Merge tag 'trace-fixes-3.11-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace

Pull tracing fixes from Steven Rostedt:
 "Oleg Nesterov has been working hard in closing all the holes that can
  lead to race conditions between deleting an event and accessing an
  event debugfs file.  This included a fix to the debugfs system (acked
  by Greg Kroah-Hartman).  We think that all the holes have been patched
  and hopefully we don't find more.  I haven't marked all of them for
  stable because I need to examine them more to figure out how far back
  some of the changes need to go.

  Along the way, some other fixes have been made.  Alexander Z Lam fixed
  some logic where the wrong buffer was being modifed.

  Andrew Vagin found a possible corruption for machines that actually
  allocate cpumask, as a reference to one was being zeroed out by
  mistake.

  Dhaval Giani found a bad prototype when tracing is not configured.

  And I not only had some changes to help Oleg, but also finally fixed a
  long standing bug that Dave Jones and others have been hitting, where
  a module unload and reload can cause the function tracing accounting
  to get screwed up"

* tag 'trace-fixes-3.11-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  tracing: Fix reset of time stamps during trace_clock changes
  tracing: Make TRACE_ITER_STOP_ON_FREE stop the correct buffer
  tracing: Fix trace_dump_stack() proto when CONFIG_TRACING is not set
  tracing: Fix fields of struct trace_iterator that are zeroed by mistake
  tracing/uprobes: Fail to unregister if probe event files are in use
  tracing/kprobes: Fail to unregister if probe event files are in use
  tracing: Add comment to describe special break case in probe_remove_event_call()
  tracing: trace_remove_event_call() should fail if call/file is in use
  debugfs: debugfs_remove_recursive() must not rely on list_empty(d_subdirs)
  ftrace: Check module functions being traced on reload
  ftrace: Consolidate some duplicate code for updating ftrace ops
  tracing: Change remove_event_file_dir() to clear "d_subdirs"->i_private
  tracing: Introduce remove_event_file_dir()
  tracing: Change f_start() to take event_mutex and verify i_private != NULL
  tracing: Change event_filter_read/write to verify i_private != NULL
  tracing: Change event_enable/disable_read() to verify i_private != NULL
  tracing: Turn event/id->i_private into call->event.type

1 files changed, 127 insertions, 73 deletions

diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 898f868833f2..29a7ebcfb426 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -409,33 +409,42 @@ static void put_system(struct ftrace_subsystem_dir *dir)
 	mutex_unlock(&event_mutex);
 }
 
-/*
- * Open and update trace_array ref count.
- * Must have the current trace_array passed to it.
- */
-static int tracing_open_generic_file(struct inode *inode, struct file *filp)
+static void remove_subsystem(struct ftrace_subsystem_dir *dir)
 {
-	struct ftrace_event_file *file = inode->i_private;
-	struct trace_array *tr = file->tr;
-	int ret;
+	if (!dir)
+		return;
 
-	if (trace_array_get(tr) < 0)
-		return -ENODEV;
+	if (!--dir->nr_events) {
+		debugfs_remove_recursive(dir->entry);
+		list_del(&dir->list);
+		__put_system_dir(dir);
+	}
+}
 
-	ret = tracing_open_generic(inode, filp);
-	if (ret < 0)
-		trace_array_put(tr);
-	return ret;
+static void *event_file_data(struct file *filp)
+{
+	return ACCESS_ONCE(file_inode(filp)->i_private);
 }
 
-static int tracing_release_generic_file(struct inode *inode, struct file *filp)
+static void remove_event_file_dir(struct ftrace_event_file *file)
 {
-	struct ftrace_event_file *file = inode->i_private;
-	struct trace_array *tr = file->tr;
+	struct dentry *dir = file->dir;
+	struct dentry *child;
 
-	trace_array_put(tr);
+	if (dir) {
+		spin_lock(&dir->d_lock);	/* probably unneeded */
+		list_for_each_entry(child, &dir->d_subdirs, d_u.d_child) {
+			if (child->d_inode)	/* probably unneeded */
+				child->d_inode->i_private = NULL;
+		}
+		spin_unlock(&dir->d_lock);
 
-	return 0;
+		debugfs_remove_recursive(dir);
+	}
+
+	list_del(&file->list);
+	remove_subsystem(file->system);
+	kmem_cache_free(file_cachep, file);
 }
 
 /*
@@ -679,15 +688,25 @@ static ssize_t
 event_enable_read(struct file *filp, char __user *ubuf, size_t cnt,
 		  loff_t *ppos)
 {
-	struct ftrace_event_file *file = filp->private_data;
+	struct ftrace_event_file *file;
+	unsigned long flags;
 	char buf[4] = "0";
 
-	if (file->flags & FTRACE_EVENT_FL_ENABLED &&
-	    !(file->flags & FTRACE_EVENT_FL_SOFT_DISABLED))
+	mutex_lock(&event_mutex);
+	file = event_file_data(filp);
+	if (likely(file))
+		flags = file->flags;
+	mutex_unlock(&event_mutex);
+
+	if (!file)
+		return -ENODEV;
+
+	if (flags & FTRACE_EVENT_FL_ENABLED &&
+	    !(flags & FTRACE_EVENT_FL_SOFT_DISABLED))
 		strcpy(buf, "1");
 
