Merge tag 'trace-tools-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace

Pull tracing tooling updates from Steven Rostedt:

 - Add cgroup support for rtla via the -C option

 - Add --house-keeping option that tells rtla where to place the
   housekeeping threads

 - Have rtla/timerlat have its own tracing instance instead of using the
   top level tracing instance that is the default for other tracing
   users to use

 - Add auto analysis to timerlat_hist

 - Have rtla start the tracers after creating the instances

 - Reduce rtla hwnoise down to 75% from 100% as it runs with preemption
   disabled and can cause system instability at 100%

 - Add support to run timerlat_top and timerlat_hist threads in
   user-space instead of just using the kernel tasks

 - Some minor clean ups and documentation changes

* tag 'trace-tools-v6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  Documentation: Add tools/rtla timerlat -u option documentation
  rtla/timerlat_hist: Add timerlat user-space support
  rtla/timerlat_top: Add timerlat user-space support
  rtla/hwnoise: Reduce runtime to 75%
  rtla: Start the tracers after creating all instances
  rtla/timerlat_hist: Add auto-analysis support
  rtla/timerlat: Give timerlat auto analysis its own instance
  rtla: Automatically move rtla to a house-keeping cpu
  rtla: Change monitored_cpus from char * to cpu_set_t
  rtla: Add --house-keeping option
  rtla: Add -C cgroup support

Author: torvalds

Documentation/tools/rtla/common_options.rst

@@ -2,6 +2,10 @@
 
         Set the osnoise tracer to run the sample threads in the cpu-list.
 
+**-H**, **--house-keeping** *cpu-list*
+
+        Run rtla control threads only on the given cpu-list.
+
 **-d**, **--duration** *time[s|m|h|d]*
 
         Set the duration of the session.
@@ -42,6 +46,10 @@
         - *f:prio* - use SCHED_FIFO with *prio*;
         - *d:runtime[us|ms|s]:period[us|ms|s]* - use SCHED_DEADLINE with *runtime* and *period* in nanoseconds.
 
+**-C**, **--cgroup**\[*=cgroup*]
+
+        Set a *cgroup* to the tracer's threads. If the **-C** option is passed without arguments, the tracer's thread will inherit **rtla**'s *cgroup*. Otherwise, the threads will be placed on the *cgroup* passed to the option.
+
 **-h**, **--help**
 
         Print help menu.

Documentation/tools/rtla/common_timerlat_aa.rst

@@ -5,10 +5,3 @@
 **--no-aa**
 
         disable auto-analysis, reducing rtla timerlat cpu usage
-
-**--aa-only** *us*
-
-        Set stop tracing conditions and run without collecting and displaying statistics.
-        Print the auto-analysis if the system hits the stop tracing condition. This option
-        is useful to reduce rtla timerlat CPU, enabling the debug without the overhead of
-        collecting the statistics.

Documentation/tools/rtla/common_timerlat_options.rst

@@ -26,3 +26,10 @@
         Set the /dev/cpu_dma_latency to *us*, aiming to bound exit from idle latencies.
         *cyclictest* sets this value to *0* by default, use **--dma-latency** *0* to have
         similar results.
+
+**-u**, **--user-threads**
+
+        Set timerlat to run without a workload, and then dispatches user-space workloads
+        to wait on the timerlat_fd. Once the workload is awakes, it goes to sleep again
+        adding so the measurement for the kernel-to-user and user-to-kernel to the tracer
+        output.

Documentation/tools/rtla/rtla-timerlat-hist.rst

@@ -29,15 +29,18 @@ OPTIONS
 
 .. include:: common_options.rst
 
+.. include:: common_timerlat_aa.rst
+
 EXAMPLE
 =======
 In the example below, **rtla timerlat hist** is set to run for *10* minutes,
 in the cpus *0-4*, *skipping zero* only lines. Moreover, **rtla timerlat
 hist** will change the priority of the *timerlat* threads to run under
 *SCHED_DEADLINE* priority, with a *10us* runtime every *1ms* period. The
-*1ms* period is also passed to the *timerlat* tracer::
+*1ms* period is also passed to the *timerlat* tracer. Auto-analysis is disabled
+to reduce overhead ::
 
-  [root@alien ~]# timerlat hist -d 10m -c 0-4 -P d:100us:1ms -p 1ms
+  [root@alien ~]# timerlat hist -d 10m -c 0-4 -P d:100us:1ms -p 1ms --no-aa
   # RTLA timerlat histogram
   # Time unit is microseconds (us)
   # Duration:   0 00:10:00

