[PATCH 00/14] [GIT PULL] tracing: fixes, clean ups and optimizations

[PATCH 00/14] [GIT PULL] tracing: fixes, clean ups and optimizations

[Steven Rostedt]

Linus,

This contains fixes, optimizations and some clean ups

Some of the fixes need to go back to 3.10. They are minor, and deal mostly
with incorrect ref counting in accessing event files.

There was a couple of optimizations that should have perf perform a bit
better when accessing trace events.

And some various clean ups. Some of the clean ups are necessary to help
in a fix to a theoretical race between opening a event file and
deleting that event.  That fix isn't there yet, we are still working on
a clean solution.

Please pull the latest trace-3.11-rc2 tree, which can be found at:

  git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git
trace-3.11-rc2

Tag SHA1: 8876f82e516564d3581e1572f10d5f229beed1b8
Head SHA1: e70e78e3c83b536730e31231dd9b979768d8df3c


Alexander Z Lam (2):
      tracing: Fix error handling to ensure instances can always be removed
      tracing: Miscellaneous fixes for trace_array ref counting

Masami Hiramatsu (1):
      tracing/kprobe: Wait for disabling all running kprobe handlers

Oleg Nesterov (7):
      tracing/function: Avoid perf_trace_buf_*() if event_function.perf_events is empty
      tracing/syscall: Avoid perf_trace_buf_*() if sys_data->perf_events is empty
      tracing/perf: Move the PERF_MAX_TRACE_SIZE check into perf_trace_buf_prepare()
      tracing: Simplify the iteration logic in f_start/f_next
      tracing: Do not (ab)use trace_seq in event_id_read()
      tracing: Kill trace_array->waiter
      tracing: Kill the unbalanced tr->ref++ in tracing_buffers_open()

Steven Rostedt (Red Hat) (1):
      tracing: Add ref_data to function and fgraph tracer structs

zhangwei(Jovi) (3):
      tracing: Use correct config guard CONFIG_STACK_TRACER
      tracing: Use trace_seq_puts()/trace_seq_putc() where possible
      tracing: Typo fix on ring buffer comments

----
 include/trace/ftrace.h               |    4 --
 kernel/trace/ring_buffer.c           |   26 ++++-----
 kernel/trace/trace.c                 |   34 ++++++++----
 kernel/trace/trace.h                 |   10 +++-
 kernel/trace/trace_event_perf.c      |   10 +++-
 kernel/trace/trace_events.c          |   98 ++++++++++++++++------------------
 kernel/trace/trace_events_filter.c   |    4 +-
 kernel/trace/trace_functions.c       |    2 +-
 kernel/trace/trace_functions_graph.c |   54 +++++++++----------
 kernel/trace/trace_kprobe.c          |   29 +++++-----
 kernel/trace/trace_mmiotrace.c       |    8 +--
 kernel/trace/trace_output.c          |   14 ++---
 kernel/trace/trace_syscalls.c        |   26 ++++-----
 kernel/trace/trace_uprobe.c          |    2 -
 14 files changed, 166 insertions(+), 155 deletions(-)

[PATCH 01/14] tracing: Use correct config guard CONFIG_STACK_TRACER

[Steven Rostedt]

[-- Attachment #1: 0001-tracing-Use-correct-config-guard-CONFIG_STACK_TRACER.patch --]
[-- Type: text/plain, Size: 1260 bytes --]

From: "zhangwei(Jovi)" <jovi.zhangwei@huawei.com>

We should use CONFIG_STACK_TRACER to guard readme text
of stack tracer related file, not CONFIG_STACKTRACE.

Link: http://lkml.kernel.org/r/51E3B3A2.8080609@huawei.com

Signed-off-by: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 0cd500b..25b91af 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -3537,14 +3537,14 @@ static const char readme_msg[] =
 	"\n  snapshot\t\t- Like 'trace' but shows the content of the static snapshot buffer\n"
 	"\t\t\t  Read the contents for more information\n"
 #endif
-#ifdef CONFIG_STACKTRACE
+#ifdef CONFIG_STACK_TRACER
 	"  stack_trace\t\t- Shows the max stack trace when active\n"
 	"  stack_max_size\t- Shows current max stack size that was traced\n"
 	"\t\t\t  Write into this file to reset the max size (trigger a new trace)\n"
 #ifdef CONFIG_DYNAMIC_FTRACE
 	"  stack_trace_filter\t- Like set_ftrace_filter but limits what stack_trace traces\n"
 #endif
-#endif /* CONFIG_STACKTRACE */
+#endif /* CONFIG_STACK_TRACER */
 ;
 
 static ssize_t
-- 
1.7.10.4

[PATCH 02/14] tracing: Use trace_seq_puts()/trace_seq_putc() where possible

[Steven Rostedt]

[-- Attachment #1: 0002-tracing-Use-trace_seq_puts-trace_seq_putc-where-poss.patch --]
[-- Type: text/plain, Size: 12933 bytes --]

From: "zhangwei(Jovi)" <jovi.zhangwei@huawei.com>

For string without format specifiers, use trace_seq_puts()
or trace_seq_putc().

Link: http://lkml.kernel.org/r/51E3B3AC.1000605@huawei.com

Signed-off-by: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
[ fixed a trace_seq_putc(s, " ") to trace_seq_putc(s, ' ') ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/ring_buffer.c           |   10 +++----
 kernel/trace/trace_events_filter.c   |    4 +--
 kernel/trace/trace_functions_graph.c |   52 +++++++++++++++++-----------------
 kernel/trace/trace_mmiotrace.c       |    8 +++---
 kernel/trace/trace_output.c          |   14 ++++-----
 kernel/trace/trace_syscalls.c        |    2 +-
 6 files changed, 45 insertions(+), 45 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index e444ff8..eef2e56 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -36,11 +36,11 @@ int ring_buffer_print_entry_header(struct trace_seq *s)
 {
 	int ret;
 
