[linux-2.6-block.git] / net / sched / act_csum.c

/*
 * Checksum updating actions
 *
 * Copyright (c) 2010 Gregoire Baron <baronchon@n7mm.org>
 *
 * 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.
 *
 */

#include <linux/types.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/spinlock.h>

#include <linux/netlink.h>
#include <net/netlink.h>
#include <linux/rtnetlink.h>

#include <linux/skbuff.h>

#include <net/ip.h>
#include <net/ipv6.h>
#include <net/icmp.h>
#include <linux/icmpv6.h>
#include <linux/igmp.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <net/ip6_checksum.h>

#include <net/act_api.h>

#include <linux/tc_act/tc_csum.h>
#include <net/tc_act/tc_csum.h>

#define CSUM_TAB_MASK 15
static struct tcf_common *tcf_csum_ht[CSUM_TAB_MASK + 1];
static u32 csum_idx_gen;
static DEFINE_RWLOCK(csum_lock);

static struct tcf_hashinfo csum_hash_info = {
	.htab	= tcf_csum_ht,
	.hmask	= CSUM_TAB_MASK,
	.lock	= &csum_lock,
};

static const struct nla_policy csum_policy[TCA_CSUM_MAX + 1] = {
	[TCA_CSUM_PARMS] = { .len = sizeof(struct tc_csum), },
};

static int tcf_csum_init(struct net *n, struct nlattr *nla, struct nlattr *est,
			 struct tc_action *a, int ovr, int bind)
{
	struct nlattr *tb[TCA_CSUM_MAX + 1];
	struct tc_csum *parm;
	struct tcf_common *pc;
	struct tcf_csum *p;
	int ret = 0, err;

	if (nla == NULL)
		return -EINVAL;

	err = nla_parse_nested(tb, TCA_CSUM_MAX, nla, csum_policy);
	if (err < 0)
		return err;

	if (tb[TCA_CSUM_PARMS] == NULL)
		return -EINVAL;
	parm = nla_data(tb[TCA_CSUM_PARMS]);

	pc = tcf_hash_check(parm->index, a, bind, &csum_hash_info);
	if (!pc) {
		pc = tcf_hash_create(parm->index, est, a, sizeof(*p), bind,
				     &csum_idx_gen, &csum_hash_info);
		if (IS_ERR(pc))
			return PTR_ERR(pc);
		ret = ACT_P_CREATED;
	} else {
		if (bind)/* dont override defaults */
			return 0;
		tcf_hash_release(pc, bind, &csum_hash_info);
		if (!ovr)
			return -EEXIST;
	}

	p = to_tcf_csum(pc);
	spin_lock_bh(&p->tcf_lock);
	p->tcf_action = parm->action;
	p->update_flags = parm->update_flags;
	spin_unlock_bh(&p->tcf_lock);

	if (ret == ACT_P_CREATED)
		tcf_hash_insert(pc, &csum_hash_info);

	return ret;
}

static int tcf_csum_cleanup(struct tc_action *a, int bind)
{
	struct tcf_csum *p = a->priv;
	return tcf_hash_release(&p->common, bind, &csum_hash_info);
}

/**
 * tcf_csum_skb_nextlayer - Get next layer pointer
 * @skb: sk_buff to use
 * @ihl: previous summed headers length
 * @ipl: complete packet length
 * @jhl: next header length
 *
 * Check the expected next layer availability in the specified sk_buff.
 * Return the next layer pointer if pass, NULL otherwise.
 */
static void *tcf_csum_skb_nextlayer(struct sk_buff *skb,
				    unsigned int ihl, unsigned int ipl,
				    unsigned int jhl)
{
	int ntkoff = skb_network_offset(skb);
	int hl = ihl + jhl;

	if (!pskb_may_pull(skb, ipl + ntkoff) || (ipl < hl) ||
	    (skb_cloned(skb) &&
	     !skb_clone_writable(skb, hl + ntkoff) &&
	     pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
		return NULL;
	else
		return (void *)(skb_network_header(skb) + ihl);
}

static int tcf_csum_ipv4_icmp(struct sk_buff *skb,
			      unsigned int ihl, unsigned int ipl)
{
	struct icmphdr *icmph;

	icmph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*icmph));
	if (icmph == NULL)
		return 0;

	icmph->checksum = 0;
	skb->csum = csum_partial(icmph, ipl - ihl, 0);
	icmph->checksum = csum_fold(skb->csum);

	skb->ip_summed = CHECKSUM_NONE;

	return 1;
}

static int tcf_csum_ipv4_igmp(struct sk_buff *skb,
			      unsigned int ihl, unsigned int ipl)
{
	struct igmphdr *igmph;

	igmph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*igmph));
	if (igmph == NULL)
		return 0;

	igmph->csum = 0;
	skb->csum = csum_partial(igmph, ipl - ihl, 0);
	igmph->csum = csum_fold(skb->csum);

	skb->ip_summed = CHECKSUM_NONE;

	return 1;
}

static int tcf_csum_ipv6_icmp(struct sk_buff *skb,
			      unsigned int ihl, unsigned int ipl)
{
	struct icmp6hdr *icmp6h;
	const struct ipv6hdr *ip6h;

	icmp6h = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*icmp6h));
	if (icmp6h == NULL)
		return 0;

