[PATCH stable] tick/nohz: Prevent bogus softirq pending warning

[PATCH stable] tick/nohz: Prevent bogus softirq pending warning

[Thomas Gleixner]

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());

Re: [PATCH stable] tick/nohz: Prevent bogus softirq pending warning

[Frederic Weisbecker]

On Thu, Aug 30, 2018 at 05:05:19PM +0200, Thomas Gleixner wrote:
> 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>

Acked-by: Frederic Weisbecker <frederic@kernel.org>

Thanks for cooking the patch!

Re: [PATCH stable] tick/nohz: Prevent bogus softirq pending warning

[Geert Uytterhoeven]

Hi Thomas,

On Thu, Aug 30, 2018 at 5:06 PM Thomas Gleixner <tglx@linutronix.de> wrote:
> 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

Fixes: 2d898915ccf4838c ("nohz: Fix missing tick reprogram when
interrupting an inline softirq")

Issue in v4.14.x bisected to the above commit, and fixed by your patch.

Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH stable] tick/nohz: Prevent bogus softirq pending warning

[Geert Uytterhoeven]

Hi Greg,

On Wed, Sep 12, 2018 at 3:56 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> On Thu, Aug 30, 2018 at 5:06 PM Thomas Gleixner <tglx@linutronix.de> wrote:
> > 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
>
> Fixes: 2d898915ccf4838c ("nohz: Fix missing tick reprogram when
> interrupting an inline softirq")
>
> Issue in v4.14.x bisected to the above commit, and fixed by your patch.
>
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>

This issue is still present in v4.14.72. Can you please apply Thomas' fix?
Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH stable] tick/nohz: Prevent bogus softirq pending warning

[Greg KH]

On Thu, Sep 27, 2018 at 10:07:39AM +0200, Geert Uytterhoeven wrote:
> Hi Greg,
> 
> On Wed, Sep 12, 2018 at 3:56 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > On Thu, Aug 30, 2018 at 5:06 PM Thomas Gleixner <tglx@linutronix.de> wrote:
> > > 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
> >
> > Fixes: 2d898915ccf4838c ("nohz: Fix missing tick reprogram when
> > interrupting an inline softirq")
> >
> > Issue in v4.14.x bisected to the above commit, and fixed by your patch.
> >
> > Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> 
> This issue is still present in v4.14.72. Can you please apply Thomas' fix?

Oops, sorry for the delay, now queued up.

greg k-h