Commit 0a0e0829f990 ("nohz: Fix missing tick reprogram when interrupting an inline softirq") got backported to stable trees and now causes the NOHZ softirq pending warning to trigger. It's not an upstream issue as the NOHZ update logic has been changed there.
The problem is when a softirq disabled section gets interrupted and on return from interrupt the tick/nohz state is evaluated, which then can observe pending soft interrupts. These soft interrupts are legitimately pending because they cannot be processed as long as soft interrupts are disabled and the interrupted code will correctly process them when soft interrupts are reenabled.
Add a check for softirqs disabled to the pending check to prevent the warning.
Reported-by: Grygorii Strashko grygorii.strashko@ti.com Reported-by: John Crispin john@phrozen.org Signed-off-by: Thomas Gleixner tglx@linutronix.de Tested-by: Grygorii Strashko grygorii.strashko@ti.com Tested-by: John Crispin john@phrozen.org Cc: Frederic Weisbecker frederic@kernel.org Cc: Ingo Molnar mingo@kernel.org Cc: Anna-Maria Gleixner anna-maria@linutronix.de Cc: Greg Kroah-Hartman gregkh@linuxfoundation.org Cc: stable@vger.kernel.org --- kernel/time/tick-sched.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
--- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -888,7 +888,7 @@ static bool can_stop_idle_tick(int cpu, if (unlikely(local_softirq_pending() && cpu_online(cpu))) { static int ratelimit;
- if (ratelimit < 10 && + if (ratelimit < 10 && !in_softirq() && (local_softirq_pending() & SOFTIRQ_STOP_IDLE_MASK)) { pr_warn("NOHZ: local_softirq_pending %02x\n", (unsigned int) local_softirq_pending());