	ip6h = ipv6_hdr(skb);
	icmp6h->icmp6_cksum = 0;
	skb->csum = csum_partial(icmp6h, ipl - ihl, 0);
	icmp6h->icmp6_cksum = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
					      ipl - ihl, IPPROTO_ICMPV6,
					      skb->csum);

	skb->ip_summed = CHECKSUM_NONE;

	return 1;
}

static int tcf_csum_ipv4_tcp(struct sk_buff *skb,
			     unsigned int ihl, unsigned int ipl)
{
	struct tcphdr *tcph;
	const struct iphdr *iph;

	tcph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*tcph));
	if (tcph == NULL)
		return 0;

	iph = ip_hdr(skb);
	tcph->check = 0;
	skb->csum = csum_partial(tcph, ipl - ihl, 0);
	tcph->check = tcp_v4_check(ipl - ihl,
				   iph->saddr, iph->daddr, skb->csum);

	skb->ip_summed = CHECKSUM_NONE;

	return 1;
}

static int tcf_csum_ipv6_tcp(struct sk_buff *skb,
			     unsigned int ihl, unsigned int ipl)
{
	struct tcphdr *tcph;
	const struct ipv6hdr *ip6h;

	tcph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*tcph));
	if (tcph == NULL)
		return 0;

	ip6h = ipv6_hdr(skb);
	tcph->check = 0;
	skb->csum = csum_partial(tcph, ipl - ihl, 0);
	tcph->check = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
				      ipl - ihl, IPPROTO_TCP,
				      skb->csum);

	skb->ip_summed = CHECKSUM_NONE;

	return 1;
}

static int tcf_csum_ipv4_udp(struct sk_buff *skb,
			     unsigned int ihl, unsigned int ipl, int udplite)
{
	struct udphdr *udph;
	const struct iphdr *iph;
	u16 ul;

	/*
	 * Support both UDP and UDPLITE checksum algorithms, Don't use
	 * udph->len to get the real length without any protocol check,
	 * UDPLITE uses udph->len for another thing,
	 * Use iph->tot_len, or just ipl.
	 */

	udph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*udph));
	if (udph == NULL)
		return 0;

	iph = ip_hdr(skb);
	ul = ntohs(udph->len);

	if (udplite || udph->check) {

		udph->check = 0;

		if (udplite) {
			if (ul == 0)
				skb->csum = csum_partial(udph, ipl - ihl, 0);
			else if ((ul >= sizeof(*udph)) && (ul <= ipl - ihl))
				skb->csum = csum_partial(udph, ul, 0);
			else
				goto ignore_obscure_skb;
		} else {
			if (ul != ipl - ihl)
				goto ignore_obscure_skb;

			skb->csum = csum_partial(udph, ul, 0);
		}

		udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
						ul, iph->protocol,
						skb->csum);

		if (!udph->check)
			udph->check = CSUM_MANGLED_0;
	}

	skb->ip_summed = CHECKSUM_NONE;

ignore_obscure_skb:
	return 1;
}

static int tcf_csum_ipv6_udp(struct sk_buff *skb,
			     unsigned int ihl, unsigned int ipl, int udplite)
{
	struct udphdr *udph;
	const struct ipv6hdr *ip6h;
	u16 ul;

	/*
	 * Support both UDP and UDPLITE checksum algorithms, Don't use
	 * udph->len to get the real length without any protocol check,
	 * UDPLITE uses udph->len for another thing,
	 * Use ip6h->payload_len + sizeof(*ip6h) ... , or just ipl.
	 */

	udph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*udph));
	if (udph == NULL)
		return 0;

	ip6h = ipv6_hdr(skb);
	ul = ntohs(udph->len);

	udph->check = 0;

	if (udplite) {
		if (ul == 0)
			skb->csum = csum_partial(udph, ipl - ihl, 0);

		else if ((ul >= sizeof(*udph)) && (ul <= ipl - ihl))
			skb->csum = csum_partial(udph, ul, 0);

		else
			goto ignore_obscure_skb;
	} else {
		if (ul != ipl - ihl)
			goto ignore_obscure_skb;

		skb->csum = csum_partial(udph, ul, 0);
	}

	udph->check = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr, ul,
				      udplite ? IPPROTO_UDPLITE : IPPROTO_UDP,
				      skb->csum);

	if (!udph->check)
		udph->check = CSUM_MANGLED_0;

	skb->ip_summed = CHECKSUM_NONE;

ignore_obscure_skb:
	return 1;
}

static int tcf_csum_ipv4(struct sk_buff *skb, u32 update_flags)
{
	const struct iphdr *iph;
	int ntkoff;

	ntkoff = skb_network_offset(skb);

	if (!pskb_may_pull(skb, sizeof(*iph) + ntkoff))
		goto fail;

	iph = ip_hdr(skb);

