Re: possible deadlock in flush_workqueue (2)

[Tejun Heo]

Hello,

On Tue, Oct 23, 2018 at 10:28:04AM -0400, Theodore Y. Ts'o wrote:
> The other, and I think better way, we could fix this is to have an
> alternate destroy_never_used_workqueue() function which skips the
> drain_workqueue().

Yeah, I'm okay with this approach.  Maybe name it
destroy_workqueue_skip_drain()?

Thanks.

-- 
tejun