 PE	596	goto recheck;
1da177e4	597
6c94d361 DM	598	if (!freq.flr_label) {
	599	if (copy_to_user(&((struct in6_flowlabel_req __user *) optval)->flr_label,
	600	&fl->label, sizeof(fl->label))) {
	601	/* Intentionally ignore fault. */
	602	}
	603	}
1da177e4	604
04028045	605	fl_link(np, sfl1, fl);
1da177e4 LT	606	return 0;
	607
	608	default:
	609	return -EINVAL;
	610	}
	611
	612	done:
	613	fl_free(fl);
	614	kfree(sfl1);
	615	return err;
	616	}
	617
	618	#ifdef CONFIG_PROC_FS
	619
	620	struct ip6fl_iter_state {
5983a3df	621	struct seq_net_private p;
1da177e4 LT	622	int bucket;
	623	};
	624
	625	#define ip6fl_seq_private(seq) ((struct ip6fl_iter_state *)(seq)->private)
	626
	627	static struct ip6_flowlabel ip6fl_get_first(struct seq_file seq)
	628	{
	629	struct ip6_flowlabel *fl = NULL;
	630	struct ip6fl_iter_state *state = ip6fl_seq_private(seq);
5983a3df	631	struct net *net = seq_file_net(seq);
1da177e4 LT	632
1da177e4 LT	633	for (state->bucket = 0; state->bucket <= FL_HASH_MASK; ++state->bucket) {
5983a3df BT	634	fl = fl_ht[state->bucket];
5983a3df BT	635
09ad9bc7	636	while (fl && !net_eq(fl->fl_net, net))
5983a3df BT	637	fl = fl->next;
5983a3df BT	638	if (fl)
1da177e4	639	break;
1da177e4 LT	640	}
	641	return fl;
	642	}
	643
	644	static struct ip6_flowlabel ip6fl_get_next(struct seq_file seq, struct ip6_flowlabel *fl)
	645	{
	646	struct ip6fl_iter_state *state = ip6fl_seq_private(seq);
5983a3df	647	struct net *net = seq_file_net(seq);
1da177e4 LT	648
1da177e4 LT	649	fl = fl->next;
5983a3df	650	try_again:
09ad9bc7	651	while (fl && !net_eq(fl->fl_net, net))
5983a3df BT	652	fl = fl->next;
5983a3df BT	653
1da177e4	654	while (!fl) {
5983a3df	655	if (++state->bucket <= FL_HASH_MASK) {
1da177e4	656	fl = fl_ht[state->bucket];
5983a3df BT	657	goto try_again;
5983a3df BT	658	} else
bcd62075	659	break;
1da177e4 LT	660	}
	661	return fl;
	662	}
	663
	664	static struct ip6_flowlabel ip6fl_get_idx(struct seq_file seq, loff_t pos)
	665	{
	666	struct ip6_flowlabel *fl = ip6fl_get_first(seq);
	667	if (fl)
	668	while (pos && (fl = ip6fl_get_next(seq, fl)) != NULL)
	669	--pos;
	670	return pos ? NULL : fl;
	671	}
	672
	673	static void ip6fl_seq_start(struct seq_file seq, loff_t *pos)
9a429c49	674	__acquires(ip6_fl_lock)
1da177e4 LT	675	{
	676	read_lock_bh(&ip6_fl_lock);
	677	return pos ? ip6fl_get_idx(seq, pos - 1) : SEQ_START_TOKEN;
	678	}
	679
	680	static void ip6fl_seq_next(struct seq_file seq, void v, loff_t pos)
	681	{
	682	struct ip6_flowlabel *fl;
	683
	684	if (v == SEQ_START_TOKEN)
	685	fl = ip6fl_get_first(seq);
	686	else
	687	fl = ip6fl_get_next(seq, v);
	688	++*pos;
	689	return fl;
	690	}
	691
	692	static void ip6fl_seq_stop(struct seq_file seq, void v)
9a429c49	693	__releases(ip6_fl_lock)
1da177e4 LT	694	{
