Skip to contentSkip to conversation

pull request (net-next): ipsec-next 2021-06-28

PATCH Applied Steffen Klassert 2 participants 20 messages Jun 28, 2021

Subsystems (4)
lore
Share & tools
lore

Steffen Klassert

author maintainer
#0 lore raw

1) Remove an unneeded error assignment in esp4_gro_receive(). From Yang Li.

2) Add a new byseq state hashtable to find acquire states faster. From Sabrina Dubroca.

3) Remove some unnecessary variables in pfkey_create(). From zuoqilin.

4) Remove the unused description from xfrm_type struct. From Florian Westphal.

5) Fix a spelling mistake in the comment of xfrm_state_ok(). From gushengxian.

6) Replace hdr_off indirections by a small helper function. From Florian Westphal.

7) Remove xfrm4_output_finish and xfrm6_output_finish declarations, they are not used anymore.From Antony Antony.

8) Remove xfrm replay indirections.
From Florian Westphal.

Please pull or let me know if there are problems.

Thanks!

The following changes since commit ea89c862f01e02ec459932c7c3113fa37aedd09a:

net: mana: Use struct_size() in kzalloc() (2021-05-13 15:58:46 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next.git master

for you to fetch changes up to b5a1d1fe0cbb9d20ba661134a09561af1dc9ebf5:

xfrm: replay: remove last replay indirection (2021-06-21 09:55:06 +0200)

----------------------------------------------------------------
Antony Antony (1):
xfrm: delete xfrm4_output_finish xfrm6_output_finish declarations

Florian Westphal (12):
xfrm: remove description from xfrm_type struct
xfrm: ipv6: add xfrm6_hdr_offset helper
xfrm: ipv6: move mip6_destopt_offset into xfrm core
xfrm: ipv6: move mip6_rthdr_offset into xfrm core
xfrm: remove hdr_offset indirection
xfrm: merge dstopt and routing hdroff functions
xfrm: avoid compiler warning when ipv6 is disabled
xfrm: replay: avoid xfrm replay notify indirection
xfrm: replay: remove advance indirection
xfrm: replay: remove recheck indirection
xfrm: replay: avoid replay indirection
xfrm: replay: remove last replay indirection

Sabrina Dubroca (1):
xfrm: add state hashtable keyed by seq

Yang Li (1):
esp: drop unneeded assignment in esp4_gro_receive()

gushengxian (1):
xfrm: policy: fix a spelling mistake

zuoqilin (1):
net: Remove unnecessary variables

 include/net/netns/xfrm.h |   1 +
 include/net/xfrm.h       |  37 +++++-----
 net/ipv4/ah4.c           |   1 -
 net/ipv4/esp4.c          |   1 -
 net/ipv4/esp4_offload.c  |   4 +-
 net/ipv4/ipcomp.c        |   1 -
 net/ipv4/xfrm4_tunnel.c  |   1 -
 net/ipv6/ah6.c           |   2 -
 net/ipv6/esp6.c          |   2 -
 net/ipv6/esp6_offload.c  |   1 -
 net/ipv6/ipcomp6.c       |   2 -
 net/ipv6/mip6.c          |  99 ---------------------------
 net/ipv6/xfrm6_output.c  |   7 --
 net/ipv6/xfrm6_tunnel.c  |   1 -
 net/key/af_key.c         |   6 +-
 net/xfrm/xfrm_hash.h     |   7 ++
 net/xfrm/xfrm_input.c    |   6 +-
 net/xfrm/xfrm_output.c   |  83 ++++++++++++++++++++++-
 net/xfrm/xfrm_policy.c   |   2 +-
 net/xfrm/xfrm_replay.c   | 171 ++++++++++++++++++++++++++++-------------------
 net/xfrm/xfrm_state.c    |  67 +++++++++++++++----
 21 files changed, 266 insertions(+), 236 deletions(-)
author maintainer
#1 lore raw

PATCH 1/17 esp: drop unneeded assignment in esp4_gro_receive()

From: Yang Li <[email protected]>

Making '!=' operation with 0 directly after calling the function xfrm_parse_spi() is more efficient, assignment to err is redundant.

Eliminate the following clang_analyzer warning:
net/ipv4/esp4_offload.c:41:7: warning: Although the value stored to 'err' is used in the enclosing expression, the value is never actually read from 'err'

No functional change, only more efficient.

---
 net/ipv4/esp4_offload.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
net/ipv4/esp4_offload.c +1 -2
index 33687cf58286..be019a1fe3af 100644
@@ -33,12 +33,11 @@ static struct sk_buff *esp4_gro_receive(struct list_head *head,
 	struct xfrm_state *x;
 	__be32 seq;
 	__be32 spi;
-	int err;
 
 	if (!pskb_pull(skb, offset))
 		return NULL;
 
-	if ((err = xfrm_parse_spi(skb, IPPROTO_ESP, &spi, &seq)) != 0)
+	if (xfrm_parse_spi(skb, IPPROTO_ESP, &spi, &seq) != 0)
 		goto out;
 
 	xo = xfrm_offload(skb);
Signature · 1 line
2.25.1
#2 lore raw

Re: [PATCH 01/17] esp: drop unneeded assignment in esp4_gro_receive()

This series was applied to netdev/net-next.git (refs/heads/master):

Hello:

This series was applied to netdev/net-next.git (refs/heads/master):

Quoted 6 lines · On Mon, 28 Jun 2021 07:45:06 +0200 you wrote:

From: Yang Li <[email protected]>

Making '!=' operation with 0 directly after calling the function xfrm_parse_spi() is more efficient, assignment to err is redundant.

Quoted 6 lines

Eliminate the following clang_analyzer warning:
net/ipv4/esp4_offload.c:41:7: warning: Although the value stored to 'err' is used in the enclosing expression, the value is never actually read from 'err'

[...]

Here is the summary with links:
- [01/17] esp: drop unneeded assignment in esp4_gro_receive()

    https://git.kernel.org/netdev/net-next/c/335a2a1fcefc
  - [02/17] xfrm: add state hashtable keyed by seq
    https://git.kernel.org/netdev/net-next/c/fe9f1d8779cb
  - [03/17] net: Remove unnecessary variables
    https://git.kernel.org/netdev/net-next/c/a925316af80a
  - [04/17] xfrm: remove description from xfrm_type struct
    https://git.kernel.org/netdev/net-next/c/152bca090243
  - [05/17] xfrm: policy: fix a spelling mistake
    https://git.kernel.org/netdev/net-next/c/7a7ae1eba24a
  - [06/17] xfrm: ipv6: add xfrm6_hdr_offset helper
    https://git.kernel.org/netdev/net-next/c/9acf4d3b9ec1
  - [07/17] xfrm: ipv6: move mip6_destopt_offset into xfrm core
    https://git.kernel.org/netdev/net-next/c/37b9e7eb5565
  - [08/17] xfrm: ipv6: move mip6_rthdr_offset into xfrm core
    https://git.kernel.org/netdev/net-next/c/848b18fb7fbd
  - [09/17] xfrm: remove hdr_offset indirection
    https://git.kernel.org/netdev/net-next/c/d1002d2490e3
  - [10/17] xfrm: merge dstopt and routing hdroff functions
    https://git.kernel.org/netdev/net-next/c/3ca5ca83e206
  - [11/17] xfrm: delete xfrm4_output_finish xfrm6_output_finish declarations
    https://git.kernel.org/netdev/net-next/c/1b50dd478f49
  - [12/17] xfrm: avoid compiler warning when ipv6 is disabled
    https://git.kernel.org/netdev/net-next/c/30ad6a84f60b
  - [13/17] xfrm: replay: avoid xfrm replay notify indirection
    https://git.kernel.org/netdev/net-next/c/cfc61c598e43
  - [14/17] xfrm: replay: remove advance indirection
    https://git.kernel.org/netdev/net-next/c/c7f877833c9f
  - [15/17] xfrm: replay: remove recheck indirection
    https://git.kernel.org/netdev/net-next/c/25cfb8bc97c2
  - [16/17] xfrm: replay: avoid replay indirection
    https://git.kernel.org/netdev/net-next/c/adfc2fdbae30
  - [17/17] xfrm: replay: remove last replay indirection
    https://git.kernel.org/netdev/net-next/c/b5a1d1fe0cbb

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

author maintainer
#3 lore raw

PATCH 2/17 xfrm: add state hashtable keyed by seq

In reply to Steffen Klassert

From: Sabrina Dubroca <[email protected]>

When creating new states with seq set in xfrm_usersa_info, we walk through all the states already installed in that netns to find a matching ACQUIRE state (__xfrm_find_acq_byseq, called from xfrm_state_add). This causes severe slowdowns on systems with a large number of states.

