[NETFILTER]: Fix conntrack seq_file handling.

Am travelling, but this passed simple tests here.  If this isn't going
in, the current seqfile stuff should be ripped out; it's a mess.

/proc/net/ip_conntrack was changed over to seq_file.  However,
seq_file isn't a great fit (a linked list which is changing is not a
good candidate for seq file), and the conversion was done badly.

1) Don't do allocation: simply hand the pointer head of the correct chain.
2) Actually output the original tuple.
3) Lock only when actually traversing hash chain.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/linux/netfilter_ipv4/ip_conntrack_protocol.h

@@ -3,10 +3,7 @@
 #define _IP_CONNTRACK_PROTOCOL_H
 #include <linux/netfilter_ipv4/ip_conntrack.h>
 
-/* length of buffer to which print_tuple/print_conntrack members are
- * writing */
-
-#define IP_CT_PRINT_BUFLEN 100
+struct seq_file;
 
 struct ip_conntrack_protocol
 {
@@ -31,13 +28,12 @@ struct ip_conntrack_protocol
 	int (*invert_tuple)(struct ip_conntrack_tuple *inverse,
 			    const struct ip_conntrack_tuple *orig);
 
-	/* Print out the per-protocol part of the tuple. */
-	unsigned int (*print_tuple)(char *buffer,
-				    const struct ip_conntrack_tuple *);
+	/* Print out the per-protocol part of the tuple. Return like seq_* */
+	int (*print_tuple)(struct seq_file *,
+			   const struct ip_conntrack_tuple *);
 
 	/* Print out the private part of the conntrack. */
-	unsigned int (*print_conntrack)(char *buffer,
-					const struct ip_conntrack *);
+	int (*print_conntrack)(struct seq_file *, const struct ip_conntrack *);
 
 	/* Returns verdict for packet, or -1 for invalid. */
 	int (*packet)(struct ip_conntrack *conntrack,

net/ipv4/netfilter/ip_conntrack_proto_generic.c

@@ -34,15 +34,15 @@ static int generic_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int generic_print_tuple(char *buffer,
-					const struct ip_conntrack_tuple *tuple)
+static int generic_print_tuple(struct seq_file *s,
+			       const struct ip_conntrack_tuple *tuple)
 {
 	return 0;
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int generic_print_conntrack(char *buffer,
-					    const struct ip_conntrack *state)
+static int generic_print_conntrack(struct seq_file *s,
+				   const struct ip_conntrack *state)
 {
 	return 0;
 }

net/ipv4/netfilter/ip_conntrack_proto_icmp.c

@@ -12,6 +12,7 @@
 #include <linux/netfilter.h>
 #include <linux/in.h>
 #include <linux/icmp.h>
+#include <linux/seq_file.h>
 #include <net/ip.h>
 #include <net/checksum.h>
 #include <linux/netfilter.h>
@@ -70,18 +71,18 @@ static int icmp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int icmp_print_tuple(char *buffer,
-				     const struct ip_conntrack_tuple *tuple)
+static int icmp_print_tuple(struct seq_file *s,
+			    const struct ip_conntrack_tuple *tuple)
 {
-	return sprintf(buffer, "type=%u code=%u id=%u ",
-		       tuple->dst.u.icmp.type,
-		       tuple->dst.u.icmp.code,
-		       ntohs(tuple->src.u.icmp.id));
+	return seq_printf(s, "type=%u code=%u id=%u ",
+			  tuple->dst.u.icmp.type,
+			  tuple->dst.u.icmp.code,
+			  ntohs(tuple->src.u.icmp.id));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int icmp_print_conntrack(char *buffer,
-				     const struct ip_conntrack *conntrack)
+static int icmp_print_conntrack(struct seq_file *s,
+				const struct ip_conntrack *conntrack)
 {
 	return 0;
 }

net/ipv4/netfilter/ip_conntrack_proto_sctp.c

@@ -22,6 +22,7 @@
 #include <linux/ip.h>
 #include <linux/sctp.h>
 #include <linux/string.h>
+#include <linux/seq_file.h>
 
 #include <linux/netfilter_ipv4/ip_conntrack.h>
 #include <linux/netfilter_ipv4/ip_conntrack_protocol.h>
@@ -178,20 +179,20 @@ static int sctp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int sctp_print_tuple(char *buffer,
-				     const struct ip_conntrack_tuple *tuple)
+static int sctp_print_tuple(struct seq_file *s,
+			    const struct ip_conntrack_tuple *tuple)
 {
 	DEBUGP(__FUNCTION__);
 	DEBUGP("\n");
 
-	return sprintf(buffer, "sport=%hu dport=%hu ",
-		       ntohs(tuple->src.u.sctp.port),
-		       ntohs(tuple->dst.u.sctp.port));
+	return seq_printf(s, "sport=%hu dport=%hu ",
+			  ntohs(tuple->src.u.sctp.port),
+			  ntohs(tuple->dst.u.sctp.port));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int sctp_print_conntrack(char *buffer,
-					 const struct ip_conntrack *conntrack)
+static int sctp_print_conntrack(struct seq_file *s,
+				const struct ip_conntrack *conntrack)
 {
 	enum sctp_conntrack state;
 
