ftrace: Have set_ftrace_pid use the bitmap like events do

Convert set_ftrace_pid to use the bitmap like set_event_pid does. This
allows for instances to use the pid filtering as well, and will allow for
function-fork option to set if the children of a traced function should be
traced or not.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

kernel/trace/ftrace.c

@@ -89,16 +89,16 @@ struct ftrace_ops *function_trace_op __read_mostly = &ftrace_list_end;
 /* What to set function_trace_op to */
 static struct ftrace_ops *set_function_trace_op;
 
-/* List for set_ftrace_pid's pids. */
+static bool ftrace_pids_enabled(struct ftrace_ops *ops)
-LIST_HEAD(ftrace_pids);
-struct ftrace_pid {
-	struct list_head list;
-	struct pid *pid;
-};
-
-static bool ftrace_pids_enabled(void)
 {
-	return !list_empty(&ftrace_pids);
+	struct trace_array *tr;
+
+	if (!(ops->flags & FTRACE_OPS_FL_PID) || !ops->private)
+		return false;
+
+	tr = ops->private;
+
+	return tr->function_pids != NULL;
 }
 
 static void ftrace_update_trampoline(struct ftrace_ops *ops);
@@ -179,7 +179,9 @@ int ftrace_nr_registered_ops(void)
 static void ftrace_pid_func(unsigned long ip, unsigned long parent_ip,
 			    struct ftrace_ops *op, struct pt_regs *regs)
 {
-	if (!test_tsk_trace_trace(current))
+	struct trace_array *tr = op->private;
+
+	if (tr && this_cpu_read(tr->trace_buffer.data->ftrace_ignore_pid))
 		return;
 
 	op->saved_func(ip, parent_ip, op, regs);
@@ -417,7 +419,7 @@ static int __register_ftrace_function(struct ftrace_ops *ops)
 	/* Always save the function, and reset at unregistering */
 	ops->saved_func = ops->func;
 
-	if (ops->flags & FTRACE_OPS_FL_PID && ftrace_pids_enabled())
+	if (ftrace_pids_enabled(ops))
 		ops->func = ftrace_pid_func;
 
 	ftrace_update_trampoline(ops);
@@ -450,7 +452,6 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops)
 
 static void ftrace_update_pid_func(void)
 {
-	bool enabled = ftrace_pids_enabled();
 	struct ftrace_ops *op;
 
 	/* Only do something if we are tracing something */
@@ -459,8 +460,8 @@ static void ftrace_update_pid_func(void)
 
 	do_for_each_ftrace_op(op, ftrace_ops_list) {
 		if (op->flags & FTRACE_OPS_FL_PID) {
-			op->func = enabled ? ftrace_pid_func :
+			op->func = ftrace_pids_enabled(op) ?
-				op->saved_func;
+				ftrace_pid_func : op->saved_func;
 			ftrace_update_trampoline(op);
 		}
 	} while_for_each_ftrace_op(op);
@@ -5324,134 +5325,46 @@ ftrace_func_t ftrace_ops_get_func(struct ftrace_ops *ops)
 	return ops->func;
 }
 
-static void clear_ftrace_swapper(void)
+static void
+ftrace_filter_pid_sched_switch_probe(void *data, bool preempt,
+		    struct task_struct *prev, struct task_struct *next)
 {
-	struct task_struct *p;
+	struct trace_array *tr = data;
+	struct trace_pid_list *pid_list;
+
+	pid_list = rcu_dereference_sched(tr->function_pids);
+
+	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
+		       trace_ignore_this_task(pid_list, next));
+}
+
+static void clear_ftrace_pids(struct trace_array *tr)
+{
+	struct trace_pid_list *pid_list;
 	int cpu;
 
-	get_online_cpus();
+	pid_list = rcu_dereference_protected(tr->function_pids,
-	for_each_online_cpu(cpu) {
+					     lockdep_is_held(&ftrace_lock));
-		p = idle_task(cpu);
+	if (!pid_list)
-		clear_tsk_trace_trace(p);
+		return;
-	}
+
-	put_online_cpus();
+	unregister_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
+
+	for_each_possible_cpu(cpu)
+		per_cpu_ptr(tr->trace_buffer.data, cpu)->ftrace_ignore_pid = false;
+
+	rcu_assign_pointer(tr->function_pids, NULL);
+
+	/* Wait till all users are no longer using pid filtering */
+	synchronize_sched();
+
+	trace_free_pid_list(pid_list);
 }
 
