[PATCHv4 net-next 0/6] xfrm: Add ESN support for AH

[PATCHv4 net-next 0/6] xfrm: Add ESN support for AH

[Fan Du]

Hi,

This is initial Extended Sequence Number support for AH based on IPv4/6.
The rationale is totally by the RFC 4302, which states:

3.3.3.2.2.  Implicit Packet Padding and ESN

   If the ESN option is elected for an SA, then the high-order 32 bits
   of the ESN must be included in the ICV computation.  For purposes of
   ICV computation, these bits are appended (implicitly) immediately
   after the end of the payload, and before any implicit packet padding.

So we attach the high-order 32bits as a scatterlist right after the packet
payload to compute ICV value. 

Test:
I add a knob in iproute2/ip/xfrm_state.c to enable esn when setting SA,
which make it possible to test with-esn and without-esn scenarios, both
cases works ok with ping using packetsize(-s) from default to 32768. 

v2:
  - Patch3/5 and Patch4/5 add IPv6 part as requested by Steffen.
  - Patch5/5 restrict ESN feature only to ESP and AH.
v3:
  - Fix double parens spotted by Sergei, and thanks for reporting.
v4:
  - Incorperate feedbacks from Steffen by simplify the code flow.
  - Add Patch1/6 to introduce skb_to_sgvec_nomark
  - Patch2/6 remove rebundant inclusion crypto/scatterwalk.h

Fan Du (6):
  skbuff: Introduce skb_to_sgvec_nomark to map skb without mark new end
  {IPv4,xfrm} Add ESN support for AH egress part
  {IPv4,xfrm} Add ESN support for AH ingress part
  {IPv6,xfrm} Add ESN support for AH egress part
  {IPv6,xfrm} Add ESN support for AH ingress part
  xfrm: Don't prohibit AH from using ESN feature

 include/linux/skbuff.h |    2 ++
 net/core/skbuff.c      |   26 ++++++++++++++++++++++
 net/ipv4/ah4.c         |   53 +++++++++++++++++++++++++++++++++++----------
 net/ipv6/ah6.c         |   56 ++++++++++++++++++++++++++++++++++++++----------
 net/xfrm/xfrm_user.c   |    3 ++-
 5 files changed, 117 insertions(+), 23 deletions(-)

-- 
1.7.9.5

[PATCHv4 net-next 1/6] skbuff: Introduce skb_to_sgvec_nomark to map skb without mark new end

[Fan Du]

As compared with skb_to_sgvec, skb_to_sgvec_nomark only map skb to given sglist
without mark the sg which contain last skb data as the end. So the caller can
mannipulate sg list as will when padding new data after the first call without
calling sg_unmark_end to expend sg list.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 include/linux/skbuff.h |    2 ++
 net/core/skbuff.c      |   26 ++++++++++++++++++++++++++
 2 files changed, 28 insertions(+)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index d97f2d0..c2eb3c4 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -691,6 +691,8 @@ struct sk_buff *skb_realloc_headroom(struct sk_buff *skb,
 				     unsigned int headroom);
 struct sk_buff *skb_copy_expand(const struct sk_buff *skb, int newheadroom,
 				int newtailroom, gfp_t priority);
+int skb_to_sgvec_nomark(struct sk_buff *skb, struct scatterlist *sg,
+			int offset, int len);
 int skb_to_sgvec(struct sk_buff *skb, struct scatterlist *sg, int offset,
 		 int len);
 int skb_cow_data(struct sk_buff *skb, int tailbits, struct sk_buff **trailer);
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 1d641e7..466c0ef 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -3308,6 +3308,32 @@ __skb_to_sgvec(struct sk_buff *skb, struct scatterlist *sg, int offset, int len)
 	return elt;
 }
 
