Re: Subject: Warning in workqueue.c

[Tejun Heo <tj@kernel.org>]

On Mon, Feb 24, 2014 at 10:01:55AM -0500, Jason J. Herne wrote:
> On 02/14/2014 11:25 AM, Tejun Heo wrote:
> >...
> >Hmmm... weird, p's rq shouldn't have changed without its cpus_allowed
> >busted.  Anyways, let's wait for Jason's test results and see whether
> >this is a regression at all.
> 
> I was unable to determine exactly when this behavior was introduced.
> The reason for this is because I keep hitting other bugs that
> prevent the test case from running, the most notable of which causes
> an immediate system hang. I also hit other warnings and bug messages
> and I'm not sure if they are related or if they will influence the
> probability of hitting the problem we are trying to solve here.
> What I did find is the following:
> 
> We hit this problem as far back as v3.10.
> The warning was introduced after v3.5 but before v3.6.

That's a bummer but it at least isn't a very new regression.  Peter,
any ideas on debugging this?  I can make workqueue to play block /
unblock dance to try to work around the issue but that'd be very
yucky.  It'd be great to root cause where the cpu selection anomaly is
coming from.

Thanks.

-- 
tejun