Re: Sleeping BUG in khugepaged for i586

[Larry Finger]

On 06/23/2017 08:25 AM, Michal Hocko wrote:
> On Fri 23-06-17 15:13:45, Vlastimil Babka wrote:
> > On 06/23/2017 02:08 PM, Michal Hocko wrote:
> > > On Thu 08-06-17 16:48:31, Michal Hocko wrote:
> > > > On Wed 07-06-17 13:56:01, David Rientjes wrote:
> > > >
> > > > I suspect, so cond_resched seems indeed inappropriate on 32b systems.
> > >
> > > The code still seems to be in the mmotm tree.
> >
> > Even mainline at this point - 338a16ba1549
> >
> > > Are there any plans to fix
> > > this or drop the patch?
> >
> > https://lkml.kernel.org/r/alpine.DEB.2.10.1706191341550.97821@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> Ahh, I have missed that. Thanks!

I also missed that patch. Applying it to my box fixes the scheduling while 
atomic splats and no downside has been detected.

You may add "Reported-and-tested-by: Larry Finger <Larry.Finger@xxxxxxxxxxxx>".

Thanks for everyone's efforts in fixing this problem.

Larry