-	if (file->flags & FTRACE_EVENT_FL_SOFT_DISABLED ||
-	    file->flags & FTRACE_EVENT_FL_SOFT_MODE)
+	if (flags & FTRACE_EVENT_FL_SOFT_DISABLED ||
+	    flags & FTRACE_EVENT_FL_SOFT_MODE)
 		strcat(buf, "*");
 
 	strcat(buf, "\n");
@@ -699,13 +718,10 @@ static ssize_t
 event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt,
 		   loff_t *ppos)
 {
-	struct ftrace_event_file *file = filp->private_data;
+	struct ftrace_event_file *file;
 	unsigned long val;
 	int ret;
 
-	if (!file)
-		return -EINVAL;
-
 	ret = kstrtoul_from_user(ubuf, cnt, 10, &val);
 	if (ret)
 		return ret;
@@ -717,8 +733,11 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt,
 	switch (val) {
 	case 0:
 	case 1:
+		ret = -ENODEV;
 		mutex_lock(&event_mutex);
-		ret = ftrace_event_enable_disable(file, val);
+		file = event_file_data(filp);
+		if (likely(file))
+			ret = ftrace_event_enable_disable(file, val);
 		mutex_unlock(&event_mutex);
 		break;
 
@@ -825,7 +844,7 @@ enum {
 
 static void *f_next(struct seq_file *m, void *v, loff_t *pos)
 {
-	struct ftrace_event_call *call = m->private;
+	struct ftrace_event_call *call = event_file_data(m->private);
 	struct list_head *common_head = &ftrace_common_fields;
 	struct list_head *head = trace_get_fields(call);
 	struct list_head *node = v;
@@ -857,7 +876,7 @@ static void *f_next(struct seq_file *m, void *v, loff_t *pos)
 
 static int f_show(struct seq_file *m, void *v)
 {
-	struct ftrace_event_call *call = m->private;
+	struct ftrace_event_call *call = event_file_data(m->private);
 	struct ftrace_event_field *field;
 	const char *array_descriptor;
 
@@ -910,6 +929,11 @@ static void *f_start(struct seq_file *m, loff_t *pos)
 	void *p = (void *)FORMAT_HEADER;
 	loff_t l = 0;
 
+	/* ->stop() is called even if ->start() fails */
+	mutex_lock(&event_mutex);
+	if (!event_file_data(m->private))
+		return ERR_PTR(-ENODEV);
+
 	while (l < *pos && p)
 		p = f_next(m, p, &l);
 
@@ -918,6 +942,7 @@ static void *f_start(struct seq_file *m, loff_t *pos)
 
 static void f_stop(struct seq_file *m, void *p)
 {
+	mutex_unlock(&event_mutex);
 }
 
 static const struct seq_operations trace_format_seq_ops = {
@@ -929,7 +954,6 @@ static const struct seq_operations trace_format_seq_ops = {
 
 static int trace_format_open(struct inode *inode, struct file *file)
 {
-	struct ftrace_event_call *call = inode->i_private;
 	struct seq_file *m;
 	int ret;
 
@@ -938,7 +962,7 @@ static int trace_format_open(struct inode *inode, struct file *file)
 		return ret;
 
 	m = file->private_data;
-	m->private = call;
+	m->private = file;
 
 	return 0;
 }
@@ -946,14 +970,18 @@ static int trace_format_open(struct inode *inode, struct file *file)
 static ssize_t
 event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos)
 {
-	struct ftrace_event_call *call = filp->private_data;
+	int id = (long)event_file_data(filp);
 	char buf[32];
 	int len;
 
 	if (*ppos)
 		return 0;
 
-	len = sprintf(buf, "%d\n", call->event.type);
+	if (unlikely(!id))
+		return -ENODEV;
+
+	len = sprintf(buf, "%d\n", id);
+
 	return simple_read_from_buffer(ubuf, cnt, ppos, buf, len);
 }
 
@@ -961,21 +989,28 @@ static ssize_t
 event_filter_read(struct file *filp, char __user *ubuf, size_t cnt,
 		  loff_t *ppos)
 {
-	struct ftrace_event_call *call = filp->private_data;
+	struct ftrace_event_call *call;
 	struct trace_seq *s;
-	int r;
+	int r = -ENODEV;
 
 	if (*ppos)
 		return 0;
 
 	s = kmalloc(sizeof(*s), GFP_KERNEL);
+
 	if (!s)
 		return -ENOMEM;
 
 	trace_seq_init(s);
 
-	print_event_filter(call, s);
-	r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len);
+	mutex_lock(&event_mutex);
+	call = event_file_data(filp);
+	if (call)
+		print_event_filter(call, s);
+	mutex_unlock(&event_mutex);
+
+	if (call)
+		r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len);
 
 	kfree(s);
 