-static void set_ftrace_swapper(void)
+static void ftrace_pid_reset(struct trace_array *tr)
 {
-	struct task_struct *p;
-	int cpu;
-
-	get_online_cpus();
-	for_each_online_cpu(cpu) {
-		p = idle_task(cpu);
-		set_tsk_trace_trace(p);
-	}
-	put_online_cpus();
-}
-
-static void clear_ftrace_pid(struct pid *pid)
-{
-	struct task_struct *p;
-
-	rcu_read_lock();
-	do_each_pid_task(pid, PIDTYPE_PID, p) {
-		clear_tsk_trace_trace(p);
-	} while_each_pid_task(pid, PIDTYPE_PID, p);
-	rcu_read_unlock();
-
-	put_pid(pid);
-}
-
-static void set_ftrace_pid(struct pid *pid)
-{
-	struct task_struct *p;
-
-	rcu_read_lock();
-	do_each_pid_task(pid, PIDTYPE_PID, p) {
-		set_tsk_trace_trace(p);
-	} while_each_pid_task(pid, PIDTYPE_PID, p);
-	rcu_read_unlock();
-}
-
-static void clear_ftrace_pid_task(struct pid *pid)
-{
-	if (pid == ftrace_swapper_pid)
-		clear_ftrace_swapper();
-	else
-		clear_ftrace_pid(pid);
-}
-
-static void set_ftrace_pid_task(struct pid *pid)
-{
-	if (pid == ftrace_swapper_pid)
-		set_ftrace_swapper();
-	else
-		set_ftrace_pid(pid);
-}
-
-static int ftrace_pid_add(int p)
-{
-	struct pid *pid;
-	struct ftrace_pid *fpid;
-	int ret = -EINVAL;
-
 	mutex_lock(&ftrace_lock);
-
+	clear_ftrace_pids(tr);
-	if (!p)
-		pid = ftrace_swapper_pid;
-	else
-		pid = find_get_pid(p);
-
-	if (!pid)
-		goto out;
-
-	ret = 0;
-
-	list_for_each_entry(fpid, &ftrace_pids, list)
-		if (fpid->pid == pid)
-			goto out_put;
-
-	ret = -ENOMEM;
-
-	fpid = kmalloc(sizeof(*fpid), GFP_KERNEL);
-	if (!fpid)
-		goto out_put;
-
-	list_add(&fpid->list, &ftrace_pids);
-	fpid->pid = pid;
-
-	set_ftrace_pid_task(pid);
-
-	ftrace_update_pid_func();
-
-	ftrace_startup_all(0);
-
-	mutex_unlock(&ftrace_lock);
-	return 0;
-
-out_put:
-	if (pid != ftrace_swapper_pid)
-		put_pid(pid);
-
-out:
-	mutex_unlock(&ftrace_lock);
-	return ret;
-}
-
-static void ftrace_pid_reset(void)
-{
-	struct ftrace_pid *fpid, *safe;
-
-	mutex_lock(&ftrace_lock);
-	list_for_each_entry_safe(fpid, safe, &ftrace_pids, list) {
-		struct pid *pid = fpid->pid;
-
-		clear_ftrace_pid_task(pid);
-
-		list_del(&fpid->list);
-		kfree(fpid);
-	}
 
 	ftrace_update_pid_func();
 	ftrace_startup_all(0);
@@ -5459,44 +5372,52 @@ static void ftrace_pid_reset(void)
 	mutex_unlock(&ftrace_lock);
 }
 
+/* Greater than any max PID */
+#define FTRACE_NO_PIDS		(void *)(PID_MAX_LIMIT + 1)
+
 static void *fpid_start(struct seq_file *m, loff_t *pos)
+	__acquires(RCU)
 {
+	struct trace_pid_list *pid_list;
+	struct trace_array *tr = m->private;
+
 	mutex_lock(&ftrace_lock);
+	rcu_read_lock_sched();
 
-	if (!ftrace_pids_enabled() && (!*pos))
+	pid_list = rcu_dereference_sched(tr->function_pids);
-		return (void *) 1;
 
-	return seq_list_start(&ftrace_pids, *pos);
+	if (!pid_list)
+		return !(*pos) ? FTRACE_NO_PIDS : NULL;
+
+	return trace_pid_start(pid_list, pos);
 }
 
 static void *fpid_next(struct seq_file *m, void *v, loff_t *pos)
 {
-	if (v == (void *)1)
+	struct trace_array *tr = m->private;
+	struct trace_pid_list *pid_list = rcu_dereference_sched(tr->function_pids);
+
+	if (v == FTRACE_NO_PIDS)
 		return NULL;
 
-	return seq_list_next(v, &ftrace_pids, pos);
+	return trace_pid_next(pid_list, v, pos);
 }
 
 static void fpid_stop(struct seq_file *m, void *p)
+	__releases(RCU)
 {
+	rcu_read_unlock_sched();
 	mutex_unlock(&ftrace_lock);
 }
 
 static int fpid_show(struct seq_file *m, void *v)
 {
-	const struct ftrace_pid *fpid = list_entry(v, struct ftrace_pid, list);
+	if (v == FTRACE_NO_PIDS) {
-
-	if (v == (void *)1) {
 		seq_puts(m, "no pid\n");
 		return 0;
 	}
 