-	ret = trace_seq_printf(s, "# compressed entry header\n");
-	ret = trace_seq_printf(s, "\ttype_len    :    5 bits\n");
-	ret = trace_seq_printf(s, "\ttime_delta  :   27 bits\n");
-	ret = trace_seq_printf(s, "\tarray       :   32 bits\n");
-	ret = trace_seq_printf(s, "\n");
+	ret = trace_seq_puts(s, "# compressed entry header\n");
+	ret = trace_seq_puts(s, "\ttype_len    :    5 bits\n");
+	ret = trace_seq_puts(s, "\ttime_delta  :   27 bits\n");
+	ret = trace_seq_puts(s, "\tarray       :   32 bits\n");
+	ret = trace_seq_putc(s, '\n');
 	ret = trace_seq_printf(s, "\tpadding     : type == %d\n",
 			       RINGBUF_TYPE_PADDING);
 	ret = trace_seq_printf(s, "\ttime_extend : type == %d\n",
diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c
index 0d883dc..0c7b75a 100644
--- a/kernel/trace/trace_events_filter.c
+++ b/kernel/trace/trace_events_filter.c
@@ -646,7 +646,7 @@ void print_event_filter(struct ftrace_event_call *call, struct trace_seq *s)
 	if (filter && filter->filter_string)
 		trace_seq_printf(s, "%s\n", filter->filter_string);
 	else
-		trace_seq_printf(s, "none\n");
+		trace_seq_puts(s, "none\n");
 	mutex_unlock(&event_mutex);
 }
 
@@ -660,7 +660,7 @@ void print_subsystem_event_filter(struct event_subsystem *system,
 	if (filter && filter->filter_string)
 		trace_seq_printf(s, "%s\n", filter->filter_string);
 	else
-		trace_seq_printf(s, DEFAULT_SYS_FILTER_MESSAGE "\n");
+		trace_seq_puts(s, DEFAULT_SYS_FILTER_MESSAGE "\n");
 	mutex_unlock(&event_mutex);
 }
 
diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c
index 8388bc9..d56ae9b 100644
--- a/kernel/trace/trace_functions_graph.c
+++ b/kernel/trace/trace_functions_graph.c
@@ -446,7 +446,7 @@ print_graph_proc(struct trace_seq *s, pid_t pid)
 
 	/* First spaces to align center */
 	for (i = 0; i < spaces / 2; i++) {
-		ret = trace_seq_printf(s, " ");
+		ret = trace_seq_putc(s, ' ');
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -457,7 +457,7 @@ print_graph_proc(struct trace_seq *s, pid_t pid)
 
 	/* Last spaces to align center */
 	for (i = 0; i < spaces - (spaces / 2); i++) {
-		ret = trace_seq_printf(s, " ");
+		ret = trace_seq_putc(s, ' ');
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -503,7 +503,7 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data)
  ------------------------------------------
 
  */
-	ret = trace_seq_printf(s,
+	ret = trace_seq_puts(s,
 		" ------------------------------------------\n");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
@@ -516,7 +516,7 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data)
 	if (ret == TRACE_TYPE_PARTIAL_LINE)
 		return TRACE_TYPE_PARTIAL_LINE;
 
-	ret = trace_seq_printf(s, " => ");
+	ret = trace_seq_puts(s, " => ");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
@@ -524,7 +524,7 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data)
 	if (ret == TRACE_TYPE_PARTIAL_LINE)
 		return TRACE_TYPE_PARTIAL_LINE;
 
-	ret = trace_seq_printf(s,
+	ret = trace_seq_puts(s,
 		"\n ------------------------------------------\n\n");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
@@ -645,7 +645,7 @@ print_graph_irq(struct trace_iterator *iter, unsigned long addr,
 			ret = print_graph_proc(s, pid);
 			if (ret == TRACE_TYPE_PARTIAL_LINE)
 				return TRACE_TYPE_PARTIAL_LINE;
-			ret = trace_seq_printf(s, " | ");
+			ret = trace_seq_puts(s, " | ");
 			if (!ret)
 				return TRACE_TYPE_PARTIAL_LINE;
 		}
@@ -657,9 +657,9 @@ print_graph_irq(struct trace_iterator *iter, unsigned long addr,
 		return ret;
 
 	if (type == TRACE_GRAPH_ENT)
-		ret = trace_seq_printf(s, "==========>");
+		ret = trace_seq_puts(s, "==========>");
 	else
-		ret = trace_seq_printf(s, "<==========");
+		ret = trace_seq_puts(s, "<==========");
 
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
@@ -668,7 +668,7 @@ print_graph_irq(struct trace_iterator *iter, unsigned long addr,
 	if (ret != TRACE_TYPE_HANDLED)
 		return ret;
 
-	ret = trace_seq_printf(s, "\n");
+	ret = trace_seq_putc(s, '\n');
 
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
@@ -705,13 +705,13 @@ trace_print_graph_duration(unsigned long long duration, struct trace_seq *s)
 		len += strlen(nsecs_str);
 	}
 
-	ret = trace_seq_printf(s, " us ");
+	ret = trace_seq_puts(s, " us ");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
 	/* Print remaining spaces to fit the row's width */
 	for (i = len; i < 7; i++) {
-		ret = trace_seq_printf(s, " ");
+		ret = trace_seq_putc(s, ' ');
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -731,13 +731,13 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s,
 	/* No real adata, just filling the column with spaces */
 	switch (duration) {
 	case DURATION_FILL_FULL:
-		ret = trace_seq_printf(s, "              |  ");
+		ret = trace_seq_puts(s, "              |  ");
 		return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE;
 	case DURATION_FILL_START:
-		ret = trace_seq_printf(s, "  ");
+		ret = trace_seq_puts(s, "  ");
 		return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE;
 	case DURATION_FILL_END:
-		ret = trace_seq_printf(s, " |");
+		ret = trace_seq_puts(s, " |");
 		return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE;
 	}
 
@@ -745,10 +745,10 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s,
 	if (flags & TRACE_GRAPH_PRINT_OVERHEAD) {
 		/* Duration exceeded 100 msecs */
 		if (duration > 100000ULL)
-			ret = trace_seq_printf(s, "! ");
+			ret = trace_seq_puts(s, "! ");
 		/* Duration exceeded 10 msecs */
 		else if (duration > 10000ULL)
-			ret = trace_seq_printf(s, "+ ");
+			ret = trace_seq_puts(s, "+ ");
 	}
 
 	/*
@@ -757,7 +757,7 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s,
 	 * to fill out the space.
 	 */
 	if (ret == -1)
-		ret = trace_seq_printf(s, "  ");
+		ret = trace_seq_puts(s, "  ");
 
 	/* Catching here any failure happenned above */
 	if (!ret)
@@ -767,7 +767,7 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s,
 	if (ret != TRACE_TYPE_HANDLED)
 		return ret;
 
-	ret = trace_seq_printf(s, "|  ");
+	ret = trace_seq_puts(s, "|  ");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
@@ -817,7 +817,7 @@ print_graph_entry_leaf(struct trace_iterator *iter,
 
 	/* Function */
 	for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) {
-		ret = trace_seq_printf(s, " ");
+		ret = trace_seq_putc(s, ' ');
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -858,7 +858,7 @@ print_graph_entry_nested(struct trace_iterator *iter,
 
