1 files changed, 65 insertions, 36 deletions

diff --git a/net/unix/garbage.c b/net/unix/garbage.c
index 2a27b84f740b..9bc73f87f64a 100644
--- a/net/unix/garbage.c
+++ b/net/unix/garbage.c
@@ -74,12 +74,12 @@
 #include <linux/un.h>
 #include <linux/net.h>
 #include <linux/fs.h>
-#include <linux/slab.h>
 #include <linux/skbuff.h>
 #include <linux/netdevice.h>
 #include <linux/file.h>
 #include <linux/proc_fs.h>
 #include <linux/mutex.h>
+#include <linux/wait.h>
 
 #include <net/sock.h>
 #include <net/af_unix.h>
@@ -91,21 +91,22 @@
 static LIST_HEAD(gc_inflight_list);
 static LIST_HEAD(gc_candidates);
 static DEFINE_SPINLOCK(unix_gc_lock);
+static DECLARE_WAIT_QUEUE_HEAD(unix_gc_wait);
 
 unsigned int unix_tot_inflight;
 
 
-static struct sock *unix_get_socket(struct file *filp)
+struct sock *unix_get_socket(struct file *filp)
 {
 	struct sock *u_sock = NULL;
-	struct inode *inode = filp->f_path.dentry->d_inode;
+	struct inode *inode = file_inode(filp);
 
 	/*
 	 *	Socket ?
 	 */
-	if (S_ISSOCK(inode->i_mode)) {
-		struct socket * sock = SOCKET_I(inode);
-		struct sock * s = sock->sk;
+	if (S_ISSOCK(inode->i_mode) && !(filp->f_mode & FMODE_PATH)) {
+		struct socket *sock = SOCKET_I(inode);
+		struct sock *s = sock->sk;
 
 		/*
 		 *	PF_UNIX ?
@@ -124,7 +125,7 @@ static struct sock *unix_get_socket(struct file *filp)
 void unix_inflight(struct file *fp)
 {
 	struct sock *s = unix_get_socket(fp);
-	if(s) {
+	if (s) {
 		struct unix_sock *u = unix_sk(s);
 		spin_lock(&unix_gc_lock);
 		if (atomic_long_inc_return(&u->inflight) == 1) {
@@ -141,7 +142,7 @@ void unix_inflight(struct file *fp)
 void unix_notinflight(struct file *fp)
 {
 	struct sock *s = unix_get_socket(fp);
-	if(s) {
+	if (s) {
 		struct unix_sock *u = unix_sk(s);
 		spin_lock(&unix_gc_lock);
 		BUG_ON(list_empty(&u->link));
@@ -152,15 +153,6 @@ void unix_notinflight(struct file *fp)
 	}
 }
 
-static inline struct sk_buff *sock_queue_head(struct sock *sk)
-{
-	return (struct sk_buff *) &sk->sk_receive_queue;
-}
-
-#define receive_queue_for_each_skb(sk, next, skb) \
-	for (skb = sock_queue_head(sk)->next, next = skb->next; \
-	     skb != sock_queue_head(sk); skb = next, next = skb->next)
-
 static void scan_inflight(struct sock *x, void (*func)(struct unix_sock *),
 			  struct sk_buff_head *hitlist)
 {
@@ -168,7 +160,7 @@ static void scan_inflight(struct sock *x, void (*func)(struct unix_sock *),
 	struct sk_buff *next;
 
 	spin_lock(&x->sk_receive_queue.lock);
-	receive_queue_for_each_skb(x, next, skb) {
+	skb_queue_walk_safe(&x->sk_receive_queue, skb, next) {
 		/*
 		 *	Do we have file descriptors ?
 		 */
@@ -186,8 +178,17 @@ static void scan_inflight(struct sock *x, void (*func)(struct unix_sock *),
 				 */
 				struct sock *sk = unix_get_socket(*fp++);
 				if (sk) {
-					hit = true;
-					func(unix_sk(sk));
+					struct unix_sock *u = unix_sk(sk);
+
+					/*
+					 * Ignore non-candidates, they could
+					 * have been added to the queues after
+					 * starting the garbage collection
+					 */
+					if (test_bit(UNIX_GC_CANDIDATE, &u->gc_flags)) {
+						hit = true;
+						func(u);
+					}
 				}
 			}
 			if (hit && hitlist != NULL) {
@@ -215,7 +216,7 @@ static void scan_children(struct sock *x, void (*func)(struct unix_sock *),
 		 * and perform a scan on them as well.
 		 */
 		spin_lock(&x->sk_receive_queue.lock);
-		receive_queue_for_each_skb(x, next, skb) {
+		skb_queue_walk_safe(&x->sk_receive_queue, skb, next) {
 			u = unix_sk(skb->sk);
 
