2005-04-16 18:20:36 -04:00
|
|
|
/*
|
|
|
|
* INET An implementation of the TCP/IP protocol suite for the LINUX
|
|
|
|
* operating system. INET is implemented using the BSD Socket
|
|
|
|
* interface as the means of communication with the user level.
|
|
|
|
*
|
|
|
|
* The IP fragmentation functionality.
|
2007-02-09 09:24:47 -05:00
|
|
|
*
|
2005-04-16 18:20:36 -04:00
|
|
|
* Version: $Id: ip_fragment.c,v 1.59 2002/01/12 07:54:56 davem Exp $
|
|
|
|
*
|
|
|
|
* Authors: Fred N. van Kempen <waltje@uWalt.NL.Mugnet.ORG>
|
|
|
|
* Alan Cox <Alan.Cox@linux.org>
|
|
|
|
*
|
|
|
|
* Fixes:
|
|
|
|
* Alan Cox : Split from ip.c , see ip_input.c for history.
|
|
|
|
* David S. Miller : Begin massive cleanup...
|
|
|
|
* Andi Kleen : Add sysctls.
|
|
|
|
* xxxx : Overlapfrag bug.
|
|
|
|
* Ultima : ip_expire() kernel panic.
|
|
|
|
* Bill Hawes : Frag accounting and evictor fixes.
|
|
|
|
* John McDonald : 0 length frag bug.
|
|
|
|
* Alexey Kuznetsov: SMP races, threading, cleanup.
|
|
|
|
* Patrick McHardy : LRU queue of frag heads for evictor.
|
|
|
|
*/
|
|
|
|
|
2005-12-14 02:14:27 -05:00
|
|
|
#include <linux/compiler.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/jiffies.h>
|
|
|
|
#include <linux/skbuff.h>
|
|
|
|
#include <linux/list.h>
|
|
|
|
#include <linux/ip.h>
|
|
|
|
#include <linux/icmp.h>
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/jhash.h>
|
|
|
|
#include <linux/random.h>
|
|
|
|
#include <net/sock.h>
|
|
|
|
#include <net/ip.h>
|
|
|
|
#include <net/icmp.h>
|
|
|
|
#include <net/checksum.h>
|
2005-12-14 02:14:27 -05:00
|
|
|
#include <net/inetpeer.h>
|
2007-10-15 05:24:19 -04:00
|
|
|
#include <net/inet_frag.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <linux/tcp.h>
|
|
|
|
#include <linux/udp.h>
|
|
|
|
#include <linux/inet.h>
|
|
|
|
#include <linux/netfilter_ipv4.h>
|
|
|
|
|
|
|
|
/* NOTE. Logic of IP defragmentation is parallel to corresponding IPv6
|
|
|
|
* code now. If you change something here, _PLEASE_ update ipv6/reassembly.c
|
|
|
|
* as well. Or notify me, at least. --ANK
|
|
|
|
*/
|
|
|
|
|
2008-01-22 08:58:31 -05:00
|
|
|
static int sysctl_ipfrag_max_dist __read_mostly = 64;
|
2005-12-14 02:14:27 -05:00
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
struct ipfrag_skb_cb
|
|
|
|
{
|
|
|
|
struct inet_skb_parm h;
|
|
|
|
int offset;
|
|
|
|
};
|
|
|
|
|
|
|
|
#define FRAG_CB(skb) ((struct ipfrag_skb_cb*)((skb)->cb))
|
|
|
|
|
|
|
|
/* Describe an entry in the "incomplete datagrams" queue. */
|
|
|
|
struct ipq {
|
2007-10-15 05:24:19 -04:00
|
|
|
struct inet_frag_queue q;
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
u32 user;
|
2006-09-27 01:19:02 -04:00
|
|
|
__be32 saddr;
|
|
|
|
__be32 daddr;
|
|
|
|
__be16 id;
|
2005-04-16 18:20:36 -04:00
|
|
|
u8 protocol;
|
2005-12-14 02:14:27 -05:00
|
|
|
int iif;
|
|
|
|
unsigned int rid;
|
|
|
|
struct inet_peer *peer;
|
2005-04-16 18:20:36 -04:00
|
|
|
};
|
|
|
|
|
2007-10-15 05:31:52 -04:00
|
|
|
static struct inet_frags ip4_frags;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-01-22 09:06:23 -05:00
|
|
|
int ip_frag_nqueues(struct net *net)
|
2007-10-15 05:31:52 -04:00
|
|
|
{
|
2008-01-22 09:06:23 -05:00
|
|
|
return net->ipv4.frags.nqueues;
|
2007-10-15 05:31:52 -04:00
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-01-22 09:07:25 -05:00
|
|
|
int ip_frag_mem(struct net *net)
|
2007-10-15 05:31:52 -04:00
|
|
|
{
|
2008-01-22 09:07:25 -05:00
|
|
|
return atomic_read(&net->ipv4.frags.mem);
|
2007-10-15 05:31:52 -04:00
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev,
|
|
|
|
struct net_device *dev);
|
|
|
|
|
2007-10-17 22:46:47 -04:00
|
|
|
struct ip4_create_arg {
|
|
|
|
struct iphdr *iph;
|
|
|
|
u32 user;
|
|
|
|
};
|
|
|
|
|
2006-09-27 01:19:02 -04:00
|
|
|
