Merge branch 'tracing/sysprof' into auto-ftrace-next

6 files changed, 404 insertions, 0 deletions

diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index 5c2295b..263e9e6 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -75,6 +75,14 @@ config PREEMPT_TRACER
 	  enabled. This option and the irqs-off timing option can be
 	  used together or separately.)
 
+config SYSPROF_TRACER
+	bool "Sysprof Tracer"
+	depends on X86
+	select TRACING
+	help
+	  This tracer provides the trace needed by the 'Sysprof' userspace
+	  tool.
+
 config SCHED_TRACER
 	bool "Scheduling Latency Tracer"
 	depends on HAVE_FTRACE
diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile
index c44a7dc..71d17de 100644
--- a/kernel/trace/Makefile
+++ b/kernel/trace/Makefile
@@ -14,6 +14,7 @@ obj-$(CONFIG_FTRACE) += libftrace.o
 
 obj-$(CONFIG_TRACING) += trace.o
 obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o
+obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o
 obj-$(CONFIG_FTRACE) += trace_functions.o
 obj-$(CONFIG_IRQSOFF_TRACER) += trace_irqsoff.o
 obj-$(CONFIG_PREEMPT_TRACER) += trace_irqsoff.o
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 4a87560..e46de64 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -2913,6 +2913,9 @@ static __init void tracer_init_debugfs(void)
 		pr_warning("Could not create debugfs "
 			   "'dyn_ftrace_total_info' entry\n");
 #endif
+#ifdef CONFIG_SYSPROF_TRACER
+	init_tracer_sysprof_debugfs(d_tracer);
+#endif
 }
 
 static int trace_alloc_page(void)
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 4966e6a..8cb215b 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -193,6 +193,8 @@ struct trace_iterator {
 void tracing_reset(struct trace_array_cpu *data);
 int tracing_open_generic(struct inode *inode, struct file *filp);
 struct dentry *tracing_init_dentry(void);
+void init_tracer_sysprof_debugfs(struct dentry *d_tracer);
+
 void ftrace(struct trace_array *tr,
 			    struct trace_array_cpu *data,
 			    unsigned long ip,
@@ -294,6 +296,10 @@ extern int trace_selftest_startup_wakeup(struct tracer *trace,
 extern int trace_selftest_startup_sched_switch(struct tracer *trace,
 					       struct trace_array *tr);
 #endif
+#ifdef CONFIG_SYSPROF_TRACER
+extern int trace_selftest_startup_sysprof(struct tracer *trace,
+					       struct trace_array *tr);
+#endif
 #endif /* CONFIG_FTRACE_STARTUP_TEST */
 
