xref: /kernel/linux/linux-5.10/net/ipv4/esp4.c (revision 8c2ecf20)
1// SPDX-License-Identifier: GPL-2.0-only
2#define pr_fmt(fmt) "IPsec: " fmt
3
4#include <crypto/aead.h>
5#include <crypto/authenc.h>
6#include <linux/err.h>
7#include <linux/module.h>
8#include <net/ip.h>
9#include <net/xfrm.h>
10#include <net/esp.h>
11#include <linux/scatterlist.h>
12#include <linux/kernel.h>
13#include <linux/pfkeyv2.h>
14#include <linux/rtnetlink.h>
15#include <linux/slab.h>
16#include <linux/spinlock.h>
17#include <linux/in6.h>
18#include <net/icmp.h>
19#include <net/protocol.h>
20#include <net/udp.h>
21#include <net/tcp.h>
22#include <net/espintcp.h>
23
24#include <linux/highmem.h>
25
26struct esp_skb_cb {
27	struct xfrm_skb_cb xfrm;
28	void *tmp;
29};
30
31struct esp_output_extra {
32	__be32 seqhi;
33	u32 esphoff;
34};
35
36#define ESP_SKB_CB(__skb) ((struct esp_skb_cb *)&((__skb)->cb[0]))
37
38/*
39 * Allocate an AEAD request structure with extra space for SG and IV.
40 *
41 * For alignment considerations the IV is placed at the front, followed
42 * by the request and finally the SG list.
43 *
44 * TODO: Use spare space in skb for this where possible.
45 */
46static void *esp_alloc_tmp(struct crypto_aead *aead, int nfrags, int extralen)
47{
48	unsigned int len;
49
50	len = extralen;
51
52	len += crypto_aead_ivsize(aead);
53
54	if (len) {
55		len += crypto_aead_alignmask(aead) &
56		       ~(crypto_tfm_ctx_alignment() - 1);
57		len = ALIGN(len, crypto_tfm_ctx_alignment());
58	}
59
60	len += sizeof(struct aead_request) + crypto_aead_reqsize(aead);
61	len = ALIGN(len, __alignof__(struct scatterlist));
62
63	len += sizeof(struct scatterlist) * nfrags;
64
65	return kmalloc(len, GFP_ATOMIC);
66}
67
68static inline void *esp_tmp_extra(void *tmp)
69{
70	return PTR_ALIGN(tmp, __alignof__(struct esp_output_extra));
71}
72
73static inline u8 *esp_tmp_iv(struct crypto_aead *aead, void *tmp, int extralen)
74{
75	return crypto_aead_ivsize(aead) ?
76	       PTR_ALIGN((u8 *)tmp + extralen,
77			 crypto_aead_alignmask(aead) + 1) : tmp + extralen;
78}
79
80static inline struct aead_request *esp_tmp_req(struct crypto_aead *aead, u8 *iv)
81{
82	struct aead_request *req;
83
84	req = (void *)PTR_ALIGN(iv + crypto_aead_ivsize(aead),
85				crypto_tfm_ctx_alignment());
86	aead_request_set_tfm(req, aead);
87	return req;
88}
89
90static inline struct scatterlist *esp_req_sg(struct crypto_aead *aead,
91					     struct aead_request *req)
92{
93	return (void *)ALIGN((unsigned long)(req + 1) +
94			     crypto_aead_reqsize(aead),
95			     __alignof__(struct scatterlist));
96}
97
98static void esp_ssg_unref(struct xfrm_state *x, void *tmp)
99{
100	struct esp_output_extra *extra = esp_tmp_extra(tmp);
101	struct crypto_aead *aead = x->data;
102	int extralen = 0;
103	u8 *iv;
104	struct aead_request *req;
105	struct scatterlist *sg;
106
107	if (x->props.flags & XFRM_STATE_ESN)
108		extralen += sizeof(*extra);
109
110	extra = esp_tmp_extra(tmp);
111	iv = esp_tmp_iv(aead, tmp, extralen);
112	req = esp_tmp_req(aead, iv);
113
114	/* Unref skb_frag_pages in the src scatterlist if necessary.
115	 * Skip the first sg which comes from skb->data.
