Re: [BUG 2.6.23-rc6-mm1] NMI Watchdog detected LOCKUP on CPU 0

[Fengguang Wu]

On Sat, Sep 22, 2007 at 10:35:45PM -0700, Andrew Morton wrote:
> On Sun, 23 Sep 2007 13:30:40 +0800 Fengguang Wu <wfg@xxxxxxxxxxxxxxxx> wrote:
> 
> > > That's interensting.  serial_in().  We have had NMI watchdog expiries when
> > > the kernel is printing a large amount of stuff out a slow serial port with
> > > interrutps disabled.  But I thought we'd pretty much plugged those problems
> > > by sprinkling touch_nmi_watchdog() in various places.
> > > 
> > > Do you think this is what was happening on your system?
> > 
> > Very likely. I'm running linux with cmdline
> > "root=/dev/sda1 ro nmi_watchdog=1 console=ttyS0,115200 console=tty0",
> > and doing a lot of printks at the time ;-)
> 
> OK.  We need to find a suitable place to poke yet another
> touch_nmi_watchdog().  Maybe we should give up and put one in
> printk().
> 
> And you oopsed for different reasons in the nmi-watchdog handling
> code too.  I think I'll pretend I didn't see that.

Let's forget it for now.
I can try Ingo's latency tracing patches at some convenient time.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/