-	if (fpid->pid == ftrace_swapper_pid)
+	return trace_pid_show(m, v);
-		seq_puts(m, "swapper tasks\n");
-	else
-		seq_printf(m, "%u\n", pid_vnr(fpid->pid));
-
-	return 0;
 }
 
 static const struct seq_operations ftrace_pid_sops = {
@@ -5509,58 +5430,103 @@ static const struct seq_operations ftrace_pid_sops = {
 static int
 ftrace_pid_open(struct inode *inode, struct file *file)
 {
+	struct trace_array *tr = inode->i_private;
+	struct seq_file *m;
 	int ret = 0;
 
+	if (trace_array_get(tr) < 0)
+		return -ENODEV;
+
 	if ((file->f_mode & FMODE_WRITE) &&
 	    (file->f_flags & O_TRUNC))
-		ftrace_pid_reset();
+		ftrace_pid_reset(tr);
 
-	if (file->f_mode & FMODE_READ)
+	ret = seq_open(file, &ftrace_pid_sops);
-		ret = seq_open(file, &ftrace_pid_sops);
+	if (ret < 0) {
+		trace_array_put(tr);
+	} else {
+		m = file->private_data;
+		/* copy tr over to seq ops */
+		m->private = tr;
+	}
 
 	return ret;
 }
 
+static void ignore_task_cpu(void *data)
+{
+	struct trace_array *tr = data;
+	struct trace_pid_list *pid_list;
+
+	/*
+	 * This function is called by on_each_cpu() while the
+	 * event_mutex is held.
+	 */
+	pid_list = rcu_dereference_protected(tr->function_pids,
+					     mutex_is_locked(&ftrace_lock));
+
+	this_cpu_write(tr->trace_buffer.data->ftrace_ignore_pid,
+		       trace_ignore_this_task(pid_list, current));
+}
+
 static ssize_t
 ftrace_pid_write(struct file *filp, const char __user *ubuf,
 		   size_t cnt, loff_t *ppos)
 {
-	char buf[64], *tmp;
+	struct seq_file *m = filp->private_data;
-	long val;
+	struct trace_array *tr = m->private;
-	int ret;
+	struct trace_pid_list *filtered_pids = NULL;
+	struct trace_pid_list *pid_list;
+	ssize_t ret;
 
-	if (cnt >= sizeof(buf))
+	if (!cnt)
-		return -EINVAL;
+		return 0;
 
-	if (copy_from_user(&buf, ubuf, cnt))
+	mutex_lock(&ftrace_lock);
-		return -EFAULT;
 
-	buf[cnt] = 0;
+	filtered_pids = rcu_dereference_protected(tr->function_pids,
+					     lockdep_is_held(&ftrace_lock));
+
+	ret = trace_pid_write(filtered_pids, &pid_list, ubuf, cnt);
+	if (ret < 0)
+		goto out;
+
+	rcu_assign_pointer(tr->function_pids, pid_list);
+
+	if (filtered_pids) {
+		synchronize_sched();
+		trace_free_pid_list(filtered_pids);
+	} else if (pid_list) {
+		/* Register a probe to set whether to ignore the tracing of a task */
+		register_trace_sched_switch(ftrace_filter_pid_sched_switch_probe, tr);
+	}
 
 	/*
-	 * Allow "echo > set_ftrace_pid" or "echo -n '' > set_ftrace_pid"
+	 * Ignoring of pids is done at task switch. But we have to
-	 * to clean the filter quietly.
+	 * check for those tasks that are currently running.
+	 * Always do this in case a pid was appended or removed.
 	 */
-	tmp = strstrip(buf);
+	on_each_cpu(ignore_task_cpu, tr, 1);
-	if (strlen(tmp) == 0)
-		return 1;
 
-	ret = kstrtol(tmp, 10, &val);
+	ftrace_update_pid_func();
-	if (ret < 0)
+	ftrace_startup_all(0);
-		return ret;
+ out:
+	mutex_unlock(&ftrace_lock);
 
-	ret = ftrace_pid_add(val);
+	if (ret > 0)
+		*ppos += ret;
 
-	return ret ? ret : cnt;
+	return ret;
 }
 
 static int
 ftrace_pid_release(struct inode *inode, struct file *file)
 {
-	if (file->f_mode & FMODE_READ)
+	struct trace_array *tr = inode->i_private;
-		seq_release(inode, file);
 
-	return 0;
+	trace_array_put(tr);
+
+	return seq_release(inode, file);
 }
 
 static const struct file_operations ftrace_pid_fops = {
@@ -5571,24 +5537,17 @@ static const struct file_operations ftrace_pid_fops = {
 	.release	= ftrace_pid_release,
 };
 