116	 */
117	if (req->src != req->dst)
118		for (sg = sg_next(req->src); sg; sg = sg_next(sg))
119			put_page(sg_page(sg));
120}
121
122#ifdef CONFIG_INET_ESPINTCP
123struct esp_tcp_sk {
124	struct sock *sk;
125	struct rcu_head rcu;
126};
127
128static void esp_free_tcp_sk(struct rcu_head *head)
129{
130	struct esp_tcp_sk *esk = container_of(head, struct esp_tcp_sk, rcu);
131
132	sock_put(esk->sk);
133	kfree(esk);
134}
135
136static struct sock *esp_find_tcp_sk(struct xfrm_state *x)
137{
138	struct xfrm_encap_tmpl *encap = x->encap;
139	struct esp_tcp_sk *esk;
140	__be16 sport, dport;
141	struct sock *nsk;
142	struct sock *sk;
143
144	sk = rcu_dereference(x->encap_sk);
145	if (sk && sk->sk_state == TCP_ESTABLISHED)
146		return sk;
147
148	spin_lock_bh(&x->lock);
149	sport = encap->encap_sport;
150	dport = encap->encap_dport;
151	nsk = rcu_dereference_protected(x->encap_sk,
152					lockdep_is_held(&x->lock));
153	if (sk && sk == nsk) {
154		esk = kmalloc(sizeof(*esk), GFP_ATOMIC);
155		if (!esk) {
156			spin_unlock_bh(&x->lock);
157			return ERR_PTR(-ENOMEM);
158		}
159		RCU_INIT_POINTER(x->encap_sk, NULL);
160		esk->sk = sk;
161		call_rcu(&esk->rcu, esp_free_tcp_sk);
162	}
163	spin_unlock_bh(&x->lock);
164
165	sk = inet_lookup_established(xs_net(x), &tcp_hashinfo, x->id.daddr.a4,
166				     dport, x->props.saddr.a4, sport, 0);
167	if (!sk)
168		return ERR_PTR(-ENOENT);
169
170	if (!tcp_is_ulp_esp(sk)) {
171		sock_put(sk);
172		return ERR_PTR(-EINVAL);
173	}
174
175	spin_lock_bh(&x->lock);
176	nsk = rcu_dereference_protected(x->encap_sk,
177					lockdep_is_held(&x->lock));
178	if (encap->encap_sport != sport ||
179	    encap->encap_dport != dport) {
180		sock_put(sk);
181		sk = nsk ?: ERR_PTR(-EREMCHG);
182	} else if (sk == nsk) {
183		sock_put(sk);
184	} else {
185		rcu_assign_pointer(x->encap_sk, sk);
186	}
187	spin_unlock_bh(&x->lock);
188
189	return sk;
190}
191
192static int esp_output_tcp_finish(struct xfrm_state *x, struct sk_buff *skb)
193{
194	struct sock *sk;
195	int err;
196
197	rcu_read_lock();
198
199	sk = esp_find_tcp_sk(x);
200	err = PTR_ERR_OR_ZERO(sk);
201	if (err)
202		goto out;
203
204	bh_lock_sock(sk);
205	if (sock_owned_by_user(sk))
206		err = espintcp_queue_out(sk, skb);
207	else
208		err = espintcp_push_skb(sk, skb);
209	bh_unlock_sock(sk);
210
211out:
212	rcu_read_unlock();
213	return err;
214}
215
216static int esp_output_tcp_encap_cb(struct net *net, struct sock *sk,
217				   struct sk_buff *skb)
218{
219	struct dst_entry *dst = skb_dst(skb);
220	struct xfrm_state *x = dst->xfrm;
221
222	return esp_output_tcp_finish(x, skb);
223}
224
225static int esp_output_tail_tcp(struct xfrm_state *x, struct sk_buff *skb)
226{
227	int err;
228
229	local_bh_disable();
230	err = xfrm_trans_queue_net(xs_net(x), skb, esp_output_tcp_encap_cb);
231	local_bh_enable();
232
233	/* EINPROGRESS just happens to do the right thing.  It
234	 * actually means that the skb has been consumed and
235	 * isn't coming back.
236	 */
237	return err ?: -EINPROGRESS;
238}
239#else
240static int esp_output_tail_tcp(struct xfrm_state *x, struct sk_buff *skb)
241{
242	kfree_skb(skb);
243
244	return -EOPNOTSUPP;
245}
246#endif
247
248static void esp_output_done(struct crypto_async_request *base, int err)
249{
250	struct sk_buff *skb = base->data;
251	struct xfrm_offload *xo = xfrm_offload(skb);
252	void *tmp;
253	struct xfrm_state *x;
254
255	if (xo && (xo->flags & XFRM_DEV_RESUME)) {
256		struct sec_path *sp = skb_sec_path(skb);
257
258		x = sp->xvec[sp->len - 1];
259	} else {
260		x = skb_dst(skb)->xfrm;
261	}
262
263	tmp = ESP_SKB_CB(skb)->tmp;
264	esp_ssg_unref(x, tmp);
265	kfree(tmp);
266
267	if (xo && (xo->flags & XFRM_DEV_RESUME)) {
268		if (err) {
269			XFRM_INC_STATS(xs_net(x), LINUX_MIB_XFRMOUTSTATEPROTOERROR);
270			kfree_skb(skb);
271			return;
272		}
273
274		skb_push(skb, skb->data - skb_mac_header(skb));
275		secpath_reset(skb);
276		xfrm_dev_resume(skb);
277	} else {
278		if (!err &&
279		    x->encap && x->encap->encap_type == TCP_ENCAP_ESPINTCP)
280			esp_output_tail_tcp(x, skb);
281		else
282			xfrm_output_resume(skb->sk, skb, err);
283	}
284}
285
286/* Move ESP header back into place. */
287static void esp_restore_header(struct sk_buff *skb, unsigned int offset)
288{
289	struct ip_esp_hdr *esph = (void *)(skb->data + offset);
290	void *tmp = ESP_SKB_CB(skb)->tmp;
291	__be32 *seqhi = esp_tmp_extra(tmp);
292
293	esph->seq_no = esph->spi;
294	esph->spi = *seqhi;
295}
296
297static void esp_output_restore_header(struct sk_buff *skb)
298{
299	void *tmp = ESP_SKB_CB(skb)->tmp;
300	struct esp_output_extra *extra = esp_tmp_extra(tmp);
301
302	esp_restore_header(skb, skb_transport_offset(skb) + extra->esphoff -
303				sizeof(__be32));
304}
305
306static struct ip_esp_hdr *esp_output_set_extra(struct sk_buff *skb,
307					       struct xfrm_state *x,
308					       struct ip_esp_hdr *esph,
309					       struct esp_output_extra *extra)
310{
311	/* For ESN we move the header forward by 4 bytes to
312	 * accomodate the high bits.  We will move it back after
313	 * encryption.
