Re: [PATCH v2 12/14] torture: Replace cpu_up/down with device_online/offline

From: Paul E. McKenney
Date: Fri Nov 29 2019 - 15:39:03 EST


On Fri, Nov 29, 2019 at 09:13:45AM +0000, Qais Yousef wrote:
> On 11/28/19 13:02, Paul E. McKenney wrote:
> > On Thu, Nov 28, 2019 at 05:00:26PM +0000, Qais Yousef wrote:
> > > On 11/28/19 16:56, Qais Yousef wrote:
> > > > On 11/27/19 13:47, Paul E. McKenney wrote:
> > > > > On Mon, Nov 25, 2019 at 11:27:52AM +0000, Qais Yousef wrote:
> > > > > > The core device API performs extra housekeeping bits that are missing
> > > > > > from directly calling cpu_up/down.
> > > > > >
> > > > > > See commit a6717c01ddc2 ("powerpc/rtas: use device model APIs and
> > > > > > serialization during LPM") for an example description of what might go
> > > > > > wrong.
> > > > > >
> > > > > > This also prepares to make cpu_up/down a private interface for anything
> > > > > > but the cpu subsystem.
> > > > > >
> > > > > > Signed-off-by: Qais Yousef <qais.yousef@xxxxxxx>
> > > > > > CC: Davidlohr Bueso <dave@xxxxxxxxxxxx>
> > > > > > CC: "Paul E. McKenney" <paulmck@xxxxxxxxxx>
> > > > > > CC: Josh Triplett <josh@xxxxxxxxxxxxxxxx>
> > > > > > CC: linux-kernel@xxxxxxxxxxxxxxx
> > > > >
> > > > > Looks fine from an rcutorture viewpoint, but why not provide an API
> > > > > that pulled lock_device_hotplug() and unlock_device_hotplug() into the
> > > > > online/offline calls?
> > > >
> > > > I *think* the right way to do what you say is by doing lock_device_hotplug()
> > > > inside device_{online, offline}() - which affects all drivers not just the CPU.
> >
> > Or there could be a CPU-specific wrapper function that did the needed
> > locking. (Whether this is worth it or not of course depends on the
> > number of invocations.)
>
> Okay I see what you mean now. driver/base/memory.c have {add,remove}_memory()
> that does what you say. I think we can replicate this in driver/base/cpu.c too.
>
> I can certainly do that, better as an improvement on top as I need to audit the
> code to make sure the critical sections weren't relying on this lock to protect
> something else beside the online/offline operation.

Works for me!

Thanx, Paul