@@ -202,7 +203,7 @@ static unsigned int sctp_print_conntrack(char *buffer,
 	state = conntrack->proto.sctp.state;
 	READ_UNLOCK(&sctp_lock);
 
-	return sprintf(buffer, "%s ", sctp_conntrack_names[state]);
+	return seq_printf(s, "%s ", sctp_conntrack_names[state]);
 }
 
 #define for_each_sctp_chunk(skb, sch, offset, count)	\

net/ipv4/netfilter/ip_conntrack_proto_tcp.c

@@ -316,17 +316,17 @@ static int tcp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int tcp_print_tuple(char *buffer,
-				    const struct ip_conntrack_tuple *tuple)
+static int tcp_print_tuple(struct seq_file *s,
+			   const struct ip_conntrack_tuple *tuple)
 {
-	return sprintf(buffer, "sport=%hu dport=%hu ",
-		       ntohs(tuple->src.u.tcp.port),
-		       ntohs(tuple->dst.u.tcp.port));
+	return seq_printf(s, "sport=%hu dport=%hu ",
+			  ntohs(tuple->src.u.tcp.port),
+			  ntohs(tuple->dst.u.tcp.port));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int tcp_print_conntrack(char *buffer,
-					const struct ip_conntrack *conntrack)
+static int tcp_print_conntrack(struct seq_file *s,
+			       const struct ip_conntrack *conntrack)
 {
 	enum tcp_conntrack state;
 
@@ -334,7 +334,7 @@ static unsigned int tcp_print_conntrack(char *buffer,
 	state = conntrack->proto.tcp.state;
 	READ_UNLOCK(&tcp_lock);
 
-	return sprintf(buffer, "%s ", tcp_conntrack_names[state]);
+	return seq_printf(s, "%s ", tcp_conntrack_names[state]);
 }
 
 static unsigned int get_conntrack_index(const struct tcphdr *tcph)

net/ipv4/netfilter/ip_conntrack_proto_udp.c

@@ -12,6 +12,7 @@
 #include <linux/netfilter.h>
 #include <linux/in.h>
 #include <linux/udp.h>
+#include <linux/seq_file.h>
 #include <net/checksum.h>
 #include <linux/netfilter.h>
 #include <linux/netfilter_ipv4.h>
@@ -46,17 +47,17 @@ static int udp_invert_tuple(struct ip_conntrack_tuple *tuple,
 }
 
 /* Print out the per-protocol part of the tuple. */
-static unsigned int udp_print_tuple(char *buffer,
-				    const struct ip_conntrack_tuple *tuple)
+static int udp_print_tuple(struct seq_file *s,
+			   const struct ip_conntrack_tuple *tuple)
 {
-	return sprintf(buffer, "sport=%hu dport=%hu ",
-		       ntohs(tuple->src.u.udp.port),
-		       ntohs(tuple->dst.u.udp.port));
+	return seq_printf(s, "sport=%hu dport=%hu ",
+			  ntohs(tuple->src.u.udp.port),
+			  ntohs(tuple->dst.u.udp.port));
 }
 
 /* Print out the private part of the conntrack. */
-static unsigned int udp_print_conntrack(char *buffer,
-					const struct ip_conntrack *conntrack)
+static int udp_print_conntrack(struct seq_file *s,
+			       const struct ip_conntrack *conntrack)
 {
 	return 0;
 }

net/ipv4/netfilter/ip_conntrack_standalone.c

@@ -57,18 +57,13 @@ static int kill_proto(const struct ip_conntrack *i, void *data)
 }
 
 #ifdef CONFIG_PROC_FS
-static unsigned int
-print_tuple(char *buffer, const struct ip_conntrack_tuple *tuple,
+static int
+print_tuple(struct seq_file *s, const struct ip_conntrack_tuple *tuple,
 	    struct ip_conntrack_protocol *proto)
 {
-	int len;
-
-	len = sprintf(buffer, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
-		      NIPQUAD(tuple->src.ip), NIPQUAD(tuple->dst.ip));
-
-	len += proto->print_tuple(buffer + len, tuple);
-
-	return len;
+	seq_printf(s, "src=%u.%u.%u.%u dst=%u.%u.%u.%u ",
+		   NIPQUAD(tuple->src.ip), NIPQUAD(tuple->dst.ip));
+	return proto->print_tuple(s, tuple);
 }
 