static unsigned int ipqhashfn(__be16 id, __be32 saddr, __be32 daddr, u8 prot)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2006-09-27 01:19:02 -04:00
|
|
|
return jhash_3words((__force u32)id << 16 | prot,
|
|
|
|
(__force u32)saddr, (__force u32)daddr,
|
2007-10-15 05:31:52 -04:00
|
|
|
ip4_frags.rnd) & (INETFRAGS_HASHSZ - 1);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2007-10-15 05:38:08 -04:00
|
|
|
static unsigned int ip4_hashfn(struct inet_frag_queue *q)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2007-10-15 05:38:08 -04:00
|
|
|
struct ipq *ipq;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-15 05:38:08 -04:00
|
|
|
ipq = container_of(q, struct ipq, q);
|
|
|
|
return ipqhashfn(ipq->id, ipq->saddr, ipq->daddr, ipq->protocol);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2007-10-17 22:47:21 -04:00
|
|
|
static int ip4_frag_match(struct inet_frag_queue *q, void *a)
|
|
|
|
{
|
|
|
|
struct ipq *qp;
|
|
|
|
struct ip4_create_arg *arg = a;
|
|
|
|
|
|
|
|
qp = container_of(q, struct ipq, q);
|
|
|
|
return (qp->id == arg->iph->id &&
|
|
|
|
qp->saddr == arg->iph->saddr &&
|
|
|
|
qp->daddr == arg->iph->daddr &&
|
|
|
|
qp->protocol == arg->iph->protocol &&
|
|
|
|
qp->user == arg->user);
|
|
|
|
}
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
/* Memory Tracking Functions. */
|
2008-01-22 09:07:25 -05:00
|
|
|
static __inline__ void frag_kfree_skb(struct netns_frags *nf,
|
|
|
|
struct sk_buff *skb, int *work)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
if (work)
|
|
|
|
*work -= skb->truesize;
|
2008-01-22 09:07:25 -05:00
|
|
|
atomic_sub(skb->truesize, &nf->mem);
|
2005-04-16 18:20:36 -04:00
|
|
|
kfree_skb(skb);
|
|
|
|
}
|
|
|
|
|
2007-10-17 22:46:47 -04:00
|
|
|
static void ip4_frag_init(struct inet_frag_queue *q, void *a)
|
|
|
|
{
|
|
|
|
struct ipq *qp = container_of(q, struct ipq, q);
|
|
|
|
struct ip4_create_arg *arg = a;
|
|
|
|
|
|
|
|
qp->protocol = arg->iph->protocol;
|
|
|
|
qp->id = arg->iph->id;
|
|
|
|
qp->saddr = arg->iph->saddr;
|
|
|
|
qp->daddr = arg->iph->daddr;
|
|
|
|
qp->user = arg->user;
|
|
|
|
qp->peer = sysctl_ipfrag_max_dist ?
|
|
|
|
inet_getpeer(arg->iph->saddr, 1) : NULL;
|
|
|
|
}
|
|
|
|
|
2007-10-15 05:39:14 -04:00
|
|
|
static __inline__ void ip4_frag_free(struct inet_frag_queue *q)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2007-10-15 05:39:14 -04:00
|
|
|
struct ipq *qp;
|
|
|
|
|
|
|
|
qp = container_of(q, struct ipq, q);
|
|
|
|
if (qp->peer)
|
|
|
|
inet_putpeer(qp->peer);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Destruction primitives. */
|
|
|
|
|
2007-10-15 05:41:09 -04:00
|
|
|
static __inline__ void ipq_put(struct ipq *ipq)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2007-10-15 05:41:56 -04:00
|
|
|
inet_frag_put(&ipq->q, &ip4_frags);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Kill ipq entry. It is not destroyed immediately,
|
|
|
|
* because caller (and someone more) holds reference count.
|
|
|
|
*/
|
|
|
|
static void ipq_kill(struct ipq *ipq)
|
|
|
|
{
|
2007-10-15 05:37:18 -04:00
|
|
|
inet_frag_kill(&ipq->q, &ip4_frags);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2007-02-09 09:24:47 -05:00
|
|
|
/* Memory limiting on fragments. Evictor trashes the oldest
|
2005-04-16 18:20:36 -04:00
|
|
|
* fragment queue until we are back under the threshold.
|
|
|
|
*/
|
2008-01-22 09:07:25 -05:00
|
|
|
static void ip_evictor(struct net *net)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2007-10-15 05:40:06 -04:00
|
|
|
int evicted;
|
|
|
|
|
2008-01-22 09:07:25 -05:00
|
|
|
evicted = inet_frag_evictor(&net->ipv4.frags, &ip4_frags);
|
2007-10-15 05:40:06 -04:00
|
|
|
if (evicted)
|
|
|
|
IP_ADD_STATS_BH(IPSTATS_MIB_REASMFAILS, evicted);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Oops, a fragment queue timed out. Kill it and send an ICMP reply.