	switch (iph->frag_off & htons(IP_OFFSET) ? 0 : iph->protocol) {
	case IPPROTO_ICMP:
		if (update_flags & TCA_CSUM_UPDATE_FLAG_ICMP)
			if (!tcf_csum_ipv4_icmp(skb, iph->ihl * 4,
						ntohs(iph->tot_len)))
				goto fail;
		break;
	case IPPROTO_IGMP:
		if (update_flags & TCA_CSUM_UPDATE_FLAG_IGMP)
			if (!tcf_csum_ipv4_igmp(skb, iph->ihl * 4,
						ntohs(iph->tot_len)))
				goto fail;
		break;
	case IPPROTO_TCP:
		if (update_flags & TCA_CSUM_UPDATE_FLAG_TCP)
			if (!tcf_csum_ipv4_tcp(skb, iph->ihl * 4,
					       ntohs(iph->tot_len)))
				goto fail;
		break;
	case IPPROTO_UDP:
		if (update_flags & TCA_CSUM_UPDATE_FLAG_UDP)
			if (!tcf_csum_ipv4_udp(skb, iph->ihl * 4,
					       ntohs(iph->tot_len), 0))
				goto fail;
		break;
	case IPPROTO_UDPLITE:
		if (update_flags & TCA_CSUM_UPDATE_FLAG_UDPLITE)
			if (!tcf_csum_ipv4_udp(skb, iph->ihl * 4,
					       ntohs(iph->tot_len), 1))
				goto fail;
		break;
	}

	if (update_flags & TCA_CSUM_UPDATE_FLAG_IPV4HDR) {
		if (skb_cloned(skb) &&
		    !skb_clone_writable(skb, sizeof(*iph) + ntkoff) &&
		    pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
			goto fail;

		ip_send_check(ip_hdr(skb));
	}

	return 1;

fail:
	return 0;
}

static int tcf_csum_ipv6_hopopts(struct ipv6_opt_hdr *ip6xh,
				 unsigned int ixhl, unsigned int *pl)
{
	int off, len, optlen;
	unsigned char *xh = (void *)ip6xh;

	off = sizeof(*ip6xh);
	len = ixhl - off;

	while (len > 1) {
		switch (xh[off]) {
		case IPV6_TLV_PAD1:
			optlen = 1;
			break;
		case IPV6_TLV_JUMBO:
			optlen = xh[off + 1] + 2;
			if (optlen != 6 || len < 6 || (off & 3) != 2)
				/* wrong jumbo option length/alignment */
				return 0;
			*pl = ntohl(*(__be32 *)(xh + off + 2));
			goto done;
		default:
			optlen = xh[off + 1] + 2;
			if (optlen > len)
				/* ignore obscure options */
				goto done;
			break;
		}
		off += optlen;
		len -= optlen;
	}

done:
	return 1;
}

static int tcf_csum_ipv6(struct sk_buff *skb, u32 update_flags)
{
	struct ipv6hdr *ip6h;
	struct ipv6_opt_hdr *ip6xh;
	unsigned int hl, ixhl;
	unsigned int pl;
	int ntkoff;
	u8 nexthdr;

	ntkoff = skb_network_offset(skb);

	hl = sizeof(*ip6h);

	if (!pskb_may_pull(skb, hl + ntkoff))
		goto fail;

	ip6h = ipv6_hdr(skb);

	pl = ntohs(ip6h->payload_len);
	nexthdr = ip6h->nexthdr;

	do {
		switch (nexthdr) {
		case NEXTHDR_FRAGMENT:
			goto ignore_skb;
		case NEXTHDR_ROUTING:
		case NEXTHDR_HOP:
		case NEXTHDR_DEST:
			if (!pskb_may_pull(skb, hl + sizeof(*ip6xh) + ntkoff))
				goto fail;
			ip6xh = (void *)(skb_network_header(skb) + hl);
			ixhl = ipv6_optlen(ip6xh);
			if (!pskb_may_pull(skb, hl + ixhl + ntkoff))
				goto fail;
			ip6xh = (void *)(skb_network_header(skb) + hl);
			if ((nexthdr == NEXTHDR_HOP) &&
			    !(tcf_csum_ipv6_hopopts(ip6xh, ixhl, &pl)))
				goto fail;
			nexthdr = ip6xh->nexthdr;
			hl += ixhl;
			break;
		case IPPROTO_ICMPV6:
			if (update_flags & TCA_CSUM_UPDATE_FLAG_ICMP)
				if (!tcf_csum_ipv6_icmp(skb,
							hl, pl + sizeof(*ip6h)))
					goto fail;
			goto done;
		case IPPROTO_TCP:
			if (update_flags & TCA_CSUM_UPDATE_FLAG_TCP)
				if (!tcf_csum_ipv6_tcp(skb,
						       hl, pl + sizeof(*ip6h)))
					goto fail;
			goto done;
		case IPPROTO_UDP:
			if (update_flags & TCA_CSUM_UPDATE_FLAG_UDP)
				if (!tcf_csum_ipv6_udp(skb, hl,
						       pl + sizeof(*ip6h), 0))
					goto fail;
			goto done;
		case IPPROTO_UDPLITE:
			if (update_flags & TCA_CSUM_UPDATE_FLAG_UDPLITE)
				if (!tcf_csum_ipv6_udp(skb, hl,
						       pl + sizeof(*ip6h), 1))
					goto fail;
			goto done;
		default:
			goto ignore_skb;
		}
	} while (pskb_may_pull(skb, hl + 1 + ntkoff));

done:
ignore_skb:
	return 1;

fail:
	return 0;
}

static int tcf_csum(struct sk_buff *skb,
		    const struct tc_action *a, struct tcf_result *res)
{
	struct tcf_csum *p = a->priv;
	int action;
	u32 update_flags;

	spin_lock(&p->tcf_lock);
	p->tcf_tm.lastuse = jiffies;
	bstats_update(&p->tcf_bstats, skb);
	action = p->tcf_action;
	update_flags = p->update_flags;
	spin_unlock(&p->tcf_lock);

