linux/net/ipv4/ah4.c
Michael Bommarito ec54093e6a xfrm: ah: account for ESN high bits in async callbacks
AH allocates its temporary auth/ICV layout differently when ESN is enabled:
the async ahash setup appends a 4-byte seqhi slot before the ICV or
auth_data area, but the async completion callbacks still reconstruct the
temporary layout as if seqhi were absent.

With an async AH implementation selected, that makes AH copy or compare
the wrong bytes on both the IPv4 and IPv6 paths. In UML repro on IPv4 AH
with ESN and forced async hmac(sha1), ping fails with 100% packet loss,
and the callback logs show the pre-fix drift:

  ah4 output_done: esn=1 err=0 icv_off=20 expected_off=24
  ah4 input_done: esn=1 auth_off=20 expected_auth_off=24 icv_off=32 expected_icv_off=36

Reconstruct the callback-side layout the same way the setup path built it
by skipping the ESN seqhi slot before locating the saved auth_data or ICV.
Per RFC 4302, the ESN high-order 32 bits participate in the AH ICV
computation, so the async callbacks must account for the seqhi slot.

Post-fix, the same IPv4 AH+ESN+forced-async-hmac(sha1) UML repro shows
the corrected offset (ah4 output_done: esn=1 err=0 icv_off=24
expected_off=24) and ping succeeds; net/ipv4/ah4.o and net/ipv6/ah6.o
build clean at W=1. IPv6 AH+ESN was not exercised at runtime, and the
change has not been tested against a real async hardware AH engine.

Fixes: d4d573d033 ("{IPv4,xfrm} Add ESN support for AH egress part")
Fixes: d8b2a8600b ("{IPv4,xfrm} Add ESN support for AH ingress part")
Fixes: 26dd70c3fa ("{IPv6,xfrm} Add ESN support for AH egress part")
Fixes: 8d6da6f325 ("{IPv6,xfrm} Add ESN support for AH ingress part")
Cc: stable@vger.kernel.org
Assisted-by: Codex:gpt-5-4
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: Michael Bommarito <michael.bommarito@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
2026-04-20 09:28:34 +02:00

613 lines
14 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
#define pr_fmt(fmt) "IPsec: " fmt
#include <crypto/hash.h>
#include <crypto/utils.h>
#include <linux/err.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <net/ip.h>
#include <net/xfrm.h>
#include <net/ah.h>
#include <linux/crypto.h>
#include <linux/pfkeyv2.h>
#include <linux/scatterlist.h>
#include <net/icmp.h>
#include <net/protocol.h>
struct ah_skb_cb {
struct xfrm_skb_cb xfrm;
void *tmp;
};
#define AH_SKB_CB(__skb) ((struct ah_skb_cb *)&((__skb)->cb[0]))
static void *ah_alloc_tmp(struct crypto_ahash *ahash, int nfrags,
unsigned int size)
{
unsigned int len;
len = size + crypto_ahash_digestsize(ahash);
len = ALIGN(len, crypto_tfm_ctx_alignment());
len += sizeof(struct ahash_request) + crypto_ahash_reqsize(ahash);
len = ALIGN(len, __alignof__(struct scatterlist));
len += sizeof(struct scatterlist) * nfrags;
return kmalloc(len, GFP_ATOMIC);
}
static inline u8 *ah_tmp_auth(void *tmp, unsigned int offset)
{
return tmp + offset;
}
static inline u8 *ah_tmp_icv(void *tmp, unsigned int offset)
{
return tmp + offset;
}
static inline struct ahash_request *ah_tmp_req(struct crypto_ahash *ahash,
u8 *icv)