 extern void *head_page(struct trace_array_cpu *data);
diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c
index 18c5423..0911b7e 100644
--- a/kernel/trace/trace_selftest.c
+++ b/kernel/trace/trace_selftest.c
@@ -538,3 +538,26 @@ trace_selftest_startup_sched_switch(struct tracer *trace, struct trace_array *tr
 	return ret;
 }
 #endif /* CONFIG_CONTEXT_SWITCH_TRACER */
+
+#ifdef CONFIG_SYSPROF_TRACER
+int
+trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long count;
+	int ret;
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* Sleep for a 1/10 of a second */
+	msleep(100);
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check the trace buffer */
+	ret = trace_test_buffer(tr, &count);
+	trace->reset(tr);
+
+	return ret;
+}
+#endif /* CONFIG_SYSPROF_TRACER */
diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c
new file mode 100644
index 0000000..2301e1e
--- /dev/null
+++ b/kernel/trace/trace_sysprof.c
@@ -0,0 +1,363 @@
+/*
+ * trace stack traces
+ *
+ * Copyright (C) 2004-2008, Soeren Sandmann
+ * Copyright (C) 2007 Steven Rostedt <srostedt@redhat.com>
+ * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
+ */
+#include <linux/kallsyms.h>
+#include <linux/debugfs.h>
+#include <linux/hrtimer.h>
+#include <linux/uaccess.h>
+#include <linux/ftrace.h>
+#include <linux/module.h>
+#include <linux/irq.h>
+#include <linux/fs.h>
+
+#include <asm/stacktrace.h>
+
+#include "trace.h"
+
+static struct trace_array	*sysprof_trace;
+static int __read_mostly	tracer_enabled;
+
+/*
+ * 1 msec sample interval by default:
+ */
+static unsigned long sample_period = 1000000;
+static const unsigned int sample_max_depth = 512;
+
+static DEFINE_MUTEX(sample_timer_lock);
+/*
+ * Per CPU hrtimers that do the profiling:
+ */
+static DEFINE_PER_CPU(struct hrtimer, stack_trace_hrtimer);
+
+struct stack_frame {
+	const void __user	*next_fp;
+	unsigned long		return_address;
+};
+
+static int copy_stack_frame(const void __user *fp, struct stack_frame *frame)
+{
+	int ret;
+
+	if (!access_ok(VERIFY_READ, fp, sizeof(*frame)))
+		return 0;
+
+	ret = 1;
+	pagefault_disable();
+	if (__copy_from_user_inatomic(frame, fp, sizeof(*frame)))
+		ret = 0;
+	pagefault_enable();
+
+	return ret;
+}
+
+struct backtrace_info {
+	struct trace_array_cpu	*data;
+	struct trace_array	*tr;
+	int			pos;
+};
+
+static void
+backtrace_warning_symbol(void *data, char *msg, unsigned long symbol)
+{
+	/* Ignore warnings */
+}
+
+static void backtrace_warning(void *data, char *msg)
+{
+	/* Ignore warnings */
+}
+
+static int backtrace_stack(void *data, char *name)
+{
+	/* Don't bother with IRQ stacks for now */
+	return -1;
+}
+
+static void backtrace_address(void *data, unsigned long addr, int reliable)
+{
+	struct backtrace_info *info = data;
+
+	if (info->pos < sample_max_depth && reliable) {
+		__trace_special(info->tr, info->data, 1, addr, 0);
+
+		info->pos++;
+	}
+}
+
+const static struct stacktrace_ops backtrace_ops = {
+	.warning		= backtrace_warning,
+	.warning_symbol		= backtrace_warning_symbol,
+	.stack			= backtrace_stack,
+	.address		= backtrace_address,
+};
+
+static int
+trace_kernel(struct pt_regs *regs, struct trace_array *tr,
+	     struct trace_array_cpu *data)
+{
+	struct backtrace_info info;
+	unsigned long bp;
+	char *stack;
+
+	info.tr = tr;
+	info.data = data;
+	info.pos = 1;
+
+	__trace_special(info.tr, info.data, 1, regs->ip, 0);
+
+	stack = ((char *)regs + sizeof(struct pt_regs));
+#ifdef CONFIG_FRAME_POINTER
+	bp = regs->bp;
+#else
+	bp = 0;
+#endif
+
+	dump_trace(NULL, regs, (void *)stack, bp, &backtrace_ops, &info);
+
+	return info.pos;
+}
+
+static void timer_notify(struct pt_regs *regs, int cpu)
+{
+	struct trace_array_cpu *data;
+	struct stack_frame frame;
+	struct trace_array *tr;
+	const void __user *fp;
+	int is_user;
+	int i;
+
+	if (!regs)
+		return;
+
+	tr = sysprof_trace;
+	data = tr->data[cpu];
+	is_user = user_mode(regs);
+
+	if (!current || current->pid == 0)
+		return;
+
+	if (is_user && current->state != TASK_RUNNING)
+		return;
+
+	__trace_special(tr, data, 0, 0, current->pid);
+
+	if (!is_user)
+		i = trace_kernel(regs, tr, data);
+	else
+		i = 0;
+
+	/*
+	 * Trace user stack if we are not a kernel thread
+	 */
+	if (current->mm && i < sample_max_depth) {
+		regs = (struct pt_regs *)current->thread.sp0 - 1;
+
+		fp = (void __user *)regs->bp;
+