	if (unlikely(action == TC_ACT_SHOT))
		goto drop;

	switch (skb->protocol) {
	case cpu_to_be16(ETH_P_IP):
		if (!tcf_csum_ipv4(skb, update_flags))
			goto drop;
		break;
	case cpu_to_be16(ETH_P_IPV6):
		if (!tcf_csum_ipv6(skb, update_flags))
			goto drop;
		break;
	}

	return action;

drop:
	spin_lock(&p->tcf_lock);
	p->tcf_qstats.drops++;
	spin_unlock(&p->tcf_lock);
	return TC_ACT_SHOT;
}

static int tcf_csum_dump(struct sk_buff *skb,
			 struct tc_action *a, int bind, int ref)
{
	unsigned char *b = skb_tail_pointer(skb);
	struct tcf_csum *p = a->priv;
	struct tc_csum opt = {
		.update_flags = p->update_flags,
		.index   = p->tcf_index,
		.action  = p->tcf_action,
		.refcnt  = p->tcf_refcnt - ref,
		.bindcnt = p->tcf_bindcnt - bind,
	};
	struct tcf_t t;

	if (nla_put(skb, TCA_CSUM_PARMS, sizeof(opt), &opt))
		goto nla_put_failure;
	t.install = jiffies_to_clock_t(jiffies - p->tcf_tm.install);
	t.lastuse = jiffies_to_clock_t(jiffies - p->tcf_tm.lastuse);
	t.expires = jiffies_to_clock_t(p->tcf_tm.expires);
	if (nla_put(skb, TCA_CSUM_TM, sizeof(t), &t))
		goto nla_put_failure;

	return skb->len;

nla_put_failure:
	nlmsg_trim(skb, b);
	return -1;
}

static struct tc_action_ops act_csum_ops = {
	.kind		= "csum",
	.hinfo		= &csum_hash_info,
	.type		= TCA_ACT_CSUM,
	.capab		= TCA_CAP_NONE,
	.owner		= THIS_MODULE,
	.act		= tcf_csum,
	.dump		= tcf_csum_dump,
	.cleanup	= tcf_csum_cleanup,
	.init		= tcf_csum_init,
};

MODULE_DESCRIPTION("Checksum updating actions");
MODULE_LICENSE("GPL");

static int __init csum_init_module(void)
{
	return tcf_register_action(&act_csum_ops);
}

static void __exit csum_cleanup_module(void)
{
	tcf_unregister_action(&act_csum_ops);
}

