From: Harald Welte <[EMAIL PROTECTED]>
Date: Sat, 23 Jul 2005 16:15:52 -0400
> The attached patch adds support for refcounting of modules implementing
> netlink protocols. The idea is that you prevent the module from
> disappearing as long as someone in userspace has still a socket talking
> to you.
Ok, the changes look mostly fine. I've made a few slight
modifications before integrating, some of which I've mentioned
already:
1) I keep nl_table[] dynamically allocated
2) I fixed up some white spacing, very minor stuff
3) I fixed a socket leak in netlink_kernel_create(). If
netlink_lookup() returns non-NULL, you have a reference
to that socket thus have to release it.
The rest I've left as-is.
I'm only including the af_netlink.c part of that patch I integrated
since that's the only part I modified compared to your original
patch.
I think there is a slight hole in this code though, which we can
fixup as a followon patch. We probably need to grab the netlink
table from the point at which we set p_ops all the way to where
we full commit and netlink_insert() the kernel netlink socket.
In fact, I think we may need to do the whole operation atomically, and
thus with the netlink table grabbed, from existence test all the way
to the final socket insert. This would require versions of
netlink_create() and friends ala. __netlink_create() that operate with
the table already grabbed.
Thanks.
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -13,7 +13,12 @@
* added netlink_proto_exit
* Tue Jan 22 18:32:44 BRST 2002 Arnaldo C. de Melo <[EMAIL PROTECTED]>
* use nlk_sk, as sk->protinfo is on a diet 8)
- *
+ * Fri Jul 22 19:51:12 MEST 2005 Harald Welte <[EMAIL PROTECTED]>
+ * - inc module use count of module that owns
+ * the kernel socket in case userspace opens
+ * socket of same protocol
+ * - remove all module support, since netlink is
+ * mandatory if CONFIG_NET=y these days
*/
#include <linux/config.h>
@@ -92,6 +97,7 @@ struct netlink_table {
struct nl_pid_hash hash;
struct hlist_head mc_list;
unsigned int nl_nonroot;
+ struct proto_ops *p_ops;
};
static struct netlink_table *nl_table;
@@ -341,7 +347,21 @@ static int netlink_create(struct socket
if (protocol<0 || protocol >= MAX_LINKS)
return -EPROTONOSUPPORT;
- sock->ops = &netlink_ops;
+ netlink_table_grab();
+ if (!nl_table[protocol].hash.entries) {
+ netlink_table_ungrab();
+#ifdef CONFIG_KMOD
+ /* We do 'best effort'. If we find a matching module,
+ * it is loaded. If not, we don't return an error to
+ * allow pure userspace<->userspace communication. -HW
+ */
+ request_module("net-pf-%d-proto-%d", PF_NETLINK, protocol);
+ netlink_table_grab();
+#endif
+ }
+ netlink_table_ungrab();
+
+ sock->ops = nl_table[protocol].p_ops;
sk = sk_alloc(PF_NETLINK, GFP_KERNEL, &netlink_proto, 1);
if (!sk)
@@ -394,6 +414,22 @@ static int netlink_release(struct socket
};
notifier_call_chain(&netlink_chain, NETLINK_URELEASE, &n);
}
+
+ /* When this is a kernel socket, we need to remove the owner pointer,
+ * since we don't know whether the module will be dying at any given
+ * point - HW
+ */
+ if (!nlk->pid) {
+ struct proto_ops *p_tmp;
+
+ netlink_table_grab();
+ p_tmp = nl_table[sk->sk_protocol].p_ops;
+ if (p_tmp != &netlink_ops) {
+ nl_table[sk->sk_protocol].p_ops = &netlink_ops;
+ kfree(p_tmp);
+ }
+ netlink_table_ungrab();
+ }
sock_put(sk);
return 0;
@@ -1023,8 +1059,9 @@ static void netlink_data_ready(struct so
*/
struct sock *
-netlink_kernel_create(int unit, void (*input)(struct sock *sk, int len))
+netlink_kernel_create(int unit, void (*input)(struct sock *sk, int len),
struct module *module)
{
+ struct proto_ops *p_ops;
struct socket *sock;
struct sock *sk;
@@ -1034,22 +1071,63 @@ netlink_kernel_create(int unit, void (*i
if (unit<0 || unit>=MAX_LINKS)
return NULL;
+ /* Do a quick check, to make us not go down to netlink_insert()
+ * if protocol already has kernel socket.
+ */
+ sk = netlink_lookup(unit, 0);
+ if (unlikely(sk)) {
+ sock_put(sk);
+ return NULL;
+ }
+
if (sock_create_lite(PF_NETLINK, SOCK_DGRAM, unit, &sock))
return NULL;
+ sk = NULL;
+ if (module) {
+ /* Every registering protocol implemented in a module needs
+ * it's own p_ops, since the socket code cannot deal with
+ * module refcounting otherwise. -HW
+ */
+ p_ops = kmalloc(sizeof(*p_ops), GFP_KERNEL);
+ if (!p_ops)
+ goto out_sock_release;
+
+ memcpy(p_ops, &netlink_ops, sizeof(*p_ops));
+ p_ops->owner = module;
+ } else
+ p_ops = &netlink_ops;
+
+ netlink_table_grab();
+ nl_table[unit].p_ops = p_ops;
+ netlink_table_ungrab();
+
if (netlink_create(sock, unit) < 0) {
- sock_release(sock);
- return NULL;
+ sk = NULL;
+ goto out_kfree_p_ops;
}
+
sk = sock->sk;
sk->sk_data_ready = netlink_data_ready;
if (input)
nlk_sk(sk)->data_ready = input;
if (netlink_insert(sk, 0)) {
- sock_release(sock);
- return NULL;
+ sk = NULL;
+ goto out_kfree_p_ops;
+ }
+
+ return sk;
+
+out_kfree_p_ops:
+ netlink_table_grab();
+ if (nl_table[unit].p_ops != &netlink_ops) {
+ kfree(nl_table[unit].p_ops);
+ nl_table[unit].p_ops = &netlink_ops;
}
+ netlink_table_ungrab();
+out_sock_release:
+ sock_release(sock);
return sk;
}
@@ -1413,6 +1491,8 @@ enomem:
for (i = 0; i < MAX_LINKS; i++) {
struct nl_pid_hash *hash = &nl_table[i].hash;
+ nl_table[i].p_ops = &netlink_ops;
+
hash->table = nl_pid_hash_alloc(1 * sizeof(*hash->table));
if (!hash->table) {
while (i-- > 0)
@@ -1438,21 +1518,7 @@ out:
return err;
}
-static void __exit netlink_proto_exit(void)
-{
- sock_unregister(PF_NETLINK);
- proc_net_remove("netlink");
- kfree(nl_table);
- nl_table = NULL;
- proto_unregister(&netlink_proto);
-}
-
core_initcall(netlink_proto_init);
-module_exit(netlink_proto_exit);
-
-MODULE_LICENSE("GPL");
-
-MODULE_ALIAS_NETPROTO(PF_NETLINK);
EXPORT_SYMBOL(netlink_ack);
EXPORT_SYMBOL(netlink_broadcast);
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