Documentation/tools/rtla/rtla-timerlat-top.rst

@@ -32,6 +32,13 @@ OPTIONS
 
 .. include:: common_timerlat_aa.rst
 
+**--aa-only** *us*
+
+        Set stop tracing conditions and run without collecting and displaying statistics.
+        Print the auto-analysis if the system hits the stop tracing condition. This option
+        is useful to reduce rtla timerlat CPU, enabling the debug without the overhead of
+        collecting the statistics.
+
 EXAMPLE
 =======
 

tools/tracing/rtla/src/osnoise.c

@@ -841,6 +841,67 @@ static void osnoise_put_irq_disable(struct osnoise_context *context)
 	context->orig_opt_irq_disable = OSNOISE_OPTION_INIT_VAL;
 }
 
+static int osnoise_get_workload(struct osnoise_context *context)
+{
+	if (context->opt_workload != OSNOISE_OPTION_INIT_VAL)
+		return context->opt_workload;
+
+	if (context->orig_opt_workload != OSNOISE_OPTION_INIT_VAL)
+		return context->orig_opt_workload;
+
+	context->orig_opt_workload = osnoise_options_get_option("OSNOISE_WORKLOAD");
+
+	return context->orig_opt_workload;
+}
+
+int osnoise_set_workload(struct osnoise_context *context, bool onoff)
+{
+	int opt_workload = osnoise_get_workload(context);
+	int retval;
+
+	if (opt_workload == OSNOISE_OPTION_INIT_VAL)
+		return -1;
+
+	if (opt_workload == onoff)
+		return 0;
+
+	retval = osnoise_options_set_option("OSNOISE_WORKLOAD", onoff);
+	if (retval < 0)
+		return -1;
+
+	context->opt_workload = onoff;
+
+	return 0;
+}
+
+static void osnoise_restore_workload(struct osnoise_context *context)
+{
+	int retval;
+
+	if (context->orig_opt_workload == OSNOISE_OPTION_INIT_VAL)
+		return;
+
+	if (context->orig_opt_workload == context->opt_workload)
+		goto out_done;
+
+	retval = osnoise_options_set_option("OSNOISE_WORKLOAD", context->orig_opt_workload);
+	if (retval < 0)
+		err_msg("Could not restore original OSNOISE_WORKLOAD option\n");
+
+out_done:
+	context->orig_opt_workload = OSNOISE_OPTION_INIT_VAL;
+}
+
+static void osnoise_put_workload(struct osnoise_context *context)
+{
+	osnoise_restore_workload(context);
+
+	if (context->orig_opt_workload == OSNOISE_OPTION_INIT_VAL)
+		return;
+
+	context->orig_opt_workload = OSNOISE_OPTION_INIT_VAL;
+}
+
 /*
  * enable_osnoise - enable osnoise tracer in the trace_instance
  */
@@ -908,6 +969,9 @@ struct osnoise_context *osnoise_context_alloc(void)
 	context->orig_opt_irq_disable	= OSNOISE_OPTION_INIT_VAL;
 	context->opt_irq_disable	= OSNOISE_OPTION_INIT_VAL;
 
+	context->orig_opt_workload	= OSNOISE_OPTION_INIT_VAL;
+	context->opt_workload		= OSNOISE_OPTION_INIT_VAL;
+
 	osnoise_get_context(context);
 
 	return context;
@@ -935,6 +999,7 @@ void osnoise_put_context(struct osnoise_context *context)
 	osnoise_put_print_stack(context);
 	osnoise_put_tracing_thresh(context);
 	osnoise_put_irq_disable(context);
+	osnoise_put_workload(context);
 
 	free(context);
 }

tools/tracing/rtla/src/osnoise.h

@@ -42,6 +42,10 @@ struct osnoise_context {
 	/* -1 as init value because 0 is off */
 	int			orig_opt_irq_disable;
 	int			opt_irq_disable;
+
+	/* -1 as init value because 0 is off */
+	int			orig_opt_workload;
+	int			opt_workload;
 };
 
 /*
@@ -84,6 +88,7 @@ int osnoise_set_print_stack(struct osnoise_context *context,
 			    long long print_stack);
 
 int osnoise_set_irq_disable(struct osnoise_context *context, bool onoff);
+int osnoise_set_workload(struct osnoise_context *context, bool onoff);
 
 /*
  * osnoise_tool -  osnoise based tool definition.