ipsec: Put dumpers on the dump list

Herbert Xu came up with the idea and the original patch to make
xfrm_state dump list contain also dumpers:

As it is we go to extraordinary lengths to ensure that states
don't go away while dumpers go to sleep.  It's much easier if
we just put the dumpers themselves on the list since they can't
go away while they're going.

I've also changed the order of addition on new states to prevent
a never-ending dump.

Timo Teräs improved the patch to apply cleanly to latest tree,
modified iteration code to be more readable by using a common
struct for entries in the list, implemented the same idea for
xfrm_policy dumping and moved the af_key specific "last" entry
caching to af_key.

Signed-off-by: Herbert Xu <[email protected]>
Signed-off-by: Timo Teras <[email protected]>
Signed-off-by: David S. Miller <[email protected]>

Author: herbertx

include/linux/netlink.h

@@ -220,7 +220,7 @@ struct netlink_callback
 	int		(*dump)(struct sk_buff * skb, struct netlink_callback *cb);
 	int		(*done)(struct netlink_callback *cb);
 	int		family;
-	long		args[7];
+	long		args[6];
 };
 
 struct netlink_notify

include/net/xfrm.h

@@ -117,12 +117,21 @@ extern struct mutex xfrm_cfg_mutex;
       metrics. Plus, it will be made via sk->sk_dst_cache. Solved.
  */
 
+struct xfrm_state_walk {
+	struct list_head	all;
+	u8			state;
+	union {
+		u8		dying;
+		u8		proto;
+	};
+	u32			seq;
+};
+
 /* Full description of state of transformer. */
 struct xfrm_state
 {
-	struct list_head	all;
 	union {
-		struct list_head	gclist;
+		struct hlist_node	gclist;
 		struct hlist_node	bydst;
 	};
 	struct hlist_node	bysrc;
@@ -136,12 +145,8 @@ struct xfrm_state
 
 	u32			genid;
 
-	/* Key manger bits */
-	struct {
-		u8		state;
-		u8		dying;
-		u32		seq;
-	} km;
+	/* Key manager bits */
+	struct xfrm_state_walk	km;
 
 	/* Parameters of this state. */
 	struct {
@@ -449,10 +454,20 @@ struct xfrm_tmpl
 
 #define XFRM_MAX_DEPTH		6
 
+struct xfrm_policy_walk_entry {
+	struct list_head	all;
+	u8			dead;
+};
+
+struct xfrm_policy_walk {
+	struct xfrm_policy_walk_entry walk;
+	u8 type;
+	u32 seq;
+};
+
 struct xfrm_policy
 {
 	struct xfrm_policy	*next;
-	struct list_head	bytype;
 	struct hlist_node	bydst;
 	struct hlist_node	byidx;
 
@@ -467,13 +482,12 @@ struct xfrm_policy
 	struct xfrm_lifetime_cfg lft;
 	struct xfrm_lifetime_cur curlft;
 	struct dst_entry       *bundles;
-	u16			family;
+	struct xfrm_policy_walk_entry walk;
 	u8			type;
 	u8			action;
 	u8			flags;
-	u8			dead;
 	u8			xfrm_nr;
-	/* XXX 1 byte hole, try to pack */
+	u16			family;
 	struct xfrm_sec_ctx	*security;
 	struct xfrm_tmpl       	xfrm_vec[XFRM_MAX_DEPTH];
 };
@@ -1245,20 +1259,6 @@ struct xfrm6_tunnel {
 	int priority;
 };
 
-struct xfrm_state_walk {
-	struct list_head list;
-	unsigned long genid;
-	struct xfrm_state *state;
-	int count;
-	u8 proto;
-};
-
-struct xfrm_policy_walk {
-	struct xfrm_policy *policy;
-	int count;
-	u8 type, cur_type;
-};
-
 extern void xfrm_init(void);
 extern void xfrm4_init(void);
 extern void xfrm_state_init(void);
@@ -1410,24 +1410,10 @@ static inline int xfrm4_udp_encap_rcv(struct sock *sk, struct sk_buff *skb)
 
 struct xfrm_policy *xfrm_policy_alloc(gfp_t gfp);
 
