Trace IPI-related IRQs with ft_irq_fired()

IPIs have some special cases where irq_enter() is not called. This
caused ft_irq_fired() to "miss" some rescheduling-related interrupts,
which in turn may cause outliers.

This patch makes sure ft_irq_fired() is called on scheduling-related
IPIs.

1 files changed, 5 insertions, 1 deletions

diff --git a/kernel/sched.c b/kernel/sched.c
index cbb1b5f57117..1d7b38a7ff25 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -2597,8 +2597,12 @@ void scheduler_ipi(void)
         struct rq *rq = this_rq();
         struct task_struct *list = xchg(&rq->wake_list, NULL);
 
-        if (!list)
+        if (!list) {
+                /* If we don't call irq_enter(), we need to trigger the IRQ
+                 * tracing manually. */
+                ft_irq_fired();
                 return;
+        }
 
         /*
          * Not all reschedule IPI handlers call irq_enter/irq_exit, since