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: <20201119110036.19959-1-aisheng.dong@nxp.com>
In-Reply-To: <20201119110036.19959-1-aisheng.dong@nxp.com>
From:   "Rafael J. Wysocki" <rafael@kernel.org>
Date:   Thu, 19 Nov 2020 13:11:38 +0100
Message-ID: <CAJZ5v0gRmmkxG7aLRHXw1Gx7GGd1PyZ6LNqChuQCVKPV9X_8zg@mail.gmail.com>
Subject: Re: [PATCH 1/3] driver core: simply go out if the same device_link is
 added again
To:     Dong Aisheng <aisheng.dong@nxp.com>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        dl-linux-imx <linux-imx@nxp.com>,
        Dong Aisheng <dongas86@gmail.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Saravana Kannan <saravanak@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Nov 19, 2020 at 12:18 PM Dong Aisheng <aisheng.dong@nxp.com> wrote:
>
> It's possible that the same device link may be added by parsing the
> function dependecy in DT. e.g. clock/gpio/regulators.
> Simply go out for this case.

Why?

> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: "Rafael J. Wysocki" <rafael@kernel.org>
> Cc: Saravana Kannan <saravanak@google.com>
> Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
> ---
>  drivers/base/core.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/drivers/base/core.c b/drivers/base/core.c
> index 4c03bdd3a268..7d91d4074136 100644
> --- a/drivers/base/core.c
> +++ b/drivers/base/core.c
> @@ -567,6 +567,9 @@ struct device_link *device_link_add(struct device *consumer,
>                 if (link->consumer != consumer)
>                         continue;
>
> +               if (flags == link->flags)
> +                       goto out;

But this prevents rpm_count from being incremented if
DL_FLAG_RPM_ACTIVE is set in flags, which is necessary, because the
supplier's PM-runtime usage counter has been incremented already.

Moreover, every attempt to create a stateless link must cause a new
reference on the existing link to be acquired, or the removal will not
work correctly.

> +
>                 if (flags & DL_FLAG_PM_RUNTIME) {
>                         if (!(link->flags & DL_FLAG_PM_RUNTIME)) {
>                                 pm_runtime_new_link(consumer);
> --