net: dynamically allocate fqdir structures

Following patch will add rcu grace period before fqdir
rhashtable destruction, so we need to dynamically allocate
fqdir structures to not force expensive synchronize_rcu() calls
in netns dismantle path.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 28 insertions, 27 deletions

diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
index b6f7385ed93c..c5d59fa568d6 100644
--- a/net/ipv6/netfilter/nf_conntrack_reasm.c
+++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
@@ -90,12 +90,12 @@ static int nf_ct_frag6_sysctl_register(struct net *net)
 			goto err_alloc;
 	}
 
-	table[0].data	= &net->nf_frag.fqdir.timeout;
-	table[1].data	= &net->nf_frag.fqdir.low_thresh;
-	table[1].extra2	= &net->nf_frag.fqdir.high_thresh;
-	table[2].data	= &net->nf_frag.fqdir.high_thresh;
-	table[2].extra1	= &net->nf_frag.fqdir.low_thresh;
-	table[2].extra2	= &init_net.nf_frag.fqdir.high_thresh;
+	table[0].data	= &net->nf_frag.fqdir->timeout;
+	table[1].data	= &net->nf_frag.fqdir->low_thresh;
+	table[1].extra2	= &net->nf_frag.fqdir->high_thresh;
+	table[2].data	= &net->nf_frag.fqdir->high_thresh;
+	table[2].extra1	= &net->nf_frag.fqdir->low_thresh;
+	table[2].extra2	= &init_net.nf_frag.fqdir->high_thresh;
 
 	hdr = register_net_sysctl(net, "net/netfilter", table);
 	if (hdr == NULL)
@@ -162,7 +162,7 @@ static struct frag_queue *fq_find(struct net *net, __be32 id, u32 user,
 	};
 	struct inet_frag_queue *q;
 
-	q = inet_frag_find(&net->nf_frag.fqdir, &key);
+	q = inet_frag_find(net->nf_frag.fqdir, &key);
 	if (!q)
 		return NULL;
 
@@ -489,23 +489,24 @@ static int nf_ct_net_init(struct net *net)
 {
 	int res;
 
-	net->nf_frag.fqdir.high_thresh = IPV6_FRAG_HIGH_THRESH;
-	net->nf_frag.fqdir.low_thresh = IPV6_FRAG_LOW_THRESH;
-	net->nf_frag.fqdir.timeout = IPV6_FRAG_TIMEOUT;
-
 	res = fqdir_init(&net->nf_frag.fqdir, &nf_frags, net);
 	if (res < 0)
 		return res;
+
+	net->nf_frag.fqdir->high_thresh = IPV6_FRAG_HIGH_THRESH;
+	net->nf_frag.fqdir->low_thresh = IPV6_FRAG_LOW_THRESH;
+	net->nf_frag.fqdir->timeout = IPV6_FRAG_TIMEOUT;
+
 	res = nf_ct_frag6_sysctl_register(net);
 	if (res < 0)
-		fqdir_exit(&net->nf_frag.fqdir);
+		fqdir_exit(net->nf_frag.fqdir);
 	return res;
 }
 
 static void nf_ct_net_exit(struct net *net)
 {
 	nf_ct_frags6_sysctl_unregister(net);
-	fqdir_exit(&net->nf_frag.fqdir);
+	fqdir_exit(net->nf_frag.fqdir);
 }
 
 static struct pernet_operations nf_ct_net_ops = {
diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c
index f3e3118393c4..0bbefc440bcd 100644
--- a/net/ipv6/proc.c
+++ b/net/ipv6/proc.c
@@ -48,8 +48,8 @@ static int sockstat6_seq_show(struct seq_file *seq, void *v)
 	seq_printf(seq, "RAW6: inuse %d\n",
 		       sock_prot_inuse_get(net, &rawv6_prot));
 	seq_printf(seq, "FRAG6: inuse %u memory %lu\n",
-		   atomic_read(&net->ipv6.fqdir.rhashtable.nelems),
-		   frag_mem_limit(&net->ipv6.fqdir));
+		   atomic_read(&net->ipv6.fqdir->rhashtable.nelems),
+		   frag_mem_limit(net->ipv6.fqdir));
 	return 0;
 }
 
diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c
index a6f26aa648fb..836ea964cf14 100644
--- a/net/ipv6/reassembly.c
+++ b/net/ipv6/reassembly.c
@@ -98,7 +98,7 @@ fq_find(struct net *net, __be32 id, const struct ipv6hdr *hdr, int iif)
 					    IPV6_ADDR_LINKLOCAL)))
 		key.iif = 0;
 
-	q = inet_frag_find(&net->ipv6.fqdir, &key);
+	q = inet_frag_find(net->ipv6.fqdir, &key);
 	if (!q)
 		return NULL;
 
@@ -443,11 +443,11 @@ static int __net_init ip6_frags_ns_sysctl_register(struct net *net)
 			goto err_alloc;
 
 	}
-	table[0].data	= &net->ipv6.fqdir.high_thresh;
-	table[0].extra1	= &net->ipv6.fqdir.low_thresh;
-	table[1].data	= &net->ipv6.fqdir.low_thresh;
-	table[1].extra2	= &net->ipv6.fqdir.high_thresh;
-	table[2].data	= &net->ipv6.fqdir.timeout;
+	table[0].data	= &net->ipv6.fqdir->high_thresh;
+	table[0].extra1	= &net->ipv6.fqdir->low_thresh;
+	table[1].data	= &net->ipv6.fqdir->low_thresh;
+	table[1].extra2	= &net->ipv6.fqdir->high_thresh;
+	table[2].data	= &net->ipv6.fqdir->timeout;
 
 	hdr = register_net_sysctl(net, "net/ipv6", table);
 	if (!hdr)
@@ -510,24 +510,24 @@ static int __net_init ipv6_frags_init_net(struct net *net)
 {
 	int res;
 
-	net->ipv6.fqdir.high_thresh = IPV6_FRAG_HIGH_THRESH;
-	net->ipv6.fqdir.low_thresh = IPV6_FRAG_LOW_THRESH;
-	net->ipv6.fqdir.timeout = IPV6_FRAG_TIMEOUT;
-
 	res = fqdir_init(&net->ipv6.fqdir, &ip6_frags, net);
 	if (res < 0)
 		return res;
 
+	net->ipv6.fqdir->high_thresh = IPV6_FRAG_HIGH_THRESH;
+	net->ipv6.fqdir->low_thresh = IPV6_FRAG_LOW_THRESH;
+	net->ipv6.fqdir->timeout = IPV6_FRAG_TIMEOUT;
+
 	res = ip6_frags_ns_sysctl_register(net);
 	if (res < 0)
-		fqdir_exit(&net->ipv6.fqdir);
+		fqdir_exit(net->ipv6.fqdir);
 	return res;
 }
 
 static void __net_exit ipv6_frags_exit_net(struct net *net)
 {
 	ip6_frags_ns_sysctl_unregister(net);
-	fqdir_exit(&net->ipv6.fqdir);
+	fqdir_exit(net->ipv6.fqdir);
 }
 
 static struct pernet_operations ip6_frags_ops = {