Re: lockdep splat in v5.11-rc1 involving console_sem and rq locks

[Peter Zijlstra]

On Wed, Jan 06, 2021 at 06:46:21AM -0800, Paul E. McKenney wrote:
> Huh.  The WARN does not always generate the lockdep complaint.  But
> fair enough.

Any printk()/WARN/BUG with rq lock held ought to generate that splat,
sometimes we die before we splat. The printk rewrite should eventually
fix that.

> >   https://lkml.kernel.org/r/20201226025117.2770-1-jiangshanlai@xxxxxxxxx
> 
> Thomas pointed me at this one a couple of weeks ago.  Here is an
> additional fix for rcutorture: f67e04bb0695 ("torture: Break affinity
> of kthreads last running on outgoing CPU").  I am still getting WARNs
> and lockdep splats with both applied.

That patch looks racy, what avoids the task being shuffled right back
before the CPU goes offline?

> What would break if I made the code dump out a few entries in the
> runqueue if the warning triggered?

There was a patch around that did that, Valentin might remember where
that was.