{
struct ahash_request *req;
req = (void *)PTR_ALIGN(icv + crypto_ahash_digestsize(ahash),
crypto_tfm_ctx_alignment());
ahash_request_set_tfm(req, ahash);
return req;
}
static inline struct scatterlist *ah_req_sg(struct crypto_ahash *ahash,
struct ahash_request *req)
{
return (void *)ALIGN((unsigned long)(req + 1) +
crypto_ahash_reqsize(ahash),
__alignof__(struct scatterlist));
}
/* Clear mutable options and find final destination to substitute
* into IP header for icv calculation. Options are already checked
* for validity, so paranoia is not required. */
static int ip_clear_mutable_options(const struct iphdr *iph, __be32 *daddr)
{
unsigned char *optptr = (unsigned char *)(iph+1);
int l = iph->ihl*4 - sizeof(struct iphdr);
int optlen;
while (l > 0) {
switch (*optptr) {
case IPOPT_END:
return 0;
case IPOPT_NOOP:
l--;
optptr++;
continue;
}
optlen = optptr[1];
if (optlen<2 || optlen>l)
return -EINVAL;
switch (*optptr) {
case IPOPT_SEC:
case 0x85: /* Some "Extended Security" crap. */
case IPOPT_CIPSO:
case IPOPT_RA:
case 0x80|21: /* RFC1770 */
break;
case IPOPT_LSRR:
case IPOPT_SSRR:
if (optlen < 6)
return -EINVAL;
memcpy(daddr, optptr+optlen-4, 4);
fallthrough;
default:
memset(optptr, 0, optlen);
}
l -= optlen;
optptr += optlen;
}
return 0;
}
static void ah_output_done(void *data, int err)
{
u8 *icv;
struct iphdr *iph;
struct sk_buff *skb = data;
struct xfrm_state *x = skb_dst(skb)->xfrm;
struct ah_data *ahp = x->data;
struct iphdr *top_iph = ip_hdr(skb);
struct ip_auth_hdr *ah = ip_auth_hdr(skb);
int ihl = ip_hdrlen(skb);
int seqhi_len = 0;
__be32 *seqhi;
if (x->props.flags & XFRM_STATE_ESN)
seqhi_len = sizeof(*seqhi);
iph = AH_SKB_CB(skb)->tmp;
seqhi = (__be32 *)((char *)iph + ihl);
icv = ah_tmp_icv(seqhi, seqhi_len);
memcpy(ah->auth_data, icv, ahp->icv_trunc_len);
top_iph->tos = iph->tos;
top_iph->ttl = iph->ttl;
top_iph->frag_off = iph->frag_off;
if (top_iph->ihl != 5) {
top_iph->daddr = iph->daddr;
memcpy(top_iph+1, iph+1, top_iph->ihl*4 - sizeof(struct iphdr));
}
kfree(AH_SKB_CB(skb)->tmp);
xfrm_output_resume(skb->sk, skb, err);
}
static int ah_output(struct xfrm_state *x, struct sk_buff *skb)
{
int err;
int nfrags;
int ihl;
u8 *icv;
struct sk_buff *trailer;
struct crypto_ahash *ahash;
struct ahash_request *req;
struct scatterlist *sg;
struct iphdr *iph, *top_iph;
struct ip_auth_hdr *ah;
struct ah_data *ahp;
int seqhi_len = 0;
__be32 *seqhi;
int sglists = 0;
struct scatterlist *seqhisg;
ahp = x->data;
ahash = ahp->ahash;
if ((err = skb_cow_data(skb, 0, &trailer)) < 0)
goto out;
nfrags = err;
skb_push(skb, -skb_network_offset(skb));
ah = ip_auth_hdr(skb);
ihl = ip_hdrlen(skb);
if (x->props.flags & XFRM_STATE_ESN) {
sglists = 1;
seqhi_len = sizeof(*seqhi);
}
err = -ENOMEM;
iph = ah_alloc_tmp(ahash, nfrags + sglists, ihl + seqhi_len);
if (!iph)
goto out;
seqhi = (__be32 *)((char *)iph + ihl);
icv = ah_tmp_icv(seqhi, seqhi_len);
req = ah_tmp_req(ahash, icv);
sg = ah_req_sg(ahash, req);
seqhisg = sg + nfrags;
memset(ah->auth_data, 0, ahp->icv_trunc_len);
top_iph = ip_hdr(skb);
iph->tos = top_iph->tos;
iph->ttl = top_iph->ttl;