 #ifdef CONFIG_IP_NF_CT_ACCT
@@ -85,48 +80,29 @@ seq_print_counters(struct seq_file *s, struct ip_conntrack_counter *counter)
 
 static void *ct_seq_start(struct seq_file *s, loff_t *pos)
 {
-	unsigned int *bucket;
-
-	/* strange seq_file api calls stop even if we fail,
-	 * thus we need to grab lock since stop unlocks */
-	READ_LOCK(&ip_conntrack_lock);
-  
 	if (*pos >= ip_conntrack_htable_size)
 		return NULL;
-
-	bucket = kmalloc(sizeof(unsigned int), GFP_KERNEL);
-	if (!bucket) {
-		return ERR_PTR(-ENOMEM);
-	}
-  
-	*bucket = *pos;
-	return bucket;
+	return &ip_conntrack_hash[*pos];
 }
   
-static void *ct_seq_next(struct seq_file *s, void *v, loff_t *pos)
+static void ct_seq_stop(struct seq_file *s, void *v)
 {
-	unsigned int *bucket = (unsigned int *) v;
+}
 
-	*pos = ++(*bucket);
-	if (*pos >= ip_conntrack_htable_size) {
-		kfree(v);
+static void *ct_seq_next(struct seq_file *s, void *v, loff_t *pos)
+{
+	(*pos)++;
+	if (*pos >= ip_conntrack_htable_size)
 		return NULL;
-	}
-	return bucket;
+	return &ip_conntrack_hash[*pos];
 }
   
-static void ct_seq_stop(struct seq_file *s, void *v)
-{
-	READ_UNLOCK(&ip_conntrack_lock);
-}
-
 /* return 0 on success, 1 in case of error */
 static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 			    struct seq_file *s)
 {
 	struct ip_conntrack *conntrack = hash->ctrack;
 	struct ip_conntrack_protocol *proto;
-	char buffer[IP_CT_PRINT_BUFLEN];
 
 	MUST_BE_READ_LOCKED(&ip_conntrack_lock);
 
@@ -147,12 +123,12 @@ static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 		      ? (conntrack->timeout.expires - jiffies)/HZ : 0) != 0)
 		return 1;
 
-	proto->print_conntrack(buffer, conntrack);
-	if (seq_puts(s, buffer))
+	if (proto->print_conntrack(s, conntrack))
 		return 1;
   
-	print_tuple(buffer, &conntrack->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
-		    proto);
+	if (print_tuple(s, &conntrack->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
+			proto))
+		return 1;
 
  	if (seq_print_counters(s, &conntrack->counters[IP_CT_DIR_ORIGINAL]))
 		return 1;
@@ -161,9 +137,8 @@ static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 		if (seq_printf(s, "[UNREPLIED] "))
 			return 1;
 
-	print_tuple(buffer, &conntrack->tuplehash[IP_CT_DIR_REPLY].tuple,
-		    proto);
-	if (seq_puts(s, buffer))
+	if (print_tuple(s, &conntrack->tuplehash[IP_CT_DIR_REPLY].tuple,
+			proto))
 		return 1;
 
  	if (seq_print_counters(s, &conntrack->counters[IP_CT_DIR_REPLY]))
@@ -179,17 +154,18 @@ static int ct_seq_real_show(const struct ip_conntrack_tuple_hash *hash,
 	return 0;
 }
 
-
 static int ct_seq_show(struct seq_file *s, void *v)
 {
-	unsigned int *bucket = (unsigned int *) v;
+	struct list_head *list = v;
+	int ret = 0;
 
-	if (LIST_FIND(&ip_conntrack_hash[*bucket], ct_seq_real_show,
-		      struct ip_conntrack_tuple_hash *, s)) {
-		/* buffer was filled and unable to print that tuple */
-		return 1;
-	}
-	return 0;
+	/* FIXME: Simply truncates if hash chain too long. */
+	READ_LOCK(&ip_conntrack_lock);
+	if (LIST_FIND(list, ct_seq_real_show,
+		      struct ip_conntrack_tuple_hash *, s))
+		ret = -ENOSPC;
+	READ_UNLOCK(&ip_conntrack_lock);
+	return ret;
 }
 	
 static struct seq_operations ct_seq_ops = {
@@ -255,7 +231,6 @@ static void exp_seq_stop(struct seq_file *s, void *v)
 static int exp_seq_show(struct seq_file *s, void *v)
 {
 	struct ip_conntrack_expect *expect = v;
-	char buffer[IP_CT_PRINT_BUFLEN];
 
 	if (expect->expectant->helper->timeout)
 		seq_printf(s, "%lu ", timer_pending(&expect->timeout)
@@ -266,9 +241,8 @@ static int exp_seq_show(struct seq_file *s, void *v)
 	seq_printf(s, "use=%u proto=%u ", atomic_read(&expect->use),
 		   expect->tuple.dst.protonum);
 
-	print_tuple(buffer, &expect->tuple,
-		    __ip_ct_find_proto(expect->tuple.dst.protonum));
-	return seq_printf(s, "%s\n", buffer);
+	return print_tuple(s, &expect->tuple,
+			   __ip_ct_find_proto(expect->tuple.dst.protonum));
 }
 
 static struct seq_operations exp_seq_ops = {