[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <95acab94-ae80-b43c-673c-9dfc82f3b40b@linaro.org>
Date: Mon, 16 Aug 2021 20:08:39 +0300
From: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>
To: Stephen Boyd <sboyd@...nel.org>
Cc: Andy Gross <agross@...nel.org>,
Bjorn Andersson <bjorn.andersson@...aro.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Michael Turquette <mturquette@...libre.com>,
Pavel Machek <pavel@....cz>,
"Rafael J. Wysocki" <rjw@...ysocki.net>,
Taniya Das <tdas@...eaurora.org>,
Linux PM <linux-pm@...r.kernel.org>,
open list <linux-kernel@...r.kernel.org>,
"open list:COMMON CLK FRAMEWORK" <linux-clk@...r.kernel.org>,
"open list:DRM DRIVER FOR MSM ADRENO GPU"
<linux-arm-msm@...r.kernel.org>
Subject: Re: [PATCH v3 0/3] PM: add two devres helpers and use them in qcom cc
On 06/08/2021 10:08, Dmitry Baryshkov wrote:
> On Fri, 6 Aug 2021 at 02:53, Stephen Boyd <sboyd@...nel.org> wrote:
>>
>> Quoting Dmitry Baryshkov (2021-07-31 12:50:31)
>>> Most of the drivers using using pm_runtime_enable() or pm_clk_create()
>>> follow the same pattern: call the function in the probe() path and call
>>> correspondingly pm_runtime_disable() or pm_clk_destroy() from the
>>> probe()'s error path and from the remove() function. This common code
>>> pattern has several drawbacks. I.e. driver authors have to ensure that
>>> the disable/destroy call in the error path really corresponds to the
>>> proper error clause. Or that the disable/destroy call is not missed in
>>> the remove() callback.
>>>
>>> Add two devres helpers replacing these code patterns with relevant devm
>>> function call, removing the need to call corresponding disable/destroy
>>> functions. As an example modify Qualcomm clock controller code to use
>>> new helpers. In this case we are able to drop error path and remove
>>> functions completely, simplifying the drivers in question.
>>
>> There are lots of folks on the To: line so I'm not sure who is supposed
>> to apply this. Please indicate which maintainer tree you're planning to
>> land a series through if it touches different areas of the tree.
>
> I'd prefer for them to go through the clock tree.
Stephen, since Rafael has acked the patched, could you please take a
look and hopefully merge them?
--
With best wishes
Dmitry
Powered by blists - more mailing lists