This patch introduces a hashtable using x->km.seq as key, so that the corresponding state can be found in a reasonable time.

---
 include/net/netns/xfrm.h |  1 +
 include/net/xfrm.h       |  1 +
 net/xfrm/xfrm_hash.h     |  7 +++++
 net/xfrm/xfrm_state.c    | 65 ++++++++++++++++++++++++++++++++--------
 4 files changed, 61 insertions(+), 13 deletions(-)
include/net/netns/xfrm.h +1 -0
index e816b6a3ef2b..e946366e8ba5 100644
@@ -42,6 +42,7 @@ struct netns_xfrm {
 	struct hlist_head	__rcu *state_bydst;
 	struct hlist_head	__rcu *state_bysrc;
 	struct hlist_head	__rcu *state_byspi;
+	struct hlist_head	__rcu *state_byseq;
 	unsigned int		state_hmask;
 	unsigned int		state_num;
 	struct work_struct	state_hash_work;
include/net/xfrm.h +1 -0
index c58a6d4eb610..6e11db6fa0ab 100644
@@ -154,6 +154,7 @@ struct xfrm_state {
 	};
 	struct hlist_node	bysrc;
 	struct hlist_node	byspi;
+	struct hlist_node	byseq;
 
 	refcount_t		refcnt;
 	spinlock_t		lock;
net/xfrm/xfrm_hash.h +7 -0
index ce66323102f9..d12bb906c9c9 100644
@@ -131,6 +131,13 @@ __xfrm_spi_hash(const xfrm_address_t *daddr, __be32 spi, u8 proto,
 	return (h ^ (h >> 10) ^ (h >> 20)) & hmask;
 }
 
+static inline unsigned int
+__xfrm_seq_hash(u32 seq, unsigned int hmask)
+{
+	unsigned int h = seq;
+	return (h ^ (h >> 10) ^ (h >> 20)) & hmask;
+}
+
 static inline unsigned int __idx_hash(u32 index, unsigned int hmask)
 {
 	return (index ^ (index >> 8)) & hmask;
net/xfrm/xfrm_state.c +52 -13

164 lines not on this page · raw message

Signature · 1 line
2.25.1
author maintainer
#4 lore raw

PATCH 3/17 net: Remove unnecessary variables

In reply to Steffen Klassert

From: zuoqilin <[email protected]>

It is not necessary to define variables to receive -ENOMEM, directly return -ENOMEM.

---
 net/key/af_key.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
net/key/af_key.c +1 -5
index ef9b4ac03e7b..de24a7d474df 100644
@@ -141,7 +141,6 @@ static int pfkey_create(struct net *net, struct socket *sock, int protocol,
 	struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
 	struct sock *sk;
 	struct pfkey_sock *pfk;
-	int err;
 
 	if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
 		return -EPERM;
@@ -150,10 +149,9 @@ static int pfkey_create(struct net *net, struct socket *sock, int protocol,
 	if (protocol != PF_KEY_V2)
 		return -EPROTONOSUPPORT;
 
-	err = -ENOMEM;
 	sk = sk_alloc(net, PF_KEY, GFP_KERNEL, &key_proto, kern);
 	if (sk == NULL)
-		goto out;
+		return -ENOMEM;
 
 	pfk = pfkey_sk(sk);
 	mutex_init(&pfk->dump_lock);
@@ -169,8 +167,6 @@ static int pfkey_create(struct net *net, struct socket *sock, int protocol,
 	pfkey_insert(sk);
 
 	return 0;
-out:
-	return err;
 }
 
 static int pfkey_release(struct socket *sock)
Signature · 1 line
2.25.1
author maintainer
#5 lore raw

PATCH 4/17 xfrm: remove description from xfrm_type struct

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

Its set but never read. Reduces size of xfrm_type to 64 bytes on 64bit.

---
 include/net/xfrm.h      | 2 --
 net/ipv4/ah4.c          | 1 -
 net/ipv4/esp4.c         | 1 -
 net/ipv4/esp4_offload.c | 1 -
 net/ipv4/ipcomp.c       | 1 -
 net/ipv4/xfrm4_tunnel.c | 1 -
 net/ipv6/ah6.c          | 1 -
 net/ipv6/esp6.c         | 1 -
 net/ipv6/esp6_offload.c | 1 -
 net/ipv6/ipcomp6.c      | 1 -
 net/ipv6/mip6.c         | 2 --
 net/ipv6/xfrm6_tunnel.c | 1 -
 12 files changed, 14 deletions(-)
include/net/xfrm.h +0 -2
index 6e11db6fa0ab..1aad78c5f2d5 100644
@@ -388,7 +388,6 @@ void xfrm_flush_gc(void);
 void xfrm_state_delete_tunnel(struct xfrm_state *x);
 
 struct xfrm_type {
-	char			*description;
 	struct module		*owner;
 	u8			proto;
 	u8			flags;
@@ -410,7 +409,6 @@ int xfrm_register_type(const struct xfrm_type *type, unsigned short family);
 void xfrm_unregister_type(const struct xfrm_type *type, unsigned short family);
 
 struct xfrm_type_offload {
-	char		*description;
 	struct module	*owner;
 	u8		proto;
 	void		(*encap)(struct xfrm_state *, struct sk_buff *pskb);
net/ipv4/ah4.c +0 -1
index 36ed85bf2ad5..2d2d08aa787d 100644
@@ -554,7 +554,6 @@ static int ah4_rcv_cb(struct sk_buff *skb, int err)
 
 static const struct xfrm_type ah_type =
 {
-	.description	= "AH4",
 	.owner		= THIS_MODULE,
 	.proto	     	= IPPROTO_AH,
 	.flags		= XFRM_TYPE_REPLAY_PROT,
net/ipv4/esp4.c +0 -1
index 35803ab7ac80..f5362b9d75eb 100644
@@ -1198,7 +1198,6 @@ static int esp4_rcv_cb(struct sk_buff *skb, int err)
 
 static const struct xfrm_type esp_type =
 {
-	.description	= "ESP4",
 	.owner		= THIS_MODULE,
 	.proto	     	= IPPROTO_ESP,
 	.flags		= XFRM_TYPE_REPLAY_PROT,
net/ipv4/esp4_offload.c +0 -1
index be019a1fe3af..8e4e9aa12130 100644
@@ -342,7 +342,6 @@ static const struct net_offload esp4_offload = {
 };
 
 static const struct xfrm_type_offload esp_type_offload = {
-	.description	= "ESP4 OFFLOAD",
 	.owner		= THIS_MODULE,
 	.proto	     	= IPPROTO_ESP,
 	.input_tail	= esp_input_tail,
net/ipv4/ipcomp.c +0 -1
index b42683212c65..2e69e81e1f5d 100644
@@ -152,7 +152,6 @@ static int ipcomp4_rcv_cb(struct sk_buff *skb, int err)
 }
 