module_init(csum_init_module);
module_exit(csum_cleanup_module);
Commit	Line	Data
eb4d4065 GB	1	/*
	2	* Checksum updating actions
	3	*
	4	* Copyright (c) 2010 Gregoire Baron <baronchon@n7mm.org>
	5	*
	6	* This program is free software; you can redistribute it and/or modify it
	7	* under the terms of the GNU General Public License as published by the Free
	8	* Software Foundation; either version 2 of the License, or (at your option)
	9	* any later version.
	10	*
	11	*/
	12
	13	#include <linux/types.h>
	14	#include <linux/init.h>
	15	#include <linux/kernel.h>
	16	#include <linux/module.h>
	17	#include <linux/spinlock.h>
	18
	19	#include <linux/netlink.h>
	20	#include <net/netlink.h>
	21	#include <linux/rtnetlink.h>
	22
	23	#include <linux/skbuff.h>
	24
	25	#include <net/ip.h>
	26	#include <net/ipv6.h>
	27	#include <net/icmp.h>
	28	#include <linux/icmpv6.h>
	29	#include <linux/igmp.h>
	30	#include <net/tcp.h>
	31	#include <net/udp.h>
2436243a	32	#include <net/ip6_checksum.h>
eb4d4065 GB	33
	34	#include <net/act_api.h>
	35
	36	#include <linux/tc_act/tc_csum.h>
	37	#include <net/tc_act/tc_csum.h>
	38
	39	#define CSUM_TAB_MASK 15
	40	static struct tcf_common *tcf_csum_ht[CSUM_TAB_MASK + 1];
	41	static u32 csum_idx_gen;
	42	static DEFINE_RWLOCK(csum_lock);
	43
	44	static struct tcf_hashinfo csum_hash_info = {
0eec32ff CG	45	.htab = tcf_csum_ht,
	46	.hmask = CSUM_TAB_MASK,
	47	.lock = &csum_lock,
eb4d4065 GB	48	};
	49
	50	static const struct nla_policy csum_policy[TCA_CSUM_MAX + 1] = {
	51	[TCA_CSUM_PARMS] = { .len = sizeof(struct tc_csum), },
	52	};
	53
c1b52739	54	static int tcf_csum_init(struct net n, struct nlattr nla, struct nlattr *est,
eb4d4065 GB	55	struct tc_action *a, int ovr, int bind)
	56	{
	57	struct nlattr *tb[TCA_CSUM_MAX + 1];
	58	struct tc_csum *parm;
	59	struct tcf_common *pc;
	60	struct tcf_csum *p;
	61	int ret = 0, err;
	62
	63	if (nla == NULL)
	64	return -EINVAL;
	65
cc7ec456	66	err = nla_parse_nested(tb, TCA_CSUM_MAX, nla, csum_policy);
eb4d4065 GB	67	if (err < 0)
	68	return err;
	69
	70	if (tb[TCA_CSUM_PARMS] == NULL)
	71	return -EINVAL;
	72	parm = nla_data(tb[TCA_CSUM_PARMS]);
	73
	74	pc = tcf_hash_check(parm->index, a, bind, &csum_hash_info);
	75	if (!pc) {
0eec32ff CG	76	pc = tcf_hash_create(parm->index, est, a, sizeof(*p), bind,
0eec32ff CG	77	&csum_idx_gen, &csum_hash_info);
eb4d4065 GB	78	if (IS_ERR(pc))
eb4d4065 GB	79	return PTR_ERR(pc);
eb4d4065 GB	80	ret = ACT_P_CREATED;
eb4d4065 GB	81	} else {
1a29321e JHS	82	if (bind)/* dont override defaults */
	83	return 0;
	84	tcf_hash_release(pc, bind, &csum_hash_info);
	85	if (!ovr)
eb4d4065	86	return -EEXIST;
eb4d4065 GB	87	}
eb4d4065 GB	88
1a29321e	89	p = to_tcf_csum(pc);
eb4d4065 GB	90	spin_lock_bh(&p->tcf_lock);
	91	p->tcf_action = parm->action;
	92	p->update_flags = parm->update_flags;
	93	spin_unlock_bh(&p->tcf_lock);
	94
	95	if (ret == ACT_P_CREATED)
	96	tcf_hash_insert(pc, &csum_hash_info);
	97
	98	return ret;
	99	}
	100
	101	static int tcf_csum_cleanup(struct tc_action *a, int bind)
	102	{
	103	struct tcf_csum *p = a->priv;
	104	return tcf_hash_release(&p->common, bind, &csum_hash_info);
	105	}
	106
	107	/**
	108	* tcf_csum_skb_nextlayer - Get next layer pointer
	109	* @skb: sk_buff to use
	110	* @ihl: previous summed headers length
	111	* @ipl: complete packet length
	112	* @jhl: next header length
	113	*
	114	* Check the expected next layer availability in the specified sk_buff.
	115	* Return the next layer pointer if pass, NULL otherwise.
	116	*/
	117	static void tcf_csum_skb_nextlayer(struct sk_buff skb,
	118	unsigned int ihl, unsigned int ipl,
	119	unsigned int jhl)
	120	{
	121	int ntkoff = skb_network_offset(skb);
	122	int hl = ihl + jhl;
	123
	124	if (!pskb_may_pull(skb, ipl + ntkoff) \|\| (ipl < hl) \|\|
	125	(skb_cloned(skb) &&
	126	!skb_clone_writable(skb, hl + ntkoff) &&
	127	pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
	128	return NULL;
	129	else
	130	return (void *)(skb_network_header(skb) + ihl);
	131	}
	132
	133	static int tcf_csum_ipv4_icmp(struct sk_buff *skb,
	134	unsigned int ihl, unsigned int ipl)
	135	{
	136	struct icmphdr *icmph;
	137
	138	icmph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*icmph));
	139	if (icmph == NULL)