 			/*
@@ -249,24 +250,36 @@ static void inc_inflight_move_tail(struct unix_sock *u)
 {
 	atomic_long_inc(&u->inflight);
 	/*
-	 * If this is still a candidate, move it to the end of the
-	 * list, so that it's checked even if it was already passed
-	 * over
+	 * If this still might be part of a cycle, move it to the end
+	 * of the list, so that it's checked even if it was already
+	 * passed over
 	 */
-	if (u->gc_candidate)
+	if (test_bit(UNIX_GC_MAYBE_CYCLE, &u->gc_flags))
 		list_move_tail(&u->link, &gc_candidates);
 }
 
-/* The external entry point: unix_gc() */
+static bool gc_in_progress = false;
+#define UNIX_INFLIGHT_TRIGGER_GC 16000
 
-void unix_gc(void)
+void wait_for_unix_gc(void)
 {
-	static bool gc_in_progress = false;
+	/*
+	 * If number of inflight sockets is insane,
+	 * force a garbage collect right now.
+	 */
+	if (unix_tot_inflight > UNIX_INFLIGHT_TRIGGER_GC && !gc_in_progress)
+		unix_gc();
+	wait_event(unix_gc_wait, gc_in_progress == false);
+}
 
+/* The external entry point: unix_gc() */
+void unix_gc(void)
+{
 	struct unix_sock *u;
 	struct unix_sock *next;
 	struct sk_buff_head hitlist;
 	struct list_head cursor;
+	LIST_HEAD(not_cycle_list);
 
 	spin_lock(&unix_gc_lock);
 
@@ -282,10 +295,14 @@ void unix_gc(void)
 	 *
 	 * Holding unix_gc_lock will protect these candidates from
 	 * being detached, and hence from gaining an external
-	 * reference.  This also means, that since there are no
-	 * possible receivers, the receive queues of these sockets are
-	 * static during the GC, even though the dequeue is done
-	 * before the detach without atomicity guarantees.
+	 * reference.  Since there are no possible receivers, all
+	 * buffers currently on the candidates' queues stay there
+	 * during the garbage collection.
+	 *
+	 * We also know that no new candidate can be added onto the
+	 * receive queues.  Other, non candidate sockets _can_ be
+	 * added to queue, so we must make sure only to touch
+	 * candidates.
 	 */
 	list_for_each_entry_safe(u, next, &gc_inflight_list, link) {
 		long total_refs;
@@ -298,7 +315,8 @@ void unix_gc(void)
 		BUG_ON(total_refs < inflight_refs);
 		if (total_refs == inflight_refs) {
 			list_move_tail(&u->link, &gc_candidates);
-			u->gc_candidate = 1;
+			__set_bit(UNIX_GC_CANDIDATE, &u->gc_flags);
+			__set_bit(UNIX_GC_MAYBE_CYCLE, &u->gc_flags);
 		}
 	}
 
@@ -325,21 +343,31 @@ void unix_gc(void)
 		list_move(&cursor, &u->link);
 
 		if (atomic_long_read(&u->inflight) > 0) {
-			list_move_tail(&u->link, &gc_inflight_list);
-			u->gc_candidate = 0;
+			list_move_tail(&u->link, &not_cycle_list);
+			__clear_bit(UNIX_GC_MAYBE_CYCLE, &u->gc_flags);
 			scan_children(&u->sk, inc_inflight_move_tail, NULL);
 		}
 	}
 	list_del(&cursor);
 
 	/*
+	 * not_cycle_list contains those sockets which do not make up a
+	 * cycle.  Restore these to the inflight list.
+	 */
+	while (!list_empty(&not_cycle_list)) {
+		u = list_entry(not_cycle_list.next, struct unix_sock, link);
+		__clear_bit(UNIX_GC_CANDIDATE, &u->gc_flags);
+		list_move_tail(&u->link, &gc_inflight_list);
+	}
+
+	/*
 	 * Now gc_candidates contains only garbage.  Restore original
 	 * inflight counters for these as well, and remove the skbuffs
 	 * which are creating the cycle(s).
 	 */
 	skb_queue_head_init(&hitlist);
 	list_for_each_entry(u, &gc_candidates, link)
-		scan_children(&u->sk, inc_inflight, &hitlist);
+	scan_children(&u->sk, inc_inflight, &hitlist);
 
 	spin_unlock(&unix_gc_lock);
 
@@ -351,6 +379,7 @@ void unix_gc(void)
 	/* All candidates should have been detached by now. */
 	BUG_ON(!list_empty(&gc_candidates));
 	gc_in_progress = false;
+	wake_up(&unix_gc_wait);
 
  out:
 	spin_unlock(&unix_gc_lock);