 static const struct xfrm_type ipcomp_type = {
-	.description	= "IPCOMP4",
 	.owner		= THIS_MODULE,
 	.proto	     	= IPPROTO_COMP,
 	.init_state	= ipcomp4_init_state,
net/ipv4/xfrm4_tunnel.c +0 -1
index fb0648e7fb32..f4555a88f86b 100644
@@ -42,7 +42,6 @@ static void ipip_destroy(struct xfrm_state *x)
 }
 
 static const struct xfrm_type ipip_type = {
-	.description	= "IPIP",
 	.owner		= THIS_MODULE,
 	.proto	     	= IPPROTO_IPIP,
 	.init_state	= ipip_init_state,
net/ipv6/ah6.c +0 -1
index 20d492da725a..e9705c256068 100644
@@ -755,7 +755,6 @@ static int ah6_rcv_cb(struct sk_buff *skb, int err)
 }
 
 static const struct xfrm_type ah6_type = {
-	.description	= "AH6",
 	.owner		= THIS_MODULE,
 	.proto		= IPPROTO_AH,
 	.flags		= XFRM_TYPE_REPLAY_PROT,
net/ipv6/esp6.c +0 -1
index 393ae2b78e7d..be2c0ac76eaa 100644
@@ -1243,7 +1243,6 @@ static int esp6_rcv_cb(struct sk_buff *skb, int err)
 }
 
 static const struct xfrm_type esp6_type = {
-	.description	= "ESP6",
 	.owner		= THIS_MODULE,
 	.proto		= IPPROTO_ESP,
 	.flags		= XFRM_TYPE_REPLAY_PROT,
net/ipv6/esp6_offload.c +0 -1
index 40ed4fcf1cf4..a349d4798077 100644
@@ -377,7 +377,6 @@ static const struct net_offload esp6_offload = {
 };
 
 static const struct xfrm_type_offload esp6_type_offload = {
-	.description	= "ESP6 OFFLOAD",
 	.owner		= THIS_MODULE,
 	.proto	     	= IPPROTO_ESP,
 	.input_tail	= esp6_input_tail,
net/ipv6/ipcomp6.c +0 -1
index daef890460b7..491aba66b7ae 100644
@@ -172,7 +172,6 @@ static int ipcomp6_rcv_cb(struct sk_buff *skb, int err)
 }
 
 static const struct xfrm_type ipcomp6_type = {
-	.description	= "IPCOMP6",
 	.owner		= THIS_MODULE,
 	.proto		= IPPROTO_COMP,
 	.init_state	= ipcomp6_init_state,
net/ipv6/mip6.c +0 -2
index 878fcec14949..bc560e1664aa 100644
@@ -324,7 +324,6 @@ static void mip6_destopt_destroy(struct xfrm_state *x)
 }
 
 static const struct xfrm_type mip6_destopt_type = {
-	.description	= "MIP6DESTOPT",
 	.owner		= THIS_MODULE,
 	.proto		= IPPROTO_DSTOPTS,
 	.flags		= XFRM_TYPE_NON_FRAGMENT | XFRM_TYPE_LOCAL_COADDR,
@@ -456,7 +455,6 @@ static void mip6_rthdr_destroy(struct xfrm_state *x)
 }
 
 static const struct xfrm_type mip6_rthdr_type = {
-	.description	= "MIP6RT",
 	.owner		= THIS_MODULE,
 	.proto		= IPPROTO_ROUTING,
 	.flags		= XFRM_TYPE_NON_FRAGMENT | XFRM_TYPE_REMOTE_COADDR,
net/ipv6/xfrm6_tunnel.c +0 -1
index f696d46e6910..2b31112c0856 100644
@@ -291,7 +291,6 @@ static void xfrm6_tunnel_destroy(struct xfrm_state *x)
 }
 
 static const struct xfrm_type xfrm6_tunnel_type = {
-	.description	= "IP6IP6",
 	.owner          = THIS_MODULE,
 	.proto		= IPPROTO_IPV6,
 	.init_state	= xfrm6_tunnel_init_state,
Signature · 1 line
2.25.1
author maintainer
#6 lore raw

PATCH 5/17 xfrm: policy: fix a spelling mistake

In reply to Steffen Klassert

From: gushengxian <[email protected]>

Fix a spelling mistake.

---
 net/xfrm/xfrm_policy.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/xfrm/xfrm_policy.c +1 -1
index ce500f847b99..1e24b21457f7 100644
@@ -3247,7 +3247,7 @@ xfrm_state_ok(const struct xfrm_tmpl *tmpl, const struct xfrm_state *x,
 
 /*
  * 0 or more than 0 is returned when validation is succeeded (either bypass
- * because of optional transport mode, or next index of the mathced secpath
+ * because of optional transport mode, or next index of the matched secpath
  * state with the template.
  * -1 is returned when no matching template is found.
  * Otherwise "-2 - errored_index" is returned.
Signature · 1 line
2.25.1
author maintainer
#7 lore raw

PATCH 6/17 xfrm: ipv6: add xfrm6_hdr_offset helper

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

This moves the ->hdr_offset indirect call to a new helper.

A followup patch can then modify the new function to replace the indirect call by direct calls to the required hdr_offset helper.

---
 net/xfrm/xfrm_output.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
net/xfrm/xfrm_output.c +7 -2
index e4cb0ff4dcf4..6b44b6e738f7 100644
@@ -77,6 +77,11 @@ static int xfrm4_transport_output(struct xfrm_state *x, struct sk_buff *skb)
 	return 0;
 }
 
+static int xfrm6_hdr_offset(struct xfrm_state *x, struct sk_buff *skb, u8 **prevhdr)
+{
+	return x->type->hdr_offset(x, skb, prevhdr);
+}
+
 /* Add encapsulation header.
  *
  * The IP header and mutable extension headers will be moved forward to make
@@ -92,7 +97,7 @@ static int xfrm6_transport_output(struct xfrm_state *x, struct sk_buff *skb)
 	iph = ipv6_hdr(skb);
 	skb_set_inner_transport_header(skb, skb_transport_offset(skb));
 
-	hdr_len = x->type->hdr_offset(x, skb, &prevhdr);
+	hdr_len = xfrm6_hdr_offset(x, skb, &prevhdr);
 	if (hdr_len < 0)
 		return hdr_len;
 	skb_set_mac_header(skb,
@@ -122,7 +127,7 @@ static int xfrm6_ro_output(struct xfrm_state *x, struct sk_buff *skb)
 
 	iph = ipv6_hdr(skb);
 
-	hdr_len = x->type->hdr_offset(x, skb, &prevhdr);
+	hdr_len = xfrm6_hdr_offset(x, skb, &prevhdr);
 	if (hdr_len < 0)
 		return hdr_len;
 	skb_set_mac_header(skb,
Signature · 1 line
2.25.1
author maintainer
#8 lore raw

PATCH 7/17 xfrm: ipv6: move mip6_destopt_offset into xfrm core

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

This helper is relatively small, just move this to the xfrm core and call it directly.

Next patch does the same for the ROUTING type.