	140	return 0;
	141
	142	icmph->checksum = 0;
	143	skb->csum = csum_partial(icmph, ipl - ihl, 0);
	144	icmph->checksum = csum_fold(skb->csum);
	145
	146	skb->ip_summed = CHECKSUM_NONE;
	147
	148	return 1;
	149	}
	150
	151	static int tcf_csum_ipv4_igmp(struct sk_buff *skb,
	152	unsigned int ihl, unsigned int ipl)
	153	{
154	struct igmphdr *igmph;
155
156	igmph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*igmph));
157	if (igmph == NULL)
158	return 0;
159
160	igmph->csum = 0;
161	skb->csum = csum_partial(igmph, ipl - ihl, 0);
162	igmph->csum = csum_fold(skb->csum);
163
164	skb->ip_summed = CHECKSUM_NONE;
165
166	return 1;
167	}
168
d14a489a	169	static int tcf_csum_ipv6_icmp(struct sk_buff *skb,
eb4d4065 GB	170	unsigned int ihl, unsigned int ipl)
	171	{
	172	struct icmp6hdr *icmp6h;
d14a489a	173	const struct ipv6hdr *ip6h;
eb4d4065 GB	174
	175	icmp6h = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*icmp6h));
	176	if (icmp6h == NULL)
	177	return 0;
	178
d14a489a	179	ip6h = ipv6_hdr(skb);
eb4d4065 GB	180	icmp6h->icmp6_cksum = 0;
	181	skb->csum = csum_partial(icmp6h, ipl - ihl, 0);
	182	icmp6h->icmp6_cksum = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
	183	ipl - ihl, IPPROTO_ICMPV6,
	184	skb->csum);
	185
	186	skb->ip_summed = CHECKSUM_NONE;
	187
	188	return 1;
	189	}
	190
d14a489a	191	static int tcf_csum_ipv4_tcp(struct sk_buff *skb,
eb4d4065 GB	192	unsigned int ihl, unsigned int ipl)
	193	{
	194	struct tcphdr *tcph;
d14a489a	195	const struct iphdr *iph;
eb4d4065 GB	196
	197	tcph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*tcph));
	198	if (tcph == NULL)
	199	return 0;
	200
d14a489a	201	iph = ip_hdr(skb);
eb4d4065 GB	202	tcph->check = 0;
	203	skb->csum = csum_partial(tcph, ipl - ihl, 0);
	204	tcph->check = tcp_v4_check(ipl - ihl,
	205	iph->saddr, iph->daddr, skb->csum);
	206
	207	skb->ip_summed = CHECKSUM_NONE;
	208
	209	return 1;
	210	}
	211
d14a489a	212	static int tcf_csum_ipv6_tcp(struct sk_buff *skb,
eb4d4065 GB	213	unsigned int ihl, unsigned int ipl)
	214	{
	215	struct tcphdr *tcph;
d14a489a	216	const struct ipv6hdr *ip6h;
eb4d4065 GB	217
	218	tcph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*tcph));
	219	if (tcph == NULL)
	220	return 0;
	221
d14a489a	222	ip6h = ipv6_hdr(skb);
eb4d4065 GB	223	tcph->check = 0;
	224	skb->csum = csum_partial(tcph, ipl - ihl, 0);
	225	tcph->check = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
	226	ipl - ihl, IPPROTO_TCP,
	227	skb->csum);
	228
	229	skb->ip_summed = CHECKSUM_NONE;
	230
	231	return 1;
	232	}
	233
d14a489a	234	static int tcf_csum_ipv4_udp(struct sk_buff *skb,
eb4d4065 GB	235	unsigned int ihl, unsigned int ipl, int udplite)
	236	{
	237	struct udphdr *udph;
d14a489a	238	const struct iphdr *iph;
eb4d4065 GB	239	u16 ul;
eb4d4065 GB	240
0eec32ff CG	241	/*
	242	* Support both UDP and UDPLITE checksum algorithms, Don't use
	243	* udph->len to get the real length without any protocol check,
eb4d4065 GB	244	* UDPLITE uses udph->len for another thing,
	245	* Use iph->tot_len, or just ipl.
	246	*/
	247
	248	udph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*udph));
	249	if (udph == NULL)
	250	return 0;
	251
d14a489a	252	iph = ip_hdr(skb);
eb4d4065 GB	253	ul = ntohs(udph->len);
	254
	255	if (udplite \|\| udph->check) {
	256
	257	udph->check = 0;
	258
	259	if (udplite) {
	260	if (ul == 0)
	261	skb->csum = csum_partial(udph, ipl - ihl, 0);
eb4d4065 GB	262	else if ((ul >= sizeof(*udph)) && (ul <= ipl - ihl))
eb4d4065 GB	263	skb->csum = csum_partial(udph, ul, 0);
eb4d4065 GB	264	else
	265	goto ignore_obscure_skb;
	266	} else {
	267	if (ul != ipl - ihl)
	268	goto ignore_obscure_skb;
	269
	270	skb->csum = csum_partial(udph, ul, 0);
	271	}
	272
	273	udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
	274	ul, iph->protocol,
	275	skb->csum);
	276
	277	if (!udph->check)
	278	udph->check = CSUM_MANGLED_0;
	279	}
	280
	281	skb->ip_summed = CHECKSUM_NONE;
	282
	283	ignore_obscure_skb:
	284	return 1;
	285	}
	286
d14a489a	287	static int tcf_csum_ipv6_udp(struct sk_buff *skb,
eb4d4065 GB	288	unsigned int ihl, unsigned int ipl, int udplite)
	289	{
	290	struct udphdr *udph;
d14a489a	291	const struct ipv6hdr *ip6h;
eb4d4065 GB	292	u16 ul;
eb4d4065 GB	293
0eec32ff CG	294	/*
