Re: [RFC] workqueue: Fix irq inversion deadlock in manage_workers()

[Tejun Heo]

On Mon, Oct 09, 2017 at 08:42:57AM +0200, Peter Zijlstra wrote:
> On Sun, Oct 08, 2017 at 12:03:47PM -0700, Tejun Heo wrote:
> > So, if I'm not mistaken, this is a regression caused by b9c16a0e1f73
> > ("locking/mutex: Fix lockdep_assert_held() fail") which seems to
> > replace irqsave operations inside mutex to unconditional irq ones.
> 
> No, it existed before that. You're looking at the DEBUG_MUTEX case, the
> normal case looked like:

Ah, I see.  That mutex usage was a stretch anyway.  I'll get rid of
it.

Thanks.

-- 
tejun