 	/* Function */
 	for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) {
-		ret = trace_seq_printf(s, " ");
+		ret = trace_seq_putc(s, ' ');
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -917,7 +917,7 @@ print_graph_prologue(struct trace_iterator *iter, struct trace_seq *s,
 		if (ret == TRACE_TYPE_PARTIAL_LINE)
 			return TRACE_TYPE_PARTIAL_LINE;
 
-		ret = trace_seq_printf(s, " | ");
+		ret = trace_seq_puts(s, " | ");
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -1117,7 +1117,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s,
 
 	/* Closing brace */
 	for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) {
-		ret = trace_seq_printf(s, " ");
+		ret = trace_seq_putc(s, ' ');
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
@@ -1129,7 +1129,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s,
 	 * belongs to, write out the function name.
 	 */
 	if (func_match) {
-		ret = trace_seq_printf(s, "}\n");
+		ret = trace_seq_puts(s, "}\n");
 		if (!ret)
 			return TRACE_TYPE_PARTIAL_LINE;
 	} else {
@@ -1179,13 +1179,13 @@ print_graph_comment(struct trace_seq *s, struct trace_entry *ent,
 	/* Indentation */
 	if (depth > 0)
 		for (i = 0; i < (depth + 1) * TRACE_GRAPH_INDENT; i++) {
-			ret = trace_seq_printf(s, " ");
+			ret = trace_seq_putc(s, ' ');
 			if (!ret)
 				return TRACE_TYPE_PARTIAL_LINE;
 		}
 
 	/* The comment */
-	ret = trace_seq_printf(s, "/* ");
+	ret = trace_seq_puts(s, "/* ");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
@@ -1216,7 +1216,7 @@ print_graph_comment(struct trace_seq *s, struct trace_entry *ent,
 		s->len--;
 	}
 
-	ret = trace_seq_printf(s, " */\n");
+	ret = trace_seq_puts(s, " */\n");
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c
index a5e8f48..b3dcfb2 100644
--- a/kernel/trace/trace_mmiotrace.c
+++ b/kernel/trace/trace_mmiotrace.c
@@ -90,7 +90,7 @@ static int mmio_print_pcidev(struct trace_seq *s, const struct pci_dev *dev)
 	if (drv)
 		ret += trace_seq_printf(s, " %s\n", drv->name);
 	else
-		ret += trace_seq_printf(s, " \n");
+		ret += trace_seq_puts(s, " \n");
 	return ret;
 }
 
@@ -107,7 +107,7 @@ static void mmio_pipe_open(struct trace_iterator *iter)
 	struct header_iter *hiter;
 	struct trace_seq *s = &iter->seq;
 
-	trace_seq_printf(s, "VERSION 20070824\n");
+	trace_seq_puts(s, "VERSION 20070824\n");
 
 	hiter = kzalloc(sizeof(*hiter), GFP_KERNEL);
 	if (!hiter)
@@ -209,7 +209,7 @@ static enum print_line_t mmio_print_rw(struct trace_iterator *iter)
 			(rw->value >> 0) & 0xff, rw->pc, 0);
 		break;
 	default:
-		ret = trace_seq_printf(s, "rw what?\n");
+		ret = trace_seq_puts(s, "rw what?\n");
 		break;
 	}
 	if (ret)
@@ -245,7 +245,7 @@ static enum print_line_t mmio_print_map(struct trace_iterator *iter)
 			secs, usec_rem, m->map_id, 0UL, 0);
 		break;
 	default:
-		ret = trace_seq_printf(s, "map what?\n");
+		ret = trace_seq_puts(s, "map what?\n");
 		break;
 	}
 	if (ret)
diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c
index bb922d9..34e7cba 100644
--- a/kernel/trace/trace_output.c
+++ b/kernel/trace/trace_output.c
@@ -78,7 +78,7 @@ enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter)
 
 	trace_assign_type(field, entry);
 
-	ret = trace_seq_printf(s, "%s", field->buf);
+	ret = trace_seq_puts(s, field->buf);
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
@@ -558,14 +558,14 @@ seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s,
 			if (ret)
 				ret = trace_seq_puts(s, "??");
 			if (ret)
-				ret = trace_seq_puts(s, "\n");
+				ret = trace_seq_putc(s, '\n');
 			continue;
 		}
 		if (!ret)
 			break;
 		if (ret)
 			ret = seq_print_user_ip(s, mm, ip, sym_flags);
-		ret = trace_seq_puts(s, "\n");
+		ret = trace_seq_putc(s, '\n');
 	}
 
 	if (mm)
@@ -579,7 +579,7 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags)
 	int ret;
 
 	if (!ip)
-		return trace_seq_printf(s, "0");
+		return trace_seq_putc(s, '0');
 
 	if (sym_flags & TRACE_ITER_SYM_OFFSET)
 		ret = seq_print_sym_offset(s, "%s", ip);
@@ -964,14 +964,14 @@ static enum print_line_t trace_fn_trace(struct trace_iterator *iter, int flags,
 		goto partial;
 
 	if ((flags & TRACE_ITER_PRINT_PARENT) && field->parent_ip) {
-		if (!trace_seq_printf(s, " <-"))
+		if (!trace_seq_puts(s, " <-"))
 			goto partial;
 		if (!seq_print_ip_sym(s,
 				      field->parent_ip,
 				      flags))
 			goto partial;
 	}
-	if (!trace_seq_printf(s, "\n"))
+	if (!trace_seq_putc(s, '\n'))
 		goto partial;
 
 	return TRACE_TYPE_HANDLED;
@@ -1210,7 +1210,7 @@ static enum print_line_t trace_stack_print(struct trace_iterator *iter,
 
 		if (!seq_print_ip_sym(s, *p, flags))
 			goto partial;
-		if (!trace_seq_puts(s, "\n"))
+		if (!trace_seq_putc(s, '\n'))
 			goto partial;
 	}
 
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index 322e164..0611562 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -175,7 +175,7 @@ print_syscall_exit(struct trace_iterator *iter, int flags,
 	entry = syscall_nr_to_meta(syscall);
 
 	if (!entry) {
-		trace_seq_printf(s, "\n");
+		trace_seq_putc(s, '\n');
 		return TRACE_TYPE_HANDLED;
 	}
 
-- 
1.7.10.4

[PATCH 03/14] tracing: Typo fix on ring buffer comments

[Steven Rostedt]

[-- Attachment #1: 0003-tracing-Typo-fix-on-ring-buffer-comments.patch --]
[-- Type: text/plain, Size: 3534 bytes --]

From: "zhangwei(Jovi)" <jovi.zhangwei@huawei.com>

There have some mismatch between comments with
real function name, update it.

This patch also add some missed function arguments
description.