-static __init int ftrace_init_tracefs(void)
+void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d_tracer)
 {
-	struct dentry *d_tracer;
+	/* Only the top level directory has the dyn_tracefs and profile */
-
+	if (tr->flags & TRACE_ARRAY_FL_GLOBAL) {
-	d_tracer = tracing_init_dentry();
+		ftrace_init_dyn_tracefs(d_tracer);
-	if (IS_ERR(d_tracer))
+		ftrace_profile_tracefs(d_tracer);
-		return 0;
+	}
-
-	ftrace_init_dyn_tracefs(d_tracer);
 
 	trace_create_file("set_ftrace_pid", 0644, d_tracer,
-			    NULL, &ftrace_pid_fops);
+			    tr, &ftrace_pid_fops);
-
-	ftrace_profile_tracefs(d_tracer);
-
-	return 0;
 }
-fs_initcall(ftrace_init_tracefs);
 
 /**
  * ftrace_kill - kill ftrace

kernel/trace/trace.c

@@ -7233,6 +7233,7 @@ init_tracer_tracefs(struct trace_array *tr, struct dentry *d_tracer)
 	for_each_tracing_cpu(cpu)
 		tracing_init_tracefs_percpu(tr, cpu);
 
+	ftrace_init_tracefs(tr, d_tracer);
 }
 
 static struct vfsmount *trace_automount(void *ingore)

kernel/trace/trace.h

@@ -156,6 +156,9 @@ struct trace_array_cpu {
 	char			comm[TASK_COMM_LEN];
 
 	bool			ignore_pid;
+#ifdef CONFIG_FUNCTION_TRACER
+	bool			ftrace_ignore_pid;
+#endif
 };
 
 struct tracer;
@@ -247,6 +250,7 @@ struct trace_array {
 	int			ref;
 #ifdef CONFIG_FUNCTION_TRACER
 	struct ftrace_ops	*ops;
+	struct trace_pid_list	__rcu *function_pids;
 	/* function tracing enabled */
 	int			function_enabled;
 #endif
@@ -840,12 +844,9 @@ extern struct list_head ftrace_pids;
 
 #ifdef CONFIG_FUNCTION_TRACER
 extern bool ftrace_filter_param __initdata;
-static inline int ftrace_trace_task(struct task_struct *task)
+static inline int ftrace_trace_task(struct trace_array *tr)
 {
-	if (list_empty(&ftrace_pids))
+	return !this_cpu_read(tr->trace_buffer.data->ftrace_ignore_pid);
-		return 1;
-
-	return test_tsk_trace_trace(task);
 }
 extern int ftrace_is_dead(void);
 int ftrace_create_function_files(struct trace_array *tr,
@@ -855,8 +856,9 @@ void ftrace_init_global_array_ops(struct trace_array *tr);
 void ftrace_init_array_ops(struct trace_array *tr, ftrace_func_t func);
 void ftrace_reset_array_ops(struct trace_array *tr);
 int using_ftrace_ops_list_func(void);
+void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d_tracer);
 #else
-static inline int ftrace_trace_task(struct task_struct *task)
+static inline int ftrace_trace_task(struct trace_array *tr)
 {
 	return 1;
 }
@@ -871,6 +873,7 @@ static inline void ftrace_destroy_function_files(struct trace_array *tr) { }
 static inline __init void
 ftrace_init_global_array_ops(struct trace_array *tr) { }
 static inline void ftrace_reset_array_ops(struct trace_array *tr) { }
+static inline void ftrace_init_tracefs(struct trace_array *tr, struct dentry *d) { }
 /* ftace_func_t type is not defined, use macro instead of static inline */
 #define ftrace_init_array_ops(tr, func) do { } while (0)
 #endif /* CONFIG_FUNCTION_TRACER */

kernel/trace/trace_functions.c

@@ -43,7 +43,7 @@ static int allocate_ftrace_ops(struct trace_array *tr)
 
 	/* Currently only the non stack verision is supported */
 	ops->func = function_trace_call;
-	ops->flags = FTRACE_OPS_FL_RECURSION_SAFE;
+	ops->flags = FTRACE_OPS_FL_RECURSION_SAFE | FTRACE_OPS_FL_PID;
 
 	tr->ops = ops;
 	ops->private = tr;

kernel/trace/trace_functions_graph.c

@@ -319,7 +319,7 @@ int trace_graph_entry(struct ftrace_graph_ent *trace)
 	int cpu;
 	int pc;
 
-	if (!ftrace_trace_task(current))
+	if (!ftrace_trace_task(tr))
 		return 0;
 
 	/* trace it when it is-nested-in or is a function enabled. */