|
|
|
|
*/
|
|
|
|
static void ip_expire(unsigned long arg)
|
|
|
|
{
|
2007-10-17 22:45:23 -04:00
|
|
|
struct ipq *qp;
|
|
|
|
|
|
|
|
qp = container_of((struct inet_frag_queue *) arg, struct ipq, q);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
spin_lock(&qp->q.lock);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-03-28 19:35:27 -04:00
|
|
|
if (qp->q.last_in & INET_FRAG_COMPLETE)
|
2005-04-16 18:20:36 -04:00
|
|
|
goto out;
|
|
|
|
|
|
|
|
ipq_kill(qp);
|
|
|
|
|
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMTIMEOUT);
|
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS);
|
|
|
|
|
2008-03-28 19:35:27 -04:00
|
|
|
if ((qp->q.last_in & INET_FRAG_FIRST_IN) && qp->q.fragments != NULL) {
|
2007-10-15 05:24:19 -04:00
|
|
|
struct sk_buff *head = qp->q.fragments;
|
2008-03-24 18:31:00 -04:00
|
|
|
struct net *net;
|
|
|
|
|
|
|
|
net = container_of(qp->q.net, struct net, ipv4.frags);
|
2005-04-16 18:20:36 -04:00
|
|
|
/* Send an ICMP "Fragment Reassembly Timeout" message. */
|
2008-03-24 18:31:00 -04:00
|
|
|
if ((head->dev = dev_get_by_index(net, qp->iif)) != NULL) {
|
2005-04-16 18:20:36 -04:00
|
|
|
icmp_send(head, ICMP_TIME_EXCEEDED, ICMP_EXC_FRAGTIME, 0);
|
|
|
|
dev_put(head->dev);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
out:
|
2007-10-15 05:24:19 -04:00
|
|
|
spin_unlock(&qp->q.lock);
|
2007-10-15 05:41:09 -04:00
|
|
|
ipq_put(qp);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2007-10-17 22:47:21 -04:00
|
|
|
/* Find the correct entry in the "incomplete datagrams" queue for
|
|
|
|
* this IP datagram, and create new one, if nothing is found.
|
|
|
|
*/
|
2008-01-22 09:02:14 -05:00
|
|
|
static inline struct ipq *ip_find(struct net *net, struct iphdr *iph, u32 user)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2007-10-17 22:46:47 -04:00
|
|
|
struct inet_frag_queue *q;
|
|
|
|
struct ip4_create_arg arg;
|
2007-10-17 22:47:21 -04:00
|
|
|
unsigned int hash;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-17 22:46:47 -04:00
|
|
|
arg.iph = iph;
|
|
|
|
arg.user = user;
|
2007-10-17 22:47:21 -04:00
|
|
|
hash = ipqhashfn(iph->id, iph->saddr, iph->daddr, iph->protocol);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-01-22 09:02:14 -05:00
|
|
|
q = inet_frag_find(&net->ipv4.frags, &ip4_frags, &arg, hash);
|
2007-10-17 22:46:47 -04:00
|
|
|
if (q == NULL)
|
|
|
|
goto out_nomem;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-17 22:46:47 -04:00
|
|
|
return container_of(q, struct ipq, q);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
out_nomem:
|
2005-08-09 23:50:53 -04:00
|
|
|
LIMIT_NETDEBUG(KERN_ERR "ip_frag_create: no memory left !\n");
|
2005-04-16 18:20:36 -04:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-12-14 02:14:27 -05:00
|
|
|
/* Is the fragment too far ahead to be part of ipq? */
|
|
|
|
static inline int ip_frag_too_far(struct ipq *qp)
|
|
|
|
{
|
|
|
|
struct inet_peer *peer = qp->peer;
|
|
|
|
unsigned int max = sysctl_ipfrag_max_dist;
|
|
|
|
unsigned int start, end;
|
|
|
|
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
if (!peer || !max)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
start = qp->rid;
|
|
|
|
end = atomic_inc_return(&peer->rid);
|
|
|
|
qp->rid = end;
|
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
rc = qp->q.fragments && (end - start) > max;
|
2005-12-14 02:14:27 -05:00
|
|
|
|
|
|
|
if (rc) {
|
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS);
|
|
|
|
}
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ip_frag_reinit(struct ipq *qp)
|
|
|
|
{
|
|
|
|
struct sk_buff *fp;
|
|
|
|
|
2008-01-22 09:09:37 -05:00
|
|
|
if (!mod_timer(&qp->q.timer, jiffies + qp->q.net->timeout)) {
|
2007-10-15 05:24:19 -04:00
|
|
|
atomic_inc(&qp->q.refcnt);
|
2005-12-14 02:14:27 -05:00
|
|
|
return -ETIMEDOUT;
|
|
|
|
}
|
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
fp = qp->q.fragments;
|