---
 net/ipv6/mip6.c        | 49 ------------------------------------
 net/xfrm/xfrm_output.c | 57 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 57 insertions(+), 49 deletions(-)
net/ipv6/mip6.c +0 -49
index bc560e1664aa..fba3b56a7dd2 100644
@@ -247,54 +247,6 @@ static int mip6_destopt_reject(struct xfrm_state *x, struct sk_buff *skb,
 	return err;
 }
 
-static int mip6_destopt_offset(struct xfrm_state *x, struct sk_buff *skb,
-			       u8 **nexthdr)
-{
-	u16 offset = sizeof(struct ipv6hdr);
-	struct ipv6_opt_hdr *exthdr =
-				   (struct ipv6_opt_hdr *)(ipv6_hdr(skb) + 1);
-	const unsigned char *nh = skb_network_header(skb);
-	unsigned int packet_len = skb_tail_pointer(skb) -
-		skb_network_header(skb);
-	int found_rhdr = 0;
-
-	*nexthdr = &ipv6_hdr(skb)->nexthdr;
-
-	while (offset + 1 <= packet_len) {
-
-		switch (**nexthdr) {
-		case NEXTHDR_HOP:
-			break;
-		case NEXTHDR_ROUTING:
-			found_rhdr = 1;
-			break;
-		case NEXTHDR_DEST:
-			/*
-			 * HAO MUST NOT appear more than once.
-			 * XXX: It is better to try to find by the end of
-			 * XXX: packet if HAO exists.
-			 */
-			if (ipv6_find_tlv(skb, offset, IPV6_TLV_HAO) >= 0) {
-				net_dbg_ratelimited("mip6: hao exists already, override\n");
-				return offset;
-			}
-
-			if (found_rhdr)
-				return offset;
-
-			break;
-		default:
-			return offset;
-		}
-
-		offset += ipv6_optlen(exthdr);
-		*nexthdr = &exthdr->nexthdr;
-		exthdr = (struct ipv6_opt_hdr *)(nh + offset);
-	}
-
-	return offset;
-}
-
 static int mip6_destopt_init_state(struct xfrm_state *x)
 {
 	if (x->id.spi) {
@@ -332,7 +284,6 @@ static const struct xfrm_type mip6_destopt_type = {
 	.input		= mip6_destopt_input,
 	.output		= mip6_destopt_output,
 	.reject		= mip6_destopt_reject,
-	.hdr_offset	= mip6_destopt_offset,
 };
 
 static int mip6_rthdr_input(struct xfrm_state *x, struct sk_buff *skb)
net/xfrm/xfrm_output.c +57 -0

65 lines not on this page · raw message

Signature · 1 line
2.25.1
author maintainer
#9 lore raw

PATCH 8/17 xfrm: ipv6: move mip6_rthdr_offset into xfrm core

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

Place the call into the xfrm core. After this all remaining users set the hdr_offset function pointer to the same function which opens the possiblity to remove the indirection.

---
 net/ipv6/mip6.c        | 48 -----------------------------------------
 net/xfrm/xfrm_output.c | 49 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 49 insertions(+), 48 deletions(-)
net/ipv6/mip6.c +0 -48
index fba3b56a7dd2..aeb35d26e474 100644
@@ -333,53 +333,6 @@ static int mip6_rthdr_output(struct xfrm_state *x, struct sk_buff *skb)
 	return 0;
 }
 
-static int mip6_rthdr_offset(struct xfrm_state *x, struct sk_buff *skb,
-			     u8 **nexthdr)
-{
-	u16 offset = sizeof(struct ipv6hdr);
-	struct ipv6_opt_hdr *exthdr =
-				   (struct ipv6_opt_hdr *)(ipv6_hdr(skb) + 1);
-	const unsigned char *nh = skb_network_header(skb);
-	unsigned int packet_len = skb_tail_pointer(skb) -
-		skb_network_header(skb);
-	int found_rhdr = 0;
-
-	*nexthdr = &ipv6_hdr(skb)->nexthdr;
-
-	while (offset + 1 <= packet_len) {
-
-		switch (**nexthdr) {
-		case NEXTHDR_HOP:
-			break;
-		case NEXTHDR_ROUTING:
-			if (offset + 3 <= packet_len) {
-				struct ipv6_rt_hdr *rt;
-				rt = (struct ipv6_rt_hdr *)(nh + offset);
-				if (rt->type != 0)
-					return offset;
-			}
-			found_rhdr = 1;
-			break;
-		case NEXTHDR_DEST:
-			if (ipv6_find_tlv(skb, offset, IPV6_TLV_HAO) >= 0)
-				return offset;
-
-			if (found_rhdr)
-				return offset;
-
-			break;
-		default:
-			return offset;
-		}
-
-		offset += ipv6_optlen(exthdr);
-		*nexthdr = &exthdr->nexthdr;
-		exthdr = (struct ipv6_opt_hdr *)(nh + offset);
-	}
-
-	return offset;
-}
-
 static int mip6_rthdr_init_state(struct xfrm_state *x)
 {
 	if (x->id.spi) {
@@ -413,7 +366,6 @@ static const struct xfrm_type mip6_rthdr_type = {
 	.destructor	= mip6_rthdr_destroy,
 	.input		= mip6_rthdr_input,
 	.output		= mip6_rthdr_output,
-	.hdr_offset	= mip6_rthdr_offset,
 };
 
 static int __init mip6_init(void)
net/xfrm/xfrm_output.c +49 -0

61 lines not on this page · raw message

Signature · 1 line
2.25.1
author maintainer
#10 lore raw

PATCH 9/17 xfrm: remove hdr_offset indirection

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

After previous patches all remaining users set the function pointer to the same function: xfrm6_find_1stfragopt.

So remove this function pointer and call ip6_find_1stfragopt directly.

Reduces size of xfrm_type to 64 bytes on 64bit platforms.

---
 include/net/xfrm.h      | 3 ---
 net/ipv6/ah6.c          | 1 -
 net/ipv6/esp6.c         | 1 -
 net/ipv6/ipcomp6.c      | 1 -
 net/ipv6/xfrm6_output.c | 7 -------
 net/xfrm/xfrm_output.c  | 2 +-
 6 files changed, 1 insertion(+), 14 deletions(-)
include/net/xfrm.h +0 -3
index 1aad78c5f2d5..c8890da00b8a 100644
@@ -402,7 +402,6 @@ struct xfrm_type {
 	int			(*output)(struct xfrm_state *, struct sk_buff *pskb);
 	int			(*reject)(struct xfrm_state *, struct sk_buff *,
 					  const struct flowi *);
-	int			(*hdr_offset)(struct xfrm_state *, struct sk_buff *, u8 **);
 };
 
 int xfrm_register_type(const struct xfrm_type *type, unsigned short family);
@@ -1605,8 +1604,6 @@ __be32 xfrm6_tunnel_alloc_spi(struct net *net, xfrm_address_t *saddr);
 __be32 xfrm6_tunnel_spi_lookup(struct net *net, const xfrm_address_t *saddr);
 int xfrm6_output(struct net *net, struct sock *sk, struct sk_buff *skb);
 int xfrm6_output_finish(struct sock *sk, struct sk_buff *skb);
-int xfrm6_find_1stfragopt(struct xfrm_state *x, struct sk_buff *skb,
-			  u8 **prevhdr);
 
 #ifdef CONFIG_XFRM
 void xfrm6_local_rxpmtu(struct sk_buff *skb, u32 mtu);
net/ipv6/ah6.c +0 -1
index e9705c256068..828e62514260 100644
@@ -762,7 +762,6 @@ static const struct xfrm_type ah6_type = {
 	.destructor	= ah6_destroy,
 	.input		= ah6_input,
 	.output		= ah6_output,
-	.hdr_offset	= xfrm6_find_1stfragopt,
 };
 
 static struct xfrm6_protocol ah6_protocol = {
net/ipv6/esp6.c +0 -1
index be2c0ac76eaa..37c4b1726c5e 100644
@@ -1250,7 +1250,6 @@ static const struct xfrm_type esp6_type = {
 	.destructor	= esp6_destroy,
 	.input		= esp6_input,
 	.output		= esp6_output,
-	.hdr_offset	= xfrm6_find_1stfragopt,
 };
 
 static struct xfrm6_protocol esp6_protocol = {
net/ipv6/ipcomp6.c +0 -1
index 491aba66b7ae..15f984be3570 100644
@@ -178,7 +178,6 @@ static const struct xfrm_type ipcomp6_type = {
 	.destructor	= ipcomp_destroy,
 	.input		= ipcomp_input,
 	.output		= ipcomp_output,
-	.hdr_offset	= xfrm6_find_1stfragopt,
 };
 
 static struct xfrm6_protocol ipcomp6_protocol = {
net/ipv6/xfrm6_output.c +0 -7
index 8b84d534b19d..57fa27c1cdf9 100644
@@ -16,13 +16,6 @@
 #include <net/ip6_route.h>
 #include <net/xfrm.h>
 
