From: Vincent Sanders <vincent.sanders@collabora.co.uk>
To: netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        "David S. Miller" <davem@davemloft.net>
Cc: Javier Martinez Canillas <javier.martinez@collabora.co.uk>,
        Vincent Sanders <vincent.sanders@collabora.co.uk>
Subject: [PATCH net-next 09/15] net: bus: Add garbage collector for AF_BUS sockets.
Date: Fri, 29 Jun 2012 17:45:48 +0100
Message-Id: <1340988354-26981-10-git-send-email-vincent.sanders@collabora.co.uk>
In-Reply-To: <1340988354-26981-1-git-send-email-vincent.sanders@collabora.co.uk>
References: <1340988354-26981-1-git-send-email-vincent.sanders@collabora.co.uk>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 9044
Lines: 347

From: Javier Martinez Canillas <javier.martinez@collabora.co.uk>

This patch adds a garbage collector for AF_BUS sockets.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Vincent Sanders <vincent.sanders@collabora.co.uk>
---
 net/bus/garbage.c |  322 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 322 insertions(+)
 create mode 100644 net/bus/garbage.c

diff --git a/net/bus/garbage.c b/net/bus/garbage.c
new file mode 100644
index 0000000..2435f38
--- /dev/null
+++ b/net/bus/garbage.c
@@ -0,0 +1,322 @@
+/*
+ * Garbage Collector For AF_BUS sockets
+ *
+ * Based on Garbage Collector For AF_UNIX sockets (net/unix/garbage.c).
+ */
+
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/socket.h>
+#include <linux/un.h>
+#include <linux/net.h>
+#include <linux/fs.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_bus.h>
+#include <net/scm.h>
+#include <net/tcp_states.h>
+
+/* Internal data structures and random procedures: */
+
+static LIST_HEAD(gc_inflight_list);
+static LIST_HEAD(gc_candidates);
+static DEFINE_SPINLOCK(bus_gc_lock);
+static DECLARE_WAIT_QUEUE_HEAD(bus_gc_wait);
+
+unsigned int bus_tot_inflight;
+
+
+struct sock *bus_get_socket(struct file *filp)
+{
+	struct sock *u_sock = NULL;
+	struct inode *inode = filp->f_path.dentry->d_inode;
+
+	/*
+	 *	Socket ?
+	 */
+	if (S_ISSOCK(inode->i_mode) && !(filp->f_mode & FMODE_PATH)) {
+		struct socket *sock = SOCKET_I(inode);
+		struct sock *s = sock->sk;
+
+		/*
+		 *	PF_BUS ?
+		 */
+		if (s && sock->ops && sock->ops->family == PF_BUS)
+			u_sock = s;
+	}
+	return u_sock;
+}
+
+/*
+ *	Keep the number of times in flight count for the file
+ *	descriptor if it is for an AF_BUS socket.
+ */
+
+void bus_inflight(struct file *fp)
+{
+	struct sock *s = bus_get_socket(fp);
+	if (s) {
+		struct bus_sock *u = bus_sk(s);
+		spin_lock(&bus_gc_lock);
+		if (atomic_long_inc_return(&u->inflight) == 1) {
+			BUG_ON(!list_empty(&u->link));
+			list_add_tail(&u->link, &gc_inflight_list);
+		} else {
+			BUG_ON(list_empty(&u->link));
+		}
+		bus_tot_inflight++;
+		spin_unlock(&bus_gc_lock);
+	}
+}
+
+void bus_notinflight(struct file *fp)
+{
+	struct sock *s = bus_get_socket(fp);
+	if (s) {
+		struct bus_sock *u = bus_sk(s);
+		spin_lock(&bus_gc_lock);
+		BUG_ON(list_empty(&u->link));
+		if (atomic_long_dec_and_test(&u->inflight))
+			list_del_init(&u->link);
+		bus_tot_inflight--;
+		spin_unlock(&bus_gc_lock);
+	}
+}
+
+static void scan_inflight(struct sock *x, void (*func)(struct bus_sock *),
+			  struct sk_buff_head *hitlist)
+{
+	struct sk_buff *skb;
+	struct sk_buff *next;
+
+	spin_lock(&x->sk_receive_queue.lock);
+	skb_queue_walk_safe(&x->sk_receive_queue, skb, next) {
+		/*
+		 *	Do we have file descriptors ?
+		 */
+		if (BUSCB(skb).fp) {
+			bool hit = false;
+			/*
+			 *	Process the descriptors of this socket
+			 */
+			int nfd = BUSCB(skb).fp->count;
+			struct file **fp = BUSCB(skb).fp->fp;
+			while (nfd--) {
+				/*
+				 *	Get the socket the fd matches
+				 *	if it indeed does so
+				 */
+				struct sock *sk = bus_get_socket(*fp++);
+				if (sk) {
+					struct bus_sock *u = bus_sk(sk);
+
+					/*
+					 * Ignore non-candidates, they could
+					 * have been added to the queues after
+					 * starting the garbage collection
+					 */
+					if (u->gc_candidate) {
+						hit = true;
+						func(u);
+					}
+				}
+			}
+			if (hit && hitlist != NULL) {
+				__skb_unlink(skb, &x->sk_receive_queue);