	695	read_unlock_bh(&ip6_fl_lock);
	696	}
	697
1b7c2dbc	698	static int ip6fl_seq_show(struct seq_file seq, void v)
1da177e4	699	{
1b7c2dbc JM	700	if (v == SEQ_START_TOKEN)
	701	seq_printf(seq, "%-5s %-1s %-6s %-6s %-6s %-8s %-32s %s\n",
	702	"Label", "S", "Owner", "Users", "Linger", "Expires", "Dst", "Opt");
	703	else {
	704	struct ip6_flowlabel *fl = v;
1da177e4	705	seq_printf(seq,
4b7a4274	706	"%05X %-1d %-6d %-6d %-6ld %-8ld %pi6 %-4d\n",
1da177e4 LT	707	(unsigned)ntohl(fl->label),
	708	fl->share,
	709	(unsigned)fl->owner,
	710	atomic_read(&fl->users),
	711	fl->linger/HZ,
	712	(long)(fl->expires - jiffies)/HZ,
b071195d	713	&fl->dst,
1da177e4	714	fl->opt ? fl->opt->opt_nflen : 0);
1da177e4	715	}
1da177e4 LT	716	return 0;
	717	}
	718
56b3d975	719	static const struct seq_operations ip6fl_seq_ops = {
1da177e4 LT	720	.start = ip6fl_seq_start,
	721	.next = ip6fl_seq_next,
	722	.stop = ip6fl_seq_stop,
	723	.show = ip6fl_seq_show,
	724	};
	725
	726	static int ip6fl_seq_open(struct inode inode, struct file file)
	727	{
5983a3df BT	728	return seq_open_net(inode, file, &ip6fl_seq_ops,
5983a3df BT	729	sizeof(struct ip6fl_iter_state));
1da177e4 LT	730	}
1da177e4 LT	731
9a32144e	732	static const struct file_operations ip6fl_seq_fops = {
1da177e4 LT	733	.owner = THIS_MODULE,
	734	.open = ip6fl_seq_open,
	735	.read = seq_read,
	736	.llseek = seq_lseek,
5983a3df	737	.release = seq_release_net,
1da177e4	738	};
1da177e4	739
2c8c1e72	740	static int __net_init ip6_flowlabel_proc_init(struct net *net)
0a3e78ac	741	{
5983a3df BT	742	if (!proc_net_fops_create(net, "ip6_flowlabel",
5983a3df BT	743	S_IRUGO, &ip6fl_seq_fops))
0a3e78ac DL	744	return -ENOMEM;
	745	return 0;
	746	}
1da177e4	747
2c8c1e72	748	static void __net_exit ip6_flowlabel_proc_fini(struct net *net)
1da177e4	749	{
0a3e78ac DL	750	proc_net_remove(net, "ip6_flowlabel");
	751	}
	752	#else
	753	static inline int ip6_flowlabel_proc_init(struct net *net)
	754	{
	755	return 0;
	756	}
	757	static inline void ip6_flowlabel_proc_fini(struct net *net)
	758	{
0a3e78ac	759	}
1da177e4	760	#endif
0a3e78ac	761
2c8c1e72	762	static void __net_exit ip6_flowlabel_net_exit(struct net *net)
60e8fbc4 BT	763	{
60e8fbc4 BT	764	ip6_fl_purge(net);
5983a3df	765	ip6_flowlabel_proc_fini(net);
60e8fbc4 BT	766	}
	767
	768	static struct pernet_operations ip6_flowlabel_net_ops = {
5983a3df	769	.init = ip6_flowlabel_proc_init,
60e8fbc4 BT	770	.exit = ip6_flowlabel_net_exit,
	771	};
	772
0a3e78ac DL	773	int ip6_flowlabel_init(void)
0a3e78ac DL	774	{
5983a3df	775	return register_pernet_subsys(&ip6_flowlabel_net_ops);
1da177e4 LT	776	}
	777
	778	void ip6_flowlabel_cleanup(void)
	779	{
	780	del_timer(&ip6_fl_gc_timer);
60e8fbc4	781	unregister_pernet_subsys(&ip6_flowlabel_net_ops);
1da177e4	782	}