Merge branch 'tip/perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace into perf/core

Documentation/kernel-parameters.txt

@@ -2437,6 +2437,14 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 	stacktrace	[FTRACE]
 			Enabled the stack tracer on boot up.
 
+	stacktrace_filter=[function-list]
+			[FTRACE] Limit the functions that the stack tracer
+			will trace at boot up. function-list is a comma separated
+			list of functions. This list can be changed at run
+			time by the stack_trace_filter file in the debugfs
+			tracing directory. Note, this enables stack tracing
+			and the stacktrace above is not needed.
+
 	sti=		[PARISC,HW]
 			Format: <num>
 			Set the STI (builtin display/keyboard on the HP-PARISC

include/linux/compiler-gcc.h

@@ -50,6 +50,11 @@
 # define inline		inline		__attribute__((always_inline))
 # define __inline__	__inline__	__attribute__((always_inline))
 # define __inline	__inline	__attribute__((always_inline))
+#else
+/* A lot of inline functions can cause havoc with function tracing */
+# define inline		inline		notrace
+# define __inline__	__inline__	notrace
+# define __inline	__inline	notrace
 #endif
 
 #define __deprecated			__attribute__((deprecated))

include/linux/ftrace.h

@@ -133,6 +133,8 @@ struct ftrace_func_command {
 int ftrace_arch_code_modify_prepare(void);
 int ftrace_arch_code_modify_post_process(void);
 
+void ftrace_bug(int err, unsigned long ip);
+
 struct seq_file;
 
 struct ftrace_probe_ops {
@@ -161,7 +163,6 @@ extern int ftrace_text_reserved(void *start, void *end);
 
 enum {
 	FTRACE_FL_ENABLED	= (1 << 30),
-	FTRACE_FL_FREE		= (1 << 31),
 };
 
 #define FTRACE_FL_MASK		(0x3UL << 30)
@@ -172,10 +173,7 @@ struct dyn_ftrace {
 		unsigned long		ip; /* address of mcount call-site */
 		struct dyn_ftrace	*freelist;
 	};
-	union {
+	unsigned long		flags;
-		unsigned long		flags;
-		struct dyn_ftrace	*newlist;
-	};
 	struct dyn_arch_ftrace		arch;
 };
 
@@ -190,6 +188,56 @@ void ftrace_set_global_notrace(unsigned char *buf, int len, int reset);
 int register_ftrace_command(struct ftrace_func_command *cmd);
 int unregister_ftrace_command(struct ftrace_func_command *cmd);
 
