Re: [PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

Re: [PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Markus Elfring]

> In-Reply-To: <cover.1592888591.git.shengjiu.wang@nxp.com>

I guess that it should be sufficient to specify such a field once
for the header information.


> Because clk_prepare_enable and clk_disable_unprepare should
> check input clock parameter is NULL or not internally,

I find this change description unclear.


> then we don't need to check them before calling the function.

Please use an imperative wording for the commit message.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?id=dd0d718152e4c65b173070d48ea9dfc06894c3e5#n151

Regards,
Markus

Re: [PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Shengjiu Wang]

On Tue, Jun 23, 2020 at 3:38 PM Markus Elfring <Markus.Elfring@web.de> wrote:
>
> > In-Reply-To: <cover.1592888591.git.shengjiu.wang@nxp.com>
>
> I guess that it should be sufficient to specify such a field once
> for the header information.

seems it's caused by my "git format-patch" command, I will update
it, hope it is better next time.

>
>
> > Because clk_prepare_enable and clk_disable_unprepare should
> > check input clock parameter is NULL or not internally,
>
> I find this change description unclear.

    clk_prepare_enable and clk_disable_unprepare check the input
    clock parameter in the beginning of the function, if the parameter
    is NULL, clk_prepare_enable and clk_disable_unprepare will
    return immediately.

    So Don't need to check input clock parameters before calling clk
    API.

Do you think this commit message is better?

best regards
wang shengjiu

Re: [PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Markus Elfring]

>     clk_prepare_enable and clk_disable_unprepare check the input
>     clock parameter in the beginning of the function,

These functions call further functions which perform null pointer checks.


>                                                       if the parameter
>     is NULL, clk_prepare_enable and clk_disable_unprepare will
>     return immediately.

The interpretation of these function implementations seems to be reasonable.
Would you like to achieve any improvements for the corresponding software documentation?


>     So Don't need to check input clock parameters before calling clk API.

What do you find imperative in this wording?

Another wording alternative:
   Thus omit extra null pointer checks before four function calls.

Regards,
Markus

Re: [PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Shengjiu Wang]

On Tue, Jun 23, 2020 at 4:55 PM Markus Elfring <Markus.Elfring@web.de> wrote:
>
> >     clk_prepare_enable and clk_disable_unprepare check the input
> >     clock parameter in the beginning of the function,
>
> These functions call further functions which perform null pointer checks.
>
>
> >                                                       if the parameter
> >     is NULL, clk_prepare_enable and clk_disable_unprepare will
> >     return immediately.
>
> The interpretation of these function implementations seems to be reasonable.
> Would you like to achieve any improvements for the corresponding software documentation?

Which document do you mean?

>
>
> >     So Don't need to check input clock parameters before calling clk API.
>
> What do you find imperative in this wording?
>
> Another wording alternative:
>    Thus omit extra null pointer checks before four function calls.
>
> Regards,
> Markus

Re: [v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Markus Elfring]

>>>                                                       if the parameter
>>>     is NULL, clk_prepare_enable and clk_disable_unprepare will
>>>     return immediately.
>>
>> The interpretation of these function implementations seems to be reasonable.
>> Would you like to achieve any improvements for the corresponding software documentation?
>
> Which document do you mean?

Example:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/include/linux/clk.h?id=dd0d718152e4c65b173070d48ea9dfc06894c3e5#n905
https://elixir.bootlin.com/linux/v5.7.2/source/include/linux/clk.h#L905

Regards,
Markus

[PATCH v2 0/2] Fix unchecked return value for clk_prepare_enable

[Shengjiu Wang]

First patch is to remove the check of clock pointer before calling
clk API.

Second patch is to fix the issue that the return value of
clk_prepare_enable is not checked.

changes in v2:
- split the patch to separate patches

Shengjiu Wang (2):
  ASoC: fsl_mqs: Don't check clock is NULL before calling clk API
  ASoC: fsl_mqs: Fix unchecked return value for clk_prepare_enable

 sound/soc/fsl/fsl_mqs.c | 23 ++++++++++++++---------
 1 file changed, 14 insertions(+), 9 deletions(-)

-- 
2.21.0

[PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Shengjiu Wang]

Because clk_prepare_enable and clk_disable_unprepare should
check input clock parameter is NULL or not internally, then
we don't need to check them before calling the function.

Fixes: 9e28f6532c61 ("ASoC: fsl_mqs: Add MQS component driver")
Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
---
 sound/soc/fsl/fsl_mqs.c | 13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)

diff --git a/sound/soc/fsl/fsl_mqs.c b/sound/soc/fsl/fsl_mqs.c
index 0c813a45bba7..b44b134390a3 100644
--- a/sound/soc/fsl/fsl_mqs.c
+++ b/sound/soc/fsl/fsl_mqs.c
@@ -266,11 +266,9 @@ static int fsl_mqs_runtime_resume(struct device *dev)
 {
 	struct fsl_mqs *mqs_priv = dev_get_drvdata(dev);
 
-	if (mqs_priv->ipg)
-		clk_prepare_enable(mqs_priv->ipg);
+	clk_prepare_enable(mqs_priv->ipg);
 
-	if (mqs_priv->mclk)
-		clk_prepare_enable(mqs_priv->mclk);
+	clk_prepare_enable(mqs_priv->mclk);
 
 	if (mqs_priv->use_gpr)
 		regmap_write(mqs_priv->regmap, IOMUXC_GPR2,
@@ -292,11 +290,8 @@ static int fsl_mqs_runtime_suspend(struct device *dev)
 		regmap_read(mqs_priv->regmap, REG_MQS_CTRL,
 			    &mqs_priv->reg_mqs_ctrl);
 
-	if (mqs_priv->mclk)
-		clk_disable_unprepare(mqs_priv->mclk);
-
-	if (mqs_priv->ipg)
-		clk_disable_unprepare(mqs_priv->ipg);
+	clk_disable_unprepare(mqs_priv->mclk);
+	clk_disable_unprepare(mqs_priv->ipg);
 
 	return 0;
 }
-- 
2.21.0

Re: [PATCH v2 1/2] ASoC: fsl_mqs: Don't check clock is NULL before calling clk API

[Nicolin Chen]

On Tue, Jun 23, 2020 at 02:01:11PM +0800, Shengjiu Wang wrote:
> Because clk_prepare_enable and clk_disable_unprepare should
> check input clock parameter is NULL or not internally, then
> we don't need to check them before calling the function.
> 
> Fixes: 9e28f6532c61 ("ASoC: fsl_mqs: Add MQS component driver")
> Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>

Acked-by: Nicolin Chen <nicoleotsuka@gmail.com>