2005-12-14 02:14:27 -05:00
|
|
|
do {
|
|
|
|
struct sk_buff *xp = fp->next;
|
2008-01-22 09:07:25 -05:00
|
|
|
frag_kfree_skb(qp->q.net, fp, NULL);
|
2005-12-14 02:14:27 -05:00
|
|
|
fp = xp;
|
|
|
|
} while (fp);
|
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.last_in = 0;
|
|
|
|
qp->q.len = 0;
|
|
|
|
qp->q.meat = 0;
|
|
|
|
qp->q.fragments = NULL;
|
2005-12-14 02:14:27 -05:00
|
|
|
qp->iif = 0;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
/* Add new segment to existing queue. */
|
2007-10-14 03:38:15 -04:00
|
|
|
static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
struct sk_buff *prev, *next;
|
2007-10-14 03:38:15 -04:00
|
|
|
struct net_device *dev;
|
2005-04-16 18:20:36 -04:00
|
|
|
int flags, offset;
|
|
|
|
int ihl, end;
|
2007-10-14 03:38:15 -04:00
|
|
|
int err = -ENOENT;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-03-28 19:35:27 -04:00
|
|
|
if (qp->q.last_in & INET_FRAG_COMPLETE)
|
2005-04-16 18:20:36 -04:00
|
|
|
goto err;
|
|
|
|
|
2005-12-14 02:14:27 -05:00
|
|
|
if (!(IPCB(skb)->flags & IPSKB_FRAG_COMPLETE) &&
|
2007-10-14 03:38:15 -04:00
|
|
|
unlikely(ip_frag_too_far(qp)) &&
|
|
|
|
unlikely(err = ip_frag_reinit(qp))) {
|
2005-12-14 02:14:27 -05:00
|
|
|
ipq_kill(qp);
|
|
|
|
goto err;
|
|
|
|
}
|
|
|
|
|
2007-04-21 01:47:35 -04:00
|
|
|
offset = ntohs(ip_hdr(skb)->frag_off);
|
2005-04-16 18:20:36 -04:00
|
|
|
flags = offset & ~IP_OFFSET;
|
|
|
|
offset &= IP_OFFSET;
|
|
|
|
offset <<= 3; /* offset is in 8-byte chunks */
|
2007-03-12 19:09:15 -04:00
|
|
|
ihl = ip_hdrlen(skb);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
/* Determine the position of this fragment. */
|
2007-02-09 09:24:47 -05:00
|
|
|
end = offset + skb->len - ihl;
|
2007-10-14 03:38:15 -04:00
|
|
|
err = -EINVAL;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
/* Is this the final fragment? */
|
|
|
|
if ((flags & IP_MF) == 0) {
|
|
|
|
/* If we already have some bits beyond end
|
|
|
|
* or have different end, the segment is corrrupted.
|
|
|
|
*/
|
2007-10-15 05:24:19 -04:00
|
|
|
if (end < qp->q.len ||
|
2008-03-28 19:35:27 -04:00
|
|
|
((qp->q.last_in & INET_FRAG_LAST_IN) && end != qp->q.len))
|
2005-04-16 18:20:36 -04:00
|
|
|
goto err;
|
2008-03-28 19:35:27 -04:00
|
|
|
qp->q.last_in |= INET_FRAG_LAST_IN;
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.len = end;
|
2005-04-16 18:20:36 -04:00
|
|
|
} else {
|
|
|
|
if (end&7) {
|
|
|
|
end &= ~7;
|
|
|
|
if (skb->ip_summed != CHECKSUM_UNNECESSARY)
|
|
|
|
skb->ip_summed = CHECKSUM_NONE;
|
|
|
|
}
|
2007-10-15 05:24:19 -04:00
|
|
|
if (end > qp->q.len) {
|
2005-04-16 18:20:36 -04:00
|
|
|
/* Some bits beyond end -> corruption. */
|
2008-03-28 19:35:27 -04:00
|
|
|
if (qp->q.last_in & INET_FRAG_LAST_IN)
|
2005-04-16 18:20:36 -04:00
|
|
|
goto err;
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.len = end;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
if (end == offset)
|
|
|
|
goto err;
|
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
err = -ENOMEM;
|
2005-04-16 18:20:36 -04:00
|
|
|
if (pskb_pull(skb, ihl) == NULL)
|
|
|
|
goto err;
|
2007-10-14 03:38:15 -04:00
|
|
|
|
|
|
|
err = pskb_trim_rcsum(skb, end - offset);
|
|
|
|
if (err)
|
2005-04-16 18:20:36 -04:00
|
|
|
goto err;
|
|
|
|
|
|
|
|
/* Find out which fragments are in front and at the back of us
|
|
|
|
* in the chain of fragments so far. We must know where to put
|
|
|
|
* this fragment, right?
|
|
|
|
*/
|
|
|
|
prev = NULL;
|
2007-10-15 05:24:19 -04:00
|
|
|
for (next = qp->q.fragments; next != NULL; next = next->next) {
|
2005-04-16 18:20:36 -04:00
|
|
|
if (FRAG_CB(next)->offset >= offset)
|
|
|
|
break; /* bingo! */
|
|
|
|
prev = next;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* We found where to put this one. Check for overlap with
|
|
|
|
* preceding fragment, and, if needed, align things so that
|
|
|
|
* any overlaps are eliminated.