Link: http://lkml.kernel.org/r/51E3B3B2.4080307@huawei.com

Signed-off-by: zhangwei(Jovi) <jovi.zhangwei@huawei.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/ring_buffer.c |   16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index eef2e56..cc2f66f 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -1066,7 +1066,7 @@ static int rb_check_list(struct ring_buffer_per_cpu *cpu_buffer,
 }
 
 /**
- * check_pages - integrity check of buffer pages
+ * rb_check_pages - integrity check of buffer pages
  * @cpu_buffer: CPU buffer with pages to test
  *
  * As a safety measure we check to make sure the data pages have not
@@ -1258,7 +1258,7 @@ static int rb_cpu_notify(struct notifier_block *self,
 #endif
 
 /**
- * ring_buffer_alloc - allocate a new ring_buffer
+ * __ring_buffer_alloc - allocate a new ring_buffer
  * @size: the size in bytes per cpu that is needed.
  * @flags: attributes to set for the ring buffer.
  *
@@ -1607,6 +1607,7 @@ static void update_pages_handler(struct work_struct *work)
  * ring_buffer_resize - resize the ring buffer
  * @buffer: the buffer to resize.
  * @size: the new size.
+ * @cpu_id: the cpu buffer to resize
  *
  * Minimum size is 2 * BUF_PAGE_SIZE.
  *
@@ -3956,11 +3957,11 @@ EXPORT_SYMBOL_GPL(ring_buffer_consume);
  * expected.
  *
  * After a sequence of ring_buffer_read_prepare calls, the user is
- * expected to make at least one call to ring_buffer_prepare_sync.
+ * expected to make at least one call to ring_buffer_read_prepare_sync.
  * Afterwards, ring_buffer_read_start is invoked to get things going
  * for real.
  *
- * This overall must be paired with ring_buffer_finish.
+ * This overall must be paired with ring_buffer_read_finish.
  */
 struct ring_buffer_iter *
 ring_buffer_read_prepare(struct ring_buffer *buffer, int cpu)
@@ -4009,7 +4010,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_read_prepare_sync);
  * an intervening ring_buffer_read_prepare_sync must have been
  * performed.
  *
- * Must be paired with ring_buffer_finish.
+ * Must be paired with ring_buffer_read_finish.
  */
 void
 ring_buffer_read_start(struct ring_buffer_iter *iter)
@@ -4031,7 +4032,7 @@ ring_buffer_read_start(struct ring_buffer_iter *iter)
 EXPORT_SYMBOL_GPL(ring_buffer_read_start);
 
 /**
- * ring_buffer_finish - finish reading the iterator of the buffer
+ * ring_buffer_read_finish - finish reading the iterator of the buffer
  * @iter: The iterator retrieved by ring_buffer_start
  *
  * This re-enables the recording to the buffer, and frees the
@@ -4346,6 +4347,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu);
 /**
  * ring_buffer_alloc_read_page - allocate a page to read from buffer
  * @buffer: the buffer to allocate for.
+ * @cpu: the cpu buffer to allocate.
  *
  * This function is used in conjunction with ring_buffer_read_page.
  * When reading a full page from the ring buffer, these functions
@@ -4403,7 +4405,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_free_read_page);
  * to swap with a page in the ring buffer.
  *
  * for example:
- *	rpage = ring_buffer_alloc_read_page(buffer);
+ *	rpage = ring_buffer_alloc_read_page(buffer, cpu);
  *	if (!rpage)
  *		return error;
  *	ret = ring_buffer_read_page(buffer, &rpage, len, cpu, 0);
-- 
1.7.10.4

[PATCH 04/14] tracing/function: Avoid perf_trace_buf_*() if event_function.perf_events is empty

[Steven Rostedt]

[-- Attachment #1: 0004-tracing-function-Avoid-perf_trace_buf_-if-event_func.patch --]
[-- Type: text/plain, Size: 1348 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

perf_trace_buf_prepare() + perf_trace_buf_submit(head, task => NULL)
make no sense if hlist_empty(head). Change perf_ftrace_function_call()
to check event_function.perf_events beforehand.

Link: http://lkml.kernel.org/r/20130617170204.GA19803@redhat.com

Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace_event_perf.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
index 84b1e04..12df557 100644
--- a/kernel/trace/trace_event_perf.c
+++ b/kernel/trace/trace_event_perf.c
@@ -266,6 +266,10 @@ perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip,
 	struct pt_regs regs;
 	int rctx;
 
+	head = this_cpu_ptr(event_function.perf_events);
+	if (hlist_empty(head))
+		return;
+
 #define ENTRY_SIZE (ALIGN(sizeof(struct ftrace_entry) + sizeof(u32), \
 		    sizeof(u64)) - sizeof(u32))
 
@@ -279,8 +283,6 @@ perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip,
 
 	entry->ip = ip;
 	entry->parent_ip = parent_ip;
-
-	head = this_cpu_ptr(event_function.perf_events);
 	perf_trace_buf_submit(entry, ENTRY_SIZE, rctx, 0,
 			      1, &regs, head, NULL);
 
-- 
1.7.10.4

[PATCH 05/14] tracing/syscall: Avoid perf_trace_buf_*() if sys_data->perf_events is empty

[Steven Rostedt]

[-- Attachment #1: 0005-tracing-syscall-Avoid-perf_trace_buf_-if-sys_data-pe.patch --]
[-- Type: text/plain, Size: 2169 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

perf_trace_buf_prepare() + perf_trace_buf_submit(head, task => NULL)
make no sense if hlist_empty(head). Change perf_syscall_enter/exit()
to check sys_data->{enter,exit}_event->perf_events beforehand.

Link: http://lkml.kernel.org/r/20130617170207.GA19806@redhat.com

Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace_syscalls.c |   12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index 0611562..ac00857 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -566,6 +566,10 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id)
 	if (!sys_data)
 		return;
 
+	head = this_cpu_ptr(sys_data->enter_event->perf_events);
+	if (hlist_empty(head))
+		return;
+
 	/* get the size after alignment with the u32 buffer size field */
 	size = sizeof(unsigned long) * sys_data->nb_args + sizeof(*rec);
 	size = ALIGN(size + sizeof(u32), sizeof(u64));
@@ -583,8 +587,6 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id)
 	rec->nr = syscall_nr;
 	syscall_get_arguments(current, regs, 0, sys_data->nb_args,
 			       (unsigned long *)&rec->args);
-
-	head = this_cpu_ptr(sys_data->enter_event->perf_events);
 	perf_trace_buf_submit(rec, size, rctx, 0, 1, regs, head, NULL);
 }
 
@@ -642,6 +644,10 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret)
 	if (!sys_data)
 		return;
 