+enum {
+	FTRACE_UPDATE_CALLS		= (1 << 0),
+	FTRACE_DISABLE_CALLS		= (1 << 1),
+	FTRACE_UPDATE_TRACE_FUNC	= (1 << 2),
+	FTRACE_START_FUNC_RET		= (1 << 3),
+	FTRACE_STOP_FUNC_RET		= (1 << 4),
+};
+
+enum {
+	FTRACE_UPDATE_IGNORE,
+	FTRACE_UPDATE_MAKE_CALL,
+	FTRACE_UPDATE_MAKE_NOP,
+};
+
+enum {
+	FTRACE_ITER_FILTER	= (1 << 0),
+	FTRACE_ITER_NOTRACE	= (1 << 1),
+	FTRACE_ITER_PRINTALL	= (1 << 2),
+	FTRACE_ITER_DO_HASH	= (1 << 3),
+	FTRACE_ITER_HASH	= (1 << 4),
+	FTRACE_ITER_ENABLED	= (1 << 5),
+};
+
+void arch_ftrace_update_code(int command);
+
+struct ftrace_rec_iter;
+
+struct ftrace_rec_iter *ftrace_rec_iter_start(void);
+struct ftrace_rec_iter *ftrace_rec_iter_next(struct ftrace_rec_iter *iter);
+struct dyn_ftrace *ftrace_rec_iter_record(struct ftrace_rec_iter *iter);
+
+int ftrace_update_record(struct dyn_ftrace *rec, int enable);
+int ftrace_test_record(struct dyn_ftrace *rec, int enable);
+void ftrace_run_stop_machine(int command);
+int ftrace_location(unsigned long ip);
+
+extern ftrace_func_t ftrace_trace_function;
+
+int ftrace_regex_open(struct ftrace_ops *ops, int flag,
+		  struct inode *inode, struct file *file);
+ssize_t ftrace_filter_write(struct file *file, const char __user *ubuf,
+			    size_t cnt, loff_t *ppos);
+ssize_t ftrace_notrace_write(struct file *file, const char __user *ubuf,
+			     size_t cnt, loff_t *ppos);
+loff_t ftrace_regex_lseek(struct file *file, loff_t offset, int origin);
+int ftrace_regex_release(struct inode *inode, struct file *file);
+
+void __init
+ftrace_set_early_filter(struct ftrace_ops *ops, char *buf, int enable);
+
 /* defined in arch */
 extern int ftrace_ip_converted(unsigned long ip);
 extern int ftrace_dyn_arch_init(void *data);
@@ -284,6 +332,25 @@ static inline int ftrace_text_reserved(void *start, void *end)
 {
 	return 0;
 }
+
+/*
+ * Again users of functions that have ftrace_ops may not
+ * have them defined when ftrace is not enabled, but these
+ * functions may still be called. Use a macro instead of inline.
+ */
+#define ftrace_regex_open(ops, flag, inod, file) ({ -ENODEV; })
+#define ftrace_set_early_filter(ops, buf, enable) do { } while (0)
+
+static inline ssize_t ftrace_filter_write(struct file *file, const char __user *ubuf,
+			    size_t cnt, loff_t *ppos) { return -ENODEV; }
+static inline ssize_t ftrace_notrace_write(struct file *file, const char __user *ubuf,
+			     size_t cnt, loff_t *ppos) { return -ENODEV; }
+static inline loff_t ftrace_regex_lseek(struct file *file, loff_t offset, int origin)
+{
+	return -ENODEV;
+}
+static inline int
+ftrace_regex_release(struct inode *inode, struct file *file) { return -ENODEV; }
 #endif /* CONFIG_DYNAMIC_FTRACE */
 
 /* totally disable ftrace - can not re-enable after this */

kernel/trace/trace_events_filter.c

@@ -1738,11 +1738,121 @@ static int replace_system_preds(struct event_subsystem *system,
 	return -ENOMEM;
 }
 