|
|
|
|
*/
|
|
|
|
if (prev) {
|
|
|
|
int i = (FRAG_CB(prev)->offset + prev->len) - offset;
|
|
|
|
|
|
|
|
if (i > 0) {
|
|
|
|
offset += i;
|
2007-10-14 03:38:15 -04:00
|
|
|
err = -EINVAL;
|
2005-04-16 18:20:36 -04:00
|
|
|
if (end <= offset)
|
|
|
|
goto err;
|
2007-10-14 03:38:15 -04:00
|
|
|
err = -ENOMEM;
|
2005-04-16 18:20:36 -04:00
|
|
|
if (!pskb_pull(skb, i))
|
|
|
|
goto err;
|
|
|
|
if (skb->ip_summed != CHECKSUM_UNNECESSARY)
|
|
|
|
skb->ip_summed = CHECKSUM_NONE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
err = -ENOMEM;
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
while (next && FRAG_CB(next)->offset < end) {
|
|
|
|
int i = end - FRAG_CB(next)->offset; /* overlap is 'i' bytes */
|
|
|
|
|
|
|
|
if (i < next->len) {
|
|
|
|
/* Eat head of the next overlapped fragment
|
|
|
|
* and leave the loop. The next ones cannot overlap.
|
|
|
|
*/
|
|
|
|
if (!pskb_pull(next, i))
|
|
|
|
goto err;
|
|
|
|
FRAG_CB(next)->offset += i;
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.meat -= i;
|
2005-04-16 18:20:36 -04:00
|
|
|
if (next->ip_summed != CHECKSUM_UNNECESSARY)
|
|
|
|
next->ip_summed = CHECKSUM_NONE;
|
|
|
|
break;
|
|
|
|
} else {
|
|
|
|
struct sk_buff *free_it = next;
|
|
|
|
|
2006-12-12 13:48:59 -05:00
|
|
|
/* Old fragment is completely overridden with
|
2005-04-16 18:20:36 -04:00
|
|
|
* new one drop it.
|
|
|
|
*/
|
|
|
|
next = next->next;
|
|
|
|
|
|
|
|
if (prev)
|
|
|
|
prev->next = next;
|
|
|
|
else
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.fragments = next;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.meat -= free_it->len;
|
2008-01-22 09:07:25 -05:00
|
|
|
frag_kfree_skb(qp->q.net, free_it, NULL);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
FRAG_CB(skb)->offset = offset;
|
|
|
|
|
|
|
|
/* Insert this fragment in the chain of fragments. */
|
|
|
|
skb->next = next;
|
|
|
|
if (prev)
|
|
|
|
prev->next = skb;
|
|
|
|
else
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.fragments = skb;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
dev = skb->dev;
|
|
|
|
if (dev) {
|
|
|
|
qp->iif = dev->ifindex;
|
|
|
|
skb->dev = NULL;
|
|
|
|
}
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.stamp = skb->tstamp;
|
|
|
|
qp->q.meat += skb->len;
|
2008-01-22 09:07:25 -05:00
|
|
|
atomic_add(skb->truesize, &qp->q.net->mem);
|
2005-04-16 18:20:36 -04:00
|
|
|
if (offset == 0)
|
2008-03-28 19:35:27 -04:00
|
|
|
qp->q.last_in |= INET_FRAG_FIRST_IN;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-03-28 19:35:27 -04:00
|
|
|
if (qp->q.last_in == (INET_FRAG_FIRST_IN | INET_FRAG_LAST_IN) &&
|
|
|
|
qp->q.meat == qp->q.len)
|
2007-10-14 03:38:15 -04:00
|
|
|
return ip_frag_reasm(qp, prev, dev);
|
|
|
|
|
2007-10-15 05:31:52 -04:00
|
|
|
write_lock(&ip4_frags.lock);
|
2008-01-22 09:11:48 -05:00
|
|
|
list_move_tail(&qp->q.lru_list, &qp->q.net->lru_list);
|
2007-10-15 05:31:52 -04:00
|
|
|
write_unlock(&ip4_frags.lock);
|
2007-10-14 03:38:15 -04:00
|
|
|
return -EINPROGRESS;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
err:
|
|
|
|
kfree_skb(skb);
|
2007-10-14 03:38:15 -04:00
|
|
|
return err;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Build a new IP datagram from all its fragments. */
|
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev,
|
|
|
|
struct net_device *dev)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
struct iphdr *iph;
|
2007-10-15 05:24:19 -04:00
|
|
|
struct sk_buff *fp, *head = qp->q.fragments;
|
2005-04-16 18:20:36 -04:00
|
|
|
int len;
|
|
|
|
int ihlen;
|
2007-10-14 03:38:15 -04:00
|
|
|
int err;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
ipq_kill(qp);
|
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