-int xfrm6_find_1stfragopt(struct xfrm_state *x, struct sk_buff *skb,
-			  u8 **prevhdr)
-{
-	return ip6_find_1stfragopt(skb, prevhdr);
-}
-EXPORT_SYMBOL(xfrm6_find_1stfragopt);
-
 void xfrm6_local_rxpmtu(struct sk_buff *skb, u32 mtu)
 {
 	struct flowi6 fl6;
net/xfrm/xfrm_output.c +1 -1
index 1734339b6dd0..10842d5cf6e1 100644
@@ -185,7 +185,7 @@ static int xfrm6_hdr_offset(struct xfrm_state *x, struct sk_buff *skb, u8 **prev
 		break;
 	}
 
-	return x->type->hdr_offset(x, skb, prevhdr);
+	return ip6_find_1stfragopt(skb, prevhdr);
 }
 
 /* Add encapsulation header.
Signature · 1 line
2.25.1
author maintainer
#11 lore raw

PATCH 10/17 xfrm: merge dstopt and routing hdroff functions

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

Both functions are very similar, so merge them into one.

The nexthdr is passed as argument to break the loop in the ROUTING case, this is the only header type where slightly different rules apply.

While at it, the merged function is realigned with ip6_find_1stfragopt(). That function received bug fixes for an infinite loop, but neither dstopt nor rh parsing functions (copy-pasted from ip6_find_1stfragopt) were changed.

---
 net/xfrm/xfrm_output.c | 80 ++++++++++++------------------------------
 1 file changed, 22 insertions(+), 58 deletions(-)
net/xfrm/xfrm_output.c +22 -58
index 10842d5cf6e1..e14fca1fb003 100644
@@ -78,24 +78,30 @@ static int xfrm4_transport_output(struct xfrm_state *x, struct sk_buff *skb)
 }
 
 #if IS_ENABLED(CONFIG_IPV6_MIP6)
-static int mip6_destopt_offset(struct xfrm_state *x, struct sk_buff *skb,
-			       u8 **nexthdr)
+static int mip6_rthdr_offset(struct sk_buff *skb, u8 **nexthdr, int type)
 {
-	u16 offset = sizeof(struct ipv6hdr);
-	struct ipv6_opt_hdr *exthdr =
-				   (struct ipv6_opt_hdr *)(ipv6_hdr(skb) + 1);
 	const unsigned char *nh = skb_network_header(skb);
-	unsigned int packet_len = skb_tail_pointer(skb) -
-		skb_network_header(skb);
+	unsigned int offset = sizeof(struct ipv6hdr);
+	unsigned int packet_len;
 	int found_rhdr = 0;
 
+	packet_len = skb_tail_pointer(skb) - nh;
 	*nexthdr = &ipv6_hdr(skb)->nexthdr;
 
-	while (offset + 1 <= packet_len) {
+	while (offset <= packet_len) {
+		struct ipv6_opt_hdr *exthdr;
+
 		switch (**nexthdr) {
 		case NEXTHDR_HOP:
 			break;
 		case NEXTHDR_ROUTING:
+			if (type == IPPROTO_ROUTING && offset + 3 <= packet_len) {
+				struct ipv6_rt_hdr *rt;
+
+				rt = (struct ipv6_rt_hdr *)(nh + offset);
+				if (rt->type != 0)
+					return offset;
+			}
 			found_rhdr = 1;
 			break;
 		case NEXTHDR_DEST:
@@ -116,59 +122,18 @@ static int mip6_destopt_offset(struct xfrm_state *x, struct sk_buff *skb,
 			return offset;
 		}
 
-		offset += ipv6_optlen(exthdr);
-		*nexthdr = &exthdr->nexthdr;
-		exthdr = (struct ipv6_opt_hdr *)(nh + offset);
-	}
-
-	return offset;
-}
-
-static int mip6_rthdr_offset(struct xfrm_state *x, struct sk_buff *skb,
-			     u8 **nexthdr)
-{
-	u16 offset = sizeof(struct ipv6hdr);
-	struct ipv6_opt_hdr *exthdr =
-				   (struct ipv6_opt_hdr *)(ipv6_hdr(skb) + 1);
-	const unsigned char *nh = skb_network_header(skb);
-	unsigned int packet_len = skb_tail_pointer(skb) -
-		skb_network_header(skb);
-	int found_rhdr = 0;
-
-	*nexthdr = &ipv6_hdr(skb)->nexthdr;
-
-	while (offset + 1 <= packet_len) {
-		switch (**nexthdr) {
-		case NEXTHDR_HOP:
-			break;
-		case NEXTHDR_ROUTING:
-			if (offset + 3 <= packet_len) {
-				struct ipv6_rt_hdr *rt;
-
-				rt = (struct ipv6_rt_hdr *)(nh + offset);
-				if (rt->type != 0)
-					return offset;
-			}
-			found_rhdr = 1;
-			break;
-		case NEXTHDR_DEST:
-			if (ipv6_find_tlv(skb, offset, IPV6_TLV_HAO) >= 0)
-				return offset;
-
-			if (found_rhdr)
-				return offset;
-
-			break;
-		default:
-			return offset;
-		}
+		if (offset + sizeof(struct ipv6_opt_hdr) > packet_len)
+			return -EINVAL;
 
+		exthdr = (struct ipv6_opt_hdr *)(skb_network_header(skb) +
+						 offset);
 		offset += ipv6_optlen(exthdr);
+		if (offset > IPV6_MAXPLEN)
+			return -EINVAL;
 		*nexthdr = &exthdr->nexthdr;
-		exthdr = (struct ipv6_opt_hdr *)(nh + offset);
 	}
 
-	return offset;
+	return -EINVAL;
 }
 #endif
 
@@ -177,9 +142,8 @@ static int xfrm6_hdr_offset(struct xfrm_state *x, struct sk_buff *skb, u8 **prev
 	switch (x->type->proto) {
 #if IS_ENABLED(CONFIG_IPV6_MIP6)
 	case IPPROTO_DSTOPTS:
-		return mip6_destopt_offset(x, skb, prevhdr);
 	case IPPROTO_ROUTING:
-		return mip6_rthdr_offset(x, skb, prevhdr);
+		return mip6_rthdr_offset(skb, prevhdr, x->type->proto);
 #endif
 	default:
 		break;
Signature · 1 line
2.25.1
author maintainer
#12 lore raw

PATCH 11/17 xfrm: delete xfrm4_output_finish xfrm6_output_finish declarations

In reply to Steffen Klassert

From: Antony Antony <[email protected]>

These function declarations are not needed any more. The definitions were deleted.