iph->frag_off = top_iph->frag_off;
if (top_iph->ihl != 5) {
iph->daddr = top_iph->daddr;
memcpy(iph+1, top_iph+1, top_iph->ihl*4 - sizeof(struct iphdr));
err = ip_clear_mutable_options(top_iph, &top_iph->daddr);
if (err)
goto out_free;
}
ah->nexthdr = *skb_mac_header(skb);
*skb_mac_header(skb) = IPPROTO_AH;
top_iph->tos = 0;
top_iph->tot_len = htons(skb->len);
top_iph->frag_off = 0;
top_iph->ttl = 0;
top_iph->check = 0;
if (x->props.flags & XFRM_STATE_ALIGN4)
ah->hdrlen = (XFRM_ALIGN4(sizeof(*ah) + ahp->icv_trunc_len) >> 2) - 2;
else
ah->hdrlen = (XFRM_ALIGN8(sizeof(*ah) + ahp->icv_trunc_len) >> 2) - 2;
ah->reserved = 0;
ah->spi = x->id.spi;
ah->seq_no = htonl(XFRM_SKB_CB(skb)->seq.output.low);
sg_init_table(sg, nfrags + sglists);
err = skb_to_sgvec_nomark(skb, sg, 0, skb->len);
if (unlikely(err < 0))
goto out_free;
if (x->props.flags & XFRM_STATE_ESN) {
/* Attach seqhi sg right after packet payload */
*seqhi = htonl(XFRM_SKB_CB(skb)->seq.output.hi);
sg_set_buf(seqhisg, seqhi, seqhi_len);
}
ahash_request_set_crypt(req, sg, icv, skb->len + seqhi_len);
ahash_request_set_callback(req, 0, ah_output_done, skb);
AH_SKB_CB(skb)->tmp = iph;
err = crypto_ahash_digest(req);
if (err) {
if (err == -EINPROGRESS)
goto out;
if (err == -ENOSPC)
err = NET_XMIT_DROP;
goto out_free;
}
memcpy(ah->auth_data, icv, ahp->icv_trunc_len);
top_iph->tos = iph->tos;
top_iph->ttl = iph->ttl;
top_iph->frag_off = iph->frag_off;
if (top_iph->ihl != 5) {
top_iph->daddr = iph->daddr;
memcpy(top_iph+1, iph+1, top_iph->ihl*4 - sizeof(struct iphdr));
}
out_free:
kfree(iph);
out:
return err;
}
static void ah_input_done(void *data, int err)
{
u8 *auth_data;
u8 *icv;
struct iphdr *work_iph;
struct sk_buff *skb = data;
struct xfrm_state *x = xfrm_input_state(skb);
struct ah_data *ahp = x->data;
struct ip_auth_hdr *ah = ip_auth_hdr(skb);
int ihl = ip_hdrlen(skb);
int ah_hlen = (ah->hdrlen + 2) << 2;
int seqhi_len = 0;
__be32 *seqhi;
if (err)
goto out;
if (x->props.flags & XFRM_STATE_ESN)
seqhi_len = sizeof(*seqhi);
work_iph = AH_SKB_CB(skb)->tmp;
seqhi = (__be32 *)((char *)work_iph + ihl);
auth_data = ah_tmp_auth(seqhi, seqhi_len);
icv = ah_tmp_icv(auth_data, ahp->icv_trunc_len);
err = crypto_memneq(icv, auth_data, ahp->icv_trunc_len) ? -EBADMSG : 0;
if (err)
goto out;
err = ah->nexthdr;
skb->network_header += ah_hlen;
memcpy(skb_network_header(skb), work_iph, ihl);
__skb_pull(skb, ah_hlen + ihl);
if (x->props.mode == XFRM_MODE_TUNNEL)
skb_reset_transport_header(skb);
else
skb_set_transport_header(skb, -ihl);
out:
kfree(AH_SKB_CB(skb)->tmp);
xfrm_input_resume(skb, err);
}
static int ah_input(struct xfrm_state *x, struct sk_buff *skb)
{
int ah_hlen;
int ihl;
int nexthdr;
int nfrags;
u8 *auth_data;
u8 *icv;
struct sk_buff *trailer;
struct crypto_ahash *ahash;
struct ahash_request *req;
struct scatterlist *sg;
struct iphdr *iph, *work_iph;
struct ip_auth_hdr *ah;
struct ah_data *ahp;
int err = -ENOMEM;
int seqhi_len = 0;
__be32 *seqhi;
int sglists = 0;
struct scatterlist *seqhisg;
if (!pskb_may_pull(skb, sizeof(*ah)))
goto out;
ah = (struct ip_auth_hdr *)skb->data;
ahp = x->data;