314	 */
315	if ((x->props.flags & XFRM_STATE_ESN)) {
316		__u32 seqhi;
317		struct xfrm_offload *xo = xfrm_offload(skb);
318
319		if (xo)
320			seqhi = xo->seq.hi;
321		else
322			seqhi = XFRM_SKB_CB(skb)->seq.output.hi;
323
324		extra->esphoff = (unsigned char *)esph -
325				 skb_transport_header(skb);
326		esph = (struct ip_esp_hdr *)((unsigned char *)esph - 4);
327		extra->seqhi = esph->spi;
328		esph->seq_no = htonl(seqhi);
329	}
330
331	esph->spi = x->id.spi;
332
333	return esph;
334}
335
336static void esp_output_done_esn(struct crypto_async_request *base, int err)
337{
338	struct sk_buff *skb = base->data;
339
340	esp_output_restore_header(skb);
341	esp_output_done(base, err);
342}
343
344static struct ip_esp_hdr *esp_output_udp_encap(struct sk_buff *skb,
345					       int encap_type,
346					       struct esp_info *esp,
347					       __be16 sport,
348					       __be16 dport)
349{
350	struct udphdr *uh;
351	__be32 *udpdata32;
352	unsigned int len;
353
354	len = skb->len + esp->tailen - skb_transport_offset(skb);
355	if (len + sizeof(struct iphdr) > IP_MAX_MTU)
356		return ERR_PTR(-EMSGSIZE);
357
358	uh = (struct udphdr *)esp->esph;
359	uh->source = sport;
360	uh->dest = dport;
361	uh->len = htons(len);
362	uh->check = 0;
363
364	*skb_mac_header(skb) = IPPROTO_UDP;
365
366	if (encap_type == UDP_ENCAP_ESPINUDP_NON_IKE) {
367		udpdata32 = (__be32 *)(uh + 1);
368		udpdata32[0] = udpdata32[1] = 0;
369		return (struct ip_esp_hdr *)(udpdata32 + 2);
370	}
371
372	return (struct ip_esp_hdr *)(uh + 1);
373}
374
375#ifdef CONFIG_INET_ESPINTCP
376static struct ip_esp_hdr *esp_output_tcp_encap(struct xfrm_state *x,
377						    struct sk_buff *skb,
378						    struct esp_info *esp)
379{
380	__be16 *lenp = (void *)esp->esph;
381	struct ip_esp_hdr *esph;
382	unsigned int len;
383	struct sock *sk;
384
385	len = skb->len + esp->tailen - skb_transport_offset(skb);
386	if (len > IP_MAX_MTU)
387		return ERR_PTR(-EMSGSIZE);
388
389	rcu_read_lock();
390	sk = esp_find_tcp_sk(x);
391	rcu_read_unlock();
392
393	if (IS_ERR(sk))
394		return ERR_CAST(sk);
395
396	*lenp = htons(len);
397	esph = (struct ip_esp_hdr *)(lenp + 1);
398
399	return esph;
400}
401#else
402static struct ip_esp_hdr *esp_output_tcp_encap(struct xfrm_state *x,
403						    struct sk_buff *skb,
404						    struct esp_info *esp)
405{
406	return ERR_PTR(-EOPNOTSUPP);
407}
408#endif
409
410static int esp_output_encap(struct xfrm_state *x, struct sk_buff *skb,
411			    struct esp_info *esp)
412{
413	struct xfrm_encap_tmpl *encap = x->encap;
414	struct ip_esp_hdr *esph;
415	__be16 sport, dport;
416	int encap_type;
417
418	spin_lock_bh(&x->lock);
419	sport = encap->encap_sport;
420	dport = encap->encap_dport;
421	encap_type = encap->encap_type;
422	spin_unlock_bh(&x->lock);
423
424	switch (encap_type) {
425	default:
426	case UDP_ENCAP_ESPINUDP:
427	case UDP_ENCAP_ESPINUDP_NON_IKE:
428		esph = esp_output_udp_encap(skb, encap_type, esp, sport, dport);
429		break;
430	case TCP_ENCAP_ESPINTCP:
431		esph = esp_output_tcp_encap(x, skb, esp);
432		break;
433	}
434
435	if (IS_ERR(esph))
436		return PTR_ERR(esph);
437
438	esp->esph = esph;
439
440	return 0;
441}
442
443int esp_output_head(struct xfrm_state *x, struct sk_buff *skb, struct esp_info *esp)
444{
445	u8 *tail;
446	int nfrags;
447	int esph_offset;
448	struct page *page;
449	struct sk_buff *trailer;
450	int tailen = esp->tailen;
451
452	/* this is non-NULL only with TCP/UDP Encapsulation */
453	if (x->encap) {
454		int err = esp_output_encap(x, skb, esp);
455
456		if (err < 0)
457			return err;
458	}
459
460	if (ALIGN(tailen, L1_CACHE_BYTES) > PAGE_SIZE ||
461	    ALIGN(skb->data_len, L1_CACHE_BYTES) > PAGE_SIZE)
462		goto cow;
463
464	if (!skb_cloned(skb)) {
465		if (tailen <= skb_tailroom(skb)) {
466			nfrags = 1;
467			trailer = skb;
468			tail = skb_tail_pointer(trailer);
469
470			goto skip_cow;
471		} else if ((skb_shinfo(skb)->nr_frags < MAX_SKB_FRAGS)
472			   && !skb_has_frag_list(skb)) {
473			int allocsize;
474			struct sock *sk = skb->sk;
475			struct page_frag *pfrag = &x->xfrag;
476
477			esp->inplace = false;
478
479			allocsize = ALIGN(tailen, L1_CACHE_BYTES);
480