Fixes 2ab6096db2f1 ("xfrm: remove output_finish indirection from xfrm_state_afinfo")
---
 include/net/xfrm.h | 2 --
 1 file changed, 2 deletions(-)
include/net/xfrm.h +0 -2
index c8890da00b8a..3a01570410ab 100644
@@ -1579,7 +1579,6 @@ static inline int xfrm4_rcv_spi(struct sk_buff *skb, int nexthdr, __be32 spi)
 }
 
 int xfrm4_output(struct net *net, struct sock *sk, struct sk_buff *skb);
-int xfrm4_output_finish(struct sock *sk, struct sk_buff *skb);
 int xfrm4_protocol_register(struct xfrm4_protocol *handler, unsigned char protocol);
 int xfrm4_protocol_deregister(struct xfrm4_protocol *handler, unsigned char protocol);
 int xfrm4_tunnel_register(struct xfrm_tunnel *handler, unsigned short family);
@@ -1603,7 +1602,6 @@ int xfrm6_tunnel_deregister(struct xfrm6_tunnel *handler, unsigned short family)
 __be32 xfrm6_tunnel_alloc_spi(struct net *net, xfrm_address_t *saddr);
 __be32 xfrm6_tunnel_spi_lookup(struct net *net, const xfrm_address_t *saddr);
 int xfrm6_output(struct net *net, struct sock *sk, struct sk_buff *skb);
-int xfrm6_output_finish(struct sock *sk, struct sk_buff *skb);
 
 #ifdef CONFIG_XFRM
 void xfrm6_local_rxpmtu(struct sk_buff *skb, u32 mtu);
Signature · 1 line
2.25.1
author maintainer
#13 lore raw

PATCH 12/17 xfrm: avoid compiler warning when ipv6 is disabled

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

with CONFIG_IPV6=n:
xfrm_output.c:140:12: warning: 'xfrm6_hdr_offset' defined but not used

Fixes 9acf4d3b9ec1 ("xfrm: ipv6: add xfrm6_hdr_offset helper")
---
 net/xfrm/xfrm_output.c | 2 ++
 1 file changed, 2 insertions(+)
net/xfrm/xfrm_output.c +2 -0
index e14fca1fb003..0b2975ef0668 100644
@@ -137,6 +137,7 @@ static int mip6_rthdr_offset(struct sk_buff *skb, u8 **nexthdr, int type)
 }
 #endif
 
+#if IS_ENABLED(CONFIG_IPV6)
 static int xfrm6_hdr_offset(struct xfrm_state *x, struct sk_buff *skb, u8 **prevhdr)
 {
 	switch (x->type->proto) {
@@ -151,6 +152,7 @@ static int xfrm6_hdr_offset(struct xfrm_state *x, struct sk_buff *skb, u8 **prev
 
 	return ip6_find_1stfragopt(skb, prevhdr);
 }
+#endif
 
 /* Add encapsulation header.
  *
Signature · 1 line
2.25.1
author maintainer
#14 lore raw

PATCH 13/17 xfrm: replay: avoid xfrm replay notify indirection

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

replay protection is implemented using a callback structure and then called via

x->repl->notify(), x->repl->recheck(), and so on.

all the differect functions are always built-in, so this could be direct calls instead.

This first patch prepares for removal of the x->repl structure. Add an enum with the three available replay modes to the xfrm_state structure and then replace all x->repl->notify() calls by the new xfrm_replay_notify() helper.

The helper checks the enum internally to adapt behaviour as needed.

---
 include/net/xfrm.h     | 11 ++++++++++-
 net/xfrm/xfrm_replay.c | 45 ++++++++++++++++++++++++++----------------
 net/xfrm/xfrm_state.c  |  2 +-
 3 files changed, 39 insertions(+), 19 deletions(-)
include/net/xfrm.h +10 -1
index 3a01570410ab..9a79e41defa7 100644
@@ -145,6 +145,12 @@ enum {
 	XFRM_MODE_FLAG_TUNNEL = 1,
 };
 
+enum xfrm_replay_mode {
+	XFRM_REPLAY_MODE_LEGACY,
+	XFRM_REPLAY_MODE_BMP,
+	XFRM_REPLAY_MODE_ESN,
+};
+
 /* Full description of state of transformer. */
 struct xfrm_state {
 	possible_net_t		xs_net;
@@ -218,6 +224,8 @@ struct xfrm_state {
 	/* The functions for replay detection. */
 	const struct xfrm_replay *repl;
 
+	/* replay detection mode */
+	enum xfrm_replay_mode    repl_mode;
 	/* internal flag that only holds state for delayed aevent at the
 	 * moment
 	*/
@@ -305,7 +313,6 @@ struct xfrm_replay {
 	int	(*recheck)(struct xfrm_state *x,
 			   struct sk_buff *skb,
 			   __be32 net_seq);
-	void	(*notify)(struct xfrm_state *x, int event);
 	int	(*overflow)(struct xfrm_state *x, struct sk_buff *skb);
 };
 
@@ -1715,6 +1722,8 @@ static inline int xfrm_policy_id2dir(u32 index)
 }
 
 #ifdef CONFIG_XFRM
+void xfrm_replay_notify(struct xfrm_state *x, int event);
+
 static inline int xfrm_aevent_is_on(struct net *net)
 {
 	struct sock *nlsk;
net/xfrm/xfrm_replay.c +28 -17

158 lines not on this page · raw message

net/xfrm/xfrm_state.c +1 -1
index 8f6058e56f7f..c2ce1e6f4760 100644
@@ -2177,7 +2177,7 @@ static void xfrm_replay_timer_handler(struct timer_list *t)
 
 	if (x->km.state == XFRM_STATE_VALID) {
 		if (xfrm_aevent_is_on(xs_net(x)))
-			x->repl->notify(x, XFRM_REPLAY_TIMEOUT);
+			xfrm_replay_notify(x, XFRM_REPLAY_TIMEOUT);
 		else
 			x->xflags |= XFRM_TIME_DEFER;
 	}
Signature · 1 line
2.25.1
author maintainer
#15 lore raw

PATCH 14/17 xfrm: replay: remove advance indirection

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

Similar to other patches: add a new helper to avoid an indirection.

v2: fix 'net/xfrm/xfrm_replay.c:519:13: warning: 'seq' may be used uninitialized in this function' warning.

---
 include/net/xfrm.h     |  2 +-
 net/xfrm/xfrm_input.c  |  2 +-
 net/xfrm/xfrm_replay.c | 24 +++++++++++++++---------
 3 files changed, 17 insertions(+), 11 deletions(-)
include/net/xfrm.h +1 -1
index 9a79e41defa7..a7f997b13198 100644
@@ -306,7 +306,6 @@ struct km_event {
 };
 
 struct xfrm_replay {
-	void	(*advance)(struct xfrm_state *x, __be32 net_seq);
 	int	(*check)(struct xfrm_state *x,
 			 struct sk_buff *skb,
 			 __be32 net_seq);
@@ -1722,6 +1721,7 @@ static inline int xfrm_policy_id2dir(u32 index)
 }
 
 #ifdef CONFIG_XFRM
+void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq);
 void xfrm_replay_notify(struct xfrm_state *x, int event);
 
 static inline int xfrm_aevent_is_on(struct net *net)
net/xfrm/xfrm_input.c +1 -1
index 1158cd0311d7..c8971e4b33ab 100644
@@ -665,7 +665,7 @@ int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
 			goto drop_unlock;
 		}
 
-		x->repl->advance(x, seq);
+		xfrm_replay_advance(x, seq);
 
 		x->curlft.bytes += skb->len;
 		x->curlft.packets++;
net/xfrm/xfrm_replay.c +15 -9
index 5feeb65f00b3..9565b0f7d380 100644
@@ -150,14 +150,26 @@ static int xfrm_replay_check(struct xfrm_state *x,
 	return -EINVAL;
 }
 