/* Make the one we just received the head. */
|
|
|
|
if (prev) {
|
|
|
|
head = prev->next;
|
|
|
|
fp = skb_clone(head, GFP_ATOMIC);
|
|
|
|
if (!fp)
|
|
|
|
goto out_nomem;
|
|
|
|
|
|
|
|
fp->next = head->next;
|
|
|
|
prev->next = fp;
|
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
skb_morph(head, qp->q.fragments);
|
|
|
|
head->next = qp->q.fragments->next;
|
2007-10-14 03:38:15 -04:00
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
kfree_skb(qp->q.fragments);
|
|
|
|
qp->q.fragments = head;
|
2007-10-14 03:38:15 -04:00
|
|
|
}
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
BUG_TRAP(head != NULL);
|
|
|
|
BUG_TRAP(FRAG_CB(head)->offset == 0);
|
|
|
|
|
|
|
|
/* Allocate a new buffer for the datagram. */
|
2007-03-12 19:09:15 -04:00
|
|
|
ihlen = ip_hdrlen(head);
|
2007-10-15 05:24:19 -04:00
|
|
|
len = ihlen + qp->q.len;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
err = -E2BIG;
|
2007-03-08 23:44:43 -05:00
|
|
|
if (len > 65535)
|
2005-04-16 18:20:36 -04:00
|
|
|
goto out_oversize;
|
|
|
|
|
|
|
|
/* Head of list must not be cloned. */
|
|
|
|
if (skb_cloned(head) && pskb_expand_head(head, 0, 0, GFP_ATOMIC))
|
|
|
|
goto out_nomem;
|
|
|
|
|
|
|
|
/* If the first fragment is fragmented itself, we split
|
|
|
|
* it to two chunks: the first with data and paged part
|
|
|
|
* and the second, holding only fragments. */
|
|
|
|
if (skb_shinfo(head)->frag_list) {
|
|
|
|
struct sk_buff *clone;
|
|
|
|
int i, plen = 0;
|
|
|
|
|
|
|
|
if ((clone = alloc_skb(0, GFP_ATOMIC)) == NULL)
|
|
|
|
goto out_nomem;
|
|
|
|
clone->next = head->next;
|
|
|
|
head->next = clone;
|
|
|
|
skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list;
|
|
|
|
skb_shinfo(head)->frag_list = NULL;
|
|
|
|
for (i=0; i<skb_shinfo(head)->nr_frags; i++)
|
|
|
|
plen += skb_shinfo(head)->frags[i].size;
|
|
|
|
clone->len = clone->data_len = head->data_len - plen;
|
|
|
|
head->data_len -= clone->len;
|
|
|
|
head->len -= clone->len;
|
|
|
|
clone->csum = 0;
|
|
|
|
clone->ip_summed = head->ip_summed;
|
2008-01-22 09:07:25 -05:00
|
|
|
atomic_add(clone->truesize, &qp->q.net->mem);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
skb_shinfo(head)->frag_list = head->next;
|
2007-04-10 23:50:43 -04:00
|
|
|
skb_push(head, head->data - skb_network_header(head));
|
2008-01-22 09:07:25 -05:00
|
|
|
atomic_sub(head->truesize, &qp->q.net->mem);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
for (fp=head->next; fp; fp = fp->next) {
|
|
|
|
head->data_len += fp->len;
|
|
|
|
head->len += fp->len;
|
|
|
|
if (head->ip_summed != fp->ip_summed)
|
|
|
|
head->ip_summed = CHECKSUM_NONE;
|
2006-08-29 19:44:56 -04:00
|
|
|
else if (head->ip_summed == CHECKSUM_COMPLETE)
|
2005-04-16 18:20:36 -04:00
|
|
|
head->csum = csum_add(head->csum, fp->csum);
|
|
|
|
head->truesize += fp->truesize;
|
2008-01-22 09:07:25 -05:00
|
|
|
atomic_sub(fp->truesize, &qp->q.net->mem);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
head->next = NULL;
|
|
|
|
head->dev = dev;
|
2007-10-15 05:24:19 -04:00
|
|
|
head->tstamp = qp->q.stamp;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-04-21 01:47:35 -04:00
|
|
|
iph = ip_hdr(head);
|
2005-04-16 18:20:36 -04:00
|
|
|
iph->frag_off = 0;
|
|
|
|
iph->tot_len = htons(len);
|
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMOKS);
|
2007-10-15 05:24:19 -04:00
|
|
|
qp->q.fragments = NULL;
|
2007-10-14 03:38:15 -04:00
|
|
|
return 0;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
out_nomem:
|
2007-02-09 09:24:47 -05:00
|
|
|
LIMIT_NETDEBUG(KERN_ERR "IP: queue_glue: no memory for gluing "
|
2005-08-09 23:50:53 -04:00
|
|
|
"queue %p\n", qp);
|
2007-10-18 00:37:22 -04:00
|
|
|
err = -ENOMEM;
|
2005-04-16 18:20:36 -04:00
|
|
|
