tracing: Have function_graph selftest call cond_resched()

When all kernel debugging is enabled (lockdep, KSAN, etc), the function
graph enabling and disabling can take several seconds to complete. The
function_graph selftest enables and disables function graph tracing
several times. With full debugging enabled, the soft lockup watchdog was
triggering because the selftest was running without ever scheduling.

Add cond_resched() throughout the test to make sure it does not trigger
the soft lockup detector.

Link: https://lkml.kernel.org/r/20230528051742.1325503-6-rostedt@goodmis.org
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>

diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c
index a931d9a..5295904 100644 (file)
--- a/kernel/trace/trace_selftest.c
+++ b/kernel/trace/trace_selftest.c
@@ -848,6 +848,12 @@ trace_selftest_startup_function_graph(struct tracer *trace,
        }
 
 #ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
+       /*
+        * These tests can take some time to run. Make sure on non PREEMPT
+        * kernels, we do not trigger the softlockup detector.
+        */
+       cond_resched();
+
        tracing_reset_online_cpus(&tr->array_buffer);
        set_graph_array(tr);
 
@@ -869,6 +875,8 @@ trace_selftest_startup_function_graph(struct tracer *trace,
        if (ret)
                goto out;
 
+       cond_resched();
+
        ret = register_ftrace_graph(&fgraph_ops);
        if (ret) {
                warn_failed_init_tracer(trace, ret);
@@ -891,6 +899,8 @@ trace_selftest_startup_function_graph(struct tracer *trace,
        if (ret)
                goto out;
 
+       cond_resched();
+
        tracing_start();
 
        if (!ret && !count) {