Re: Linux 2.6.21-rc1

From: Thomas Gleixner
Date: Wed Feb 21 2007 - 12:02:59 EST

On Wed, 2007-02-21 at 08:24 -0800, Daniel Walker wrote:
> > The most interesting core change may be the dyntick/nohz one, where timer
> > ticks will only happen when needed. It's been brewing for a _loong_ time,
> > but it's in the standard kernel now as an option.
> On i386 I get the following,
> TCP cubic registered
> NET: Registered protocol family 1
> NET: Registered protocol family 17
> Testing NMI watchdog ... CPU#0: NMI appears to be stuck (24->24)!
> CPU#1: NMI appears to be stuck (0->0)!
> CPU#2: NMI appears to be stuck (0->0)!
> CPU#3: NMI appears to be stuck (0->0)!
> when I add nmi_watchdog=1 to my boot args which worked on prior kernels.
> On closer inspection it looks like arch/i386/kernel/io_apic.c :
> check_timer() --> timer_irq_works() depends on IRQ0 incrementing jiffies
> which is no longer the case AFAIK.

At this point the PIT / HPET _is_ active and incrementing jiffies. The
switch to local apic timers happens afterwards.

> I'm not sure exactly how that relates to the NMI, but the check_timer()
> function disabled the NMI through the io-apic if it can't get the
> "timer" working through the io-apic.

Boot log please.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at