+static int create_filter_start(char *filter_str, bool set_str,
+			       struct filter_parse_state **psp,
+			       struct event_filter **filterp)
+{
+	struct event_filter *filter;
+	struct filter_parse_state *ps = NULL;
+	int err = 0;
+
+	WARN_ON_ONCE(*psp || *filterp);
+
+	/* allocate everything, and if any fails, free all and fail */
+	filter = __alloc_filter();
+	if (filter && set_str)
+		err = replace_filter_string(filter, filter_str);
+
+	ps = kzalloc(sizeof(*ps), GFP_KERNEL);
+
+	if (!filter || !ps || err) {
+		kfree(ps);
+		__free_filter(filter);
+		return -ENOMEM;
+	}
+
+	/* we're committed to creating a new filter */
+	*filterp = filter;
+	*psp = ps;
+
+	parse_init(ps, filter_ops, filter_str);
+	err = filter_parse(ps);
+	if (err && set_str)
+		append_filter_err(ps, filter);
+	return err;
+}
+
+static void create_filter_finish(struct filter_parse_state *ps)
+{
+	if (ps) {
+		filter_opstack_clear(ps);
+		postfix_clear(ps);
+		kfree(ps);
+	}
+}
+
+/**
+ * create_filter - create a filter for a ftrace_event_call
+ * @call: ftrace_event_call to create a filter for
+ * @filter_str: filter string
+ * @set_str: remember @filter_str and enable detailed error in filter
+ * @filterp: out param for created filter (always updated on return)
+ *
+ * Creates a filter for @call with @filter_str.  If @set_str is %true,
+ * @filter_str is copied and recorded in the new filter.
+ *
+ * On success, returns 0 and *@filterp points to the new filter.  On
+ * failure, returns -errno and *@filterp may point to %NULL or to a new
+ * filter.  In the latter case, the returned filter contains error
+ * information if @set_str is %true and the caller is responsible for
+ * freeing it.
+ */
+static int create_filter(struct ftrace_event_call *call,
+			 char *filter_str, bool set_str,
+			 struct event_filter **filterp)
+{
+	struct event_filter *filter = NULL;
+	struct filter_parse_state *ps = NULL;
+	int err;
+
+	err = create_filter_start(filter_str, set_str, &ps, &filter);
+	if (!err) {
+		err = replace_preds(call, filter, ps, filter_str, false);
+		if (err && set_str)
+			append_filter_err(ps, filter);
+	}
+	create_filter_finish(ps);
+
+	*filterp = filter;
+	return err;
+}
+
+/**
+ * create_system_filter - create a filter for an event_subsystem
+ * @system: event_subsystem to create a filter for
+ * @filter_str: filter string
+ * @filterp: out param for created filter (always updated on return)
+ *
+ * Identical to create_filter() except that it creates a subsystem filter
+ * and always remembers @filter_str.
+ */
+static int create_system_filter(struct event_subsystem *system,
+				char *filter_str, struct event_filter **filterp)
+{
+	struct event_filter *filter = NULL;
+	struct filter_parse_state *ps = NULL;
+	int err;
+
+	err = create_filter_start(filter_str, true, &ps, &filter);
+	if (!err) {
+		err = replace_system_preds(system, ps, filter_str);
+		if (!err) {
+			/* System filters just show a default message */
+			kfree(filter->filter_string);
+			filter->filter_string = NULL;
+		} else {
+			append_filter_err(ps, filter);
+		}
+	}
+	create_filter_finish(ps);
+
+	*filterp = filter;
+	return err;
+}
+
 int apply_event_filter(struct ftrace_event_call *call, char *filter_string)
 {
-	struct filter_parse_state *ps;
 	struct event_filter *filter;
-	struct event_filter *tmp;
 	int err = 0;
 
 	mutex_lock(&event_mutex);
@@ -1759,49 +1869,30 @@ int apply_event_filter(struct ftrace_event_call *call, char *filter_string)
 		goto out_unlock;
 	}
 
-	err = -ENOMEM;
+	err = create_filter(call, filter_string, true, &filter);
-	ps = kzalloc(sizeof(*ps), GFP_KERNEL);
-	if (!ps)
-		goto out_unlock;
 
-	filter = __alloc_filter();
-	if (!filter) {
-		kfree(ps);
-		goto out_unlock;
-	}
-
-	replace_filter_string(filter, filter_string);
-
-	parse_init(ps, filter_ops, filter_string);
-	err = filter_parse(ps);
-	if (err) {
-		append_filter_err(ps, filter);
-		goto out;
-	}
-
-	err = replace_preds(call, filter, ps, filter_string, false);
-	if (err) {
-		filter_disable(call);
-		append_filter_err(ps, filter);
-	} else
-		call->flags |= TRACE_EVENT_FL_FILTERED;
-out:
 	/*
 	 * Always swap the call filter with the new filter
 	 * even if there was an error. If there was an error
 	 * in the filter, we disable the filter and show the error
 	 * string
 	 */
-	tmp = call->filter;
+	if (filter) {
-	rcu_assign_pointer(call->filter, filter);
+		struct event_filter *tmp = call->filter;
-	if (tmp) {
+
-		/* Make sure the call is done with the filter */
+		if (!err)
-		synchronize_sched();
+			call->flags |= TRACE_EVENT_FL_FILTERED;
-		__free_filter(tmp);
+		else
+			filter_disable(call);
+
+		rcu_assign_pointer(call->filter, filter);
+
+		if (tmp) {
+			/* Make sure the call is done with the filter */
+			synchronize_sched();
+			__free_filter(tmp);
+		}
 	}
-	filter_opstack_clear(ps);
-	postfix_clear(ps);
-	kfree(ps);
 out_unlock:
 	mutex_unlock(&event_mutex);
 
