[NET]: make netlink user -> kernel interface synchronious

This patch make processing netlink user -> kernel messages synchronious.
This change was inspired by the talk with Alexey Kuznetsov about current
netlink messages processing. He says that he was badly wrong when introduced 
asynchronious user -> kernel communication.

The call netlink_unicast is the only path to send message to the kernel
netlink socket. But, unfortunately, it is also used to send data to the
user.

Before this change the user message has been attached to the socket queue
and sk->sk_data_ready was called. The process has been blocked until all
pending messages were processed. The bad thing is that this processing
may occur in the arbitrary process context.

This patch changes nlk->data_ready callback to get 1 skb and force packet
processing right in the netlink_unicast.

Kernel -> user path in netlink_unicast remains untouched.

EINTR processing for in netlink_run_queue was changed. It forces rtnl_lock
drop, but the process remains in the cycle until the message will be fully
processed. So, there is no need to use this kludges now.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Acked-by: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 38 insertions, 83 deletions

diff --git a/drivers/connector/connector.c b/drivers/connector/connector.c
index 569070997cc1..0e328d387af4 100644
--- a/drivers/connector/connector.c
+++ b/drivers/connector/connector.c
@@ -235,18 +235,6 @@ out:
 }
 
 /*
- * Netlink socket input callback - dequeues the skbs and calls the
- * main netlink receiving function.
- */
-static void cn_input(struct sock *sk, int len)
-{
-	struct sk_buff *skb;
-
-	while ((skb = skb_dequeue(&sk->sk_receive_queue)) != NULL)
-		cn_rx_skb(skb);
-}
-
-/*
  * Notification routing.
  *
  * Gets id and checks if there are notification request for it's idx
@@ -442,7 +430,7 @@ static int __devinit cn_init(void)
 	struct cn_dev *dev = &cdev;
 	int err;
 
-	dev->input = cn_input;
+	dev->input = cn_rx_skb;
 	dev->id.idx = cn_idx;
 	dev->id.val = cn_val;
 
diff --git a/drivers/scsi/scsi_netlink.c b/drivers/scsi/scsi_netlink.c
index 163acf6ad2d3..40579edca101 100644
--- a/drivers/scsi/scsi_netlink.c
+++ b/drivers/scsi/scsi_netlink.c
@@ -64,7 +64,7 @@ scsi_nl_rcv_msg(struct sk_buff *skb)
 
 		if (nlh->nlmsg_type != SCSI_TRANSPORT_MSG) {
 			err = -EBADMSG;
-			goto next_msg;
+			return;
 		}
 
 		hdr = NLMSG_DATA(nlh);
@@ -99,27 +99,6 @@ next_msg:
 
 
 /**
- * scsi_nl_rcv_msg -
- *    Receive handler for a socket. Extracts a received message buffer from
- *    the socket, and starts message processing.
- *
- * @sk:		socket
- * @len:	unused
- *
- **/
-static void
-scsi_nl_rcv(struct sock *sk, int len)
-{
-	struct sk_buff *skb;
-
-	while ((skb = skb_dequeue(&sk->sk_receive_queue))) {
-		scsi_nl_rcv_msg(skb);
-		kfree_skb(skb);
-	}
-}
-
-
-/**
  * scsi_nl_rcv_event -
  *    Event handler for a netlink socket.
  *
@@ -168,7 +147,7 @@ scsi_netlink_init(void)
 	}
 
 	scsi_nl_sock = netlink_kernel_create(&init_net, NETLINK_SCSITRANSPORT,
-				SCSI_NL_GRP_CNT, scsi_nl_rcv, NULL,
+				SCSI_NL_GRP_CNT, scsi_nl_rcv_msg, NULL,
 				THIS_MODULE);
 	if (!scsi_nl_sock) {
 		printk(KERN_ERR "%s: register of recieve handler failed\n",
diff --git a/drivers/scsi/scsi_transport_iscsi.c b/drivers/scsi/scsi_transport_iscsi.c
index 4916f01230dc..5428d15f23c6 100644
--- a/drivers/scsi/scsi_transport_iscsi.c
+++ b/drivers/scsi/scsi_transport_iscsi.c
@@ -1097,61 +1097,49 @@ iscsi_if_recv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
 }
 