	295	* Support both UDP and UDPLITE checksum algorithms, Don't use
	296	* udph->len to get the real length without any protocol check,
eb4d4065 GB	297	* UDPLITE uses udph->len for another thing,
	298	* Use ip6h->payload_len + sizeof(*ip6h) ... , or just ipl.
	299	*/
	300
	301	udph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*udph));
	302	if (udph == NULL)
	303	return 0;
	304
d14a489a	305	ip6h = ipv6_hdr(skb);
eb4d4065 GB	306	ul = ntohs(udph->len);
	307
	308	udph->check = 0;
	309
	310	if (udplite) {
	311	if (ul == 0)
	312	skb->csum = csum_partial(udph, ipl - ihl, 0);
	313
	314	else if ((ul >= sizeof(*udph)) && (ul <= ipl - ihl))
	315	skb->csum = csum_partial(udph, ul, 0);
	316
	317	else
	318	goto ignore_obscure_skb;
	319	} else {
	320	if (ul != ipl - ihl)
	321	goto ignore_obscure_skb;
	322
	323	skb->csum = csum_partial(udph, ul, 0);
	324	}
	325
	326	udph->check = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr, ul,
	327	udplite ? IPPROTO_UDPLITE : IPPROTO_UDP,
	328	skb->csum);
	329
	330	if (!udph->check)
	331	udph->check = CSUM_MANGLED_0;
	332
	333	skb->ip_summed = CHECKSUM_NONE;
	334
	335	ignore_obscure_skb:
	336	return 1;
	337	}
	338
	339	static int tcf_csum_ipv4(struct sk_buff *skb, u32 update_flags)
	340	{
d14a489a	341	const struct iphdr *iph;
eb4d4065 GB	342	int ntkoff;
	343
	344	ntkoff = skb_network_offset(skb);
	345
	346	if (!pskb_may_pull(skb, sizeof(*iph) + ntkoff))
	347	goto fail;
	348
	349	iph = ip_hdr(skb);
	350
	351	switch (iph->frag_off & htons(IP_OFFSET) ? 0 : iph->protocol) {
	352	case IPPROTO_ICMP:
	353	if (update_flags & TCA_CSUM_UPDATE_FLAG_ICMP)
0eec32ff CG	354	if (!tcf_csum_ipv4_icmp(skb, iph->ihl * 4,
0eec32ff CG	355	ntohs(iph->tot_len)))
eb4d4065 GB	356	goto fail;
	357	break;
	358	case IPPROTO_IGMP:
	359	if (update_flags & TCA_CSUM_UPDATE_FLAG_IGMP)
0eec32ff CG	360	if (!tcf_csum_ipv4_igmp(skb, iph->ihl * 4,
0eec32ff CG	361	ntohs(iph->tot_len)))
eb4d4065 GB	362	goto fail;
	363	break;
	364	case IPPROTO_TCP:
	365	if (update_flags & TCA_CSUM_UPDATE_FLAG_TCP)
d14a489a	366	if (!tcf_csum_ipv4_tcp(skb, iph->ihl * 4,
0eec32ff	367	ntohs(iph->tot_len)))
eb4d4065 GB	368	goto fail;
	369	break;
	370	case IPPROTO_UDP:
	371	if (update_flags & TCA_CSUM_UPDATE_FLAG_UDP)
d14a489a	372	if (!tcf_csum_ipv4_udp(skb, iph->ihl * 4,
0eec32ff	373	ntohs(iph->tot_len), 0))
eb4d4065 GB	374	goto fail;
	375	break;
	376	case IPPROTO_UDPLITE:
	377	if (update_flags & TCA_CSUM_UPDATE_FLAG_UDPLITE)
d14a489a	378	if (!tcf_csum_ipv4_udp(skb, iph->ihl * 4,
0eec32ff	379	ntohs(iph->tot_len), 1))
eb4d4065 GB	380	goto fail;
	381	break;
	382	}
	383
	384	if (update_flags & TCA_CSUM_UPDATE_FLAG_IPV4HDR) {
	385	if (skb_cloned(skb) &&
	386	!skb_clone_writable(skb, sizeof(*iph) + ntkoff) &&
	387	pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
	388	goto fail;
	389
d14a489a	390	ip_send_check(ip_hdr(skb));
eb4d4065 GB	391	}
	392
	393	return 1;
	394
	395	fail:
	396	return 0;
	397	}
	398
	399	static int tcf_csum_ipv6_hopopts(struct ipv6_opt_hdr *ip6xh,
0eec32ff	400	unsigned int ixhl, unsigned int *pl)
eb4d4065 GB	401	{
	402	int off, len, optlen;
	403	unsigned char xh = (void )ip6xh;
	404
	405	off = sizeof(*ip6xh);
	406	len = ixhl - off;
	407
	408	while (len > 1) {
0eec32ff	409	switch (xh[off]) {
1de5a71c	410	case IPV6_TLV_PAD1:
eb4d4065 GB	411	optlen = 1;
	412	break;
	413	case IPV6_TLV_JUMBO:
	414	optlen = xh[off + 1] + 2;
	415	if (optlen != 6 \|\| len < 6 \|\| (off & 3) != 2)
	416	/* wrong jumbo option length/alignment */
	417	return 0;
	418	pl = ntohl((__be32 *)(xh + off + 2));
	419	goto done;
	420	default:
	421	optlen = xh[off + 1] + 2;
	422	if (optlen > len)
	423	/* ignore obscure options */
	424	goto done;
	425	break;
	426	}
	427	off += optlen;
	428	len -= optlen;
	429	}
	430
	431	done:
	432	return 1;
	433	}
	434
	435	static int tcf_csum_ipv6(struct sk_buff *skb, u32 update_flags)
	436	{
	437	struct ipv6hdr *ip6h;
	438	struct ipv6_opt_hdr *ip6xh;
	439	unsigned int hl, ixhl;
	440	unsigned int pl;
	441	int ntkoff;
	442	u8 nexthdr;
	443
	444	ntkoff = skb_network_offset(skb);
	445
	446	hl = sizeof(*ip6h);
	447
	448	if (!pskb_may_pull(skb, hl + ntkoff))
	449	goto fail;
	450
	451	ip6h = ipv6_hdr(skb);
	452
	453	pl = ntohs(ip6h->payload_len);