481			spin_lock_bh(&x->lock);
482
483			if (unlikely(!skb_page_frag_refill(allocsize, pfrag, GFP_ATOMIC))) {
484				spin_unlock_bh(&x->lock);
485				goto cow;
486			}
487
488			page = pfrag->page;
489			get_page(page);
490
491			tail = page_address(page) + pfrag->offset;
492
493			esp_output_fill_trailer(tail, esp->tfclen, esp->plen, esp->proto);
494
495			nfrags = skb_shinfo(skb)->nr_frags;
496
497			__skb_fill_page_desc(skb, nfrags, page, pfrag->offset,
498					     tailen);
499			skb_shinfo(skb)->nr_frags = ++nfrags;
500
501			pfrag->offset = pfrag->offset + allocsize;
502
503			spin_unlock_bh(&x->lock);
504
505			nfrags++;
506
507			skb->len += tailen;
508			skb->data_len += tailen;
509			skb->truesize += tailen;
510			if (sk && sk_fullsock(sk))
511				refcount_add(tailen, &sk->sk_wmem_alloc);
512
513			goto out;
514		}
515	}
516
517cow:
518	esph_offset = (unsigned char *)esp->esph - skb_transport_header(skb);
519
520	nfrags = skb_cow_data(skb, tailen, &trailer);
521	if (nfrags < 0)
522		goto out;
523	tail = skb_tail_pointer(trailer);
524	esp->esph = (struct ip_esp_hdr *)(skb_transport_header(skb) + esph_offset);
525
526skip_cow:
527	esp_output_fill_trailer(tail, esp->tfclen, esp->plen, esp->proto);
528	pskb_put(skb, trailer, tailen);
529
530out:
531	return nfrags;
532}
533EXPORT_SYMBOL_GPL(esp_output_head);
534
535int esp_output_tail(struct xfrm_state *x, struct sk_buff *skb, struct esp_info *esp)
536{
537	u8 *iv;
538	int alen;
539	void *tmp;
540	int ivlen;
541	int assoclen;
542	int extralen;
543	struct page *page;
544	struct ip_esp_hdr *esph;
545	struct crypto_aead *aead;
546	struct aead_request *req;
547	struct scatterlist *sg, *dsg;
548	struct esp_output_extra *extra;
549	int err = -ENOMEM;
550
551	assoclen = sizeof(struct ip_esp_hdr);
552	extralen = 0;
553
554	if (x->props.flags & XFRM_STATE_ESN) {
555		extralen += sizeof(*extra);
556		assoclen += sizeof(__be32);
557	}
558
559	aead = x->data;
560	alen = crypto_aead_authsize(aead);
561	ivlen = crypto_aead_ivsize(aead);
562
563	tmp = esp_alloc_tmp(aead, esp->nfrags + 2, extralen);
564	if (!tmp)
565		goto error;
566
567	extra = esp_tmp_extra(tmp);
568	iv = esp_tmp_iv(aead, tmp, extralen);
569	req = esp_tmp_req(aead, iv);
570	sg = esp_req_sg(aead, req);
571
572	if (esp->inplace)
573		dsg = sg;
574	else
575		dsg = &sg[esp->nfrags];
576
577	esph = esp_output_set_extra(skb, x, esp->esph, extra);
578	esp->esph = esph;
579
580	sg_init_table(sg, esp->nfrags);
581	err = skb_to_sgvec(skb, sg,
582		           (unsigned char *)esph - skb->data,
583		           assoclen + ivlen + esp->clen + alen);
584	if (unlikely(err < 0))
585		goto error_free;
586
587	if (!esp->inplace) {
588		int allocsize;
589		struct page_frag *pfrag = &x->xfrag;
590
591		allocsize = ALIGN(skb->data_len, L1_CACHE_BYTES);
592
593		spin_lock_bh(&x->lock);
594		if (unlikely(!skb_page_frag_refill(allocsize, pfrag, GFP_ATOMIC))) {
595			spin_unlock_bh(&x->lock);
596			goto error_free;
597		}
598
599		skb_shinfo(skb)->nr_frags = 1;
600
601		page = pfrag->page;
602		get_page(page);
603		/* replace page frags in skb with new page */
604		__skb_fill_page_desc(skb, 0, page, pfrag->offset, skb->data_len);
605		pfrag->offset = pfrag->offset + allocsize;
606		spin_unlock_bh(&x->lock);
607
608		sg_init_table(dsg, skb_shinfo(skb)->nr_frags + 1);
609		err = skb_to_sgvec(skb, dsg,
610			           (unsigned char *)esph - skb->data,
611			           assoclen + ivlen + esp->clen + alen);
612		if (unlikely(err < 0))
613			goto error_free;
614	}
615
616	if ((x->props.flags & XFRM_STATE_ESN))
617		aead_request_set_callback(req, 0, esp_output_done_esn, skb);
618	else
619		aead_request_set_callback(req, 0, esp_output_done, skb);
620
621	aead_request_set_crypt(req, sg, dsg, ivlen + esp->clen, iv);
622	aead_request_set_ad(req, assoclen);
623
624	memset(iv, 0, ivlen);
625	memcpy(iv + ivlen - min(ivlen, 8), (u8 *)&esp->seqno + 8 - min(ivlen, 8),
626	       min(ivlen, 8));
627
628	ESP_SKB_CB(skb)->tmp = tmp;
629	err = crypto_aead_encrypt(req);
630
631	switch (err) {
632	case -EINPROGRESS:
633		goto error;
634
635	case -ENOSPC:
636		err = NET_XMIT_DROP;
637		break;
638
639	case 0:
640		if ((x->props.flags & XFRM_STATE_ESN))
641			esp_output_restore_header(skb);
642	}
643
644	if (sg != dsg)
645		esp_ssg_unref(x, tmp);
646
647	if (!err && x->encap && x->encap->encap_type == TCP_ENCAP_ESPINTCP)
648		err = esp_output_tail_tcp(x, skb);
649
650error_free:
651	kfree(tmp);
652error:
653	return err;
654}
655EXPORT_SYMBOL_GPL(esp_output_tail);
656
657static int esp_output(struct xfrm_state *x, struct sk_buff *skb)
658{
659	int alen;
660	int blksize;
661	struct ip_esp_hdr *esph;
662	struct crypto_aead *aead;
663	struct esp_info esp;
664
665	esp.inplace = true;
666
667	esp.proto = *skb_mac_header(skb);
668	*skb_mac_header(skb) = IPPROTO_ESP;
669
670	/* skb is pure payload to encrypt */
671
672	aead = x->data;
673	alen = crypto_aead_authsize(aead);
674
675	esp.tfclen = 0;
676	if (x->tfcpad) {
677		struct xfrm_dst *dst = (struct xfrm_dst *)skb_dst(skb);
678		u32 padto;
679
680		padto = min(x->tfcpad, xfrm_state_mtu(x, dst->child_mtu_cached));
681		if (skb->len < padto)
682			esp.tfclen = padto - skb->len;
683	}
684	blksize = ALIGN(crypto_aead_blocksize(aead), 4);
685	esp.clen = ALIGN(skb->len + 2 + esp.tfclen, blksize);
686	esp.plen = esp.clen - skb->len - esp.tfclen;
687	esp.tailen = esp.tfclen + esp.plen + alen;
688
689	esp.esph = ip_esp_hdr(skb);
690
691	esp.nfrags = esp_output_head(x, skb, &esp);
692	if (esp.nfrags < 0)
693		return esp.nfrags;
694
695	esph = esp.esph;
696	esph->spi = x->id.spi;
697
698	esph->seq_no = htonl(XFRM_SKB_CB(skb)->seq.output.low);
699	esp.seqno = cpu_to_be64(XFRM_SKB_CB(skb)->seq.output.low +
700				 ((u64)XFRM_SKB_CB(skb)->seq.output.hi << 32));
701
702	skb_push(skb, -skb_network_offset(skb));
703
704	return esp_output_tail(x, skb, &esp);
705}
706
707static inline int esp_remove_trailer(struct sk_buff *skb)
708{
709	struct xfrm_state *x = xfrm_input_state(skb);
710	struct xfrm_offload *xo = xfrm_offload(skb);
711	struct crypto_aead *aead = x->data;
712	int alen, hlen, elen;
713	int padlen, trimlen;
714	__wsum csumdiff;
715	u8 nexthdr[2];
716	int ret;
717
718	alen = crypto_aead_authsize(aead);
719	hlen = sizeof(struct ip_esp_hdr) + crypto_aead_ivsize(aead);
720	elen = skb->len - hlen;
721
722	if (xo && (xo->flags & XFRM_ESP_NO_TRAILER)) {
723		ret = xo->proto;
724		goto out;
725	}
726
727	if (skb_copy_bits(skb, skb->len - alen - 2, nexthdr, 2))
728		BUG();
729
730	ret = -EINVAL;
731	padlen = nexthdr[0];
732	if (padlen + 2 + alen >= elen) {
733		net_dbg_ratelimited("ipsec esp packet is garbage padlen=%d, elen=%d\n",
734				    padlen + 2, elen - alen);
735		goto out;
736	}
737
738	trimlen = alen + padlen + 2;
739	if (skb->ip_summed == CHECKSUM_COMPLETE) {
740		csumdiff = skb_checksum(skb, skb->len - trimlen, trimlen, 0);
741		skb->csum = csum_block_sub(skb->csum, csumdiff,
742					   skb->len - trimlen);
743	}
744	ret = pskb_trim(skb, skb->len - trimlen);
745	if (unlikely(ret))
746		return ret;
747
748	ret = nexthdr[1];
749
750out:
751	return ret;
752}
753
754int esp_input_done2(struct sk_buff *skb, int err)
755{
756	const struct iphdr *iph;
757	struct xfrm_state *x = xfrm_input_state(skb);
758	struct xfrm_offload *xo = xfrm_offload(skb);
759	struct crypto_aead *aead = x->data;
760	int hlen = sizeof(struct ip_esp_hdr) + crypto_aead_ivsize(aead);
761	int ihl;
762
763	if (!xo || (xo && !(xo->flags & CRYPTO_DONE)))
764		kfree(ESP_SKB_CB(skb)->tmp);
765
766	if (unlikely(err))
767		goto out;
768
769	err = esp_remove_trailer(skb);
770	if (unlikely(err < 0))
771		goto out;
772
773	iph = ip_hdr(skb);
774	ihl = iph->ihl * 4;
775
776	if (x->encap) {
777		struct xfrm_encap_tmpl *encap = x->encap;
778		struct tcphdr *th = (void *)(skb_network_header(skb) + ihl);
779		struct udphdr *uh = (void *)(skb_network_header(skb) + ihl);
780		__be16 source;
781
782		switch (x->encap->encap_type) {
783		case TCP_ENCAP_ESPINTCP:
784			source = th->source;
785			break;
786		case UDP_ENCAP_ESPINUDP:
787		case UDP_ENCAP_ESPINUDP_NON_IKE:
788			source = uh->source;
789			break;
790		default:
791			WARN_ON_ONCE(1);
792			err = -EINVAL;
793			goto out;
794		}
795
796		/*
797		 * 1) if the NAT-T peer's IP or port changed then
798		 *    advertize the change to the keying daemon.
