Merge tag 'trace-v5.17' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace

Pull tracing updates from Steven Rostedt:
 "New:

   - The Real Time Linux Analysis (RTLA) tool is added to the tools
     directory.

   - Can safely filter on user space pointers with: field.ustring ~
     "match-string"

   - eprobes can now be filtered like any other event.

   - trace_marker(_raw) now uses stream_open() to allow multiple threads
     to safely write to it. Note, this could possibly break existing
     user space, but we will not know until we hear about it, and then
     can revert the change if need be.

   - New field in events to display when bottom halfs are disabled.

   - Sorting of the ftrace functions are now done at compile time
     instead of at bootup.

  Infrastructure changes to support future efforts:

   - Added __rel_loc type for trace events. Similar to __data_loc but
     the offset to the dynamic data is based off of the location of the
     descriptor and not the beginning of the event. Needed for user
     defined events.

   - Some simplification of event trigger code.

   - Make synthetic events process its callback better to not hinder
     other event callbacks that are registered. Needed for user defined
     events.

  And other small fixes and cleanups"

* tag 'trace-v5.17' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (50 commits)
  tracing: Add ustring operation to filtering string pointers
  rtla: Add rtla timerlat hist documentation
  rtla: Add rtla timerlat top documentation
  rtla: Add rtla timerlat documentation
  rtla: Add rtla osnoise hist documentation
  rtla: Add rtla osnoise top documentation
  rtla: Add rtla osnoise man page
  rtla: Add Documentation
  rtla/timerlat: Add timerlat hist mode
  rtla: Add timerlat tool and timelart top mode
  rtla/osnoise: Add the hist mode
  rtla/osnoise: Add osnoise top mode
  rtla: Add osnoise tool
  rtla: Helper functions for rtla
  rtla: Real-Time Linux Analysis tool
  tracing/osnoise: Properly unhook events if start_per_cpu_kthreads() fails
  tracing: Remove duplicate warnings when calling trace_create_file()
  tracing/kprobes: 'nmissed' not showed correctly for kretprobe
  tracing: Add test for user space strings when filtering on string pointers
  tracing: Have syscall trace events use trace_event_buffer_lock_reserve()
  ...

2 files changed, 37 insertions, 1 deletions

diff --git a/samples/trace_events/trace-events-sample.c b/samples/trace_events/trace-events-sample.c
index 1a72b7d95cdc..4d34dc0b0fee 100644
--- a/samples/trace_events/trace-events-sample.c
+++ b/samples/trace_events/trace-events-sample.c
@@ -21,6 +21,7 @@ static const char *random_strings[] = {
 
 static void simple_thread_func(int cnt)
 {
+	unsigned long bitmask[1] = {0xdeadbeefUL};
 	int array[6];
 	int len = cnt % 5;
 	int i;
@@ -43,6 +44,8 @@ static void simple_thread_func(int cnt)
 	trace_foo_with_template_cond("prints other times", cnt);
 
 	trace_foo_with_template_print("I have to be different", cnt);
+
+	trace_foo_rel_loc("Hello __rel_loc", cnt, bitmask);
 }
 
 static int simple_thread(void *arg)
diff --git a/samples/trace_events/trace-events-sample.h b/samples/trace_events/trace-events-sample.h
index e61471ab7d14..cbbbb83beced 100644
--- a/samples/trace_events/trace-events-sample.h
+++ b/samples/trace_events/trace-events-sample.h
@@ -155,7 +155,7 @@
  *
  *         To assign this string, use the helper macro __assign_str_len().
  *
- *         __assign_str(foo, bar, len);
+ *         __assign_str_len(foo, bar, len);
  *
  *         Then len + 1 is allocated to the ring buffer, and a nul terminating
  *         byte is added. This is similar to:
@@ -506,6 +506,39 @@ DEFINE_EVENT_PRINT(foo_template, foo_with_template_print,
 	TP_ARGS(foo, bar),
 	TP_printk("bar %s %d", __get_str(foo), __entry->bar));
 
+/*
+ * There are yet another __rel_loc dynamic data attribute. If you
+ * use __rel_dynamic_array() and __rel_string() etc. macros, you
+ * can use this attribute. There is no difference from the viewpoint
+ * of functionality with/without 'rel' but the encoding is a bit
+ * different. This is expected to be used with user-space event,
+ * there is no reason that the kernel event use this, but only for
+ * testing.
+ */
+
+TRACE_EVENT(foo_rel_loc,
+
+	TP_PROTO(const char *foo, int bar, unsigned long *mask),
+
+	TP_ARGS(foo, bar, mask),
+
+	TP_STRUCT__entry(
+		__rel_string(	foo,	foo	)
+		__field(	int,	bar	)
+		__rel_bitmask(	bitmask,
+			BITS_PER_BYTE * sizeof(unsigned long)	)
+	),
+
+	TP_fast_assign(
+		__assign_rel_str(foo, foo);
+		__entry->bar = bar;
+		__assign_rel_bitmask(bitmask, mask,
+			BITS_PER_BYTE * sizeof(unsigned long));
+	),
+
+	TP_printk("foo_rel_loc %s, %d, %s", __get_rel_str(foo), __entry->bar,
+		  __get_rel_bitmask(bitmask))
+);
 #endif
 
 /***** NOTICE! The #if protection ends here. *****/