ahash = ahp->ahash;
nexthdr = ah->nexthdr;
ah_hlen = (ah->hdrlen + 2) << 2;
if (x->props.flags & XFRM_STATE_ALIGN4) {
if (ah_hlen != XFRM_ALIGN4(sizeof(*ah) + ahp->icv_full_len) &&
ah_hlen != XFRM_ALIGN4(sizeof(*ah) + ahp->icv_trunc_len))
goto out;
} else {
if (ah_hlen != XFRM_ALIGN8(sizeof(*ah) + ahp->icv_full_len) &&
ah_hlen != XFRM_ALIGN8(sizeof(*ah) + ahp->icv_trunc_len))
goto out;
}
if (!pskb_may_pull(skb, ah_hlen))
goto out;
/* We are going to _remove_ AH header to keep sockets happy,
* so... Later this can change. */
if (skb_unclone(skb, GFP_ATOMIC))
goto out;
skb->ip_summed = CHECKSUM_NONE;
if ((err = skb_cow_data(skb, 0, &trailer)) < 0)
goto out;
nfrags = err;
ah = (struct ip_auth_hdr *)skb->data;
iph = ip_hdr(skb);
ihl = ip_hdrlen(skb);
if (x->props.flags & XFRM_STATE_ESN) {
sglists = 1;
seqhi_len = sizeof(*seqhi);
}
work_iph = ah_alloc_tmp(ahash, nfrags + sglists, ihl +
ahp->icv_trunc_len + seqhi_len);
if (!work_iph) {
err = -ENOMEM;
goto out;
}
seqhi = (__be32 *)((char *)work_iph + ihl);
auth_data = ah_tmp_auth(seqhi, seqhi_len);
icv = ah_tmp_icv(auth_data, ahp->icv_trunc_len);
req = ah_tmp_req(ahash, icv);
sg = ah_req_sg(ahash, req);
seqhisg = sg + nfrags;
memcpy(work_iph, iph, ihl);
memcpy(auth_data, ah->auth_data, ahp->icv_trunc_len);
memset(ah->auth_data, 0, ahp->icv_trunc_len);
iph->ttl = 0;
iph->tos = 0;
iph->frag_off = 0;
iph->check = 0;
if (ihl > sizeof(*iph)) {
__be32 dummy;
err = ip_clear_mutable_options(iph, &dummy);
if (err)
goto out_free;
}
skb_push(skb, ihl);
sg_init_table(sg, nfrags + sglists);
err = skb_to_sgvec_nomark(skb, sg, 0, skb->len);
if (unlikely(err < 0))
goto out_free;
if (x->props.flags & XFRM_STATE_ESN) {
/* Attach seqhi sg right after packet payload */
*seqhi = XFRM_SKB_CB(skb)->seq.input.hi;
sg_set_buf(seqhisg, seqhi, seqhi_len);
}
ahash_request_set_crypt(req, sg, icv, skb->len + seqhi_len);
ahash_request_set_callback(req, 0, ah_input_done, skb);
AH_SKB_CB(skb)->tmp = work_iph;
err = crypto_ahash_digest(req);
if (err) {
if (err == -EINPROGRESS)
goto out;
goto out_free;
}
err = crypto_memneq(icv, auth_data, ahp->icv_trunc_len) ? -EBADMSG : 0;
if (err)
goto out_free;
skb->network_header += ah_hlen;
memcpy(skb_network_header(skb), work_iph, ihl);
__skb_pull(skb, ah_hlen + ihl);
if (x->props.mode == XFRM_MODE_TUNNEL)
skb_reset_transport_header(skb);
else
skb_set_transport_header(skb, -ihl);
err = nexthdr;
out_free:
kfree (work_iph);
out:
return err;
}
static int ah4_err(struct sk_buff *skb, u32 info)
{
struct net *net = dev_net(skb->dev);
const struct iphdr *iph = (const struct iphdr *)skb->data;
struct ip_auth_hdr *ah = (struct ip_auth_hdr *)(skb->data+(iph->ihl<<2));
struct xfrm_state *x;
switch (icmp_hdr(skb)->type) {
case ICMP_DEST_UNREACH:
if (icmp_hdr(skb)->code != ICMP_FRAG_NEEDED)
return 0;
break;
case ICMP_REDIRECT:
break;
default:
return 0;
}
x = xfrm_state_lookup(net, skb->mark, (const xfrm_address_t *)&iph->daddr,
ah->spi, IPPROTO_AH, AF_INET);
if (!x)
return 0;
if (icmp_hdr(skb)->type == ICMP_DEST_UNREACH)
ipv4_update_pmtu(skb, net, info, 0, IPPROTO_AH);
else
ipv4_redirect(skb, net, 0, IPPROTO_AH);
