When the function graph tracer is activated, it iterates over the task_list
to allocate a stack to store the return addresses.
But the per cpu idle tasks are not iterated by using
do_each_thread / while_each_thread.
So we have to iterate on them manually.
This fixes somes weirdness in the traces and many losses of traces.
Examples on two cpus:
0) Xorg-4287 | 2.906 us | }
0) Xorg-4287 | 3.965 us | }
0) Xorg-4287 | 5.302 us | }
------------------------------------------
0) Xorg-4287 => <idle>-0
------------------------------------------
0) <idle>-0 | 2.861 us | }
0) <idle>-0 | 0.526 us | set_normalized_timespec();
0) <idle>-0 | 7.201 us | }
0) <idle>-0 | 8.214 us | }
0) <idle>-0 | | clockevents_program_event() {
0) <idle>-0 | | lapic_next_event() {
0) <idle>-0 | 0.510 us | native_apic_mem_write();
0) <idle>-0 | 1.546 us | }
0) <idle>-0 | 2.583 us | }
0) <idle>-0 | + 12.435 us | }
0) <idle>-0 | + 13.470 us | }
0) <idle>-0 | 0.608 us | _spin_unlock_irqrestore();
0) <idle>-0 | + 23.270 us | }
0) <idle>-0 | + 24.336 us | }
0) <idle>-0 | + 25.417 us | }
0) <idle>-0 | 0.593 us | _spin_unlock();
0) <idle>-0 | + 41.869 us | }
0) <idle>-0 | + 42.906 us | }
0) <idle>-0 | + 95.035 us | }
0) <idle>-0 | 0.540 us | menu_reflect();
0) <idle>-0 | ! 100.404 us | }
0) <idle>-0 | 0.564 us | mce_idle_callback();
0) <idle>-0 | | enter_idle() {
0) <idle>-0 | 0.526 us | mce_idle_callback();
0) <idle>-0 | 1.757 us | }
0) <idle>-0 | | cpuidle_idle_call() {
0) <idle>-0 | | menu_select() {
0) <idle>-0 | 0.525 us | pm_qos_requirement();
0) <idle>-0 | 0.518 us | tick_nohz_get_sleep_length();
0) <idle>-0 | 2.621 us | }
[...]
1) <idle>-0 | 0.518 us | touch_softlockup_watchdog();
1) <idle>-0 | + 14.355 us | }
1) <idle>-0 | + 22.840 us | }
1) <idle>-0 | + 25.949 us | }
1) <idle>-0 | | handle_irq() {
1) <idle>-0 | 0.511 us | irq_to_desc();
1) <idle>-0 | | handle_edge_irq() {
1) <idle>-0 | 0.638 us | _spin_lock();
1) <idle>-0 | | ack_apic_edge() {
1) <idle>-0 | 0.510 us | irq_to_desc();
1) <idle>-0 | | move_native_irq() {
1) <idle>-0 | 0.510 us | irq_to_desc();
1) <idle>-0 | 1.532 us | }
1) <idle>-0 | 0.511 us | native_apic_mem_write();
------------------------------------------
1) <idle>-0 => cat-5073
------------------------------------------
1) cat-5073 | 3.731 us | }
1) cat-5073 | | run_local_timers() {
1) cat-5073 | 0.533 us | hrtimer_run_queues();
1) cat-5073 | | raise_softirq() {
1) cat-5073 | | __raise_softirq_irqoff() {
1) cat-5073 | | /* nr: 1 */
1) cat-5073 | 2.718 us | }
1) cat-5073 | 3.814 us | }
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
static int start_graph_tracing(void)
{
struct ftrace_ret_stack **ret_stack_list;
- int ret;
+ int ret, cpu;
ret_stack_list = kmalloc(FTRACE_RETSTACK_ALLOC_SIZE *
sizeof(struct ftrace_ret_stack *),
if (!ret_stack_list)
return -ENOMEM;
+ /* The cpu_boot init_task->ret_stack will never be freed */
+ for_each_online_cpu(cpu)
+ ftrace_graph_init_task(idle_task(cpu));
+
do {
ret = alloc_retstack_tasklist(ret_stack_list);
} while (ret == -EAGAIN);