@@ -986,9 +1021,9 @@ static ssize_t
 event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt,
 		   loff_t *ppos)
 {
-	struct ftrace_event_call *call = filp->private_data;
+	struct ftrace_event_call *call;
 	char *buf;
-	int err;
+	int err = -ENODEV;
 
 	if (cnt >= PAGE_SIZE)
 		return -EINVAL;
@@ -1003,7 +1038,12 @@ event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt,
 	}
 	buf[cnt] = '\0';
 
-	err = apply_event_filter(call, buf);
+	mutex_lock(&event_mutex);
+	call = event_file_data(filp);
+	if (call)
+		err = apply_event_filter(call, buf);
+	mutex_unlock(&event_mutex);
+
 	free_page((unsigned long) buf);
 	if (err < 0)
 		return err;
@@ -1225,10 +1265,9 @@ static const struct file_operations ftrace_set_event_fops = {
 };
 
 static const struct file_operations ftrace_enable_fops = {
-	.open = tracing_open_generic_file,
+	.open = tracing_open_generic,
 	.read = event_enable_read,
 	.write = event_enable_write,
-	.release = tracing_release_generic_file,
 	.llseek = default_llseek,
 };
 
@@ -1240,7 +1279,6 @@ static const struct file_operations ftrace_event_format_fops = {
 };
 
 static const struct file_operations ftrace_event_id_fops = {
-	.open = tracing_open_generic,
 	.read = event_id_read,
 	.llseek = default_llseek,
 };
@@ -1488,8 +1526,8 @@ event_create_dir(struct dentry *parent,
 
 #ifdef CONFIG_PERF_EVENTS
 	if (call->event.type && call->class->reg)
-		trace_create_file("id", 0444, file->dir, call,
-		 		  id);
+		trace_create_file("id", 0444, file->dir,
+				  (void *)(long)call->event.type, id);
 #endif
 
 	/*
@@ -1514,33 +1552,16 @@ event_create_dir(struct dentry *parent,
 	return 0;
 }
 
-static void remove_subsystem(struct ftrace_subsystem_dir *dir)
-{
-	if (!dir)
-		return;
-
-	if (!--dir->nr_events) {
-		debugfs_remove_recursive(dir->entry);
-		list_del(&dir->list);
-		__put_system_dir(dir);
-	}
-}
-
 static void remove_event_from_tracers(struct ftrace_event_call *call)
 {
 	struct ftrace_event_file *file;
 	struct trace_array *tr;
 
 	do_for_each_event_file_safe(tr, file) {
-
 		if (file->event_call != call)
 			continue;
 
-		list_del(&file->list);
-		debugfs_remove_recursive(file->dir);
-		remove_subsystem(file->system);
-		kmem_cache_free(file_cachep, file);
-
+		remove_event_file_dir(file);
 		/*
 		 * The do_for_each_event_file_safe() is
 		 * a double loop. After finding the call for this
@@ -1692,16 +1713,53 @@ static void __trace_remove_event_call(struct ftrace_event_call *call)
 	destroy_preds(call);
 }
 
+static int probe_remove_event_call(struct ftrace_event_call *call)
+{
+	struct trace_array *tr;
+	struct ftrace_event_file *file;
+
+#ifdef CONFIG_PERF_EVENTS
+	if (call->perf_refcount)
+		return -EBUSY;
+#endif
+	do_for_each_event_file(tr, file) {
+		if (file->event_call != call)
+			continue;
+		/*
+		 * We can't rely on ftrace_event_enable_disable(enable => 0)
+		 * we are going to do, FTRACE_EVENT_FL_SOFT_MODE can suppress
+		 * TRACE_REG_UNREGISTER.
+		 */
+		if (file->flags & FTRACE_EVENT_FL_ENABLED)
+			return -EBUSY;
+		/*
+		 * The do_for_each_event_file_safe() is
+		 * a double loop. After finding the call for this
+		 * trace_array, we use break to jump to the next
+		 * trace_array.
+		 */
+		break;
+	} while_for_each_event_file();
+
+	__trace_remove_event_call(call);
+
+	return 0;
+}
+
 /* Remove an event_call */
-void trace_remove_event_call(struct ftrace_event_call *call)
+int trace_remove_event_call(struct ftrace_event_call *call)
 {
+	int ret;
+
 	mutex_lock(&trace_types_lock);
 	mutex_lock(&event_mutex);
 	down_write(&trace_event_sem);
-	__trace_remove_event_call(call);
+	ret = probe_remove_event_call(call);
 	up_write(&trace_event_sem);
 	mutex_unlock(&event_mutex);
 	mutex_unlock(&trace_types_lock);
+
+	return ret;
 }
 
 #define for_each_event(event, start, end)			\
@@ -2270,12 +2328,8 @@ __trace_remove_event_dirs(struct trace_array *tr)
 {
 	struct ftrace_event_file *file, *next;
 
-	list_for_each_entry_safe(file, next, &tr->events, list) {
-		list_del(&file->list);
-		debugfs_remove_recursive(file->dir);
-		remove_subsystem(file->system);
-		kmem_cache_free(file_cachep, file);
-	}
+	list_for_each_entry_safe(file, next, &tr->events, list)
+		remove_event_file_dir(file);
 }
 
 static void