xfrm_state_put(x);
return 0;
}
static int ah_init_state(struct xfrm_state *x, struct netlink_ext_ack *extack)
{
struct ah_data *ahp = NULL;
struct xfrm_algo_desc *aalg_desc;
struct crypto_ahash *ahash;
if (!x->aalg) {
NL_SET_ERR_MSG(extack, "AH requires a state with an AUTH algorithm");
goto error;
}
if (x->encap) {
NL_SET_ERR_MSG(extack, "AH is not compatible with encapsulation");
goto error;
}
ahp = kzalloc_obj(*ahp);
if (!ahp)
return -ENOMEM;
ahash = crypto_alloc_ahash(x->aalg->alg_name, 0, 0);
if (IS_ERR(ahash)) {
NL_SET_ERR_MSG(extack, "Kernel was unable to initialize cryptographic operations");
goto error;
}
ahp->ahash = ahash;
if (crypto_ahash_setkey(ahash, x->aalg->alg_key,
(x->aalg->alg_key_len + 7) / 8)) {
NL_SET_ERR_MSG(extack, "Kernel was unable to initialize cryptographic operations");
goto error;
}
/*
* Lookup the algorithm description maintained by xfrm_algo,
* verify crypto transform properties, and store information
* we need for AH processing. This lookup cannot fail here
* after a successful crypto_alloc_ahash().
*/
aalg_desc = xfrm_aalg_get_byname(x->aalg->alg_name, 0);
BUG_ON(!aalg_desc);
if (aalg_desc->uinfo.auth.icv_fullbits/8 !=
crypto_ahash_digestsize(ahash)) {
NL_SET_ERR_MSG(extack, "Kernel was unable to initialize cryptographic operations");
goto error;
}
ahp->icv_full_len = aalg_desc->uinfo.auth.icv_fullbits/8;
ahp->icv_trunc_len = x->aalg->alg_trunc_len/8;
if (x->props.flags & XFRM_STATE_ALIGN4)
x->props.header_len = XFRM_ALIGN4(sizeof(struct ip_auth_hdr) +
ahp->icv_trunc_len);
else
x->props.header_len = XFRM_ALIGN8(sizeof(struct ip_auth_hdr) +
ahp->icv_trunc_len);
if (x->props.mode == XFRM_MODE_TUNNEL)
x->props.header_len += sizeof(struct iphdr);
x->data = ahp;
return 0;
error:
if (ahp) {
crypto_free_ahash(ahp->ahash);
kfree(ahp);
}
return -EINVAL;
}
static void ah_destroy(struct xfrm_state *x)
{
struct ah_data *ahp = x->data;
if (!ahp)
return;
crypto_free_ahash(ahp->ahash);
kfree(ahp);
}
static int ah4_rcv_cb(struct sk_buff *skb, int err)
{
return 0;
}
static const struct xfrm_type ah_type =
{
.owner = THIS_MODULE,
.proto = IPPROTO_AH,
.flags = XFRM_TYPE_REPLAY_PROT,
.init_state = ah_init_state,
.destructor = ah_destroy,
.input = ah_input,
.output = ah_output
};
static struct xfrm4_protocol ah4_protocol = {
.handler = xfrm4_rcv,
.input_handler = xfrm_input,
.cb_handler = ah4_rcv_cb,
.err_handler = ah4_err,
.priority = 0,
};
static int __init ah4_init(void)
{
if (xfrm_register_type(&ah_type, AF_INET) < 0) {
pr_info("%s: can't add xfrm type\n", __func__);
return -EAGAIN;
}
if (xfrm4_protocol_register(&ah4_protocol, IPPROTO_AH) < 0) {
pr_info("%s: can't add protocol\n", __func__);
xfrm_unregister_type(&ah_type, AF_INET);
return -EAGAIN;
}
return 0;
}
static void __exit ah4_fini(void)
{
if (xfrm4_protocol_deregister(&ah4_protocol, IPPROTO_AH) < 0)
pr_info("%s: can't remove protocol\n", __func__);
xfrm_unregister_type(&ah_type, AF_INET);
}
module_init(ah4_init);
module_exit(ah4_fini);
MODULE_DESCRIPTION("IPv4 AH transformation library");
MODULE_LICENSE("GPL");
MODULE_ALIAS_XFRM_TYPE(AF_INET, XFRM_PROTO_AH);