799		 *    This is an inbound SA, so just compare
800		 *    SRC ports.
801		 */
802		if (iph->saddr != x->props.saddr.a4 ||
803		    source != encap->encap_sport) {
804			xfrm_address_t ipaddr;
805
806			ipaddr.a4 = iph->saddr;
807			km_new_mapping(x, &ipaddr, source);
808
809			/* XXX: perhaps add an extra
810			 * policy check here, to see
811			 * if we should allow or
812			 * reject a packet from a
813			 * different source
814			 * address/port.
815			 */
816		}
817
818		/*
819		 * 2) ignore UDP/TCP checksums in case
820		 *    of NAT-T in Transport Mode, or
821		 *    perform other post-processing fixes
822		 *    as per draft-ietf-ipsec-udp-encaps-06,
823		 *    section 3.1.2
824		 */
825		if (x->props.mode == XFRM_MODE_TRANSPORT)
826			skb->ip_summed = CHECKSUM_UNNECESSARY;
827	}
828
829	skb_pull_rcsum(skb, hlen);
830	if (x->props.mode == XFRM_MODE_TUNNEL)
831		skb_reset_transport_header(skb);
832	else
833		skb_set_transport_header(skb, -ihl);
834
835	/* RFC4303: Drop dummy packets without any error */
836	if (err == IPPROTO_NONE)
837		err = -EINVAL;
838
839out:
840	return err;
841}
842EXPORT_SYMBOL_GPL(esp_input_done2);
843
844static void esp_input_done(struct crypto_async_request *base, int err)
845{
846	struct sk_buff *skb = base->data;
847
848	xfrm_input_resume(skb, esp_input_done2(skb, err));
849}
850
851static void esp_input_restore_header(struct sk_buff *skb)
852{
853	esp_restore_header(skb, 0);
854	__skb_pull(skb, 4);
855}
856
857static void esp_input_set_header(struct sk_buff *skb, __be32 *seqhi)
858{
859	struct xfrm_state *x = xfrm_input_state(skb);
860	struct ip_esp_hdr *esph;
861
862	/* For ESN we move the header forward by 4 bytes to
863	 * accomodate the high bits.  We will move it back after
864	 * decryption.
865	 */
866	if ((x->props.flags & XFRM_STATE_ESN)) {
867		esph = skb_push(skb, 4);
868		*seqhi = esph->spi;
869		esph->spi = esph->seq_no;
870		esph->seq_no = XFRM_SKB_CB(skb)->seq.input.hi;
871	}
872}
873
874static void esp_input_done_esn(struct crypto_async_request *base, int err)
875{
876	struct sk_buff *skb = base->data;
877
878	esp_input_restore_header(skb);
879	esp_input_done(base, err);
880}
881
882/*
883 * Note: detecting truncated vs. non-truncated authentication data is very
884 * expensive, so we only support truncated data, which is the recommended
885 * and common case.
886 */
887static int esp_input(struct xfrm_state *x, struct sk_buff *skb)
888{
889	struct crypto_aead *aead = x->data;
890	struct aead_request *req;
891	struct sk_buff *trailer;
892	int ivlen = crypto_aead_ivsize(aead);
893	int elen = skb->len - sizeof(struct ip_esp_hdr) - ivlen;
894	int nfrags;
895	int assoclen;
896	int seqhilen;
897	__be32 *seqhi;
898	void *tmp;
899	u8 *iv;
900	struct scatterlist *sg;
901	int err = -EINVAL;
902
903	if (!pskb_may_pull(skb, sizeof(struct ip_esp_hdr) + ivlen))
904		goto out;
905
906	if (elen <= 0)
907		goto out;
908
909	assoclen = sizeof(struct ip_esp_hdr);
910	seqhilen = 0;
911
912	if (x->props.flags & XFRM_STATE_ESN) {
913		seqhilen += sizeof(__be32);
914		assoclen += seqhilen;
915	}
916
917	if (!skb_cloned(skb)) {
918		if (!skb_is_nonlinear(skb)) {
919			nfrags = 1;
920
921			goto skip_cow;
922		} else if (!skb_has_frag_list(skb)) {
923			nfrags = skb_shinfo(skb)->nr_frags;
924			nfrags++;
925
926			goto skip_cow;
927		}
928	}
929
930	err = skb_cow_data(skb, 0, &trailer);
931	if (err < 0)
932		goto out;
933
934	nfrags = err;
935
936skip_cow:
937	err = -ENOMEM;
938	tmp = esp_alloc_tmp(aead, nfrags, seqhilen);
939	if (!tmp)
940		goto out;
941
942	ESP_SKB_CB(skb)->tmp = tmp;
943	seqhi = esp_tmp_extra(tmp);
944	iv = esp_tmp_iv(aead, tmp, seqhilen);
945	req = esp_tmp_req(aead, iv);
946	sg = esp_req_sg(aead, req);
947
948	esp_input_set_header(skb, seqhi);
949
950	sg_init_table(sg, nfrags);
951	err = skb_to_sgvec(skb, sg, 0, skb->len);
952	if (unlikely(err < 0)) {
953		kfree(tmp);