goto out_fail;
|
|
|
|
out_oversize:
|
|
|
|
if (net_ratelimit())
|
|
|
|
printk(KERN_INFO
|
|
|
|
"Oversized IP packet from %d.%d.%d.%d.\n",
|
|
|
|
NIPQUAD(qp->saddr));
|
|
|
|
out_fail:
|
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS);
|
2007-10-14 03:38:15 -04:00
|
|
|
return err;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Process an incoming IP datagram fragment. */
|
2007-10-14 03:38:32 -04:00
|
|
|
int ip_defrag(struct sk_buff *skb, u32 user)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
struct ipq *qp;
|
2008-01-22 09:02:14 -05:00
|
|
|
struct net *net;
|
2007-02-09 09:24:47 -05:00
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMREQDS);
|
|
|
|
|
2008-03-25 08:47:49 -04:00
|
|
|
net = skb->dev ? dev_net(skb->dev) : dev_net(skb->dst->dev);
|
2005-04-16 18:20:36 -04:00
|
|
|
/* Start by cleaning up the memory. */
|
2008-01-22 09:10:13 -05:00
|
|
|
if (atomic_read(&net->ipv4.frags.mem) > net->ipv4.frags.high_thresh)
|
2008-01-22 09:07:25 -05:00
|
|
|
ip_evictor(net);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
/* Lookup (or create) queue header */
|
2008-01-22 09:02:14 -05:00
|
|
|
if ((qp = ip_find(net, ip_hdr(skb), user)) != NULL) {
|
2007-10-14 03:38:15 -04:00
|
|
|
int ret;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
spin_lock(&qp->q.lock);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-14 03:38:15 -04:00
|
|
|
ret = ip_frag_queue(qp, skb);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2007-10-15 05:24:19 -04:00
|
|
|
spin_unlock(&qp->q.lock);
|
2007-10-15 05:41:09 -04:00
|
|
|
ipq_put(qp);
|
2007-10-14 03:38:32 -04:00
|
|
|
return ret;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
IP_INC_STATS_BH(IPSTATS_MIB_REASMFAILS);
|
|
|
|
kfree_skb(skb);
|
2007-10-14 03:38:32 -04:00
|
|
|
return -ENOMEM;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2008-01-22 08:58:31 -05:00
|
|
|
#ifdef CONFIG_SYSCTL
|
|
|
|
static int zero;
|
|
|
|
|
|
|
|
static struct ctl_table ip4_frags_ctl_table[] = {
|
|
|
|
{
|
|
|
|
.ctl_name = NET_IPV4_IPFRAG_HIGH_THRESH,
|
|
|
|
.procname = "ipfrag_high_thresh",
|
2008-01-22 09:10:13 -05:00
|
|
|
.data = &init_net.ipv4.frags.high_thresh,
|
2008-01-22 08:58:31 -05:00
|
|
|
.maxlen = sizeof(int),
|
|
|
|
.mode = 0644,
|
|
|
|
.proc_handler = &proc_dointvec
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.ctl_name = NET_IPV4_IPFRAG_LOW_THRESH,
|
|
|
|
.procname = "ipfrag_low_thresh",
|
2008-01-22 09:10:13 -05:00
|
|
|
.data = &init_net.ipv4.frags.low_thresh,
|
2008-01-22 08:58:31 -05:00
|
|
|
.maxlen = sizeof(int),
|
|
|
|
.mode = 0644,
|
|
|
|
.proc_handler = &proc_dointvec
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.ctl_name = NET_IPV4_IPFRAG_TIME,
|
|
|
|
.procname = "ipfrag_time",
|
2008-01-22 09:09:37 -05:00
|
|
|
.data = &init_net.ipv4.frags.timeout,
|
2008-01-22 08:58:31 -05:00
|
|
|
.maxlen = sizeof(int),
|
|
|
|
.mode = 0644,
|
|
|
|
.proc_handler = &proc_dointvec_jiffies,
|
|
|
|
.strategy = &sysctl_jiffies
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.ctl_name = NET_IPV4_IPFRAG_SECRET_INTERVAL,
|
|
|
|
.procname = "ipfrag_secret_interval",
|
2008-01-22 09:11:04 -05:00
|
|
|
.data = &ip4_frags.secret_interval,
|
2008-01-22 08:58:31 -05:00
|
|
|
.maxlen = sizeof(int),
|
|
|
|
.mode = 0644,
|
|
|
|
.proc_handler = &proc_dointvec_jiffies,
|
|
|
|
.strategy = &sysctl_jiffies
|
|
|
|
},
|
|
|
|
{
|
|
|
|
.procname = "ipfrag_max_dist",
|
|
|
|
.data = &sysctl_ipfrag_max_dist,
|
|
|
|
.maxlen = sizeof(int),
|
|
|
|
.mode = 0644,
|
|
|
|
.proc_handler = &proc_dointvec_minmax,
|
|
|
|
.extra1 = &zero
|
|
|
|
},
|
|
|
|
{ }
|
|
|
|
};
|
|
|
|
|
|
|
|
static int ip4_frags_ctl_register(struct net *net)
|
|
|
|
{
|
2008-01-22 09:08:36 -05:00