+	head = this_cpu_ptr(sys_data->exit_event->perf_events);
+	if (hlist_empty(head))
+		return;
+
 	/* We can probably do that at build time */
 	size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64));
 	size -= sizeof(u32);
@@ -661,8 +667,6 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret)
 
 	rec->nr = syscall_nr;
 	rec->ret = syscall_get_return_value(current, regs);
-
-	head = this_cpu_ptr(sys_data->exit_event->perf_events);
 	perf_trace_buf_submit(rec, size, rctx, 0, 1, regs, head, NULL);
 }
 
-- 
1.7.10.4

[PATCH 06/14] tracing/perf: Move the PERF_MAX_TRACE_SIZE check into perf_trace_buf_prepare()

[Steven Rostedt]

[-- Attachment #1: 0006-tracing-perf-Move-the-PERF_MAX_TRACE_SIZE-check-into.patch --]
[-- Type: text/plain, Size: 4597 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

Every perf_trace_buf_prepare() caller does
WARN_ONCE(size > PERF_MAX_TRACE_SIZE, message) and "message" is
almost the same.

Shift this WARN_ONCE() into perf_trace_buf_prepare(). This changes
the meaning of _ONCE, but I think this is fine.

	- 4947014 2932448 10104832  17984294  1126b26 vmlinux
	+ 4948422 2932448 10104832  17985702  11270a6 vmlinux

on my build.

Link: http://lkml.kernel.org/r/20130617170211.GA19813@redhat.com

Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 include/trace/ftrace.h          |    4 ----
 kernel/trace/trace_event_perf.c |    4 ++++
 kernel/trace/trace_kprobe.c     |    6 ------
 kernel/trace/trace_syscalls.c   |   12 ------------
 kernel/trace/trace_uprobe.c     |    2 --
 5 files changed, 4 insertions(+), 24 deletions(-)

diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h
index d615f78..41a6643 100644
--- a/include/trace/ftrace.h
+++ b/include/trace/ftrace.h
@@ -670,10 +670,6 @@ perf_trace_##call(void *__data, proto)					\
 			     sizeof(u64));				\
 	__entry_size -= sizeof(u32);					\
 									\
-	if (WARN_ONCE(__entry_size > PERF_MAX_TRACE_SIZE,		\
-		      "profile buffer not large enough"))		\
-		return;							\
-									\
 	entry = (struct ftrace_raw_##call *)perf_trace_buf_prepare(	\
 		__entry_size, event_call->event.type, &__regs, &rctx);	\
 	if (!entry)							\
diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c
index 12df557..80c36bc 100644
--- a/kernel/trace/trace_event_perf.c
+++ b/kernel/trace/trace_event_perf.c
@@ -236,6 +236,10 @@ __kprobes void *perf_trace_buf_prepare(int size, unsigned short type,
 
 	BUILD_BUG_ON(PERF_MAX_TRACE_SIZE % sizeof(unsigned long));
 
+	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
+			"perf buffer not large enough"))
+		return NULL;
+
 	pc = preempt_count();
 
 	*rctxp = perf_swevent_get_recursion_context();
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 7ed6976..ae6ce83 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -1087,9 +1087,6 @@ kprobe_perf_func(struct trace_probe *tp, struct pt_regs *regs)
 	__size = sizeof(*entry) + tp->size + dsize;
 	size = ALIGN(__size + sizeof(u32), sizeof(u64));
 	size -= sizeof(u32);
-	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
-		     "profile buffer not large enough"))
-		return;
 
 	entry = perf_trace_buf_prepare(size, call->event.type, regs, &rctx);
 	if (!entry)
@@ -1120,9 +1117,6 @@ kretprobe_perf_func(struct trace_probe *tp, struct kretprobe_instance *ri,
 	__size = sizeof(*entry) + tp->size + dsize;
 	size = ALIGN(__size + sizeof(u32), sizeof(u64));
 	size -= sizeof(u32);
-	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
-		     "profile buffer not large enough"))
-		return;
 
 	entry = perf_trace_buf_prepare(size, call->event.type, regs, &rctx);
 	if (!entry)
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index ac00857..8fd0365 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -575,10 +575,6 @@ static void perf_syscall_enter(void *ignore, struct pt_regs *regs, long id)
 	size = ALIGN(size + sizeof(u32), sizeof(u64));
 	size -= sizeof(u32);
 
-	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
-		      "perf buffer not large enough"))
-		return;
-
 	rec = (struct syscall_trace_enter *)perf_trace_buf_prepare(size,
 				sys_data->enter_event->event.type, regs, &rctx);
 	if (!rec)
@@ -652,14 +648,6 @@ static void perf_syscall_exit(void *ignore, struct pt_regs *regs, long ret)
 	size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64));
 	size -= sizeof(u32);
 
-	/*
-	 * Impossible, but be paranoid with the future
-	 * How to put this check outside runtime?
-	 */
-	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
-		"exit event has grown above perf buffer size"))
-		return;
-
 	rec = (struct syscall_trace_exit *)perf_trace_buf_prepare(size,
 				sys_data->exit_event->event.type, regs, &rctx);
 	if (!rec)
diff --git a/kernel/trace/trace_uprobe.c b/kernel/trace/trace_uprobe.c
index d5d0cd3..a23d2d7 100644
--- a/kernel/trace/trace_uprobe.c
+++ b/kernel/trace/trace_uprobe.c
@@ -818,8 +818,6 @@ static void uprobe_perf_print(struct trace_uprobe *tu,
 
 	size = SIZEOF_TRACE_ENTRY(is_ret_probe(tu));
 	size = ALIGN(size + tu->size + sizeof(u32), sizeof(u64)) - sizeof(u32);
-	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE, "profile buffer not large enough"))
-		return;
 
 	preempt_disable();
 	head = this_cpu_ptr(call->perf_events);
-- 
1.7.10.4

[PATCH 07/14] tracing/kprobe: Wait for disabling all running kprobe handlers

[Steven Rostedt]

[-- Attachment #1: 0007-tracing-kprobe-Wait-for-disabling-all-running-kprobe.patch --]
[-- Type: text/plain, Size: 2211 bytes --]

From: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>

Wait for disabling all running kprobe handlers when a kprobe
event is disabled, since the caller, trace_remove_event_call()
supposes that a removing event is disabled completely by
disabling the event.
With this change, ftrace can ensure that there is no running
event handlers after disabling it.