+/* As compared with skb_to_sgvec, skb_to_sgvec_nomark only map skb to given
+ * sglist without mark the sg which contain last skb data as the end.
+ * So the caller can mannipulate sg list as will when padding new data after
+ * the first call without calling sg_unmark_end to expend sg list.
+ *
+ * Scenario to use skb_to_sgvec_nomark:
+ * 1. sg_init_table
+ * 2. skb_to_sgvec_nomark(payload1)
+ * 3. skb_to_sgvec_nomark(payload2)
+ *
+ * This is equivalent to:
+ * 1. sg_init_table
+ * 2. skb_to_sgvec(payload1)
+ * 3. sg_unmark_end
+ * 4. skb_to_sgvec(payload2)
+ *
+ * When mapping mutilple payload conditionally, skb_to_sgvec_nomark
+ * is more preferable.
+ */
+int skb_to_sgvec_nomark(struct sk_buff *skb, struct scatterlist *sg,
+			int offset, int len)
+{
+	return __skb_to_sgvec(skb, sg, offset, len);
+}
+EXPORT_SYMBOL_GPL(skb_to_sgvec_nomark);
+
 int skb_to_sgvec(struct sk_buff *skb, struct scatterlist *sg, int offset, int len)
 {
 	int nsg = __skb_to_sgvec(skb, sg, offset, len);
-- 
1.7.9.5

[PATCHv4 net-next 2/6] {IPv4,xfrm} Add ESN support for AH egress part

[Fan Du]

This patch add esn support for AH output stage by attaching upper 32bits
sequence number right after packet payload as specified by RFC 4302.

Then the ICV value will guard upper 32bits sequence number as well when
packet going out.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 net/ipv4/ah4.c |   26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)

diff --git a/net/ipv4/ah4.c b/net/ipv4/ah4.c
index 7179026..c6accac 100644
--- a/net/ipv4/ah4.c
+++ b/net/ipv4/ah4.c
@@ -155,6 +155,10 @@ static int ah_output(struct xfrm_state *x, struct sk_buff *skb)
 	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;
@@ -167,14 +171,19 @@ static int ah_output(struct xfrm_state *x, struct sk_buff *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, ihl);
+	iph = ah_alloc_tmp(ahash, nfrags + sglists, ihl + seqhi_len);
 	if (!iph)
 		goto out;
-
-	icv = ah_tmp_icv(ahash, iph, ihl);
+	seqhi = (__be32 *)((char *)iph + ihl);
+	icv = ah_tmp_icv(ahash, 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);
 
@@ -210,10 +219,15 @@ static int ah_output(struct xfrm_state *x, struct sk_buff *skb)
 	ah->spi = x->id.spi;
 	ah->seq_no = htonl(XFRM_SKB_CB(skb)->seq.output.low);
 
-	sg_init_table(sg, nfrags);
-	skb_to_sgvec(skb, sg, 0, skb->len);
+	sg_init_table(sg, nfrags + sglists);
+	skb_to_sgvec_nomark(skb, sg, 0, skb->len);
 
-	ahash_request_set_crypt(req, sg, icv, skb->len);
+	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;
-- 
1.7.9.5

[PATCHv4 net-next 3/6] {IPv4,xfrm} Add ESN support for AH ingress part

[Fan Du]

This patch add esn support for AH input stage by attaching upper 32bits
sequence number right after packet payload as specified by RFC 4302.

Then the ICV value will guard upper 32bits sequence number as well when
packet getting in.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 net/ipv4/ah4.c |   27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/net/ipv4/ah4.c b/net/ipv4/ah4.c
index c6accac..c94ef19 100644
--- a/net/ipv4/ah4.c
+++ b/net/ipv4/ah4.c
@@ -309,6 +309,10 @@ static int ah_input(struct xfrm_state *x, struct sk_buff *skb)
 	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;
@@ -349,14 +353,22 @@ static int ah_input(struct xfrm_state *x, struct sk_buff *skb)
 	iph = ip_hdr(skb);
 	ihl = ip_hdrlen(skb);
 
-	work_iph = ah_alloc_tmp(ahash, nfrags, ihl + ahp->icv_trunc_len);
+	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)
 		goto out;
 
-	auth_data = ah_tmp_auth(work_iph, ihl);
+	seqhi = (__be32 *)((char *)work_iph + ihl);
+	auth_data = ah_tmp_auth(seqhi, seqhi_len);
 	icv = ah_tmp_icv(ahash, 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);
@@ -375,10 +387,15 @@ static int ah_input(struct xfrm_state *x, struct sk_buff *skb)
 
 	skb_push(skb, ihl);
 
-	sg_init_table(sg, nfrags);
-	skb_to_sgvec(skb, sg, 0, skb->len);
+	sg_init_table(sg, nfrags + sglists);
+	skb_to_sgvec_nomark(skb, sg, 0, skb->len);
 
-	ahash_request_set_crypt(req, sg, icv, skb->len);
+	if (x->props.flags & XFRM_STATE_ESN) {
+		/* Attach seqhi sg right after packet payload */
+		*seqhi = htonl(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;
-- 
1.7.9.5

[PATCHv4 net-next 4/6] {IPv6,xfrm} Add ESN support for AH egress part

[Fan Du]

This patch add esn support for AH output stage by attaching upper 32bits
sequence number right after packet payload as specified by RFC 4302.