+				__skb_queue_tail(hitlist, skb);
+			}
+		}
+	}
+	spin_unlock(&x->sk_receive_queue.lock);
+}
+
+static void scan_children(struct sock *x, void (*func)(struct bus_sock *),
+			  struct sk_buff_head *hitlist)
+{
+	if (x->sk_state != TCP_LISTEN)
+		scan_inflight(x, func, hitlist);
+	else {
+		struct sk_buff *skb;
+		struct sk_buff *next;
+		struct bus_sock *u;
+		LIST_HEAD(embryos);
+
+		/*
+		 * For a listening socket collect the queued embryos
+		 * and perform a scan on them as well.
+		 */
+		spin_lock(&x->sk_receive_queue.lock);
+		skb_queue_walk_safe(&x->sk_receive_queue, skb, next) {
+			u = bus_sk(skb->sk);
+
+			/*
+			 * An embryo cannot be in-flight, so it's safe
+			 * to use the list link.
+			 */
+			BUG_ON(!list_empty(&u->link));
+			list_add_tail(&u->link, &embryos);
+		}
+		spin_unlock(&x->sk_receive_queue.lock);
+
+		while (!list_empty(&embryos)) {
+			u = list_entry(embryos.next, struct bus_sock, link);
+			scan_inflight(&u->sk, func, hitlist);
+			list_del_init(&u->link);
+		}
+	}
+}
+
+static void dec_inflight(struct bus_sock *usk)
+{
+	atomic_long_dec(&usk->inflight);
+}
+
+static void inc_inflight(struct bus_sock *usk)
+{
+	atomic_long_inc(&usk->inflight);
+}
+
+static void inc_inflight_move_tail(struct bus_sock *u)
+{
+	atomic_long_inc(&u->inflight);
+	/*
+	 * 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_maybe_cycle)
+		list_move_tail(&u->link, &gc_candidates);
+}
+
+static bool gc_in_progress = false;
+#define BUS_INFLIGHT_TRIGGER_GC 16000
+
+void wait_for_bus_gc(void)
+{
+	/*
+	 * If number of inflight sockets is insane,
+	 * force a garbage collect right now.
+	 */
+	if (bus_tot_inflight > BUS_INFLIGHT_TRIGGER_GC && !gc_in_progress)
+		bus_gc();
+	wait_event(bus_gc_wait, gc_in_progress == false);
+}
+
+/* The external entry point: bus_gc() */
+void bus_gc(void)
+{
+	struct bus_sock *u;
+	struct bus_sock *next;
+	struct sk_buff_head hitlist;
+	struct list_head cursor;
+	LIST_HEAD(not_cycle_list);
+
+	spin_lock(&bus_gc_lock);
+
+	/* Avoid a recursive GC. */
+	if (gc_in_progress)
+		goto out;
+
+	gc_in_progress = true;
+	/*
+	 * First, select candidates for garbage collection.  Only
+	 * in-flight sockets are considered, and from those only ones
+	 * which don't have any external reference.
+	 *
+	 * Holding bus_gc_lock will protect these candidates from
+	 * being detached, and hence from gaining an external
+	 * 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;
+		long inflight_refs;
+
+		total_refs = file_count(u->sk.sk_socket->file);
+		inflight_refs = atomic_long_read(&u->inflight);
+
+		BUG_ON(inflight_refs < 1);
+		BUG_ON(total_refs < inflight_refs);
+		if (total_refs == inflight_refs) {
+			list_move_tail(&u->link, &gc_candidates);
+			u->gc_candidate = 1;
+			u->gc_maybe_cycle = 1;
+		}
+	}
+
+	/*
+	 * Now remove all internal in-flight reference to children of
+	 * the candidates.
+	 */
+	list_for_each_entry(u, &gc_candidates, link)
+		scan_children(&u->sk, dec_inflight, NULL);
+
+	/*
+	 * Restore the references for children of all candidates,
+	 * which have remaining references.  Do this recursively, so
+	 * only those remain, which form cyclic references.
+	 *
+	 * Use a "cursor" link, to make the list traversal safe, even
+	 * though elements might be moved about.
+	 */
+	list_add(&cursor, &gc_candidates);
+	while (cursor.next != &gc_candidates) {
+		u = list_entry(cursor.next, struct bus_sock, link);
+
+		/* Move cursor to after the current position. */
+		list_move(&cursor, &u->link);
+
+		if (atomic_long_read(&u->inflight) > 0) {
+			list_move_tail(&u->link, &not_cycle_list);
+			u->gc_maybe_cycle = 0;
+			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 bus_sock, link);
+		u->gc_candidate = 0;
+		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);
+
+	spin_unlock(&bus_gc_lock);
+
+	/* Here we are. Hitlist is filled. Die. */
+	__skb_queue_purge(&hitlist);
+
+	spin_lock(&bus_gc_lock);
+
+	/* All candidates should have been detached by now. */
+	BUG_ON(!list_empty(&gc_candidates));
+	gc_in_progress = false;
+	wake_up(&bus_gc_wait);
+
+ out:
+	spin_unlock(&bus_gc_lock);
+}
-- 
1.7.10

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/