Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
In-Reply-To: <20180302164317.10554-8-ilina@codeaurora.org>
References: <20180302164317.10554-1-ilina@codeaurora.org> <20180302164317.10554-8-ilina@codeaurora.org>
From:   Evan Green <evgreen@chromium.org>
Date:   Mon, 5 Mar 2018 12:44:28 -0800
Message-ID: <CAE=gft59Jkb+X0fOcdEZN_8moUPvptZ-xuQaX1mg-VxBzHrSXw@mail.gmail.com>
Subject: Re: [PATCH v3 07/10] drivers: qcom: rpmh: cache sleep/wake state requests
To:     Lina Iyer <ilina@codeaurora.org>
Cc:     Andy Gross <andy.gross@linaro.org>,
        David Brown <david.brown@linaro.org>,
        linux-arm-msm@vger.kernel.org, linux-soc@vger.kernel.org,
        Rajendra Nayak <rnayak@codeaurora.org>,
        Bjorn Andersson <bjorn.andersson@linaro.org>,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Lina,

On Fri, Mar 2, 2018 at 8:43 AM, Lina Iyer <ilina@codeaurora.org> wrote:
> Active state requests are sent immediately to the mailbox controller,
> while sleep and wake state requests are cached in this driver to avoid
> taxing the mailbox controller repeatedly. The cached values will be sent
> to the controller when the rpmh_flush() is called.
>
> Generally, flushing is a system PM activity and may be called from the
> system PM drivers when the system is entering suspend or deeper sleep
> modes during cpuidle.
>
> Also allow invalidating the cached requests, so they may be re-populated
> again.
>
> Signed-off-by: Lina Iyer <ilina@codeaurora.org>
> ---
>
> Changes in v3:
>         - Remove locking for flush function
>         - Improve comments
> ---
>  drivers/soc/qcom/rpmh.c | 208 +++++++++++++++++++++++++++++++++++++++++++++++-
>  include/soc/qcom/rpmh.h |  10 +++
>  2 files changed, 217 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/soc/qcom/rpmh.c b/drivers/soc/qcom/rpmh.c
> index d95ea3fa8b67..8a04009075b8 100644
> --- a/drivers/soc/qcom/rpmh.c
> +++ b/drivers/soc/qcom/rpmh.c
[...]
> +
> +/**
> + * rpmh_invalidate: Invalidate all sleep and active sets
> + * sets.
> + *
> + * @rc: The RPMh handle got from rpmh_get_dev_channel
> + *
> + * Invalidate the sleep and active values in the TCS blocks.
> + */
> +int rpmh_invalidate(struct rpmh_client *rc)
> +{
> +       struct rpmh_ctrlr *rpm = rc->ctrlr;
> +       int ret;
> +       unsigned long flags;
> +
> +       if (IS_ERR_OR_NULL(rc))
> +               return -EINVAL;
> +
> +       spin_lock_irqsave(&rpm->lock, flags);
> +       rpm->dirty = true;
> +       spin_unlock_irqrestore(&rpm->lock, flags);

Thanks for removing the locking from the flush path. I was hoping to
see the locking removed around this statement as well. The way I
understand it, all of the racy bits are attempting to set dirty to
true, so you don't need a lock to protect multiple threads from
setting the same value. The only time dirty is read or cleared is in
the single-threaded PM path, so there are no potentially dangerous
interactions.

If no one has any other comments on the series, then I don't need to
hold everything up based on this one tweak alone. But if you end up
spinning it again for other reasons, consider making this change as
well.

Thanks,
-Evan