-static inline void xfrm_policy_walk_init(struct xfrm_policy_walk *walk, u8 type)
-{
-	walk->cur_type = XFRM_POLICY_TYPE_MAIN;
-	walk->type = type;
-	walk->policy = NULL;
-	walk->count = 0;
-}
-
-static inline void xfrm_policy_walk_done(struct xfrm_policy_walk *walk)
-{
-	if (walk->policy != NULL) {
-		xfrm_pol_put(walk->policy);
-		walk->policy = NULL;
-	}
-}
-
+extern void xfrm_policy_walk_init(struct xfrm_policy_walk *walk, u8 type);
 extern int xfrm_policy_walk(struct xfrm_policy_walk *walk,
 	int (*func)(struct xfrm_policy *, int, int, void*), void *);
+extern void xfrm_policy_walk_done(struct xfrm_policy_walk *walk);
 int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl);
 struct xfrm_policy *xfrm_policy_bysel_ctx(u8 type, int dir,
 					  struct xfrm_selector *sel,

net/key/af_key.c

@@ -58,6 +58,7 @@ struct pfkey_sock {
 			struct xfrm_policy_walk	policy;
 			struct xfrm_state_walk	state;
 		} u;
+		struct sk_buff	*skb;
 	} dump;
 };
 
@@ -76,6 +77,10 @@ static int pfkey_can_dump(struct sock *sk)
 static void pfkey_terminate_dump(struct pfkey_sock *pfk)
 {
 	if (pfk->dump.dump) {
+		if (pfk->dump.skb) {
+			kfree_skb(pfk->dump.skb);
+			pfk->dump.skb = NULL;
+		}
 		pfk->dump.done(pfk);
 		pfk->dump.dump = NULL;
 		pfk->dump.done = NULL;
@@ -308,12 +313,25 @@ static int pfkey_broadcast(struct sk_buff *skb, gfp_t allocation,
 
 static int pfkey_do_dump(struct pfkey_sock *pfk)
 {
+	struct sadb_msg *hdr;
 	int rc;
 
 	rc = pfk->dump.dump(pfk);
 	if (rc == -ENOBUFS)
 		return 0;
 
+	if (pfk->dump.skb) {
+		if (!pfkey_can_dump(&pfk->sk))
+			return 0;
+
+		hdr = (struct sadb_msg *) pfk->dump.skb->data;
+		hdr->sadb_msg_seq = 0;
+		hdr->sadb_msg_errno = rc;
+		pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
+				&pfk->sk);
+		pfk->dump.skb = NULL;
+	}
+
 	pfkey_terminate_dump(pfk);
 	return rc;
 }
@@ -1744,9 +1762,14 @@ static int dump_sa(struct xfrm_state *x, int count, void *ptr)
 	out_hdr->sadb_msg_satype = pfkey_proto2satype(x->id.proto);
 	out_hdr->sadb_msg_errno = 0;
 	out_hdr->sadb_msg_reserved = 0;
-	out_hdr->sadb_msg_seq = count;
+	out_hdr->sadb_msg_seq = count + 1;
 	out_hdr->sadb_msg_pid = pfk->dump.msg_pid;
-	pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ONE, &pfk->sk);
+
+	if (pfk->dump.skb)
+		pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
+				&pfk->sk);
+	pfk->dump.skb = out_skb;
+
 	return 0;
 }
 
@@ -2245,7 +2268,7 @@ static int pfkey_spdadd(struct sock *sk, struct sk_buff *skb, struct sadb_msg *h
 	return 0;
 
 out:
-	xp->dead = 1;
+	xp->walk.dead = 1;
 	xfrm_policy_destroy(xp);
 	return err;
 }
@@ -2583,9 +2606,14 @@ static int dump_sp(struct xfrm_policy *xp, int dir, int count, void *ptr)
 	out_hdr->sadb_msg_type = SADB_X_SPDDUMP;
 	out_hdr->sadb_msg_satype = SADB_SATYPE_UNSPEC;
 	out_hdr->sadb_msg_errno = 0;
-	out_hdr->sadb_msg_seq = count;
+	out_hdr->sadb_msg_seq = count + 1;
 	out_hdr->sadb_msg_pid = pfk->dump.msg_pid;
-	pfkey_broadcast(out_skb, GFP_ATOMIC, BROADCAST_ONE, &pfk->sk);
+
+	if (pfk->dump.skb)
+		pfkey_broadcast(pfk->dump.skb, GFP_ATOMIC, BROADCAST_ONE,
+				&pfk->sk);
+	pfk->dump.skb = out_skb;
+
 	return 0;
 }