Then the ICV value will guard upper 32bits sequence number as well when
packet going out.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 net/ipv6/ah6.c |   26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/net/ipv6/ah6.c b/net/ipv6/ah6.c
index 81e496a..8929812 100644
--- a/net/ipv6/ah6.c
+++ b/net/ipv6/ah6.c
@@ -346,6 +346,10 @@ static int ah6_output(struct xfrm_state *x, struct sk_buff *skb)
 	struct ip_auth_hdr *ah;
 	struct ah_data *ahp;
 	struct tmp_ext *iph_ext;
+	int seqhi_len = 0;
+	__be32 *seqhi;
+	int sglists = 0;
+	struct scatterlist *seqhisg;
 
 	ahp = x->data;
 	ahash = ahp->ahash;
@@ -359,15 +363,22 @@ static int ah6_output(struct xfrm_state *x, struct sk_buff *skb)
 	if (extlen)
 		extlen += sizeof(*iph_ext);
 
+	if (x->props.flags & XFRM_STATE_ESN) {
+		sglists = 1;
+		seqhi_len = sizeof(*seqhi);
+	}
 	err = -ENOMEM;
-	iph_base = ah_alloc_tmp(ahash, nfrags, IPV6HDR_BASELEN + extlen);
+	iph_base = ah_alloc_tmp(ahash, nfrags + sglists, IPV6HDR_BASELEN +
+				extlen + seqhi_len);
 	if (!iph_base)
 		goto out;
 
 	iph_ext = ah_tmp_ext(iph_base);
-	icv = ah_tmp_icv(ahash, iph_ext, extlen);
+	seqhi = (__be32 *)((char *)iph_ext + extlen);
+	icv = ah_tmp_icv(ahash, seqhi, seqhi_len);
 	req = ah_tmp_req(ahash, icv);
 	sg = ah_req_sg(ahash, req);
+	seqhisg = sg + nfrags;
 
 	ah = ip_auth_hdr(skb);
 	memset(ah->auth_data, 0, ahp->icv_trunc_len);
@@ -411,10 +422,15 @@ static int ah6_output(struct xfrm_state *x, struct sk_buff *skb)
 	ah->spi = x->id.spi;
 	ah->seq_no = htonl(XFRM_SKB_CB(skb)->seq.output.low);
 
-	sg_init_table(sg, nfrags);
-	skb_to_sgvec(skb, sg, 0, skb->len);
+	sg_init_table(sg, nfrags + sglists);
+	skb_to_sgvec_nomark(skb, sg, 0, skb->len);
 
-	ahash_request_set_crypt(req, sg, icv, skb->len);
+	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, ah6_output_done, skb);
 
 	AH_SKB_CB(skb)->tmp = iph_base;
-- 
1.7.9.5

[PATCHv4 net-next 5/6] {IPv6,xfrm} Add ESN support for AH ingress part

[Fan Du]

This patch add esn support for AH input stage by attaching upper 32bits
sequence number right after packet payload as specified by RFC 4302.

Then the ICV value will guard upper 32bits sequence number as well when
packet going in.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 net/ipv6/ah6.c |   30 ++++++++++++++++++++++++------
 1 file changed, 24 insertions(+), 6 deletions(-)

diff --git a/net/ipv6/ah6.c b/net/ipv6/ah6.c
index 8929812..558c549 100644
--- a/net/ipv6/ah6.c
+++ b/net/ipv6/ah6.c
@@ -530,6 +530,10 @@ static int ah6_input(struct xfrm_state *x, struct sk_buff *skb)
 	int nexthdr;
 	int nfrags;
 	int err = -ENOMEM;
+	int seqhi_len = 0;
+	__be32 *seqhi;
+	int sglists = 0;
+	struct scatterlist *seqhisg;
 
 	if (!pskb_may_pull(skb, sizeof(struct ip_auth_hdr)))
 		goto out;
@@ -566,14 +570,22 @@ static int ah6_input(struct xfrm_state *x, struct sk_buff *skb)
 
 	skb_push(skb, hdr_len);
 