-static void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq)
+static void xfrm_replay_advance_bmp(struct xfrm_state *x, __be32 net_seq);
+static void xfrm_replay_advance_esn(struct xfrm_state *x, __be32 net_seq);
+
+void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq)
 {
-	u32 diff;
-	u32 seq = ntohl(net_seq);
+	u32 diff, seq;
+
+	switch (x->repl_mode) {
+	case XFRM_REPLAY_MODE_LEGACY:
+		break;
+	case XFRM_REPLAY_MODE_BMP:
+		return xfrm_replay_advance_bmp(x, net_seq);
+	case XFRM_REPLAY_MODE_ESN:
+		return xfrm_replay_advance_esn(x, net_seq);
+	}
 
 	if (!x->props.replay_window)
 		return;
 
+	seq = ntohl(net_seq);
 	if (seq > x->replay.seq) {
 		diff = seq - x->replay.seq;
 		if (diff < x->props.replay_window)
@@ -695,42 +707,36 @@ static int xfrm_replay_overflow_offload_esn(struct xfrm_state *x, struct sk_buff
 }
 
 static const struct xfrm_replay xfrm_replay_legacy = {
-	.advance	= xfrm_replay_advance,
 	.check		= xfrm_replay_check,
 	.recheck	= xfrm_replay_check,
 	.overflow	= xfrm_replay_overflow_offload,
 };
 
 static const struct xfrm_replay xfrm_replay_bmp = {
-	.advance	= xfrm_replay_advance_bmp,
 	.check		= xfrm_replay_check_bmp,
 	.recheck	= xfrm_replay_check_bmp,
 	.overflow	= xfrm_replay_overflow_offload_bmp,
 };
 
 static const struct xfrm_replay xfrm_replay_esn = {
-	.advance	= xfrm_replay_advance_esn,
 	.check		= xfrm_replay_check_esn,
 	.recheck	= xfrm_replay_recheck_esn,
 	.overflow	= xfrm_replay_overflow_offload_esn,
 };
 #else
 static const struct xfrm_replay xfrm_replay_legacy = {
-	.advance	= xfrm_replay_advance,
 	.check		= xfrm_replay_check,
 	.recheck	= xfrm_replay_check,
 	.overflow	= xfrm_replay_overflow,
 };
 
 static const struct xfrm_replay xfrm_replay_bmp = {
-	.advance	= xfrm_replay_advance_bmp,
 	.check		= xfrm_replay_check_bmp,
 	.recheck	= xfrm_replay_check_bmp,
 	.overflow	= xfrm_replay_overflow_bmp,
 };
 
 static const struct xfrm_replay xfrm_replay_esn = {
-	.advance	= xfrm_replay_advance_esn,
 	.check		= xfrm_replay_check_esn,
 	.recheck	= xfrm_replay_recheck_esn,
 	.overflow	= xfrm_replay_overflow_esn,
Signature · 1 line
2.25.1
author maintainer
#16 lore raw

PATCH 15/17 xfrm: replay: remove recheck indirection

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

Adds new xfrm_replay_recheck() helper and calls it from xfrm input path instead of the indirection.

---
 include/net/xfrm.h     |  4 +---
 net/xfrm/xfrm_input.c  |  2 +-
 net/xfrm/xfrm_replay.c | 22 ++++++++++++++++------
 3 files changed, 18 insertions(+), 10 deletions(-)
include/net/xfrm.h +1 -3
index a7f997b13198..3a219b34cb8c 100644
@@ -309,9 +309,6 @@ struct xfrm_replay {
 	int	(*check)(struct xfrm_state *x,
 			 struct sk_buff *skb,
 			 __be32 net_seq);
-	int	(*recheck)(struct xfrm_state *x,
-			   struct sk_buff *skb,
-			   __be32 net_seq);
 	int	(*overflow)(struct xfrm_state *x, struct sk_buff *skb);
 };
 
@@ -1723,6 +1720,7 @@ static inline int xfrm_policy_id2dir(u32 index)
 #ifdef CONFIG_XFRM
 void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq);
 void xfrm_replay_notify(struct xfrm_state *x, int event);
+int xfrm_replay_recheck(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
 
 static inline int xfrm_aevent_is_on(struct net *net)
 {
net/xfrm/xfrm_input.c +1 -1
index c8971e4b33ab..8046ef1a6680 100644
@@ -660,7 +660,7 @@ int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
 		/* only the first xfrm gets the encap type */
 		encap_type = 0;
 
-		if (x->repl->recheck(x, skb, seq)) {
+		if (xfrm_replay_recheck(x, skb, seq)) {
 			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
 			goto drop_unlock;
 		}
net/xfrm/xfrm_replay.c +16 -6
index 9565b0f7d380..59391dc80fa3 100644
@@ -519,6 +519,22 @@ static int xfrm_replay_recheck_esn(struct xfrm_state *x,
 	return xfrm_replay_check_esn(x, skb, net_seq);
 }
 
+int xfrm_replay_recheck(struct xfrm_state *x,
+			struct sk_buff *skb, __be32 net_seq)
+{
+	switch (x->repl_mode) {
+	case XFRM_REPLAY_MODE_LEGACY:
+		break;
+	case XFRM_REPLAY_MODE_BMP:
+		/* no special recheck treatment */
+		return xfrm_replay_check_bmp(x, skb, net_seq);
+	case XFRM_REPLAY_MODE_ESN:
+		return xfrm_replay_recheck_esn(x, skb, net_seq);
+	}
+
+	return xfrm_replay_check(x, skb, net_seq);
+}
+
 static void xfrm_replay_advance_esn(struct xfrm_state *x, __be32 net_seq)
 {
 	unsigned int bitnr, nr, i;
@@ -708,37 +724,31 @@ static int xfrm_replay_overflow_offload_esn(struct xfrm_state *x, struct sk_buff
 
 static const struct xfrm_replay xfrm_replay_legacy = {
 	.check		= xfrm_replay_check,
-	.recheck	= xfrm_replay_check,
 	.overflow	= xfrm_replay_overflow_offload,
 };
 
 static const struct xfrm_replay xfrm_replay_bmp = {
 	.check		= xfrm_replay_check_bmp,
-	.recheck	= xfrm_replay_check_bmp,
 	.overflow	= xfrm_replay_overflow_offload_bmp,
 };
 
 static const struct xfrm_replay xfrm_replay_esn = {
 	.check		= xfrm_replay_check_esn,
-	.recheck	= xfrm_replay_recheck_esn,
 	.overflow	= xfrm_replay_overflow_offload_esn,
 };
 #else
 static const struct xfrm_replay xfrm_replay_legacy = {
 	.check		= xfrm_replay_check,
-	.recheck	= xfrm_replay_check,
 	.overflow	= xfrm_replay_overflow,
 };
 
 static const struct xfrm_replay xfrm_replay_bmp = {
 	.check		= xfrm_replay_check_bmp,
-	.recheck	= xfrm_replay_check_bmp,
 	.overflow	= xfrm_replay_overflow_bmp,
 };
 
 static const struct xfrm_replay xfrm_replay_esn = {
 	.check		= xfrm_replay_check_esn,
-	.recheck	= xfrm_replay_recheck_esn,
 	.overflow	= xfrm_replay_overflow_esn,
 };
 #endif
Signature · 1 line
2.25.1
author maintainer
#17 lore raw

PATCH 16/17 xfrm: replay: avoid replay indirection

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

Add and use xfrm_replay_check helper instead of indirection.