954		goto out;
955	}
956
957	skb->ip_summed = CHECKSUM_NONE;
958
959	if ((x->props.flags & XFRM_STATE_ESN))
960		aead_request_set_callback(req, 0, esp_input_done_esn, skb);
961	else
962		aead_request_set_callback(req, 0, esp_input_done, skb);
963
964	aead_request_set_crypt(req, sg, sg, elen + ivlen, iv);
965	aead_request_set_ad(req, assoclen);
966
967	err = crypto_aead_decrypt(req);
968	if (err == -EINPROGRESS)
969		goto out;
970
971	if ((x->props.flags & XFRM_STATE_ESN))
972		esp_input_restore_header(skb);
973
974	err = esp_input_done2(skb, err);
975
976out:
977	return err;
978}
979
980static int esp4_err(struct sk_buff *skb, u32 info)
981{
982	struct net *net = dev_net(skb->dev);
983	const struct iphdr *iph = (const struct iphdr *)skb->data;
984	struct ip_esp_hdr *esph = (struct ip_esp_hdr *)(skb->data+(iph->ihl<<2));
985	struct xfrm_state *x;
986
987	switch (icmp_hdr(skb)->type) {
988	case ICMP_DEST_UNREACH:
989		if (icmp_hdr(skb)->code != ICMP_FRAG_NEEDED)
990			return 0;
991	case ICMP_REDIRECT:
992		break;
993	default:
994		return 0;
995	}
996
997	x = xfrm_state_lookup(net, skb->mark, (const xfrm_address_t *)&iph->daddr,
998			      esph->spi, IPPROTO_ESP, AF_INET);
999	if (!x)
1000		return 0;
1001
1002	if (icmp_hdr(skb)->type == ICMP_DEST_UNREACH)
1003		ipv4_update_pmtu(skb, net, info, 0, IPPROTO_ESP);
1004	else
1005		ipv4_redirect(skb, net, 0, IPPROTO_ESP);
1006	xfrm_state_put(x);
1007
1008	return 0;
1009}
1010
1011static void esp_destroy(struct xfrm_state *x)
1012{
1013	struct crypto_aead *aead = x->data;
1014
1015	if (!aead)
1016		return;
1017
1018	crypto_free_aead(aead);
1019}
1020
1021static int esp_init_aead(struct xfrm_state *x)
1022{
1023	char aead_name[CRYPTO_MAX_ALG_NAME];
1024	struct crypto_aead *aead;
1025	int err;
1026
1027	err = -ENAMETOOLONG;
1028	if (snprintf(aead_name, CRYPTO_MAX_ALG_NAME, "%s(%s)",
1029		     x->geniv, x->aead->alg_name) >= CRYPTO_MAX_ALG_NAME)
1030		goto error;
1031
1032	aead = crypto_alloc_aead(aead_name, 0, 0);
1033	err = PTR_ERR(aead);
1034	if (IS_ERR(aead))
1035		goto error;
1036
1037	x->data = aead;
1038
1039	err = crypto_aead_setkey(aead, x->aead->alg_key,
1040				 (x->aead->alg_key_len + 7) / 8);
1041	if (err)
1042		goto error;
1043
1044	err = crypto_aead_setauthsize(aead, x->aead->alg_icv_len / 8);
1045	if (err)
1046		goto error;
1047
1048error:
1049	return err;
1050}
1051
1052static int esp_init_authenc(struct xfrm_state *x)
1053{
1054	struct crypto_aead *aead;
1055	struct crypto_authenc_key_param *param;
1056	struct rtattr *rta;
1057	char *key;
1058	char *p;
1059	char authenc_name[CRYPTO_MAX_ALG_NAME];
1060	unsigned int keylen;
1061	int err;
1062
1063	err = -EINVAL;
1064	if (!x->ealg)
1065		goto error;
1066
1067	err = -ENAMETOOLONG;
1068
1069	if ((x->props.flags & XFRM_STATE_ESN)) {
1070		if (snprintf(authenc_name, CRYPTO_MAX_ALG_NAME,
1071			     "%s%sauthencesn(%s,%s)%s",
1072			     x->geniv ?: "", x->geniv ? "(" : "",
1073			     x->aalg ? x->aalg->alg_name : "digest_null",
1074			     x->ealg->alg_name,
1075			     x->geniv ? ")" : "") >= CRYPTO_MAX_ALG_NAME)
1076			goto error;
1077	} else {
1078		if (snprintf(authenc_name, CRYPTO_MAX_ALG_NAME,
1079			     "%s%sauthenc(%s,%s)%s",
1080			     x->geniv ?: "", x->geniv ? "(" : "",
1081			     x->aalg ? x->aalg->alg_name : "digest_null",
1082			     x->ealg->alg_name,
1083			     x->geniv ? ")" : "") >= CRYPTO_MAX_ALG_NAME)
1084			goto error;
1085	}
1086
1087	aead = crypto_alloc_aead(authenc_name, 0, 0);
1088	err = PTR_ERR(aead);
1089	if (IS_ERR(aead))
1090		goto error;
1091
1092	x->data = aead;
1093
1094	keylen = (x->aalg ? (x->aalg->alg_key_len + 7) / 8 : 0) +
1095		 (x->ealg->alg_key_len + 7) / 8 + RTA_SPACE(sizeof(*param));
1096	err = -ENOMEM;
1097	key = kmalloc(keylen, GFP_KERNEL);
1098	if (!key)
1099		goto error;
1100
1101	p = key;
1102	rta = (void *)p;
1103	rta->rta_type = CRYPTO_AUTHENC_KEYA_PARAM;
