Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

No conflicts.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>

4 files changed, 90 insertions, 41 deletions

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 48bae58d240e..07d51f95d6a6 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -6253,10 +6253,10 @@ again:
 
 		if (!atomic_inc_not_zero(&event->rb->mmap_count)) {
 			/*
-			 * Raced against perf_mmap_close() through
-			 * perf_event_set_output(). Try again, hope for better
-			 * luck.
+			 * Raced against perf_mmap_close(); remove the
+			 * event and try again.
 			 */
+			ring_buffer_attach(event, NULL);
 			mutex_unlock(&event->mmap_mutex);
 			goto again;
 		}
@@ -11829,14 +11829,25 @@ err_size:
 	goto out;
 }
 
+static void mutex_lock_double(struct mutex *a, struct mutex *b)
+{
+	if (b < a)
+		swap(a, b);
+
+	mutex_lock(a);
+	mutex_lock_nested(b, SINGLE_DEPTH_NESTING);
+}
+
 static int
 perf_event_set_output(struct perf_event *event, struct perf_event *output_event)
 {
 	struct perf_buffer *rb = NULL;
 	int ret = -EINVAL;
 
-	if (!output_event)
+	if (!output_event) {
+		mutex_lock(&event->mmap_mutex);
 		goto set;
+	}
 
 	/* don't allow circular references */
 	if (event == output_event)
@@ -11874,8 +11885,15 @@ perf_event_set_output(struct perf_event *event, struct perf_event *output_event)
 	    event->pmu != output_event->pmu)
 		goto out;
 
+	/*
+	 * Hold both mmap_mutex to serialize against perf_mmap_close().  Since
+	 * output_event is already on rb->event_list, and the list iteration
+	 * restarts after every removal, it is guaranteed this new event is
+	 * observed *OR* if output_event is already removed, it's guaranteed we
+	 * observe !rb->mmap_count.
+	 */
+	mutex_lock_double(&event->mmap_mutex, &output_event->mmap_mutex);
 set:
-	mutex_lock(&event->mmap_mutex);
 	/* Can't redirect output if we've got an active mmap() */
 	if (atomic_read(&event->mmap_count))
 		goto unlock;
@@ -11885,6 +11903,12 @@ set:
 		rb = ring_buffer_get(output_event);
 		if (!rb)
 			goto unlock;
+
+		/* did we race against perf_mmap_close() */
+		if (!atomic_read(&rb->mmap_count)) {
+			ring_buffer_put(rb);
+			goto unlock;
+		}
 	}
 
 	ring_buffer_attach(event, rb);
@@ -11892,20 +11916,13 @@ set:
 	ret = 0;
 unlock:
 	mutex_unlock(&event->mmap_mutex);
+	if (output_event)
+		mutex_unlock(&output_event->mmap_mutex);
 
 out:
 	return ret;
 }
 
-static void mutex_lock_double(struct mutex *a, struct mutex *b)
-{
-	if (b < a)
-		swap(a, b);
-
-	mutex_lock(a);
-	mutex_lock_nested(b, SINGLE_DEPTH_NESTING);
-}
-
 static int perf_event_set_clock(struct perf_event *event, clockid_t clk_id)
 {
 	bool nmi_safe = false;
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index b49c6ff6dca0..a1a81fd9889b 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -3380,6 +3380,7 @@ static bool __pr_flush(struct console *con, int timeout_ms, bool reset_on_progre
 		diff = 0;
 
 		console_lock();
+
 		for_each_console(c) {
 			if (con && con != c)
 				continue;
@@ -3389,11 +3390,19 @@ static bool __pr_flush(struct console *con, int timeout_ms, bool reset_on_progre
 			if (printk_seq < seq)
 				diff += seq - printk_seq;
 		}
-		console_unlock();
 
-		if (diff != last_diff && reset_on_progress)
+		/*
+		 * If consoles are suspended, it cannot be expected that they
+		 * make forward progress, so timeout immediately. @diff is
+		 * still used to return a valid flush status.
+		 */
+		if (console_suspended)
+			remaining = 0;
+		else if (diff != last_diff && reset_on_progress)
 			remaining = timeout_ms;
 
+		console_unlock();
+
 		if (diff == 0 || remaining == 0)
 			break;
 