+		__trace_special(tr, data, 2, regs->ip, 0);
+
+		while (i < sample_max_depth) {
+			frame.next_fp = 0;
+			frame.return_address = 0;
+			if (!copy_stack_frame(fp, &frame))
+				break;
+			if ((unsigned long)fp < regs->sp)
+				break;
+
+			__trace_special(tr, data, 2, frame.return_address,
+					(unsigned long)fp);
+			fp = frame.next_fp;
+
+			i++;
+		}
+
+	}
+
+	/*
+	 * Special trace entry if we overflow the max depth:
+	 */
+	if (i == sample_max_depth)
+		__trace_special(tr, data, -1, -1, -1);
+
+	__trace_special(tr, data, 3, current->pid, i);
+}
+
+static enum hrtimer_restart stack_trace_timer_fn(struct hrtimer *hrtimer)
+{
+	/* trace here */
+	timer_notify(get_irq_regs(), smp_processor_id());
+
+	hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
+
+	return HRTIMER_RESTART;
+}
+
+static void start_stack_timer(int cpu)
+{
+	struct hrtimer *hrtimer = &per_cpu(stack_trace_hrtimer, cpu);
+
+	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	hrtimer->function = stack_trace_timer_fn;
+	hrtimer->cb_mode = HRTIMER_CB_IRQSAFE_NO_SOFTIRQ;
+
+	hrtimer_start(hrtimer, ns_to_ktime(sample_period), HRTIMER_MODE_REL);
+}
+
+static void start_stack_timers(void)
+{
+	cpumask_t saved_mask = current->cpus_allowed;
+	int cpu;
+
+	for_each_online_cpu(cpu) {
+		set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
+		start_stack_timer(cpu);
+	}
+	set_cpus_allowed_ptr(current, &saved_mask);
+}
+
+static void stop_stack_timer(int cpu)
+{
+	struct hrtimer *hrtimer = &per_cpu(stack_trace_hrtimer, cpu);
+
+	hrtimer_cancel(hrtimer);
+}
+
+static void stop_stack_timers(void)
+{
+	int cpu;
+
+	for_each_online_cpu(cpu)
+		stop_stack_timer(cpu);
+}
+
+static void stack_reset(struct trace_array *tr)
+{
+	int cpu;
+
+	tr->time_start = ftrace_now(tr->cpu);
+
+	for_each_online_cpu(cpu)
+		tracing_reset(tr->data[cpu]);
+}
+
+static void start_stack_trace(struct trace_array *tr)
+{
+	mutex_lock(&sample_timer_lock);
+	stack_reset(tr);
+	start_stack_timers();
+	tracer_enabled = 1;
+	mutex_unlock(&sample_timer_lock);
+}
+
+static void stop_stack_trace(struct trace_array *tr)
+{
+	mutex_lock(&sample_timer_lock);
+	stop_stack_timers();
+	tracer_enabled = 0;
+	mutex_unlock(&sample_timer_lock);
+}
+
+static void stack_trace_init(struct trace_array *tr)
+{
+	sysprof_trace = tr;
+
+	if (tr->ctrl)
+		start_stack_trace(tr);
+}
+
+static void stack_trace_reset(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		stop_stack_trace(tr);
+}
+
+static void stack_trace_ctrl_update(struct trace_array *tr)
+{
+	/* When starting a new trace, reset the buffers */
+	if (tr->ctrl)
+		start_stack_trace(tr);
+	else
+		stop_stack_trace(tr);
+}
+
+static struct tracer stack_trace __read_mostly =
+{
+	.name		= "sysprof",
+	.init		= stack_trace_init,
+	.reset		= stack_trace_reset,
+	.ctrl_update	= stack_trace_ctrl_update,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_sysprof,
+#endif
+};
+
+__init static int init_stack_trace(void)
+{
+	return register_tracer(&stack_trace);
+}
+device_initcall(init_stack_trace);
+
+#define MAX_LONG_DIGITS 22
+
+static ssize_t
+sysprof_sample_read(struct file *filp, char __user *ubuf,
+		    size_t cnt, loff_t *ppos)
+{
+	char buf[MAX_LONG_DIGITS];
+	int r;
+
+	r = sprintf(buf, "%ld\n", nsecs_to_usecs(sample_period));
+
+	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
+}
+
+static ssize_t
+sysprof_sample_write(struct file *filp, const char __user *ubuf,
+		     size_t cnt, loff_t *ppos)
+{
+	char buf[MAX_LONG_DIGITS];
+	unsigned long val;
+
+	if (cnt > MAX_LONG_DIGITS-1)
+		cnt = MAX_LONG_DIGITS-1;
+
+	if (copy_from_user(&buf, ubuf, cnt))
+		return -EFAULT;
+
+	buf[cnt] = 0;
+
+	val = simple_strtoul(buf, NULL, 10);
+	/*
+	 * Enforce a minimum sample period of 100 usecs:
+	 */
+	if (val < 100)
+		val = 100;
+
+	mutex_lock(&sample_timer_lock);
+	stop_stack_timers();
+	sample_period = val * 1000;
+	start_stack_timers();
+	mutex_unlock(&sample_timer_lock);
+
+	return cnt;
+}
+
+static struct file_operations sysprof_sample_fops = {
+	.read		= sysprof_sample_read,
+	.write		= sysprof_sample_write,
+};
+
+void init_tracer_sysprof_debugfs(struct dentry *d_tracer)
+{
+	struct dentry *entry;
+
+	entry = debugfs_create_file("sysprof_sample_period", 0644,
+			d_tracer, NULL, &sysprof_sample_fops);
+	if (entry)
+		return;
+	pr_warning("Could not create debugfs 'dyn_ftrace_total_info' entry\n");
+}