---
 include/net/xfrm.h     |  4 +---
 net/xfrm/xfrm_input.c  |  2 +-
 net/xfrm/xfrm_replay.c | 27 ++++++++++++++++++---------
 3 files changed, 20 insertions(+), 13 deletions(-)
include/net/xfrm.h +1 -3
index 3a219b34cb8c..0206d80ec291 100644
@@ -306,9 +306,6 @@ struct km_event {
 };
 
 struct xfrm_replay {
-	int	(*check)(struct xfrm_state *x,
-			 struct sk_buff *skb,
-			 __be32 net_seq);
 	int	(*overflow)(struct xfrm_state *x, struct sk_buff *skb);
 };
 
@@ -1719,6 +1716,7 @@ static inline int xfrm_policy_id2dir(u32 index)
 
 #ifdef CONFIG_XFRM
 void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq);
+int xfrm_replay_check(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
 void xfrm_replay_notify(struct xfrm_state *x, int event);
 int xfrm_replay_recheck(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
 
net/xfrm/xfrm_input.c +1 -1
index 8046ef1a6680..3df0861d4390 100644
@@ -612,7 +612,7 @@ int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
 			goto drop_unlock;
 		}
 
-		if (x->repl->check(x, skb, seq)) {
+		if (xfrm_replay_check(x, skb, seq)) {
 			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
 			goto drop_unlock;
 		}
net/xfrm/xfrm_replay.c +18 -9
index 59391dc80fa3..e8703aa8d06a 100644
@@ -118,8 +118,8 @@ static int xfrm_replay_overflow(struct xfrm_state *x, struct sk_buff *skb)
 	return err;
 }
 
-static int xfrm_replay_check(struct xfrm_state *x,
-		      struct sk_buff *skb, __be32 net_seq)
+static int xfrm_replay_check_legacy(struct xfrm_state *x,
+				    struct sk_buff *skb, __be32 net_seq)
 {
 	u32 diff;
 	u32 seq = ntohl(net_seq);
@@ -507,6 +507,21 @@ static int xfrm_replay_check_esn(struct xfrm_state *x,
 	return -EINVAL;
 }
 
+int xfrm_replay_check(struct xfrm_state *x,
+		      struct sk_buff *skb, __be32 net_seq)
+{
+	switch (x->repl_mode) {
+	case XFRM_REPLAY_MODE_LEGACY:
+		break;
+	case XFRM_REPLAY_MODE_BMP:
+		return xfrm_replay_check_bmp(x, skb, net_seq);
+	case XFRM_REPLAY_MODE_ESN:
+		return xfrm_replay_check_esn(x, skb, net_seq);
+	}
+
+	return xfrm_replay_check_legacy(x, skb, net_seq);
+}
+
 static int xfrm_replay_recheck_esn(struct xfrm_state *x,
 				   struct sk_buff *skb, __be32 net_seq)
 {
@@ -532,7 +547,7 @@ int xfrm_replay_recheck(struct xfrm_state *x,
 		return xfrm_replay_recheck_esn(x, skb, net_seq);
 	}
 
-	return xfrm_replay_check(x, skb, net_seq);
+	return xfrm_replay_check_legacy(x, skb, net_seq);
 }
 
 static void xfrm_replay_advance_esn(struct xfrm_state *x, __be32 net_seq)
@@ -723,32 +738,26 @@ static int xfrm_replay_overflow_offload_esn(struct xfrm_state *x, struct sk_buff
 }
 
 static const struct xfrm_replay xfrm_replay_legacy = {
-	.check		= xfrm_replay_check,
 	.overflow	= xfrm_replay_overflow_offload,
 };
 
 static const struct xfrm_replay xfrm_replay_bmp = {
-	.check		= xfrm_replay_check_bmp,
 	.overflow	= xfrm_replay_overflow_offload_bmp,
 };
 
 static const struct xfrm_replay xfrm_replay_esn = {
-	.check		= xfrm_replay_check_esn,
 	.overflow	= xfrm_replay_overflow_offload_esn,
 };
 #else
 static const struct xfrm_replay xfrm_replay_legacy = {
-	.check		= xfrm_replay_check,
 	.overflow	= xfrm_replay_overflow,
 };
 
 static const struct xfrm_replay xfrm_replay_bmp = {
-	.check		= xfrm_replay_check_bmp,
 	.overflow	= xfrm_replay_overflow_bmp,
 };
 
 static const struct xfrm_replay xfrm_replay_esn = {
-	.check		= xfrm_replay_check_esn,
 	.overflow	= xfrm_replay_overflow_esn,
 };
 #endif
Signature · 1 line
2.25.1
author maintainer
#18 lore raw

PATCH 17/17 xfrm: replay: remove last replay indirection

In reply to Steffen Klassert

From: Florian Westphal <[email protected]>

This replaces the overflow indirection with the new xfrm_replay_overflow helper. After this, the 'repl' pointer in xfrm_state is no longer needed and can be removed as well.

xfrm_replay_overflow() is added in two incarnations, one is used when the kernel is compiled with xfrm hardware offload support enabled, the other when its disabled.

---
 include/net/xfrm.h     |  8 +------
 net/xfrm/xfrm_output.c |  2 +-
 net/xfrm/xfrm_replay.c | 51 +++++++++++++++++++++---------------------
 3 files changed, 28 insertions(+), 33 deletions(-)
include/net/xfrm.h +1 -7
index 0206d80ec291..d2a0559c255f 100644
@@ -221,9 +221,6 @@ struct xfrm_state {
 	struct xfrm_replay_state preplay;
 	struct xfrm_replay_state_esn *preplay_esn;
 
-	/* The functions for replay detection. */
-	const struct xfrm_replay *repl;
-
 	/* replay detection mode */
 	enum xfrm_replay_mode    repl_mode;
 	/* internal flag that only holds state for delayed aevent at the
@@ -305,10 +302,6 @@ struct km_event {
 	struct net *net;
 };
 
-struct xfrm_replay {
-	int	(*overflow)(struct xfrm_state *x, struct sk_buff *skb);
-};
-
 struct xfrm_if_cb {
 	struct xfrm_if	*(*decode_session)(struct sk_buff *skb,
 					   unsigned short family);
@@ -1718,6 +1711,7 @@ static inline int xfrm_policy_id2dir(u32 index)
 void xfrm_replay_advance(struct xfrm_state *x, __be32 net_seq);
 int xfrm_replay_check(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
 void xfrm_replay_notify(struct xfrm_state *x, int event);
+int xfrm_replay_overflow(struct xfrm_state *x, struct sk_buff *skb);
 int xfrm_replay_recheck(struct xfrm_state *x, struct sk_buff *skb, __be32 net_seq);
 
 static inline int xfrm_aevent_is_on(struct net *net)
net/xfrm/xfrm_output.c +1 -1
index 0b2975ef0668..527da58464f3 100644
@@ -525,7 +525,7 @@ static int xfrm_output_one(struct sk_buff *skb, int err)
 			goto error;
 		}
 
-		err = x->repl->overflow(x, skb);
+		err = xfrm_replay_overflow(x, skb);
 		if (err) {
 			XFRM_INC_STATS(net, LINUX_MIB_XFRMOUTSTATESEQERROR);
 			goto error;
net/xfrm/xfrm_replay.c +26 -25

83 lines not on this page · raw message

Signature · 1 line
2.25.1
#19 lore raw
In reply to Steffen Klassert
This pull request was applied to netdev/net-next.git (refs/heads/master):

Hello:

This pull request was applied to netdev/net-next.git (refs/heads/master):

Quoted 10 lines · On Mon, 28 Jun 2021 07:45:05 +0200 you wrote:

1) Remove an unneeded error assignment in esp4_gro_receive(). From Yang Li.

2) Add a new byseq state hashtable to find acquire states faster. From Sabrina Dubroca.

3) Remove some unnecessary variables in pfkey_create(). From zuoqilin.

[...]

Here is the summary with links:
- pull request (net-next): ipsec-next 2021-06-28

    https://git.kernel.org/netdev/net-next/c/1b077ce1c5be

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html