	454	nexthdr = ip6h->nexthdr;
	455
	456	do {
	457	switch (nexthdr) {
	458	case NEXTHDR_FRAGMENT:
	459	goto ignore_skb;
	460	case NEXTHDR_ROUTING:
	461	case NEXTHDR_HOP:
	462	case NEXTHDR_DEST:
	463	if (!pskb_may_pull(skb, hl + sizeof(*ip6xh) + ntkoff))
	464	goto fail;
	465	ip6xh = (void *)(skb_network_header(skb) + hl);
	466	ixhl = ipv6_optlen(ip6xh);
	467	if (!pskb_may_pull(skb, hl + ixhl + ntkoff))
	468	goto fail;
d14a489a	469	ip6xh = (void *)(skb_network_header(skb) + hl);
eb4d4065 GB	470	if ((nexthdr == NEXTHDR_HOP) &&
	471	!(tcf_csum_ipv6_hopopts(ip6xh, ixhl, &pl)))
	472	goto fail;
	473	nexthdr = ip6xh->nexthdr;
	474	hl += ixhl;
	475	break;
	476	case IPPROTO_ICMPV6:
	477	if (update_flags & TCA_CSUM_UPDATE_FLAG_ICMP)
d14a489a	478	if (!tcf_csum_ipv6_icmp(skb,
eb4d4065 GB	479	hl, pl + sizeof(*ip6h)))
	480	goto fail;
	481	goto done;
	482	case IPPROTO_TCP:
	483	if (update_flags & TCA_CSUM_UPDATE_FLAG_TCP)
d14a489a	484	if (!tcf_csum_ipv6_tcp(skb,
eb4d4065 GB	485	hl, pl + sizeof(*ip6h)))
	486	goto fail;
	487	goto done;
	488	case IPPROTO_UDP:
	489	if (update_flags & TCA_CSUM_UPDATE_FLAG_UDP)
d14a489a	490	if (!tcf_csum_ipv6_udp(skb, hl,
0eec32ff	491	pl + sizeof(*ip6h), 0))
eb4d4065 GB	492	goto fail;
	493	goto done;
	494	case IPPROTO_UDPLITE:
	495	if (update_flags & TCA_CSUM_UPDATE_FLAG_UDPLITE)
d14a489a	496	if (!tcf_csum_ipv6_udp(skb, hl,
0eec32ff	497	pl + sizeof(*ip6h), 1))
eb4d4065 GB	498	goto fail;
	499	goto done;
	500	default:
	501	goto ignore_skb;
	502	}
	503	} while (pskb_may_pull(skb, hl + 1 + ntkoff));
	504
	505	done:
	506	ignore_skb:
	507	return 1;
	508
	509	fail:
	510	return 0;
	511	}
	512
	513	static int tcf_csum(struct sk_buff *skb,
dc7f9f6e	514	const struct tc_action a, struct tcf_result res)
eb4d4065 GB	515	{
	516	struct tcf_csum *p = a->priv;
	517	int action;
	518	u32 update_flags;
	519
	520	spin_lock(&p->tcf_lock);
	521	p->tcf_tm.lastuse = jiffies;
bfe0d029	522	bstats_update(&p->tcf_bstats, skb);
eb4d4065 GB	523	action = p->tcf_action;
	524	update_flags = p->update_flags;
	525	spin_unlock(&p->tcf_lock);
	526
	527	if (unlikely(action == TC_ACT_SHOT))
	528	goto drop;
	529
	530	switch (skb->protocol) {
	531	case cpu_to_be16(ETH_P_IP):
	532	if (!tcf_csum_ipv4(skb, update_flags))
	533	goto drop;
	534	break;
	535	case cpu_to_be16(ETH_P_IPV6):
	536	if (!tcf_csum_ipv6(skb, update_flags))
	537	goto drop;
	538	break;
	539	}
	540
	541	return action;
	542
	543	drop:
	544	spin_lock(&p->tcf_lock);
	545	p->tcf_qstats.drops++;
	546	spin_unlock(&p->tcf_lock);
	547	return TC_ACT_SHOT;
	548	}
	549
	550	static int tcf_csum_dump(struct sk_buff *skb,
	551	struct tc_action *a, int bind, int ref)
	552	{
	553	unsigned char *b = skb_tail_pointer(skb);
	554	struct tcf_csum *p = a->priv;
	555	struct tc_csum opt = {
	556	.update_flags = p->update_flags,
eb4d4065 GB	557	.index = p->tcf_index,
	558	.action = p->tcf_action,
	559	.refcnt = p->tcf_refcnt - ref,
	560	.bindcnt = p->tcf_bindcnt - bind,
	561	};
	562	struct tcf_t t;
	563
1b34ec43 DM	564	if (nla_put(skb, TCA_CSUM_PARMS, sizeof(opt), &opt))
1b34ec43 DM	565	goto nla_put_failure;
eb4d4065 GB	566	t.install = jiffies_to_clock_t(jiffies - p->tcf_tm.install);
	567	t.lastuse = jiffies_to_clock_t(jiffies - p->tcf_tm.lastuse);
	568	t.expires = jiffies_to_clock_t(p->tcf_tm.expires);
1b34ec43 DM	569	if (nla_put(skb, TCA_CSUM_TM, sizeof(t), &t))
1b34ec43 DM	570	goto nla_put_failure;
eb4d4065 GB	571
	572	return skb->len;
	573
	574	nla_put_failure:
	575	nlmsg_trim(skb, b);
	576	return -1;
	577	}
	578
	579	static struct tc_action_ops act_csum_ops = {
0eec32ff CG	580	.kind = "csum",
	581	.hinfo = &csum_hash_info,
	582	.type = TCA_ACT_CSUM,
	583	.capab = TCA_CAP_NONE,
	584	.owner = THIS_MODULE,
	585	.act = tcf_csum,
	586	.dump = tcf_csum_dump,
	587	.cleanup = tcf_csum_cleanup,
0eec32ff	588	.init = tcf_csum_init,
eb4d4065 GB	589	};
	590
	591	MODULE_DESCRIPTION("Checksum updating actions");
	592	MODULE_LICENSE("GPL");
	593
	594	static int __init csum_init_module(void)
	595	{
	596	return tcf_register_action(&act_csum_ops);
	597	}
	598
	599	static void __exit csum_cleanup_module(void)
	600	{
	601	tcf_unregister_action(&act_csum_ops);
	602	}
	603
	604	module_init(csum_init_module);
	605	module_exit(csum_cleanup_module);