@@ -1811,7 +1902,6 @@ out_unlock:
 int apply_subsystem_event_filter(struct event_subsystem *system,
 				 char *filter_string)
 {
-	struct filter_parse_state *ps;
 	struct event_filter *filter;
 	int err = 0;
 
@@ -1835,48 +1925,19 @@ int apply_subsystem_event_filter(struct event_subsystem *system,
 		goto out_unlock;
 	}
 
-	err = -ENOMEM;
+	err = create_system_filter(system, filter_string, &filter);
-	ps = kzalloc(sizeof(*ps), GFP_KERNEL);
+	if (filter) {
-	if (!ps)
+		/*
-		goto out_unlock;
+		 * No event actually uses the system filter
-
+		 * we can free it without synchronize_sched().
-	filter = __alloc_filter();
+		 */
-	if (!filter)
+		__free_filter(system->filter);
-		goto out;
+		system->filter = filter;
-
+	}
-	/* System filters just show a default message */
-	kfree(filter->filter_string);
-	filter->filter_string = NULL;
-
-	/*
-	 * No event actually uses the system filter
-	 * we can free it without synchronize_sched().
-	 */
-	__free_filter(system->filter);
-	system->filter = filter;
-
-	parse_init(ps, filter_ops, filter_string);
-	err = filter_parse(ps);
-	if (err)
-		goto err_filter;
-
-	err = replace_system_preds(system, ps, filter_string);
-	if (err)
-		goto err_filter;
-
-out:
-	filter_opstack_clear(ps);
-	postfix_clear(ps);
-	kfree(ps);
 out_unlock:
 	mutex_unlock(&event_mutex);
 
 	return err;
-
-err_filter:
-	replace_filter_string(filter, filter_string);
-	append_filter_err(ps, system->filter);
-	goto out;
 }
 
 #ifdef CONFIG_PERF_EVENTS
@@ -1894,7 +1955,6 @@ int ftrace_profile_set_filter(struct perf_event *event, int event_id,
 {
 	int err;
 	struct event_filter *filter;
-	struct filter_parse_state *ps;
 	struct ftrace_event_call *call;
 
 	mutex_lock(&event_mutex);
@@ -1909,33 +1969,10 @@ int ftrace_profile_set_filter(struct perf_event *event, int event_id,
 	if (event->filter)
 		goto out_unlock;
 
-	filter = __alloc_filter();
+	err = create_filter(call, filter_str, false, &filter);
-	if (!filter) {
-		err = PTR_ERR(filter);
-		goto out_unlock;
-	}
-
-	err = -ENOMEM;
-	ps = kzalloc(sizeof(*ps), GFP_KERNEL);
-	if (!ps)
-		goto free_filter;
-
-	parse_init(ps, filter_ops, filter_str);
-	err = filter_parse(ps);
-	if (err)
-		goto free_ps;
-
-	err = replace_preds(call, filter, ps, filter_str, false);
 	if (!err)
 		event->filter = filter;
-
+	else
-free_ps:
-	filter_opstack_clear(ps);
-	postfix_clear(ps);
-	kfree(ps);
-
-free_filter:
-	if (err)
 		__free_filter(filter);
 
 out_unlock:
@@ -1954,43 +1991,6 @@ out_unlock:
 #define CREATE_TRACE_POINTS
 #include "trace_events_filter_test.h"
 