Link: http://lkml.kernel.org/r/20130709093526.20138.93100.stgit@mhiramat-M0-7522

Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace_kprobe.c |   23 +++++++++++++++++------
 1 file changed, 17 insertions(+), 6 deletions(-)

diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index ae6ce83..3811487 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -243,11 +243,11 @@ find_event_file_link(struct trace_probe *tp, struct ftrace_event_file *file)
 static int
 disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file)
 {
+	struct event_file_link *link = NULL;
+	int wait = 0;
 	int ret = 0;
 
 	if (file) {
-		struct event_file_link *link;
-
 		link = find_event_file_link(tp, file);
 		if (!link) {
 			ret = -EINVAL;
@@ -255,10 +255,7 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file)
 		}
 
 		list_del_rcu(&link->list);
-		/* synchronize with kprobe_trace_func/kretprobe_trace_func */
-		synchronize_sched();
-		kfree(link);
-
+		wait = 1;
 		if (!list_empty(&tp->files))
 			goto out;
 
@@ -271,8 +268,22 @@ disable_trace_probe(struct trace_probe *tp, struct ftrace_event_file *file)
 			disable_kretprobe(&tp->rp);
 		else
 			disable_kprobe(&tp->rp.kp);
+		wait = 1;
 	}
  out:
+	if (wait) {
+		/*
+		 * Synchronize with kprobe_trace_func/kretprobe_trace_func
+		 * to ensure disabled (all running handlers are finished).
+		 * This is not only for kfree(), but also the caller,
+		 * trace_remove_event_call() supposes it for releasing
+		 * event_call related objects, which will be accessed in
+		 * the kprobe_trace_func/kretprobe_trace_func.
+		 */
+		synchronize_sched();
+		kfree(link);	/* Ignored if link == NULL */
+	}
+
 	return ret;
 }
 
-- 
1.7.10.4

[PATCH 08/14] tracing: Fix error handling to ensure instances can always be removed

[Steven Rostedt]

[-- Attachment #1: 0008-tracing-Fix-error-handling-to-ensure-instances-can-a.patch --]
[-- Type: text/plain, Size: 1198 bytes --]

From: Alexander Z Lam <azl@google.com>

Remove debugfs directories for tracing instances during creation if an error
occurs causing the trace_array for that instance to not be added to
ftrace_trace_arrays. If the directory continues to exist after the error, it
cannot be removed because the respective trace_array is not in
ftrace_trace_arrays.

Link: http://lkml.kernel.org/r/1373502874-1706-2-git-send-email-azl@google.com

Cc: stable@vger.kernel.org # 3.10
Cc: Vaibhav Nagarnaik <vnagarnaik@google.com>
Cc: David Sharp <dhsharp@google.com>
Cc: Alexander Z Lam <lambchop468@gmail.com>
Signed-off-by: Alexander Z Lam <azl@google.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 25b91af..7c3da7b 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5973,8 +5973,10 @@ static int new_instance_create(const char *name)
 		goto out_free_tr;
 
 	ret = event_trace_add_tracer(tr->dir, tr);
-	if (ret)
+	if (ret) {
+		debugfs_remove_recursive(tr->dir);
 		goto out_free_tr;
+	}
 
 	init_tracer_debugfs(tr, tr->dir);
 
-- 
1.7.10.4

[PATCH 09/14] tracing: Miscellaneous fixes for trace_array ref counting

[Steven Rostedt]

[-- Attachment #1: 0009-tracing-Miscellaneous-fixes-for-trace_array-ref-coun.patch --]
[-- Type: text/plain, Size: 4962 bytes --]

From: Alexander Z Lam <azl@google.com>

Some error paths did not handle ref counting properly, and some trace files need
ref counting.

Link: http://lkml.kernel.org/r/1374171524-11948-1-git-send-email-azl@google.com

Cc: stable@vger.kernel.org # 3.10
Cc: Vaibhav Nagarnaik <vnagarnaik@google.com>
Cc: David Sharp <dhsharp@google.com>
Cc: Alexander Z Lam <lambchop468@gmail.com>
Signed-off-by: Alexander Z Lam <azl@google.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.c        |   24 ++++++++++++++++++------
 kernel/trace/trace_events.c |   21 +++++++++++++++++++--
 2 files changed, 37 insertions(+), 8 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 7c3da7b..7d9ceab 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -3008,7 +3008,6 @@ static int tracing_release(struct inode *inode, struct file *file)
 
 	iter = m->private;
 	tr = iter->tr;
-	trace_array_put(tr);
 
 	mutex_lock(&trace_types_lock);
 
@@ -3023,6 +3022,9 @@ static int tracing_release(struct inode *inode, struct file *file)
 	if (!iter->snapshot)
 		/* reenable tracing if it was previously enabled */
 		tracing_start_tr(tr);
+
+	__trace_array_put(tr);
+
 	mutex_unlock(&trace_types_lock);
 
 	mutex_destroy(&iter->mutex);
@@ -3447,6 +3449,7 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf,
 static int tracing_trace_options_open(struct inode *inode, struct file *file)
 {
 	struct trace_array *tr = inode->i_private;
+	int ret;
 
 	if (tracing_disabled)
 		return -ENODEV;
@@ -3454,7 +3457,11 @@ static int tracing_trace_options_open(struct inode *inode, struct file *file)
 	if (trace_array_get(tr) < 0)
 		return -ENODEV;
 
-	return single_open(file, tracing_trace_options_show, inode->i_private);
+	ret = single_open(file, tracing_trace_options_show, inode->i_private);
+	if (ret < 0)
+		trace_array_put(tr);
+
+	return ret;
 }
 
 static const struct file_operations tracing_iter_fops = {
@@ -3958,6 +3965,7 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp)
 	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
 	if (!iter) {
 		ret = -ENOMEM;
+		__trace_array_put(tr);
 		goto out;
 	}
 
@@ -4704,21 +4712,24 @@ static int tracing_snapshot_open(struct inode *inode, struct file *file)
 			ret = PTR_ERR(iter);
 	} else {
 		/* Writes still need the seq_file to hold the private data */
+		ret = -ENOMEM;
 		m = kzalloc(sizeof(*m), GFP_KERNEL);
 		if (!m)
-			return -ENOMEM;
+			goto out;
 		iter = kzalloc(sizeof(*iter), GFP_KERNEL);
 		if (!iter) {
 			kfree(m);
-			return -ENOMEM;
+			goto out;
 		}
+		ret = 0;
+
 		iter->tr = tr;
 		iter->trace_buffer = &tc->tr->max_buffer;
 		iter->cpu_file = tc->cpu;
 		m->private = iter;
 		file->private_data = m;
 	}
-
+out:
 	if (ret < 0)
 		trace_array_put(tr);
 
