Re: [PATCH 0/6] workqueue: convert to raw_spinlock_t

From: Sebastian Andrzej Siewior
Date: Wed Jun 26 2019 - 10:09:25 EST


On 2019-06-26 06:49:57 [-0700], Tejun Heo wrote:
> On Wed, Jun 26, 2019 at 09:17:19AM +0200, Sebastian Andrzej Siewior wrote:
> > On 2019-06-13 16:50:21 [+0200], To linux-kernel@xxxxxxxxxxxxxxx wrote:
> > > Hi,
> > >
> > > the workqueue code has been reworked in -RT to use raw_spinlock_t based
> > > locking. This change allows to schedule worker from preempt_disable()ed
> > > or IRQ disabled section on -RT. This is the last patch. The previous
> > > patches are prerequisites or tiny cleanup (like patch #1 and #2).
> >
> > a gentle *ping*
>
> I don't now what to make of the series. AFAICS, there's no benefit to
> mainline. What am I missing?

Is there something specific you don't like? #1 and #2 are cleanups so we
don't argue about those right?
#5 makes use of swake_up which is slightly smaller compared to wake_up()
so that should be fine.
That last one makes no change to !RT because the difference between
raw_spin and spinlock is not existing. However I'm working a lockdep
patch which complains about wrong context so without it would complain
if anyone would try to schedule a workqueue from IRQ or preemption
disabled region.

> Thanks.

Sebastian