-	work_iph = ah_alloc_tmp(ahash, nfrags, hdr_len + ahp->icv_trunc_len);
+	if (x->props.flags & XFRM_STATE_ESN) {
+		sglists = 1;
+		seqhi_len = sizeof(*seqhi);
+	}
+
+	work_iph = ah_alloc_tmp(ahash, nfrags + sglists, hdr_len +
+				ahp->icv_trunc_len + seqhi_len);
 	if (!work_iph)
 		goto out;
 
-	auth_data = ah_tmp_auth(work_iph, hdr_len);
-	icv = ah_tmp_icv(ahash, auth_data, ahp->icv_trunc_len);
+	auth_data = ah_tmp_auth((u8 *)work_iph, hdr_len);
+	seqhi = (__be32 *)(auth_data + ahp->icv_trunc_len);
+	icv = ah_tmp_icv(ahash, seqhi, seqhi_len);
 	req = ah_tmp_req(ahash, icv);
 	sg = ah_req_sg(ahash, req);
+	seqhisg = sg + nfrags;
 
 	memcpy(work_iph, ip6h, hdr_len);
 	memcpy(auth_data, ah->auth_data, ahp->icv_trunc_len);
@@ -588,10 +600,16 @@ static int ah6_input(struct xfrm_state *x, struct sk_buff *skb)
 	ip6h->flow_lbl[2] = 0;
 	ip6h->hop_limit   = 0;
 
-	sg_init_table(sg, nfrags);
-	skb_to_sgvec(skb, sg, 0, skb->len);
+	sg_init_table(sg, nfrags + sglists);
+	skb_to_sgvec_nomark(skb, sg, 0, skb->len);
+
+	if (x->props.flags & XFRM_STATE_ESN) {
+		/* Attach seqhi sg right after packet payload */
+		*seqhi = htonl(XFRM_SKB_CB(skb)->seq.input.hi);
+		sg_set_buf(seqhisg, seqhi, seqhi_len);
+	}
 
-	ahash_request_set_crypt(req, sg, icv, skb->len);
+	ahash_request_set_crypt(req, sg, icv, skb->len + seqhi_len);
 	ahash_request_set_callback(req, 0, ah6_input_done, skb);
 
 	AH_SKB_CB(skb)->tmp = work_iph;
-- 
1.7.9.5

[PATCHv4 net-next 6/6] xfrm: Don't prohibit AH from using ESN feature

[Fan Du]

Clear checking when user try to use ESN through netlink keymgr for AH.
As only ESP and AH support ESN feature according to RFC.

Signed-off-by: Fan Du <fan.du@windriver.com>
---
 net/xfrm/xfrm_user.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c
index 3348566..9a80aa9 100644
--- a/net/xfrm/xfrm_user.c
+++ b/net/xfrm/xfrm_user.c
@@ -142,7 +142,8 @@ static inline int verify_replay(struct xfrm_usersa_info *p,
 	if (!rt)
 		return 0;
 
-	if (p->id.proto != IPPROTO_ESP)
+	/* As only ESP and AH support ESN feature. */
+	if ((p->id.proto != IPPROTO_ESP) && (p->id.proto != IPPROTO_AH))
 		return -EINVAL;
 
 	if (p->replay_window != 0)
-- 
1.7.9.5

Re: [PATCHv4 net-next 3/6] {IPv4,xfrm} Add ESN support for AH ingress part

[Steffen Klassert]

On Wed, Jan 15, 2014 at 03:13:12PM +0800, Fan Du wrote:
>  
> -	ahash_request_set_crypt(req, sg, icv, skb->len);
> +	if (x->props.flags & XFRM_STATE_ESN) {
> +		/* Attach seqhi sg right after packet payload */
> +		*seqhi = htonl(XFRM_SKB_CB(skb)->seq.input.hi);

XFRM_SKB_CB(skb)->seq.input.hi is already in network byte order,
please remove the htonl(). The ipv6 patch has the same problem.