@@ -5328,9 +5339,10 @@ tracing_stats_read(struct file *filp, char __user *ubuf,
 }
 
 static const struct file_operations tracing_stats_fops = {
-	.open		= tracing_open_generic,
+	.open		= tracing_open_generic_tc,
 	.read		= tracing_stats_read,
 	.llseek		= generic_file_llseek,
+	.release	= tracing_release_generic_tc,
 };
 
 #ifdef CONFIG_DYNAMIC_FTRACE
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 7d85429..7a75cb2 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -1218,6 +1218,7 @@ show_header(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos)
 
 static int ftrace_event_avail_open(struct inode *inode, struct file *file);
 static int ftrace_event_set_open(struct inode *inode, struct file *file);
+static int ftrace_event_release(struct inode *inode, struct file *file);
 
 static const struct seq_operations show_event_seq_ops = {
 	.start = t_start,
@@ -1245,7 +1246,7 @@ static const struct file_operations ftrace_set_event_fops = {
 	.read = seq_read,
 	.write = ftrace_event_write,
 	.llseek = seq_lseek,
-	.release = seq_release,
+	.release = ftrace_event_release,
 };
 
 static const struct file_operations ftrace_enable_fops = {
@@ -1323,6 +1324,15 @@ ftrace_event_open(struct inode *inode, struct file *file,
 	return ret;
 }
 
+static int ftrace_event_release(struct inode *inode, struct file *file)
+{
+	struct trace_array *tr = inode->i_private;
+
+	trace_array_put(tr);
+
+	return seq_release(inode, file);
+}
+
 static int
 ftrace_event_avail_open(struct inode *inode, struct file *file)
 {
@@ -1336,12 +1346,19 @@ ftrace_event_set_open(struct inode *inode, struct file *file)
 {
 	const struct seq_operations *seq_ops = &show_set_event_seq_ops;
 	struct trace_array *tr = inode->i_private;
+	int ret;
+
+	if (trace_array_get(tr) < 0)
+		return -ENODEV;
 
 	if ((file->f_mode & FMODE_WRITE) &&
 	    (file->f_flags & O_TRUNC))
 		ftrace_clear_events(tr);
 
-	return ftrace_event_open(inode, file, seq_ops);
+	ret = ftrace_event_open(inode, file, seq_ops);
+	if (ret < 0)
+		trace_array_put(tr);
+	return ret;
 }
 
 static struct event_subsystem *
-- 
1.7.10.4

[PATCH 10/14] tracing: Add ref_data to function and fgraph tracer structs

[Steven Rostedt]

[-- Attachment #1: 0010-tracing-Add-ref_data-to-function-and-fgraph-tracer-s.patch --]
[-- Type: text/plain, Size: 2903 bytes --]

From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>

The selftest for function and function graph tracers are defined as
__init, as they are only executed at boot up. The "tracer" structs
that are associated to those tracers are not setup as __init as they
are used after boot. To stop mismatch warnings, those structures
need to be annotated with __ref_data.

Currently, the tracer structures are defined to __read_mostly, as they
do not really change. But in the future they should be converted to
consts, but that will take a little work because they have a "next"
pointer that gets updated when they are registered. That will have to
wait till the next major release.

Link: http://lkml.kernel.org/r/1373596735.17876.84.camel@gandalf.local.home

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Reported-by: Chen Gang <gang.chen@asianux.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.h                 |    9 +++++++++
 kernel/trace/trace_functions.c       |    2 +-
 kernel/trace/trace_functions_graph.c |    2 +-
 3 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 4a4f6e1..57b7bb0 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -680,6 +680,15 @@ extern int trace_selftest_startup_sched_switch(struct tracer *trace,
 					       struct trace_array *tr);
 extern int trace_selftest_startup_branch(struct tracer *trace,
 					 struct trace_array *tr);
+/*
+ * Tracer data references selftest functions that only occur
+ * on boot up. These can be __init functions. Thus, when selftests
+ * are enabled, then the tracers need to reference __init functions.
+ */
+#define __tracer_data		__refdata
+#else
+/* Tracers are seldom changed. Optimize when selftests are disabled. */
+#define __tracer_data		__read_mostly
 #endif /* CONFIG_FTRACE_STARTUP_TEST */
 
 extern void *head_page(struct trace_array_cpu *data);
diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c
index b863f93..38fe148 100644
--- a/kernel/trace/trace_functions.c
+++ b/kernel/trace/trace_functions.c
@@ -199,7 +199,7 @@ static int func_set_flag(u32 old_flags, u32 bit, int set)
 	return 0;
 }
 
-static struct tracer function_trace __read_mostly =
+static struct tracer function_trace __tracer_data =
 {
 	.name		= "function",
 	.init		= function_trace_init,
diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c
index d56ae9b..b5c0924 100644
--- a/kernel/trace/trace_functions_graph.c
+++ b/kernel/trace/trace_functions_graph.c
@@ -1448,7 +1448,7 @@ static struct trace_event graph_trace_ret_event = {
 	.funcs		= &graph_functions
 };
 
-static struct tracer graph_trace __read_mostly = {
+static struct tracer graph_trace __tracer_data = {
 	.name		= "function_graph",
 	.open		= graph_trace_open,
 	.pipe_open	= graph_trace_open,
-- 
1.7.10.4

[PATCH 11/14] tracing: Simplify the iteration logic in f_start/f_next

[Steven Rostedt]

[-- Attachment #1: 0011-tracing-Simplify-the-iteration-logic-in-f_start-f_ne.patch --]
[-- Type: text/plain, Size: 3667 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

f_next() looks overcomplicated, and it is not strictly correct
even if this doesn't matter.

Say, FORMAT_FIELD_SEPERATOR should not return NULL (means EOF)
if trace_get_fields() returns an empty list, we should simply
advance to FORMAT_PRINTFMT as we do when we find the end of list.

1. Change f_next() to return "struct list_head *" rather than
   "ftrace_event_field *", and change f_show() to do list_entry().

   This simplifies the code a bit, only f_show() needs to know
   about ftrace_event_field, and f_next() can play with ->prev
   directly

2. Change f_next() to not play with ->prev / return inside the
   switch() statement. It can simply set node = head/common_head,
   the prev-or-advance-to-the-next-magic below does all work.

While at it. f_start() looks overcomplicated too. I don't think
*pos == 0 makes sense as a separate case, just change this code
to do "while" instead of "do/while".

The patch also moves f_start() down, close to f_stop(). This is
purely cosmetic, just to make the locking added by the next patch
more clear/visible.