1104	rta->rta_len = RTA_LENGTH(sizeof(*param));
1105	param = RTA_DATA(rta);
1106	p += RTA_SPACE(sizeof(*param));
1107
1108	if (x->aalg) {
1109		struct xfrm_algo_desc *aalg_desc;
1110
1111		memcpy(p, x->aalg->alg_key, (x->aalg->alg_key_len + 7) / 8);
1112		p += (x->aalg->alg_key_len + 7) / 8;
1113
1114		aalg_desc = xfrm_aalg_get_byname(x->aalg->alg_name, 0);
1115		BUG_ON(!aalg_desc);
1116
1117		err = -EINVAL;
1118		if (aalg_desc->uinfo.auth.icv_fullbits / 8 !=
1119		    crypto_aead_authsize(aead)) {
1120			pr_info("ESP: %s digestsize %u != %hu\n",
1121				x->aalg->alg_name,
1122				crypto_aead_authsize(aead),
1123				aalg_desc->uinfo.auth.icv_fullbits / 8);
1124			goto free_key;
1125		}
1126
1127		err = crypto_aead_setauthsize(
1128			aead, x->aalg->alg_trunc_len / 8);
1129		if (err)
1130			goto free_key;
1131	}
1132
1133	param->enckeylen = cpu_to_be32((x->ealg->alg_key_len + 7) / 8);
1134	memcpy(p, x->ealg->alg_key, (x->ealg->alg_key_len + 7) / 8);
1135
1136	err = crypto_aead_setkey(aead, key, keylen);
1137
1138free_key:
1139	kfree_sensitive(key);
1140
1141error:
1142	return err;
1143}
1144
1145static int esp_init_state(struct xfrm_state *x)
1146{
1147	struct crypto_aead *aead;
1148	u32 align;
1149	int err;
1150
1151	x->data = NULL;
1152
1153	if (x->aead)
1154		err = esp_init_aead(x);
1155	else
1156		err = esp_init_authenc(x);
1157
1158	if (err)
1159		goto error;
1160
1161	aead = x->data;
1162
1163	x->props.header_len = sizeof(struct ip_esp_hdr) +
1164			      crypto_aead_ivsize(aead);
1165	if (x->props.mode == XFRM_MODE_TUNNEL)
1166		x->props.header_len += sizeof(struct iphdr);
1167	else if (x->props.mode == XFRM_MODE_BEET && x->sel.family != AF_INET6)
1168		x->props.header_len += IPV4_BEET_PHMAXLEN;
1169	if (x->encap) {
1170		struct xfrm_encap_tmpl *encap = x->encap;
1171
1172		switch (encap->encap_type) {
1173		default:
1174			err = -EINVAL;
1175			goto error;
1176		case UDP_ENCAP_ESPINUDP:
1177			x->props.header_len += sizeof(struct udphdr);
1178			break;
1179		case UDP_ENCAP_ESPINUDP_NON_IKE:
1180			x->props.header_len += sizeof(struct udphdr) + 2 * sizeof(u32);
1181			break;
1182#ifdef CONFIG_INET_ESPINTCP
1183		case TCP_ENCAP_ESPINTCP:
1184			/* only the length field, TCP encap is done by
1185			 * the socket
1186			 */
1187			x->props.header_len += 2;
1188			break;
1189#endif
1190		}
1191	}
1192
1193	align = ALIGN(crypto_aead_blocksize(aead), 4);
1194	x->props.trailer_len = align + 1 + crypto_aead_authsize(aead);
1195
1196error:
1197	return err;
1198}
1199
1200static int esp4_rcv_cb(struct sk_buff *skb, int err)
1201{
1202	return 0;
1203}
1204
1205static const struct xfrm_type esp_type =
1206{
1207	.description	= "ESP4",
1208	.owner		= THIS_MODULE,
1209	.proto	     	= IPPROTO_ESP,
1210	.flags		= XFRM_TYPE_REPLAY_PROT,
1211	.init_state	= esp_init_state,
1212	.destructor	= esp_destroy,
1213	.input		= esp_input,
1214	.output		= esp_output,
1215};
1216
1217static struct xfrm4_protocol esp4_protocol = {
1218	.handler	=	xfrm4_rcv,
1219	.input_handler	=	xfrm_input,
1220	.cb_handler	=	esp4_rcv_cb,
1221	.err_handler	=	esp4_err,
1222	.priority	=	0,
1223};
1224
1225static int __init esp4_init(void)
1226{
1227	if (xfrm_register_type(&esp_type, AF_INET) < 0) {
1228		pr_info("%s: can't add xfrm type\n", __func__);
1229		return -EAGAIN;
1230	}
1231	if (xfrm4_protocol_register(&esp4_protocol, IPPROTO_ESP) < 0) {
1232		pr_info("%s: can't add protocol\n", __func__);
1233		xfrm_unregister_type(&esp_type, AF_INET);
1234		return -EAGAIN;
1235	}
1236	return 0;
1237}
1238
1239static void __exit esp4_fini(void)
1240{
1241	if (xfrm4_protocol_deregister(&esp4_protocol, IPPROTO_ESP) < 0)
1242		pr_info("%s: can't remove protocol\n", __func__);
1243	xfrm_unregister_type(&esp_type, AF_INET);
1244}
1245
1246module_init(esp4_init);
1247module_exit(esp4_fini);
1248MODULE_LICENSE("GPL");
1249MODULE_ALIAS_XFRM_TYPE(AF_INET, XFRM_PROTO_ESP);
1250