On Thu, Apr 17, 2014 at 10:06:55AM +0100, Mel Gorman wrote:
> On Thu, Apr 17, 2014 at 10:59:12AM +0800, Fengguang Wu wrote:
> > On Tue, Apr 15, 2014 at 03:41:13PM +0100, Mel Gorman wrote:
> > > Fengguang Wu found that an earlier version crashed on his
> > > tests. This version passed tests running with DEBUG_VM and
> > > DEBUG_PAGEALLOC. Fengguang, another test would be appreciated and
> > > if it helps this series is the mm-numa-use-high-bit-v4r3 branch in
> > > git://git.kernel.org/pub/scm/linux/kernel/git/mel/linux-balancenuma.git
> > 
> > Hi Mel,
> > 
> > We noticed the below changes. The last_state.is_incomplete_run 0=>1 change
> > means the test box failed to boot up. Unfortunately we don't have
> > serial console output of this testbox, it may be hard to check the
> > root cause. Anyway, I'll try to bisect it to make the debug easier.
> > 
> 
> The bisection will be pretty small and probably point to the last patch.

Sorry I find it is lkp-a04 that is not reliable: by increasing test
count of v3.14, it boot hangs, too..

> I assume that lkp-04 is the machine name. What sort of machine is it?

It's an Atom machine.

> Does it have an unusual Kconfig that I might be missing a case for?

We use the attached default Kconfig for testing.

> What userspace is it running? Maybe there is a chance I can duplicate
> it. I assume fake/boot/1 is a test case that just boots the machine but
> does it do anything else?

Right, it's a simple boot test. The user space is a small 64bit debian
unstable system.

Thanks,
Fengguang