-static int test_get_filter(char *filter_str, struct ftrace_event_call *call,
-			   struct event_filter **pfilter)
-{
-	struct event_filter *filter;
-	struct filter_parse_state *ps;
-	int err = -ENOMEM;
-
-	filter = __alloc_filter();
-	if (!filter)
-		goto out;
-
-	ps = kzalloc(sizeof(*ps), GFP_KERNEL);
-	if (!ps)
-		goto free_filter;
-
-	parse_init(ps, filter_ops, filter_str);
-	err = filter_parse(ps);
-	if (err)
-		goto free_ps;
-
-	err = replace_preds(call, filter, ps, filter_str, false);
-	if (!err)
-		*pfilter = filter;
-
- free_ps:
-	filter_opstack_clear(ps);
-	postfix_clear(ps);
-	kfree(ps);
-
- free_filter:
-	if (err)
-		__free_filter(filter);
-
- out:
-	return err;
-}
-
 #define DATA_REC(m, va, vb, vc, vd, ve, vf, vg, vh, nvisit) \
 { \
 	.filter = FILTER, \
@@ -2109,12 +2109,13 @@ static __init int ftrace_test_event_filter(void)
 		struct test_filter_data_t *d = &test_filter_data[i];
 		int err;
 
-		err = test_get_filter(d->filter, &event_ftrace_test_filter,
+		err = create_filter(&event_ftrace_test_filter, d->filter,
-				      &filter);
+				    false, &filter);
 		if (err) {
 			printk(KERN_INFO
 			       "Failed to get filter for '%s', err %d\n",
 			       d->filter, err);
+			__free_filter(filter);
 			break;
 		}
 

kernel/trace/trace_stack.c

@@ -13,6 +13,9 @@
 #include <linux/sysctl.h>
 #include <linux/init.h>
 #include <linux/fs.h>
+
+#include <asm/setup.h>
+
 #include "trace.h"
 
 #define STACK_TRACE_ENTRIES 500
@@ -133,7 +136,6 @@ stack_trace_call(unsigned long ip, unsigned long parent_ip)
 static struct ftrace_ops trace_ops __read_mostly =
 {
 	.func = stack_trace_call,
-	.flags = FTRACE_OPS_FL_GLOBAL,
 };
 
 static ssize_t
@@ -311,6 +313,21 @@ static const struct file_operations stack_trace_fops = {
 	.release	= seq_release,
 };
 
+static int
+stack_trace_filter_open(struct inode *inode, struct file *file)
+{
+	return ftrace_regex_open(&trace_ops, FTRACE_ITER_FILTER,
+				 inode, file);
+}
+
+static const struct file_operations stack_trace_filter_fops = {
+	.open = stack_trace_filter_open,
+	.read = seq_read,
+	.write = ftrace_filter_write,
+	.llseek = ftrace_regex_lseek,
+	.release = ftrace_regex_release,
+};
+
 int
 stack_trace_sysctl(struct ctl_table *table, int write,
 		   void __user *buffer, size_t *lenp,
@@ -338,8 +355,13 @@ stack_trace_sysctl(struct ctl_table *table, int write,
 	return ret;
 }
 
+static char stack_trace_filter_buf[COMMAND_LINE_SIZE+1] __initdata;
+
 static __init int enable_stacktrace(char *str)
 {
+	if (strncmp(str, "_filter=", 8) == 0)
+		strncpy(stack_trace_filter_buf, str+8, COMMAND_LINE_SIZE);
+
 	stack_tracer_enabled = 1;
 	last_stack_tracer_enabled = 1;
 	return 1;
@@ -358,6 +380,12 @@ static __init int stack_trace_init(void)
 	trace_create_file("stack_trace", 0444, d_tracer,
 			NULL, &stack_trace_fops);
 
+	trace_create_file("stack_trace_filter", 0444, d_tracer,
+			NULL, &stack_trace_filter_fops);
+
+	if (stack_trace_filter_buf[0])
+		ftrace_set_early_filter(&trace_ops, stack_trace_filter_buf, 1);
+
 	if (stack_tracer_enabled)
 		register_ftrace_function(&trace_ops);