Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <6543936.FbWAdBN1tG@kreacher> <1708806.S9fAqql2gf@kreacher> <e9466a28-5c6a-f88a-51ab-547ecc21e5d0@hisilicon.com>
In-Reply-To: <e9466a28-5c6a-f88a-51ab-547ecc21e5d0@hisilicon.com>
From:   "Rafael J. Wysocki" <rafael@kernel.org>
Date:   Thu, 22 Oct 2020 16:16:25 +0200
Message-ID: <CAJZ5v0hdWtUyDFBJpXH5tkFUtRF4y7P6SvcepEA24H1K4dy+iQ@mail.gmail.com>
Subject: Re: [PATCH 3/3] PM: runtime: Resume the device earlier in __device_release_driver()
To:     "chenxiang (M)" <chenxiang66@hisilicon.com>
Cc:     "Rafael J. Wysocki" <rjw@rjwysocki.net>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Linux PM <linux-pm@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Lukas Wunner <lukas@wunner.de>,
        Saravana Kannan <saravanak@google.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

On Thu, Oct 22, 2020 at 3:40 PM chenxiang (M) <chenxiang66@hisilicon.com> wrote:
>
> Hi Rafael,
>
> 在 2020/10/22 3:14, Rafael J. Wysocki 写道:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >
> > Since the device is resumed from runtime-suspend in
> > __device_release_driver() anyway, it is better to do that before
> > looking for busy managed device links from it to consumers, because
> > if there are any, device_links_unbind_consumers() will be called
> > and it will cause the consumer devices' drivers to unbind, so the
> > consumer devices will be runtime-resumed.  In turn, resuming each
> > consumer device will cause the supplier to be resumed and when the
> > runtime PM references from the given consumer to it are dropped, it
> > may be suspended.  Then, the runtime-resume of the next consumer
> > will cause the supplier to resume again and so on.
> >
> > Update the code accordingly.
> >
> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > Fixes: 9ed9895370ae ("driver core: Functional dependencies tracking support")
> > Cc: All applicable <stable@vger.kernel.org> # All applicable
> > ---
> >   drivers/base/dd.c |    4 ++--
> >   1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > Index: linux-pm/drivers/base/dd.c
> > ===================================================================
> > --- linux-pm.orig/drivers/base/dd.c
> > +++ linux-pm/drivers/base/dd.c
> > @@ -1117,6 +1117,8 @@ static void __device_release_driver(stru
> >
> >       drv = dev->driver;
> >       if (drv) {
> > +             pm_runtime_get_sync(dev);
> > +
> >               while (device_links_busy(dev)) {
> >                       __device_driver_unlock(dev, parent);
> >
> > @@ -1132,8 +1134,6 @@ static void __device_release_driver(stru
>
> pm_runtime_put_sync() is required to be called if existed from here.
>
> >                               return;

Right, I overlooked this, thanks!

> >               }
> >
> > -             pm_runtime_get_sync(dev);
> > -
> >               driver_sysfs_remove(dev);
> >
> >               if (dev->bus)