|
|
|
struct ctl_table *table;
|
2008-01-22 08:58:31 -05:00
|
|
|
struct ctl_table_header *hdr;
|
|
|
|
|
2008-01-22 09:08:36 -05:00
|
|
|
table = ip4_frags_ctl_table;
|
|
|
|
if (net != &init_net) {
|
|
|
|
table = kmemdup(table, sizeof(ip4_frags_ctl_table), GFP_KERNEL);
|
|
|
|
if (table == NULL)
|
|
|
|
goto err_alloc;
|
|
|
|
|
2008-01-22 09:10:13 -05:00
|
|
|
table[0].data = &net->ipv4.frags.high_thresh;
|
|
|
|
table[1].data = &net->ipv4.frags.low_thresh;
|
2008-01-22 09:09:37 -05:00
|
|
|
table[2].data = &net->ipv4.frags.timeout;
|
2008-01-22 09:08:36 -05:00
|
|
|
table[3].mode &= ~0222;
|
|
|
|
table[4].mode &= ~0222;
|
|
|
|
}
|
|
|
|
|
|
|
|
hdr = register_net_sysctl_table(net, net_ipv4_ctl_path, table);
|
|
|
|
if (hdr == NULL)
|
|
|
|
goto err_reg;
|
|
|
|
|
|
|
|
net->ipv4.frags_hdr = hdr;
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_reg:
|
|
|
|
if (net != &init_net)
|
|
|
|
kfree(table);
|
|
|
|
err_alloc:
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ip4_frags_ctl_unregister(struct net *net)
|
|
|
|
{
|
|
|
|
struct ctl_table *table;
|
|
|
|
|
|
|
|
table = net->ipv4.frags_hdr->ctl_table_arg;
|
|
|
|
unregister_net_sysctl_table(net->ipv4.frags_hdr);
|
|
|
|
kfree(table);
|
2008-01-22 08:58:31 -05:00
|
|
|
}
|
|
|
|
#else
|
|
|
|
static inline int ip4_frags_ctl_register(struct net *net)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
2008-01-22 09:08:36 -05:00
|
|
|
|
|
|
|
static inline void ip4_frags_ctl_unregister(struct net *net)
|
|
|
|
{
|
|
|
|
}
|
2008-01-22 08:58:31 -05:00
|
|
|
#endif
|
|
|
|
|
|
|
|
static int ipv4_frags_init_net(struct net *net)
|
|
|
|
{
|
2008-01-22 09:10:13 -05:00
|
|
|
/*
|
|
|
|
* Fragment cache limits. We will commit 256K at one time. Should we
|
|
|
|
* cross that limit we will prune down to 192K. This should cope with
|
|
|
|
* even the most extreme cases without allowing an attacker to
|
|
|
|
* measurably harm machine performance.
|
|
|
|
*/
|
|
|
|
net->ipv4.frags.high_thresh = 256 * 1024;
|
|
|
|
net->ipv4.frags.low_thresh = 192 * 1024;
|
2008-01-22 09:09:37 -05:00
|
|
|
/*
|
|
|
|
* Important NOTE! Fragment queue must be destroyed before MSL expires.
|
|
|
|
* RFC791 is wrong proposing to prolongate timer each fragment arrival
|
|
|
|
* by TTL.
|
|
|
|
*/
|
|
|
|
net->ipv4.frags.timeout = IP_FRAG_TIME;
|
|
|
|
|
2008-01-22 09:06:23 -05:00
|
|
|
inet_frags_init_net(&net->ipv4.frags);
|
|
|
|
|
2008-01-22 08:58:31 -05:00
|
|
|
return ip4_frags_ctl_register(net);
|
|
|
|
}
|
|
|
|
|
2008-01-22 09:12:39 -05:00
|
|
|
static void ipv4_frags_exit_net(struct net *net)
|
|
|
|
{
|
|
|
|
ip4_frags_ctl_unregister(net);
|
|
|
|
inet_frags_exit_net(&net->ipv4.frags, &ip4_frags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pernet_operations ip4_frags_ops = {
|
|
|
|
.init = ipv4_frags_init_net,
|
|
|
|
.exit = ipv4_frags_exit_net,
|
|
|
|
};
|
|
|
|
|
2007-04-19 19:16:32 -04:00
|
|
|
void __init ipfrag_init(void)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2008-01-22 09:12:39 -05:00
|
|
|
register_pernet_subsys(&ip4_frags_ops);
|
2007-10-15 05:38:08 -04:00
|
|
|
ip4_frags.hashfn = ip4_hashfn;
|
2007-10-17 22:46:47 -04:00
|
|
|
ip4_frags.constructor = ip4_frag_init;
|
2007-10-15 05:39:14 -04:00
|
|
|
ip4_frags.destructor = ip4_frag_free;
|
|
|
|
ip4_frags.skb_free = NULL;
|
|
|
|
ip4_frags.qsize = sizeof(struct ipq);
|
2007-10-17 22:47:21 -04:00
|
|
|
ip4_frags.match = ip4_frag_match;
|
2007-10-17 22:45:23 -04:00
|
|
|
ip4_frags.frag_expire = ip_expire;
|
2008-01-22 09:11:04 -05:00
|
|
|
ip4_frags.secret_interval = 10 * 60 * HZ;
|
2007-10-15 05:31:52 -04:00
|
|
|
inet_frags_init(&ip4_frags);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
EXPORT_SYMBOL(ip_defrag);
|