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 619b02443522..b233714a1c78 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -2135,6 +2135,17 @@ static struct ctl_table vm_table[] = {
 		.extra1		= SYSCTL_ZERO,
 		.extra2		= SYSCTL_TWO_HUNDRED,
 	},
+#ifdef CONFIG_NUMA
+	{
+		.procname	= "numa_stat",
+		.data		= &sysctl_vm_numa_stat,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= sysctl_vm_numa_stat_handler,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_ONE,
+	},
+#endif
 #ifdef CONFIG_HUGETLB_PAGE
 	{
 		.procname	= "nr_hugepages",
@@ -2151,15 +2162,6 @@ static struct ctl_table vm_table[] = {
 		.mode           = 0644,
 		.proc_handler   = &hugetlb_mempolicy_sysctl_handler,
 	},
-	{
-		.procname		= "numa_stat",
-		.data			= &sysctl_vm_numa_stat,
-		.maxlen			= sizeof(int),
-		.mode			= 0644,
-		.proc_handler	= sysctl_vm_numa_stat_handler,
-		.extra1			= SYSCTL_ZERO,
-		.extra2			= SYSCTL_ONE,
-	},
 #endif
 	 {
 		.procname	= "hugetlb_shm_group",
diff --git a/kernel/watch_queue.c b/kernel/watch_queue.c
index 230038d4f908..bb9962b33f95 100644
--- a/kernel/watch_queue.c
+++ b/kernel/watch_queue.c
@@ -34,6 +34,27 @@ MODULE_LICENSE("GPL");
 #define WATCH_QUEUE_NOTE_SIZE 128
 #define WATCH_QUEUE_NOTES_PER_PAGE (PAGE_SIZE / WATCH_QUEUE_NOTE_SIZE)
 
+/*
+ * This must be called under the RCU read-lock, which makes
+ * sure that the wqueue still exists. It can then take the lock,
+ * and check that the wqueue hasn't been destroyed, which in
+ * turn makes sure that the notification pipe still exists.
+ */
+static inline bool lock_wqueue(struct watch_queue *wqueue)
+{
+	spin_lock_bh(&wqueue->lock);
+	if (unlikely(wqueue->defunct)) {
+		spin_unlock_bh(&wqueue->lock);
+		return false;
+	}
+	return true;
+}
+
+static inline void unlock_wqueue(struct watch_queue *wqueue)
+{
+	spin_unlock_bh(&wqueue->lock);
+}
+
 static void watch_queue_pipe_buf_release(struct pipe_inode_info *pipe,
 					 struct pipe_buffer *buf)
 {
@@ -69,6 +90,10 @@ static const struct pipe_buf_operations watch_queue_pipe_buf_ops = {
 
 /*
  * Post a notification to a watch queue.
+ *
+ * Must be called with the RCU lock for reading, and the
+ * watch_queue lock held, which guarantees that the pipe
+ * hasn't been released.
  */
 static bool post_one_notification(struct watch_queue *wqueue,
 				  struct watch_notification *n)
@@ -85,9 +110,6 @@ static bool post_one_notification(struct watch_queue *wqueue,
 
 	spin_lock_irq(&pipe->rd_wait.lock);
 
-	if (wqueue->defunct)
-		goto out;
-
 	mask = pipe->ring_size - 1;
 	head = pipe->head;
 	tail = pipe->tail;
@@ -203,7 +225,10 @@ void __post_watch_notification(struct watch_list *wlist,
 		if (security_post_notification(watch->cred, cred, n) < 0)
 			continue;
 
-		post_one_notification(wqueue, n);
+		if (lock_wqueue(wqueue)) {
+			post_one_notification(wqueue, n);
+			unlock_wqueue(wqueue);
+		}
 	}
 
 	rcu_read_unlock();
@@ -462,11 +487,12 @@ int add_watch_to_object(struct watch *watch, struct watch_list *wlist)
 		return -EAGAIN;
 	}
 
-	spin_lock_bh(&wqueue->lock);
-	kref_get(&wqueue->usage);
-	kref_get(&watch->usage);
-	hlist_add_head(&watch->queue_node, &wqueue->watches);
-	spin_unlock_bh(&wqueue->lock);
+	if (lock_wqueue(wqueue)) {
+		kref_get(&wqueue->usage);
+		kref_get(&watch->usage);
+		hlist_add_head(&watch->queue_node, &wqueue->watches);
+		unlock_wqueue(wqueue);
+	}
 
 	hlist_add_head(&watch->list_node, &wlist->watchers);
 	return 0;
@@ -520,20 +546,15 @@ found:
 
 	wqueue = rcu_dereference(watch->queue);
 
-	/* We don't need the watch list lock for the next bit as RCU is
-	 * protecting *wqueue from deallocation.
-	 */
-	if (wqueue) {
+	if (lock_wqueue(wqueue)) {
 		post_one_notification(wqueue, &n.watch);
 
-		spin_lock_bh(&wqueue->lock);
-
 		if (!hlist_unhashed(&watch->queue_node)) {
 			hlist_del_init_rcu(&watch->queue_node);
 			put_watch(watch);
 		}
 
-		spin_unlock_bh(&wqueue->lock);
+		unlock_wqueue(wqueue);
 	}
 
 	if (wlist->release_watch) {