 /*
- * Get message from skb (based on rtnetlink_rcv_skb).  Each message is
- * processed by iscsi_if_recv_msg.  Malformed skbs with wrong lengths or
- * invalid creds are discarded silently.
+ * Get message from skb.  Each message is processed by iscsi_if_recv_msg.
+ * Malformed skbs with wrong lengths or invalid creds are not processed.
  */
 static void
-iscsi_if_rx(struct sock *sk, int len)
+iscsi_if_rx(struct sk_buff *skb)
 {
-	struct sk_buff *skb;
-
 	mutex_lock(&rx_queue_mutex);
-	while ((skb = skb_dequeue(&sk->sk_receive_queue)) != NULL) {
-		if (NETLINK_CREDS(skb)->uid) {
-			skb_pull(skb, skb->len);
-			goto free_skb;
+	while (skb->len >= NLMSG_SPACE(0)) {
+		int err;
+		uint32_t rlen;
+		struct nlmsghdr	*nlh;
+		struct iscsi_uevent *ev;
+
+		nlh = nlmsg_hdr(skb);
+		if (nlh->nlmsg_len < sizeof(*nlh) ||
+		    skb->len < nlh->nlmsg_len) {
+			break;
 		}
 
-		while (skb->len >= NLMSG_SPACE(0)) {
-			int err;
-			uint32_t rlen;
-			struct nlmsghdr	*nlh;
-			struct iscsi_uevent *ev;
+		ev = NLMSG_DATA(nlh);
+		rlen = NLMSG_ALIGN(nlh->nlmsg_len);
+		if (rlen > skb->len)
+			rlen = skb->len;
 
-			nlh = nlmsg_hdr(skb);
-			if (nlh->nlmsg_len < sizeof(*nlh) ||
-			    skb->len < nlh->nlmsg_len) {
-				break;
-			}
-
-			ev = NLMSG_DATA(nlh);
-			rlen = NLMSG_ALIGN(nlh->nlmsg_len);
-			if (rlen > skb->len)
-				rlen = skb->len;
-
-			err = iscsi_if_recv_msg(skb, nlh);
-			if (err) {
-				ev->type = ISCSI_KEVENT_IF_ERROR;
-				ev->iferror = err;
-			}
-			do {
-				/*
-				 * special case for GET_STATS:
-				 * on success - sending reply and stats from
-				 * inside of if_recv_msg(),
-				 * on error - fall through.
-				 */
-				if (ev->type == ISCSI_UEVENT_GET_STATS && !err)
-					break;
-				err = iscsi_if_send_reply(
-					NETLINK_CREDS(skb)->pid, nlh->nlmsg_seq,
-					nlh->nlmsg_type, 0, 0, ev, sizeof(*ev));
-			} while (err < 0 && err != -ECONNREFUSED);
-			skb_pull(skb, rlen);
+		err = iscsi_if_recv_msg(skb, nlh);
+		if (err) {
+			ev->type = ISCSI_KEVENT_IF_ERROR;
+			ev->iferror = err;
 		}
-free_skb:
-		kfree_skb(skb);
+		do {
+			/*
+			 * special case for GET_STATS:
+			 * on success - sending reply and stats from
+			 * inside of if_recv_msg(),
+			 * on error - fall through.
+			 */
+			if (ev->type == ISCSI_UEVENT_GET_STATS && !err)
+				break;
+			err = iscsi_if_send_reply(
+				NETLINK_CREDS(skb)->pid, nlh->nlmsg_seq,
+				nlh->nlmsg_type, 0, 0, ev, sizeof(*ev));
+		} while (err < 0 && err != -ECONNREFUSED);
+		skb_pull(skb, rlen);
 	}
 	mutex_unlock(&rx_queue_mutex);
 }