Link: http://lkml.kernel.org/r/20130718184710.GA4783@redhat.com

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace_events.c |   60 ++++++++++++++++---------------------------
 1 file changed, 22 insertions(+), 38 deletions(-)

diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 7a75cb2..76defd9 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -826,59 +826,33 @@ enum {
 static void *f_next(struct seq_file *m, void *v, loff_t *pos)
 {
 	struct ftrace_event_call *call = m->private;
-	struct ftrace_event_field *field;
 	struct list_head *common_head = &ftrace_common_fields;
 	struct list_head *head = trace_get_fields(call);
+	struct list_head *node = v;
 
 	(*pos)++;
 
 	switch ((unsigned long)v) {
 	case FORMAT_HEADER:
-		if (unlikely(list_empty(common_head)))
-			return NULL;
-
-		field = list_entry(common_head->prev,
-				   struct ftrace_event_field, link);
-		return field;
+		node = common_head;
+		break;
 
 	case FORMAT_FIELD_SEPERATOR:
-		if (unlikely(list_empty(head)))
-			return NULL;
-
-		field = list_entry(head->prev, struct ftrace_event_field, link);
-		return field;
+		node = head;
+		break;
 
 	case FORMAT_PRINTFMT:
 		/* all done */
 		return NULL;
 	}
 
-	field = v;
-	if (field->link.prev == common_head)
+	node = node->prev;
+	if (node == common_head)
 		return (void *)FORMAT_FIELD_SEPERATOR;
-	else if (field->link.prev == head)
+	else if (node == head)
 		return (void *)FORMAT_PRINTFMT;
-
-	field = list_entry(field->link.prev, struct ftrace_event_field, link);
-
-	return field;
-}
-
-static void *f_start(struct seq_file *m, loff_t *pos)
-{
-	loff_t l = 0;
-	void *p;
-
-	/* Start by showing the header */
-	if (!*pos)
-		return (void *)FORMAT_HEADER;
-
-	p = (void *)FORMAT_HEADER;
-	do {
-		p = f_next(m, p, &l);
-	} while (p && l < *pos);
-
-	return p;
+	else
+		return node;
 }
 
 static int f_show(struct seq_file *m, void *v)
@@ -904,8 +878,7 @@ static int f_show(struct seq_file *m, void *v)
 		return 0;
 	}
 
-	field = v;
-
+	field = list_entry(v, struct ftrace_event_field, link);
 	/*
 	 * Smartly shows the array type(except dynamic array).
 	 * Normal:
@@ -932,6 +905,17 @@ static int f_show(struct seq_file *m, void *v)
 	return 0;
 }
 
+static void *f_start(struct seq_file *m, loff_t *pos)
+{
+	void *p = (void *)FORMAT_HEADER;
+	loff_t l = 0;
+
+	while (l < *pos && p)
+		p = f_next(m, p, &l);
+
+	return p;
+}
+
 static void f_stop(struct seq_file *m, void *p)
 {
 }
-- 
1.7.10.4

[PATCH 12/14] tracing: Do not (ab)use trace_seq in event_id_read()

[Steven Rostedt]

[-- Attachment #1: 0012-tracing-Do-not-ab-use-trace_seq-in-event_id_read.patch --]
[-- Type: text/plain, Size: 1578 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

event_id_read() has no reason to kmalloc "struct trace_seq"
(more than PAGE_SIZE!), it can use a small buffer instead.

Note: "if (*ppos) return 0" looks strange and even wrong,
simple_read_from_buffer() handles ppos != 0 case corrrectly.

And it seems that almost every user of trace_seq in this file
should be converted too. Unless you use seq_open(), trace_seq
buys nothing compared to the raw buffer, but it needs a bit
more memory and code.

Link: http://lkml.kernel.org/r/20130718184712.GA4786@redhat.com

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace_events.c |   17 ++++-------------
 1 file changed, 4 insertions(+), 13 deletions(-)

diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 76defd9..898f868 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -947,23 +947,14 @@ 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;
-	struct trace_seq *s;
-	int r;
+	char buf[32];
+	int len;
 
 	if (*ppos)
 		return 0;
 
-	s = kmalloc(sizeof(*s), GFP_KERNEL);
-	if (!s)
-		return -ENOMEM;
-
-	trace_seq_init(s);
-	trace_seq_printf(s, "%d\n", call->event.type);
-
-	r = simple_read_from_buffer(ubuf, cnt, ppos,
-				    s->buffer, s->len);
-	kfree(s);
-	return r;
+	len = sprintf(buf, "%d\n", call->event.type);
+	return simple_read_from_buffer(ubuf, cnt, ppos, buf, len);
 }
 
 static ssize_t
-- 
1.7.10.4

[PATCH 13/14] tracing: Kill trace_array->waiter

[Steven Rostedt]

[-- Attachment #1: 0013-tracing-Kill-trace_array-waiter.patch --]
[-- Type: text/plain, Size: 713 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

Trivial. trace_array->waiter has no users since 6eaaa5d5
"tracing/core: use appropriate waiting on trace_pipe".

Link: http://lkml.kernel.org/r/20130719142036.GA1594@redhat.com

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.h |    1 -
 1 file changed, 1 deletion(-)

diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 57b7bb0..e7d643b 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -214,7 +214,6 @@ struct trace_array {
 	struct dentry		*event_dir;
 	struct list_head	systems;
 	struct list_head	events;
-	struct task_struct	*waiter;
 	int			ref;
 };
 
-- 
1.7.10.4

[PATCH 14/14] tracing: Kill the unbalanced tr->ref++ in tracing_buffers_open()

[Steven Rostedt]

[-- Attachment #1: 0014-tracing-Kill-the-unbalanced-tr-ref-in-tracing_buffer.patch --]
[-- Type: text/plain, Size: 1194 bytes --]

From: Oleg Nesterov <oleg@redhat.com>

tracing_buffers_open() does trace_array_get() and then it wrongly
inrcements tr->ref again under trace_types_lock. This means that
every caller leaks trace_array:

	# cd /sys/kernel/debug/tracing/
	# mkdir instances/X
	# true < instances/X/per_cpu/cpu0/trace_pipe_raw
	# rmdir instances/X
	rmdir: failed to remove `instances/X': Device or resource busy

Link: http://lkml.kernel.org/r/20130719153644.GA18899@redhat.com

Cc: Ingo Molnar <mingo@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: stable@vger.kernel.org # 3.10
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.c |    2 --
 1 file changed, 2 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 7d9ceab..3f24777 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -4959,8 +4959,6 @@ static int tracing_buffers_open(struct inode *inode, struct file *filp)
 
 	mutex_lock(&trace_types_lock);
 
-	tr->ref++;
-
 	info->iter.tr		= tr;
 	info->iter.cpu_file	= tc->cpu;
 	info->iter.trace	= tr->current_trace;
-- 
1.7.10.4